A Guide Book to Mechanism in Organic Chemistry_OCR

A guidebook to mechanismin organic chemistry

Peter Sykes M.Sc., Ph.D., F.R.S.C., C.Chem.Fellow and Vice-Master, Christ's College, Cambridge

~ .. ~ Logan»=== 50 "lie \~Technical

Copublished in the United States withjohn Wiley & Sons, Inc., New York

Contents*

Page

Foreword by Professor Lord Todd, O.M., P.P.R.S. ixPreface to sixth edition xi

1 Structure, reactivity, and mechanism I

2 Energetics, kinetics, and the investigation of mechanism 33

3 The strengths of acids and bases 53

4 Nucleophilic substitution at a saturated carbon atom 77

5 Carbocations, electron-deficient Nand 0 atoms and their 101reactions

6 Electrophilic and nucleophilic substitution in aromatic 130systems

7 Electrophilic and nucleophilic addition to C=C 178

8 Nucleophilic addition to C=O 203

9 Elimination reactions 246

10 Carbanions and their reactions 270

II Radicals and their reactions 299

12 Symmetry controlled reactions 340

13 Linear free energy relationships 358

Select bibliography 396

Index 399

• A detailed list of contents will be found at the beginning of each chapter.

Foreword

Fifty years ago the student taking up organic chemistry-and I speakfrom experience-was almost certain to be referred to one or other ofa few textbooks generally known by the name of their authors-e.g.Holleman, Bernthsen, Schmidt, Karrer and Gattermann. On thesetexts successive generations of chemists were nurtured, and not in onecountry alone, for they were translated into several languages. These,the household names of fifty years ago, have for the most part gone. Inpast times of course the total number of books available was rathersmall and it is only in the last quarter of a century that we have seena veritable flood of organic chemical textbooks pouring into booksellers' lists. The increase in the number of texts may be in part due tothe rise in student numbers but the primary reason for it is the revolutionary impact of mechanistic studies on our approach to organicchemistry at the elementary level. With the plethora of books available, however, it is now much more difficult for an author to becomea household name wherever the subject is taught. Yet this has indeedhappened to Dr. Peter Sykes through his Guidebook to Mechanism inOrganic Chemistry.

In the Foreword which I was privileged to write for the FirstEdition in 1961 I described not only my own view of what was happening in organic chemistry but also the type of approach to teaching itwhich was favoured by Dr. Sykes. Having known and watched himover many years first as student, then as colleague, and always asfriend, I was confident that he had written an excellent book which,in my view at least, would add new interest to the study of organicchemistry. But its success has far exceeded even my high expectationsand in its later editions it has been revised and refined without everlosing the cutting edge of the original.

The present volume continues the tradition. Once again the recentliterature has been combed for new examples the better to exemplifyprinciples of reactions. Of particular interest is an admirable chapterdealing with reactions controlled by orbital symmetry. Until I read itI was not convinced that this very important new development in thetheory of organic reactions could be simply yet usefully communicatedto students at an elementary level. To have succeeded in doing so onlyunderlines further Dr. Sykes' gifts as a teacher and writer and I am surethat this new edition of the Guidebook will more than equal the successof its predecessors.

Cambridge TODD

Preface to sixth editionIt is now twenty-five years since this Guidebook first appeared and,hardly surprisingly, the current version is vastly different in bothcontent and physical appearance from that first offering of so longago. Over the years a real endeavour has been made to incorporatenew, and to delete old, material not to reflect current trends andfashions, but to encompass significant changes in our fundamentalunderstanding of organic chemistry; more particularly, to decidehow these changes can best be conveyed to a largely undergraduateaudience. At the same time care has been taken to retain theunderlying framework and structure of the book for the excellent,pragmatic reason that this has been found to work well in practice.

The current version incorporates no new chapter but a number ofnew topics have been introduced, e.g. ipso aromatic substitution; themechanistic borderline in nucleophilic substitution; more use ofactivation parameters, particularly in ester hydrolysis; Dimroth's Erparameter; correlation of spectroscopic data with Hammett's 0":<;

13e n.m.r. in biogenesis, etc. The now outmoded term 'carboniumion' has been replaced throughout by 'carbocation', which has theadvantage of being the natural antithesis to carbanion, and avoidsthe rather dubious alternative of carbenium ion. Apart from thesemore obvious changes, the whole text has been gone through, lineby line, in an effort to remove ambiguities, to provide clearer, morecogent explanations, and more telling examples. The overall result,in garage parlance, has been a very thorough overhaul and extensivere-tune!

It has always been my feeling that many textbooks fall short oftheir full potential because the author has never entirely made uphis or her mind whether the subject matter is addressed wholly tostudents or, in part at least, to their mentors; and the requirementsof the two are, after all, different. This new edition is directed, aswere its predecessors, unequivocally at the student; I trust thereforethat it will continue to prove helpful to chemistry students ingeneral, irrespective of the particular institution in which theyhappen to be studying.

As always I am greatly indebted to many correspondents whohave pointed out errors, infelicities, and made suggestions forimprovements; wherever feasible these have been incorporated inthis revision. I should greatly appreciate similar kind assistance fromreaders in the future.

xii Preface to sixth edition

Finally, acknowledgement is made to the copyright holders forpermission to reprint diagrams as follows: the American ChemicalSociety for Fig. 13.1 (Hammett, L. P. and Pfluger, H. L., J. Arner.Chern. Soc., 1933,55,4083), Fig. 13.2 (Hammett, L. P. and PflugerH. L., J. Arner. Chern. Soc., 1933, 55, 4086), Fig. 13.3 (Hammett,L. P., Chern. Rev., 1935,17, 131), Fig. 13.4 (Taft, R. W. and Lewis,I. c., J. Arner. Chern. Soc., 1958,80,2437), Fig. 13.5 (Brown, H. C.and Okamoto, Y., J. Arner. Chern. Soc., 1957,79, 1915), Fig. 13.6(Brown, H. c., Schleyer, P. von R. et al., J. Arner. Chern. Soc.,1970, 92, 5244), Fig. 13.8 (Hart, H. and Sedor, F. A, J. Arner.Chern. Soc., 1967,89,2344); the Chemical Society and Professor J.A. Leisten for Fig. 13.7 (Leisten, J. A and Kershaw, D. N., Proc.Chern. Soc., 1960, 84).

CambridgeSeptember 1985

PETER SYKES

1Structure, reactivity, and mechanism

1.1 ATOMIC ORBITALS, p. 1.1.2 HYBRIDISATION, p. 4.1.3 BONDING IN CARBON COMPOUNDS, p. 5:

1.3.1 Carbon-earbon single bonds, p. 6; 1.3.2 Carbon-carbondouble bonds, p. 8; 1.3.3 Carbon-earbon triple bonds, p. 9; 1.3.4Carbon-ilxygen and carbon-nitrogen bonds, p. 10; 1.3.5Conjugation, p. 11; 1.3.6 Benzene and aromaticity, p. 14; 1.3.7Conditions necessary for delocalisation, p. 18.

1.4 THE BREAKING AND FORMING OF BONDS, p. 20.1.5 FACTORS INFLUENCING EL.ECrRON-AVAILABILITY, p. 21:

1.5.1 Inductive and field effects, p. 21; 1.5.2 Mesomeric(conjugative) effects, p. 23; 1.5.3 Time-variable effects, p. 24; 1.5.4Hyperconjugation, p. 25.

1.6 STERIC EFFECTS, p. 26.1.7 REAGENT TYPES, p. 28.1.8 REACTION TYPES, p. 30.

The chief advantage of a mechanistic approach, to the vast array ofdisparate information that makes up organic chemistry, is the way inwhich a relatively small number of guiding principles can be used, notonly to explain and interrelate existing facts, but to forecast the outcome of changing the conditions under which already known reactionsare carried out, and to foretell the products that may be expectedfrom new ones. It is the business of this chapter to outline some ofthese guiding principles, and to show how they work. As it is thecompounds of carbon with which we shall be dealing, somethingmust be said about the way in which carbon atoms can form bondswith other atoms, especially with other carbon atoms.

1.1 ATOMIC ORBITAlS

The carbon atom has, outside its nucleus, six electrons which, on theBOhr theory of atomic structure. were believed to be arranged inorbits at increasing distance from the nucleus. These orbits corres-

x

3

y

18

2p~, 2p, and 2pzcombined

Is

1.1 Atomic orbitals

JP }M shell3s _

2P }L shell18 _

Is K shell

- 1, respectively), all of the same shape and energy level (orbitalshaving the same energy level are described as degenerate), but differingfrom each other in their spatial orientation. They are in fact arrangedmutually at right-angles along notional x, y and z axes and, therefore,designated as 2px, 2py and 2pz, respectively. Further, these three 2porbitals are found not to be spherically symmetrical, like the Is and2s, but 'dumb-bell' shaped with a plane, in which there is zero probabilityof finding an electron (nodal plane~ passing through the nucleus (atright-angles to the x, y and z axes, respectively), and so separating thetwo halves of each dumb-bell:

The six electrons of the carbon atom are then accommodated inatomic orbitals of increasing energy level until all are assigned (theaufbau, or build-up, principle~ Thus two electrons, with paired spins,will go into the Is orbital, a further two into the 2s orbital, but at the2p level the remaining two electrons could be accommodated either inthe same, e.g. 2px' or different, e.g. 2px and 2py, orbitals. Hund's rule,which states that two electrons will avoid occupying the same orbitalso long as there are other energetically equivalent, i.e. degenerate,orbitals still empty, will apply, and the electron configuration of carbonwill thus be IS22s22P~2P~, with the 2pz orbital remaining unoccupied.This represents the ground state of the free carbon atom in whichonly two unpaired electrons (in the 2px and 2py orbitals) are available

Nodal yplane

z

x

NodJplane

2p~ 2p, 2pz

Structure, reactivity, and mechanism2

• n can have values of 1,2,3, ... ,/ values of 0, 1,2, ... , n - I, and m values of0, ± I, ±2, ... , ± I. We shall normally be concerned only with / values of 0 and I, thecorresponding orbitals being referred to (from spectroscopic terminology) as sand porbitals, respectively, e.g. Is, 2s, 2p orbitals, etc.

t One electron with spin quantum number +t, the other -to

ponded to gradually increasing levels of energy, that of lowest energy,the Is, accommodating two electrons, the next, the 2s, also accommodating two electrons, and the remaining two electrons of the carbonatom going into the 2p level, which is actually capable of accommodating a total of six electrons.

The Heisenberg indeterminacy principle, and the wave-mechanicalview of the electron, have made it necessary to do away with anythingso precisely defined as actual orbits. Instead the wave-like electronsare now symbolised by wave functions, t/J, and the precise, classicalorbits of Bohr are superseded by three-dimensional atomic orbitalsof differing energy level. The size, shape and orientation of theseatomic orbitals-regions in which there is the greatest probability offinding an electron corresponding to a particular, quantised energylevel-are each delineated by a wave function, t/JA' t/JB' t/Jc' etc. Theorbitals are indeed rather like three-dimensional electronic contourmaps, in which t/J2 determines the relative probability of finding anelectron at a particular point in the orbital.

The relative size of atomic orbitals, which is found to increase astheir energy level rises, is defined by the principal quantum number, n,their shape and spatial orientation (with respect to the nucleus andeach other) by the subsidiary quantum numbers, I and m, respectively.·Electrons in orbitals also have a further designation in terms of thespin quantum number, which can have the values +t or -to Onelimitation that theory imposes on such orbitals is that each mayaccommodate not more than two electrons, these electrons beingdistinguished from each other by having opposed (paired) spins. t Thisfollows from the Pauli exclusion principle, which states that no twoelectrons in any atom may have exactly the same set of quantumnumbers.

It can be shown, from wave-mechanical calculations, that the Isorbital (quantum numbers n = 1, I = 0, m = 0, corresponding to theclassical K shell) is spherically symmetrical about the nucleus of theatom, and that the 2s orbital (quantum numbers n = 2, I = 0, m = 0)is similarly spherically symmetrical, but at a greater distance from thenucleus; there is a region between the two latter orbitals where theprobability of finding an electron approaches zero (a spherical nodalswface).

As yet, this marks no radical departure from the classical pictureof orbits, but with the 2p level (the continuation of the L shell) adifference becomes apparent. Theory now requires the existence ofthree 2p orbitals (quantum numbers n = 2, I = 1, with m = + 1, 0, and

1.2 HYBRIDISATION

5J.3 Bonding in carbon compounds

Similar, but different, redeployment is envisaged when a carbonatom combines with three other atoms, e.g. in ethene (ethylene) (p. 8):three Sp2 hybrid atomic orbitals disposed at 120° to each other inthe same plane (plane trigonal hybridisation) are then employed. Finally,when carbon combines with two other atoms, e.g. in ethyne (acetylene)(p. 9): two Spl hybrid atomic orbitals disposed at 1800 to each .oth~r

(digonal hybridisation) are employed. In each case the s orbItal IS

always involved as it is the one of lowest energy level.These are all valid ways of deploying one 2s and three 2p atomic

orbitals-in the case of Sp2 hybridisation there will be one unhybridisedp orbital also available (p. 8), and in the case of Spl hybridisation therewill be two (p. 10). Other, equally valid, modes of hybridisation arealso possible in which the hybrid orbitals are not necessarily identicalwith each other, e.g. those used in CH2Cl2 compared with the onesused in CCl4 and CH4 • Hybridisation takes place so that the atomconcerned can form as strong bonds as possible, and so that theother atoms thus bonded (and the electron pairs constituting thebonds) are as far apart from each other as possible, i.e. so that thetotal intrinsic energy of the resultant compound is at a minimum.

s = 1·00

p = 1·72

Spl = 1.93

Sp2 = 1.99

Sp3 = 2.00

lt will thus be apparent why the use of hybrid orbitals, e.g. Sp3 hybridorbitals in the combination of one carbon and four hydrogen atomsto form methane, results in the formation of stronger bonds.

When the atoms have come sufficiently close together, it can beshown that their two atomic orbitals are replaced by two molecularorbitals, one being at a lower, and the other at a higher, energy levelthan the two original atomic orbitals. These two new molecularorbitals spread over both atoms and either may contain the twoelectrons (Fig. 1.1):

1.3 BONDING IN CARBON COMPOUNDS

Bond formation between two atoms is then envisaged as the progressiveoverlapping of an atomic orbital from each of the participatingatoms, the greater the overlap achieved (the overlap integral), thestronger the bond so formed. The relative overlapping powers ofatomic orbitals have been calculated as follows:

Sp3 hybrids

+-GG+

Structure, reactivity, and mechanism

2s

4

A carbon atom combining with four other atoms clearly does not usethe one 2s and the three 2p atomic orbitals that would now be available,for this would lead to the formation of three directed bonds, mutuallyat right angles (with the three 2p orbitals), and one different, nondirected bond (with the spherical 2s orbital). Whereas in fact, the fourC-H bonds in, for example, methane are known to be identical andsymmetrically (tetrahedrally) disposed at an angle of 109° 28' to eachother. This may be accounted for on the basis of redeploying the 2sand the three 2p atomic orbitals so as to yield four new (identical)orbitals, which are capable of forming stronger bonds (cf p. 5). Thesenew orbitals are known as Sp3 hybrid atomic orbitals, and the processby which they are obtained as hybridisation:

for the formation of bonds with other atoms, i.e. at first sight carbonmight appear to be only divalent.

This, however, is contrary to experience, for though compounds areknown in which carbon is singly bonded to only two other atoms,e.g. CCl2 (p. 267), these are highly unstable: in the enormous majorityof its compounds carbon exhibits quadrivalency, e.g. CH4 • This canbe achieved by uncoupling the 2s2 electron pair, and promoting oneof them to the vacant 2p= orbital. The carbon atom is then in a higherenergy (excited) state, Is22s12p~2p;'2Pl, but as it now hasfour unpairedelectrons it is able to form four, rather than only two, bonds withother atoms or groups. The large amount of energy produced byforming these two extra bonds considerably outweighs that required[;:::: 406 kJ (97 kcal) mol-I] for the initial 2s 2 uncoupling, and 2s --+ 2pzpromotion.

It should, however, be emphasised, despite the diagram above, thathybridisation is an operation carried out not actually on orbitalsthemselves but on the mathematical functions that define them.

1.3.1 Carbon--carbon single bonds

The combination of two carbon atoms, for example in ethane, resultsfrom the axial overlap of two Sp3 atomic orbitals, one from each

·The anti-bonding molecular orbital is referred to as a <1. orbital.

~he mole~ular orbital of lower energy is called the bondingorbital, and Its occupancy results in the formation of a stable bondb~tw~en the two atoms. In the above case, the pair of electrons constItutmg the bond tend to be concentrated between the two positivelycharged atomic nuclei, which can thus be thought of as being heldtogether by the negative charge between them. The molecular orbitalof higher energy is called the anti-bonding orbital; this correspondsto a state in which the internuclear space remains largely empty ofelectrons, an~ thus ~esults in .repulsion between the two positivelycharged atomIc nucleI. The antI-bonding orbital remains empty in theground state of the molecule, and need not here be further consideredin the formation of stable bonds between atoms.

If overlap of the two atomic orbitals has taken place along theirmajor axes, the resultant bonding molecular orbital is referred to as au orbital,· and the bond formed as a u bond. The u molecular orbital,and the electrons occupying it, are found to be localised symmetricallyabout the internuclear axis of the atoms that are bonded to each other.Thus on combining with hydrogen, the four hybrid Sp3 atomic orbitalsof carbon overlap with the Is atomic orbitals of four hydrogen atomsto form four identical, strong u bonds, making angles of 109° 28' witheach other (the regular tetrahedral angle~ in methane. A similar,exactly regular, tetrahedral structure will result with, for example,CCI4 , but where the atoms bonded to carbon are not all the same,e.g. CH2CI2 , the spatial arrangement may depart slightly from theexactly symmetrical while remaining essentially tetrahedral (ct. p. 5).

Fig. 1.1

7

H~HH~H

H

1.3.1 Carbon-carbon single bonds

~HH

carbon atom, to form a strong u bond between them. The carboncarbon bond length in saturated compounds is found to be prettyconstant---O·154 nm (1·54 A). This refers, however, to a carbon--earbonsingle bond between Sp3 hybridised carbons. A similar single bondbetween two Sp2 hybridised carbons, =CH-CH=, is found on averageto be about 0·147 nm (1·47 A) in length, and one between two Splhybridised carbons, =C-C ,about 0·138 nm (1·38 A). This is notreally surprising, for an s orbital and any electrons in it are held closerto, and more tightly by, the nucleus than is a p orbital and any electronsin it. The same effect will be observed with hybrid orbitals as theirs component increases, and for two carbon atoms bonded to eachother the nuclei are drawn inexorably closer together on going from

Sp3_Sp3 -+ Sp2_Sp2 -+ Spl_Spl.

We have not, however, defined a unique structure for ethane; the(J bond joining the two carbon atoms is symmetrical about a linejoining the two nuclei, and, theoretically, an infinite variety of differentstructures is still possible, defined by the position of the hydrogenson one carbon atom relative to the position of those on the other.The two extremes, of all the possible species, are known as the eclipsedand staggered forms:

k1H 0H

H H H H

Eclipsed Staggered

The above quasi three-dimensional representations are known as'sawhorse' and Newman projections, respectively. The eclipsed andstaggered forms, and the infinite variety of possible structures lyingbetween them as extremes, are known as conformations of the ethanemolecule; conformations being defined as different arrangements ofthe same group of atoms that can be converted into one anotherwithout the breaking of any bonds.

The staggered conformation is likely to be the more stable of thetwo as hydrogen atoms on one carbon are then as far away fromthose on the other as they can get (0,310 nm; 3·1 A), and anyso-called 'non-bonded' interaction between them is thus at a

1Atomic orbital B

1!Bonding (<1)

molecular orbital

Anti-bonding (<1-)molecular orbital

~ t~ Atomic orbital A


9J.3.3 Carbon-carbon triple bonds

* An anti-bonding, It*. molecular orbital is also fonned (ef p.12).

1.3.3 CarboD---(:arbon triple bonds

In ethyne each carbon atom is bonded to only two other atoms, onehydrogen and one carbon. Strong C1 bonds are formed with these twoatoms by the use of two hybrid orbitals derived by hybridising the 2sand, this time, one only of the carbon atom's 2p atomic orbitals. Theresultant digonal Spl hybrid orbitals are co-linear. Thus, in forming themolecule of ethyne, these hybrid orbitals are used to form strong C1

bonds between each carbon atom and one hydrogen atom, andbetween the two carbon atoms themselves, resulting in a linear molecule

This new bonding molecular orbital is known as a 7[ orbital,· and theelectrons that occupy it as 7[ electrons. The new 7[ bond that is thusformed has the effect of drawing the carbon atoms closer together,and the C=C distance in ethene is found to be 0·133 nm (1·33 A),compared with a C-C distance of 0·154 nm (1·54 A) in ethane. Thelateral overlap of the p atomic orbitals that occurs in forming a 7[

bond is less effective than the axial overlap that occurs in forming aC1 bond, and the former is thus weaker than the latter. This is reflectedin the fact that the energy of a carbon-earbon double bond, thoughmore than that of a single bond is, indeed, less than twice as much.Thus the C-C bond energy in ethane is 347 kJ (83 kcal) mol-I, whilethat of C=C in ethene is only 598 kJ (143 kcal) mol- I.

The lateral overlap of the two 2p atomic orbitals, and hence thestrength of the 7[ bond, will clearly be at a maximum when the twocarbon and four hydrogen atoms are exactly coplanar, for it is only inthis position that the p atomic orbitals are exactly parallel to eachother, and will thus be capable of maximum overlap. Any disturbance of this coplanar state, by twisting about the (T bond joiningthe two carbon atoms, would lead to reduction in 7T overlapping,and hence a decrease in the strength of the 7T bond: it will thus beresisted. A theoretical justification is thus provided for the longobserved resistance to rotation about a carbon-carbon double bond.The distribution of the 7T electrons in two lobes, above and belowthe plane of the molecule, and extending beyond the carbon-carbonbond axis, means that a region of negative charge is effectivelywaiting there to welcome any electron-seeking reagents (e.g. oxidising agents); so that it comes as no surprise to realise that thecharacteristic reactions of a carbon-earbon double bond are predominantly with such reagents (cf. p, 178). Here the classical pictureof a double bond has been replaced by an alternative, in which thetwo bonds joining the carbon atoms, far from being identical,are considered to be different in nature, strength and position.

-+


minimum' whereas in the eclipsed conformation they are sufferingthe max~um of crowding [0·230 nm (2,3 A), slightly less than thesum of their van der Waals radii]. The long cherished principle of I

free rotation about a carbon-earbon single bond is not contravened,however as it has been shown that the eclipsed and staggered i

conform~tions differ by only = 12 kJ (3 kcal) mol-1 in energy contentat 250 and this is small enough to allow their ready interconversionthrou~h the agency of ordinary thermal motions at roomtemperature-the rotation frequency at 25° being =10 12 sec- l

. Thatsuch crowding can lead to a real restriction of rotation about acarbon-carbon single bond has been confirmed by the isolation oftwo forms of CHBr2CHBr2 , though admittedly only at low temperatures where collisions between molecules do not provide enoughenergy to effect the interconversion.

8

1.3.2 Carbon-carbon double bonds

In ethene each carbon atom is bonded to only three other atoms, twohydrogens and one carbon. Strong C1 bonds are formed with thesethree atoms by the use of three orbitals derived by hy~ridisi~g the 2sand this time, two only of the carbon atom's 2p atomic orbitals-anato~ will normally only mobilise as many hybrid orbitals as it hasatoms or groups to form strong C1 bonds with. The resultant Sp2 hybridorbitals all lie in the same plane, and are inclined at 1200 to each other(plane trigonal orbitals). In forming the molecule of ethen~, two.of theSp2 orbitals of each carbon atom are seen as overlappmg with theIs orbitals of two hydrogen atoms to form two strong C1 C-H bonds,while the third Sp2 orbital of each carbon atom overlap axially to forma strong C1 C-C bond between them. It is found experimentally thatthe H-C-H and H-C-C bond angles are in fact 116·,]" and 121·6°,respectively. The departure from 120" is hardly surprising seeing thatdifferent trios of atoms are involved. .

This then leaves, on each carbon atom, one unhybridised 2p atomicorbital at right angles to the plane containing the carbon and hydrogenatoms. When these two 2p orbitals become parallel to each otherthey can themselves overlap, resulting in the formation of a b~:mding

molecular orbital spreading over both carbon atoms and Situatedabove and below the plane (i.e. it has a node in the plane of themolecule) containing the two carbon and four hydrogen atomsC', indicates bonds to atoms lying behind the plane of the paper,and" bonds to those lying in front of it):

II

(2b)

(I b)

1.3.5 Conjugation

-+

-

(2a)

(Ia)

0" 0 I)CH· OH V~. 'clf i.', 'CH" 6) 6 2

The C-C-O bond angle is found to be :::::; 120°, the C=O bondlength, Q.122 nm (1·22 A~ and the bond energy, 750 kJ (179 kcal) mol-I.The fact that this is very slightly greater than twice the C-O bondenergy, whereas the C=C bond energy is markedly less than twicethat of C-C, may be due in part to the fact that the lone pairs onoxygen are further apart, and so more stable, in C=O than in C-O;there being no equivalent circumstance with carbon. The fact thatcarbon-oxygen, unlike carbon-earbon, bonds are polar linkages alsoplays a part.

A nitrogen atom, with the electron configuration ls22s22p~2PPp~,

can also be looked upon as using hybrid orbitals in forming slOgle,C-N, double, C=N, and triple, C=N, bonds with carbon. In eachcase one such orbital is used to accommodate the nitrogen lone pairof electrons; in double and triple bond formation one and two 1t bonds,respectively, are also formed by lateral overlap of the unhybridised porbitals on nitrogen and carbon. Average bond lengths and bondenergies are single, 0·147 nm (1·47 A) and 305 kJ (73 kcal) mol- I,double, 0·129nm (1.29A) and 616kJ (147kcal)mol- l , and triple,0·116 om (1·16 A) and 893 kJ (213 kcal) mol- I.

1.3.5 Conjugation

When we come to consider molecules that contain more than onemultiple bond, e.g. dienes with two C=C bonds, it is found thatcompounds in which the bonds are conjugated (alternating multipleand single; 1) are slightly more stable than those in which they areisolated (2) :

-+


having two unhybridised 2p atomic orbitals, at right angles to eachother, on each of the two carbon atoms. The atomic orbitals on onecarbon atom are parallel to those on the other, and can thus overlapwith each other resulting in the formation of two 1t bonds in planesat right angles to each other:

The ethyne molecule is thus effectively sheathed in a cylinder ofnegative charge. The C=C bond energy is 812kJ (194kcal)mol- l

, sothat the increment due to the third bond is less than that occurring ongoing from a single to a double bond. The C=C bond distance is0·120nm (I.20A) so that the carbon atoms have been drawn stillfurther together, but here again the decrement on going C=C -+ C=Cis smaller than that on going C-C --+ C=C.

1.3.4 Carbon-oxygen and carbon-nitrogen bonds

An oxygen atom has the electron configuration ls22s22p;2p~2p~, andit too, on combining with other atoms, can be looked upon as utilisinghybrid orbitals so as to form the strongest possible bonds. Thus oncombining with the carbon atoms of two methyl groups, to formmethoxymethane (dimethyl ether), CH3-Q-CH3 , the oxygen atomcould use four Sp3 hybrid orbitals: two to form (T bonds by overlapwith an Sp3 orbital of each of the two carbon atoms, and the othertwo to accommodate its two lone pairs of electrons. The C-O-Cbond angle is found to be 110°, the C-O bond length, 0·142 nm(1·42 A), and the bond energy, 360 kJ (86 kcal) mol-I.

An oxygen atom can also form a double bond to carbon; thus inpropanone (acetone), Me2C=Q: , the oxygen atom could use three Sp2hybrid orbitals: one to form a a bond by overlap with an Sp2 orbitalof the carbon atom, and the other two to accommodate the two lonepairs of electrons. This leaves an unhybridised p orbital on bothoxygen and carbon, and these can overlap with each other laterally(cf C=C, p. 9) to form a 1t bond:

BaseMeCH=CH-CHz-C=O -----., MeCHz-CH=CH-C=O

I ~.~ IMe Me

It will be seen from Fig. 1.2 that accommodating the four electronsof the conjugated diene (1a) in the two bonding orbitals as shownleads to a lower total energy for the compound than-by analogywith ethene-accommodating them in two localised 7t bonds. The

This greater thermodynamic stability (lower energy content) of conjugated molecules is revealed in (1) having a lower heat of combustion,and a lower heat of hydrogenation than (2); and also in the generalobservation that isolated double bonds can often be made to migratequite readily so that they become conjugated:

13/.3.5. Conjugation

Colour

colourlessyelloworangered

C6 H 5(CH=CH).C6 H 5

n=1n = 2-4n=Sn=8

electrons are said to be delocalised, as they are now held in commonby the whole of the conjugated system rather than being localisedover two carbon atoms in 7t bonds, as in ethene or in (1 b). Accommodation of ~he four elec~ro~s in. the. bonding molecular orbitals 1/11 and 1/12results m electron dIstributIOn m a charge cloud as in (3) :

/~Me~Hz

(3)

For such delocalisation to occur the four p atomic orbitals in (1a)woul.d have to be. es.sentially parallel, and this would clearly imposeco~slderable restrictIOns on rotation about the C2-e3 bond in (3),whIch is indeed observed in practice as highly preferred conformations.It might also be expected that the 7t electron density between C2 andC3 would result in this bond having some double bond charactere.g. in its being shorter than a C-C single bond. The observed bondlength is indeed short~·147 nm (1·47 A}-though no shorter thanmight be expected for a single bond between Sp2 hybridised carbonatoms (cf p. 9). The stabilisation energy of a simple conjugated diene,compared with the corresponding isolated one, is relatively smallca. 17 kJ (4 kcal) mol- '-and even this cannot be ascribed wholly toelectron delocalisation: the state of hybridisation of the carbon atomsinvolved, and the consequent differing strengths of the (1 bonds betweenthem, must also be taken into account.

Delocal.isation is, however, much involved in stabilising the excitedstates of dl~nes, ~nd of polyenes in general, i.e. in lowering the energylevel of theIr excIted states. The effect of this is to reduce the energygap between ground and excited states of conjugated molecules, ascompared with those containing isolated double bonds, and' this~nergy gap i~ progressively lessened as the extent of conjugationmcreases. ~hls means that the amount of energy required to effectth.e p~omotlOn of an electron, from ground to excited state, decreasesWlt~ I~cre~sing conjugation, i.e. the wavelength at which the necessaryr~dlatlOn IS absorbed increases. Simple dienes absorb in the ultraVIOlet region, but as the extent of conjugation increases the absorptiongradually mo~es.to~ards the visible range, i.e. the compound becomescoloured. ThIS IS Illustrated by the series of (Xw-diphenylpolyenesbelow:

i!I/IzrT} Bonding1/11--

MOs

Butadiene

1/14-}Anti-bonding

i ! i ! 1/13-

pAOs

Fig. 1.2

Ethylene (x2)

It*--

ItMOs

i !

p AOs i!It--


Itenergy

12

Conjugation is not of course confined to carbon-earbon multiplebonds.

With both (1a) and (2a) above, lateral overlap of the p atomicorbitals on adjacent carbon atoms could lead to the formation of twolocalised 7t bonds as shown, and the compounds would thus beexpected to resemble ethene, only twice as it were! This is indeedfound to be the case with (2), but (1) is found to behave differentlyin terms of its slightly greater stability (referred to above), in spectroscopic behaviour (see below), and in undergoing addition reactionsmore readily than does an isolated diene (p.194). On looking moreclosely, however, it is seen that with (1 a), but not with (2a), lateraloverlap could take place between all four p atomic orbitals on adjacentcarbon atoms. Such overlap will result in the formation of fourmolecular orbitals (Fig. 1.2), two bonding (1/1 1 and 1/12) and two 'antibonding (1/13 and 1/14}-the overlap of n atomic orbitals always givesrise to n molecular orbitals:

151.3.6 Benzene and aromaticity

Nodal,Plane.

HHuGHM()~OOaIHH H H HHplane

(Sa) (5b) (Sci

H

H

The net result is an annular electron cloud, above and below theplane of the ring (6):

(6)

each of these two MOs accommodates two more electrons-thusmaking six in all :

The influence of this cloud of negative charge on the type of reagentsthat will attack benzene is discussed below (p. 131).

Support for the above view is provided by the observation that allthe carbon--earbon bond lengths in benzene are exactly the same,·O·I40nm (1·40A~ i.e. benzene is a regular hexagon with bond lengthssomewhere in between the normal values for a single (0,154 nm;1·54 A) and a double (0,133 nm; 1·33 A) bond. This regularity maybe emphasised by avoiding writing Kekule structures for benzene, asthese are clearly an inadequate representation, and using instead:

There remains, however, the question of the much remarkedthermodynamic stability of benzene. Part of this no doubt arises fromthe disposition of the three plane trigonal u bonds about each carbonat their optimum angle of 1200 (the regular hexagonal angle~ but alarger part stems from the use of cyclic, delocalised molecular orbitalsto accommodate the six residual electrons; this is a considerably morestable (lower energy) arrangement than accommodating the electrons

• As also are all the c-H bond lengths at ().108 nm (1·08 M

(4b)(4)

MOs

"'6- }. d"Anti-bon 109"'4-- ~5

i! i!"'2- ~)}i! Bonding"'.--

It*--

(40)

i !

Atoms Ethylene (x3) Benzene

Fig. 1.3

i!pAOs n

ItMOs

It

Energy


The bonding MO of lowest energy (1/1 d is cyclic and embraces allsix carbon atoms, i.e. is delocalised. It has a nodal plane in the planeof the ring, so that there are two annular lobes, one above and onebelow the plane of the ring, of which only the upper one (looking downfrom above) is shown in (5a): two electrons are thus accommodated.The two further bonding MOs (1/12 and 1/1J~ ofequal energy (degenerate),also encompass all six carbon atoms, but each has a further nodalplane, at right angles to the plane of the ring, in addition to the onein the plane of the ring. Each MO thus has four lobes of which onlythe upper pair (looking down from above) are shown in (5b) and (5c);

Overlapping could, of course, take place, 1,2; 3,4; 5,6; or 1,6; 5,4; 3,2,leading to formulations corresponding to the Kekule structures (40and 4b); but, as an alternative, all six adjacent p orbitals could overlap,as with conjugated dienes (p. 12), resulting in the formation of sixmolecular orbitals, three bonding (1/1 I --+ 1/1 J) and three anti-bonding(1/14 --+ 1/16~ with energy levels as represented below (Fig. 1.3):


One of the major problems of elementary organic chemistry is thedetailed structure of benzene. The known planar structure of themolecule implies Sp2 hybridisation with p atomic orbitals, at rightangles to the plane of the nucleus, on each of the six carbon atoms (4):

14

II

17

(14)


(12)

surprising for cyclic p orbital overlap, as with benzene, would require(9) to be flat with a consequent C-C-C bond angle of 135°, resultingin very considerable ring strain for an array of Sp2 hybridised carbons(preferred angle 120°). Such strain can be relieved by puckering of thering, but only at the expense of sacrificing the possibility of overall porbital overlap. That such puckering occurs can be seen from X-raycrystallographic measurements, which show cyclooctatetraene to havethe 'tub' structure (9a) with alternating double (0,133 nm; 1·33 A),and single (Sp2_Sp 2, 0·146 nm; 1·46 A; cf. p. 7), carbon~arbon

honds:

~()'133nm

~146nm(9a)

Conditions necessary for cyclic polyenes to possess aromatic characterare referred to below.

The amount by which benzene is stabilised compared with thehypothetical 'cyclohexatriene' should properly be called its stabilisationenergy, it is, however, often called its delocalisation energy, whichimmediately begs the question as to how much of the stabilisation isactually due to delocalisation of the 671 electrons in benzene. The termresonance energy, though still widely used, is highly unsatisfactory onsemantic grounds as it immediately conjures up visions of rapidoscillation between one structure and another, e.g. the Kekulestructures, thus entirely misrepresenting the real state of affairs (c;{.p. 19).

The requirements necessary for the occurrence of aromatic stabilisation, and character, in cyclic polyenes appear to be: (a) that the moleculeshould be flat (to allow of cyclic overlap of p orbitals); and (b) thatall the bonding orbitals should be completely filled. This latter conditionIS fulfilled in cyclic systems with 4n + 271 electrons· (Hiickel's rule),and the arrangement that occurs by far the most commonly in aromaticcompounds is when n = I, i.e. that with 671 electrons. IOn electrons(n = 2) are present in naphthalene [12, stabilisation energy, 255 kJ(61 kcal) mol-I], and 1471 electrons (n = 3) in anthracene (13) andphenanthrene (14)----stabilisation energies, 352 and 380 kJ (84 and91 kcal) mol- I, respectively:

10000e(n = 2) 14lte(n = 3) 14lte(n = 3)• Significantly, cyclooctatetraene with 81t electrons (4n, n = 2) has already been shown

not to be aromatic (p. 16).


O~O~O(9) (10) (11)

~H = -410 kJ( -98 kcal) mol- t ~H = -96 kJ (-23kcal) mo)-l

The difference between ~H for (9) and 4 x ~H for (11) is thus minus26 kJ (-6 kcal) mol- I : cyclooctatetraene, unlike benzene, exhibits nocharacteristic stabilisation when compared with the relevant hypothetical cyclic polyene (it is in fact slightly destabilised), i.e. it is notaromatic. This lack of aromatic character is, on reflection, not really

in three localised 7l molecular orbitals, as is apparent from Fig. 1.3(p. 14). The much greater stabilisation in benzene than in, for exa~ple,conjugated dienes (cf p. 13) presumably stems from benzene bemg acyclic, i.e. closed, symmetrical system. .

A rough estimate of the stabilisation of benze~e, compared w.lthsimple cyclic unsaturated structures, can be obtamed by companngits heat of hydrogenation with those of cyclohexene (7) and cyclohexal,3-diene (8):o + H

2- 0 ~H = -120kJ(-28·6kcal)mol-

1

(7)o +2H2

- 0 ~H = -232kJ(-SS·6kcal)mol-1

(8)

@ +3H2 - 0 ~H= _208kJ(-49·8kcal)mol-1

The heat of hydrogenation of the cyclic diene (8) is very nearly twicethat of cyclohexene (7), and the heat of hydrogenation of the threedouble bonds in a Kekule structure might thus be expected to be ofthe order of 3 x - 120 kJ (- 28·6 kcal) mol- I = - 360 kJ (- 85·8 kcal)mol-I. but when 'real' benzene is hydrogenated only - 208 kJ ( - 49·8kcal) ~ol-I are evolved. 'Real' benzene is th us thermodynamically morestable than the hypothetical 'cyclohexatriene' ~r 151 kJ .(36 kcal) ~ol- 1 ;

this compares with only ~ 17 kJ (4 kcal) mol b~ whlc~ a conJu~ateddiene is stabilised, with respect to its analogue 10 whIch there IS nointeraction between the electrons of the double bonds.

In marked contrast to benzene above. the heat of hydrogenation of I

cyclooctatetraene (9) to cyclooctane (10) is -410 kJ (-98 kcal) mol-I,while that of cyclooctene (11) is -96 kJ (-23 kcal) mol-I:

16

Further, the ring structure need not be purely carbocyclic, and pyridine(18, cf. p. 165~ for example, with a nitrogen atom in the ring and 6ne(n = 1), is as highly stabilised as benzene:

©N

(18)

19J.3.7 Conditions necessary for delocalisation

where, in flat contradiction of the above formula, X-ray crystallography shows that the two oxygen atoms are indistinguishable fromeach other, the two carbon-oxygen bond distances being the same, i.e.involving the same number of electrons.

These difficulties have led to the convention of representing moleculesthat cannot adequately be written as a single classical structure by acombination of two or more classical structures, the so-called canonicalstructures, linked by a double-headed arrow. The way in which oneof these structures can .be related to another often being indicated bycurved arr?ws, the tall of the curved arrow indicating where anelectron paIr moves frotn and the head ofthearrow where it moves to. :

twO, four and six electrons, respectively, is clearly inadequate: somebonds involve other, even fractional, numbers of electrons. This isseen very clearly in the ethanoate (acetate) anion (19),

o~

CHJC"

0 9

(19)

~o r [ /..0]9CHJC - CHJC == CHJC'

"-~ '\.~ ""o 0 0

(19a) (l9b) (l9ab)

It cannot be too firmly emphasised, however, that the ethanoate aniondoe.s n~t have two. possible, and alternative, structures which arerapIdly mterconvertlble, but a single, real structure (I9ab)-sometimesreferred to as a hybrid-for which the classical (canonical) structures(19a) and (19b) are less exact, limiting approximations.

A certain number of limitations must be borne in mind, howeverwhen consider.iog delocalisation and its representation through tw~or m~re classIcal structures as above. Broadly speaking,. the morecanODlcal structures that can be written for a compound, the greaterthe delocalisation of electrons, and the more stable the compound willbe. These structures must not vary too widely from each other in~nergy content, however, or those of higher energy will contribute sohttle to the hybrid as to make their contribution virtually irrelevant.The stabilising effect is particularly marked when the structures havethe same energy content, as with (19a) and (19b) above. Structures

. • We shall, however, subsequently write canonical structures, e.g. (19a) and (19b)hnked by a double-headed arrow, but without curved arrows. These will be reservedfor indicat!ng a real mo~ement of electron pairs, i.e. as happens during the fonning,and breakmg, of bonds In the course of a real reaction.

(17)(16)(IS)


Though these substances are not monocyclic like benzene-and:Hiickel's rule should not, strictly, apply to them-the introduction ofthe transannular bond, that makes them bi- and tricyclic, respectively"seems to cause relatively little perturbation, so far as delocalisationof the n electrons over the cyclic group of ten or fourteen carbonatoms is concerned.

Quasi-aromatic structures are also known in which the stabilisedcyclic species is an ion, e.g. the cycloheptatrienyl (tropylium) cation(15, cf. p. 106), the cyclopentadienyl anion (16, cf. p. 275~ both ofwhich have 6ne (n = n and even more surprisingly the cyclopropenylcation (17, cf. p. 106) which has 2ne (n = 0):

A useful experimental criterion of aromatic character, in addition!to those already mentioned, arises from the position of the signal.from hydrogen atoms attached to the ring carbons in the compound'snuclear magnetic resonance (n.m.r.) spectrum.· The position of then.m.r. signal from a hydrogen atom depends on the nature, i.e. localenvironment, of the carbon (or other atom) to which it is attached.Thus the proton signal of cyclooctatetraene is seen at 85·6, and such aposition is typical of protons in a non-aromatic cyclic polyene, whilethe proton signal of benzene is seen at 82·8, which is found to be typical'of aromatic compounds in general.

1.3.7 Conditions necessary for delocalisation

The difficulty in finding a satisfactory representation for the carbon- .carbon bonding in benzene brings home to us the fact that our normalway of writing bonds between atoms as single, double or triple, involving

- A useful simple account of the use or n.m.r. (and other) spectra in an organic contextmay be found in: Williams, D. H. and Fleming, I. Spectroscopic Methods in OrganicChemistry, McGraw-Hili, 3rd Edition, 1980

R' + Br-Br -. R-Br + Br'

R· + H10 -. R-OH+H.

1.4 THE BREAKING AND FORMING OF BONDS

A covalent bond between two atoms can be broken in essentiallyfollowing ways:

211.5.1 Inductive and field effects

1.5 FACfORS INFLUENCING ELECfRON-AVAILABILITY

In the lignt of what has been said above, any factors that influence therelative availability of electrons (the electron density) in partiCUlarbonds, or at particular atoms, in a compound might be expected toaffect very considerably its reactivity towards a particular reagent:a position of high electron availability will be attacked with difficultyif at all by, for example, SOH, whereas a position of low electronavailability is likely to be attacked with ease, and vice versa with apositively charged reagent. A number of such factors have beenrecognised.

solvents, because of the greater ease of separation of charge therein,and very often because of the stabilisation of the resultant ion pairsthrough solvation. Many of these ionic intermediates can be consideredas carrying their charge on a carbon atom, though the ion is oftenstabilised by delocalisation of the charge, to a greater or lesser extent,

COHvez~octHhe-ercHaZr~o?n.Hat:~s, or atoms of diff:ren~:::m[e~t::=CH~HZJ. +=! CHz=CH-CHz-OHz +=! • t

CHz-CH=CHz

?i 6o CH 3-C-CHzII SOH

CH 3-C-CH 3 +=! t + HzO0 6

ICH 3-C=CHz

When a positive charge is carried on carbon the entity is known as acarbocation, and when a negative charge, a carbanion. Though suchions may be formed only transiently and be present only in minuteconcentration, they are nevertheless often of paramount importancein controlling the reactions in which they participate.

These three types, radicals, carbocations and carbanions, by nomeans exhaust the possibilities of transient intermediates in whichcarbon is the active centre; others include the electron-deficient speciescarbenes, R2C: (p. 266), nitrenes, R~ (p. 122); and also arynes (p.174).

1.5.1 Inductive and field effects

In a covalent single bond between unlike atoms, the electron pairforming the (J bond is never shared absolutely equally between thetwo atoms; it tends to be attracted a little more towards the moreelectronegative atom of the two. Thus in an alkyl chloride (20), the

R· ·x/'

R:X -. R: 6 X·~

R· :X6


involving separation of charge (cf p. 24) may be written but, other"things being equal, these are usually of higher energy content thad;those in which such separation has not taken place, and hence contribute,correspondingly less to the hybrid. The structures written must am:contain the same number of paired electrons, and the constituenttatoms must all occupy essentially the same positions relative to eacMother in each canonical structure. If delocalisation is to be significant,;all atoms attached to unsaturated centres must lie in the same planeor nearly so; examples where delocalisation, with consequent stabilisa-I,tion, is actually prevented by steric factors are discussed subsequently~~ .

Such radicals or ion pairs are formed transiently as reactive intermediates in a very wide variety of organic reactions, as will be shown Ibelow. Reactions involving radicals tend to occur in the gas phase'and in solution in non-polar solvents, and to be catalysed by lightand by the addition of other radicals (p. 300~ Reactions involving';ionic intermediates take place more readily in solution in polar'

In the first case each atom separates with one electron, leading to theformation of highly reactive entities called radicals, owing theirreactivity to their unpaired electron; this is referred to as homolyticfission of the bond. Alternatively, one atom may hold on to bothelectrons, leaving none for the other, the result in the above case beinga negative and a positive ion, respectively. Where R and X are noridentical, the fission can, of course, take place in either of two ways, .~'as shown above, depending on whether R or X retains the e1ectrori,pair. Either of these processes is referred to as heterolytic fission, the',!result being the formation of an ion pair. Formation of a covalene;bond can take place by the reversal of any of these processes, and also, •of course, by the attack of first-formed radicals or ions on otherspecies:

• The metal atoms in, for example, lithium alkyls and Grignard reagents, both ofwhich compounds are largely covalent, are also electron-donating, leading to negativelypolarised carbon atoms in each case: R-+ Li and R-+ MgHal (el. p.221).

electron density tends to be greater nearer chlorine than carbon, asthe former is the more electronegative; this is generally representedas in (20a) or (20b):

If the carbon atom bonded to chlorine is itself attached to furthercarbon atoms, the effect can be transmitted further:

C-C-C+-C~CI

432 1

23

:; [ /CH"\:.. /CH"\;. JMe CH 0 J

(22ab)

"Ell e [""'+" '~-JC+-O:; C+-O/ /

(21b) (21ab)

1.5.2 Mesomeric (conjugative) effects

"C=i=O +-+/

(21a)

The actual structure is somewhere in between, i.e. (21ab) a hybrid ofwhich (2Ia) and (21 b) are the canonical forms. There will also be aninductive effect, as shown in (21ab) but this will be much smaller thanthe mesomeric effect as (T electrons are much less polarisable, andhence less readily shifted, than 7T electrons. . .

If the C=O group is conjugated with C=C, the above polansatlOncan be transmitted further via the TC electrons, e.g. (22):

Delocalisation takes place (cf 1,3-dienes, p. 13), so that an electrondeficient atom results at C3 ' as well as at C t as in a simple c~rbonyl

compound. The difference between this transmission via a conjugatedsystem, and the inductive effect in ~ saturated s.ys~em, is that here t~eeffect suffers much less diminution by its transmiSSIOn, and the polantyat adjacent carbon atoms alternates. ., ..

The stabilisation that can result by delocahsatlOn of a positive ornegative charge in an ion, via its TC orbitals, can be a potent feature inmaking the formation of the ion possible in the fir~t pla~ (cf p. 55).It is, for instance, the stabilisation of the phenoxlde amon (23), bydelocalisation of its charge via the delocalised TC orbitals of the nucleus,that is largely responsible for the acidity of phenol (cf p. 56):

~ +Hp ~ l6~6~6~·6tH'O'(23.) (2") (23') (23d) IJ

An apparently similar delocalisation can take place. in undissociatedphenol (24) itself, involving an unshared electron pair on the oxygen

1.5.2 Mesomeric (conjugative) effects

These are essentially electron redistributions that can take place inunsaturated, and especially in conjugated, systems via their TC orbitals.An example is the carbonyl group (p. 203), whose properties are notaccounted for entirely satisfactorily by the classical formulation (21a),nor by the extreme dipole (21b) obtainable by shift of the TC electrons:

Ell

/CH'\. /CH'\. +-+ /CH" ~CH"Me CH 0 Me CH Oe

(22a) (22b)

"-C+-C1/

(20b)

"H ~-C-C1/

(20a)


The effect of the chlorine atom's partial appropriation of the electronsof the carbon-chlorine bond is to leave C t slightly electron-deficient;this it seeks to rectify by, in turn, appropriating slightly more than itsshare of the electrons of the (T bond joining it to Cz, and so on downthe chain. The effect of C t on Cz is less than the effect of Cion C t ,

however, and the transmission quickly dies away in a saturated chain,usually being too small to be noticeable beyond Cz. These influenceson the electron distribution in (T bonds are known as inductiveeffects.

In addition to any inductive effect operating through the bonds ina compound, an essentially analogous effect can operate eitherthrough the space surrounding the molecule or, in solution, via themolecules of solvent that surround it. In many cases, however, it isnot possible to distinguish between the operation of an inductiveeffect as such, and this closely similar (and parallel) field effect.Subsequently, reference to an inductive effect will, therefore, normally be taken to include any such field effect.

Most atoms and groups attached to carbon exert such inductiveeffects in the same direction as chlorine, i.e. they are electron-withdrawing, owing to their being more electronegative than carbon, themajor exception being alkyl groups which are electron-donating.·Though the effect is quantitatively rather small, it is responsible forthe increase in basicity that results when one of the hydrogen atoms ofammonia is replaced by an alkyl group (p. 66), and, in part at any rate,for the readier substitution of the aromatic nucleus in methylbenzenethan in benzene itself (p. 153).

All inductive effects are permanent polarisations in the ground stateof a molecule, and are therefore manifested in its physical properties,for example, its dipole moment.

but this involves separation of charge, and will thus be correspondinglyless effective than the stabilisation of the phenoxide ion which does not.

Mesomeric, like inductive, effects are permanent polarisations in theground state ofa molecule, and are therefore manifested in the physicalproperties of the compounds in which they occur. The essentialdifference between inductive and mesomeric effects is that whileinductive effects can operate in both saturated and unsaturatedcompounds, mesomeric effects can operate only in unsaturated,especially in conjugated, compounds. The former involve the electrons in u bonds, the latter those in "IT bonds and orbitals. Inductiveeffects are transmitted over only quite short distances in saturatedchains before dying away, whereas mesomeric effects may be transmitted from one end to the other of quite large molecules providedthat conjugation (i.e. delocalised "IT orbitals) is present, throughwhich they can proceed.

25

(26d)(26c)

1.5.4 Hyperconjugation

(26b)

H H~

I eH-C-CH=CH ...... H-C=CH-CHI 2 I 2

H H

(25a) (25b)

(200)

This effect has been called hyperconjugation, and has been usedsuccessfully to explain a number of otherwise unconnected phenomena.It should be emphasised that it is not suggested that a proton actuallybecomes free in (25) or (26), for if it moved from its original positionone of the conditions necessary for delocalisation to occur would becontroverted (p. 20).

Reversal of the expected (inductive) order of electron-donation toCH3 > MeCHz> MezCH > Me3 C could be explained on the basis of

Me""Me+-C+- >

7'Me

variable factors, the actual close approach of a reagent may have aprofound effect in enhancing reactivity in a reactant molecule, and soin promoting reaction.

1.5.4 Hyperconjugation

The relative magnitude of the inductive effect ofalkyl groups is normallyfound to follow the order,

as would be expected. When, however, the alkyl groups are attachedto an unsaturated system, e.g. a double bond or a benzene nucleus,this order is found to be disturbed, and in the case of some conjugatedsystems actually reversed. It thus appears that alkyl groups are capable,in these circumstances, of gi ving rise to electron release by a mechanismdifferent from the inductive effect. This has been explained asproceeding by an extension of the conjugative or mesomeric effect,delocalisation taking place in the following way:


atom,

1.5.3 Time-variable effects

Some workers have sought to distinguish between effects such as thetwo considered above, which are permanent polarisations manifestedin the ground state of a molecule, and changes in electron distributionthat may result either on the close approach of a reagent or, moreespecially, in the transition state (p. 38) that is formed from itsinitial attack. The time-variable factors, by analogy with the permanent effects discussed above, have been named the inductomeric andelectromeric effects, respectively. Any such effects can be lookedupon as polarisabilities rather than as polarisations, for the distribution of electrons reverts to that of the ground state of the moleculeattacked either if the reagent is removed without reaction beingallowed to take place, or if the transition state, once reached,decomposes to yield the starting materials again.

Such time-variable effects, being only temporary, will not, of course,be reflected in the physical properties of the compounds concerned.It has often proved impossible to distinguish experimentally betweenpermanent and time-variable effects, but it cannot be too greatlyemphasised that, despite the difficulties in distinguishing what proportions of a given effect are due to permanent and what to time-

This results in the preferential formation of non-terminal alkenes inreactions which could lead to either these or their terminal iso~ers?n intr.od~ction of the double bond (p. 256), and to the fairly readyIsomensatlon of the less to the more stable compound e.g. (31) ~(30). '

27

(33)

The most common steric effect, however, is the classical sterichindrance, in which it is apparently the sheer bulk of groups that isinfluencing the reactivity of a site in a compound directly: by impedingapproach of a reagent to the reacting centre, and by introducingcrowding in the transition state (cf p. 38), and not by promoting orinhibiting electron-availability. This has been investigated closely inconnection with the stability of the complexes formed by trimethylboron with a wide variety of amines. Thus the complex (34) formedwith triethylamine dissociates extremely readily, whereas the complex(35) with quinuclidine, which can be looked upon as having threeethyl groups on nitrogen that are 'held back' from assuming a conformation that would interfere sterically with attack on the nitrogen

The 2,6-dimethyl derivative (33) does not couple under theseconditions, however, despite the fact that the methyl groups that havebeen introduced are much too far away for their bulk to interferedirectly with attack at the p-position. The failure to couple at thisposition is, in fact, due to the two methyl groups, in the a-positions tothe NMe

2, interfering sterically with the two methyl groups attached

to nitrogen, and so preventing these lying in the same plane as thebenzene nucleus. This means that the p orbitals on nitrogen, and onthe ring carbon atom to which it is attached, are prevented frombecoming parallel to each other, and their overlapping is thus inhibited.Electronic interaction with the nucleus is thus largely prevented, andtransfer of charge, as in (32), does not take place (cf p. 71):

1.6 Steric effects

(32)

leading to substitution at the p-position (cf p. 153):

Me Me Me Me Me Me'\" / '\,,$ / '\" .. /q '"',., Q~ ~

PhN=N PhN=N H PhN=N$(29)

~eMe-C-CH=CHI 2

Me

(28)

1H3

MeCH2 -C=CH2

(31)

~eMe-C-CH=CHI 2

H

(27)

1H3

CH3-C=CH-CH3

(30)

(25)


hyperconjugation being dependent on the presence of hydrogen onthe ~arbon a.toms a- to the unsaturated system. This is clearly at amaXImum With CH3 (25) and non-existent with Me3 C (29),

H MeI I

H-C-CH=CH2 H-C-CH=CHI I 2

H H

26

hence the increased electron-donating ability of CH3 groups underthese conditions. Hyperconjugation could, however, involve C-Cas well as C-H bonds, and the differences in relative reactivityobserved in a series of compounds may actually result from theoperation of solvent, as well as hyperconjugative, effects.

Hyperconjugation has also been invoked to account for thegreater thermodynamic stability of alkenes in which the doublebond is not terminal, e.g. (30), compared with isomeric compoundsin which it is, e.g. (31): in (30) there are nine 'hyperconjugable'a-hydrogen atoms, compared with only five in (31):

1.6 STERIC EFFECTS

We have to date been discussing factors that may influence the relativeavailability ofelectrons in bonds, or at particular atoms, in a compound,and hence affect that compound's reactivity. The operation of thesefactors may, however, be modified or even nullified by the influenceof steric factors; th us effective delocalisation via 1t orbitals can onlytake place if the p or 1t orbitals, on the atoms involved in the delocalisation, can become parallel or fairly nearly so. If this is preventedsignificant overlapping cannot take place, and delocalisation may b~inhibited. A good example of this is provided by a comparison betweenthe behaviour of N,N-dimethylaniline (32) and its 2,6-dialkyl derivatives, e.g. (33). The NMe2 group in (32), being electron-donating (dueto the unshared electron pair on nitrogen interacting with thedelocalised 1t orbitals of the nucleus), activates the nucleus towardsattack by the diazonium cation PhN2 al, i.e. towards azo-coupling,

atom, is very stable:

will tend to be most readily attacked by positively charged cations suchas C6 HsN/ll, a diazon!um cation (p. 146), or by other species which,thou~ not actually catIOns, possess an atom or centre that is electrondeficIent; for example, the sulphur atom of sulphur trioxide (37) in

291.7 Reagent types

0'·,. ,L,'O~Si+++

.\"0'·

(37)

sulphonation (p. 140):

Such reagents, because they tend to attack the substrate at a position(or positions) of high electron density, are referred to as electrophilicreagents or electrophiles.

Conversely, an electron-deficient centre, such as the carbon atomin chloromethane (38)

which acts as an acid by accepting the electron pair on nitrogen intrimethylamine to form the complex (40), and is therefore referred

e eF38 + :NMe] +=! F38: NMe]

(39) (40)

'+ ,H3C...-Cl

(38)

will tend to be attacked most readily by (negatively charged) anionssuch as eOH, eCN, etc., or by other species which, though not actuallyanions, possess an atom or centre which is electron-rich; for example,the nitrogen atom in ammonia or amines, H3 N: or R3 N:. Suchreagents, because they tend to attack the substrate at a position (orpositions) of low electron density, i.e. where the atomic nucleus isshort of its normal complement of orbital electrons, are referred to asnucleophilic reagents or nucleophiles.

It must be emphasised that only a slightly unsymmetrical distributionof electrons is required for a reaction's course to be dominated: thepresence of a full-blown charge on a reactant certainly helps, but isfar from being essential. Indeed the requisite unsymmetrical chargedistribution may be induced by the mutual polarisation of reagentand substrate on their close approach, as when bromine adds toethene (p. 180).

This electrophile/nucleophile dichotomy can be looked upon as aspecial case of the acid/base idea. The classical definition of acids andbases is that the former are proton donors, and the latter protonacceptors. This was made more general by Lewis, who defined acidsas compounds prepared to accept electron pairs, and bases as substances that could provide such pairs. This would include a numberof compounds not previously thought of as acids and bases, e.g. borontrifluoride (39),

(35)


6~6(36a) (36b)

That this difference is not due to differing electron availability at thenit~ogen .atom in t~e two cases is confirmed by the fact that the twoammes dIffer very !lttI~ in their strengths as bases (cf p. 72): the uptakeof a proton constltutmg very much less of a steric obstacle than theuptake of the rel~tively bU.lky BMe3 • Esterification and ester hydrolysisare other reactions partIcularly susceptible to steric inhibition (cf.~Mn .

It should be emphasised that such steric inhibition is only anex~reme. case, and any factors which disturb or inhibit a particular~ment~tlOn of the reactants with respect to each other, short of prevent109 theIr close ~pproac~, can also profoundly affect the rate of reactions:a state of affaIrs that IS often encountered in reactions in biologicalsystems.

1.7 REAGENT TYPES

R~ference. has already ~en made t? electron-donating and electronwlthdrawl.ng groups, theIr effect bemg to render a site in a molecule~Iectron-nch or electron-deficient, respectively. This will clearlymflu~nce the type of reagent with which the compound will mostreadIly react. An electron-rich species such as phenoxide anion (36)

1.8 REACTION TYPES

*He, BH4e , HS03e, HOe, ROe, RSe , eCN, RCOze, RC=Ce, eCH(COzEt)z

0:, N:, S:, RMgBr, RLi (41)

Electrophiles:

HEll, H30Ell, EllN02 , EllNO, PhN2Ell , R3C Ell

* • • • •S03' CO2 , BF3, AICI3, rCI, Brz' 0 3

N ucleophiles :

311.8 Reaction types

OH""- / e

C +CN/ ""-

CN

oe""-H 6- sCN ""- / HCN

C=o 4 ~ C 4 ~/c V Slow / ""- Fa"CN

eCN

Elimination reactions are, of course, essentially the reversal ofaddition reactions; the most common type is the loss of hy.drogen andanother atom or group from adjacent carbon atoms to yIeld alkenes(p. 246):

H H'I / ""- / -HO ""-I /, -HOr C=C -"- C-CC-C ---. / ""- / I"

/ 1""- OHBr

which can be initiated by the attack of either H(J) (p.l~4) or Br: ~p. 317)on the double bond. By contrast, the addition reactIOns exhIbIted bythe carbon-oxygen double bond, in simple aldehydes and ~etones, areusually nucleophilic in character. (p .. 20~) ..An exa~ple IS the basecatalysed formation of cyanohydnns III hqUId HeN.

NCe + R-Br _ NC-R + Bre

Br""- / HOr ""- 1/

C=C - C-C/ ""- /1 ""-

H

but nucleophilic displacement of hydrogen is also known (p. 167).Radical-induced displacement is also known, for example the halo-genation of alkanes (cf. p. 323). . . d' I

Addition reactions, too, can be electrophilic, nu~l~~phlhc or ra Icain character, depending on the type of species that m!tIates the pro.cess.Addition to simple carbon--earbon doubl~.bonds IS normally eItherelectrophile-, or radical-, induced; e.g. addItIOn of HBr,

In nucleophilic substitution it is often an atom other than hydrogenthat is displaced (p. 77),

In electrophilic substitution it is often .hydrogen that is d~splaced,classical aromatic substitution (p. 132) bemg a good example.

H N02

@+@NO, ~ @+H'


to as a Lewis acid. Electrophiles and nucleophiles in organic reactionscan be looked upon essentially as acceptors and donors, respectively,of electron pairs, from and to other atoms-most frequently carbon.Electrophiles and nucleophiles also, of course, bear a relationship tooxidising and reducing agents, for the former can be looked upon aselectron acceptors and the latter as electron donors. A number of themore common electrophiles and nucleophiles are listed below:

Where a reagent is starred, the star indicates the atom that acceptselectrons from, or donates electrons to, the substrate as the case maybe. No clear distinction can necessarily be made between what constitutes a reagent and what a substrate, for though (J)N02 , eOH, etc.,are normally thought of as reagents, the carbanion (41) could, at will,be either reagent or substrate, when reacted with, for example, an alkylhalide. The reaction of the former on the latter is a nucleophilic attack,while that of the latter on the former would be looked upon as anelectrophilic attack; but no matter from which reactant's standpointa reaction is viewed, its essential nature is not for a moment in doubt.

It should be remembered that reactions involving radicals as thereactive entities are also known. These are much less susceptible tovariations in electron density in the substrate than are reactionsinvolving polar intermediates, but they are greatly affected by theaddition of small traces of substances that either liberate or removeradicals. They are considered in detail below (p. 313).

There are essentially four general types of reaction which organiccompounds can undergo:

(a) Displacement (substitution)(b) Addition(c) Elimination(d) Rearrangement

In (a) it is displacement from carbon that is normally referred to, butthe atom displaced can be either hydrogen or another atom or group.

32 Structure, reactivity, and mechanism

The actual r~arrangement step in such reactions is often followed bya further displacement, addition or elimination, before a stableend-product is obtained.

2.1 ENERGETICS OF REACTION

2.1 ENERGEITCS OF REAcnON, p. 33.2.2 KINETICS OF REAcnoN, p. 36:

2.2.1 Reaction rate and free energy of activation, p. 37; 2.2.2Kinetics and the rate-limiting step, p. 39; 2.2.3 Kinetic v.thermodynamic control, p. 42.

2.3 INVESTIGATION OF REAcnON MECHANISMS, p. 43:2.3.1 The nature of the products, p. 43; 2.3.2 Kinetic data, p. 44;2.3.3 The use of isotopes, p. 46; 2.3.4 The study of intermediates,p. 49; 2.3.5 Stereochemical criteria, p. 51.

Energetics, kinetics, and theinvestigation of mechanism

When we consider the conversion of starting materials into products, which constitutes an organic reaction, one of the things thatwe particularly want to know is 'how far will the reaction go overtowards products?' Systems tend to move towards their most stablestate, so we might expect that the more stable the products are,compared with the starting materials, the further over in theformer's favour any equilibrium between them might be expected tolie, i.e. the larger l1stability is in the diagram (Fig. 2.1) below, thegreater the expected conversion into products:

We have now listed a number of electronic and steric factors that caninfluence the reactivity of a compound in a given situation, and alsothe types of reagent that might be expected to attack particular centresin such a compound especially readily. We have as yet, however, hadlittle to say directly about how these electronic and steric factors,varying from one structure to another, actually operate in energeticand kinetic terms to influence the course and rate of a reaction. Theseconsiderations are of major importance, not least for the light theymight be expected to throw on the detailed pathway by which areaction proceeds.

2

(43)(42)

Rearrangements may also proceed via intermediates that areessential~y cations, anions, or radicals, though those involvingcarbocatlons, or other electron-deficient species, are by far the mostcommon. They may involve a major rearrangement of the carbonskeleton of a compound, as during the conversion of 2 3dirnethylbutan-2,3-diol (pinacol, 42) into 2,2-dirnethylbutan-3-dne(pinacolone, 43, cf. p. 113):

.1G = .1H - T.1S

• H is a measure of the heat content, or enthalpy, of a compound, and .1H is precededby a minus sign if the products have a lower heat content than the starting materials:when there is such a decrease in enthalpy the reaction is exothermic.

3S2.1 Energetics of reaction

and it is found that the change in free energy in going from startingmaterials to products, AGe- (AGe- refers to the change under standardconditions: at unit activity; less exactly at unit, i.e. molar, concentration), is related to the equilibrium constant, K, for the change by therelation,

-.1G e = 2·303RTlogK

i.e. the larger the decrease in free energy (hence, minus AGe-) on goingfrom starting materials to products, the larger the value of K, and thefurther over the equilibrium lies in favour of products. The positionof minimum free energy thus corresponds to the attainment ofequilibrium by starting materials/products. In a reaction for whichthere is no free energy change (AGe = 0) K = 1, which correspondsto 50% conversion of starting materials into products. Increasingpositive values of AGe imply rapidly decreasing fractional values ofK (the relationship is a logarithmic one), corresponding to extremely little conversion into products, while increasing negativevalues of AGe imply correspondingly rapidly increasing values of K.Thus a AGe of -42 kJ (-10 kcal) mol- 1 corresponds to an equilibrium constant of =107

, and essentially complete conversion intoproducts. A knowledge of the standard free energies of startingmaterials and of products, which have been measured for a largenumber of organic compounds, thus enables us to predict theexpected extent of the conversion of the former into the latter.

The AH factor for the change can be equated with the difference inenergies between the bonds in the starting materials and the bonds inthe products, and an approximate value of AH for a reaction canoften be predicted from tables of standard bond energies: which ishardly unexpected, as it is from AH data that the average bond energieswere compiled in the first place!

The entropy factor cannot be explained quite so readily, buteffectively it relates to the number of possible ways in which theirtotal, aggregate energy may be shared out among an assembly ofmolecules; and also to the number of ways in which an individualmolecule's quanta of energy may be shared out for translational,rotational, and vibrational purposes, of which the translational islikely to be by far the largest in magnitude. Thus for a reaction inwhich there is an increase in the number of molecular species ongoing from starting materials to products,

A ~B+C

there is likely to be a sizeable increase in entropy because of the gain intranslational freedom. The - TAS term may then be large enough tooutweigh the + AH term of an endothermic reaction, thus leading toa negative value for AG, and an equilibrium that lies well over infavour of products. If the reaction is exothermic anyway (AH negative),

Products

L\ sla bil ily

Fig. 2.1

rEii Starting materials

~Oll

.~

~

~

Energetics, kinetics, and the investigation ofmechanism34

However, it quickly becomes apparent that the simple energychange that OCcurs on going from starting materials to products, andthat may readily be measured as the heat of reaction, AH*, is not anadequate measure of the difference in stability between them, for thereis often found to be no correlation between AH and the equilibriumconstant for the reaction, K. Highly exothermic reactions are knownwith only small equilibrium constants (little conversion of startingmaterials into products), and some reactions with large equilibriumconstants are known that are actually endothermic (enthalpy ofproducts higher than that of starting materials): clearly some factorin addition to enthalpy must be concerned in the relative stability ofchemical species.

That this should be so is a corollary of the Second Law of Thermodynamics which is concerned essentially with probabilities, and withthe tendency for ordered systems to become disordered: a measure ofthe degree of disorder of a system being provided by its entropy, S. 'In seeking their most stable condition, systems tend towards minimumenergy (actually enthalpy, H) and maximum entropy (disorder orrandomness), a measure of their relative stability must thus embrace acompromise between Hand S, and is provided by the Gibb's freeenergy, G, which is defined by,

G = H - TS

where T is the absolute temperature. The free energy change during areaction, at a particular temperature, is thus given by,

37

Products

1

x

Starting>. materialsl>O.....c..~u.

Products

Fig. 2.3

Fig. 2.2

1 S"rt;..~ materialsg..~

2.2.1 Reaction rate andfree energy ofactivation

2.2.1 Reaction rate and free energy of activationThe position x in the energy profile above. (Fig. 2.3) c~rresponds. tothe least stable configuration through whIch th~ startIng materIalspass during their conversion into products, and IS generally referred

AGe is negative and large in magnitude, so that the equilibrium liesessentially completely over in favour of CO2 and ~20; but a ~ewspaper (very largely cellulose) can be read, in the. aIr (?r even In .anoxygen tent!) for long periods of time WIthout .It ~otIceably fadIngaway to gaseous products: the rate of the converSIOn IS extreme~y .slowunder these conditions despite the very large _AGe, t~ough It IS,. ofcourse, speeded up at higher t~mperature~. The cgn~erslOn o~ startIngmaterials into products, despIte a negatIve AG ,~s rarely if ever amere run down-hill (Fig. 2.2), there is generally a bamer to be overcome

en route (Fig. 2.3):

H ...... H.. H /H'0 0...... ····if 0/

" I...... ,I 1/......C-C, .....C-C,

inter-molecular

H,O···H-....O,I I ............C-C,

intra-molecular


It should n~t be overlooked that the entropy term involves temperature (TA~) whIle the enthalpy (AH) term does not, and their relativecontrIbutIOns to the free energy change may be markedly differentfor the same reaction carried out at widely differing temperatures.

AG will of course be even more negative, and the equilibrium constant,K, ~orrespondingly larger still. Where the number of participatingspecIes decreases on g~ing from starting materials to products there islIkely to be a decrease In entropy (AS negative); hence,

A+B~C .:1G=.:1H-(-)T.:1S

and unless the reaction is sufficiently exothermic (AH negative andlarg~. e~ough) to counterbalance this, AG will be positive, and theeqUIlIbrIUm thus well over in favour of starting materials.

Cyclisation reactions may also be attended by a decrease in entropy,

H 2C

H c/ 'CHCHicH2lJCH=CH2 +=t 2I I 2

H2C, /CH2CH 2

for tho~gh. th~re is no necessary change in translational entropy, aconstraInt. I~ Impos~ on rot.ation about the carbon-earbon single?onds: thIs IS essentIally free m the open chain starting material, butIS gre~t1y restricted in th~ c~c1ic product. This rotational entropy~erm IS, h~wever, ~maller In SIze than the translational entropy termInvolved In reactIons where the number of participating speciesdecreases on forming products-a fact that is reflected in thepreference for intra- rather than inter-molecular hydrogen bondingin 1,2-diols:

2.2 KINETICS OF REACfION

Thou.gh a negative value for AGe is a necessary condition for areactIOn to take place at all under a given set of conditions, furtherinformation is still needed as the -AGe value tells us nothing abouthow fast the starting materials are converted into products. Thus forthe oxidation of cellulose,

392.2.2 Kinetics and the rate-limiting step

it comes as nO surprise to find a rate equation,

where k is known as the rate constant for the reaction. The reaction issaid to be second order overall; first order with respect to CH 3Br,and first order with respect to eOH.

Such coincidence of stoichiometry and rate law is fairly uncommon,the former is commonly no guide at all to the latter, which can onlybe obtained by experiment. Thus for the base catalysed bromination


Experimentally, the measurement of reaction rates consists in investigating the rate at which starting materials disappear andlor productsappear at a particular (constant) temperature, and seeking to relatethis to the concentration of one, or all, of the reactants. The reactionmay be monitored by a variety of methods, e.g. directly by the removalof aliquots followed by their titrimetric determination, or indirectlyby observation of colorimetric, conductimetric, spectroscopic, etc.,changes. Whatever method is used the crucial step normally involvesmatching the crude kinetic data against variable possible functions ofconcentration, either graphically or by calculation, until a reasonablefit is obtained. Thus for the reaction,

and subsequent calculation.The L\S* term (the entropy ofactivation) again relates to randomness.

It is a measure of the change in degree of organisation, or ordering, ofboth the reacting molecules themselves and of the distribution ofenergywithin them, on going from starting materials to the transition state;L\S* is related to the A factor in the Arrhenius equation above. Ifformation of the transition state requires the imposition of a highdegree of organisation in the way the reactant molecules must approacheach other, and also of the concentration of their energy in particularlinkages so as to allow of their ultimate breakage, then the attainmentof the transition state is attended by a sizeable decrease in entropy(randomness), and the probability of its formation is correspond-

ingly decreased.

of Eact

may then be obtained graphically by plotting values of logl 0 kagainst liT, or by conversion of the above equation into,

HH~,

C-Br-+/

H

HOe +


to as an activated complex or transitionthat th~s is merely a highly unstable sta~~a~~a~t .should be emphas~seddynamic process, and not a discret IS ~assed .through m athat can actually be detected or e m?lecular speCies, an Intermediate,(I) in the alkaline hydrolysis of ~ven Isolat~ (cf..p. 49): An example isbond is being formed at the same ti romo~etcane, m whl.ch t~e HO-Cme as e - Br bond IS being broken,

[

H H J* HlJ- , / lJ I H

HO ....·f·.. B~ -+ Ho-d~H + Bre

(1)

and the three hydrogen atoms attached b .a configuration in which they all Ii . to car on are pa.ssmg throughthe plane of the paper) Th' e. m ~ne plane (at nght-angles to(p. 77). . IS reactIOn IS discussed in detail below

The height of the barrier in (Fi 23) L\G* .of activation for the reaction (the gh' h' .. '~s called the free energyand can be considered as being mad Ig err Isht e slower the reaction),(TL\S*) terms: e up 0 ent alpy (L\H*) and entropy

~G* = ~H* - T~S*

L\H* (the enthalpy of activation) corr dto effect the stretching or even br k~sponf ~ to the en~rgy necessaryprerequisite for reaction to take ep~ mg(o onds t~at IS an essentialbond in 1) Thu' ace e.g. stretchmg of the C-Br

. s reactmg molecules must b . . h hcollision a certain minimum th h I nng Wit t em to anypossible (often called simply th rest.o d. of energy for reaction to beL\H*); the well-known increa~aci~v~~on energy, Eacl' bu~ related totemperature is raised is indeed derate of ~ reactIOn as themolecules with an ener~y abov~ t~~ to. t~e growmg proportion ofrises. IS mmlmum as the temperature

The magnitude of E for a reaction m beof k, the rate constant'(cj p 39) d t 3:Y calcul.ated from valuesdifferent temperatures T' a~d T e e~mmed expenm~ntally at twowhich relates k to T the albsol t t2 ' usmg the Arrhemus expression, u e emperature:

k = Ae- EIRT or log k - ~ I,0 - -2.303RT + ogJOA

Where R is the gas constant (8'32' I -1-1constant for the reaction-inde end JOU es mol deg ), and A is ato the proportion of the total ~um~nt o~tem~e.rature-that is relatedmolecules that result in successful co er 0 . col~lslOns between reactantnverslOn mto products. The value

* The symbol * will often be applied to a str ..an attempted representation of a tran 't' ucture to mdlcate that it is intended as

Sl Ion state (T.S.).


of the two steps, and hence will be the slower, i.e. the step whose rateour kinetic experiments will actually be measuring. It is followed by afast (less energy-demanding), non rate-limiting conversion of theintermediate into products. The above bromination of propanone,can,under certain conditions, be said to follow an idealised pattern corresponding to Fig. 2.4, in which slow, rate-limiting removal of protonby base results in the formation of the carbanion intermediate (2),which then undergoes rapid, non rate-limiting attack by Br2 to yieldbromopropanone and bromide ion as the products:

It should be emphasised that though this explanation is a reasonablededuction from the experimentalIy established rate equation, the lattercannot be claimed to prove the former. Our experimentally determinedrate equation will give us information about the species that areinvolved up to and including the rate-limiting step of a reaction: therate equation does indeed specify the composition but not, other thanby inference, the structure of the transition state for the rate-limitingstep. It gives no direct information about intermediates nor, exceptby default as it were, about the species that are involved in rapid,non rate-limiting processes beyond this rate-limiting step,

In considering the effect that a change of conditions, e.g. of solventor in the structure of the starting material, might be expected to haveon the rate of a reaction, we need to know what effect such changeswilI have on the stability (free energy level) of the transition state:any factors which serve to stabilise it will lead to its more rapidformation, and the opposite will also apply. It is seldom possible toobtain such detailed information about these high-energy transitionstates: the best we can commonly do is to take the relevant intermediatesas models for them, and see what effect such changes might be expectedto have on these. Such a model is not unreasonable; the transientlyformed intermediate in Fig. 2.4 closely resembles, in terms of freeenergy level, the transition state that precedes it, and might be expectedto resemble it in structure as well. Certainly such an intermediate isnormalIy likely to be a better model for the transition state than thestarting material would be. Thus (1 complexes (Wheland intermediates)in aromatic e1ectrophilic substitutions are used as models for thetransition states that are their immediate precursors (p. 151).

The effect of a catalyst is to increase the rate at which a reactionwill take place; this is done by making available an alternative path ofless energetic demand, often through the formation of a new, and more

Products

Startingmaterials

~

11'-----------=====

Energetics, kinetics, and the investigation o/mechanism40

of propanone,

Fig, 2.4

C eOHHJCOCHJ + Brz -.. CHJCOCHzBr + HBr

we find the rate equation,

Rate = k[CH JCOCHJ][60H]

i.e. bromine does not appear though [SOH]bromine must be involved ; , does (cf p. 295)., Clearlyis incorporated into th fi t ~me stag~ In the overalI reactIOn as itin the ste wh e na pr uct, but It patently cannot be involvedmust thu~ inv~~:~~ ~~s~r~:ctuaIIY ~easu~ing. ~he overa~1 re~ctioninvolved ( h 0 steps. one In which bromme IS notver few 0 w o~ rate ~e are measuring1 and one in which it is In fact

Thl'Y' br~amc reactlO~s are one-step processes as depicted in 'Fig 23's IS 0 VIOUS enough In • . .

of hexamine, an extreme example such as the formation

stable (lower energy), intermediate:

Thus the rate of hydration of an alkene, directly with water,

432.3.1 The nature of the products

2.3 INVESTIGATION OF REACTION MECHANISMS

It is seldom, if ever, possible to provide complete and entire information,structural, energetic, and stereochemical, about the pathway that istraversed by any chemical reaction: no reaction mechanism can everbe proved to be correct! Sufficient data can nevertheless usually begathered to show that one or more theoretically possible mechanismsare just not compatible with the experimental results, and/or todemonstrate that of several remaining alternatives one is a good dealmore likely than the others.

alternative reactions is reversible, or if the products are readily interconvertible directly under the conditions of the reaction, the composition of the final product mixture may be dictated not by the relativerates offormation of the different products, but by their relative thermodynamic stabilities in the reaction system: we are then seeing thermodynamic or equilibrium control. Thus the nitration of methyl benzene isfound to be kinetically controlled, whereas the Friedel-erafts alkylationof the same species is often thermodynamically controlled (p. 163).The form of control that operates may also be influenced by thereaction condition, thus the sulphonation of naphthalene with concentrated H 2S04 at 800 is essentially kinetically controlled, whereasat 1600 it is thermodynamically controlled (p. 164).

2.3.1 The nature of the products

Perhaps the most fundamental information about a reaction is providedby establishing the structure of the products that are formed duringits course, and relating this information to the structure of the startingmaterial. Where, as is often the case with organic reactions, more thanone product is obtained then it is usually an advantage to know alsothe relative proportions in which the products are obtained, e.g. inestablishing, among other things, whether kinetic or thermodynamiccontrol is operating(cf p.42). In the past this had to bedonelaboriouslyand often imprecisely-by manual isolation of the products, but maynow often be achieved more easily, and precisely, by sophisticatedchromatographic methods or, indirectly, by suitable spectroscopicones.

The importance of establishing the correct structure of the reactionproduct is best illustrated by the confusion that can result when thishas been assumed, wrongly, as self-evident, or established erroneously.Thus the yellow triphenylmethyl radical (3, cf p. 300), obtained fromthe action of silver on triphenylmethyl chloride in 1900, readilyforms a colourless dimer (m.w. = 486) which was-reasonablyenough-assumed to be hexaphenylethane (4) with thirty 'aromatic'

Products

T.S. for rate-limiting stepof catalysed reaction

Fig. 2.5

T.S. for uncatalysed reaction

~ Starting;g materials<Ll

~u..

i


OH"/ '" //

C=C + H2 0 --+ C-C

" / I"H

is often ~xtremely slow,. but it can ~ ~reatly speeded up by the presenceof an aCid catalyst, which effects mltIal protonation of the alkene to acarbocationic intermediate. This is then followed by easy and rapidattack on t~e now positively charged carbocation by a watermolecule actmg as a nucleophile, and finally by liberation of aproton which is able to function again as a catalyst (p. 187):

H20:) Ell

" " OH2 OHC=C/~ cEIl C/ H,O "I / -H~ "I /

~ - ~. C-C to C-C/ " / I" / I" / '"H H H

The details of acid/base catalysis are discussed subsequently (p.74).

2.2.3 Kinetic versus thermodynamic control

Where a starting material may be converted into two or more alternative products,. e.g. in el.ectrophilic attack on an aromatic speciesthat alr~adycarnes a substItuent (p. 150), the proportions in which thealternative pr~ducts are formed are often determined by their relativerate of formatIon: the faster a product is formed the more of it therewil! ~ in the final product mixture; this is known as kinetic control.ThiS IS not always what is observed however, for if one or more of the

Expected Unexpected

The latter cle~rly cannot be obtained from (6) by a simple substitutionprocess.. and eIther must be formed from (6) via a different pathway than(7), Or if the two products are formed through some common intermediate then clearly (7) cannot be formed by a direct substitutioneither.

452.3.2 Kinetic data

attacking species is usually lIlNOz (p. 134), but it is HN03 that weput into the reaction mixture, and whose changing concentration weare measuring; the relationship between the two may well be complexand so, therefore, may be the relation between the rate of reactionand [HN0 3]. Despite the fact that the essential reaction is a simpleone, it may not be easy to deduce this from the quantities that we canreadily measure.

Then again, if the hydrolysis in aqueous solution of the alkylhalide, RHal, is found to follow the rate equation,

Rate = k,[RHal]

it is not necessarily safe to conclude that the rate-determining stepdoes not involve the participation of water, simply on the grounds that[HzO] does not appear in the rate equation; for if wa~er is being u~ed

as the solvent it will be present in very large excess, and Its concentratIOnwould remain virtually unchanged whether or not it actually participated in the rate-limiting stage. The point could perhaps be settledby carrying out the hydrolysis in another solvent, e.g. HCOzH, andby using a much smaller concentration of water as a potentialnucleophile. The hydrolysis may then be found to follow the rateequation,

Rate = kz[RHal][HzO]

but the actual mechanism of hydrolysis could well have changed onaltering the solvent, so that we are not, of necessity, any the wiserabout what actually went on in the original aqueous solution.

The vast majority of organic reactions are carried out in solution,and quite small changes in the solvent used can have the profoundesteffects on reaction rates and mechanisms. Particularly is this so whenpolar intermediates, for example carbocations or carbanions asconstituents of ion pairs, are involved, for such species normally carryan envelope of solvent molecules about with them. This greatly affectstheir stability (and their ease of formation), and is strongly influencedby the composition and nature of the solvent employed, particularlyits polarity and ion-solvating capabilities. By contrast, reactions thatinvolve radicals (p. 299) are much less influenced by the nature of thesolvent (unless this is itself capable of reacting with radicals), but aregreatly influenced by the addition of radical sources (e.g. peroxides)or radical absorbers (e.g. quinones), or by light which may initiatereaction through the production of radicals by photochemical activation, e.g. Brz~ Br' . Br.

A reaction that is found, on kinetic investigation, to proceedunexpectedly faster or slower than the apparently similar reactions,under comparable conditions, of compounds of related structuresuggests the operation of a different, or modified, pathway from the

(8)(7)

9NH1 in----+liq. NH,

Me

(6)


hydrogen atoms. Only after nearly seventy years (in 1968) did then.m.r. spectrum (cf. p. 18) of the dimer (with only twenty-five'aromatic' (H), four 'dienic' (H), and one 'saturated' (H), hydrogenatoms) demonstrate that it could not have the hexaphenylethanestructure (4) and was, in fact (5):

H H

(C6 H shCAHXClC6Hsh ~ 2(C6HshC'*(C6HshC-ClC6H,lj

H (5) H (3) (4)

At which point numerous small details of the behaviour of (3) and ofits dimer, that had previously appeared anomalous, promptly becameunderstandable., Infor~ation about the products of a reaction can be particularlymformatIve when one of them is quite unexpected. Thus the reactionof chloro-4-methylbenzene (p-chlorotoluene, 6) with amide ion eNHin liquid ammonia (p. 173) is found to lead not only to the ~xpected4-methylphenylamine (p-toluidine, 7), but also to the quite unexpected3-methylphenylamine (m-toluidine, 8), which is in fact the majorproduct:

2.3.2 Kinetic data

The la~gest body of information about reaction pathways has comeand stIll does comefrom kinetic studies as we shall see but theinterpretat,ion of ,kinetic dat~ in mechanistic terms (cl p. 39) is notalwa~s qUIte ~ slmpl~ as mIght at first sight be supposed. Thus theeffectIve reactmg speCIes, whose concentration really determines there~ction rate, may differ from the species that was put into the reactionmIxture to start with, and whose changing concentration we areactually seeking to measure. Thus in aromatic nitration the effective

thus cannot be involved in the rate-limiting step (cf p. 136).

472.3.3 The use of isotopes

If the reaction is carried out in water enriched in the heavier oxygenisotope 180, (a) will lead to an alcohol which is 180 enriched and anacid which is not, while (b) will lead to an 180 enriched acid but anormal alcohol. Most simple esters are in fact found to yield an 180

enriched acid indicating that hydrolysis, under these conditions,proceeds via (b) acyl/oxygen fission (p. 238), It should of course beemphasised that these results are only valid provided that neither acidnor alcohol, once formed, can itself exchange its oxygen with waterenriched in 180, as has indeed been shown to be the case.

Heavy water, O 2°' has often been used in a rather similar way.Thus in the Cannizzaro reaction of benzaldehyde (p. 216),

o 0 0 OHII II SOH II I

PhC-H + PhC-H --+ PhC-Oe + PhC-HH,O I

H

(9)

the question arises of whether the second hydrogen atom that becomesattached to carbon, in the molecule of phenylmethanol (benzyl alcohol,

It should be emphasised that primary kinetic isotope effects areobserved experimentally with values intermediate between the maximum calculated value and unity (i.e. no isotope effect): these too canbe useful, as they may supply important information about the breakingof particular bonds in the transition state.

Isotopes can also be used to solve mechanistic problems that arenon-kinetic. Thus the aqueous hydrolysis of esters to yield an acidand an alcohol could, in theory, proceed by cleavage at (a) alkyl/oxygen fission, or (b) acyl/oxygen fission:

o(a) 7(a) R~-O-H + H 180-R'

o IIII I

RC+O-}-R' H,"OI I ~

(b) (bJ'.. 0II

RC_ 180H + H-OR'

Primary kinetic isotope effects are also observable with pairs ofisotopes other than hydrogen/deuterium, but as the relative massdifference must needs be smaller their maximum values will becorrespondingly smaller. Thus the following have been observed:

(14) 114) k ,HOe + 12CH3-I --+ HO- 1ZCH 3 + Ie ~ = 1·09 (25°)

kl4C


general one that might otherwise have been assumed for the seriesThus the observed rates of hydrolysis of the chloromethanes withstrong bases are found, under comparable conditions to vary as I

follows, ,CH 3CI » CH2C12 « CHCI3 » CCI4

cl.earlY suggesting that trichloromethane undergoes hydrolysisdIfferent manner from the other compounds (cf p. 267).

2.3.3 The use of isotopes

It is often a matter of some concern to know whether a particularbon~ .h~s~ or has not, been ~roken. in a step up to and including thera~e lImItIng step of a reactIOn: sImple kmetic data cannot tell usthIS, and further refinements have to be resorted to If fo Ith b d ed . . , r examp e,

e on concern. IS C-H, the question may be settled by comparinthe. rates of r~actIOn, under the same conditions, of the compound i~WhICh we are mterested, and its exact analogue in which this bond hasbeen ~eplaced by a c;-O linkage. The two bonds will have the sameche.mIc.al n~ture as Iso~opes of the same element are involved, buttheIr. vIbratI<?n frequencIes, and hence their dissociation energies willbe slIghtly dIfferent because atoms of different mass are involved'· thegr~ater the mass, the stronger the bond. This difference in bond stre~gthwIll, of course, be reflected in different rates of breaking of the twobonds under comp.arable conditions: the weaker C-H bond beingbroken. more rapI~ly than the stronger C-D bond; quantummechamcal calculatIOn suggests a maximum rate difference k /kof =7 at 250. ' H 0,

Thus in the oxidationOH

/ MnO e

Ph2C" eOH'~ Ph2C=O

H

it is found. tha~ Ph2~HOH is oxidised 6·7 times as rapidly as Ph COOH'the r~actIOn IS saId to exhibit a primary kinetic isotope effect, and~re~~mg of the C-H bond must clearly be involved in the rate~m~tmg step of the reaction. By contrast benzene, C 6 H 6 , and hexa-~ terobenzene, C6 0 6 , are found to u~dergo nitration at essentially

t e s~me rate, and C-H bond-breakmg, that must occur at somestage m the overall process,

found in cultures of several fungi, was built up stepwise frommolecules of ethanoic acid. General confirmation of this hypothesiswas obtained through feeding suitable fungal cultures, in separateexperiments, with 14CH3C02H and CH3

14C02H, respectively. Itwas then found from radioactive counting measurements e4C is a fjemitter), on the two extracted samples of sterigmatocystin(C1sH l20 6 ) that: (i) 14CH3C02H led to the introduction of ~14C

atoms, and CH314C02H to the introduction of 214C atoms. But that

stiIlleaves open the question of exactly where in the sterigmatocystin molecule these two sets of labelled carbon atoms are located.

Not long ago this could have been determined only by extremelylaborious, and often equivocal, selective degradation experiments;but the coming of carbon n.m.T. spectroscopy has now made all thedifference. Neither the l2C nor the 14C carbon isotopes produce ann.m.r. signal but the l3C isotope, which occurs in ordinary carbon to

OH

492.3.4 The study o/intermediates

(b) 913c- atoms(a) .!l13C· atoms -

bon atoms in CH3C02H moleculesKnowing which ~f the t~o car'(ons in sterigmatocystin, it becomesare incorporated mto ~hlch POSI I (ons about the synthetic pathwaypossible to make pertment sugges

II .d ntally it also shows that the

employed by the fungal cultures. nCI e , f CH CO H.h I b of the *CH 0 group does not come rom 3 2

met y car on 3

. . I the N_bromoamide, RCONHBr, itsit is with care, possIble to ISO .ate RNCO' thus going some

. ' CONB e nd an Isocyanate" .amo~, R r ,a. the overall mechanism of the reactl<;>n.conSIderable way to elUCIdate bl" h b yond all doubt that any specIesIt is of course necessar~ to esta

d. IS e d not merely an alternative

isolated really ~ ~n mter~e ::~eb~nconverted, under the normalproduct-by ~~owm.g that It rreaction products at a rate at leastreaction condItions, mto th~ usua der the same conditions. It is alsoas fast as the overall reactIOn un . . lated really is on the directimportant to establish that the specIes ISO

2 3 4 The study of intermediates .. . 'd ce obtainable about the mechamsm

Among the most concret~ eVI ~n the actual isolation of one or moreof a reaction is that provld~d y. Thus in the Hofmann reaction. mediates from the reactIon mIxture.. .(~~e~22), by which amides are converted mto ammes,

°II Br, RNHRC-NH 2 eOM- 2

• 0 d It is thus possible, with suitabl~ ~n-the extent of 1 11 Yo, ~~s. tra of all carbon-contammg

. to record C n.m.r. spec h bostrumentatlOn, . 1.11% l3C content): eac car ncompounds (becafu~de o:ic~~tylrsituated carbon atoms, in a moleculeatom, or group 0 I en . .produci~g a distinguishtbly;~e;~::i:~:t~~ystin can thus be com-

The C spectrum 0 no olecules resulting from separatepared with the spectra. of l~hCe m . h d (a) l3CH C02H and (b)

. . ts with ennc e 3 . hfeedmg experunen . I Those carbon atoms, in each case, whlcCH l3CO H respective y. . 'fi d'

3 2' h d 13C signals can thereby be Identl e .nOW show en ance

(i) g'4C atoms(ii) 2'4C atoms

OH

OCH3

sterigmatocystin

Energetics, kinetics, and the investigation 0/mechanism

(i) 14CH3C02H(ii) CH3'4C02H --

°

48

9) that is formed, comes from the solvent (H20) or from a secondmolecule of benzaldehyde. Carrying out the reaction in D 20 is foundto lead to the formation of no PhCHDOH, thus demonstrating thatthe second hydrogen atom could not have come from water, and musttherefore have been provided by direct transfer from a second moleculeof benzaldehyde.

A wide range of other isotopic labels, e.g. 3H (or T), l3C, 14C, 15N,32p, 358, 37CI, 131 1, etc., have also been used to provide importantmechanistic information. The major difficulties encountered in suchlabelling studies have always been: (a) ensuring that the label isincorporated only into the desired position(s) in the test compound;and (b) finding exactly where the label has gone to in the product(s)after the reaction being studied has taken place.

The enormous increase in selectivity of modern synthetic methodshas all but eliminated (a), but (b) long remained a major problem;particularly when isotopes of carbon were being used: these beingespecially valuable because carbon atoms are present in all organiccompounds. The 14C isotope has been much used in investigatingbiosynthetic pathways: the routes by which living organisms build upthe highly elaborate molecules that may be obtained from them.

Thus there was reason to believe that the pentacyclic compoundsterigmatocystin,

51

(16)

QBr Br

~Br

(15)

Br,--+

2.3.5 Stereochemical criteria

(14)

QBr-Br

(14)

o

PhCOCHMeEt~ PhCOCBrMeEtSOH (+)

(+) -

(13)

, b d' t tep addition of the brominethe reaction cannot sImply e /rech,·one-s t lead to the cis dibromidemolecule to the double bond, lor t IS mus(16) :

2.3.5 Stereochemical criteriaInformation about the stereochemical cO,urse ,followed b~ a par~culareaction can also provide useful insight Into Its mechanism, an

h~~y

r . 'h ggested mec anlstlcwell introduce stringent CrIterIa t at any su I dscheme will have to meet. Thus the fa~t that the base-cata ysebromination of an optically active stereOIsomer of the ketone (13)

The addition must be at least a two~step process (cf p. l79)h~l~a~~~like this, which proceed so as to gIve largely-or even w y

, 2 5 ' d' t thatleads to an optically inactive racemIc produc~ (p. 9 d); In lcah·ecsh can

, d th gh a planar mterme late, w I

~~ed~~:~t~~a~u:~J:I~;e:ell fr~~ efithher sidedleatdiTnghetno ~qg~~~ a~~~~~~, 'age forms 0 t e pro uc. ,

of the two mlrror-I~4 adds on bromine under polar conditions tothat cYclopentedn~b( )'d (15) only indicates that the mechanism ofyield the trans I roml e ,

The successful study of intermediates not only provides one orsi osts which help define the detailed pathway, tra.versed ?y

m~~~cti~Pthe intermediates themselves may also provIde :nferen~lal:vidence ~bout the transition states for which they are 0 ten ta en

as models (cf p. 41).

(12)(11)

1\Me Me


reaction pathway, and not merely in equilibrium with the true intermediate.

