Metal triflate catalysed organic transformations
by
Michelle Claire Lawton
Thesis submitted in fulfillment of the requirements for the degree
Philosophiae Doctor
in
Chemistry
in the
Faculty of Science
of the
University of Johannesburg
Promoter Prof DBG Williams
June 2009
Contents
Synopsis i ndash ii
Abbreviations iii ndash v
List of figures and tables vi ndash xix
Chapter 1 Lewis acids-A Literature Overview
Section Heading Page
11 The history of the acid-base theory 1 ndash 6
12 Lewis acids in catalysis ndash a focus on metal triflates 6 ndash 44
13 The ranking of Lewis acids - introduction 44 ndash 45
131 Thermodynamic measurements 45 ndash 51
132 Spectroscopic measurements 51 ndash 56
133 Comparative studies 57 ndash 64
14 Lewis acids in aqueous media 64 ndash 75
15 Conclusions 75 ndash 76
16 References 77 ndash 82
Chapter 2
Metal triflates in protection group chemistry
Section Heading Page
21 Introduction 83
22 Acetal formation using aluminium triflate 84 ndash 91
23 Catalyst recycling 91 ndash 92
24 Deprotections 92 ndash 97
25 Other protecting groups 97 ndash 99
26 Other metal triflates 99 ndash 100
27 Tetrahydropyranyl ethers (THP) 100 ndash 104
28 Two protection groups 104 ndash 106
29 Conclusions 107
210 References 108 ndash 109
Chapter 3
The drying of organic solvents
Section Heading Page
31 Introduction 110 ndash 111
32 Choice of analytical method ndash Karl Fischer 111 ndash 114
33 Limits of the Karl Fischer method 114 ndash 116
34 Choice of solvents 116 ndash116
35 Methods of drying the solvent 116 ndash 117
351 Traditional drying methods 117 ndash 119
352 Drying reagents 119
353 Preparation of desiccants 119
36 The alcohols 120 ndash 126
37 Acetonitrile 126 ndash 128
38 Dichloromethane (DCM) toluene and tetrahydrofuran
(THF) 128 ndash 134
39 Discussion 134
310 Conclusions 134 ndash 135
311 References 136
Chapter 4
Investigations on the role of water in metal triflate catalysed reactions
Section Heading Page
41 Introduction 137 ndash 138
42 The arguments for and against triflic acid as the true
catalyst taken from the literature
138 ndash 145
43 The Mukaiyama aldol Reaction 146
431 Introduction 146
432 Metal salt catalysed Mukaiyama aldol reaction 146 ndash 152
433 The possible role of water in the Mukaiyama Aldol
reaction
152 ndash 158
44 A perspective of metal triflates in organic solvents 158
441 On solvation of the metal triflates in dry organic
solvents
159 ndash 162
442 The possible role of water in organic solvents 162 ndash 163
45 The case for 26-di-tert-butyl-4-methyl-pyridine 163 ndash 165
46 Summary 165 ndash 165
47 Carbocation formation 165
471 Carbocation formation in wet and dry solvents 165 ndash 175
472 The proton and the sterically hindered base ndash Xray
crystallography
176 ndash 177
48 Friedel-Crafts alkenylation reactions of arenes 177
481 Optimising the reaction 177 ndash 179
482 Reactions in dry solvent 179 ndash 181
49 Conclusions 182 ndash 183
410 References 184 ndash 185
Chapter 5
Ranking of Lewis acids
Section Heading Page
51 Introduction 186
52 Lewis acidity from NMR resonance shifts 187 ndash 202
53 Infrared Spectroscopy (IR) 202 ndash 206
54 Conclusions 207
55 References 208
Summary of conclusions and future research 209 ndash 211
Chapter 6
Experimental data and characterisation
Section Heading Page
61 Standard experimental techniques 212
611 Chromatography 212
612 Anhydrous solvents and reagents 212
62 Spectroscopical and spectrometrical methods 212
621 Nuclear Magnetic Resonance Spectroscopy (NMR) 212 ndash 213
622 Mass spectroscopy (mz) 213 ndash 213
623 Infrared spectroscopy (IR) 200 ndash 200
63 Melting Points 213
64 Chemical methods 214
641 General procedure for acetal formation 214
642 General procedure for TBDMS protection of alcohols 226
643 General procedure for the tetrahydropyranylation of
alcohols
227
644 General procedure for TBDMS protection of alcohols 229
645 Standard procedure for Karl Fischer measurements 232
646 TMS protection of propiophenone 234
647 General procedure for the Mukaiyama Aldol reaction 234
648 General procedure for carbocation formation 236
649 Synthesis for [bmim][OTf] 236 ndash 237
6410 Crystal data 238 ndash 239
6411 General procedure for Friedel Crafts alkenylation 239
6412 Ranking of Lewis acids using NMR spectroscopy 240
6413 Determination of equilibrium constants 241
6414 Ranking of Lewis acids by Infrared Spectroscopy (IR) 241
65 References 242
Appendix A 243 - 283
Synopses The research described in this thesis was directed at advancing the application of metal
triflates Al(OTf)3 in particular in organic synthesis on the one hand and to contribute to
the understanding of the underlying basis for their catalytic activity The study was
undertaken against the background and on the bases of a detailed literature study of metal
triflates their chemical and catalytic properties and applications thereof Amongst others
it deals with the possible role of metal-bound water that give rise to Broslashnsted type acidity
and that this induced Broslashnsted acidity may be responsible for the catalytic activity that is
observed
The study was prompted by the realisation that Al(OTf)3 was largely neglected as a
potential reusable catalyst This is in marked contrast to the attention paid to other metal
triflates the rare earth metals in particular Earlier work in this laboratory has shown that
Al(OTf)3 is stable in water from which it can be recovered easily for reuse In addition it
showed promise as a Lewis acid catalyst and is relatively soluble in several organic
solvents
New applications for the use of Al(OTf)3 have now been demonstrated These include the
efficient formation of acetals from aldehydes and ketones The conversions can be carried
out in an alcoholorthoester mixture or preferably in neat orthoester Other metal triflates
notably Sc(OTf)3 and In(OTf)3 are useful alternative catalysts Al(OTf)3 can be easily
recycled without loss of activity This methodology also can be applied to aldehydes and
ketones containing TBDMS groups without effecting deprotection of the ethers In view
of the sensitivity of the TBDMS groups to hydrolysis in the presence of triflic acid the
results suggest little hydrolysis (or alcoholysis) of the metal triflates in the protic solvents
used which would generate trifluoromethanesulfonic acid as a consequence of such
metal based hydrolysis Al(OTf)3 was also found to be a good catalyst for the formation
of THP ethers It proved to be excellent for Friedel Crafts reactions using alkynes as
substrates Al(OTf)3 together with other triflates offers a mild alternative to the more
traditional water sensitive Lewis acids eg BF3 AlCl3 and TiCl4 which are difficult to
recover and require the use of extremely dry solvents
i
ii
On the basis of published literature it was known that metal triflates showed catalytic
activity in both aqueous and non-aqueous solutions In aqueous solutions the possibility
of catalysis by a dual mechanism namely Lewis and Broslashnsted acid catalysis cannot be
ruled out Nevertheless some metal triflates can be regarded as essentially Lewis acids
while others can be regarded as essentially Broslashnsted acids when operating in aqueous
environments Evidence was obtained that Al(OTf)3 tended to function to a significant
extent as a Broslashnsted acid in aqueous environments Here generation of retinyl cations in
the presence of Al(OTf)3 was used as the probe reaction These cations are intensely blue
coloured and consequently easily detected It was found that the cation generation
proceeded exclusively as a Broslashnsted-type acid-promoted process In order to establish if
Al(OTf)3 can function purelyessentially as a Lewis acid in an organic solvent it was
necessary to investigate methods for the drying of organic solvents to preclude the
formation of H2O-M(OTf)3 adducts which would be strongly Broslashnsted acidic The most
successful methods were then applied in further work It could thus be demonstrated that
Al(OTf)3 shows Lewis catalytic activity in extremely dry (low ppm water content)
organic solvents specifically for the Friedel-Crafts reactions mentioned above
Finally several approaches towards the quantitative measurement of Lewis acidity were
applied and evaluated The application of these approaches to establish a ratingranking
of Lewis acidity of metal triflates was complicated by the low solubility of these
compounds in most aprotic organic solvents Another difficulty was to identify suitable
probes that could be applied over a wide range of Lewis acidity While the perennial
problem of Lewis acid-base theory namely the ranking of Lewis acids has still to be
solved it appears that induced chemical shifts (NMR) offer a valuable basis for this
rating Here IR UV-Vis and NMR techniques were all employed with variable success
Abbreviations
Aring angstrom
Bmim 1-butyl-3-methylimidazolium
Bn benzyl
Bu butyl
C coulomb
degC degrees Celcius
CIMS chemical ionization mass spectroscopy
CL confidence level
COSY correlation spectroscopy
CPL ε-caprolactone
Cy cyclohexyl
d doublet
dba dibenzylidene acetone
DCM dichloromethane
dd doublet of doublets
DHP dihydropyran
DL detection limit
DMF NNndashdimethylformamide
DTBP 26-di-tert-butylpyridine
DTMP 26-di-tert-butyl-4-methylpyridine
Eq equivalents
EIMS electron ionization mass spectroscopy
EtOH ethanol
Et3N triethylamine
GC gas chromatography
h hour
HRMS high resolution mass spectroscopy
HSAB hard-soft-acid-base
IR infrared
iii
i-Pr iso-propyl
Kh hydrolysis constant
mp melting point
MSA methanesulfonic acid
MeOH methanol
Me methyl
n number of measurements taken
NMR nuclear magnetic resonance
OAc acetate
OTf trifluoromethanesulfonate (triflate)
PCL poly(ε-caprolactone)
PDI polydispersity
Ph phenyl
ppm parts per million
p-TsOH para-toluenesulfonic acid
q quartet
ROP ring opening polymerisation
RSD relative standard deviation
rt room temperature
s singlet
Std dev standard deviation
t triplet
t-Bu tert-butyl
TBDMS tert-butyldimethlsilyl
TES triethylsilyl
THF tetrahydrofuran
THP tetrahydropyran
TIPS triisopropylsilyl
TLC thin layer chromatography
TMS trimethylsilyl
Triflimide bis(trifluoromethane)sulfonimide
iv
v
Triflate trifluoromethanesulfonate (the contraction lsquotriflatersquo is used throughout this
thesis)
UVVis ultra violetvisible light spectrophotometry
VL valerolactone
WERC water exchange rate constant
wrt with respect to
wv weight per volume
List of figures and schemes
Chapter 1 Lewis acids-A Literature Overview
Scheme
Figure Heading Page
Scheme 11 Regioselective deacetylation of compound 1 9
Scheme 12 Anomeric deacetylation using Nd(OTf)3 10
Scheme 131 Reaction between difluoropropargyl bromide (15) and
benzaldehyde (16) 11
Scheme 132 Indium-mediated Barbier-type reaction of 19 with
different aldehydes in aqueous media 12
Scheme 14 Proposed mechanism 13
Scheme 15 Lewis acid mediated alkylation of 3-hydroxy-2-napthoate
with 1-(2-naphthyl)ethanol 14
Scheme 16 Alkylation of 26 using La(OTf)3 15
Scheme 17 Proposed mechanism of alkylation reaction 17
Scheme 18 Metal triflate catalysed ROP of CPL or VL 18
Scheme 19 A plausible mechanism of the ROP of CPL by Sc(OTf)3
via an activated monomer mechanism 20
Scheme 110 Synthesis of β-enaminones catalysed by Yb(OTf)3 21
Scheme 111 Cleavage of benzylidene acetals using Er(OTf)3 21
Scheme 112 Allylation of acetophenone with diallyldibuyltin
catalysed In(OTf)3 23
Scheme 113 Metal triflate catalysed allylation 24
Scheme 114 Proposed mechanism of acyclic transition state of
allylation reaction 26
Scheme 115 Reaction of 7-methyl-1-phenyloct-6-en-3-one oxime 34
catalysed by aluminium triflate 26
Figure 11 Active oxime-derived intermediates 27
vi
Scheme 116 Al(OTf)3 catalysed ring opening of styrene oxide with
alcohols 29
Scheme 117 Opening of glycidyl ether with alcohol and Al(OTf)3 30
Figure 12 Chelation of Al(OTf)3 to glycidyl ether 31
Scheme 118 Reaction between 2-phenylenediamine 40 and benzoin 41
catalysed by Ga(OTf)3 33
Scheme 119 Proposed mechanism for Ga(OTf)3-catalysed reaction of
2-hydrochalcones with o-phenylenediamine 35
Scheme 120 Reaction between iastoic anhydride p-
methylbenzaldehyde and ammonium acetate 36
Scheme 121 One pot synthesis of 23-dihydroquinazolin-4(1H)-ones
catalysed by Ga(OTf)3 36
Scheme 122 Proposed mechanism for the formation 0f 23-
dihydroquinazolin-4(1H)-ones and quinazolin-4(3H)-ones 39
Scheme 123 [3+2] cycloaddition reaction of 2-aryl-N-tosylaziridines
with different nitriles catalysed by Cu(OTf)2 39
Scheme 124 Mechanism for the [3+2] cycloaddition reaction of 2-aryl-
N-tosylaziridine with nitriles 41
Scheme 125 Claisen rearrangement of allyl 1-naphthyl ethers catalysed
by Bi(OTf)3 41
Scheme 126 Claisen rearrangement of 2-substituted allyl 1-naphthyl
ethers catalysed by Bi(OTf)3 41
Scheme 127 Protection of carbonyl compounds catalysed by Y(OTf)3 42
Figure 13 Electrostatic or covalent nature of Lewis acids 50
Figure 14 Complexation of acid to acyl oxygen atom 51
Figure 15 Phenalen-1-one 52
Figure 16 Lewis acid bound to Lewis base 53
Figure 17 C5H9O3P 54
Figure 18 C6H9O3P 54
Figure 19 Crotonaldehyde 55
vii
Figure 110
1H chemical shift differences of protons of
crotonaldehyde versus the H3 chemical shift difference
on complexation with various Lewis acids
56
Figure 111 Plot of δcis against K 59
Figure 112 Least square plot of Childsrsquo Lewis acid scale against AN
values 61
Figure 113 Perfluorophenyl boron Lewis acids 63
Figure 114 Graph comparing relationship between Gutmanns method
and that of Childs 63
Scheme 128 Ln(OTf)3 catalysed aldol reaction of 63 with
formaldehyde 65
Scheme 129 Aldol reaction between 1-trimethylsiloxycyclohexene
(64) and benzaldehyde catalysed by Ln(OTf)3 66
Figure 115 Yield of aldol product 67
Scheme 130
Aldol reaction between benzaldehyde and 1-phenyl-1-
(trimethylsiloxy)propene catalysed by various Lewis
acids
67
Scheme 131 Aldol reaction of glucose-derived silyl ether with aqueous
formaldehyde catalysed by Lewis acids 69
Scheme 132 Tin-mediated allylation of carbonyl compounds catalysed
by InCl3 69
Scheme 133 Transmetallation from allytin with InCl3 70
Scheme 134 Michael addition of β-ketoesters with methyl vinyl ketone
catalysed by Yb(OTf)3 71
Scheme 135 Direct-type catalytic Mannich reaction of benzaldehyde
aniline and cyclohexanone in water 72
Scheme 136 Biginelli reaction of p-anisaldehyde ethyl acetoacetate
and urea 73
viii
Chapter 2 Metal triflates in protection group chemistry
Scheme
Figure Heading Page
Scheme 21 Mechanism of acid catalysed acetal formation 84
Scheme 22 Al(OTf)3 catalysed acetal formation in the presence of
alcohol andor drying reagent 84
Scheme 23 Acetalisation of benzaldehyde with
trimethylorthoformate catalysed by Al(OTf)3 92
Scheme 24 Acetal deprotection in an aqueous solution of
trifluoroacetic acid 93
Scheme 25 Deprotection of acetals catalysed by Al(OTf)3 93
Scheme 26 Hydrolysis of acetal 96
Scheme 27 TBDMS protection of 2-phenylethanol and para-
bromophenol 98
Scheme 28 Mechanism of acid catalysed THP ether formation 101
Scheme 29 THP protection catalysed by Al(OTf)3 102
Scheme 210 Mechanism for the hydrolysis of 227 103
Chapter 3 The drying of organic solvents
Scheme
Figure Heading Page
Scheme 31 Interaction of metal and a water molecule to form a
Broslashnsted acid 110
Scheme 32 Standard reaction of Karl Fischer titration 112
Figure 31 Karl Fischer measuring cell and electrodes 113
Figure 32 Karl Fischer anode and cathode 114
Figure 33 Graph indicating decreasing water content in methanol
over time 122
ix
Figure 34 Graph indicating decreasing water content in ethanol
over time 124
Figure 35 Graph indicating decreasing water content with
increasing silica mesh size 132
Figure 36 Graph of residual water content in THF after drying with
various desiccants under given conditions 133
Chapter 4 The role of water in metal triflate catalysed reactions
Scheme
Figure Heading Page
Scheme 41 Hydrolysis of a metal 138
Scheme 42 Hydrolysis of organic compounds 138
Figure 41 Effect on the pKa of increasing steric requirements of the
alkyl groups in the 26 positions (in 50 ethanol) 139
Scheme 42 Ionisation of 26-di-tert-butyl pyridine 139
Scheme 44 Direct type catalytic Mannich reaction of benzaldehyde
aniline and cyclohexanone in water 140
Scheme 45 Metal triflate catalysed acylation of alcohols 141
Scheme 46 Benzoylation of hex-2-en-1-ol using metal triflates 142
Scheme 47 Possible source of triflic acid in acetylation reaction 144
Figure 42 Resorcinarene isomers 144
Figure 43 Enhanced Broslashnsted acid 145
Scheme 48 Base catalysed aldol reaction 146
Scheme 49 Mukaiyama aldol reaction 146
Scheme 410 Mukaiyama aldol reaction catalysed by a Lewis acid 147
Scheme 411 Mukaiyama aldol reaction showing silyl ether removed
by water 153
Scheme 412 The Mukaiyama aldol reaction 154
Figure 44 Benzaldehyde showing 1H chemical shifts in CD2Cl2 156
x
Figure 45 Activation of benzaldehyde by metal triflate 157
Figure 46 1H NMR of benzaldehyde complexed to ScCl3 157
Figure 47 NMR peaks showing the difference between fast
intermediate and slow exchange of ligands 158
Figure 48 Possible cluster formations versus free ion formation of
metal triflate in organic solvents 160
Scheme 413 Typical SN2 reaction 161
Scheme 414 Activation of the aldehyde by the Lewis acid in dry
organic solvent 161
Scheme 415 Formation of Broslashnsted acidity through polarisation of
water by a Lewis acid 163
Scheme 416 Interaction of DTBMP with protic solvent to form
nucleophile 164
Scheme 417 Carbocation formation using retinyl acetate 166
Figure 49
UVVis scan of retinyl acetate and its corresponding
carbocation showing the typical wavelengths and colours
of the solutions
167
Figure 410 Solvation of carbocation by nitrobenzene 168
Figure 411 1-butyl-3-methylimidazolium triflate ([bmim][OTf]) 170
Scheme 418 Common ion effect on Al(OTf)3 in [bmim][OTf] 171
Figure 412
UVVis scan showing the different intensities of
carbocation formation with triflic acid and Al(OTf)3 in
DCM
173
Figure 413 UVVis scan of Al(OTf)3 and retinyl acetate in dry DCM 173
Figure 414 DSC scan of standard Al(OTf)3 174
Figure 415 DSC scan of dried Al(OTf)3 175
Figure 416 Crystal structure of protonated DTBMP with OTf-
counterion 176
Scheme 419 Friedel-Crafts reaction of p-xylene with phenylacetylene 178
xi
Chapter 5 Ranking of Lewis acids
Scheme
Figure Heading Page
Figure 51 Proton numbering used on crotonaldehyde 188
Figure 52 1H chemical shift differences of crotonaldehyde versus
the various Lewis acids 189
Figure 53 Carbon numbering used on croton aldehyde 190
Figure 54 13C chemical shift differences of crotonaldehyde versus
the various Lewis acids 191
Figure 55 Atom numbering on trans-cinnamaldehyde 192
Figure 56 13C chemical shift differences of trans-cinnamaldehyde
versus the various Lewis acids 193
Figure 57 Diphenylphosphinobenzaldehyde chelated to a metal
triflate 195
Figure 58 UVVis spectrum of adduct formation between Lewis
base and Lewis acid 199
Figure 59 UVVis spectrum of 4-methyl-3-aniline in DCM 200
Figure 510 UVVis spectrum of 4-nitrodiphenylaniline in THF 201
Figure 511 UVVis spectrum of 4-nitrodiphenylaniline coordinated
to Lewis acid 201
xii
List of tables
Chapter 1 Lewis acids-A Literature Overview
Table Heading Page
Table 11 Yields obtained for the anomeric deacetylation of
compound 1 using lanthanide triflates 9
Table 12 Yields obtained for anomeric deacetylation of selected
sugars using metal triflates 10
Table 13
Yield of reaction between difluoropropargyl bromide
and benzaldehyde in different solvent systems with
different additives
12
Table 14 Indium mediated Barbier-type reaction of 19 with
different aldehydes in aqueous media 13
Table 15 Effects of catalysts on the reaction between 3-hydroxy-2-
naphthoate and 1-(2-naphthyl)ethanol 15
Table 16 Alkylation of 26 with acid-sensitive functional group
substituted benzylic alcohols 16
Table 17 Alkylation with various catalysts 16
Table 18 Sc(OTf)3 catalysed ROP of CPL and VL in the presence
of various amounts of n-BnOH 18
Table 19 The ROP of CPL by M(OTf)3 in ionic liquids 19
Table 110 In(OTf)3 catalysed peracetylation of carbohydrates 22
Table 111 Effect of different triflates on the allylation of
acetophenone with diallyldibuyltin 23
Table 112 Allylation of ketones with diallyldibutyltin catalysed
with In(OTf)3 25
Table 113 Cyclisation of non-activated unsaturated oximes
catalysed by Al(OTf)3 28
xiii
Table 114 Yields of products obtained from reactions with selected
epoxides in various alcohols 30
Table 115 Products obtained from reactions with selected epoxides
with various amines 32
Table 116 N-Methylaniline reactions with epoxides 32
Table 117
Yield quinoxaline derivative from the reaction 12
diamines were reacted with 12 diketones catalysed by
Ga(OTf)3
34
Table 118 Yields obtained in the one pot synthesis of 23-
dihydroquinazolin-4(1H)-ones catalysed by Ga(OTf)3 37
Table 119 Yields obtained in the one pot synthesis of quinazolin-
4(3H)-ones catalysed by Ga(OTf)3 38
Table 120 Cu(OTf)2 mediated [3+2] cycloaddition reactions of
aziridines and nitrile 40
Table 121 Amination of benzyl alcohols with various sulfonamides
catalysed by AgOTf 43
Table 122 Equilibrium constants and spectral data for complexes of
Lewis acids with ketones 46
Table 123 Free energies of complex formation 47
Table 124 Equilibrium constants for complex formation 47
Table 125 Comparison of Lewis acid ranking by different methods 49
Table 126 Boron trihalides 51
Table 127 Δ Carbonyl shifts of metal halides with phenalen-1-one 52
Table 128 Relative acidity of Lewis acids towards various reference
bases at 28 degC 54
Table 129 1H and 13C NMR chemical shift differences on
complexation with crotonaldehyde 55
Table 130 ΔνC=O and pK values of metal halides with phenalen-1-
one 57
Table 131 Chemical shifts and equilibrium constants for equation 9 58
xiv
Table 132 Summary of spectroscopic and structural trends for MX3
(9-fluorenone) 62
Table 133 Yields reaction of 63 with formaldehyde catalysed by
Ln(OTf)3 65
Table 134 Yields reaction between 1-trimethylsiloxycyclohexene
(45) and benzaldehyde catalysed by Ln(OTf)3 66
Table 135 Mukaiyama aldol reactions in water catalysed by InCl3 68
Table 136 Indium trichloride promoted tin mediated allylation of
aldehydes 70
Table 137 Direct-type catalytic Mannich reaction of benzaldehyde
aniline and cyclohexanone in water 72
Table 138 Reaction between p-anisaldehyde ethyl acetoacetate and
urea catalysed by a variety of Lewis acids 74
Table 139
Yields reaction between p-anisaldehyde ethyl
acetoacetate and urea catalysed by a variety of Lewis
acids in the presence of a Broslashnsted acid
74
Table 140 Reactions between p-anisaldehyde ethyl acetoacetate
and urea catalysed by a variety of Broslashnsted acids 75
Chapter 2 Metal triflates in protection group chemistry
Table Heading Page
Table 21 Al(OTf)3 catalysed acetal formation of aldehydes and
ketones 85
Table 21 cont Al(OTf)3 catalysed acetal formation of aldehydes and
ketones 86
Table 22 Al(OTf)3 catalysed acetal formation of aldehydes and
ketones 90
xv
Table 23 Al(OTf)3 recycling experiments using benzaldehyde as
the substrate 92
Table 24 Yield deprotection of acetals catalysed by Al(OTf)3 94
Table 25 Yield deprotection of acetals catalysed by a variety of
metal triflates 94
Table 26 Yield deprotection of acetals in water 95
Table 27 Yield deprotection of acetals in water after one hour 96
Table 28 Yield of alcohol for TBDMS deprotection 100
Table 29 THP protected primary alcohols in reactions catalysed by
Al(OTf)3 102
Table 210 Deprotection of THP ethers using different metal triflates 103
Table 211 Yield TBDMS protected substrates 104
Table 212 Yield of acetal formation with a variety of metal
triflates 105
Table 213 Acetal formation in the absence of added alcohol 106
Chapter 3 The drying of organic solvents
Table Heading Page
Table 31 Result of Karl Fisher titration of hydranal 100 ppm water
standard 115
Table 32 Results of Karl Fischer titration of lsquowet solventsrsquo 117
Table 33 Results of Karl Fischer titration of solvents dried by
traditional methods 118
Table 34 Results of Karl Fischer titration of THF dried by
SilicaSodium 118
Table 35 Results of Karl Fischer titrations of methanol dried over
3Aring molecular sieves (5 wv) 120
Table 36 Results of Karl Fischer titrations of methanol dried over
3Aring molecular sieves (10 wv) 121
xvi
Table 37 Results of Karl Fischer titrations of methanol dried over
3Aring molecular sieves (20 wv) 121
Table 38 Results of Karl Fischer titrations of ethanol dried over
3Aring molecular sieves (5 wv) 122
Table 39 Results of Karl Fischer titrations of ethanol dried over
3Aring molecular sieves (10 wv) 123
Table 310 Results of Karl Fischer titrations of ethanol dried over
3Aring molecular sieves (20 wv) 123
Table 311 Results of Karl Fischer titrations of methanol and ethanol
dried over KOH powder 125
Table 312 Results of Karl Fisher titrations of acetonitrile dried over
3Aring molecular sieves and neutral alumina 127
Table 313 Results of Karl Fisher titrations of toluene DCM and
THF dried over 3Aring molecular sieves and silica 129
Table 314 Results of Karl Fischer titrations of THF dried over 3Aring
molecular sieves (20 wv) 129
Table 315
Results of Karl Fischer titrations for THF dried by
passing over columns of various types of silica and
alumina
131
Chapter 4 The role of water in metal triflate catalysed reactions
Table Heading Page
Table 41 Conversions () of metal triflate catalysed benzoylations 142
Table 42 Yield of acetylation reactions 143
Table 43 Yield of aldol product from different metal triflates 148
Table 44 Results of Mukaiyama aldol reaction 149
Table 45 Average pH readings of Lewis acids in THF H2O
mixture 150
Table 46 Comparison of pH and pKh values 151
xvii
Table 47 Yield aldol reactions in dry THF 154
Table 48 Yield aldol reactions carried out in dry DCM 155
Table 49 Change in chemical shift of benzaldehyde on
complexation of metal triflate 156
Table 410 Dielectric constants of several organic solvents 160
Table 411 Rate of SN2 reaction in Scheme 413 in different
solvents 161
Table 412 Yield aldol reactions carried out in the presence of
DTBMP 164
Table 413 pH measurements of Lewis acids in THFH2O with and
without DTBMP 164
Table 414 Variation of λmax and absorptivity of the retinyl
carbocation 167
Table 415 Results of carbocation formation in nitrobenzene 169
Table 416 Results of carbocation formation in DCM 172
Table 417 Crystal data of protonated 26-di-tertbutyl-4-methyl
pyridine 178
Table 418 Yield of Friedel-Crafts alkenylation reactions
catalysed by various M(OTf)x 178
Table 419 Yield of Friedel-Crafts alkenylation reactions in
various amounts of p-xylene 179
Table 420 Yield of Friedel-Crafts alkenylation reactions with
alternative aromatic systems 179
Table 421 Friedel-Crafts alkenylation reaction in dry solvent 180
xviii
xix
Chapter 5 Ranking of Lewis acids
Table Heading Page
Table 51
1H NMR chemical shift differences (Δδ) of
crotonaldehyde on complexation with various Lewis
acids
188
Table 52
13C NMR chemical shift differences (Δδ) of
crotonaldehyde on complexation with various Lewis
acids
190
Table 53
1H and13C chemical shift differences (Δδ) of trans-
cinnamaldehyde in [bmim][OTf] on complexation with
various metal triflates
192
Table 54 31P NMR chemical shift differences (Δδ) of phosphorus
compounds on complexation with various metal triflates 194
Table 55 1H chemical shift differences (Δδ) of crotonaldehyde on
complexation with various metal triflates 197
Table 56 Δλ of probe on addition of a Lewis acid 201
Table 57 Δν (cm-1) of pyridine on complexation with various