It is much more common not to be able to isolate any intermediatesat all, but this does not necessarily mean that none are formed, merelythat they may be too labile or transient to permit of their isolation.Their occurrence may then often be inferred from physical, particularlyspectroscopic, measurements made on the system. Thus in the formation of oximes from a number of carbonyl compounds by reactionwith hydroxylamine (p. 219),

R R"C=O NH,OH ~ "C=N' + H

20

/ /"R R OH

the infra-red absorption band characteristic of C=O in the startingmaterial disappears rapidly, and may have gone completely beforethe band characteristic of C-N in the product even begins to appear.Clearly an intermediate must be formed, and further evidence suggeStsthat it is the carbinolamine (10),

R OH" /C

R/ 'NHOH

(10)

which forms rapidly and then breaks down only slowly to yield theproducts, the oxime and water,

Where we have reason to suspect the involvement of a particularspecies as a labile intermediate in the course of a reaction, it may bepossible to confirm our suspicions by introducing into the reactionmixture, with malice aforethought, a reactive species which we shouldexpect our postulated intermediate to react with particularly readily.It may then be possible to divert the labile intermediate from themain reaction pathway-to trap it-and to isolate a stable speciesinto which it has been unequivocally incorporated. Thus in thehydrolysis of trichloromethane with strong bases (cf p. 46), the highlyelectron-deficient dichlorocarbene, CCI2 , which has been suggested asa labile intermediate (p. 267), was 'trapped' by introducing into thereaction mixture the electron-rich species cis but-2-ene (11), and thenisolating the resultant stable cyclopropane derivative (12), whoseformation can hardly be accounted for in any other way:

~~Me Me

52 Energetics, kinetics, and the investigation ofmechanism

This clearly sets limitations to which any mechanism advanced forthe reaction will have to conform. and gives the lie to that prime tenetof 'lasso chemistry': groups are eliminated most readily when closesttogether:

ACIDS, p. 54: 2 The ori in of acidity in organic compounds,3.1.1 pKQ , p. 54;.3.1. f h

gI t P 56' 31.4 Simple aliphatic

p 55; 3.1.3 The mfluence 0 t e so ve~,: .' . 3 16 Phenols"d p 57' 3 1 5 Substituted aliphatIc aCids, p. 59, :. I"

aCI s,. ,.. "d 62' 3 1 8 Dlcarboxy ICp 61; 3.1.7 Aromatic carboxyhc act s, p. ,"

. 'ds p 63' 319 pV and temperature, p. 64.act,. ," ~

BASES, p. 65~ and pK,., p. 65; 3.2.2 Aliphatic bases, p. 66;~:;:; ~~:ati~};ases,p. 69; 3.2.4 Heterocyclic bases, p. 72.

Aero/BASE CATALYSIS, p. 74: . 32 S ecific and3.3.1 Specific and general acid catalysts, p. 74; 3.. Pgeneral base catalysis, p. 75.

b ting a proton and is therebyHere water is acting as a base ~ accep 'd H OlD 'while the acid,converted into its. so-called con/ugate ~:ed in~o it~ conjugate base,H

2S0

4, by donatmg a proton IS conve

HS04 e.

. f . h mistry have been highlyModem electronic theones 0 organt~ c e . beh . r with

. wide variety of fields 10 correlatmg avlOu .successful 10 a. . for the relative strengths of orgamcst~udctured' ~~~~asl~o~~~~~~~ntte definition of Arrh.enius, ~cids areact s an '. . ns HlD in solutIon whIle basescompounds ~hat. Ylel~ hYdrog~~ ~~fi~ition~ are reasonably adequateyield hydrOXIde Ions, OH. Su be 'd ed but the acid/base. . . t only are to conSI er ,if re~cUOl~s 10 wa e~ so useful in practice that the concel?ts of bothrel~tlOnshlp has pro ed considerably more generahsed. Thus

~~:£~~~;:h;~~E~~~=;:'t~n~~:;'~~s~~eUJ',sf\~~::ii~:of sulphuric 'acid in aqueous solution IS then looked upon as.

H2S0

4+ H

20: i=!: H 30e + HS04

e

Acid Base Con- .Con-jugate JUgateacid base

3.3

3.2

3.1

3The strengths of acids and bases

Ph Hpr much I... "C/eOH C~ II

readily' III .NN . "

<>COMe

Syn

Ph H" /C

~/ .'MeCOO

Anti

stereoisomer out of the two alternatives possible, are said to be stereoselective.

Then again, many elimination reactions are found to occur much','more readily in that member of a pair of geometrical isomerides iD~

which the atoms or groups to be eliminated are trans to each other, ithan in the isomer in which they are cis (p. 255). As is seen in the'"relative ease of elimination from anti and syn aldoxime acetates toyield the same cyanide:

The degree of success with which a suggested mechanism can besaid to delineate the course of a particular reaction is not determinedsolely by its ability to account for the known facts; the acid test ishow successful it is at forecasting a change in rate, or even in thenature of the products formed, when the conditions under which thereaction is carried out, or the structure of the starting material, arechanged. Some of the suggested mechanisms we shall encountermeasure up to these criteria better than do others, but the overallsuccess ofa mechanistic approach to organic reactions is demonstratedby the way in which the application of a few relatively simple guidingprinciples can bring light and order to bear on a vast mass of disparateinformation about equilibria, reaction rates, and the relative reactivityof organic compounds. We shall now go on to consider some simpleexamples of this.

. r t" n with consequent stabilisa-There is extremely effectlv~de1~ I~i~o ~ it does two canonical

tion, in the metha~oate anion m~otho~gh delocalisation can takestructures of identIcal energy, an

!oe

/HC~~o

°~HC

"Q-H!

oe/

HC~O-HID

3 1 2 The origin of acidity in organic compounds .. • 'd' f organIc com-

Among the factors that may influence the aCI Ity 0 an

pound, HA, are: d(a) The strength of t~~ H-A bon.(b) The e1ectronegatlVlty of A. ed 'th HA(c) Factors stabilising Ae compar WI .

(d) The nature of the solven:. d t be a limiting factor, but theOf these (a) is not n~rmally oun 0 K of methanol, CH 30-H,effect of (b) is reflected m the fact~~t th~P isa ~ 43, oxygen being conis ~ 16 while that of metha~e, 3 rbon By contrast the pKa ofsiderably more electro~e~atlve th~. ~ . part due to the electronmethanoic (formic) aCId IS 3·77. h IS. IS ~he electron affinity of thewithdrawing carbonyl group ~n anc~g is attached, but much moreoxygen atom to which th~ ~nci'plent ~:fbf:in the resultant methanoateimportant is (c): th~ stablhsadtl.on ~ ted methanoic acid molecule:anion compared WIth the un ISSOCIa

°~HC

"oe

ds 553.1.2 The origin of acidity in organic compoun

. .. . ) thus cannot be measured in water at all.Their relative aCld~tles (pK.,sffi' t1y strong (low enough pKa), theyFurther, when a~lds are s~ ~Ien. water and will thus all appearwill all be essentially fully Ion;;~ ~NO HCI0

4, etc. This is known

to be of the same strength, e.g. , 3'

as the levelling effect of wat~r. K measurement can, however, beThe range of comparatIve P '~r a stronger and the latter case

extended by, in the first case, P[OV\ ~; carrying o'ut measurements i~a weaker, base than H~O as s.o ve: . 'city (and by using an acid that ISa range of solvents. of. mc;e~m:cid~~~ range in one solvent and nearnear the bottom hmIt 0 t e ommon reference in each case)

f h g in the next as a c d tthe top 0 t e ran e . t" of acid strengths on own 0"t is possible to carry determma Ion~cids as weak as methane (pKa ~ 43).

54 The strengths of acids and bases

3.1 ACIDS

The more generalised picture provided by Lewis, who definedacids as molecules or ions capable of coordinating with unsharedelectron pairs, and bases as molecules or ions which have such unsharedelectron pairs available for coordination, has already been referred to(p. 29). Lewis acids include such species as boron trifluoride (1) whichreacts with trimethylamine to form a solid salt (m.p. 128°):

"\ IDeMe3N: + BF3 ~ Me3N:BF3

(I)

Other common examples are aluminium chloride, tin(IV) chloride,zinc chloride, etc. We shall, at this point, be concerned essentially withproton acids, and the effect of structure on the strength of a number oforganic acids and bases will now be considered in turn. Compounds inwhich it is a C-H bond that is ionised will be considered subsequently(p. 270), however.

3.1.1 pK..

The strength of an acid, HA, in water, i.e. the extent to which it isdissociated, may be determined by considering the equilibrium:

H 20: + HA ~ H30ID + Ae

Then the equilibrium constant, in water, is given by:

_ [H30ID][A e]K. - [HA]

The [H20] term is incorporated into K .. because water is present insuch excess that its concentration does not change significantly. Itshould be emphasised that K .. , the acidity constant of the acid inwater, is only approximate (as above) if concentrations are usedinstead of the more correct activities; it is a reasonable assumption,however, provided the solution is fairly dilute. The acidity constantis influenced by the composition of the solvent in which the acid isdissolved (see below) and by other factors, but it does, nevertheless,serve as a useful guide to comparative acid strength. In order to avoidwriting negative powers of 10, K .. is generally converted into pKa(pKa = -loglO K a ); thus while K a for ethanoic (acetic) acid in waterat 25° is 1·79 x 10- 5, pKa = 4·76. The smaller the numerical valueof pKa , the stronger the acid to which it refers.

Very weak acids, those with pKa greater than ~ 16, will not bedetectable as acids at all in water, as the [H30Gl] they wiIl producetherein wiIl be less than that produced by the autolysis of wateritself:

3.1.3 The influence of the solvent

Despite the above discussion on the influence of internal structuralfeatures on a compound's acidity, the real determining role is oftenexerted by the solvent, and this is particularly the case when, ascommonly, the solvent is water.

Water has the initial disadvantage as an ionising solvent for organiccompounds that some of them are insufficiently soluble in theirunionised form to dissolve in it in the first place. That limitation apart,

573.1.4 Simple aliphatic acids

. I I a ctive ionising solvent on account (a) of itswater i.s a sl~gu ar Y e e{ _ 80) and (b) of its ion-solvating ability.high dIelectrIc constant ". - , b the higher the dielectric

fi t exerts Its effect ecauseThe rst prope~ Y f solvent the lower the electrostatic energy. ofconstant (polanty) 0 a ., 'Il b the more readily will such Ionan~ pairs of iotS preJe~~ m~~;:1 sta~;e will they be in solution, andpaIrs thus be o~e, e re to recombine with each other.the less ready ~Ill they b~, the[e~~ ~earby solvent molecules, thereby

lon~ in solutIOn, strong Ypo ~: solvent molecules around them: ,thecollectmg a solvatIOn e~vel~ take place the greater the stabihsa-greater the extent to whIch t IS can b'I' " 'tself by spreading or

. h' h' in effect sta Ilsmg Ition of, t~e I.on, w IC IS eculiar effectiveness of water, as an iO,n-delocahsmg It~ charg~. T~e p th £ ct that H 0 is extremely readIlysolvating medIUm, anses .rom e I~ 'n size' 2because of this it canpolarised, and also relau.v.ely sma I ions ~nd anions. The effect issolvate, and thereby s~ablhs,e, both c~~erful 'hydrogen-bonded' typepartic':llarly marked WIth a~~~~:~~rSfmilar H-bonded type solvationsolvatlo~ can occur (see 'th c tions but in the particular case ofcannot m ~e~~ral oc~ur w~EIl aan also solvate through hydrogenacids the mltIaI catIOn, ' cbond'ing with the solvent water molecules:

H HO I:I-OH$/ I: e H

o p H-O + H"'Y'"H-Y + nHz " : I

H Ho-Ii OH

Alcohols, just so long as they are not too bulky, e'f~ ~~~HtoS~r~something of water's abilities and, for example

h, HCI be forgotten

.' h I Iso It should not, owever,strong aCId m met. ano a . .s that it should be capablethat th~ pr!me reqbUlre~ehnt ofa~:r ~~~v~~~el the smaller the dissociationoffuncuonmgasa ase.t ewe, I )of the acid. Thus we find that in, for example, m~thyl.benzene (to ueneHCI occurs as such, i.e. it is almost wholly undlssoclated.

3.1.4 Simple aliphatic acids .The replacement of the non-hydroxylic edhYdrOgendautceomao~~~~~a~~~

Ik I ' ht be expect to pro 'acid by an a Y group ml.g d t' ffect of the alkyl group wouldas the electron-donatmg m uc I~e e en atom carrying thereduce the residual electr~ a~~ty ~~e~;t~:;lthe acid. In the alkyl-

~~~~tii~~~~r~~~~~~~~ ~~c;ea;: ~Ie~t~?nh availability ~~~=:~w~~~~~serve to promote its recomb.matl.on WIt p~oton, as c

methanoate anion/methanOIc aCI]desYs~tem . o]e/0 /"

r'+c;"o "-c"·o

(2e1)(2c)(2b)(2a)

The strengths of acids and bases56

place in the methanoic acid molecule also, this involves separation ofcharge and will consequently be much less effective as a stabilisinginfluence (cf p. 20). The effect of this differential stabilisation is somewhat to discourage the recombination of proton with the methanoateanion, the equilibrium is to this extent displaced to the right, andmethanoic acid is, by organic standards, a moderately strong acid.

With alcohols there is no such factor stabilising the alkoxide anion'ROe, relative to the alcohol itself, and alcohols are thus very muchless acidic than carboxylic acids. With phenols, however, there isagain the possibility of relative stabilisation of the anion (2), bydelocalisation of its negative charge through interaction with the 1t '

orbitals of the aromatic nucleus:

Delocalisation also occurs in the undissociated phenol molecule (cfp. 23) but, involving charge separation, this is less effective than in theanion (2), thus leading to some reluctance on the part of the latter torecombine with a proton. Phenols are indeed found to be strongeracids than alcohols (the pKa of phenol itself is 9·95) but considerablyweaker than carboxylic acids. This is due to the fact that delocalisationof the negative charge in the carboxylate anion involves structures ofidentical energy content (see above), and of the centres involved twoare highly electronegative oxygen atoms; whereas in the phenoxideanion (2) the structures involving negative charge on the nuclearcarbon atoms are likely to be of higher energy content than the one inwhich it is on oxygen and, in addition, of the centres involved hereonly one is a highly electronegative oxygen atom. The relativestabilisation of the anion, with respect to the undissociatedmolecule, is thus likely to be less effective with a phenol than with acarboxylic acid, leading to the lower relative acidity of the former.

59

o

""CH-+C02HoJ" 1·25

3.1.5 Substituted aliphatic acids

F-+CH2-+-c°2H2·57

O-+CH2~02H

2"86

3.1.5 substituted aliphatic acids

The eff~ of~tr~ucinge1ectr~:tw~~a~~~~;~:,s~~n~~n~;~:f~~aliphatic aCids IS more ~rd. .. to alkyl might be expected to

- . th oppoSite uecUon .. . - . d ed~ffect actmg 10 e id so substituted, and thiS IS 10 emcrease the strength of an acobserved as pKa values show:

Me(CH ) CO H Me(CH2hC02HH2 2 4-86

" bonded carbon atom adjacent to the ~rbo~ylIf there IS a doubly .. sed Thus propenoic (acrylIc) aCid,group the acid strength IS IOcrea . ed"th 4 88 ~or the saturated

H h K of 4·25 compar WI .CH 2 =CHC02 ,.asa~ This is due to the fact that the unsaturatedanalogue, propa~OlC~CI~ b -d·sed which means that electrons areex-carbon atom IS Sp Y n I, "a saturated Sp3 hybridiseddrawn closer to the carbon nucleus t~~!1b:ion in the ;p2 hybrid. Theatom ~ue to th~ rat~~~!::le:a:~nat~ms are less e1ectron-donatil!gresult IS that sp hy "dOsed and so propenoic acid though stIlih ted Sp3 hybn I ones, . fti"

t at satura . cid is stronger than propanoIc. The e ect ISweaker than methanOl~~ h 1 h bridised carbon atom of a triplemuch more marked WI t e SP. Y . 1-) ·d HC=CCO H is

K f pynOic (proplo IC aCi , - 2'bond, thus the p a_ 0 ~ro with the hydrogen atoms of ethene1-84_ An analogous sltua;lOn OCCU~ittle more acidic than the hydroge!1sand ethyne; those ofthhe orm

fet~aynree are sufficiently acidic to be readIly

in ethane whereas t ose 0 ereplaceable by a number of metals (cf. p.272).

and other influences playing a part; pKa values are observed as

foIlows:


We should thus expect the equilibrium to be shifted to the leftcompared with that for methanoic acid/methanoate anion, and it is infact found that the pKa of ethanoic acid is 4-76, compared with 3·77for methanoic acid" However, the degree of structural changeeffected in so small a molecule as methanoic acid by replacement of \H by CH3 makes it doubtful whether so simple an argument is reallyvalid; it could well be that the relative solvation possibilities in thetwo cases are markedly affected by the considerably different shapesof, as weIl as by the relative charge distribution in, the two smallmolecules.

It is important to remember that the value of the acidity constant,K a , of an acid is related to the standard free energy change for theionisation, L\G"\ by the relation

_~Ge = 2"J03RTlogK.

and that L\GB includes both enthalpy and entropy terms:

Thus it is found for the ionisation of ethanoic acid in water at 25°(Ka = )·79 x 10- 5) that L\G B = 27·2kJ (6·5kcal), L\H B = -O·5kJ(-0'13 kcal), and ~SB = -92 J (-22 cal) deg-1 [i.e. T~Se =-27,6 kJ (-6·6 kcaI)J; while for methanoic acid (Ka = 17·6x 10-5

)

the corresponding figures are: ~Ge=21kJ (5'lkcal), ~He=

-O·3kJ (-O'07kcal), and ~Se=-74J (-18ca1)deg- 1 [i.e.T~Se=-21'3kJ (-5·17kcal)J. The surprisingly small ~He valuesalmost certainly arise from the fact that the energy required fordissociation of the O-H bond in the undissociated carboxylic acidsis cancelled out by that evolved in solvating the resultant ions.

The differing L\GBS, and hence the differing K,.s. for the two acidsthus result from the different values of the two entropy (L\SB) terms.There are two species on each side of the equilibrium and differencesin translational entropy on dissociation will thus be small. However,the two species are neutral molecules on one side of the equilibriumand ions on the other. The main feature that contributes to L\SB isthus the solvation sheaths of water molecules that surround RC02 eand H 3 0ED, and the consequent restriction, in terms of increasedorderliness, that is thereby imposed on the solvent water molecules;the increase in orderliness not being quite so great as might have beenexpected as there is already a good deal of orderliness in liquid wateritself. The difference in strength between methanoic and ethanoic acidsthus does indeed relate to the differential solvation of their anions,as was suggested above.

Further substitution of alkyl groups in ethanoic acid has much lesseffect than this first introduction and, being now essentially a secondorder effect, the influence on acid strength is not always regular, steric

613.1. 6 Phenols

pKa

C6H50H 9·95o-02NC6H40H 7·23m-02NC6H40H 8·35p-O NC6H40H 7·14

2,4-(02~hC6HJOH 4·012,4,6-(02NhC6H20H 1·02

Here again ~H~ is found. to .vary °rwv:~;ess~~~:}~re%eO;hr~~and p- nitroph~nols, the ?Iffenng ~G~ t ms i e from variations inarising from dIfferences m thhe T~S. e~ue't~ the differing distributhe solvation patterns of the t ree amons,tion of negative charge in .them

l-d ting alkyl groups into the

The effect of introducmg e ectron onabenzene nucleus is found to be smaIl :

HO+-CH 2-+-c°2H3·83

3.1.6 Phenolsbe observed with substituted phenols, the

Analogous effects can. . s in the nucleus increasingpre~en~ ?f electron-wl:;d;t:l~ftr~r~~bstituent,the inductive effectthelf aCIdIty. In the ca ff.th d'stance on going 0- --+ m- --+ pwould be expected to fall ~d ~I bel an electron-withdrawing mesonitrophenol, but there ~ou a so. . the 0- or p_ but not in the mmeric effect when the mtro group IS m . . tio~ by stabilisation

. . d thO too would promote IOmsaposItIOn; an . IS . sultant anion We might therefore(though delocahs~tlOn) of the re acidic than the m-compoundex~ect ?- ~nd p-mtrophenol::~~ :~:.e Introduction of furthe~ N<?2whIch IS, m fact, f~dU?td to kedly thus 24 6-trinitrophenol (picrIC aCId)groups promotes aCI I y mar .' , ,is found to be a very strong aCId:

6-0eo/~-:::::-o eo/~ ......oe

s The unshared electrons on the oxygenhydroxyl and methoxyl group. t able to exert a mesomeric effect,atoms of the .last ~wo ~oups ~r~ n.od ctive effect, owing to the interin the opposIte dIrectIOn tOt t elf~l ~his is seen in the pKa values:vening saturated carbon a oms.

02N+-CHio...-C02H EtOp.".-cH io"--C02H

1.68 3·35

MeJN+-cHi....-c02H MeC(}+-CHf,,-c°2H

1.83 3·58


[o]e/"

F-+-CH 2+-<::<°

The relative effect of the different halogens is in the expected order,fluorine being the most electronegative (electron-withdrawing) andproducing a hundredfold increase in strength of fluoroethanoic acid ascompared with ethanoic acid itself. The effect is very much greaterthan that produced, in the opposite direction, by the introduction of .an alkyl group, and the introduction of further halogens stiII produceslarge increases in acid strength: trichloroethanoic is thus a very strongacid.

Here again it is important to remember that K a (and hence pKa) isrelated to ~G~ for the ionisation, and that ~G~ includes both ~H~and ~S~ terms. In this series of halogen-substituted ethanoic acids~H~ is found to differ little from one compound to another, theobserved change in ~G~ along the series being due largely to variationin ~S~. This arises from the substituent halogen atom effecting delocalisation of the negative charge over the whole of the anion,

11 11

CH2CH2CH2C02H MeCHCH 2C02H4·52 4·06

Other electron-withdrawing groups, e.g. R3NE9, CN, N02, S02R,CO, C02R increase the strength of simple aliphatic acids, as also do

MeCH 2CH2C02H4·82

the latter thus imposes correspondingly less powerful restriction onthe water molecules surrounding it than does the unsubstitutedethanoate anion whose charge is largely concentrated, being confinedsubstantiaIly to CO2e. There is therefore a smaIler decrease in entropyon ionisation of the halogen-substituted ethanoic acids than withethanoic acid itself. This is particularly pronounced with CF3C02H(pKa 0·23) for whose ionisation ila B = 1·3 kJ (0,3 kcal) comparedwith 27·2 kJ (6'5 kcal) for CH3C02H, while the ilHB values forthese two acids differ very little from each other.

The introduction of a halogen atom further away from the carboxylgroup than the adjacent IX-position has much less influence. Its inductive effect quickly dies away down a saturated chain, with the resultthat the negative charge becomes progressively less spread, i.e. moreconcentrated, in the carboxylate anion. The acid thus increasinglyresembles the corresponding simple aliphatic acid itself, as the foIlowingpKa values show:

3.1.8 Dicarboxylic acidsAs the carboxyl group itself has an electron-withdrawing inductiveeffect, the presence of a second such group in an acid might be expectedto make it stronger, as shown by the following pKa values:

HCO,H HO,CCO,H3.77 t·23

CH]CO,H HOFCH,CO, H4.76 2·83

CH]CH,CO,H HO,CCH,CH,CO,H4-88 4-19

C6H

sCO, H HO, CC6H4CO, H

4-t7 0-2·98/11-3·46p-3·51

3.1. 8 Dicarboxylic acids 63

©(o'-

o ~C

V

6l H

(3) (4)

Intramolecular hydrogen bonding can, of course, operate in theundissociated acid as well as in the anion, but it is likely to be considerably more effective in the latter than in the former-with consequentrelative stabilisation-because the negative charge on oxygen in theanion will lead to stronger hydrogen bonding. The effect is evenmore pronounced where hydrogen bonding can occur with hydroxylgroups in both o_positions, and 2,6_dihydroxybenzoic acid is found

to have pKa = 1·30.

Itwill benoticed that thiscompensatingeffect becomes more pronouncedin going CI ~ Br -.-.OH, i.e. in increasing order of readiness withwhich the atom attached to the nucleus will part with its electron

pairs.It is important to emphasise, however, that here-as in the casesabove-it is probably the effect of differing charge distributions inthe anions on their patterns of solvation, i.e. on the TAS

Bterm relating

to the degree of ordering induced locally in the assembly of solventmolecules, that is responsible for the observed differences in pK

a•

The behaviour of o-substituted acids is, as seen above, oftenanomalous. Their strength is sometimes found to be considerablygreater than expected due to direct interaction between the adjacentgroups. Thus intramolecular hydrogen bonding (cf p. 36) stabilisesthe anion (4) from o_hydroxybenzoic(salicyclic) acid (3) by delocalisingits charge, an advantage not shared by its m- and p-isomers, nor by

o_methoxybenzoic acid:

OH2,98

4-084·58

pK. or XC6 H4 CO, H

a Dr OMe2·94 2·85 4-093·83 J.81 4-093·99 4-00 447

pK.4·202-173453432·83

H0- 4-20

/11- 4-20p- 4-20

62

o"'C,;.:.o e

oeo.....~'-oe

The particularly marked effect .thshort distance over which th WI ~-N.02 m~y be due to the verybut some direct interaction e power ul mdu~lVe effect is operatingroups cannot be ruled out between the adjacent N02 and CO fiTh' 2

e presence of groups such as OHelec!ron-withdrawing inductiv fIi b' OMe, or halogen having anmenc t:

ffectwhen in the 0- an~e _ecl,. ~t an electron-donating meso-

p-substttuted acids to be weaker fh~sltlOns. may, howeve~, cause theeven than the unsubstituted acid it Ifthe m- and, on occaSIOn, weakerse ,e.g. p-hydroxybenzoic acid'

0", :,;;-0 0..0 e .

¢~6x: x·

The strengths of acids and bases

The resulting substituted ph Isth "'ect' eno are very slightl ke ell' IS marginal and irr I . ., y wea er acids. but~ubstjtu~nts in destabilising e~~ear'hmdl~ttng. that the effect of suchmteractlOn of its negativ h p ~noxlde Ion, by disturbing ththe aromatic nucleus is s~ ~I arge ~Ith the delocalised 11: orbitals ;

, a , as might have been expected.

3.1.7 Aromatic carboxylic acids

Benzoic acid, with a pKa of 4·20 is .analogue cyclohexane carbo I" ~ stronger acid than its saturateda ph I . xy IC aCid (pK - 4·87)' .~ny group, hke a double b a -:- ' suggestmg that~onatmg--eomparedwith a saturated con~, IS here less electron-

oxyl group, due to the Sp2 h b 'd' ar on atom~towardsthe car~arboxyl group is attached (cf / 5~) ~~ .carbon a.tom to which themt.o the benzene nucleus has v r' r' e mtroductlOn of alkyl groupsaCid (c/. similar introduction~J pl~t1e eflil eet on the strength of benzoiceno s, p. 61),

3.1.9 pK.. and temperature

W~ h.ave already seen (p. 56) that the K and haCid IS !!£! an intrinsic attribute of the;' . ~nce pKa , valu.e for anfrom one so!vent to another: the valupe~es Itse.lf, because It variessystem of which the acid is a . e ependmg on the overallfor aqueous solution unless :~stIt~ent. Values are normally quotedare available for that'solvent. M ertl~ specified, because most dataagain because most data were obt~~ne:t ues ~re also quoted as at 25°,te.mperature has to be specified as K at thIS .t~mperature. A constantWith temperature. We hav be a' an eqUilibrIum constant, varies

e en concerned above with the relative

653.2 Bases

acidity of various categories of acids, and in trying to correlate relativeacidity sequences with structure in a rational way-with some degreeof success. It is, however, pertinent to point out that not only doindividual Ka values vary with temperature, they also vary relative toeach other: thus ethanoic is a weaker acid than Et2CHC02H below30°, but a stronger acid above that temperature. Such reversals ofrelative acidity with change of temperature are found to be fairlycommon; it thus behoves us not to split too many fine hairs aboutcorrelating relative acidity with structure at 25°!

[B:][H3 0Ill]K a [BHIIl]

3.2 BASES

3.2.1 pI(", pKBH'" and pK..

The strength of a base, B:, in water, may be determined by consideringthe equilibrium:

B: + HOH P BHIIl + eOH

The equilibrium constant in water, K b , is then given by:

K_ [BHIIl][eOH]

b - [B:]

The [H20] term is incorporated into K b , because water is present insuch excess that its concentration does not change significantly; hereagain, concentrations can commonly be used instead of the morecorrect activities provided the solution is reasonably dilute.

It is, however, now more usual to describe the strength of basesalso in terms of Ka and pKa , thereby establishing a single continuous scale for both acids and bases. To make this possible we use, asour reference reaction for bases, the equilibrium

where Ka (and pKa ) is a measure of the acid strength of theconjugate acid, BHEB

, of the base, B:. This measure of the readinesswith which BHEB will part with a proton is, conversely, a measure ofthe lack of readiness with which the base, B:, will accept one: thestronger BHEB is as an acid, the weaker B: w21 be as a base. Thus thesmaller the numerical value of pK" for BH , the weaker B: is as abase. When using pK to quote the strength of a base, B:, pKBH•

should actually be specified but it has become common-though incorrect-to write it simply as pKa •

BHIIl + H 2 0: ~ B: + HlOIll

for which we can then write,

e

(7)

°H II'c......c ......

II °C \H/ \ .H

C-O'II

°


The effect is very pronounced, but f: IIcarboxyl groups are separated bas off sharply as SOon as theat<:>m. Cis-butenedioic(maleic) acl fO~ ~h~n one. saturated carbon~cld than trans-butenedioic(fumari~)'1'cil - 1.9~) IS a much strongerIntramolecular hydrogen bonding that (6, pKa = 3'?2), due to thebut not with the latter lead' can. take place With the former(maleate, 7) mono-anion'(cf o_~ngd to rbelatI v~ sta.bilisation of the ci;

. y roxy enZOIC aCid above):

°H II'c......c......

II ° -H·C \--

H/ \ .HC=O'

6l..H(5)

(6)

The second dissociation of t b . .occurs more readily than that O/~:s~i~_~~edlOlc :cid (pKa

2 = 4·38)because of the greater difficulty . . d (pKa = 6·23), howeverc~a.rged cyclic system in the a~~~eD)vIn~ a proton from the negativel;dl<:>Ic(oxalic), propane-l,3-dioiC(maloni~;mvr from the la~t~r. EthaneaCids ~re each weaker in their second d~n b~t~ne-l,4-dlOlc(succinic)ethanolc and propanoic acid . ISsoclatlOns than methanoicproton has to be removed f s, respectl~ely. This is because the secondan electron-donating SUbst:~~n~n.ega~~l~char~ed s~ecies containingto destabilise the anion with ,I.e. 2, which .mlght be expectedcompared with the unsubstitute~t~~~=~~ the undlssociated acid, as

3.2.2 Aliphatic bases

Taking as an example N~, with a pKa value of 9·25,

673.2.2 Aliphatic bases

. NH --+RNH --+R NH--+R 3N,Thus on going along the senes, 3 . ~ 2 iv Iy

the inductive .effect will te~d to increarS:ti~:~~I~I~~~~;~~~~gr;.~~te~dless stabilisation of .t~e ca;~nnb~ ~~~ct of introducing'successive alkylto decrease tho;: baSICity. e ~ I mailer and an actual changeovergroupS thus beco~esfprogressive ~:ry to a 'tertiary amine. If this is thetakes place on gOlOg rom a secon should be observed if measurereal explanat~o~, no suchdch~~~~~~~ent in which hydrogen-bondingments of baSICity ~re rna. e

ded been found that in chlorobenzene the

cannot take place; It has, 10 e? ,order of basicity of the butylamlOeS IS

BuNH 2 < Bu2NH < BU3N

. K alues in water are 10-61,11-28 and 9·87.though their related'p a V R NEllIe, are known, on treatment

Tetraalkylammon~umsalts, e.g. 4. Id basic solutions comparablewith moist silver oXld~, AgOH, I~o r~· . readily understandable forin strength with t.he mlOera~~okaHIs.. b~~~d to be completely ionisedh b so obtalOed, R4N , IS .tease. 'bTt as with tertiary ammes, etc.,as there IS no poSSI I I y,

R3~1H + eOH - R3N: + H 2 0

ci reverting to an union~sed form. . hdrawin groups, e.g. CI, N02 ,

The effect of introduClOg electron-w~t b ·citygdue to their electron-close to a.bas.ic cent.re iSftito d

te(cf.r:~~s~it:t~S~nilines below, p. 70); thus

withdrawmg mductlve e eC c.the amine

Decreasing stabilisation by solvation

F3C

"F 3C-+-N:/

F3C

. d th three powerfully eIectron-is found to be virtually non-baSIC, ue to ewithdrawing CF J gfClupS.

b h xtent to which the cation, form~ bynitrogen atom, but also y~ e e solvation, and so become stabilIsed.uptake of a proton, can un trg~ed to nitrogen in the cation, the greaterThe more hydrogen atomsfa

laCI tion via hydrogen bonding between

the possibilities of power u so vathese and water:

Me Me~ ~

Me+NH2 NH Me+N;r ;r

Me Me10·64 10·77 9·80

NH39·25

Et Et~ ~

Et+NH2 NH Et+N;r ;r

Et Et10·67 10·93 10'88


it is found that ~Ge- = 52·7 kJ (12·6 kcal), ~He- = 51·9 kJ (12·4 kcal),and LlSB = -2·9 J (-0·7 cal) deg-1 [i.e. TLlSB = -0·8 kJ (-0·2 kcal))at 25°. Thus the position of the above equilibrium is effectivelydetermined by LlHB

, the effect of LlSB being all but negligible: aresult that is in marked contrast to the behaviour of many acids aswe have seen above (p. 58). The reason for the small effect of LlSB

is that here there is one charged species (a positive ion) on each sideof the equilibrium, and these ions have closely comparable effects inrestricting the solvent water molecules that surround them, so thattheir entropies of solvation tend to cancel each other out.

As increasing strength in nitrogenous bases is related to the readinesswith which they are prepared to take up protons and, therefore, to theavailability of the unshared electron pair on nitrogen, we mightexpect to see an increase in basic strength on going: NH 3 --+ RNH 2 --+R2 NH --+ R3N, due to the increasing inductive effect of successivealkyl groups making the nitrogen atom more negative. An actual seriesof amines was found to have related pKa values as follows, however:

It will be seen that the introduction of an alkyl group into ammoniaincreases the basic strength markedly as expected. The introductionof a second alkyl group further increases the basic strength, but the neteffect of introducing the second alkyl group is very much less markedthan with the first. The introduction of a third alkyl group to yield atertiary amine, however, actually decreases the basic strength in boththe series quoted. This is due to the fact that the basic strength of anamine in water is determined not only by electron-availability on the

II

69

(Btl)

3.2.3 Aromatic bases

R RI , NeutralC +-+

6 C~$ moleculeHN-7 'NH 2 HN ...... "NH 2

(Ila) (lIb) ilHe

R RI I

Cation$hC ..-......C~$

H N 7 'NH H N "NH22 •• 2 2 ••

(12a) (l2b)

$ $

o~6~66

(13a) (13b) (Be)

(14)

A somewhat analogous situation occurs with the amidines,RC(=NH)NH2 (11):


While stabilisation by delocalisation in the cation (12) would not beexpected to be as effective as that in the guanidine cation (10) above,ethanamidine, CH3C( NH)NH2(pKa = 12·4), is found to be a muchstronger base than ethylamine, MeCH2NH2 (pKa = 10·67).

The exact reverse of the above is seen with aniline (13), which is a veryweak base (pKa = 4·62) compared with ammonia (pKa =~·25) ?rcyclohexylamine (pKa = 10·68). In aniline the nitroge.n a.tom IS agambonded to an Sp2 hybridised carbon atom but, more slgmficantly, theunshared electron pair on nitrogen can interact with the delocalised7T orbitals of the nucleus:

If aniline is protonated, any such interaction, with resultant stabilisation, in the anilinium cation (14) is prohibited, as the electron pair onN is no longer available:


:NH2$NH 1 :NH2I II I

NeutralC - 6 C - 6 C~$HN-7 'NH2 HN...... 'NH HN ...... "NH

2 molecule'. 2

(9a) (9b) (9c) HH e

:NH2$NH2 :NH2I - II I$hC C - C~$ CationH 2N 7 'NH2 H N ...... 'NH H 2N ...... "NH 22 •• 2

(lOa) (lOb) (We)

68

. The change is .also pronounced with C=O, for not only is theOItro~en atom, wIth its electron pair, bonded to an electron-withdrawmg group throu~ an Sp2 hybridised carbon atom (cf p. 59), butan electron-wIthdrawmg mesomeric effect can also operate:

L 0 06

]II .. I $R-C+-NH 1 ..... R-C=NH

2

Thus amides. are found. to b.e only very weakly basic in water [pKafor ethanamlde(aceta~l~e) IS =0·5], and if two C=O groups are

p.resent th~ ~esultant Imldes, far from being basic, are often suffic~ently aCIdIc to form alkali metal salts, e.g. benzene-12-dlcarboximide (phthalimide, 8): '

(8)

The. effect. of del~c~lisation in increasing the basic strength of anamme IS seen m guamdl~e, HN=C(~H2h (9), which, with the exceptionof the .tetr~alkylammomum hydroxIdes above, is among the strongestorgamc mtrogenous bases known, with a related pK

aof =13.6.

Both .the neut~al molecul~, and the cation, H2~=C(N~)2 (10),resultmg from Its protonatlon, are stabilised by delocalisation;

but i? t~e cation the P?sitive charge is spread symmetrically bycontnbutlon to the hybnd of three exactly equivalent structures ofequal energy. No comparably effective delocalisation Occurs in theneutral. molecule (in w~ich two of the contributing structures involveseparatIOn of ch~rge), WIth the result that the cation is greatly stabilisedWIth resp~~ to It, thus making protonation 'energetically profitable'and guamdme an extremely strong base.

The extra base-weakening effect, when the substituent is in theo-position, is due in part to the short distance over which its inductiveeffect is operating, and also to direct interaction, both steric and byhydrogen bonding, with the NHz group (cf the case of o-substitutedbenzoic acids, p. 63). o-Nitroaniline is such a weak base that its salts

The aniline molecule is thus stabilised with respect to the aniliniumcation, and it is therefore 'energetically unprofitable' for aniline totake up a proton; it thus functions as a base with the utmost reluctance(pKa = 4'62,comparedwithcyclohexylamine,pKa = 10,68). The baseweakening effect is naturally more pronounced when further phenylgroups are introduced on the nitrogen atom; thus diphenylamine,PhzNH, is an extremely weak base (pKa = 0'8), while triphenylamine,Ph3 N, is by ordinary standards not basic at all.

Introduction of alkyl, e.g. Me, groups on to the nitrogen atom ofaniline results in small increases in pKa :


MeOC6 H4 NH 2

0- 4·49m- 4·20p- 5·29

HOC6 H4 NH 2

0- 4·72m- 4·17p- 5·30

PhNH 2

4·62

......N~eO ID 0

(15)

An interesting case is provided by 2,4,6-trinitro-l'!,N-dimethylaniline(15) and 2,4,6-trinitroaniline (16), where the former IS found to be about40 000 times (ApKa 4·60) stronger a base than the lat~er (?y co~tra'st N,N-dimethylaniline and aniline itself differ very ~Ittle m .basIcstrength). This is due to the fact that the NMez group IS su~cIentlylarge to interfere sterically with the very large. NOz groups m botho-positions. Rotation about ring-carbon to mtrogen bonds allowsthe 0 atoms of NOz and the Me groups of NMez to move out ofeach other's way, but the p orbitals on t~e N atoms are now nolonger parallel to the p orbitals of the nng-carbon atoms..As a

consequence mesomeric shift of the unshared electron p~Ir on, . h bItals ofNMe to the oxygen atoms of the NOz groups, vIa t e p or

the ring-carbon atoms (cf. p. 70), is inhi?ited, and the expectedbase-weakening-by mesomeric electron-wIthdrawal-does not t~ke

I ( f P 27) The base-weakening influence of the three mtropace c.. . " . ffgroups in (15) is thus confined essentially to theIr mductIve e ects:

~~N~N,

ecrID ~ ID oe

are largely hydrolysed in aqueous solutio~,.while. ~,4-dinitroanilineis insoluble in aqueous acids, and 2,4,6,-tnn~troamhne resembles a~

'de' it is indeed called picramide and readIly undergoes hydrolysIsamI ,to picric acid (2,4,6-trinitrophenol).

With substituents such as OH and OMe that h~ve unshared ~Iectronpairs an electron-donating, i.e. base-strengthenmg, meso~e~Ic eff~ct

be exerted from the 0- and p-, but not from the m-posItIon, wIth~~~ result that the p-substituted aniline is a st~onger base than thecorresponding m-compound. The m-compound IS a weaker base thananiline itself, due to the electron-withdrawing inductive effect exertedb the oxygen atom in each case. As so often, the. effec~ of the. 0

s~bstituent remains somewhat anomalous, due to dIrect mter~ctIOnwith the NHz group by both steric and polar effects. The substItutedanilines are found to have related pKa values as follows:

[g, ~ QJ

N02C6H4 NH 20- -0,28m- 2·45p- 0'98


PhNH2

4·62

70

C6HsNH2 C6HsNHMe C6HsNMe2 MeC6H4 NH24·62 4·84 5·15 0- 4·38

m- 4·67p- 5·10

Unlike on such introduction in aliphatic amines (p. 66), this smallincrease is progressive: suggesting that cation stabilisation throughhydrogen-bonded solvation, responsible for the irregular behaviour ofaliphatic amines, here has less influence on the overall effect. Themajor determinant of basic strength in alkyl-substituted anilines remains mesomeric stabilisation of the aniline molecule (13) withrespect to the cation (14); borne out by the irregular effect of introducing Me groups into the 0-, m- and p-positions in aniline. Similarirregular effects on the pKa of phenol were observed when Megroups were introduced into its 0-, m- and p-positions (p. 61).

A group with a more powerful (electron-withdrawing) inductiveeffect, e.g. NOz, is found to have rather more influence. Electronwithdrawal is intensified when the nitro group is in the 0- orp-position, for the interaction of the unshared pair of the aminonitrogen with the delocalised 7T orbital system of the benzenenucleus is then enhanced. The neutral molecule is thus stabilisedeven further with respect to the cation, resulting in further weakening as a base. Thus the nitro-anilines are found to have related pKavalues:

(17)

733.2.4 Heterocyclic bases

(20)

fO)Ne

n"H~-"\~ H

(19)

f[5)NH

(180)

f[5)NI

H

(180)

For such aromaticity to be achieved, six 1t electrons (4n + 2, n = 1)from the ring atoms must fill the three bonding molecular orbitals(cf p. 17). This necessitates the contribution of two el~ctrons by thenitrogen atom and, though the resultant electron clou? WIll be deformedtowards nitrogen because of the more electroneg~tlve nature o! th~t

atom as compared with the four carbons, nitrogen s electron paIr wdlnot be readily available for taking up a proton (18a):

oNH

(18)

again in an Sp3 orbital and its related pKa (10,95) is found to bevery little different from that of triethylamine (10·75).

Pyrrole (18) is found to exhibit some aromatic character (thoughthis is not so pronounced as with benzene or pyridine), and does notbehave like a conjugated diene as might otherwise have been expected:

Protonation, if forced upon pyrrole, is found to take place not onnitrogen but on the a:-carbon atom (19). This occurs because incorpor~

tion of the nitrogen atom's lone pair of electrons into the aromatic61te system leaves the N atom positively polarised; protons tend to berepelled by it, and are thus taken up by the adjacent a:-carbon ato.m.The basicity situation rather resembles that already encountered WIthaniline (p. 70) in that the cation (19) is destabilised with respect to theneutral molecule (18a). The effect is much more pronounced withpyrrole, however, for to function as a base it has to lose all aromaticcharacter and consequent stabilisation: this is reflected in its relatedpKa (-0.'27) compared with aniline's of 4,62, i.e. pyrrole is a veryweak base indeed. It can in fact function as an acid, albeit a veryweak one, in that the H atom of the NH group may be removed bystrong bases, e.g. 6NH2 ; the resultant anion (20) then retains thearomatic character of pyrro]e, unlike the cation (19):

(l6b)

1>·V··R..~"-H···rN N

.......... o~e I I e':::"o

N60..-e ......0 6

(160)


In 2,4,6-trinitr~an.ilin~ (16), h~wever, the NH2 group is sufficientlysmall for no such lImitatIOn to be Imposed; hydrogen-bonding betweenthe oxygen atoms of the o-N02 groups and the hydrogen atoms of theNH2 gro~p m~y indeed help to hold these groups in the required,~Ianar, OrIentatIOn. The p orbitals may thus assume a parallel orientation, and the strength of (I6) as a base is enormously reduced by thepowerful electron-withdrawing mesomeric effect of the three N0

2groups:

3.2.4 Heterocyclic bases

~yri~ine, ~ ~ is an aromatic compound (cf p. 18), the N atomIS sp hybrIdlsed~ and contributes one electron to the 61te (4n + 2,n = 1) system; thIS leaves a lone pair of electrons available on nitrogen(accommodated in an Sp2 hybrid orbital), and pyridine is thus foundto be basic (pKa = 5,21). It is, however, a very much weaker basethan ~]iphatic tertiary amines (e.g. Et3N, pKa = 10,75), and this weak~ess IS ~ound to be cha~a~teristic of bases in which the nitrogen atomIS multIply bonded. ThIS IS due to the fact that as the nitrogen atom?ecomes progressi.vely more multiply bonded, its lone pair of electronsIS accommodated m an orbital that has progressively more s character.The electro~ pair is ~hus drawn closer to the nitrogen nucleus, andheld mo~e tIghtly by It, ~hereby becoming less available for forming abond wIth a proton, wIth consequent decline in the basicity of the

compound(cj.p.59).Ongoing )N: -+ )N: -+ =N: in,forexample,

R3N: ~ CsHsN: ~ RC=N:, the unshared pairs are in Sp3, Sp2 andSpl orbItals, respectIvely, and the declining basicity is reflected in thetwo pKa ~alue.s quoted above, and in the fact that the basicity ofalkyl. cyam~es I~ ~ery small indeed (MeCN, pK

a= -4'3).

WIth qumuchdme (17), however, the unshared electron pair is

which is found to have a related pKa of 11,27, closely resemblingthat of diethylamine (11'04).

75

+ EtOH

3.3.2 Specific and general base catalysis

COEt OEt 0/' .,(.- H 20 -I"

MeC-OEt 4=.==~. MeC ED - MeC,,~ n. sloW "'.. fasl "9 H-A (Ii~i~~~g) OEI OEt

Et +EtOH + Ae

Rate = k[eOH] [Me2C(OH)CH 2COMe]

and general acid catalysis is characteristic of reactions in whichprotonation of the substrate is slow, i.e. rate-limiting, and is followedby rapid conversion of the intermediate into products.

3.3.2 Specific and general base catalysis

Exactly the same distinction can be made over catalysis by bases aswas made above for acids. Thus in specific base catalysis the reactionrate is again found to be oc pH, this time rising as the pH rises, i.e. isoc [eOH]. Thus in the reversal of an aldol condensation (cf p. 224) itis found that,

HOYHU;> 0 90H 9~ ~j ?eMe} ~ +=! Me2C C ~ Me2C=O + /C"

" /" r.st "....--.y" (r.t.') "CH2 Me CH2 Me IImltmg CH 2 Me

and the reaction is believed to proceed:

This is known as general acid catalysis, general because the catalysisis by proton donors in general, and not by H 30 Ell alone. Generalacid catalysis often only becomes important at higher pHs, e.g. ~ pH 7when [H

30 Ell ] ~ 10- 7, while [HA] may be 1-2, molar; general acid

catalysis will still occur at lower pHs, but may then be masked by thegreater contribution by H3 0 Ell . The above orthoester hydrolysis isbelieved to proceed (only HA is shown here but H30 Ell will do thesame thing),

which there is rapid, reversible protonation of the substrate before theslow, rate-limiting step.

Reactions are also known which are catalysed not only by H3 0 Ell ,but by other acids in the system as well; e.g. in the hydrolysis of orthoesters such as MeC(OEth in the presence of an acid, HA, where it isfound that:

lD •MeCH=OEt ""Ta;t MeCH=O

+EtOH +EtOH


oNH

(21)

No such considerations can, of course, apply to the fully reducedpyrrole, pyrrolidine (21),

Rate = k[H]OlD][MeCH(OEth]

This is known as specific acid catalysis, specific in that H 3 0Ell is theonly acidic species that catalyses the reaction: the reaction rate isfound to be unaffected by the addition of other potential proton donors~acids) such as NH4 $, provided that [H 3 0 Ell ], i.e. pH, is not changed,mdirectly, by their addition. The mechanism of the above acetalhydrolysis is believed to be,

3.3.1 Specific and general acid catalysis

The simplest case is that in which the reaction rate is found to bea [HEiJ), i.e. to [H3 0EiJ) in aqueous media; the rate rising as pH falls.A common example (cf. p. 210) is the hydrolysis of simple acetals,e.g. MeCH(OEt)z, where it is found that:

3.3 ACID/BASE CATALYSIS

Catalysis in homogeneous solution has already been referred to (p. 41)as operating by making available an alternative reaction path oflower energetic demand, often via a new and more stable (lowerenergy) intermediate: by far the most common, and important,catalysts in organic chemistry are acids and bases.

and specific acid catalysis is found to be characteristic of reactions in-The symbol ~ will be used subsequently to indicate an overall conversion

that proceeds via more than one step.

and the reaction is believed to proceed:

76 The strengths of acids and bases

A type of reaction that has probably received more detailed studythan any other-largely due to the monumental work 6f Ingold andhis school-is nucleophilic substitution at a saturated carbon atom:the classical displacement reaction exemplified by the conversion ofan alkyl halide into an alcohol by the action of aqueous base:

HOe + R-Hal --. HO-R + Hale

Kinetic measurements on reactions in which alkyl halides areattacked by a wide variety of different nucleophiles, Nu:, have revealedtwo, essentially extreme, types; one in which,

Rate = k2[RHal][Nu:] [I]

4.1 RELATION OF KINETICS TO MECHANISM, p. 77.4.2 EFFECT OF SOLVENf, p. 80.4.3 EFFECT OF STRUCIURE, p. 82.4.4 STERE<XEEMICAL IMPUCATIONS OF MECHANISM, p. 87:

4.4.1 SN2 mechanism: inversion of configuration, p. 87;4.4.2 Determination of relative configuration, p. 88; 4.4.3 SN 1mechanism: racemisation?, p. 90; 4.4.4 The mechanistic borderline,p. 91; 4.4.5 SNi mechanism: retention of configuration, p. 92; 4.4.6Neighbouring group participation: 'retention', p. 93.

4.5 EFFECT OF ENTERING AND LEAVING GROUPS, p. 96:4.5.1 The entering group, p. 96; 4.5.2 The leaving group, p. 98.

4.6 OTHER NUCLEOPlllLIC DISPLACEMENTS, p. 99.

4.1 RELATION OF KINETICS TO MECHANISM

Hydrolysis of the primary halide bromomethane (methyl bromide) inaqueous base has been shown to proceed according to equation [I]

Rate = k1[RHal] [2]

i.e. the rate is independent of [Nu:]. In some cases the rate equationsare found to be 'mixed' or are otherwise complicated, but examplesare known which exactly follow the simple relations above.

and another in which,

Nucleophilic substitution at asaturated carbon atom

4

~r +Bre

P CHz-C-Mefast ~

By anal0¥Y, with acids above, specific basic catalysis is found to becharacterIstic of reactions in which there is rapid, reversible protonremoval from the substrate before the slow, rate-limiting step.. In general base catalysis, bases other than sOH are involved. ThusIII the ?a.se cataIysed bromination of acetone (cf. p. 295) in an acetatebuffer It IS found that,

Again by analogy with acids above, general base catalysis is found tobe chara~teristic of reactions in which removal of proton from thesubstrate IS slow, i.e. rate-limiting, and is followed by rapid conversionof the intermediate into products.

Transition state

H [H J'" H" b- I b- /

+ "C-Br -+ HO·····C·····Br -+ HO-C +H" /\ \--H

H HH H(I) (2)

above, and this has been interpreted as involving the participation ofboth alkyl halid~ and hydroxyl ion in the rate-limiting (i.e. slowest)step of the reactIOn. Ingold has suggested a transition state in whichthe attacking hydroxyl ion becomes partially bonded to the reactingcarbon atom bc:fore the incipient bromide ion has become whollydetac~ed from It; thus part of the energy necessary to effect thebrea~Ing of the C-Br bond is then supplied by that produced informIng the HO-C bond. Quantum mechanical calculation showsthat an approach by the hydroxyl ion along the line of centres of thecarbon and bromine atoms is that of lowest energy requirement. Thiscan be represented:

794.1 Relation ofkinetics to mechanism

Rate = kJ[Af[B] Third order overall

Rate = k2[Af Second order overall

Generally, however, it is the order with respect to a particular reactant(or reactants) that is of more interest and significance than the overallorder, i.e. that the above reactions are first order, second order, andsecond order, respectively, with respect to A. Examples of both zeroorder, and non-integral orders, with respect to a particular reactantare also known.

The molecularity refers to the number of species (molecules, ions,etc.) that are undergoing bond-breaking and/or bond-making in onestep of the reaction, usually in the rate-limiting step. It is importantto realise that the molecularity is not an experimentally determinedquantity, and has significance onVin the light of the particularmechanism chosen for the reaction: it is an integral part of themechanistic interpretation of the reaction and is susceptible to re-

Rate = kJ[A][B][C] Third order overall

This type of mechanism has been designated SN 1: Substitution~ucleophilic unirnolecular. The energy necessary to effect the initialionisation is largely recovered from the energy evolved throughsolvation of the resultant ion pair. The entropy of activation, as"',for such a dissociative process (cf. p. 39) is also advantageous; thusas"" for the hydrolysis of Me3CCI is found to be +51 J K-1 mol-I,compared with -17 J K-1 mol-1 for hydrolysis of CH3CI. The cationin the ion pair (4), in which the central carbon atom carries the +vecharge, is of course a carbocation intermediate, and during itsformation the initially Sp3 hybridised carbon atom collapses to amore stable planar (Sp2) state, in which the three methyl groups areas far apart from each other as they can get. Attack by GOH orsolvent (e.g. H 20:) can then take place from either side of thisplanar intermediate. If attainment of this planar state is inhibited bysteric or other factors (cf. p. 87), the carbocation intermediate willbe formed only with difficulty, if at all; i.e. ionisation, and hencereaction by the SNI pathway, may then not take place.

Thus the salient difference between reaction by the SN2 and SN 1pathways is that SN2 proceeds in one step only, via a transitionstate; while SN1 proceeds in two steps, via an actual (carbocation)intermediate.

A certain element of confusion is to be met with both in textbooks,and in the literature, over the use and meaning of the terms order(cf p. 39) and molecularity as applied to reactions. The order is anexperimentally determined quantity, the overall order of a reactionbeing the sum of the powers of the concentration terms that appearin the rate equation:

Me\,C-OH

Me"Me

Me\ <ll -H""C-OH2 +==t

Me"Me

Me/

HO-CX'Me

Me

MeI ~

C<ll 60H

/\Me Me He "'-.tI,o

CI raSl~

(4)

Nucleophilic substitution at a saturated carbon atom

(3)

Me\,C-C1

Me"Me

78

The negative charge is spread in the transition state in the course ofbeing transferred from hydroxyl to bromine, and the hydrogen atomsattached to the carbon atom attacked pass through a position in whichthey all lie in one plane (at right angles to the plane of the paper asdrawn above). The initially Sp3 hybridised carbon atom becomes Sp2h~bridised in the transition state, the HO and Br being associatedwIth the two lobes of the unhybridised p orbital that is therebymade available. This type of mechanism has been designated byIngold as SN2: Substitution ~ucleophilic bimolecular.

By contrast, hydrolysis of the tertiary halide 2-chloro-2-methylpropane (3,t-butyl chloride) in base is found kinetically to followequati0t.I [2], i.e. a~ t~~ rate is independent of [eOH], this can playno part ill th~ rate-lImItIng step. This has been interpreted as indicatingthat the halI.de ~ndergoes slow ionisation (in fact, completion of theR+-CI polansatlOn that has already been shown to be present in sucha molecule) as t~e rate-limiting step to yield the ion pair RfaCIe (4):followed by rapId, non rate-limiting attack by eOH or, if that issuitable, by solvent, the latter often predominating because of itsvery high concentration:

Rate = kob.[RHal]

4.2 EFFECT OF SOLVENT

814.2 Effect of solvent

the T.S. compared with the starting material:

[

6- 6- ]+Nue +R-Hal- Nu'" R ... Hal - Nu-R +Hale

thus solvation of the T.S. is likely to be somewhat less effective thanthat 'of the initial nucleophile-hence the slight decrease. This differingbehaviour of SNI and SN2 modes to changes of solvent can be usedto some extent diagnostically.

A very marked effect on the rate of SN2 reactions is, however,effected on transferring them from polar hydroxylic solvents to polarnon-hydroxylic solvents. Thus the reaction rate of the primary halide,Mel with Neat 0° increased 4· 5 x 104-fold on transfer from methanol(£ ='33) to N,N-dimethylmethanamide(dimet~ylformamide, ~MF),HCONMe2, with very much the same polanty (£ ~ 37). ThIS v~ry

large rate difference stems from the fact that the attackmg nucleophIie,N

3e, is highly solvated through hydrogen-bonding in MeOH (cf. p. 57)

whereas it is very much less strongly solvated-and not bJ hy~roge~bonding-in HCONMe2' The largely unsolvated N.3 amon (10HCONMe2) is a very much more powerful nucleophIi~ .than w~ensurrounded (as in MeOH) by a very much less nucleophIlIc solvatIOnenvelope, hence the rise in reaction rate. Rate increases.of as much as109-fold have been observed on transferring SN2 reactIOns from, e.g.MeOH, to another polar non-protic solvent, dimethyl sulphoxide(DMSO), Me2SO (£ = 46). .

So far as actual changes of mechanistic pathway WIth change ofsolvent are concerned, increase in solvent polarity and ion-solvatingability may (but not necessarily will) change the reaction ~ode fromSN2 ---+ SNI. Transfer from hydroxylic to polar, non-protIc solvents(e.g. DMSO) can, and often do,. change the .reaction mode fromSN I ---+ SN2 by enormously increasmg the effectIveness of the nucleo-phile in the system.

The energy required to effect such a process decreases as £ rises; theprocess is also facilitated by increasing solvation, a.nd consequ~nt

stabilisation of the developing ion pair compared wIth the startmgmaterial. That such effects, particularly solvation, are of primeimportance is borne out by the fact that SNI type reactions are extremelyuncommon in the gas phase.

For the S 2 mode, however, increasing solvent polarity is found tohave a much less marked effect, resulting in a slight decrease in reactionrate. This occurs because in this particular example new charge is notdeveloped, and existing charge is dispersed, in the T.S. compared withthe starting materials;

Nucleophilic substitution at a saturated carbon atom80

and simple kinetic measurements in aqueous solution will thus suggest,erroneously, that the reaction is of the SNI type.

A kinetic distinction between the operation of the SNI and SN2modes can often be made by observing the effect on the overall reactionrate of adding a competing nucleophile, e.g. azide anion, N 3e. Thetotal nucleophile concentration is thus increased, and for the SN2mode where [Nu:] appears in the rate equation, this will result in anincreased reaction rate due to the increased [Nu:]. By contrast, forthe SNI mode [Nu:] does not appear in the rate equation, i.e. is notinvolved in the rate-limiting step, and addition of N 3e will thusbew~thout significant effect on the observed reaction rate, though itwill naturally influence the composition of the product.

evaluation, in the light of additional experimental information aboutthe reaction, in a way that the order cannot be. The molecularity ofthe reaction as a whole only has meaning if the reaction proceeds ina single step(an elementary reaction), as is believed to be the casewith the hydrolysis of bromomethane above (p. 78); order andmolecularity then coincide, the reaction being second order overall(first order in each reactant) and bimolecular. Order and molecularitydo not always, or necessarily, have the same value, however.

Simple kinetic measurements can, however, be an inadequate guide~o which of the above two mechanisms, SNI or SN2, is actually operating10, for example, the hydrolysis of a halide. Thus, as we have seen(p. 45), where the solvent can act as a nucleophile (solvolysis), e.g. H 20,we would expect for an SN2 type reaction,

but as [H 20] remains effectively constant the rate equation actuallyobserved will be,

Changing the solvent in which a reaction is carried out often exerts aprofound effect on its rate and may, indeed, even result in a changein its mechanistic pathway. Thus for a halide that undergoes hydrolysisby the SNI mode, increase in the polarity of the solvent (i.e. increasein £, the dielectric constant) and/or its ion-solvating ability is found toresult in a very marked increase in reaction rate. Thus the rate ofsolvolysis of the tertiary halide, Me3CBr, is found to be 3 X 104 timesfaster in 50 % aqueous ethanol than in ethanol alone. This occursbecause, in the SNI mode, charge is developed and concentrated in

834.3 Effect ofstructure

H, '" /H Me", '" /H Me-..C'" ..... Me

< C < C <t + +Me Me Me

For S 1 attack, considerable charge separation has taken place inthe T.S. (ct. p. 81), and the ion pair intermediate to which it gi~esrise is therefore often taken as a model for it. As the above halideseries is traversed, there is increasing stabilisation of the carbocationmoiety of the ion pair, i.e. increasing rate of forma~ion of the T.S.This increasing stabilisation arises from the operatIOn of both aninductive effect,

RelativeSN2 rate:

and hyperconjugation (p. 25), e.g.

bears the bromine progressively less positively polarised, and hen~e

less readily attacked by sOH. This effect is probably small, and stencfactors are of more importance; thus SOH will find it progressivelymore difficult to attack the bromine-carrying carbon as the latterbecomes more heavily substituted. More significantly, the resultantS 2 transition state will have jive groups around this carbon atom(:c,mpared with only four in the initial ?~l~de), t~ere will thus ~. anincrease in crowding on going from the Initial halide. to the tran~I~lOn

state, and this relative crowding will increase as the size of the ongl~al

substituents increases (H --+ Me). The more crowded the T.S. relativeto the starting materials, the higher its energy will be, and the slowertherefore will it be formed. We would thus expect the purely SN2reaction rate to decrease as the above series is traversed. It is in f~ct

possible to effect nucleophilic s.ubst~tution (Br8 +R-Cl) on a ~~nesof halides analogous to those In Fig. 4.1 (p. 82), under condltIo~s

such that a strictly second order rate equation (SN2 pathway) ISfollowed throughout. We then observe:

via the hydrogen atoms attached to the IX-carbons, the above seriesof carbocations having 0, 3, 6 and 9 such hydrogen atoms, respec-tively. .

Support for such an interaction of the H-C bonds ~Ith. the carbonatom carrying the positive charge is provided by subs~ltutIng H by 0in the original halide, the rate of formation <;>f the ion p~ir is then foundto be slowed down by ~ 10% per deutenum atom Incorporated: aresult compatible only with the H-C bonds being involved in theionisation. This is known as a secondary kinetic isotope effect, secondary


Fig. 4.1

The first and last members are described in the literature as undergoingready hydrolysis, the two intermediate members being more resistant.Measurement of their rates of hydrolysis with dilute, aqueous ethanolicsodium hydroxide solution gives the plot· (Fig. 4.1),

4.3 EFFECT OF STRUCTURE

An interesting sequence is provided by the reaction with base of theseries of halides:

and further kinetic investigation reveals a change in order of reaction,and hence presumably, of mechanism, as the series is traversed. Thusbromomethane (5) and bromoethane (6) are found to follow a secondorder rate equation, 2-bromopropane (7) a mixed second and first~)f.d~r equation, the relative proportion of the two depending on theInitial [SOH] (the higher the initial concentration the greater thesecond order proportion) and the total rate here being a minimum forthe series, while 2-bromo-2-methylpropane (8) is found to follow afirst order rate equation.

In seeking an explanation for the implied changeover in mechanisticpathway we need to consider, in each case, the effect on the transition~tate o~ both electronic and steric factors. For SN2 attack, the enhancedInductIVe effect of an increasing number of methyl groups, as we goacross the series, might be expected to make the carbon atom that

·Based on Ingold, Structure and Mechanism in Organic Chemistry, by permission ofCornell University Press.

854.3 Effect ofstructure

SN 1 attack is thus promoted and allyl, like benzyl, halides arenormally more reactive than species, e.g. CH3CH2CH2Cl andC6HsCH2CH2CH2CI, in which such carbocation stabilisation cannottake place. SN2 attack is also speeded up, compared withCH3CH2CH2Cl, presumably because any electronic effect of thedouble bond-promoting reaction-is not here nullified by an adverse steric effect, as with the bulky C6Hs group in C6HsCH2Cl (cf.above). The proportion of the total reaction proceeding by each ofthe two pathways is found to depend on the conditions: morepowerful nucleophiles promoting the SN2 mode (cf. p. 96).

By contrast, vinyl halides such as chloroethene, CH2=CHCl, andhalogenobenzenes are very unreactive towards nucleophiles. Thisstems from the fact that the halogen atom is now bonded to an Sp2

hybridised carbon, with the result that the electron pair of the C-Clbond is drawn closer to carbon than in the bond to an Sp3 hybridised carbon. The C-Cl is found to be stronger, and thus less easilybroken, than in, for example, CH3CH2Cl, and the C-Cl dipoleis smaller; there is thus less tendency to ionisation (SN1) and a lesspositive carbon for SOH to attack (SN2); the n electrons of the doublebond also inhibit the close approach of an attacking nucleophile. Thedouble bond would not help to stabilise either the SN2 transition stateor the carbocation involved in the SN 1 pathway. Very much thesame considerations apply to halogenobenzenes, with their Sp2

hybridised carbons and the 7r orbital system of the benzene nucleus;their reactions, which though often bimolecular are not in factsimply SN2 in nature, are discussed further below (p. 170).

The influence of steric factors on the reaction pathway is particularlyobserved when substitution takes place at the tJ-position. Thus for the

This is a classical example of an ion stabilised by charge delocalisationvia the agency of the delocalised n orbitals of the benzene nucleus(cf. the negatively charged phenoxide ion, p. 23). The effect willbecome progressively more pronounced, and SN 1 attack furtherfacilitated, with (C6HshCHCl(11) and (C6HsbCCl(12), as the possibilities for delocalising the positive charge are increased in thecarbocations to which these latter halides give rise.

SN2 attack on the CH2 in (10) is found to proceed at very muchthe same rate as on that in MeCH2Cl, suggesting that any adversesteric crowding in the T.S. by the bulky C6Hs group is compensatedby a small electronic (inductive?) effect promoting reaction.

Similar carbocation stabilisation can also occur in the hydrolysisof allyl halides, e.g. 3-chloropropene:

(C6H,hCH -Cl

(ll)


because it is a bond other than that carrying the isotopic label that isbeing broken (cf p. 46). The relative contributions of hyperconjugationand inductive effects to the stabilisation of carbocations is open todebate, but it is significant that a number of carbocations will onlyfonn at all if they can take up a planar arrangement, the state inwhich hyperconjugation will operate most effectively (cf. p. 104).

In steric terms there is a relief of crowding on going from theinitial halide, with a tetrahedral disposition of four substituents aboutthe Sp3 hybridised carbon atom, to the carbocation, with a planardisposition of only three substituents (cffive for the SN2 T.S.) about thenow Sp2 hybridised carbon atom. The three substituents are as farapart from each other as they can get in the planar carbocation, andthe relative relief of crowding (halide~ carbocation) will increase asthe substituents increase in size (H~ Me~ Me3C). The SN 1 reaction rate would thus be expected to increase markedly (on bothelectronic and steric grounds) as the series of halides is traversed. Ithas not, however, proved possible to confinn this experimentally bysetting up conditions such that the four halides of Fig. 4.1 (p. 82) allreact via the SN 1 pathway.

Thus, as the SN2 rate is expected to decrease, and the SN 1 rate toincrease, across the series in Fig. 4.1, the reason for the observedpattern of reaction rates, and changeover in reaction pathway,becomes apparent.

A similar mechanistic changeover is observed, though considerablysooner, in traversing the series:

Thus for hydrolysis in 50 % aqueous acetone, a mixed second andfirst order rate equation is observed for phenylchloromethane (benzylchloride, 10}-moving over almost completely to the SNI mode inwater alone. Diphenylchloromethane (11) is found to follow a firstorder rate equation, with a very large increase in total rate, whilewith triphenylchloromethane (trityl chloride, 12) the ionisation is sopronounced that the compound exhibits electrical conductivity whendissolved in liquid S02. The main reason for the greater promotionof ionisation-with consequent earlier changeover to the SNI pathwayin this series-is the considerable stabilisation of the carbocation, bydelocalisation of its positive charge, that is now possible:

87

(1)

oJa~

10- 23

(18)

1

4.4 Stereochemical implications of mechanism

(+)

R'\.

C-Dr -+

/JR'R"

• A chiral compound is one that is not superimposable on its mirror image.

4.4 STEREOCHEMICAL IMPLICAnONS OF MECHANISM

Hydrolysis of an optically active form of a chiral· halide presents so~einteresting stereochemical features. Thus considering each pathway In

turn :

4.4.1 SN2 mechanism: inversion of configuration

in which the bromine atom is found to be virtually inert to nucleophiles.Despite the formal resemblance in the environment of the bromineatom in (19) to that in (18), they are found to differ in their rate ofreaction under parallel conditions by a factor of ~ 10- 23

: I! This isbecause stabilisation of the carbocation from (8) can occur bydelocalisation of its charge through the 7r orbital systems of thethree benzene rings; whereas the extremely rigid structure of (19)will hold the cation's empty orbital (from loss of Bre ) all but at rightangles to these 7r orbital systems, thus preventing such delocalisation.

also inhibited because the resultant carbocations from (16) and (17)would be unable, because of their rigid frameworks, to stabilisethemselves by collapsing to the stable planar state. These carbocation intermediates are thus of very much higher energy level thanusual, and therefore are formed only slowly and with reluctance.The very greatly reduced solvolysis rate of (17) compared with (16)reflects the greater rigidity about the bridgehead (cationic) carbonwith a one-carbon (17), than with a two-carbon (16), bridge.

This rigidity is carried even further in I-bromotriptycene (19),

Me)C-CH2-Dr

(15) 4·2 x 10- 6

(17)

( ISal

Me2CH-CH 2-Dr

(14) 3·0 x 10- 2

(16)

(I4al

(8)

Dr

~

MeCH2-CH2 - Dr

(13) 2·8 x 10- 1


CH)-CH 2-Dr

(6) 1·0

series,

86

the figures quoted are relative rates of reaction (SN2 throughout) withEtOS in EtOH at 55°. Any differences in electronic effect of the Megroups through two saturated carbon atoms would be very small, andthe reason for the rate differences is steric: increased difficulty ofapproach of EtOS 'from the back' of the carbon atom carryingBr, and increased crowding in the resultant T.S. The reason for theparticularly large drop in rate between l-bromo-2-methylpropane (14)and I-bromo-2,2-dimethylpropane(neopentyl bromide, 15) is that theT.S. for the former, though somewhat crowded, can, by rotation aboutthe C",-Cp bond, adopt one conformation (l4a) in which the attackingEtOS is interfered with only by H, while no such relief of crowding isopen in the T.S. (15a) for the latter (but see also, p. 110):

1 ::::: 10- 6 ::::: 10- 14

All are tertiary halides so that attack by the SN2 mode would not beexpected to occur on (16) or (17) any more than it did on (8) (cf. p. 82).SN2 attack 'from the back' on the carbon atom carrying Br would inany case be prevented in (16) and (17) both sterically by their cagelike structure, and also by the impossibility of forcing their fairly rigidframework through transition states with the required planar distribution of bonds to the bridgehead carbon atom (cf. p. 84). Solvolysis viarate-limiting formation of the ion pair (SN 1), as happens with (8) is

The T.s. (I Sa) will thus be at a much higher energy level, aG* (p.38)will be larger and the reaction rate correspondingly lower.

The effect of structure on relative reactivity may be seen particularlyclearly when a halogen atom is located at the bridgehead of a bicyclicsystem. Thus the following rates were observed for solvolysis in 80%aqueous ethanol at 25°:

formation of an ester with 4-methylbenzenesulphonyl(tosyl) chloride

4.4.2 Determination of relative configuration

This turns essentially on the fact that if a chiral compound undergoesa reaction in which a bond joining one of the groups to the chiralcentre is broken, then the centre may-though it need not of necessityundergo inversion of configuration; while if the compound undergoesreaction in which no such bond is broken then the chiral centre willpreserve its configuration intact.

Thus in the series of reactions on the optically active( +) alcohol(20),

89

/C6H 13

IZ81_C(-)'\Me

H

(25a)

4.4.2 Determination of relative configuration

• That such is the case may be shown by using an alcohol labelled with 180 in itsOH group. and demonstrating that this atom is not eliminated on forming the tosylate;it is, however, eliminated when the tosylate is reacted with MeCOze.

t Hydrolysis of an acetate in which the alcohol-oxygen atom is 180 labelled failsto result in the laUer's replacement, thus showing that the alkyl-oxygen bond of theacetate is not broken during its hydrolysis (cf. p.47).

is known not to break the c-o bond of the alcohol,* hence thetosylate (21) must have the same configuration as the original alcohol.Reaction of this ester (21) with MeCOz

e is known to be a displacementin which ArS03

6 (Ar = p-MeC6H 4 ) is expelled and MeC02 6

introduced,* hence the c-o bond is broken in this reaction, andinversion of configuration can thus take place in forming the acetate(22). Alkaline hydrolysis of the acetate (22 --+ 23) can be shown notto involve fission of the alkyl-oxygen C-O linkage,t so the alcohol(23) must have the same configuration as the acetate (22). As (23) isfound to be the mirror image of the starting material (20}-oppositedirection of optical rotation-an inversion of configuration must havetaken place during the series of reactions, and can have occurred onlyduring reaction of MeC02

6 with the tosylate (21). Reaction of thistosylate (21) with a number of nucleophiles showed that inversion ofconfiguration occurred in each case: it may thus be concluded withsome confidence that it occurs on reaction with Br8 to yield thebromide (24), i.e. that the bromide (24), like the acetate (22), hasthe opposite configuration to the original alcohol (20).

The general principle-that bimolecular (SN2) displacement reactions are attended by inversion of configuration-was established inan elegant and highly ingenious experiment, in which an opticallyactive alkyl halide undergoes displacement by the same-thoughisotopically labelled-halide ion as nucleophile, e.g. radioactive 12816

on (+ )2-iodooctane (25) :

The displacement was monitored by observing the changing distribution of radioactive 1281 between the inorganic (sodium) iodide and2-iodooctane, and it was found, under these conditions, to be secondorder overall (first order with respect to 12816 and to 2-iodooctane)with k2 = 3·00 ± 0·25 x 10- 5 (at 30°).

If inversion takes place, as SN 2 requires, the optical activity of thesolution will decline to zero, i.e. racemisation will occur. This willhappen because inversion of the configuration of a molecule of (+)(25) results in formation of a molecule of its mirror image (-)

(24)

R/

Dr-C.\'R'

R"

[

R=PhCH2 ]R'=MeR"=HAr= p-MeC6 H 4

R/

MeCO~O-C ..\"R'R"

(22)

RArSO,C1~ \ + Bra

..C OSOzAr -R"/

R"

(21) 1MeCO,a

R

HO-C: ~\"R'

(-) R"

(23)

R\.C-OH

R"jR"(+)

(20)


It will be seen that the spatial arrangement of the three residualgrQups attached to the carbon atom attacked has been effectivelyturned inside out. The carbon atom is said to have undergone inversionof its configuration (the arrangement in space of the groups attachedto it). Indeed, if the product could be the bromide, instead of, as here,the corresponding alcohol, it would be found to rotate the plane ofpolarisation of plane polarised light in the opposite direction, i.e. ( - ),to the starting material, (+), for it would, of course, be its mirrorimage (cf p. 89). The actual product is the alcohol, however, and weare unfortunately not able to tell, merely by observing its direction ofoptical rotation, whether it has the same or the opposite configurationto the bromide from which it was derived: compounds, other thanmirror images, that have opposite configurations do not necessarilyexhibit opposite directions of optical rotation, while compounds thathave the same configuration do not necessarily exhibit the same direction of optical rotation. Thus in order to confirm that the above SN2reaction is, in practice, attended by an inversion of configuration, astheory requires, it is necessary to have an independent method forrelating the configuration of starting material and product, e.g. thebromide and corresponding alcohol above.

4.4.3 SN1 mechanism: racemisatioo?

(25a), which 'pairs off' with a second molecule of (+) (25) to fonn a(±) racemate: the observed rate of racemisation will thus be twicethe rate of inversion. The reaction was monitored polarimetrically,the rate of racemisation measured thereby, and the rate of inversioncalculated from it: it was found to have k=2·88±0·03xlO-5 (at30°).

The rate of displacement and of inversion are thus identical withinthe limits of experimental error, and it thus follows that each act ofbimolecular displacement must thus proceed with inversion of configuration. Having shown that SN2 reactions are attended by inversionof configuration, independent demonstration that a particular reactionoccurs via the SN2 mode is often used to correlate the configurationof product and starting material in the reaction.

914.4.4 The mechanistic borderline

Rate = k[R-X)

This is so because in the SN2 pathway the concentration of nucleophile will remain essentially constant throughout the react!onas-being also the solvent-it is present in very large, unchang10g

4.4.4 The mechanistic borderlineReference has already been made (p. 82) to the fact that thereactions of some substrates, e.g. secondary halides, may follow amixed first/second order rate equation. The question then ariseswhether such a reaction is proceeding via both SN2 and SN1pathways simultaneously (their relative proportions depend.ing ~n

the solvent, etc.) or whether it is proceeding via some speCific, '10between' mechanistic pathway.

In solvolytic reactions like those we have just been considering,where the solvent itself is the nucleophile, such mixed kinetics maynot be detectable, irrespective of what is actually happening, as bothSN 1 and SN2 pathways are likely to follow a rate equation of thefonn:

Here (26) is an intimate ion pair in which the jointly solvated gegenions are in very close association with no solvent molecules betweenthem, (27) is a solvent-separated ion pair, and (28) represents the nowdissociated, and separately solvated, pair of ions.

In a solvolysis reaction, attack on R$ by a solvent molecule, e.g.H20:, in (26) is likely to lead to inversion, as attack can take place(by the solvent envelope) on the 'back' side of R$, but not on the 'front'side where there are no solvent molecules, and which is shielded bythe Bra gegen ion. Attack in (27) is more likely to lead to attack fromeither side, leading to racemisation, while attack on (28) can clearlyhappen with equal facility from either side. Thus the longer the lifeof R$, i.e. the longer it escapes nucleophilic attack, the greater theproportion of racemisation that we should expect to occur. The lifeof R $ is likely to be longer the more stable it is-{a) above-but theshorter the more powerfulIy nucleophilic the solvent-{b) above.

Thus solvolysis of (+ )C6 HsCHMeCI, which can form a stabilisedbenzyl type carbocation (cf. p. 84), leads to 98% racemisation while(+ )C6 H 13CHMeCI, where no comparable stabilisation can occ~r,

leads to only 34 % racemisation. Solvolysis of (+ )C6 HsCHMeCI 1080 %acetone/20 %water leads to 98 %racemisation (above), but in themore nucleophilic water alone to only 80 % racemisation. The samegeneral considerations apply to nucleophilic displacement reactionsby Nu: as to solvolysis, except that R$ may persist a little furtheralong the sequence because part at least of the solvent envelope has tobe stripped away before Nu: can get at R$. It is important to noticethat racemisation is clearly very much less of a stereochemicalrequirement for SNI reactions than inversion was for SN2.

(±)

(28)

R/

HO-C.\,R'R"

50%

(27)

R "OH/tel //\

R' R"~

Bre "OH~

(26)

R\o.C-Br ~

R'o,Roo (+)


R\.C-OH

R'"R"

50%

As the carbocation formed in the slow, rate-limiting step of thereaction is planar, it might be expected that subsequent attack by anucleophile such as eOH, or the solvent (H20:), would take placewith equal readiness from either side of this planar carbocation;leading to a 50/50 mixture of species having the same, and theopposite, configuration as the starting material, i.e. that racemisationwould take place yielding an optically inactive (±) product.

In practice, however, the expected racemisation-and nothing butracemisation-is rarely observed, it is almost always accompanied bysome degree of inversion. The relative proportions of the two are foundto depend on: (a) the structure of the halide, in particular the relativestability of the carbocation to which it can give rise; and (b) thesolvent, in particular on its ability as a nucleophile. The more stablethe carbocation, the greater is the proportion of racemisation; themore nucleophilic the solvent, the greater is the proportion ofinversion. These observations become understandable if the ratelimiting SN1 ionisation follows the sequence:

4.4.5 SNi mechanism: retention of configuration

De~pite v.,rhat has been said above about displacement reactions leadingto InverSIOn of configuration, to racemisation, or to a mixture ofboth, a number of cases are known of reactions that proceed withactual retention of configuration, i.e. in which the starting materialand product have the same configuration. One reaction in which thishas been shown to occur is in the replacement of OH by CI throughthe use of thionyl chloride, SOClz :

excess. This raises the question whether the mixture ofracemisation/inversion observed in such cases stems from the simultaneou~ operation.of SN 1 and SN2 pathways for solvolysis, ratherthan via the relatively elaborate, variable ion pair hypothesis advanced above.

In some cases at least it is possible to demonstrate that a 'mixed'SN 1+ SN2 pathway is n?t operating. Thus solvolysis of the halide,(+)CfjHsCHMeCI, mentIOned above, but this time in MeCOzH,

MeCo,HC6 HsCH-CI -----=--+ C6 HsCH-OCOMe

I IMe Me(+) 88% racemisation

12% net inversion

was found to lead to 88% racemisation, and 12% net inversion.Adding the much more powerfully nucleophilic MeCOz8 (asMeCOz8NaEll) to the reaction mixture was found to result in: (a) noincrease in the overall reaction rate, and (b) no increase in theproportion of net inversion. This strongly suggests that the inversionthat is observed does not stem from part of the overall reactionproceeding via an SN2 pathway simultaneously with the (major)SN 1 mode. If it did, we would expect the change to a much morepowerful nucleophile (MeCOzH --. MeCOz8) to lead to markedincreases in both (a) and (b) above.. A good deal of interest, and controversy, has centred on whetherIn the ~a~t analysis there is perhaps a continuous spectrum ofmechamstIc pathways intermediate between SN2 and S 1: these~~perceptibly shading into each other via gradually varying transltI~n states from ~he pure SN2 side, and via gradually varying ionpau/solvent combInations from the pure SN 1 side. It is an area inwhich theory has shaded over into semantics if indeed not eveninto theology' ' ,

93

Me\

---+ C-C1Ph)

H

(30a)

+ S02MeI

CEll CIe

/\Ph H

(33)

Me 0I II

CEll eOSCl

,/ \Ph H

(32)

4.4.6 Neighbouring group participation: 'retention'

Me 0 r Me 0 ]+ Me\ II 6- I 116- /,C-OSCI --. CI .. · .. ·C ...... OSCI --. CI-C..

Ph'" .: \ \'PhH Ph H H

(31) (30b)

(31)

Me 0\ II.C-OSCI

Ph"H

Cle +

The reaction has been shown to follow a second order rate equation,rate = kz[ROH][SOClz], but clearly cannot proceed by the simpleSN2 mode for this would lead to inversion of configuration (p. 87) inthe product, which is not observed,

Carrying out the reaction under milder conditions allows of theisolation of an alkyl chlorosulphite, ROSOCI (31), and this can beshown to be a true intermediate, The chlorosulphite is formed withretention of configuration, the R-0 bond not being broken duringthe reaction. The rate at which the alkyl chlorosulphite intermediate(31) breaks down to the product, RCI (30a), is found to increase withincreasing polarity of the solven t, and also with increasing stabilityof the carbocation REll: an ion pair, R E98 0SOCl (32), is almostcertainly involved, Provided collapse of the ion pair to products thenoccurs rapidly, i.e. in the intimate ion pair (33) within a solvent cage(cf. p. 90), then attack by Cl8 is likely to occur on the same side ofRE9 from which 80S0Cl departed, i.e, with retention of configuration:

Whether the breaking of the C-O and the S-CI bonds occurssimultaneously, or whether the former occurs first, is still a matter ofdebate.

It is interesting that if the SOClz reaction on ROH (29) is carriedout in the presence of pyridine, the product RCI is found now to haveundergone inversion of configuration (30b), This occurs because theHCI produced during the formation of (31) from ROH and SOClz isconverted by pyridine into CsHsNHEilCle and Cle, being an effectivenucleophile, attacks (31) 'from the back' in a normal SN2 reactionwith inversion of configuration:


Me\,C-OH

Ph"H

(29)

MeSOCI't t-CI

Ph'jH(30a)

+ S02 + HCI


There are also some examples of retention of configuration in nucleophilic displacement reactions where the common feature is an atom orgroup-dose to the carbon undergoing attack-which has an electron

(39) (41) (40)

Whether the intermediate (41) is a zwitterion as shown or a highly

95

(44)

oII MC e

/ , ...O-C

(4Ia)\

Els0cH2E'CI, /

CH 2

(42)


MeI

EtS CHOH, /Expected

Me Me ~CH 2

I I faSIEll

EIs0cH-cf1 EIS-CH, / --+ e' /CH 2 slow CI CH 2

~(45) (46) H,O MeI

EtS-CH/

UnexpectedCHzOH

By contrast, 0: in (43) is ~ufficiently electronegative not to donate a!'electron pair (unlike Oe in ROe and RC02e above), and hydrolySISof EtOCH2CH2CI thus proceeds via ordinary SN2 attack by an ~xternalnucleophile-which is likely to be very much ~Iower tha? the Internalnucleophilic attack in (42) -+ (44). That a cychc sul~hontum salt suchas (44) is involved is demonstrated by the hydrolySIS .of the anal~gue(45), which yields two alcohols (the unexpect,ed ~ne m gr~ater YI~ld)indicating the participation of the unsymmetncal mtermedlate (46).

N: can act as a neighbouring group in similar, circumstances, e.g.the hydrolysis of Me2NCH2CH2CI, but the rate IS markedly slower,

has not been clearly established. As the co~centration of nucle?phile,[eOH], is increased an increasing proportl.on ~f normal SN2 attackfrom the back', with inversion of configuratIOn, IS observed..

Neighbouring group effects have also been o~ser~ed ~Ith atomsother than oxygen, e.g. sulphur and nitrogen, and m sltuatlOn~ where,though no stereochemical point is at issue, unexpectedly rapid rat~ssuggest a change in reaction path~ay. Thus EtSCH 2CH2CI (42) ISfound to undergo hydrolysis 104 times fa.ster than Et,OCH2CH2C1(43) under comparable conditions: and thiS has been mterpreted asinvolving S: acting as a neighbounng group:

labile IX-lactone (4Ia)

oIIC

~ eO 'C-OHMe'}

H

~iOn(i)

CEt280H / \- Oa1,C!-.CI

Me"H

(36)

oIIC

e /, ~O:···C-BrMe'}

H

HOCEt2~C!-C1

Me"H

(34)


t Preferential attack takes place on this, rather than the other, carbon of the threemembered ring as it will be the more positive of the two, carrying as it does only oneelectron-donating alkyl group.

HOC(t2 H 0 F~t2 HOy (37),C~H"""'!"" eo .C!.....OH /inversion (ii)

Me" Me"H H(35) (38)

Initial attack by base on (34) yields the alkoxide anion (36),internal attack by this ROe then yields the epoxide (37) withinversion of configuration at C* (these cyclic intermediates canactually be isolated in many cases); this carbon atomt, in turn,undergoes ordinary SN2 attack by eOH, with a second inversion ofconfiguration at C*. Finally, this second alkoxide anion (38) abstracts a proton from the solvent to yield the product 1,2-diol (35)with the same configuration as the starting material (34). Thisapparent retention of configuration has, however, been broughtabout by two successive inversions.

Another example of oxygen as a neighbouring group occurs in thehydrolysis of the 2-bromopropanoate anion (39) at low [eOH], whichis also found to proceed with retention of configuration (40). The rateis found to be independent of [eOH], and the reaction is believed toproceed:

94

pair available. This neighbouring group can use its electron pair tointeract with the 'backside' of the carbon atom undergoing substitution, thus preventing attack by the nucleophilic reagent; attack canthus take place only 'from the frontside', leading to retention ofconfiguration. Thus base hydrolysis of the 1,2-chlorohydrin (34) isfound to yield the 1,2-diol (35) with the same configuration (retention):

EIOe > PhOe > MeC02 e > NOJe

A shift in mechanistic type can also occur with change of nucleophile,thus a displacement that is SNI with, for example, H 20:, HC0 3

e ,MeC02e , etc., may become SN2 with eOR or EtOe .

Nucleophilicity is found to be very much affected by the size of theattacking atom in the nucleophile, at least for comparisons within

974.5.1 The entering group

the same group or sub-group of the periodic table; thus we find:

[B- B-]+

NCe +R-Br~ NC'" R· .. Br ~N==C-R+Bre

This is understandable as, unlike SNI, bond formation is now takingplace in the T.S. for the rate-limiting step, for which ready polarisabilityof the bonding atom of the nucleophile is clearly important-the

T.S.

In the absence of such promotion by Ag EB , e.g. with NaEB[CN]e, theresulting SN2 reaction is found to proceed with preferential attack onthe atom in the nucleophile which is the more polarisable:

Size as well as electronegativity, governs polarisability (cf soft bases,above): as the atom increases in size the hold the nucleus has on theperipheral electrons decreases, with the result that they become morereadily polarisable, leading to the initiation of bonding at increasingnuclear separations. Also the larger the nucleophilic ion or group theless its solvation energy, i.e. the more readily is it converted into theeffective, largely non-solvated, nucleophile; thus heats of hydrationof Ie and Fe are 284 and 490 kJ mol-I, respectively. This combination of factors makes the large, highly polarisable, weakly solvatediodide ion, Ie, a very much better nucleophile than the small,difficulty polarisable, strongly solvated (R-bonding with a hydroxylicsolvent) fluoride ion, Fe. We should, on this basis, expect theincrease in reaction rate on transfer from a hydroxylic to a polarnon-pro tic solvent (cf. p. 81) to be much less for Ie than, forexample, for Bre or Cle : as is indeed found to be the case (Bre is abetter nucleophile than Ie in Me2CO).

A further interesting point arises with nucleophiles which havemore than one-generally two-suitable atoms through which theycan attack the substrate, ambident nucleophiles:

[ex=y ...... x=ye]

It is found in practice that in (highly polar) SNI reactions attack takesplace on the carbocationic intermediate, REB, through the atom inthe nucleophile on which electron density is the higher. With, forexample, halides that do not readily undergo SN 1 attack this can bepromoted by use of the silver salt of the anion, e.g. AgCN, as AgEB

promotes REB formation by precipitation of AgRal (cf. p. 102):

[eC=N ...... C=N e]


under comparable conditions, than that for (42) above, because of thegreater stability of the cyclic immonium ion intermediate correspondingto (44). Such cyclic species are formed during the hydrolysis of mustardgas, S(CH 2CH 2Clh and the related nitrogen mustards, such asMeN(CH 2 CH 2Clh: the cyclic immonium salts derived from the latterare also powerful neurotoxins. The 1t orbital system of the benzenering can also act as a neighbouring group (cf. pp. 105, 376).

4.5 EFFECT OF ENTERING AND LEAVING GROUPS

4.5.1 The entering group

Changing the nucleophilic reagent employed, i.e. the entering group,will not directly alter the rate of an SNI displacement reaction for thisreagent does not take part in the rate-limiting step of the overallreaction. In an SN2 displacement, however, the more strongly nucleophilic the reagent the more the reaction will be promoted. The nucleophilicity of a reagent might perhaps be expected to correlate with itsbasicity, as both involve the availability of electron pairs and the easewith which they are donated. The parallel is by no means exact, however,in that basicity involves electron pair donation to hydrogen, whereasnucleophilicity involves electron pair donation to another atom, veryoften carbon; basicity involves an equilibrium (thermodynamic), i.e.L1G B , situation, whereas nucleophilicity usually involves a kinetic, i.e.L1G *, one; basicity is likely to be little affected by steric influences,whereas nucleophilicity may be markedly affected.

This distinction follows to some extent the recently introducedone between hard and soft bases: a hard base is one in which thedonor atom is of high electronegativity, low polarisability, and ishard to oxidise, i.e. eOH, eOR, R3 N: ; while a soft base is one inwhich the donor atom is of low electronegativity, high polarisability,and is easy to oxidise, e.g. RSe , Ie, SCNe ; for a given degree ofbasicity, softness promotes nucleophilicity. Basicity data are oftenthe more readily available, however, and can be used as a guide tonucleophilicity provided like is being compared with like. Thus if theattacking atom is the same (cf. electronegativity above), then thetwo run reasonably in parallel, and we find the stronger the base themore powerful the nucleophile:

96

R-I> R-Dr > R-CI > R-F

Ie + R-CI'"ra..

i I-R + CleIt"" rast

H<J)le + R-OH

beginning of bonding at as great an internuclear separation as possible(cf above). This AgCN/NaCN dichotomy has long been exploitedpreparatively. Similarly, nitrite ion [NOz]e is found to result in theformation of ~Ikyl nitrites, R-<?-N =0, under SN I conditions (0 isthe atom of higher electron denSity) and nitroalkanes, R- NO under8N2 conditions (N is the more readily polarisable atom). z'

994.6 Other nucleophilic displacements

<J)Dre + Me-NMe3 --+ Me-Dr + :NMe3

<J)Ie + C6 H I3-OH --+ C6 H 13 -1 + H20:

H

<J)Et2S: + Me-Dr --+ Et2 SMe + Dre

nucleophilic anions on positively charged species,

<J)Me3 N: + Et-Dr --+ Me3 NEt + Dre

4.6 OTHER NUCLEOPHILIC DISPLACEMENTS

H'"!i

This is known as nucleophilic catalysis. The stronger, and harder, asa base a leaving group is, the less readily can it be displaced; thusgroups such as eOH, eOR, eNH z bonded to carbon by small, highlyelectronegative atoms of low polarisability (cf hard bases, above)cannot normally be displaced directly by other nucleophiles.

Displacements that are otherwise difficult, or even impossible, toaccomplish directly may sometimes be effected by modification of thepotential leaving group-often through protonation-so as to makeit weaker, and/or softer, as a base. Thus eOH cannot be displaceddirectly by Bre , but is displaced readily if protonated first:

Dre + R-OH -- Dr-R + eOH

H'" <J) IeR-OPh ~ R-OPh --+ RI + PhOH

H

In this discussion of nucleophilic displacement at a saturated carbonatom, interest has tended to centre on attack by nucleophilic anionsNu: e, es~ecially eOH, on polarised neutral species, especially alkylhalides, + R- Hal6

-. In fact this general type of displacement isextremely common involving, in addition to the above, attack by noncharged nucleophiles Nu: on polarised neutral species,

<J)Dre + R-~H --+ Dr-R + H 20

H

There are two main reasons for this: (a) Bre is now attacking a positivelycharged, as opposed to a neutral, species, and (b) the very weakly basicHzO is a very much better leaving group than the strongly basic eOH.The well known use of HI to cleave ethers results from Ie being aboutthe most nucleophilic species that can be generated in the strongly acidsolution required for the initial protonation:


4.5.2 The leaving group

Changing the leaving group will clearly alter the rate of both S 1 and~N2 reactions, as breaking the bond to the leaving group is in~olved10 .t~e slow, rate-limiting step of both. We might expect the relativeabilIty of Y as a leaving group, in R-Y, to be influenced by: (a) thestrength of the R-Y bond; (b) the polarisability of this bond· (C)the .s!abi.lity of y e ; and (r~lated to the latter) (d) the degre~ ofstabilIsatIon, through solvatIon, of the forming y e in the T.S. foreither SN1 or SN2.

The observed reactivity (SN2 or SN1) sequence for halides

suggests that here (a) and (b) above are probably more important~han (c) and (d). For a wider range of potential leaving groups,mvolvement of (c) would suggest that the weaker y e is as a base (orthe. st.ronger H-Y is as an acid) the better a leaving group it will be.Th~s IS borne ou~ to some extent over a series of leaving groups inwhich the atom 10 Y through which it is attached to R remains thesame. Thus the anions of strong 'oxygen acids' such as p_Me~6H4S03 e (tos~late, cf. p. 88), CF3S03

8 (triflate) are goodleavmg groups (as mdeed are halide anions); with such O-Ieavinggroups, (c) and (d) above are of increased importance. Relativeleaving group ability may however vary with change of solventreflecti~g the influence of (d). This variation in relative ability ca~b~ particularly m~rked on changing from a hydroxylic solvent to abipolar, non-protlc one (e.g. Me2S0, HCONMe2' etc.) as initial(c )/(d) domination may then shift to (a )/(b) control.

~igh polari~ability makes Ie both a good entering and a goodleavmg group, It can thus often be used as a catalyst to promote anotherwise slow displacement reaction, e.g.:

100 Nucleophilic substitution at a saturated carbon atom

and non-charged nucleophiles on positively charged species (N2 isprobably the best leaving group there is):

H20: + PhN2Ell -+ PhOH + N2 + HEll

We have also seen good leaving groups other than halide ion, e.g.tosylate anion (cf p. 88),

and 'internal' leaving groups (c/. p. 94):

Cle\ CH2-CH2 -+ CICH2CH20 e

(p/

There are also nucleophilic displacement reactions, of considerable synthetic importance, in which the attacking atom in the nuc1eophile is carbon in either a carbanion (p. 288) or a source ofnegatively polarised carbon (c/. p. 221); new carbon-carbon bondsare thus formed:

8NHHC=CH ~2 HC=Ce + Pr-Br -+ HC=C-Pr + Bre

Et08

CH2(C02Et)2 ~ (Et02C)2CHe + PhCH2-Br -+ (Et02ChCH-CH2Ph + Bre

6+ 6-

BrMgPh + C6H 13-Br -+ MgBr 2 + Ph-C6H I3

It should be remembered that in the above examples what is nucleophilic attack from the viewpoint of one participant is electrophilicattack from the viewpoint of the other. Any designation of the processas a whole tends therefore to be somewhat arbitrary, reflecting as itdoes our preconceptions about what constitutes a reagent as opposedto a substrate (cf p. 30).

Hardly surprisingly, not all nucleophilic displacement reactionsproceed so as to give 100% yields of the desired products! Here, aselsewhere, side-reactions occur yielding unexpected, and in preparativeterms unwanted, products. A major side-reaction is elimination toyield unsaturated compounds: this is discussed in detail below (p. 246).

5Carbocations, electron-deficient Nand 0 atoms and their reactions

5.1 MEIHODS OF FORMING CARBOCATIONS, p. 101.5.1.1 Heterolytic fission of neutral species, p. 101; 5.1.2 Additionof cations to neutral species, p. 103; 5.1.3 From other cations,p.104.

5.2 STABD...ITY AND STRUCTURE OF CARBOCATIONS, p. 104.5.3 CARBOCATION REACTIONS, p. 107.5.4 CARBOCATION REARRANGEMENTS, p. 109.

5.4.1 Without change in carbon skeleton, p. 109: 5.4.1.1 Allylicrearrangements, p. 109; 5.4.2 With change in carbon skeleton,p. no: 5.4.2.1 Neopentyl rearrangements, p. no, 5.4.2.2Rearrangement of hydrocarbons, p. 112, 5.4.2.3 Pinacol/pinacolonerearrangements, p. 113, 5.4.2.4 Stereochemistry of rearrangements,p. 116, 5.4.2.5 Wolff rearrangements, p. 119.

5.5 DIAZONIUM CATIONS, p. 119.5.6 MIGRATION TO El..ECTRON-DEFICIENT N, p. 122:

5.6.1 Hofmann, Curtius, Lossen and Schmidt reactions, p. 122;5.6.2 Beckmann rearrangements, p. 123.

5.7 MIGRATION TO El..ECTRON-DEFICIENT 0, p. 127:5.7.1 Baeyer-Villiger oxidation of ketones, p. 127; 5.7.2Hydroperoxide rearrangements, p. 128.

Reference has already been made in the last chapter to the generationof carbocations, in ion pairs, as intermediates in some displacementreactions at a saturated carbon atom, e.g. the solvolysis of an alkylhalide via the SN1 mechanism. Carbocations are, however, fairlywidespread in occurrence and, although their e~istence. is ofte~ onlytransient, they are of considerable importance 10 a wIde varIety ofchemical reactions.

5.1 METHODS OF FORMING CARBOCATIONS

5.1.1 Heterolytic fission of neutral species

The obvious example is simple ionisation, the group attached tocarbon departing with its bonding electrons to form an ion pair,REIl y8:

1035.1.2 Addition of cations to neutral species

(I)

where the intermediate (1) is a delocalised carbocation.

thus providing a more positive carbon atom for subsequent attackby a nucleophile, in this case H20: in acid-catalysed hydration ofcarbonyl compounds (cf. p. 207). That such protonation does indeedoccur may be demonstrated, in the absence of water, by the two-folddepression of freezing point observed with ketones in concentratedsulphuric acid:

"- "-$c=o + H SO i=! C-OH + HSO e/ z. / •

Carbocations may also be generated by protonation of lone pairelectrons, if the protonated atom is converted into a better leavinggroup thereby and ionisation thus promoted:

H,SO. $ H,SO.Ph)C-OH 4 • HSO.e + Ph)C-OH. ' Ph)C$ + H)O$ + 2HSO.e

H

c/. protonation of OH above; but here there is no H that can be lost(as HlD) from an adjacent carbon. Lewis acids may also be used,

"- "-$c=o: + AICI) i=! C-OAICI)e

/ /

and other cations, e.g. lDN02 in the nitration of benzene (p. 134),

The reaction is reversible, the reverse being the perhaps better knownacid-catalysed dehydration of alcohols (p. 247). Protonation can alsooccur on oxygen in a carbon-oxygen double bond,

[ JOH

....... 1+ 1- H" "- $ "-$ H,O "- /c=o i=! C=OH +-+ C-OH i=! C + H$

/ / / /"-OH

5.1.2 Addition of cations to neutral species

The most common cation is H lD , adding to unsaturated linkages, i.e.protonation, in for example the acid-catalysed hydration of alkenes(p. 187):

H H " HH" I H,O I -H I

-CH=CH-i=! -CH-CH-i=! -CH-CH-i=!-CH-CH-$ I I

$OHz OH

PhzCH-CI +='- PhzCH$Cle

MeOCHz-CI i=! MeOCHz$Cle

Carbocations and electron-deficient Nand 0 atoms

In each case a highly polar (high f), powerful ion-solvating medium isgenerally necessary. In a similar context the effect of AglD in catalysingreactions, often by a shift from SN2 -. SNI mode,

has alrea~y been referred to (p. 97). The catalytic effect of AglD canbe complIcated, however, by the fact that the precipitated silver halidemay itself act as a heterogeneous catalyst.

Ionisation may also be induced by Lewis acids, e.g. BF3'

to yie.ld in this case an acyl cation; the equilibrium here being considerably mfluenced by the very great stability of the anion, BF4 e. Alsowith AICI 3 ,

Ag$ + R-Br -+ AgBr,J.. + R$

here the relatively unstable acyl cation decomposes to yield the verystable Me3 ClD, the equilibrium being driven over to the right by theescape of CO.

Particularly striking examples are provided by the work of Olahwith SbFs as a Lewis acid, with either liquid S02 or excess SbFs assolvent,

102

leading to the formation of simple alkyl cations in conditions thatallow of their detailed study by n.m.r. spectroscopy and other means.The use of the same investigator's 'super acids', such as SbFs/FS03 H,allows of the formation of alkyl cations even from alkanes:

The relative stability of Me3 ClD is shown by the fact that under these

conditions the isomeric carbocation, Me~HCH2Me, obtained fromMeCH2CH2Me, was found to rearrange virtually instantaneously toMe3ClD. !he struct~re/stabilityrelationships, and rearrangements, ofcarbocahons are dIscussed below (pp. 104 and 109, respectively).

105

---+

Ell Ell

Me-O-CH 2 +-+ Me-O=CH 2

Ell Ell

CH 2 =CH-CH 2 +-+ CH 2-CH=CH 2

5.2 Stability and structure of carbocations

MeO MeO) Me<Y+

~ 0 SbF,/SO, QSbF,Cle4-*- --- ••+ '

_70' . .Ell 1

CH 2-CH 2H C-CH -CI H 2C-CH 22 2 U'

(4) (2) (3)

Such species with a bridging phenyl group are known as phen?niumions. The neighbouring group effect is even more p~o.nounced Wlt~ ~OH rather than an OMe substituent in the p-posltlOn. Solvol~~ls IS

found to occur ~ 106times more rapidly under comparable conditIons,and matters can be so arranged as to make possible the isolation of abridged intermediate (5), albeit not now a carbocation:

(5)

can involve 7t orbitals:

Thus the S 1 reactivity of allyl and benzyl halides has already beenreferred to,Nand the particular effectiveness of the lone yair on theoxygen atom above is reflected in the fact that MeOCH2Clis solvolysedat least 1014 times faster than CH3CI.

Stabilisation can also occur, again by delocalisation, through theoperation of a neighbouring group effect resulting in ~he .for~ation

of a 'bridged' carbocation. Thus the action of SbFs In lIqUid S02on p-MeOC6H4 CH 2CH2CI (2) results in the form~tion of .(3) ratherthan the expected cation (4), phenyl acting as a nelghbounng group(cf. pp. 93, 376):

Carbocations and electron-deficient Nand 0 atoms104

Ell Ell

[R-N=N +-+ R~=N] ---+ REil + N=Ni

and also by the use of a readily available carbocation to generateanother that is not so accessible (cf. p. 106):

H H

Ph,Co +0 .. Ph,C-H + 05.2 STABILITY AND STRUCTURE OF CARBOCATIONS

The simple alkyl carbocations have already been seen (p. 83) tofollow the stability sequence,

5.1.3 From other cations

Carbocations may be obtained from the decomposition of othercations, e.g. diazonium cations from the action of NaN02/HCI onRNH2 (cf. p. 119),

Me3CEIl > Me2CHEil > MeCH2

E1l > CH 3 E1l

resulting from increasing subs'titution of the cationic carbon atomleading to increasing delocalisation of the positive charge (withconsequent progressive stabilisation) by both inductive and hyperconjugative effects. The particular stability of Me3C

Ell is borne outby the fact that it may often be formed, under vigorous conditions,by the isomerisation of other first-formed carbocations (cf. p. 102),and also by the observation that it remained unchanged afterheating at 1700 in SbFs/FS03H for four weeks!

An essential requirement for such stabilisation is that the carbocation should be planar, for it is only in this configuration thateffective delocalisation can occur. Quantum mechanical calculationsfor simple alkyl cations do indeed suggest that the planar (Sp2)configuration is more stable than the pyramidal (Sp3) by =84 kJ(20 kcal) mol-1

. As planarity is departed from, or its attainmentinhibited, instability of the cation and consequent difficulty in itsformation increases very rapidly. This has already been seen in theextreme inertness of I-bromotriptycene (p. 87) to SNI attack, dueto inability to assume the planar configuration preventing formationof the carbocation. The expected planar structure of even simplecations has been confirmed by analysis of the n.m.I. and i.r. spectraof species such as Me3CEllSbF6e; they thus parallel the trialkylborons, R3B, with which they are isoelectronic.

A major factor influencing the stability of less simple cations isagain the possibility of delocalising the charge, particularly where this

Stabilisation, through delocalisation, can also occur througharomatisation. Thus l-bromocyclohepta-2,4,6-triene(tropylium bromide, 6),

1075.3 Carbocation reactions

Carbocations are found to undergo four basic types of reaction:(a) Combination with a nucleophile.(b) Elimination of a proton.(c) Addition to an unsaturated linkage.(d) Rearrangement of their structure.

The first two reaction types often lead to the formation of stable endproducts, but (c) and (d) lead to the formation of new carbocationsto which the whole spectrum of reaction types is still open. Most ofthese possibilities are neatly illustrated in the reaction of I-aminopropane (II) with sodium nitrite and dilute hydrochloric acid [thebehaviour of diazonium cations, e.g. (12), will be discussed furtherbelow, p. 119]:

Thus reaction of the I-propyl cation (13) with water (reaction type a)will yield propan-I-ol (14), elimination of a proton from (13) will yieldpropene (15, reaction type b), while rearrangement of (13, reactiontype d~in this case migration of He-will yield the 2-propyl cation(16). Type (b) reaction on this rearranged cation (16) will yield morepropene (15), while type (a) reaction with water will yield propan-2-ol(17). The product mixture obtained in a typical experiment was 7 %propan-I-ol, 28 %propene, and 32 %propan-2-ol : the relative proportions of propan-I-ol and propan-2-ol reflecting the relative stabilityof the two cations (13) and (16).

The sum of the above products still represents only 67 %conversionof the original I-aminopropane, however, and we have clearly not

5.3 CARBOCATION REACI10NS

an ion pair containing the parent cyclopropenyl cation (10) itself, asa white crystalline solid. 13C n.m.r. (cf. p. 48) has proved useful inthis field as the position of the signal from the +ve carbon correlateswith the electron density at this atom (ct. p. 393).


A~A yePr Pc Pc Pc

(8) (9) (10)

and the.la.tter.is found to be even more stable (=103 times) than (7)above: It IS still present as a carbocation to the extent of =50% inwater at pH 7! More recently it has also proved possible to isolate

(6)

which is isomeric with C6HsCH 2 Br, is found, unlike the latter compound, to be a crystalline solid (m.p. 208°) which is highly soluble inwater yielding bromide ions in solution, i.e. it has not the above covalentstructure but is an ion pair. The reason for this behaviour resides inthe fact that the cyclic cation (7) has 61te which can be accommodatedin three delocalised molecular orbitals spread over the seven atoms.It is. t~lUS a H.tickel 4n.+ 2 system (n = 1) like benzene (cf p. 17) andexhibits quasI-aromatic stability:

106

H~HH H

(7)

thus the planar carbocation is here stabilised by aromatisation. Theabove delocalised structure is confirmed by the fact that its n.m.r.spectrum exhibits only a single proton signal, i.e. all seven hydrogenatoms are equivalent. The effectiveness of such aromatic stabilisation is ~eflected in its being =1011 times more stable than the highlydelocahsed Ph3CEIl

• The generation of (7) by the action of Ph3CEIl oncycloheptatriene itself has already been referred to (p. 104).

A particularly interesting case of carbocation stabilisation occurswith Huckel 4n + 2 systems when n = 0, i.e. cyclic systems with 2'lTe(p. 18). Thus derivatives of 1,2,3-tripropylcyclopropene (8) arefound to yield ion pairs containing the corresponding cyclopropenylcation (9) extremely readily,

(18)

An explanation that would also account for the similar statisticalscrambling of the 13C label that is found to occur (over several hours)

1095.4 Carbocation rearrangements

This reflects the greater stability of a secondary rather than aprimary carbocation; shifts in the reverse direction can, however,take place where this makes available the greater delocalisationpossibilities of the 7T orbital system of a benzene ring (i.e. tertiary---.secondary):

Thus in the SNI solvolysis in EtOH of 3-chlorobut-l-ene (19), notone but a mixture of two isomeric ethers is obtained; and the samemixture (i.e. the same ethers in approximately the same proportion)

5.4.1.1 Allylic rearrangements

There are more interesting rearrangement possibilities inherent indelocalised cations, e.g. allylic rearrangements.

5.4.1 Without change in carbon skeleton

We have already seen one example of this type (p. 107), in which theI-propyl cation rearranged to the 2-propyl cation by the migrationof a hydrogen atom, with its electron pair (i.e. as He), from C2 tothe carbocationic C t , a 1,2-hydride shift:

H HI (II (II I

CH 3CHCH z ---. CH 3CHCH z

in the initial 2-propyl cation, CH313CH E9 CH 3' generated from

CH313CH(Cl)CH3 with SbFs at -60°.

The elimination reactions of carbocations (type b) will bediscussed in more detail subsequently (p. 248), but the rearrangementreactions (type d) are of sufficient interest and importance to meritfurther study now.

5.4 CARBOCATION REARRANGEMENTS

Despite the apparent confusion above, rearrangements involvingcarbocations may be usefully divided into those in which an actualchange in carbon skeleton itself does, or does not, take place; theformer are much the more important but the latter will first be brieflyreferred to.


exhausted the reaction possibilities. There are indeed other nucleophilespresent in the system, e.g. Cl e and N02e, capable of reacting witheither cation, (13) or (16), the latter nucleophile leading to the possibleformation of both RN02 and RONO (nitrite esters may also arisefrom direct esterification of first formed ROH). The cations (13) and(16) may also react with first formed ROH to yield ethers, ROR, orwith as yet unchanged RNH2to yield RNHR (which may itself undergofurther alkylation, or nitrosation cf. p. 121). Finally, either cationmay add to the double bond of first formed propene, MeCH=CH2

(reaction type c, c/. p.188), to yield further cations, Mel:H-CH2R,which can themselves undergo the whole gamut of reactions. Themixture of products actually obtained is considerably influenced bythe reaction conditions, but it will come as no surprise that thisreaction is seldom a satisfactory preparative method for the conversion:RNH2 -+ROH!

Reaction type (d) also complicates the Friedel-Crafts alkylation ofbenzene (a type clb reaction, p. 141) by 1-bromopropane, MeCHrCH2Br, in the presence of gallium bromide, GaBr3, as Lewis acidcatalyst. The attacking electrophile is here a highly polarised complex,6+ +RGaBri- -, and the greater stability of the complex in which R6 + +

carries its positive charge mainly on the secondary, rather than on a6+ + 6+ +

primary, carbon atom, i.e. Me2CHGaBri- - rather than MeCH2CHrGaBri- -, again results in a hydride shift (cf. above) so that the majorproduct of the reaction is Me2CHC6Hs.

That such rearrangements are not necessarily quite as simple asthey look, i.e. mere migration of He, is illustrated by the behaviour of13CH3CH2CH3 with AlBr 3, when the label is found to becomestatistically scrambled: the product consists of 2 parts 13CH3CH2CH3to 1 part CH313CH2CH3' as determined by analysis of the fragmentsproduced in the mass spectrometer. It is possible that the scramblingproceeds through the agency of a protonated cyclopropane intermediate (18):

108

is also obtained from the similar solvolysis of l-chlorobut-2-ene (20).

$ $

[MeCH-CH=CHz """ MeCH=CH-CHz] CIa

(23)

This clearly reflects formation of the same, delocalised allylic cation(23, ct.~. 105) as an ion pair intermediate from each halide, capable ofundergomg subsequent rapid nucleophilic attack by EtOH at eitherC t or C 3 :

111

(26)

(25)

(28)

(27)

l

5.4.2.1 Neopentyl rearrangements

(29)

(24)

Me Me MeI SNI I $ $ I

Me-C-CHMe -'--+ Me-C-CHMe -+ Me-C-CHMe ---+ ProductsI I I I

Me Dr Me Me

(30)

Me Me Me OH" / -He $ I H 20 I

C=C .- Me-C-CH -+ Me-C-CH Me/" I z I

Me H Me Me

Me Me MeI S I I $ Hoi

Me-C-CH Dr ~ Me-C-CH ~ Me-C-CH OHI z I z I

Me Me Me

The greater stability of the tertiary carbocation (28), compared withthe initial primary one (27), provides the driving force for the C-Cbond-breaking involved in migration of the Me group, with itselectron pair. Such changes in carbon skeleton-involvingcarbocations-are known collectively as Wagner-Meerwein rearrangements. Further confirmation of the involvement of (28) is thesimultaneous formation of the alkene, 2-methylbut-2-ene (29) byloss of proton: a product not obtainable from (27).

The possible occurrence of such major rearrangement of a compound's carbon skeleton, during the course of apparently unequivocalreactions, is clearly of the utmost significance in interpreting the resultsof experiments aimed at structure elucidation: particularly when theactual product is isomeric with the expected one. Some rearrangementsof this type are highly complex, e.g. in the field of natural productssuch as terpenes, and have often made the unambiguous elucidationof reaction pathways extremely difficult. The structure of reactionproducts should never be assumed but always confirmed as a routinemeasure: IH and l3C n.m.r. spectroscopy have proved of enormouSvalue in this respect.

It is interesting to note that while the neopentyl-type bromide (30)undergoes rearrangement during SNI hydrolysis, no such rearrangement takes place with its phenyl analogue (31):

2,2-dimethylpropanol (neopentyl alcohol, 25); a neopentyl rearrangement has taken place:


It is interesting that when EtOe , in fairly high concentration, isused as the nucleophile in preference to EtOH, the reaction of (19)becomes SN2 in type and yields only the one ether (21). Allylic rearrangeme~ts have been observed, however, in the course of displacement. reactIons that are proceeding by a bimolecular process. SuchreactIOns are referred to as SN2' and are believed to proceed:

R R() _~ I ~ I

Nu. CHz=CH-CH-CI -+ Nu-CHz-CH=CH + CIa

This process tends to occur when substituents (R) on the (X-carbona~om are bulky enough to reduce markedly the rate of direct SN2dIsplacement at Ca' A1lylic rearrangements are of quite commonoccurren~, but disentangli~g the detailed pathway by which theyproceed IS a matter of consIderable difficulty.

OEtI

MeCHCH=CHzCI (21)

M CIHCH EtOH EtOHe =CH z -- + +-- MeCH=CHCHzCI

(19) MeCH=CHCHzOEt (20)

(22)

110

5.4.2 With change in carbon skeleton

5.4.2.1 Neopentyl rearrangements

We ~ave already noticed (p. 86) that the SN2 hydrolysis of l-bromo2,.2-dlmethylpropane (neopentyl bromide, 24) is slow due to sterichmdrance. Carrying o~t the .reaction under conditions favouring theSNI mode can result 10 an Increased reaction rate but the productalcohol is found to be 2-methylbutan-2-01 (26) and not the expected

This reflects the greater stability of the benzylic cation (32), thoughonly secondary, compared with the tertiary cation (33) that wouldbe-but in fact is not-obtained by its rearrangement (cf. p. 105).

Me Me. I He I III

Me-C-CH=CH2 +:! Me-C-CH-CH]I I

Me Me

This relatively ready rearrangement can be a nuisance in thepreparative addition of acids, e.g. hydrogen halides (p. 184) to alkeneso~ in their acid-catalysed hydration (p. 187): mixed products that ar~difficult to separate may result or, in unfavourable cases, practically

113

(36)

~

(37)(35)

(34)

5.4.2.3 Pinacol/pinacolone rearrangements

5.4.2.3 Pinacol/pinacolone rearrangements

Another example of migration of a group, in the original case Me,to a cationic carbon atom occurs in the acid-catalysed rearrangement of 1,2-diols, e.g. pinacol (ct. p. 218) Me2C(OH)C(OH)Me2(34) to ketones, e.g. pinacolone, MeCOCMe3 (35):

Me Me MeI He I -H,O I III

MeC-CMe2 +:! MeC-CMe2 4 ~ MeC-CMe2I I I d I

HO OH HO \'H2 HO

The fact that a 1,2-shift of Me takes place in (36), which is alreadya tertiary carbocation, results from the extra stabilisation conferredon the rearranged carbocation (37) by delocalisation of char~ethrough an electron pair on the oxygen atom; (37) ~an also readIlylose a proton to yield a stable end-product (35). It mIght be expectedthat an analogous reaction would occur with othe~ c.ompoundscapable of forming the crucial carbocation (36): thIs IS, I~ fact,found to be the case. Thus the corresponding 1,2-bromohydnn (38)and 1,2-amino-alcohol (39) are found to yield pinacolone (35) when

Me [Me Me lI -He I III IMeC-CMe2~ MeC-CMe2 44 MeC-CMe2

II II Io o Ell HO:

/H

none of the desired product may be obtained. Further, addition ofcarbocations to initial, or product, alkenes may also take place(p. 188).

Rearrangement of di- and poly-alkylbenzenes also takes. placereadily in the presence of Lewis acid catalysts (p. 163), and III thedienone/phenol rearrangement (p. 115).

(33)(32)

1Products

(31)

Me Me MeI Ell 5N I I Ell Ell I

Me-C-CHPh --+ Me-C-CHPh -- Me-C-CHPhI I I I

Me Dr Me Me


~Me Me Me

" / -He Ell IC=C ~ Me-C-CH-CH/" I ]

Me CH3 Me

5.4.2.2 Rearrangement of hydrocarbons

Wagner-Meerwein type rearrangements are also encountered in thecracking of petroleum hydrocarbons when catalysts of a Lewis acidtype are used. These generate carbocations from the straight chainalkanes (ct. the isomerisation of l3C labelled propane, p. 108), whichthen tend to rearrange to yield branched-chain products. Fission alsotakes place, but this branching is important because the resultantalkanes cause less knocking in the cylinders of internal combustionengines than do their straight-chain isomers. It should be mentioned,however, that petroleum cracking can also be induced by catalysts thatpromote reaction via radical intermediates (p. 305).

Rearrangement of alkenes takes place readily in the presence ofacids:

115

0·30·7

~.R

(47)

1·0

ArI

PhC-C-PhII Io Ar

PhI

ArC-C-PhII I

/

0 Ar

(44)

Ph Ph" $/C-C

/1 "Ar ArHO

ooR R

(46)

~12) -H,O

15·7

5.4.2.3 Pinacollpinacolone rearrangements

500

Ph Ph" /C-C

A/ I I"ArHO OH

(43)

(45)

i.e. by determining the relative proportions of the two ketones (44)and (45) that are produced, the relative migratory aptitude sequence:

This particular problem can be avoided by chosing symmetrical1,2-diols such as PhArC(OH)C(OH)PhAr (43) and it has been possibleto establish by experiments on such compounds,

This sequence could be interpreted (except for o-MeOC6 H4 ) in termsofdecreasing potential electron-donation in the group that is migrating,with its electron pair, to a positive centre, the cationic carbon atom.A similar simple theory of potential electron release could alsoaccount for the observed alkyl group sequence mentioned above.The o-MeOC6H4 group, despite its being electron-donating, isactually found to migrate slower than C6Hs, and there is evidencethat both the relative crowding of possible alternative transitionstates, and the conformation adopted by the starting material atreaction (see below) are also of considerable importance. Thesemay, as with o-MeOC6H4 above, outweigh electronic effects.

A rearrangement essentially akin to a reversal of the pinacol/pinacolone change, a retro pinacol reaction, is the dienone/phenolrearrangement,

(35)(36)

t(38)

(39)


treated with Agel and NaN02 /HCI, respectively:

A.num~er of experiments have been carried out to determine therelatIVe migratory ~ptitude of groups i.n pinacol/pinacolone type rearrangements, and In general the relatIve ease of migration is foundto be:

Pb PhI Hoi $

PhC-CMe2~ PhC-CMeI GI I 2

HO OH 2 HO81

(42)

Ph > Me3C > MeCH2 > Me

!t shoul~ be r~alised that there are considerable difficulties involvedIn choOSIng sUitable models for such experiments, and in interpretingthe r~sults when we have got them. Thus in the rearrangement of the1,2-dlOl, Ph2 C(OH)C(OH)Me2 (40), it is Me that is found to migrateand not C6 Hs as m~ght.have been expected from the sequence above.However, the reactIOn IS here controlled by preferential protonation~n that OH group which will lead to the more stable initial carbocatlOn (41 rather than 42), and the migration of Me rather than Ph isthereby predetermined:

in which protonation of the initial dienone (46) allows reattainment ofthe wholly aromatic condition (47) through 1,2-migration of an alkylgroup:

117

1~ 1)rI?' "R"

R" tNu:

I-N2

R*1"'---'"

MeCH-CHz

(~==NEll

-

R"-Nu

5.4.2.4 Stereochemistry of rearrangements

R'

R*I NaNo,

MeCH-CHz~INHz

(+)

R* R*I H 20 Ell I

MeCH-CHz~ MeCH-CHzI

OHfRO = Me2CHCH2CHMe)

I

The inversion is often found to be almost complete in cyclic compounds where rotation about the C1-C2 bond is largely prevented,but also to a considerable extent in acyclic compounds. This featurecould be explained on the basis of a 'bridged' intermediate (cf.

in which the chiral R* group does indeed migrate with retention ofits configuration.

On the other two points, the evidence supports predominantinversion of configuration at both migration origin (a) and terminus(b ):

Similarly, if rearrangements in which there is a hydride shift (cf. p. 109)are carried out in a deuteriated solvent (e.g. 0 20, MeOD, etc.), nodeuterium is incorporated into the new C-H(D) bond in the finalrearranged product. In both cases the rearrangement is thus strictlyintramolecular, i.e. the migrating group does not become detachedfrom the rest of the molecule, as opposed to intermolecular where itdoes.

This suggests very close association of the migrating group, R, withthe migration terminus before it has completely severed its connectionwith the migration origin; we should thus expect no opportunity forits configuration to change, i.e. retention of configuration in a chiralR* group. This has been confirmed in the reaction below (cf. p.118),

01.R

(47)

_He--+6~~:

R R R

EllOHo

R R

Me Me EtI I I

Ph1C-CMe Ph1C-TiMe Ph1C-TiMeI IHO OH 0 0

(48)NOHe or

+ --+ +Et Et MeI I I

Ph1C-CEt Ph1C-CEt Ph1C-TiEtI I"HO OH 0 0

(49)

Hep


ooR R

(46)

5.4.2.4 Stereochemistry of rearrangements

116

There are essentially three points of major stereochemical interestin carbocationic rearrangements: what happens to the configurationat the carbon atom from which migration takes place (the migrationorigin), to the configuration at the carbon atom to which migrationtakes place (the cationic carbon atom, the migration terminus), andto the configuration of the migrating group, if that is chiral, e.g.PhMeCH. Interestingly enough, these three questions have neverbeen answered for one and the same compound, despite the enormous body of work that has now been done on carbocations.

It has been shown that the migrating group does not become freeduring the rearrangement by, for example, taking two pinacols (48and 49) that are very similar in structure (and that rearrange at verymuch the same rate) but that have different migrating groups, andrearranging them simultaneously in the same solution (a crossoverexperiment): no cross migration is ever observed:

(S2c) (S2d) (SId)

The ratio of inversion (Slab) to retention (SId) in the productketone would then be determined by the relative rate of rotationabout Cl-~ in (52c) compared with the rate of migration of Ph.

It was actually found, however, that though inversion was predominant (Slab: 88%), the product ketone contained a significantamount of the mirror image (SId: 12%): thus 12% of the totalreaction can not have proceeded via a bridged carbocation. Thesimplest explanation is that part at least of the total rearrangementis proceeding via a non-bridged carbocation (52c), in which somerotation about the Cc -(; bond can take place (52c ~ 52d),thereby yielding ketone (SId) in which the original configurationhas been retained [cf. (50a) or (SOb) with (SId)]:

1195.5 Diazonium cations

H As,oL -N,

I H,ORCH-C=O +- RCH=C=O

IOH (54)

As well as in water, the reaction can be carried out in ammonia or inan alcohol when addition again takes place across the C=C bond ofthe ketene to yield an amide or an ester, respectively, of the homologousacid.

The Wolff rearrangement has a close formal resemblance to theHofmann and related reactions (p. 122), in which migration takesplace to an electron-deficient nitrogen atom to form an isocyanate,RN=C=O, intermediate.

o 0 0

R~-OH SOCl', R~-C1 CH,N" R~-CHN2(53)

5.4.2.5 Wolff rearrangements

This rearrangement has been separated from carbocationic rearrangements proper as it involves migration to an uncharged, albeitelectron-deficient, carbene-Iike carbon (ct. p. 266) atom rather thanto a positively charged one. The reaction involves the loss ofnitrogen from a-diazoketones (53), and rearrangement to highlyreactive ketenes (54):

R R R" e (/l -N, ,,---.. /.,f'C-CH-Jl=N -+ .,f'C1:,C;H -+ O=C=C"

o (53) 0 (55) (54) H

The ketenes will then react readily with any nucleophiles present inthe system, e.g. H 2 0 below. The reaction can be brought about byphotolysis, thermolysis, or by treatment with silver oxide. In the firsttwo cases an actual carbene intermediate (55) is probably formed asshown above, in the silver catalysed reaction loss of nitrogen andmigration of R may be more or less simultaneous. In the case whereR is chiral, e.g. C4 H9 C*MePh, it has been shown to migrate withretention of its configuration (ct. p. 117).

Diazoketones (53) may be obtained by the reaction ofdiazomethane,CH 2 N 2 , on acid chlorides, and a subsequent Wolff rearrangement inthe presence of water is of importance because it constitutes part ofthe Arndt-Eistert procedure, by which an acid may be converted intoits homologue:

5.5 DIAZONIUM CATIONSThe nitrosation of primary amines, RNH 2 , with, for example, sodiumnitrite and dilute acid (ct. p. 107) leads to the formation of diazonium

Ph

o IiMeYH(-)'Ph

R

R: ,R"~rR' 'R"

Nu

-R.$...•

J:4"R"R'T R"

Nu:

-


Me H Ph Me Hn NaNo,

~~ ~2-- o --NH2HO Me H HO

HO Ph (-) (+) Ph OH (-)Ph

(SOu) (Slab) (SOb)

Me H

0H- 0Me--HO Ph HO Ph

bromium ion, p. 180) or transition state:

118

Actual bridging during rearrangement is not, however, by anymeans universal even when the migrating group is C6Hs, whose 7T

orbital system might well be expected to assist in the stabilisation ofa bridged carbocation through delocalisation (ct. p. 105).

This is clearly demonstrated in the pinacolinic deamination (ct. p.114) of an optically active form of the amino-alcohol (50). Suchreactions proceed from a conformation (antiperiplanar; 50a or SOb)in which the migrating (Ph) and leaving (NH2 : as N2 ; ct. p. 114)groups are TRANS to each other. Rearrangement via a bridgedcarbocation would necessarily lead to 100% inversion at themigration terminus in the product ketone (Slab), whichever initialconformation, (50a) or (SOb), was involved:

1215.5 Diazonium cations

aromatic nucleus:

R2 N-@-N=0

As might be expected, substituents in the aromatic nucleus have amarked effect on the stability of ArN2 al, electron-donating groupshaving a marked stabilising effect:

Because primary aromatic amines are weaker bases/nucleophilesthan aliphatic (due to interaction of the electron pair on N with then orbital system of the aromatic nucleus), a fairly powerful nitrosatingagent is required, and the reaction is thus carried out at relativelyhi~ acidity. Sufficient equilibrium concentration of unprotonated

ArNH2 remains (as it is a weak base), but the concentration is lowenough to prevent as yet undiazotised amine undergoing a couplingreaction with the first formed ArN 2 al (cf. p.147). Aromatic diazoniumchlorides, sUlphates, nitrates, etc., are reasonably stable in aqueoussolution at room temperature or below, but cannot readily be isolatedwithout decomposition. Fluoroborates, ArN2 alBF4

8 , are more stable(cf. stabilising effect of BF4 8 on other ion pairs, p. 136) and can beisolated in the dry solid state: thermolysis of the dry solid is an importantpreparative method for fluoroarenes:

Nitrosation also occurs with secondary amines but stops at thestable N-nitroso stage, R2N-N=O. Tertiary aliphatic aminesare converted initiaIly into the nitrosotrialkylammonium cation,

alR3 N-N=O, but this then readily undergoes C-N fission to yieldrelatively complex products. With aromatic tertiary amines, ArNR 2 ,

nitrosation can take place not on N but at the activated p-positionof the nucleus (cf p. 137) to yield a C-nitroso compound:

tl..

RN=N-OH

HI

RN-N=O

Hj$ -H$

RN-N=O ---+I

H xe

[

R-N=Nj! 4 (I) +H$

Ell (2) -H,o

R-Cr::::N(56)

HI~

RN: N=O----'"I hH X,.I


cations (56):

The instability of aliphatic diazonium cations, in the absence ofany stabilising structural feature, is due very largely to the effectivenessof N 2 as a leaving group; in aromatic diazonium cations, however,such a stabilising feature is provided by the n orbital system of the

If R contains a powerful electron-withdrawing group, however, lossof Hal-rather than loss of N2---ean take place to yield a substituteddiazoalkane, e.g. ethyl aminoacetate -+ ethyl diazoacetate:

120

The effective nitrosating agent is probably never HN02 itself; atrelatively low acidity it is thought to be N 20 3 (X = ONO) obtainedby,

2HN02 ~ ONO-NO + H20

while as the acidity is increased this is replaced by the more effective

species, protonated nitrous acid Hl>-NO (X = H 20), and finaIlyby the nitrosonium ion, alNO (cf. p. 137). A compromise has to bestruck in nitrosation, however, between an increasingly effectivenitrosating agent as the acidity of the solution is increased, and

decreasing [RNH2], as the amine becomes increasingly protonated andso rendered unreactive.

With simple aliphatic amines, the initial diazonium cations (56)will break down extremely readily to yield carbocations (ct. p. 107)which are, for reasons that are not wholly clear, markedly morereactive than those obtained from other fission processes, e.g.RBr -. R$Br8

. Where the prime purpose is the formation of carbocations, the nitrosation is better carried out on a derivative of theamine (to avoid formation of H 20) under anhydrous conditions:

5.6 MIGRAnON TO ELECTRON-DEFICIENT N

5.6.1 Hofmann, Curtius, Lossen and Schmidt reactions

A typical example is the conversion of an amide (57). to an amine (~8),

containing one carbon less, by the action of alkalIne hypobromIte,the Hofmann reaction:

The rearrangements that we have considered to date all hav~ 0!lefeature in common: the migration of an alkyl or aryl group, wIth Itselectron pair, to a carbon atom which, whether it be a carbocationor not, is electron-deficient. Another atom that can similarly. ~ecome electron-deficient is nitrogen in, for example, R2NEJ) or RN (anitrene, cf. carbenes above), and it might be expected that alkylor aryl migration to such centres would take place, just as it did toR3Cal and R2~ ; this is indeed found to be the case.

1235.6.2 Beckmann rearrangements

Lassen:

Curtius:

Schmidt:

the rearrangement is strictly intramolecular, and it is further foundthat when R is chiral, e.g. C6HsMeCH, it migrates with its configuration unchanged.

There are a· group of reactions very closely related to that ofHofmann, all of which involve the formation of an isocyanate (61)by rearrangement of an intermediate analogous to (60):

R R" eOH.,/"

HNb~~R~ --+ ~~:~R~(63) (64) ~

R-N=C=O

(61)

R" /C=O

HN'H2 ~.NO, RHa~ ~

(65) .,/ 'C=O

R Nt

"C=O~~=N/ fIl

HO (67)

(66)

The Lossen reaction involves the action of base on O-acyl derivatives(63) of hydroxamic acids, RCONHOH, and involves R'C02 a as theleaving group from the intermediate (64), as compared with Bra from(60). The reaction also occurs with the hydroxamic acids themselves,but not as well as with their O-acyl derivatives as R'C02 a is a betterleaving group than aOH. The concerted nature of the rearrangementis supported by the fact that not only is the reaction facilitated bye1ectron-donating substituents in R (cf Hofmann), but also by e1ectronwithdrawing substituents in R', i.e. both are involved in the rate-limiting step of the reaction. --

The Curtius and Schmidt reactions both involve N2 as the leavinggroup from the azide intermediate (67), and here again the migrationof R occurs in a concerted process. The azide may be obtained eitherby nitrosation of an acid hydrazide (65)-Curtius reaction-or by thereaction of hydrazoic acid, HN3 , on a carboxylic acid (66)-theSchmidt reaction.


The most famous of the rearrangements in which R migrates fromcarbon to nitrogen is undoubtedly the conversion of ketoximes to

t

R-N=C=O

(61)

e (JJ

R-N-C=O

(57)

R R R" "eOHJ}..c=o BrO

e• c=o --+ c=o --+

/ / /NH 2 N~ N~

Br eBr

(59) (60)

HeOH I H,O

RNH 2 + HC03e +- RN-C=O +-I

(58) OH

(62)


A salient intermediate in this reaction is the isocyanate (61), corresponding closely to the ketene intermediate in the Wolff reaction (p.117); this, too, then undergoes addition of water, but the resultantcarbamic acid (62) is unstable and decarboxylates readily to yieldthe amine (58). By careful control of conditions it is possibleactually to isolate N-bromoamide (59), its anion (60), and isocyanate (61) as intermediates: the suggested reaction pathway is thusunusually well documented. The rate-limiting step is probably lossof Bre from (60), and the question arises whether this loss isconcerted with the migration of R, or whether a carbonylnitreneintermediate, RCON, is formed, which then rearranges. The factthat the rearrangement of ArCONH2 is speeded up when Arcontains electron-donating substituents (cf. the pinacol/pinacolonerearrangement, p. 115), and that the formation of hydroxamic acids,RCONHOH (that would be expected from attack of solvent H2 0 onRCON), has never been detected, both support a concertedmechanism. Crossover experiments lead to no mixed products, i.e.

122

125

(75)(76)

HO" -HeC=NR +--

/R'


(77)

o~

C-NHR <::::::;-/

R'

In strong acid the rearrangement involves O-protonation to yield (73a)followed by loss of water to (74), while with acid chlorides, PCIs , etc.,

R",-----,.

C=N'R He~ / CA

" ~ R' ~H2"_H,0C=N" (730) ~ $

R/ "OH R R'C=NR

""'RCOCI "'c=N' / (74)

RS?,C1~ / CA ~oxe jele. R' OX H,O:

(73b)

That simple interchange of R and OR does not take place hasbeen demonstrated by carrying out the rearrangement of benzophenone oxime, Ph2C-NOR, to benzanilide, PhCONRPh, inR 2

180. Provided that neither the initial oxime nor the productanilide exchanges its oxygen for 180 when dissolved in R 2

111Q-ashas been confirmed-direct, intramolecular interchange of Ph andOR would result in the incorporation of no 180 in the rearrangedproduct. In fact, the product benzanilide is found to contain thesame proportion of 180 as did the original water, so the rearrangement must involve loss of the oxime OR group and subsequentre-introduction of oxygen from a water molecule. The main functionof the acidic catalyst is, indeed, to convert the OR of the oxime intoa better leaving group by protonation, esterification, etc.

The rearrangement is believed to proceed as follows:

Ar Me Ar Me" / ?i " / ~C TIII -+ ArC-NHMe -+ MeC-NHArN. .N

/. ."HO (68) (71) (69) OH (72)

oxime pairs whose configuration has already been established. Onceit had been clearly demonstrated that the anti-R group migrates inBeckmann rearrangements, however, the structure of the amideproduced is quite often used to establish the configuration of theoxime from which it was derived. Thus, as expected, (68) is found toyield only the N-methyl substituted benzamide (71), while (69)yields only the aryl substituted acetanilide (72):

(i.e. R'CONHR only)

(70)

o R'~/

CI

N

/ "R H

eOH-+cold

HO R'" /CII 4.N./.

R


R R'" /CII Acid.

.N."OH

The reaction is catalysed by a wide variety of acidic reagents, e.g.H2S04 , S03' SOCI2 , P 20 S ' PCIs , BF3, etc., and takes place notonly with ketoximes themselves but also with their O-esters. Only avery few aldoximes rearrange under these conditions, but more canbe made to do so by use of polyphosphoric acid as catalyst. Perhaps themost interesting feature of the rearrangement is that, unlike those wehave already considered, it is not the nature (e.g. relative electronreleasing ability) but the stereochemicalarrangement of the Rand R'groups that determines which of them migrates. Almost withoutexception it is found to be the R group anti to the OH group thatmigrates C ----+ N :

(69)

N-substituted amides, the Beckmann rearrangement:

RR'C=NOH -+ R'CONHR or RCONHR'

Confirmation that this is indeed the case requires an initial, unambiguous assignment of configuration to a pair of oximes. This waseffected by working with the pair of oximes (68) and (69}-one of themcyclised to the benzisoxazole (70) with base even in the cold, whilethe other was little affected even under much more vigorous conditions.The one undergoing easy cyclisation was, on this basis, assigned theconfiguration (68), in which the nuclear carbon carrying Br and theo of the OR group (08 in base) that attacks it are close together:

In (69) these atoms are far apart, and can be brought within reactingdistance only by cleavage of the C=N bond in the oxime.

Subsequently, configuration may be assigned to other pairs ofketoximes by correlation of their physical constants with those of

124

the intermediate ester (73b) is formed; the anion XOe constitutes agood leaving group so that, again, (74) is obtained. A number ofintermediate esters (73b) have indeed been prepared independentlyand shown to undergo rearrangement to the expected amides, in theabsence of added catalysts and in neutral solvents. The stronger theacid XOH is, i.e. the more capable the anion is of independent existence,the better leaving group XOe should be and hence the faster therearrangement should occur. This is observed in the series where XOe

is CH 3C02 e < CICH 2C02 e < PhS03 e. That such ionisation is therate-limiting step in the rearrangement is also suggested by theobservation that the rate of reaction increases as the solvent polarityincreases.

As with- the rearrangements we have discussed previously, loss ofleaving group and migration of R are believed to proceed essentiallysimultaneously in the conversion of (73) into (74). This is borne outby the strict intramolecularity of the reaction (no crossover products,cf. p. 116), the high stereoselectivity already referred to (i.e. only R,not R' migrates), and the fact that R, if chiral, e.g. PhCHMe, retainsits configuration on migration. This probably also reflects the greater

Ellstability of the cationic intermediate, R'C=NR (74, which has beendetected by n.m.I. spectroscopy), compared with the oneRR'C=NEIl

, that would be obtained if loss of leaving group precededthe mi~ation of R. The rearrangement is completed by attack ofwater (it is, of course, at this stage that 180 is introduced in therearrangement of benzophenone oxime in H/80 referred to above)on the cationic carbon atom of (74) to yield (75), followed by loss ofproton to form the enol (76) of the product amide (77).

The stereochemical use ofthe Beckmann rearrangement in assigningconfiguration to ketoximes has already been referred to, and it alsohas a large-scale application in the synthesis of the textile polymerNylon-6 from cyclohexanone oxime (78) via the cyclic amide (lactam,79):

127

l-R'co,eO-H

~ 1;/~R-C$

IOR

(82)

oII

R-CI

OR

(83)

OH OH~ (I) R'CO,H I

R-~-R (2) -H~ I R-C-(R

R'cQT 6 (81)IIV'o

5.7 Migration to electron-deficient 0

oII

R-C-R

(80)

Cyclic ketones are converted into lactones (cyclic esters):

o 0II H,O, II

R-C-R~ R-C-OR

5.7 MIGRATION TO ELECTRON-DEFICIENT 0

It might reasonably be expected that similar rearrangements wouldalso occur in which the migration terminus was an electron-deficientoxygen atom: such rearrangements are indeed known.

The reaction is believed to proceed as follows:

5.7.1 Baeyer-ViUiger oxidation of ketones

Oxidation of ketones with hydrogen peroxide or with a peroxyacid,RCOzOH (cf. p. 330) results in. their conversion into esters:

Initial protonation of the ketone (80) is followed by addition of theperacid to yield the adduct (81), loss of the good leaving group R'C02 eand migration of R to the resulting electron-deficient oxygen atomyields (82), the protonated form of the ester (83). The above mechanismis supported by the fact that oxidation of Ph2C=180 yields onlyPhC 180·OPh, i.e. there is no 'scrambling' of the 180 label in theproduct ester. That loss of R'C02e and migration of R are concerted is(78)(79)

!NH,OH

o~cSH(0 )II ~

-NH(CH 2),C- • +-Base

Nylon-6


5.7.2 Hydroperoxide rearrangements

A very similar rearrangement takes place during the acid-catalyseddecomposition of hydroperoxides, RO-OH, where R is a secondary ortertiary carbon atom carrying alkyl or aryl groups. A good example isthe decomposition of the hydroperoxide (84) obtained by the airoxidation of cumene [(l-methylethyl)benzene]; this is used on thelarge scale for the preparation of phenol and acetone:

supported by the fact that the reaction is speeded up by electronwithdrawing substituents in R' of the leaving group, and by electrondonating substituents in the migrating group R: the concerted conversion of (81) into (82) thus appears to be the rate-limiting stepof the reaction. Further, a chiral R is found to migrate with its configuration unchanged. When an unsymmetrical ketone, ReOR', isoxidised either group could migrate, but it is found in practice that itis normally the more nucleophilic group, i.e. the group better able tostabilise a positive charge, that actually migrates, cf the pinacol/pinacolone rearrangement (p.115). As with the latter reaction, however,steric effects can also play a part, and may occasionally changemarkedly the expected order of migratory aptitude based on electronreleasing ability alone.

group are found to increase the rate of reaction, and to promote themigratory ability of a particular group with respect to its unsubstitutedanalogue. It may be that the superior migratory aptitude of Ph aboveresults from its migrating via a bridged transition state:

1295.7.2 Hydroperoxide rearrangements

[

Me JtHIH ..C-MePb:' I H···H......OH 2

Intermediates of the general form of (86) have been detected insuper acid solution (ct. p. 102), and their structure confirmed byn.m.r. spectroscopy.

In these examples we have been considering the essentially heterolytic fission of peroxide linkages, -0:0- ---+ -0$:08 -, in polarsolvents; homolytic fission can also occur, under suitable conditions,to yield radicals, -0:0- ---+ -0' '0-, as we shall see below (p. 304).

Carbocations and electron -deficient Nand 0 atoms128

MeI H'"

Ph-C-Me PIO-OH

Me MeI -H ° IPh-C-Me ----.l.......+ EIlC-Me

) I Ell I0"G0H2 PhO

(84) (85) (86)

1(1) H 20(2) -H'"

Me MeI H'"/H,O I

O=C-Me~ HO-C-MeI

+ PbO

PbOH (87)

Here again loss of the leaving group (H20), and migration of Ph tothe resulting electron-deficient oxygen atom in (85), are almost certainlyconcerted. Addition of water to the carbocation (86) yields thehemi-ketal (87), which undergoes ready hydrolysis under the reactionconditions to yield phenol and acetone. It will be observed that Phmigrates in preference to Me in (85) as, from previous experience, wewould have expected. Electron-donating substituents in a migrating

6Electrophilic and nucleophilicsubstitution in aromatic systems

6.1 ELECrROPHIUC ATIACK ON BENZENE, p. 1316.1.1 7T and u complexes, p. 131.

6.2 NITRATION, p. 133.6.3 HALOGENATION, p. 138.6.4 SULPHONATION, p. 140.6.5 FRIEDEL--eRAFTS REACTIONS, p. 141:

6.5.1 Alkylation, p. 141; 6.5.2 Acylation, p. 143.6.6 DIAZO COUPLING, p. 146.6.7 ELEcrROPHll.JC ATTACK ON C6H sY, p. 150:

6.7.1 Electronic effects of Y, p. 151: 6.7.1.1 Y=~R3' CCh,N02 , CHO, C02H, etc., p. 151,6.7.1.2 Y = Alkyl, phenyl, p. 152,6.7.1.3 Y=OCOR, NHCOR, OR, OH, NH2 , NR2 , p. 153,6.7.1.4Y = Cl, Br, I, p. 155; 6.7.2 Partial rate factors and selectivity,p. 156; 6.7.3 o-Ip-Ratios, p. 159; 6.7.4 Ipso substitution, p. 161.

6.8 KINETIC V. THERMODYNAMIC CONTROL, p. 163.6.9 ELEcrROPHll.JC SUBSTITUTION OF OTHER AROMATIC SPECIES,

p.l64.6.10 NUCLEOPHll.JC ATTACK ON AROMATIC SPECIES, p. 167:

6.10.1 Substitution of hydrogen, p. 167; 6.10.2 Substitution ofatoms other than hydrogen, p. 169; 6.10.3 'Substitution' via aryneintermediates p. 173.

Reference has already been made to the structure of benzene and, inparticular, to its delocalised 1T orbitals (p. 15); the concentration ofnegative charge above and below the plane of the ring-carbon atomsis thus benzene's most accessible feature:

6.1 Electrophilic attack on benzene 131

This concentration of charge might be expected to shield the ringcarbon atoms from the attack of nucleophilic reagents and, by contrast,to promote attack by cations, XEe, or electron-deficient species, i.e.by electrophilic reagents; this is indeed found to be the case.

6.1 ELECfROPHILIC ArrACK ON BENZENE

6.1.1 'IT and a complexes

It might be expected that the first phase of reaction would be interaction between the approaching 'electrophile and the delocalised 1C

orbitals and, in fact, so-called 1C complexes such as (1) are formed:

(I)

Thus methylbenzene (toluene) forms a 1: 1 complex with hydrogenchloride at -78°, the reaction being readily reversible. That no actualbond is formed between a ring-carbon atom and the proton from HCIis confirmed by repeating the reaction with DCI; this also yields a1C complex, but its formation and decomposition do not lead to theexchange of deuterium with any of the hydrogen atoms of the nucleus,showing that no C-D bond has been formed in the complex. Aromatichydrocarbons have also been shown to form 1C complexes with speciessuch as the halogens, AgEe, and, better known, with picric acid, 2,4,6(02NhC6H20H, to form stable coloured crystalline adducts whosemelting points may be used to characterise the hydrocarbons. Theseadducts are also known as charge transfer complexes. In the complexthat benzene forms with bromine, it has been shown that the halogenmolecule is located centrally, and at right angles to the plane of thebenzene ring.

In the presence of a compound having an electron-deficient orbitale.g. a Lewis acid such as AICI 3 , a different complex is formed, however.If DCI is now employed in place of HCI, rapid exchange of deuteriumwith the hydrogen atoms ofthe nucleus is found to take place indicatingthe formation of a (T complex* (2), also called a Wheland intermediate(cf. p. 41), in which H Ee or DEe, as the case may be, has actually becomecovalently bonded to a ring-carbon atom. The positive charge isshared over the remaining five carbon atoms of the nucleus via the1C orbitals and the deuterium and hydrogen atoms are in a plane at

* These species are also referred to as arenium or arenonium ions, as well as by themore general term of carbocation intermediate.

(2)

XLJ

1336.2 Nitration

as has indeed been implied by writing its structure as (2). The use of(2) should not, however, be taken to imply a uniform d.istrib~tion ofelectron density in the ion-that this could not be so IS pIam whenthe separate canonical structures (2a, 2b and 2c, p. 132) contribut-ing to (2) are written out. . . ..

Ifwe are correct in our assumption that the electrophIllc substItutIOnof aromatic species involves such (T complexes as intermediates-andit has proved possible actually to isolate them in the course of somesuch substitutions (p. 136)-then what we commonly refer to asaromatic 'substitution' really involves initial addition followed bysubsequent elimination. How this basic theory is borne ,?ut in thecommon e1ectrophilic substitution reactions of benzene wIll now beconsidered.

6.2 NITRATION

The aromatic substitution reaction that has received by far the closeststudy is nitration and, as a result, it is the one th~t pr?ba~ly p!ovidesthe most detailed mechanistic picture. PreparatIve nItratIOn IS mostfrequently carried out with a mixture of concentrated nitric andsulphuric acids, the so-called nitrating mixture. The 'classical' explanation for the presence of the sulphuric acid is that it absorbs the waterformed in the nitration proper

It might perhaps be expected that conversion of benzene into the(T complex (2), which has forfeited its aromatic stabilisation, wouldinvolve the expenditure of a considerable amount of energy, i.e. thatthe activation energy for the process would be high and the reactionrate correspondingly low: in fact, many aromatic electrophilic substitutions are found to proceed quite rapidly at room temperature.This is because there are two factors operating in (2) that serve toreduce the energy barrier that has to be surmounted in order to effectits formation: first, the energy liberated by the complete formationof the new bond to the attacking electrophile, and, second, the factthat the positively charged (T complex can stabilise itself, i.e. lower itsenergy level, by delocalisation

H D H D D

6<' AICI,S (:) AICI,S @- -(+CISj ............ (-He)

(3) (2) (4)

Addition Substitution

H D H D H D

66 +--+ 0 +--+ -0 = AICI.e:····.1····:

ED

(2a) (2b) (2c) (2)

right angles to that of the ring in the carbocation intermediate:

132 Electrophilic and nucleophilic substitution in aromatic systems

That the 7[ and (T complexes with, e.g. methylbenzene and HCl, reallyare different from each other is confirmed by their differing behaviour.Thus formation of the former leads to a solution that is a non-conductorof electricity, to no colour change, and to but little difference in U.v.spectrum, indicating that there has been little disturbance of electrondistribution in the original methylbenzene ; while if AlCl 3 is presentthe solution becomes green, will conduct electricity and the u.v.spectrum of the original methylbenzene is modified, indicating theformation of a complex such as (2) as there is no evidence that aluminium chloride forms complexes of the type, HB Al04e.

The reaction may be completed by AlCl4 e removing a protonfrom the (T complex (2) -+ (4). This can lead only to exchange ofhydrogen atoms when HCI is employed but to some substitution ofhydrogen by deuterium with DCI, i.e. the overall process is electrophilicsubstitution. In theory, (2) could, as an alternative, react by removingCle from AlCI4 e resulting in an overall electrophilic addition reaction(2) -+ (3) as happens with a simple carbon-earbon double bond(p. 181); but this would result in permanent loss of the stabilisationconferred on the molecule by the presence of delocalised 7[ orbitalsinvolving all six carbon atoms of the nucleus, so that the product, anaddition compound, would no longer be aromatic with all that implies.By expelling WB, i.e. by undergoing overall substitution rather thanaddition, the completely filled, delocalised 1T orbitals are reattainedin the product (4) and characteristic aromatic stability recovered:

The gain in stabilisation in going from (2) -+ (4) helps to providethe energy required to break the strong C-H bond that expulsionof H(!) necessitates; in the reaction of, for example, HCl with alkenes(p. 184) there is no such factor promoting substitution and additionreactions are therefore the rule.

(5)

in which C-N02 bond-formation and C-H bond-breaking areoccurring simultaneously; or two-step pathways [2] involving aWheland intermediate or (J complex (6),

135

.... [2J

..... [IJ

6.2 Nitration

H H N02 NO

©+"No,~6 ~@+H'(6)

in which either step (a)- C-N02 bond-formation-or step (b)C-H bond-breaking-eould be slow and rate-limiting. The C-Hbond must, of course, be broken at some stage in all three of the abovepathways, but a partial distinction between them is that it must bebroken in the slow, rate-limiting step in [1] (only one step, anyway)and in [2b], but not in [2a]. If the C-H bond is, in fact, broken in arate-limiting step, then the reaction will exhibit a primary kinetic

the overall nitration reaction. That EIlN02, once formed, is a highlyeffective nitrating agent is borne out by the rapid nitration that maybe effected, of even relatively unreactive aromatic species, by thesalt EIlN02BF4e at room temperature or below.

Many nitration reactions with nitrating mixture are, however,found to follow an 'idealised' rate equation of the form,

Rate = k[Ar-HJ[ EIlN02J

but, in practice, the actual kinetics are not always easy to follow or tointerpret for a variety of reasons. Thus the solubility of, for example,benzene itself in nitrating mixture is sufficiently low for the rate ofnitration to be governed by the rate at which the immiscible hydrocarbon dissolves in the acid layer. With nitrating mixture, [EIlN02] isrelated directly to [HN03] added, as HN03 is converted rapidly andcompletely into EIlN02 , but with nitrations in other solvents complexequilibria may be set up. The relation of the concentration of theeffective electrophile (nearly always N02Ell) to the concentration ofHN0

3, or other potential nitrating agent, actually added may then

be far from simple.The above general, idealised rate law is compatible with at least

three different potential pathways for nitration: one-step, concertedpathway [1] that involves a single transition state (5),

~nd so p~events the reverse reaction from proceeding. This explanationIS unsatIsfactory in a number of respects, not least in that nitrobenz~~e, once formed, appears not to be attacked by water under thecondItIOns of the reaction! What is certain is that nitration is slow inthe absence of sulphuric acid, yet sulphuric acid by itself has virtualIyno effect on benzene under the conditions normalIy employed. Itwould thus appear that the sulphuric acid is acting on the nitric acidrather tha.n the benzene in the system. This is borne out by the factthat solutIons of nitric acid in pure sulphuric acid show an almostfou.r-fold mole~ular freezing-point depression (actually i ~ 3·82),WhICh has been mterpreted as being due to formation of the four ions:

•. H,SO. [) H,SO.HO-N02 , • H 20-N02 4 • H OEil + HSO e + EIlNOEll 3 4 2

+HS04 e


i.e. HN03 + 2H 2S04 _',EIlN02 + H3o.EIl + 2HS04 e ,

~he slight shortfall of i below 4 is probably due to incomplete protonatIOn of H 20 under these conditions.. The presence of lBN02, the nitronium ion, both in this solution andm a n.umber of salts (some of which, e.g. lBN02 CI04 e, have actually?een Isolated) has been confirmed spectroscopically: there is a linem .t~e Raman spectru~ of ea~h of the~ at 1400 cm - 1 which can onlyongmate from a speCIes that IS both lInear and triatomic. Nitric aciditself is converted in concentrated sulphuric acid virtualIy entirely intolBN02, and there can be little doubt left that this is the effective e1ectroph.ile .in nitration under ~heseconditions. If the purpose of the sulphuricaCId IS merely to functIon as a highly acid medium in which lBNOcan be re!eased from HO-N02, it would be expected that othe:~trong aCIds, e.g. HCI04 , would also promote nitration. This ismdeed found to b~ t~e ca~e, and. HF plus BF3 are also effective. The poorperformance of mtnc aCId by Itself in the nitration of benzene is thusexplained for it contains but little lBN02; the smalI amount that ispresent is obtained by the two-stage process

........... slowH2~-N02 + HN03 P H30Ell + N03

e + EIlN02

in which nitric acid is first co?verted rapidly into its conjugate acid,and that then more slowly mto EIlN02. The rate of nitration ofaro~atic species more reactive than benzene itself is often found tobe mdependent of [Ar-H], indicating that here it is the actualformation of EIlN02 that is the slow, and hence rate-limiting, step in

137

~+H'

(11)

6.2 Nitration

¢~~...{ NO NO

(9) (10)

Fig. 6.1

It is often very difficult to obtain d~tailed infor~~tion about suchspecies, and the intermediates, of which the transItion states are theimmediate predecessors, are thus often ~aken as. m~els for them,because detailed information about such mtermedlates IS much morereadily come by. This may be )ustifi~d on the basi~ of Ha~mond's

principle that in a sequence, Immediately succee?mg species thatclosely resemble each other in ene~gy level. are hkel.y to res.embleeach other in structure also; certamly the mtermedlate (6) I~ thesequence above is likely to be a better model for T.S. 1 , than IS thestarting material. We shall see a number of examples subsequ~~tly

where (f complexes are used in this way as models for the transitIOnstates that precede them (cf. p. 151). .. .

A further point of preparative significance stIli requires explanatIOn,however. Highly reactive aromatic compou.nds, s~c~ as phenol, arefound to undergo ready nitration even in dIlute m.tnc aCid, and at afar more rapid rate than can b~ explain~d on the ~asls of the concentration of EIlN02 that is present m the mixture. ThiS has b~en s?own tobe due to the presence of nitrous acid in the system which mtrosa~es

the reactive nucleus via the nitrosonium ion, EIlNO (or other speciescapable of effecting nitrosation, cf p. 120):

HNOz + 2HN03 P H 30Ell + 2N0 3e + EIlNO

OH

~ r:$J + HNO,

NOz

that exerts the controlling influence (Fig. 6.1):

H 6°' ~+©+ EIlNOz --+ --+ HEll [2a]slow ............ rast

(6)


isotope effect (cf p. 46) if C6 H6 is replaced by C6 D6 . The comparisonwas in fact (for experimental reasons) made on C6 HsN02 andC6 DsN02-this makes no difference to the argument-and it wasfound that kttlkD at 25° ~ 1·00, i.e. that there is no primary kineticisotope effect. The C-H bond is thus not being broken in the ratelimiting step of the reaction: pathways [1] and [2b] are thereforeruled out. This does not of course prove that nitration proceeds bypathway [2a]: slow, rate-limiting formation of the C-NOz bond,followed by fast, non rate-limiting breaking of the C-H bond;

but this is the only pathway, of those considered, that is compatiblewith our experimental data.

That the cleavage of the C-H bond-a strong one-should befast seems less surprising when we realise that by loss of HEll theintermediate (6) is able to reattain the highly stabilised aromaticcondition in the product nitrobenzene. The incipient proton is removedfrom (6) by the attack of bases, e.g. probably HS04 e in nitratingmixture, but sometimes by solvent molecules. The credibility of speciessuch as (6) as intermediates is enhanced by the actual isolation ofanalogous species, e.g. (8) in the nitration of trifluoromethylbenzene(7) with NOzF/BF3 :

(8) is quite stable below -50°, but is converted into the normalnitration product of (7) on warming. The relative stability of (8) isdue in part to the BF4 e anion in the ion pair (ct. p. 102). Theisolation of (8) does not, of course prove that similar intermediatesare necessarily formed in nitration reactions with nitrating mixturebut, coupled with the kinetic and other evidence, it does make theinvolvement of such species seem very much more plausible.

In discussing rates of aromatic substitution reactions it is, of course,the formation of the transition state (T.S. 1) immediately preceding (6)

In the presence of strong acid, however, H~-Hal ~comes a ~erypowerful halogenating agent due to the formatIon of a hIghly polansed

complex (15):

.. <!l H (<!lHO-Hal + H<!l --. H 20JHai -- H 20 + a

(15)

The evidence is that this species is the effective electrop~ile underthese conditions, and does not support the further converSIOn of (1~)into Hale, i.e. unlike the case with H 20 e -N02 (p.134); HOCI + aCId

1396.3 Halogenation

~+he¢,·~~+mJ-i I I

(14)

Thus kHlk for the iodination of phenol and 2,4,6-trideute~iophenolisfound to be ~4, i.e. pathway [2bJ. Iodination is often aSSIsted by thepresence of bases or of oxidising agents, .which re.m.o~e HI and thusdisplace the above equilibrium to the nght. OXIdIsmg agents ~Isotend to produce I®, or a complex conta.ining positi~elY polansediodine from I thus providing a more effectIve electrophIle. Halogenation ~ay als~ be carried out by use of interhalogen compounds0+ 0- 0+ 0- I .Br-CI, I-C\' etc., attack occurring through the less e ectronegatIvehalogen as this will constitute the 'electrophilic' e~d o~ the m<?lec.ule.The two species above are thus found to effect brommatlOn and IOdma-tion, respectively. . o· 0+

Halogenation may be effected by hypohalous aCIds, HO-Hal,also. This is markedly slower than with molecular halogens as H~eo. 0+ 0+ 0

is a poorer leaving group from HO-Hal than Hale is from Hal-Hal.The reaction is speeded in the presence of Hale, however, as HO-Halis then converted into the more reactive Ha1 2 , e.g.:

Kinetic isotope effects have not been observed for chlorination, andonly rarely for bromination, i.e. the reactions normally follo": pa~h":ay[2aJ like nitration. In iodinat~on,. w~ich only takes place wIth IOdm.eitself on activated species, kmetIc ISOtO~ effec~s are th~ rule. T.hISpresumably arises because the rea~tion IS readIly reverSIble (unlIkeother halogenations), loss of I OCCUrrIng more often from the (1 complex(14) than loss of H, i.e. L 1 :> k2 :

Br1fu\ FeBrJH Br Br

Br, @ 0 ©-+ -+ -+............


The nitrosophenol (10), which may be isolated, is oxidised veryrapidly by nitric acid to yield the p-nitrophenol (11) and nitrousacid; more nitrous acid is produced thereby and the process isprogressively speeded up. No nitrous acid need be present initiaIly I

i~ the nitric acid for a little of the latter attacks phenol oxidatively toyIeld HN02. The rate-determining step is again believed to be theformation of the intermediate (9). Some direct nitration of suchreactive aromatic compounds by ®N02 also takes place simultaneously, the relative amount by the two routes depending onthe conditions.

Many other aromatic electrophilic substitution reactions are found~o follow the general pathway [2J discussed above, usually correspond109 to [2aJ though a number of [2bJ examples are known. The majorpoint still requiring elucidation is very often the exact nature of theelectrophilic species that is involved in attack on the aromatic nucleus.

In contrast to nitration, halogenation can involve a variety of differentelectrophiles in attack on the aromatic system. The free halogens, e.g.Cl2 and Br2, will readily attack an activated nucleus (cf. p. 150) suchas phenol, but are unable to substitute benzene itself (photochemicalactivation can lead to addition, however, through the agency of freehalogen atoms, p. 316): a Lewis acid catalyst such as AlCl3 is requiredto assist in polarising the attacking halogen molecule, therebyproviding it with an 'electrophilic end'; the energy required to formCl® is prohibitive. The rate equation found is often of the form:

Rate = k[Ar- H][HaI 2][Lewis acid]

It seems likely that benzene forms a n complex (12) with, for example,Br2 (ct. p. 131), and that the Lewis acid then interacts with this. Thecatalyst probably polarises Br-Br, assists in the formation of a(1 bond between the bromine molecule's now electrophilic end and aring carbon atom, and finally helps to remove the incipient bromideion so as to form a (1 complex (13) :

6.3 HALOGENATION

(12) FeBr4e + HBr + FeBrJ

(13)

The anion FeBr~e assists in the removal of a proton from the (1 complex(13). The claSSIcal halogen 'carrier' iron filings does, of course, actonly after it has been converted into the Lewis acid, FeX3 .

141

~OCMe] HBr

L8J + AlBr]--+

XMe]~AIBr.e

(18)

In other cases it seems more likely that the attacking electrophileis a polarised complex (19), the degree of polarisation in a particularcase depending on R in R- Hal and the Lewis acid employed:

0+ 0-The carbon atom of alkyl halides, R- Hal, is electrophilic, but rarelyis it sufficiently so to effect the substitution of aromatic species: thepresence of a Lewis acid catalyst, e.g. AIHal3 is also required. Thatalkyl halides do react with Lewis acids has been demonstrated by theexchange of radioactive bromine into EtBr from AlBr~ on mixing andre-isolation; also the actual isolation of solid 1: 1 complexes, e.g.CH 3Br' AlBr3' at low temperatures (- 78°). These complexes, thoughpolar, are only faintly conducting. Where R is capable of forming aparticularly stable carbocation, e.g. with Me3C-Br, it is probablethat the attacking electrophile in alkylation is then the actualcarbocation, Me3CIl

• as part of an ion pair:

6.5.1 Alkylation

6.5 Friedel-Crafts reactions

6.S FRIEDEL-CRAFfS REACTIONS

This can be conveniently divided into alkylation and acylation.

0°'" @+ @"1+ 1- H R RR-CI...FeCI]

~@+@ + so]k, : "EIl :

k, 0= (19) HCIP P HEll • t I --+ 6 FeCI]k_ , ............ rast d-

............

(16) (17)FeCI. e

(20)

can still be a more effective chlorinating agent than CI + AICIh F

· 2 3·owe~er. 2 reacts vIgorously with benzene, but C-C bond-

breakmg occurs and the reaction is of no preparative significance (ct.p. 170).

6.4 SULPHONATION

The mechanisti~ det~ils of sulphonation have been less closely exploredth~n those of mtratIOn or halogenation. Benzene itself is sulphonatedfaIrly slowly by hot concentrated sulphuric acid, but rapidly by?Ieum (the rate then being related to its S03 content) or by S03 inmert.s?lvents. T~e nature of the actual electrophile depends on the~ond.ItI~nS, but IS probably always S03: either free or linked to acarner, e.g. H2S04·S03 (H2S20 7 ) in sUlphuric acid. A small conc~ntration of S03 is developed in H 2S04 itself through the equilibflum:

Attack takes place through S as this is highly positively polarised, i.e.electron-deficient:

Ob-b- ,(.-O~Sb+ + +

'\.Ob-


SuIphonation, like iodination, is reversible and is believed to takeplace in concentrated sulphuric acid via the pathway:

In oleum, the (J complex (16) is believed to undergo protonation ofthe S03 e. b~(ore.un~ergoing C-H fission to yield the S03H analogue?f (! 7).. LIke IOdmatIOn, sulphonation exhibits a kinetic isotope effect,mdicatmg that C-H bond-breaking is involved in the rate-limitingstep of the reaction, i.e. that L I :> k2'

Practical use is made of the reversibility of the reaction in orderto replace S~3H b~ H on treating sulphonic acids with steam. It maythus be pOSSIble to mtroduce an S03H group for its directive influence(cf p. 150), and then eliminate it subsequently. The sulphonation ofnaphthalene presents some interesting features (p. 164).

Either pathway is, of course, compatible with the commonly observedrate law:

Rate = k[ArH)[RX)[MX])

The order of effectiveness of Lewis acid catalysts has been shown to be:

AICI] > FeCI] > BF] > TiCI] > ZnCl2 > SnCI.

The validity of Wheland intermediates such as (18) and (20) inFriedel-Crafts alkylation has been established by the actual isolation

143

~M'

Me

45%

6.5.2 Acylation

10%45%

(5 Me(7H -He

~~

Me Me

(22)

Me(23) (25) (24)

Apart from the possibility of rearrangement, .the.main drawb~ckin the preparative use of this Friedel-Crafts reactIOn. IS polyal.kylatl~n(cf. p. 153). The presence of an electron-withdrawIng. substItuent .ISgenerally sufficient to inhibit Friedel-Crafts alk.ylatlOn ; thus nItrobenzene is often used as a solvent for the reactIon becau~e AlChdissolves readily in it, thus avoiding a heterogeneous reactIOn.

HeMeCHCH) +=! MeCHCH)

I IOH (iJOH 2

6.5.2 Acylation

Friedel-Crafts acylation, in cases where the kinetics can readily bemonitored, is often found to follow the same general rate law as

He Ell CoHoMeCH=CH 2 +=! MeCHCH) -+ Me2CHPh

BFJ

It-H,O

This reaction must of course be intermolecular, but rearrangementsinvolving change in the relative positions of substi.tuents in the benzenering are also known, and these are found t? be Intramolecular. Thusheating p-dimethylbenzene (p-xylene, 23) wIth AICl 3 and HCl resultsin the conversion of the majority of it into the more stable .(cf. p. 1~3)m-dimethylbenzene (m-xylene, 24). The. pre~ence of HCl IS ess~ntlal,and the change is believed to involve mIgratIOn of an Me group In theinitially protonated species (25):

Lewis acid catalysts can also effect dealkylation, i.e. the reaction isreversible. Thus ethylbenzene (22) with BF3 and HF, is found todisproportionate:

to protonate the alkene or alcohol; BF 3 is then often used as theLewis acid catalyst:

MehMe

~+Me

--+_IS"

(21)

of some of them, e.g. (21), at low temperatures (the stabilising effect ofBF4 e on ion pairs has already been referred to p. 136):

r'8YCH2CHMe2 AICI

~ ~


Thus (21) is an orange, crystalline solid that melts with decompositionat - 15° to yield the expected alkylated product in essentially quantitative yield (cf. p. 136).

In a number of cases of Friedel--erafts alkylation the final productis found to contain a rearranged alkyl group. Thus the action ofMe3CCH2CljAICl3 on benzene is found to yield almost wholly therearranged product, PhCMe2CH2Me, which would be explainable onthe basis of the initial electrophilic complex being polarised enough toallow the rearrangement of [Me3CCH2]6+ ... Cl··· AICI/- to themore stable [Me2CCH2Me]6+ ... Cl··· AICI/- (cf. relative stabilityof the corresponding carbocations, p. 104). By contrastMe3CCH2Cl/FeCh on benzene is found to yield almost wholly theunrearranged product, PhCH2CMe3; the presumption being that thecomplex with the weaker Lewis acid, FeCh, is not now polarisedenough to allow of isomerisation taking place. Temperature is alsofound to have an effect, the amount of rearranged product from agiven halide and Lewis acid being less at lower temperatures.

The actual proportions of products obtained in many cases are notnecessarily found to reflect the relative stabilities of the incipientcarbocations, unrearranged and rearranged, however. This followsfrom the fact that their relative rates of reaction with the aromaticspecies almost certainly do not follow the order of their relativestabilities, and may well be diametrically opposed to it. Attack on thearomatic species by the first formed polarised complex may be fasterthan its rearrangement. The study of these rearrangements is alsocomplicated by the fact that Lewis acids are found to be capable ofrearranging both the original halides, and the final, alkylated endprod ucts, e.g. :

Alkenes and alcohols can also be used in place of alkyl halides foralkylating aromatic species. The presence of a proton acid is required

145

HC]AlCI,

6.5.2 Acylation

EllHCO

I 'AICl,e

Ell Ell C.H. hCMMe,C-C=O -----. CO + Me,C ~ P e,

R 6+ 6-

~O~CI'

(29)

The reaction is in fact an equilibrium that lies unfavourably for productformation, but is pulled over to the right by complexing of the aldehyde(30) with the Lewis acid catalyst. . _

Acylation may also be effected by aCid anhydndes, (RCOhO, andE9

Lewis acids (the effective nucleophile here may be RC=O or in somecases RCOCI is formed by the action ofAICI 3on the original anhydri?e),and also by acids themselves. This latter is promote? by strong ac~ds,

e.g. H2S0

4, HF, as well as by Lewis acids and may mvolve formatIOn

of acyl cations through protonation:

Formylation may be carried out by use of CO, HCI, and AIC~3

(the Gattermann-Koch reaction); it is d,?ubtfu~ whether HC0C:I IS

ever formed, the most likely electrophde bemg the acyl catIOn,

RC=O (i.e. protonated CO) in the ion pair, RCOal AlCl4E9

:

and is thereby removed from further participation in the reaction.No polyacylation occurs (cf. alkylation,. p. 143).as the productketone is much less reactive than startmg matenal (cf. p. 151).Rearrangement of R does not take place, as in alkylation, butdecarbonylation can take place, especially. where R wou!d form astable carbocation, so that the end result IS then alkylatIOn ratherthan the expected acylation:

because the Lewis acid complexes (29) with the product ketone (28) asit is formed,

+ HCIAICl,

(28)

6+

RC:..:..:O· --AICI/-I (27)

Cl

Ell

RC=O AICI4 e(26)

alkylation:

Rate = k(ArH](RCOCI](AICI,]

There is also the similar general dilemma of whether the effective e1ectrophile is the acyl cation (26) as a constituent of an ion pair, or a ;polarised complex (27):


Acyl cations have been detected in a number of solid complexes, inthe liquid complex between MeCOCI and AlCl3 (by Lr. spectroscopy), in solution in polar solvents, and in a number of cases whereR is very bulky. In less polar solvents, and under a number of othercircumstances, acyl cations are not detectable, however, and it mustbe the polarised complex that acts as the electrophile.

The direct chemical evidence clearly indicates that either (26) or(27) can be involved depending on the circumstances. Thus in thebenzoylation of toluene, the same mixture of products (l %m-, 9 %0

and 90% p-) is obtained no matter what the Lewis acid catalyst is,and with either benzoyl chloride or benzoyl bromide, though thereaction rates do of course differ: this suggests a common attacking

E9species in all cases, i.e. Ph-C=O. On the other hand, in many casesthe proportion of o-product is very small compared with other e1ectrophilic substitutions, e.g. nitration, suggesting a very bulky electrophile:

E9a role better filled by the complex (27) than by the linear R-C=O(26). The nature of the electrophile in any given case clearly dependsvery much on the conditions.

The reaction may thus be represented:

One significant difference of acylation from alkylation is that in theformer rather more than one mole of Lewis acid is required, comparedwith the catalytic quantity only that is required in the latter. This is

147

~N=NAr

(33)

6.6 Diazo coupling

(0

~()H N=NAr

--+

eO

Q~=NAr

$

H 2 N:1. N=NAr

©

Rate = k[ArN 2 $][PhOe]

Coupling with phenoxide ion could take place either on oxygen oron carbon, and though relative electron-density might be expected tofavour the former, the strength of the bond that is formed is also ofsignificance. Thus here, as with other electrophilic attacks on phenols,it is found to be the C-substituted product (31) that is formed:

With most primary amines this is virtually the sole product, but withsecondary amines (i.e. N-alkylanilines) some coupling may also takeplace on a carbon atom of the nucleus, while with tertiary amines (i.e.N,N-dialkylanilines) only the product coupled on carbon (34) is

(32) (31)

Removal of the proton (usually non rate-limiting) from (32) is assistedby one or other of the basic species present in solution. Couplingnormally takes place largely in the p-, rather than the 0-, position(ct. p. 154)-provided this is available-because of the considerablebulk of.the attacking electrophile, ArN2ED (cf. p. 159).

Aromatic amines are in general somewhat less readily attacked thanphenols and coupling is often carried out in slightly acid solution,thus ensuring a high [PhN2ED] without markedly converting the amine,

.. EDArNH2, into the unreactive, protonated cation, ArNH 3-sucharomatic amines are very weak bases (cf p. 69). The initial diazotisationof aromatic primary amines is carried out in strongly acid mediato ensure that as yet unreacted amine is converted to the cation and soprevented from coupling with the diazonium salt as it is formed.

With aromatic amines there is the possibility of attack on eithernitrogen or carbon, and, by contrast with phenols, attack is found totake place largely on nitrogen, with primary and secondary (i.e. Nalkylanilines) amines, to yield diazo-amino compounds (33):eo...... o $

/~ ~=NeO -

6.6 DIAZO COUPLING

Another classical electrophilic aromatic substitution reaction is diazocoupling, in which the effective e1ectrophile has been shown to be thediazonium cation (ct. p. 120):


This latter acylation is used particularly in ring-closures:

000

~o*~~©C)§Jo 0 0

Because polyacylation does not occur (cf. p.145), it is often preferableto prepare alkyl-benzenes by acylation, followed by Clemmensen orother reduction, rather than by direct alkylation:

Thus the 2,4-dinitrophenyldiazonium cation will couple withPhOMe and the 2,4,6-compound with even the hydrocarbon 2,4,6trimethylbenzene (mesitylene). Diazonium cations exist in acid andslightly alkaline solution (in more strongly alkaline solution they areconverted first into diazotic acids, PhN=N-OH, and then intodiazotate anions, PhN=N-08

) and coupling reactions are'therefore carried out under these conditions, the optimum pH dependingon the species being attacked. With phenols this is at a slightlyalkaline pH as it is Ph08

, and not PhOH, that undergoes attack by

$ $

PhN=~"'" Ph~=~

This is, however, a weak electrophile compared with species such asEDN02 and will normally only attack highly reactive aromatic compounds such as phenols and amines; it is thus without effect on theotherwise highly reactive PhOMe. Introduction of electron-withdrawing groups into the 0- or p-positions of the diazonium cationenhances its electrophilic character, however, by increasing the positivecharge on the diazo group:

148 Electrophilic and nucleophilic substitution in aromatic systems 6.6 Diazo coupling 149

EfJOr>.·HH2rt .- N=NAr

H

~ @-~';N=NAr~ (39)

1_

_ He H2N I(y EfJ~ + N=NAr

@-~-N=NArH

(33)

compounds (38) by warming in acid:

H 2N-@-N=NAr

(38)

The rearrangement has been shown under these conditions to be anintermolecular process, i.e. the diazonium cation becomes free, forthe latter may be transferred to phenols, aromatic amines or othersuitable species added to the solution. It is indeed found that therearrangement proceeds most readily with an acid catalyst plus anexcess of the amine that initially underwent coupling to yield thediazoamino compound (33). It may then be that this amine attacKsthe protonated diazoamino compound (39) directly with expulsion ofPhNH 2 and loss of a proton:

H 2N-@-N=NAr

(38)

In conclusion, it should be mentioned that though the great majorityof aromatic electrophilic substitution reactions involve displacementof hydrogen, other atoms or groups can be involved. Thus we havealready seen the displacement of S03H in the reversal of sulphonation(p. 140), of alkyl in dealkylation (p. 143), and a further, less common,displacement is that of SiR3 in protodesilylation (cf. also p. 161):

Displacements such as this show all the usual characteristics of electrophilic aromatic substitution (substituent effects, etc., see below), butthey are normally of much less preparative significance than theexamples we have already considered. In face of all the foregoingdiscussion of polar intermediates it is pertinent to point out thathomolytic aromatic substitution reactions, i.e. by radicals, are alsoknown (p. 331); as too is attack by nucleophiles (p. 167).

(37)

(34)

~N=NAr

B:--+*,

(35)

NaNO,---+/HCI

oN=NAr6l

(36)

obtained:

The reaction is usually found to follow the general rate law:

Rate = k[ArN 2 EfJ][PhNR2]

In some cases the coupling reaction is found to be base-catalysed, andthis is found to be accompanied by a kinetic isotope effect, i.e. L t :> k2 ,

and the breaking of the C-H bond in (35) is now involved in the ratelimiting step of the reaction.

An interesting example of an internal coupling reaction is providedby the diazotisation of o-diaminobenzene (36):

Benzotriazole (37) may be obtained preparatively (75 % yield) in thisway.

The difference in position of attack on primary and secondaryaromatic amines, compared with phenols, probably reflects the relativeelectron-density of the various positions in the former compoundsexerting the controlling influence for, in contrast to a number ofother aromatic electrophilic substitution reactions, diazo coupling issensitive to relatively small differences in electron density (reflecting therather low ability as an electrophile of PhN 2 $). Similar differences inelectron-density do of course occur in phenols but here control overthe position of attack is exerted more by the relative strengths of thebonds formed in the two products: in the two alternative coupledproducts derivable from amines, this latter difference is much lessmarked.

The formation of diazoamino compounds, on coupling ArN2 $

with primary amines, does not constitute a total preparative bar toobtaining products coupled on the benzene nucleus for diazoaminocompounds (33) may be rearranged to the corresponding amino-azo

151

OHH/

PhC~O~-

Oe$/

PhN~o

CI~

~+++/PhC-CI~-

"CI~-

$ $ $

Os" 0< 0"~ .o-attack: +-+ +-+ I E::::::,... $E $ h E #

(400) (4Ob) (40c)

$ $ $

m-allack: & 0" Q"•.H +-+ +-+::::::,... .

E E $ E(410) (41b) (41c)

6.7.1 Electronic effects of Y

What we shall be doing in the discussion that follows is comparingthe effect that a particular Y would be expected to have on the rateof attack on positions a-jp- and m-, respectively, to the substituent Y.This assumes that the proportions of isomers formed are determinedentirely by their relative rates of formation, i.e. that the control iswholly kinetic (cf. p. 163). Strictly we should seek to compare theeffect of Y on the different transition states for a-, m- and p-attack,but this is not usually possible. Instead we shall use Wheland intermediates as models for the transition states that immediately precedethem in the rate-limiting step, just as we have done already indiscussing the individual electrophilic substitution reactions (cf. p.136). It will be convenient to discuss several different types of Y inturn.

6.7.1.1 Y = E&NRJ, CCIJ, NO z, CHO, COzH, etc.

These groups, and other such as S03H, CN, COR, etc., all have incommon a positively charged, or positively polarised, atom adjacentto a carbon atom of the benzene ring:

They are thus all electron-withdrawing with respect to the benzenering, i.e. aromatic species containing them all have a dipole with thepositive end located on the benzene nucleus. Taking Y = E&NR3 asexemplar of the rest, we can write the a complexes for attack by anelectrophile, EE& (e.g. E&N02 ), a-, m- and p- to the original E&NR3substituent:

6.7 ELECfROPHILIC ATTACK ON C6 HsY

When a mono-substituted benzene derivative, C6 H sY, undergoesfurther electrophilic substitution, e.g. nitration, the incoming substituentmay be incorporated at the a-, m- or p-position, and the overallrate at which substitution takes place may be faster or slower thanwith benzene itself. What is found in practice is that substitution


occurs so as to yield either predominantly the m-isomer, or predominantly a mixture of a- and p-isomers; in the former case theoverall rate of attack is always slower than on benzene itself, in thelatter case the overall rate of attack is usually faster than on benzeneitself. The major controlling influence is found to be exerted by Y, thesubstituent already present, and this can be explained in detail on thebasis of the electronic effects that Y can exert. It can, of course, alsoexert a steric effect, but the operation of this factor is confined essentiallyto attack at the a-position; this influence will be discussed separatelybelow (p. 159).

Substituents, Y, are thus classed as being m-, or a-jp-directing;if they induce faster overall attack than on benzene itself they aresaid to be activating, ifslower, then deactivating. It should be emphasisedthat these directing effects are relative rather than absolute: some ofall three isomers are nearly always formed in a substitution reaction,though the proportion of m-product with an a-jp-directing Y or ofa-jp-products with am-directing Y may well be very small. Thusnitration of nitrobenzene (Y = N02 ) is found to result in a mixtureof 93 % m-, 6% a- and 1% p-isomers, i.e. N0 2 is classed as a mdirecting (deactivating) substituent. By contrast nitration of methoxybenzene (anisole, Y = OMe) yields 56 %p-, 43 %a- and I %m-isomers,i.e. OMe is an a-jp-directing (activating) substituent.

152 Electrophilic and nucleophilic substitution in aromatic systems 6.7.1.3 Y = OCOR, NHCOR, OR, OH, NH2 , NR 2 153

e e e

~'ll,d Q. ~ Q~09H E H E H E(42a) (42b) (42c)

The (BNR3 substituent will, of course, exert, overall, a powerfulelectron-withdrawing, i.e. destabilising, inductive (polar) effect onall three positively charged (T complexes (40, 41 and 42), comparedwith the (T complex for similar attack on benzene itself (cf. p. 132).

(BThus attack on any position in C6HsNR3 (0-, p- and m-) will beslower than comparable attack on benzene, e.g. k4,Hsy/k4,H., =1·6 x lO-s for bromination when R = Me.

The (BNR3 group will exert a selective destabilising effect on oneof the canonical structures (40c) of the (T complex for o-attack, andon one of the structures for p-attack (42b); for in each of thesestructures two E9 charges are located on adjacent atoms. The ring E9charge will thus be delocalised less well in (40) and (42) than in (41),in which there is no such disability. The transition state for which(41) is taken as a model will thus be at a lower energy level thanthose corresponding to (40) and (42); its free energy of activation(AG''') will be lower and it will therefore be formed more rapidly:the m-isomer will thus predominate in the reaction product.

Where the positive charge on the atom adjacent to the nucleus isreal rather than formal, i.e. EElNR 3 rather than N02 , there is evidencethat its effect on (J complex stability is exerted through a field effect(cf. p. 22) operating through space, in addition to any polar (inductive) effect operating through the bonds. The deactivating effect ofY on the nucleus declines, i.e. the overall rate of substitutionincreases, in the approximate order:

eNRJ < N02 < CN < SOJH < c=o < C02 H

The order is approximate only as it is found to vary slightly from onesubstitution process to another, depending to some extent on thenature of the attacking e1ectrophile. Thus, hardly surprisingly, substituents such as EElNR 3 will be particularly deactivating in substitutionreactions where the attacking electrophile is itself positively charged,e.g. EElN0 2 (kC6Hsy/kc6H6 = 1·5 x I0- ~ for nitration when R = Me).

6.7.1.2 Y = Alkyl, phenyl

Alkyl groups are electron-donating compared with hydrogen, andthose canonical states for 0- and p-attack, respectively, in which apositive charge is located on the adjacent nuclear carbon atom-(43c)

and (~}-will thus be selectively stabilised;

CH J

QH E

(~) (44b)

in contrast to (4Oc) and (42b) above which were selectively destabilised.No such factor operates in the a-complex for m-attack, cf (41a~ 4Ic),and o-/p- substitution is thus promoted at the expense of m-. Becauseof the overall electron-donating inductive (polar) effect, attack on anyposition will be faster than in benzene itself (kC6HsMe/kc6H6 = 3·4 x 102

for chlorination).We would thus expect C6HsCMe3 to be attacked faster than

C6HsCH3, because of the greater electron-donating inductive effectof Me3C. This is observed for nitration, but the order is reversed forchlorination-suggesting control, of this less polar reaction, byelectron-donation through hyperconjugation which is much greaterin C6HsCH3 (45b-45d) than in C6HsCMe3 (46b). The relative sizeof these alkyl groups also plays a part, however (cf. p. 159):

H H MeI I I

H-C-H H-C He Me-C-Me

Q+-+ ¢ QH E H E H E

I (45b) (45d) (46b)

Specific stabilisation of canonical forms of the (J-complexes for 0- andp-attack can also be effected by a phenyl group, e.g. (47b +-+ 47d),

~ ...H +-+ (:'\--/\...H

~E ~E(47b) (47d)

and the overall rate of attack on biphenyl is found to be faster thanon benzene itself (kC6Hsy/kc6H6 = 4·2 X )02 for chlorination).

6.7.1.3 Y = OCOR, NHCOR, OR, OH, NH%, NR%

These groups all have in common an atom adjacent to the nucleus thatcan exert an electron-withdrawing inductive (polar) effect (cf N in,

154 Electrophilic and nucleophilic substitution in aromatic systems 6.7.1.4 Y = Cl, Br, I 155

Ell

¢H E

(52d)

o 0 6

.. II IHN-C-R ...... HN=C-R

I I

(5Id)

EO

6< ¢H Cl

(52b)(5Ic)

o 0 6

.. II IO-C-R ...... O=C-RI I

the inductive effect will be reversed in direction because of the negativecharge now carried by the oxygen atom, thereby making it even morerapidly attacked than phenol. Even the m-position will now be attackedmore readily than benzene itself (little or no m-product is formed,however). Many electrophilic substitution reactions take place underacid conditions so that the phenoxide anion cannot be involved, butan exception is diazo-coupling (p. 146) which is carried out onphenols in slightly basic solution (cf. p. 147).

The activating effect of Y on the nucleus is found to increase, i.e.the overall rate of substitution increases, in the approximate order:

OCOR < NHCOR < OR <OH <NH2 < NR2

NRzis more powerfully activating than NHz because of the electrondonating effect of the R groups. It should not be forgotten, however,that in acid solution, e.g. in nitration, these two groups will be converted into alNHR z and alNH 3 , respectively; the nucleus will then bedeactivated and substitution will be predominantly m- (cf alNR 3 ,

p. 151). The OH group is sufficiently activating to cause phenol tobe brominated instantaneously to the 2,4,6-tribromo derivative (thep- and both o-positions all attacked) by bromine water at roomtemperature. The groups OCOR and NHCOR are less powerfullyactivating than OH and NHz, respectively, because of the reductionin electron-availability on 0 and N by delocalisation over theadjacent, electron-withdrawing carbonyl group:

6.7.1.4 Y = CI, Br, I

The halobenzenes also have an atom adjacent to the nucleus thatcarries an electron-pair; thus specific stabilisation of the (1 complexesfor 0- and p-attack, (51c++51d) and (52c++52d) respectively, canagain take place, ~~ ~~

6<~

The NHCOR group is not protonated in acid solution, and nitrationof aniline to yield o-Ip-products can thus be carried out by using, forexample, COMe as a protecting group which is subsequently removed.

(~)

~E H(~)

(48c)

QE H

(49c)

QHEll E

(50c)

A::-:~ A::-:~U: $V"(48a) (48b)

a-attack:

p-'''"k 9~ .Q' ~E H E H

(49a) (49b)

e.g. NOz), but they also possess an electron pair (e.g. OMe) that caneffect the specific stabilisation of the (j complexes for 0- andp-attack, (48c ..... 48d) and (49c ..... 49d), respectively, but not thatfor m-attack, (50a=.50c): -

The stabilisation is particularly marked in that not only is an extra(fourth) canonical state involved in the stabilisation of the 0- and p(1 complexes, but these forms (1fu! and 49d, respectively), in whichthe positive charge is located on oxygen, are inherently more stablethan their other three complementary forms, in which the positivecharge is located on carbon (cf. 48a~48c, and 49a~49c). Thiseffect is sufficiently pronounced to outweigh by far the electronwithdrawing inductive (polar) effect also operating in these two (j

complexes, substitution is thus almost completely o-Jp- (<< 1% of them-isomer is obtained in the nitration of PhOMe), and much morerapid than on benzene itself (kc..H,oMeJkc..~ = 9·7 X 106 for chlorination).

The operation of the electron-withdrawing inductive effect can,however, be seen in the fact that the very small amount of m-attack(for which there is no specific stabilisation of the (1 complex by delocalisation) occurs more slowly than attack on benzene itself (cfp. 158). In the case of the phenoxide anion,

m-attack:


1~~' r

i.e. the halogens are o-Ip-directing. The electron-withdrawing inductiveeffects of the halogens are such that attack is slower than on benzeneitself, i.e. they are deactivating substituents (kC6H,Cl/kc6H6 = 3 x 10- 2

for nitration). This net electron-withdrawal by the halogens is reflectedin the ground state by a dipole in chlorobenzene (53) with its + veend on the nucleus, compared with anisole (54) in which the dipoleis in the opposite direction:

1576. 7.2 Partial rate factors and selectivity

Me Me Me

©r, ~ @::60 820 2400

Nitration Chlorination Bromination

kp• kC6H,V/1 '.%,p-isomerf. = - =-- x (I p-position u.6 H positions)po k H kC6H6/6 100

f. =k.. kC6H,V/2 x '.%,m-isomer

(2 m-positions u. 6 H positions)... k H kC6H6/6 100

conditions (or by a competition experiment in which equimolarquantities of C6 H sY and C 6 H 6 compete for an inadequate supplyof an e1ectrophile, thus giving the ratio kC6H,ylkc6H6)' and analysis ofthe relative amounts of 0-, m-, and p-products obtained from C 6 H sYthe isomer distribution (generally quoted as percentages of the totalsubstitution product obtained). Then, remembering that there are6 positions available for attack in C 6 H 6 compared with 20- 2m- andlp-positions in C6 H sY, we have:

k.. kC6H,V/2 %o-isomer/, = - = -- x (2 o-positions u. 6 H positions)

n· k H k C6H6'6 100

Thus for the nitration of toluene by nitric acid in acetic anhydrideat 0° kC6H'MelkC6H6 was found to be 27, and the isomer distribution(%): 0-, 61·5; m-, 1·5; p-, 37·0; the partial rate factors for nitration,under these conditions, are thus:

Comparison of the partial rate factors for nitration of toluene withthose for chlorination and bromination (above) show that these differ,both absolutely and relatively, with the attacking electrophile: inother words relative directive effects in C6 H sY do depend on EEil aswell as on Y. We notice above that the absolute values of the partialrate factors, i.e. kylkH, increase in the order,

Jl = 1·2D(54)(53)Jl = 1·6D

The overall effect exerted by a substituent is, of course, made upfrom inductive/field and mesomeric contributions. With OMe (p.154), the balance is such that the selective stabilisation of thepositively charged intermediates for 0- and p-attack (48 and 49,respectively) is much greater than stabilisation of the correspondingintermediate [(2), p. 132] for attack on benzene itself: o/p-attack onC6H sOMe is thus much faster than attack on C6 H 6 • With a halogensubstituent, e.g. CI, however, the balance-because of a powerfulelectron-withdrawing inductive/field effect-is such that the selective stabilisation of the intermediates for 0- and p-attack (51 and52, respectively) is slightly less than stabilisation of the corresponding intermediate for attack on benzene itself: o-/p-attack on C6H sCIis thus slightly slower than attack on C6 H 6 •

A very similar situation is encountered in the electrophilic additionof unsymmetrical adducts (e.g. HBr) to vinyl halides (e.g. CH2=CHBr),where the inductive effect of halogen controls the rate, but relativemesomeric stabilisation of the carbocationic intermediate controlsthe orientation, of addition (p. 185).

6.7.2 Partial rate factors and selectivity

More refined kinetic methods, and the ability to determine veryprecisely the relative proportions of 0-, m- and p-isomers formed-by,for example, spectroscopic methods rather than by isolation as in thepast-now allow of a much more quantitative approach to aromaticsubstitution. One very useful concept here is that of partial rate/actors:the rate at which one position, e.g. the p-, in C6 H sY is attacked compared with the rate of attack on one position in benzene; it is writtenas/p-o -

Partial rate factors may be obtained by separate kinetic measurements of the overall rate constants k C6H,y and kC6H6 under analogous

Nitration < Chlorination < Bromination

i.e. as the reactivity of the attacking electrophile decreases. Thisapparent paradox is seen on reflection to be reasonable enough: ifE(!) was reactive enough every collision would lead to substitution,the attacking reagent wouIOtllus be quite undiscriminating, and eachpartial rate factor would be unity. As the reactivity of EEil decreases,however, every collision will no longer lead to reaction, which willincreasingly depend on the relative ability of 0-, m- and p-positions inC6 HsY, and positions in C 6 H 6 , to supply an electron pair to bond withE(!). The reagent will thus become increasingly more discriminating-

158 Electrophilic and nucleophilic substitution in aromatic systems 6.7.3 o-/p-Ratios 159

its selectivity will rise-the absolute values of the partial rate factorswill increase, as will the relative difference between these values:exactly what is seen in the figures quoted above. This relative selectivityis best considered by comparing};'" and!m_. only, as!o_, will be influencedby steric effects (size of Y, and relative size of attacking reagent,cf p. 159) in addition to the electronic effects that influence all three.

The use of partial rate factors allows of a more precise investigationof directive effects than has been possible to date. Thus all the partialrate factors for toluene above are> I, indicating that the CH3 group(p. 153) activates all positions in the nucleus compared with benzene.The same is true for Y = CMe3 but here!m- for nitration is 3·0, comparedwith 1·3 for toluene, indicating that CMe3 exerts a larger electrondonating inductive (polar) effect than does CH3. By contrast, whenY = C6 H 5 in biphenyl (p.153),fm- for chlorination is found to be 0·7,i.e. attack on this position is slower than on benzene (althoughkC.H,y/kc•H• = 4·2 x 102), because the Sp2 carbon atom by which theC6 H 5 substituent is attached to the benzene ring exerts an electronwithdrawing inductive (polar) effect (55):

A ..HU E

(55)

A similar effect is also seen with o-/p-directing, activating substituentswhen a reaction can be investigated that produces enough m-productto measure, e.g. deuteration (deuterium exchange) with the strong acid,CF3C02D, on C6 H 50Ph(56):

~6~0 0'12

33000

(56)

The enormous 10- and!~ values reflect the ability of the electron pairon 0 to stabilise, selectively, the transition states for 0- and p-attack(cf p. 154), while the!m- value of < 1 reflects the destabilisation(compared with attack on benzene) of the transition state for m-attackby the electron-withdrawing inductive (polar) effect of the oxygenatom.

Partial rate factors, and hence the isomer distribution in a particularsubstitution reaction, are also affected by temperature. Increasingtemperature has the greatest relative effect on the substitution reactionof highest AG* (out of the three possible, alternative attacks on C6 H 5V),

i.e. on the slowest. The effect of a rise in temperature is thus, like theeffect of an increase in the reactivity of E$, to 'iron out' differencesbetween partial rate factors, and to make the isomer distribution inthe product move a little more towards the statistical.

6.7.3 o-/p-RatiosAfter what we have seen to date, it surely comes as no great surpriseto find that the ratio of 0- to p-product obtained from substitutionof C6 HsY, where Y is o-/p-directing, is seldom, if ever, the statistical ratio of 2: 1. There is found to be very close agreement betweencalculation and n.m.r. data for the distribution of +ve charge-p- >0-» m--around the ring in the cyclohexadienyl cation (57), whichis the Wheland intermediate for proton exchange in benzene (cf. p.133):

H H

0.260",.0·26: Ell :

0·09 : ...... : 0·09

0·30

(57) (57a) (57b)

On this basis an electron-donating substituent, Y, should be somewhat better at promoting attack by HEll p-(57a, R = H), rather than0-(57 b, R = H), to Y because of the slightly more effective delocalisation of +ve charge that thereby results. The figures quoted in (57)would point to an expected value for the log partial rate ratio,log foJlog fp -, of =0'87, and values very close to this have indeedbeen observed for protonation of a number of different C6 HsYspecies.

The steric demand of HEll is, however, extremely small, and whenattack on C6 HsY is by any other electrophile, EEll, which willnecessarily be larger, there will be increasing interaction between Eand Y in the transition state for attack at the position 0- to Y (57b,R = E) as attacking electrophile and substituent increase in size;there can be no such interaction in the transition state for p-attack(57 a, R = E). This will be reflected in an increasing AG+ foro-attack, a consequently slower reaction, and the relative proportion of o-product will thus fall as the size of E and/or Y increase.This is illustrated by the falling foHp - ratios which are observed forthe nitration of alkylbenzenes (Y = CH3 ~ CMe3) under comparableconditions;

y %0- %p- foHp -ICH,

58 37 0·78

Increase in CH2Me 45 49 0·46sire of Y CHMe2 30 62 0·24

CMe3 16 73 0·11

160 Electrophilic and nucleophilic substitution in aromatic systems 6.7.4 Ipso substitution 161

H

@+H,SO.

An Me3Si substituent may be displaced particularly readily in thisway (protodesilylation) , but we have already seen similar displacement of a more familiar substituent (protodesulphonylation), in thereversal of sulphonation (p. 140):

The net result, if the reaction is to involve more than merelyreversible formation of an intermediate, would thus be displacementof yEll (by EEIl). Such an overall reaction is referred to as ipsosubstitution.

A number of such reactions are known in which the attackingelectrophile is HEll:

intermediate

MC$J~~OH OH

6.7.4 Ipso substitutiouIn addition to 0-, m- and p-attack on C6HsY there is, in theory atleast, the possibility of attack by an electrophile occurring on thering carbon atom to which the substituent Y is already attached:

Finally it should be said that o-/p-ratios can be considerablyinfluenced by the solvent in which the reaction is carried out. Thiscan arise from changes in the relative stabilisation by solvent moleculesof the transition states for 0- and p-attack, but it may also involve theactual attacking electrophile being different in two different solvents:the species actually added complexing with solvent molecules to formthe electrophile proper-a different one in each case. This almostcertainly occurs in halogenation without Lewis acid catalysts, e.g. inthe chlorination of toluene at 25°, where fo-Ifp - ratios between 0·75and O' 34 have been observed depending on the solvent.

and for attack on chlorobenzene by several different electrophiles:

I~:~:~~:tion o/;~_ o~~_ t~;;p-Increase in N' . 30size of E'" ItratIOn 70 O· 21

Bromination 11 87 0·06Sulphonation 1 99 0·005

That the steric factor is not the sole determinant is, however, seenin the figures for the nitration of the halobenzenes, which areo-/p-directing but on which overall attack is slightly slower than onbenzene (p. 155):

Y %0- %p- foJfp _

IF 12 88 0·07

Increase in Cl 30 69 O· 22size of Y Br 37 62 0.30

I 38 60 0·32

Despite the increase in size of the substituent y from F~ I, theproportion of o-isomer, and thus the fo-lfp - ratio, is actually found toincrease. An increasing steric effect will, as with the alkyl benzenes,be operating to inhibit o-attack, but this must here be outweighedby the electron-withdrawing inductive/field effect exerted by thehalogen atom (Y). This effect will tend to decrease with distancefrom Y, being exerted somewhat less strongly on the distant pposition compared with the adjacent o-position. Electronwithdrawal will be particularly marked 0- to the highly electronegative F, and relatively little o-attack thus takes place on C6HsF,despite the small size of F. The electron-withdrawing effect of thehalogen (Y) decreases considerably from F to I (the biggest changebeing between F and CO, resulting in increasing attack at theo-position despite the increasing bulk of Y.

There are some cases where o-substitution occurs to the almosttotal exclusion of any p-attack. These commonly arise from complexingof the substituent already present with the attacking electrophile sothat the latter is 'steered' into the adjacent o-position. Thus when theether l-methoxy-2-pheny!ethane (58) is nitrated with nitrating mixture,32 %0- and 59 %p-isomers are obtained (quite a normal distribution);but nitration with N 20 S in MeCN results in the formation of 69 %0

and 28 % p-isomers. This preferential o-attack in the second case isbelieved to proceed:

(58)

162 Electrophilic and nucleophilic substitution in aromatic systems 6.8 Kinetic versus thermodynamic control 163

A major feature promoting overall ipso substitution will be easeof formation of y$, and we might thus expect to see some suchdisplacement of secondary and tertiary alkyl substituents, because ofthe relative stability of the resultant carbocations, R$. This is foundto happen in the nitration (nitrodealkylation) reactions below:

substitution, however, is not to overlook its possible occurrencewhen contemplating preparative electrophilic substitution of moreheavily substituted benzene derivatives.

(71%)

3931

%p-

2146

%m-

40

23

~oo-

0·01

10

Time (sec)

Ex perimental Calculated

o/~o_ 19 18

~~m· 60 58

%p- 21 24

6.8 KINETIC versus THERMODYNAMIC CONTROL

In all that has gone before a tacit assumption has been made: thatthe proportions of alternative products formed in a reaction, e.g. 0-,

m- and p-isomers, are determined by their relative rates of formation,i.e. that the control is kinetic (p. 42). This is not, however, always whatis observed in practice; thus in the Friedel-Crafts alkylation of methylbenzene (Me: o-/p-directing) with benzyl bromide and GaBr3 (asLewis acid catalyst) at 25°, the isomer distribution is found to be:

Even after a very short reaction time (0·01 sec) it is doubtful whetherthe isomer distribution (in the small amount of product that has asyet been formed) is purely kinetically controlled-the proportion ofm-isomer is already relatively large-and after 10 sec it clearly is not:m-benzyltoluene, the thermodynamically most stable isomer, predominating and the control now clearly being equilibrium or thermodynamic (p. 43).

This is a situation that must rise where the alternative products aremutually interconvertible under the conditions of the reaction, eitherby direct isomerisation or by reversal of the reaction to form thestarting material which then undergoes new attack to yield a morethermodynamically stable isomer. It is important to emphasise thatthe relative proportions of alternative products formed will be definedby their relative thermodynamic stabilities under the conditions of thereaction, which may possibly differ from those ofthe isolated molecules.Thus if m-dimethylbenzene is heated at 82° with HF and a catalyticamount of BF3 the proportions ofthe three isomeric dimethylbenzenesin the product resemble very closely those calculated thermodynamically:

If, however, an excess of BF 3 is used the reaction product is found tocontain > 97 % of m-dimethylbenzene; this is because the dimethyl-

~~o,,~CHMe2 CHMe2

(44%) (56%)

Groups other than alkyl may also be displaced, however (e.g.nitrodehalogenation):

No,·",", ~Me3C CMe3

(100%)

~ :~'~NO'+~[y= BrG>.310/;]Y~IG>.40% J

Analogous nitrodechlorination is not observed, however, owing tothe greater resistance to the formation of CI$, compared with Br$and 1$. Though many of the ipso substitutions that have beenobserved are nitrations, it does also occur during attack by otherelectrophiles (e.g. bromodealkylation):

M,.C@CM', :;;.. B,@.,No doubt ipso attack is also promoted here, and in some other of

the dealkylations above, through the inhibition of normal electrophilic attack at positions in the ring which are flanked by massivealkyl groups. Perhaps the most important point to note about ipso

164 Electrophilic and nucleophilic substitution in aromatic systems 6.9 Electrophilic substitution of other aromatic species 165

benzenes can now be converted to the corresponding salts, e.g.

6.9 ELECfROPHILIC SUBSTITUTION OF OTHER

AROMATIC SPECIES

With naphthalene, electrophilic substitution (e.g. nitration) is foundto take place preferentially at the 1- (Ct-), rather than the alternative2- (P-), position. This can be accounted for by the more effectivedelocalisa~ion, an~ hence stabilisation, that can take place in theWheland mtermedlate for 1- attack (60a +--+ 60b) compared with thatfor 2-attack (61):

H E

~ON

(65)

@Nr

o;H?" "

3-- '>-.. e E

N

(63)

2-sulphonic acid in concentrated H2S04 at 160° results in the formationof exactly the same equilibrium mixture as above, containing 80 %of2-, 20% of 1-, su1phonic acids. The greater stability of the 2-acidstems from the destabilising effect, in the I-acid, of steric interactionbetween the very bulky S03H and the H atom in the adjacent8-position; the 1- and 3-H atoms, in the 2-acid, are both furtheraway.

The interconversion of 1- and 2-acids in H 2S04 at 160° could resulteither from a direct intramolecular isomerisation, or by reversal ofsuIphonation to yield naphthalene which undergoes new attack at theother position. It should be possible to distinguish between thesealternatives by carrying out the reaction in H 2BS04 , for the formershould lead to no incorporation of 3SS in the product suIphonic acids,whereas the latter should lead to such incorporation. Experimentallyit is found that incorporation of 3 Ss does take place but at a rateslower than that at which the conversion Occurs. This could implyeither that both routes are operative simultaneously, or that, afterreversal of suIphonation, new attack takes place on the resultantnaphthalene by the departing H 2S04 molecule faster than by surrounding H 2

3sS04 molecules-the question is still open.Pyridine (62), like benzene, has six 1t electrons (one being supplied

by nitrogen) in delocalised 1t orbitals but, unlike benzene, the orbitalswill be deformed by being attracted towards the nitrogen atom becauseof the latter's being more electronegative than carbon. This is reflectedin the dipole of pyridine, which has the negative end on N and thepositive end on the nucleus:

Jl = 2·3 D (62)

Pyridine is thus referred to as a 1t-deficient heterocycle and, by analogywith a benzene ring that carries an electron-withdrawing substituent,e.g. N0 2 (p. 151), one would expect it to be deactivated towardselectrophilic attack. Substitution takes place, with difficulty, at the3-position because this leads to the most stable Wheland intermediate(63); the intermediates for 2- and 4-attack (64 and 65, respectively)each has a canonical state in which the E9 charge is located on divalentN-a highly unstable, i.e. high energy, state:

(j) .H

2- ~N02

Me

~HFjBF,IXS)

+- •

Me[

Me Me J~.: +-+ ~': BF4

e

MeV' Me~'(59)

The equilibrium will therefore be shifted towards the most basic isomeri.e.. the one (m-) that forms the most stabilised cation (59) in the io~palT. Cases are also known in which the type of control that is operativeIS dependent on temperature (see below).

~ro'~ooO'(j)

(60a) (60b) (61)

More f,?rms can also. be written in each case in which the positivecharge IS now delocahsed over the second ring, leading to a total ofseven forms for the I-intermediate as against six for the 2- but theabo.ve, in which the second ring retains intact, fully delo~alised 1t

orbItals, are probably the most important and the contrast betweentwo ~ontributing forms in the one case and one in the oth~r, corresp,?ndmgly mo~e m~rked. The possibility of the charge becoming morewIdely delocahsed 10 the naphthalene intermediate, as compared withbenzene, would lead us to expect more ready electrophilic attack onnaphthalene which is indeed observed.. The suIphonation of naphthalene with concentrated H

2S0

4at 80°

IS found to lead to almost complete I-substitution, the rate of formation?f th~ al~ernative2-sulphonic acid being very slow at this temperature,I.e. kmetlc control. Sulphonation at 160°, however, leads to the formation o~ no less than 80 %of the 2-sulphonic acid, the remainder beingthe I-Isomer. That we are now seeing thermodynamic control isconfirmed by the observation that heating pure naphthalene 1_ or

166 Electrophilic and nucleophilic substitution in aromatic systems 6.10 Nucleophilic attack on aromatic species 167

Pyrrole. is thus r~ferred to as a n-excessive heterocycle and behavesrather lIke a reactIve ~~nzene deriva!ive, e.g. aniline (p.1S3), undergoingvery .ready electro~hIhc at~ack. ThIS m~y be complicated by the factthat In strongly aCId solutIon protonatlOn (69) is forced even on theweakly basic pyrrole (it takes place on the 2-carbon atom rather thanon N, cf. p. 73):

Q(H~ H

(69)

~rom.atic charac~er is thus lost, and the cation behaves like a conjugateddlene m undergoIng very ready pOlymerisation.

There are certain formal analogies here to m- attack on nitrobenzene(cf p. 152), but pyridine is very much more difficult to substitute thanthe former. Thus nitration, chlorination, bromination and FriedelCrafts rea.ctions cannot really be made to take place usefully, andsulphonatIon only OCCurs on heating with oleum for 24 hours at 2300with an ~?2E1l catalyst. T.his difficulty of attack is due partly to the factthat PyridIne has an avaIlable electron pair on nitrogen, and can thusprotonate (66), or interact with an electrophile (67):

2- r::xH U H[)(H.......... ® " +---+

N E N E N EH H H

(70a) (70b) (70c)

.H .H

3- OE .......... UE"'/-N NH H

(71a) (71 b)

Electrophilic substitution of pyrrole can, however, be carried outunder specialised conditions (e.g. acylation with (MeCOhO/BF3'sulphonation with a pyridine/S03 complex, CsH sN·S03, cf. .(~7))

leading to preferential attack at the 2-, rather than the 3-, pos.ItIOn.This reflects the slightly greater stabilisation of the Wheland mtermediate for the former (70) compared with that for the latter (71):

The difference in stability between the two is not very marked, howe~er,reflecting the highly activated state of the nucleus, and read~ s~bstItu

tion will take place at the 3-position if the 2- is blocked. It IS, mdeed,not uncommon to get substitution on all four carbon atoms, e.g. onbromination with bromine in ethanoic acid.

(72) (73)

Both (a) and (b) would be overcome to some extent if a sufficientlypowerful electron-withdrawing substit.uent was prese~t.' and nucleophilic attack might then become pos.sIble (cf: the add~tlOn of nucleophiles to alkenes carrying electron-wIthdrawmg SubstItuents, p. 198).

6.10 NUCLEOPHILIC ATTACK ON AROMATIC SPECIES

6.10.1 Substitution of hydrogen

It is to be expected that attack by nucleophiles on an unsubstit~ted

benzene nucleus will be much more difficult than attack by electrophlles.This is so (a) because the n electron cloud of the nucleus. (p. 130~ islikely to repel an approaching nucleophil.e,. and (b) because I~~ ~ orbItalsystem is much less capable of delocahsmg (and so stabIhsm?~ thetwo extra electrons in the negatively charged (72), than the posItIvelycharged Wheland intermediate (73):

6' 6

©",N

IE

(67)(66)

©",N

H

!{J)NH

JJ = 1·80 (68)

The positive ~~argewi.ll cl~arly further destabilise any of the (J complexesfor electrophlhc substItutIon, as did a substituent such as EIlNR3 on thebenzene nucleus (p. 152), but the destabilisation will be much more"!ar~ed than with ttlNR 3 as the EB charge is now on an atom of thering Itself and not merely on a substituent.

P~rrole (68) also has 6n electrons in delocalised n orbitals, but herethe mtrogen. atom has. to contribute two electrons to make up the six(t~1Us becomIng es~entIally non-basic in the process, cf p. 73), and thedIP?I~ of 'pyrr~le IS found to be in the opposite direction to that ofpyrldme, I.e. wIth the positive end on nitrogen and the negative endon the nucleus:

1

168 Electrophilic and nucleophilic substitution in aromatic systems 6.10.2 Substitution of atoms other than hydrogen 169

It is found in practice that nitrobenzene can be fused with KOH, inthe presence of air, to yield 0- (plus a little p-) nitrophenol (74):

amino group into its anion (77a). This anio~ may ul~imatel~ .beconverted on treatment with H 20, into the deSIred 2-amInopyndme(77), which is a useful starting material for further synthetic activity.

ya--+

fast

(78)

t::+slow

ArN 2 Gl + ye -+ ArY + N 2

This is found to follow the rate law,

Rate = k[ArN 2Gl

]

i.e. the rate is independent of [ye], and analogies to SN I (~. 78)immediately spring to mind. The observed rate law has been In~erpreted in terms of the slow, rate-limiting, formation of an. aryl catIon,e.g. (78), followed by its rapid reaction with any nucleophlle present:

6.10.2 Substitution of atoms other than hydrogenH6 is in contrast to HEll, a very poor leaving group indeed, with theresult'that in simple aromatic nucleophilic substitution ipso attack(cf. p. 161) is the rule rather than the exception. Cl6 ; Br

6, ~2'

SO 26 6NR etc. are found to be among the more effectIve leavmggro~p~ and, ~ith them, certain analogies to nucleophilic substitutionat a saturated carbon atom (p. 77) may now be observed.

One very common example is the displacement of N2 in the reactionsof diazonium salts, ArN 2Ell, a very useful preparative series:

The S I analogy is reinforced by the fact that added nucleophiles,Cle, MeOH, etc., are found to affect the product compositi.on but notthe rate of reaction-just as the above rate law would reqUIre.

The formation of the highly unstable phenyl cation (78, the EElcharge cannot be delocalised b~ t.he 1t orbi.tal sys~em) is at first sightsomewhat surprising, but the dnvIng force IS provIded by the extremeeffectiveness of N as a leaving group [N N bond energy = 946 kJ(226 kcal) mol- 1]. 2It is significant that this appears to ~ the ~nlyreaction by which simple aryl cations can be generated. In solutIon.The aryl cations are highly reactive, and thus unselectlve, towar~snucleophiles: thus the selectivity between Cle and ~20 (kCle/kH20~ IS

.only 3 for C6H sEll compared w!th. 180 ~~r Me 3C . T~e ve~y hIgh

reactivity of C6H sEll is reflect~ In I~ abllIt! to. recom~Ine Wlt~ N2'i.e. the decomposition of the dlazon!um catIOn. IS reversl~I~; thIS w~sdemonstrated by observing the partIal scramblIng of the N label In

eO'GlJ;o eo 'Oe eo,lIl-?ON (1 NareaH _He ©rOHeOH

I 'ca~(2)

+=t P(1) 4

(75) (74)

The leaving group, H6, subsequently removes a proton from theintroduced NH2 group, thereby evolving H2 and converting the

Other canonical states can be written for the anionic species (75,cf Wheland intermediates), but by far the most significant one is thatshown above in which the e charge is accommodated (and stabilised)by an oxygen atom of the nitro group. This can occur only if theattacking eOH enters the positions 0- and p- to the N02 group (cfspecific stabilisation of a-complexes for electrophilic attack 0- andp- to OMe, p. 154). The species (75) can regain the aromatic conditionby either eOH (I) or eH (2) acting as a leaving group: the formerresulting in recovery of the starting material (nitrobenzene), the latterresulting in the formation of product (74). He is a poor leaving group(contrast the very much better leaving group HEll in electrophilic attack)so the equilibrium tends to lie over to the left-eOH, being a betterleaving group, is lost a lot more often than He-unless an oxidisingagent, e.g. air, KN0 3 , or K 3Fe(CN)6' is present to encourage theelimination of hydride ion, and to destroy it as formed. Some conversion does occur in the absence of any added oxidising agent becausenitrobenzene can act as its own oxidising agent (being reduced toazoxybenzene in the process), but the yield of nitrophenol is then verypoor.

As we might have expected, the electron-withdrawing substituent,N02, that we have already seen to direct electrophilic attack m- toitself (p. 151), directs nucleophilic attack into the 0- and p-positions.

Pyridine (76) requires no more than its own in-built capacity forelectron withdrawal and is itself attacked by powerful nucleophiles,e.g. by 6NH2 (sodamide, NaNH2) in N,N-dimethylaniline assolvent-the Tschitschibabin reaction:

1716.10.2 Substitution of atoms other than hydrogen

In support of the latter interpretation it has proved possible toisolate, and to characterise by n.m.r. spectroscopy and by X-raydiffraction, a number of species closely analogous to (81), e.g. (82);

OMe MeO OEt OEt

02NL9rN02 0

02NQ"NO z

MeOe OZNL9r

NOZEIO " 0

l I : e : l IH~ : H~

NO z N02 NOz

Y, OMeO'NqNa,N02

(Y = H.MeO)(82) (84) (83) (85)

and including the so-called Meisenheimer complex (83), a red crystallinesolid obtainable by the action of EtOe on the methyl eth~r (84). or ofMeOe on the ethyl ether (85). Acidification of the reactIOn mixturefrom either substrate results in the formation of exactly the sameequilibrium mixture of (84) + (85). ~his does not, of course, pro~ethat the normal displacement reactions of, for example, aromatichalides proceed via intermediates but it does make IS seem more

likely. . ' 'd d bDirect support for a stepwise pathway IS, however, proVI e . y

comparison of the rates of reaction of a series of. substrates, ~a~mgdifferent leaving groups, with the same nucleophlle, e.g. 2,4-dlnItrohalogenobenzenes (86) with piperidine (87):

N02 N02

a,N-@-X + HNJ ~ a'N-@-:f)(86) (87) X

S

The relative rates for X = CI, Br and I were found to be 4.3,. 4·3 and1.0, respectively; breaking of the c;-X bond thus cannot ~e I~volvedin the rate-limiting step of the reaction, or we should expect slgnIfica~tlybigger rate differences and in the sequence ~ > Br > Cl. The reaction,in this case, cannot therefore be one-step, I.e. concerted (ef SN2), andin the two-step pathway suggested ~b?~e, step. (~}-att~ck by thenucleophile-would have to be rate-h~ltIng. It IS mtere~tmg too, t?observe that the rate of the above reaction when X. = F IS 330<? ThiSresults from the very powerfully electron-withdrawl~g ~ speedmg upstep (I): (a) by making the nuclea~ carbon to which It IS a~tachedmore positive and hence more r~ad.t1~ attack~ by a n~c1eophlle, and(b) by heiping to stabilise the aniOniC mtermedlate (88).

~..... N~;

02N -:: e ;1....... ~D(88)

e 1S 's IseN==:N N==:N N==:N

@ e

@+=::t @ +=::t

(79a) (79b)

CI CI Y Y

c9Jyell) 0 -Cle(2)

c9Jl I f I

'"N eN EilNso"'-; ~O so"'-- 'Os sO"'-- ~O

(SO) (81)


The fluoroborates are unusual among diazonium salts in beingrelatively stable. They may be isolated, and then heated in the drystate to yield pure ArF; the other products being lost as gases.

A number of the reactions of diazonium salts, particularly in lesspolar solvents, may proceed via the initial generation of an arylradical, however (cf. p. 334).

Probably the most common aromatic nucleophilic displacementreactions involve the displacement of Hale from a halide activatedby electron-withdrawing groups, e.g. (80):

A particularly useful displacement reaction on ArN2 E!J is theintroduction of F into the benzene nucleus (not possible by directreaction with F2 , cf. p. 140):

(79):

Rate = k[ArX][YS]

These reactions are generally found to follow the rate law.

so that there is some formal resemblance to SN2. The above pathwaymust, however, differ in that attack by ye cannot take place from theback of the carbon atom carrying the leaving group (ef SN2, p. 78),but must occur from the side; it is thus often referred to as SN2(aromatie).Further, on the basis of the above rate law the reaction could beconcerted (like SN2}-in which case (81) is a transition state-or itcould proceed by a stepwise pathway with either step (I) or step (2)as the slow, rate-limiting one-in which case (81) is an intermediate.

DMSO 8Me)C0 8 + Ph-C1 ----+ Ph-OCMe) + CI

This surprising difference in reactivity suggests the possibility of ~reaction pathway other than SN2(aromatic), and some clue to what Itmight be is provided by the observation that p-~hlorom~thylbenzene

(93) undergoes the same reaction (equally readIly) to gIve not onlythe expected p-aminomethylbenzene (94), but also the unexpected m-

1736.10.3 'Substitution' via aryne intermediates

are observed:


The relative inertness of unactivated aromatic halides towards nucleophiles, under normal conditions, is in sharp contrast to their markedreactivity towards nucleophiles that are also very strong ba~es. T~uschlorobenzene is readily converted into aniline by reactIOn wItheNHz (NaN Hz) in liquid ammonia at - 33°:

N N80 ° 80 @ ° III III

~ ~'~(' ~" M'*M'Br Br Br Br

(89) (90) (91) (92)

40 1 4 I

The rate difference between (91) and (92) is very small as the Me groupsdo not prevent mesomeric electron-withdrawal by the linear CN group.The rate difference is much more pronounced between (89) and (90),however, as the Me groups prevent the oxygen atoms of the nitrogroup lying in the same plane as the nucleus, p overlap between Nand the adjacent C is thus markedly reduced. . .

Finally it should be mentioned that a number of nucleophIlIcsubstitution reactions of unactivated halides can be made to proceedin bipolar non-protic solvents such as dimethyl sulphoxide (DMSO),MezSal-Oe . No hydrogen-bonded solvent envelo~e, as in for ~xampleMeOH then needs to be stripped from ye before It can functIOn as anUcleophile'l\G* is thus much lower and the reaction correspondinglyfaster. Rate'differences of as much as )09 have been obser~ed onchanging the solvent from MeOH to MezSO. Chlorobenzene wIll thusreact readily under these conditions with Me)COe :

2,4-Dinitroftuorobenzene (86, X = F) is, because of its reactivity,much used for 'tagging' the NHz group of terminal amino-acids inprotein end group analysis. Once it has reacted with the NHz it isvery difficult to remove again and will thus withstand the subsequenthydrolysis of the protein to its constituent amino-acids.

Such rate difference as there is for attack on (86) depends on theability of X, through electron-withdrawal, to influence the relativeease of attack on the substrate by the nucleophile: it is in the reverseorder of the relative ability of the halide ions as leaving groups.When the same series of halides is reacted with C6H sNHMe (innitrobenzene at 120°), however, the relative rates for X = F, CI andBr were found to be 1, 15 and 46, e.g. in the order of their relativeability as leaving groups, so that in this latter reaction it wouldappear that step (2) is now involved, to some extent at least, in therate-limiting step overall.

The first pathway above is much the more common, however, andwe can add it [SN2(aromatic)]-bond-breaking by the leaving groupafter bond-formation to the nucleophile-to the SN2-bond-breakingby the leaving group and bond-formation to the nucleophile simultaneous-and the SNI-bond-breaking by the leaving group before bondformation to the nucleophile-pathways that we have already encountered. Thus nucleophilic aromatic 'substitution' is in fact anaddition/elimination process very similar to electrophilic aromatic'substitution', except for the different attacking species. Other importantexamples of nucleophilic aromatic substitutions of preparative significance are the displacement of SO)ze from the alkali-metal salts ofsulphonic acids, e.g. ArSO) 8Naal, by eOH and eCN and, lessimportantly, the displacement of eNRz from p-nitroso-N,N-dialkylanilines by eOH.

Significant electron-withdrawal by a substituent to stabilise theanionic intermediate, e.g. (81), only occurs through a mesomericeffect, i.e. when the nitro group, for example, is 0- and/or p- to theleaving group. Thus we observe the reactivity sequence:

~ , ~NO'" c9J < ~NO' < O'N*NO'

NO, NO, NO,

2- and 4-, but not 3-, halogenopyridines undergo ready nucleophilicdisplacement reactions for exactly the same reason. Mesomericinteraction with an electron-withdrawing substituent will be reducedor inhibited if the p orbital on the atom adjacent to the nucleus, e.g. Nin NOz, is prevented from becoming parallel to the p orbital on thenuclear carbon to which it is attached (steric inhibition of delocalisation, cf p. 71). Thus the following relative rates of nucleophilic attack


J; MeWMe rnJMe~Me Me~Me Me~Me

Br . Br. I

not a symmetrical intermediate, Le. the two possible positions ofattack by 8NHz are not identical.

This is clearly an elimination/addition mechanism [in contrastto the addition/elimination of SN2 (aromatic)] and formally parallels,in its genesis, the elimination reactions of simple alkyl halides that weshall consider subsequently (p. 246). Direct evidence in support ofthe aryne pathway is provided by the fact that the halides (98), (99)and (100),

175

(100)

(I Ole)

(99)

(lOlb)(lOla)

(98)


react with 8NHz only under conditions much more vigorous thancan be explained by any steric effect exerted by their a-Me gro~ps.

None, however, possesses an H atom 0- to the hal<,>gen: a r~qUlre

ment essential for initiating reaction via an aryne mtermedlate, aswe saw above.

Arynes present structural features of s~me interest. ,They clearlycannot be acetylenic in the usual sense as this would reqUire enormousdeformation of the benzene ring in order to accommodate the 1800

bond angle required by the Spl hybrid!sed carbo~s in an alkyne (p: 9).lt seems more likely that the delocahsed n orbitals of the arom~ttc

system are left largely untouched (aromatic stability thereby bel~g

conserved), and that the two available electrons are accommodated 10

the original Sp2 hybrid orbitals (10 I) :

Me Me Me

c9J6NJ.l 1

c9J ©NH'----. +Iiq. NH J

_ 33°

CI NH l

(93) (94) (95)

Expected: 38 % Unexpected: 62 ~~

Me

Me Me MeeNH7 c9JH

r$lHCGNH' ~NHl

--+ --+ @ (94)

MeCI eCI eNH2~

r$lNH'(93) (96) (97)

H

aminomethylbenzene (95), and that in the larger relative yield:


No a-isomer is ever obtained, and (94) and (95) are found not to beinterconvertible under the conditions of the reaction. This coupled

'h e 'Wit the fact that NHz is known to be able to remove protons(deuterons) from a benzene ring [it removes proton (deuteron) 106times faster from fluorobenzene with an a-deuterium substituent thanfrom deuteriobenzene itself],

suggests that here, too, attack by 8NHz may be as a base on H 0- toCI, rather than as a nucleophile on C of the C-Cl bond:

(95)

The los~ofpr~ton from (93) c?uld be concerted with, or followed by,loss of CI to yield the aryne mtermediate (97). The latter has twoalternative positions (cf. 101b/c) which 8NHz could attack, productformation then being completed by abstraction of a proton from thesolvent NH3 ; the net effect is of the formal addition of NH

3in two

alternative ways round. We should not expect the relative proportions of the two alternative products to be the same because (97) is

Overlap between these orbitals will, on spatial grounds, be very poor,and the resultant bonding correspondingly weak: arynes are ~hus

likely to be highly reactive towards nucleophile~ (~nd e,lectrophIies),though they are found not to be entirely unselecttve 10 this.

Benzyne itself has been isolated in solid argon at 8 K, ~nd m~c~

evidence for the existence of arynes has come from trappmgexperiments and spectroscopy. Thus generation of. benzyne (10~) inthe presence of furan (102) leads to the formation of. the DlelsAlder (p. 197) adduct (103), which undergoes ready aCld-catalysed

involve both aryne intermediates and SN2(aromatic) pathways; therelative proportions of the overalI conversion proceeding by eachpathway are found to depend on the nucleophile/base, the structureof the aromatic substrate, and on the reaction conditions.


ring fission to yield the more familiar I-naphthol (104):

6.10.3 'Substitution' via aryne intermediates 177

(101) (102) (103) (104)

If benzyne is produced under conditions where there is no suitablespecies for it to react with, then it dimerises ('self-trapping') veryrapidly to the stable biphenylene (105):

A very convincing demonstration of the existence of benzyne byphysical methods involves the introduction into the heated inlet of amass spectrometer of the zwitterion ion (106), a salt of diazotised 0

aminobenzoic(anthranilic) acid. The mass spectrum is found to be avery simple one exhibiting m/e peaks at 28, 44, 76 and 152:

(105)1101)

CO2 m/e44

N 2 m/e 28

(107)

©t~~:, ~ [©tiNl ~ @) +2N,( lOll

(101)

m/e 76

The m/e 76 peak declined and the m/e 152 peak increased rapidly withtime, indicating the progressive dimerisation of benzyne to the morestable biphenylene (105, above).

Methods such as the pyrolysis of (106), that do not require stronglybasic conditions, have been used to generate arynes in bulk for preparative purposes, and another, even better, method is oxidation ofI-aminobenzotriazole (107) with lead tetraacetate:

Reactions of unactivated halides with the weaker base eOH, thatonly proceed under considerably more vigorous conditions, may welI

i

7Electrophilic and nucleophilic additiontoC=C

7.1 Addition of halogens 179

radicals homolytic, fission of the n bond. The former is usually foundto predominate in polar solvents, the latter in non-polar solventsespecially in the presence of light. Radical induced additions arediscussed subsequently (p. 313).

As we have already seen (p. 8), a carbon-earbon double bond consistsof a strong (J bond plus a weaker n bond differently situated (I):

H.. A ... H"C-C'

H..,....~'H(1)

The pair of electrons ~ the n orbital are more diffuse and less firmlyheld by the carbon n~clel, and so more readily polarisable, than thoseof the (J bond, leadmg to the characteristic reactivity of such unsaturated compounds. As the n electrons are the most readily accessiblefe~ture of the carbon-earbon double bond, we should expect them to~tlleld the molecule from attack by nucleophilic reagents and this ismdee? found to be the case (cJ, p.198, however). The most characteristicreactIOns of the system are, hardly surprisingly, found to be initiatedby e~ectron-deficient species such as XED and X· (radicals can beconsIdered. elect~on-deficient species as they are seeking a furtherelectron WIth whIch to form a bond~ cations inducing heterolytic, and

(3) (4)

Br1 + yeCHz=CH z~ CHz-CHz + CHz-CHz

I I I IDr Dr Dr Y

(2)

~r JKH' 'H

H H

--+

(1)

7.1 ADDITION OF HALOGENS

The decolorisation of bromine, usually in CCl4 solution, is one of theclassical tests for unsaturation, and probably constitutes the mostfamiliar of the addition reactions of alkenes. It normally proceedsreadily in the absence of added catalysts, and one is tempted to assumethat it proceeds by a simple, one-step pathway;

Th~s is clearly incompatible with a one-step pathway like the above,in which there would be no opportunity for attack by yeo It is, ofcourse, important to establish that (4) does not arise merely by subsequent attack of ye on first formed (3), but it is found in practice thatthe formation of (4) is much more rapid than nucleophilic substitutionreactions would be under these conditions. A possible explanation iscompetition by y8 and Bre (derived from Br2) for a common intermediate (see below).

Secondly it is found-with those simple alkenes in which it canbe detected, e.g, trans 2-butene (5)--that the two bromine atoms add

there are, however, two highly significant pieces of experimentalevidence that serve to refute this.

Firstly, if bromine addition is carried out in the presence of addednucleophiles ye or Y: (e.g. Cle , N03

e , H20:) then, in addition tothe expected 1,2-dibromide (3), products are also obtained in whichone bromine atom and one Y atom, or group, have been added tothe double bond (4):

ADDITION OF HALOGENS, p. 179.EFFEcr OF SUBSTITUENTS ON RAIE OF ADDITION, p. 182.ORIENTATION OF ADDITION, p. 184.OTHER ADDITION REAcrIONS, p. 186:704.1 Further halogen derivatives, p. 186; 704.2 Hydration, p. 187;704.3 Carbocations, p. 188; 70404 Hydroxylation, p. 189; 704.5Hydrogenation, p. 191; 704.6 Ozonolysis, p. 192.ADDITION TO CONJUGATED DIENES, p. 194:7.5.1 Electrophilic addition, p. 195: 7.5.2 Diels-Alder reaction,p. 197.NUCLEOPHILIC ADDITION, p. 198:7.6.1 Cyanoethylation, p. 199; 7.6.2 Michael reaction, p. 200; 7.6.3Addition to C=C-C=O, p. 200.

7.5

7.6

7.17.27.3704

GlBr

-(13) (14)

This is possible only because further attack by Bra, on the firstformed intermediate (14), is prevented completely by the extremelybulky, cage-like structures at each end of the original double bond:completion of normal, overall addition of Br2 thus cannot OCCUT.

7.1 Addition of halogens 181

Enough mutual polarisation can apparently result, in (8), for (9) toform, but polarisation of the bromine molecule may be greatlyincreased by the addition of Lewis acids, e.g. AIBr 3 (cf brominationof benzene, p. 138), with consequent rise in the rate of reaction.Formation of (9) usually appears to be the rate-limiting step of thereaction.

The suggestion of cyclic bromonium ions as intermediates, toaccount for the highly stereoselective (ANTI) addition often ob-

, served with simple acyclic alkenes, was made as long ago as 1938.Evidence supporting the existence of such intermediates has comefrom a number of different fields: thus it has proved possible todetect one by physical methods using the 'super' acids of Olah (p.102) and n.m.T. spectroscopy. Thus reaction of the 1,2-dibromide(11) with SbFs in liquid S02 at -600 led to the formation of an ionpair, but this exhibited not the two signal (one from each of twodifferent groups of six equivalent protons) n.m.T. spectrum expectedof (12). Instead one signal only (82·9) was observed, indicatingthat all twelve protons were equivalent, i.e. what is being observedis almost certainly the bromonium ion (9a):

klCH' ~i?: ;pt:H' ~ ftCH'CH 3 CH 3 CH3 CH3 CH 3 CH 3

(II) (I2) (9a)

82·0 82·9

This neighbouring group participation by bromine (cf p. 93) does notofcourse prove that addition to alkenes proceeds via cyclic bromoniurnions, but it does mean that such species are no longer merely ad hocassumptions, and to that extent are correspondingly more plausibleas intermediates.

In attempting to add Br2 to the highly unusual alkene (13), it hasproved possible actually to isolate the cyclic bromonium ion (14):

(6)

~r XMe

A ,~!Me H (Y)

(10)

~r ~K Me

Me H

~~r HBr

Me r Me

H (5)

on from opposite sides of the planar alkene, i.e. ANTI addition:

M'~'~ ~M' ~HH Me H Me H (ANTI addition)

(5) (6) (meso) ---

The product is the symmetrical meso dibromide (6), whereas if additionhad been SYN (both bromine atoms adding from the same side) itwould have been the unsymmetrical (±) dibromide (7):

MeJMe ~ ~r ~HBr Mer K H" (SYN addition)

H Me H

180 Electrophilic and nucleophilic addition to C=C

(5) (7) (±)

It is found in practice with (5), and with other simple acyclic alkenes,that the addition is almost completely stereoselective, i.e. ~ I{)() %ANTI addition. This result also is incompatible with a one-steppathway, as the atoms in a bromine molecule are too close to eachother to be able to add, simultaneously, ANTI.

These observations are explainable by a pathway in which one endof a bromine molecule becomes positively polarised through electronrepulsion by the 7[ electrons of the alkene, thereby forming a 7[ complexwith it (8; cf Br2 + benzene, p. 131). This then breaks down to forma cyclic bromonium ion (9)-an alternative canonical form of thecarbocation (10). Addition is completed through nucleophilic attackby the residual Br8 (or added y8) on either of the original doublebond carbon atoms, from the side opposite to the large bromoniumion Br(!), to yield the meso dibromide (6):

MyfM' (6)Y Br III

(Y)

183

Dr"$ 1/C-C

,;< "Y

Me Me

"" ,;<C=C

;<' "Me Me

9·6xlO

<

$Dr,,/ ,,/

C--C~Y';< "

" /C=C,;< "Y (19)

$Dr,,/ ,,/

C--C;<' "Y

Dr"$ 1/C-C;<' "Y

7.2 Effect of substituents on rate of addition

3 X 10-2

Me Et Me Me

""C=C/ < ""C=C,;<

/ " / 'lo\.H H H Et

<

y" /C=C;<' "Y (18)

O~ <::H-CH +--+_ I 2

Dr

(20)

rate-limiting we should expect, by analogy with electrophilic aromatic substitution (p. 153), that it-and the transition state that precedes it-would be stabilised by electron-donating substituents; i.e.that such substituents (18) would speed up the rate of electrophilicaddition, and vice versa with electron-withdrawing substituents (19):

The following relative rates are actually observed in practice underanalogous conditions;

these relative rates are very susceptible to variation in the reactionconditions, however. The observed rate increase arising from increasingelectron donation by introduction of the later alkyl groups isperhaps smaller than might have been expected; this is due to theincreasing crowding in the transition state introduced by these lateralkyl groups. A phenyl group also increases -the rate of electrophilicaddition considerably (4 x 103

), due to the stabilisation that it can induce in the intermediate (20), and in the transition state that precedes it:

30% SYN addition70% ANTI addition

Dr C02H" /TIC

/ "Dr C02H

30%

+

(16)

Electrophilic and nucleophilic addition to C=C182

The degree of ANTI stereoselectivity exhibited in the addition ofhalogens to alkenes will clearly depend on the relative stability, underthe reaction conditions, of any cyclic halonium ion intermediate, e.g.(9a), compared with the corresponding carbocationic intermediate,e.g. (12). Thus, because of the higher electronegativity of chlorinethan bromine, with corresponding reluctance to share its electronpairs, it might be expected that with some alkenes the addition ofchlorine would be less stereoselective than that of bromine: this isfound to be the case. It might also be expected that structuralfeatures leading to specific stabilisation of carbocations, might alsolead to less ANTI stereoselectivity; this is observed with, for example, trans I-phenylpropene (15):

C.H'~M'H

(15)

The possible formation of a delocalised benzyl type carbocation (16)results in much lower (70%) ANTI stereoselectivity than with trans2-butene (5; =100% ANTI stereoselectivity, p. 180), where no suchdelocalisation is possible. It is also found that increasing the polarity, and ion-solvating ability, of the solvent also stabilises thecarbocation, relative to the bromium ion, intermediate with consequent decrease in ANTI stereoselectivity. Thus addition of bromineto 1,2-diphenylethene (stilbene) was found to proceed 90-100%ANTI in solvents of low dielectric constant, but =50% ANTI onlyin a solvent with f:: = 35.

It is not normally possible to add fluorine directly to alkenes asthe reaction is so exothermic that bond fission occurs. Many alkeneswill not add iodine directly either, and when the reaction does occurit is usually readily reversible. Alkynes are also found to undergopreferential, though not exclusive, ANTI addition of halogens, e.g.with butyne-I,2-dioic acid (17):

C1

0 2H Dr C0 2H" /C CIII -+ IIC C

1 / "C02 H H02C Dr

(17) 70%

7.2 EFFECf OF SUBSTITUENTS ON RATE OF ADDITIONThe intermediate in bromine addition, whether bromonium ion orcarbocation, is indeed positively charged. In so far as its formation is

H HI el Br" I

Me-CH ......CH2- Me-CH-CH2

IDr

185

Me HI I

Me2C-CH-CH2II

unexpectedproduct

7.3 Orientation of addition

tertiarycarbocation

proceed via carbocation intermediates is provided by the formation,from some alkenes, of abnormal (i.e. unexpected) addition productsthat could only have arisen through rearrangement of a carbocation(cf. p. 112), e.g. with 3,3-dimethylbutene (26) and HI:

Me Me H Me HI H~ I I I" , I

Me2C-CH=CH2 ----.. Me2C~H-eH2 ---'--+ Me2C-rH-CH2

I(26) secondary expected

carbocation product

1

Other preparative snags also occur in the addition of HHal toalkenes. Thus in solution in H20, or in other hydroxylic solvents,acid-catalysed hydration (p. 187) or solvation may constitute acompeting reaction; while in less polar solvents radical formationmay be encouraged, resulting in anti-Markownikov addition to giveI-bromopropane (MeC~2CH2Br), via the preferentially formedradical intermediate, MeCHCH2Br. This is discussed in detail below(p. 316).

Electrophilic addition to I-haloalkenes (e.g. 27), presents a numberof parallels to the electrophilic substitution of halobenzenes (p. 155).Thus it is the involvement of the electron pairs on Br that controls theorientation of addition (cf o-jp-direction in C

6H sBr) ;

H HHe el I Br" I

CH =CH-Br -- CH -CH-Br --+ CH -CH-Br2 2 I 2

(27) (28) Br

(23)

H HEll I Br" I

Me ......CH-+-CH2 --+ Me-CH-CH2I

Dr

(22)

Me-CH=CH2

Electrophilic and nucleophilic addition to C=C

(21)

184

7.3 ORIENTATION OF ADDITION

When the electrophile being added is, unlike the halogens, nonsymmetrical then with a non-symmetrical alkene, e.g. propene, theproblem of orientation of addition arises: this will be the case withthe hydrogen halides. These are found to add to a given alkene in therate order: HI>HBr>HCl»HF, i.e. in order of their acidstrengths. This suggests rate-limiting addition of proton to thealkene, followed by rapid nucleophilic attack by Hale to completethe addition. In non-polar solvents the proton is no doubt providedby HHal, but in polar, and especially in hydroxylic solvents, morelikely by its conjugate acid, e.g. by H 30 ffi in H 20.

A bridged intermediate exactly analogous to a bromonium ioncannot be formed as H has no electron pair available, but it may bethat in some cases a 7t complex (21) is the intermediate. We shall,however, normally write the intermediate as a carbocation, and it isthe relative stability of possible, alternative, carbocations (e.g. 23and 24) that determines the overall orientation of addition, e.g. inthe addition of HBr to propene (22) under polar conditions:

As we have seen already (p. 104) secondary carbocations aremore stable than primary, and in so far as this also applies to thetransition states that precede them, (24) will be formed in preference to (23). In fact it appears to be formed exclusively, as the onlyaddition product obtained is 2-bromopropane (25). Addition, ashere, in which halogen (or the more negative moiety of any otherunsymmetrical adduct) becomes attached to the more highly substituted of the two alkene carbon atoms is known as Markownikovaddition.

Support for the suggestion that addition of HHal does normally

(29) is stabilised compared with (28), is therefore formed preferentially,and I, I-dibromoethane (30) is in fact the only product obtained. Therate of addition is, however, controlled by the electron-withdrawinginductive effect of the halogen atom, and (27) is found to add HBr about

(24) (25)

(29a) (29b) (30)

1877.4.2 Hydration

BrI

CH 2-CH 2IOH (35b)

initial bromonium ion intermediate (36):

EElOH 2 OH

H~ Ell H 20 I -H~ IMeCH=CH 2 P MeCH-CH 2 P MeCH-CH 2 P MeCH-CH 2

I I IH H H

(37)

The formation of the carbocationic intermediate (37), either directlyor via an initial 7T complex, appears to be rate-limiting, and theoverall orientation of addition is Markownikov. There is evidence ofsome ANTI stereoselectivity, but this is not very marked and isdependent on the alkene and on the reaction conditions.

Acids that have weakly nucleophilic anions, e.g. HS04 e from diluteaqueous H 2S04 , are chosen as catalysts, so that their anions willoffer little competition to H 20; any ROS0 3 H formed will in any casebe hydrolysed to ROH under the conditions of the reaction. Rearrangement of the carbocationic intermediate may take place, and electrophilic addition of it to as yet unprotonated alkene is also known(p. 185). The reaction is used on the large scale to convert 'cracked'petroleum alkene fractions to alcohols by vapour phase hydrationwith steam over heterogeneous acid catalysts. Also under acidcatalysis, ROH may be added to alkenes to yield ethers, andRC02H to yield esters.

Anti-Markownikov hydration of alkenes may be effected indirectlyby addition of 8 2 H6 (hydroboration), followed by oxidation of the

7.4.2 Hydration

Acid-catalysed hydration of an alkene is the reversal of the similarlyacid-catalysed dehydration (by the El pathway, ct. p. 248) of alcoholsto alkenes:

BrI

Me 2C-CH 2I

CI(34)

(31)(29)

(32)

H

ar•••- ·:I·.Dr-CI ~++ .' . 6+

Me2C=CH 2~ Me2C-CH 2 -+

C8C)

(33)


Addition is initiated by the positively polarised end (the less electronegative halogen atom) of the unsymmetrical molecule, and a cyclichalonium ion intermediate probably results. Addition of I-Cl isparticularly stereoselective (ANTI) because of the ease of formation(and relative stability compared with carbocations) of cycliciodonium ions. With an unsymmetrical alkene, e.g. 2methylpropene (32), the more heavily alkyl-substituted carbon willbe the more carbocationic (i.e. the less bonded to Br in 33), and willtherefore be attacked preferentially by the residual nucleophile,Cle . The overall orientation of addition will thus be Markownikovto yield (34):

BrCI> Br2 > ICI > IBr > 12

The addition of halogen hydracids to simple alkenes is found tobe somewhat less stereoselective than was the addition of halogens,being rather more dependent on the particular alkene, and on thereaction conditions.

30 times more slowly than does ethene (cf bromobenzene is attackedmore slowly by electrophiles than is benzene), i.e. (29) is less stable,and is formed more slowly, than (31):

7.4 OTHER ADDITION REACTIONS

7.4.1 Further halogen derivatives

Interhalogen compounds, hardly surprisingly, add to alkenes verymuch as do the halogens themselves, and the following order ofreactivity has been observed:

Hypohalous acids, e.g. HOI>--Brl>+ (bromine water), werethought to add on in very much the same way, but there is someevidence that the actual electrophile may well be the halogen itself,e.g. Br2 , and that both 1,2-dibromide (35a) and 1,2-bromhydrin(35b) are then obtained by competition of Bre and H20: for the

186

(40)

189

(47) meso

HQ 3H

KM~ 'H

Me H

7. 4. 4 Hydroxylation

H,O--o~ -'/0

-;...os-f-o

WHMe H

(46)(480)

7.4.4 Hydroxylation

There are a number of reagents that, overall, add two OH groups toalkenes. Thus osmium tetroxide, OS04' adds to yield cyclic osmicesters (46), which can be made to undergo ready hydrolytic cleavageof their Os-O bonds to yield the 1,2-diol (47):

It will be noticed that protonation, and subsequent addition, occursto give the most stable cation in each case.

2-Methylpropene can be made to continue the process to yieldhigh polymers-eationic polymerisation-but most simple alkenes willgo no further than di- or tri-meric structures. The main alkenemonomers used on the large scale are 2-methylpropene (- 'butylrubber'), and vinyl ethers, ROCH=CH2 (_ adhesives). Cationicpolymerisation is often initiated by Lewis acid catalysts, e.g. BF3 ,

plus a source of initial protons, the co-catalyst, e.g. traces of H20etc.; polymerisation occurs readily at low temperatures and is usually very rapid. Many more alkenes are polymerised by a radicalinduced pathway, however (p. 320).

Cis 2-butene (48a) thus yields the meso 1,2-diol (47), i.e. the overallhydroxylation is stereoselectively SYN, as would be expected fromOs-O cleavage in a necessarily cis cyclic ester (46). The disadvantageof this reaction as a preparative method is the expense and toxicityof OS04' This may, however, be overcome by using it in catalyticquantities only, but in association with H20 2 which re-oxidises theosmic acid, (HOhOs02, formed to OS04'

Alkaline permanganate, Mn04e, a reagent used classically to testfor unsaturation, will also effect stereoselective SYN addition andthis by analogy with the above, is thought to proceed via cyclic (cis)permanganic esters. It has not proved possible actually to isolatesuch species (some of them are detectable spectroscopically), but useof Mn 180 4e was found to lead to a 1,2-diol (e.g. 47) in which bothoxygen atoms were lS0 labelled. Thus both were derived fromMn04e, and neither from the H2 0 solvent, which provides supportfor a permanganic analogue of (46) as an intermediate, providedthat Mn 1S0 4e undergoes no 180 exchange with the solvent H 20

(44)

<IlMe3C-CH2-CMe2-CH2-CMe2 ---+

Me3C-CH=CMe2

-~<Il

Me3C-CH2-CMe2

(43) CH,=C~(41)

(45)


7.4.3 Carbocations

Protonation of alkenes yields carbocations, as we have seen, and inthe absence of other effective nucleophiles (e.g. H 20, p. 187) theseions can act as electrophiles towards as yet unprotonated alkene (cf.p. 108), e.g. with 2-methylpropene (41):

The diborane is generated (in situ, or separately, from NaBH4 andEt20 Ell-BF3e), and probably complexes, as the monomeric BH

3, with

the ethereal solvent used for the reaction. BH3 is a Lewis acid andadds to the least substituted carbon atom of the alkene (Markownikovaddition), overall addition is completed by hydride transfer to theadjacent, positively polarised carbon atom:

MeCH=CH 2 B,H" (MeCH 2CH 2)3 B H,O" MeCH 2CH 20H + B(OH)3

(38) (39)

resultant trialkylboron (38) with alkaline H20 2:

It may be that (40) has some cyclic character as the overall additionof BH3 is found, in suitable cases, to be stereoselectively SYN. Thefirst-formed RBH 2 then reacts further with the alkene to yield thetrialkylboron, R3 B (38). H20 2 oxidation results in fission of the C-Bbond to yield the alcohol (39), the net result being overall antiMarkownikov hydration that is often stereoselectively SYN; yieldsare usually very good.

The first-formed cation (42) can add to a second molecule of 2-methylpropene (41) to yield the new (dimeric) cation (43); this in turn canlose a proton to yield the Cs alkene (44) or, alternatively, add to athird molecule of alkene to yield the (trimeric) cation (45), and so on.

(42) (41)

Me2C=CH2i~ He (41)

Me3C<Il~CMe2 --.

191

not

7.4.5 Hydrogenation

H HH "'- /

Me3CC=CCMe3 Lind:a,' / C=C "'-catalyst Me3C CMe3

7.4.5 Hydrogenation

The addition of hydrogen to unsaturated compounds is ~mon~ ~he

commonest and almost certainly the most useful, of all theIr addItIonreactions' because of this it is considered here-though it is notpolar in ~ature-rather than under the reactions of radicals. ~irectaddition of hydrogen normally involves heterogeneous catalY~Is byfinely divided metals such as Ni, Pt, Pd, Ru, Rh. The atoms In thesurface of a metal crystal will clearly differ from atoms in the bodyof the crystal in having 'residual combining power' directed awayfrom the surface. It is significant in this context that both alkenes,e.g. ethene, and hydrogen, react exothermically, and r~versibly, withthe catalytic metals, e.g. nickel. With the alkene thIS presumablyinvolves its 7T electrons as alkanes are not similarly adsorbed. No 7T

electrons are available in the hydrogen molecule either, and itsadsorption must involve considerable weakening of its II bond,though not necessarily complete fission to ,yield H· atoms.,

The actual spacings of the metal atoms In the surface wIll clearlybe of importance in making one face of a metal crystal catalyticallyeffective, and another not, depending on how closely the actual atomspacings approximate to the bond distances in alkene and hydrogenmolecules. In practice only a relatively sm~1 proportion of the totalmetal surface is found to be catalytically\effective-the so-called'active points'. These adsorb alkene strongly, and then desorbimmediately the resultant alkane, thus becoming free for furtheralkene adsorption.

In agreement with this 'lining-up' of alkene molecules on the catalystsurface, and the probable approach of activated hydrogen molec~les

from the body of the metal, it might be expected that hydrogenatIOnwould proceed stereoselectively SYN. This is broadly true, and hasoften been of synthetic/structural importance, e.g. :

H

NilH" ~e

Me

(52) (53)

Alkynes can often be reduced selectively to the alkene by use of theLindlar catalyst [Pd on CaC03 , partly 'poisoned' with Pb(OAc)2].Here again SYN stereoselectivity is observed des~ite the fact thatthis will lead to the more crowded, thermodynamIcally less stable,cis-alkene, i.e, (52) rather than (53).

RC02 H

°~Me

Me H

(49)

---+

~ Mt; .HJHrHo 'Me

OH(51)

---+

(48b)


AM, 0"" ~M'~ j?CMe H 80H Me H Me H

(49)III

(51)U H~

H

°/~ H,OK H Me---+

Me i H•. (50)OH2

190

under these conditions-as was shown to be the case. The disadvantage of Mn04 e for hydroxylation is that the resultant 1,2-diol (47) isvery susceptible to further oxidation by it.

Peroxyacids, RCO·OOH, will also oxidise alkenes, e.g. trans2-butene (48b), by adding an oxygen atom across the double bond toform an epoxide (49):

Epoxides, though uncharged, have a formal resemblance to cyclicbromonium ion intermediates (cf. p. 180), but unlike them are stableand may readily be isolated. They do, however, undergo nucleophilicattack under either acid- or base-catalysed conditions to yield the1,2-diol. In either case attack by the nucleophile on a carbon atomwill be on the side opposite to the oxygen bridge in (49); such attackon the epoxide will involve inversion of configuration (ct. p. 94):

Attack has been shown on only one of the two possible carbon atomsin (49) and (50), though on different ones in the two cases. Attack onthe other carbon, in each case, will lead to the same product, themeso 1,2-diol (51). By comparing the configuration of (51) with thatof the original alkene (48b), it will be seen that-in overall termsstereoselective ANTI hydroxylation has been effected.

Thus by suitable choice of reagent, the hydroxylation of alkenescan be made stereoselectively SYN or ANTI at will.

192 Electrophilic and nucleophilic addition to C=C 7.4.6 Ozonolysis 193

-

Its structure has been confirmed by n.m.r. spectroscopy, and by itsreduction with sodium and liquid ammonia to the 1,2-diol (56). It isdifficult to see how catalytic reduction of (54) could lead directly tothe normal carbonyl end.products, however, and on raising the temperature it is found that (54) is converted into (57): which can (and does)yield the normal carbonyl products on catalytic reduction. .

(54) is referred to as a rnolozonide, (57) as the normal ozomde, andthe conversion of the former into the latter is believed to proceed bythe pathway:

(54a) (58) (59) (59) (57a)

The suggested fragments from (54a) are a carbonyl compound (58)and a peroxy zwitterion (59), the latter then. effecting a 1,3-dipo!araddition on the former to yield the ozomde (57 a). AlternatIvereactions of the zwitterion (59), including its polymerisation, lead tothe formation of the 'abnormal' products that are sometimes observed in addition to the ozonide" If ozonolysis is carried out inMeOH as solvent then (59) is 'tr~ped', as it is formed, by itsconversion into the relatively stable a'--hydroperoxy ether (60):

./OMe

R2C"u-oH(60\

The zwitterion (59) is thereby prevented from reacting with theketone (58) to form the ozonide in the normal way, and both (58)and (60) may now be isolated and identified. In preparativeozonolysis it is important to decompose the ozonide (57a) by asuitable reductive process, as otherwise H20 2 is produced (ondecomposition of the ozonide with H20, for example) which canfurther oxidise sensitive carbonyl compounds, e.g. aldehydes~carboxylic acids.

The above pathway accounts satisfactorily for the main features ofozonolysis but requires modification in detail to account for theobserved stereochemistry of the reaction. Thus while a trans- (orcis-) alkene is often found to lead to a mixture of cis- and transozonides as might have been expected, the trans-alkene (55) leadsonly to the trans-ozonide (57). The latter example .d.emands a highdegree of stereoselectivity in both the decompositIon of \54). toaldehyde + peroxyzwitterion and in their subsequent recombInationto (57): a demand that is not implicit in the pathway as we havewritten it.

. Stereoselectivity is often short of being 100 % SYN, and can beInfluenced by reaction conditions, sometimes being very far short of100% SYN. The actual mechanism of hydrogenation has received agood deal of detailed study, by use of D 2 • etc., and is in fact highlycomplex; among other things, hydrogen exchange takes place withthe alkene. It has been established that the two hydrogens are notadded to the alkene simultaneously, however, and the reason for< 100% SY N stereoselectivity thus becomes apparent. It has also beenshown that cis. alkenes, e.g. cis 2-butene, are usually hydrogenatedmuch more rapIdly than trans, e.g. trans 2-butene; in either case therate of hydrogenation falls with increasing substitution in the alkene.

More recently homogeneous hydrogenation catalysts, such asRhCI(Ph3Ph, have been developed which are soluble in the reactionmedium. These are believed to transfer H to an alkene via a metalhydride intermed.ia.te; .they, too, lead to a considerable degree ofSYN stereoselectIVIty In hydrogen addition.

(57)

7.4.6 Ozonolysis

The additi?~ of Ozone to alkenes to form ozonides, and the subsequentdecOmpOSItIon of the latter to yield carbonyl compounds, has longbeen known;

[

R2C

=CR

2j, 0, H,/PtR2C=CR2 -- ;3 -- R2C=O + O=CR; + H20

ozonide

bu~ the structure. of the ozonide has been a matter of some debate.It IS easy. ~o envIsage 1,3-dipolar addition of ozone initiated by its~Iectrophlhc end, and. the crystalline adduct (54) has actually beenIsolated from the reactIon of ozone at - 70° with the alkene (55):

7.5 ADDITION TO CONJUGATED DIENES

195

(69)

(64)

BrI 6)

CH2-CH-CH=CH2Br 2 ~ !

e1eclrophilic

Br, • CH -CHelectrophilic ",2 / 2

Br6)

(66)

(64)

7.5.1 Electrophilic addition

Br,~H2-CH2 +-,-ra""""di,..o.c.a'--l- CH2=CH2

Br(67)

(68)

(65)

7.5.1 Electrophilic addition

BrI .

CH 2-CH-CH=CH2

!Br

'radic:., CH2=CH-CH=CH2

CH -CH=CH-CH (63)I 2 2

Br

Both products are commonly obtained, but their relative proportions depend very much on the reaction conditions, e.g. temperature. Thus HCI with butadiene (63) at -600 yields only 20-25% ofthe l,4-adduct (the rest being the 1,2-adduct), while at highertemperatures =75% of the l,4-adduct was obtained. It is believedthat with bromination at lower temperature the control is kinetic (ct.p. 42), the 1,2-adduct being formed more rapidly from (64) than isthe l,4-adduct; while at higher temperatures, and/or with longerreaction times, equilibrium or thermodynamic control operates, and

or radicals (65) areofthe allylic type (ct. pp. 105,311), and are stabilisedby delocalisation to a considerably greater extent than was the initialdiene. They are also stabilised with respect to the correspondingintermediates (66 and 67) obtained on similar addition to a simplealkene:

Initial attack will always take place on a terminal carbon atom of theconjugated system, otherwise the carbocationic intermediate (64),that is stabilised by delocalisation, would not be obtained. It isbecause of this stabilisation that a carbocation intermediate isformed rather than a cfCIiC--uromonium ion (ct. 66). Completionof overall addition by nucleophilic attack of Br8 on (64) can thentake place at Cz [1,2-addition, (a) ~ (68)] or C4 [1,4-addition,(b)~(69)]:

Br Br BrI (aJ I H H Ib) I

CH2-CH-CH=CH2 ....--'-'--- CH2-CH~CH~CH2 14 dd" • CH2-CH=CH-CH2I 1,2 addition ~ "J .• Ilion IBr (u) Br (b)eBr Br

N~Ph,l'/ 5KR' 'H

R H(62)


The reaction may be carried out in one stage, the sodium metaperiodateused to cleave the 1,2-diol being present in sufficient excess to reoxidisethe catalytic quantity only of Mn04e or OS04 needed for the faststage.

1,3-Dipolar addition to alkenes also occurs with species other thanozone, often to give products much more stable than the labilemolozonides (54), e.g. addition of azides (61) to give dihydrotriazoles(62) :

Ozonolysis was once used to locate the position of a double bond(or bonds) in unsaturated compounds of unknown structure-largelybecause of the ease of characterisation of the carbonyl productsbut has now been superseded by physical methods, e.g. n.m.r.spectroscopy, which are easier and quicker. Benzene forms atriozonide which decomposes to yield three molecules of glyoxal,OHC-CHO: the sole reaction of benzene that suggests it maycontain three 'real' double bonds in a Kekule structure! Alkynes alsoundergo ozonolysis, but at a much slower rate than alkenes.

For preparative cleavage of alkenes, it may be preferable to usethe sequence:

1,3-Dipolar addition to alkenes is considered further subsequently(p. 351).

Conjugated dienes, e.g. butadiene (63) are somewhat more stable thanotherwise similar dienes in which the double bonds are not conjugated(cf p. II). This is reflected in their respective heats of hydrogenation(p. 16), though the delocalisation energy consequent on the extended1t orbital system is only of the order of 17 kJ (4 kcal) mol- 1 ; conjugateddienes are found nevertheless to undergo addition reactions somewhatmore rapidly than non-conjugated dienes. This occurs because theintermediates (and, more importantly, the transition states thatprecede them) arising from initial attack by either electrophiles (64)

197

Me1Il ICH 2-C=CH-CH 2

IH

Me HI Ql I

CH 2 =C-CH-CH 2

7.5.2 Diels-Alder reaction

MeHe I He

.- CH 2=C-CH=CH 2 --

(74b)

MeI

CH 2-C=CH-CH 2I Ql

H

H MeI I

CH 2 -C-CH=CH 21Il

(81) (80) (82)

Cyclic dienes which are locked in the cisoid conformation, e.g. (82),are found to react very much faster than acyclic dienes in which therequired conformation has to be attained by rotation about thesingle bond (the transoid conformation is normally the m.ore stableof the two). Thus cyclopentadiene (82) is suffic!ently !eactlve to .addto itself to form a tricyclic dimer, whose formation-lIke most DlelsAlder reactions-is reversible.

(~o~~oo 0

(63) (79)

involves the I 4-addition of an alkene to a conjugated diene. Thereaction is usu~lIy easy and rapid, of very broad scope, and involvesthe formation of carbon-carbon bonds, hence its synthetic utility. Thediene must react in the cisoid (80), rather than the transoid (81), conformation:

7.5.2 Diels-Alder reactioo

This reaction, of which the classical example is between butadiene(63) and maleic anhydride (79),

(77) (78)

Among other addition reactions dienes undergo catalytic hydrogeneration (1,2- and 1,4-), epoxidation (1,2- only, and more slowlythan the corresponding simple alkenes), but they seldom undergohydration.

(76)

[see over]

1Il

MeCH-CH=CH -CH 2I

H

(72)

CH 2Br/

HC=CH/

BrCH 2

(69)

BrI

P CH 2-CH=CH-CH 2I

Br

(71)

(73)

6+ lH

CH 2-CH'-'-'CH'-'-'CH 2I

Br

(70)


(75)

(68)

BrI

CH 2-CH-CH=CH 2 +='I

Br

With unsymmetrical dienes (74a and 74b) and unsymmetricaladducts, the problem of orientation of addition (cf. p. 184) arises.Initial attack will still be on a terminal carbon atom of the conjugatedsystem so that a delocalised allylic intermediate is obtained, butpreferential attack will be on the terminal carbon that will yield themore stable of the two possible cations; i.e. (75) rather than (76), and(77) rather than (78):

the thermodynamically more stable l,4-adduct is then the majorproduct. This is borne out by the fact that at higher temperaturespure 1,2- or l,4-adduct can each be converted into the sameequilibrium mixture of 1,2- + 1,4- under the conditions of the reaction, l,4-Addition is also favoured by increasing solvent polarity.

There is just the possibility that in adding bromine to butadienethe l,4-adduct might be obtained via an unstrained, five-memberedcyclic bromonium ion (70). This would lead, on nucleophilic cleavage by Br8

, to the cis l,4-dibromide (71):

in fact, only the trans 1,4-dibromide (72) is obtained. Species such as(70) thus cannot be involved, and it seems likely that the commonintermediate is the ion pair (73) involving a delocalised carbocation;interconversion of 1,2- and l,4-adducts, (68) and (69) respectively,could also proceed via such an intermediate:

196

H H1Il I He He I 1Il

MeCH=CH-CH-CH 2 .- MeCH=CH-CH=CH 2 -- MeCH-CH-CH=CH 2

(74a)1Il

MeCH-CH=CH-CH 2

IH

. Diels-Alder reactions are found to be little influenced by theIntroduction of radicals (ef p. 300), or by changes in the polarity ofthe solvent: they are thus unlikely to involve either radical or ionp~ir intermediates. They are found to proceed stereoselectively SYNwIth respect both to the diene and to the dienophile, and are believedto take place via a concerted pathway in which bond-formation andbond-breaking Occur more or less simultaneously, though notneces.s~rily to t?e same extent, in the transition state. This cyclict~anSItIon state IS a planar, aromatic type, with consequent stabilisatIOn because of the cyclic overlap that can occur between the sixp orbitals of the constituent diene and dienophile. Such perieycliereactions are considered further below (p. 341).

T~ese rea~tions a~e found to be promoted by electron-donating~ubstItuents In the dlene, and by electron-withdrawing substituentsIn the al.kene, the dienophile. Reactions are normally poor with simple,unsubstItuted alkenes; thus butadiene (63) reacts with ethene only at2~0 under ~ress.ure, an.d even t~en to the ~xtent of but 18 %, comparedwIth :::::: 100 % yIeld wIth maleIc anhydnde (79) in benzene at 15°.Other common dienophiles include cyclohexadiene-I,4-dione (p_benzoquinone, 83), propenal (acrolein, 84), tetracyanoethene (85),b~nzyne (86, cf p. 175), and also suitably substituted alkynes, e.g.dlethyl butyne-1,4-dioate ('acetylenedicarboxylic ester', 87);

° °

Or C02EtNc CN @ I

I I I H 1 I 0 ccl,11NC CN

° C0 2Et(83) (84) (85) (86) (87)

The reaction is also sensitive to steric effects; thus of the three isomerides of 1,4-diphenylbutadiene (88a -. 88e), only the trans-transform (88a) will undergo a Diels-Alder reaction;

199

F F

M0 + sOMe

Cl CI

CN Ph

~ 0+~NPh H

7.6.1 Cyanoethylation

(89)

<fN ~C=NK Ph

Ph H

eeN--+

FMeOe Me? ~ MeOH-K F-

CI C\

CHO > COR> C0 2R > CN > N02

(90)

The orientation of addition of an unsymmetrical adduct, HY or XV,to an unsymmetrically substituted alkene will be defined by thepreferential formation of the more stabilised car?~nion, as seen .abo~e

(cf. preferential formation of the more stablhsed carbocatI~n melectrophilic addition, p. 184). There is little evidence aval1ab~e

about stereoselectivity in such nucleophilic additions to acychcalkenes. Nucleophilic addition also occurs with suitable alkynes,generally more readily than with the corresponding alkenes:

A number of these nucleophilic addition reactions are ofconsiderablesyn the tic importance;

but SOR, SOzR and F also act in the same way. Such substituentsoperate by reducing 1t electron density at the alk~ne carbon atoms,thereby facilitating the approach of a n~cIeophIle, y e , but m?reparticularly by delocalising the -ve charge l~ th~ re~ultant carbamonintermediate, e.g. (89) and (90). This delocahsatIon IS generally moreeffective when it involves mesomeric delocalisation (89), rather thanonly inductive electron-withdrawal (90):

(p. 151), and to make nucleophilic substitution possible (p. 167);exactly the same is true of addition to al~enes. T~lUs we have alreadyseen that the introduction of electron-wIthdrawmg groups tends toinhibit addition initiated by electrophiles (p. 183); the same groupsare also found to promote addition initiated by nucleophiles. A partialorder of effectiveness is found to be,

Ph Ph

< (Ph (Ph~ Ph

Ph(88a) (88b) (88c)

trans-trans trans-cis cis-cis


7.6 NUCLEOPHILIC ADDITION

:'s we saw ab~ve, the introduction of electron-withdrawing groupsInto an aromatic nucleus tended to inhibit electrophilic substitution

7.6.1 Cyanoethylation

Among the more important of these .reactions of gen.eral syntheticsignificance is one in which ethene carnes a cyano-substItuent (acrylonitrile, 91). Attack of ye or y: on the unsubstituted carbon, followed

200 Electrophilic and nucleophilic addition to C=C 7.6.3 Addition to C=C-C=O 201

EtOHI • R2CCHO

ICH2CH 2CN

by abstraction of a proton from the solvent, leads overall to theattachment of a 2-cyanoethyl group to the original nucleophile;

PbOCH 2CH 2CN '" ~ ROCH2CH2CN~H ROY

CH 2 =CH-CN

~NH, (91) H,~RNHCH2CH 2CN HSCH2CH2CN

the procedure is thus referred to as cyanoethylation. It is often carriedout in the presence of base in order to convert HY into the morepowerfully nucleophilic yeo The synthetic utility of cyanoethylationresides in the incorporation ofa three carbon unit, in which the terminalcyano group may be modified by reduction, hydrolysis, etc., preparatoryto further synthetic operations. Addition of ye, e.g. eNHz, will ofcourse, form a carbanion, YCHz-eCHCN, and, in the absence of aproton donor, this can add to a further molecule of CHz=CHCNresulting, on subsequent repetition, in anionic polymerisation (cf, p.226).

7.6.2 Michael reaction

Where the nucleophile attacking the substituted alkene is a carbanion(cf p. 284) the process is referred to as a Michael reaction; its particularsynthetic utility resides in its being a general method of carboncarbon bond formation; e.g. with (91):

HI EtOe

R2CCHO +===! R2CCHO (92)c:CH 2 =SHCN

(91)

The reaction is promoted by a variety of bases, usually in catalyticquantities only, which generate an equilibrium concentration ofcarbanion (92); it is reversible, and the rate-limiting step is believedto be carbon-earbon bond formation, i.e. the reaction of the carbanion(92) with the substituted alkene (91). Its general synthetic utility stemsfrom the wide variety both of substituted alkenes and of carbanionsthat may be employed; the most common carbanions are probablythose from CHiC02Et)2-see below, MeCOCH2C02Et, NCCH2COzEt, RCHzN02 , etc. Many Michael reactions involve C=C-C=Oas the substituted alkene.

7.6.3 Addition to C=C-C=O

Among the commonest substituents 'activating' an alkene to nucleophilic attack is the C=O group, in such IXp-unsaturated compounds

as RCH=CHCHO, RCH=CHCOR', RCH=CHCOzEt, etc. As ~~ecarbonyl group in such compounds can itself undergo ?~c1e~phI1lcattack (cf, p. 204), the question arises as to whether addItIon IS predominantly to C=C, to C=O, or conjugate (1,4-) to the overallC=C-C=O system. In fact, the last type of ~ddition (93) n?~mallyyields the same product (94) as would be obtamed from addItIon ,toC=C, owing to tautomerisation of the first fo!~ed e.nol (95), e.g. wIththe Grignard reagent PhMgBr followed by aCIdIficatIon:

PIIMg"Sr $R C=CH-C=O ----+ R2C-CH=C-OSMgBr

2 I I IR Ph R

(93)

H 1H"jH,O

IR2T-CH-T=O ~ R2T-CH=C-OH

Ph R Ph ~(94) (95)

Incidentally, 1,4-electrophilic addition (e.g. HBr) also yields the C=Cadduct (96) for the same reason, and can be looked upon formallyas acid-catalysed (97) addition of the nucleophile Bre :

H" <ll

R2C=CH-T

=O +=! R2C-CH=T-OH

R R(97)

H ~eI

R C-CH-C=O q R2C-CH=C-OH2 I I I IDr R Dr R

(96)

Less powerful nucleophiles such as ROH can also be made to add(l ,4-) under acid catalysis.

Whether nucleophilic addition is predominantly conjugate (1,4-) orto C=O may depend on whether the reaction is reversible or no~;

if it is reversible, then the control of product can be thermodynamIC(equilibrium cf p. 43), and this will favour 1,4-addition. This is sobecause the C=C adduct (98) obtained from 1,4-addition will tendto be thermodynamically more stable than the C=O adduct (99),because the former contains a residual C=O 1t bond, and this isstronger than the residual C=C 1t bond in the latter:

H YI I

R2C-CH-C=O R2C=CH-C-OHI I IY R R

(98) (99)Steric hindrance at one site can, however, be very potent at promoting

202 Electrophilic and nucleophilic addition to C=C

1I' = 2·8 D

°IIC/ "-

Me Me1I' = 2·3 D

°IIC/ "-

H H

Carbonyl compounds exhibit dipole moments (/L) because the oxygenatom of the C=O group is more electronegative than the carbon:

As well as the C+- 0 inductive effect in the (J bond joining the twoatoms, the more readily polarisable 1C electrons are also affected (cfp. 22) so that the carbonyl group is best represented by a hybrid

8.1 STRUcruRE AND REACTIVITY, p. 205.8.2 SIMPLE ADDITION REACTIONS, p. 207.

8.2.1 Hydration, p. 207; 8.2.2 Alcohols, p. 209; 8.2.3 Thiols,p. 211; 8.2.4 Hydrogen cyanide, p. 212; 8.2.5 Bisulphite and otheranions, p. 213; 8.2.6 Hydride ions, p. 214: 8.2.6.1 Complex metalhydride ions, p. 214; 8.2.6.2 Meerwein-Ponndorf reaction, p. 215,8.2.6.3 Cannizzaro reaction, p. 216; 8.2.7 Electrons, p. 217.

8.3 ADDITION/ELIMINATION REACTIONS, p. 219:8.3.1 Derivatives of NH3 , p. 219.

8.4 CARBON NUCLEOPHILE ADDITIONS, p. 221:8.4.1 Grignard, etc., reagents, p. 221; 8.4.2 Acetylide anions,p. 223; 8.4.3 Carbanions (general), p. 223; 8.4.4 Aldol reactions,p. 224; 8.4.5 Nitroalkanes, p. 226; 8.4.6 Perkin reaction, p. 227;8.4.7 Knoevenagel and Stobbe reactions, p. 228; 8.4.8 Oaisen estercondensation, p. 229; 8.4.9 Benzoin condensation, p. 231; 8.4.10Benzilic acid rearrangement, p. 232; 8.4.11 Wittig reaction, p. 233.

8.5 S'IEREOSELECTIVITY IN CARBONYL ADDITION REACTIONS, p. 234.8.6 ADDITION/ELIMINATION REACTIONS OF CARBOXYLIC DERIVATIVES,

p.236:8.6.1 Grignard, etc., reagents, p. 238; 8.6.2 Some othernucleophiles, p. 238; 8.6.3 Acid-catalysed reactions, p. 240.

8.7 ADDITION TO O=N, p. 244.

8Nucleophilic addition to C=O

p PH2F-C\ 80Et H2F-~,

Me2C, CH 3 -- Me2C, ~CH2CHC02Et HC-C-OEtI I ~

C02Et C02Et ° (103)

(102) fOEt 0

P ;p ;pH 2C-C H2C-C H2C-C

1 , 1 , (I) Hydrolysis 1,Me2C ;,CH 1_ Me2C CH2 I Me2C CH2

H t-c H t-c (2) Decarboxylation Ht-C2 \ 2 ~ I ~O

OH ° C0 2Et

(lOOa) (100) (104)

The Michael reaction as such is complete on formation of the adduct(102), but treatment of this with base (80Et) yields the carbanion(103), which can, in turn, attack the carbonyl carbon atom of one ofthe C02Et groups; 80Et, a good leaving group, is expelled resultingin cyclisation to (104~reminiscent of a Dieckmann reaction (cf.p. 230). Hydrolysis and decarboxylation of the residual C02Et groupthen yields the desired end-product dimedone (1 (0), which exists tothe extent of ~ 100 % in the enol form (100a).

Dimedone is ofvalue as a reagent for the differential characterisation,and separation, of aldehydes and ketones as it readily yields derivatives(105) with the former, but not with the latter, from a mixture of thetwo:

>Q+b<OH HO

(105)

This selectivity is no doubt due largely to steric reasons.

addition at the other; thus PhCH=CHCHO was found to undergo~ 100% C=O addition with PhMgBr, whereas PhCH=CHCOCMe3underwent ~ 100% C=C addition with the same reagent. This alsoreflects decreasing 'carbonyl' reactivity of the C=O group in thesequence aldehyde> ketone> ester (cf p. 205), with consequentincrease in the proportion of C=C addition.

Amines, thio1s, GOR (p. 226), etc., will also add to the (3-carbonatom of a(3-unsaturated carbonyl compounds and esters, but the mostimportant reactions of C=C-C=O systems are in Michael reactionswith carbanions: reactions in which carbon--earbon bonds are formed.A good example is the synthesis of 1,I-dimethylcyclohexan-3,5-dione(dimedone, 100) starting from 2-methylpent-2-ene-4-one (mesityl oxide,101) and the carbanion 8CH(C02Et)2:

°1/HC-C1/) \ EtOH

Me2C~ CH 3 --

(l0l) 6CHCO EtI 2

C0 2Et

(2)

205

R

@-{_oe

y

(10)

(9)

RI

-. R2C=CH-C-OeI

y

R

'"c=o;<'

R

8.1 Structure and reactivity

R

'"c=o>/

H

H""c=o>/

H

(8)

[O-~ 0·....~ JII_ ~ c=o +--+ <... ~. c-oe

-+

~R R JI Ell I

R2C=CH-C=O ...... R2C-CH=C-Oe

(7)

6+ 6- yO [ 6- 6-J* HYR C=O P R C:.:.:O P R C-oe P R C-OH + ye2 2 : 2 I 2 I

(5) y6- Y Y

(6)

We should thus expect the rate of addition to be reduced by electrondonating R groups and enhanced by electron-withdrawing ones; thisis borne out by the observed sequence:

R groups in which the C=O group is conjugated with C=C (1,4addition can also compete here, ct. p. 200), or with a benzene ring,also exhibit slower addition reactions than their saturated analogues.This is because the stabilisation, through delocalisation, in the initialcarbonyl compounds (7 and 8) is lost on proceeding to the adducts(9 and 10), and to the transition states that precede them:

In the above examples steric, as well as electronic, effects could beinfluencing relative rates of reaction, but the influence of electroniceffects alone may be seen in the series of compounds (II):

H

XO~=o Relative rates: X=N0 2 > H> OMe

(II)

8.1 STRUcrURE AND REACTIVITY

In simple nucleophilic additions where the rate-limiting step is attackby ye, the positive character of the carbonyl carbon atom is reducedon going from the starting material (5) to the transition state (6):

R'6-/

H-O

(4)

(lab)

6+ 6- ,~

i,e, R2C+=0 R2 C+-O'.

(3)

Ell eR2C=0 ...... R2C-0

(Ia) (Ib)

Nucleophilic addition to C=O204

structure (I):

H~ Ell Ell

R2 C=0: o=t R2C=OH ...... R2C-OH

We would expect the C=O linkage, by analogy with C=C (p. 178),to undergo addition reactions; but whereas polar attack on the latteris normally initiated only by electrophiles, attack on the formerbecause of its bipolar nature--eould be initiated either by electrophilicattack of Xal or X on oxygen or by nucleophilic attack of ye or y:on carbon (radical-induced addition reactions of carbonyl compoundsare rare). In practice, initial electrophilic attack on oxygen is of littlesignificance except where the electrophile is an acid (or a Lewis acid),when rapid, reversible protonation may be a prelude to slow, ratelimiting attack by a nucleophile on carbon, to complete the addition,i.e. the addition is then acid-catalysed.

Protonation will clearly increase the positive character of thecarbonyl carbon atom (2),

and thereby facilitate nucleophilic attack upon it. Similar activation,though to a lesser extent, can also arise through hydrogen-bondingof an acid (3), or even of a hydroxylic solvent (4), to the carbonyloxygen atom:

In the absence of such activation, weak nucleophiles, e.g. H 20:, mayreact only very slowly, but strong ones, e.g. eCN, do not require suchaid. Additions may also be base-catalysed, the base acting by convertingthe weak nucleophile HY into the stronger one, y e , e.g. HCN +base -. eCN. Further, while acids may activate the carbonyl carbonatom to nucleophilic attack, they may simultaneously reduce theeffective concentration of the nucleophile, e.g. eCN + HA -. HCN +Ae, RNH 2 + HA -. RNH 3 Eil + Ae. Many simple addition reactionsof carbonyl compounds are thus found to have an optimum pH; thiscan be of great importance for preparative purposes.

yaR....\ ~

'C-O

R.......' "ya

(12)

S? fa~ as steric effects are concerned, the least energy-demandingdirectIOn of approach by the nucleophile to the carbonyl carbonatom will be from above, or below, the substantially planar carbonylcompound. It is also likely to be from slightly to the rear of the carbonatom (cf 12), because of potential coulombic repulsion between theapproaching nucleophile and the high electron density at the carbonyloxygen atom:

eCN > RNH 2 > ROH

A number of the more characteristic addition reactions will now bestudied in greater detail; they will be grouped under the heads: (a)simple addition, (b) addition/elimination, and (c) addition of carbon

nucleophiles.

2078.2 Simple addition reactions

For a given carbonyl compound, K will be influenced by the sizeof the nucleophile: lhus the value of K for addition of the verybulky bisulphite anion (SzO/8, p. 213) to (MeCHzhC=O is only4 x 10-4 , compared with K = 38 for addition of HCN (above) to thevery similar ketone, MeCHzCOMe. The value of K is also influenced by the nature of the atom in the nucleophile that attacksthe carbonyl carbon atom, and of the bond that is thereby formed;as is observed in the following sequence for reaction with the samecarbonyl compound:


Increasing bulk in the R groups will slow the reaction as the Sp2

hybridised carbon atom in the original carbonyl compound (R-C-Rbond angle ~ 120°) is converted to an Sp3 hybridised carbon atom inthe adduct-and in the preceding T.S.-{R-C-R bond angle~ 109°). The R groups thus move closer together as the reactionproceeds, i.e. the T.S. becomes more crowded, its energy level thereforeincreases and the reaction rate drops, as R increases in size. Theobserved drop in reaction rate, H 2C=0 > RHC=O > R 2C=0, isthus determined by both electronic and steric effects. Increase in sizeof the nucleophile, with a given carbonyl compound, may also leadto a drop in reaction rate for the same reason.

Apart from reaction with the strongest nucleophiles, e.g. AIH 4 e6- 6+

(p. 214), RMgBr (p. 221), many additions to C=O are reversible. Ingeneral, the factors that we have seen to affect the rate of reaction (k)influence the position of equilibrium (K) in much the same way; thisis because the T.S. for simple addition reactions probably resemblesthe adduct a good deal more closely than it does the original carbonylcompound. Thus the Ks for cyanohydrin formation (cf p. 212) arefound to reflect this operation of both steric and electronic factors:

K

CHJCHO very largep-N02C.H4 CHO 1420

C.HsCHO 210p-MeOC.H 4 CHO 32CH JCOCH 2CH J 38C.HsCOCH J 0·8

C.HsCOC.H s very small indeed

Highly hindered ketones, such as Me3CCOCMe3' may not react atall except possibly with very small, highly reactive nucleophiles.

8.2 SIMPLE ADDITION REACTIONS

8.2.1 Hydration

Many carbonyl compounds undergo reversible hydration in aqueous

solution.

thus the K values at 20° for H 2C=0, MeHC=O and Me 2C=0 are2 x 103, 1·4, and 2 x 10- 3, respectively; this sequence reflects theprogressive effect of increasing electron-donation. The readyreversibility of hydration is reflected in the fact that H 2C=0 can bedistilled, as such, out of its &queous solution. That adynamic equilibriumactually is set up with Me2C=0, though the ambient conc:ntrationof the hydrate is sO low (its presence has been demonstrated In fro~enMe

2CO/H

20 mixtures, however), may be demonstrated by working

in H/ 80:

'SOH/

P Me 2C"OH

(13)

Incorporation of 18 0 into the ketone occurs hardly at all u~der theseconditions, i.e. at pH 7, but in the presence of a trace of aCid or baseit occurs [via the hydrate (13)J very rapidly indeed. The fact that acarbonyl compound is hydrated will not influence nucleo~h.ilicadditi~~sthat are irreversible; it may, however, influence the positIOn of eqUilibrium in reversible addition reactions, and also the reaction rate, as

the effective concentration of free carbonyl compound, [R2C=O], is

naturally reduced.Hydration is found to be susceptible to both general acid and

~eneral base (p. 74) catalysis, i.e. the rate-limiting step of the reactionInvolves either protonation of the carbonyl compound (general acid14), or conversion of H 20 into the more nucleophilic eOH (generaibase, 15):

2098.2.2 Alcohols

H-O\ P~H: C .

O"=::C' 'c~OI IPh Ph

(19)

H,O

-.=-+

(20)

HO.~OHHO •+-++-

o

A(21)

(18)

8.2.2 AlcoholsThe reactions of carbonyl compounds with alcohols, R'OH, to yieldhemi-acetals (22),

OR'/

R2C=0 + R'OH ~ R2C

"OH

(22)

follows-hardly surprisingly-a very similar pattern to hydrate formation. It also is subject to general acid catalysis, but K forMeCHO/EtOH is only 0·50 compared with a value of 1·4 for H 20;stable hemi-acetals may, however, be isolated from carbonyl compounds carrying electron-withdrawing groups, e.g. Br3CCHO withEtOH. Conversion of the hemi-acetal to the acetal proper (23)requires specific acid catalysis, however (cf. p. 73), Le. it is.lo.s~ ofH

20 (SN 1, cf. p. 80) from (24) that is slow and rate IImltmg,

Another example of a readily isolable hydrate is the one (20) fromcyclopropanone (21),

where the driving force is provided by the measure of relief in bondstrain on going from carbonyl compound (C-C-C bond angle = 60°,compared with normal Sp2 value of 120°) to hydrate (C-C-C bondangle = 60°, compared with normal Sp3 value of 109°).

through H-bonding (17a) between its OH groups (as shown by i.r.spectroscopy) and the highly electronegative chlorine substituen~.

Carbonyl groups can also be effective in stabilising h~drates, pOSSIbly through H-bonding as well as through electron-wIthdrawal, e.g.with diphenylpropantrione (18) which crystallises from water as thehydrate (19):

-0.,0

*H H"../QR2C.

'\;..QH"'A~

(14) T.S'(G.A.)

Nucleophilic addition to C=O

slowR2C=0 P

H-A

H H" /o

208

B H H [~B"'H H] *" / " /o 0 OHslow; /

R2C=0 P R2C. ~ R2C'\;.. "O~- 0 8

(15) T,S'(G.B.)

In contrast to Me2CO, H 2CO hydrates quite readily at pH 7, reflectingthe fact that its more positive carbonyl carbon atom undergoes attackby H 2?: without first requiring protonation of its carbonyl oxygenatom: It nevertheless hydrates very much fa~ter at pH 4 or II!

Just as electron-donating substituents inhibit hydrate formationelectron-withdrawing ones promote it. Thus K for the hydration ofCI 3CCHO (16) is 2·7 x 104

, and this aldehyde (tri-chloroethanalchloral) does indeed form an isolable, crystalline hydrate (17). Th~powerfully electron-withdrawing chlorine atoms destabilise the originalcarbonyl compound, but not the hydrate whose formation is thuspromoted:

...H .....CI 0

[CI-+~~C=O +-+ CI-+~~~-oe] C1-+~~~:OH a-f~-Htl tl tl CIOCI H CI H CI H "'W

(16) (17) (l7a)

For the hydrate to revert to the original carbonyl compound -it hasto. lose e~H or H20:, which is rendered more difficult by the electronwlthdrawmg group. The hydrate from chloral is also stabilised

followed by fast nucleophilic attack by R'OH:

The reaction does not normally take place with ketones underthese conditions (i.e. with simple alcohols), but they can often be madeto react with 1,2-diols, e.g. (25), to form cyclic acetals (26):

2118.2.3 Thiols

rated acetal (28) to the unsaturated car~onyl compound (29); areaction that could not have been carrle~ o~t directly on thebromoaldehyde (30), because of its polymensatlOn by base:

Br BrI EtOH I

~)'-rHO ... CH'-CT:O~:~

CH 2=CHCHO l~~ CH2=CHCH(OEth

(29) (28)

Acetals exhibit the three major requirements of an effective p~otectinggroup: (a) is easy to put on, (b) stays on firmly when reqUIred, and(c) is easy to take off finally.

~20

R R C/SJ"'-.. HgCI:z/CdC03 .....

c=o ~(---- D /' 'sD/ H 20 ,/

8.2.3 ThiolsCarbonyl compounds react with thiols, RSH.. to form he~i-thioacetalsand thioacetals, rather more readily than with ~OH: t~IS refl~cts thegreater nucleophilicity of sulphur comp~red w~th slmtlar~y situated

Thl'oacetals offer with acetals, differential protectIOn for theoxygen. , d'l'd thC=O group as they are relatively stable to I ute aCI; ~y ma~,

however be decomposed readily by H 20/HgClz/CdC03 . It IS pOSSIble usin~ a thioacetal to reverse the polarity of the carbonyl carb.o.nato~ in an aldehyde; thereby convertin~ this initially electrophlhccentre into a nucleophilic one in the amon (31):

thioacetal

This reversal of polarity at an atom, which is r~ferred to as .anumpolung, cannot be effected directly on RCHO Itself. The amon

R'OHi~ faS!

/O.....CH 1R1C I

"o~CHl

(26)

(23)

OR' OR'/ -H$ /

RCH ~ RCH" fast "<J)

OR' OR'H

OR'H~ / -H 2 0 <Il

i=! RCH i=! RCH -OR'fasl " EEl slow

OHH

(24)(22)

OR'/

RCH

"OH

HO-CH1 H~

R1C=O + I i=!HO-CH1

(25)


The fact that reaction can be made to go with (25), but not with thesimple R'OH, is due to the l\SB- (cf p. 36) value for the former beingmore favourable than that for the latter, which involves a decreasein the number of molecules on going from starting material to product.Both aldehydes and ketones that are otherwise difficult to convertinto acetals may often be transformed by use of orthoesters, e.g.HqOEth, triethoxymethane ('ethyl orthoformate'), with NH4CI ascatalyst.

Acetal formation is reversible (K for MeCHO/EtOH is 0·0125)but the position of equilibrium will be influenced by the relativeproportions of R'OH and H20 present. Preparative acetal formationis thus normally carried out in excess R'OH with an anhydrous acidcatalyst. The equilibrium may be shifted over to the right either byazeotropic distillation to remove H20 as it is formed, or by usingexcess acid catalyst (e.g. passing HCl gas continuously) to convertH20: into the non-nucleophilic H 30(!). Hydrolysis of an acetal backto the parent carbonyl compound may be effected readily with diluteacid. Acetals are, however, resistant to hydrolysis induced bybases-there is no proton that can be removed from an oxygenatom, cf. the base-induced hydrolysis of hydrates: this results inacetals being very useful protecting groups for the C=O function,which is itself very susceptible to the attack of bases (cf. p. 224).Such protection thus allows base-catalysed elimination of HBr fromthe acetal (27), followed by ready hydrolysis of the resultant unsatu-

R c=o~ R C(SR') H,/Ni~ R CH2 2 2 2 2

This is a conversion that is usually difficult to effect directly for preparative purposes.

2138.2.5 Bisulphite and other anions

Me,SiCNR 2C=O ~

8.2.5 Bisulphite and other anions

Another classic anion reaction is that with bisulphite ion to yieldcrystalline adducts. The structure of these was long a matter of dispu~ebefore it was established that they were indeed salts of sulphoDlcacids (33), reflecting the greater nucleophilicity of sulphur rather thanoxygen in the attacking anion. The effective nucleophile is almostcertainly S032e (34) rather than HS03e (HOe + HS03e +=t H20 +S032e), as though the latter will be present in higher relative concentration the former is a much more effective nucleophile :

The attacking anion is already present in solution as such so nobase catalysis is required, and S032e is a sufficiently powerful nucleophile not to require activation (by protonation) of the. c~bonyl group,so no acid catalysis is required either. This nucleophJie IS a large one,however, and the K values for product formation are normally

this possibility stems from the large amount of energy that isevolved by formation of the very strong O-Si bond. The preparative intent of initial cyanohydrin formation is usually the furthertransformation of the CN group (e.g. by reduction, hydrolysis, etc.),and this can still be achieved-in high yield~n the Me3Si derivative. This further transformation must, however, be carried outunder conditions (as shown above) such that any backward reactionto the initial carbonyl compound is prevented from taking place. Anadded advantage of Me3SiCN, over HCN itself, is that reaction withC=C-C=O is strictly 1,2-(cf. p. 200), and with ArCHO nobenzoin reaction (p. 231) can take place.

Those carbonyl compounds for which the equilibrium with HCNdoes not lie over in favour of cyanohydrin formation may often beconverted satisfactorily into a derivative of the cyanohydrin throughreaction with Me3SiCN;

+ y8


Rale "" k[R 2C=0][8CN]

The addition of eCN is reversible, and tends to lie over in favour ofstarting materials unless a proton donor is present; this pulls the~eaction over to the right, as t~e equilibrium involving the cyanohydrinIS more favourable than that Involving the intermediate anion (32):

8.2.4 Hydrogen cyanide

Although addition of HCN could be looked upon as a carbanionreaction, it is commonly regarded as involving a simple anion. It isof unusual interest in that it was almost certainly the first organicreaction to have its mechanistic pathway established (Lapworth 1903).HCN is not itself a powerful enough nucleophile to attack C=O, andthe reaction requires base-catalysis in order to convert HCN into themore nucleophilic eCN; the reaction then obeys the rate law:

(31) on treatment with D 20, followed by hydrolysis, results inconversion of the original aldehyde, RCHO, into its deuteriolabelled analogue, RCDO, selectively and in high yield. Alternatively, the anion (31) may be alkylated (e.g. with R'!), and theoriginal aldehyde, RCHO, then converted into a ketone RR'CO.

Thioacetals and thioketals can also be made to undergo desulphurisation with Raney nickel catalyst, thus effecting, overall, theindirect conversion of C=0~CH2:

0 8 OH9CN / HY /

R2C=0 i=! R2C i=! R2Cslow "- fa.. "-

CN CN

(32)

Attack by eCN is slow (rate-limiting), while proton transfer fromHCN or a protic solvent, e.g. H20, is rapid. The effect of the structureof the carbonyl compound on the position of equilibrium in cyanohydri? formatio~.has a.lready been referred to (p. 206): it is a preparatIve propoSItIon WIth aldehydes, and with simple aliphatic andcyclic ketones, but is poor for ArCOR, and does not take place at allwith ArCOAr. With ArCHO the benzoin reaction (p.231) may competewith cyanohydrin formation; with C=C-C=O, 1,4-addition maycompete (cf. p. 200).

215

(39)(38)(37)

8.2.6.2 Meerwein-Ponndorf reaction

OR' OR' "I "" AIH, e (I ,-.., - OR' (I) AIH, e I

RC=O --+ RC-iOe----. RC=O-- RC-OH~ I I (2) R'OH I"-AIH)" " "

e (40) (41) (42)

A less powerful complex metal hydride is Nala BH4 e which willreduce aldehydes and ketones only, and does not attack carboxylicacid derivatives; nor does it-as LilaAIH4 e does-attack NO z orC=N present in the same compound. It has the great advantage ofbeing usable in hydroxylic solvents. A wide variety of other reagentsof the MH

4e, MH 30Re, MHz(ORhe type have been ~e.v~loped:

their relative effectiveness is related to both the nucleophdlclty andsize of MH4 e, etc.

8.2.6.2 Meerwein-Ponndorf reaction

Hydride transfer from carb~n to a c.arbonyl car.bon atom o~curs,

reversibly, in the above reactIOn of whIch the classlcal.example IS thereduction of ketones, e.g. (43), with AI(OCHMezh (44) In propan-2-01,

this then transfers He to three more R2C=O molecules to form thecomplex (38), which finally yields the product alcohol on treatmentwith a protic solvent: thus one of the two H atoms in (39) isprovided by AlH4 e and the other by R'OH. . .

In the reduction of acids there is a tendency for the lIthIUm salt,RCOzeLila to separate from the ethereal solution, and. thus bri~g

reduction to a halt; this can be avoided by first convertIng the aCIdto a simple, e.g. Me or Et, ester. In the redu.ction ofth~ I~tter, ~h~ ini~ialnucleophilic attack by AIH4 e results In an addltIon/elImmatlOnreaction-OR' is a good leaving group in (40)-followed by normalattack, as above, on the resultant carbonyl compound (41) to yield theprimary alcohol (42):

the case, reduction cannot be carried out in protic solvents, e.g. HzO,ROH, as preferential proton abstraction would then take place.Ethers, in a number of which LilaAIH4 e is significantly soluble, arethus commonly employed as solvents.

The nucleophilic AIH4 e donates He, irreversibly, to. th~ carbonylcarbon atom, and the residual AIH 3 then complexes wIth Its oxygenatom to form (37);


8.2.6.1 Complex metal hydride ions

Among the most powerful of these is lithium aluminium hydride,LilaAIH4 e, which will reduce the C=O group in aldehydes, ketones,acids, esters, and amides to CHz, while leaving untouched any C=Cor C-C linkages also present in the compound (C=C conjugatedwith C=O is sometimes affected). The effective reducing agent isAIH4 e which acts as a powerful hydride ion, He, donor; such being

considerably smaller than those for cyanohydrin formation with thesame carbonyl compound (cf p.206). Preparative bisulphite compoundformation is indeed confined to aldehydes, methyl ketones and somecyclic ketones. Such carbonyl compounds can be separated frommixtures and/or purified by isolation, purification, and subsequentdecomposition of their bisulphite adducts.

Halide ions will also act as nucleophiles towards aldehydes underacid catalysis, but the resultant, for example, I, I-hydroxychlorocompound (35) is highly unstable, the equilibrium lying over in favourof starting material. With HCI in solution in an alcohol, ROH, theequilibrium is more favourable, and I,I-alkoxychloro compounds maybe prepared, e.g. I-chloro-I-methoxymethane (36, 'tx-chloromethylether') from CHzO and MeOH (cf acetal formation, p.209), providedthe reaction mixture is neutralised before isolation is attempted :

(35)

tl He

(I) MeOH CI(2) -He Ell -H,O /

, • H 2C-CI ~'===i. H2C"'-Ell

OHH

8.2.6 Hydride ions

Carbonyl groups may be hydrogenated catalytically, as carboncarbon unsaturated linkages were (p. 191). It is, however, normallymore difficult to effect the catalytic reduction of C=O than ofC=c, C=C, C=N, or C===N, so that selective reduction of theformer-in the presence of any of the latter---eannot normally beachieved catalytically. This can, however, be done with various,usually complex, metal hydrides.

(48) (49) (50) (51)

Rapi?, reversible addition of eOH to PhCHO yields the potentialhydnde donor (48), this is followed by slow, rate-limiting hydride

2178.2. 7 Electrons

liqNa' ~ Na Ell + eS(NH 3).

NH,

8.2.7 ElectronsAtoms of a number of the more strongly electropositive metals, e.g.Na, K, etc., can under suitable conditions yield solvated electrons insolution:

SOH OH OS'-.:; 90H I 90H IHC-H ~ HC-H ~ HC-H

C~ s6 s6(52) (53) (54)

transfer to the carbonyl carbon atom (49) of a second molecule ofPhCHO, and the reaction is completed by rapid proton exchange toyield the more stable pair (50) and (51). Mutual oxidation/reductionof two molecules of aldehyde has thus taken place to yield onemolecule each of the corresponding carboxylate anion (50) and ofthe primary alcohol (51).

When PhCHO is made to react in D 20, no D is incorporated intothe CH

2group of (51); showing that H(D) must be transferred (as

the above mechanism requires) directly, from one aldehyde moleculeto the other, and not by any indirect sequence involving the solvent.In very concentrated base, the rate equation may, e.g. with HCHO(52), approach the form:

Rate = k[HCHO]2[SOHf

This corresponds to the removal of a second proton from the species(53), corresponding to (48), to yield the dianion (54), which will clearlybe a much more powerful hydride donor than (53)-or (48):

Suitable dialdehydes can also undergo intramolecular hydridetransfer, as in the Cannizzaro reaction of ethan-l,2-dial (55, 'glyoxal')....... hydroxyethanoate ('glycollate,' 56) anion,

H HI 90H I

o=c-c=o -----+ O=C-C-OHI I I

H sO H

(55) (56)

for which the observed rate law is found, as expected, to be:

Rate = k[OHCCHO)[SOH]

Aldehydes that possess H atoms on the carbon atom adjacent to theCHO group (the IX-carbon atom) do not undergo the Cannizzaroreaction with base, as they undergo the aldol reaction (p. 224) verymuch faster.

OH (0 OH So OS HOI II PhCHO I I I I

PhC,H, CPh~ PhC + H-CPh --,---+ PhC + H-CPhr:: 1 '-..../ I sow II I 4- II IsO H 0 H 0 H


Rate = k[PhCHO)2[SOH)

and the reaction is believed to follow the pathway:

SOHC. 80H

PhC-H. •C~ fa51

216

an equilibrium being set up:

~e2 l Me2 J* Me2 (45)0/ ....... H if,C"'H -7

CI ~ .. p 0 H

(2-PrOlzAI -7CR2 ~ (2-PrOlzAI"'0,;;-CR2 (2-PrOlzAI ....... /~R2o 0

(44) (43) (47) (46)

Propan??e .(45) is the lowest boiling constituent of the system, andthe eqUilIbnum can be displaced, essentially completely, to the rightby di~tiIling this continuously out of the system. An excess of propan2-01 IS employed, and this exchanges with the mixed AI alkoxidepro?uct (46) to liberate the desired reduction product R2CHOH:agam one hydrogen atom has been supplied by hydride transfer andone by a hydroxylic solvent. Because of the specific nature of theequilibr~um, an~ ?f the way in which it is set up, no other groupspresent 10 the ongmal ketone are reduced.

That specific hydride transfer from carbon to carbon does occurwas established by showing that use of labelled (Me CDO) AI ledto t~e formation of R2CDOH. The reaction probably ~roceeJs via acyclIc T.S. such as (47), though some cases have been observed in~hich t~o moles of alkoxide are involved-one to transfer hydridelOn, ~hde the other complexes with the carbonyl oxygen atom. ThereactIOn has ~ow been essentially superseded by MH4 e reductions,but can sometimes be made to operate in the reverse direction (oxidation)by us~ of AI(OC~~3h.as catalyst, and with a large excess of propanoneto dnve. the equilIbnum over to the left. This reverse (oxidation)process IS generally referred to as the Oppenauer reaction.

8.2.6.3 Cannizzaro reaction

This involves hydride transfer from an aldehyde molecule lacking anIX~H atom, e.g. HCHO, R3CCHO, ArCHO, to a second molecule ofeither the same aldehyde (disproportionation) or sometimes to amole.cule of a different aldehyde ('crossed' Cannizzaro). The reactionreqUires the presence of strong bases, and with, for example, PhCHOthe rate law is found to be,

218 Nucleophilic addition to C=O8.3 Addition/elimination reactions 219

Such electrons may act as nucleophiles, and add to the carbonyl~arbo~ at~m of a c=o group to yield a radical anion (57), often as anIOn paIr with the metal cation, ME9:

(71)

Vmax 1400 em - t

(72)(70)

vmax 1710em- t

The end-product is the 2-hydroxyketone, or acyloin (64). Thereaction possibily proceeds as above via a 1,2-diketone (65) whichcan itself accept further electrons from sodium. The end-product ofthe reaction in xylene is the Na salt of (66), but subsequent additionof R'OH effects protonation to yield the 1,2-enediol (67); the finalacyloin (64) is merely the more stable tautomeric form of this. Thereaction is of considerable preparative value for the cyclisation oflong chain diesters, Et02C(CH2)nC02Et, in the synthesis of largering hydroxy-ketones. The yields are very good: 60-95 % over therange n = 8-18, i.e. for 10-20 membered rings, respectively.

8.3.1 Derivatives of NH3

If, for example, the reaction at pH 7 between pyruvate anion (70) andhydroxylamine, NH20H, is followed by monitoring the infra-redspectrum of the reaction mixture, then the absorption characteristicofC=O (vmax 1710cm- 1)-in the starting material (70)-is found todisappear completely before any absorption characteristic of C=N(vmax 1400 cm I)-in the product oxime (71)-appears at all. Clearlyan intermediate must thus be formed, and it seems probable thatthis is a carbinolamine (72; such a species has actually been detectedby n.m.r. spectroscopy, in the reaction of MeCHO with NH20H):

Me Me OH Me, NH,OH 'c/ ,-H20. 'C=NOH/C=O, • /, /

S02C S02C NHOH S02C

8.3 ADDmON/ELIMINATlON REACTIONS

There are a number of nucleophilic additions to C=O known inwhich the added nucleophile still carries an acidic proton (68); asubsequent elimination of H20 then becomes possible, leadingoverall to (69), a net replacement of the oxygen atom:

OHNuH, / -H,O.

R2C=0 , • R2C ~ R2C=Nu,NuH

(68) (69)

By far the most common examples of this are with derivatives of NH 3 ,

particularly those like HONH2> NH2CONHNH2> PhNHNH2, etc.,which have long been used to convert liquid carbonyl compoundsinto solid derivatives, for their characterisation; 2,4dinitrophenylhydrazine, (N02bC6 H3NHNH2> is particularly usefulin this respect.

oMe2C"'" \ H'"/H,o Me2C-OHI Mg-----. I

Me2C-.o' Me2C-OH

(61)

Mg:Mg +=!

(57)

Thus when Na is dissolved, in the absence of air, in ethereal solutionsof aro~atic ketones, a blue colour is seen, due to the presence of thedel,?cahsed (over Ar as well as over the c=o system) species (58), asodIUm ketyl;

[Ar c-osJNa Ell [Ar2c;-OS .-. Ar2C-0'] +=! 21 2NaEil

S Ar2C-Os

(58) (59)

this l.atter is also in equilibrium with its dimer (59), the dianion of a1,2-dlOl or pinacol. Under the right conditions, addition of a protondonor, e.g. ROH, can lead to the preparative formation of the pinacolitself. This tends to work better with aromatic rather than with aliphatic~etones, ~ut propanone (60) may be converted readily by magnesiummto 2,3-dlmethylbutan-2,3-diol (61), so-called pinacol itself:

Me2C=0

OEI OEI S

C'=0 ~ q'-os -OEte Q,IJ Na' (c/O

c=o +---'- I---()6 I' ~ tI ~." sOEI OEI 0 0

(62) (63) (65) II;0 /OH /08

(T ~O g(~CH C C'OH 'oH 'os

(64) (67) (66)

Me2C=0

(60)

P!eparative conversion of ketones (particularly aromatic ones) intopmacols can also be effected photochemically by u.v. irradiation inthe pr~sence of a hydrogen donor, e.g. Me2CHOH.

Similar nucleophilic addition of electrons can also occur to thecarbonyl carbon atom of diesters such as (62), e.g. from sodium in~olvents su:ch as xylene, but the resultant dianion (63) differs from (59)10 possessmg excellent leaving groups, e.g. aOEt, and the overallresult is the acyloin condensation:

220 Nucleophilic addition to C=o 8.4 Carbon nucleophile additions 221

n.R2C=0

.) 4 •

H 2N-Y

Enamines are of some importance as synthetic intermediates.

OH/

CIJCCH"NH 2

(73)

p.209] :

8.4 CARBON NUCLEOPHILE ADDITIONS

In discussing this group of reactions no formal distinction will besought between those which are simple additions and those whichare addition/eliminations. They are considered together, as a groupon their own, because they result in the formation of carbon--earbonbonds, i.e. many of them are of great use and importance in syntheticorganic chemistry. Before considering carbanion reactions in general,however, two specific nucleophilic additions will first be mentioned.

0 0 R,NH OOH -H,O

•• NR 2 • •

(74)

With RNH2

the products are also Imllles; these, too, are usuallyunstable unless one of the substituents on the carbonyl carbon atomis aromatic, e.g. ArCH=NR-the stable products are then known asSchiff bases. With R2NH, the initial adduct (74) cannot lose water inthe normal way; some such species have been isolated but they arenot particularly stable. If, however, the adduct has any IX-H atomsthen a different dehydration can be made to take place yielding anenamine (75) :

Increasing the acidity of the reaction mixture is found to decreasethe rate at which C=O absorption disappears-: NH 20H is being

Ellprogressively converted into HNH 20H, which is not nucleophilic-and to increase very markedly the rate at which the C=N absorptionappears-increasing acid catalysis of the dehydration of (72)~ (71).This is compatible with a reaction pathway of the general form:

.. HOH ((!JOH

/ H$ / H,O4 • R2C 4 • R2C • ! R2C=NY" ~HNY C7Y

H

Strong nucleophiles such as NH 20H (Y = OH) do not requirecatalysis of their initial addition to C=O, but weaker ones such asPhNHNH2 (Y = PhNH) and NH2CONHNH 2 (Y = NHCONH 2)

often require acid catalysis to activate the C=O group (cf p. 204, it is infact general acid catalysis). Often, either the initial addition step or thedehydration step can be made rate-limiting at will, depending onthe pH of the solution. At neutral and alkaline pHs it is generally thedehydration, e.g. (72) -+ (71), that is slow and rate-limiting (cf above),while at more acid pHs it is generally the initial addition of the nucleophile, e.g. (70) -+ (72), that is slow and rate-limiting. This clearly hassignificance in preparative terms, and formation of such derivativesof carbonyl compounds tends to exhibit pH optima-the value depending on the nature of the particular carbonyl compound and of theammonia derivative employed: thus for the formation of an oximefrom propanone, Me 2CO, the optimum pH is found to be ~4·5.

With aldehydes (and with unsymmetrical ketones, RCOR') thereis, of course, the possibility of forming alternative syn and antigeometrical isomers:

It is found in practice that the syn isomer usually predominates;with RCOR' this is the isomer in which Y is nearest to the smallerof the groups, R or R'.

Ammonia itself yields imines, R2C=NH, with carbonyl compoundsbut these derivatives are unstable and react with each other to formpolymers of varying size. The classical 'aldehyde ammonias' are foundto be hydrated cyclic trimers, but from aldehydes carrying powerfullyelectron-withdrawing substituents it is possible to isolate the simpleammonia adduct [73, ct. (72), and hydrates, p. 208, hemi-acetals,

syn anti

8.4.1 Grignard, etc., reagents

The actual composition/structure of Grignard reagents---eommonlywritten as RMgX-is still a matter of some dispute. It appears todepend on the nature of R, and also on the solvent in which the reagentis, or has been, dissolved. Thus the nuclear magnetic resonance (n.m.f.)spectrum of MeMgBr in Et 20 indicates that it is present largely asMgMe2 + MgBr2' while X-ray measurements on crystals of PhMgBr,isolated from Et0 2 solution, indicate that it has the compositionPhMgBr· 2Et20, with the four ligands arranged tetrahedrally roundthe Mg atom. Whatever the details may be, Grignard reagents may beregarded as acting as sources of negatively polarised carbon, i.e. asb-RMgXH.

2238.4.3 Carbanions (general)

method of joining different groups of carbon atoms t~geth~r, i.e. ~he

original alcohol products can then be further modified In a WIdevariety of reactions. In the past ~rgano-zinc c0IJ.lpounds were u~~d

in a similar way, being largely dIsplaced by Gngnard. re~gents, III

turn, Grignard reagents are tendi?g to be displaced by Itthtum alkylsand aryls, RLi and ArLi, respectIvely. These ~atter r~agents .tend to

ive more of the normal addition product WIth stencally hIndered~etones than do Grignard reagents, and also more 1,2- and less 1,4addition with C=C-C=O than do Grignard reagents (cf. p. 201).

8.4.2 Acetylide aniolLS

Acetylenes RC=CH and HC-CH, are markedly acidic and. may beconverted 'by strong bases, e.g. eNH2 in liquid ammonia, Into thecorresponding anions (cf p. 273), which are somewh~ more nucleophilic than eCN. Though these species, e.g. RC-.C , are palpablycarbanions, they are considered separatel~ .as, .unltke the grouP. weshall consider below, they require no stablltsatlOn by electron-wIthdrawing groups such as C=O, C==N, NOz, etc. A ~sef~l grou!, ofcarbon atoms may thus be added to C=O, and the reaction IS especIallyuseful synthetically in that the C-C linkage now present may befurther modified in a variety of ways, e.g. reduced to the alkene (83)by H2 with the Lindlar catalyst (cf p. 191):

XR Mg'RI

R~C, /MgXof

X J+..Mg..R' 'R

-- R~~ ...~gX--"0'

(77)(76)

Nucleophilic addition to C=o222

There is evidence of complexing of the Mg atom of the Grignardreagent with the carbonyl oxygen atom (76), and it is found that twomolecules of RMgX are involved in the addition reaction, in somecases at least, possibly via a cyclic T.S. such as (77):

X/Mg

R Rd++ Id-

R~C ",d-. MgX"'0'

The second molecule of RMgX could be looked upon as a Lewis acidcatalyst, increasing the positive polarisation of the carbonyl carbonatom through complexing with oxygen. It is indeed found in practicethat the addition of Lewis acids, e.g. MgBrz, does speed up the rateof Grignard additions. Reliable details of the mechanism of Grignardaddition to C=O are surprisingly scanty for so well-known a reaction,but pathways closely analogous to the above (i.e. via 77a and 77b)can be invoked to explain two important further observations: (a)that Grignard reagents having H atoms on their p-carbon atom(RCH 2CH 2MgX, 78) tend to reduce C=O ----+ CHOH (79), beingthemselves converted to alkenes (80) in the process (transfer of Hrather than RCH2 CH 2 taking place):

(b) that sterically hindered ketones having H atoms on their IX-carbons,e.g. (81), tend to be converted to their enols (82), the Grignard reagent,RMgX being lost as RH in the process:

Grignard reagents act as strong nucleophiles and the additionreaction is essentially irreversible. The end-products of addition,after aqueous hydrolysis (of, for example, R 3C-OMgX), are alcohols (R3C-OH). It is, however, important to emphasise that theutility of Grignard, and similar, additions to C=O is as a l!eneral

(861

R2C=CYZ

(i,e. x= H)

Ill,\

OH/

R1C ,.

"CXYZ

(QA\

,. 0 6

R2C=0 / R'OHr • · R2C ,.

6CXYZ "CXYZ

OH OH/ H, R C/

R2C Lindta! 1 "

"C=CR' calalyst CH=CHR'

(83)

8.4.3 CarbaniolLS (general)

In general these reactions are base-catalysed in that it is necessar~ toremove a proton from HCXYZ in order to generate the carbamon,eCXYZ, the effective nucleophile; one or ~?re.of X, :. ~nd Z areusually electron-withdrawing in order to stablltse It. The InItial ad~uct(84) acquires a proton from the solvent (often H 2 0 or R.OH) to yIeldthe simple addition product (85). Whether or n~t t~I.S undergoessubsequent dehydration (86) depends on the avatlablhty of an Hatom, either on an IX-carbon or where X, Y or Z = H, and ~lso onwhether the C=C so introduced would, or would not, be conjugatedwith other C=C or C=O linkages in the product:

(79)

R H" /C""CH (80)

2HI

R~C, /MgXo

HR C 'R2

11R'C, /MgX

o(82)

~2C H R J+

-- I --R'C",O MgX

(77b)(81)

HR c/ R

2 I IR'C M X'::::-0' g

R H\. ,/

CH/ 'CH

I 2

R~C.::::- .. ' MgXo (78)

224 Nucleophilic addition to C=O 8.4.4 Aldol reactions 225

The initial carbon-earbon bond formation (---+ 84) is often reversible,and a subsequent step-such as dehydration-may be necessary todisplace the equilibrium. The many different (often named) reactionsreally differ from each other only in the nature ofthe particular carbonylcompound (aldehyde, ketone, ester, etc.), and in the type of carbanion,employed.

8.4.4 Aldol reactions

Here the carbanion (87), obtained from the action of base (usuallyeOH) on an IX-H atom of one molecule of a carbonyl compound (88),adds to the carbonyl carbon of another (88) to yield a fJ-hydroxycarbonyl compound. Thus with ethanal, CH3CHO, the product is3-hydroxybutanal (89)-aldol itself:

HI

(88) CH 2CHO

(I) t!eOH

o oe(II (; (2) I H,O

Mel eCH 2-l

=0 ~ MeCH-CH2CHO ~ MeCH(OH)CH 2CHO

H H (89)(88)

CH2=C-OeI

H

(87)

In the case of CH3CHO the equilibrium is found to lie right over infavour of aldol. The forward reaction of step (2) and the reversal ofstep (I) are essentially competing with each other for the carbanion(87). Carrying out the reaction in D 20 fails to result in the incorporation of any deuterium into the CH3 group of as yet unchangedethanal, however, so that step (2) must be so much more rapid thanthe reverse of step (1) as to make the latter virtually irreversible.

For even simple ketones, e.g. propanone (90), the equilibrium isfound to lie far over to the left (~2 %of 91 )-reflecting the less readyattack ofthe carbanion (92) on a 'keto' (90), rather than on an 'aldehydo'(88, above), carbonyl carbon atom:

HI

(90) CH 2COMe

(I)t! eOH

,0 eoLII C' (2) I H,O

Me2C eCH 2COMe ~ Me 2C-CH2COMe ~ Me2C(OH)CH 2COMe(90) (92) (91)

Thus it is found in the case of propanone (90) that carrying out thereaction in D 20 does result in the incorporation of deuterium intothe CH3 group of as yet unchanged propanone, i.e. step (2) is no longerrapid with respect to the reversal of step (I).

The reaction can, however, be made preparative for (91) by acontinuous distillation/siphoning process in a Soxhlet apparatus:equilibrium is effected in hot propanone over solid Ba(OHh (as basecatalyst), the equilibrium mixture [containing ~2% (91)] is thensiphoned off. This mixture is then distilled back on to the Ba(OHh,but only propanone (b.p. 56°) will distil out, the ~2%of 2-methyl-2hydroxypentan-4-one ('diacetone alcohol', 91, b.p. 164°) being leftbehind. A second siphoning will add a further ~2% 'equilibrium'sworth' of (91) to the first 2 %, and more or less total conversion of(90) ---+ (91) can thus ultimately be effected. These 'poor' aldol reactionscan, however, be accomplished very much more readily under acidcatalysis. The acid promotes the formation ofan ambient concentrationofthe enol form (93) of, for example, propanone (90), and this undergoesattack by the protonated form of a second molecule of carbonylcompound, a carbocation (94):

(90) Me2C=0

t! H"'

(I)~ (I) -H"'Me2C CH 2=CMe« I Me2C-CH2-CMe« I Me2C(OH)CH 2CMe

I I I 17) IIOH OH OH O-H 0

(94) (93) (91)

Under acid conditions the tertiary alcohol (91) almost alwaysundergoes acid-catalysed dehydration (cf p. 247) to yield the IXfJunsaturated ketone, 2-methylpent-3-ene-2-one (mesityl oxide, 95):

H (I) +H"'I (2) -H,O

Me 2C-CHCOMe « • Me2C=CHCOMeI

OH

(91) (95)

Dehydration of aldols may also be effected under the influence ofbase, e.g. with aldol itself (89) to but-2-eneal (crotonaldehyde, 96):

HI eOH C'e -OHe

MeCH-CHCHO « I MeCH-CHCHO « • MeCH=CHCHO

6H C6H(89) (97) (96)

Base-catalysed dehydrations. are relatively unusual, and that oneoccurs here stems from the facts: (a) that (89) contains IX-H atomsremovable by base to yield an ambient concentration of the carbanion(97), and (b) that this carbanion possesses a goodish leaving group-

226 Nucleophilic addition to C=O 8.4.6 Perkin reactions 227

8.4.6 Perkin reaction

l-Meco2H

OH oe. I ~!H20 I

Ph~O~ Ph~':::::::,O

nos) nOSa)

(102)

eoI H,O

P R2C-CH 2N02 P R2C(OH)CH 2N0 2

o 0(II n ,f"

R C eCH-N2 2 Ell "-

Oe

!

compounds, e.g. nitromethane (10 I):

HI

(101) CH 2N02

t~eoH

In this reaction the carbanion (103) is obtained by removal of antx-H atom from a molecule of an acid anhydride (104), the anion of thecorresponding acid acting as the necessary base; the carbonyl acceptoris pretty well confined to aromatic aldehydes. The products are txfJunsaturated acids, e.g. 3-phenylpropenoic(cinnamic) acid (l05) fromPhCHO/excess (MeCO)zO/MeCOz

8 at 140°:(Qe

~ U X'6 rr~ recH;lo-PIi O-Ph~O-Ph¥O