Lewis acids in [bmim][OTf] 203
Table 58 Δν (cm-1) of electron-poor pyridine derivatives on
complexation with various Lewis acids in [bmim][OTf] 205
Table 59 Δν (cm-1) of electron rich pyridine derivatives on
complexation with various Lewis acids in [bmim][OTf] 206
Chapter 1
Lewis acids-A Literature Overview
11 The history of the acid-base theory
Introduction
In the middle of the seventeenth century acids and bases were classified by the fact
that when they were added to each other they formed a salt usually resulting in the
liberation of water1 Towards the end of the century Boyle23 placed more significance
on the properties of the acid and stressed characteristics such as acids will dissolve
many substances they will precipitate sulfur they change blue plant dyes red and
they lose these properties when they come into contact with alkalis
The first ldquomodernrdquo theory was put forward at the end of the eighteenth century by
Lavoisier24 He regarded oxygen as the necessary element for acidity His views led
to the assumption that acids were formed by a reaction of a lsquoradicalrsquo with oxygen
Davy (in around 1810-15)23 showed that hydrochloric acid contained no oxygen and
soon recognised that hydrogen was the essential element for acidity Liebig (in
1838)2-4 followed up this work and defined an acid as a compound that contained
hydrogen and in which the hydrogen could be replaced by a metal
The next theory that was generally accepted was that of Arrhenius (in 1880)24 His
definition of acids and bases was based on the assumption that when dissolved in an
aqueous solution acids will dissociate into hydronium ions and anions and bases into
hydroxyl ions and cations Even though this theory had limitations when it came to
non-aqueous systems it was generally accepted for the next thirty to forty years4
Development of theories
The latter end of this period saw the development of three main theories of acids and
bases1349 namely the solvent-system theory the electron pair theory and that of the
proton The solvent system theory was established by Franklin567 in 1905 He
extrapolated from the theory that water ionises into hydronium and hydroxyl ions
reasoning that ammonia must then ionise into ammonium and amide ions Thus
compounds like ammonium chloride in ammonia are acids and those like sodium
amide in ammonia are bases (equations 1 and 2)
NH4Cl + NaNH2 NaCl + 2NH3
HCl + NaOH NaCl + (1)
(2)
acid base salt solvent
H2O
Cady and Elsey9 who are often accredited with the solvent-system theory13 described
an acid as a solute that produces a cation that has characteristics of the solvent and a
base as a solute which will produce the anion that is characteristic of the solvent
Smith10 later extended this definition with his work on selenium oxychloride His
definition of an acid as an electron pair acceptor from the solvent and the base as an
electron pair donor to the solvent clearly shows for the first time3 the influence of
Lewisrsquos12 1923 definition of acids and bases
The electron pair theory first put forward by G N Lewis11 in 1916 was again
published in his more famous monograph in 192312 In this book Lewis also
independently put forward the proton idea and the generalised solvent-system theory
for acids and bases5
The initial theory put forward by Lewis failed to create any interest there is almost no
reference made to it in the literature for the next fifteen years5 During this period
there was however extensive developments made on the proton and the solvent-
system theories5
In the same year that Lewisrsquos monograph appeared Broslashnsted13 and Lowry14
independently put forward their definitions on the proton theory of acids and bases
This definition is still in use today and defines an acid as a proton donor and a base as
a proton acceptor
During the next fifteen years several scientists developed electronic classifications
which were along similar lines as those developed by Lewis In 1927 Sidgwick15 put
forward his electronic theory for coordination chemistry Here he noted that transition
2
metal atoms generally complete a stable electronic configuration by accepting
electron pairs from the ligands He named the process coordinate bond formation and
created the terms ldquodonorrdquo and ldquoacceptorrdquo
In 1925 Lapworth1617 divided reactants into electron poor (cationoid) and electron
rich (anionoid) This work was later further developed by Robinson18 who extended
the category of electron poor reactants to include neutral molecules with incomplete
octets as well as oxidising agents Similarly the electron rich category was extended
to include neutral molecules with lone pairs and reducing agents
Ingold1920 made a classification based on redox chemistry between 1933 and 1934
He suggested that earlier work that had been done in this area by workers such as H
S Fry and J Stieglitz should include those reactions where there is a degree of
electron transfer due to the limited donation or sharing of electrons and not just
include the reactions where there is complete electron transfer
It is in this work where Ingold first proposed the term electrophile for such
generalised oxidising agents or electron acceptors and the term nucleophile for
generalised reducing agents or electron donors5
In 1938 Lewis published another paper on acids and bases21 This paper contains the
details that his first paper lacked5 Here definitions and examples along with
supporting data21 were also provided
Lewisrsquos second attempt at his definition did not meet the same destiny as the first and
was widely brought to the attention of the scientific community through symposia
The resulting symposium papers were published in two volumes2223 the second of
which was published largely due to the interest that was stimulated by the first23
Usanovich24 put forward his definition of acids and bases in 1939 This theory is not
well-known as it was published in a Russian journal that was relatively inaccessible to
non Russian speakers25 His definition reads as follows An acid is any chemical
species which reacts with bases gives up cations or accepts anions or electrons and
conversely a base is a chemical species which reacts with acids gives up anions or
3
electrons or combines with cations The greatest criticism of this definition is that it
is too sweeping as it includes all Lewis acid-base reactions as well as redox chemistry
and that one may by the same token simply use the term ldquoreactantrdquo
A review was published in 1940 by Luder3 outlining the comparisons between the
newer electronic theories and the older theories of acids and bases He also added to
Lewisrsquos work by way of examples and showed how the theory could be used as a
systematising tool for chemical reactions This work was later expanded into a book
entitled ldquoThe Electronic Theory of Acids and Basesrdquo26
It was around this time that workers started to describe Lewisrsquos theory in terms of
quantum mechanics all previous theories had been based on the Bohr atom model
Mulliken was one of the first to attempt a quantum mechanical explanation in his
series of papers beginning in 195127 His original attraction to the subject was to
explain a class of weak Lewis adducts known as charge transfer complexes
The wave function for a one-to-one charge transfer adduct can be expressed by
equation 3
ψ ψ ψAB = a 0 (AB) + b 1 (A-B+) (3)
Where
ψ0 (AB) = the wave function in the absence of charge transfer
ψ1 (A-B
+) = the wave function after the net transfer of one electron from base B to acid
A
By applying the second-order perturbation theory to equation (3) Mulliken was able
to derive the expression (4) for energy EAB of a weak AB complex27
β 01 EoSo1( )_ 2
E1_ Eo( )
(4)
I II
EAB = Eo_
4
Where
E0 = the energy of the state ψ0 (AB)
E1 = the energy of the excited state ψ1 (A-B
+)
β01 = is the resonance integral between ψ0 (AB) and ψ1 (A-B
+)
S01 = the overlap integral
In general this equation is the lsquosumrsquo of an electrostatic energy term I and the charge
transfer or covalent term energy term II (in equation 4)
In 1967 Hundson and Klopman28-32 used the perturbation molecular orbital theory to
derive a version of equation (4) to explain the role that certain ground state properties
of the acids and bases play in determining the course of adduct formation (given by
equation 5)
ψ ψAB = a A + b Bψ (5)
Varying degrees of donation can be shown by the ratio a2b2 The expression for
energy is shown by equation (6)
Σ ΣΔΕ β_ qsqt
Rstε+ 2 (CsmCtn st )2
(Em _ En )occupied orbtals mof speciesB
unoccupied orbtals nof speciesA
(6)
I II
x x=
Where
ΔE = energy gained or lost
q = electron density in atomic orbitals
R = distance between atoms s and t
csm= coefficient of atomic orbital m in molecular orbital r where r refers to the
molecular orbitals on one molecule ans s refers to those of the other
β = resonance intergral
E = energy of the molecular orbital
5
Again the first term represents electrostatic effects and is dependent on the net charge
densities and the radii of the donor and acceptor atoms (s and t) The second term
represents covalency combining the functions of overlap symmetry and energy of the
donor and acceptor orbitals (m and n) It is assumed that these lsquofrontierrsquo orbitals
correspond to the traditional acceptor and donor orbitals of the original Lewis
definition5 Klopman suggested that on the basis of equation (6) acid-base reactions
could be divided into those which are dominated by lsquocharge controlrsquo (term I) and
those dominated by lsquoorbital controlrsquo (term II)5
Conclusions
The use of the Lewis acid definition is widespread in both teaching and research
however it is not without criticism143334 There is a view that the definition is too
extensive and because of this it includes all chemical reactants Unlike the Broslashnsted
definitions the Lewis definitions can not be quantified It has also been expressed that
the properties of Lewis acids arise due to their action on any H2O present in the
system thus releasing a proton33 This particular aspect is important because it forms
the basis of one aspect of the present study
The acid-base theories most commonly used by chemists today are those of Broslashnsted-
Lowry (usually referred to as a Broslashnsted acid or base) and that of Lewis (referred to
as a Lewis acid or base) The remainder of this literature review shall focus mainly on
Lewis acids and bases their roles in catalysis how they have been used in aqueous
media and previous attempts at ranking them Much of the review will look at metal
triflates a relatively new family of Lewis acids that have started to replace the more
traditional Lewis acids
12 Lewis acids in catalysis ndash a focus on metal triflates
Lewis acid catalysis is a well documented practice in chemistry today35 They are
used for industrial36 and for pharmaceutical processes37 Lewis acids are important
because they enable reactions to be done under milder conditions and offer unique
reactivities and selectivities3839
6
Traditional Lewis acids (for example AlCl3 BF3 and TiCl4) have been used to
successfully catalyse well known transformations such as the aldol and the Friedel-
Crafts reactions4041 However these catalysts are often required in stoichiometric or
sub-stoichiometric amounts and are decomposed or deactivated by even small
amounts of water4142 These traditional Lewis acids are also non recoverable from
reaction mixtures
The strict definition of a catalyst is a substance that can cause a change in the rate of a
chemical reaction and is itself not consumed in the reaction A catalysts works by
lowering the activation energy of a reaction Because a catalyst is not used up in the
reaction it is essentially recyclable and can therefore be used in sub-stoichiometric
amounts
Activators and promoters work in similar ways to catalysts ie by lowering the
activation energy of the reaction However they are consumed or altered during the
reaction and are usually used in stoichiometric amounts Despite this activators and
promoters are often referred to as catalyst in the literature For example In
Vorbruumlggenrsquos many papers on glycosylation trimethylsilyl trifluoromethanesulfonate
(TMSOTF) is called the ldquocatalystrdquo but it is used in stoichiometric amounts as it is
changed in the reaction often into trimethylsilyl acetate (TMSOAc)43
In this thesis the term lsquocatalystrsquo is used to typically denote an entity which facilitates a
given reaction typically where the activator does not participate in the reaction and
where it is used in small amounts relative to the substrates (ie a ldquocatalystrsquo in ther
strict sense of the word) but instances will be cited where the activator is present in
larger amounts up to stoichiometric amounts
Many of the more sensitive catalysts may be catalyst precursors For example BCl3
may hydrolyse in the presence of water to form Cl2BOH and HCl and the former may
be the active catalyst or the active catalyst may be a more complex system such as
such as H+ [Cl3B-OH]- In such cases the BCl3 is a catalyst precursor With respect to
metal triflates (triflates is an accepted contraction for trifluoromethanesulfonates that
is used throughout this thesis) in particular in coordinating solvents or water the
system provides solvent in the inner sphere and triflate counter ions in the outer
7
sphere44 It is these coordinated protic solvents or water molecules that appear to be
the active catalyst in at least some cases as will be seen in later parts of this thesis
(see Chapter 4)
In the past three decades there has been a growing interest in metal triflates Unlike
the more traditional Lewis acid catalysts metal triflates can be added in catalytic
amounts they can be recovered and are reusable without loss of activity In most
cases they are stable and can work as Lewis acids in water
Kobayashi42 wrote a review in 2002 on rare earth metal triflates in organic synthesis
Included in this review are many comparative studies of other triflates This
comprehensive work contains over four hundred references and is an indication of the
growing interest in the field of metal triflates in catalysis
The following discussion is a review of the literature on catalysis using metal triflates
during the past five years It also serves to show the diversity of organic
transformations for which metal triflates have been used
Rare earth metal triflates are dealt with as a group mdash this literature is by far the most
voluminous Group III triflates are dealt with separately as a lot of work can be found
on In(OTf)3 and because of our own successes with Al(OTf)3 There are of course
numerous other triflates that have been used in many other transformations these
have been discussed under the heading Other
Rare earth metal triflates
Rare earth metal triflates have been employed in a plethora of organic
transformations42 They are readily prepared by heating the corresponding metal
oxides or chlorides in an aqueous trifluoromethanesulfonic acid (TfOH) solution
These metal triflates are stable and work as Lewis acids in aqueous media42
Recently the lanthanide triflates were found to effectively catalyse the regioselective
deacetylation of anomeric acetates45 (Scheme 11)
8
O
OAc
OBn
OAc
OAcMeOOC O
OAc
OBn
OAc
MeOOC OHLn(OTf)3MeOH 0 oC
1 2
OMeOOC
AcO
BnO OAc
O
O(OTf)3M
Scheme 11 Regioselective deacetylation of compound 1
Compound 1 (methyl-123-tri-O-acetyl-3-O-benzyl-β-L-idopyranuronate) was
selected to screen a variety of lanthanide triflates for their efficiency in catalysing the
deacetylation of anomeric acetates The reaction was carried out using 5 mol of the
triflate in anhydrous methanol (Table 11) The selectivity in the reaction for
methanolysis of the anomeric acetate may be explained by a favoured six-membered
chelate structure involving the ring oxygen
Table 11 Yields () obtained for the anomeric deacetylation of compound 1 using
lanthanide triflates
Lanthanide Reaction
timea (min) Yield ()b
Yb(OTf)3 180 70
Eu(OTf)3 90 85
Sm(OTf)3 90 92
Nd(OTf)3 90 95 a Reactions were carried out in anhydrous MeOH (24 mL) containing starting material
(50 mg 011 mmol) and 5 mol Mx(OTf)3 at 0 degC b Isolated yield after purification
by flash chromatography
When using Nd(OTf)3 it was found that the catalyst could be recycled from the
aqueous layer and used without loss of reactivity at least three times This catalyst
was then used with other substrates (Scheme 12) and found to effectively deactylate
the anomeric position under similar conditions to afford excellent yields
9
O
AcOAcOAcO
COOMe
OAcO
AcOAcOAcO
COOMe
O
AcOAcO
AcO COOMe
OAc Nd(OTf)3
MeOH rt 4 h
Nd(OTf)3
MeOH rt 4 h
O
AcOAcO
AcO COOMe
3 4
5 6
OH
OH
Scheme 12 Anomeric deacetylation using Nd(OTf)3
This method was also tested on sugars that contained an acetyl group in the C-6
primary position Again the selectivity for the anomeric acetate was maintained with
all of the triflates used (Table 12) good yields were obtained in all cases but
Nd(OTf)3 again showed the highest yields
Table 12 Yields obtained for anomeric deacetylation of selected sugars using
metal triflates
Starting
material Product
Yb(OTf)3a
Yield ()
Eu(OTf)3a
Yield ()
Sm(OTf)3a
Yield ()
Nd(OTf)3a
Yield ()
O O
AcOAcAcOAcOAcO H
7
OAcOAcO
OH
8
85 81 85 81
O
OAcAcOAcO
AcO CH2OAc
H
9
O
AcOAcO
AcO CH2OAc
OH 10
68 78 82 82
O
OAcAcOAcOcO
CH2OAc
HA
11
O
AcOAcOAcO
CH2OAc
OH
12
61 62 79 81
O
HAcOAcOAcO
CH2OAc
OAc
13
O
AcOAcOAcO
CH2OAc
OH
14
67 62 67 68
a Reactions carried out in anhydrous MeOH (6 mL) containing starting material (100 mg) and 5 mol Lx(OTf)3
at rt Yields were obtained after purification by flash chromatography
10
Propargyl alcohols are known to be versatile building blocks in organic synthesis46
Previously Wang and Hammond47 reported a process for making ββ-
difluorohomopropargyl alcohols Following on from this work Armitsu and
Hammond48 have now reported the use of lanthanide metal triflates in an indium-
mediated preparation of ββ-difluorohomopropargyl alcohols
The indium metal-mediated reaction between benzaldehyde and difluoropropargyl
bromide (Scheme 13) was examined to determine the effect of the solvent on the
reaction and then the effect of adding a metal triflate as a Lewis acid catalyst was
investigated (Table 13) The solvent system THFH2O was found to be the most
effectivemdashwhen anhydrous THF was used the reaction did not proceed at least in the
case of the uncatalysed reaction (Table 13 entry 5) Eu(OTf)3 afforded highest yields
and was used in the subsequent studies
TIPSF
BrF +
Ph H
OIn (10 eq)
Lewis acid additive (5 mol)
THFH2O 40 oC 20h
TIPSF
F
HOPh
+ TIPS CF2
215 16 17 18
3
Scheme 131 Reaction between difluoropropargyl bromide (15) and benzaldehyde
(16)
11
Table 13 Yield of reaction between difluoropropargyl bromide and benzaldehyde in
different solvent systems with different additives
Entry Solvent Eq of 16 Additive Yield ()a
1718
1 H2OTHF (41) 11 - 206
2 H2OTHF (41) 11 - 3112
3 H2OTHF (41) 11 - 257
4 Satd NH4Cl(aq)THF (41) 11 - 204
5 THF 11 - No reaction
6 DMF 11 - Complex mixture
7 H2OTHF (41) 22 - 298
8 H2OTHF (41) 33 - 364
9 H2OTHF (41) 11 Sc(OTf)3 424
10 H2OTHF (41) 11 Er(OTf)3 647
11 H2OTHF (41) 11 Eu(OTf)3 788
12 H2OTHF (41) 11 Tb(OTf)3 7610
13 H2OTHF (41) 11 Sm(OTf)3 489
14 H2OTHF (41) 11 Y(OTf)3 4713
15 H2OTHF (41) 11 Ce(OTf)3 778 a Yield determined by 19F NMR
In another study (Scheme 132) compounds of the type 19 were used as substrates
The effects of the substituents R were then examined by using the optimal conditions
that had been found in the previous experiments
RF
BrF + R
FF
HOR
21
In (10 eq)Eu(OTf)3 (5 mol)
H2OTHF (41) 40 oC 20h
R H
O
19 20 Scheme 13 Indium-mediated reaction of 19 with different aldehydes in aqueous
media
Triethylsilyl and triisopropylsilyl were found to be the most effective groups (Table
14) and TES was subsequently used in the reactions with the other aldehydes
12
Table 14 Indium mediated reaction of 19 with different aldehydes in aqueous media
Entry R Rrsquo Yield ()a
1 TIPS Ph 68
2 TES Ph 72
3 TMSb Ph 41
4 n-Hexb Ph 55 5 Phb Ph 35 6 TESb 4-Me-C6H4 65 7 TESb 4-MeOC6H4 61 8 TESb 3-MeO-C4H4 60 9 TESb 24-(MeO)2-C6H3 73 10 TESb 4-OH-C6H4 62 11 TESb 4-Cl-C6H4 71 12 TESb 2-F-C6H4 65 13 TESb 4-NO2-C6H4 No reaction 14 TESb Et 52 15 TESb (CH3)2CH 69 16 TESb BzOCH2 65
a Isolated yield b The reaction was sonicated for 12 h c The reaction was sonicated for 6 h
To date no satisfactory explanation for the results in particular the role of the Lewis
acid has been proposed Studies to elucidate the mechanism are still underway4950
However it is tempting to suggest that the Lewis acid increases the rate of the reaction
of a carbanion type intermediate with the aldehyde at the expense of dimer formation
(18 scheme 14)
TIPSF
BrF
In (reducing agent)
TIPSF
F
Ph H
O
M+3 (Lewis acid) TIPSF
F
HOPh
17
(18)
H2O
15
Scheme 14 Proposed mechanism
13
The Friedel-Crafts alkylations using alcohols and triflates has been previously
reported51-53 However these reactions usually required more than 5 mol of the
triflate and the functional groups of the substrates have been limited to a few
examples
The lanthanide triflates as well as scandium triflate were found to effectively
catalyse benzylation using secondary benzyl alcohols in Friedel-Crafts type
reactions54 The combination of a secondary benzyl alcohol and a metal triflate in
nitromethane gave products in high yields The reaction presumably proceeds via the
carbocation intermediate shown in Scheme 15
OHHO CO2Me
+Lewis acid
CH2Cl2 or CH3NO2
HO CO2Me
22 23 24Carbocation intermediate
Scheme 15 Lewis acid mediated alkylation of 3-hydroxy-2-napthoate with 1-(2-
naphthyl)ethanol
The reaction between 3-hydroxy-2-naphthoate and 1-(2-naphthyl)ethanol was
examined to compare the efficacy of a variety of Lewis acids (Table 15) Titanium
tetrachloride failed to catalyse the reaction and a stoichiometric amount of BF3OEt2
had to be added before an acceptable yield of the product was obtained On the other
hand the metal triflates all showed promising results and La(OTf)3 even afforded a
98 yield when 10 equivalents of water were added to the reaction mixture
14
Table 15 Effects of catalysts on the reaction between 3-hydroxy-2-napthoate and 1-
(2-naphthyl)ethanol
Entry Lewis acid Loading
mol Time h Yield ()
1a TiCl4 100 1 6
2a BF3OEt2 100 1 96
3a BF3OEt2 10 1 6
4b La(OTf)3 1 05 99
5b La(OTf)3 01 1 98
6b La(OTf)3 001 24 93
7b La(OTf)3 1 15c 98
8b Yb(OTf)3 1 033 96
9b Yb(OTf)3 01 025 94 a Reaction conditions DCM (250 mmolL) 0 degC b Reaction conditions
nitromethane (250 mmolL) 100 degC c H2O (10 eq) was added
Once the optimum conditions had been established the same reaction was used to
examine the effect of groups R (reagent 25) on the yield of the reaction (Table 16)
as well as the effects of acid sensitive functional groups in position Rrsquo(reagent 26) on
the outcome (Table 17) of the reaction (Scheme 16)
OHHO CO2R
+1 mol La(OTf)3
R
HO CO2R
R25 26 27
CH3NO2
Scheme 16 Alkylation of 26 using La(OTf)3
15
Table 16 Alkylation of 26 with acid-sensitive functional group substituted benzylic
alcohols
Entry R Rrsquo time Yield ()
1 H Me 15 min 99
2 OSiR3a Me 2 h 93
3 OAc Me 6 h 95
4 OBn Me 14 h 97
5 OH Me 25 h 65 a SiR3 = tert-butyldimethylsilyl
Table 17 Alkylation with various catalysts
Entry R Rrsquo triflate time Yield ()
1 H Me Hf 5 min 89
2 H Me H+a 15 min 87
3 H Me Sc 15 min 99
4 H Me Yb 25 min 99
5 H Bn La 25 h 99
6 H t-Bu La 19 h 59
7 H t-Bu Sc 20 min 89
a TfOH 3 mol
The catalytic activity of Sc(OTf)3 (1 mol) was found to be almost the same as that
of TfOH at 3 mol La(OTf)3 and Yb(OTf)3 at 1 mol were found to be less reactive
than 1 mol of TfOH The catalytic activity increased in order La(OTf)3 lt Yb(OTf)3
lt TfOH (1 mol) lt TfOH (3 mol) asymp Sc(OTf)3 ltlt Hf(OTf)3
It is postulated by the authors that the metal triflates may hydrolyse to form triflic acid
which may aid in the catalysis of these reactions They do however point out that it
has been documented by Kobayashi et al55 that the rare earth metal triflates are stable
in water Also La(OTf)3 and Yb(OTf)3 showed poorer catalytic activity than triflic
acid for these reasons the authors deduce that the formation of triflic acid in these
reactions generates such small amounts of TfOH as to be inconsequential
16
A proposed mechanism for the reaction is shown in Scheme 17 The mechanism
proposes the intermediate of a benzylic carbocation which may form the dibenzyl
ether eliminate H+ to generate the corresponding styrene or react with the nucleophile
to yield the product Both the styrene and the dibenzyl ether products were found in
trace amounts lending credence to the presence of the carbocation intermediate
Ar
Nucleophile
Ar
Olefin
Triflate
H2OOHAr
OHAr
Triflate Ar O Ar
Ar Nu
Products
2-naphthyl 29Byproducts
Ph
Ph
Ar = Ph
28
Ar = 2-naphthyl
Scheme 17 Proposed mechanism of alkylation reaction
Research on environmentally friendly aliphatic polyesters has received considerable
attention because of their biocompatibility and biodegradability56 However there are
only a few reports where metal triflates have been used to catalyse the polymerisation
process5758
The ring opening polymerisation (ROP) of lactones by rare earth metal triflates and
Sc(OTf)3 has been reported59 The immobilisation of the triflates in ionic liquids was
also successfully investigated for this purpose
Various metal triflates were examined for their ability to catalyse the ROP of ε-
caprolactone (CPL Scheme 18) Sc(OTf)3 was the most effective and gave a
quantitative conversion to the desired poly(ε-caprolactone) (PCL) in just two hours
The polydispersity (PDI MwMn) was small being 11 Sc(OTf)3 was then used in
further investigating the ring opening of lactones
17
O
O
CPL
2 mol M(OTf)x
toluene 25 oCO
O
n
O
O
VL
or
Scheme 18 Metal triflate catalysed ROP of CPL or VL
The role of protic additives was investigated in the ROP of CPL as well as δ-
valerolactone (VL Table 18) Various amounts of n-BnOH were then added to the
reaction mixture When an excess of n-BnOH was added to the reaction mixture
polymerisation was accelerated rather than terminated and it was also seen that Mn
decreases as the amount of n-BnOH increases Interestingly Sc(OTf)3 catalyses the
ROP of VL which has a lower ring strain than CPL at a higher rate than that of CPL
Table 18 Sc(OTf)3 catalysed ROP of CPL and VL in the presence of various amounts
of n-BnOH
Entry Monomer n-BnOH
Mol Time (h)
Yield
() Mn
b
1 CPL 0 4 gt99 6900
2 CPL 2 23 gt99 3500
3 CPL 5 2 gt99 2400
4 CPL 10 2 gt99 1600
5 VL 0 15 96 4300
6 VL 2 1 95 2900
7 VL 5 1 97 1900
8 VL 10 1 95 1400 a Sc(OTf)3 2 mol (002 mmol) CPL or VL 10 mmol toluene 10 mL temp 25 degC (CPL) or 0 degC (VL) b
Estimated by size exclusion chromatography (CHCl3 polystyrene standards)
18
Table 19 The ROP of CPL by M(OTf)3 in ionic liquidsa
Entry M(OTf)3 X- Time (h) Yield () Mn
1 Sc [BF4]- 2 d 0 -
2 Y [BF4]- 7 d 29 500
3 La [BF4]- 2 d 29 300
4 Ceb [BF4]- 6 d 32 600
5 Nd [BF4]- 6 d 30 500
6 Eu [BF4]- 2 d 0 -
7 Gd [BF4]- 5 d 30 600
8 Yb [BF4]- 4 d 27 500
9 Lu [BF4]- 3 d 26 500
10 Sc [PF6]- 42 80 2500
11 Y [PF6]- 53 83 2800
12 La [PF6]- 46 100 3700
13 Ceb [PF6]- 47 100 3500
14 Nd [PF6]- 35 100 2700
15 Eu [PF6]- 49 100 2400
16 Gd [PF6]- 48 100 3400
17 Yb [PF6]- 48 43 1600
18 Lu [PF6]- 47 100 4400
19 Sc [SbF6]- 30 87 900
20 Y [SbF6]- 48 99 1800
21 La [SbF6]- 42 100 1800
22 Ceb [SbF6]- 29 100 1700
23 Nd [SbF6]- 43 100 1700
24 Eu [SbF6]- 48 4 1500
25 Gd [SbF6]- 29 100 2500
26 Yb [SbF6]- 48 82 1400
27 Lu [SbF6]- 80 80 1100 a Polymerisation conditions M(OTf)3 20 micromol ionic liquid ([Bmim]X) 050 mL CPL 011 mL (099 mmol)
toluene 10 mL temp 25 degC N2 b Ce(OTf)4
19
Ionic liquids were then screened with the idea to use them to immobilize the triflates
(Table 19) All reactions in [Bmim][PF6] (Bmim = n-butyl-3-methylimidazolium)
proceed smoothly compared to those in the other two ionic liquids and afforded high
yields of PCL in average time periods [Bmim][SbF6] showed similar results to
[Bmim][PF6] However with Sc Eu Gd and Lu the phases became inseparable
suggesting to those authors some kind of interaction between the cation of the ionic
liquid and the triflate
The formation of benzyl esters in an early stage of polymerisation was confirmed by
the consumption of BnOH as shown by TLC analysis and the 1H NMR spectrum of
the reaction mixture A possible reaction mechanism was proposed by the authors
(Scheme 19) in which the Sc(OTf)3 coordinates to the CPL to afford the cationic
complex 30 which is attacked by the alcohol to produce the corresponding ester A
rapid equilibrium exists between 31 32 and 33 under which Sc(OTf)3 preferentially
binds to CPL to form 30 The free hydroxy ester 33 can serve as an alcohol
nucleophile and attacks the electrophilic 30
O
O
Sc(OTf)3
O
O(TfO)3Sc
HORO
O(TfO)3Sc
H+
HOOR
O(TfO)3Sc
HOOR
OSc(OTf)3
HOOR
O
Sc(OTf)3
HOR
Fast
Fast
Fast
30
31
32
33
Propagation
OR
Scheme 19 A plausible mechanism of the ROP of CPL by Sc(OTf)3 via an activated
monomer mechanism
20
Ytterbium triflate was reported to catalyse the synthesis of β-enaminones under
solventless conditions (Scheme 110)60 The method was efficient on both cyclic and
acylic ketones aromatic as well as aliphatic amines and in all cases excellent yields
were achieved The reactions were allowed to proceed at room temperature for 12
hours in the presence of only 001 mol of the catalyst
O
O
R NH2+
O
NHR
001 molYb(OTf)3
12 h rt Scheme 110 Synthesis of β-enaminones catalysed by Yb(OTf)3
Benzylidene acetals have been effectively cleaved in the presence of Er(OTf)361
Reactions were carried out in nitromethane using 1-5 mol of the triflate and the
reactions proceeded quickly and with high yields (Scheme 111) The solvent was
then changed to acetic anhydride and the reactions repeated with the view of isolating
the preacetylated product These reactions were also successful and showed good
yields
O
OOH
Er(OTf)3 1 - 5 mol HO
HOOH
CH3NO2 Scheme 111 Cleavage of benzylidene acetals using Er(OTf)3
Group III metal triflates
Indium triflate
Peracetylation of carbohydrates catalysed by indium triflate under solvent-free
conditions has been reported62a It was found that 005 mol of In(OTf)3 successfully
catalysed the peracetylation of various sugars in acetic anhydride after a period of 1
hour at 0 degC (Table 110)
21
Table 110 In(OTf)3 catalysed peracetylation of carbohydrates
Entries Carbohydrate Product Yield ()
1 99
2 O
OH
HO
HOHO
HO
O
OAc
OAc
AcOAcO
AcO
96
3 O
OHHOHO
HO
HO
O
OAcAcOAcO
AcO
AcO
88
4 63
5 O
OHHOHO
HOH3C
O
OAcAcOAcO
AcOH3C
96
6 89
7 O
OO
HO
OH
OHOH
HO
HO
OH HO O
OO
AcO
OAc
OAcOAc
AcO
AcO
OAc AcO 97
O
OAcAcOAcO
AcO
AcO
O
OHHOHO
HO
HO
O
OAcAcOAcO
AcOO
OHHOHO
HO
O
OHAcHNHO
HO
HOO
OAcAcHNAcO
AcO
AcO
a Reaction conditions In(OTf)3 (005 eq) in Ac2O (30 equiv) 1 h 0 degC (reactions1-5) or 0 degC to rt (reactions 6-7)
The compatibility of this reaction with commonly used protection groups was
investigated TBDPS and benzylidene acetals were stable under the reaction
conditions while TMS-and-TBS were deprotected The mechanism of the
peracetylation of galactose was examined The reaction proceeds in the presence of
005 mol of TfOH However the reaction does not proceed when 26-di-tert-
butylpyridine (DTBP) is added to the reaction mixture with In(OTf)3 This suggested
to the authors that triflic acid is the catalytic species
The reaction mixture containing the In(OTf)3 and DTBP was allowed to stir for one
day and 11 yield of the peracetylated product was obtained indicating to the
authors that there could be a dual pathway the In(OTf)3ndashcatalysed pathway being the
slower of the two The possibility of the pyridinium acid catalysing the reaction
(albeit slowly) was ignored by the authors Given the anomalous base behaviour of
DTBP as shown by Brown62b the latter hypothesis remains feasible Brown showed
that DTBP is an unusually weak base compared to other 26-dialkylpyridines
22
The effects of other catalysts on the acetylation of glucose were also determined
In(OTf)3 was compared to InBr3 and InCl3 Both of the alternative Lewis acids
catalysed the reaction at a much lower rate than the triflate InBr3 being the slower of
the two It was concluded that the relative strengths of the acids generated in the
reaction TfOH gt HBr gt HCl accounted for the differential activity observed in the
reactions The generation of these protic acids assumes the hydrolysis of the InX3
salts at least to some extent
The allylation of various ketones with diallyldibutyltin was successfully catalysed by
In(OTf)364 The reaction between acetophenone and diallyldibutyltin (Scheme 129)
was investigated to determine the effects of different solvents on the reaction The
results showed that DCM gave superior results over the other solvents such as MeCN
ether and THF
CH3Ph
O+ SnBu2
2 H3C OH
Ph10 mol M(OTf)3
Solvent rt 12h
Scheme 112 Allylation of acetophenone with diallyldibuyltin catalysed In(OTf)3
This reaction (Scheme 112) was also used to determine the efficacy of other metal
triflates to catalyse this reaction (Table 111) In(OTf)3 and Y(OTf)3 showed the most
promising results and afforded 95 and 86 yields of the allylated product
respectively
Table 111 Effect of different triflates on the allylation of acetophenone with
diallyldibuyltin
Entry Metal triflate Yield
()a
1 In(OTf)3 95 2 Yb(OTf)3 74 3 AgOTf 71 4 Y(OTf)3 89 5 Cu(OTf)2 51 6 Zn(OTf)2 48
a Isolated yields
23
Once the optimal conditions were obtained (10 mol In(OTf)3 DCM rt 12 h) a
variety of ketones was examined (Scheme 113 Table 112) High yields were
obtained in the majority of cases
R2R1
O+ SnBu2
2 R2 OH
R110 mol M(OTf)3
Solv rt 12 h Scheme 113 Metal triflate catalysed allylation
Interestingly meta-substituted aromatic compounds underwent allylation more
efficiently than the para-substituted analogue aliphatic ketones afforded the allylation
products in moderate to good yields Benzophenone also reacted smoothly to give
products in a reasonable yield To confirm that the reaction system could be applied to
aldehydes p-anisaldehyde was used which usually reacts slowly in allylation
reactions Under these conditions an 80 yield of the homoallylic alcohol was
obtained (Table 112 entry 16)
24
Table 112 Allylation of ketones with diallyldibutyltin catalysed with In(OTf)3
Entry R1 R2 Yield
()a
1 Ph CH3 95
2 p-FC6H4 CH3 81
3 p-ClC6H4 CH3 75
4 p-BrC6H4 CH3 86
5 p-NO2C6H4 CH3 66
6 m-BrC6H4 CH3 93
7 m-CF3C6H4 CH3 90
8 p-MeC6H4 CH3 58
9 p-NH2C6H5 CH3 -
10 Ph Ph 42
11 78
12 82
13
90
14 CH2CH2CH3 CH3 45
15 CH2CH(CH3)2 CH3 27
16 p-MeOC6H4 H 80
O
O
O
a Isolated yields
The mechanism of the reaction was studied by 1H NMR and ESI mass spectrometry
and an acyclic transition state has been proposed as follows in Scheme 114
25
In(OTf)3
O
R1 R2
In(OTf)3
R1
R2
O(TfO)2In
SnBu2
OTf
R1
R2
O(TfO)2In
+
SnBu2OTf
R1
R2
OBu2Sn
O
R1 R2
SnBu2
2 Scheme 114 Proposed mechanism of acyclic transition state
Aluminium triflate
Regioselective cycloisomerisation of non-activated unsaturated oximes catalysed by
aluminium triflate has been reported65 The reaction afforded 5- 6- and 7- membered
rings containing oxygen and nitrogen The model reaction using oxime 34 (Scheme
115) was used to determine the efficacy of different metal triflates on the
cycloisomerisation reaction and the effect of different solvents
N
R
HO
NO
NO
R
O
34
35
36
37
Catalyst (20 mol)Refluxing solvent
Scheme 115 Reaction of 7-methyl-1-phenyloct-6-en-3-one oxime 34 catalysed by
aluminium triflate
No cyclisation occurred when Cu Fe and Sn triflates were used in which cases only
ketone 37 was formed However the desired seven membered ring 35 formed in good
yield when Al(OTf)3 was used as the catalyst in boiling nitromethane When Al(OTf)3
and dichloroethane were used a 73 conversion was seen of which 59 was the six
26
membered ring 36 Triflic acid failed to catalyse this reaction and gave only 22 of
the ketone 37 bringing an argument against TfOH being the true catalyst
Al(OTf)3 in nitromethane or dichloroethane was then used with a variety of oximes to
determine the scope of the reaction (Table 113) Oximes bearing trisubstituted double
bonds (Table 113 Entries 1 and 2) afforded seven membered ring structures in high
yields When the oxime contained disubstituted olefins (Entries 3-6) the reactions
resulted in 5 or 6 membered cyclic structures depending on the chain length between
the oxime moiety and the double bond of the starting material When the double bond
in the original material was terminal no cyclic structure was formed the reaction
resulted in amides through a type of Beckman rearrangement
The regioselectivity of the reaction is explained through electrophilic attack by the
oxime proton once the ndashOH as been activated by the highly electrophilic Al3+
species followed by nucleophilic attack of the oxime oxygen atom on the carbocation
so generated via the intermediates shown in Figure 11
R
NO
H
Al(OTf)3
R
NOAl(OTf)3
O-H activated intermediate Alkene-promotedCarbocation intermediate
Figure 11 Active oxime-derived intermediates
27
Table 113 Cyclisation of non-activated unsaturated oximes catalysed by Al(OTf)3
Entry Substrate Solvent Product Isolated
Yield ()
1 N
OH
MeNO2
ON
84
2 N OH
(CH2Cl)2
ON
80
3 N
OH
MeNO2
ON
82
4 N
OH
MeNO2
ON
81
5 N
OH
MeNO2
ON
73
6 N
OH
MeNO2
ON
84
7 N
OH
MeNO2
O
NH
HN
O
82 (11)
Acetylations of alcohols phenols and thiophenols could be carried out with catalytic
amount of Al(OTf)366 A variety of primary secondary benzylic and cyclic alcohols
was successfully acetylated using 01 mol Al(OTf)3 with acetic anhydride under
solventless conditions in excellent yields in under a minute The system was also used
successfully to acetylate thiophenols
28
Al(OTf)3 was found to be a highly effective Lewis acid catalyst for the ring-opening
of a variety of epoxides by a range of alcohols The reactions provided products in
very high yields and often with high selectivity67a
Initially work was carried out on styrene oxide was used with Al(OTf)3 and varying
amounts of ethanol (Scheme 116) This system was subjected to various temperatures
until optimal conditions were established
O
ROH
Al(OTf)338
OROH
39 Scheme 116 Al(OTf)3 catalysed ring opening of styrene oxide with alcohols
The scope of Al(OTf)3 as a Lewis acid catalyst for epoxide opening was then
determined by changing the catalyst loading the types of alcohols used and the
substrates employed (Table 114)
The reactions with styrene oxide proceeded smoothly and in the majority of the cases
were regioselective This is due to the stabilisation of the carbocation intermediate by
resonance the phenyl ring In contrast butylene oxide forms an almost 5050 ratio of
isomers where the nucleophile attacks at either end of the epoxide
When styrene oxide is reacted with ethanol in the absence of a catalyst only 5 yield
of the glycol ether is obtained after 24 hours Triflic acid also failed to catalyse this
reaction when 0005 mol TfOH was added to the reaction mixture only 4 of the
glycol ether was formed after 24 hours
29
Table 114 Yields of products obtained from reactions with selected epoxides in
various alcohols
Entry Product
Yield
()
00005
cat
Yield
()
0001
cat
Yield
()
0002
cat
Yield
()
0003
cat
1 39 R = Me 0 1 8 94
2 39 R = Et 94 95 -a -a
3 39 R = nPr 93 97 92 (4)b -a
4 39 R= iPr 91 92 -a -a
5 39 R = 2-Bu -a 14 97 96 (4)b
6 39 R = tBu -a 77 77 -a
7
-a 41 (34)b -a -a
8
-a 31 (24)b -a -a
9
-a 55 88 -a
10
-a 21 42 62
11
-a -a 89 -a
OH
OCH3
OH
OCH2CH3
OHOCH2CH2CH3
OHOCH(CH3)2
OHOCH2CH2CH2CH3
a Reactions not performed b Yields in parenthesis refer to the other regioisomer
To test the Al(OTf)3 on epoxides containing different structural motifs glycidyl ethers
were used (Scheme 117) In contrast to the above results the main product formed
from these reactions was the secondary alcohol where the nucleophile had favoured
the attack on the less hindered side of the epoxide electrophile
O
OR
EtOHAl(OTf)3 R
OOEt
OHR
OOH
OEt+
1o alcohol 2o alcohol Scheme 117 Opening of glycidyl ether with alcohol and Al(OTf)3
30
This is presumably due to Al(OTf)3 forming a chelate structure with the oxygen atom
of the epoxide and the oxygen of the glycidyl ether (Figure 12) This would decrease
the Lewis acidity of the metal which would usually activate the internal carbon atom
such that it plays a dominant role This is supported by the notion that acid-catalysed
epoxide ring opening reactions are borderline SN2 reactions67b implying significant
SN1 type character in the transition state The proposed chelate structure should have
the effect of shifting the mechanism to favour the SN2 like transition state more fully
accounting for the shift to regioselective from the more hindered to the less hindered
position
O AlO
O Et
H
(OTf)3 Figure 12 Chelation of Al(OTf)3 to glycidyl ether
The aminolysis of epoxides was reported to be catalysed efficiently by Al(OTf)3 to
afford a variety of amino alcohols in high yields68 Using aniline as the nucleophile
with a variety of epoxides reactions were carried out using 10 mol Al(OTf)3 in
toluene DCM and ether to determine the effect of solvents on the reaction From the
results it was clear that reactions carried out in toluene were far superior to those
carried out in the other solvents
It was also clear from these initial experiments that higher concentrations of Al(OTf)3
were required for aminolysis than for alcoholysis of epoxides67a Presumably the
nitrogen atom of the amine group and the oxygen atom of the epoxide compete for
complexation to the Al(OTf)3 Since oxygen and nitrogen are both hard but different
Lewis bases there would be a differential competition for the metal centre A variety
of alkyl and arylamine nucleophiles were then used in reactions with selected
epoxides to determine the steric and electronic effects they would have on the reaction
(Table 115)
31
Table 115 Products obtained from reactions with selected epoxides with various
aminesa
Product Yield ()
1 mol
Al(OTf)3
Yield ()
2 mol
Al(OTf)3
Yield ()
10 mol
Al(OTf)3
OOH
N
48 75b -
OOH
N
45 80b -
OH HN
14 - 43c
O
OH
NH
31 (31)cd
- 35 (34)cd
a 12 eq amine 100 degC 5 h b Isolated yields () c 24 h reaction time d Yields in parentheses refer to other
regioisomer
In general the alkyl-amines (Table 115) were found to be harder nucleophiles and
therefore compete more efficiently for the Al(OTf)3 than the aromatic amines (Table
116) Higher catalyst loadings andor longer reaction times were required for reaction
with diethylamine and isopropylamine than the aromatic amines
Table 116 N-Methylaniline reactions with epoxidesa
Product Yield ()
1 mol
Al(OTf)3
Yield ()
2 mol
Al(OTf)3 OH
N
87 89
OOH
N
50 93
a 12 eq amine 100 degC 5 h
32
Reactions with less active nucleophiles were also successful and several reactions
showed that the catalyst was suitable for recovery and reuse without loss of activity
After the reaction the catalyst was extracted in to the water layer which was
subsequently removed under vacuum at elevated temperature
Gallium triflate
Gallium triflate has been used to catalyse the synthesis of quinoxaline derivatives69
An initial solvent study was carried out using the condensation reaction between 12-
phenylenediamine 40 and benzoin 41 (Scheme 118) Reactions were complete in
polar solvents such as ethanol methanol and acetonitrile in as short a period as five
minutes with conversions to product 42 being quantitative The reaction carried out in
water although slower (30 minutes) was also successful and gave an 85 yield
Ethanol was chosen as the solvent for further work due to its wide availability and
relatively low toxicity Both 5 mol and 1 mol of the catalyst gave quantitative
conversion to product 42 in 5 minutes while 01 mol gave only 85 conversion
NH2
NH2
+
O O5 mol Ga(OTf)3
solvent
N
N
40 41 42 Scheme 118 Reaction between 2-phenylenediamine 40 and benzoin 41 catalysed by
Ga(OTf)3
Using these optimised conditions a variety of 12-diamines was reacted with 12-
diketones (Table 117) The quinoxaline derivatives were formed in high yields
33
Table 117 Yield () quinoxaline derivative from the reaction 12-diamines were
reacted with 12-diketones catalysed by Ga(OTf)3
Entry Product Time
(min)
Yield
()
1 N
N
5 gt99
2 N
N
O
20 95
3 N
N
Cl
10 94
4 N
N
O2N
360 90
5 N
N
Cl
Cl
10 gt99
6 N
N
O
O 10 gt99
7 N
N
O
OO2N
240 90
8 N
N
O
OO 10 98
9 N
N
Cl
Cl
80 90
10 N
N
O
O
20 92
34
The Ga(OTf)3 catalyst could be recovered from the solvent and was reused without
loss of activity in ten repetitive reaction cycles
A mechanism for the reaction of o-phenylenediamine with 2-hydroxychalcone
derivatives to form 24-disubstituted 15-benzodiazepine was proposed (Scheme
119) Ga(OTf)3 forms a chelate structure with the 2-hydroxychalcone which
facilitates the dehydration process The 2-hydroxy group in the chalcone makes the
αβ-unsaturated carbonyl more reactive towards the addition of NH
OOH
R1 R2
43
OHO
R1 R2
44
(OTf)3Ga
Ga(OTf)3
H2N NH2
OHO
R1 R2
45
(OTf)3Ga
NH2
NH2
HO
R1 R2
46
(OTf)3Ga
N NH2
-H2O
HO
R1 R2
N NH
47
Scheme 119 Proposed mechanism for Ga(OTf)3-catalysed reaction of 2-
hydrochalcones with o-phenylenediamine
The one pot synthesis of 23-dihydroquinazolin-4(1H)-ones and quinazolin-4(3H)-
ones was found to be effectively catalysed by Ga(OTf)3 in a highly selective
reaction70 A model reaction was carried out using iastoic anhydride p-
35
methylbenzaldehyde and ammonium acetate under various different reaction
conditions (Scheme 120)
NH
O
O
O+
CHO
CH3
+ NH4OAcNH
NHCatalyst
O
CH3 Scheme 120 Reaction between iastoic anhydride p-methylbenzaldehyde and
ammonium acetate
A range of different triflates was examined using ethanol as the solvent Ga(OTf)3
was by far the superior Lewis acid catalyst 1 mol was sufficient to catalyse the
reaction When other solvents were used such as THF water and acetonitrile the
yields dropped dramatically
The scope of the study was then expanded by looking at the effect of electron
withdrawing or donating groups on the aromatic rings (Scheme 121)
NH
O
O
O+ +
NH4OAc or R3NH2 N
H
NCatalyst
O
R2
R1
R2CHOR1 R3
Scheme 121 One pot synthesis of 23-dihydroquinazolin-4(1H)-ones catalysed by
Ga(OTf)3
The reaction afforded the desired products in high yields (Table 118) The effect of
substitutuent groups on the aromatic ring had no obvious effects on the yield Good
yields were also obtained with reactions that were carried out with aromatic
heterocyclic aldehydes and anthranilamide
36
Table 118 Yields obtained in the one pot synthesis of 23-dihydroquinazolin-4(1H)-
ones catalysed by Ga(OTf)3
Entry R1 R2
NH4OAc
or
R3NH2
Time
(min)
Yield
()a
1 H p-(CH3)C6H4 NH4OAc 50 86
2 H 24-(OCH3)2C6H3 NH4OAc 45 90
3 H p-(N(CH3)2)C6H4 NH4OAc 35 88
4 H p-(OH)C6H4 NH4OAc 50 86
5 H m-(F)C6H4 NH4OAc 50 85
6 H p-(Cl)C6H4 NH4OAc 35 89
7 H o-(NO2)C6H4 NH4OAc 70 71
8 H p-(NO2)C6H4 NH4OAc 60 73
9 H 2-Furyl NH4OAc 40 91
10 H 2-Pyridyl NH4OAc 50 88
11 Cl p-(CH3)C6H5 NH4OAc 50 86
12 Cl p-(OCH3)C6H4 NH4OAc 40 83
13 Cl p-(F)C6H4 NH4OAc 55 83
14 Cl p-(NO2)C6H4 NH4OAc 50 80
15 H p-(OH)C6H4 sBuNH2 55 83
16 H p-(Cl)C6H4 nBuNH2 50 89
17 H p-(NO2)C6H4 nPrNH2 50 87
18 H p-(Cl)C6H4 C6H5NH2 60 82 a Isolated yields
During the initial solvent study it was discovered that if the reactions were carried out
in DMSO the condensation reaction is followed by oxidation and the products
formed are the quinazolin-4(3H)-ones Reactions with the various aldehydes
containing electron donating or electron withdrawing groups on the aromatic ring
were carried out (Table 119) The products were obtained in high yields in all cases
37
Table 119 Yields obtained in the one pot synthesis of quinazolin-4(3H)-ones
catalysed by Ga(OTf)3
Entry R1 R2 Time
(min)
Yield
()a
1 H p-(CH3)C6H4 50 84
2 H C6H4 55 83
3 H 24-(OCH3)2C6H3 50 89
4 H 34-(CH2O2)C6H3 55 84
5 H p-(N(CH3)2)C6H4 55 87
6 H p-(OH)C6H4 50 92
7 H m-(F)C6H4 55 84
8 H m-(Cl)C6H4 55 86
9 H p-(Cl)C6H4 55 89
10 H m-(NO2)C6H4 70 82
11 H p-(NO2)C6H4 65 81
12 H 2-Furyl 55 80
13 Cl p-(Cl)C6H4 60 79 a Isolated yields
A tentative mechanism was proposed for the formation of 23-dihydroquinazolin-
4(1H)-ones and quinazolin-4(3H)-ones (Scheme 122) The first step probably
involves the condensation of isatoic anhydride 48 with ammonia and then
anthranilamide 49 could be produced with the liberation of carbon dioxide The
intermediate 50 is obtained by the condensation of 49 with aldehydes promoted by
Ga(OTf)3 The amide in intermediate 50 is tautomerised in the presence of Ga(OTf)3
while the imine part of this intermediate is activated by Ga(OTf)3 Thus intermediate
51 is converted to 52 by intramolecular nucleophilic attack of the nitrogen atom on
the imine carbon Subsequently 23-dihydroquinazolin-4(1H)-ones 53 could be
formed by a 15-proton transfer Finally product 54 is obtained by oxidation using
DMSO as a solvent
38
NH
O
O
O
NH4OAc
CO2 NH2
O
RCHOGa(OTf)3
NH2
N
O
R
Ga(OTf)3
NH2
N
O
R
Ga(OTf)3
Ga(OTf)3
2 Ga(OTf)3N
NH
O
RNH
NH
O
R
DMSO
N
NH
O
R
48 49 50
51 52 53
54
(Ga(OTf)3)
H+
Scheme 122 Proposed mechanism for the formation 0f 23-dihydroquinazolin-4(1H)-
ones and quinazolin-4(3H)-ones
Other triflates
Copper(II) triflate or zinc(II) triflate promotes the cycloaddition reactions of α-alkyl
or aryl substituted N-tosylaziridines with nitriles71 A [3+2] cycloaddition reaction
was carried out on 2-phenyl-N-tosylaziridine in acetonitrile at 60 degC for 30 minutes
with a number of Lewis acids (Scheme 123)
NTs
Ar
RCN Cu(OTf)2
65 oC 30 min N
NAr R
Ts Scheme 123 [3+2] cycloaddition reaction of 2-aryl-N-tosylaziridines with different
nitriles promoted by Cu(OTf)2
Although both triflates promoted the reaction Cu(OTf)2 was the superior promoter
Further studies showed that optimal reaction conditions required 1 equivalent of the
metal triflate and reactions to be performed at 60 degC
Using this approach various nitriles were reacted with 2-aryl-N-tosylaziridines (Table
120) Good yields were obtained in all cases The study was extended to cycloalkyl
benzyl and n-octyl substituted aziridines Previously reported methods for these
substrates have been unsuccessful72 when the reactions were carried out using
39
copper(II) triflate good yields of the products were obtained (Table 120 Entries 8-
10)
Table 120 Cu(OTf)2 promoted [3+2] cycloaddition reactions of aziridines and nitrilea
Entry Aziridine Nitrile Product Yield ()b
1
CH3CN
82 (91)c
2
PhCN
67
3 N
N
NPh CH3
TsN
TsPh
N
NPh Ph
Ts
NTsPh
Ts(C6H4)Me-4 CH3CN N N
(C6H4)Me-4
CH3
Ts
77
4 NTs
(C6H4)Me-4 PhCN N N
(C6H4)Me-4
Ph
Ts
62
5
CH3CN N N
(C6H4)Cl-4
CH3
Ts
72
6
PhCN N N
(C6H4)Cl-4
Ph
Ts
61
7 N Ts CH3CN N
NCH3
Ts 62 (93)c
8 N Ts PhCN N
NPh
Ts 60
NTsC6Cl4H4
NTsC6Cl4H4
a In all cases the nitrile served as a solvent b Isolated yields after column chromatography c Yield was determined
by 1H NMR analysis of crude reaction mixture
The mechanism for the [3+2] cycloaddition reaction between 2-aryl-N-tosylaziridine
with nitriles is illustrated in Scheme 124 Copper is coordinated to the nitrogen atom
of 1a generating 56 the subsequent cyclcoaddition reaction results in the formation
of the imidazoline 58
N
Ph
TsCu(OTf)2
N C R
N
Ph
Ts Cu(OTf)2Ph
NCu(OTf)2
Ts
NCR
N
NPh
R
Ts
55 56 57 58
40
Scheme 124 Mechanism for the [3+2] cycloaddition reaction of 2-aryl-N-
tosylaziridine with nitriles
Bismuth triflate has been reported to catalyse the Claisen rearrangement of allyl
naphthyl ethers to afford the corresponding ortho-allyl naphthol derivatives in good to
moderate yields73 The naphthol derivatives were allylated according to Scheme 125
and the rearrangement was then carried out using 20 mol Bi(OTf)3 in acetonitrile
Allyl 1-naphthyl ethers with various substituents on the aromatic ring reacted
smoothly to give the corresponding ortho-allyl naphthols in good yields Similarly
Reactions carried out with 14-di(allyloxy)naphthalene gave clean doubly rearranged
products The other di(allyloxy) naphthalenes that were tested gave mixtures of di and
mono rearranged products (Scheme 125)
OH
R3
R4
Br
Cs2CO3 DMF60 - 100
O
R3
R4
Bi(OTf)3 xH2O 20 mol
OH R3R4
R1 R1 R1R2R2R2
O
O OH
OH
Bi(OTf)3 xH2O 20 mol
MeCN reflux
Scheme 125 Claisen rearrangement of ally 1-naphthyl ethers catalysed by Bi(OTf)3
The study was extended to include 2-substited 1-naphthyl ethers and diallyloxy
naphthalenes The 2-substituted ethers afforded the para-allyl naphthols in excellent
yields in these cases the ortho-Claisen rearrangement is followed by a [3+3]
rearrangement (Scheme 126)
OH
Cs2CO3 DMF74 - 95
Br
OR R
Bi(OTf)3 xH2O 20 mol
MeCN reflux
OHR
Scheme 126 Claisen rearrangement of 2-substituted allyl 1-naphthyl ethers catalysed
by Bi(OTf)3
41
Carbonyl compounds have been chemoselectively protected using Y(OTf)3 as a
catalyst74 By using 2-mercaptoethanol 12-ethanedithiol and 13-propanedithiol
aldehydes and ketones were converted into their corresponding oxathiolanes
dithiolanes and dithiane derivatives in the presence of 5 mol Y(OTf)3 in acetonitrile
and no drying reagent is necessary (Scheme 127)
R1
O
R2Y(OTf)3 5 mol
CH3CN rt
SH SH
HS SH
HO SH
S S
R2R1
S S
R1 R2
O S
R1 R2 Scheme 127 Protection of carbonyl compounds catalysed by Y(OTf)3
Various aliphatic aromatic and heterocyclic aldehydes underwent protection at room
temperature to give the corresponding products in high yields Ketones required reflux
conditions before satisfactory results were obtained The Y(OTf)3 catalyst could be
recovered from the aqueous layer and reused
AgOTf catalysed the direct amination of benzyl alcohols with sulfonamides75 This
process required no preactivation of the primary alcohols into better leaving groups
The reaction is also relatively environmentally friendly as water is the by-product
An initial solvent study on the reaction between p-chlorobenzyl alcohol and p-
toluenesulfonamide in the presence of 5 mol AgOTf showed that the reactions
proceeded optimally in nitromethane From this study it was also seen that by
changing the ratio of alcohol or amine yields and selectivities also changed If 15 or
2 equivalents of the amine were used mono-substituted products were produced
However if 2 equivalents of alcohol were added the product was disubstituted
A range of Lewis acids was also examined using the above reaction Cu(OTf)2
Sc(OTf)3 and Bi(OTf)3 all catalysed the reaction and gave the corresponding products
in high yields and selectivities However the workers chose AgOTf to continue the
42
investigations Even though it was not the top-performing catalyst it showed the best
selectivity overall
Under the optimal conditions found (2 equivalent amine 5 mol AgOTf
nitromethane) a variety of primary alcohols were reacted with different sulfonamides
(Table 121) When toluene-4-sulfonamide was reacted with structurally and
electronically diverse alcohols (Entries 1-4) the desired products were formed in
good yields
Table 121 Amination of benzyl alcohols with various sulfonamides catalysed by
AgOTfa
Entry Productb Yield
()
1 90
2 80
3 85
4 72
5 0
6 45
7 85
8 90
9
82
10 64
Cl
NH
Ts
NH
Ts
Cl
NH
Ts
F
NH
Ts
F3C
NH
Ts
O2N
NH
Ts
NH
SO2
Cl
Ph
NH
Ms
Cl
NH
SO2
Cl
N Ts
ClPh
a Reaction conditions 1 mmol alcohol 2 mmol sulfonamide 5 mol AgOTf in 5 mL nitromethane 100 degC 8 h b
Isolated yields
43