Ph~ MeC026 \H

(103) (106a) (107) (l06b)

-~I

01

MeC02e")~ 1CH~

IHl4\

(100)

Bases such as eOH and eOEt are used to obtain the carbanion, andwhether or not the fJ-hydroxynitro compound (102) undergoes subsequent dehydration to RzC=CHNO z depends on the conditions.Where the carbonyl compound is an aldehyde there is some dangerof its undergoing an aldol reaction with itself, but the de10calised

ecarbanion (100) usually forms more readily than does RCHCHO, andthe danger is thus relatively small. The product nitro compoundsmay be reduced to amines, and also modified in other ways.

(99)

e~:>l~-~n the adjacent (fJ-) carbon atom. The possibility of such anelImmatlOn may displace the equilibrium over to the right in a numberof simple aldol additions, where it would otherwise lie far over to theleft. It is ~~portant to remember, however, that the overall processaldol addItIOn + dehydration is reversible, i.e. (88)~ (96), and thattxfJ.-unsaturat~~carbonyl compounds are thus cleaved by base undersUItable condItIOns. It is also pertinent that (96) is still an aldehyde andcan under~o.furthercarbanion addition, followed by dehydration, andso on. ThIs IS how low molecular weight polymers are produced onheating simple aliphatic aldehvdes with aqueous NaOH: to stop atthe aldol, the best catalysts are basic ion-exchange resins.

Crossed aldol condensations, where both aldehydes (or othersuitable car~onyl compoun~s) have tx-H atoms, are not normally ofany preparatIve value as a mIxture of four different products can result.Crossed aldol reactions can be of synthetic utility, where one aldehydehas no tx-H, .however, and can thus act only as a carbanion acceptor.An ex~mpl~ IS the C~aisen.-Schmidtcondensation ofaromatic aldehydes(98) wIth sImple alIphatIc aldehydes or (usually methyl) ketones inthe presence of 10 %aqueous KOH (dehydration always takes placesubsequent to the initial carbanion addition under these conditions):

ArCH=CHCHOeCH,CHO /'

r' 90H

(98) ArCHO

eCH,co~OHArCH=CHCOMe

As would be expected, electron-donating groups in Ar slow the reaction,e.g. p-MeOC6H 4 CHO reacts at only about one-seventh the rate ofC6H5C~O.Self-condensation of the aliphatic aldehyde can, of course,be an Important competing reaction under these conditions butCannizzaro reaction of ArCHO (cf p. 216) is so much slower a rea~tionas not to be a significant competitor. The condensation can also beeffected under acid catalysis (ct. p. 225).

Finally, a.ldol reactions c.an, with suitable dicarbonyl compounds,e.g. (99), be mtramolecular, I.e. cyclisations:

8.4.5 NitroaIkanes

Another synthetically useful reaction involves the addition to aldehydesand ketones of carbanions. e.g. (I (0), derived from aliphatic nitro

229

(3) tl.90EI

[eO 0 ]I II e

MeC=CHC0 2 Et ..... MeC-CHC0 2Et

(115)

8.4.8 Claisen ester condensation

t

o (I) tl. 90El ~O 0 H(II ~ (2) I II I

MeC ~ eCH 2-C=0 ~ MeC-CH 2C0 2Et P MeC-CHC0 2EtI I ,I

OEt OEt "'OEt (114)

(112) (113)

CH 2=C-OeI

OEt

(111)

One significant difference from the simple aldol reaction, however, isthat the original adduct (113) now possesses a good leaving group(OEt); thus instead of adding a proton, as in the aldol reaction proper(p. 224), aOEt is lost to yield a fJ-ketoester, ethyl 3-ketobutanoate(ethyl acetoacetate, 114). This is finally converted by base (aOEt)into its stabilised (delocalised) carbanion, (115).

Classically the base catalyst, aOEt, is introduced by adding justover one mole of sodium (as wire, or in other finely divided form)plus just a little EtOH to generate an initial small concentration ofNaEBaOEt. Further EtOH is generated in step (I), which yields furtherNaEBaOEt with sodium, and the concentration of aOEt is therebymaintained. A whole mole is required as it is essential for the fJ-ketoester (114) to be converted (step 3) into its anion (l15}-MeCOCHzC02Et is more acidic than EtOH (cf p. 272}-if the overall successionof equilibria is to be displaced to the right. This is necessary becausethe carbanion-formation equilibrium-step (I)-lies even further overto the left than that with, for example, CH3CHO; this reflects theless effective stabilisation through delocalisation in the ester carbanion(III) than in that from the aldehyde (116):

eCH -c=o ..... eCH -c-oe eCH -c=o ..... CH 2=C-Oe2 I 2 II 2 I I

: OEt 190Et H H

(111) (116)

This requirement to pull the equilibrium of step (I) over to theright is reflected in the fact that no reaction occurs with R2CHCOzEt

8.4.8 Claisen ester condensation

This is another reaction that involves carbanions derived from esters,e.g. (III), but this time adding to the carbonyl carbon atom of anotherester molecule. The reason for considering it here rather than undercarboxylic derivatives (p. 237) is that it can, in its initiation, be regardedas something of an analogue, for esters, of the aldol condensation onaldehydes (ct. p. 224), e.g. with ethyl ethanoate (acetate, 112):

H1

(112) CH 2C0 2Et

....Q:'C02EtR2C CH 2

\ I ~o-c-oe

C6Et

tl. -OEt 9

CHC0 2EtR2C"" 'CH 2

1 I n.°eJ 1=0

OEt(108)


C0 2 Et1

R2C=CCH 2C0 2 e q

(109)

228

8.4.7 Knoevenagel and Stobbe reactions

This addition involves carbanions from a wide variety of CH2XY

types but particularly where X and/or Yare C02R groups, e.g.

CH2(C0 2Eth; organic bases are often used as catalysts. In mostcases the intermediate aldol is dehydrated to the cxfJ-unsaturatedpro~uct (ester). An interesting example is with carbanions, e.g. (108),derIved from esters of l,4-butandioic(succinic) acid, e.g. (CH

2C0

2Eth,

and aldehydes or ketones, employing alkoxide ions as base catalysts:the Stobbe condensation. These esters react a great deal more readily~han might, a priori, have been expected, and one of the C0

2R groups

IS always, and unexpectedly, converted to CO2a in the course of thereaction; the product is always the cxfJ-unsaturated derivative (109),never the aldol. A pathway that will account for all these facts involvesa cyclic (lactone) intermediate (110):

The carbanion (103) attacks the carbonyl carbon of the aldehyde inthe usual way to yield the alkoxide anion (106a). Internal transferof the acetyl group in this anion is believed to take place: from thecarboxyl oxygen atom (in 106a) to the alkoxy oxygen atom (in106b), via the cyclic intermediate (107); thereby forming a morestable species. Removal of an a-H from this anion by MeC0

2e

results in loss of the good leaving group MeC02 e from the adjacentl3-position to yield the anion (105a) of the al3-unsaturated acid.Work up of the reaction mixture with dilute acid leads to theproduct (105).

Some support for this mechanism is provided by the observationthat on reaction with anhydrides of the form (R2CHCOhO-wherethere would be no a-H to remove in the intermediate correspondingto (106b)-it is possible to isolate the analogue of (106b) as theactual end-product of the reaction.

....q::'C02 Et90Et R2C CH

2• , \ Io-c

~o

(110)

It has, in a few cases, proved possible actually to isolate cyclic intermediates such as (110).

in the presence of eO£1, despite the fact that a normal p-ketoester,R2CHCOCR2C02Et, could be formed. It is significant, however, thatthis product p-ketoester has no oc-H atom, and so cannot be convertedinto a carbanion corresponding to (115), i.e. step (3) cannot take place!Use of a base, e.g. Ph 3Ce NaEll , that is sufficiently strong to make step(I) virtually irreversible in the forward direction

8R2CHC02Et + Ph 3C8 P RCC02Et + Ph 3CH

is found to induce a normal Claisen reaction in R2CHC02Et, despitethe fact that step (3) is still impossible.

It is important to emphasise the complete reversibility of norma]Claisen reactions under suitable conditions, e.g. the so-called 'aciddecomposition' (because both products-( 117) and (l18)-are derivatives of acids) of p-ketoesters (119):

Q? <:9 ~ 8RC-CHR'C02Et +=! RC~CHR'C02Et +=! RC + R'CHC0 2Et

c;"OEt dEt dEt(119) (117) (118)

1,3-(i.e. p-)diketones, e.g. (120), are also cleaved under these conditionsto yield a derivative of an acid (121), and one of a simple ketone (122):

o 80 0(II <"';1 II 8RC-CH 2COR +=! RC~CH2COR +=! RC + CH 2COR

C;OEtdEt dEt(120) (121) (122)

'Crossed' Claisen reactions with two ditlerent esters, each of whichhas oc-H atoms, are seldom useful synthetically as there are, of course,four possible products. Crossed Claisen reactions are, however, oftenuseful when one of the two esters has no oc-H atoms, e.g. HC0 2Et,ArC02Et, (C02Eth, etc., as this can act only as a carbanion acceptor.Such species are in fact good acceptors, and the side reaction of theself-condensation of the other, e.g. RCH 2C02Et, ester is not normallya problem. Intramolecular Claisen reactions, where both C02Etgroups are part of the same molecule [e.g. (123)], are referred to asDieckmann cyclisations. These work best, under simple conditions, forthe formation of the anions of 5-,6- or 7-membered cyclic p-ketoesters[e.g. (124)], i.e. with Et02C(CH2)nC02Et where n = 4-6:

EtO, EtO, C;o Et«r (08C=O SOEt c:7 C

H2Co'" CH 2C02Et l ,H2Co'" C,8CHC02Et +==t H2Co'" 'CHC02Et\ I \ I \ I

H2C-(CH 2). H2C-(CH 2). H2C-(CH 2).

(x = 1-3) (123) i! -OE,6

~ 0 0I II II

o"'C:".. o"'C,8 60E, CH2C "'CC02Et +-+ H2C CC02Et!:::::::::i H2Co'" 'CHC02Et

\ I \ I \ IH2C-(CH 2lx H2C-(CH 2). H2C-(CH 2).

(124)

Rate = k[ArCHO]2[8CN]

and the reaction is believed to follow the general pathway:

Big ring ketones (cf the acyloin condensation, p..218) may be obtainedalso by working at high dilution, i.e. the carbamon carbon atom thenhas a greater chance of reacting with the .ester carbon~1 carbon atomat the other end of its own chain than wIth one that IS attached to adifferent molecule (intermolecular reaction).

231

(3) tl. A,CHO

HO 0 8

I IArC-CAr

I ICN H

(129)

+...

H? ~)ArC8""\ CAr

1 1C=N H

(l27a) (128)

HO1

ArC~=N8 (127b)

8.4.9 Benzoin condensation

80 OH<"'1 I

ArC-CAr( I 1

CN H

(130)

o 80(II SCN I

ArC-H· , ArC- H +=!l. (I) I (21

8CN CN

(125) (126)

o OHII I _CNS

ArC-CAr· ,1

+ CN8

H

(131)

When the reaction is carried out in MeOH neither step (2), the formation of the carbanion (127), nor step (3), addition of this carbanion tothe carbonyl carbon of the acceptor molecule (128), is completelyrate-limiting in itself. These steps are followed by rapid proton trans~er,(129)~ (130), and, finally, by rapid loss of eCN-a good leavmge:rouo--i.e. reversal ofcyanohydrin formation (ef p. 212) on the product

8.4.9 Benzoin condensation

This reaction ofaromatic aldehydes, ArCHO, resembles the Cannizzaroreaction in that the initial attack [rapid and reversible-step (I)] isby an anion-this time eCN---on the carbonyl carbon atom of onemolecule, the 'donor' (125); but instead of hydride transfer lef.Cannizzaro, p. 216) it is now carbanion addition by (127) to the carbonylcarbon atom of the second molecule of ArCHO, the 'acceptor' (128),that occurs. This, in common with cyanohydrin formation (p. 212)was one of the earliest reactions to have its pathway establishedcorrectly!-in 1903. The rate law commonly observed is, as might beexpected,


232 Nucleophilic addition to C=O 8.4.11 Wittig reaction 233

2-hydroxyketone (131). Where Ar = Ph, the product is called benzoin.The reaction is completely reversible.

eCN was for long the only species known to catalyse this reaction.It was thought to owe this capacity to: (a) its ability as a nucleophile; (b) its ability as a leaving group; but, most particularly, (c)its ability, through electron-withdrawal, to increase the acidity of theC-H bond in (26) and to stabilise the carbanion 027a~127b)that results from loss of this proton. More recently it has been foundthat the ylids (32), formed in low concentration at pH=7 insolutions of N-alkylthiazolium salts (lacking a 2-substituent),

H

R_~As

ye )=<(132)

are excellent catalysts for the benzoin condensation. Ylids, whichare species which have charges of opposite sign on adjacent atoms,seem a far cry from eCN, but it is significant that they would indeedbe expected to fulfil roles (a), (b) and, most notably, (c) particularlywell.

8.4.10 Benzilic acid rearrangement

Oxidation of benzoin, PhCH(OH)COPh (above) yields benzil,PhCOCOPh (33), and this, in common with non-enolisable 1,2diketones in general, undergoes base-catalysed rearrangement toyield the. anion of an a-hydroxy acid, benzilate anion,PhzC(OH)COze (34). This is, almost certainly, the first molecularrearrangement to be recognised as such. The rate equation is foundto be,

Rate = k[PhCOCOph][eOH]

and the reaction is believed to follow the general pathway:

Ph Ph Ph Phf") I e'l k.. (">. I I

O=C-C=O i=! O-C-C=O - O=C-C-Oe p O=C-C-OH~ I I I I I I I

HOe Ph HO Ph HO Ph eO Ph

(133) (135) (134)

The slow, rate-limiting step is almost certainly the migration of phenylthat occurs in the initial eOH adduct (135). This is essentially theanalogue for 1,2-diketones of the intramolecular Cannizzaro reaction on the 1,2-dialdehyde glyoxal, OHCCHO (P. 217). In the latter

it was an H atom that migrated with its electron pair, i.e. as hydrideion, to the adjacent C=O group, whereas in benzil (33) it is Phthat migrates with its electron pair, i.e. as a carbanion; hence thereason for considering this reaction as an (intramolecular) carbanionaddition to C=O.

the 1,2-ketoaldehyde, PhCOCHO, also reacts with eOH,thereby being converted into PhCH(OH)COz

e . This same productcould, however, be formed by migration of either H or Ph; experiments with suitably D, and I4C, labelled PhCOCHO show that, infact, it is only H that migrates. There seem to be no examples of t.heequivalent of intermolecular Cannizzaro reactions on ketones, 10

volving, as they necessarily must, migration of R with its electronpair from one molecule to another, i.e. 2RzCO~RCOze + R3COH.

8.4.11 Wittig reaction

This is an extremely useful reaction for the synthesis of alkenes. Itinvolves the addition of a phosphonium ylid, e.g. (36), also knownas a phosphorane, to the carbonyl group of an aldehyde or ketone;the ylid is indeed a carbanion having an adjacent hetero atom. Suchspecies are generated by the reaction of an alkyl halide, RR'CHX(37), on a trialkyl- or triaryl-phosphine (138)-very often Ph3P-toyield a phosphonium salt (39), followed by abstraction of a protonfrom it by a very strong base, e.g. PhLi:

<ll PhLi <ll ePh 3P + RR'CHX -- Ph 3P-CHRR' -- Ph 3P-CRR'

(138) (137) xe (139) t

Ph3P=CRR'

(136)

Addition of the Wittig reagent (136) to C=O, e.g. (140), is believedto follow the pathway:

(140) R;~fo ~ R;l-~ ~ R;l.I~ ~ R;rr ~RR'C-PPh 3 RR'C-PPh 3 RR'CI"PPh 3 RR'C PPh 3

<ll <ll

(136) (141) (142)

All Wittig reactions do not, however, follow the same detailedpathway: step 0) may, or may not, be an equilibrium, and therate-limiting step may also differ. It is, in some cases, possible todetect (at -80°) the oxaphosphetane intermediate (41), which thendecomposes to products on raising the temperature (_0°). Thedriving force underlying this rather unusual reaction is the largeamount of energy that is evolved through forming the very strongP=O bond (535 kJ mol-I). If the original phosphonium salt (39) is

235

(±)

+ ya

+ ya

M S

~y:YY%

R (154)

Hi'-Hi'-

M S

HO SR' l ,R-XM~ "HOy: L '%

R L R (153)

More favoured product

~~OH "

R

Less favoured product

(148)

8.5 Stereoselectivity in carbonyl addition reactions

M~R'MgBryR'MgBr

R(152)

Addition of HY to RR'C=O introduces a chiral centre into the adduct(151), but the product will always be the (±) form-the racemate(151 ab}-because initial nucleophilic attack from above (a), or below(b), on the planar carbonyl compound (148) will be statistically equallylikely:

This is saying, reasonably enough, that preferential reaction willtake place via the less crowded (lower energy) TS. We should thus

If, however, R or R' is chiral-and particularly if this stems fromthe a-carbon atom-then the two faces of the carbonyl compound(148) are no longer equivalent, and addition from above and belownot therefore statistically equally likely. Where the reaction is reversible it is likely that there will be a greater proportion of the thermodynamically more stable of the two alternative products in the reactionmixture (thermodynamic or equilibrium control, cf p. 43). Foressentially irreversible reactions, e.g. with RMgX, LiAl~, etc., theproduct that is formed the more rapidly is likely to preponderate(kinetic control); which this is likely to be can often be forecast fromCram's rule: a ketone will react in that conformation in which the°of the C=O group is anti to the largest of the three substituentson the a-carbon atom (152). Preferential nucleophilic attack (e.g. byR'MgBr) will then take place from the least hindered side of thecarbonyl carbon atom, i.e. (a). This is best seen using Newmanprojection formulae (cf. p. 7):

HY Ik'0

R R'

(149)

III

y °K~

R R'

(150)

~RyO MY

R' TRA~

(148)

yl.R R'

(146)

0'R R'

(147)


Ell 8R2C=0 + Ph3P-CHCH 2C028 --+ R2C=CHCH2C028 + Ph3P=0

(144)

by other methods. In the case of (144), most methods tend to induceisomerisation to the thermodynamically more stable, conjugated (afJunsaturated) acid. The Wittig reaction has also been used intramolecularly to prepare cyclic alkenes containing 5-16 carbon atoms.

chiral (at phosphorus), e.g. RR'R"plIlCHzR, the configuration at thephosphorus atom is found to be retained in the related phosphineoxide, RR'R"P=O, product.

Because of the variations possible in the R groups of the originalhalide (137) and in the carbonyl component (140), this is an extremelyuseful and versatile method for the synthesis of substituted alkenes.The presence of C=C or C=C, even when conjugated with the C=Ogroup, does not interfere. A COzR group, though it will react with theylid (136) does so very much more slowly than with C=O, and thusdoes not interfere either. The reaction is particularly valuable forgetting a double bond into positions that are difficult, e.g. exocyclicmethylene groups (143),

0 0 OCH2+ Ph3~-~H2 --+ + Ph3P=0

(143)

or all but impossible, e.g. fJy-unsaturated acids (144),

8.5 STEREOSELECTIVITY IN CARBONYL ADDITIONREACTIONS

Whether nucleophilic addition of HY to C=O proceeds stereoselectively CIS or TRANS clearly has no meaning-unlike e1ectrophilic addition of HY to C=C [(145) ---+ (146) or (l47)]-for withC=O [(148) ---+ (149) or (150)] the alternative products are identical,because of free rotation about their C-O bonds:

R

~cy

R~R'r Hi'

R' TRA~

(145)

MJyL~I

R

(155)

irrespective of the size, relative to Cl, of the other groups attached tothat a-carbon atom. Either of these types of effect may outweighpurely steric considerations in determining the geometry of thepreferred T.S.

8.6 ADDmON/ELIMINATION REACTIONS OFCARBOXYLIC DERIVATIVES

The general reactions of this series are of the form:

o eo 0(II ye c.. I 0! _xe II

R-C-X ~ R-C-X ~ R-C + xe

v;;> ~ ~(156) (157)

The reaction pathway is normally nUcleophilic addition/elimination,via a so-called tetrahedral intermediate (157), leading to overallsubstitution. The difference between the reactions of carboxylic

expect the ratio x/y to increase: (i) as the difference in size betweenM and S increases, and (ii) as the size of R' in R'MgBr increases. Inpractice, it is found that for attack of MeMgI on the aldehyde C

6Hs

(Me)CHCHO (152, L = C6 Hs' M = Me, S = H, R = H) x/y = 2: I,while replacing Me by the rather bulkier Et raises x/y to 2·5: I. Similarly,replacing MeMgI by the much bulkier C6 HsMgBr for attack on theMe compound (152, M=Me) is found to raise the x/y ratio to >4: l.

The operation of Cram's rule has been investigated very largely forGrignard additions, and some hydride additions, to C=O ; in generalit works quite well at forecasting which will be the more favouredproduct, but there are a number of exceptions. This is hardly surprising~or the. rule assumes that product control depends only on stericmteractlOns, whereas complex formation-between groups in thesubstrate, e.g. hydrogen bonding, or between substrate and attackingnucleophile, e.g. RMgX and carbonyl oxygen atom-and dipole/dipoleinteraction may also playa part. As an example of the latter effecta-chioro-aldehydes and -ketones are found to react (because ofelectrostatic repulsion) in that conformation, e.g. (155), in which Cland carbonyl oxygen atom are anti to each other,

237

eoI

CF -C-OEt) IOEt

(160) (159)

This adduct (=100% yield) may be isolated and characterised; theless nucleophilic H20 or EtOH does not add on. On going from theoriginal carboxylic derivative (156) to the tetrahedral intermediate

8.6 Addition/elimination reactions of carboxylic derivatives

The rate law followed by these reactions is generally of the form,

Rate = k[RCOX][ye]

and the question arises whether they might perhaps proceed by adirect, one step (cj. SN2) displacement on the carbonyl carbon atom.It is not normally possible to isolate tetrahedral intermediates suchas (157), but it has proved possible to obtain evidence of the formationof one where R carries strongly electron-withdrawing atoms orgroups (ct. 03CCHO, p. 208), Le. (159) from the addition of eOEt(in dibutyl ether) to CF3C02Et (160):

o<: II SOBICF)-C-OEt 4 ~

C;OEt

Thus acid chlorides and anhydrides react readily with ROH and NH3to yield esters and amides, respectively, while esters react with NH3or amines to give amides, but the simple reversal of any of these reactions on an amide, though not impossible, is usually pretty difficult.The relative reactivity will also depend on both the electronic and,more particularly, the steric effect of R. A slightly unusual leavinggroup is eCX3 (e.g. eCI3) in the haloform (158) reaction (cj. p.297):

o eo 0 0(II eOH c.. I II II

R-C-CX)4 ~R-C-Q:X) -+ R-C + ecx) ~ R-C + HCX)C I I IeOH OH OH Oe

(158)

derivatives (156) and those of simple carbonyl compounds (aldehydesand ketones) stems from the fact that in carboxylic derivatives thereis, attached to the carbonyl carbon atom, a group X which is a goodpotential leaving group (as Xe); whereas in simple carbonyl compounds the potential leaving group (Re or He) is very poor indeed.The relative reactivity of the series (156, with differing X) towards aparticular nucleophile ye (e.g. eOH) depends on: (a) the relativeelectron-donating or -withdrawing power of X towards the carbonylcarbon atom, and (b) the relative ability of X as a leaving group.The reactivity series is not necessarily exactly the same for everyye, but in general it follows the order:

o 0 0 0 0II II II II II

RC-C1> RC-OCOR > RC-OR' > RC-NH 2 > RC-NR~


(157), the carbonyl carbon atom changes its hybridisation fromSp2~ Sp3 and, in so far as the T.S. of the rate-limiting stage of theoverall reaction resembles (157), we might expect these reactions tobe susceptible to steric effects: such is indeed the case (see below).

We have already discussed carbanion addition to RC0 2Et (Claisenester condensation, p. 229), and also their reduction by LiEil AlH

4e

(p. 214); some further examples of nucleophilic attack will now beconsidered.

established that this normally undergoes acyl-oxygen cleavage (c/.p. 47), i.e. 180 label is found only in EtOH. This supports thetetrahedral intermediate pathway, via (165), discussed above:

eo 0 0(.1 9 II II

R-C-eDEt -DEI. R-C + 180Ete~ R-C + H 180EtI I \OH OH 0e

239

(167)(166)

8.6.2 Some other nucleophiles

(165)

o(II

R-C_ 180Et

C;OH

(164)


8.6.1 Grignardt etc. t reagents

Attack of.Gr.ignard reagents on esters, e.g. (161), follows the generalp~th.wa~ lOdlcate? abov~, so that the initial product of addition/elImmatlOn (eOR as leavmg group) is a ketone (162):

III III

c.~ (161)" ~? MgX 0 oe MgXR-C-OR' ~ R-C-OR' ~R-~ R"MgX. R-t-R"~ IV I I

R"-MgX R" R" R"

(162) (163)

The carbonyl carbon atom of (162) is, however, more reactive towardsnucleophiles than that of the original ester (161), because of the electrondonating mesomeric effect, in the latter, of the ester oxygen atom:

As soon as it is formed, (162) thus competes preferentially with as yetunreacted ester (161) for Grignard reagent, R"MgX, and the actualend-product of the reaction is the salt of a tertiary alcohol (163);two of its alkyl groups having been supplied by the Grignard reagent.H~rdly ~urprisingly, acyl halides, e.g. RCOCI, yield the same productswith Gngnard reagents, but the reaction can in this case be stoppedat the ke~one stage by use of organo-cadmium compound, CdR;.The reactIOn also stops at the ketone stage with esters when R"Li isused at higher temperatures in place of R"MgX.

oII

R-C + OEte

6R(168)

Attack by eOH on amides, RCONH 2 , follows an analogous courseto that with esters (above) except that here eNH 2-rather than eOEtis the leaving group. This removes a proton from (166) to form themore stable pair of carboxylate anion (167) +NH3 ; loss of the latterfrom the hot, basic solution tends to drive the reaction over to theright. The attack of amines, RNH2 , on esters, e.g. (164), to formamides (169) follows very much the same general course as theexamples above (it has been shown that RNHe , the conjugate baseof RNH2 , is not involved in nucleophilic attack on the ester):

The rate-limiting step is almost certainly attack of eOH on theoriginal ester (164). This is borne out by the activation parametersfor the base-induced hydrolysis of MeC02Et: aH+ = 112 kJ mol-I;as+ = -109 J K-I mol-I. The relatively large -ve value of as+indicates the decrease in translational entropy (cf. p. 35) characteristic of two separate species (MeC02Et +eOH) combining (an associative process) to form the T.S. in the rate-limiting step of theoverall reaction: formation of (165). The overall reaction is essentially irreversible as eOEt would remove a proton from (166) ratherthan attack its carbonyl carbon atom, while the carboxylate anion(167) will be insusceptible to nucleophilic attack by EtOH or EtOe .This mechanism is generally referred to as BAC2 (lJase-catalysed,acyl-oxygen cleavage, bimolecular). Where nucleophilic attack is byeOR, rather than eOH, transesterification occurs, and an equilibrium mixture of both esters, (164) + (168), is obtained; the positionof equilibrium depends on the relative concentrations and nucleophilic abilities of eOEt and eOR:

o eo(II 90R (.1 ~ -OEI9

R-C-OEt • • R-C-OEt • •

CeOR 6R(164)

O~-

IIR-C-OR'

H0'(161)(162)

O~-

IIR-C-R"

H+

8.6.2 Some other nucleophiles

A reaction that has been much investigated is the hydrolysis of esters,e.g. (164), by aqueous base, i.e. eOH. It is found to be kineticallysecond order, and 18 0 isotopic labelling experiments on (164) have

o(II R'NH,

R-C-OEt. •lo'

H1NR'(164)

eoI R'NH

R-C-OEt • ~I

H2~R'

eo 0I BH II

R-C-OEt~ R-C + HOEt + BeI I

HNR' HNR'

(170) (169)

o HOEll HOjII -He II -H,O I Ell

R-C , • R-C - R-C~()HI 1-- 11j"2OEt OEt OEt

(172) (174) (176)

N.m.r. spectra support preferential protonation of the carbonyloxygen of the acid (173) in the forward reaction (esterification), and

8.6.3 Acid~atalysed reactions

It is difficult to effect attack on the carbonyl carbon atom of RCO~H,(171), with nucleophiles of the general type ye, as they commonlyremove proton instead, and the resultant RCOze is then insusceptibleto nucleophilic attack. Weaker nucleophiles of the form YH, e.g. ROH,do not suffer this inability, but their reactions with the relativelyunreactive carbonyl carbon atom of RCOzH are slow. The carbonylcharacter may be enhanced by protonation, however, i.e. by acidcatalysis in, for example, esterification [(171) --+ (172)]:

The slow, rate-limiting step seems to be loss of the leaving groupfrom (170), and this normally needs the assistance of a proton donorBH, e.g. H 20.

Acid chlorides, RCOCI, undergo ready attack by weaker nucleophiles, e.g. H20, ROH. The question then arises whether, withthe better potential leaving group C1 e , the reactions of acidchlorides could proceed either via a single step 'SN2 type' pathway(cf. p. 78) involving a T.S., in which attack by ye and loss of a e areessentially synchronous; or via an 'SN 1 type' pathway (cf. p. 79) inwhich the slow step is RCOCI~ RCO E9C\e, followed by fast attackby ye on the acyl cation, RCO E9

• In fact, most reactions of acidchlorides probably proceed via the now familiar 'tetrahedral intermediate' pathway, though there may be some exceptions.

Acid anhydrides, (RCO)20, will also often react with weakernucleophiles, though more slowly than acid chlorides; neither 'SN 1nor SN2 types' of reaction pathway normally occurs. Anhydrides areessentially intermediate in reactivity-towards a particular nucleophile-between acid chlorides and esters, reflecting the leaving groupability sequence:

241

(177)

t1. H,O

8.6.3 Acid-eatalysed reactions

(178)

HO HOI -He I Ell

RC= 180 + HO-CMe3~ RC= 180 + H 20-CMe3(180) (179)

The activation parameters for the acid-catalysed hydrolysis ofMeC02CMe3 are found to be: li.H+ = 112 kJ mol-I; li.S+ =+55 J K-1mol-I. The now +ve value of li.S+ (indicating an increasein translational entropy in forming the T.S. for the rate-limitingstep) suggests that this step is a dissociative process: as exemplifiedin the reaction pathway above by breakdown of the protonated esterinto two separate species, the carboxylic acid and the carbocation(177). This mechanism is generally referred to as AAL 1 (~cid

catalysed, alkyl-oxygen cleavage, unimolecular), it also occurs withester alkyl groups such as Ph2CH, etc. When attempts are made totransesterify (178) with R'OH, the product is not now the expectedester, RC02R', but RC02H plus, the ether R'OCMe3; the latterarises from attack of R'OH on the carbocationic intermediate (177),cf. the conversion of (177) to (179) by H20 above.

Where the acid alkyl group, R in RC02R', is sufficiently bulky,e.g. R3C (181), that bimolecular hydrolysis via a tetrahedral intermediate is inhibited (because of the degree of crowding there would be

of the ester (174) in the reverse reaction (hydrolysis). Acid catalysisalso has the effect of promoting the loss of the leaving group, i.e. it iseasier to lose H 20 from (I 76)--esterification-or EtOH from (l75}hydrolysis-than it is, for example, to lose eOEt from (165) above.The formation of a tetrahedral intermediate (an associative process)in the rate-limiting step (174~ 176, for hydrolysis) is borne out bythe activation parameters observed for acid-catalysed hydrolysis of asimple ethanoate ester: Ii.H+ = 75 kJ mol-I; Ii.S+ = -105 J K-1mol-1

(cf. p. 239). The equilibrium is normally displaced in the desireddirection by using an excess of ROH (or of H20 for hydrolysis).This mechanism is generally referred to as AAc2 (~cid-catalysed,

acyl-oxygen cleavage, bimolecular). Reaction of R'OH withRC02R", under these conditions, results in transesterification, theposition of equilibrium being determined by the relative proportionsof R'OH and R"OH. Acid anhydrides and amides undergo acidcatalysed hydrolysis in very much the same way as esters.

Esters, RC02R', where the alkyl group R' can form a relativelystable carbocation, e.g. (177) from (178), have been shown-by 180labelling experiments-to undergo alkyl-oxygen cleavage:

o HO HOII He I slow I

RC-180CMe3 +=! RC-~-CMe3 +=! RC= 180 + EIlCMe3Ell

0 HO HOII He I EtOH I

R-C-OH , .R-CEil-OH , . R-C-OHr:.. slow IHOEt HOEt

Ell(171) (173) (175)

iJ.


243

Ph

1,3-diphenylfluorenone

Ph

8. 6.3 Acid-eataLysed reactions

~g ~

Ph I~o

Ph

(188a)

cone.-H 2 S04

Ph~~

Ph

(187a)

If the trisubstituted acid (186) were protonated in the normalposition (on the carbonyl oxygen atom, cf. 190), the two bulky o-Megroups would force the two adjacent OH groups into a planevirtually at right angles to the plane of the ring, i.e. (190a):

Me

(190a)

dissolution of the hindered (186) results in four-fold depression:

EllArC0 2H + 2H 2S04 P ArC=O + H30Ell + 2HS04 e

(186) (188)

Nucleophilic attack on the cationic carbon atom by, for example,MeOH is thereby prevented from taking place from all directions.By contrast, abnormal protonation (cf. 185), on the hydroxyl oxygenatom in (186), allows formation (through loss of H 20) of the planaracyl cation (188). Easy, unhindered attack on the cationic carbonatom by MeOH can now take place from either of two directions atright angles to the plane of the ring. That two different pathways,A AC2 and AAC 1, are indeed operating in acid-catalysed hydrolysisof simple esters of (a) benzoic acid and (b) 2,4,6-trisubstitutedbenzoic acids, respectively, is borne out by the relevant activation

Furthermore, if the 2,4,6-triphenyl ester (187a) is dissolved in cone.H 2S04 the brilliant colour of 1,3-diphenylfluorenone is at onceobserved--obtained via ring-closure (intramolecular Friedel-Craftsacylation) of the acyl cation (188a):f! H,O

(184)

°II EllR3CC-OH 2 + HOR'

(185)

(183)

°II -HeR3CC-OH + R'OH l •

(182)

(181)


°HO, -r0 ~EIl MeO, hOC c~

M'l$(' M"r9rM' MeOH M'L$rM'+===tl •H,O

Me Me Me

(186) (188) (187)

242

~ ?HPhC-OH + H 2 S04 P Ph~-OH + HS04 e

(189) (190)

in the T.S.), a further, relatively rare, acid-catalysed mechanism isfound to operate-AAC 1 (~cid-catalysed, acyl-oxygen cleavage, unimolecular) ; it occurs only in powerful ionising solvents:

Exactly the same considerations apply to the esterification ofhindered acids (182) in the reverse direction. It will be noticed thatthis mechanism requires protonation on the less favoured (cf. p. 240)hydroxyl oxygen atom (185) to allow the formation of the acylcarbocationic intermediate (184). Apart from a number of R3Ctypes, a very well known example is 2,4,6-trimethylbenzoic(mesitoic) acid (186), which will not esterify under ordinary acidcatalysis conditions-and nor will its esters (187) hydrolyse. Dissolving acid or ester in cone. H 2S04 and pouring this solution into toldalcohol or water, respectively, is. found to effect essentially quantitative esterification or hydrolysis as required; the reaction proceedsvia the acyl cation (188):

Evidence for the formation of (188) is provided by the observation that while dissolution of unhindered benzoic acid itself (189) incone. H2S04 results in the expected two-fold freezing point depression;

8.7 ADDITION TO C=N

2458. 7 Addition to C==N

RC=NH~RC-NH2I II

00. OH 0H,O

(197)

6lRC=NH

H~ ~H,O

(196)

RC=N

(193) ~eOH

RC=NeIOH

He 6l EIOH 6lRC=N P ~C=NHI t RC=NH P RC=NH2

(193) HOEt HJEt JEt6l

The initial product is an amide (197), but this also undergoes readyacid- or base-catalysed hydrolysis (see above), and the actual reactionproduct is often the carboxylic acid, RCOzH, or its anion.

The addition ofHzO (hydrolysis) may be both acid- and base-catalysed:

e.g. EtOH, yields salts of iminoethers (196, cf. hemiacetals, p. 209):

liS

+S7(AAc l)

6+ 6- ~i.e. RC=N = RC±N

(193ab) ~

6l eRC=N +-+ RC=N

(I93a) (193b)

aH+ (kJ mol-'): 79

as+ (JK-' mol-'): -11O(AAc 2)


parameters :

244

RCfN -----+ RC=Ne~gX He/H,O. RC=O

R:'lMgX~' ~,(193) (194) (195)

HO, -::;.0C

HO~O M'~M'Me~Me lYJ

Me Me(191) (192)

That the major factor responsible for this shift in reaction pathway is indeed a steric one is demonstrated by the observation thatthe acids (191) and (192), and their simple esters, undergo readyesterification/hydrolysis by the normal AAC2 mode:

The C N linkage bears an obvious formal resemblance to C=O,

and might be expected to undergo a number of analogous nucleophilicaddition reactions. Thus they add Grignard reagents to yield salts ofketimines (I94), which may be hydrolysed to ketones (I 95):

With RCHzCN, however, there is a tendency for Grignard reagentsto remove a proton from the CHz group, leading to more complexreactions. Reduction with LieAlH4e (cf. p. 214) yields RCHzNHz,NH3 adds to (193), in the presence of NH4 eCle to yield salts ofamidines, RC(NHz)=NHzeCle . Acid-catalysed addition of alcohols,

I I -HY "- /H-C'-C"-Y --. c=c

I I / "-

(5)

B:~HI~ Ell k,

H 2C-CR2 -+

ya

HI k,

H C-CR +=!2 12k(y .,

eOHRCH 2CH 2Br--'RCH=CH2 + H 20 + Bra

(I)

acid-catalysed dehydration of alcohols (2);

~ [6. ]*B. H B···H BHEilI :

Ri\:-rH2 -+ R2t:":':~H2 -+ R2C=CH 2(Y y6- ya

(4)

i.e. a one-step process, passing through a single T.S. (4); this is referredto as the E2 mechanism (,£limination, ~molecular) and is somewhatreminiscent of SN2 (cf p. 78). Alternatively, the H-C and C-Ybonds can be broken separately in two-step processes. If the C- Ybond is broken first, (b), a carbocationic intermediate (5) is involved;

RCH 2CR 20H ~ RCH=CR 2 + HJOEil

(2)

and Hofmann degradation of quaternary alkylammonium hydroxides(3):

9.1 1,2-«(3-)Elimination 247

Many other leaving groups are known, however, e.g. SR2, S02R,OS02Ar, etc. 1,2-Eliminations are, of course, the major route toalkenes.

Three different, simple mechanisms can be envisaged for 1,2eliminations, differing from each other in the timing ofH-C and C-Ybond-breaking. This could (a) be concerted,

9.1 1,2-(IHELIMINATION

In 1,2-eliminations involving carbon atoms (i.e. most), the atom fromwhich Y is lost is usually designated as the I-(a-) carbon and thatlosing (usually) H as the 2-(fJ-) carbon; in the older afJ-terminology,the a- is commonly omitted, and the reactions are referred to as fJeliminations. Among the most familiar examples are base-inducedelimination of hydrogen halide from alkyl halides-this almostcertainly the most common elimination of all-particularly frombromides (I) ;

-M.CO,H, ArC=N

Ar OCOMe"- /

C=N/ .'H

9.1 1,2-(t3-)FLIMINATION, p. 247.9.2 El MECHANISM, p. 248.9.3 ElcB MECHANISM, p. 249.9.4 E2 MECHANISM, p. 251:

9.4.1 Stereoselectivity in E2, p. 253; 9.4.2 Orientation in E2:Saytzev v. Hoffmann, p. 256.

9.5 FLIMINATION V. SUBSTITUTION, p. 260.9.6 EFFECT OF ACTIVATING GROUPS, p. 262.9.7 OTHER 1,2-FLIMINATIONS, p. 263.9.8 1,1-(a-)FLIMINATIoN, p. 266.9.9 PYROLYTIC SYN ELIMINATION, p. 267.

H"- / -HYc=c --. -c=c-/ "-y

Eliminations from atoms other than carbon are also known;

9Elimination reactions

Elimination reactions involve the removal from a molecule of twoatoms or groups, without their being replaced by other atoms orgroups. In the g~eat majority of such reactions the atoms or groupsare lost from adjacent carbon atoms, one of them very often beinga proton and the other a nucleophile, y: or ye, resulting in theformation of a multiple bond, a l,2-(or afJ-) elimination:

H H HI I -HCN I

R-C-O • R-C=OI

CN

as are eliminations both from the same atom, I, I-(a-) eliminations(cf. p. 266), and from atoms further apart than 1,2-, i.e. reversal of1,4-~d~ition (ct. p..195), also 1,5- and 1,6-eliminations leading tocychsatlOn. 1,2-Ehmmauons are by far the most common and important, however, and most of our discussion will be concerned with them.

248 Elimination reactions 9.3 BleB mechanism 249

this is referred to as the El mechanism (Elimination, unimoleeular). It isreminiscent of SN 1 (ef. p. 79), and the carbocationic intermediatesfor SN 1 and El are, of course, identical. FinaIly, the H-C bondcould be broken first, (e), involving a carbanion intermediate (6);

B0 H BHel BHelI ., ~ .,

XzC-yHz e XzC-rlCH z -+ XzC=CH zY ~y ya

(6)

this is referred to as the ElcB mechanism [Elimination, from conjugate!!ase, i.e. (6)]. Examples of reactions proceedmg by all three mechanismsare known: ElcB is the least, and E2 probably the most, common. Thethree mechanisms will now be considered in turn, but it should berealised that they are only limiting cases (ef SN1/SN2), and that in facta continuous mechanistic spectrum, in the relative time of breaking ofthe two bonds, is available and is indeed observed in practice.

9.2 Et MECHANISM

If, as is normally the case, carbocation, e.g. (5), formation is slowand rate-limiting (i.e. k 2 > L 1), then the rate law observed with, forexample, the bromide MeCH2CMe2Br is;

Rate = k[MeCHzCMezBr]

the overaIl elimination is then completed (7) by rapid, non rate-limitingremoval of a proton from (8) usuaIly by a solvent molecule, in thiscase EtOH:

Me Me MeI EIOH (EtOH /

MeCHzC-OEt -5 MeCHzC ~ MeCH=CI .1 "- EI "-

Me Me Me

(p. 79), and the latter reaction to yield the substitution product (9)is commonly a competitor with E I elimination. Some evidence thatthe two processes do have a common intermediate is provided bythe fact that the EI/SN I ratio is reasonably constant for a given alkylgroup irrespective of the leaving group, yS. The two processes do,however, proceed from (8) to products-(7) and (9), respectively-viadifferent T.S.s, and the factors that influence elimination v. substitutionare discussed subsequently (p. 260).

The factors that promote unimolecular, as opposed to bimolecular(E2), elimination are very much the same as those that promote SN Iwith respect to SN2, namely: (a) an alkyl group in the substrate thatcan give rise to a relatively stable carbocation, and (b) a goodionising, ion-solvating medium. Thus (a) is reflected in the fact thatwith halides, increasing El elimination occurs along the series,

primary < secondary < tertiary

reflecting the relative stability of the resultant carbocations; primaryhalides hardly ever undergo El elimination. Branching at the ~

carbon atom also favours El elimination; thus MeCH2CMe2CI isfound to yield only 34% of alkene, while Me2CHCMe2CI yields62%. This is probably related to the fact that Me2CHCMe2CI canlead to a more heavily substituted, and hence thermodynamicaIlymore stable (ef. p. 26), alkene than the first. This is, with Elreactions, also the major controlling factor (Saytzeff elimination, p.256) in orientation of elimination, where more than one alkene canbe derived by loss of different ~-protons from a carbocationicintermediate (8):

HQ)I

uCHz (DH CH z Me~ Q) I / <D /

MeCHz-C - MeCH-C ftl-'-+ MeCH=C

"- -He "- -He "-Me Me Me

(9) (8) (7) (10) (8) (7)

It could be claimed that such an EI solvolytic elimination would beindistinguishable kinetically from a bimolecular (E2) elimination, inwhich EtOH was acting as base, because the [EtOH] term in the E2rate law,

Rate = k[MeCHzCMezBr][EtOH]

would remain constant. The two can often be distinguished, however,by adding a little of the conjugate base of the solvent, i.e. SOEt inthis case. If no significant change in rate is observed, an E2mechanism cannot be operating, for if SOEt is not participating as abase the much weaker EtOH certainly cannot be.

The carbocation (8) is identical with that from SN 1 solvolysis

Thus in the above case the elimination product is found to contain82 /~ of (7). Unexpected alkenes may arise, however, from rearrangement of the initial carbocationic intermediate before loss of proton.El elimination reactions have been shown as involving a dissociatedcarbocation; they may in fact often involve ion pairs, of varyingdegrees of intimacy depending on the nature of the solvent (ef. SN 1,p.90).

9.3 EtcH MECHANISM

If as might be expected for this pathway, formation of the carbanioni~termediate (6) is fast and reversible, while subsequent loss of the

. I

251

(19)

9.4 E2 mechanism

(18)

This has all the right attributes in the substrate: (a) electronegativehalogen atoms on the p-carbon to make the P-H more acid, (b) stabilisation of the carbanion (16) through electron-withdrawal by the halogenatoms on the carbanion carbon atom, and (c) a poor leaving groupin F. An attempt has been made to correlate the relative leavinggroup ability of a series of different Y groups in the EIcB reaction:

B:·HI eO-CR2 p e"QLCR 2 -+ O=CR 2 + CN

I (ICN CN

(20)

Y: PhSe> PhO> PhS = PhS02> PhSO> MeO» CN

The observed order of ability did not, however, correlate with thepKa of YH, with the strength of the C-Y bond, or with the polareffect of Y! Clearly, leaving group ability even in this simplereaction is a highly complex attribute.

Other examples of the E1cB pathway are benzyne formation fromC6H sF (cf. p. 174), reversal of simple nucleophilic addition to C=O,e.g. base-induced elimination of HCN from cyanohydrins (20; cf. p.212),

By far the commonest elimination mechanism is the one-step concerted(E2) pathway exhibiting, e.g. for the base-induced elimination ofHBr from the halide RCH 2CH 2 Br (21), the rate law:

9.4 E2 MECHANISM

and base-induced dehydration of aldols to a~-unsaturated carbonylcompounds (cf. p. 225).

(13)(12)

Elimination reactions250

leavi~g gr,?up, ye, is slow and rate-limiting, i.e. L I > k2

, then thisreactIOn will follow the rate law,

Rale = k[RY][B]

and will be kinetically indistinguishable from the concerted (E2)pathway. It should be possible to distinguish between them howeverby observing exchange of isotopic label, between as yet ~nchangedsubstr~te and solv~nt, arising during fast, reversible carbanion (6)formatlOn-somethmg that clearly could not happen in the one-step,concerted (E2) pathway. A good example to test would be PhCH

2

CH 2Br (II), as the Ph group on the p-carbon would be expected topromote acidity in the P-H atoms, and also to stabilise the resultantcarbanion (12) by delocalisation:

EtOe"") HI

PhCH-CH2I

Br(11) ~OEI

EtOH~

Ph~iP-CH2 -+ PhCH=CH2qr slow

EtOOP

? 00EIPhCH-CH2

IBr

(14)

The reaction was carried out with eOEt in EtOD, and (11) re-isolatedafter ~ half-conversion to (13): it was found to contain no deuteriumi.e. no (14); nor did the alkene (13) contain any deuterium, as migh{have been expected by elimination from any (14) formed. Thispotentially favour~ble case thus does not proceed by an ElcB pathwayof the form descnbed above; though we have not ruled out the casewhere k 2 » k_ l , i.e. essentially irreversible carbanion formation.. In fact reac~io~s proceeding by this carbanion pathway are exceedmgly rare; this IS not altogether surprising as calculations suggestthat the energy of activation for E2 is generally more favourable thanthat for EIcB, in most cases by ~ 30-60 kJ (7-14 kcal) mol- I (therevers~ of step 2 would requ~re addition of ye to C=C, which certainly~oesn t happen at all eastly). One example that almost certainlymvolves the latter pathway, however, is X2CHCF3 (15, X = Hal):

B:"") H,X2C'-C"F2 P X2~CF2 --+ X2C=CF2

~ rast '~SloW

(15) (16) (17)As B is often a nucleophile as well as a base, elimination is frequentlyaccompanied by one-step, concerted (SN2) nucleophilic substitution

(cf p. 78):

253

PhCH 2CH 21

2·7 x 104

PhCH 2CH 2 Br

4·2 x 10)

9.4.1 Stereoselectivity in E2

ReI. rate:

its oxy-anion as a leaving group. Thus p-MeC6ILS03e ('tosylate',Ts) is a very much better leaving group than eOH, reflectingp-MeC6~S03H being a very much stronger acid (lower pKa value)than H 20. Where the atom through which Y is bonded to C inR-Y does not stay the same, however, this inverse correlation withpK;. often breaks down. Thus the importance of the strength of theC-Y bond (rather than the pKa of H-Y) is borne out by therelative rate sequence observed for PhCH2CH2Hal witheOEt/EtOH:

Incipient solvation of the developing ye in the transition state (e.g.22), through hydrogen-bonding or other means, can also play itspart in determining relative leaving group ability, and this mayormay not follow the same general sequence as acid strength of HYand/or C-Y bond strength. Change of solvent thus can, and does,change the sequence of relative leaving group ability in a series ofdifferent yes.

Finally, the major structural features in the substrate promotingE2 elimination are those that serve to stabilise the resultant alkeneor, more particularly, the T.S. that precedes it. Such features includeincreasing alkyl substitution at both IX- and fJ-carbon atoms (leadingto alkenes of increasing thermodynamic stability), or introduction ofa phenyl group that can become conjugated with the developingdouble bond.

There is an obvious advantage in elimination taking place from aconformation in which H, CfJ, C~ and Yare in the same plane as thep orbitals that are developing on CfJ and C~, as HEll and ye are departing,will then be parallel to each other, and thus capable of maximumoverlap in the forming 7t bond. It will be energetically advantageousfor the attacking atom of the base B to lie in this common plane also.


With acylic molecules elimination could be envisaged as taking placefrom one or other of two limiting conformations-the anti-periplanar(24a) or the syn-periplanar (24b):

R' R R' B¥ R

B~ -HY 0· -HY .~.-+-. R R' -+

R R' R R'R' R'

(24a) (25) (24b) (26)


The factors that influence elimination v. substitution are discussedsub~eq~ently (p. 260). Evidence for the involvement of C-H bond~sslOn ~n the rate-limiting st~p-as a concerted pathway requiresIS provided by the observation of a primary kinetic isotope effect(cf p. 46) when H is replaced by D on the fJ-carbon.

O~~ of the factors that affects the rate of E2 reactions is, hardlysurpnsIngly, the strength of the base employed; thus we find:

E2:

eNH 2 > eOR > eOH

Some studies have been made with bases of the type ArOe , as thisallows stud.y ofthe ~ffectsof v~iati?n in basic strength (by introductionof ~-subst~tuents In C6HsO ) without concomitant change in thestenc reqUirements of the base. With a given base, transfer from ahydroxyhc solvent, e.g. H2 0 or EtOH, to a bipolar aprotic one, e.g.HCONMe 2 (DMF) or Me2SEIl -Oe (DMSO), can have a very pro?ounced effect as the s.tren.gth ofthe base, e.g. eOH, eOR, is enormouslyIncreased thereby. This anses because the base has in the latter solventsno. envelope of hydrogen-bonded solvent molec'ules that have to b~~tnpped away before it can act as a base (cf effect on nucleophilicityIn SN2, p. 81). Such change of solvent may result in a shift of mechanisticpathway fr0!D E I to E2 for some substrate/base pairs.

To. explain the effect change of Y may have on the rate ofreac~lon of R-Y (in which R remains the same) we need toconsider: (a) any effect Y may have on C-H bond-breaking (E2 isa co~~erted r~action), (b) the strength of the C-Y bond, and (c) thestabil~ty of Y , as reflected in the pKa of H-Y. It thus comes as nosurpns~ to find that forecasting the relative ability of Y as a leaving~oup IS far from easy! If the atom in Y that is directly bonded to CIn R-Y (and to. H In H-Y) remains the same, e.g. oxygen, then~he rate of reactIon of R-Y may correlate not too badly with theInverse of the pKa of H-Y: the stronger the oxy-acid, the better is

nm

E9For Y = Br, Ts or NMe3' elimination was essentially 100% stereo-selectively ANTI-only (28) was obtained from (26), and only (29) wasobtained from (27). There are, however, numerous exceptions with

E9longer chain NR3 compounds, perhaps because of some SYN elim-ination of the Ii-H via a cyclic transition state involving the quaternary ammonium hydroxide ion pair (30);

R~,RR'

7. ....~MeJceO'

H IH

255

Q~OTS

(32)

%SYN elimination

90464

37

UIH H

CICI CI

CI CIH H

(35)


Q Ph ANTI~elimil\8.tion h e~n

H(33)

(34)

;;:::JY

Ring size

CyclobutylCyclopentylCyclohexylCycloheptyl

Q:iOTS

(31)

Thus of the geometrical isomers of hexachlorocyclohexane, C6 H6 Cl6 ,

one is found to undergo elimination of HCl at a rate slower, by afactor of 7-24 x 103, than any of the others; it is found to be the one(35) that cannot assume the above trans-diaxial conformation.

The relative lack of stereoselectivity with cyclopentyl compounds isreflected in the behaviour of the trans- and cis-isomerides, (31) and(32). Each, if it eliminates by E2 at all, will be converted into the samealkene (33H31) via SYN elimination, and (32) via ANTI elimination:

The degree of stereoselectivity may be influenced to some extent bythe polarity and ion-solvating ability of the solvent.

In cyclic compounds the conformation from which elimination cantake place may to a considerable extent be enforced by the relativerigidity of the ring structure. Thus for a series of eliminations fromdifferent sized rings, the following degrees of stereoselectivity wereobserved for HY elimination from the cyclic compounds(CHz)nCHY:

ANTI elimination [(32) -+ (33)] was found to proceed only 14 timesfaster than SYN elimination [(31) -+ (33)] reflecting the fact that theenergy needed to distort the ring, so that (32) can assume an approximately anti-periplanar conformation, almost outweighs the normalenergetic advantage of the staggered conformation over the, synperiplanar, eclipsed one, i.e. (31).

The marked ANTI stereoselectivity observed with cyclohexylsystems (see above) reflects the ability to achieve, and the very markedpreference to eliminate from, the so-called trans-diaxial conformation(34):

(29)

60EIEIOH

(27)

0"o Me

(28)

Elimination reactions

(26)

254

Having established the desirability of elimination taking place froma planar conformation, there remains the question of whether either(24a) or (24b) is preferred over the other.

Three possible grounds can be stated for favouring eliminationfrom the anti-periplanar conformation (24a): (a) elimination wouldthen be taking place from the lower energy 'staggered' conformation(24a), rather than from the higher energy 'eclipsed' conformation(24b; cf. p. 7), and this energy differential is likely to be reflected inthe corresponding transition states; (b) the attacking base, B:, andthe departing leaving group, ye, would be as far apart from eachother as possible in the T.S.; and (c) the electron pair developingfrom the initial C-H bond would be attacking the a-carbon atomfrom the side opposite to that from which the electron pair of theinitial C-Y bond will be departing (ct. the favoured 'backside'attack in the SN2 pathway, p. 78). It seems likely that (a) will be themost significant of these features, however. We would thus forecasta preference for ANTI ('opposite side') elimination of Hand Y(from 24a), rather than SYN ('same side') elimination (from 24b).

Where, as with (24) above, both cP and C" are chiral, eliminationfrom the two conformations will lead to different products-the transalkene (25) from (24a) and the cis-alkene (26) from (24b). Thus knowingthe configuration ofthe original diastereoisomer (e.g. 24), and establishing the configuration of the geometrical isomeride(s) that is formed,enables us to establish the degree of stereoselectivity of the eliminationprocess. In most simple acyclic cases, ANTI elimination is found tobe very much preferred, e.g. in about the simplest system, (26) and (27),that permits of stereochemical distinction:

~ Xy

Me 60E JK y EtO~' DyMe

o Me Me

9.4.2 Orientation in E2: Saytzev v. Hofmann

In substrates which have alternative P-hydrogen atoms available itis possible to obtain more than one alkene on elimination, e.g. (36)where there are two possibilities:

III IIIy= Dr SMe2 NMe)MeCH 2CH=CH 2 19% 74% 95%

Q)H HCD 0/ (37)

I IMeCH-CH-CH2 60E,

Iy ~(36)

MeCH=CHCH) 81% 26% 5%(38)


expected from an E2 pathway-so the real question is why a +velycharged Y should prompt a divergence from this apparent norm? AY group such as lDNMe3 will exert a powerful, electron-withdrawing,inductive/field effect on both the ~-carbon atoms, and thus on the Hatoms attached to them,

<%>H H<Dt t

Me-+-CH-+CH+CH2•(BNMe3

(36al

thus making these hydrogens markedly more acidic. They will thusbe much more readily removed by base than when Y was Br, andthe powerful electron-withdrawal by lDNMe3 will also stabilise theincipient carbanion forming as either H is being removed. Thiseffect will, in the case of <2IH, be reduced to some extent throughelectron-donation by the Me substituent on this ~-carbon; such anacid-weakening effect does not occur with <DfI, which is thus moreacidic than ~, and hence the proton that is more easily removedby base. This effect of lDNMe3 is apparently sufficient to makerelative proton acidity, rather than potential alkene stabilisation, thecontrolling factor. The reaction now proceeds through a T.S. (37a)

l 8-1+H- .. ·OEt

MecH2TH-1~EllNMe3

(37a)

possessing some degree of 'carbanion character', but in which littleor no 'alkene character' has yet developed. E2 eliminations can thusinvolve transition states along a whole spectrum of 'character',whose nature is determined in considerable part by Y.

It is interesting in this respect that when Y is F, despite this notbeing +vely charged, there is a marked tendency towards theHofmann product: thus EtCH2CH(F)CH3Ieads to no less than 85%of EtCH2 CH=CH2 • This 'unexpected' result stems from the extremely powerful electron-withdrawing effect of F (cf. lDNMe3); andalso that Fe is an extremely poor leaving group, thereby delayingC-F bond-breaking in the T.S. 'spectrum'. Support for the importance of proton acidity, and the development of 'carbanion character' in the T.S., for Hofmann elimination is provided by the observation that increase in the strength of the base attacking RY (whetherY is +vely charged or not) also leads to increasing formation of theHofmann product. ~-Substituents that would help stabilise a developing -ve charge promote formation of the Hofmann product,


To help in forecasting which alkene is the more likely to be producedthere have long been two empirical rules that can be summarised as

lDfollo~s: (a) Hofmann (1851; working on RNMeJ compounds, i.e.

Y = NMeJ ) stated 'that alkene will predominate which has leastalkyl substituents on the double bond carbons', i.e. (37) above;(1ijSaytzev (1875; working on RBr compounds, i.e. Y = Br) stated 'thatalkene will predominate which has most alkyl substituents on thedouble bond carbons', i.e. (38) above. Both generalisations are validas the figures quoted above indicate. It is thus clear that the compositionof the alkene mixt~re obtained on elimination is influenced by Y, thenature of the leavlOg group, and an explanation is required abouthow this influence may be exerted.

Saytzev elimination, which appears to occur when Y is neutral(e.g. with Y = OTs, etc., as well as Br), leads to the more stable (i.e.more heavily substituted, cf. p. 26) alkene. It seems reasonable tosuppose, therefore, that reaction here proceeds via a T.S. in which anot inconsiderable degree of 'alkene character' has already beendeveloped; the alkyl substituents thereby being able to begin exerting their stabilising (energy lowering) effect quite early in the singlestep of the E2 pathway, e.g. (38a):

lEtb~"'H 1+

M,tH=r.CH,(38a)

The preference for Saytzev elimination in the E1 pathway hasalready been referred to (p. 249).

This appears to be wholly logical-just what we would have

259

Me

H~H"H

(45)

[B+'0CH3

],~ '~CH'R R R

(44)

--


<D Me CHR2

B~}1

H H(46b)

@H H<DI I

R C-CMe-CH2 I 2

Y

(46)

In many cases it is all but impossible to distinguish, separately, theoperation of electronic and steric effects, as they often both operatetowards the same end result. Except where crowding becomes extreme,however, it seems likely that the electronic effects are commonly incontrol.

In cyclic systems, the usual simple requirements of Saytzev orHofmann rules may be overridden by other special requirements of thesystem, e.g. the preference for elimination from the trans-diaxialconformation in cyclohexane derivatives (cf p. 255). Another suchlimitation is that it is not normally possible to effect an eliminationso as to introduce a double bond on a bridgehead carbon atom in afused ring system (Bredt's rule), e.g. (47)~ (48):

'these several steric effects are explainable on the basis that ~crowding, irrespective of its origin, will make the T.S. (44) that involvesthe removal of proton (D from (46a)-Saytzev elimination-relativelymore crowded than the T.S. (45) that involves removal of proton CDfrom (46b)-Hofmann elimination. The differential wi.1I .incr~ase ~s

the crowding increases (in R, Y or B), and Hofmann ehmlOatlOn wIllthus be progressively favoured over Saytzev:

@ Me CH 3

B=-~ X __Saytzev K 'y

R R (460)

III

Hormann

%Hormann(i.e. CD)

Hofmann

%Hormann (i.e. CD)

Me$/N-Me

'"Me

98

Me$/

Y = Dr S",

Me

31 87


but substituents such as Ph, C=C, etc., promote formation ofwhichever alkene has its double bond conjugated with them.Another manifestation of Hofmann elimination is that where, as in(39), there are alternative potential RNMez leaving groups, the leastsubstituted alkene is always formed preferentially, i.e. (40) ratherthan (41):

<DHICH2-C~2$ r CH 2=CH 2 + MeCH 2CH 2NMe2

NMe2 90E, (40)

Me+-yH-C'2 ~ MeCH=CH 2 + CH l CH 2NMe2@H (41)

(39)

The effect of Y on the mode of elimination may also involve asteric element. Thus it is found that increase in the size of Y and, moreparticularly branching in it, leads to an increasing proportion ofHofmann elimination with the same alkyl group, e.g. with (42):

The proportion of Hofmann elimination is also found to increasewith increasing branching in the alkyl group of the substrate (constantY and base), and with increasing branching in the base, e.g. with (43),a bromide where preferential Saytzev elimination would normally beexpected:

(43)

It may be mentioned in passing that the volume, and quantitativeprecision, of data available in this field owes much to the use of gas/liquid chromatography for the rapid, and accurate, quantitativeanalysis of alkene mixtures.

chHBase ttz ~H~

Br H H

(47) (48) (49)

(50) (51)

This is presumably the case because the developing p orbitals in an E2reaction, far from being coplanar (cf p. 253), would be virtually atright angles to each other (49), and so could not overlap significantlyto allow development of a double bond. The relatively small ringsystem is rigid enough to make the distortion required for effectivep orbital overlap energetically unattainable; there seems no reasonwhy an E I or E IcB pathway would be any more successful: the bicycloheptene (48) has, indeed, never been prepared. With bigger rings, e.g.the bicyclononene (50), or a more flexible system (51), sufficientdistortion is now possible to allow the introduction of a double bondby an elimination reaction:

9.S ELIMINATION v. SUBSTITUTION

E I elimination reactions are normally accompanied by SN I substitution,~s both have ~ common-earbocationic-intermediate; though thisIS converted mto either elimination or substitution products viadifferent T.S.s in a fast, non rate-limiting step. Similarly, E2 eliminationis often accompanied by SN2 substitution, though in this case theparallel, concerted processes involve entirely separate pathwaysthroughout. Thus considering elimination v. substitution there arereally three main issues: (a) factors influencing EI/SNI product ratios,(b) factors influencing E2/SN2 product ratios, and (c) factors influencingch~nge of pathway, i.e. El/SNI ~ E2/SN2 (or vice versa), as such ashift often changes the proportion of elimination to substitution.

The last of these, (c), may weB be the most potent. Thus EI/SNIsolvolysis of Me 3CBr, and of EtMe2CBr, in EtOH (at 25°) was foundto yield 19 %and 36 %, respectively, of alkene; while introduction of2M EtOe-which shifts the mechanism in part at least to E2/SN2resulted in the alkene yields rising to 93 % and 99 %, respectively.It is indeed found generaBy, for a given substrate, that the E2/SN2ratio is substantially higher than the E I/SNI ratio. A point that isworth bearing in mind when contemplating preparative, syntheticoperations is the use of a less polar solvent (the El/SN1 process isfavoured by polar, ion-solvating medial-the classical alcoholic,rather than aqueous, potash for elimination of HBr from alkylbromides. A shift in mechanism may also be induced by increasingthe concentration of the base employed, e.g. eOH; hence theclassical use of concentrated, rather than dilute, potash for elimination.

2619.5 Elimination v. Substitution

RCH RCH1 RCH 2II Be I ...!:. 'C-BC - CEIl

.:' \EI .:' \ SNI R")

R R R R R

(53) (52) (54)

Crowding strain is thus re-introduced in the T.S. for substitution, butmuch less so if at all, in the T.S. for elimination, and the differentialbetween then't will become greater-increasingly favouring elimination-as the size and degree of branching in the R groups increases; butonly becoming significant when larger/more branched than Me3C-Y.A related, but slightly different, point is that the peripheral Hwill be much more accessible than the relatively hindered carbocationic carbon; we should thus expect the proportion of elimination to rise as the size of the attacking base/nucleophile increases: asis indeed observed, Le. Me3COe is usually better than EtO

efor

carrying out elimination reactions with. This discussion has tendedto centre on the El/SNI case, but essentially analogous steric effectsare involved in the differential stabilisation of the T.S. for E2 withrespect to the T.S. for SN2.

The EI/SNI ratio is, of course, substantiaBy independent of theleaving group Y, but this is not the case with E.2/SN2, where break~ngof the C-Y bond is involved in each alternatIve T.S. The followmgrough sequence, in order of increasing promotion of elimination, is

In either (a) or (b), the carbon structure of the substrate is ofconsiderable importance, the proportion of elimination rising ongoing: primary < secondary < tertiary. In electronic terms this stemsfrom increasing relative stabilisation of the T.S. for elimination as thenumber of alkyl groups on the carbon atoms of the developing doublebond increases (cf. p. 256). Thus with EtOe in EtOHon alkyl bromides,we find: primary ---+ ca. 10 % alkene, secondary ---+ ca. 60 %, andtertiary ---+ > 90 %. This stems not only from an increasing rate ofelimination but also from a decreasing rate of substitution. Similarly,sUbstituent~ such as C=C and Ar that can stabilise the developingdouble bond through conjugation (cf. p. 253) also strongly favourelimination: under comparable conditions, CH 3CH2Br yielded::::: I %alkene, while PhCH2CH2Br yielded::::: 99 %.

In El/SNI increasing branching in R-Y leads to an increase inthe proportion of elimination. This arises from increasing stability ofthe progressively more highly substituted alkene product ~nd, .moreimportantly, of the T.S. leading to it from the carbo~at~on .mte~mediate. A steric factor may also operate to favour elImmatIon mthat the Sp2 hybridised carbon atom in the carbocation (52) remainsSp2 hybridised (= 120° bond angles) on eliminatio~ (~3), but becomes Sp3 hybridised (=109° bond angles) on substItutIOn (54):


262 Elimination reactions 9.7 Other 1,2-eliminations 263

observed: iXfJ-unsaturated aldehyde (59, cf p. 225):

..Dehydrations are normally acid-catalysed (protonation of OH turningit into EllOH 2 , H20 being a better leaving group than eOH), and abase-catalysed elimination is here made possible by the CHO groupmaking the fJ-H atoms more acidic, and stabilising the resultantcarbanion, i.e. (a )/(b) on p. 262. Stabilisation, by conjugation, of thedeveloping double bond [(c) above] has been included in the T.S.(60) above, but how large a part this plays is not wholly clear. It is,however, significant that electron-withdrawing substituents are usually very much more effective in promoting elimination when theyare on the (3-, rather than the a-, carbon atom: they could conjugate with a developing double bond equally well from either position, but can only increase acidity of (3-H, and stabilise a carbanionfrom the (3-position. This is clearly seen in base-induced eliminationof HBr from 1- and 2-bromoketones, (61) and (62), respectively,

Ell EllTosylate < Br < SMe2 < NMe3

The attacking base/nucleophile is obviously of importance also; werequire, ideally, a species that is a strong base but a poor nucleophile. Preparatively, tertiary amines, e.g. Et3N, pyridine, areoften used to promote elimination. Though these are not particularly strong bases, they are poor nucleophiles because of stericeffects, e.g. branching in Et3N, impeding nucleophilic attack oncarbon, but not basic attack on a peripheral hydrogen. The use of abase of relatively high b.p. is also advantageous (see below).

Finally, elimination-whether E I or E2~is favoured with respectto substitution by rise in temperature. This is probably due to elimination leading to an increase in the number of particles, whereas substitution does not. Elimination thus has a more +ve entropy term (cf. p.241), and because this (~S+) is multiplied by T in the relation forthe free energy of activation, ~G+ (~G+ = ~H+-T~S+, cf. p. 38), itwill incre.asingly outweigh a less favourable ~H+ term as the temperature rIses.

9.6 EFFECT OF ACTIVATING GROUPS

[

6+ ]*B:. H BOO'HI :

O=C-CH-CHMe P O:':":C:':":CH:':":CHMeI I I:H OH H 6-0H

(58) (60)

BHEil

P O=C-CH=CHMeIH eOH

(59)

We have to date considered the effect of alkyl substituents in promotingelimination reactions in suitable substrates, and also, in passing, thatof Ar and C=c. Elimination is, in general, promoted by most e1ectronwithdrawing substituents, e.g. CF3 , N02 , ArS02 , CN, c=o, C0

2Et,

etc. Their effect can be exerted: (a) through making the fJ-H atomsmore acidic (55), and hence more easily removable by a base, (b)through stabilisation of a developing carbanion by electron-withdrawal(56), or in some cases, (c) through stabilisation ofthe developing doublebond by conjugation with it (57):

(63)

0=C-CH=CH 2I

Me

Me Hr.BI I

O=C-CH -CH 2IBr

(61)

B~ ¥ [ B6+"'J:I ]*O=C~CH-CH --+ 0:':":C:':":CH:':":<;H 2I I 2 fast I :

Me Br Me Br6-

(62)

where both give the same iXfJ-unsaturated (i.e. conjugated) ketone (63),but (62) is found to eliminate HBr very much faster than (61), underanalogous conditions. Such fJ-substituents are often effective enoughto promote loss of more unusual-and poor-leaving groups such asOR, NH2 , etc. (OH above).

6+B"'I;I

O:':":C:':":CH:':":CHI : 2

OEt y6-(57)

6+B···Heo...... .

: N--'-CH-CH./Ell I 2

6-0 Y

(56)(55)

The more powerfully electron-withdrawing the substituent the greaterthe chance that the T.S. in an E2 elimination will be 'carbanion-like'(cf p. 257), or even that the reaction pathway may be shifted to theElcB mode (cf p.249), e.g. possibly with N02 or ArS0 2 , especiallyif the leaving group, Y, is a poor one.

A good example of elimination promotion is by the CHO groupin aldol (58) making possible a base-catalysed dehydration to an

9.7 OTHER 1,2-ELIMINATIONS

Attention has to-date been devoted almost entirely to eliminationsin which it has been H that has been lost, as a proton, from the fJ-carbonatom. These are certainly the most important eliminations, but examplesare known that involve the departure of an atom or group other than

264 Elimination reactions 9.7 Other 1,2-eliminations 265

(71)

~CCOO'z JHH~ Ph

Brs

Br

MeZnBr<ll JH~H

r BrsMe

~c:-?O ;~/lh

K ~rH Br

(72)

Me HZn:\~r A

r~ c.sr fl.H Me

(69)

(70)

Strict ANTI stereoselectivity is, however, departed from with longerchain 1,2-dibromides, i.e. above C4 . The reaction may also be inducedby Mg, hence the impossibility of making Grignard reagents fromsimple 1,2-dibromides. Metal cations have also been used to inducedehalogenation, the reaction then has the advantage over that withmetals of occurring homogeneously. Debromination is rarely apreparatively useful reaction as the 1,2-dibromide starting materialhas usually been prepared by adding bromine to the product alkene!Brominationjdebromination is, however, sometimes used for 'protecting' double bonds, e.g. in the oxidation of (70) ----. (71), which couldnot be carried out directly because the double bond would be attackedoxidatively at the same time.

Eliminations have also been carried out on a number of compoundsof the form HalCHzCHzY, where Y = OH, OR, OCOR, NHz, etc.;these eliminations normally require conditions more drastic than for1,2-dihalides, and metals or metal cations are found to be moreeffective than Ie. These eliminations are often found to be somewhat indiscriminate in their stereochemistry. The elimination ofCOz/Bre from the diastereoisomer (72) of 2,3-dibromo-3phenylpropanoate in MezCO is, however, found to proceed 100%ANTI, and under extremely mild conditions:

renewing the active surface by removing the metal halide that isformed there. With simple examples,like those above, e.g. (69), thereis a high degree of ANTI stereoselectivity, and the reaction pathway isprobably simple E2, though the metal surface is certainly involved.

Dr DrBr, I HNO, I Zn

RCH=CHCHzOH -----. RCH-CHCHzOH -----. RCH-CHCOzH -----. RCH=CHCOzHI I

Dr Dr

E2 H~~yD

D

H D

=InH D

(65)

Me

~D~Dr Br

s

Me

Me D

I~D Me

(64)

which would be compatible with a simple E2 pathway.This is borne out by the high degree of ANTI stereoselectivity that

is observed in acyclic examples (cf p.254), when either or both thebromine atoms are attached to secondary or tertiary carbon atoms,e.g. (64):

H from CfJ, the commonest probably being 1,2-dehalogenations and,in particular, 1,2-debromination. This can be induced by a number ofdifferent species includin% iodide ion, Ie, metals such as zinc, andsome metal ions, e.g. Fez . The reaction with Ie in acetone is foundto follow the rate law (after allowance has been made for the Iecomplexed by the 12 produced in the reaction),

Rate = k[I,2-dibromide][IS]

only the trans-alkene (65) is obtained. When either or both the bromineatoms are attached to primary carbon atoms, e.g. (66), however, theoverall reaction is found to proceed stereoselectively SYN, i.e. thecis-alkene (67) is the only product. This somewhat surprising resultis believed not to represent a stereochemical change in the eliminationitself, but to result from a composite SN2jE2 mechanism; in whichSN2 displacement of Br by Ie, with inversion of configuration (68),is followed by a stereoselective ANTI elimination on the l-iodo-2bromide (68) to yield (67}-the overall reaction being an apparentSYN elimination [(66) ----. (67)]:

H D

0r 5H2slow~

Br

D )H1

(66) (68) (68) (67)

Support for the actual elimination step, in each case, being E2 isprovided by the fact that changing the alkyl substituents on C« andCfJ results in reaction rates that, in general, increase with the relativethermodynamic stability of the product alkene.

Bre and Cle are much less effective at inducing 1,2-dehalogenationthan Ie, but metals-particularly Zn-have long been used. Reactiontakes place heterogeneously at the surface of the metal, the solvent

266 Elimination reactions9.9 Pyrolytie SYN elimination 267

9.8 l,l-{cx-)ELIMINATION

A relatively small number of examples are known of I,I-eliminationsin which both H and the leaving group, Y, are lost from the same(IX-) carbon atom, e.g. (73)~ (74). They tend to be favoured: (a) bypowerfully electron-withdrawing Y groups-these increase the acidityof the IX-H atoms, and stabilise a developing -ve charge on the IXcarbon atom, (b) by using very strong bases, B, and (e) by the absenceof fJ-H atoms-though this is not a requirement (ef 73):

In some, though not necessarily all, cases loss of HGl and CI 8 is thoughtto be concerted, leading directly to the earbene (ef p. 50) intermediate(75); fo~ma.tion of the pro.duct alkene from (75) then requires migrationof H, wIth Its electron paIr, from the fJ-carbon atom. A I,I-elimination(EIX) .will be indistinguishable kinetically from 1,2-(E2), and evidencefo~ Its occurrence rests on isotopic labelling, and on inferentialeVIdence for the formation of carbenes, e.g. (75).

Thu.s introduction of 20 atoms at the IX-position in (73) is found toresu.lt In .one of th~m .being lo~t in going to (74}-both would beretaIne~ In E2; ~hIle l.ntroductlOn of 20 at the fJ-position in (73)resul~s In both beIng still present in (74), though one is now on thetermmal (IX- in 73) carbon atom-one would have been lost in E2.From such isotopic labelling data it is possible to determine howmuch of a given elimination proceeds by the 1,1-, and how much bythe I,2-path~ay ..Use of C6 H 5 8NaGl-an enormously strong base-indecane solutIon IS found to result in 94 ~/ I I-elimination from (73)

h'l Gl 8 0 , ,~ leNa NJ:l2 caused much less, and Na GlOMe8 hardly any at all,I.e. the operatIOn. offactor. (b). above. It was also found that, for a givenbase, alkyl bromIdes and IodIdes underwent much less I,I-eliminationthan th~ cor~esponding chlorides, i.e. operation of factor (a), above.~nferen~lal eVIdence .for t~e formation of the carbene intermediate (75)IS provIded by the IsolatIOn from the reaction mixture of the cyclopropane (76),

(76)

of(73), but no less than 32 /~ of(76) was isolated from the I,I-eliminationof the isomeric chloride, MeCH(CI)CH 2CH 3 .

The most familiar, and most studied, example of I,I-eliminationoccurs where no fJ-H atoms are available-the operation of factor (e)above-in the hydrolysis of haloforms, e.g. CHCI3 (77), with strongbases. This involves an initiall,l-elimination, probably via a two-step,i.e. 1,I-ElcB, pathway, to yield a dichlorocarbene intermediate (78);

HOS""< H H20Ie.. eOH/H 0CCI] +:! CCI 2 -I CCI2~[, CO + HC02e

fast 4. S ow ast

(77) (78)

The hydrolysis, as expected, follows the rate law,

Rate = k[CHCI]][eOH]

and the fast, reversible first step is supported by the fact that deuteratedchloroform, CDCI3 , is found to undergo base-catalysed exchangewith H 20 (loss of D) much faster than it undergoes hydrolysis.Further support for the above mechanism comes from the observation that HCCI3 is relatively inert towards PhSe alone; but will, ifeOH is added, then react very rapidly to form HC(SPhh, i.e. PhS8

while not nucleophilic enough to attack HCCh will attack the highlyreactive CCl2 . This dichlorocarbene is a highly electron-deficientspecies and (if generated in a non-protic solvent) will add to thedouble bond of (electron-rich) alkenes, e.g. cis 2-butene (79), toform cyclopropanes, e.g. (80), a 'trapping' reaction (ct. p. 50):

Me Me(79) Me Me ~"=I C

Me,COeCHCI]~ CCI2 -----.

ID C.H. (78) Cl (80)

Under suitable conditions, this can be a useful preparative methodfor cyclopropanes; another preparative 'trapping' reaction of CCI2is its electrophilic attack on phenols in the Reimer-Tiemann reaction(p.290).

It should however, be emphasised that in protic solvents, with thecommon bases, and with substrates containing fJ-H atoms 1,1elimination occurs to only a small extent if at all.

(74)

CH 2 CH 2

MeC~21 - Mec~ I~H CH2

(75)

B~H H~ I~

MeCH zCH 2CH-cP - MeCH 2CH 2l:;.<;H

(73) (75)

such intramol~cular 'insertions' to form cyclopropanes being acommon reactIOn of suitable carbenes; it is an example of 'internaltrapping' (ef p. 50). Only 4 /~ of (76) was isolated from the reaction

9.9 PYROLYTIC SYN ELIMINATION

There are a number of organic compounds including esters-especiallyacetates, xanthates (see below}-amine oxides, and halides that undergo

268 Elimination reactions 9.9 Pyrolytic SYN elimination 269

pyrolytic elimination of HY, in the absence of added reagents, eitherin inert solvents or in the absence of solvent-in some cases in thegas phase. In general these eliminations follow the rate law,

Rate = k[substrate]

but are usually distinguishable from E I eliminations (that follow thesame rate law) by the degree of SYN stereoselectivity that they exhibit.They are sometimes referred to as Ei eliminations (elimination, intramolecular), and the degree of SYN stereoselectivity reflects the extentto which they proceed via cyclic transition states, e.g. (81) below, thatwould dictate a SYN pathway.

The reaction that is perhaps ofthe greatest synthetic utility-becauseit proceeds at relatively low temperatures-is the Cope reaction oftertiary amine oxides, e.g. (82):

A~HH H ~ fAr ...u~....~y.....:. H J* ----. ArcJH (83)

(J /

NMe2"'1000 ~"'\ ~ H: NMe 2

H eO : / H H NMe2

H·······O '0"""'-

(82) (81)

The leaving groups, Hand NMe 20, must assume a syn-periplanarconformation, with respect to each other, to be close enough togetherto permit the development of the O .. ·H bond in the T.S. (81); theproducts are the alkene (83) and N,N-dimethylhydroxylamine. TheCope reaction, proceeding via this tight, essentially planar fivemembered T.S., exhibits the greatest degree of SYN stereoselectivityof any of these reactions.

The pyrolysis of xanthates (84)-the Chugaev reaction-and ofcarboxylic esters (85) differ from the above in proceeding via sixmembered, cyclic transition states, e.g. (86) and (87), respectively:

-Ii" ~ [_~H ]~ RJHH C "I r 0 COS~ 'SMe H····S·yc" H /I --+ +

SMe H,S_C\ MeSH(84) (86) SMe

The six-membered rings in these T.S.s are more flexible than thefive-membered T.S.-(81) above-and need not be planar (cf cyclohexanes v. cyclopentanes). Elimination may thus proceed, in part atleast, from conformations other than the syn-periplanar, with theresult that the degree of SYN stereoselectivity in these eliminationsmay sometimes be lower than that observed in the Cope reaction.Both reactions require higher temperatures than for the Cope reaction,carboxylic esters particularly so.

One of the major advantages of this group of elimination reactions,as a preparative method for alkenes, is that the conditions are relativelymild, in particular any acidity/basicity is low. This means that it ispossible to synthesise alkenes that are labile, i.e. which isomeriseduring the course of alternative methods of synthesis through bondmigration (into conjugation with others), or molecular rearrangement.Thus pyrolysis of the xanthate (88) of the alcohol (89) results in theformation of the unrearranged terminal alkene (90), whereas the moreusual acid-catalysed dehydration of (89) results in rearrangement inthe carbocationic intermediate (91, cf. p. 111), and thus in formationof the thermodynamically more stable, rearranged alkene (92):

MeS"-C=S/. '.:::...

XH 0) H

CS,jeOH '-< C( IiMe)C- H-CH)~ Me)C-CH-CH2 --+ Me)CCH=CH 2

(89) (88) (90)

(I) +H~ !t(2) -H,O

Me MeI $ $ I -H"

Me2C-CHMe --+ Me 2C'-TMe -- Me2C=CMe2

H

(9Ia) (9Ib) (92)

Pyrolysis of alkyl chlorides and bromides (alkyl fluorides are toostable; alkyl iodides lead to some alkane, as well as alkene, throughreduction by the eliminated HI) also results in the formation of alkenes,but temperatures up to 600° are required, and the elimination isseldom of preparative use; paradoxically it is the type that has receivedthe most detailed study. A wholly concerted 1,2-elimination ofhydrogen halide would involve a highly strained, four-memberedT.S. It seems not unlikely therefore that a good deal of C-Halbond-breaking takes place in advance of the C-H bond-breaking: ahigh degree of 'carbocationic character' thereby being developed atthe C-Hal carbon atom. It thus comes as no surprise to find thateliminations of HHal are observed to exhibit less SYN stereoselectivity than the others. Further mention will be made of Ei concertedeliminations, and of other reactions involving cyclic T.S.s, subsequently (p. 340).(87)

Ii",5000

(85)

R••HXH

Y (?H d~'R'

10Carbanions and their reactions

10.1 CARBANION FORMATION, p. 271.10.2 CARBANION STABILISATION, p. 273.10.3 CARBANION CONFIGURATION, p. 276.10.4 CARBANIONS AND TAUTOMERISM, p. 277:

10.4.1 Mechanism of interconversion, p. 278; 10.4.2 Rate andstructure, p. 279; 10.4.3 Position of equilibrium and structure,p.280.

10.5 CARBANION REACTIONS, p. 284:10.5.1 Addition, p. 284: 10.5.1.1 Carbonation, p. 284; 10.5.2Elimination, p. 285: 10.5.2.1 Decarboxylation, p. 285; 10.5.3Displacement, p. 287: 10.5.3.1 Deuterium exchange, p. 288;10.5.3.2 Carbanion nucleophiles, p. 288; 10.5.3.3 ReimerTiemann reaction, p. 290; 10.5.4 Rearrangement, p. 292; 10.5.5Oxidation, p. 294; 10.5.6 Halogenation of ketones, p. 295.

In theory any organic compound such as (I) that contains a C- H bond,i.e. nearly all of them, can function as an acid in the classical sense bydonating a proton to a suitable base, the resultant conjugate acid (2)being a carbanion (cf p. 21) :

R 3C-H + B: ~ R 3Ce + BHEil

(I) (2)

In considering relative acidity, classically it is only the thermodynamics of the situation that are of interest in that the pKa

value for the acid (cf. p. 54) can be derived from the equilibriumabove. The kinetics of the situation are normally of little significance, as proton transfer from atoms such as 0, N, etc., is extremelyrapid in solution. With carbon acids such as (1), however, the rate atwhich proton is transferred to the base may well be sufficiently slowas to constitute the limiting factor: the acidity of (1) is thencontrolled kinetically rather than thermodynamically (cf. p. 280).

There are, however, other methods of generating carbanions thanby proton removal as we shall see below. Carbanion formation isimportant-apart from the inherent interest of the species-because of

10.1 Carbanion formation 271

their participation in a wide variety of reactions of synthetic utility:many of them of especial value in that they result in the formation ofcarbon-earbon bonds (cf p.221).

10.1 CARBANION FORMATION

The most general method of forming carbanions is by removal of anatom or group X from carbon, X leaving its bonding electron pairbehind:

By far the most common leaving group is X = H where, as above, it isa proton that is removed, (I) ---. (2), though other leaving groups arealso known, e.g. CO2 from the decarboxylation (p. 285) of RC02 e(3), or CIS from Ph3C-C1 to yield the blood-red, ether soluble salt(4):

Ph3C-Cl ~ Ph3CeNaEil

(4)

Hardly surprisingly the tendency of alkanes to lose proton and formcarbanions is not marked, as they possess no structural features thateither promote acidity in their H atoms, or are calculated to stabilisethe carbanion with respect to the undissociated alkane (cf carboxylicacids, p. 55). Thus CH 4 has been estimated to have a pKa value of ~43,compared with 4·76 for MeC02 H. The usual methods for determiningpKa do not, of course, work so far down the acidity scale as this, andthese estimates have been obtained from measurements on theiodide/organo-metallic equilibria:

RM + R'I ~ RI + R'M

The assumption is made that the stronger an acid, RH, is the greaterwill be the proportion of it in the form RM (e.g. M = Li) rather thanas RI. Determination of the equilibrium constant K allows a measureof the relative acidity of RH and R'H, and by suitable choice of pairsit is possible to ascend the pKa scale until direct comparison can bemade with an RH compound whose pKa has been determined byother means.

Thus Ph 3C-H (5) is found to have a pK a value of 33, i.e. it is a

272 Carbanions and their reactions 10.2 Carbanion stabilisation 273

very much stronger acid than CH4 , and the carbanion (4) may beobtained from it, preparatively, by the action of sodamide, i.e. eNH2 ,

in liquid ammonia:

Ph3Ce may also be obtained, as we saw above, by the action of sodiumon Ph3C-Cl (3) in an inert solvent; the resulting solution of sodiumtriphenylmethyl is used as a very strong organic base (cf. p. 230)because of the proton-appropriating ability of the carbanion (4).Alkenes are slightly stronger acids than the alkanes-CH2 =CH2 hasa pKa value of 37-but the alkynes are very much more strongly acidic,and HC_CH itself has a pKa value of 25. The carbanion HC=C8

(or of course RC C8) may be generated from the hydrocarbon with8NH2 in liquid ammonia: these acetylenic anions are of some syntheticimportance (cf p. 223).

Hardly surprisingly, the introduction of electron-withdrawing substituents also increases the acidity of hydrogen atoms on carbon.Thus we have already seen the formation of a somewhat unstablecarbanion, 8CC13, in the action of strong bases on chloroform (cfp. 267), and the pKa values of HCF3 and HC(CF3)3 are found to be=28 and 11, respectively. The effects with substituents that candelocalise a -ve charge, as well as having an electron-withdrawinginductive effect, are even more marked; thus thepKa values ofCH3CN,CH3COCH3and CH3N0 2 are found to be 25, 20 and 10·2, respectively.With CH3N0 2 , the corresponding carbanion, 8CH2N0 2 , may beobtained by the action of 80Et in EtOH, or even of 80H in H2 0(cf. p. 227); but small concentrations of carbanion must be developedin aqueous solution even from the less acidic carbonyl compounds toenable the aldol reaction (cf p. 224) to take place.

A table of some pKa values for carbon acids is appended for convenience, before going on to discuss the factors that can contributeto the relative stabilisation of carbanions :

pK. pK.CH 4 43 CH 2(C02Eth 13-3CH 2=CH 2 37 CH 2(CNh 12C.H. 37 HC(CFJh 11PhCH J 37 MeCOCH 2C02Et 10·7Ph 3CH 33 CH 3 N02 10·2CFJH 28 (MeCOhCH2 8·8HC=CH 25 (MeCOhCH 6CHJCN 25 CH 2(N02)2 4CH JCOCH 3 20 CH(N02h 0C.HsCOCH 3 19 CH(CNh 0

10.2 CARBANION STABILISATION

There are a number of structural features in R- H that promote theremoval of H by bases through making it more acidic, and also featuresthat serve to stabilise the resultant carbanion, R8 ; in some cases botheffects are promoted by the same feature. The main features that serveto stabilise carbanions are (cf. factors that serve to stabilise carbocations, p. 104): (a) increase in s character at the carbanion carbon,(b) electron-withdrawing inductive effects, (c) conjugation of thecarbanion lone pair with a polarised multiple bond, and (d) aromatisation.

The operation of (a) is seen in the increasing acidity of the hydrogenatoms in the sequence: CH3-CH3 < CH2 =CH2 < HC CH; theincrease in acidity being particularly marked (see table above) ongoing from alkene to alkyne. This reflects the increasing s characterof the hybrid orbital involved in the (J bond to H, i.e. Sp3 < Sp2 < Spl.

The s orbitals are closer to the nucleus than the corresponding porbitals, and they are at a lower energy level; this difference is carriedthrough into the hybrid orbitals resulting from their deployment.The electron pair in an Spi orbital is thus held closer to, and moretightly by, the carbon atom than an electron pair in an Sp2 or Sp3

orbital (effectively, the apparent electronegativity of the carbon atomincreases). This serves not only to make the H atom more easily lostwithout its electron pair, i.e. more acidic, but also to stabilise theresultant carbanion.

The operation of (b) is seen in HCF3 (pKa = 28) and HC(CF3h(pKa = 11), where the change from CH4 (pKa = 43) is brought aboutby the powerful electron-withdrawing inductive effect of the fluorineatoms making the H atom more acidic, and also stabilising the resultantcarbanions, 8CF3 and 8C(CF3h by electron-withdrawal. The effectis naturally more marked in HC(CF3h where nine F atoms are involved-<:ompared with only three in HCF3-despite the fact that they arenot now operating directly on the carbanion carbon atom. We havealready referred to the formation of 8CCl 3from HCCl3(p.267), wherea similar electron-withdrawing inductive effect must operate. This islikely to be less effective with Cl than with the more electronegative F,but the deficiency may be overcome to some extent by the delocalisationof the carbanion electron pair into the vacant d orbitals of the secondrow element chlorine-this is, of course, not possible with the firstrow element fluorine.

The destabilising influence of the electron-donating inductive effectof alkyl groups is seen in the observed carbanion stability sequence:

CH 3 8 > RCH 28 > R2CH 8 > RJC8

Hardly surprisingly, it is the exact reverse of the stability sequence forcarbocations (p. 83).

274 Carbanions and their reactions 10.2 Carbanion stabilisation 275

The operation of (c) is by far the most common stabilising feature,e.g. with CN (6), C=O (7), N02 (8), C02 Et (9), etc. :

B~HI H 6-

CH2~C=N P

(6)

[6CH 2-C=N.-+ CH2=C=N6] + BHe:>

(lOa) (lOb)

pK. = 25

The operation of (d) is seen in cyclopentadiene (14) which is foundto have a pKtJ value of 16 compared with ~ 37 for a simple alkene.This is due to the resultant carbanion, the cyclopentadienyl anion (15),being a 6'7T electron delocalised system, i.e. a 4n +2 Hiickel systemwhere n = I (cf p. 18). The 6 electrons can be accommodated in threestabilised 7t molecular orbitals, like benzene, and the anion thusshows quasi-aromatic stabilisation; it is stabilised by aromatisation:

(t5)

MeCOCI----.Alel,

QB:'\ H H

(t4)

(16)

(17) (t8)

Its aromaticity cannot, of course, be tested by attempted electrophilicsubstitution, for attack by XEll would merely lead to direct combinationwith the anion. True aromatic character (e.g. a Friedel-Crafts reaction)is, however, demonstrable in the remarkable series of extremely stable,neutral compounds obtainable from (15), and called metallocenes, e.g.ferrocene (16), in which the metal is held by 7t bonds in a kind ofmolecular 'sandwich' between the two cyclopentadienyl structures:

It is also possible to add two electrons to the non-planar,non-aromatic (cf p. 17) cyclooctatetrane (17) by treating it withpotassium, thereby converting it into the isolable, crystalline salt ofthe cyclooctatetraenyl dianion (18):

This too is a Hiickel 4n + 2 p electron system (n = 2, this time) andshows quasi-aromatic stability; stabilisation by aromatisation hasagain taken place, remarkably this time in a doubly charged carbanion(18).

pK. = 20

pK. = 24

[

Me Me ]eCH2-t=0'-+ CH 2=t-Oe + BH81

(tta) (tlb)

B~H MeI ICH2~C=0 ~

(7)

B~ H [ ]I 81 81 81CH2~N=0 ~ eCH2-N=0.-+ CH2=N-Oe + BH81 pK. = 1Q.2

I I IOe Oe Oe

(8) (t2a) (t2b)

There is in each case an electron-withdrawing inductive effect increasingthe acidity of the H atoms on the incipient carbanion atom, but thestabilisation of the resultant carbanion by delocalisation is likely tobe of considerably greater significance. Overall, N02 is much themost powerful as might have been expected. The marked effect ofintroducing more than one such group on to a carbon atom may beseen from the table of pKtJ values above (p. 272); thus CH(CNh andCH(N02)3 are as strong acids in water as HCl, HN03 etc. The questiondoes arise however, about whether (lOab), (l1ab) and (l2ab) ought tobe described as carbanions: °and N are more electronegative than Cand (lOb), (lIb) and (l2b) are likely to contribute markedly more tothe hydrid anion structure than (lOa), (lla) and (l2a), respectively.

The carboxylate group, e.g. C02Et (9), is less effective in carbanionstabilisation than the C=O group in simple aldehydes and ketones,as may be seen from the sequence of pKtJ values: CH2(C02Et)2' 13·3;MeCOCH 2C02Et, 10·7; and CH2(COMeh, 8·8. This is due to theelectron-donating conjugative ability of the lone pair of electrons onthe oxygen atom of the OEt group:

B~H [ ]I 6+ 1-

CH+-C=O P eCH2-C=0 CH2=C-Oe + BH81~I c:1.-+"...1\: OEt : OEt \: OEt

(9) (l3a) (13b)

With second row elements, as we saw above, any inductive effectthey exert may be complemented by delocalisation, through use oftheir empty d orbitals to accommodate the carbanion carbon atom'slone pair of electrons; this can happen with S in, for example, anArS02 substituent, and also with P in an R 3PEll substituent.

276 Carbanions and their reactions 10.4 Carbanions and tautomerism 277

10.3 CARBANION CONFIGURATION

In theory a simple carbanion ofthe type R3Cs could assume a pyramidal(Sp3) or a planar (Sp2) configuration, or possibly something in betweendepending on the nature of R. The pyramidal configuration wouldbe preferred on energy grounds, as the unshared electron pair wouldthen be accommodated in an Sp3 orbital (19) rather than in thehigher energy, unhybridised p orbital of the planar configuration.The. pyram~dal configuration is, of course, the one adopted by~ertlary aO?lOes, R 3N:, with which simple carbanions, R3ce areIs?electroOl~; no doubt ready inversion of configuration takes placewith carbaOlons (l9a~ 19b) just as it does with the amines:

o OH» I

MeC-CHC02 Et i=! MeC=CHC02EtIH

(23a) Keto (23b) Enol

HI ~ ~CH2 -N=0 i=! CH 2=N-OH

I IOe 0e

(24a) (24b)

dione (20), with a pKa value of 8·8, and cyciohexan-l,3-dione (21) areboth readily soluble in aqueous NaOH (though not in water), andgive a red colour with FeC1 3 solution (cf phenol), the formally similar1,3-diketone (22) does neither:

0:X;0 oQo oJfzo(20) (21) (22)

The H atom flanked by the two C=O groups in (22) exhibits hardlyany more acidic character than the analogous one in the correspondinghydrocarbon. The different behaviour of (22) stems from the fact thatafter proton removal, the carbanion's lone pair would be in an Sp3orbital more or less at right angles to the p orbitals on each of theadjacent carbonyl carbon atoms (cf. p. 259): no Sp3/p overlap couldthus take place, consequently there would be no stabilisation of the-ve charge through delocalisation, and the (unstabilised) carbaniondoes not, therefore, form.

Tautomerism, strictly defined, could be used to describe the reversibleinterconversion of isomers, in all cases and under all conditions.In practice, the term has increasingly been restricted to isomers thatare fairly readily interconvertible, and that differ from each otheronly (a) in electron distribution, and (b) in the position of a relativelymobile atom or group. The mobile atom is, in the great majority ofexamples, hydrogen, and the phenomenon is then referred to asprototropy. Familiar examples are p-ketoesters, e.g. ethy12-ketobutanoate (ethyl acetoacetate, 23), and aliphatic nitro compounds, e.g. nitromethane (24):

10.4 CARBANIONS AND TAUTOMERISM

(0)(4)

E1c~ P

/ \"R"R R'

(l9a) 09b)

Evidence in support ofa preferred Sp3 configuration is provided bythe observation that reactions which involve the formation of carbanion intermediates at bridgehead positions often take place readily; while those that would have involved the corresponding carbocation (Sp2) intermediates do not (cf. p. 86).

In organometallic compounds of the form RR'R"C- M, pretty wellthe whole spectrum of bonding is known from the essentially covalentvia the polar-covalent, RR'R"C6 - - MH , to the essentially ionic'RR'R"CSMal. In their reactions, predominant retention racemisation'and inversion of configuration have all been observed" the outcom~in a particular case depending not only on the alkyl re~idue, but alsoon the metal, and particularly on the solvent. Even with the mostionic examples it seems unlikely that we are dealing with a simplecarb~nion; thus in the reaction of Ed with [PhCOCHMe]S Mal, therelative rates under analogous conditions are found to differ over arange of ::::: 1<f for M = Li, Na and K.

Car~ani.ons which have su~stitu.ents capable of conjugativedelocahsatlOn of the electron pair wIll perforce be planar (Sp2), inorder to allow the maximum orbital overlap of the p orbital with thoseof the substituent, e.g. (4) and (l 0):

Such interconversions are catalysed by both acids and bases.Where such alignment is prevented by structural or steric featuresthe expected stabilisation may not take place. Thus while pentan-2,4~

Pseudo-acid Aci-form

10.4.1 Mechanism of interconversion

carbanionintermediate

279

~H H

(29)

10.4.2 Rate and structure

(30)

0 6 -

E 6-MeC:..:..:CHC02Et

(27)

(28)

Me", H [ Me J~• Et,N: ","',©if ~ ©=fUN"" ~

(23a) and (23b) are tautomers: quite distinct, chemically distinguishableand different species, readily interconverted but, in this case, actuallyseparately isolable in a pure state. The two structures written for thecarbanion intermediate (27) are mesomers: they have no real existenceat all, they are merely somewhat inaccurate attempts to represent theelectron distribution in the carbanion, which is a single individualonly. It is perhaps better to represent (27) by a single structure of theform,

but this is still not wholly satisfactory in that it does not convey theimportant fact that more ofthe negative charge on the anion is locatedon the more electronegative oxygen, rather than on the carbonatom. Indeed, though we have referred (and will, for convenience,continue to refer) to species such as (27) as carbanions, they arealso-and perhaps more correctly-referred to as eno/ate anions. Itis very common to find a pair of tautomers, such as (23a) and (23b),'underlain' as it were by a single, stabilised carbanion/enolate anionsuch as (27).

The other extreme case, i.e. wholly intramolecular proton transferpathway (b), is seen in the Et 3N: catalysed conversion of the opticallyactive substrate (28) into (29) :

Here it is found that the rate of loss of optical activity and the rate ofisomerisation are identical, and if the reaction is carried out in thepresence of D 20 (five moles per mole of substrate) no deuterium isincorporated into the product. The reaction is thus wholly intramolecular under these conditions-no carbanion is involved-and isbelieved to proceed via a bridged T.S. such as (30). With a number ofsubstrates features of both inter- and intra-molecular pathways areobserved, the relative proportions being dependent not only on thesubstrate, but to a considerable extent on the base and solventemployed also.

10.4.2 Rate and structure

In virtually all the examples we have been talking about, the slow,rate-limiting stage is the breaking-or forming-of a C- H bond;

(Intramolecular)

(23b) Enol

OHR'OH I+:::t MeC=CHC02 EtB:

[.,cJ:vJ~". .;/

CH(26)

T.S.

Carbanions and their reactions

HI B:

R2C-CH=Y +:::t

~ B:MeC-CHC02 Et +:::t

I ROHH

(23a) Keto

(b)

(a)

278

Many of the compounds that undergo ready base-catalysed keto+=t enol prototropic changes, e.g. fJ-keto esters, I,3-(fJ-) diketones,aliphatic nitro compounds, etc., fonn relatively stable carbanions, e.g.(25), that can often be isolated. Thus it is possible to obtain carbanionsfrom the 'keto' forms of the fJ-keto ester (23a) and nitromethane (24a)and, under suitable conditions, to protonate them so as to obtain thepure enol forms (23b) and (24b), respectively. It thus seems extremelyprobable that their interconversion follows the intermolecular pathway (a). The more acidic the substrate, i.e. the more stable thecarbanion to which it gives rise, the greater the chance that prototropic interconversion will involve the carbanion as an intennediate.

The mechanism (a) nicely illustrates the difference between tautomerism and mesomerism that often gives rise to confusion. Thus takingethyl 2-ketobutanoate (23) as an exam-pIe,

[OJII

MeC-CHC02 Et0861

Met=CHC02 Et

(27)

Prototropic interconversions have been the subject of much detailedstudy, as they lend themselves particularly well to investigation bydeuterium labelling, both in solvent and substrate, and by chartingthe stereochemical fate of optically active substrates having a chiralcentre at the site of proton departure. Possible limiting mechanisms(cf SN I/SN2) are those: (a) in which proton removal and protonacceptance (from the solvent) are separate operations, and a carbanionintermediate is involved, i.e. an intermolecular pathway; and (b) inwhich one and the same proton is transferred intramolecularly:

B:'\ H 6 HI B: [R 2C-CH=Y ] R'OH . I

R C-CH=Y +:::t ! +:::t R2C=CH-Y (Intermolecular)2 R'O" B:

R2C=CH-y6

(25)carbanion

intermediate

k,

R3C-H + B: ~ R3Ce + BH ED (K = kdL ,)L,

10.4.3 Position of equilibrium and structure

In this context it is keto/enol systems that have been investigated byfar the most closely, and most of our discussion will centre on them.The relative proportion of the two forms was commonly determined

for the former involves a AG* term and the .latter a AG B term, andthere is no necessary relation between the two. It is very broadly truethat structural changes in the substrate that lead to greater (thermodynamic) acidity also tend to lead to its more rapid conversion intothe carbanion, but there are many exceptions as may be seen below:

281

%Enol in liquid

1·5 x 10- 4

7·7 X 10- 3

2·5 X 10- 1

1·28·0

30·076·489·2

(3) Vm.. 1650 em - 1

MeCOCH 3

CHz(COzEt)zNCCHzCOzEtCyclohexanoneMeCOCHzCOzEtMeCOCHPhCOzEtMeCOCHzCOMePhCOCHzCOMe

10.4.3 Position of equilibrium and structure

(I) vma,1718em- 1

(2) Vma, 1742 em-I

In simple carbonyl compounds, e.g. MeCOMe, the proportion of enolat equilibrium is extremely small; the main structural features thatresult in its increase may be seen in the table below:

chemically, e.g. by titration of the enol form with bromine underconditions such that the rate of keto/enol interconversion was verylow; it is, however, more accurate and more convenient to do thisspectroscopically, e.g. in the i.r. for ethyl 3-ketobutanoate:

o OH,,(I) (2) I (3)

MeC-CHz-C=O ~ MeC=CH-C=OI I

OEt OEt

(23a) Keto (23b) Enol

O"H··O o/H.··oI II I II

/c-,:::- /c" /c-,:::- /C,Me CH Me Me CH OEt

(31) (23)

Apart from any stabilisation effected with respect to the keto form, suchintramolecular hydrogen-bonding will lead to a decrease in the polarcharacter ofthe enol, and to a more compact, 'folded-up' conformationof the molecule, compared with the more extended conformation ofthe keto form. This has the rather surprising result that where keto

The major feature is a multiple bond, or a 1t orbital system such asPh, which can become conjugated with the C=C double bond in theenol form. C=O is clearly effective in this respect, with an ordinarycarbonyl C=O group being considerably more effective than theC=O in an ester group, cf. MeCOCHzCOzEt (8%) andCHz(COzEt}Z (7·7 x 10-3 °/0). The added effect of Ph may be seen incomparing MeCOCHzCOzEt (8%) with MeCOCHPhCOzEt (30%),and MeCOCHzCOMe (76,4%) with PhCOCHzCOMe (89·2%).

Another feature that will serve to stabilise the enol, with respect tothe keto, form is the possibility of strong, intramolecular hydrogenbonding, e.g. in MeCOCHzCOMe (31) and MeCOCHzCOzEt (23):

pK. k,(see-')

4 8·3 x 10- 1

8·8 1·7 x lO- z

10·2 4·3 x 10- 8

10·7 1·2 x 10- 3

12 1·5 x lO- z

13·3 2·5 x 1O-~

20 4·7 x 10- 10

CHz(NOzhCHz(COMehCH 3 NOzMeCOCHzCOzEICHz(CNhCHz(COzElhCH3COCH 3

Carbanions and their reactions280

this is one major respect in which carbon acids differ from those acidsin which the incipient proton is attached to 0, N, etc. The rate of suchC-H bond-breaking can often be measured by determining the rateof hydrogen isotope exchange with suitable proton (deuteron) donorssuch as DzO, EtOD, etc. It is interesting, though hardly surprising,to find that this kinetic acidity scale (defined by k 1 ) does not correlatedirectly with the thermodynamic acidity scale (defined by K) that wehave considered to date, i.e. pKa values;

Simple nitro compounds are particularly slow in their rate ofionisation, considering their relatively high acid strength; thusCH3NOz and MeCOCHzCOzEt have very much the same pKa , butthe former ionises more slowly by a factor of nearly lOs. This probablyreflects a greater degree of delocalisation of charge in the carbanionderived from CH3NOz than in that from CH 3COCHzCOzEt. Insuch cases both proton abstraction and donation tend to be slow,compared with those carbon acids in which the charge is more concentrated on carbon in their carbanions. This is borne out by theeffect ofC=N substituents on carbanions, where less charge delocalisation would be expected than with a C=O substituent; thus CHz(CN)zis found to have very much the same k 1 value as CHz(COMe)z,despite the fact that its pKa value is larger (i.e. acidity lower) by 3·2units. The relation between pKa and k 1 can be much affected by thesolvent, however.

282 Carbanions and their reactions 10.4.3 Position of equilibrium and structure 283

and enol forms can actually be separated, the latter usually has thelower b.p. despite its hydroxyl group. The effectiveness ofintramolecularhydrogen-bonding in stabilising the enol, with respect to the keto, formis seen on varying the solvent, and particularly on transfer to ahydroxylic solvent, e.g. with MeCOCH2COMe (31):

two forms under the particular conditions being studied. An interestingsituation arises with aliphatic nitro compounds, e.g. phenylnitromethane (34), however. With this, the nitro-form-a yellow oil(34a) is the more stable of the two and, at equilibrium, predominates to the almost total exclusion of the aci-form (34b):

(34b)

Aci-form

Fig. 10.1

(35)

T,S' I

Nitro-form

~......5!!

LL. HI (fJJJ

PhCH-N'O

(34a)

1

Despite this fact, acidification of the isolable sodium salt of thecarbanion intermediate (35) yields only the less stable aci-form(34b)-a colourless solid. This happens because more rapid protonation takes place at the position of higher electron density, i.e.product formation under these conditions is kinetically controlled.The energy profile for the system has the fonn (Fig. 10.1),

HOG\ H HI Ql eOH J- Ql J- H IO I

PhCH-N=O +=! PhCH:.:..:N:.:..:O +=! PhCH=N-OI H IO I eOH I0e 0e 0e

(34a) (35) (34b)

i.e. the transition state between (35) and (34b)-T.S. 2-is at a lowerenergy level than that between (35) and (34a)-T.S. 1 , reflecting thegreater ease of breaking an O-H than a C-H bond. Although theimmediate result of the acidification of (35) is thus the formation of(34b), which will, however, undergo spontaneous re-ionisation: theaci-fonn (an oxygen acid) will lose its proton faster than will thenitro-fonn (a carbon acid). Equilibrium is thus gradually established,leading to the slow, but inexorable, formation of the more stable(34a): the ultimate composition of the product is thus thermodynamically controlled.

H

B

9292765815

% EnolSolvent

Gas phaseHexaneLiquidMeCNH 20

(32a)

Thus the proportion of enol in the non-polar solvent hexane is thesame as in the gas phase, and higher than in the liquid itself, the latteracting as a somewhat polar auto-solvent; the proportion drops againin the more polar MeCN, and more dramatically in water. What ishappening is the increasing relative stabilisation of the keto form bysolvation, this being particularly marked in water where intermolecularhydrogen bonding of the keto form's C=O group can now take placeas an alternative to its enolisation. The behaviour of MeCOCH2CO 2Et(23) is closely analogous; thus the 8 %enol present in the liquid risesto 46 %in hexane and to 50 %in the gas phase, but drops to 0·4 %indilute aqueous solution. The percentage of enol present is alsodependent on the temperature.

A particularly interesting-and extreme~xample is provided bya comparison of MeCOCOMe (32) and the cylic 1,2-diketone, cyclopentan-I,2-dione (33):

/H.° "0PHCH 2 Me

(32b) (33a) (33b)

5.6 x 10- 3 % 'l::IOO%

With (32), despite the intramolecular hydrogen-bonding possible inthe enol form (32b), the equilibrium lies essentially completely over infavour of the keto form (32a), because this can take up an anti-conformation in which the two electronegative oxygen atoms are as farfrom each other as possible, and in which the carbonyl dipoles areopposed. With (33), the C=O groups are 'locked' in the syn-conformation in both keto (33a) and enol (33b) forms, and the intramolecular hydrogen-bonding open to (33b), but not to (33a), thendecides the issue.

In the above examples the composition of the equilibrium mixtureis, of course, governed by the relative thermodynamic stability of the

~MEB /o8MEB

o=c=o -+ R-C\J ~

°(36)

10.5.1.1 Carbonation

A further interesting, and synthetically useful, reaction of carbanionsand of organometallic compounds acting as sources of negativecarbon-is addition to the very weak electrophile CO 2 , to form thecorresponding carboxylate anion (36}-carbonation :

285

Me CO 8Lia:>

" / zC=C

/ "H Me

(39)

H$-+ R-Hfast

co,-+

Me" .8Lia:>

C=C

/ "H Me

(37)

Li-+

e<P-C...C"'R -+ cOz + R8II slow

°(41) (42)

Me Br" /C=C/ "H Me

(38)

10.5.2.1 Decarboxylation

lithium, by converting it into (39):

(40)

Rate = k[RCO z8]

Another example is decarboxylation.

The yield of (39) was ~ 75/~, while that of its geometrical isomeridewas <5/~'

10.5.2.1 Decarboxylation

Loss of CO 2 from carboxylate anions (41) is believed to involve acarbanion intermediate (42) that subsequently acquires a proton fromsolvent, or other source:

subsequent proton abstraction being rapid. Decarboxylation shouldthus be promoted by electron-withdrawing substituents in R thatcould stabilise the carbanion intermediate (42) by delocalisation of itsnegative charge. This is borne out by the very much readier decarboxylation of the nitro-substituted carboxylate anion (43) than of

Loss of CO 2 is normally rate-limiting, i.e. the rate law is,

10.5.2 Elimination

We have already seen examples ofcarbanions involved as intermediates,e.g. (40), in elimination reactions, i.e. those that proceed by the E IcBpathway (p. 251), for example:


10.5.1 Addition

We have already discussed a large group of reactions in which carbanions add to the C=O group (cf. pp. 221-234), including examplesof intramolecular carbanion addition, e.g. an aldol reaction (p. 226),Dieckmann reaction (p. 230), and the benzilic acid rearrangement(p. 232), and also to the C=C-C=O system, the Michael reaction(p.200).

10.5 CARBANION REACfIONS

Carbanions can take part in most of the main reaction types, e.g.addition, elimination, displacement, rearrangement, etc. They are alsoinvolved in reactions, such as oxidation, that do not fit entirelysatisfactorily into this classification, and as specific-ad hoc-intermediates in a number of other processes as well. A selection of thereactions in which they participate will now be considered; many areof particular synthetic utility, because they result in the formation ofcarbon--earbon bonds.

It occurs with the alkyls, aryls or acetylides of metals more electropositive t.han magnesium, but including Grignard reagents, and isoften carried out by adding a solution ofthe organometallic compoundin an inert solvent to a large excess of powdered, solid CO 2 ; it is aparticularly useful method for the preparation of acetylenic acids.The Kolbe-Schmidt reaction (p. 291) is another example of carbanioncarbonation.

This reaction has been used a good deal in the study of carbanions,to detect their formation by converting them into stable, identifiableproducts. Thus substantial retention of configuration in an alkenylcarbanion (37) has been demonstrated, in the reaction of (38) with

287

°IIC

0-:/

10.5.3 Displacement

~ hO --+ ~e -.=+~~~~C' ~~,) ~~~HH ---ol H

(eO 1(50)PhCOMc

--+

(49)

~ -:/0 +==2N C

IH-O

(51)

IXp-unsaturated acids, R2CHCR=CHC02H, probably decarboxylateby this pathway also, as it has been shown that they isomerise to thecorresponding py-unsaturated acid prior to decarboxylation.

Another example in which the free acid undergoes ready decarboxylation, but this time via a carbanion intermediate (50, actuallyan ylid), is pyridine-2-carboxylic acid (51), which is decarboxylatedvery much more readily that its 3- or 4-isomers :

The ylid intermediate (50) can be 'trapped' by carrying out the decarboxylation in the presence of carbonyl compounds, e.g. PhCOMe,to yield the carbanion addition product, e.g. (52); this process canindeed be used preparatively. The reason for the much easier decarboxylation of (51), than of its 3-, and 4-isomers, is the stabilisationthat the NEIl can effect on the adjacent carbanion carbon atom in theintermediate ylid (50).

10.5.3 Displacement

Carbanions, or similar species, are involved in a variety of displacementreactions, either as intermediates or as attacking nuc1eophiles.

Some evidence for this mode of decarboxylation of the free acid hasbeen obtained by 'trapping' the enol intermediate (48). ~'Y

Unsaturated acids (49) probably also decarboxylate by an analogouspathway:

(48)

(43)


e<Y--CL"CMe NO --+ CO2 +II 2 2

°

(47)

286

e(jL.c-I(:H 2COMe --+ cO2 +II

° (46)

1IlMe2C=N-Oe

I°e

(44)

Similar ease of decarboxylation is seen in Hal 3CCH 2C0 2e, 2,4,6(NOzh C6HzCOz

e , etc., but the reaction is not normally of preparative value with the anions of simple aliphatic acids other thanMeCOz

e .Evidence that carbanion intermediates, e.g. (44), are involved is

provided by carrying out the decarboxylation in the presence ofbromine. This is without effect on the overall rate of the reaction butthe end-product is now MezCBrNOz rather than MezCHNOzunder conditions where neither MezC(NOz)COze nor MezCHNOzundergoes bromination. The bromo product (45) arises from rapidattack of Brz on the carbanion intermediate (44), which is thereby'trapped' (cf. the base-catalysed bromination of ketones, p. 295):

B'P-Br Br~ I

Me2CN02 --+ Me2CN02 + Bre

(44) (45)

C=O can also act like N02 , and the anions of p-ketoacids (46) aredecarboxylated very readily:

The overall rate law is, however, found to contain a term involving[ketoacid] (47) as well as the term involving [ketoacid anion]. Theready decarboxylation of the ~-ketoacid itself is probably due toincipient proton transfer to C=O through hydrogen-bonding in(47):/-H.

0) \0I II

-:/c~/c,° CH 2 Me

eOB e Rfu(MeCO)2CH2 i=! (MeCO)2CH + R-Br - (MeCOhCH-R + Bre

(55) (56)

The SN2 character of the process has been confirmed kinetically,and in suitable cases inversion of configuration has been demonstrated

10.5.3.2 CarbanioD Ducleophiles

Both overt carbanions and organometallic compounds, such asGrignard reagents, are powerful nucleophiles as we have seen in theiraddition reactions with C=O (p. 221 et seq.); they tend therefore topromote an SN2 pathway in their displacement reactions. Particularlyuseful carbanions, in preparative terms, are those derived fromCH2 (C02Eth, (3-ketoesters, 1,3-((3-)diketones, e.g. (55), acyanoesters, nitroalkanes, etc.-the so-called 'reactive methylenes':

10.5.3.1 Deuterium exchange

The ketone (53) is found to undergo exchange of its (X-hydrogen atomfor deuterium when treated with base (90D) in solution in D20.When the reaction was carried out on an optically active form of (53),it underwent loss of optical activity (racemisation) at the same rate asdeuterium exchange. When the analogous compound containing Din place of H underwent exchange in H 20, there was found to be akinetic isotope effect (kH!ko) on comparing the rates of exchange forthe two compounds:

289

(60)(59)

RCH1CH 1I

(62) H

- RCH=CH 1 + NalDCl e

(63)

10.5.3.2 Carbanion nucleophiles

He/H 0- RCH(OEth~ RCHO

d- d+

~MgBr

CH(OEth

c6Et

(58)

RCH 1CH 2'(61) H

RC~CH2~l Na lD

e

at the carbon atom attacked in RBr above. The alkylated product (56)still contains an acidic hydrogen, and the process may be repeated toyield the dialkyl product, (MeCOhCRR'. Synthetically useful alkylations can also be effected on acetylide anion (57):

eNH, RBrHC:=CH i=! HC:=Ce + R-Br - HC:=C-R + Bre

(57)

Here too, a second alkylation can be made to take place yieldingRC=CR or R'C-CR. It should, however, be remembered that theabove carbanions-particularly the acetylide anion (57}-are theanions of very weak acids, and are thus themselves strong bases, aswell as powerful nucleophiles. They can thus induce elimination(p. 260) as well as displacement, and reaction with tertiary halidesis often found to result in alkene formation to the exclusion ofalkylation.

Grignard reagents can also act as sources of negative carbon indisplacement reactions, e.g. in the synthetically useful reaction withtriethoxymethane (ethyl orthoformate, 58) to yield acetals (59) and,subsequently, their parent aldehydes (60):

Thus leading to the disproportionation-alkane (62) + alkene (63}that is often observed as a side-reaction to the normal Wurtz coupling.

2Na' R'BrRCH 2CH 2-CI -. RCH 2CH 1

e Na lD -. RCH 1CH 2R'

(61)

This is, ofcourse, the Wurtz reaction, and support for such a mechanisminvolving carbanions (radicals may be involved under some conditions,however) is provided by the observation that in some cases it is possible,with optically active halides, to demonstrate inversion of configurationat the carbon atom undergoing nucleophilic attack. The carbanion,e.g. (61), can also act as a base and promote elimination:

It is also possible, under suitable conditions, to generate the alkyls(61) of more electropositive metals, e.g. sodium, and then subsequently to react these with alkyl halides:

o0,0 II- MeEtC-CPhfast I

D (±)

oII

MeEtC-CPhe

1oII eoo

MeEtC-CPh -I slow

H (+)

(53)


eoI

MeEtC=CPh

(54)

This all suggests slow, rate-limiting breaking of the C-H bond toform the stabilised carbanion intermediate (54), followed by fastuptake of DEll from the solvent D 20. Loss of optical activity occursat each C-H bond breakage, as the bonds to the carbanion carbonatom will need to assume a planar configuration if stabilisation bydelocalisation over the adjacent C=O is to occur. Subsequent additionof DEll is then statistically equally likely to occur from either side. Thisslow, rate-limiting formation of a carbanion intermediate, followedby rapid electrophilic attack to complete the overall substitution, isformalIy similar to rate-limiting carbocation formation in the SNlpathway; it is therefore referred to as the SE 1 pathway.

291

~CHOMe(68)

Cfvv---+above

10.5.3.3 Reimer-Tiemann reaction

anion of p-hydroxytoluene (p-cresol, 67):

.Njl®

66-<l "~6-

: 6-... <:=0I H

#

(72)

¢~¢tH~Me Me

(67a) (67b)

!o 0 0

O~O ~OMe Me CCI 2 Me CHCI 2

(67c) (70) (69)

In addition to the expected o-aldehyde (68), it is also possible to isolatethe unhydrolysed dichloro compound (69). Attack by CCI 2 at thep-position in (67c) yields the mtermediate (70) which, unlike theintermediate for o-attack, has no H atom that can be lost, as He, toallow the ring to re-aromatise; (70) thus just acquires a proton, onfinal acidification, to yield (69). The dicWoro compound (69) owesits resistance to hydrolysis partly to its insolubility in the aqueousbase medium, but also to the sterically hindered, neopentyl-typeenvironment (cf. p. 86) of the cWorine atoms.

The somewhat analogous Kolbe-Schmidt reaction involves CO 2 asthe electrophile in attack on powdered sodium phenoxide (64b):

o 0

OH II)

IeJC

II # 0

(64b)

The product is almost exclusively sodium o-hydroxybenzoate (salicylate, 71) only traces of the p-isomer being obtained; if, however, thereaction is carried out on potassium phenoxide the salt of the p-acidbecomes the major product. It has been suggested that the preferentialo-attack with sodium phenoxide may result from stabilisation of theT.S. (72) through chelation by NdB in the ion pair:

1

enolate anionintermediate

carbamon

+--

af3 -epoxyester

a -chloroester

R 0 R'

~Et02C R'


Oe 0 (0 eo

6 6~co' O~C1'©CHCI,

+--+ - -0(64a) (64b) (66) leoH

HO eO

©CHO H$ ©CHO+--

An interesting intramolecular displacement occurs in the Darzensreaction, in which carbanions derived from a-haloesters react withcarbonyl compounds to yield a-epoxyesters:

290

It is sometimes possible, e.g. with a-cWoroesters, actually to isolatethe enolate anion intermediate.

(65)

The product from phenoxide ion (64) is, after acidification, verylargely the o-aldehyde (salicylaldehyde, 65) plus just a small amountof the p-isomer. If both o-positions in the initial phenoxide anionare substituted, however, reaction then yields the p-aldehyde.

Some support for the reaction pathway suggested above is providedby what is observed when the analogous reaction is carried out on the

10.5.3.3 Reimer-Tiemann reactionThis involves an aryl carbanion/enolate anion (64), and also eC03

derived from the action of strong bases on HCCl3 (p. 267), thoughthe latter has only a transient existence decomposing to CCh, ahighly electron-deficient electrophile that attacks the aromatic nucleus:

The K E9 cation is larger and likely to be less effective in this role,so that attack on the p-position will therefore become more competitive.

The product is a sodium alkyl, as expected, but protonation andcarbonation yield the rearranged products (74) and (75~ respectively.It is not known whether the unrearranged sodium alkyl (76) is formedwhich then rearranges, or whether loss of CI and migration of Ph aresubstantially concerted, !l0 that the rearranged sodium alkyl (77) is

29310.5.4 Rearrangement

H Li lllIII I PhLi III e

Me2N-CHPh~Me2N-CHPh -+ Me2N-CHPhI I I

Me Me Me

(78)

i.e. in the order ofdecreasing ionic character of the carbon-metal bond.This coupled with a study of the relative migratory aptitude of psubstituted Ar groups-suggesting it is ArE9 rather than Ar' thatmigrates-strongly support the view that the 1,2-shift is carbanionic,rather than radical, in character.

Simple I,2-shifts ofalkyl, from carbon to carbon, that are carbanionicin character are essentially unknown. Examples are known, however,in which alkyl is involved in a 1,2-shift from other atoms such as NandS to a carbanion atom-the Stevens rearrangement:

K ~ Na > Li > Mg

formed directly. With Li in place of Na, however, it is possible toform the unrearranged lithium alkyl, corresponding to (76), as witnessedby the products of its protonation and carbonation, and then rearrangeit subsequently by raising the temperature. The tendency to rearrangement on reacting (73) with metals, or metal derivatives, is found todecrease in the order,

HIII I eOH III e

MeS-CHCOPh~ MeS-CHCOPh -+ MeS-CHCOPhI I I

PbCH2 PbCH2 PbCH2(79)

There is, however, some evidence which suggests that certain ofthese reactions may involve radical, rather than carbanion, intermediates. Very strong bases, e.g. PhLi, are required to remove aproton from the positively charged species (78), unless an electronwithdrawing substituent, such as C=O, is present, e.g. (79). PhCHzis found to migrate preferentially to Me (ct. 79), being the morestable of the two without an electron pair (cf. p. 105). Allyl andbenzyl ethers, e.g. (80), undergo the analogous Wittig rearrangement (to be distinguished from the Wittig reaction for the synthesisof alkenes, p. 233):

H Li lllI PhLi e He/H 20

O-CHPh~ O-CHPh -+ LillleO-CHPh ------ HO-CHPhI I I I

Me Me Me Me

(80)

Finally, there are base-induced rearrangements involving carbanions that proceed via 1,3-elimination to form cyclopropanone inter-

HI

Ph2C-CH2Pb

(74)

~co,Ph2C-CH2Pb

IC02

e Na lll

(75)

(77)

110/Pb

e I-+ Ph2C-CH2

Nalll

(76)

[JiJ ~jiJCarbocation T.S. carbanion T.S.

(Ze) (4e)


(73)

292

The former involves the accommodation of two electrons (those ofthe original R-C single bond), while the latter involves the accommodation of four electrons. Two electrons can be accommodated inthe available bonding molecular orbital, but the additional twoelectrons in the carbanion T.S. can be accommodated only in ananti-bonding molecular orbital of much higher energy. 1,2-Shifts ofaryl groups are known, however, e.g. in the reaction of the chloride(73) with sodium, but here some stabilisation of the carbanion T.S.is possible through delocalisation of the extra electrons by themigrating phenyl group:


Rearrangements that involve carbanions are found to be very muchless common than formally similar rearrangements that involvecarbocations (p. 109). This becomes more understandable if wecompare the T.S. for a 1,2-alkyl shift in a carbocation with that forthe same shift in a carbanion:

mediates, e.g. (8I)-the Favorskii rearrangement of a-haloketones,e.g. (82):

10.5.5 Oxidation

Carbanions can, under suitable conditions, be oxidised; thus thetriphenylmethyl anion (85) is oxidised, fairly slowly, by air:

295

°IIMeEtC-CPh

IHal (±)

(92)

10.5.6 Halogenation of ketones

° O·-~·.\}/'• II eOH .-:1 /"

MeEtC-CPh i=! MeEtC..:..:CPh fa"I slow "-. °H (+) D,O"" II(90) (91) MeEtT-CPh

o (±)

(93)

This intermediate is then attacked in a fast, non rate-limiting stepby anyone of the series of electrophiles-Oz, Brz, Iz, HzO, DzO,etc.-to yield end-products such as (92), (93), etc.; all of which willnecessarily be produced at the same rate. This process has a fonnalresemblance to slow, rate-limiting fonnation of a carbocationicintennediate, followed by rapid nucleophilic attack, in the SN1pathway; it is therefore referred to as an SE 1 process.

With ketones such as (94), that have alternative groups of IX-Hatoms to attack, two questions arise: (a) which group, the CHz orthe CH3 , is attacked preferentially, and (b) when one H has beensubstituted by halogen, will a second halogen become attached tothe same or to the other IX-carbon atom. So far as (a) is concerned,it is found that bromination of, for example, MeCHzCOCH3 , yields1- and 3-bromobutanones in virtually equal amount (both thesebromoketones then undergo very rapid further reaction, cf. p. 296).The inductive effect exerted by a simple alkyl group, R, thusappears to have relatively little effect on the acidity of ~, or on the


One of the earliest observations relating to the possible occurrenceof carbanions as reaction intermediates was that the bromination ofacetone, in the presence of aqueous base, followed the rate law,

Rate = k[MeCOMe][60H]

i.e. was independent of [Brz]' Subsequently it was shown that, underanalogous conditions, iodination took place at the same rate asbromination; as was to be expected from the above rate law. We havealready seen (p. 288) that base-induced deuterium exchange (in DzO),and racemisation, of the optically active ketone (90) occur at the samerate, and are subject to a kinetic isotope effect (kH > kD) when theIX-H atom is replaced by D, i.e. C-H bond-breaking is involved inthe slow, rate-limiting step. All these observations make the involvement of a common carbanion intermediate, e.g. (91), virtually inescapable:

H06

\ ~jC

~ 1\PhHC-CH2

! (81)

°IIC

I \PhH~.J CH2.Lb


°IIC eOH

I \ +=tPhHC CH 2CI

IHO~H (82)

294

HO" ;/'0

C -6 IPhCH 2CH 2C026 -- PhCH-CH 2

(84) (83)

The cyclopropanone intermediate (81) undergoes subsequent additionof eOH, followed by ring-opening to yield the more stable of the twopossible carbanions (83, benzyl> primary), followed by protonexchange to yield the rearranged carboxylate anion end-product (84).

The resultant radical (86) can, in turn, be reduced back to the carbanionby shaking with sodium amalgam. In suitable cases, e.g. (87), theoxidation of carbanions with one-electron oxidising agents, usuallyiodine, can be useful synthetically for forming a carbon--earbon bond,through dimerisation (-+ 88) of the resultant radical (89):

(MeCObCH6 I, (MeCObCH (MeCO)2CH-+ : -+ I

(MeCObCH 6 (MeCObCH (MeCObCH

(87) (89) (88)

Another useful synthetic reaction is the oxidative coupling of alkynes, RC==CH, induced by Cu(n) salts (e.g. acetate) in pyridinesolution:

2RO=ce~ 2RO=C· __ RO=C----C==CR

Almost certainly, the acetylide anion-fonned in the basicsolution-is oxidised by Cu(n) (another one-electron oxidisingagent) to the corresponding radical, which then undergoes dimerisation.

CX3 is a good leaving group because of the electron-withdrawinginductive effect of the three halogen atoms; this activates the carbonylcarbon atom in (100) to nucleophilic attack, and also stabilises the

stability of the resultant carbanion/enolate anion, (96):

~RCH2C-CH2 - RCH2C=CH2]

'" r,-,<D II I-b~ ~\.!.I e~ 0 Os

R.....CH*-C-#CH2 slow (95)II

(94~ (j)e~ [R .....CH-~CH3 - R.....CH=~:H3J

(96)

So far as (b) is concerned, an introduced halogen substituent, e.g.Br in (97), is found to exert very considerable influence on theposition at which further halogenation occurs:

297

o H-AII

CH 2-CMeIBr


OH 0I Br, II

RCH=CMe -+ RCH-CMeIBr

This then undergoes rapid, non rate-limiting attack by Br2 or anyother electrophile present.

To discover which of the groups of a-H atoms would be expectedto undergo preferential substitution in RCHzCOCH3 requires comparison of the formation of the relevant enols, (105) and (106):

Rate = k[ketone][acid]

and, as with the base-catalysed reaction, the rates of bromination,iodination, deuterium exchange and racemisation are identical. Thistime the common intermediate, whose formation is slow and ratelimiting, is the enol (104):

departing carbanion (10 I). The end-product, apart from the carboxylateanion (102), is the haloform (103), and the overall processRCH 2COCH3 --+ RCH 2C02e + HCX 3-is known as the haloformreaction. It has been employed as a diagnostic test for methyl ketones,using 12 and aqueous base as the resultant CHI3 ('iodoform') is yellow,has a highly characteristic smell, and is insoluble in the reactionmedium.

The halogenation of ketones is also catalysed by acids (generalacid catalysis, cf p. 74), the rate law observed is,

(105) (107) (106) (108)

Of these (105) is likely to be more stable than (106) as it has themore heavily substituted double bond of the two (cf. p. 26); thefavoured bromination product is thus expected to be (107). In fact,the acid-catalysed bromination of MeCHzCOCH3 is indeed found toyield about three times as much 3- as 1-bromobutanone.

It is also found, in contrast to bromination under base-catalysedconditions, that introduction of a further bromine into a monobromoketone is more difficult than was introduction of the initialone. It is thus normally possible, under acid conditions, to stopbromination so as to obtain the mono-bromo product, e.g. (107),preparatively. This is, of course, in contrast to under base conditions, where further bromination cannot be prevented and is followed, in suitable cases, by haloform cleavage (p. 296).

o 0II s II

-+ RCH 2C + CX3 -p RCH 2C + HCX3I I

OH Os

(101) (102) (103)

[RCH2~-CH..... Br - RCH2~:CH ..... BJ

(98)

~R .....CH-CCH2Br - R..... CH=CCH 2Br]

II Io sO

(99)


(~ eOH (? ()RCH 2C*-CX3 +=! RCH 2C-CX3

1e"0H 6H(100)

@H HQ) <De~

t tR.....CH*-C-#CH ..... Br slow

IIo ~e~

(97)

296

The powerful electron-withdrawing inductive/field effect exerted byBr makes the a-H atoms of the CHzBr group more acidic thanthose of the RCHz group, and may also help stabilise the resultantcarbanion (98), compared with (99). The former will thus be formedpreferentially, and further bromination will thus be expected onCIIzBr rather than on RCl:Iz. Further, because of this electronwithdrawal by the Br atom, (98) will be formed more rapidly thanwas, for example, (95), i.e. the second bromination will be fasterthan the first; and the third bromination of CH3 will be correspondingly faster still. We might thus expect the end-product of thisbase-catalysed halogenation to be RCHzCOCX3 (100). Reversibleaddition of eOH to the C=O group of the ketone can, however,take place at any time, and in CX3 we now have an excellent leavinggroup; the result is thus C-C bond fission (ct. p. 237):

298 Carbanions and their reactions

The reason for the greater difficulty of further attack, under acidconditions, is that the intermediate (and T.S.) involved in formationof the enol, e.g. (104) from CH3COMe, carries a +ve charge. Thecorresponding +vely charged intermediate involved in formation ofthe enol from BrCH2 COMe will therefore be destabilised (relativeto the one from CH3COMe) by the electron-withdrawinginductive/field effect exerted by its Br atom. As yet unreactedCH3COMe will thus undergo enolisation, and subsequent (rapid)bromination, in preference to BrCH2COMe. When further bromination is made to take place, the major product is found to be thel,l-dibromo compound Br2CHCOMe, but the issue is complicatedby the fact that, under the reaction conditions, this isomerises tosome extent to the 1,3-derivative, BrCH2COCH2Br.

11Radicals and their reactions

11.1 INIRODUCTION, p. 299.11.2 RADICAL FORMATION, p. 303.

11.2.1 Photolysis, p. 303; 11.2.2 Thermolysis, p. 304; 11.2.3Redox reactions, p. 306.

11.3 RADICAL DETECTION, p. 308.11.4 RADICAL SHAPE AND STABILISATION, p. 309.11.5 RADICAL REACTIONS, p. 313

11.5.1 Addition, p. 313: 11.5.1.1 Halogens, p. 313; 11.5.1.2Hydrogen bromide, p. 316; 11.5.1.3 Other additions, p. 319;11.5.1.4 Vinyl polymerisation, p. 320; 11.5.2 Substitution, p. 323;11.5.2.1 Halogenation, p. 323; 11.5.2.2 Autoxidation, p. 328;11.5.2.3 Aromatic substitution, p. 331; 11.5.3 Rearrangement,p.335.

11.6 BIRADICALS, p. 337.

11.1 INTRODUCTION

Most of the reactions that have been considered to-date have involvedthe participation of polar reactants and intermediates, Le. carbocations and carbanions, Or related highly polarised species, involvingthe heterolytic fission, and formation, of covalent bonds:

But homolytic fission can also take place, thus generating speciespossessing an unpaired electron-radicals, e.g. (I) and (2) :

R)C-X +=t R)C' ,X

(1) (2)

Homolytic fission of an R3C-X bond is, in the gas phase, alwaysless energy-demanding than heterolytic fission. This energetic advantage is, however, often reversed in polar solvents, because of theenergy then developed-in heterolytic fission-from solvation of thedeveloping ions.

Reactions involving radicals occur widely in the gas phase: thecombustion of any organic compound is nearly always a radicalreaction, and the oxidative breakdown of alkanes in internal combus-

799

Br-Br

·Br + R-Br

30111.1 Introduction

Another nitrogen radical, of considerable importance, is I, I-diphenyl2-picrylhydrazyl (10) obtained by Pb02 oxidation of the triaryl-

PbMe4 P Pb + 4Me'

MnO e2Ph2NH~ Ph 2N-NPh 2 P Ph 2N' + 'NPh 2

(8) (9)

Once recognised in this way, alkyl radicals were invoked as intermediates in a number of reactions (see below).

Radicals, of varying degrees of stability, involving atoms other thancarbon-heteroradicals-were also recognised. Thus it was discoveredin 1911 that on warming N,N,N',N'-tetraarylhydrazines, e.g. (8), innon-polar solvents resulted in the development of a green colour dueto the radical (9) :

and conveyed along a glass tube in a stream of an inert carrier gas,e.g. nitrogen. It was found that thin lead mirrors deposited at variousdistances along the inner wall of the tube were attacked by the streamof radicals. By measurements of how far along the tube mirrorscontinued to be attacked, coupled with a known rate of flow of carriergas, it was possible to make accurate estimates of the half-life of alkylradicals; for Me this was found to be 8 x 10- 3 sec. The fate of suchalkyl radicals, in the absence of metal mirrors to attack, is very largelydimerisation:

tion, and with rise of temperature. Thus a dilute solution of thedimer in benzene contains =2% of Ph3C· at 20° and =10% at 80°;on removal of the solvent only the dimer was obtained. This was,not unnaturally, assumed to be hexaphenylethane, Ph3C-CPh3 ,

and, as mentioned previously (p. 44), it was only 70 years later thatthe dimer was shown (by proton n.m.r. spectroscopy) not to be this,but to have the structure (7):

Ph3COCPh2

H -(7)

Hexaphenylethane has not, indeed, ever been prepared, and may wellbe not capable ofexisting under normal conditions due to the enormoussteric crowding that would be present. The reasons for the relativelyhigh stability of Ph 3C' are discussed below (p. 311).

Simple alkyl radicals are very much more reactive, and were firststudied systematically only in 1929. The radicals were generated bythe thermal decomposition of organometallic compounds, such asPbMe4 ,

Radicals and their reactions300

R-H + ·Br - R· + H-Br

(3) r !Br,

!hll

tion engines is the largest scale, and most widespread, chemicalreaction of all! Radical reactions also occur in solution, particularlyif carried out in non-polar solvents, and if catalysed by light or thesimultaneous decomposition of substances known to produce radicalsthemselves, e.g. organic peroxides. Radicals, once formed in solution,are generally found to be less selective in their attack on other species,or on alternative positions within the same species, than are carbocations or carbanions.

Another characteristic of many radical reactions is that, onceinitiated, they often proceed with great rapidity owing to the establishment of fast chain reactions of low energy requirement, e.g. in thehalogenation of alkanes (3, cf p.323):

In this case, the radical obtained photochemically, a bromine atomBr', generates another, R', on reaction with the neutral substrate,R- H (3). This radical reacts in turn with a further neutral molecule,Br2 , generating Br' once again: the cycle thus proceeds without theneed for further photochemical generation of Br', i.e. it is self-perpetuating. It is also characteristic of such radical reactions that theycan be inhibited by the introduction of substances that themselvesreact particularly readily with radicals (inhibitors, or radical'scavengers'), e.g. phenols, quinones, diphenylamine, iodine, etc. Theseand similar substances can also be used to bring a radical reaction,already in progress, to a stop (terminators).

The first radicals to be studied were, hardly surprisingly, those thatwere somewhat less reactive, and thus capable of rather longer independent existence. The first such radical to be detected unequivocallywas Ph3C' (4), obtained in 1900 on reacting Ph 3CCI with finelydivided silver (cf p. 43). The radical reacted with halogens to reformthe triphenylmethyl halide (5), or with oxygen from the air to fonn(6), a peroxide (all radicals react readily with O2 from the air):

Ph,C'Ph3C' + x-x - Ph 3C-X + x· -----+ 2Ph3C-X

(4) (5) (5)Ph C'

Ph3C' + O 2 - Ph 3COO· ~ Ph3COOCPh3(4) (6)

The yellow radical (4) was in equilibrium in solution in inert solventswith a colourless dimer, the proportion of radical increasing on dilu-

As its solutions are highly coloured, its reaction with other radicalsto form colourless products can be followed colorimetrically.

Solutions of diphenyl disulphide (/3) are found to become yellowon heating, the colour disappearing again on cooling:

exists in this form, i.e. not as the dimer, both in solution and in thesolid state; it is a dark blue solid (m.p. 97°). The reason for its relativeunreactivity is almost certainly hindrance, by the bulky CMe3 in botha-positions, to the approach of either another molecule of (15), or ofother species, to the radical oxygen atom.

30311.2 RadicaL formation

which can then initiate, for example, the halogenation of alkanes(p. 323), or addition to alkenes (p. 313).

The two major advantages of photolysis over thermolysis (see below)for the generation of radicals are: (a) it is possible to cleave strongbonds that do not break readily-or at all-at reasonable temperatures,

Dr-Dr ~ Dr' + ·Dr

CI-CI ~ C1' + 'CI

Another very useful photolytic homolysis is that of halogen moleculesto yield atoms,

RO-NO ~ RO' + .NO

(19) (20)

this happens because carbonyl compounds have an absorption bandin this region. The photochemical decomposition yields the initialpair of radicals, (/6) and (/7), and the latter then breaks down spontaneously to yield another methyl radical and the stable species <;0.Other species that undergo ready photolysis are alkyl hypochloTltes(/8) and nitrites (/9), both of which can be used to generate alkoxylradicals (20) :

° °II hv IIMe-C-Me -+ Me' + ·C-Me -+ CO + ·Me

(16) (17) (16)

h,RO-CI -+ RO' + ·CI

(18) (20)

11.2 RADICAL FORMAnON

There are a number of ways in which radicals may be generated fromneutral molecules, several of which we have already seen; the mostimportant are (a) photolysis, (b) thermolysis, and (c) redox reactionsby inorganic ions, metals or electrolysis-that involve one-electrontransfers.

11.2.1 Photolysis

The prerequisite of this method is the ability of the molecule concernedto absorb radiation in the ultra-violet or visible range. Thus acetonein the vapour phase is decomposed by light having a wave-length of=320 nm (3200 A= 375 kJ mol- 1

);

(12)

NO,

PbO, -~o-- PhlNN~NOl

NOl

(10)

(10)

Radicals and their reactions

hydrazine (11):

Picryl ~NOolPh,NNH, -- Ph,NNH NO,

chloride

NOl

(111

11PhS-SPh ~ PhS· + ·SPh

(13) (14)

The radicals (/4) formed may be trapped with, for example, (/0) above.Simple alkyl thiyl radicals such as MeS' have been detected as reactionintermediates; they are highly reactive. Relatively stable oxygencontaining radicals are also known. Thus the phenoxy radical (/5),

oMO'Ct$rCMO'

CMeJ

(15)

302

This is sufficiently stable (the reasons for its stability are discussedbelow, p. 312) to be recrystallised from various solvents, and obtainedas violet prisms that may be kept more or less indefinitely. It is relativelyunreactive towards other neutral molecules, but reacts readily withother radicals; it is indeed used as a 'trap', forming stable products,e.g. (/2), with almost any other radical:

e.g. azoalkanes (21),

30511. 2. 2 Thermolysis

but also by disproportionation :

Thus MeN=NMe, despite the driving force supplied by N-N asamong the best of all leaving groups, is stable up to ~2oo°, while (24)has a half-life of only ~ 5 min at 100°.

In the absence of other species with which a radical can react (e.g.abstraction of H from a suitable solvent), their life is terminated largelyby dimerisation,

promoting initial decomposition of the peroxide. Thus (Me3CCOO)2has a half-life of =200 hr at 100°, while (PhCOOh has one of only=0·5 hr at the same temperature. As was mentioned above, simplealkyl azo compounds, e.g. (21), are too stable to undergo thermolysis at reasonable temperatures, but can be made useful sourcesof radicals by the introduction of suitable substituents, e.g. (24):

Me2C-N=N-CMe 2 -4 2[Me 2C-C=N +-+ Me2C=C=N] + N=NI ICN CN

(24)

The use of PbEt4 as an anti-knock agent in petrol depends in parton the ability ofthe ethyl radicals, generated on its thermal decomposition, to combine with radicals produced in the over-rapid combustionof petroleum hydrocarbons; chain reactions which are building upto explosion (knocking) are thus terminated short of this. The completedetails of how PbEt4 operates are not known, but there is some evidencethat minute Pb02 particles derived from it can also act as 'chainstoppers'.

Radical formation through carbon-earbon bond-fission is seen inthe radical-induced 'cracking' at ~6oo° of long-chain alkanes. Theradicals introduced initially into the system probably act by abstractinga hydrogen atom from a CH2 group of the chain; the resultant longchain, non-terminal radical (25) then undergoes fission p- to theradical carbon atom to yield a lower molecular weight alkene (26)plus a further radical (27) to maintain a chain reaction:

Ra-H

RCH-CH 2R' --+ RCH=CH 2 + 'R'

(25) (26) (27)

Termination of the reaction by radical/radical interaction is unlikelyto occur to any significant extent, until the concentration of longchain alkane has dropped to a very low level.


and (b) energy at only one particular level is transferred to a moleculeso that it is a more specific method of effecting homolysis than i~pyrolysis. Thus the cleavage of diacyl peroxides, e.g. (22), occurscleanly on photolysis,

000II II hv II

RCO-OCR --+ 2R-C-O' --+ 2R· + 2C0 2

(22)

whereas in a number of cases thermolysis gives rise to other sidereactions.

A very interesting technique for radical generation isjlash photolysis,which employs a very intense pulse of radiation (visible or u.v.) ofvery short duration. This produces a very high immediate concentration of radicals, which may be detected-and whose fate may befollowed-by spectroscopy through one or more subsequent pulsesof lower intensity radiation of suitable wavelength. This is, of course,primarily a technique for the study of radicals rather than for theiruse in preparative procedures. Radicals may also be generated, insuitable cases, by irradiation of neutral molecules with X-rays orwith 'Y-rays: radiolysis.

11.2.2 Thermolysis

PbR 4 +:t Pb + 4R·

(23)

Much of the early work on alkyl radicals of short life was, as we haveseen (p. 301), carried out in the vapour phase through decompositionof metal alkyls, e.g. (23):

hvR-N=N-R --+ R· + N=N + 'R

(21)

This stems from the weakness, i.e. ease of thermal fission, of the Pb-Rbond, and radicals may be generated in solution in inert solvents aswell as in the vapour phase, through such thermolysis of weak eno~ghbonds, e.g. those with a bond dissociation energy of < ~ 165 kJ(40 kcal) mol- I. Such bonds very often involve elements other thancarb~n,and the ~ajorsources of radicals in solution are the thermolysisof sUitable peroxides (0+0) and azo compounds (C+N). Relativelyvigorous conditions may, however, be necessary if the substrate doesnot contain substituents capable of stabilising the product radical, or

HO· + H-CH2CMe20H -> H20 + 'CH2CMe20H -> HOCMe2CH2CH2CMe20H

(30) (31)

Direct reduction of carbocations is not common but has beenobserved, e.g. with vanadium(IJ) chloride:

307

6MeJC~CMeJ~ + Fe(CN).48

CMeJ

(32)

11. 2. 3 Redox reactions

(36)(35)

+ee R C-08 He R C-OH2R2C=0 c----'athode 2R2<;:-08 -+ 2 I -+ 2 I

R2C-08 R 2C-OH

(37)

0 8

MeJC~OCMeJlYJ + Fe(CN).J8 --.

CMeJ

radical (32) through a one-electron oxidation by Fe(CN)63e,

and also the dimeric oxidation of carbanions, e.g. (33), with iodine(p.294):

2(MeCOhCH 8 ~ 2(MeCO)2CH· --+ (MeCOhCH-CH(COMeh

(33)

R-H + Ra· -> R· + H-Ra

CX2-CX2+Ra' -> ·Cx2-CX2-Ra

2RC028 ~ 2RC0

2' -co,. 2R. -+ R-R

anode(34)

Conversely, electrolysis of ketones, (35), results in their cathodicreduction to radical anions (36), which dimerise to the dianions ofpinacols (37):

We have seen similar radical anions generated from ketones in pinacolreduction with sodium or magnesium (p. 218), and also from esterswith sodium in the acyloin condensation (p.218).

It should, however, be emphasised that the methods of radicalformation we have been discussing all involve the generation ofradicals ab initio from neutral molecules, or from ions. In fact,radicals in which we may be interested are often produced viaattack on suitable species by pre-formed radicals, Ra·, generatedspecially for this purpose-with malice aforethought, as it werefrom precursors such as peroxides or azoalkanes:

Radicals, (34), that subsequently dimerise, are also obtained throughthe anodic oxidation of carboxylate anions, RC0 2e, in the Kolbeelectrolytic synthesis of hydrocarbons:


The mixture is known as Fenton's reagent, and the effective oxidisingagent in the system is the hydroxyl radical, HO·. This is particularlygood as an abstractor of H, and can be used either to generate ther~sultant radical, e.g. (30), for further study, or, in some cases preparatIvely through the latter's dimerisation, e.g. (31):

( ~ ) ~ArCO 2 + Cu'" -+ ArC-O' + ArC02

8 + Cu 2'"

(28) (29)

This constitutes a useful method for generating ArC0 2 ·, as in thethermolysis of (28) there is a danger of the further decomposition of(29) to Ar· + CO 2 , Cu$ is also involved in the conversion ofdiazoniumsalts, ArN 2 $Cle , to ArCI + N2 (Sandmeyer reaction), where Ar· isvery probably formed transiently as an intermediate:

11.2.3 Redox reactions

These reactions all involve one-electron transfers in generating theradical, and it is therefore no surprise to find metal ions such asFe 2$/Fe3 $ and Cu$/Cu 2$ involved. Thus Cu$ ions are found toaccelerate greatly the decomposition of acyl peroxides, e.g. (28):

° °" "PhC-H + FeJ'" -+ PhC· + H'" + Fe2'"

We have already seen (p. 302) the generation of a stable phenoxy

Both of these reactions ani reductions, another is the use of Fe2$ tocatalyse the oxidation reactions of aqueous hydrogen peroxidesolution:

Generation of a radical through an oxidative process probablyoccurs in the initiation of the autoxidation of benzaldehyde (p. 319),which is catalysed by a number of heavy metal ions capable of oneelectron transfers, e.g. Fe3 $ :

11.3 RADICAL DETECI10N

We have already seen how the high chemical reactivity of short-livedradicals can be enlisted to aid in their detection through their abilityto etch metal mirrors (p. 301). The fact that the transition of an unpairedelectron between the energy levels of a radical involves less energythan the transition ofthe paired electrons in the stable parent moleculemeans that the radical tends to absorb at longer wavelength. A numberof radicals are thus coloured-where their precursors are not-andmay readily be detected in this way, e.g. (II, p.302) and (15, p.302).Radicals may also be detected by their rapid discharge of the colourof solutions containing species such as I, l-diphenyl-2-picrylhydrazyl(II).

Another useful, and quite sensitive, test is the initiation of polymerisation (cf p. 320). Polymerisation can be initiated, in suitablesubstrates, by cations and anions as weIl as by radicals, but the effectof these several species can be differentiated by using a 50/50 mixtureof phenylethene (styrene), PhCH=CH2 , and methyl 2-methylpropenoate (methyl methacrylate), CH2 =C(Me)C02Me, as substrate:cationic initiators are found to produce polystyrene only, anionspolymethyl methacrylate only, while radicals produce a copolymercontaining equal amounts of the two monomers.

By far the most useful method for detecting radicals is, however,electron spin resonance (e.s.r.) spectroscopy, which utilises thepermanent magnetic moment conferred on a radical by virtue of thespin of its unpaired electron (radicals are paramagnetic, species containing only electron pairs are diamagnetic). The electron spin canhave one of two values (+tor - t, cf p. 2) and, in the presence of anapplied magnetic field, these correspond to different energy levels;transitions are possible between them resulting in a characteristic, anddetectable, absorption spectrum. E.s.r. spectroscopy of unpairedelectrons is thus the analogue of n.m.r. spectroscopy of nuclei thathave a permanent magnetic moment, e.g. IH, l3C, etc.; hardly surprisingly, they occur in different energy ranges (an unpaired electron hasa much larger magnetic moment than a proton-1H-and moreenergy is required to reverse its spin).

In e.s.r. spectroscopy, interaction ('splitting') occurs between theunpaired electron and neighbouring magnetic nuclei-especiaIly I Hleading to quite complex patterns oflines; analysis of these can providea great deal of detailed information about the structure and shape ofa radical. Thus hydrogen abstraction from cycloheptatriene (38) by,OH is found to lead to a radical having a very simple e.s.r. spectrum:eight equally spaced lines, indicating interaction of the unpairedelectron with seven equivalent I H nuclei. The product radical thuscannot have the expected structure (39)-which would have a verymuch more complex e.s.r. spectrum-but must be the delocalised

11.4 RADICAL SHAPE AND STABR.ISATION

A~ with carbocations (p. 104) and carbanions (p. 276), the questionarISes of whether simple radicals-of the type R3C·-accommodate

Radicals have been detected by e.s.r. spectroscopy, under the bestconditions, in concentrations as low as 10-8 M. Radicals to bestudied may sometimes be generated (by irradiation) actuaIly in thecavity of the spectrometer; failing that, they may be generated justoutside, and a continuous flow technique then used to maintain a'stan~ing' concentration in the spectrometer cavity. A disadvantageof thIS method is that it requires relatively large volumes, andquantities of starting material. The longer the life of the radical thegreater the chance of observing its spectrum; thus species such asPh3C· are easily observed, but species like Ph" PhCHz', CzHs', etc.,are a little more difficult. A technique that has been used to'prolong' the life of short-lived species is to introduce a suitablediamagnetic substance, e.g. (41), which will react with the transientradical, and convert it into a longer-lived radical (42) that can bedetected quite readily:

309

H~HH9HH H

(40)

11.4 Radical shape and stabilisation

(39)

H

~6(38)

H H

ospecies (40, cf p. 106):

RaI

Ra' + Me)C-N=O --+ Me)C-N-O'

(41) (42)

This is known as 'spin trapping'. Another technique, that has beenus~d to study very short-lived radicals, is to generate them photolytIcaIly, from precursors, in a solid inert matrix, e.g. frozen argon.Their life is thus artificially prolonged because they are shielded fromcoIlision either with each other, or with other species that couldterminate their existence.

Quite apart from such specific physical methods for the detectionof radicals, it should be emphasised that more general indicationsthat radical intermediates are involved in a particular reaction areprovided by its high susceptibility to the addition of radical initiators(cf. p. 314) or inhibitors (cf. p. 300), and (compared with polarreactions) its relative insusceptibility to change of solvent.


310 Radicals and their reactions 11.4 Radical shape and stabilisation 311

CH)' < CH 2F· < CHF2' < CF)·

however, being essentially Sp3 in CF3" i.e. the latter radical is thuspyramidal (44, R = F); the radicals ·CH20H and ·CMe20H are alsosubstantially 'bent'. Comparison ofthe ease offormation, and reactivityonce formed, of bridged radicals-such as (45) and (46}-

(49b)(49a)