The presence of a strong electron withdrawing (NO2) group was detrimental to the
reaction (Entry 5) and no product was formed Reactions of p-chlorobenzyl alcohol
(Entries 7-9) were also successful
Conclusions
From the above discussion it is clear that metal triflates can be used efficiently as
Lewis acid catalysts in a plethora of organic transformations Metal triflates are
reported to be water tolerant and can even be recovered from the aqueous layer to be
recycled and reused with out loss of activity More traditional Lewis acids are
deactivated or decomposed by even the smallest amounts of water This clearly
demonstrates the advantages of metal triflates as Lewis acid catalysts It is also clear
that some differences of opinion exist as to the hydrolysis of metal triflates in water to
form triflic acid This issue is important as it forms part of the present study
13 The ranking of Lewis acids
Introduction
Lewis acids play an important role in all areas of chemistry there are literally
thousands of references to Lewis acids in the literature (Sci finder search as of
January 2009 produces ca 50000 hits) New kinds of Lewis acids are being
synthesised on a continual basis76 and the yields of existing reactions are being
improved by their use76 The ldquotrickrdquo is to try to determine the correct Lewis acid for
the reaction at hand7778 When a Broslashnsted acid is needed pKa measurements inform
us of the acid strength and in this way we are able to determine at least to some
extent if the acid is suitable for our reaction When using Lewis acids matters are
more complex
The quantitative measurement of Lewis acidity appears to be one of the perennial
problems of the Lewis acid-base theory Lewis himself pointed out that a given series
of compounds would not exhibit a uniform relative acid (or base) strength79 Relative
acidity (or basicity) would be dependent on the reference acid (or base) used80
44
Despite this attempts by many researchers from as far back as the 1950rsquos have been
made to quantify the strength of Lewis acids However there is still no uniform
quantitative measurement for Lewis acidity across a broad range of Lewis acids
Perhaps the closest method for actual values for Lewis acidity would be the
calculation of the enthalpy change accompanying the formation of the acid-base
adduct in the gas phase81 so as to eliminate any solvent effects This is however not
possible for a wide range of compounds and alternative more manageable methods
have to be found
Measurements that have been carried out to try to rank Lewis acids in order can be
broadly broken down into two categories namely a) thermodynamic measurements
and b) measurements carried out spectroscopically This review deals with literature
that covers both of these categories separately as well as the literature that deals with
comparative studies
131 Thermodynamic measurements
Equilibrium constants
Interest in this work started with the investigation of the interaction of iodine with
aromatic hydrocarbons such as benzene82 In this investigation the equilibrium
constants were calculated in different solvents by using ultraviolet absorption making
use of the slight shifts in the UV spectra between the iodine-complexed benzene and
the free benzene The equilibrium constant is calculated by K = (Br2middotA)(Br2)(A)
where A represents the aromatic compound The concentrations are determined from
the UV readings While UV spectroscopy represents a spectroscopic method the
results being in the form of equilibrium constants represent thermodynamic data
Keefer and Andrews83 then used this method to determine the equilibrium constants
for bromine with benzene and its derivatives It was not the intent of that paper to
rank the Lewis acidity of the halides However the results along with the data found
in a paper by Blake et al84 which deals with the Kc of complexes of I2 and Br2 with
45
naphthalene have been cited for this purpose in a later paper by Scott85 In his work
the heats of formation are emphasised
Complexes of iodine monochloride with benzene and certain derivatives were
investigated by examining equilibrium constants86 It was found that ICl functions as
a much stronger acid in these types of reactions than Br2 or I2 The order was found to
be ICl gt I2 gt Br2
Work started by Moodie87 in which he complexed BF3 to three different types of
acetophenones and calculated equilibrium constants was continued by Mohammad et
al88 A range of Lewis acids was complexed to different ketones and by calculating
their Kc values their relative acidities were determined (Table 122)
Table 122 Equilibrium constants and spectral data for complexes of Lewis acids with
ketones
Acid 4-Methoxyacetophenone 44rsquo-Dimethoxybenzophenone
Δλ pK Δλ pK
GaBr3 66 -057 91 -061
GaCl3 63 -057 89 -051
ZnBr2 37 -052 48 -051
ZnCl2 39 -052 50 -039
BF3 59 +063 90 +251
Table 122 shows that in ether the order of acidity based on the Kc calculated by the
formula K = [Adduct][MXn][Ketone] for each acid-ketone combination used the
observed equilibrium was formulated by equation 7 was found to be GaBr3 ~ GaCl3 ge
ZnBr2 ~ ZnCl2 gt BF3 Moodiersquos87 data for BF3 with 4-methoxyacetophenone were not
comparable to the data found in the later study None of the other adducts is similar to
those studied by Moodie and there is consequently no bases for their comparison
Interestingly the benzophenone base gave more consistent results than the
acetophenone adducts possibly because itrsquos less basic than acetophenone
O1R2RC + MXn(Et2O)m O1R2RC MXn(Et2O)m-1 + Et2O 7
46
When an aniline is used as the reference base we see that the order of Lewis acidity
changes89 The equilibrium constants for Lewis acids complexed to unhindered
aniline bases show an order of acidity BF3 ge GaBr3 ~ GaCl3 ~ SnCl4 gt SnBr4 gt ZnBr2
~ ZnI2 gt SbCl3 In this series it can be seen that BF3 is ranked as the strongest acid
Heat of complex formation
On the basis of free energy of complex formation (Table 123) the decreasing acid
strength was found to be ICl gtgt BrCl gt IBr gt gtI2 gt Br2 gtgt Cl2 which is in agreement
with the previous results85 The acid strength of the various halogens was determined
by calculating the heats of formation of trihalogens where the halide ion acts as the
base and the halogen acts as the acid85 The order of ranking obtained on the bases of
free energy of complex formation were found to be comparable to those previously
reported where equilibrium constants have been calculated (Table 124)838490
Table 123 Free energies of complex formation
Base rarr
Aciddarr I-(aq)a Br-(aq)a Cl-(aq)a H2O(l)a
ICl(g) -143 -86 -57 -27 BrCl(g) - -578 - (-20) IBr(g) -1066 -530 -404 -181 I2(g) -459 -216 -114 -070
Br2(g) - -144 +004 +0226 Cl2(g) - - +44 +165
aΔFdeg in Kcalmole at 25 degC
Table 124 Equilibrium constants for complex formation838490
Base rarr
Aciddarr Benzene p-Xylene Hexamethyl
-benzene
Naphalen
e
ICl 054 151 227 139 I2 015 031 135 025
Br2 011 023 - -
By using NN-dimethylacetamide heats of complex formations were used to determine
the Lewis acidities of ICl SO2 Br and I290 The order of acidity compares well with
47
the previous data90 The data were also compared to the polarity and the dipole
moment of the different acids9293
The heat released on formation of an acid-base complex should be increased by an
acid with a large polarisability and a high dipole moment91 The polarisabilities92 of
these acids decrease in the order I2 gt ICl gt Br2 gt gt SO2 The dipole moments93 for I2
and Br2 are zero for ICl 12 Debye and for SO2 16 Debye
From the heats of complex formation the relative acidities were determined to be ICl
gt I2 gt SO2 gt Br2 The data show the importance of both polarisability and the dipole
moment when determining relative acidities ICl has a high dipole moment and a large
polarisability and because of this is the strongest acid Iodine has no dipole moment
but is still a stronger acid than SO2 as it is more polarisable Bromine also has no
dipole moment is more polarisable than SO2 but much less than I2 so it is the
weakest acid The relative acidities of SO2 and Br2 are interchangeable towards a
polar and non-polar donor respectively
Other thermodynamic data
Lewis acids have also been ranked according to a quantity known as the selectivity
parameter9495 The selectivity parameter is the relative ability of an acid to catalyse
the isomerisation of methylpentanes to 22-dimethylbutane and n-hexane and the
simultaneous deprotonation of carbonium ion intermediates A ldquoperfectrdquo acid allows
the isomerisation to occur without hydrogen exchange while a ldquopoorrdquo acid is unable
to stabilise the ion and hydrogen exchange occurs The selectivity parameter is the
ratio of isomerisation to exchange rate constants kisokex equiv (IE) This parameter
allows the ranking of strong acids
In the first paper that appeared on the matter94 the selectivity parameter was used to
rank the acidities of SbF5 TaF5 and NbF5 in HBr HCl HF HSO3F and CF3SO3H In
general the order found was SbF5 gt TaF5 gt NbF5 But the acidity of these systems was
found to be concentration dependent
The second paper by that author95 ranks 2M mixtures of Lewis acids in HBr while a
number were also reported in HF HBr was chosen as the common solvent for the
48
range of Lewis acids as metal fluorides would be stable and not undergo fluoride ndash
bromide exchange The relative strength of the Lewis acids in HBr was found to be
AlBr3 gt GaBr3 gt TaF5 gt BBr3 gt (TiF4 BF3 HF4)
The relative acidities determined in this study were compared to those found by other
methods (Table 125) The data are in general agreement for a wide variety of
techniques
Table 125 Comparison of Lewis acid ranking by different methods
Order Method Ref
BF3 gt TaF5 gt NbF5 gt TiF4 gt PF5 gt SbF3 gt WF6
gtgt SiF4 ~ CrF3
Solvent extraction of
ArH+MFn+1-
96
AsF5 ~ BF3 gt PF3 ~ WF6 gt NbF5 gt SiF4 ~ CrF3 Solubility of Lewis acids 97
SbF5 gt AsF5 gt BF3 gt PF5 Decomposition of
complex
98
AsF5 gt PF5 gt BF3 Displacement reaction 99
BF3 = SbF5 = AsF5 = PF5 gt GeF4 gt TeF6 gt InF5 gt
SeF5
Salt formation 100
AsF3 gt BF3 gt SiF4 gt AsF5 gt PF3 F- transfer from SF6- 101
AsF5 gt PF5 gt BF3 gt SiF4 gt AsF3 gt SF4 SF5 Ion cyclotron
spectroscopy
102
SF5 gt TaF5 gt NbF5 BF3 gt TiF4 gt HfF4 Selectivity 103
SbF5 gt TaF5 ~ NbF5 Conductivity 104
SbF5 ~ PF5 gt BF3 Solvolysis constants salt
formation
105
SbF5 gt AsF5 = BF3 gt PF5 = SnF4 = ReF6 = WF6 =
MoF6 = VF6 gt IF5 = TeF6 = GeF4 = TaF5 = NbF5
gt SeF4 = SiF4 = TiF4 gt SbF3 = AlF3 = CrF3 =
BeF2
Solubility salt formation 106
The methods referred to in Table 125 are not discussed in further detail in this review
as they have found only limited application in the literature
49
A scale for Lewis acidity was calculated by using a dual parameter equation107 This
equation is based on a combination of electrostatic and covalent factors that contribute
to adduct formation108 Equation (8) was used to determine Lewis acidity in terms of
the parameter
Ž = Z rk2 ndash 77Xz + 80 (8)
where
Ž = charge number of atomic core
rk = ionic radius
Zrk2 = related to electrostatic forces
Xz = electronegativity of elements in the valence state (related covalent
bond strength)
The electronegativity (Xz) of the metal ions is plotted against the electrostatic forces
parameter (Zrk2) (Figure 13) Three clear areas of acidity can be seen for metal ions
in which electrostatic forces dominate Ž values are gt 066 Acids which are clearly
dominated by electronegativity (large covalent property) are found in the right bottom
of the graph Their Ž values are below zero The borderline acids lie between these
two and have Ž value that range between zero and 066
Figure 13 Electrostatic or covalent nature of Lewis acids
50
The calculated Ž value gives a quantitative classification for Pearsonrsquos qualitative
description of hard and soft acid and bases107 and is in good agreement with Pearsonrsquos
data
132 Spectroscopic measurements
IR spectroscopy
Ethyl acetate was used as an infrared probe by Lappert109 to rank the acidity of a
number of group III and IV Lewis acids Complexation of the acid to the acyl oxygen
atom (Figure 14) requires perturbation of the C=O bond and the strength of the
donor-acceptor bond will be reflected in the extent of polarisation of the C=O bond
This may be measured by the change in the C=O bond stretching frequency (ΔνC=O)
CO
MeEtO
LA
Figure 14 Complexation of acid to acyl oxygen atom
Table 126 shows the results of the change in wavenumber when the boron trihalide
complexes to ethyl acetate110111112
Table 126 Boron trihalides
X in base
BX3
Ethyl
acetate
Δν (cm-1)
- ΔHf
(kcal mol-1)
F 119 317
Cl 176 395 (379)9
Br 191 445 (435)9
` The results show the order BBr3 gt BCl3 gt BF3 for Lewis acidity This method was
also used to determine the acidities of the group III chlorides and bromides Using
ethyl acetate as the reference base the relative acceptor strengths for the trichlorides
51
was found to be B gt Ga gt Al gt In and for the tribromides B gt Al gt In (results for
GaBr3 were not available) This order appears reasonable as it reflects the metalrsquos
ability to accept electron density from the base109 Paulingrsquos electronegativity order
for these acids is B 20 Ga 16 Al 15
Silicon(IV) germanium(IV) and tin(IV) chlorides showed only weak interactions with
the ethyl acetate especially for SiCl4 (1SiCl4 2EtOAc) and GeCl4 (1GeCl4
2EtOAc) Tin chloride showed more distinct peaks The relative strengths were
ranked in the order SnIV gtgt GeIV gt SiIV
Phenalen-1-one (Figure 15) was used as an infrared probe to measure the acidities of
several acid halides113 Initially the spectra were measured using dilute solutions in
ether However the results were confused by strong solvent absorption The samples
were then made up in their solid form No correlation could be found between the
ΔνC=O and acid strength (Table 127) which may have been due to the fact that they
were solid samples
O
Figure 15 Phenalen-1-one
Table 127 Δ Carbonyl shifts of metal halides with phenalen-1-one
Acid ΔνC=O Solid state
stoichiometry
BF3 156 1 1
PhSnCl3 150 1 1
ZnCl2 120 1 1
SnBr4 157 2 1
ZnBr2 118 2 1
SnCl4 158 2 1
GaCl3 157 1 1
52
If these results are to be taken on face value the ranking of the Lewis acids would be
SnCl4 gt SnBr4 ~ GaCl3 ~ BF3 gt ZnCl2 gt ZnBr2 These results are contradictory to
those found elsewhere in the literature (see above discussions) and are consequently
not held to be reliable
Nuclear magnetic resonance spectroscopy
The ranking of Lewis acidity by NMR spectroscopy is based on the theory that
binding of the electron acceptor (Lewis acid) to the electron donor (Lewis base)
results in the reduction of electron density in the base resulting in a down-field shift
in the signals of the basic compound used (Figure 16) The more acidic the Lewis
acid the greater the shifts in the spectra and thus the Lewis acids can be ranked
OC
HH3C
HLA
H Figure 16 Lewis acid bound to Lewis base
A scale for Lewis acids was set up using a number of in organic halides and a variety
of ethers by looking at their 1H NMR shifts on complexation114 A value of 100 was
assigned to BCl3 the strongest Lewis acid in the study the other acids being
expressed as percentage of the effect observed for BCl3 The results in Table 128
show the trend in Lewis acidity towards the different ethers used The trend appears to
follow a Lewis acidity ranking of for BCl3 gt AlCl3 gt InCl3
53
Table 128 Relative acidity of Lewis acids towards various reference bases at 28 degC
Acid THF Di-n-butylether Di-ethylether
Hz Acidity Hz Acidity Hz Acidity
BCl3 640 100 752 100 790 100
AlCl3 518 81 640 85 672 85
i-C4H9CCl2 500 78 511 68 536 68
BiCl3a 480 75 - - - -
TiCl4 - - 526 70 600 76
InCl3b 410 64 - - - -
BF3 391 61 458 61 498 63
SnCl4 - - 384 51 - -
SnBr4c 256 40 00 0 - 0
(i-C4H9)2BCl 252 40 00 0 00 -
AsCl3 110 11 - - - 0
SnI4 00 0 - - 00 0
SiCl4 00 0 00 0 00 0
GeCl4 00 0 00 0 00 0
PCl3 00 0 00 0 00 0
(i-C4H9)3B 00 0 00 0 00 0 a By extrapolation from 03 acidbase ratio b By extrapolation from 025 acidbase ratio c At 80 acidbase ratio
The polycyclic phosphites C5H9O3P (Figure 17) and C6H9O3P (Figure 18) were used
to determine the acidities of boron Lewis acids through 1H NMR spectroscopy115 In
that study the acid strengths were found to be BH3 gt B(CH3)3 asymp BF3
O
P
O
O CH3
PO OO
C5H9O3P C6H9O3P
Figure 17 Figure 18
One of the most comprehensive studies carried out on ranking Lewis acidities by
NMR spectroscopy was that by Childs et al116 The Lewis acids employed were BF3
54
BBr3 SnCl4 SbCl5 TiCl4 BCl3 SiCl4 AlCl3 EtAlCl2 Et2AlCl Et3Al2Cl3 and Et3Al
The reference bases that were selected were examples of important classes of
compounds that undergo typical organic acid catalysed reactions eg crotonaldehyde
(Figure 19 Table 129) Here Δδ is the difference in chemical shift between the free
crotonaldehyde and its Lewis acid complex for a given signal
H1
O
H3C
H2
H3 Figure 19 Crotonaldehyde
Table 129 1H and 13C NMR chemical shift differences on complexation with
crotonaldehydea
Lewis acid Proton Δδ (ppm) Carbon Δδ (ppm)
H1 H2 H3 H4 C1 C2 C3 C4
BBr3 011 093 149 051
BCl3 -065 085 135 049 68 -37 317 43
SbCl5 017 078 132 048 69 -40 276 37
AlCl3 -020 076 123 047
EtAlCl2bc -020 077 125 047
BF3 -027 074 117 044 83 -33 261 31
EtAlCl2bcd -017 067 115 038
Et3Al2Cl3b -015 069 114 039
TiCl4 003 060 103 036
Et2AlClbc -015 055 091 030 94 -20 201 23
SnC4 -002 050 087 029 78 -28 192 23
Et3Al -034 042 063 023 a In ppm chemical shift of free base 1H δ = 947 (d H1) 610 (ddq H2) 693 (m H3) 202 (dd CH3) 13C δ =
1947 (C1) 1344 (C2) 1552 (C3) 190 (C4) Δδ is positive for a downfield shift Approximately 03 M solution
in DCM at -20 degC unless otherwise stated b At -60 degC chemical shift of uncomplexed base δ = 984 (H1) 61
(H2) 698 (H3) 205 (H4) c Written in monomeric form for simplicity d 21 crotonaldehyde ndash EtAlCl2
The H-2 H-3 and H-4 resonances of the crotonaldehyde were all found to be shifted
downfield on complexation to the Lewis acid The downfield shifts seen in H-3 were
the greatest while those seen in H-2 and H-4 were smaller in magnitude but showed a
55
linear relationship to those of H-3 The shifts of H-1 appear to be random This can be
seen in the plot (Figure 110) where Δδ of H-2 and H-4 are plotted against Δδ H-3 as
the Lewis acid varies
Figure 110 1H chemical shift differences of protons of crotonaldehyde versus the H3
chemical shift difference on complexation with various Lewis acids (+ Δδ H2 Δδ
H3 Δδ H4 Δδ H3 H1 Δδ H3)
What is also clear from Table 129 is that the Δδ of the signals in the 13C NMR
spectra follow the same trends The number of acids used in this study was lower than
that used in the 1H NMR study due to problems with solubility The shifts of C-1 (the
carbonyl carbon) appear to be random and unrelated to the shifts of the C-3 carbon
There is an upfield shift of the C-2 carbon in complexation which is a striking
difference These shifts are linearly related to the shifts of C-3 and C-4 but in the
opposite direction
Similar results were found for the other bases that were used in this study for example
cyclohex-2-enone All of the bases led to linear relationships for the shifts induced by
complexation with the Lewis acid This was interpreted to mean that it was possible to
set up a Lewis acid scale based on the shifts of the H3 protons
56
133 Comparative studies
In order to determine the relevance of a particular method for determining relative
Lewis acidities many comparative studies have been carried out The results of the
UVVis experiments carried out on phenalen-1-one113 that were discussed above
(Section 131) were compared to those obtained from equilibrium constants of the
adducts
No correlation could be found between the ΔνC=O and the pK values (Table130)
However the order of acidity according to the pK values is the same that is obtained
when making use of aniline bases91 with the exception of BF3 This acid comes at the
bottom of the list and not at the top It is clear from this result that boron has a higher
propensity to bind to nitrogen over oxygen It would appear then that the pK values
are more reliable for comparative purposes
Table 130 ΔνC=O and pK values of metal halides with phenalen-1-one
Acid ΔνC=O pK
BF3 156 -113
PhSnCl3 150 -114
ZnCl2 120 -148
SnBr4 157 -155
ZnBr2 118 -172
SnCl4 158 -266
GaCl3 157 lt-40
Satchell and Satchell117 carried out a comparative study on metal halides with amides
using NMR and equilibrium measurements The equilibrium constants and NMR data
were correlated for the 11 adducts of metal halides and two benzamides (Table 131)
in ether solution (Equation 9) The bases employed are free from steric effects
RC6H4CONH2 + M(Hal)n RC6H4C(NH2) O M(Hal)n
(9)
57
Table 131 Chemical shifts and equilibrium constants for equation 9
M(Hal)n m-Methoxybenzamide p-Nitrobenzamide
pK K -δcis -δtrans pK K -δcis
AsCl3 ca -
080 ca 63 ca 80 - - - -
SbCl3 -176 58 801 ca 74 - - -
BF3 -293 850 930 815 -186 72 ca 96
ZnBr2 -294 870 860 800 -217 148 890
ZnCl2 -294 870 558 802 -217 148 892
GaBr3 -340 2510 960 864 -249 310 1010
GaCl3 - - - - -253 340 1012
The existence of two NH signals indicated that in each case the acid bound to the
oxygen of the amide (Equation 9) rather than to the nitrogen atom This is typical of
systems in which there is double bound character in the C-NH2 bond which inhibits
free rotation and leads to chemically distinct NH atoms If the Lewis acid is bound to
the carbonyl oxygen significant electron density is donated to the C=O system by the
nitrogen atom thereby increasing the C-NH2 bond character The results in Table 131
show that a linear relationship was found between the shifts of the NH signal and the
equilibrium constants of the adducts This indicates that for these systems the NMR
shifts can be used as a guideline to relative Lewis acidities
In a follow up paper to this initial communication the authors found results that were
contrary to their initial findings118 In this study metal halides were complexed to
benzamides in diethyl ether The NMR shifts and equilibrium constants were
calculated
The values of K show the relative acidities for adducts with benzamides be AlCl3 gt
SnCl4 gt GaCl3 asymp GaBr3 gt ZnCl2 asymp ZnBr2 ge BF3gt PhSnCl3 gt SbCl3 gt AsCl3 Towards
substituted benzamides the order was BF3 asymp AlCl3 asymp GaBr3 asymp GaCl3 asymp SnCl4 gt ZnCl2
asymp ZnBr2 gt PhSnCl3 gt SbCl3 and towards perinaphthenone GaCl3 gtgt SnCl4 gt ZnBr2 gt
ZnCl2 gt PhSnCl3 gt BF3
58
While the lists show similar relative Lewis acidities with all bases the benzamides
differentiate the acids to a larger degree According the authors the lower position of
BF3 when complexed to the ketone is possibly due to boron preferring to bind to the
nitrogen over oxygen
The chemical shifts for the NH protons in the 1H NMR spectra for the adducts of 3-
methoxy and 4-nitro-benzamide with the metal halides could not be correlated with
the equilibrium constants (Figure 111) This is contradictory to previous
findings117and would suggest that for wider ranges of acids the heats of reactions and
the measurements of chemical shifts are not a good indication to relative acidities118
Figure 111 Plot of δcis against K
A bull 3-methoxybenzamide B 4-nitrobenzamide 1 AsCl3 2 SbCl3 3 PhSnCl3 4 BF3 5 ZnBr2 6 ZnCl2 7
GaBr3 8 GaCl3 9 AlCl3
In another study the Lewis acidities of trimethylchlorosilane -germane and -stannane
were investigated using two independent techniques calorimetry and 31P NMR
spectroscopy114 The heats of reaction and the equilibrium constants were determined
for the adduct of these Lewis acids with triphenylphosphine oxide and the data were
correlated to the shifts on the phosphorous NMR spectra using triphenylphosphine
oxide as the base
59
The enthalpies and equilibrium constants calculated showed the tin adducts to be
stronger Lewis acids than the germanium derivatives The results of silane adducts
were not reproducible until the experiments were carried out in strictly anhydrous
conditions at which stage they exhibited Lewis acidity similar to germanium
The order of acidity according to the calorimetric data was found to be Sn gt Ge gt Si
These findings correspond to most other findings in the literature83 However they are
in the opposite order to the findings of Graddon and Rana120121 who concluded that
trimethylsilane is a stronger acid than trimethylstannanne The excessive heats of
formation found for the silane adducts in the experiments carried out by Graddon and
Ranna are likely due to hydrolysis of the silane in the presence of moisture119
The shifts in the 31P NMR spectra for the complexation to triphenylphosphine oxide
in benzene are inconclusive For trimethylchlorostanne the shift obtained can be
correlated with the enthalpy data but is contrary to data found in literature122 No shift
is seen for (CH3)3GeCl This adduct also shows a very small ΔH value and the
experiment is not reproducible for (CH3)3SiCl Again this dataset brings into question
the validity of using NMR shifts for ranking Lewis acidity
A linear relationship was found between Gutmannrsquos acceptor number method123124
and Childsrsquos116 NMR method of ranking Lewis acidity when B(C6F5)3bullL L =
CH3COOEt (1) Et3PO (2) were synthesised and ranked along with other Lewis
acids125
The acceptor number method (AN) defined by Gutmann is based on the 31P chemical
shifts of triphenylphosphine oxide measured in solution The AN represents the
electrophilic properties of the solvent relative to hexane123124
These results confirmed the relative Lewis acidities BCl3 ~ AlCl3 gt B(C6F5)3 with
TiCl4 and SnCl4 being much weaker acids125125 The excellent correlation between
Gutmannrsquos AN and Childsrsquos 1H NMR method for these acids can be seen in Figure
112 Furthermore Lappert109 also gives two almost identical ΔνC=O values for EtOAc
adducts of these compounds
60
Figure 112 Least square plot of Childsrsquo Lewis acid scale against AN values 1 CH2Cl2
2 SnCl4 3 TiCl4 4 B(C6F5)3 5 AlCl3 6 BF3 7 SbCl5 8 BCl3 and 9BBr3 Best line fit AN = 9221C +
1598 (R2 = 097) AN 2-5 [ref 123] 1[refs 123-124 82ndash84] [ref 127] Childsrsquos data [ref 116] except 4 [ref 128]
Group 13 trihalide complexes of 9-fluorenone were prepared and characterised by
NMR IR and UVVis spectroscopy127 Where crystallisation was possible the
compounds were characterised by single crystal XRD These complexes were then
used to examine the suitability of certain parameters for the measuring of Lewis
acidity (Table 132)
The shift in carbonyl IR frequencies follows the same trends as those seen by
Lappert109 The UVVis spectra can be correlated with the theoretical data calculated
by Laszo129 However these data cannot be correlated with each other The shifts in
the 13C NMR spectra have a reverse order to all other parameter studies and seem to
have an almost inverse relationship with the ΔG
Solution dissociation energies are an indication of BDE (bond dissociation energy) for
simple Lewis acid-base complexes however in this study it is clear that ΔH is not an
indication of Lewis acid strength For example the ΔH for GaBr3 is the largest but has
the highest dissociation constant This suggests that the entropy term is important in
defining the Lewis acid ndash base interaction in solution
61
An alternative approach to determining the Lewis acidity is to study the distortion on
the crystal structure from the ideal tetrahedron expected from many systems129 From
these results it became clear that the deviation from planarity was related to the M-O
bond distance which is not a good measure for Lewis acidity for complexes with
dissimilar metals129
In summary it is clear from this study that none of these parameters gave a clear
indication of the Lewis acidity However the authors proposed that these parameters
together with kinetic data would make an ideal essential component in an analysis of
Lewis acidity
Table 132 Summary of spectroscopic and structural trends for MX3 (9-fluorenone)
Technique Parameter Trend
IR Δvc=o BCl3 gt AlBr3 gt GaCl3 gt GaBr3 gt AlCl3
UV-vis Δλmax BCl3 gt AlBr3 gt AlCl3 gt GaBr3 gt GaCl3 13C NMR ΔδC=O GaCl3 gtGaBr3 gt AlCl3 gt AlBr3 gt BCl3 1H NMR Keq at 298 K BCl3 gt AlBr3 gt AlCl3 gt GaBr3 gt GaCl3
Δ H GaBr3 gt BCl3 gt AlBr3 gtAlCl3 gt GaCl3
Δ G BCl3 gt AlBr3 gt AlCl3 gt GaBr3 gt GaCl3
XRD O(1)-C(9) AlBr3 gt AlCl3 gt BCl3 gt GaBr3 gt GaCl3
Σ(X-M-X) AlBr3 gt AlCl3 gt BCl3 gt GaBr3 gt GaCl3
Δox AlBr3 gt AlCl3 gt BCl3 gt GaBr3 gt GaCl3
The Lewis acids 59-62 (Figure 113) below were synthesised and their relative Lewis
acidities determined through two different NMR spectroscopic methods130 The first
method employed was that by Gutmann123124 The solvent used was changed from
THF to benzene due to polymerisation of the THF The second the method used was
that of Childs116
62
F
F
FF
F
BFF
FF
FF
F F
FF
FFF
FF
B
F
F
FF
F
FF
F
FF
O
F
F
FF
F
B
F
F
FF
F
FF
F
FF
OO
F
F
FF
F
B
F
F
FF
F
FF
F
FF
OO
O
59 60
61 62
Figure 113 Perfluorophenyl boron Lewis acids 59-62
Two non-fluorinated boron acids B(C6H5)3 and B(OC6H5)3 were also compared in the
study The results using the Gutmann method showed that Lewis acidity increases
with the following order 59 lt 60 lt 61 lt 62 Childsrsquos method resulted in the order
been reversed 59 gt 60 gt 61 gt 62 Moreover the linear relationship between these two
methods that was previously reported124 was not seen in this study (Figure 114) This
suggests that the linear relationship does not exist for all Lewis acids but that different
bases will give different trends
Figure 114 Graph comparing relationship between Guttmannrsquos
method and that of Childs
63
Conclusions
What is evident from the discussion is that despite the challenges many attempts have
been made to quantify Lewis acidity Lappert109 suggested the use of the change in the
carbonyl stretching frequencies using IR spectroscopy The work of Childs111 is based
on NMR spectroscopy and the shift in 1H and 13C NMR spectra the greater the shift
the more acidic the Lewis acid Thermodynamic data have been calculated for many
Lewis acidndashbase adducts Amongst those data are equilibrium constants83 and heats of
complex formation85
The comparative studies that have been carried out test these methods for their
relevance for determining Lewis acidities Some of them find correlation between
methods but most of them find that the thermodynamic data based on equilibrium
constants or the energy released on the formation of an acidndashbase complex for
example are the most reliable In particular the equilibrium constants are found to be
especially consistent particularly when a wide variety of Lewis acids is being used
This said it would appear that spectroscopic data show limited application for ranking
Lewis acidity and may be primarily useful when fewer Lewis acids of a similar type
are being used
The statement of Lewis remains true and the quest for an absolute ranking of Lewis
acidity or for predicting Lewis acid behaviour in different solvents and with different
acceptors remains elusive The problem of Lewis acidity in aqueous media has yet to
be addressed
14 Lewis acids in aqueous media
Metal triflates have been found to act as Lewis acid catalysts in aqueous media or
water-containing solvents131-132 This is advantageous from an environmental point of
view and also eliminates the need for time-consuming drying of solvents and
reagents
64
The Mukaiyama Aldol reaction
The Mukaiyama aldol reaction was carried out using commercially available
formaldehyde and a variety of silyl enol ethers The reactions were catalysed by 5-10
mol Yb(OTf)3134 and the aldol products were obtained in high yields What makes
this reaction exceptional is the fact that commercial formaldehyde is an aqueous
solution A range of metal triflates was tested using the reaction between 63 and
formaldehyde (Scheme 128) After 24 hours all the lanthanum triflates used had
successfully catalysed the reaction to afford the desired aldol product yields (Table
133)
Ph
OSiMe3Ln(OTf)3THF
CH2O aq Ph OH
O
+
63 Scheme 128 Ln(OTf)3 catalysed aldol reaction of 63 with formaldehyde
Table 133 Yields () reaction of 63 with formaldehyde catalysed by Ln(OTf)3
Yield ()
100 mol 20 mol
Entry Ln(OTf)3 24 h 1 h 36h
1 La(OTf)3 90 23 88
2 Pr(OTf)3 92 40 80
3 Nd(OTf)3 74 6 89
4 Sm(OTf)3 92 51 91
5 Eu(OTf)3 92 28 93
6 Gd(OTf)3 92 20 79
7 Dy(OTf)3 89 20 85
8 Ho(OTf)3 91 38 86
9 Er(OTf)3 90 44 83
10 Yb(OTf)3 94 5 94
The study was then expanded to include other aldehydes135 The model reaction
examined was 1-trimethylsiloxycyclohexene (64) with benzaldehyde (Scheme 129)
The reactions were carried out in a mixture of THFH2O (41)
65
OSime3
+PhCHOLn(OTf)3 10 mol
THFH2O (41) rt 20h
Ph
OH O
64 Scheme 129 Aldol reaction between 1-trimethylsiloxycyclohexene (64) and
benzaldehyde catalysed by Ln(OTf)3
The results of the reactions when different triflates were used can be seen in Table
134 The majority of the Ln(OTf)3 systems catalysed the reaction to afford the aldol
product in good yields Interestingly when the reactions were carried out in THF or
water alone the yields were low
Table 134 Yields () reaction between 1-trimethylsiloxycyclohexene (45) and
benzaldehyde catalysed by Ln(OTf)3
Ln(OTf)3 Yield () Ln(OTf)3 Yield ()
La(OTf)3 8 Dy(OTf)3 73
Pr(OTf)3 28 Ho(OTf)3 47
Nd(OTf)3 83 Er(OTf)3 52
Sm(OTf)3 46 Tm(OTf)3 20
Eu(OTf)3 34 Yb(OTf)3 91
Gd(OTf)3 89 Lu(OTf)3 88
The effect of different Yb3+ salts was also investigated to determine the role of the
counterion Only low yields of the aldol product were obtained when the Cl- OAc-
NO3- and SO4
2- salts were used in the reactions This suggests that the higher Lewis
acidity bought on by the less nucleophilic counterion promoted the desired reaction
To determine the effect of the amount of water on these reactions a model reaction
was used (Figure 115) where increasing amounts of water are added to sequential
reactions136
66
Figure 115 Yield () of aldol product
As can be seen from the Figure 114 above the best yields are obtained when 10 ndash
20 water is present relative to the THF The yields drop dramatically if the water
content is increased beyond this point
In a later study carried out by Kobayashi et al137 on the aldol reaction numerous
Lewis acids were used in a model reaction (Scheme 130) in an attempt to correlate
the catalytic activity of the Lewis acids in water and their hydrolysis constants and
their water exchange rate constants
PhCOH + Ph
OSiMe3
Ph PH
OH OMXn (02eq)
H2OTHF (91)rt 12 h
Scheme 130 Aldol reaction between benzaldehyde and 1-phenyl-1-
(trimethylsiloxy)propene catalysed by various Lewis acids
The reactions were carried out in a mixture of THFH2O (91) The reactions were
only really successfully catalysed by metal triflates the exception to this was a few of
the perchlorates where the yields were below 50 for the expected aldol adduct
The water exchange rate (represented by the water exchange rate constant WERC)
and the extent of hydrolysis (represented by the hydrolysis constant pKh) of the Lewis
acid in question play a large role in whether or not these reactions will proceed Metal
compounds that gave yields greater than 50 of the aldol product all have WERC gt
32 x 106 M-1s-1 and pKh values from 43 to 1008 In this context the word
67
ldquohydrolysisrdquo is taken to define the reaction of water with the Lewis acid to form the
corresponding metal hydroxide and H+
In summary it is believed that the pKh values allow the dissociation and hydration of
the metal as soon as it is added to the aqueous media and the fast WERC enables the
aldehyde to bind to the metal causing it to be activated and allowing attack by the silyl
ether
The Mukaiyama aldol reaction was investigated by Loh et al138139 By using
Kobayashirsquos conditions they investigated the effects of InCl3 in H2O on the reaction
The results (Table 134) were inconsistent and showed a strong dependence on the
order in which the reagents and catalyst were added The heterogeneous nature of the
reaction mixture may have contributed to the inconsistencies of the results
Table 135 Mukaiyama aldol reactions in water catalysed by InCl3
Entry Aldehyde Silyl ether SequenceYield
()
1 A 51 OSime3
2 B 74
3 C 79
4
C 82
5 HCOCH2OH H2O C 80
CHO
N
CHO
Sequence A aldehyde + InCl3 then H2O then Silyl ether (15 h)
Sequence B aldehyde + InCl3 then H2O (15 h) then Silyl ether
Sequence C aldehyde + InCl3 then Silyl ether then H2O (15 h)
The aldol reaction of glucose-derived silyl enol ethers with commercially available
formaldehyde was investigated (Scheme 131)140 A range of Lewis acid catalysts was
investigated in which InCl3 afforded good yields and excellent selectivities The
corresponding triflate showed considerably lower yields as did Yb(OTf)3
68
OTMSO
TBDMSO
BnO OO
OO
TBDMSO
BnOO
O
OH
Lewis acidCH2O (37 aq) rt
Scheme 131 Aldol reaction of glucose-derived silyl ether with aqueous formaldehyde
catalysed by Lewis acids
Allylation reactions
The allylation reaction of carbonyl compounds was reported to take place successfully
using a catalytic amount of Sc(OTf)3 in aqueous THF141 The reactions proceeded
smoothly in either a 91 or 41 mixture of THFH2O to afford the expected products
in high yields Unprotected sugars reacted directly to form the required adducts It
was found that Yb(OTf)3 was also an effective catalyst for these reactions
The tin-mediated allylation of carbonyl compounds was investigated in the presence
of InCl3 in water (Scheme 132)142 Yields obtained for the reactions were moderate to
high after 15 hours and in most cases the diastereoselectivities were good The InCl3
is believed not to play a large role in yields but plays a significant role in the outcome
of the of the final isomer ratio
R1 H
O
+
R2 Br
Sn InCl3
H2O rtR1
R2R1
R2
OH OH+
anti syn Scheme 132 Tin-mediated allylation of carbonyl compounds catalysed by InCl3
Especially noteworthy is the high anti diastereoselectivity (98 de) observed (Table
135 entry 3 and 8) The level of diastereoselectivity of the reaction is lower when it is
carried out without InCl3 as the catalyst
69
Table 136 Indium trichloride promoted tin mediated allylation of aldehydes
Entry Aldehyde Halide Conditions Yield
()c
Isomer
ratio
(antisyn)b
1 C6H5CHO Me Br Sn InCl3 H2O (15 h) 80 5050
2 C6H5CHO Ph Br Sn InCl3 H2O (15 h) 45d 991
3 C6H5CHO EtO2C Br Sn InCl3 H2O (15 h) 96 855
4 CH2BrCHO EtO2C Br Sn InCl3 H2O (15 h) 55 8020
5 3-C5H4NCHO EtO2C Br Sn InCl3 H2O (24 h) 51 8020
6 C6H11CHO EtO2C Br Sn H2O (18 h) 60 8515
7 C6H11CHO CO2Et
Br Sn InCl3 H2O (15 h) 65 6832
8 C6H11CHO EtO2C Br Sn InCl3 H2O (15 h) 65 991 aall reactions were carried out in 05-1 mmol scale b The isomer ratio was determined by 1H or 13C NMR c
Isolated yields d Balance of material is unreacted aldehyde
The strong preference of the reaction for the anti adduct suggests that transmetallation
is involved Transmetallation from allyltin with indium trichloride as Lewis acid
would proceed via SE2 process to produce compound 65 (Scheme 133) which would
further rearrange to compound 66 and its isomers 67 the former being favoured due to
its relative thermodynamic stability No isomerisation was observed by the authors
during the reactions The high anti selectivity can be explained by a six membered
transition state depicted in Scheme 133
SnBrR InCl3H2O
R
InCl2
R InCl2 +
R
InCl265 66 67
R InCl266
RCHOO
InCl2RH
RR
R
OH
antiR = Ph 3-Pyr Cyclohexyl BrCH2 Scheme 133 Transmetallation from allytin with InCl3
Araki et al143 reported the allylation of aldehydes and ketones using catalytic
amounts of InCl3 in combination with aluminium or zinc metal These reactions were
70
carried out in a THFH2O (52) mixture at room temperature One disadvantage of
these reactions is that they are slow when compared to those that are catalysed by
stoichiometric amounts of indium they require days to complete Interestingly when
the reactions are carried out in anhydrous THF the yields drop dramatically and side
reactions occur
Michael additions
Ytterbium triflate was found to catalyse the Michael addition of β-ketoesters to αβ
unsaturated ketones in water143 Several Michael donors were used in a reaction with
methyl vinyl ketone (Scheme 134) in the presence of 10 mol Yb(OTf)3 in water
The reactions were stirred at room temperature for 5 days and quantitative yields were
obtained in all cases If the reactions are conducted without a catalyst a yield of only
40 is obtained after 14 days
R OR
O O
O R
ORR
OO
OR
+Yb(OTf)3
H2O
Scheme 134 Michael addition of β-ketoesters with methyl vinyl ketone catalysed by
Yb(OTf)3
Other reactions
Erbium triflate has been used for the aminolysis of epoxides in water145 Both
aliphatic and aromatic amines could be used and the corresponding β-amino alcohols
were afforded in high yields The reactions required only 5 mol of Er(OTf)3 and
typical reactions times were between 2-8 hours
The catalyst could be recycled from the aqueous layer of the workup mixture dried
under reduced pressure and heat and reused without loss of activity up to three times
The pH of a 01 M solution of Er(OTf)3 was found to be 59 only mildly acidic The
aqueous layer from the work up was found to be even less acidic with pH 66 It was
also determined that triflic acid was not the active catalyst by carrying out a reaction
using 10 mol triflic acid The yield of this reaction was only 23 after 7 hours at 25
71
degC in contrast to the same reaction carried out with 5 mol Er(OTf)3 which shows an
88 yield after 8 hours at room temperature
Bi(OTf)3 catalysed the direct-type Mannich reaction of cyclohexanone an aromatic
aldehyde and an aromatic amine146 These reactions proceeded smoothly in water to
give the corresponding β-amino ketone
H
O O
Ph NH2+ + Conditions
ONHPh
Scheme 135 Direct-type catalytic Mannich reaction of benzaldehyde aniline and
cyclohexanone in water
Initial reactions were carried out using benzaldehyde aniline and cyclohexanone in
water (Scheme 135) These reactions were conducted using several different acids
(Table 136) in order to determine optimal reaction conditions
Table 137 Direct-type catalytic Mannich reaction of benzaldehyde aniline and
cyclohexanone in watera
Entry Conditions antisynb Yieldc ()
1 15 TfOH 7723 92
2 5 Bi(O2CCF3)3 7723 77
3 1 Bi(OTf)3 7228 94
4 5 Bi(OTf)3 8614 84
5 10 Bi(OTf)3 8317 97 a The reaction was conducted at 25 degC for 7 h in water b antisyn ratio calculated by 1H NMR c Isolated yield
When the catalyst loading of Bi(OTf)3 is changed from 1 mol to 5 mol it has a
positive effect on the stereoselectivity of the product By further increasing the
catalyst loading more the stereoselectivity is not improved 5 mol was thus chosen
as part of the standard conditions Triflic acid catalyses this reaction effectively and it
may be possible that the true catalyst when Bi(OTf)3 is used in the reaction is triflic
acid The model reaction (Scheme 135) was carried out using 5 mol Bi(OTf)3 to
which was added 15 equivalents (compared to Bi(OTf)3) of the sterically hindered
72
base 26-di-tert-butylpyridine The reaction was left to proceed for 7 hours at room
temperature after which the yield of the reaction was 83 (antisyn 7525) This
result indicates that a Lewis acid is involved in the process but as previously
indicated the pyridinium salt itself may catalyse the reaction a possibility for which
has to be specifically tested
When the reaction is repeated with no other catalyst besides the pyridinium base
itself the yield is 76 If the initial reaction is carried out with 5 mol Bi(OTf)3 and
a different proton scavenger K2CO3 the yield of the reaction is only 44 This
indicates that a Broslashnsted acid is involved in the process The optimal conditions were
then used on a variety of aldehydes and anilines Various substituted benzaldehydes
reacted with aniline or p-chloroaniline to give good yields
The Biginelli reaction is a condensation reaction between a β-ketoester an aldehyde
and urea under strongly acidic conditions146 The catalytic activities of metal
triflimides (triflimide is a contraction for bis(trifluoromethane)sulfonimide) as Lewis
acid catalysts for the Biginelli reaction were investigated and found to be effective147
The catalytic activity of different metal triflimides was investigated using the reaction
between p-anisaldehyde ethyl acetoacetate and urea (Scheme 136) Metal triflates
which are known to catalyse the Biginelli reaction148 were also used in this study for
comparative reasons
MeEtO
OO
H2N NH2
O+NH
NH
REtO2C
O
MeO CHOMeO
M(NTf2)norM(OTf)n
water rt 24 h
Scheme 136 Biginelli reaction of p-anisaldehyde ethyl acetoacetate and urea
The reactions were carried out using 5 mol of the Lewis acid in water The metal
triflimides catalysed the reactions to afford moderate to good yields in all cases but
one The reaction with Zn(NTf2)2 only gave 8 of the reaction product (Table 18)
The reactions carried out with the metal triflates were very poor the yields were
below 10 in all cases
73
Table 138 Reaction between p-anisaldehyde ethyl acetoacetate and urea catalysed by
a variety of Lewis acidsa
Entry Lewis acids Yields ()
1 Ni(NTf2)2 59 65c 40d
2 Ni(OTf)2 Complex mixture
3 Cu(NTf2)2 65 70c 62d
4 Cu(OTf)2 lt10b
5 Zn(NTf2)2 8
6 Yb(NTf2)3 88 90c
7 Yb(OTf)3 lt5b
a The reactions conditions p-anisaldehyde (10 mmol) ethyl acetoacetate (10 mmol) urea (15 mmol) and a Lewis
acid in water (1 mL) 24 h at rt b Considerable amounts of inseparable by-products were accompanied by the
Biginelli product c The reaction was carried out for 72 h d The reaction was carried out at 70 degC
In an attempt to improve the yields of the reactions with the metal triflimides a
catalytic amount of a Broslashnsted acid was added to the reaction mixture (Table 139)
Interestingly the results showed considerable improvements in the yields
Table 139 Reaction between p-anisaldehyde ethyl acetoacetate and urea catalysed by
a variety of Lewis acids in the presence of a Broslashnsted acida
Entry Lewis acid Additive Yield ()
1 Ni(NTf2)2 CH3CO2H 84
2 Ni(NTf2)2 CF3CO2H 90
3 Ni(NTf2)2 HNTf2 71
4 Ni(NTf2)2 HCl 71
5 Ni(NTf2)2 CH3SO2H 92
6 Cu(NTf2)2 CF3CO2H gt95
7 Cu(NTf2)2 HCl gt95
8 Cu(NTf2)2 HNTf2 65
9 Cu(NTf2)2 HNTf2 gt95
10 Yb(NTf2)3 HCl 85 a The reactions conditions p-anisaldehyde (10 mmol) ethyl acetoacetate (10 mmol) urea (15 mmol) Broslashnsted
acid (5 mol) and a Lewis acid (5 mol) in water (1 mL) 24 h at rt
74
To determine whether the Broslashnsted acids would catalyse the reactions independently
the reactions were repeated with the Broslashnsted acids only as the catalysts The results
in Table 139 clearly indicate that the Lewis acid is required for improved activation
Table 140 Reactions between p-anisaldehyde ethyl acetoacetate and urea catalysed
by a variety of Broslashnsted acidsa
Entry Broslashnsted acid (pKa) Yield ()
1 CH3CO2H (47) lt5
2 CF3CO2H (023) 56
3 HCl (-23) lt5
4 HNTf2 (12) 33
5 CH3SO2H (-12) lt5
6 p-TsOH lt5 a The reactions conditions p-anisaldehyde (10 mmol) ethyl acetoacetate (10 mmol) urea (15 mmol) Broslashnsted
acid (5 mol) in water (1 mL) 24 h at rt b Values in parenthesis indicate the pKa of the acids
The pKa values of the acids are not indicative of their ability to catalyse the reactions
and more complex modes of activation are indicated As an example of this entries 4
7 and 10 in Table 138 are instructive The metal triflimides and HCl were used in
reactions in the presence of different aldehydes ethyl acetoacetate and urea In
reactions of varying levels of success
15 Conclusions
This literature overview has demonstrated the applicability of metal triflates to a
significant number of organic transformations It was demonstrated that these Lewis
acids in the majority of instances outperform traditional Lewis acid catalysts as
measured against various performance characteristics Amongst others many of the
metal triflates have been shown to be useful in catalytic quantities and also to be
readily recyclable They are not only water tolerant by may also be used to effect
catalysed reactions in binary (waterorganic solvent) solvent systems The call for
further investigation into the application of metal triflate-based Lewis acids is clear
75
This overview furthermore highlighted the many excellent efforts expended in
attempts to rank Lewis acids in general It points to the various successes and to some
of the difficulties experienced in executing this particular task This aspect also
indicates that there remains much work to be done in this area
The present study aims to address some of the issues raised Firstly it seeks to expand
on the types of reactions that may be effected by metal triflates with a focus on
Al(OTf)3 as catalyst Particular attention is given to the acetalisation reaction of
aldehydes and ketones in which the compatibility of some acid sensitive protecting
groups with the reaction in question is investigated Secondly work performed to
investigate the possible role of water in Lewis acid-mediated transformations is
detailed This aspect of the study raised secondary questions along the way and the
efforts to address these questions are also provided In particular one aspect of the
study called for what is termed herein ldquosuper dry solventsrdquo in which the water
content of the solvent is demonstrably in the low single-digit ppm region This work
required accurate measurement techniques as well as simple yet efficient method for
drying various solvents
This study also touches on the ranking of Lewis acids with a particular focus on the
metal triflates This investigation involved the use of NMR UVVis and infrared
spectroscopy in an attempt to determine a suitable ranking
76
16 References
1 Kolthoff I M J Phys Chem 1944 48 51
2 Walden P Salts Acids and Bases McGraw-Hill Book company New York N
Y 1929
3 Luder W F Chem Rev 1940 27 547
4 Bell R P Q Rev Chem Soc 1947 1 113
5 Jensen W B Chem Rev 1978 78 1-22
6 Franklin E C J Am Chem Soc 1905 27 820
7 Franklin E C J Am Chem Soc 1912 47 285
8 Franklin E C J Am Chem Soc 1924 46 2137
9 Cady H P Elsey H M J Chem Ed 1928 5 1425
10 Smith G B L Chem Rev 1938 23 165
11 Lewis G N J Am Chem Soc 1916 38 762
12 Lewis G N Valence and the Structure of Atoms and Molecules The Chemical
Catalogue Co New York NY 1923 pp 144-142
13 Broslashnsted J Recl Trav Chim Pays-Bas 1923 42 718
14 Lowry T Chem Ind (London) 1923 42 43
15 Sidgwick N V The Electronic Theory of Valency Clarendon Press Oxford
1927 pp 60 and 116
16 Lapworth A Nature (London) 1925 115 625
17 Lapworth A Mem Proc Manchester Lit Phil Soc 1925 69 xviii
18 Robinson Outline of an Electrochemical (Electronic) Theory of the Course of
Organic Reactions Institute of Chemistry London 1932 pp 12-15
19 Ingold C K J Chem Soc 1933 1120
20 Ingold C K Chem Rev 1934 15 225
21 Lewis G N J Franklin Inst 1938 226 293
22 Acids and Bases a Collection of Papers Journal of Chemical Education Easton
Pa 1941
23 More Acids and Bases a Collection of Papersrdquo Journal of Chemical Education
Easton Pa 1944
24 Luder W F Zuffanti S The Electronic Theory of Acids and Bases Willey
New York NY 1946
25 Usanovich M Zhur Obschei Khim 1939 9 182
77
26 Huheey J E Keiter A E Keiter R L Inorganic Chemistry Principles of
Structure and Reactivity Harper Collins New York 1993
27 Mulliken R S Pearson W B Molecular Complexes A Lecture and Reprint
Volume Willey-Interscience New York NY 1969 (Contains complete reprints of
Mulliken papers)
28 Hudson R F Klopman G Tetrahedron Lett 1967 12 1103
29 Hudson R F Klopman G Theor Chim Acta 1967 8 165
30 Klopman G J Am Chem Soc 1968 90 223
31 Klopman G Sigma Molecular Orbital Theory Sinaoglu O Wiberg K B Ed
Yale University press New Haven Connecticut 1970 pp 115
32 Klopman G Chemical Reactivity and Reaction Paths Klopman G Ed Wiley-
Interscience New York NY 1974 Chapter 4
33 Bell R P Acids and Bases Meuthuenrsquos Monographs of Chemical Subjects
London 1969
34 Bell R P The Proton in Chemistry Meuthuenrsquos Monographs of Chemical
Subjects London 1959
35 Denmark S E Beutner G L Angw Chem Int Ed 2008 47 1560
36 Yoshida A Hao X Yamazaki O Nishikido J QSAR amp Comb Sci 2006 25
697
37 Tarasow T M Eaton B E Cell Mol Life Sci 1999 55 1463
38 Schinzer D Selectivities in Lewis acid Promoted Reactions Kluwer Academic
Publishers Dordrecht 1989
39 Lewis acids in Organic Synthesis Yamamoto H Ed Wiley-VCH Weinheim
2000
40 Mukaiyama T Narasaka K Banno T Chem Lett 1973 1011
41 Olah G A Friedel-Crafts and Realated Reactions Wiley Interscience New
York 1973
42 Kobayashi S Sugiura M Kitagawa H Lam WWL Chem Rev 2002 102
2227
43 Vorbruumlggen H Krolikiewiez K Angew Chem Int Ed 1975 14 421
44 Persson I Drsquo Angelo P De Panfilis S Sandstroumlm M Eriksson L Chem
Eur J 2008 14 3056
45 Tran A T Deydier S Bonnaffeacute D Le Narvor C Tetrahedron Lett 2008 49
2163
78
46 (a) Mirinda P O Ramirez M A Martin V S Padron J I Org Lett 2006
8 1633 (b) Miura K Wang D Matsumoto Y Hosomi A Org Lett 2005
7 503
47 Wang Z Hammond G B J Org Chem 2000 65 6547
48 Arimitsum S Hammond G B J Org Chem 2006 71 8665
49 Alcaide B almendros P Mart T Angew Chem Int Ed 2006 45 4501
50 Miyabe H Ueda M Nishimura A Naito T Tetrahedron 2004 60 4227
51 Kobayashi S Eur J Org Chem 1999 15
52 Kawada A Mitamura A Kobayashi S Synlett 1994 545
53 Shiina I Suzuki M Tetrahedron Lett 2002 43 6391
54 Noji M Ohno T Fuji K Futaba N Tajima H Ishii K J Org Chem
2003 68 9340
55 Kobayashi S Nagayama S Busujima T J Am Chem Soc 1998 120 8287
56 a) Hayashi T Prog Polym Sci 1994 19 663 b) Chiellini E Solaro R Adv
Matter 1996 8 305 c) Okada M Prog Polym Sci 2002 27 87
57 a) Nomura N Taira A Tomioka T Okada M Macromolecules 2000 33
1497 b) Moller M Kange R Hendrik J J Polym Sci Part A Poly Chem
2000 38 2067 c) Deng X M Yuan M L Xiong C D Li X H Appl Poly
Sci 1999 71 1941
58 Reference 52 lists a further 20 references were rare earth metal triflates have been
used to catalyse ROP reactions or other polymer reactions
59 Nomura N Taira A Nakase A Tomioka T Okada M Tretrahedron 2007
63 8478
60 Epifano F Genovese Salvatore G Curini M Tetrahedron Lett 2007 48
2717
61 Procopio A Dalpozzo R De Nino A Maiuolo L Nardi M Org Biomol
Chem 2005 3 4129
62 A)Bizier P N Atkins S R Helland L C Colvin S F Twitchell J R
Cloninger M J Carb Res 2008 343 2814 B) Brown H C Kanner B J
Am Chem Soc 1965 88 986
63 Smith B M Graham A E Tetrahedron Lett 2006 47 9317
64 Liu L Tang L Yu L Chang W Li J Tetrahedron 2005 61 10930
65 Chaminade X Chiba S Narasaka K Duntildeach E Tetrahedron Lett 2008 49
2384
79
66 Kamal A Khan M N A Reddy K S Srikanth Y V V Krishnaji T
Tetrahedron Lett 2007 48 3813
67 A)Williams D B G Lawton M Org Biomol Chem 2005 3 3269 B) Parker
R E Isaacs N S Chem Rev 1959 737-799
68 Williams D B G Lawton M Tetrahedron Lett 2006 6557
69 Cai J J Zou J P Pan X Q Zhang W Tetrahedron Lett 2008 49 5302
70 Chen J Wu D He F Liu M Wu H Ding J Su W Tetrahedron Lett
2008 49 3814
71 Ghorai M K Ghosh K Kalpataru D Tetrahedron Lett 2006 47 5399
72 a) Hiyama T Koide H Fujita S Nozaki H Tetrahedron 1973 29 3137 b)
Wu J Sun X Xia H G Tetrahedron Lett 2006 47 1509
73 Ollevier T Mwena-Mbeja T M Tetrahedron Lett 2006 47 4051
74 Kante De S Tetrahedron Lett 2004 45 2339
75 Sreedhar R Reddy P S Reddy M A Neelima B Arundhathi R