~~~c~

6··········...........

[RCH:..:..:CH'-'-'CH z]·

(47)(48)

Both are essentially planar, i.e. Sp2 hybridised, at the radical carbonatom for only in this configuration is maximum pin orbital overlapwith consequent stabilisation-possible. The stability of a radicalincreases as the extent of potential delocalisation increases; thusPh2CH' is more stable than PhCH2', and Ph3C' (cf p. 300) is apretty stable radical.

The shape of Ph 3C' (49) is a matter of some interest as it has abearing on the extent to which delocalisation of the unpaired electron,with consequent stabilisation, can occur. The radical carbon atom iscertainly Sp2 hybridised in (49), i.e. the bonds joining it to the threebenzene nuclei all lie in the same plane; but maximum stabilisationwill only occur if all three benzene nuclei can be simultaneouslycoplanar (49a),

Radicals ofallylic, RCH=CHCH2' (47), and benzylic, PhCHR (48),type are more stable, and less reactive, than simple alkyl radicals,because of delocalisation of the unpaired electron over the n orbitalsystem in each case:

for only in this conformation can the p orbital on the central carbonatom interact equally, and maximally, with the n orbital systems ofthe three' nuclei. In fact triarylmethyl radicals have been shown, byspectroscopic and X-ray crystallographic measurements, to bepropeller-shaped (49b), the benzene rings being angled at about 30°out of the common plane. Thus though delocalisation occurs in (49}as indicated by its e.s.r. spectrum-it is not maximal, and its extentis not enormously greater in Ph 3C' than in Ph 2CH " or even inPhCH 2 •• The major reason for the greater 'stability' of Ph3C·, asreflected in its greater reluctance to dimerise, must therefore belargely steric: the crowding involved when two, enormously bulky,Ph3C· radicals seek to combine with each other. A crowding that isechoed in the fact that the dimer, when formed, is found to be notthe expected hexaphenylethane, Ph3C-CPh3 (cf. p. 301), but (7)resulting from the preferential reaction of one bulky Ph3C· radicalon the much more readily accessible periphery (through electron

(46)(45)

R",0

C-R

R/O

their unpaired electron in a p orbital (planar shape, 43) or an Sp3

hybrid orbital (pyramidal shape, 44),R

R,' R\/CU

oC (~

/J'"R' R

R

(43) (44a) (44b)

or. whether the shape is somewhere between the two. Direct physicaleVIdence for CH 3' comes from the e.S.r. spectrum of 13CH3·. Analysisof the lines, resulting from interaction between the unpaired electronand the paramagnetic 13C nucleus, provides information about thedegree of s character of the orbital in which the unpaired electron isaccommodated. That in 13CH3· is found to have little or none andthe radical is thus essentially (within ~ 5 %) planar, i.e. (43, R ~ H);a conclusion that is supported by evidence derived from U.V. and i.r.spectra. The s character of the half-filled orbital is found to increaseacross the series,

with their acylic equivalents would suggest that alkyl radicals doexhibit some preference for the planar state. This is nothing like somarked as with carbocations, however, and, unlike the latter (p. 86),there is little difficulty in generating radicals at bridgehead positions.

The relative stability of simple alkyl radicals is found to follow thesequence:

R)C' > R2CH· > RCH 2 ' > CH)·

This reflects the relative ease with which the C-H bond in the alkaneprecursor will undergo homolytic fission, and more particularly,decreasing stabilisation, by hyperconjugation or other means as theser.ies is traversed. There will also be decreasing relief of strai~ (whenR IS large) on going from Sp3 hybridised precursor to essentially Sp2

hybridised radical, as the series is traversed. The relative difference instability is, however, very much less than with the correspondingcarbocations.

delocalisation) of the other:

(p-RC6H4hC' + R-F\- C(C6H 4R-p)2 -- Dimer(50a) ~ (50b)

The hetero radicals that have already been referred to-(9, p. 301),(10, p. 302), (14, p. 302) and (15, p. 302)-owe their relative stability[with respect to their dimers-apart from 1, I-diphenyl-2-picrylhydrazyl(10)] to a variety of factors: (a) the relative weakness of N-N, S-Sand 0-0 bonds, (b) the delocalisation through the agency of aromaticnuclei, and (c) steric inhibition of access to the atom with the unpairedelectron, or to an aryl p-position, cf (50). The latter factor bulks large(in addition to the weakness of 0-0 bonds) in the great stability of(15, cf p. 302); and all three factors operate to stabilise (51), which iswholly dissociated in solution:

31311. 5 Radical reactions

11.5.1 Addition

This radical has been shown, from calculations based on e.s.r. spectra,to have the p-phenyl group coplanar with the central phenoxy nucleus,but the two o-phenyl groups angled at 460 to it. The p-group can thuseffect maximum delocalisation-(b)-and also act as a bulky groupto inhibit dimerisation-(c), cf (50) above, while the two angledo-substituents inhibit access to the 0 atom, preventing formation ofan 0-0 dimer [dimerisation does occur in the solid state, but it isthen through one p-position, cf (7, p. 301)].

11.5 RADICAL REACTIONS

Additions to C=C are almost certainly the most important group ofreactions involving radicals. This is due largely to the importance ofaddition (vinyl) polymerisation (p. 320), and the consequent extentto which its mechanism has been investigated; but addition of halogensand of halogen hydracids is also of significance.

It is possible, and logical, to classify the multifarious reactions ofradicals from the point of view of the radical itself: (a) unimolecularreactions, e.g. fragmentation, rearrangement; (b) bimolecular reactionsbetween radicals, e.g. dimerisation, disproportionation; and (c)bimolecular reactions between radicals and molecules, e.g. addition,displacement, atom (often H) abstraction. Such a grouping has, forour purpose, the disadvantage of fitting much less well into the generalclassification of reaction types that has been adopted throughout.We shall therefore discuss the reactions in which radicals are involved,either as reactants or intermediates, under the general heads ofaddition,displacement and rearrangement.

It is important to emphasise that in any reaction of a radical witha neutral molecule a further radical will be formed (cf. p. 309), thusestablishing a chain reaction that does not require further input ofinitiator radicals to sustain it. Such a chain reaction is normallyterminated by the relatively rare reaction of two radicals with eachother (radicals are present in only very low ambient concentration)resulting in dimerisation or disproportionation (cf. p. 305), with nonew radical now being produced.

11.5.1.1 Halogens

In addition to the polar mechanism already considered (p. 179),halogen addition to alkenes can proceed via radical intermediates.The former is favoured by polar solvents and by the presence of Lewisacid catalysts, the latter by non-polar solvents (or in the gas phase),


PhC>Q=( ) 3 CPh

2

H -

(7)

The benzene rings are forced out of the coplanar conformationby steric interaction ofthe o-H atoms of adjacent rings with each other;as would be expected, o-substituents bulkier than H are found toincrease the out of plane dihedral angle of the aromatic nuclei-to 500

or more. Delocalisation must then be even further decreased, butsuch radicals with bulky o-substituents are nevertheless found to bemore stable, i.e. more reluctant to form their dimers than is Ph3C·itself. This must, of course, be due to a steric effect-the o-substituentsare very close to the radical carbon atom and are thus capable ofpreventing its access to other species, or other species access to it[cf (15), p. 302]. It is significant, in the light of what has been said above,that their effectiveness at 'masking' the radical carbon atom willincrease the more the benzene rings are angled out of the coplanarconformation, i.e. the greater the dihedral angle.

If each aromatic nucleus in the radical has a bulky p-substituent,e.g. (50), then, irrespective of any substitution at the o-positions,dimerisation will be greatly inhibited, or even prevented [cf (7), p. 301]:

Rate oc Jlntensity of absorbed light

The reaction is found to be inhibited by the presence of oxygen;

Cl Cl Cl CII I I I

Cl)CC1

· + ·CCCl) --+ CI)CC-CCCI)I I I

Cl Cl CI Cl

(53) (53) (55)

31511.5.1.1 Halogens

CI·0,· I

vv--+ CI)C-C=O

(56)

ClI

CI c-c·) ICl

(53)

this is because the molecule of oxygen has two unpaired electrons, andbehaves as a diradical (cf p.337), ·0-0·, albeit a not very reactiveone. It can, however, combine with the highly reactive radical intermediates in the above addition, converting them into the very muchless reactive peroxy radicals, RaO-O·, which are unable to carryonthe chain: it is thus a highly efficient inhibitor. That oxygen is reactinglargely with the pentachloroethyl radicals (53) is shown by the formation of (56),

when the normal addition reaction is inhibited by oxygen.The reactivity sequence for homolytic addition of the different

halogens to alkenes is, hardly surprisingly, the same as that for electrophilic addition, i.e. F 2 > Cl2 > Br2 > 12 • The addition of fluorinenot requiring photochemical or other activation-is too vigorous tobe of much use, and side reactions are common. Chlorination isgenerally rapid, with long reaction chains, and not readily reversibleexcept at temperatures > 200°; as the temperature rises, however,there is an increasing tendency to hydrogen abstraction leading tooverall substitution by chlorine-rather than addition-in suitablecases (cf p. 325). Bromination occurs readily, but with somewhatshorter reaction chains, and is usually reversible, while iodinationtakes place with difficulty, if at all, and is very readily reversible. Theeffect of increasing alkyl substitution at the double bond carbonatoms is found to have relatively little effect on the rate of halogenaddition, certainly a good deal less than for addition by the polarmechanism (p. 183). Halogen substitution, e.g. by CI, on the doublebond carbon atoms results in a decreased reaction rate, e.g. CI2C=CCI2adds chlorine much more slowly than CH2=CH2 •

The reversibility of addition of Br2 and 12-particularly the latterhas been made use of in the isomerisation (of the less to the morestable) of a pair of doubly bonded geometrical isomerides: in simplecases the cis to the trans e.g. (57)~ (58). This may be carried out byu.v. irridation in the presence of catalytic quantities of Br2 or 12 :

H~., ~ 0., P 0H¢HyLHDr H Dr H Dr Dr

(57) (58)

Normally, of course, an equilibrium mixture will be produced withthe more stable form preponderating. That the interconversion does


~y.s~nlight or u.v. irradiation, and by the addition of radical precursors(Imt!~tors) as cat~lysts. An example is the photochemically catalysedaddlt.lOn of chlonne to tetrachloroethene (52), which involves a chainreactIon (cf p. 300) :

CI-C1

~h'

CI 2C=CCI 2 + ·Cl --+ Cll:-CCI)

(52) j (53)

I 1C)-C!

·CI + C1)C-CCI)

(54)

Each mol~cule of ch!orine, .on photochemical fission, will give rise totwo chlonne atoms, I.e. radicals, each of which is capable of initiatinga con.tinuing reaction chain. That each quantum of energy absorbeddoes Indeed lead to the initiation of two reaction chains is confirmedby the observation that: -

Chlor~ne ato~ns are electrophilic (the element is electronegative, andCI· ~Ill readIly take up an electron to complete its octet) and thus addreadIly to the double bond of (52) to yield the radical (53). This, inturn, can abstract a chlorine atom from a second molecule (the processcan equally.well be regarded as a radical displacement reaction onCl-CI) .to Yield th~ end-product of addition (54), plus a further atomof c.hlonne ~o c?ntInue the reaction chain, i.e. a very fast, continuingcham react.lOn IS set up by each chlorine atom initiator generatedphotochemlcal.ly. Each quantum of energy absorbed is found to leadto th~ conve.rslOn o~ seve.ral thous~nd molecules of (52) into (54); thereactIOn ch~ms are, m this case, said to be long. Until the later stagesof the reac~lOn, when nearly all of (52) and Cl2 have been used up, theconcentratIons o~ (53) and. of CI· ~i.ll be very small compared withth.ose of the startIng matenals; collIsIOn of a radical with a moleculewIll thus be. very much more frequent than collision of a radical withan~ther ra~lcal. Ch~i.n termination will ultimately take place through~adlcaljradlcal collIsIOn, however, and this is generally found toInvolve (53) + (53)~ (55):

11.5.1.2 Hydrogen bromide

The addition of HBr to propene, MeCH=CHz (63), under polarconditions to yield 2-bromopropane, has already been referred to

proceed, as above, via addition and elimination of Br· has been shownby using radioactive Brz as catalyst: both (58) and (57), in the equilibrium mixture obtained, are then found to contain radioactive Br.