Tetrahedron Lett 2007 48 8174
76 Yamamoto H Futatuji K Angew Chem Int Ed 2005 44 1924
77 Acid catalysis in modern organic synthesis Volume 1 Yamamoto H Ishihara
K Ed Wiley-VCH Verlag GmbH amp Co Germany 2008
78 Acid catalysis in maodern organic synthesis Volume 2 Yamamoto H Ishihara
K Ed Wiley-VCH Verlag GmbH amp Co Germany 2008
79 Lewis G N J Franklin Inst 1938 226 293
80 Satchell D P N Satchell R S Chem Rev 1969 69 3 251
81 Branch C S Bott S G Barron A R Organomet Chem 2003 666 23
82 Cromwell T M Scott R C J Chem Soc 1950 72 3825
83 Keefer R M Andrews L J J Am Chem Soc 1950 72 4677
84 Blake N W Winston H J A Patterson J Am Chem Soc 1951 73 4437
85 Scott R L J Am Chem Soc 1953 75 1550
86 Keefer R M Andrews L J J Am Chem Soc 1950 72 5170
87 Moodie R B ChemInd 1961 1269
88 Mohammad A Satchell D P N Satchell R S J Chem Soc (B) 1967 727
89 Mohammad A Satchell D P N J Chem Soc (B) 1967 403
90 Keefer R M Andrews L J J Am Chem Soc 1952 74 4500
91 Drago R S Wenz D A J Am Chem Soc 1962 84 526
92 Fairbrother F J Chem Soc 1962 847
80
93 Smith J W Electric Dipole Moments Butterworths London 1955 p 86
94 Kramer G M J Org Chem 1975 40 298
95 Kramer G M J Org Chem 1975 40 302
96 (a) McCauley D A Lien A P J Am Chem Soc 1951 73 2013 (b)
McCauley D A Higley W S Lien A P J Am Chem Soc 1956 78 3009
97 Henis J M S Mabie C A J Chem Phys 1970 53 2999
98 Brownstein S Can J Chem 1969 47 605
99 McDonald J D Williams C H Thompson J C Margrave J L Advan
Chem Ser 1968 72 265
100Clifford A F Morris A G J Inorg Nucl Chem 1957 5 71
101Rhyne T C Dillard J G Inorg Chem 1971 10 730
102Haartz J C McDaniel D H J Am Chem Soc 1973 95 8562
103Kramer G M J Org Chem 1975 40 298
104Hyman H H Quarterman L A Klipatrick M Katz J J J Phys Chem
1961 65 123
105 Clifford A F Kongpricha S J Inorg Nucl Chem 1961 20 147
106 Clifford A F Beachell H C Jack W M J Inorg Nucl Chem 1957 5 57
107 Zhang Y Inorg Chem 1982 21 3889
108 Drago R S Wayland B B J Am Chem Soc 1965 87 3571
109 Lappert M F J Chem Soc 1961 103 542
110 Bax C M Katritzky A R Sutton L E J Chem Soc 1958 1258
111 Brown H C Holmes R R J Am Chem Soc 1956 78 2173
112 Greenwood N N Perkins P G J Chem Soc 1960 1141
113 Mohammad A Satchell D P N Satchell R S J Chem Soc 1967 723
114 Deters J F McCusker P A Pilger Jr R C J Am Chem Soc 1968 90
4583
115 Verkade J G King R W Heitsch C W Inorg Chem 1964 3 884
116 Childs R F Mulholland D L Nixon A J Can Chem 1982 60 801
117 Satchell D P N Satchell R S Chem Commun 1969 110
118 Satchell R S Bukka K Payne C J J Chem Soc 1975 541
119 Spencer J N Barton S C Cader B M Corsico C D Harrison L E
Mankuta M E Yoder C H Organometallics 1985 4 394
120 Graddon D P Rana B A J Organomet Chem 1977 140 21
121 Graddon D P Rana B A J Organomet Chem 1976 105 51
81
82
122 Zeldin M Mehta P Vernon W P Inorg Chem 1979 18 463
123 Mayer U Gutmann V Gerger W Monatshefte Chem 1975 106 1235
124 Gutmann V Coord Chem Rev 1976 18 225
125 Beckett M A Brassington D S Coles S J Hursthouse M B Inorg
Chem Commun 2000 3 530
126 Doumlring S Erker R Froumlhlich R Bergander K Organometallics 1998 17
2187
127 Beckett M A Strickland G C Holland J R Varma K S Polymer 1996
37 4629
128 Branch C S Bott S G Barron A R J Organomet Chem 2003 666 23
129 Laszo P Teston J Am Chem Soc 1990 112 8750
130 Britovsek G J P Ugolotti J White A J Organometallics 2005 24 1685
131 Kobayashi S Synlett 1994 9 689
132 Li C J Chem Rev 2005 105 3095
133 Loh T P Chua G L Chem Comm 2006 2739
134 Kobayashi S Chem Lett 1991 12 2187
135 Kobayashi S Hachiya I Tetrahedron Lett 1992 1625
136 Kobayashi S Hachiya I J Org Chem 1994 59 3590
137 Kobayashi S Nagayama S Busujima T J Am Chem Soc 1998 120 8287
138 Loh T P Pei J Cao G Q Chem Commun 1996 1819
139 Loh T P Pei J Koh S V Cao G Q Li X R Tetrahedron Lett 1997
38 3465
140 Loh T P Cao G Q Vittal J J Wong M W Chem Commun 1998 861
141 Hachiya I Kobayashi S J Org Chem1993 58 6958
142 Li X Loh T P Tetrahedron Asymm 1996 7 1535
143 Araki S Jin S J Idou Y Butsugan Y Bull Chem Soc Jpn 1992 65
1736
144 Keller E Feringa B L Tetrahedron Lett 1996 1879
145 Procopio A Gaspari M Nardi M Oliverio M Rosati O Tetrahedron
Lett 2008 2289
146 Kappe C O Stadler A Org React 2004 63 1
147 Suzuki I Suzumura Y Takeda K Tetrahedron Lett 2006 47 7861
148 a) Paraaskar A S Dewker G K Sudalai A Tetrahedron Lett 2003 44
3305 b) Ma Y Qian C Wang L Yang M J Org Chem 2000 65 3864
Chapter 2 Metal triflates in protection group chemistry
21 Introduction As stated in the introduction chapter one of the aims of the research described in this thesis is
to expand the use of metal triflates Al(OTf)3 in particular in organic synthesis To this end
this chapter focuses on the use of these compounds in protection group chemistry This
follows our earlier work on the catalytic ring opening of epoxides
Acetalisation is probably the most important protection strategy for carbonyl groups1 Acetal
formation is most often achieved by treating aldehydes or ketones with an excess of lower
alcohols or diols in the presence of a drying reagent and a Lewis or Broslashnsted acid2 or by
removing water through the formation of an azeotrope with the solvent and the use of a Dean-
Stark trap1
Copper(II) tetrafluoroborate is an effective catalyst for the formation of acetals3 but the BF4-
counter ion is harmful to the environment Metal triflates have previously been reported to
catalyse acetalisation reactions45 In particular Bi(OTf)3 and In(OTf)3 have been found to
effectively catalyse this reaction While efficient there are drawbacks with these catalysts
Bi(OTf)3 requires the reactions to be carried out under reflux and suffers from relatively long
reaction times5 When reactions are carried out using In(OTf)3 an aqueous workup cannot be
used as the acetal undergoes rapid hydrolysis back to the corresponding carbonyl4 making
recycling of the catalyst difficult On a slightly different note polystyrene-supported
Al(OTf)3 has been used to catalyse dithioacetalisation and trans-dithioacetalisation reactions6
Perhaps the biggest drawback of these Lewis acid catalysts is that they require the presence
of either a solvent such as CH2Cl2 an alcohol in excess or two or more equivalents of the
orthoester
With the increasing interest in metal triflates and successes in these laboratories working with
aluminium triflate7-9 it was decided to investigate the efficacy of Al(OTf)3 in the formation
of acetals using a variety of aryl and alkyl aldehydes and ketones
83
22 Acetal formation using aluminium triflate
All the steps in acetal formation are fully reversible (Scheme 21) For this reason an excess
of anhydrous alcohol is used and water is removed from the system pushing the equilibrium
to the right and an acid catalyst is used to promote the reaction10 Water can also be removed
from the system by using a drying reagent such as an orthoester to push the equilibrium to
the right (Le Chatelierrsquos principle) promoting the formation of the acetal
OH
ROH
H RO
H
RH O
HR+ C O
RH
OH
RH O R
H
CR
H OO R
H + OH RH
CR
H OH2
O R HC O
R
Hemiacetal
R H2O
O
+
C
oxonium cation
R
HR O R
HCR
H OO
R
HR OR H
CR
OO
RR
H ROH2+
Scheme 21 Mechanism of acid catalysed acetal formation
With Le Chatelierrsquos principle in mind the initial acetal reactions were performed in the
presences of 10 equivalents of the desired alcohol (MeOH or EtOH) in order to drive the
reaction to completion and 2 equivalents of the corresponding orthoester which acts as a
drying reagent removing water from the system (Scheme 22) Strictly speaking aldehydes
provide acetals while ketones provide ketals In the following text the word ldquoacetalrdquo has been
used to denote either product
O
R2R1R2RO
R1RO
1 mol Al(OTf)3ROH and or CH(OR)3
Scheme 22 Al(OTf)3 catalysed acetal formation in the presence of alcohol andor drying
reagent
The reactions were carried out using 1 mol Al(OTf)3 All reactions were allowed to run for
one hour at room temperature for the sake of comparison even though it was clear that some
reactions were over after a few minutes This was confirmed in a few instances by performing
84
the reaction in an NMR tube under similar conditions The reactions were quenched by
passing the reaction mixture through a plug of neutral alumina to remove the Al(OTf)3 The
volatiles were then removed under vacuum If further purification was necessary bulbndashtondash
bulb vacuum distillation was used
It is clear from Table 21 column A(1 h reaction 1mol Al(OTf)3 10 eq ROH 2 eq
orthoester) that with few exceptions the reactions afforded good to excellent yields of the
anticipated acetals Pleasingly anisaldehyde reacted smoothly under these conditions to give
products (Table 21 entries 1 and 2) The presence of an electron donating group (OCH3) on
the arene ring is known to retard acetal formation1 The electron density is donated into the
carbon of the carbonyl group making it less electrophilic and thus less available for attack by
the alcohol
Table 21 Al(OTf)3 catalysed acetal formation of aldehydes and ketones
a 1 h reaction 1mol Al(OTf)3 catalyst unless otherwise noted b 10 eq ROH 2 eq orthoester c Isolated yields d 2 eq orthoester no ROH e 12 eq of orthoester were used f Yields determined by 1H NMR spectroscopy g 2 h reaction h 5 h
Entry Product Aabc Yield ()
Bacd Yield ()
Caef Yield ()
1 H
MeO OMe
H3CO 21
88g 97h 75 g (90)
2 H
EtO OEt
H3CO 22
67g 90h 72 g (80)
3
MeO
H
OMe
Cl 23
76 88 gt98i
4
EtO
H
OEt
Cl 24
79 90 gt98i
5
MeO OMe
O2N 25
97 94h 82g (94)h
6
EtO OEt
O2N 26
92 91h 57g (75)h
85
reaction i gt98 implies quantitative reaction with no other products detected in concentrated samples in which spinning side bands are observable in the 1H NMR spectra of the products
Table 21 cont Al(OTf)3 catalysed acetal formation of aldehydes and ketones
a 1 h reaction 1mol Al(OTf)3 catalyst unless otherwise noted b 10 eq ROH 2 eq orthoester c Isolated yields d 2 eq ortho- ester no ROH e 12 eq of orthoester were used f Yields determined by 1H NMR spectroscopy g 2 h reaction h 5 h reaction i gt98 implies quantitative reaction with no other products detected in concentrated samples in which spinning side bands are
observable in this 1H NMR spectra of the product j Reaction carried out in DCM 2 h 4 eq of diol used
Entry Product Aabc Yield ()
Bacd Yield ()
Caef Yield ()
7 H
MeO OMe
NO2 27
57 96 95
8 H
EtO OEt
NO2 28
64 92 gt98i
9
MeO
H
OMe
29
92 99 gt98i
10
EtO
H
OEt
210
72 93 96
11 OMeMeO
H
211
90 96 gt98i
12 OEtEtO
H
212
85 96 gt98i
13
OMe
OMe
213
78 74g gt98i
14
OEt
OEt
214
81 73g gt98i
15 O
O 215
- 98j
Acetals of products substituted with deactivating groups were also readily formed in high
yields (Table 21 entries 3-8) Notably steric effects play a more significant role here than
those of electronics Ketones are known to react slower than aldehydes1 Nonetheless the
reactions with 4-nitroacetophenone give the corresponding acetals in gt90 yields When the
86
2-nitrobenzaldehyde is reacted the steric effect of the nitro group in the ortho position can be
clearly seen as the yields drop dramatically (Table 21 entries 7 and 8)
αβ-Unsaturated aldehydes (Table 21 entries 9-10) as well as aliphatic aldehydes (Table 21
entries 11-14) underwent smooth transformations to the corresponding methyl or ethyl acetals
in good to excellent yields The results of the reactions which are presented in Table 21
column B (2 eq orthoester no ROH) and C (12 eq of orthoester were used) shall be discussed
below
Green Chemistry
Increasingly people are becoming more aware of their environment Environmental events of
the past have illustrated the negative impacts that chemical activity can have on the
environment and human health11 Just one example of this is the negative ecological effect
that DDT had on the environment These kinds of environmental issues have shaped the
general publicrsquos opinion of chemicals in the last thirty or so years to the point where the
general belief is that all chemicals are toxic or otherwise hazardous11 These opinions have
motivated steps to be taken through legislation such that harmful or hazardous incidents do
not happen again11
As a result there is a big drive towards what is termed lsquogreen chemistryrsquo The definition of
green is ndash Green chemistry is carrying out chemical activities ndash including chemical design
manufacture use and disposal ndash such that hazardous substances will not be used and
generated11 Green chemistry includes all areas of chemistry from analytical to organic to
biochemistry It also encompasses all sectors of the chemical industry including
pharmaceutical and manufacturers of bulk chemicals11
Important areas of investigation in green chemistry11
bull Selection of starting material
The selection of the starting material is critical Is the substance benign If so will
using this substance mean having to use other toxic chemicals further on in the
process Hazards come in the form of flammability ozone depletion and ecosystem
destruction
bull Selecting reagents
87
The selection of reagents is similar to that of starting material in that they should be
assessed for their hazards In addition their effectiveness should be determined
including selectivity reaction efficiency and separation requirements Can the
reaction be done with other reagents that generate less waste Most importantly the
quantity needed in which to perform the reaction in question must be reduced
bull The selection of solvents and reaction conditions
The use of solvents in chemistry is usually unavoidable They also account for a large
amount of waste disposed of Are the solvents highly flammable or explosive Are
they hazardous to human health Chemicals such as chloroform are suspected
carcinogens Are the solvents being used environmentally benign Until several years
ago chlorofluorocarbons were used in refrigerators for cooling Later it was
discovered that these chemicals were responsible for ozone depletion and have
subsequently been phased out
bull Other
Other important areas of investigation include the design of safer chemicals Also
determining the mechanism of action to ensure that both humans and the environment
are safe and eliminating toxic functional groups
In light of the above it was therefore of interest to ascertain whether the acetalisation process
described in this chapter in any way conforms to some of the requirements of green
chemistry The manufacturing process of the triflate salts themselves may not conform to the
green chemistry requirements since it requires triflic acid which itself may not be
environmentally friendly However as will become clear the process to be described allows
recycling of the catalyst (minimising waste) is highly atom efficient (requiring only 12
equivalents of the orthoester produces virtually no by-products (gt95 yield) and avoids the
use of solvents
The orthoester has a dual role in the reaction it removes water but it can also act as a source
of the required alcohol in this way eliminating the need for the large excess of the solvent It
makes the reaction more atom efficient and eliminates the need to for large amounts of waste
to be disposed of at the end of the reaction Accordingly the reactions described above in
Table 21 Column A were repeated using only 2 equivalents of the orthoester and 1 mol of
the catalyst (Omitting the alcohol altogether)
88
Under these solventless conditions ie only 2 equivalents of the orthoester and the catalyst
the majority of the reactions afforded yields that were comparable with the reactions that had
been carried out in alcohol (compare Table 21 columns A and B) In fact in most cases the
yields are improved in the cases where the reactions are carried out under solventless
conditions This is especially evident in column B entries 5 6 and 10
Under the solventless conditions anisaldehyde (Table 21 column B) required an extended
reaction period before satisfactory yields of the acetals were obtained this is possibly due to
the effect of the methoxy deactivating group In contrast to the reactions carried out in
alcohol 4-nitroacetophenone also required a longer reaction time
In order to determine if the reaction could be made even more atom efficient the reactions
were repeated but this time only 12 equivalents of the orthoester were used in the presence of
1 mol of Al(OTf)3 Remarkably in the majority of the reactions the yields were improved
over those obtained using 2 equivalents of the orthoester (Table 21 column C) In most cases
quantitative yields were obtained In the case of 4-nitroacetophenone the reaction mixture
was biphasic under these conditions and initially the yields were poor However when the
reactions were allowed to proceed for 5 hours the yields improved considerably (Table 21
column C entries 7 and 8)
As already noted reactions that were carried out in an excess of the alcohol as solvent
afforded lower yields than those reactions that were carried out using only the orthoester
Al(OTf)3 is a hard Lewis acid and is oxophilic It is possible that in the reactions where an
excess of alcohol is used the oxygen atom from the MeOH or EtOH solvent competes with
the oxygen atom of the carbonyl group of the aldehyde or ketone for the aluminium metal
centre thereby reducing the activity of the catalyst
This competition would be largely removed by the absence of the alcohol solvent The Lewis
donor strengths of various solvents have been previously measured12 and methanol was
found to be one of the strongest donors for both hard and soft Lewis acids
The ldquotamingrdquo effect that the alcohols have on these reactions can be better seen from the
results presented in Table 22 (compare columns A and B) Here reactions carried out in the
presence of the alcohol and the corresponding orthoester gave high yields while those that
89
were carried out under solventless conditions (ie in the presence of only 2 equivalents of the
orthoester) afforded lower yields and in some cases no product at all
Table 22 Al(OTf)3 catalysed acetal formation of aldehydes and ketones
a1 hour reactions 1mol Al(OTf)3catalyst unless otherwise stated b10 eq ROH and 2 eq orthoester c2 eq orthoester no ROH dReactions carried out at -20 degC e12 eq of orthoester and 05 mol Al(OTf)3 30 min reaction time fIsolated yields
gYields determined by 1H NMR spectroscopy h5 h reaction time i2 h reaction time j5 mol Al(OTf)3 used kIf the reaction was allowed to proceed beyond one hour the product curiously reverted back to starting material
Entry Product Column A
Yield () abf Column B
Yield () acf Column C
Yield () adf
Column D Yield() degf
1 H
MeO OMe
216
96 63 98 92
2 H
EtO OEt
217
83h 75 89 92
3
MeO OMe
218
94hj 0 92 73 (92)i
4
EtO OEt
219
77 0 0 75 (82)i
5
MeO OMe
220
80 33 77 90
6
EtO OEt
221
83 33 80 93
7 O
OMe
OMe 222
86 0 gt98k 96
8 O
OEt
OEt 223
74 0 82 92
90
In cases where none of the desired products were formed (column B) the reactivity of the
system was high and led to a significant initial exotherm Presumably this increase in
temperature of the system gave rise to the substantial by-product formation that was seen
In order to circumvent this problem the reactions were carried out at ndash20ordmC and slowly
warmed up to room temperature still only using 20 equivalents of the orthoester and 1 mol
Al(OTf)3 (Table 22 column C) This move improved the yields dramatically
Even under these conditions reactions carried out with acetophenone and
triethylorthoformate were highly reactive and failed to afford any product On the other
hand furan derivatives are notoriously prone to polymerisation yet under these conditions
furfural afforded excellent yields of the desired acetals (Table 22 column entries 7 and 8)
Since high yields of the acetal products had been obtained previously by using 12 equiv of
the orthoester and 1 mol Al(OTf)3 (Table 21 Column C) this same methodology was
applied to the latter more reactive substrates using only 05 mol Al(OTf)3 and allowing the
reactions to proceed for 30 minutes Yields of greater than 90 were obtained for all but two
of the substrates (Table 22 entries 3 and 4) For these two cases the acetophenone products
slightly longer reaction times afforded high yields of the products It should be noted that
these are the only reaction conditions under which acetophenone diethylacetal could be
formed
23 Catalyst recycling
Previously Al(OTf)3 has been recycled from the aqueous layer of the workup mixture8 This
process involves removing the water under vacuum and heat and then drying the Al(OTf)3
under high vacuum This process is time consuming and once the Al(OTf)3 is dried steps
must be taken to ensure that it is not exposed to the atmosphere
In order to simplify this process recycling after acetalisation was carried out through
distillation of the product from the reaction vessel The reaction between benzaldehyde and
12 equivalents of trimethylorthoformate in the presence of 05 mol Al(OTf)3 was used as a
test reaction (Scheme 23)
91
H
O 05 mol Al(OTf)312 eq CH(OCH3)3 OMe
OMe
Scheme 23 Acetalisation of benzaldehyde with trimethylorthoformate catalysed by Al(OTf)3
The reaction was allowed to proceed for 30min after which the yield was determined by 1H
NMR spectroscopy The reaction mixture was distilled using bulb-to-bulb vacuum distillation
on a Kugelrohr vacuum apparatus Once all of the volatiles were removed the flask
containing the catalyst was allowed to cool and another reaction was performed in the same
vessel The catalyst was recycled successfully in this manner three times (Table 23) and
reused in successive reactions with little loss of activity
Table 23 Al(OTf)3 recycling experiments using benzaldehyde as the substrate
Run Yield ()abc
1 gt98
2 gt98
3 96
4 95 a30 min reaction time 05 mol Al(OTf)3 catalyst 12 eq trimethylorthoformate bDetermined by 1H NMR spectroscopy c
gt98 implies quantitative reaction with no other products detected in concentrated sample
24 Deprotections
Acyclic acetals are usually selected when mild hydrolysis is required otherwise the more
robust cyclic acetal can be used1 Their ease of hydrolysis depends on their substituents - the
relative rates follow the order CH2(OEt)2 [1] MeCH(OEt)2 [6000] Me2C(OEt)2 [18 x 107]1
The selective deprotection of an acyclic acetal using an aqueous solution of trifluoromethyl
acetic acid was reported to be successful (Scheme 24)13 This method of acetal removal is so
mild that the dioxolane and the dithiane remain intact while the aldehyde is produced in a
96 yield
92
S S
OOOMeMeO
S S
OOO
50 CF3COOHCHCl3 - H2O
0 oC 15 h96
Scheme 24 Acetal deprotection in an aqueous solution of trifluoroacetic acid
The use of metal triflates for deacetalisation has also been reported14 Erbium triflate was
found to catalyse the deprotection of benzylidene derivatives The reactions proceeded
smoothly at room temperature in the presence of 5 mol Er(OTf)3 in acetonitrile to afford
high yields of the desired products
To determine if Al(OTf)3 would be an effective Lewis acid for this process a variety of the
acetals that had being synthesised previously (Table 21 and Table 22) were used in a
reaction with 5 mol Al(OTf)3 in a mixture of THFH2O (82) at room temperature (Scheme
25) Vigorous stirring ensured that the reaction mixtures were homogeneous
O
R2R1
R2RO
R1RO
5 mol Al(OTf)3
THFH2O (82) 24 h Scheme 25 Deprotection of acetals catalysed by Al(OTf)3
The reactions were allowed to proceed for 24 hours after which the crude products were
passed through a plug of neutral alumina to remove the active catalyst The products were
isolated by removing the volatiles under vacuum (Table 24)
93
Table 24 Yield () deprotection of acetals catalysed by Al(OTf)3
Acetal Product Yield ()
H
MeO OMe
H3CO 21
H3CO
H
O
100
MeO OMe
O2N 25
O
O N2 55
H
EtO OEt
217
H
O
100
The results show that the acetals were successfully deprotected using Al(OTf)3 The 4-
nitroacetophenone dimethyl acetal (25) yield is lower which is probably due to the fact that
it is an insoluble solid in the THFH2O mixture
In order to determine if this phenomenon was unique to Al(OTf)3 the reactions were repeated
using the same conditions as above but with a variety of metal triflates amongst others also
making use of the THFH2O mixture
Table 25 Yield () deprotection of acetals catalysed by a variety of metal triflates
Acetal Product
Yield
()
In(OTf)3
Yield
()
Hf(OTf)4
Yield
()
Cu(OTf)2
Yield
()
Ca(OTf)2
H
MeO OMe
H3CO
21 H3CO
H
O
100 100 100 21
MeO OMe
O2N 25
O
O N2 89 100 18 15
H
EtO OEt
217
H
O
100 100 43 19
94
The harder metal triflates appear to catalyse the hydrolysis of the acetals to a greater degree
than the softer Ca(OTf)2 Again 4-nitroacetophenone dimethyl acetal is the most resilient
acetalmdashonly Hf(OTf)4 deprotected the acetal to afford 100 yield of the starting ketone
Since there has been a lot of interest in metal triflates acting as Lewis acids in aqueous media
it was decided to carry out the same reactions in neat water to determine if they had the same
reactivity A set of control experiments was run alongside the catalysed analogues in which
the acetal was placed in water with no catalyst and simply stirred for 24 hours (Table 26)
Table 26 Yield () deprotection of acetals in water
Acetal Product
Yield
()
Al(OTf)3
Yield
()
In(OTf)3
Yield
()
Hf(OTf)4
Yield
()
No Cat
H
MeO OMe
H3CO 21
H3CO
H
O
100 100 100 100
MeO OMe
O2N 25
O
O N2 100a 100a 100a 100a
OMeMeO
H
212
H
O16 22 26 0
H
EtO OEt
217
H
O
100 100 100 68
a reactions carried out at 80 ordmC to aid solubility
All of the aromatic acetals were completely deprotected under these conditions Conversely
the aliphatic acetal failed to undergo complete hydrolysis under any of the conditions used
What was surprising however was the results of the control experiment The aromatic
acetals underwent hydrolysis in each case in the absence of a catalyst The pH of the neat
deionised water was measured and was found to be 63 which is not out of the ordinary and
is near neutral All of the reactions were repeated and the same results were obtained in
duplicate runs The deprotection of the acetals under these conditions is most likely due to the
95
solvation of the acetal by the water leading to polarisation of the carbon oxygen bond
(Scheme 26) This polarisation allows ejection of the methoxide which either converts
rapidly to methanol or in any event remains highly solvated The carbenium intermediate
would also be stabilised by solvation Attack of water would eventually lead to the observed
carbonyl starting material
RR
O OMeR R
O Me
OH2
SS S
Where S = water (solvation)
+
S S
S S
OMeMe
OHH
HOH
R R
O
Scheme 26 Hydrolysis of acetal
In order to gauge the rate at which hydrolysis of the acetals takes place in water a study was
conducted in which the reactions were stopped after one hour (Table 27) Using the same
reaction conditions samples were taken from the reaction mixture and extracted with DCM
The solvent was then removed and the crude mixture was then analysed using 1H NMR
Table 27 Yield () deprotection of acetals in water after one hour
Acetal Product
Yield
()
Al(OTf)3
1 h
Yield
()
In(OTf)3
1 h
Yield
()
Hf(OTf)4
1 h
Yield
()
Sc(OTf)3
1 h
Yield
()
No Cat
1 h
H
MeO OMe
H3CO 21
H3CO
H
O
100 100 100 100 97
MeO OMe
O2N 25
O
O N2 100a 100a 100a 100a 100a
H
EtO OEt
217
H
O
100 100 100 100 75
96
All reactions were essentially complete after one hour The intention of the study was to take
samples periodically over a 24 hour period As can be seen from the results this was not
necessary The rate of hydrolysis appears to be rapid for the aromatic acetals even when there
is no catalyst present This ability to deprotect acyclic aromatic acetals under such mild
conditions may afford excellent opportunities in synthetic organic chemistry where chemists
are often faced with challenges of trying to remove one protecting group while retaining
another sensitive group In such deprotection reactions an acid is typically employed13 to
facilitate the reaction but the present study shows that this may not be necessary Work to
further investigate the potential of this method is underway in our laboratories
25 Other protecting groups
Silyl ethers
The stability of common protection groups has been investigated in the presence of metal
triflates One example is in the study of In(OTf)3 catalysed peracetylation of carbohydrates15
During that work it was found that the benzylidene acetal and tert-butyldiphenylsilyl
(TBDPS) ethers were stable to In(OTf)3 in acetic anhydride at 0 degC On the other hand the
authors found that trimethylsilyl (TMS) and tert-butyldimethlsilyl (TBDMS) groups were
unstable under these conditions and readily hydrolysed15
There has been much debate in the literature1617 as to the possible role that triflic acid plays
in metal triflate catalysed reactions (this aspect forms part of the present study and is
discussed in detail in Chapter 4) The central issue of this debate revolves around whether or
not the metal triflates hydrolyse with trace water found in organic solvents to form triflic
acid which in turn would then catalyse or aid in the catalysis of the reaction The use of
protection groups that are typically removed by Broslashnsted acids in the presence of metal
triflates will allow a determination of the stability of such groups to the metal triflate solution
and the extent of their hydrolysis In an indirect way this approach allows the acid sensitive
protection group to be used as a probe for Broslashnsted acidity
The TBDMS group is a commonly used silyl protection group It is easily cleaved under mild
acidic conditions such as acetic acid water THF (3 1 1)1 These conditions are so mild
that the TBDMS group is removed in an 87 yield while TBDPS remains intact By using
97
this acid sensitive group in the presence of metal triflates any Broslashnsted acidity generated by
the metal triflates will lead to hydrolysis of the TBDMS group
The primary alcohols 2-phenylethanol and para-bromophenol were TBDMS protected by
treatment with 15 equivalents of pyridine base and 20 equivalents of TBDMSCl The
reaction was allowed to stir overnight in DCM (Scheme 27) to afford high yields of products
224 and 225 respectively
OH OTBDMS
224 (78)
Br
OH
DCM
15 eq Pyridine
Br
OTBDMS
225 (80)
DCM
15 eq Pyridine
20 eq TBDMSCl
20 eq TBDMSCl
Scheme 27 TBDMS protection of 2-phenylethanol and para-bromophenol
The TBDMS protected substrates 224 and 225 were then both stirred for 5 hours in a
mixture of THFH2O (82) in the presence of 1 mol Al(OTf)3 at room temperature In the
control experiments these substrates were reacted under the same conditions in the presence
of 1 mol TfOH
After 5 hours the reactions were quenched with aqueous sodium bicarbonate and extracted
with DCM The volatiles were removed under vacuum The crude reaction mixtures were
then checked using 1H NMR spectroscopy Both reactions carried out using Al(OTf)3 showed
no signs of deprotection of the TBDMS group However reaction of substrate 224 with
TfOH showed 30 deprotection while substrate 225 showed no deprotection after being
reacted with TfOH for 5 hours
Since the reactions with Al(OTf)3 failed to show deprotection at 1 mol after 5 hours the
reactions were repeated using 5 mol Al(OTf)3 at room temperature for 10 hours As some
deprotection had been seen with TfOH these reactions were repeated at 1 mol TfOH for 10
hours
98
Surprisingly substrate 225 showed no deprotection in either the Al(OTf)3 reaction or the
TfOH reaction after the extended reaction period On the contrary substrate 224 was
completely hydrolysed to the original alcohol after the 10 hour reaction with triflic acid the
substrate in the reaction with Al(OTf)3 again showed no signs of hydrolysis
It was clear that substrate 225 was a more stable compound this is probably due to the
deactivating effect of the bromine on the aromatic ring making the TBDMS group more
stable and less susceptible to attack by the acid The reactions were repeated at 80ordmC using 5
mol Al(OTf)3 or 2 mol TfOH respectively and were monitored by TLC
After 6 hours TLC analysis showed no deprotection in either reaction After 10 hours the
reactions were analysed by TLC Again the reaction with TfOH showed some deprotection
and 1H NMR analysis showed it to be 30 deprotected Complete deprotection in the
reaction with TfOH was seen after 22 hours On the other hand the reaction with Al(OTf)3
showed no deprotection after this time
From these results it is clear that if Al(OTf)3 does at all hydrolyse in solution to produce
triflic acid it does so in such small amounts so as not to influence the outcome of the
reactions The results further indicate that Broslashnsted acidity generated whatsoever by the
presence of Al(OTf)3 (see chapter 4 for more details) is insufficient to catalyse this hydrolysis
reaction
26 Other metal triflates
It is clear from the literature that different metal triflates behave differently under the same
conditions Acetophenone can undergo allylation with diallyldibutyltin in the presence of 10
mol In(OTf)3 to afford 95 yield of the desired product The same reaction with 10 mol
Zn(OTf)2 produces only 48 yield of the product18
There are many reasons for their different behaviour ranging from ionic radius charge
density hydrolysis constants or pH of the solutions (many of these parameters are discussed
in later chapters of this thesis) In an attempt to gain further understanding of the current
system the study was extended to include other metal triflates
99
Using the same methodology as before substrate 224 was stirred in a mixture of THFH2O
(82) for 10 hours in the presence of 5 mol M(OTf)x while substrate 225 the more stable
of the two was subjected to harsher conditions namely 5 mol M(OTf)x in a mixture of
THFH2O at 80 degC for 22 hours (Table 28)
As was seen previously no deprotection occurred when 224 was subjected to the Lewis
acids over a 10 hour period A small amount of hydrolysis occurred when Sc(OTf)3 and
In(OTf)3 were allowed to react with 225 These results also point against the formation of
triflic acid by hydrolysis of the metal triflates in solution and that if indeed it occurs it is in
such small quantities that the TBDMS protection group would not be relatively unaffected by
it
Table 28 Yield () of alcohol for TBDMS deprotection
Substrate Lewis Acid OH
Yield ()a
OTBDMS
224
Sc(OTf)3 0 In(OTf)3 0 Ca(OTf)2 0 Cu(OTf)2 0
Substrate Lewis Acid Br
OH
Yield ()a
Br
OTBDMS
225
Sc(OTf)3 26 In(OTf)3 23 Ca(OTf)2 0 Cu(OTf)2 0
a Yields determined by 1H NMR spectroscopy
27 Tetrahydropyranyl ethers (THP)
Satisfied that the work on the TBDMS protection had been relatively fully investigated the
study was extended to another protection group namely the tetrahydropyranyl ether (THP)
This protection group was one of the first generally used protection groups employed for
alcohols19 and is still widely used today20 THP ethers are easily formed by acid catalysed
addition of dihydropyran (DHP) onto primary secondary and tertiary alcohols19 The reaction
is said to proceed via protonation of the enol ether carbon generating a highly electrophilic
oxonium ion which is then attacked by the alcohol (Scheme 28)1 One of the drawbacks of
100
using this protection group is the introduction of a stereogenic centre which leads to
diastereoisomers when chiral alcohols are used and may make NMR interpretation complex
ODHP
H+
O OROH
ORO
Scheme 28 Mechanism of acid catalysed THP ether formation
The cleavage of the THP group can be brought about under mildly acidic conditions such as
HOAc THF H2O (4 2 1) at 45 ordmC21 These ethers are also labile to Lewis acids such as
MgBr222 in ether and ferric chloride on silica23 THP ethers are stable to strongly basic
conditions23
Reports can be found in the literature of THP protection being carried out using metal triflate
catalysis Karimi and Maleki24 showed that LiOTf could be used under mild conditions to
form tetrahydropyranyl ethers in excellent yields The reactions were carried out in refluxing
DCM using 16-20 equivalents of DHP in the presence of 06-07 mol LiOTf
The use of Bi(OTf)3middot4H2O was also reported for THP protection25 However these reactions
were carried out under solventless conditions (for particular substrates) This catalyst was
also found to be active for the deprotection of the THP group in the solvent mixture DMF-
CH3OH
With this knowledge at hand it was decided that the THP ethers could be used in a similar
fashion to the TBDMS group That is once the THP protected alcohols had been synthesised
it would be possible to determine their stability with the different metal triflates
Selected alcohols were allowed to react with 2 equivalents of DHP in DCM for 3 hours in the
presence of 10 mol Al(OTf)3 at room temperature (Scheme 29) after which the reactions
were quenched using an aqueous sodium bicarbonate solution followed by extraction of the
mixtures with DCM The volatiles were removed under vacuum and the products purified by
column chromatography to afford high yields of the THP protected products (Table 29)
101
R OH 1 mol Al(OTf)3DCM rt 3 h
O+
RO
O
Scheme 29 THP protection catalysed by Al(OTf)3
Table 29 THP protected primary alcohols in reactions catalysed by Al(OTf)3
Entry Substrate Product Yield ()
1 OH
O O
226
98
2 Br
OH
Br
O O
227
60
3 OH
OO
228
gt98
A noteworthy point here is that the procedure is mild enough to tolerate another functional
group such as the triple bond (Table 28 entry 3) Because an aim of the investigation was to
look at the deprotection of the THP group with metal triflates no further optimisation of the
reactions was carried out Possibilities for optimising the reaction include repeating them
under solventless conditions shortening the reaction time and using less DHP to improve
overall efficiency
The THP ethers (Table 28 entry 2 and 3) were then reacted with a range of metal triflates in
order to determine if they would be stable under the variety of conditions (Table 210) The
ethers were allowed to stir in methanol (not anhydrous) for three hours in the presence of 1
mol of the metal triflate as catalyst Workup was carried out using an aqueous sodium
bicarbonate solution followed by extraction with DCM The volatiles were removed under
vacuum and the crude products were examined by 1H NMR spectroscopy
102
Table 210 Deprotection of THP ethers using different metal triflates
Entry Substrate Lewis Acid PhOH
Yield ()
1
PhO O
226
Al(OTf)3 68 Sc(OTf)3 100 In(OTf)3 100 Ca(OTf)2 0 Cu(OTf)2 0
Entry Substrate Lewis Acid Br
OH
Yield ()
2 Br
O O
227
Al(OTf)3 91 Sc(OTf)3 100 In(OTf)3 100 Ca(OTf)2 4 Cu(OTf)2 10
Entry Substrate Lewis Acid OH Yield ()
3 O
O
228
Al(OTf)3 0 Sc(OTf)3 0 In(OTf)3 0 Ca(OTf)2 0 Cu(OTf)2 0
Substrates 226 and 227 (Table 210 entry 1 and 2) were successfully deprotected by triflates
of Al Sc and In triflates to afford high yields of the alcohols Reactions with Ca and Cu
triflates resulted in very little of the free alcohol if any affording only the starting material
A possible mechanism for the hydrolysis is shown in Scheme 210 for the substrate 227
Br
O OLA
Br
OLA
Br
+O MeOH
OMeO+ H+
OLA
Br
OH
+ LA
227
H+ Scheme 210 Mechanism for the hydrolysis of 227
All of the metal triflates failed to successfully deprotect 228 (Table 210 entry 3) without
by-product formation In the reactions that had been carried out using Ca(OTf)2 and
Cu(OTf)2 only starting material was obtained The reactions that had been carried out with
103
the other metal triflates yielded a complex mixture of by-products apart from some starting
material It is possible that the alkyne functional group led to secondary reactions under these
conditions These reactions indicate that the THP protecting group could be readily installed
making use of the metal triflate catalysis The deprotections were less satisfactory but were
successful in several instances indicating a measure of catalyst selection should be applied as
and when the need arises
28 Two protection groups
Often during a synthesis an organic chemist will be required to use two or more protection
groups on a substrate at the same time When incorporating the second protecting group into
the molecule the first obviously needs to stay in place Alternatively when removing one the
other has to remain It was evident from the present study that the TBDMS group was stable
in the presence of Al(OTf)3 as well as a range of other triflates and from literature it is known
that acetals can be formed by other metal triflates45 What remained to be determined was if
an acetal could be formed on a substrate already containing a TBDMS group using the
different metal triflates as catalyst
Vanillin and salicyl aldehyde were successfully TBDMS protected by reacting them with 15
equivalents of pyridine in DCM and 2 equivalents of TBDMSCl overnight The reactions
were quenched with a brine solution containing CuCl2 (to assist in the removal of the
pyridine) and extracted with DCM The products were purified by column chromatography
to afford high yields of their respective TBDMS protected products 229 and 230 (Table
211 entry 1 and 2)
Table 211 Yield () TBDMS protected substrates
Entry Substrate Product Yield ()
1 HO
OO
TBDMSO
OO 229
90
2 OH
H
O
OTBDMS
H
O 230
80
104
Using conditions from the previous work that had been carried out on acetal formation with
Al(OTf)3 substrates 229 and 230 were reacted with 2 equivalents of trimethyl orthoester
and 10 equivalents of alcohol26 In order to optimise yields with respect to the Lewis acid the
reactions were allowed to proceed for 12 hours in the presence of 10 mol M(OTf)x after
which the mixture was passed through a small column of neutral alumina The excess solvent
was removed under vacuum and the product was analysed by 1H NMR spectroscopy (Table
212)
Acetal formation was successful only with Al(OTf)3 when substrate 229 (Table 212 entry
1) was used resulting in an 86 yield of the desired acetal Interestingly Sc(OTf)3 and
In(OTf)3 hydrolysed the TBDMS protecting group from substrate 229 to afford 100 yield
of the original aldehyde Substrate 229 (Table 212 entry 1) contains an electron donating
methoxy group ortho to the TBDMS group This group increases the electron density on the
system facilitating the loss of the TBDMS This is unlike the stable aliphatic substrate 226
and the Br-stabilised aromatic molecule 227 (Table 210 entries 1 and 2)
Table 212 Yield () of acetal formation with a variety of metal triflatesa
Entry Substrate Lewis Acid
TBDMSO
OMeO
OMe
H
231 Yield ()
O
TBDMSO
H
O Yield ()
HO
OO
Yield ()
1 O
TBDMSO
H
O 229
Al(OTf)3 86 10 4 Sc(OTf)3 0 0 100 In(OTf)3 0 0 100 Ca(OTf)2 0 80 20 Cu(OTf)2 0 92 8
Entry Substrate Lewis Acid
OTBDMS
OMe
OMe 232
Yield ()
OTBDMS
H
O Yield ()
OH
H
O Yield ()
2
OTBDMS
H
O 230
Al(OTf)3 51 49 0 Sc(OTf)3 65 18 17 In(OTf)3 63 14 23 Ca(OTf)2 85 15 0 Cu(OTf)2 89 11 0
a10 mol M(OTf)x 2 eq orthoester 10 eq MeOH 12 h
In contrast to these results only small amounts of deprotection were seen with Ca(OTf)2 and
Cu(OTf)2 while acetal formation did not take place
105
Very little hydrolysis of the TBDMS group of substrate 230 (Table 211 entry 2) was
observed in all cases This substrate has no electron donating group and the TBDMS group is
consequently stabilised by the electron-withdrawing aldehyde
The acetal formation reactions with substrate 230 were successful (Table 211) in virtually
all cases All of the metal triflates used catalysed the acetal formation reaction Again a small
amount of deprotection was seen with Sc and In triflates but no deprotection was noted with
Ca and Cu triflates suggesting that the acetal formation reaction occurs faster than the
deprotection reaction In this case as opposed to that of substrate 229 acetal formation is the
faster reaction
To determine if the acetal formation using substrate 230 could be improved with the triflates
of Al Sc and In the reactions were carried out again using solventless conditions a method
that has proved efficient previously in this study Substrate 230 was allowed to react with 2
equivalents of trimethylorthoester and 10 mol M(OTf)x for 12 hours after which the
mixture was passed through a small column of neutral alumina The excess solvent was
removed under vacuum and the product was analysed by 1H NMR spectroscopy (Table 213)
Table 213 Yield () Acetal formation in the absence of added alcohol
Substrate Lewis Acid
OTBDMS
OMe
OMe Yield ()
OTBDMS
H
O Yield ()
OH
H
O Yield ()
OTBDMS
H
O 230
Al(OTf)3 82 18 0
Sc(OTf)3 91 9 0
In(OTf)3 94 6 0
Cu(OTf)2 gt98 0 0
Ca(OTf)2 gt98 0 0
The yields of acetal product are much improved in all three cases What is also interesting to
note is that in the cases of Sc(OTf)3 and In(OTf)3 no deprotection of the TBDMS protecting
group occurred suggesting that under these conditions acetal formation takes place faster
than the hydrolysis of the TBDMS group
106
29 Conclusions
Al(OTf)3 offers a mild greener alternative for the formation of acetals from various
aldehydes and ketones The reaction may be performed in the presence of the
alcoholorthoester mixture or using neat orthoester The latter solvent-free procedure is
preferred Not only does this method improve the yields in most cases but it offers a
procedure that produces less waste Simple distillation of the product allowed for product
isolation in excellent yields The catalyst was recycled using simple techniques and could be
reused several times without loss of activity
TBDMS ethers were used as a probe in order to test for the hydrolysis of metal triflates in
protic or wet solvents and the concomitant formation of triflic acid This protection group is
known to be sensitive under acidic conditions Very little hydrolysis of the TBDMS group is
seen with metal triflates while full hydrolysis is seen with triflic acid suggesting that if the
metal triflates do hydrolyse (methanolysis) to form triflic acid in solution it is so little that it
does not affect the TBDMS protecting group This is advantageous from the point of view of
protection group chemistry where a chemist often needs a protection group to remain in place
while using a Lewis acid on another part of the molecule
Al(OTf)3 was found to be a good catalyst for the formation of THP ethers Moreover some of
the metal triflates were found to deprotect the aromatic THP ethers under mild conditions to
afford the alcohols in excellent yields
Finally the metal triflates were used to form acetals in the presence of the TBDMS group
when the reaction was carried out under solventless conditions the yields were much
improved and no hydrolysis of the TBDMS group was seen
These results as a whole are an important contribution to protection group chemistry the
metal triflates offers a mild alternative to method that have previously been used
Furthermore these mild alternatives can be used in the presence of other protection groups
This initial study into the hydrolysis of the metal triflates formed the basis for subsequent
chapters in this thesis where an in-depth investigation has been carried out into the hydrolysis
of the metal triflates and the role of water in metal triflate catalysed reactions
107
210 References
1 Kocieński P J Protecting groups Thieme New York 1994
2 Torok D S Fiueroa J J Scott W J J Org Chem 1993 58 7274
3 Kumar R Chakraborti A K Tetrahedron Lett 2005 46 8319
4 Leonard N M Oswald M C Frieberg D A Nattier B A Smith R C Mohan
R S J Org Chem 2002 67 5202
5 Smith B M Graham A E Tetrahedron Lett 2006 47 9317
6 Borujeni K P Massah A R React Funct Polym 2006 66 1126
7 Williams D B G Lawton M Org Biomol Chem 2005 3 3269
8 Williams D B G Lawton M Tetrahedron Lett 2006 47 6557
9 Williams D B G Shaw M L Green M J Holzapfel C W Angew Chem Int
Ed 2008 47 560
10 Solomons T W G Fryhle C B Organic Chemistry 7th ed Wiley New York
2000
11 Anastas P T Williamson T C Ed Green Chemistry Frontiers in Benign Chemical
Synthesis and Processes Oxford University Press New York 1998
12 Sandstroumlm M Persson I Persson P Acta Chem Scand 1990 44 653
13 Ellison R A Lukenbach E R Chiu C W Tetrahedron Lett 1975 499
14 Procopio A Dalpozzo R De Nino A Maiuolo L Nardi M Romeo G Org
Biomol Chem 2005 3 4129
15 Bizier N P Atkins S R Hellend Colvin S F Twitchell J R Cloninger M J
Carb Res 2008 343 1814
16 Ollevier T Lavie-Compin G Tetrahedron Lett 2004 45 49
108
109
17 Barret A G M Braddock D C Henschke J P Walker E D J Chem Soc
Perkin Trans 1 1999 873
18 Liu L Tang L Yu L Chang W Li J Tetrahedron 2005 61 10930
19 Parham W E Anderson E L J Am Chem Soc 1948 70 4187
20 Greene T W Wuts P G Protective Groups in Organic Synthesis 2nd ed Wiley
New York 1991 31
21 Bernardy K F Floyd M B Poletto J Weiss M J J Org Chem 1979 44 1438
22 Kim S Ho Park J Tetrahedron Lett 1987 28 439
23 Fadel A Salaun J Tetrahedron 1985 41 1267
24 Karimi B Maleki J Tetrahedron Lett 2002 43 5353
25 Stephens R J Butler P L Clow C H Oswald M C Smith R C Mohan R
S Eur J Org Chem 2003 3827
26 Williams D B G Lawton M C Green Chem 2008 10 914
Chapter 3
The drying of organic solvents
31 Introduction
Metal triflates (trifluoromethanesulfonates) in aqueous media are thought to exhibit
not only Lewis acid activity but also Broslashnsted acid activity (see chapter 4) generated
by the direct interaction of the water and the metal (Scheme 31)
M+n + H2O M OHH
+n
Lewis acid Broslashnsted acid Scheme 31 Interaction of metal and a water molecule to form a Broslashnsted acid
In order to determine whether metal triflates with a high propensity for showing
Broslashnsted acid activity in the presence of water are independently capable of Lewis
acid activity it is important to study their activity in completely dry solvents (see
chapter 4 for more details) Therefore the present study embarks on the evaluation of
the effectiveness of methods for drying a number of organic solvents
Chemists are often faced with the challenge of having to carry out reactions under dry
conditions These reactions involve solvents that require rigorous drying regimes The
literature on how to dry laboratory solvents can be contradictory1abcd For example
magnesium sulfate has been described as neutral1ab or acidic1cd as a good drying
reagent that is rapid in its action1ab or a slow drying reagent1d Aluminium oxide is
recommended mainly for the use in dessiccators1d while another group of workers
recommend it as the ultimate drying reagent for organic solvents2 The literature still
suggests procedures that are outdated such as the practice of drying solvents over
sodium which according to Plesch2 is carried out by organic chemists as more of a
ritual than an effective process Many recommendations fail to mention the
concentration of water that remains after the drying process is complete
Burfield3abcd wrote a comprehensive comparative series of papers on the drying of
solvents with different desiccants By using tritiated water he was able to determine
110
the trace amounts of water remaining in the sample after the drying process using
scintillation counting To our knowledge this is the only study of this kind that has
been done on laboratory solvents
Interest in dry solvents in the present instance stemmed from the work that was
carried out as described in chapter 4 of this thesis namely on the role of water in
Lewis acid catalysis Reactions that were carried out during the course of this work
required solvents that contained as little as 5 ppm water Although the literature
explains how to dry solvents the amount of remaining water content is generally not
stated Burfieldrsquos papers although comprehensive did not cover all of the solvents of
interest for the current work so it was decided to investigate the drying of organic
solvents more fully
The aim of this current investigation was to determine which of the desiccants was
capable of drying the desired solvents to a water content that was preferably below 5
ppm in order to use this knowledge in further investigations carried out in this thesis
32 Choice of analytical method ndash Karl Fischer
Several methods have been developed for the determination of water in organic
solvents They include the gravimetric method45 near infrared6 and the radio tracer
method3a None of these methods was completely satisfactory for the purposes of the
current investigation The gravimetric method employed by Trussel and Diehel5 and
by earlier workers4 for gas drying is limited as the results cannot be extrapolated to
the drying of solvents in the liquid phase The use of near infrared for the
determination of water content in organic solvents is a useful and rapid technique6
the detection limits of this method are reported to be 10 ppm63a Ideally for the
purpose of this study a method with lower detection limits was needed Also
measurements needed to be taken in a dry atmosphere ie inside a glove box which
discounted near infrared spectroscopy for the present investigation
Burfield3a developed a method of water determination using tritium labelled water
The method works by the addition of a specified amount of the labelled water to a
rigorously dried solvent and subsequent determination of the decrease in activity of
111
the solvent after treatment with various drying agents This method is extremely
accurate and the detection limits are as low as 01 ppm However the process is
laborious it requires two drying steps and also requires the use of scintillation
cocktails which further encumbers the process This method was found to be
inappropriate for the current investigation
Karl Fischer titration is another method for the determination of water in organic
solvents This process was first published by Fischer7 in 1935 Since then the method
has been developed and improved8 This method works on a simple standard reaction
(Scheme 32) and depending on the amount of sample used can detect lower than 2
ppm water The iodine acts as an oxidant for the alkylsulfite which becomes an
alkylsulfate In the process water is consumed and iodide is generated
ROH + SO2 + RN
2I- I2 + 2e-
[alcohol] [base](RNH)SO3R
[alkylsulfite salt]
(RNH)SO3R + H2O + I2 + 2RN 2(RNH)I + (RNH)SO4R
[alkylsulfite salt] [water] [iodine] [hyroiodic salt] [alkylsulfate salt]
Scheme 32 Standard reaction of Karl Fischer titration
What made the method more appealing for the current study was that the Karl Fischer
unit was able to be placed inside the glove box preventing the ingress of atmospheric
water into the system Atmospheric humidity is the biggest single cause of error in
Karl Fischer titration8 Water can enter the sample the titrant and the cell in this
manner This manoeuvre placing the Karl Fischer titrator inside the glove box led to
reliable data being generated for trace water determination as will become clear To
the best knowledge of the author the results presented represent the most sensitive
water detection experiments yet performed by Karl Fischer titration outside of those
conducted by the manufacture in setting up the specification sheet (See section 33)
112
Fundamentals of Karl Fischer Coulometry
As mentioned above Karl Fischer titration is based on a standard reaction (Scheme
32) Iodine generation takes place at the generator electrode which is incorporated in
the glass titration cell next to the measuring electrode (Figure 31)
Figure 31 Karl Fischer Measuring cell and electrodes