The addition of chlorine or bromine to benzene-one of the fewoverall addition reactions of a simple benzene nucleus-has also beenshown to proceed via a radical pathway, i.e. it is catalysed by lightand by the addition of peroxides, and is slowed or prevented by theusual inhibitors. With chlorine this presumably proceeds;

317

(2) XCH 2CH 2' + HX

+155 (+37)+21 (+5)-46 (-11)-113 (-27)

11. 5.1. 2 Hydrogen bromide

(1) x, + CH 2=CH 2-188 (-45)-109 (-26)-21 (-5)+29 (+ 7)

MeCH=C~2 +(63)

H-FH-C1H-BrH-I

Br' - MetH-CH2Br (65)i ~H-Br

Br' + MeCH 2-CH 2Br

(64)

The initiation is by Br·, as hydrogen abstraction by RO· from HBr(as above) is energetically much more favourable than the alternativeof bromine abstraction to form ROBr + H·. The alternative additionofBr· to (63) to form MeCH(Br)CH z· (66) does not occur, as secondaryradicals, e.g. (65), are more stable (cf p. 310) than primary, e.g. (66).

HBr is the only one ofthe four hydrogen halides that will add readilyto alkenes via a radical pathway. The reason for this is reflected inthe ~H values-in kJ (kcal) mol- I-below for the two steps of thechain reaction for addition of HX to CHz=CHz, for example:

Only for HBr are both chain steps exothermic; for HF the secondstep is highly endothermic, reflecting the strength of the H-F bondand the difficulty of breaking it; for HCl it is again the second stepthat is endothermic, though not to so great an extent; while for HIit is the first step that is endothermic, reflecting the fact that the energygained in forming the weak I-C bond is not as great as that lost inbreaking the C=C double bond. Thus a few radical additions of HClare known, but the reactions are not very rapid, and the reaction chainsare short at ordinary temperatures.

Even with HBr addition the reaction chains tend to be rather shortmuch shorter than those in halogen addition-and more than a traceof peroxide is thus needed to provide sufficient initiator radicals: forpreparative purposes up to 0·01 mol peroxide per mol of alkene. Inpractical terms, however, there may already be sufficient peroxide

(p.t84). In the presence of peroxides (or under other conditions thatpromote radical formation), however, the addition p~o~eeds via arapid chain reaction to yield l-bromopropane (64); thIS IS generallyreferred to as the peroxide effect leading to anti-Markownikov addition.This difference in orientation of HBr addition is due to the fact thatin the first (polar) case it is initiated by H$ and proceeds via the morestable (secondary) carbocation, while in the second (radical) case itis initiated by Br· and proceeds via the more stable (secondary)radical (65):

RO' + H-Br - RO-H + Br'

~

(59)


@ ~CI

OCI· : ..... " cr,--+:. --'+

...........

316

~IH~CI

'H

+ CI·

the product is a mixture of several of the eight possible geometricalisomerides of hexachlorocyclohexane (59). In the absence of light orperoxides no reaction takes place, while in the presence of Lewis acidsoverall electrophilic substitution takes place by an addition/eliminationpathway (p. 138). With radicals other than CI·, e.g. Ph·, overallhomolytic substitution can be made to take place on benzene by anaddition/elimination pathway too (p. 331).

Radical attack on methylbenzene (toluene, 60) results in preferentialhydrogen abstraction by CI· leading to overall substitution in theCH 3 group, rather than addition to the nucleus. This reflects the greaterstability of the first formed (delocalised) benzyl radical, PhCHz· (61),rather than the hexadienyl radical (62), in which the aromatic stabilisation of the starting material has been lost:

Me Me Me Br

M'~B' Br· 0 HDr 0---. -- + Br·Br

H Me H Me H

(67) (69) (68)

Essentially exclusive TRANS addition of HBr also occurred with thetrans isomeride of (67), i.e. (70), under analogous conditions.

To account for this very high TRANS stereoselectivity, it has beensuggested that addition. proceeds via a cyclic bromonium radical (71),analogous to the cyclic bromonium cations involved in the polaraddition of bromine to alkenes (p. 180):

Dr Me Br/-~ ~ ~r X

nA MeI Br A It + Br'Me H Br Me H

(71) (68)

The overall addition would then be completed through attack byHBr from the less hindered side (away from bridging Br) to yield the

31911.5. 1.3 Other additions

Me Me

J?(M'~B' Br· K0 HDr + Br·-- --Br

Me H Me HH

(67) (69) (68)

Bri~ Br Br Me

M'~M' Dr· ~HDr 0 + Br·-- --Me

Me H Me HH

(70) (73) (72)

11.5.1.3 Other additionsThiyl radicals, RS', may be obtained by hydrogen abstraction fromRSH, and will then add readily to alkenes by a chain reaction analogous

There are reasons for believing that this common product mixturefrom each of the two alkenes---(67) and (70}--does not arise fromequilibration of these startiQg materials before addition proper takesplace. It could well be that the higher degree of TRANSstereoselectivity observed (p. 318) for addition at lower temperature, and with higher concentration of HBr, results not from theintervention of a cyclic bromonium radical (71), but from slowerrotation about the central carbon--earbon bond. Relatively rapid Htransfer (by the higher concentration of HBr) could then take placeto the less hindered side of (69) or (73), leading to preferentialTRANS additional overall.

In cyclic alkenes, where such equilibration of radical intermediatescannot occur, there is a preference, but not an exclusive one (exceptfor cyclohexenes), for overall TRANS addition.

product of overall TRANS addition (68). However, addition of HBrto (67) at room temperature, and using a lower concentration of HBr,is found to result in a 78 %: 22 %mixture of the products arising fromTRANS and CIS stereoselective addition, (68) and (72), respectively.Significantly, the same mixture of products, in the same proportion,is obtained under these conditions from the trans isomeride of (67),namely (70). This strongly suggests that, under these conditions,rotation about the central carbon--earbon bond is sufficiently rapidfor conformational equilibrium to be established between the radicalintermediates, (69) and (73), before they can abstract H from HBr tocomplete overall addition:


present in the alkene-through its autoxidation (p. 329) by the oxygenin the air-to auto-initiate the radical pathway of HBr addition,~hether this is wanted or not. Once initiated, reaction by this pathwayIS very much faster than any competing addition via the polar pathway,and the anti-Markownikov product, e.g. (64), will thus predominate.If th~ M~r~ownikov product, e.g. MeCH(Br)CH3 from propene, isrequIred It IS necessary either to purify the alkene rigorously beforeus~' or to add inhibitors (good radical acceptors such as phenols,qUInones, etc.) to mop up any radicals, or potential radicals, present;the latter is much the easier to do preparatively. Essentially completecontrol of orientation of HBr addition, in either direction, can thusbe achieved, under preparative conditions, by incorporating eitherperoxides (radical initiators) or radical inhibitors in the reactionmixture. This is particularly useful as such control is not confinedpurely to alkenes themselves: CH2=CHCH2Br, for example, can beconverted into 1,2- or 1,3-dibromopropane at will.

In any consideration of stereoselectivity in radical addition to acylicsubstrates, interpretation ofthe results is complicated by the knowledgethat alkenes may be converted, at least in part, into their geometricaliso?1erides by traces of bromine (or of HBr, i.e. by Br', cf. p. 315).ThIs may, however, be minimised by working at low temperatures,and by using a high concentration of HBr. Thus addition of liquidH.Br ~t - 800 to cis 2-bromobut-2-ene (67) wa~ found to proceedwIth hIgh TRANS stereoselectivity, and to yield (68) almost exclusively:

t<:> that for ~Br. The addition is of preparative value for makingdlalkyl sulphIdes, but is reversible:

Ra'

RCH=CH 2 + R'SH +=t RCH 2CH 2SR'

S~lphen~1 chlorides, e.g. CI 3CSCI, can also be used as sources ofth.lyl radIcals, but here the addition is initiated by CI· and the R'SwIIl thus become attached to the other carbon atom of the doublebond;

11.5.1.4 Vinyl polymerisation

This re~ct.ion has been the subject of a great deal of theoretical andmechamsttc study, .Iar~ely bec.ause. of the commercial importance ofthe polym.ers to whIch It can gIve nse. Like the other radical reactionswe have dlsc.ussed, it can be said to involve three stages--(a) initiation(b) propagatton, and (c) termination: '

32111.5.1.4 Vinyl polymerisation

As the alkene monomers can absorb oxygen from the air, formingperoxides (cf p. 329) whose ready decomposition can effect autoinitiation of polymerisation, it is usual to add a smaIl quantity ofinhibitor, e.g. quinone, to stabilise the monomer during storage. Whensubsequent polymerisation is carried out, sufficient radical initiatormust therefore be added to 'saturate' the inhibitor before any polymerisation can be initiated; an induction period is thus often observed.

The radical initiators are not, strictly speaking, catalysts-thoughoften referred to as such-for each radical that initiates a polymerchain becomes irreversibly attached to it and, ifof suitable composition,may be detected in the molecules of product. The efficiency of someinitiators may be so great that, after any induction period, every radicalgenerated leads to a polymer chain.

Termination of a growing chain can result from coIlision with eitheran initiator radical (c i) or with another growing chain (c ii), but of thesethe latter is much the more frequent, as the initiator radicals will havebeen largely used up in starting the chains. Termination has beenshown above as dimerisation (c ii), but it can also involve disproportionation (cf. p. 305) between growing chains (c iii). H-abstractioncan also occur by attack of a growing chain on 'dead' (no longergrowing) polymer, leading to a new growing point and, hence, tobranching (76):

Ra(CH 2h.· H Ra(CH 2h._ICH J (CH2 h.·I -. , (CH,=CH,I. I

Ra(CH 2).CH(CH 2),Ra Ra(CH 2).CH(CH 2),Ra • Ra(CH 2).CH(CH2),Ra

(76)

The extent to which branching occurs can, hardly surprisingly, havea profound effect on the physical and mechanical properties of theresultant polymer.

Another major influence on the properties of the polymer is theaverage molecular weight, i.e. the average length of polymer molecules;this may vary from only a few monomer units to many thousand.Apart from the average length of polymer molecules, the actual spreadof lengths among the polymer molecules also has a considerableinfluence; th us the properties of two polymers of approximately thesame average m.w. will differ greatly if one is made up of molecules allof much the same length, while the other includes both very long andvery short polymer molecules in its make-up. The length of moleculesin a polymer may be controIled in a number of ways. Thus increasein the concentration of initiator, relative to that of alkene, wiIl leadto shorter chain lengths: the number of growing chains is increased.and termination thus becomes more probable relative to continuedpropagation. Alternatively, actual terminators may be added or, moreusually, chain transfer agents. These are compounds, usualIy of theform XH, that suffer H-abstraction by a growing polymer chain,


RCH CH CI·, CI CSCI= 2 -. RCH-CH 2C1~ RCHCH CI + CI,

I 2

SCCIJ

C~rbon--<:arbon bonds may be formed by the addition, among otherthmgs, of halomethyl r~dicals to al~enes, The ,CX

3(X = Br, CI) may

be generated by the actIon of peroxIdes on, or by photolysis of, CX4

:

RCH CH 'CCI" CCI= 1 ---. RCHCH 2CCl J ~ R?HCH 2CCl J + 'CCl

J

CI(74) (75)

That .t~e relatively inert CCI4 adds in this wa.y may seem a littlesurpnsmg, b~t the L\H values for both steps of the reaction chainare exothe~mlc: -75(-18), and -17(-4)kJ(kcal)mol- 1. The firstformed radIcal (74) may, however, co~pete with 'CCI

3in adding to

RCH-C~2.'so that low molecular weIght polymers are formed undersome condlttons, as well as the normal addition product (75).

(a) Initiation:

~i,> Formation of initiator, Ra·, from, e.g. peroxides or azo compounds(II) Ra·+CH2=CH2 --+RaCH

2CH

2, •

(b) Propagation:

RaCH2CH2 , -J..-llCH,==CH;, • Ra(CH2)2.'

(el Termination:

(i) Ra(CH1h. + ·Ra --+ Ra(CH2h.Ra(ii) Ra(CH2h. + ·(CH2h.Ra --+ Ra(CH2)4.Ra

(iii) Ra(CH2).CH1,+ 'CH2CH2(CH2),Ra --+ Ra(CH2).CH3 + CH2

= CH(CH2),Ra

The propagation step is usuaIly very rapid.

thereby terminating the chain but generating a new radical, X', inthe process, that is capable of initiating a new chain (77) from monomer.Thiols, RSH, are often used:

"CHz=CHzRa(CHz).CH z· + RSH --+ Ra(CH z).CH 3 + RS·"'" ~ RS(CHz)z.·

(77)

A new growing chain is thus generated without slowing down theoverall process of monomer conversion. In the case of terminators,XH is chosen so that X' is not reactive enough to initiate a new chainfrom monomer.

Radical-induced polymerisation of simple alkenes, e.g. ethene andpropene, requires vigorous conditions including very high pressure,but many other alkene monomers carrying substituents polymerisereadily. These include CH 2 =CHCI-+ polyvinyl chloride (r.v.c.) formaking pipes, etc., CHz=CMeCOzMe --+ perspex, PhCH=CHz --+

polystyrene, the expanded form for insulation, etc., andCFz==CFz --+ tefton, which has an extremely low coefficient of friction, high chemical inertness and high m.p. (lining of frying pans,etc.). The properties of a polymer may be varied even furtheralmost as required-by the copolymerisation of two differentmonomers so that both are incorporated, equally or in other proportions, in the polymer molecules; thus most of the synthetic rubbersare styrene/butadiene copolymers. Reference has already beenmade (p. 308) to the analytical use of 50: 50 copolymerisation ofPhCH=CHz and CHz=CMeCOzMe to distinguish radical-inducedpolymerisation from that initiated by anions or cations (ct. p. 188).

Radical-induced polymerisation "has some drawbacks, however;thus branching induced by H-abstraction from the growing chainhas already been referred to (p. 321). Another difficulty arises withmonomers of the form CH 2 =CHX (i.e. with all the common monomersexcept CH 2 =CH2 and CF2=CF 2) over the orientation of the substituent groups, X, with respect to the 'backbone' alkane chain ofpolymer molecules, whose conformation is 'frozen' in the final rigidsolid. In radical polymerisation, the arrangement of the X groups israndom, and such atactic polymers, e.g. atactic polypropene, arefound to be non-crystalline, low density, low melting, and mechanicallyweak. It has been found, however, that use of a TiCI 3 ' AIEt 3 catalystresults not only in polymerisation occurring under very mild conditions,but with, for example, propene, the resultant polymer has all the Megroups oriented, regularly, in the same direction. This isotactic polypropene is found to be crystalline, high density (closer packing ofchains), high melting, and mechanically strong-all desirable qualities-and branching has been largely avoided. This regular, coordinationpolymerisation is believed to result from groups of atoms in the surfaceof the heterogeneous catalyst acting as a template, so that eachsuccessive monomer molecule can be added to the growing polymer

323

trans

11.5.2.1 Halogenation

cis

We might well expect this differing stereochemistJ}'. to have amarked effect on the properties of the polymer, and thIS IS borne outby the two naturally occurring polyisoprenes, n.atu~al ~ubber andgutta percha. The former, which be~ore vu~camsatIOn IS sof~ an~tacky, has all cis junctions in ~ts c~ams; whIle the latter, WhICh IShard and brittle, has all trans JunctIons.

11.5.2.1 HalogenationAlkanes are attacked extremely readily by hal.o~en~, provided theconditions allow the formation of radicals. ThIS IS 10 ~arked contrast to their extreme resistance to attack by ~lectrophIles or nucleophiles, which stems from the very low polanty of the C-H bond

11.5.2 Substitution

Although most of the reactions to be c?ns~dered ~n~er this head arenet, i.e. overall, displacements or subStItutIO,:!S, t~IS IS n?t commonlyachieved directly, cf SN2. In some cases a radIca! IS o~tamed from thesubstrate by abstraction (usually of H), and thIS .radIcal then effectsdisplacement on, or addition t?, a ~urther specI~s: In some <;ases,however, the net displacement IS achIeved by addItIOn/abstractIon.

chain only through 'coordination', in one particular orientation, atthe catalyst surface. .

When the monomers are conjugated dienes, e.g. buta-1,3-dlene,CH =CH-CH=CHz, or 2-methylbuta-1,3-diene (isoprene),CH~=C(Me)-CH-CHz, the polym~r c~ain o~)1ain~d from ?ormal(1,4-, ct. p. 195) addition polymensatIon w~l still conta~n one

arbon-carbon double bond per monomer umt. The resultIng re~idual reactivity allows of chemical cross-linking from one polymerchain to another, e.g. the formation of S-S 'bridges' be~we~n thepolymer chains by reaction with sulphur in the vulcamsat~on ofrubber. A relatively low degree of cross-linking is fou?d t.o Impartelastic properties to the polymer aggregate, while ca.rry~g I~ furtheryields a rigid structure through. extensive. cro~s-hnkIng In th~eedimensions. A stereochemical POInt also anses, In that the .relatIveorientation of the parts of the polymer molecule on each SIde of adouble bond in the chain can be either cis or trans to each other,e.g. with polyisoprene:


H H H

" " " "H-C-H < H-C-H < -C-H < -C-H/ / / /

H

i.e. i.n. the order of weakening of the C-H bond, and of increasingstabilIty of the product radical (cf p. 310); the figures quoted are forthe relative rates of abstraction of H by Cl- at 25°. This differentialmay often be opposed by a statistical effect, Le. relative numbers ofthe different types of hydrogen atom available· thus in (CH ) CHth .. '33e~e are mne pnmary hydrogen atoms available to every one

tertiary hydrogen atom. On chlorination (CH3hCH is found to yieldmono-chloro products in the ratio of =65% (CH) CHCH Cl to35:0 (CH3)3CCl-which is only roughly in accord ~ith th~ rateratios quoted above, after 'statistical' allowance has been made. Hchlorination is carried out in solution, the product distribution isfo~~d to depend on ~he nature of the solvent, and particularly on its~billty to.C?mplex With Cl·, thereby stabilising it and thus increasingItS selectivity as compared with its reaction in the vapour phase.

325

(2) CH)· + X 2

-292 (-70)-96 (-23)-88 (-21)-75 (-18)


(I) x· + H-CH)

-134 (-32)-4 (-I)+63 (+ 15)+138 (+33)

(81)

Selectivity in halogenation is found to decrease with rise in temperature.

Halogenation, and particularly chlorination, unlike most radicalreactions, is markedly influenced by the presence in the substrate ofpolar substituents; this is because Cl-, owing to the electronegativityof chlorine, is markedly electrophilic (ct. p. 314), and will thereforeattack preferentially at sites of higher electron density. Chlorinationwill thus tend to be inhibited by the presence of electron-withdrawinggroups, as is seen in the relative amounts of substitution at the fourdifferent carbon atoms in 1-chlorobutane (78) on photochemicallyinitiated chlorination at 35° :

CH)-CH 2-CH 2-CH 2-CI (78)

25% 50% 17% 3%

The variation over the three different CH 2 groups nicely demonstratesthe falling off with distance of the electron-withdrawing inductiveeffect of Cl. The y-(3-)CH 2 group is behaving essentially analogouslyto that in CH 3CH 2CH 2CH 3 , while the lower figure for the CH 3group reflects the greater difficulty of breaking the C-H bond inCH 3 than in CH 2 (see above).

With propene, CH3CH=CH 2 (79), there is the possibility of eitheraddition of chlorine to the double bond, or of attack on the CH3group. It is found that at elevated temperatures, e.g. :::::450° (Cl- thenbeing provided by thermolysis of CI2), substitution occurs to thetotal exclusion ofaddition. This is because the allyI radical (80) obtainedby H-abstraction is stabilised by delocalisation, whereas the one (81)obtained on Cl- addition is not, and its formation is in any- casereversible at elevated temperatures, the equilibrium lying over to theleft :

Cyclohexene undergoes analogous 'allylic' chlorination for the samereasons.

So far as the other halogens are concerned, the AH values-inkJ (kcal) mol- I-for the two steps of the halogenation chain reaction(p. 324) on CH4 are as follows:

tertiary6·7

secondary4·4

primaryI


i? alkanes. The net displacement occurring at carbon on chlorinatIOn, for exaT?ple, of alkanes consists (after initial formation of Cl.)of H-abstractIon from R-H by Cl·, followed by Cl-abstraction fromCI--Cl by R· (this step can also be regarded as direct displacementat Cl), the two steps alternating in a very rapid chain reaction:

CI-C1l hy

R-H + ·CI -+ R· + H-C1

t lCI-Ci

·CI + R-C1

The chain length, i.e.. n~mber ~f RH~ RCI conversions per Clproduced by photolYSIS, IS ::::: 10 for CH4 , and the reaction can beexplosiv~ in sunlight. Chlorination can also be initiated thermolytically,but considerably elevated temperatures are required to effect Cl ~2~1-, an~ the rate of chlorination of C2 H6 in the dark at 120° isvirtually mdetectable. It becomes extremely rapid on the introductionof traces ?f PbEt4 , however, as this decomposes to yield ethyl radicals,Et -, at this temperature, and these can act as initiators: Et - + CI-Cl~ Et-Cl + Cl-. Chlorination of simple alkanes such as these isseldom useful for the preparation of mono-chloro derivatives as thisfirst t;>roduct readily undergoes further attack by the highly ~eactivechlonne, and complex product mixtures are often obtained.

Ease of attack on differently situated hydrogen atoms in an alkaneis found to increase in the sequence,

The figures for fluorination reflect the weakness of the F-F [150 kJ(36 kcal) mol-I ],and the strength of the H-F [560 kJ (134 kcal) mol-I],bonds. Fluorination normally requires no specific initiation (cf p.324),and is explosive unless carried out at high dilution. That fluorinationdoes proceed by a radical pathway, despite not requiring specificinitiation, is demonstrated by the fact that chlorination may beinitiated in the dark, and at room temperature, by the addition ofsmall traces of F z. Bromination is a good deal slower than chlorination,under comparable conditions, as step (l}-H-abstraction by Br· -iscommonly endothermic. This step is usually so endothermic for I· thatdirect iodination of alkanes does not normally take place.

The markedly lower reactivity ofBr· than Cl· towards H-abstractionmeans that bromination is much more selective than chlorination (thefigures refer to H-abstraction by Br- at 25°):

H H

" " "H-C-H < -C-H < -C-H/ / /

primary secondary tertiaryJ 80 1600

A fact that can be put to preparative/synthetic use; thus brominationof (CH 3hCH is found to yield only (CH 3hCBr (cf chlorination, p. 324).The effect is more pronounced when substituents are present that canstabilise the initial radical; thus across the series, CH4 , PhCH3 ,

PhzCHz and Ph3CH the relative rates of bromination differ over arange of 109 , but only over a range of 103 for chlorination. Selectivitydecreases with rise of temperature, however.

Halogenation of an optically active form of a chiral alkane,RR'R"CH, is normally found to yield a racemic (±) halide-a resultthat tells us nothing about the preferred conformation of the intermediate radical, RR'R"C·, as racemisation would be observed witheither a planar, or a rapidly inverting pyramidal, structure (cf p. 310).However, bromination of (+ )1-bromo-2-methylbutane (82) is foundto yield an optically active bromide, ( - )1,2-dibromo-2-methylbutane(83), i.e. the overall substitution occurs with retention of configuration.This is believed to result from the original (l-)bromo substituentinteracting with one side of the intermediate radical (84}-the oneopposite to that from which H has been abstracted-and so promotingattack by Brz on the other, thus leading to retention of configuration:

327

o

~ QNH+ B',to


(88)

-- 0 + HSr*

(87)!Br-Brt

Sr'

+ Sr'

o

.') QNB' +HB,'(85) 0

(b)Q(86)

Bromination of an optically active form of the corre~ponding c~lorocompound (1_chloro-2-methylbutane) also .results m an optically

ctive rOOuct, and retention of configuratIOn. It may be th~t ana tual ~ridged radical is formed, but a somewhat less concrete mt~raCtl'on seems more likely, as halogenation with the more reactiveac . .chlorine is found to lead wholly to racemls~tlO~.

Radical halogenation (particularly chlor~natlOn) by reag~nt~ otherthan the halogens themselves is of conSiderable synth~tlc. lffiportance because of its greater stereoselectivity. Thus ~hlormatlon mayb effected through reaction with alkyl hypochlontes, ROO (e.g..;= Me3C), in the presence of radical initiators, the latter a?stracting 0 to form RO· which has been shown to ~ the species th~tabstracts H from RH; this reagent is used part~cu1arly f?r ~lyl~cchlorination. Another useful reagent for preparatlv~ chlormatlon ISSO a the radical initiator again abstracts 0 to yIeld ·SOzO, andboth this species and the a· it yields by loss of SOz can act asH-abstractors from RH. . . b

Another reagent that is extremely useful syn~hetl~ally IS N.- romo-. . 'd (NBS 85) which is highly selective m attackmg onlySUCClmml e " . . I .

weak C-H bonds, Le. at allylic, benzylic, etc., posItIOns. t requ.lfesthe presence of radical initiators, and has been shown.to effect broml~ation through providing a constant, but very low, ambIent concentra~~nof Br -this is maintained through reaction of the HBr produce mthe r~action with NBS (c, below). There is u~u.ally a trace of Brz orHBr in the NBS that can react with the Imtlator to generate theinitial Br· to start reaction (a, below):

(a) Br2 or HSr + initiator - Br'

MeEt":::, /CH 2 Sr

~ (-)

Sr + Br

(83)

Me, Sro"C::""-"CH ~/. 2

EtBr-Br

(84)


Me [ Me Sr J*Et< /CH 2Sr 80. Et< ~ ~CH2

~ (+) -- T '--H R

(82) Sr

Control of the bromine concentration is maintained by reaction (c)which is fast, though ionic, but can be activated only by HBr producedin the chain reaction (b). The alternative reaction of addition of Br'to the double bond to form (89) is reversible,

329

(95)

(93)

11.5.2.2 Autoxidation

H<f" H to. o.00Hro o~(94)

(92)

(96)

@jH... OH

H,/Ni--

Relative reactivities towards H-abstraction by R0 2 ' at 30° areobserved as follows: PhCI:h:1, Ph2CI:h: 30, andPhC!:hCH=CH2 : 63. ... ..

With alkenes rather than alkanes, autoxidatIOn can lOvolve addItIonof RO . to the 'double bond as well as, or in place of, H-abstraction,partic~larly where there are no allylic, benzylic or tertiary. C-~linkages available. The effect of the presence of such peroXides 10

alkenes on the orientation of HBr addition to the latter has alreadybeen referred to (p. 317). Ethers are particularly prone to autoxidation,initial attack taking place at a C-H linkage ex- to the ox~gen atomto yield a stabilised radical; the first-fo~med hydroperoxlde ~eactsfurther to yield dialkyl peroxides that are highly explOSive on heatmgnot to be forgotten on evaporating ethereal solutions to dryness!Accumulated peroxides, in ether that has been standing, .may besafely decomposed before its use by washing with a solutIon of areducing agent, e.g. FeS04' .

Autoxidation may in some cases be ofpreparatIve use; thus referencehas already been made to the large-scale production of phen~l+acetone by the acid-catalysed rearrangement of the hydroperoxIdefrom 2-phenylpropane (cumene, p. 128). Another example involves thehydroperoxide (94) obtained by the air oxidation at 70° of tetrahydronaphthalene (tetralin); the action of base then yields the ketone(ex-tetralone, 95), and reductive fission of the 0-0 linkage the alcohol(ex-tetralol, 96) :

tion then becomes autocatalytic. The addition of O2 to R· is veryfast, often diffusion-controlled, but the peroxy radicals (91) areusually of relatively low reactivity (cf. ·0-0· i~self, p. 31?), and arethuS highly selective in the positions from WhICh they wIll abstractH. Thus allylic and benzylic C-H are relatively readily attacked,because the C-H bonds are slightly weaker and the resultantradicals stabilised by delocalisation, e.g. the allylic position in cyclopentene to form (92). In simple alkanes only tertiary ~-H isgenerally attacked, e.g. as in decalin, which yields the bridgeheadperoxide (93):


o ;;. 0·'(86) (89)

while formation of (87) is not; overall substitution is thus favouredover addition so long as [Br2J is kept low. The radical (87) is alsostabilised by delocalisation, while (89) is not (cf p. 311). Support forthe above interpretation of the reaction of NBS is provided: (i) by thefact t~at NBS shows exactly the same selectivity ratios as does Br2'

and (11) by the fact that cyclohexene (86) is found to undergo largelyaddition with high concentrations of bromine, but largely allylicsubstitution with low (it is necessary to remove the HBr produced-ashappens with NBS).

'0·Ra' + H-R -+ Ra-H + R· ---!.. RO-O' (91)

t LR-HR' + RO-OH (90)

Under certain conditions the hydroperoxide (90) itself breaks downto radicals, RO' + .OH which can act as initiators, and the autoxida-

11.5.2.2 Autoxidation

Autoxidation is the low temperature oxidation of organic compoundsby O2, involving a radical chain reaction; as opposed to combustionwhich happens only at higher temperature. The initial stage iscommonly the formation of hydroperoxides, RH -+ ROOH, so it isa net, overall displacement, though the actual pathway involves Habstrac~on and O2 addition (see below). The first-formed hydroperoXldes frequently undergo further reactions. Autoxidation is ofimportance in the hardening of paints, where unsaturated esters inthe oils used form hydroperoxides, whose decomposition to RO'initiates polymerisation in further unsaturated molecules to form aprotective, polymeric, surface film. But autoxidation is also responsible for deleterious changes, particularly in materials containingunsaturated linkages, e.g. rancidity in fats, and perishing of rubber.Indeed, the gradual decomposition of most organic compoundsexposed to air and sunlight is due to photosensitised autoxidation.Autoxidation may be initiated by trace metal ions (cf. below), aswell as by light and the usual radical initiators.

The main reaction pathway is a two-step chain involving Habstraction:

Aldehydes, and particularly aromatic ones, are highly susceptibleto autoxidation; thus benzaldehyde (97) is rapidly converted intobenzoic acid (98) in air at room temperature. This reaction is catalysedby light and the usual radical initiators, but is also highly susceptibleto the presence of traces of metal ions that can act as one-electronoxidising agents (ct. p. 306), e.g. Fe3e, C03e, etc:

000II II '0,' II

(a) Fe 3Ell + PhC-H -> Fe2Ell + HEll + PhC' -> PhC-O-O' (100)

(97) (99) 0

1 II (97)(b) t PhC-H

o 0II II

PhC· + PhC-O-OH (101)

(99)

o 0 0II II H" II

(e) PhC-O-OH + PhC-H -> 2PhC-OH

(10I) (97) (98)

The oxidation is initiated (a) by Fe3e to yield the benzoyl radical (99)which adds on a molecule of oxygen to form the perbenzoate radical(100), this reacts with benzaldehyde (97) to yield perbenzoic acid (101)and another benzoyl radical (99)--these two steps constituting thechain reaction (b). The actual end-product is not perbenzoic acid(101), however, as this undergoes a rapid acid-catalysed, non-radicalreaction (c) with more benzaldehyde (97) to yield benzoic acid (98).This latter reaction (c), being acid-catalysed, speeds up as the concentration of product benzoic acid (98) builds up, i.e. it is autocatalytic.That benzoyl radicals (99) are involved is borne out by the observationthat carrying out the reaction at higher temperatures (~ 100°), and atlo~ oxygen concentrations, results in the formation of CO, i.e. byPhCO-+ Ph· + CO.

The autoxidation of aldehydes, and of other organic compounds,may be lessened considerably by very careful purification-removal ofexisting peroxides, trace metal ions, etc.-but much more readily andeffectively by the addition of suitable radical inhibitors, referred to inthis context as anti-oxidants. The best ofthese are phenols and aromaticamines which have a readily abstractable H atom, the resultantradical is of relatively low reactivity, being able to act as a goodchain terminator (by reaction with another radical) but only as a poorinitiator (by reaction with a new substrate molecule).

An interesting, and slightly different, autoxidation is photooxidation of hydrocarbons such as 9,lo-diphenylanthracene (102) insolvents such as CS 2 • The light absorbed converts the hydrocarboninto the stabilised diradical (103, cf p. 337), or something rather like it,

331

*Ph

(104)

11.5.2.3 Aromatic substitution

©¢©Ph

(103)

hI'--+~

~Ph

(102)

as to make it impossible to work with in the presence of sunlight andair (cf p. 337).

(105)

~X ~r&+Ra-H.Iow V r••t l8J

(106) (107)

Loss of a hydrogen atom from the delocalised cyclohexadienyl radicalintermediate (106) to yield the substituted end-product (107) does notproceed spontaneously, however, but requires intervention by a

Similar photo-oxidation occurs with increasing readiness as the numberof benzene rings in the lin (rings joined successively in the same line)hydrocarbon increases, i.e. as its overall aromatic character decreases;this occurs so readily with, for example, the very dark green hydrocarbonhexacene (105),

11.5.2.3 Aromatic substitutionAttack on aromatic species can occur with radicals, as well as withthe electrophiles (p. 131) and nucleophiles (p. 1~7) that w~ have ~lre~dyconsidered; as with these polar species, homolytic aromatic substitutiOnproceeds by an addition/elimination pathway:

stabilisation occurring through delocalisation of its unpaired electronsand also by conversion of a partially aromatic state in (102) to acompletely aromatic one in (103). The diradical then adds on a moleculeof oxygen to yield the trans-annular peroxide (104) in a non-chainstep:


The very small spread in relative rate, as Y is varied, is in markedcontrast to electrophilic substitution, e.g. nitration, on the samesubstrates where the spread in relative rate would have been =108

•

Though it should be remembered that phenylation involves attackby a species of low polarity.

For simplicity's sake only the products of p-interaction in (106) havebeen shown above: o-interaction can also lead to an o-dihydro isomerof (109), and to both 0-/0- and o-/p-coupled isomers of (108). Productmixtures from arylation of aromatic species can thus be quite complex.

So far as the overall substitution reaction (~107) is concerned,marked differences from electrophilic and nucleophilic attack becomeapparent as soon as the behaviour of substituted benzene derivatives(C6H 5Y) is considered. Thus homolytic attack on C6H 5Y is found tobe faster than on C6H 6 , no matter whether Y is electron-attracting or-withdrawing, as shown by the relative rate data for attack by Ph-:

33311. 5. 2. 3 Aromatic substitution

It is also found, as shown by the partial rate factors (ct. p. 156),

PhY f~ f... f~PhOMe 5·60 1·23 2·31PhN0 2 5·50 0·86 4·90PhMe 4·70 1·24 3·55PhCI 3·90 1·65 2·12PhBr 3·05 1·70 1·92PhCMe3 0·70 1·64 1·81

that, irrespective of the nature of Y, the observed preference forposition of attack by Ph- is 0- > p.- > m- except, as wi~h Y = Me3C,where the steric effect of Y may lDlpede o-attack. This preferencefor 0- and p-attack can be rationalised on the basis that theelectron, brought by the attacking Ph- radical to the intermediate(106), can be delocalised (and the intermediate thereby stabi1is~d)by either an electron-withdrawing (110), or an electron~~onatmg

(111), substituent-as shown here for attack at the p-pOSltlon:

Qh ~o Q~QN(Il N(Il :OMe 'OMe

,f' " 8 / "08 (Ilo 0 '0(I lOa) (I lOb) (1110) (lllb)

There is, however, no very satisfactory explanation of why suchm-attack as does take place on ~HsY should also be faster thanattack on ~; or of why attack on the o-position in C6HsY iscommonly faster than attack on the p-position. The relatively smallspread of the partial rate factors for a particular C6HsY means thathomolytic aromatic substitution normally leads to a more complexmixture of products than does electrophilic attack on the samespecies.

The above data all refer to phenylation by Ph- derived from(PhC0

2h. This, and other, diacylperoxides have been much used for

this purpose, but as Ph- formation involves the step: PhC02-

Ph- + CO2 , it is usually impossible to stop some formation of esterseither by Ph-+PhC02 - or, more commonly, by attack of PhC02 - onthe aromatic substrate (acyloxylation). This particular difficulty maybe avoided by generating AI- from the thermal decomposition ofN-nitroso derivatives of acetylated aromatic amines,AIN(NO)COMe, or of diazonium salts, AIN2$, under slightly ~asicconditions-the latter is the Gomberg reaction for the syntheSIS ofunsymmetrical diaryls, AI-AI'. In each case the AI- precursor isdecomposed in the presence of an excess of the aromatic substrate,which is in fact often used as the solvent. The yield from theclassical Gomberg reaction may be much improved by diazotising

Ph4·0

CN3·7

-+Ph~Ph

H~H

(108)

Me1·9

Br1·8

(106)

PhD·········

H .........

OMe1·2


Ph-@

(107) __

+PhOH

H _ H(109)

332

further radical, Ra-, to abstract H-. Reaction between two radicals(106) and the H-abstractor-is likely to be fast, i.e. non ratelimiting, and no significant kH/kD kinetic isotope effect is observed,i.e. attack of Ra- on the original aromatic substrate is rate-limiting.Overall substitution reactions have been investigated in which Ra- isAI- (especially Ph-), PhC02 - (and some RC02 -), R- and HO-.Attack by HO-, hydroxylation, is of particular importance in biological systems: as the first step in the detoxification of 'foreign'aromatic molecules. There are also a few reactions known in whichit is an atom or group other than H, e.g. halogen, MeO, that isdisplaced. It is, however, the displacement of H by AI--arylationthat has been studied in by far the greatest detail.

Attack of, for example, Ph - on aromatic species such as benzeneis found to lead to products other than the one arising from overallsubstitution (107, Ra = Ph). This is because the intermediate radical(106), as well as undergoing H-abstraction to (107), can also dimeriseto (108) and/or disproportionate to (107) + (109):

dimer

(113)

This is essentially a chain reaction involving Cuo ~CuED interconversions. Similar results have been achieved by photolysis of thearyl iodide corresponding to the diazonium salt (112).

An interesting further example of a homolytic aromatic reactioninvolves the oxidation of phenols in basic solution with one-electronoxidising agents (e.g. Fe(m), H20 + the enzyme peroxidase):


[>t~J [>LJ [j~Jcarbocation T.S. radical T.S. carbanion T.S.

(2e) (;}e) (4e)

11.5.3 RearrangementRearrangements that involve radicals are found to be much lesscommon than otherwise similar rearrangements that involve carbocations. In this they resemble carbanions (cf. p. 292), and the reasonfor the resemblance becomes apparent when we compare the T.S.sfor a 1,2-alkyl shift in the three series:

These T.S.s involve two, three and four electrons, respectively (cf. p.292), and electrons in excess of two can be accommodated only. inan anti-bonding molecular orbital of much higher energy. As wIthcarbanions, however, 1,2-aryl shifts are known in radicals, whichinvolve stabilised, bridged transition states, e.g. (114). A goodexample is with the aldehyde (115), which undergoes H-abstractionfrom the CHO group by Me3CO· (ex. Me3COOCMe3) to yield theacyl radical (116), which readily loses CO to form (117). This can, inturn, abstract H from RCHO (115) to form a hydrocarbon, but theonly hydrocarbon actually obtained is not the one derivable from(117), but the one (118) from the rearranged radical (119):

H ~I Me CO· • CO I

Ph2MeCCH

2C=O~ Ph 2MeCCH 2C=O~ PhMeC-CH 2 '

(115) (116) (117)

L

[ Oi··········~]*H Ph ::.....I RCHO I " '.

PhMeC-CH Ph + RCO+--- PhMeC-CH l - PhMeC-CH 22 (II S)

(118) (116) (119) (114)

Such phenolic coupling has here been shown occurring through twoo-positions, but other (and mixed) combinations of coup~ing through0-, p- and 0 atoms have also been observed; % couplIng does .notgenerally occur because of the instability of the resultant peroXIde.The study of these reactions is, of course, complicated by the factthat the initial dimeric product can itself, in turn, be oxidised to aphenoxy radical that can react either with itself or with furthersimple phenoxy radicals. Such phenolic coupling reactions, controlled by enzymes, are of the greatest importance in the biosynthesis ofmany natural products including alkaloids, lignins, pigments andantibiotics.


the free amine, ArNH2 , with CsHuONO in solution in the aromaticsubstrate; no acid is necessary and the whole reaction is thenhomogeneous. Some Ar-Ar' syntheses work quite well, but thereaction is not really of general applicability.

Intramolecular radical arylations are found to work quite well,however, e.g. the Pschorr reaction; this involves the thermal decomposition of diazonium salts, e.g. (112), in the presence of copperpowder as catalyst, and is used in the synthesis of phenathrenes suchas (113):

~+C"O~(112) l

~+C"O~

phenoxy radical

We might well expect the resultant phenoxy radical to attackthrough the unpaired electron on its 0, or on its 0- or p-C, atom-afurther molecule of phenol or phenoxide anion. Such homolyticsubstitution on a non-radical aromatic substrate has been observedwhere the overall reaction is intramolecular (all within the singlemolecule of a complex phenol), but it is usually found to involvedimerisation (coupling) through attack on another phenoxy radical:

o 0 0 0 OH HO

6·H+H~_~~~~ 'U VHU V U

CI CI HBr' J. HBr I I

CI)C-CH=CH 2 -+ C1 2C-CH-CH 2 ---x--. CI 2C-CH-CH + Dr'I I 2

(124) (127a) 1 Br (125) Br

/C~ CI H CI

CI C . C . I HBr I I2 - H-CH 2 -+ CI 2C-CH-CH 2 -+ C1 2C-CH-CH 2 +

I I IBr Br Br

The driving force of the reaction is the formation of a more stableradi.cal, i.e. the unpaired electron is delocalised more effectively bya 10 (127b) than by H in (127a). Migration of fluorine does notoccur as its d orbitals are not accessible, and migration of Br onlyrarely as the intermediate radicals undergo elimination (to alkene)more readily than rearrangement.

While no 1,2-alkyl shifts have been observed in solution, the 'coolflame' oxidation of Me 3CH (in the gas phase at 480°) is found to

The rearranged radical (119) is more stable than the original one(117) not only because the former is tertiary and the latter primary,but also because (119) is stabilised by delocalisation of the unpairedelectron over the 1t orbital system of a benzene nucleus. It is significantthat only Ph migrates in (117), despite the fact that migration of Me

would yield.the even more stabilised radical, Ph2CCH2Me; this reflectsthe energetic advantage of migration via a bridged, delocalised T.S.such as (114). When no Ph group is present, as in EtMe2CCH2'from EtMe2CCH2CHO, no migration takes place at all and the endproduct is EtMe2CCH3'

~ryl m~grations are.not confined to carbon/carbon rearrangements,as IS seen In the behavIOur of (Ph3COh (120, cf p. 300) on heating:

337

MeIC=OI

Me--CH2

11. 6 Biradicals

yield considerable quantities of MeCH2COMe:

Me Me MeI O

2I I

Me-C-H _ Me--C-OO' ----+ Me--C-o+OH ----+I I \ I

Me CH2-H -.......CH2 •

(128)

The formation of this ketone is believed to proceed via internalabstraction of H in the initial peroxy radical (128; ct· p. 328),followed by migration of Me'. It may be that the vigorous condi.tionsemployed now make a 1,2-a1kyl shift feasible, or that the shift ofMe' may involve fragmentation followed by re-addition, rather thandirect migration.

Radical migration of hydrogen is also known, though only overlonger distances than 1,2-shifts, e.g. a 1,5-shift to oxygen via a 6membered cyclic T.S. in the photolysis of the nitrite ester (129}-anexample of the Barton reaction:

H

~~.

H

(130)

Biradicals have also been encountered as intermediates in the Mgreduction of ketones to pinacols (p. 218) and, as radical anions, inthe acyloin condensation of esters (p. 218). The thermolysis of cyclopropane (131) to propene (132) at ~ 500° is also believed to involve

11.6 BIRADICALS

The oxygen molecule, a paramagnetic species with an unpaired electronon each atom, has already been referred to as biradical, albeit anunreactive one. The photochemical excitation of an anthracene to abiradical, or to something rather like one, has also bee~ mentioned(p. 331); if this excitation is carried out in the absence of a~r or oxyge~,instead of the trans-annular peroxide-(I04}-a photo-dimer (130) IS

obtained:

(126)(127b)(123)


Ph Ph6 I I Ph C-OPh

Ph)CO-OCPh) -+ 2Ph 2C-O' -+ 2Ph 2<;::-O -+ 2 IPh2C-OPh

(120) (121) (122)

This too proceeds via a bridged T.S.; again the driving force of therearrangement is the much greater stability of (122) than (121). As wellas 1,2-aryl shifts, similar migrations of vinyl, acyl and acyloxy groupsare known, occurring via bridged transition states or intermediates,and al~o 1,2-chlorine shifts in which an empty d orbital on the halogen~tom IS .used to accommodate the unpaired electron in a bridgedmtermedlate, e.g. (123). Thus photo-catalysed addition of HBr toCCI3CH=CH2 (124) yields none of the expected CCI3CH2CH2Br(125), but 100% ofCHCI2CHCICH 2Br (126):

(135)

biradical intermediates, e.g. (133) and (134):

339

(139)

11. 6 Biradicals

(138)

both n = 3 and n = 4 species are paramagnetic in the solid state,corresponding to ~8% biradical character for n = 3 and 15 % forn = 4 at 200

•

Here there is no formal electronic bar to .interac.tion between t~eelectrons, i.e. pairing to form the diamagnetic specl~s (137); b~t thlsdoes not in fact happen, because the bulky chlonne atoms 10 ~hea-positions prevent the benzene rings from att~ining a c~nformatlOnclose enough to coplanarity to allow of sufficlent p orbltal overlapfor electron-pairing to occur.

Interestingly enough, some biradical character has been obser~edin systems similar to (136) even ~he? there ~re no bulky c~lonneatoms present to inhibit delocahsatlOn stencally. Thus wlth thesystem (138) p(139),

Ph,t~i:Ph' <=' Ph,CM.CPh'

CI CI

Ph'C«rCPh'CI CI

(137)


CI CI

Ph,t#tPh'*

CI CI

(136)

338

It behaves, hardly surprisingly, very like Ph3C· (ef p. 300), existingout of solution as a colourless solid, but this latter is probably apolymer rather than a dimer as with Ph 3 C·. The solid is dissociatedin solution to about the same extent as the Ph3C· dimer. The unpairedelectrons in the biradical form (135) cannot interact with each otherto form a wholly paired, diamagnetic species, as such interactionacross both central benzene nuclei would necessitate m-quinoid formsthat cannot exist; the electrons are thus 'internally insulated' fromeach other. Such internal insulation in biradicals may also arise throughsteric rather than electronic causes. Thus the species (136) exists insolution as a biradical to the extent of ~ 17 %, being in equilibriumwith a polymer (like 135):

H/C~2 ~ /:H2 ----. 'C~2 ----. C~3

H2C-CH2 H-HC-<;:H2 <;:H-<;:H2 CH=CH2

(131) (133) (134) (132)

In order to form the biradical (133), the cyclopropane moleculebecomes vibrationally excited by collision with another molecule;the C-C bond may then break provided the extra energy is not losttoo rapidly by further collision. There is driving force here for a 1,2shift of hydrogen-unlike in mono-radicals (p. 335)-because of theopportunity of electron-pairing to form a 7t bond (with evolution ofenergy) in (134). There is evidence that this H-migration is commonlythe rate-limiting step of the reaction.

The above biradicals, with the exception of the oxygen molecule,are all highly unstable; there are, however, a number of much morestable species that show evidence of biradical character. Thus thehydrocarbon (135) exists, in part, in solution as a biradical:

12.1 INTRODUCTION

Symmetry controlled reactions

(7)

cxMe

~ Me

~Ph Ph

(5)

...+==!130'

h.-(4)

(6)

~Me

~Me

(4)

(8)

~Me

~···Me

can be induced photochemically but not thermally. The differentialstereochemical effect is clearly seen with trans, cis, trans 2,4,6-octatriene(6); this found to cyclise on heating to give the cis 1,2-dimethylcyclohexa-3,5-diene (7) only, while on photochemical irradiation it cyclisesto give the trans 1,2-dimethylcyclohexa-3,5-diene (8) only:

12.1 Introduction 341

Such reactions are apparently concerted, i.e. the electronic rearrangements involved in bond-making/bond-breaking proceed simultaneously in a one-step process; though each bond undergoingchange need not necessarily have been made or broken t,o the sameextent by the time the T.S. has been reached. The transiti~n states arecyclic, ones involving six p electrons (cf aromaticity,jp. 17) beingpreferred though not essential, and the reactions are normally attendedby a high degree of stereoselectivity (cf p.268). Many of the reactionsare reversible, e.g. Diels-Alder reactions, though the equilibrium oftenlies well over to one side or the other. The term pericyclic has beencoined to describe such concerted reactions that proceed via cyclictransition states.

As pericyclic reactions are largely unaffected by polar reagents,solvent changes, radical initiators, etc., the only means of influencingthem is thermally or photochemically. It is a significant feature ofpericyclic reactions that these two influences often effect markedlydifferent results, either in terms of whether a reaction can be inducedto proceed readily (or at all), or in terms of the stereochemical coursethat it then follows. Thus the Diels-Alder reaction (cf above), anexample of a cycloaddition process, can normally be induced thermally but not photochemically, while the cycloaddition of twomolecules of alkene, e.g. (4) to form a cyclobutane (5),

This type of reaction, whether it involves the cyclisation of apolyene, as here, or the ring-opening of a cyclic compound to form apolyene, is known as an electrocyclic reaction.

Whether a particular reaction proceeds via a single step, concerted pathway, or in more than one step via a biradical or bipolar

RHC=CH 2(3)

-H 0\ IjS-CSMe

RHC-CH 2I \H 0

/S=CSMe

(2)

340

RHC=CH 2(3j

- H 0\ IjO-CR'

RHC-CH 2I \

H 0/

O=CR'

(I)

12

12.1 JNfRODUCTION, p. 340.12.2 PHASE AND SYMMETRY OF ORBITALS, p. 342.12.3 ELECIROCYa..IC REACTIONS, p. 342.12.4 CYCLOADDITlONS, p. 348:

12.4.1 Diels-Alder reaction (41T+21T), p. 349; 12.4.21,3-Dipolar additions, p. 351.

12.5 SIGMATROPIC REARRANGEMENTS, p. 352:12.5.1 Hydrogen shifts, p. 352; 12.5.2 Carbon shifts, p. 354.

At a time when general mechanistic considerations had brought order (and light to our understanding of the vast majority of organic reactionsthere remained a small group of apparently unrelated reactions thatappear to proceed neither by a polar, nor by a radical, pathway.Thus they do not involve polar reagents, are substantially uninfluencedby changes in solvent polarity, by the presence of radical initiators(or inhibitors) or other catalysts, and all attempts to isolate, detect,or trap intermediates were unsuccessful. Examples of such reactionsthat we have already encountered, are the Diels-Alder reaction (p.197),involving the 1,4-addition of (usually) substituted alkenes to conjugateddienes;

( "J-eX~ OX

~ .. , y y

and the pyrolytic elimination reactions ofcarboxylic esters (l), xanthates(2), etc. (p. 268), to yield alkenes (3):

Fig. 12.1

12.2 PHASE AND SYMMETRY OF ORBITAlS

We have already seen (p. 2) that the individual electrons of an atomcan be symbolised by wave functions, t/J, and some physical analogycan be drawn between the behaviour of such a 'wave-like' electronand the standing waves that can be generated in a string fastened atboth ends-the 'electron in a (one-dimensional) box' analogy. Thefirst three possible modes of vibration will thus be (Fig. 12.1):

343

Anti-bonding

12.2 Phase and symmetry of orbitals

r It is important to emphasise that rjl2, which represents the probability of findingan electron in a particular element of space, will always be positive, no matter whetherrjI is positive or negative.

and the four MOs (t/Jt, t/J2' t/J 3 and t/J4) arising from the four p AOs in

Molecular orbitals are obtained by the linear combination ofatomic orbitals, and the question of phase will, of course, arise withthem too. Thus we can write the two MOs (11' and 11'*, cf. p. 12) arisingfrom the two p atomic orbitals in ethene,

(9) (10)

..fJ_. Nodal __, __CiJ plane (j

In the first mode, t/J l' the amplitude of the wave increases from zeroto a maximurn, and then decreases to zero again; in the second, t/J 2'

the amplitude increases to a maximum, decreases through zero (a node,marked - above) to a minimum, and then back to zero again, i.e. thephase of the wave changes once; while in the third mode, t/J 3' theamplitude changes from zero to a maximum, through zero to aminimum, through zero to a maximum again, and then finally backto zero, i.e. there are two nodes (marked - above), and the phase ofthe wave changes twice. Displacements above the nodal plane areconventionally designated +, and those below -. The lobes of, forexample, a 2p atomic orbital, which has one nodal plane, thus differin phase, and are conventionally designated as + and -, i.e. (9); thiscan, however, lead to confusion because of the usual association of+ and - with charge,t and phase differences, which are purelyrelative, will therefore be designated here by shading and no shading,i.e. (10):

Symmetry controlled reactions342

intermediate, will be determined by the relative magnitude of .iGt

(cf. p. 38) for the former compared with that of .iG+ for therate-limiting step of the latter. .iG+ is, of course, the resultant of.is'1'' and .iH+ terms; it is found in practice that concerted reactionstend to have large -ve values of .is+ and small +ve values of .iH+.The former reflects the degree of ordering-of participant moleculesor groups-required by a cyclic T.S.; while the latter reflects theextent to which energy derived from bond-formation in the T.S. canassist in necessary bond-breakin!. It should, however, be stressedthat observation of high -ve .is , and low +ve .iH+, values for aparticular reaction cannot necessarily be taken as establishing that itproceeds via a concerted pathway.

So far as .iH+ is concerned, it would seem reasonable to supposethat the favoured pathway for a particular reaction would be that?ne in which t~e greatest degree of residual bonding is maintained10 the T.S. Mamtenance of bonding implies maintenance of orbitaloverlap, and it is therefore necessary to establish the conditions thatensure the maintenance of such overlap. To do this we have toconsider a property of atomic and molecular orbitals not yet referred to, namely phase.

345

(7)

Anti-bonding)( ~

(13)

(12)

12.3 Electrocyclic reactions

III

(11)

Me Me Me Me

M'~M'_Dis_rot_ato--+,Y. ~B~ U

Mef1;* ~M~onrotatory~---~~

M'~Ut M'~M'1/13 1/16

iM'~M' M'~'

i'" '"01)..

1/12 ~"C 1/15 ""'-l c"'-l

M'~M' M'~M'1/1. 1/14

Bonding Anti-bonding

As there are 6 1T electrons to accommodate-two per orbital-theHOMO will be 1/13 (11). To form the C-C (T bond on cyclisation,the orbital lobes on the terminal carbon atoms of the conjugatedsystem (C2 and C7-the C atoms carrying the Me substituents) musteach rotate through 90° if mutual overlap is to occur (piSp2~ Sp3re-hybridisation must also occur). This necessary rotation could beeither (a) both in the same direction-conrotatory (12), or (b) eachin opposite directions-disrotatory (13):

cyclohexa-3,5-diene (7) only, and photochemically to yield the. ~or~esponding trans isomer (8) only: in either case any eqUlhbnumlies all but completely over in favour of the cyclic product. Thestereoselectivity is in fact so great that the thermal cyclisation yields<0·1 % of the trans isomer (8), despite the latter being thermodynamically more stable than the cis form (7). The six MOs of(6)-1/11' 1/12, 1/13, 1/14' 1/15 and 1/16, arising from the six p AOs-may bewritten (cf. butadiene, p.344):

Bonding

Anti-bonding


butadiene (ef Fig. 1.2, p. 12) in the eisoid conformation (p. 197):

The importance of considering the phase of orbitals is that: onlyorbitals ofthe same phase will overlap, and so result in a bonding situation;orbitals of different phase lead to a repulsive, anti-bonding situation.

By a consideration of the relative phases, and hence overallsymmetry, of the orbitals involved, Woodward and Hoffmann wereable in 1965 to formulate a set of rules; these not only explained thebehaviour of the pericyclic reactions that were known to date, butalso made precise predictions about the behaviour to be expected ofmany others, that had not yet been carried out. These predictionsincluded whether reactions would be induced thermally or photochemically, and the detailed stereochemistry that would then befollowed. The achievement is all the greater in that a number of thepredictions-since proved correct- appeared at the .time to be highlyimplausible. To make these predictions it was necessary to considerthe relative phases, i.e. symmetry, of all the orbitals involved duringthe transformation of reactants into products. It is, however, possibleto obtain a reasonable understanding, much more simply, by use ofthe frontier orbital approach. In this, the electrons in the J:}.ighestQccupied Molecular Qrbital (HOMO) of one reactant are lookedupon as being analogous to the outer (valence) electrons of an atom,and reaction is then envisaged as involving the overlap of this(HOMO) orbital-a potential electron donor-with the L,owestT)noccupied Molecular Qrbital (LUMO)-a potential electronacceptor---of the other reactant. Where, as in electrocyclic reactions, only one species is involved only the HOMO need beconsidered. A variety of pericyclic reactions will now be reviewed,using this approach.

12.3 ELECfROCYCLIC REACfIONS

We have already seen (p. 341) that the cyclisation of trans,cis,trans2,4,6-octatriene (6) proceeds thermally to yield cis 1,2-dimethyl-

347

(18)

(19)

Anti-bondingH •

Bonding-

(21)

(22)

Me Me~ Anti-bonding

H )('

12.3 Electrocyclic reactions

Me Me--n.£l.- Disrotatory ~Me~Me In

(13) (25)

-t:: ~ DisrotatoryMe~Me •

(20)

This time it is conrotatory movement that results in a bondingsituation and formation of the trans dimethylcyclobutene (18). Forthe phot~hemical interconversion (which tends to I~e over in fav?urof the cyclobutene~ irradiation of the diene will result 10 the promotIOnof an electron into the orbital of next higher energy level, i.e. t/J 2 l!!+ t/J 3'

and the HOMO to be considered now therefore becomes l/!3 (23):

l/!2 (20) as there are four 'IT electrons to accommodate:

M'oe:~eBonding-

Thus disrotatory movement now results in a bonding situation, andformation of the cis dimethylcyclobutene (19).

The difference in stereochemical outcome of these reactions isdetermined, therefore, by the relative phase of the lobes-at theterminal carbon atoms-of the MOs of these (and other similar) n'ITesystems: by the symmetry of their orbitals, that is. As we have seen,the orbital lobes, at the two terminal carbon atoms, have the samephase in the HOMO (l/!3) of the triene (6'ITe), and in the. HOMOafter irradiation (l/!3) of the diene (4'ITe); while these orbital lobeshave opposite phases in the HOMO (l/!2) of the diene a?d in theHOMO after irradiation (l/!4) of the triene. Two such termmallobeswith the same phase require disrotatory movement before bondmaking/bond-breaking can occur, while two terminal lobes with

(18)

Me~Me

(17)(19)


Conrotatory movement results in the apposition of orbital lobeswith opposite phase-an anti-bonding situation, while disrotatorymovement results in the apposition of orbital lobes with the samephase-a bonding situation, leading to formation of the cyclohexadiene (7) in which the two Me groups are cis.

On photochemical ring-closure, irradiation results in the promotionof an electron into the orbital of next higher energy level, i.e. t/J 3 l!4 t/J4

and the HOMO to be considered now therefore becomes l/!4 (14):

M~irMe Me Me

~ bConrotatory Bonding• ----.

(IS) (8)

HOMO("'4): III

M'~M' gDisrotatory Anti-bonding• )( •

(14) (16)

It is now conrotatory movement that results in the apposition oforbital lobes with the same phase-the bonding situation, leading toformation of the trans isomer (8).

It is interesting to contrast the above with the hexa-2,4-diene p3,4-dimethylcyclobutene situation. Here exactly the opposite stereochemical inter-relationships are observed, i.e. trans, trans hexa-2,4-diene(I7) is associated thermally with trans 3,4-dimethylcyclobutene (18),and photochemically with the cis isomer (19):

For the thermal interconversion (the equilibrium tends to lie overtowards the diene), the HOMO for the diene (17; ct. p. 344) will be

12.4 CYCLOADDITIONS

HOMO(1£) H

349

cijJPh Ph

(5)

---+

12.4.1 Diels-Alder reaction

o(htJPh Ph

(4) (4)

The importance of (47T + 27T) thermal, concerted cycloadditions isgreat enough to warrant their separate consideration.

12.4.1 Diels---Alder reaction

By far the best known (47t + 27t) cycloaddition is the Diels-Alderreaction. This has been discussed to some extent already (p. 197),including the fact that it proceeds rigorously, stereospecifically SYN,with respect to both diene (26) and dienophile (27):

This is confirmatory evidence of a concerted pathway, implying as itdoes the simultaneous formation of both new (J bonds in the T.S.That both new bonds are not necessarily formed to the same extentin the T.S. is, however, suggested by the fact that the reaction is markedlyinfluenced by the electronic effect of substituents. It is found to bepromoted by electron-donating substituents in the diene, and by

<R.. H R' ~R...H H~ H......)( ---+ I -'-R'::0- H...... H

." R' "H " R.'

R R 'H(26) (27)

allowed, as irradiation will promote an electron, of one component,into the orbital of next higher energy level, i.e. 7T~ 7T*, and theHOMO to be considered now therefore becomes (7T*);

HOMO(~ 1£*) HLUMO(1£*) H

Many such reactions may indeed be carried out preparatively underphotochemical conditions, though, for reasons that cannot be goneinto here (the detailed mechanism of photochemical changes), theyare often not concerted but proceed via biradical intermediates. Onephotochemical (27t + 27t) cycloaddition that does, however, proceedvia a concerted process is the example we have already referred to:

Motion forbonding

conrotatorydisrotatorydisrotatoryconrota tory

~ LUMO(tjlJ)

H HOMO(1T)

Conditions forreaction

thermalphotochemicalthermalphotochemical

No.of1£electrons

4n4n4n + 24n + 2

HOMO(tjlz)

LUMO(1£*)


In cycloadditions two components are commonly involved, and thefeasibility of a concerted process will be determined by whether overlapcan take place between the HOMO of one component and the LUMOof the other. Thus for a diene plus a monoene,

Apart from their intrinsic interest, these electrocyclic reactions haveconsiderable synthetic carbon--earbon bond-forming importancebecause of their rigid stereospecificity, which is much greater than inthe vast majority of other, non-concerted reactions involving biradicalor bipolar intermediates.

LUMO(1£*)H

opposite phases require conrotatory movement before bondmaking/bond-breaking can occur. This thermal/photochemical antithesis may thus be summarised in the generalisations:

the situation is a bonding one and concerted addition will be feasible,whichever component has the HOMO, or the LUMO: the cycloaddition is said to be symmetry allowed. By contrast, for two monoenecomponents,

the situation is a non-bonding one and concerted addition will not befeasible: the cycloaddition is said to be symmetry forbidden.

This is a general situation for thermal, concerted additions. thoseinvolving 47te + 27te systems proceed readily, e.g. the Diels-Alderreaction, whereas those involving 27te + 27te systems, e.g. the cyclodimerisation of alkenes, do not. We might, however, expect thatphotochemical cyclodimerisation of alkenes would be symmetry

35112.4.2 1,3-Dipolar additions

adduct may sometimes be increased by longer reaction times: thefirst-formed endo (kinetic) product then being converted into themore stable exo (thermodynamic) product via reversal of the reaction and subsequent exo re-addition (cf. p. 283).

The great advantage of the Diels-Alder reaction. (as a c~rbon

carbon bond-forming process) is its generality; the vanety of dIfferentdienophiles that can be used preparatively is very wide indeed (possiblevariations in the diene are somewhat less wide), and conditions canusually be found to make the great majority of such cycloadditions goin satisfactory yield. Like other cycloadditions, Diels-Alder reactionsare potentially reversible and in some cases the re.verse process, theretro Diels-Alder reaction, can be made preparatively useful. Thuscyclopentadiene (32) will readily undergo an auto Di~ls-A~der rea~tion

to form a tricyclic dimer; it is commonly stored In thIS, relativelystable form and reconverted to (32) on heating, i.e. by distillation, asrequi;ed. The thermal cracking ofcyclohexene (the Diels-Alder adductof butadiene and ethene-though not prepared that way!) has beenused as a useful method for the laboratory preparation of butadiene.A few Diels-Alder reactions are known, particularly those involvinghetero atoms and/or highly polar substituents, that proceed via anon-concerted, two-step pathway involving zwitterion intermediates. Reactions proceeding via a two-step pathway involvingbiradical intermediates have not, however, yet been observed. Thepyrolytic SYN eliminations of carboxylic esters and xanthates, thathave already been referred to (p. 268), can also be considered asclose analogues of retro (4n +2n) cycloaddition reactions.

12.4.2 1,~Dipolar additions

The 47te component in a (47t + 27t) cycloaddition need be neither afour-atom system (as in 1,3-dienes), nor involve carbon atoms only,so long as the HOMO/LUMO symmetry requirements for a concertedpathway can be fulfilled. The most common of these non-dienic 47tesystems involve three atoms, and have one or more dipolar canonicalstructures, e.g. (34a), hence the term-l,3-dipolar addition. They neednot, however, possess a large permanent, i.e. residual, dipole, cfdiazomethane (34a ++ 34b ):

N Ne/ "\.a. a./ "\.e

H 2C N +-+ H 2C N

(34a) (34b)

The initial addition of ozone to alkenes to form molozonides (p. 193)can be regarded as a 1,3-dipolar addition, and many other suchadditions are of great importance in the preparation of fivemembered heterocyclic systems. Thus we have already seen the

R{ycisoidIransoid(28)


:cJ-.=>JRcisoid transoid

350

(29)

Another stereochemical point of significance is that in some DielsAlder reactions there is the possibility of two alternative modes ofaddition, the exo (30) and the endo (31), e.g. with cyclopentadiene (32),and maleic anhydride (33) as dienophile :

electron-withdrawing substituents in the dienophile; the reactiondoes indeed proceed poorly, if at all, in the absence of the latter.The effect of such suhstituents is to lower the energy level of theLUMO in the dienophile and to raise the energy level of theHOMO in the diene, thus enhancing the degree of possible interaction between them. The presence of substituents, and even of heteroatoms, in the system appears not to affect the symmetry of theorbitals involved, however.

Substituents in the diene may also affect the cycloaddition sterically,through influencing the equilibrium proportion of the diene that is inthe required cisoid conformation. Thus bulky I-cis substituents (28)slow the reaction down, whereas bulky 2-substituents (29) speed it up,through this agency:

F\ ~ (32)0(3M "0 0 0'·

yO ~, ""0 0 O~o 0

(33) (30) (33) (31)

Despite the fact that the exo adduct is likely to be the more stable ofthe two thermodynamically, it is often (though not universally)found in Diels-Alder reactions that the endo adduct is the major, ifnot the sole, product. To explain this, it has been suggested that inendo addition stabilisation of the T.S. can occur (and the rate ofreaction thereby speeds up) through secondary interaction of thoselobes of the HOMO in, e.g. (32) and of the LUMO in (33) that arenot themselves involved directly in bond-formation, provided theseare of the same phase. Such interaction would not, of course, bepossible in the T.S. for exo addition because the relevant sets ofcentres in (32) and (33) will now be too far apart from each other;the endo adduct is thus the kinetically controlled product. It issignificant in this connection that the relative proportion of exo

(38)

353

MEt· Me?,

H(41)

12.5 Sigmatropic rearrangements

&l' 0 -Me -+

H

III (40a)

electron configuration will be 'i'fr#~r#L and its HOMO is therefore r#3'This MO can be shown to have terminal lobes of the same phase, sothat overlap between the hydrogen atom's Is orbital and both theterminal lobes of (38)'s MO can be maintained in the T.S. (39)-:-

(39)

Thermal 1,5-hydrogen shifts are thus allowed and, because of thesymmetry of the T.S. (39), the H atom in the product (37, x = I) willbe on the same side of the common plane of the polyene's carbon atomsas it was in the starting material (36, x = I); this is described as a~uprafacial shift. This latter point would not be experimentally verifiable10 the above example, but that thermal 1,5-shifts (which are quitecommon) do involve strictly suprafacial migration has been demonstrated in the compound (40). This is found, on heating, to yield amixture of (41) and (42), which are produced by suprafacial shifts inthe alternative conformations (4Oa) and (40b), respectively:

~M' ~M,-qMe Et 0 Me

(40b) (42)

The terminal lobes of the HOMO will be of the same phase in anonatetraenyl radical also, i.e. for (36, x = 3), and 1,9-shifts (in atOe system overall) should thus be allowed, and suprafacial. Formationof the required to-membered T.S. could present some geometricaldifficulty, however, and it is somewhat doubtful whether any suchconcerted 1,9-shifts have actually been observed. Suprafacial thermalshifts have not been observed in other 'allowed', i.e. (4n + 2)e overall-{36, x = 3,5 ...), systems either.

It is conceivable that the spherically symmetrical Is hydrogenorbital could, alternatively, overlap across the plane of the polyene'scarbon atoms, when the terminal lobes of the latter's HOMO wereopposite in phase-antarafacial overlap. The terminal lobes of theHOMO will be opposite in phase for (36, x = 0,2,4 ... ), leading to a


12.5.1 Hydrogen shifts

Such reactions, in acyclic polyenes, can be generalised in the form:

H HI I

R2C(CH=CH)xCH=CR~ --+ R2C=CH(CH=CH)xCR~

(36) (37)

(34)

12.5 SIGMATROPIC REARRANGEMENTS

The third major category of pericyclic reactions can be looked uponas involving the migration of a (f bond-hence the name-within an-electron framework. The simplest examples involve the migrationof a (f bond that carries a hydrogen atom.

e Ellpreparation of a 1,2,3-triazole from PhN-N=N (p. 194), andanother example involves preparation of the dihydropyrazole (35)from diazomethane (34):

Consideration of the feasibility of these shifts as concerted processes,i.e. via cyclic transition states, requires as usual a consideration of thesymmetry of the orbitals involved. A model related to the transitionstate can be constructed by the device of assuming that the C-H(f bond that is migrating can be broken down into a hydrogen Isorbital and a carbon 2p orbital. For the case where x = I in (36), theT.S. can then be considered as being made up from a pentadienylradical (38), with a hydrogen atom (one electron in a Is orbital)migrating between the terminal carbon atoms of its 5ne system (i.e.a 6e system overall is involved):

By analogy with the categories of pericyclic reactions we havealready considered, the feasibility of the migration will then be decidedby the relative phase of the terminal lobes, i.e. the symmetry, of theHOMO of the pentadienyl radical (38). As this is a 5ne system, its

12.5.2 Carbon shifts

T.S. such as (43) when x = 0 (an overall 4e system):

355

(500)

cis,lrans

(SO)

Me~

Me~


[Me~]*

: : +=tMe~

[

Me J*....~H-- ~Me--

H

meso(49a)

(49)

Me~

Me~

CH2 [HC r HC H2C",0/ ......CH 03- ~H 0

2'::::"CH HO "'CH

6II . :1 rytH' ©rtH,CH 2 6~H' -'-- -- I H

#

(51) (52)

So far as thermal reactions are concerned, those that can proceedvia six-membered transition states go most readily, and are by farthe commonest. That a six-membered cyclic T.S. in the chairconfonnation is commonly preferred is shown by the fact that themeso fonn of (49) yields only (99.7%) the cis, trans forin (50a), outof the three possible geometrical isomerides (cis,cis; cis, trans ; andtrans, trans ) of (50):

~H~Me

H

Cope elimination, p. 268),

and a shift from oxygen to carbon in the Claisen rearrangement ofallyl aryl ethers (51 ~ 52):

CH 2

~ HC J HC H2C.::::,.0/ ......CH 03- '~H 0

2'::::"CH HO CH

6"". :1 atH

' ©itH'

'-':::: CH2 6~H' -'I -- -- I H# #

(5Ia) (53a) (520)

This corresponds to a shift which is suprafacial at both 'ends' of themigrating system.

The Claisen rearrangement is strictly intramolecular, and showsthe large negative value of ~S* characteristic of the degree of orderingrequired by a cyclic T.S. This latter requirement is also borne out by14C labelling, which indicates that the position of the 14C atom in theallyl group is 'inverted' during migration (51a~ 52a):

HOMe

HO(48)

R,tV\n/R' ~ R2C=CH-CR~R~R' I

H H(46) (45)

OH ,.OMe

H '

r ~ ~

(47)

(44)

R2C-CH=CR~ ~

IH


:~::(43)

Such 1,3-shifts are, indeed, found to be relatively common. 1,5-Photochemical shifts in (36, x = 1) should be antarafacial, but this is likelyto involve a strained T.S. and no examples are known. 1,7-Photochemical shifts in (36, x = 2) should be allowed and suprafacial, andthe example (47~ 48) has in fact been observed:

Such a transition state is likely to be highly strained, however, andno such 1,3-antarafacial shifts have actually been observed. A 1,7thermal antarafacial shift in (36, x = 2), where the T.S. is likely to bemuch less strained (i.e. able to adopt the required helical geometry)has, however, been observed in the vitamin D series.

1,3-Photochemical shifts should, however, be allowed and suprafacial(44 --+ 45) as the HOMO of the T.S. (1jJ3, due to ljJiljJi --+ ljJiljJj)now has terminal lobes which are of the same phase (46):

354

The occurrence of a 1,7-photochemical shift of H in this compounddoes not, of itself, establish directly that this shift proceeds via asuprafacial pathway. The relatively rigid cyclic structure of (47)must, however, rule out the possibility of the shift having proceededvia the antarafacial route.

Among the best known examples, involving a carbon moiety, is theshift from one carbon atom to another observed in the Coperearrangement of 1,5-dienes (49 ~ 50; not to be confused with the

Further confirmation of the two-fold shift, and of the double inversionof the position of the 14C label, is provided by 'trapping' (cf. p. 50)the first dienone intermediate (55a) with maleic anhydride in a DielsAlder reaction. An exactly analogous rearrangement is found tooccur in allyl ethers of aliphatic enols, e.g (58):


above, a reaction has been described as symmetry forbidden, thisapplies to the concerted pathway only: it could well be that an energetically feasible, non-concerted pathway is still available, involvingzwitterionic or biradical intermediates. Equally, the statement that areaction is symmetry allowed does not necessarily guarantee that itwill proceed readily in practice: the attainment of the required geometry in the T.S. could well be inhibited by the size of ring required,by the presence of particular substituents, or for other reasons.

oMe~MeV "4fH2

~CH

H2C(55a)

~

[

0 J*'::Me Me'Q,....tcH '!

.: '1 2 ,

H·:...· .CHH;C"

+--

-


R~~ ~[)~~' ~RJ)H H

(58) (59)

The dienone intermediate (53a), as well as enolising to the phenol(52a), is itself capable of undergoing a Cope rearrangement to yielda second dienone (cf 56a), whose enol is the p-substituted phenol (cf57a). Enolisation normally predominates, but where (51) has 0

substituents, i.e. (540), 'o-enolisation' cannot take place, and only thep-phenol (57a) is then obtained. That this product is indeed formednot by direct migration of the allyl group, but by two successive shifts,is suggested by the 'double inversion' of the position of the 14C labelin the allyl group that is found to occur:

This reaction also is concerted and proceeds via a six-memberedtransition state, but here the species (59), corresponding to the eneone intermediate (53a) in the aromatic Claisen rearrangement, is infact the end-product. This is so because there is in (59) no energeticdriving force, comparable to re-aromatisation in (53a -+ 52a), topromote its enolisation.

Finally, it must be emphasised that where, in any of the e1ectrocyclicreactions, cycloadditions or sigmatropic rearrangements considered

M'~M' -. M'«('CH2-CH::bCH2 H CH2-CHd:CH2

(.57a) (500)

8

13.2 First Hammett plots 359

-aGelog K = 2.303 RT

-aG+ [k'T]log k = 2.303 RT+ log h

[k' = Boltzmann's constant]

h = Planck's constant

Equilibrium constants, K, and rate constants, k, are e~ch r~latedto free energy changes (pp. 34, 38) in the relevant reactions 10 thefollowing way:

-logK RC02 H 2

Fig. 13.1

The fact that there is in Fig. 13.1 a straight line relation~hi~be~e~n-log k for reaction of the esters (1), and -log K, for IOnIsation 10

(1)

the rates of reaction were directly related to the .ionisation constants, in water, of the corresponding carboxylic aCids (2):

RC02H + H20 ~ RCO~ + H 30

Ell

(2)

Thus on plotting log k for reaction of the esters (1) against log K forionisation of the acids (2) (he actually plotted the -log val~es so. asto have more easily handled numbers) a reasonable straight hne

resulted (Fig. 13.1):

13.2 FIRST HAMMETT PLOTSThe first such relationship, on a thoroughly established basis, wasobserved by Hammett as long ago as 1933: He showed that for thereaction of a series of methyl esters (1) with NMe3,

RC02Me+ NMe3~RC~ + EIlNMe4

13.1 INTRODUCTION, p. 358.13.2 FIRST HAMME"IT PLaTS, p. 359.13.3 TIlE HAMMEll EQUATION, p. 362:

13.3.1 Derivation of Hammett equation, p. 362; 13.3.2 Substituentconstant, ax, p. 362; 13.3.3 Reaction constant, p, p. 363; 13.3.4Physical significance of ax, p .. 364; 13.3.5 Physical significance ofp, p. 367; 13.3.6 Through-conjugation: ax and ax, p. 368; 13.3.7Yukawa-Tsuno equation, p. 372.

13.4 USES OF HAMME"IT PLaTS, p. 374:13.4.1 Calculation of k and K values, p. 374; 13.4.2 Deviationsfrom straight line plots, p. 375; 13.4.3 Concave upwardsdeviations, p. 375: 13.4.3.1 Acetolysis of 3-aryl-2-butyl brosylates,p. 375; 13.4.3.2 Hydrolysis of ArC02R in 99·9% H 2S04 , p. 378;13.4.4 Concave downwards deviations, p. 380; 13.4.4.1Cyclodehydration of 2-phenyltriarylmethanols, p. 380.

13.5 SIERIC EFFECTS, p. 383:13.5.1 Taft equation, p. 384; 13.5.2 Steric parameters, Es and l),

p.386.13.6 SOLVENT EFFECTS, p. 388:

13.6.1 Change of p with solvent, p. 388; 13.6.2 GrunwaldWinstein equation, p. 389; 13.6.3 Dimroth's Er parameter, p. 391.

13.7 SPECIROSCOPIC CORRELATIONS, p. 392.13.8 TIlERMODYNAMIC IMPLICATIONS, p. 394.

Linear free energy relationships

358

In previous chapters we have considered the relative reactivity ofnumerous series of compounds in specific reactions-such as nucleophilic displacement by EtOe in the series of bromoalkanes below(ct. p. 86)-

and have sought to account for the reactivity sequences observed interms of the operation of electronic and sterk effects. This hasproved a useful and rewarding exercise, but a major disadvantage ofsuch studies, and explanations, is that they remain qualitative: whatis still needed is a method for relating structure and reactivity on aquantitative basis.

13.1 INTRODUcrtON

13

361

fB,e'lklimiting

aliphatic ester

OH

.-- 1_0Et~~

a-substituent

log KArC02H

Fig. 13.3

13.2 First Hammett plots

13.3, the m- or p-substituent in (3a) is far removed from thereaction centre and, in this rigid molecule, can exert no steric effectupon it. By contrast, the o-substituent in (3b) is close at hand (ct.. p.242) and leads to increasing crowding in the transition state leadlOgto the tetrahedral intermediate produced in slow, rate-limiting attack on ester (3b) by eOH; very much the same is true also for the

A reason for such non-conformity on the part of o-substitutedbenzene, and of aliphatic, derivatives is not far to seek. Thus for thebase-catalysed hydrolysis (p. 238) of the esters (3) in Figs 13.2 and

00-N02C6 H4oO-Cl C6H4

Fig. 13.2

etha~oic

Linear free energy relationships360

water of the corresponding carboxylic acids (2), implies that there isalso a straight line relationship between ~G+, the free energy ofactivation for the ester reaction, and ~Ge, the standard free energychange for ionisation in water of the acids. Because of this straightline relationship between the free energy terms for these twodifferent reaction series, straight line plots like the one in Fig. 13.1are generally referred to as linear free energy relationships.

Another early example of Hammett's is shown in Fig. 13.2, whichrepresents a plot of log k for base-catalysed hydrolysis of a group ofethyl esters (3) against log K for ionisation in water of the corresponding carboxylic acids (2). Judged

RC02Et + SOH~RCO~+EtOH

(3)

by the standards of Fig. 13.1, the plot in Fig. 13.2 is prettydisappointing: there is a straight line relationship for benzoic acidand its p-Me and p-N02 derivatives, but the o-N02 and o-C]benzoic acid derivatives then lie far off to one side of this straightline, while the aliphatic derivatives, of ethanoic and 2-hydroxypropanoic acids, lie far off to the other side. Hammettfound indeed that straight lines were not generally obtained ifreaction data for either o-substituted benzene derivatives, or aliphatic species, were included in the plot. He did, however, find that ifconsideration was restricted to reactions of m- and p-substitutedbenzene derivatives, then-as shown for ester hydrolysis in Fig. 13.3(p. 361)-excellent linearity resulted, and this held for a very widerange of different reactions of such derivatives.

more flexible molecules of the aliphatic ester (3c). Such steric effectswill be much smaller, if indeed apparent at all, in the removal of theperipheral H from the C02H group by H 20 (i.e., in acid ionisation).

[6]

36313.3.3 Reaction constant, p

Substituent, X um_x Up_x

Me3C -0,10 -0,20Me -0,07 -0·17H 0 0 (by definition)

MeO +0·12 -0,27

HO +0·12 -0·37F +0·34 +0·06a +0·37 +0·23MeCO +0·38 +0·50Br +0·39 +0·23CN +0·56 +0·66N02 +0·71 +0·78

Hardly surprisingly, the value of CTx for a particular substituent isfound to depend on the location of the substituent, having adifferent value in the m-position from that in the p-position.

(4)

13.3.3 Reaction constant, p

Having thus obtained a range of substituent constant, CTx, values it isnow possible to use them to calculate the value of p, the reactionconstant, in [6] for any further reactions in which we may beinterested: this is often done graphically. Thus to evaluate p for, say,the base-catalysed hydrolysis of m- and p-substituted ethyl 2arylethanoates (4) we would, from kinetic measurements (or from

which is the usual form of what has come to be called the Hammettequation.

By using known values of Kx (or pKa ) for aqueous ionisation ofm- and p-substituted benzoic acids (or measuring Kx [pKa ] wherethe value is not already available for a particular m- or psubstituent) it is possible to calculate CTx as required, and a selectionof values obtained in this way is shown below:

where CTx is a substituent constant, whose value will remain constantfor a specific substituent in a specific position (m - or p-), irrespective of the nature of the particular reaction in which a benzenederivative, carrying this substituent, is involved.

Substituting [5] into [4] we then get,

[4]

[2]

[3]

[5]*


13.3.2 Substituent constant, O'xHammett then designated the ionisation, in water at 25°, of m- andp-substituted benzoic acids as his standard reference reaction. Hechose this reaction because reasonably precise aqueous ionisationconstant, !<x, data were already available in the literature for quite arange ofddre~entlym- ~nd p-substituted benzoic acids. Knowing XH andXx for a vanety of differently X-substituted benzoic acids it is thenpossible to define a quantity, ux, as '

Kxux=logK H

-[5]. may, of course, also be written in the form, U x = pKa(H) - pKa(x); so that thenumencal value of U x for a particular substituent is obtained by simple subtraction ofthe pKa value for the substituted acid (where this is known) from the pK value forbenzoic acid itself. a

13.3 THE HAMMETI EQUATION

Desp.ite establishing such linear relationships for a wide range ofreactions of m- and p-substituted benzene derivatives we still lackany simple form of this quantitative relationship that c~n actually beused to investigate new situations: here again, it was Hammett whosupplied the answer.

Subtracting [2] from [1], we obtain,

log kx -log kH = p(log Kx -log KH)

which may also be written in the form:

13.3.1 Derivation of Hammett equationThe .general equation for a straight line is y = mx +c, and this can beapplIed to the straight line in Fig. 13.3 to give,

10gkx=plogKx+c [1]

whe~e p is the slope of this straight line, c the intercept, and X is theparticular m- or p-substituent in the benzene ring of the speciesconc~rned. I~ is also possible to write an exactly analogous equationthat IS restncted to the unsubstituted ester and acid, i.e. whereX=H:

The standard reaction, the aqueous ionisation of m- and psubstituted benzoic acids at 25°, will have a p value of 1·00 as anecessary concomitant of the definition of Ux in [5], and its use in[6]. The value of the reaction constant, p, for a particular reaction,carried out under specified conditions, remains constant no matterwhat the m- or p-substituents present in the compounds involved.

365

OHI

e(»-C-OEt

~~o,tetrahedral

intermediate

tetrahedralintermediate

~=0'66H

transitionstate

3-

t. <)H OEt..."",.~

~NO'transition

state

13.3.4 Physical significance of Ux

(J'.-MeO = -0'27

(J'm-No, = +0·71

(6)

3- c...eoH-e,~ /OEt

ut k-._

~~Me

COH

3-0 OEtt...-~c/

~o3+ ~~ slow

N02(5)

The m-nitro ester (5), with Um-NO, = +0·71, is hydrolysed 63·5times as fast as the unsubstituted ester (powerful electronwithdrawal markedly assisting SOH attack on the carbonyl carbonatom, and stabilising the transition state leading to the negativ~lycharged tetrahedral intermediate); while the m-Me ester (6): wIthU = -0·07, is hydrolysed 0·66 times as fast as the unsubstItutede;t~~(veryweakelectron-donation slightly inhibiting s0I:I a~tack,etc.).

If we now look at the list of U poX values (p. 363), It IS apparentthat not only does the U poX value for a particular substit~ent, ~,vary in magnitude from the U m-X value fo~ the same substItuent, Itmay differ in sign too: as is the case wIth m- and p-MeQ. Anexamination of the effect of a m-MeO (7) and a p-MeO (8)substituent on the same reaction as above (base-catalysed ester

would thus seem that Um-x represents, both in direction and magnitude, a measure of the total polar effect exerted by the substituentX on the reaction centre.

This is borne out by a comparison of the rates of base-catalysedhydrolysis (cf. p. 238) of m-NOz (5), an~ of m-Me (6), subst~tut~dethyl !;lenzoates with that of the unsubstItuted ester: a reactIOn mwhich the slow, and hence rate-limiting, step is initial attack on theester by SOH (p. 239):

p

-3'19-2,69-1,88-0·99-0,09+0·03+0·47+0·79+0·82

+ 1'00 (standard reaction)+2·01+2·14+2·51+2·73


Reaction Type

(1) ArNH2 with 2,4-(N02)2C6H3CI in EtOH(25°) k(2) ArNH2 with PhCOCI in C6H6(25°) k(3) ArCH2C1 solvolysis in aq. Me2CO(69'8°) k(4) ArOe with EtI in EtOH(25°) k(5) ArC02H with MeOH (acid-catalysed, 25°) k(6) ArC02Me hydrolysis (acid) in aq. MeOH(25°) k(7) ArCH2C02H ionisation in H20(25°) K(8) ArCH2CI with J8 in Me2CO(200) k(9) ArCH2C02Et hydrolysis (base) in k

aq. Et(OH(300)(10) ArC02H ionisation in H20(25°) K(11) ArOH ionisation in H20(25°) K(12) ArCN with H2S (base) in EtOH(60·6°) k(13) ArCO~Et hydrolysis (base) in aq. EtOH(25°) k(14) ArNH3 ionisation in H20(25°) K

the literature if we're lucky!), obtain kH for the unsubstituted ester,and kx for at least three different substituted esters. Knowing thevalue of Ux for each of these substituents, we can then plotlog (kx/kH) against Ux and, from [6], the slope of the resultingstraight line will be the value of p for this reaction: it turns out to be+0·82 for this particular hydrolysis, when carried out in aqueousethanol at 30°. The p values for quite a wide range of differentreactions of m- and p-substituted benzene derivatives are shownbelow:

13.3.4 Physical significance of ax

Before we can go on to consider the actual use that may be made ofHammett plots, it is necessary to provide some physical justificationfor Ux and p in terms of the more familiar factors that we havealready seen influencing reaction rates and equilibria.

If we consider Ux, the substituent constant, first and look at thelist of Um-x values (p. 363), we can see that m-Me3C and m-Meeach have a small -ve value, H has the value-by definition---ofzero, while all the other m-substituents have (increasing) +ve values. The change in sign (-ve- +ve) does, of course, parallel thechange in direction (electron-donating - electron-withdrawing) ofthe inductive effect exerted by these substituents. The substituentsmay also exert a field effect (p. 22), operating through the medium,but this will act in the same direction as the inductive effect. It

36713.3.5 Physical significance of p

This has p value of +2-51, the known slow, rate-limiting step in thisreaction is attended by the development of -ve charge adjacent tothe reaction centre in the transition state leading to the intermediate(12), and the overall reaction is, as we have already seen (p. 365),

X

(11)

Ph Ph Ph Ph

I n al I HN-tJ()e HN-t=o

~r~~F~~pQl~~(9) (10)

The slow, rate-limiting step of this reaction is found to be initialattack by the electron pair of the nitrogen atom of the substitutedaniline (9) on the carbonyl carbon atom of the acid chloride. Thisresults in the development of +ve charge at the reaction centre-theN atom attached directly to the substituted benzene ring in theforming intermediate (10). The reaction is thus accelerated byelectron-donating substituents, which help delocalise this forming+ve charge in the transition state leading to the intermediate (10),and correspondingly retarded by electron-withdrawing substituents;this behaviour is found to hold in general for reactions with -ve p

values.We have already had some discussion of a reaction with a +ve p

value, reaction 13 in the list (p. 364), the base-catalysed hydrolysisof m- and p-substituted ethyl benzoates (11):

13.3.5 Physical significance of pNow let us consider p, the reaction constant. Looking at the list of

p values (p. 364), we can select first a reaction with a sizeable -ve pvalue, say reaction 2-the benzoylation of m- and p-substitutedanilines (9)-with p = -2-69, and look at this reaction rather moreclosely:

?HeoZOEt

~lQlOMe

tetrahedralintennediate

OH

~ "",tOE!WCOMe

tetrahedralintennediate

transitionstate

transitionstate

(7)


3- <:OHt~ /OEt

~~

(8)

366

hydrolysis) makes plain the reason for this change in sign:

3- C:0H3- ~

~Et ~,~t~Me OMe

In the m-position, the electronegative oxygen atom of the MeOgroup exerts an electron-withdrawing inductive effect (u -M =+0·12) and hydrolysis is faster than with the unsubstituted e~ter[cf_the m-N02 ester (5)]. In the p-position, MeO will still exert an~lectron-withm:awing inductive effect, but in addition it can, throughIts el~ctron pans, exert an electron-donating mesomeric effect onthe nng carbon atom to which the C02Et group is attached. Thelatter effect, bec~use it involves the more readily polarisable 1T

electron system, IS the greater of the two, and the overall result istherefore ne~ electron-donation (Up-Meo = -0-27); as is required bythe observatIon that the p-MeO ester is hydrolysed markedly moreslowly than the unsubstituted compound (cf. p. 154).

Thus Ux can be regarded as a measure of the overall polar effectexert~d b~ a substituent, X, on the reaction centre. Its sign indicatesthe drrectton (-ve = electron-donating; +ve = electron-withdrawing)and its magnitude the extent, of the effect that X exerts-<:ompared'of course, with the effect exerted by H. Indeed, the assumedconst.ancy of a substituent's Ux value, over a wide range of differentreactIons, does ?ot necessarily imply that the absolute polar effect ofX always remams constant, but only that its effect relative to Hremains constant.

369

log~ CDFig. 13.4

p-CN.

13.3.6 Through-eonjugation: ax and ai

(18a) (I8b)

£~~ ~°

+H20. I+---+ ¢ + H

3O$

t ~/$ ecf'~~ ~~~oe

<>0 °(19a) (19b)

Two substituents however, the powerfully electron-withdrawing pN0

2and p-CN, lie above this straight line: indicating that p-N02

phenol and p-CN phenol are in fact stronger acids than we ~ight

have expected them to be. Why this is so becomes apparent If wewrite out the structures of the species involved in both ionisationequilibria for, say, the p-N02 compounds (18 and 19) and examinethe polar, electronic effects that can operate in them:

,/OH ~,;YO

6) + H 20 +====t ~ + H 3cP

~ ~~ eif~

CD XC6H4C02H +H20 +=t XC6H4C~ +H3 <J!l(13)

(2) XC6H40H +H20 +=t XC6H400 +H30 Ell

(17)

obtained for a wide range of different substituents:

p

1·00 (standard reaction)0·490·210·47

Acid ionisation (H20)

(13) XC6H4C02H(14) XC6H4CH2C02H(15) XC6H4CH2CH2C02H(16) XC6H4CH=CHC02H


Introduction of first one, and then two, CH2 groups between the "benzene ring and C02H progressively reduces the susceptibility ofthe acid's ionisation to the polar effect of the substituent X in thebenzene ring. The susceptibility, as revealed by the value of p, risesagain for (16), however, as CH=CH is a markedly better transmitter of electronic effects than is CH2-CH2 •

accelerated by electron-withdrawing, and retarded by electrondonating, substituents.

Thus p can be regarded as a measure of the susceptibility of areaction to the electron-donating or- withdrawing effect exerted by asubstituent X; relative, of course, to the susceptibility (towards sucha substituent) of the standard reaction-the aqueous dissociation ofm- and p-substituted benzoic acids at 25°-for which p = + 1,00, bydefinition. The sign of p is of diagnostic value, as we have seen, inthat a -ve value indicates the development of +ve charge (or, ofcourse, the disappearance of -ve charge) at the reaction centre (during formation of the T.S. in the rate-limiting step of the overallreaction; while, vice versa, a +ve value indicates the development of-ve charge (or the disappearance of +ve charge) at that centre. Themagnitude of p can be regarded, therefore, as a measure of thechange in charge density at the reaction centre during formation ofthe T.S., or on proceeding from one side of an equilibrium to theother.

On this basis, it might well be expected that the p value, ofotherwise similar reactions, would decrease as the reaction centre ismoved further away from the substituents that are exerting a polar,electronic effect upon it. This is borne out by the p values for theaqueous ionisation of the acids (13)-(16):

13.3.6 Through-conjugation: a; and a;Before we go on to consider the major uses of ax and p, it is firstnecessary to take a little closer look at just how constant the axvalue for a particular substituent really is. If we plot data for theaqueous ionisation of m- and p-substituted benzoic acids (B)-thestandard reaction-against that for ionisation of the correspondingsubstituted phenols (17), a very reasonable straight line (Fig. 13.4) is

371

Me;~M¢

H~ HQl

CH2

p-Me(21b)

13.3.6 Through-eonjugation: ax and a;

(20)

M~e2C-C1 S"I

~slow

X

M~~MQCOMe 6lQMe

p-MeO (21a)

The fact that development of +ve charge, in the transition state for

op-MeO

t1x

Fig. 13.5

Solvolysis of the v-MeO and v-Me chlorides is fo~nd to be fast~r(p-MeO=800 times) than would have bee? pr~dlcted from therr0" values. This stems from the stabilIsation, by through~~jugation, of the carbocationic interrnediat.es. ~21a and 21b)which are developing during the slow, rate-Imutmg step of theoverall reaction:

tertiary halides, 2-aryl-2-chloropropanes (20), shown in Fig. 13.5:

M~C-OH+HQl

M~>::5?~~

X MeC=CH2

(21) ~carbocationintermediate

X

Substituent, X a;_x V"p_x

C02Et 0·68 0·45COMe 0·84 0·50CN 0·88 0·66CHO 1·03 0·43N02 1·27 0·78


An exactly analogous situation will arise where there is thepossibility of direct through-conjugation between a suitable'electron-donating p-substituent and a reaction centre at which +vecharge is developing. A good example is solvolysis (SN 1) of the

For each species, the inductive effect of the p-N02 substituentwhich will be essentially similar in each of the sets of species-hasbeen omitted, but the mesomeric or conjugative effect has beenincluded. In (18a) ~ (18b), the standard reaction that was used toevaluate 0"p-N0

2' the conjugative effect of the p-N02 substituent is I

transmitted ultimately to the reaction centre only through an inductive effect: operating on the C02H, or COE£, group from the ringcarbon atom to which it is attached. In (19a)~(19b), however, theconjugative effect can be transmitted right through from the p-N02substituent to the electron pairs on the oxygen atom which is nowthe reaction centre. This effect will be particularly marked in (19b),where the anion will be stabilised substantially by delocalisation ofits -ve charge, and the ionisation equilibrium for p-N02 phenolthereby displaced over towards the right in the anion's favour; thusincreasing this phenol's strength as an acid.

The value for 0" p-No, obtained from the standard reaction (18a~18b) dearly does not take into account the heightened effect of this'through-conjugation', which is why the point for p-N02-and forp-CN-is off the line in Fig. 13.4. Such through-conjugation can,however, be allowed for by using the aqueous ionisation of phenolsto establish a set of new, alternative, 0" values, for p-N02 and othercomparable electron-withdrawing substituents: these new values maythen be used for reactions in which through-conjugation can occur.

This can be achieved by first plotting log Kx/KH against o"x forrn-substituted phenols only (which cannot be involved in throughconjugation), then the slope of the resulting straight line will givethe value of p, the reaction constant, for this reaction. Using thisvalue in the normal Hammett equation ([6], p. 363), enables us tocalculate the new, revised, O"p_ value for p-N02 , and for similarsubstituents capable of through-conjugation. These revised figuresare generally referred to as 0";_ values, and a number are comparedwith the normal O"p_ values below:

this slow step, is substantial is borne out by the large -ve p value,-4'54, for the reaction. By using this solvolysis as a new standardreaction, it is possible, as with 0';, to obtain in an exactly analogousmanner a set of 0'; values that make allowance for throughconjugation by powerful electron-donating p-substituents; a numberof these revised figures are compared with the corresponding O'p_values below:

[8]

[6]

373

918--.M~cCa++

P' p= -4'54

~ I r=1'00(by definition)

COMe

(24)

k xlogk=PU;(

H

13.3.7 Yukawa-Tsuno equation

To evaluate r for other reactions, we can obtain p for the reactionby measuring k x values for m-substituted compounds only, andthen measure k x for p-substituted compounds where the values ofO'p_x and O';_x, or O';_x, are already known. Using [7], r can then beevaluated by calculation, or by graphical methods. Thus for thebase-catalysed hydrolysis of p-substituted phenoxytriethylsilanes

which is reasonable enough as it was this reaction that we used (p_371) to define O':;{ in the first place, for electron-donating psubstituents capable of considerable through-conjugation! Similarly,for a reaction in which no through-conjugation occurs r will be zero,and [7] will then, of course, simplify to the original, simple Hammettequation r6]:

applicable to electron-donating p-substituents; for electronwithdrawing p-substituents O':;{ would, of course, be replaced by O'x.The new parameter, r, intended as a measure of the throughconjugation operating in a particular reaction, is given the value of1·00 for solvolysis of the tertiary halides, 2-aryl-2-chloropropanes(20)_ For this reaction [7] does, of course, then simplify to [8],

H<;>"-,$iEt3,

~P'+3."¥r=0-50

N/EII~eo 0

(23)

(22), the value of r is found to be 0·50_ This extent of throughconjugation-by a substituent such as p-NOz-suggests the development of substantial -ve charge (p=+3'52) in the transitionstate (23) tor the rate-limiting step. This will not, however, be so faradvanced as the development of + ve charge (p = -4-54) in thetransition state (24) for the standard reaction, halide solvolysis,where r = 1·00. As, in each case, the development of charge in thetransition state goes hand-in-hand with bond-breaking between the

OSiEt3 0 6

¢~~}a'~OHX X

(22)

[7]

Substituent, X a-;_x Up

_xC6Hs -0,18 -0·01Me -0,31 -0'17MeO -0-78 -0,27NH2 -1-30 -0-66NMe2 -1-70 -0-83


So for each p-substituent we now have available two, alternative,substituent constants-O';_x and 0'p-x for electron-withdrawing substituents or 0';-x and 0'pox for electron-donating substituentswhose use depends on whether through-conjugation between psubstituent and reaction centre does, or does not, take place in aparticular reaction_ It would be nice to think that these dualsubstituent constant values would now take care of all eventualities,and an analysis was therefore made of no less than eighty differentreactions to see whether use of O';_x or 0'poX' and O';_x or 0'poX,would lead to straight line plots in all cases. In fact, it was found thatthe values required for, say, p-NOz did not cluster round either0'78(0') or 1'27(0'-), but were spread more or less evenly throughout the range between these two, limiting values; and similarly forv-MeO. between -0'27(0') and -0'78(0'+).

On reflection, this is seen to be hardly surprising. The extent ofthe change in electron density at the reaction centre-an atomattached directly to the benzene ring in such reactions-during theslow, rate-limiting step will obviously differ from one reaction toanother. So too, therefore, will the degree of response (via throughconjugation) elicited from the same p-substituent towards differingreactions_ Hence the apparent need for a range of different O'p_xvalues for a particular p-substituent, reflecting the differing degreesof through-conjugation elicited from it by different reactions.

13.3.7 YokawB-Tsono eqoatiOD

There have been a number of attempts, by the introduction of afurther parameter into the Hammett equation, to quantify thisgraded response-via through-conjugation-<>n the part of a psubstituent. Among the best known of these is the Yukawa-Tsunoequation, [7], which, in the form shown here, is

13.4 USES OF HAMMETf PLOTS

37513.4.3.1 Acetolysis of 3-aryl-2-butyl brosylates

Fig. 13.6

13.4.3 Concave upwards deviations

13.4.3.1 Acetolysis of 3-aryl-2-butyl brosylatesAn interesting case in point is the acetolysis of 3-aryl-2-butyl pbromobenzenesulphonates or brosylates (25), for which the Hammett plot is shown in Fig. 13.6. The lower right-hand side of the

13.4.2 Deviations from straight fine plotsWe have already seen (p. 368) how the sign and magnitude of p, thereaction constant, can provide useful information about the development (or dissipation) of charge (+ve or -ve) on going fromstarting materials to the transition state for the rate-limiting step ofa reaction. We have also seen (p. 369) how deviations from straightline plots using normal u x , substituent constant, values led to thedefinition of u;c or Ux values to take into account throughconjugation between certain p-substituents and the reaction centre.The need to use other than the normal Ux values indicates theoccurrence of such through-conjugation in a particular reaction, andthe Yukawa-Tsuno parameter, r, then provides a measure of itsextent.

Paradoxically, Hammett plots are usually most informative at thevery point at which they depart from linearity: but the m~jorinference that can be drawn from this departure IS found to differdepending on whether the deviation is concave 'upwards' or concave'downwards'.

~NO:;_,X~ ~MeCH-CHMe MeCH-CHMe OIl

I 1.!2OBs OAc

(25) (26)

[Bs = p-BrC6H4S02]

brosylate

[6bJ

63·5log-=pxO·71 :. p=2·S4

1

k m-No,log k = PUm-No, [6aJ

H


i.e.

reaction centre and the leaving group, the magnitude of r canperhaps be construed as some indication of the extent of suchbond-breaking by the time the transition state has been reached.

It is, however, important to resist the temptation to introduce newparameters into the Hammett equation merely to achieve a better'fit' with the experimental data. This is particularly true where, as insome cases, it may be difficult to ascribe real significance, in physicalterms, to the new parameter anyway. It is in fact possible, as weshall see, to obtain much highly relevant information about reactionpathways using simple Hammett plots only.

i.e.

log k p _MeO _k

H-PUp-MeO

k MeO k MeOlog-P---=2·S4x-0·27 .. -P---=0·21kH kH

When k p _MeO subsequently came to be determined experimentally, kp_MeO/kH was indeed found to be 0,21, so the calculatedvalue was pretty satisfactory! In fact, Ux and p values are rarelyused for such a purpose, they are employed much more often inproviding salient data about reaction pathways.

Having now given some consideration to the significance that can beattached to Ux and p in more familiar physical terms, it is possibleto go on and discuss the actual uses that can be made of them inproviding information about reactions and the pathways by whichthey take place.

13.4.1 Calculation of k and K valuesThe simplest possible use that can be made of the Hammettequation is to calculate k or K for a specific reaction of a specificcompound, where this information is not available in the literature,or indeed where the actual compound has not even been preparedyet. Thus it is known that the base-catalysed hydrolysis of ethylm-nitrobenzoate is 63·5 times as fast as the hydrolysis of thecorresponding unsubstituted ester under parallel conditions; whatthen will be the comparable rate for base-catalysed hydrolysis ofethyl p-methoxybenzoate under the same conditions? Looking atthe table of Ux values (p. 363), we find that u m-NOz = 0'71, while 'u p-MeO = -0,27. Then from the Hammett equation [6] (p. 363):

plot-where the substituents are powerfully electron-withdrawingis a straight line whose slope indicates a p value for the reaction of-1-46. On moving across to the left-as the substituents becomeless electron-withdrawing-the plot now curves upwards, indicatingthat the rate of acetolysis of these species is faster than we wouldhave expected it to be on the basis of the O"x values for theirsubstituents.

What we might expect as a pathway for this reaction would besimple SN2 displacement (p. 98) of the good leaving group-brosylate anion-by acetate anion:

377

--:-,....,-.--+ ¢'MeCH-yHMe

OAc

(30)

~ ?A'+ x~ ?A'~H ~Me

Me H Me H

(32) (33)erythro acetolysis threo acetolysis

product product

~MeCH-CHMe

t90Ac

(29)cyclic phenoniumion intermediate

slow

III

/:& X~eHinternal t ...............nuc\eop/lile , .

slow

OAc

Me H e)'H Me H

~Ac (33)cyclic phenonium threo acetolysision intermediate product

13.4. 3.1 Acetolysis of 3-aryl-2-butyl brosylates

Q'MeCH-CHMe

(6BS(28)

Xh %/H

~BSMe H

(31)

threo brosylate

III

x[()] M<0~

~BSMe H

(31)threo brosylate

The two, alternative, acetolysis products (32 and 33), beingdiastereoisomers not mirror images, may then be separated, or their

to yield the normal acetolysis product (30):

Support for the suggestion that Fig. 13.6 involves a change inactual reaction pathway is provided by acetolysis of the threodiastereoisomer (31) of the brosylate. Acetolysis leads to two different distinguishable, diastereoisomers whose relative proportion willdepend on how much of the total reaction proceeds by extemalnucleophilic attack via the SN2 pathway (erythro product, 32), andhow much by internal nucleophilic attack via a cyclic phenoniumion intermediate (threo product, 33):(26)

x~ OA,

~IMeCH-CHMe

+eaBs

---+

transition statf'

X

~'~A'MeCH-CHMe,,

(27) 6-0Bs


x

~tC;;MeCH---eHMe

(6Bs(25)

376

The smallish -ve p value (-1'46) is compatible with such a path-;way, given that in the transition state (27) breaking of the C-OBs'{bond, is somewhat more fully advanced than formation of the Aco-C 'bond, resulting in the transient development of a 'small amount of';;+ve charge at the reaction centre. This is in no sense unreasonable):with (a) a secondary carbon atom as reaction centre (cf. p. 82), and)!(b) so good a leaving group (cf. p. 98); this pathway would be:increasingly aided, albeit weakly, as the substituent X becomes lesslelectron-withdrawing, i.e. the rate of acetolysis might be expected toli~crease, gradually and linearly, from right to left across the plot in;Fig. 13.6, .

To account for the departure from linearity, as X becomes moreelectron-donating, it would seem that the substituted benzene ring ..must gradually become capable of exerting some more direct effect:on the reaction centre in (25) than it does in the SN2 pathway. It is;~significant in this respect that increasing electron-donation by X willincrease the nucleophilicity of the substituted benzene ring itself,thereby enabling it to function-in competition with eOAc-as a I

neighbouring group (p. 93) or 'internal' nucleophile, e.g. when !'

X = Mea (28). This alternative reaction pathway would then involveslow, rate-limiting formation of the cyclic phenonium ion intermediate (29, cf. p. 105), followed by its rapid ring-opening by eOAc

relative yields estimated by spectroscopic methods. It is found thatthe yield of threo product (33) varies considerably as the nature ofX, the substituent in the benzene ring, is changed:

379

\·40·7(Ix

Fig. 13.7

0·0

The same AAc 1 pathway must also be operating initially for the ethylesters (34b), on the left-hand side of Fig. 13.7, as the p value(-3'25) for this reaction is the same as that for the methyl esters(34a). As the substituent in the benzene ring becomes more stronglyelectron-withdrawing, however, a sharp change in curvature is observed with the ethyl esters to a new straight line with a p value of+2.0. This now +ve p value requires a slow, rate-limiting step forhydrolysis in which +ve charge is decreased at the reaction centrethe overall reaction being increasingly accelerated as the substituentin the ring becomes more electron-withdrawing.

There is indeed yet another pathway for acid-catalysed ester

result of which is to make the concentration of water available forhydrolysis very low.

We have, however, already seen an alternative acid hydrolysispathway (AAc 1, p. 242) in which a water molecule is not involved inthe slow, rate-limiting step. In addition, this step is one in whichconsiderable +ve charge is developed at the reaction centre as theprotonated ester (35a) is converted into the acyl cation intermediate(36a): a necessary requirement for a reaction with a large -ve(-3'25) p value:

13.4.3.2 Hydrolysis of ArC02 R in 99·9% H2 S04

©:~OX(34a,b)

[R=Me,Et]

* = percentage of reactionproceeding viainternalnucleophilicattack

Yield of threoproduct* (33)

100886859391261

SubstifULnt, X

v-Meav-Mem-MeHV-CIm-CIm-CF,V-NO;


When X = p-MeO, th~ most electron-donating substituent at the top~eft-hand corner ~f. Fig. 13.6, 100% of acetolysis is proceeding viamte~nal nucleophlhc .attac.k by p-MeOC6H4 ; when X= m-CI, justcommg on to the straight hne part of the plot in Fig. 13.6, only 12%of the total reaction is proceeding via the internal route; while whenX = p-N02 , the most electron-withdrawing substituent, only 1% ofthe total reaction is now proceeding by this route.

When a simple Hammett plat exhibits an upward deviation, i.e. isconcave upwards as in Fig. 13.6, then this can usually be taken asevidence of a change in overall reaction pathway, as the nature ofthe substituent is varied. That a change in reaction pathway shouldlead t? an upwa~d deviation is reasonable enough: in Fig. 13.6,there IS, at the 'p?~nt where departure from linearity occurs, nothingto prevent the mltIal SN 2 pathway from continuing to operate (alongthe dotted extrapolation). Any change to a new pathway must offera less ~emanding, and hence faster (necessarily upward-curving),alternative or, of course, the initial pathway would continue toprevail and no departure from the original straight line would thenbe observed.

13.4.3.2 Hydrolysis of ArC02 R in 99·9% H2S04

Sometimes departure from the straight line is considerably moreabrupt than in Fig. 13.6; a particularly good example is the hydrolysis, in 99·9% H2S04 , of the substituted methyl (34a), and ethyl(34b), benzoates shown in Fig. 13.7 (p. 379).

Considering first the simpler of the two cases, the straight line forthe m~t~yl esters (34a) which has a p value of -3,25. From this pvalue It IS apparent that this reaction cannot be proceeding via thenormal (AAc2) pathway (p. 241) for acid-catalysed ester hydrolysiswhich, as we know (reaction 6, p. 364), has a p value of +0,03. Thatvalue refers, however, to hydrolysis being carried out with dilutesulphuric acid, while here 99·9% sulphuric acid is being used: one

hydrolysis (AAL 1, p. 241) that would fulfil this requirement:

381

Fig. 13.H

• p = +2·67

,Z = MeO

(OKO) * (0(9~<QC9C-OH C ~H C/ "'.. / '\H / "-AI AT Ar AT Ar Ar

(38) (40)

j~

QOO)4Q);)/C"" /C",- H

Ar Ar Ar Ar

(39)

13.4.4.1 Cyclodehydration of 2-phenyltriarylmethanols

(0(9

~<o<X Z

(38)

(0()Q>)Q5C<o<

X Z

(39)

substitution on the 2-phenyl nucleus to yield the product tetraarylmethane (39):

The question then arises-which step in the overall reaction islikely to be the slow, and hence rate-limiting, one? It's unlikely to be

(37b)


Ar-C=OI6l~H2Me

H

(35b)

380

Loss of MeC~ the ethyl cation (37b), leads to a marked decreasein +ve charge adjacent to the reaction centre (had it actually beenfrom the reaction centre itself the +ve value of p would have beenmuch larger); this carbocationic intermediate (37 b) will then reactrapidly with any available water to yield ethanol.

The question does then arise, given the observed shift in reactionpathway for the ethyl esters (34b), why does a similar shift not occurwith the corresponding methyl esters (34a)? Such a shift would, ofcourse, necessitate the formation of a methyl, C~, rather than anethyl, MeCH~ (37b), cation in the slow, rate-limiting step. C~ isknown to be considerably more difficult to form than is MeC~and this difference is apparently great enough to rule out, onenergetic grounds, such an AAC 1~ AAL 1 shift with the methylesters, despite potential assistance (to AAL 1) from increasinglyelectron-withdrawing substituents.

13.4.4 Concave downwards deviationsThere are, however, also examples of deviations from simple Hammett plots in which the curvature is in the opposite direction,concave downwards, and these deviations have a rather differentsignificance.

13.4.4.1 Cyclodehyclration of 2-.phenyltriarylmethanolsA good example is the cyclodehydration of some substituted 2phenyltriarylmethanols (38), in 80% aqueous ethanoic acid containing 4% H2S04 at 25°, to yield the corresponding tetraarylmethanes(39), as shown in Fig. 13.8 (p. 381).

Two of the benzene rings in (38) each carry a p-substituent (Xand Z, respectively), and the value of u actually plotted is ~u: thesum of the u values for X and Z. The plot in Fig. 13.8-of log kobo

for the reaction against Iu-is clearly a composite of two straightlines, one on the left with p = +2,67, and one on the right withp = -2'51.

There seems little doubt that the overall reaction follows afour-step pathway, the first two steps constituting an E1 (p. 247)elimination of water to yield a carbocationic intermediate (40),which then, in the last two steps, effects internal electrophilic

38313.5 Steric effects

13.5 STERle EFFECfS

Quite early on (p. 361) in this discussion of linear free energyrelationships consideration was restricted to the side-chain reactionsof m- and p-substituted benzene derivatives. The reactions ofo-substituted benzene derivatives, and indeed of aliphatic compounds, were excluded because of the operation of steric and othereffects, which led to non-linear, or even to apparently random,plots. . .

The success and utility of Hammett plots, and the realIsation thatthey are often of most value diagnostically when they do indeeddiverge from linearity, has emboldened a nu~ber of workers toseek, with suitable modifications, to extend theu scope to a muchwider range of compounds. The most general and successful of theseextensions was proposed by Taft.

Rate-limiting step Q) is thus speeded-up, and the overall reactionrate therefore increases, i.e. the slope of the plot is upwards fromleft to right (p is +ve). Also on moving from left to right, decreasingthrough-conjugation, as the substituents become less electrondonating, makes carbocation fonnation more difficult; thus step~ isbeing slowed down as step Q) is being speeded-up. There must,therefore, come a point at which speeding-up step Q) catches upwith the slowing-down step ~; any further decrease in electrondonation by the substituents must result in step (2) becoming slowerthan step (3), thereby making it now rate-limiting for the overallreaction. This shift in rate-limiting step from step Q)~ step (2)occurs, in Fig. 13.8, with the compound (38; X, Z= Me).

Still further decrease in electron-donation by the substituents,beyond this point, will result in still further slowing-down of stepa>----now the rate-limiting step-and hence slowing-down of theoverall reaction i.e. the slope on the right-hand side of the plot isnow downward; from left to nght (p is -ve). For a reaction in whichsuch a shift of rate-limiting step is observed (as the electrondonating/-withdrawing ability of the substituent is changed) therewill be one substituent, or narrow range of substituents, for whichthe balance between the rates of step ~ and step Q) is such as tomake the overall reaction rate a maximum.

This happens in Fig. 13.8, as we have seen, with the compound(38; X, Z = Me). On each side of this maximum the, different,rate-limiting step will be slowing down progressively, and so therefore will the overall reaction rate. Shifts in rate-limiting step, withinthe same overall reaction pathway, are thus distinguished by concave downwards deviations in Hammett plots; this in contrast to theconcave upwards deviations which, as we have already seen (p.364), are characteristic of a change in overall reaction pathway.

QO.Pt<

MeO OMe(4tb)


<ocq~¢Q

MeO OMe(41a)

382

step CD: initial protonation in acid-catalysed dehydration is generallyrapid; or step @: final loss of proton in aromatic electrophilicsubstitution is also generally rapid. This leaves steps ~ and Q) aspossible candidates for the slow step overall, and fortunately a cleardistinction can be made between them. In step (2), +ve charge isincreasing at the reaction centre (the carbon atom carrying the twosubstituted Ar groups), while in step Q), +ve charge is decreasing atthe reaction centre. How does this match up with the requirementsof Fig. 13.8 (p. 381)?

The right-hand side of the plot in Fig. 13.8 has a -ve p value(- 2·51) indicating the development of substantial +ve charge at the 'reaction centre during the overall, rate-limiting step. This would, ofcourse, be compatible with step ~ being rate-limiting, but not withstep Q). For the left-hand side of the plot in Fig. 13.8, exactly thereverse is true; here a +ve p value (+2'67) indicates a substantialdecrease of +ve charge at the reaction centre, which would becompatible with step Q) being rate-limiting, but not with step ~.

It is significant that the substituents involved at the far left-handside of the plot (38; X, Z = MeO) are powerfully electron-donating,and thus capable of stabilising the carbocation (41a ++ 41b), developing in step ~, by delocalisation of its +ve charge. It is indeed

found that the log kobs values on the left-hand side of Fig. 13.8 givea better straight line when plotted against Iu+, rather than againstIu, because of the through-conjugation (41a ++ 41b) between thesep-substituents and the reaction centre.

In (38; X, Z = MeO) this conjugative stabilisation results in easy ,fonnation of the carbocation (41), i.e. to a rapid step ~; but theconsequent delocalisation of +ve charge, away from the reactioncentre (41a ++ 41b), clearly makes (41) a less effective electrophile,i.e. step Q)-.electrophile attack on the benzene nucleus-is therefore slow. It is thus step Q) that is slow, and hence rate-limitingoverall, for compound (38; X, Z = MeO). On moving across Fig.13.8, from left to right, the substituents become less electrondonating, delocalisation of +ve charge thereby becomes less pronounced, and the reaction centre progressively more electrophilic.

Base-catalysed hydrolysis(BAc2): p = 2·51[9]

38513.5.1 Taft equation

[R-1:J*OH2 J

(42a)

T.S. for acid-catalysed

hydrolysis (AAC2)

[kR] [kR]log -- -log -- = p*a:'k o base k o acid

[~. ]!R-t;Ej

(42b)

T.S. for base-catalysedhydrolysis (BAC2)

polar substituents is concerned, and the overall p value for thereaction is thus virtually zero.

U we now extend our consideration of base-catalysed (BAc 2), andacid-catalysed (AAc 2), hydrolysis to esters in general, includingaliphatic ones (RCOzEt), we see that there is a close similaritybetween the transition states (42b or 42a) for the rate-limiting stepin each of the two pathways: they are both tetrahedral; and differ

only in the second of them having two protons more than the first.Protons, being very small, exert comparatively little steric influence;it i~ therefore a not unreasonable assumption that any steric effectstemming from the group R is, because of the close spatial similarityof the two transition states, substantially the same in both acid- andbase-catalysed hydrolysis.t It then becomes possible to write aHammett type equation, [9], to represent the operation of the polareffect only of substituent R in ester hydrolysis:

As the steric effect exerted by R is essentially the same in bothmodes of hydrolysis, the two steric terms will cancel each other out,and will thus not appear in equation [9].

Taft then gave p* in [9], the value 2·48, derived by subtractingthe p value for acid-catalysed hydrolysis of benzoate esters (0·03)from the p value for base-catalysed hydrolysis of the same esters(2,51). He took as his reference substituent R = Me, rather thanR = H, so that ko in [9] refers to MeCOzEt rather than HCOzEt.Then by kinetic measurements on the acid- and base-catalysedhydrolysis of a series of esters containing R groups other than Me, itis possible-using [9}-to evaluate 0": for each of these different Rgroups with respect to Me, for which by definition O"~e = 0 (cf. Hwith O"H = 0 for benzoic acid ionisation, p. 363). By giving p* herethe value 2·48, the resulting 0": values-which are a measure of thepolar effect only exerted by R-do not differ too greatly in mag-

t Such an assumption does, however, neglect the possibility that the degree ofsolvation of +vely and -vely charged T.S.s could be markedly different, therebygreatly influencing the relative rates of the two hydrolyses.

°IIAr-C+HOEt

6e

°-OEl& /I~Ar-C+ OEt

IOH

(0 oe/I Ij

Ar-C-OEt "O<DH, Ar-C,..-OEt

<e; IVOH OH

(42)

OHH2 0:, I ..~ Ar-C-OEt

(2) IH 2°(il

11~° ~H C.0IH/I -H* II -EtOH

Ar-C ==::; Ar-C+ HOEt ' 'Ar-C1-.tQEtI <3J I @ tt

OH HO HO

Acid-catalysed hydrolysis(AAc2): p = 0-03

0:/I H*

Ar-C-OEt , <D '(42)


13.5.1 Taft equation

Acting on a suggestion originally made by Ingold, Taft began bycomparing the relative susceptibility to polar substituent effects (thep value) of the hydrolysis-under acid-catalysed (AAc2, p. 241) andunder base-catalysed (BAc 2, p. 239) condition~f m- and psubstituted benzoate esters (42).

The p value for base-catalysed hydrolysis (+2·51) is +ve andquite large, reflecting the development of not inconsiderable -vecharge at the reaction centre in the rate-limiting step--attack on thiscentre by sOH (step ill in the BAc 2 pathway). By contrast, the pvalue for acid-catalysed hydrolysis (+0·03) is very nearly zero;which means, of course, that the rate of this hydrolysis does not varysignificantly from one ester to another, no matter what the m- or

p-substituent present. The p value for this hydrolysis is so small,despite their being considerable redistribution of +ve charge in theslow step (step ~), because the overall rate of reaction, Le. koblJ(which is plotted to evaluate p), is determined not solely by kz forthis slow step, but involves also K 1 for the preceding, reversible,step CD. These two terms all but cancel each other out, in so far assusceptibility of the two steps to electron-donation/-withdrawal by

steric effect on it either. These esters thus all undergo acid-catalysed

[11]

[12]

387

Es-0,39-1,13-1,54-1·74-1-76-3·81

R in RC02Et

Me(CH2)3Me2CHCH2Me3CMe3CCH2Ph2CHEt3C

[k ]log RCD,EI = Esk MeC02Et acid

13.5.2 Steric parameters, Es and {J

Es+1·24

o (by definition)-0-07-0-24-0-37-0-38

R in RC02Et

HMeEtCICH2ICH2PhCH2

susceptibility towards steric effects. In that sense 8 is the stericparallel to p*-which measures the reaction's susceptibility towardspolar effects. The 8 parameter is given the value 1·00 for acidcatalysed ester hydrolysis, as the standard reaction, and its value forother reactions can then be determined experimentally in the usualway.

hydrolysis at essentially the same rate. There is no reason to believethat acid-catalysed hydrolysis of aliphatic esters, RC02Et, will beany more susceptible to polar effects than was the correspondinghydrolysis of benzoate esters. If then different hydrolysis rates areobserved with aliphatic esters as R is varied, these must reflectdiffering steric effects exerted by the different R groups_ Suchaliphatic esters are indeed found to undergo hydrolysis at markedlydifferent rates, so it is possible, taking Me as the standard substituent once again, to use equation [11]

From the form of equation [11], the Es value for Me, thereference substituent, will of course be 0_ All substituents other thanH have -ve Es values because all substituents other than Harelarger than Me, and the rate of hydrolysis of any ester RC02Et(R =f H) will thus be slower than that of MeC02Et, in a reactionwhose rate is governed solely by the steric effect of R.

It is found in practice that the value of the steric parameter, Es,for a particular group, R, differs to some extent from one reaction toanother. This is not altogether surprising as both the local environment of R and the size of the attacking reagent will vary from onereaction to another. It means, however, that on incorporating Esinto the Hammett type equation, [12], it is necessary to introduce ayet further parameter, 8, as a measure of a particular reaction's

to evaluate Es, the steric substituent parameter, for R. Es values,obtained in this way for a number of different substituents, are listedbelow:

[10]


nitude from the values of ux, u;c and Ux with which we are alreadyfamiliar (p. 363).

Then, employing the more general equation [10], it is possible to ,I

use these u: values, in conjunction with suitable kinetic measure~ l

ments of kR and kMe, to evaluate p* for other .

kRlog-=p*u:kMe

reactions of a whole range of aliphatic compounds in addition toesters. Using [10] in this way, straight line plots were obtained for anumber of different reactions of aliphatic compounds.

13.5.2 Sterie parameters, Es and 8

After all the emphasis we placed earlier (p. 361) on steric effects,obtaining a straight line plot may at first sight seem a rathersurprising result; especially, given that the relation [10] takes into·account only the polar effect exerted by R. However, obtaining astraight line plot, using [10], does not necessarily mean that nosteric effects are operating in a reaction. It means only that there isno substantial change in the operation of such effects on going fromstarting materials to the transition state for the rate-limiting step ofthe overall reaction (or on going from starting materials to productsfor an equilibrium).

It is not necessary to look very far to find aliphatic reactions that 'do not yield straight line plots with [10], however; and, as withprevious deviations from linearity (p. 375), these departures arecommonly much more informative about the details of reaction .~athways than are neat straight lines. Where such departures from ;\l~near (~lar. effects only) plots are observed, suggesting the opera- ,~Ion of sIgmficant-and changing-steric effects, it is possible toIncorporate a steric substituent parameter, Es, whose evaluation is 'based on an earlier observation.

Thus we have already seen (p. 384) that the acid-catalysedhydrolysis of m- and p-substituted benzoate esters (42) is (with a pvalue of 0·03) essentially uninfluenced by any polar effect exertedby the substituent, X; and this substituent is sufficiently far removedfrom the reaction centre to be clearly incapable of exerting any

13.6 SOLVENT EFFECfS

389

[13]

13.6.2 Grunwald-Winstein equation

Me3C-Cl s1S~1 Me3CEIl Cia S M C 8v_ ~ e3 -

(46) (47) [8 = solvent]ion-pair

intermediate

The relative effectiveness of such solvation-<>f negatively chargedanion (45) with respect to neutral, undissociated acid (44)-is amajor factor in determining the position of equilibrium, i.e. Kx. Asthe solvent is changed from water, with a dielectric constant of 79,to ethanol, with a dielectric constant of only 24, there will be amarked decrease in advantageous solvation of the charged anion(45) with respect to the uncharged acid (44). The relative importance of the polar effect exerted by electron-withdrawing substituents, in overall stabilisation of the carboxylate anion (i.e. inacid-strengthening: increasing Kx), will therefore increase as thedielectric constant of the solvent decreases. The value of p, thesusceptibility of the reaction to the polar effect of a substituent, willalso increase, therefore, on changing the solvent from water toethanol.

13.6.2 Gnmwald-Winstein equation

Attempts to correlate the differing rate of a particular reaction,when carried out in a range of different solvents, with the dielectricconstant values for these solvents have not proved very rewarding.Attempts have therefore been made to establish empiricalreactivity/solvent correlations along general Hammett lines. Amongthe more significant of these attempts has been that of Grunwaldand Winstein on the solvolysis of halides. They sought to establish asolvent parameter, designated Y, which would correlate with thedifferent rate constants found for solvolysis of the same halide in arange of different solvents.

They took as their standard reaction the SN1 solvolysis of thetertiary halide, 2-chloro-2-methylpropane (46), and selected as theirstandard solvent 80% aqueous ethanol (80% EtOH/20% H 20):

It is then possible to set up the Hammett-like relation, [13],

in which the rate constants, kA and ko, refer to solvolysis of thetertiary halide (46) in a solvent A and in the standard solvent (80%aq. EtOH), respectively; while YA and Yo are the empirical solventparameters for solvent A and for this standard solvent. By settingthe value of Yo at zero and measuring kA for the solvolysis of (46) ina range of different solvents, it is then possible, using [13], to derive

P

1·00 (by definition)1·601·96


effect of substituents than is the standard reaction, the acidcatalysed hydrolysis of o-substituted esters. In general, attempts toquantify o-substitituent effects have not, however, been very successful. We are here, once again, faced with some dilemma as wewere with the Yukawa-Tsuno equation (p. 374): how far does anyadditional information gained merit the very considerable effortinvolved in the experimental evaluation of such further parametersin the first place?

Reaction

ArC02H(44) +H20 +:t ArC~(45) +H 30t& (H20)" +" +:t" +.. (50% aq. EtOH)

+ .. +:t + " (EtOH)

Now that steric parameters have been introduced in this way, thetreatment can be extended to include the reactions of o-substitutedbenzene derivatives as well. Thus for the acid-catalysed hydrolysisof o-substituted benzamides (43), the value of l) is found to be 0·81·so this reaction is apparently slightly less susceptible to the steri~

One of the things our discussion of linear free energy relationshiphas not yet made any endeavour to take into account is the roleplayed in reactions by the solvent. This despite the fact that the verygreat majority of organic reactions do take place in solution, withthe solvent often playing a crucial role.

For ionisation of m- and p-substituted benzoic acids (44), thehydroxylic solvent is capable of solvating both the undissociated acid(44) and the carboxylate anion (45) obtained from its ionisation.

ArC02Et + SOH -+ ArCdf +EtOH (70% aq. dioxan) 1·83+ " -+ + (85% aq. EtOH) 2·54

13.6.1 Change of p with solv~nt

It is, of course, true that some implicit consideration is given to thesolvent in that the p value for a particular reaction is found tochange when the solvent in which the reaction is carried outchanged:

39113.6.3 Dimroth's ET parameter

An alternative interpretation of m is that it provides some measure of the extent of ion-pair formation in the transition state for therate-limiting step of the overall solvolysis reaction: it can then beput to some diagnostic use. Thus, ion-pair formation is known to bewell advanced in the transition state for SN 1 solvolysis of 2-chloro2-methylpropane (46), the standard halide, for which m = 1·00. Notaltogether surprisingly the value of m for 1-bromo-1-phenylethane(48), in which the developing benzyl type cation, [PhCHMe]E9, isstabilised by delocalisation of its +ve charge over the 7T-system ofthe attached benzene nucleus (cf. p. 84), is even larger-at 1·20. Bycontrast, the m values for the primary halides, bromoethane (49)and 1-bromobutane (50), are much lower-0'34 and 0,33, respectively. These values, indicating low susceptibility towards the ionising power of the solvent, are characteristic of halides whose solvolysis is known to proceed via the SN2 pathwa¥. In general, an mvalue of 0·5 can be taken as an approximate indicator of anSN 1/SN2 mechanistic borderline in solvolysis reactions of this kind.

The major defect of the Grunwald-Winstein treatment is that it islimited in its scope. It has been applied to reactions other thanhalide solvolysis, but is in general restricted to those reactions forwhich the major contribution to the rate-limiting step is of the form:

(51)

Ph ~P~:e'--(~

Ph Ph

whose absorption maximum was found to vary between 450 and1000 nm, depending on the solvent: its solution being yellow inMeOH, red in Me2CHOH, and blue in CHCl3 ! Dimroth took as ameasure of solvent polarity, Ey: the excitation energy (ground~excited state) in kcal mol- l at the absorption maximum in thatsolvent. The justification for Ey is that the ground state of (51) is

13.6.3 Dimrotb's Er parameterThere have been several other attempts to define solvent polarityparameters, among the more successful being those related tosolvatochromic shifts: the shift in wave-length/frequency of a bandin the spectrum of a suitable absorbing species resulting from itsinteraction with the molecules of a series of different solvents.Particularly large shifts were observed with the zwitterion (51),

[14]

~

78·5

32·76·2

24·318·312·2

109·5

m

1·201·00 (by definition)0·940·900·890·340·33

YA

+3.49+1·97+0·60+0·59

o (by definition)-0·67-1,09-1,64-2·03-2,73-3,26

Halide

PhCH(Me)Br(48)MeJCCI(46)MeJCBrEtMe2CBrCH2=CHCH(Me)CIEtBr(49)Me(CH2)JBr(50)


Solvent, A

H 20aq. MeOH (50% H20)

HCONH2aq. EtOH (30% H20)aq. EtOH (20% H20)aq. Me2CO (20% H20)

MeOHMeC02HEtOHMe2CHOHMeJCOH

a YA value for each of them:

390

These YA values are found not to run in parallel with thedielectric constant values for the solvents concerned. Obviously thedielectric constant value for the solvent must be involved in some 'way in YA , as separation of opposite charges is a crucial feature ofthe rate-limiting step in an SN1 reaction: formation of the T.S.leading to the ion-pair intermediate (47). But specific solvation ofthe separating charges must also be involved and Y A will reflectthose, and quite possibly other properties of the solvent as well. It iscommon to describe YA as representing a measure of the 'ionisingpower' of the solvent A.

It is now possible to go a stage further, and write a not unfamiliarrelation, [14], that now covers the solvolysis of halides in general,and not merely that of the

kAlog-=mYAko

standard halide, 2-chloro-2-methylpropane (46). Here kA and koare the rate constants for solvolysis of any halide, in solvent A andin the standard solvent, respectively. YA has already been defined asa solvent parameter representing the ionising power of solvent A,while m is a compound parameter characteristic of the particularhalide: it is given the value 1·00 for the standard halide, 2-chloro-2methylpropane (46). The actual value of m can be taken as ameasure of the susceptibility of the solvolysis of a particular halidetowards the ionising power, YA' of that solvent:

13.7 SPECfROSCOPIC CORRELATIONS

393

p=-4'54defines u;_x

X:(57)

13. 7 Spectroscopic correlations

MA ,--18.'y:. needs u;_~

X:

(56a)

all, reflect the degree of electron shielding or de-shielding at therelevant atom. In fact, correlation of 8 data for IH with Ux has notbeen very impressive except, as with (55),

showed a good straight line correlation of the l3C chemical shiftdifference (for the carbocation carbon: 8~-8c;b with u;:-'_x, butnot with u;_x. Correlation of the shift differences for the psubstituted carbocations required enhanced p-substituent constants,u~, reflecting the much more powerful 'through conjugation' withp-X in the fully formed carbocation (56a), as compared with that inthe only partly formed carbocation (57) in the T.S. for cumylchloride solvolysis-the standard reaction that was chosen (p. 357) todefine u;_x:

6,--033X

(55)

Me Me"'-.Gl/

C

~X

(56)

where the relevant proton is fairly remote from the substitutedbenzene ring.

However, an atom somewhat heavier than IH might well be lesssusceptible to the perturbations that may disturb the latter; as, forexample, l3C which also generates an n.m.r. spectrum. Thus the2-arylpropyl(cumyl) carbocations (56; produced from the corresponding tertiary alcohols in 'super acid' -SOzCIF/FS03H/SbFssolution, ct. p. 181),

YA

+3·49+0·60

o-0·67-1,09-2·03-2·73-3·26

~X

(54)


Solvent ErH 20 63·1HCONH2 56·6

aq. EtOH (20% H20) 53·7aq. Me2CO (20% H 20) 52·2

MeOH 55·5EtOH 51·9Me2CHOH 48·6Me3COH 43·9CHCl3 39·1

Values of YA (ct. p. 390) for the same range of solvents are includedfor comparison; by and large the Er parameter is the more successful of the two, and has somewhat wider application.

very much more polar than the excited state to which it gives rise,and so will, of the two, be stabilised to a much greater extent bypolar solvents. Assuming that the effect of solvent variation on theenergy level of the excited state is only small, then the varyingvalues of Er observed will be a measure of the relative stabilisationof the ground state (51), and hence of the relative polarity of thesolvent involved; Er rising as the stabilisation, and hence solventpolarity, increases:

We have discussed at some length correlation of the chemicalproperties of X-substituted molecules with ux-the polar substituent constant for X-and it is pertinent to enquire whethersimilar correlations can also be established between Ux and theirphysical properties, among which spectroscopic data constitute areadily accessible example.

There have been many attempts to correlate Ux with the frequency and/or intensity of bands in the Lr. spectra of X-substitutedaromatic species. Among the most successful have been with thefrequency of the C=O band in (52) and (53), and with the intensityof the 1600 cm-1 ring vibration in (54):

We might well expect to find reasonable correlations of Ux withchemical shift, 8, data (cf. p. 18) from n.m.r. spectra, which do, after

13.8 THERMODYNAMIC IMPUCATIONS

[k' = Boltzmann's constant]Lh = Planck's constant J

hE' ~.~':,=' ~e +E,OH ~::~;':~,~ Z5° ~

X X(58)

as'F is found to be virtually constant--eondition (3) above-andaG\ ~r aH'f, is thus.tound to be proportional to Ux' Not altogethersurpnsIngly, no convIncing example is known in which condition (2)

39513.8 Thermodynamic implications

is met, but condition (1) might well be expected to be one that ismost frequently satisfied. Interestingly enough, doubt has in the pastbeen expressed as to whether even the standard, referencereaction-the aqueous dissociation of m- and p-substituted benzoicacids in water at 2So-satisfied this condition.

A major obstacle to deciding the truth or otherwise of thisassertion about benzoic acid ionisation has been the experimentaldifficulties involved in making the necessary measurements. Thesolubility of the acids in water is pretty low, and their aHB valuesare very small, with consequent imprecision in, and unreliability of,the results so obtained. Relatively recently, however, aGe, aw andaSB have been redetermined, with great precision, for a series often m- and p-substituted benzoic acids. Using these data, stringentlylinear plots were obtained for aHB against aSB

, for aGB againstaHB

, and for aGB against aSB. So it looks as though Hammett was

on to a good thing after all when he made his choice of standard,reference reaction in the first place!

It is important, however, to remember that what theoreticalinterpretation there has been of the Hammett equation has comefrom circumstantial evidence rather than by rigorous proof. Itremains an empirical relationship and, to that extent, there is nopoint in even trying to evaluate Ux and p to several places ofdecimals. The sort of information we need, as an aid to theelucidation of reaction pathways, is of an 'order of magnitude' kind:such things as whether p is +ve or -ve, whether its value is large orsmall, whether there are noticeable deviations from linearity in plotsof Ux against log k x and, if so, of what kind. This also raises thequestion of multi-parameter equations, not so much of their generalvalidity but of their actual usefulness. While they are certainly ofconsiderable interest to physical organic chemists, it is moredoubtful-so far as practising organic chemists in general areconcerned-whether the extra labour, necessarily involved inevaluating all these further parameters, is repaid by the quality ofthe additional information that is thereby gained: you pays yourmoney and you takes your choice!

Having said all that, it is equally important to remember that thenumber and variety of useful correlations to which Hammett plotshave given rise is quite astonishing, particularly when we considerthe simplicity and convenience of the approach. Indeed, linear freeenergy relationships in general constitute a testament to the theoreticalutility of concepts that are purely empirical in their genesis!


It is perhaps interesting, in view of the very considerable success ofHammett plots, to say a word finally about the thermodynamicimplications of linear free energy relationships in general. We havealready mentioned (p. 359) the relationship between free energychange, aG, and log k or log K; and each aG term is, of course,made up of an enthalpy, aH, and an entropy, as, component:

Equilibrium AGe = - 2·303 RT log Kconstant: AGe = AIF -TASe

Rate AG'f = -2,303 RT log k-.!!...-constant: - k'T

AG'f = ARf -TAS"

Looking back at one of our earliest examples-Fig. 13.3 (p. 361)in which log K for the ionisation of ArCOzH is plotted against log kfor the base-catalysed hydrolysis of ArCOzEt-thestraight line impliesthat there is also a linear relationship between the aGe valuesfor the former reaction and the aG'f values for the latter. Such astraight line relationship between these two series of aG terms is tobe expected only if, for each series, one or other of the followingconditions is satisfied:

(1) aH is linearly related to as for the series(2) aH is constant for the series(3) as is constant for the series

. Any of these conditions constitutes an extremely stringent limitatIon, and there has always been some doubt expressed over theextent to which anyone of them is indeed satisfied in reactionswhich nevertheless give quite good straight line Hammett plots:thereby making the linear relationships that are observed all themore mysterious! Examples are, however, known that can indeed beshown to conform to one or other of the above conditions. Thus forthe base-induced hydrolysis of the esters (58),

Select bibliography

Structural theory and spectroscopy

DEWAR, M. J. s. and DOUGHERTY, R. C. The PMO Theory of OrganicChemistry (Plenum, 1975).

FLEMING, I. Frontier Orbitals and Organic Chemical Reactions(Wiley, 1976).

MCWEENY, R. Coulson's Valence (GUP, 3rd Edition, 1979).TEDDER, J. M. and NECHVATAL, A. Pictorial Orbital Theory (Pitman

1985). 'WILLIAMS, D. H. and FLEMING, I. Spectroscopic Methods in Organic

Chemistry (McGraw-Hill, 3rd Edition, 1980).

Structure and reaction mechanism

ALDER, R. W., BAKER, R. and BROWN, J. M. Mechanism in OrganicChemistry (Wiley-Interscience, 1971).

AMIS, E. s. Solvent Effects on Reaction Rates and Mechanisms(Academic Press, 1966).

BUNCEL, E. Carbanions: Mechanistic and Isotopic Aspects (Elsevier,1975).

BUNCEL, E. and DURST, T. (Eds). Comprehensive Carbanion Chemistry:Parts A and B (Elsevier, 1980, 1984).

CAPON, B. and MCMANUS, S. P. Neighbouring Group Participation(Plenum, 1976).

CARPENTER, B. K. Determination of Organic Reaction Mechanisms(Wiley, 1984).

COLLINS, c. J. and BOWMAN, N. s. (Eds). Isotope Effects in ChemicalReactions (Van Nostrand Reinhold, 1970).

DE LA MARE, P. B. D. Electrophilic Halogenation (CUP, 1976).DE LA MARE, P. B. D. and BOLTON, J. Electrophilic Addition to Unsatu-

rated Systems (Elsevier, 2nd Edition, 1982). :DE MAYO, P. (Ed.). Rearrangements in Ground and Excited States

(Academic Press, Vols. I-III, 1980).ELIEL, E. L. Stereochemistry of Carbon Compounds (McGraw-Hill,

1962).FORRESTER, A. R., HAY, J. M. and THOMSON, R. H. Organic Chemistry of

Stable Free Radicals (Academic Press, 1968).GARRATT, P. J. Aromaticity (McGraw-Hill, 1971).GILCHRIST, T. L. and REES, c. w. Carbenes, Nitrenes and Arynes

(Nelson, 1969).

Select bibliography 397

GILCHRIST, T. L. and STORR, R. c. Organic Reactions and OrbitalSymmetry (CUP, 2nd Edition, 1979).

GILLIOM, R. D. Introduction to Physical Organic Chemistry (AddisonWesley, 1970).

HARTSHORN, s. R. Aliphatic Nucleophilic Substitution (CUP, 1973).HINE, J. Structural Effects on Equilibria in Organic Chemistry (Wiley,

1975).HOFFMANN, R. W. Dehydrobenzene and Cycloalkynes (Academic

Press, 1967).HOGGETf, J. G., MOODIE, R. B., PENTON, J. R. and SCHOFIELD, K.

Nitration and Aromatic Reactivity (CUP, 1971).INGOLD, c. K. Structure and Mechanism in Organic Chemistry (Bell,

2nd Edition, 1969).ISAACS, N. s. Reactive Intermediates in Organic Chemistry (Wiley,

1974).JENCKS, w. P. Catalysis in Chemistry and Enzymology (McGraw-Hill,

1969).I> ••JOHNSON, C. D. The Hammett Equation (CUP, 1973).

JONES, R. A. Y. Physical and Mechanistic Organic Chemistry (CUP,1979).

KIRMSE, w. Carbene Chemistry (Academic Press, 2nd Edition, 1971).KLUMPP, G. w. Reactivity in Organic Chemistry (Wiley, 1982).KOCHI, J. K. (Ed.). Free Radicals (Wiley-Interscience, Vols. I and II,

1973).LOWRY, T. H. and RICHARDSON, K. s. Mechanism and Theory in

Organic Chemistry (Harper and Row, 1976).LWOWSKI, w. (Ed.). Nitrenes (Interscience, 1970).MARCHAND, A. P. and LEHR, R. E. (Eds). Pericyclic Reactions

(Academic Press, Vols. I and II, 1977).MILLER, J. Aromatic Nucleophilic Substitution (Elsevier, 1968).NONHEBEL, D. c. and WALTON, J. c. Free-radical Chemistry (CUP,

1974).NORMAN, R. O. c. and TAYLOR, R. Electrophilic Substitution in Ben

zenoid Compounds (Elsevier, 1965).OLAR, G. A. and SCHLEYER, P. VON R. (Eds). Carbonium Ions (Intersci

ence, Vols. I-V, 1968-76).RITCHIE, c. D. Physical Organic Chemistry (Dekker, 1975).SAUNDERS, w. H. and COCKERILL, A. F. Mechanisms of Elimination

Reactions (Wiley-Interscience, 1973).SHORTER, J. Co"elation Analysis in Organic Chemistry (GUP, 1973).STEVENS, T. s. and WATTS, N. E. Selected Molecular Rearrangements

(Van Nostrand Reinhold, 1973).TESTA, B. Principles of Organic Stereochemistry (Dekker, 1979).THYAGARAJAN, B. s. (Ed.). Mechanisms of Molecular Migrations (In

terscience, Vol. I, 1968; Vol. II, 1969; Vol. III, 1971).VOGEL, P. Carbocation Chemistry (Elsevier, 1985).

398 Select bibliography

WENTRlJP, c. Reactive Molecules: The Neutral Reactive Intermediatein Organic Chemistry (Wiley, 1984).

Review seriesAdvances in Physical Organic Chemistry. GOLD, v. and BEIHELL, D.

(Eds) (Academic Press, Vol. I, 1963- ).Organic Reaction Mechanisms. KNIPE, A. C. and WATTS, w. E. (Eds)

(Interscience, Vol. I, 1966- ).Progress in Physical Organic Chemistry. TAFT, R. w. (Ed.) (Inter

science, Vol. I, 1963- l.Reactive Intermediates-A Serial Publication, JONES, M. and MOSS, R.

A. (Eds) (Wiley-Interscience, Vol. I, 1978- ).

Audio cassettesThe following titles in the 'Chemistry Cassette' series are publishedby, and are available from, The Chemical Society, Blackhorse Road,Letchworth, Herts., SG6 1HN, England:CC3 SYKES, P. Some Organic Reaction Pathways: (A l. EliminatiOn;

(B). Aromatic Substitution (1975).CC6 SYKES, P. Some Reaction Pathways of Double Bonds: (A).

C=C; (B). C=O (1977).CC7 SYKES, P. Some Reaction Pathways of Carboxylic Acid Deriva

tives (1979).CC9 SYKES, P. Radicals and their Reaction Pathways (1979).CCll SYKES, P. Linear Free Energy Relationships (1980).

1'I"

Index

Acetalscyclic, 210formation, 209, 289hydrolysis, 74, 210

Acid catalysisgeneral,74specific, 74

AAc1 pathway, 242, 243AAc2 pathway, 241, 244,384AAL1 pathway, 241Acidity

anion stability and, 23, 55, 271aromaticity and, 275Brlllnsted and, 53constant Ka • 54,271definitions of, 53~GB and, 58, 60, 61, 394~IP and, 58, 394~SB and, 58, 394effect of solvent, 56effect of temperature, 64H-bonding in, 63, 64hybridisation and, 59, 273in C-H comps., 251, 270kinetic control of, 270, 280Lewisand,54origin in organic comps., 55steric effects in, 58

Acid strengthalcohols, 55, 56aliphatic acids, 54, 55, 57, 59alkanes, 55, 271alkenes, 272alkynes, 223, 273, 289,294aromatic acids, 62catalysis and, 74dicarboxylic acids, 63irnides, 68nitroalkanes, 272, 280, 283phenols, 23, 56,61triphenylmethane, 271

Acrylonitrile, 199Activated complex, 38

399

Activation parametersenergy, Ea , 38enthalpy, ~H"', 38entropy, ~S"', 39free energy, ~G"', 38

Acyl cations, 102, 144, 240, 242, 379Acyloin condensation, 218Addition, 30

1,3-dipolar, 192, 194,351electrophilic, 31, 50, 178-198nucleophilic, 31, 198,207-245radical, 31, 313-323to C=C, 31, 51, 103, 113, 178, 188to C==C, 182to c=c----e=C, 195to c=c----e=O, 200to C==N, 244to C=O, 31, 50, 103,203

Adsorption, 191Aldehyde ammonias, 220Aldol reaction, 224

acid catalysed, 225crossed,226intramolecular, 226reversibility, 75, 224

Alkenescycloaddition, 348polar addition, 178-195protection in, 265radical addition, 313-323relative stability, 26, 256

Alkylchlorosulphites, decomp., 93Alkyl shifts, 111, 113, 115,293,337A1kynes

acidity, 223, 272addition to C=O, 223alkylation, 289dimerisation, 294ozonolysis, 194

Allylcation, 85, 105, 196radical, 311rearrangement, 109

Ambident nucleophiles, 97Amides

from Beckmann, 123Hofmann reaction, 49, 122hydrolysis, 239, 245, 388reduction, 214

Amidinesbasicity, 69formation, 244

Aminoazo comps, 148Anionic polymerisation, 200, 226Antarafacial shifts, 353Anthracene, 17

photodimer, 337transannular peroxide, 331

Anti-bonding orbitals, 6, 292Anti-knock agents, 305Anti-Markownikov addition, 317Anti-oxidants, 330Antiperiplanar conformation, 118, 253Arenium ions, 131Arenonium ions, 131Arndt-Eistert reaction, 119Aromaticity, 14

cycloheptatrienyl cation, 18, 104, 106cyclopentadienyl anion, 18, 275cyclopropenyl cation, 18, 106n.m.r. spectrum and, 18requirements for, 17

Aromatic substitution, 41electrophilic, 130--167internal, 334, 381nucleophilic, 167-177radical,331-335

Arrhenius equation, 38Arrows

curved,19double-headed, 19

Arylation, 332Aryne intermediates, 21,174Associative process, 239, 241Atactic polymers, 322Atomic orbitals, 1,342Aufbau principle, 3Autoxidation, 306, 318, 328

a1kenes, 329benzaldehyde, 306, 330cumene, 128ethers, 329tetralin, 329

Axial overlap, 6Azides, 123, 194Azoalkanes

photolysis, 304thermolysis, 305

cyclopentadienyl anion, 275deuterium exchange, 288displacement reactions, 287ElcB elimination and, 248, 285electronic effects in, 272enolate anions, 279, 290, 295formation, 271in carbonation, 284in Darzens reaction, 290in decarboxylation, 285in Favorskii reaction, 294in halogenation of ketones, 295in Kolbe-Schmidt reaction, 291in Michael reaction, 200in Reimer-Tiemann reaction, 290intermediates, 41, 200, 229, 295in Wurtz reaction, 289oxidation, 294, 307rearrangement, 292solvation, 45stabilisation, 251, 257, 262, 273, 2%steric effects in, 276tautomerism and, 277trapping of, 286triphenylmethyl, 271, 272

Carbenes, 21dichlorocarbene, 4, 50,267,290intermediates, 119,266

Carbinolamine intermediates, 50, 219Carbocations, 21, 101-119

acyl, 102, 144, 242, 379addition to C=C, 113, 188, 225allylic, 85, 105, 1%benzylic, 84, 91, 102, 105, 112,371bridged, 105, 118, 129,377cycloheptatrienyl, 18, 104, 106cyclohexadienyl, 159cyclopropenyl, 18, 106decomposition, 102ethyl, 83, 104, 380formation, 101, 107, 120hydride shifts in, 108, 109in E1 pathway, 248, 261in Friedel-Crafts, 108, 141, 145in SN1 pathway, 78, 81, 83, 90, 371loss ofHlIl , 107, 111methyl, 83, 104, 3802-methylpropyl, 78, 83, 102, 104n.m.r. and, 102, 181,393phenonium, 105phenyl,169propyl, 83, 104, 107, 109rearrangement, 32,102,107,109-119reduction, 306solvation, 45

400 Index

BAC2 pathway, 239, 384Baeyer-Villiger oxidation, 127Barton reaction, 337Base catalysis

general,75ion exchange resins, 226specific, 75

Basicityaromaticity and, 69, 72Brl1lnsted and, 53cation stability and, 67, 68, 72, 73constant, 65effect of solvent on, 66, 67H-bonding in, 67inductive effect and, 22, 66Lewis and, 54multiply bonded Nand, 72nucleophilicity and, 96origin in organic comps., 66ortho effects in, 71softness, 96steric effects in, 70, 71

Basic strengthaliphatic amines, 66amides,68amidines, 69aromatic bases, 69catalysis and, 75conjugate acids, 65~Wand,66guanidine, 68heterocyclic bases, 72, 166tetraalkylammonium hydroxides, 67

Beckmann rearrangement, 123acid catalysis, 124H2

180 in, 125intramolecularity, 126solvent polarity and, 126stereoselectivity, 124

Benzenearomaticity and, 14bond lengths, 15charge cloud, 15, 130delocalisation in, 14, 130heat of hydrogenation, 16Kekul~ structures, 14, 194m.o.s of, 14n.m.r. signal, 18ozonolysis, 194planarity, 14stability, 15

Benzilic acid rearrangement, 232Benzoin condensation, 231

cyanide ion in, 231thiazolium ylids in, 232

Benzotriazoles, 148, 176Benzyne, 175,251

dimerisation, 176isolation, 175mass spec., 176structure, 175

Biosynthetic pathways, 48Biphenylene, 176Bipolar non-protic solvents, 81, 98, 173,

252Biradicals, 315, 331, 337

oxygen,315,330stable, 338

Bisulphite addition comps., 207, 213Bond

angles, 4energy, 9, 304heterolytic fission, 20, 299homolytic fission, 20, 299

Bond-breaking~H!' and, 38

Bond lengths, 7hybridisation and, 7

Bonding orbitals, 6Branching, 321Bredt's rule, 259Bridged species, 105, 118, 129, 335, 377Bromodealkylation, 162Bromonium ions, cyclic, 180

detection, 181isolation, 181

1-Bromotriptycene, 87Brl1lnsted

acids, 53bases, 53

Butyl rubber, 189

Cadmium Alkyls, 238Cannizzaro reaction, 47, 216

intramolecular, 217isotopic labels in, 47, 217crossed, 216

Canonical structures, 19, 152Carbamic acids, 122Carbanions, 21, 270-298

acetylide anion, 223, 272, 289, 294acidity and, 272, 277addition to C=C, 200addition to C=O, 221-236, 284alkylation of, 288as nucleophiles, 288configuration, 276cyclooctatetraenyl dianion, 275

Index 401

402 Index

Carbocations (contd.)stabilisation, 21, 84, 87, 104, 109,371,

382stereochemistry of, 79, 84, 86, 104structure, 104triphenylmethyl, 84, 87, 102, 103, 381tropylium, 104

Carbon acids, 270Carbonation, 284Carbon, electron-deficient migration to,

108Carbonyl group, 203-244

bond length, 11characterisation, 219conjugated, 12, 23, 200, 205, 223dipole, 23, 203hydrates, 207hydrogen bonding, 204, 209i.I. spectrum, 219nucleophilic addition, 31, 200, 203-

244protection, 210, 211protonation, 103, 204reactivity, 205reduction, 212, 214-219structure, 23, 203

Carboxylic derivs., reactions, 236--244acid hydrolysis, 240base hydrolysis, 238electronic effects in, 237Grignard reagents, 238leaving groups and, 237reduction, 215steric effects in, 238tetrahedral intermediates in, 236

Catalystsacid, 42,74base, 75cyanide ion, 231heterogeneous, 143, 191,264Lewis acid, 131, 189metal, 191nucleophilic, 99reaction pathway and, 41thiazolium ylids, 232

Cationic polymerisation, 189Cellulose oxidation, 37Chain length, 314, 324Chain reactions, 300, 314Chain transfer agents, 321Chair conformation, 355Charge transfer complexes, 131Chelation, 291Chirality, 87,116,235,254,278,326Chloral hydrate, 208Chromatography, 43, 258

Chugaev reaction, 268Cisoid conformation, 197,344,350Claisen ester condensation, 229Claisen rearrangement, 355Claisen-Schmidt reaction, 226Clemmensen reduction, 146Colour

conjugation and, 13, 331, 391Combustion, heat of, 12Competition experiments, 157Complexes, dissociation, 27Concerted reactions, 341

stereoselectivity in, 341symmetry allowed, 348symmetry forbidden, 348, 357

Configuration, 38, 88apparent retention, 94carbanions, 276carbocations, 104determination of relative, 88, 90inversion, 87, 90, 190oximes,124racemisation, 89, 90, 326radicals, 309retention, 93, 326

Conformationantiperiplanar, 118, 253chair, 355cisoid, 197, 344, 350eclipsed, 7, 254staggered, 7, 254synperiplanar, 253, 269trans diaxial, 255transoid, 197, 350

Conjugateacids, 53, 65bases, 53

Conjugate addition, 201Conjugated carbonyl comps., 12,23,

200,223Conjugated dienes, 11Conjugation, 11,368Conrotation,345Coordination polymerisation, 322Copolymerisation, 322Cope reaction, 268Cope rearrangement, 354Cracking, 112, 305Cram's rule, 235Cross-linking, 323Crossover experiments, 116, 122, 126Crowding, 8, 110, 162, 165, 183,206,235,

259, 261, 301, 311Cumene, phenol from, 128Curtius rearrangement, 122Curved arrows, 19

Cyanides, addition to, 244Cyanoethylation, 199Cyanohydrins, 212Cyclic bromonium ions, 180Cycloadditions, 341, 348-352

2ne + 2ne, 3484ne + 2ne, 348, 349symmetry allowed, 349symmetry forbidden, 348

Cycloheptatrienyl cation, 18, 104, 106Cyclohexadienyl cation, 159Cyclohexadienyl radical, 331Cyclooctatetraene, 16

dianion, 275n.m.r. signal, 18shape, 17

Cyclopentadieneacidity, 275in Diels-Alder, 197,350tricyclic dimer, 351

Cyclopentadienyl anion, 18, 275Cyclopropanes

from carbenes, 50, 266thermolysis, 337

Cyclopropanonehydration, 209intermediates, 294

Cyclopropenyl cation, 18, 106

DarzeDli reaction, 290Dealkylation,143Debromination, 264Decarbonylation, 145, 335Decarboxylation, 271, 285

carbanions in, 271,285cyclic T.S.s in, 286electronic effects on, 286trapping in, 286

Dehydrationacid-catalysed, 103,220,225,247,380base-catalysed, 225, 262

Delocalisation, 13amidines, 69benzene, 14, 130canonical structures and, 19carbanions, 21, 23, 229, 274carbocations, 21,83,84,87,104,110carbonyl comps., 23, 205carboxylate anions, 19,55,57conditions necessary for, 18dienes, 13, 194energy, 13, 16guanidine, 68hyperconjugation and, 25

Index 403

in a complexes, 132, 151, 164steric inhibition of, 26, 71,172tropylium cation, 18, 104, 106

{) steric parameter, 386/1G,34/1G+, 38, 152, 158, 342, 394/1G9

, 36, 58, 66, 394/1H, 16,34

bond energy and, 35/1H+, 38,239,241,342,394/1W, 58, 66, 394/1S,34/1S+,39, 79, 239, 241,342, 355,394/1S9 , 58, 60,63,66, 210,394Desulphurisation, 212Deuterium

exchange, 131, 158, 174,211,288kinetic isotope effect, 46, 136, 139, 288

Diamagnetism, 308Diaryls, synthesis, 333Diazoamino comps., 147

rearrangement, 148Diazo coupling, 26, 28, 146

amines, 26, 147electron density in, 148intramolecular, 148kinetics, 147, 148pH and, 146phenols, 147, 155steric effects in, 27

Diazoketones, rearrangement, 119Diazonium cations, 27,119,146

alkyl, 107, 120aryl, 121, 146coupling by, 27,146decomposition, 104, 107, 120, 333stability of, 120, 146nucleophilic substitution, 169, 306

Diazotate anions, 146Diazotisation

aliphatic arnines, 107, 119aromatic amines, 121, 146

Dichlorocarbenes, 4, 50, 267, 290Dieckmann reaction, 230Diels-A1der reaction, 175, 197, 340, 349

cyclopentadiene in, 197electronic effects in, 198, 349endo v.exo addition, 3504ne + 2ne, 349reversibility, 351secondary orbital interaction, 350stereoselectivi ty in, 198, 349steric effects in, 198, 350symmetry control in, 348trapping in, 175zwitterion intermediate, 351

E., 38Eclipsed conformation, 7, 254Electrocyclic reactions, 341, 344-348Electrolytic oxidation, 307Electrolytic reduction, 307Electromeric effect, 24Electron configuration, 3Electron density, 21, 26, 29, 393Electron-donating groups, 23, 26

addition to C=C and, 183addition to C=O and, 205, 206aromatic substitution and, 153, 158pinacol change and, 115

Electronegativity, 21, 22, 95Electrons, lone pair, 10,72Electron spin, paired, 2, 308Electron-withdrawing groups, 23

acidity and, 59, 61, 62, 272

Dienes, 11addition to, 194-198cisoid conformation, 197,350conjugated, 11Cope rearrangement, 354cyclisation, 346cycloaddition to, 348Diels-Alder reaction, 197,349excited state, 13heat of hydrogenation, 16, 194isolated, 11m.o.s of, 12polymerisation, 323

Dienone intermediates, 356Dienone/phenol rearrangement, 115Dienophiles,198,350Digonal hybridisation, 5Dimedone, 202Dimroth's Er parameter, 391

solvatochromic shifts, 391solvent polarity, 391Y and, 392

Dinitrofluorobenzeneproteins and, 172

1,2-Diolsformation, 189rearrangement, 113

1,1-Diphenyl-2-picrylhydrazyl, 301Diphenylpolyenes, 131,3-Dipolar addition, 192, 194,351Dipole moments, 22, 156, 165, 203Diradicals, 315, 330, 337

stable, 338Displacement, 30Disproportionation, 216, 305, 321Disrotation, 345Dissociative process, 241

405Index

alkene stability and, 261base size and, 261basicity/nucleophilicity in, 262change of pathway, 260ElIS/V1 ratio, 260E2/S/V2 ratio, 260entropy and, 262leaving group and, 261solvent and, 260steric effects in, 261, 262structure and, 260, 261temperature and, 262

Enamines,221End group analysis, 172Endo addition, 350Energetics ofreaction, 33Energy barriers, 37Energy profiles, 37Enolate anions, 279, 290, 295Enolisation, 201, 219, 225, 280, 297, 356Enthalpy, 34

of activation, 38Entropy, 34

cyclisation and, 36energy partition and, 35hydrogen bonding and, 36of activation, 39, 79rotational, 36temperature and, 36translational, 36, 58, 239

Epoxides,as intermediates, 94, 190hydrolysis, 190

Epoxyesters, afJ-, 290E,375Equilibrium

acidity and, 54basicity and, 65constant, K, 35control, 43,163,165,171,195entropy and, 36free energy and, 35

E.s.r. spectroscopy, 308spin trapping in, 309splitting in, 308

Es steric parameter, 386Ester hydrolysis

acid-catalysed,24Oacyl-oxygen fission, 47, 239alkyl-oxygen fission, 47, 241BAC2 pathway, 239, 384base-catalysed, 239, 366, 367, 374isotopic labels in, 47, 89, 238steric effects in, 241, 361, 385

EsterificationAAc1 pathway, 242, 379

o-/p-ratios, 159n complexes in, 131partial rate factors, 156selectivity in, 158a complexes in, 41, 131, 151, 159solvent and, 161steric effects in, 153, 158, 159, 162,

165substituent effects, 150-163thermodynamic control, 163, 164transition states, 135, 151, 158

1,3-Elimination, 293E1 elimination, 248, 381

alkene stability and, 249, 256carbocations and, 247, 248orientation in, 249Saytzev mode, 249S/V1 and, 248, 260steric effects in, 261structure and, 249v. E2, 249, 252

E1cB elimination, 249activation energy, 250carbanions and, 248, 249, 257electron-withdrawal and, 251in aldol dehydration, 225, 251in benzyne formation, 251isotopic exchange and, 250leaving group and, 251structure and, 251

E2 elimination, 251-260alkene stability and, 253, 256base strength and, 252bond strength and, 252conformation and, 253, 255electronic effects in, 257Hofmann mode, 256in cyclic comps., 255kinetic isotope effect, 252leaving group and, 252, 257orientation in, 256Saytzev mode, 256S/V2 and, 252, 254, 260solvent and, 252, 255stereoselectivity in, 253, 264steric effects in, 258strength of base and, 257variable T.S. in, 256

E",(1,1-)elimination,266bases and, 266carbenesin, 50, 266isotopic labelling in, 266

Ei elimination, pyrolytic, 268, 340Cope reaction, 268stereoselectivity, 268, 269

Elimination v. substitution, 100, 260

addition to C=C and, 183addition to C=O and, 205, 226aromatic substitution and, 151, 158basicity and, 67, 70elimination and, 251, 262

Electrophiles, 29Electrophilic addition to C=C, 31,178

194bromine, 51, 179carbocations, 113, 180, 184, 187, 188,

225cationic polymerisation, 189cyclic bromonium ions, 1801,3-dipolar, 192, 194,351effect of added nucleophiles, 179hydration, 187hydroboration, 187hydrogenation, 191hydrogen bromide, 184,201hydroxylation, 189inductive effect in, 185Lewis acids, 181mesomeric effect in, 182orientation, 184ozone, 192peroxyacids, 190n complexes in, 180, 184proton, 103, 184, 187rate, 183, 185rearrangements in, 185, 187solvent and, 182stereoselectivity, 51,180,182,186,

188, 189, 191steric effects, 181vinyl halides, 185

Electrophilic addition to C=C-C=C,194

Electrophilic substitution, aromatic, 31,130-167,381

1,2-v.1,4-addition, 195as addition/elimination, 133complexing with substituent, 160deuterium exchange, 131, 158electronic effects in, 148, 158, 159energetics of, 132, 136field effect in, 152hyperconjugation in, 153inductive effect in, 22, 152, 153, 156,

160intermediates, isolation, 136, 142internal, 381ipso, 161isotope effects, 46, 139kinetic control, 151, 163, 164Lewis acids in, 131, 138, 141, 144, 163mesomeric effect in, 154, 155, 164

Index404

406 Index Index 407

Esterification (contd.)AAC2 pathway, 241, 384AAL1 pathway, 241, 380acid-catalysed, 240, 378acyl-oxygen fission, 88, 240, 242alkyl-oxygen fission, 240isotope labels in, 88, 241steric effects in, 242

Estersacyloin condensation, 218Claisen condensation, 229hydrolysis, 238-244, 365, 367, 378,

384reduction, 215

E T solvent parameter, 391Ethanoate anion, 19Excited state

carbon, 4dienes, 13

Exclusion principle, 2Exo addition, 350Exocyclic methylenes, 234

Fats, rancidity, 328Favorskii rearrangement, 294Fenton's reagent, 306Ferrocene, 275Field effects, 22, 152, 364Flash photolysis, 304Fluorination, 121, 140, 170,315,326Free energy, Gibbs, 34

change and K, 35,359of activation, 38, 359standard, 35

Freezing point depressionbenzoic acid, 242ketones, 103mesitoic acid, 243nitric acid, 134

Friedel-Crafts reactionacylation, 143acylium ions in, 144alkylation, 141carbocations in, 108, 141, 145dealkylation in, 143formylation, 145intermediates in, 141, 144intramolecular, 243Lewis acids in, 108, 141, 144polarised complexes in, 108, 141, 144polyalkylation, 143rearrangements in, 108, 142, 143, 163solvent and, 143, 144thermodynamic control in, 163with alcohols, 142with alkenes, 142

with cyclic anhydrides, 146Frontier orbital approach, 344he, 156, 159, 333

Gatterman-Koch reaction, 145Gegen ions, 91General acid catalysis, 74, 208, 209, 220

297 'General base catalysis, 75, 208Gibbs free energy, G, 341,2-Glycols

cis, 189rearrangement, 113trans, 190

Gomberg reaction, 333Grignard reagents

addition to C=C--e=O, 201addition to C==N, 244addition to c=o, 221, 235, 238structure, 221

Ground statecarbon, 3dienes, 13halogenobenzenes, 156

Grunwald-Winstein equation, 389compound parameter, m, in, 390derivation of, 389diagnostic use of, 391limitations of, 391solvent parameter, Y, in, 389standard reaction for, 389standard solvent for, 389

Gutta percha, 323

Half-life, 301Haloform reaction, 237, 296Halogenation

alkanes, 300, 323alkenes, 179, 186,313benzene, 138, 316ketones, 295

Hammett equation, 362additional parameters, 374, 388, 395derivation of, 362deviations from, 375empirical nature of, 395implications of, 394reaction pathway, and, 375solvent effects and, 388spectroscopic correlations, 392standard reaction for, 362, 395steric effects and, 361, 383thermodynamic implications of, 394

Hammett plots, 359change in rate-limiting step and, 383change in reaction pathway and, 378

downward deviations in, 380solvent effects and, 388steric effects and, 361, 383upward deviations in, 375uses of, 374

Hammond's principle, 137Hard bases, 96Heat of combustion, 12Heat of hydration, 97Heat of hydrogenation, 12, 16

benzene, 16cyclohexene,16cyclooctatetraene, 16cyclooctene, 16dienes, 16, 194

Heat of reaction, 34Heisenberg principle, 2Hemi-acetals, 209Heterolysis, 20, 129, 178,299Hexacene, 331Hofmann

elimination, 256reaction of amides, 49, 122

HOMO, 344Homolysis, 20, 129, 179,299Hybridisation, 4Hydration

C=C,42,187C=O,207heat of, 97

Hydride transfer, 108, 109,214,215,216Hydroboration, 187Hydrogen abstraction, 309, 316, 321Hydrogenation

C=C,191C==C,191C=O,214heat of, 16homogeneous catalysis in, 192stereoselectivity in, 191

Hydrogen bondingacidity and, 63, 64C=O and, 204, 209, 286intermolecular, 36, 282intramolecular, 36, 208, 281solvation and, 57, 67, 252, 282

Hydrogen peroxideFenton's reagent, 306hydroxylation of C=C, 189oxidation of ketones, 127

Hydrogen shifts1,3-photochemical, 3541,5-thermal,3531,7-photochemical,3541,7-thermal, 354

Hydrolysis,

esters, 238, 374halides, 77

Hydroperoxidesformation, 328in autoxidation, 328rearrangement, 128

Hydroxamic acids, Lossen degradation,123

Hydroxylation, 189Hyperconjugation, 25

alkene stability and, 26aromatic substitution and, 153carbocation stability and, 83

Hypochlorites, alkyl, 327Hiickel's rule, 17, 106Hund's rule, 3, 275

Imino-ethers, 245Inductive effect, 21

acidity and, 273basicity and, 66electronegativity and, 22electrophilic substitution and, 152,

156, 160Inductomeric effect, 24Insertion reactions, 50, 266, 267Intermediates

bridged, 105, 118, 129,335,377catalysts and, 42criteria for, 49cyclic, 180, 189, 227isolation, 48,136,142,171,290models for T.S., 41, 51, 78, 137, 151spectroscopic detection, 50, 144, 171,

181, 219, 237study of, 49tetrahedral, 236trapping, 50,175,286,356Wheland, 41,131,151

Intermolecular rearrangements, 143,149,278

Intimate ion pairs, 91, 249, 291Intramolecular rearrangements, 116,

122,126,127,128,217,232,279,355

Inversionin SN1, 90in SN2, 88, 89, 190

Iodoform, 297Iodonium ions, cyclic, 186Ionisation

t.GB and, 58solvent and, 56, 102, 388temperature and, 64

Ion pairs, 20, 45, 57, 64, 79,102,136,141

408 Index Index 409

Ion pairs (contd.)intimate, 91, 249, 291solvation, 45, 57, 81, 390solvent separated, 91

Ipso substitution, 161, 169steric effects in, 162

Isocyanate intermediates, 49, 122Isomer distribution, 158Isomerisation, cis-+trans, 315Isoprene, 323Isotope effects, kinetic, 46, 83, 139, 252,

288,295Isotopic labels, 47

bromide, 141carbon, 48, 108, 233, 355deuterium, 47,117,131,174,216,

217,224,233,250,264,278,288iodine, 89monitoring, 48, 89nitrogen, 170oxygen, 47, 89, 125, 127, 189,207,239scrambling of, 108, 141, 169sulphur, 165

K, 35, 359k, 38, 359, 374K., 54, 65

AGe and, 58Kb ,65

AGe and, 66Ketenes, 119p-Keto-acids, decarboxylation, 286Keto-enol tautomerism, 201, 219, 225

carbanions in, 278catalysis of, 277equilibrium and structure, 280, 297hydrogen bonding and, 281mechanism of, 278solvent and, 282

Ketones, halogenation, 295bromination, 41, 51, 76,297electronic effects in, 296enolisation in, 297general acid catalysis, 297general base catalysis, 76orientation in, 296, 297rate of, 295, 296, m

Ketoximes, Beckmann rearrangement,123

Ketyls, 218Kinetic acidity, 280Kinetic control, 42, 151

addition to C=O, 235addition to dienes, 195Diels-Alder,350

nitration, 43protonation, 283

Kinetic data, interpretation, 40, 44, 78Kinetic isotope effects, 46

carbon, 47chlorine, 47deuterium, 46, 83, 252, 288, 295prirnnary,46, 252,288, 295secondary, 83

Kinetics of reaction, 36isotope effects in, 46measurement of, 39, 89rate-limiting step in, 39

Knocking, 305Knoevenagel reaction, 228Kolbe electrolytic synthesis, 307Kolbe-Schmidt reaction, 291

Lactams,126Lactone intermediates, 94, 127, 228Lateral overlap, orbitals, 9Lead a1kyls, 301

anti-knock,305thermolysis, 301, 304, 324

Leaving groups, 98, 99, 127, 139, 237ability of, 98, 251bromobenzene sulphonate ion, 375cyanide ion, 232ethanoate anion, 228ethoxide ion, 229hydride ion, 168hydrogen bonding and, 252hydroxyl ion, 225, 253in elimination, 247, 251, 253internal, 100nitrogen, 100, 104, 107,114,121,123,

169,305protonated, 103, 125relative ability, 98, 251, 253tosylate, 89, 98,100,253triflate, 98

Levelling effect, water, 55Lewis acids, 29Lewis bases, 29Light absorption, colour and, 13Lindlar catalyst, 191, 223Linear free energy relationships, 358

395Lithium

a1kyls, 223, 238, 293aryls, 223, 233, 293

Lithium aluminium hydride, 214Localised orbitals, 6Lone pairs, 10, 72

in neighbouring groups, 94protonated, 99, 103

Lossen rearrangement, 123LUMO,344

Magnetic moment, electronic, 308Markownikov addition, 184Mass spectrometry, 176m, compound parameter, 390

diagnostic use of, 391Mechanistic borderline, 91Meerwein-Ponndorf reaction, 215Meisenheimer complexes, 171Mesomeric effect, 23,154,156,172,238,

278Meta (m )-directing groups, 150Metal alkyls

carbonation, 284thermolysis, 304Wurtz reaction and, 289

Michael reaction, 200Migration

origin, 116terminus, 116to electron-deficient C, 109, 110-119to electron-deficient N, 122-126to electron-deficient 0, 127-129

Migratory aptitude, 114, 129conformation of T.S. and, 118

Mixed SNlISN2 pathway, 92Models for T.S., 41, 51, 78,137,151,

194,254Molecularity,79Molecular orbitals, 5, 343Molownides, 193Monomers, 321Mustard gas, 96

Naphthalene, 17nitration, 164sulphonation, 164

N-bromosuccinimide,327Neighbouring group participation, 93-

96,181,377Neopentyl rearrangements, 110Newman projections, 7,235Nitration, 31, 45,133-138

dilute RN03 in, 137isotope effects in, 46kinetics of, 134, 135naphthalene, 164nitronium fluoroborate in, 135NO? in, 134

Nitrating mixture, 133Nitrenes, 21, 122

carbonyl, 122Nitriles, reactions, 244

Nitroalkanesacidity, 272, 280addition to C=O, 226tautomerism, 283

Nitrodealkylation, 162Nitrodehalogenation, 162Nitrogen, electron-deficient, migration

to, 120Nitronium ion, 45, 103, 134Nitrosating agents, 119Nitrosation

phenol,137prirnnary aliphatic amines, 107, 119primary aromatic amines, 120secondary amines, 121tertiary amines, 121

Nitrosonium ion, 120, 137Nitrosotrialkylammonium cation, 121Nitrous acid, protonated, 120N.m.r. spectroscopy, 18,48, 102, 106,

111,126,129,159,181,219aromaticity and, 1813C, 49, 106, 111,393carboxyl protonation and, 240Grignard reagents and, 221ozonolysis and, 193, 194ax and, 393triphenylmethyl dimer, 44, 301

N-nitroso compounds, 121Nodal

plane, 3, 343surface, 2

Nodes, 343Non-bonded interaction, 7Nucleophiles, 29

ambident, 97Nucleophilic addition to C=C, 198

carbanions in, 199cyanoethylation, 199hydrogen cyanide, 199methanol, 199Michael reaction, 200

Nucleophilic addition to C=C-C=O,200

conjugate addition, 201Grignard reagents, 201Me3SiCN, 213Michael reaction, 202steric effects in, 201

Nucleophilic addition to C=O, 31, 203-244

acetylide ions, 223acid catalysis, 204, 207,209,220,225addition/elimination, 219alcohols, 209aldol reaction, 224

410 Index

Nucleophilic addition to C=O (contd.)ammonia derivs., 219base c~talysis, 204, 207, 212, 216, 226benzom condensation, 231bisulphite anion, 207, 213Cannizzaro reaction, 216carbanions,221-234Claisen ester condensation, 229Claisen-Schmidt reaction, 226conjugate, 200, 213cyanide ion, 212Dieckmann reaction, 230electronic effects in, 205, 208, 226electrons, 217Grignard reagents, 221, 235halide ion, 214hydration, 207hydride ion, 214hydrogen bonding in, 204, 209in carboxylic derivs., 236-244intermediates in, 50, 219intramolecular, 217, 232irreversible, 215, 222Knoevenagel reaction, 228Lewis acids in, 204, 222Meerwein-Ponndorf reaction 215Me3SiCN, 213 'nitroalkanes, 226Perkin reaction, 227pH and, 204, 208, 219protection, 211rate-limiting step, 205, 212, 216, 220reversibility, 206, 210, 212, 224size of nUcleophile in, 207, 213spectroscopy and, 50, 219stereoselectivity, 234steric effects in, 205, 213, 222, 243Stobbe reaction, 228thiols,211T.S. in, 205Wittig reaction, 233

Nucleophilic catalysis, 99Nucleophilicity, %, 211

basicity and, 96soft bases and, 96

Nucleophilic substitution, aliphatic 3145, 77-100 ' ,

AgEB catalysis, 97allyl halides, 85ambident nucleophiles, 97benzyl halides, 84, 91bridgehead halides, 86bromomethane,782-bromopropanoate, 941-bromotriptycene,87carbanions in, 100,288

charge distribution in T.S., 78, 80, 83,84

1,2-chlorohydrins,942-chloro-2-methylpropane, 78configuration in, 88, 89, 90, 92, 93, 94electronic effects in, 82, 83, 84entering group and, 96H EB catalysis, 99Ie catalysis, 98isotopic label (I) in, 89kinetics of, 45, 77,89leaving group and, 98list of reactions, 99mechanistic borderline, 91mechanistic changeover, 82, 84, 96neighbouring group participation, 93neopentyl halides, 86rate-limiting step, 78SN1, 79, 90SN2, 78, 87, 93SN1 v. SN2, 80, 81, 91, 96SNi,92solvation in, 45, 79, 80, 90, 91, 97solvolysis, 80, 91steric effects in, 82, 84, 85, 86, 109stereochemistry of, 87-96structure and, 82-87, 90tosylates, 89triphenylmethyl halides, 84v. elimination, 100, 248, 260vinyl halides, 85

Nucleophilic substitution, aromatic,167-177

activated aryl halides, 170anionic intermediates, 168,170,171aryne intermediates, 173as addition/elimination, 172as elimination/addition, 175deuterium exchange, 174diazonium salts, 121, 169ipso, 169nitrobenzene, 168pyridine, 168steric inhibition of delocalisation, 173

Nylon-6, 126

Oleum, 140Oppenauer oxidation, 216Orbitals

anti-bonding, 6, 292atomic, 1axial overlap, 6bonding, 6degenerate, 3delocalised, 13dumb-bell, 3

filled, 17frontier, 344HOMO, 344hybridisation of, 4lasral overlap, 9lobes, 9localised, 6LUMO,344molecular, 5nodal planes and, 1,343overlap integral, 5, 342p,2phase of, 342:rr, 9,337:rr", 9, 337s,2shape, 30, 60",6size, 2secondary interaction, 350sp1, 5sp2,5sp3,4spatial orientation, 2, 3symmetry, 342

Order of reaction, 39, 79first, 39mixed, 82, 91molecularity and, 79second, 39

Organometallic compounds, structure,221,276,293

Orthoestersacetals from, 210, 289hydrolysis, 75

Ortho/para (o-/p-)directing groups, 150, 159ratios, 159

Osmic esters, cyclic, 189Osmium tetroxide, 189Overlap

axial, 6integral, 5, 342lateral, 9

Oxaphosphetanes, 233Oximes

configuration, 124, 220formation, 219rearrangement, 123

Oxygendiradical,315migration to electron-deficient, 127

Ozonides, 192Ozonolysis, 192

stereochemistry, 193

Index 411

Paramagnetism, 308Paint, hardening, 328Partial rate factors, Ix, 156, 159, 333Pauli principle, 2Pericyclic reactions, 198, 341Perkin reaction, 227Peroxide effect, 317Peroxides

as initiators, 300, 317heterolysis, 129, 306homolysis, 129,333in ozonolysis, 193photolysis, 304rearrangement, 127thermolysis, 304trans-annular, 331

Peroxyacids, 190,330Peroxy radicals, 315, 328, 337Peroxy zwitterions, 193Petroleum cracking, 112, 305Phase, orbital, 342Phenanthrene, 17

Pschorr synthesis, 334Phenol

acidity, 23, 61, 370coupling, 334diazo coupling, 147, 155from cumene, 128nitration, 137oxidation, 334

Phenonium ion, 105,376Phenylation, 332Phenylnitromethane, tautomerism, 283Phenyl radicals, 332Phosphine oxides, 234Phosphonium ylids, 233Phosphoranes,233Phosphorus-oxygen bond, 233Photochemical concerted reactions, 341,

3461,3-hydrogen shifts, 3541,7-hydrogen shifts, 3542:rre + 2:rre, 349

Photochemical initiation, 300Photo-dimerisation,337Photolysis, 303Photo-oxidation, 330Physical methods

analysis of products, 43, 156, 281detection of intermediates, 50, 144,

171,176,181,219,304,308structure determination, 102, 106,

111,134,193,221,240:rr bond, 9, 178:rr complexes, 131, 180, 184:rr-deficient heterocycles, 165

Jr-excessive heterocycles, 166Pinacolinic deamination, 113, 118Pinacol/pinacolone rearrangement, 113

migratory aptitude in, 114Pinacols

formation, 218rearrangement, 113

pK., 54,270, 362temperature and, 64

pKb ,65pKB~' 65Plane trigonal hybridisation, 5Polarisability, 24, %Polarisation, 22, 29Polarised complexes, 108, 141, 144Polar non-protic solvents, 81Polyenes

hydrogen shifts in, 352Polyisoprenes, 323Polymerisation

anionic, 200, 226branching, 321cationic, 189chain length, 321chain transfer, 321coordination, 322copolymerisation, 322cross-linking, 323induction period, 321initiation, 321propagation, 320radical, 308, 320stereochemistry, 322, 323termination, 320

Products, nature of, 43Propane, rearrangement, 108Protecting groups, 155,210,211,265

requirements, 211Protodesilylation, 149, 161Protodesulphonylation, 140, 161Proton transfer

catalysed,74to C=C, 103to lone pairs, 103, 112, 116

Prototropy, 277Pschorr reaction, 334Pyridine

basicity, 72delocalisation in, 18, 165electrophilic substitution, 165nucleophilic substitution, 168

Pyrolytic elimination, 267, 340carbocationic character T.S., 269,340Cope, 268Chugaev, 268Ei,261

types of, 30Reaction constant, p, 363

attenuation of, 368determination of, 363effect of solvent on, 388physical significance of, 367rate-limiting step and, 368, 381reaction centre and, 368reaction pathway and, 378sign of, 368values of, 364variation with solvent, 388

Reaction mechanism, investigation, 43,375

Reactive methylenes, 288Reagents, classification, 28Rearrangements, 30, 32,109-129,352-

357alkanes, 108a1kenes, 112allylic, 109aryl, 128Beckmann, 123benzilic acid, 232carbanions, 292carbocations,109-119Claisen, 355configuration in, 116, 123conformation in, 115, 118Cope, 354Curtius, 122diazoamino comps., 148diazoketones, 119dienone-phenol, 115Favorskii,294Hofmann, 122hydroperoxides, 128in Friedel-Crafts, 108, 142, 145, 163intermolecular, 117, 143, 149,278intramolecular, 116, 122, 126, 127,

279,355Lassen, 122migratory aptitude in, 114neopentyl, 110pinacol-pinacolone, 113radical, 335Schmidt, 122sigmatropic, 352stereochemistry of, 116, 117, 119steric effects in, 115Stevens, 293Wagner-Meerwein, 111Wittig, 293Wolff, 119

Redox reactions, 306Reirner-Tiemann reaction, 290

412 Index

SYN,267Pyrrole

basicity, 73electrophilic substitution, 166protonation, 73

Quantum numbersprincipal,n,2spin, 2subsidiary, I and m, 2

Quinuclidinebasicity, 72complex with Me3B, 28

RacemisatioD, 89deuterium exchange and, 288in radical reactions, 326in SN 1, 90

Radical addition, 312-323carbon tetrachloride, 320halogens, 313hydrogen bromide, 316sulphenyl halides, 320vinyl polymerisation, 320

Radical anions, 218Radical rearrangements, 335Radicals, 20,30,299-339

acyl, 306, 330, 335addition to C=C, 313-323alkoxyl,303alkyl, 301, 303, 304, 324a1lylic, 311, 325, 327, 329benzoyl, 330benzylic, 311, 316, 329biradicals, 315, 330, 337bridged, 310chain reactions, 300, 313, 314, 328combustion and, 299conformational equilibrium, 319cycloheptatrienyl, 308cyclohexadienyl,331detection, 308dimerisation, 300, 305, 311, 313, 314,

320, 332, 3341,1-diphenyl-2-picrylhydrazyl, 301disproportionation, 305, 313, 320, 332e.s.r. spectroscopy and, 308formation, 303half-life, 301, 305halomethyl, 320hetero,302hydroxyl,306in acyloxylation, 333in aromatic substitution, 331, 334in arylation, 332in autoxidation, 306, 318, 328

in Barton reaction, 337in Gomberg reaction, 333in halogenation, 323in hydroxylation, 332inhibitors, 300, 318, 321initiators, 314, 321in phenol oxidation, 334in polymerisation, 308, 320in Pschorr reaction, 334oxygen and, 300paramagnetic, 308pentachloroethyl,314perbenzoate,330phenoxy, 302, 334peroxy,315,328,337polar effects in, 325rearrangement, 325shape, 309, 326solvent and, 309, 324stability, 302, 309, 312, 324stereoselectivity and, 318, 326, 333terminators, 300, 321thiyl, 302, 319trapping, 302triphenylmethyl, 43, 300, 306, 311

Radical substitution, 323-335aromatic, 331autoxidation, 328halogenation, 323

Radiolysis, 304Raman spectra, 134Raney nickel catalyst, 212Rate constant, k, 38, 39Rate equation, 39

mixed order, 82, 91T.S. and, 41

Rate-limiting step, 39, 40, 134, 148, 171,381

Rate of reactionactivation energy and, 37catalysts and, 41measurement of, 39a and, 366temperature and, 38

Reactioncollisions in, 38concerted,341energetics of, 33, 37energy profile of, 37heat of, 34intermediates, 38, 49kinetics of, 36molecularity, 79order, 39,79rate constant, 38rate-limiting step, 39

Index 413

414 Index

Relative configuration, determination,88

Resonance energy, 17Retro Diels-Alder reaction, 351Retro pinacol rearrangement, 115p,363p·,385Rotational entropy, 36Rotation frequency, 8Rubber

natural, 323perishing of, 328synthetic, 189,322vulcanisation, 323

Sandmeyer reaction, 306Sandwich compounds, 275Sawhorse projections, 7Saytzev elimination, 249, 256Schiff bases, 221Schmidt rearrangement, 122Selectivity, 156, 169, 326a,362a-,370a+,372a;', 385a bonds, 6a complexes, 41, 131Sigmatropic rearrangements, 352-357

antarafacial, 353carbon shifts, 354hydrogen shifts, 352orbital symmetry in, 352photochemical, 354suprafacial, 353thermal, 353

Silicon-oxygen bond, 213SE I pathway, 288, 295SNI pathway, 79, 370, 389SN 2 pathway, 78, 376SN 2' pathway, 110SN 2 (aromatic) pathway, 170SNi pathway, 92Sodium borohydride, 215Soft bases, 96Solvation, 45, 56, 79, 253, 299, 388

H-bonded, 57, 67, 81, 253~Se and, 58, 60, 63, 66polar non-protic. 81. 173.252

Solvatochromic shifts, 391Solvent

bipolar non-protic, 81, 98, 173,252effect of, 45, SO, 161,252,260,388Er and, 391ionising power, 390ion-solvating ability, 80, 260, 390

parameter, Y, 389Solvent separated ion pair, 90Solvolysis, SO, 91, 109,389Specific acid catalysis, 74, 209Specific base catalysis, 75Spectroscopic correlations 39211, ,

ax and C n.m.r., 393ax and 1H n.m.r., 393ax and i.r. shifts, 392

Spin, electronic, paired, 2Stabilisation energy, 13Stability, thermodynamic, 12Staggered conformation, 7Standing waves, electronic, 342Stereochemical criteria, 51Stereoselectivity, 52

in acetolysis, 377in addition to C=C, 51, 180, 182,318in addition to c=O, 234in Beckmann, 124in concerted reactions, 341in Dieis-Alder, 198,349in elimination, 253, 264, 267in ozonolysis, 193in rearrangements, 118

Steric effects, 26crowding, 27, 115,235,301,350delocalisation and, 26, 71,172in acidity, 58in addition to C=C, 181in addition to C=O, 205, 235in aromatic substitution, 152, 159,

162,165in diazo coupling, 27in elimination, 261in ester hydrolysis, 241in rearrangements, 115non-bonded interaction, 7

Steric hindrance, 27, 79, 110, 159, 162,222,235,243,301,312,356

Steric parameters, 386Es,3866, 387

Sterigmatocystin, 48Stevens rearrangements, 293Stobbe reaction, 228Substituent constant, a, 362

'constancy' of, 368, 372determination of, 363physical significance of, 364pK. and, 362polar effects and, 366sign of, 364spectroscopic shifts and, 392values of, 363

Substituent effects, 150, 388

Substitutionelectrophilic, aromatic, 41,130-167nucleophilic, aliphatic, 31, 77-100nucleophilic, aromatic, 167-177radical,323-335

Sulphenyl compounds, 320Sulphonation, 29, 140, 164Sulphonium salts, 95, 99

cyclic, 95elimination from, 258

Sulphur trioxide, in sulphonation, 140Sulphuryl chloride, in chlorination, 327Super acids, 102, 129, 181Suprafacial shifts, 353Symmetry controlled reactions, 340-357

activation parameters, 342'allowed', 348antarafacial shifts, 353Claisen rearrangement, 355concerted v. stepwise, 342, 351conrotation,345Cope rearrangement, 354cycloaddition, 348Diels-Alder reaction, 349disrotation, 345electrocyclic, 344'forbidden', 348residual bonding in, 342secondary orbital interaction, 350sigmatropic rearrangements, 352suprafacial shifts, 353

Symmetry, orbital, 344Synperiplanar conformation, 253, 269

Taft equation, 384amide hydrolysis and, 388derivation of, 386{) parameter in, 387Es parameter in, 386ester hydrolysis and, 384polar effects in, 385p. in, 385at.. in, 385standard substituent in, 385

Tautomerism, 277catalysis of, 277equilibrium and structure, 280intermolecular, 278intramolecular, 278, 279keto-enol, 201, 219, 225, 277mechanisms of, 278nitroalkanes, 277, 283rate and structure, 279rate-limiting step, 279

Terminators, 300, 320

Index 415

Tetraalkylammonium saltsbasicity, 67elimination from, 254, 256

Tetrahedral hybridisation, 4Tetrahedral intermediates, 236, 361

isolation, 237Thermal concerted reactions, 341

Claisen rearrangement, 355Cope rearrangement, 3541,5-hydrogen shifts, 3531,7-hydrogen shifts, 3544Jre + 2Jre, 348

Thermodynamic control, 43, 163addition to c=O, 235addition to dienes, 195Diels-Alder,350Friedel-Crafts, 163napthalene sulphonation, 43, 164nitroalkane formation, 283

Thermodynamics, second law, 34Thermodynamic stability, 12

alkenes, 26, 249benzene, 15delocalisation and, 26dienes, 12hyperconjugation and, 26keto-enol forms, 282

Thiazolium ylids, 232Thioacetals, 211

desulphurisation,212Thiols, addition to c=O, 211Thionyl chloride, in chlorination, 92Through-conjugation, 368

rand,373a- and, 370a+ and, 372, 393

Transesterification, 239Transition state, 24, 38

bridged, 118, 129,292,335composition of, 41conformation and, 118, 235crowding in, 27, 83, 86, 183,206,235,

259,385cyclic, 216, 222, 254, 268, 287,352,

356energylevelof,4O,137,235,283models for, 41, 49,137,151orbital interaction in, 350organisation in, 39residual bonding in, 342solvation of, 253variable, 92, 256, 257

Translational entropy, 35, 58, 239, 241Transoid conformation, 197Trapping of intermediates

arynes, 176

416 Index

Trapping of intermediates (contd.)carbenes, 50, 267dienones,356internal, 266peroxyzwitterions, 193

Trichloromethane hydrolysis, 42, 267Triphenylmethane

acidity, 271carbanion from, 271, 272radical from, 43, 300

Triphenylmethyl cation, 84, 87, 102,103,381

reduction, 306Triphenylmethyl peroxide, 300

rearrangement, 336Triphenylmethyl radical, 43, 300

dimer,44,301,311shape, 311

Tropylium cation, 18, 104, 106stability, 106

Tschitschibabin reaction, 168

Ultra-violet absorption, 13, 132,218Umpolung, 211Unsaturated acids

exf3-, from Perkin reaction, 227f3y-, synthesis, 234

Unsaturated carbonyl compounds, exf3-,addition to, 200formation of, 251

Van der Waals radii, 8Vibrational modes, 342Vinyl ethers, polymerisation, 189Vinyl polymerisation, 320

branching in, 321chain length in, 321coordination, 322induction period in, 321

initiation, 320, 321propagation, 320termination, 320, 321

Vulcanisation, 323

Wagner-Meerwein rearrangements, 111Water

addition to C=C, 187addition to C=O, 207autolysis of, 54ion solvation and, 57, 58, 60levelling effect in, 55polarisability, 57

Wave amplitude, 343Wave functions, 2, 342Wave nodes, 343Wheland intermediates, 41, 131, 151Wittig reaction, 233Wittig rearrangement, 293Wolff rearrangement, 119Woodward-Hoffmann rules, 344Wurtz reaction, 289

Xanthates, pyrolysis, 268X-ray crystallography

ethanoate anion, 19Grignard reagents, 221Meisenheimer complexes, 171triphenylmethyl radicals, 311

1r6ds,232,233,287Y, solvent parameter, 389

values of, 390Yukawa-Tsuno equation, 372

parameter, T, and, 373

Zinc alkyls, 223Zwitterions, 94, 176

diazoanthranilate anion, 176peroxy,193

Notes

Documents

A Guide Book to Mechanism in Organic Chemistry_OCR