The glass titration cell consists of two parts namely the anode compartment and the
cathode compartment which are separated by a diaphragm in the case of the present
instrument set up The anode compartment holds the anolyte which contains the sulfur
dioxide imidazole and the iodide and methanol or ethanol can be used as solvent
The cathode compartment contains the catholyte which is a reagent containing an
ammonium salt
At the anode iodine is generated from iodide (Figure 32) The negative iodide ions
release electrons at the anode and form iodine This in turn will react (indirectly via
the sulfite sulphate oxidation) with the water present in the sample injected into the
titration cell At the cathode the positive hydrogen ions are reduced to hydrogen gas
being the main product that forms
From a coulometric point of view the two iodide ions (I-) each carry the charge of one
electron implying 96485 C per mole At the anode the two iodide ions are converted
into elemental iodine which reacts indirectly with one molecule of water (According
to the concepts of Scheme 32) From an electronic point of view 2 x 96485 C (twice
113
the charge on 1 mole of electrons) are needed for one mole of water or 1072 C for 1
mg of water
Figure 32 Karl Fischer anode and cathode
Therefore it is possible to measure the amount of iodine that has been released and
thus the amount of water that has reacted with the iodine by measuring the current
(amperes) over a period of time (seconds) This method assumes that all of the current
produced has been used for iodine production Coloumetry is an absolute method of
detection and can be used as a reference method for the determination of water
content8
33 Limits of the Karl Fischer method
In order to ensure maximum stability the Karl Fischer unit was placed in a glove box
In this way the atmosphere surrounding the equipment was controlled to contain 1
ppm or less of water throughout the experiments
The manufacturerrsquos (Mettler Toledo) documents8 for this instrument indicate a
detection limit of less than 5 ppm with good reproducibility (accuracy and precision)
for toluene and n-hexane Results were reported as low as 46 ppm plusmn 04 ppm and 15
ppm plusmn 02 ppm respectively
To determine the limits of the Karl Fischer apparatus 4 mL of a 100 ppm hydranal
water standard were used Since this standard had a water concentration an order of
114
magnitude higher than that required for the present purpose the addition of 10 ppm of
water was simulated as follows The instrument requires the entry of the mass of
sample added from which it performs its calculations to determine water content
Accordingly if the sample of say 0400 g was added and that mass was entered on
the instrument a reading of 100 ppm would be obtained If however a mass of 4000
g was entered for the same 0400 g sample a reading of about 10 ppm would be
obtained This approach of multiplying the masses by 10 was employed here (Table
31) It may be argued that a simple dilution of the analytical standard would have
sufficed However such an approach would bring about uncertainty as to the amount
of water added during the dilution process as a consequence of such water being
present in the solvent used for dilution and was therefore not applied
When 4 mL of the standard were used the readings obtained for the two samples were
9891 and 9765 ppm which values are well within the accepted range8
Table 31 Result of Karl Fischer titration of hydranal 100 ppm water standard
Reading
Hydranal
Std Mass
(g)
Water
content
(ppm)
1 0401 72
2 0391 80
3 0386 75
4 0390 74
5 0391 75
6 0411 77
7 0381 76
8 0375 80
9 0370 70
10 0391 75
Average 75 ppm
Std dev 03 ppm
RSD 39
115
Although the values obtained in this experiment are lower than expected the results
are not entirely unsatisfactory The precision of the instrument is good as the standard
deviation is only 03 ppm The fact that the instrument was slightly under reading the
water content of the hydranal standard probably stems from the relatively small
amount of sample that was injected into the cell The accuracy of the Karl Fischer
increases as the amount of sample injected increases this is particularly evident at
very low concentrations of water9
34 Choice of solvents
The next step of the investigation involved the choice the solvents for drying The
following chapters of this thesis deal with the role of water in Lewis acid catalysis In
that work aprotic solvents have been used so as to reduce the possibility of the
formation of Broslashnsted acid type species during the course of the reactions by
hydrolysis of the Lewis acid or by other means
A review of the literature also shows that other than Burfieldrsquos studies3abcd which
cover solvents such as benzene dioxane acetonitrile and the alcohols no
comprehensive comparative study has been carried out on the effects of the different
drying reagents on several other every day laboratory solvents With this in mind
toluene THF and DCM were chosen for this study These three solvents are
commonly used in research laboratories as well as forming part of work still to be
described in this thesis For comparative purposes with Burfieldrsquos work methanol
ethanol and acetonitrile were also used in the current study
35 Methods of drying the solvent
In order to set the scene the water content of each of the solvents selected namely
DCM toluene THF acetonitrile methanol and ethanol was determined for the lsquowetrsquo
solvent The same bottle of lsquoanalytical reagentrsquo solvent was then used throughout the
study to avoid variance in initial water content The results of the lsquowetrsquo solvent
measurement are summarised in Table 32 Throughout the study 3 mL of the sample
were injected directly into the anolyte of the Karl Fischer (For ease of reading all
116
tables of individual measurements can be found in appendix A of this thesis and only
worked data are presented here)
Table 32 Results of Karl Fischer titration of lsquowet solventsrsquo
Solvents
Average sample weight
(g)
n
Residual water
contenta (ppm)
Std dev ppm RSD ()
DCM 4103 6 224 120 536 THF 2815 6 1078 066 062
Toluene 2765 6 2249 128 058 Acetonitrile 2553 6 1420 118 083 Methanol 2590 6 1751 044 025 Ethanol 2622 6 14283 380 027
a Outliers have been removed confidence level 95
The above results show the precision of the Karl Fischer especially at higher
concentrations of water where the relative standard deviations are below 1 The
results for DCM show a higher relative standard deviation which is still well within
acceptable limits
351 Traditional drying methods
Each of the solvents used in the current study can be traditionally dried by a particular
method in chemical laboratories Such methods can be found in most books on drying
of organic solvents9 Tetrahydrofuran is typically pre-dried over KOH pellets after
which it is dried by heating under reflux over sodium wire9 Benzophenone is used as
an indicator the characteristic blue colour indicating dryness9
Dichloromethane is perhaps the easiest of the three solvents to dry as it is the least
hygroscopic and heating over CaH2 appears to be the method of choice for this
solvent9 There is no indicator for DCM Toluene is most commonly dried by heating
over sodium9 and it can be pre-dried using CaCl2 CaH2 or CaSO49 Again
benzophenone is used as an indicator
Methanol and ethanol are dried using magnesium turnings and iodide in the ratio 5 g
05 g respectively per one litre of the alcohol9 The alcohol is heating with the Mg and
117
I2 until the Mg has dissolved Acetonitrile can simply be dried by using molecular
sieves9
Samples of these solvents dried in these various ways were subjected to Karl Fischer
titration to determine their water content (In our laboratories THF and toluene are
passed through a column of dried silica before being heated under reflux over sodium
and benzophenone) All samples were collected under nitrogen or in Schlenk tubes
kept under nitrogen and placed directly into the glove box The Karl Fischer readings
were carried out in a glove box so as to minimise any contamination from
atmospheric water (Table 33 and 34) within two hours of having been collected (see
section 38 for discussion on the potential role of the dried glass vessels in drying the
solvents stored therein)
Table 33 Results of Karl Fischer titration of solvents dried by traditional methods
Solvent
Desiccant Average sample weight
(g)
n
Residual water
contenta (ppm)
Std dev ppm
RSD ()
DCM CaH2 4154 6 129 127 988 Toluene SilicaSodium 2781 6 314 193 613
Acetonitrile 3Aring molecular sieves 2628 6 40 065 1606
Methanol MgI2 2603 6 536 058 108 a Outliers have been removed confidence level 95
Table 34 Results of Karl Fischer titration of THF dried by SilicaSodium
Experiment number
Average sample weight
(g)
n
Residual water
contenta (ppm)
Std dev (ppm)
RSD ()
1 2836 6 1140 429 376
2 2650 6 440 066 150
3 2685 6 428 073 171 a Outliers have been removed confidence level 95
Initially THF dried over silicasodium gave the anomalous result of 114 ppm water
remaining in the solvent even though the characteristic blue colour of the ketyl
radical had formed According to this result no water had been removed by this drying
process In order to determine the accuracy of these results the drying experiments
118
were repeated (each repetition represents a separate experiment carried out
independently from each other This is the case for all experiments that were carried
out in triplicate during this work) As can be seen from Table 34 the initial result of
114 ppm of residual water that was obtained appears to be an outlier when compared
to the results of the repeat experiments where only approximately 43 ppm water
remained in the THF in both experiments These experiments show that the majority
of water can be removed using this method of drying and that the method is
repeatable
Acetonitrile was effectively dried over activated (at 250 degC for 24 hours) 3Aring
molecular sieves at a loading of 5 (wv) After 24 hours roughly 4 ppm water
remained in the solvent The drying of toluene was also found to be quite efficient
using the dual method of passing it over activated silica (dried in an oven at 250 degC
for 24 hours) and heating over sodium In this instance there was a decrease from
around 220 ppm water to approximately 30 ppm
352 Drying reagents
The next part of the study was to determine the efficacy of different drying reagents
for the solvents In all cases the drying of the solvents as well as the measurements
were carried out in a glove box Residual water content was determined by Karl
Fischer titration As before 3 mL of the solvent were injected directly into the anolyte
of the Karl Fischer cell
The following discussion relating to the investigation is broken up according to the
different solvents used Not all desiccants were used for each solvent An overview is
given at the end of the chapter
353 Preparation of desiccants
bull 3Aring molecular sieves Before the sieves were used they were washed
thoroughly with AR grade methanol dried on a rotary evaporator and then
placed in an oven at 250 degC for 24 hours
119
bull Silica and alumina All silica and alumina samples used in the study were
placed in glass beakers and covered with aluminium foil into which small
holes were punctured The beakers were then placed in an oven overnight at
300 degC to ensure that the silica and alumina samples were dry
36 The alcohols
The lower alcohols are more difficult to dry than their longer carbon chain
counterparts3d These solvents are hygroscopic and hygrophilic and pose a problem
for most bench chemists trying to work under dry conditions3d Many desiccants have
been suggested for the drying of these alcohols910 including KOH BaO CaO and
molecular sieves to name just a few
In this study ethanol and methanol were dried over activated 3Aring molecular sieves
Each of the alcohols (200 mL) was dried over 5 10 and 20 (weightvolume) of
the molecular sieves The water content of these samples was determined by Karl
Fischer titration after 24 48 and 72 hours and 5 days The experiments were repeated
three times to determine the accuracy of the methods The same procedure was carried
out using KOH pellets The results of these experiments are summarised in Tables 35
-310
Table 35 Results of Karl Fischer titrations of methanol dried over 3Aring molecular
sieves (5 wv)
Time (h) n
Average sample
weighta (g)
Residual water
content (ppm)
Std dev
(ppm)
RSD
()
24 6 2672 753 074 098 24 6 2621 789 067 085 24 6 2557 777 062 080 48 6 2488 452 055 122 48 6 2571 489 047 096 48 6 2612 464 082 177 72 6 2612 329 041 125 72 6 2488 343 040 117 72 6 2612 354 049 138
5 days 6 2523 257 124 482 5 days 6 2584 266 120 451
120
5 days 6 2495 273 092 337 a Outliers have been removed confidence level 95
Table 36 Results of Karl Fischer titrations of methanol dried over 3Aring molecular
sieves (10 wv)
Time (h)
n
Average sample weighta
(g)
Residual water
content (ppm)
Std dev (ppm)
RSD
()
24 6 2688 407 093 229 24 6 2619 392 059 151 24 6 2624 419 041 098 48 6 2561 268 071 265 48 6 2508 293 071 242 48 6 2502 322 087 270 72 6 2612 224 042 188 72 6 2592 236 084 356 72 6 2505 263 063 240
5 days 6 2589 166 057 343 5 days 6 2573 182 058 319 5 days 6 2606 199 168 844
a Outliers have been removed confidence level 95
Table 37 Results of Karl Fischer titrations of methanol dried over 3Aring molecular
sieves (20 wv)
Time (h) n
Average sample weighta
(g)
Residual water
content (ppm)
Std dev
(ppm) RSD ()
24 6 2509 280 057 204 24 6 2609 292 042 144 24 6 2634 290 031 107 48 6 2496 211 046 218 48 6 2605 239 038 159 48 6 2556 244 087 357 72 6 2591 176 051 290 72 6 2541 188 055 293 72 6 2610 212 088 415
5 days 6 2575 97 099 1021 5 days 6 2542 104 049 471
121
5 days 6 2603 115 110 957 a Outliers have been removed confidence level 95
Figure 33 Graph indicating decreasing water content in methanol over time
Table 38 Results of Karl Fischer titrations of ethanol dried over 3Aring molecular sieves
(5 wv)
Time (h) n
Average sample weighta
(g)
Residual water
content (ppm)
Std dev
(ppm) RSD
()
24 6 2699 2511 181 072 24 6 2637 2703 166 061 24 6 2662 2664 250 094 48 6 2648 1021 075 073 48 6 2639 1064 074 070 48 6 2633 1110 055 050 72 6 2664 549 093 169 72 6 2639 541 059 109 72 6 2639 563 123 218
5 days 6 2635 144 051 354 5 days 6 2631 141 049 348 5 days 6 2659 149 015 101
a Outliers have been removed confidence level 95
122
Table 39 Results of Karl Fischer titrations of ethanol dried over 3Aring molecular sieves
(10 wv)
Time (h) n
Average sample weighta
(g)
Residual water
content (ppm)
Std dev ppm
RSD ()
24 6 2668 1894 153 081 24 6 2666 1792 081 045 24 6 2638 1896 050 026 48 6 2645 678 051 075 48 6 2637 695 064 092 48 6 2640 712 058 081 72 6 2642 351 107 305 72 6 2624 359 073 203 72 6 2616 397 118 297
5 days 6 2611 114 113 991 5 days 6 2658 130 076 585 5 days 6 2619 124 089 718
a Outliers have been removed confidence level 95
Table 310 Results of Karl Fischer titrations of ethanol dried over 3Aring molecular sieves
(20 wv)
Time (h)
n
Average sample weighta
(g)
Residual water
content (ppm)
Std dev ppm
RSD ()
24 6 2639 1186 058 049 24 6 2641 1190 109 092 24 6 2630 1218 037 030 48 6 2643 231 076 329 48 6 2639 243 055 226 48 6 2666 277 079 285 72 6 2604 190 028 147 72 6 2612 194 056 289 72 6 2608 231 061 264
5 days 6 2651 69 035 507 5 days 6 2612 81 035 432
123
5 days 6 2599 95 050 526
The results show that activated molecular sieves are effective at removing water from
both methanol and ethanol However this method requires at least 72 hours and a 10
(wv) loading of molecular sieves to be effectual The trend in the precision of the
Karl Fischer is again evident in these experiments namely the higher the water
concentration the lower the standard deviation At lower water concentrations the
standard deviation increases despite which the standard deviations are mostly
acceptable
Figure 34 Graph indicating decreasing water content in ethanol over time
The graphs (Figures 33 and 34) show that the effectiveness of the molecular sieves
converges after some time and does not necessarily ultimately improve with loading
After 72 hours methanol had reached a similar level of dryness with 10 and 20
(wv) molecular sieves Leaving the methanol a further 2 days to dry showed some
improvement in the dryness of the samples in all cases The molecular sieves appear
to dry the ethanol more effectively than the methanol initially there is a rapid drop in
the water content and after 5 days the ethanol had reached a similar level of dryness in
all of the experiments
Burfield3d conducted a similar experiment in which he dried methanol and ethanol
over 3Aring molecular sieves 5 (wv) He does not indicate in his paper if the sieves
124
were activated by heat The initial water content of the methanol was found to be
1010 ppm and that of the ethanol 1500 ppm The alcohols stood for a period of 24
hours over the molecular sieves after which the tritiated water method3a showed the
residual water content to be 95 ppm and 99 ppm for methanol and ethanol
respectively The present study shows similar amounts of residual water for methanol
at the same loading of the molecular sieves while that for ethanol was found to be
higher (at 24 hours) but ultimately reaching much lower levels of residual water
These reductions in water content are remarkable Methanol showed a 90 reduction
in water content and ethanol a 93 reduction in water content after a contact period
of 72 hours In the current investigation the 5 (wv) experiment results indicated a
60 reduction in water content for methanol and an 82 reduction for ethanol after
24 hours The current study shows that the MgI2 method was more effective at
rapidly reducing the water content of the methanol the method requiring only two to
three hours for the magnesium to dissolve under refluxing conditions and a reduction
of 85 in the water content of the methanol is seen However improved ultimate
dryness is obtainable when using molecular sieves
Table 311 Results of Karl Fischer titrations of methanol and ethanol dried over KOH
powder
Solvent Desiccant Time (h)
Average sample
weight (g) n
Residual water
contenta (ppm)
Std dev ppm
RSD ()
Methanol KOH (10 wv) 24 2525 6 318 081 255
Methanol KOH (10 wv) 24 2555 6 339 099 292
Methanol KOH (10 wv) 24 2606 6 337 093 276
Ethanol KOH (10 wv) 24 2856 6 259 074 286
Ethanol KOH (10 wv) 24 2760 6 284 053 187
Ethanol KOH (10 wv) 24 2640 6 250 056 225
a Outliers have been removed confidence level 95
125
The initial results (not shown here) of the drying of the alcohols with KOH a well
known and recommended procedure were disappointing since it appeared to increase
the water content of these solvents However it transpired that this result was due to
the fact that the KOH was used in pellet form as supplied When these pellets first
powdered to a fine powder and then added to the alcohol which was left to stand for
24 hours before being distilled the expected results were obtained (Table 311)
37 Acetonitrile
Acetonitrile is a polar aprotic solvent with high solvating ability It has a high affinity
for water and can be difficult to dry11 Burfield11 found that P2O5 was particularly
efficient at removing water from acetonitrile A desiccant loading of 5 (wv) with 24
hours of static drying leaves a residual water content of 9 ppm The initial water
content was 2800 ppm which equates to a 99 reduction in water content A 5
(wv) loading of 3Aring molecular sieves resulted in a 98 reduction in water from the
acetonitrile after 24 hours
In the current investigation the acetonitrile was dried over activated 3Aring molecular
sieves using the same conditions as those used for the methanol and ethanol
experiments The acetonitrile was also dried in a separate experiment by passing it
through a column of activated neutral alumina The results of the experiments are
summarised in Table 312
126
Table 312 Results of Karl Fischer titrations of acetonitrile dried over 3Aring molecular
sieves and neutral alumina
Desiccant Time (h)
Average sample weight
(g)
n
Residual water
contenta (ppm)
Std dev ppm
RSD ()
3Aring molecular sieves 5
(wv)
24 2628 6 40 065 1606 48 2626 6 18 050 2841 72 - 6 ltdlb - -
3Aring molecular sieves 10
(wv)
24 2643 6 05 041 8367
48 - 6 ltdlb - -
Activated neutral alumina
10 (wv) -c 2966 6 61 062 1015
Activated neutral alumina
10 (wv) -c 2924 6 49 015 306
Activated neutral alumina
10 (wv) -c 2826 6 68 037 537
a Outliers have been removed confidence level 95 b ltdl below detection limits of Karl Fischer titrator C The
solvent was passed over a column of the desiccant
The results indicate that molecular sieves are indeed an effective method for the
drying of acetonitrile After just 24 hours at 10 (wv) loading the acetonitrile is
essentially lsquosuper dryrsquo (lt5 ppm water) In order to test the precision and accuracy of
the Karl Fischer titrations at these low levels of water the titrations were repeated
using the 24 hour 10 (wv) sample In this experiment however 10 mL of the
sample were injected into the cell instead of the customary 3 mL (Table312) The
results showed an average water content of 173 ppm with a standard deviation of
006 ppm This result gives us a slightly better indication of the water content of the
acetonitrile once it has been dried However using 10 mL of sample to obtain
analytical data with six repeats was not practical for our use of the Karl Fischer
apparatus as the titration cell quickly fills with the sample and needs to be emptied
and replenished with hydranal The large sample results in inefficient use of the
hydranal solution which may otherwise be used for a large number of further
titrations
127
The drying of acetonitrile by passage over activated neutral alumina (Table 312) also
showed excellent results in which a 96 reduction in water content is achieved This
method is simple and although not quite as effective as molecular sieves it is
advantageous from the point of view that it is far more rapid These experiments were
repeated three times to determine the accuracy of this method the results show that
the method is reproducible Passage through alumina and storage over molecular
sieves would seem be the optimum drying method
38 Dichloromethane (DCM) toluene and tetrahydrofuran (THF)
Toluene DCM and THF are common organic solvents that can be found in most
synthesis laboratories DCM and toluene are non polar aprotic solvents that are
relatively non hygroscopic On the other hand THF is a polar aprotic solvent that is
relatively hygroscopic and difficult to dry As mentioned above toluene and THF are
usually dried by heating over sodium and DCM is dried by distillation over CaH2
CaSO4 or other drying agents
In the current study it was observed that a solvent left in an oven-dried Schlenk tube
in the glove box was found to contain less water after a few days As glass is made
predominantly of silica it was decided to attempt to dry these three solvents by using
standard silica The solvents were also dried using 3Aring molecular sieves and the results
are summarised in Table 313
128
Table 313 Results of Karl Fischer titrations of toluene DCM and THF dried over 3Aring
molecular sieves and silica
Solvent Desiccant Time(h)
Average sample weight
(g)
n
Residual water
contenta (ppm)
Std dev ppm
RSD ()
DCM 3Aring molecular sieves 10 (wv) 24 426 6 01 005 4167
Toluene 3Aring molecular sieves 10 (wv) 24 276 6 09 026 3023
THF 3Aring molecular sieves 10 (wv) 24 278 6 277 099 32
DCM Silicab na 422 6 13 031 2441 Toluene Silicab na 214 6 21 017 796
THF Silicab na 253 6 562 253 451 a Outliers have been removed confidence level 95 b Silica grade 12 pore size 22Aring mesh 28-200
Table 314 Results of Karl Fischer titrations of THF dried over 3Aring molecular sieves
(20 wv)
Time (h) n
Average sample weighta
(g)
Residual water
content (ppm)
Std dev (ppm)
RSD ()
24 6 3029 152 029 191 24 6 3027 143 012 084 24 6 3007 146 036 247 48 6 3034 59 021 356 48 6 3000 60 021 350 48 6 2998 64 010 156 72 6 2970 43 005 116 72 6 3016 42 018 429 72 6 2907 38 010 263
a Outliers have been removed confidence level 95
Both toluene and DCM were successfully dried with the 3Aring molecular sieves After
only 24 hours both solvents show a level of residual water of less than 1 ppm Not
unexpectedly the THF was more difficult to dry with the molecular sieves after 24
hours approximately 30 ppm water still remained in the sample that had been dried
over 10 (wv) of the sieves Increasing the amount of molecular sieves to 20 (wv)
129
drying agent (Table 314) led to improved reduction of the amount of residual water
in the THF after a 24 hours period to around 15 ppm while an additional 24 hours of
contact with the drying agent reduces the water content further to roughly 6 ppm It
was noted that leaving the THF for a further time period made little difference to the
water content These experiments indicated that activated 3Aring molecular sieves are
suitable for the storage of dried THF The experiments using 20 (wv) were repeated
three times and showed good repeatability
Pleasingly the experiments conducted by passing the solvent through a column of
silica worked equally well in the case of toluene and DCM The solvents were dried
rapidly and Karl Fischer titration of DCM indicated that only 13 ppm water remained
with only 21 ppm being present in the case of toluene Again THF proved more
difficult to dry and the residual water was found to be around 56 ppm
Encouraged by the initial results obtained from the drying of the solvents by silica it
was decided to extend the study to other types of silica with the view to investigating
the influence of pore size and mesh size on the efficacy of the drying agent The
solvent used for this study was THF Since this solvent is particularly difficult to dry
small variations in results should be conspicuous In all cases 10 (wv) of the drying
agent was used All experiments were carried out in the glove box and the results are
summarised in Table 315 The THF was simply passed over a column of the drying
agent specified collected and the water content thereof directly measured using Karl
Fischer titration
130
Table 315 Results of Karl Fischer titrations for THF dried by passing over columns
of various types of silica and alumina
Silica Average sample weight
(g)
n
Residual water
contenta (ppm)
Std dev ppm
RSD ()
Grade Pore size Mesh
636 60Aring 35-60 2871 6 1057 351 332 636 60Aring 60-100 2904 6 894 283 317 634 60Aring 100-200 2873 6 746 287 385 633 60Aring 200-425 2930 6 595 366 615 644 100Aring 100-200 2913 6 690 327 473 643 100Aring 200-425 2927 6 608 193 317
Flash silica 70-230 2980 6 825 118 143 Neutral alumina - 2966 6 61 062 102 Neutral alumina - 2924 6 49 015 306 Neutral alumina - 2826 6 68 037 537
a Outliers have been removed confidence level 95
The above table of results shows that the various types of silica are only moderately
effective in the drying of THF at this weight-to-volume loading What is interesting to
note is that the pore size of the silica appears to make very little difference to the
drying efficiency in this method (Table 315) When one compares the results
obtained for the 60 Aring 100-200 and 200-425 mesh silica with those obtained for 100 Aring
silica of the same mesh size it is apparent that the residual water remaining in the THF
is approximately the same with respect to the two pore sizes A possible explanation
for this is that the contact time of the solvent on the silica in this method is minimal so
as to render the pore size irrelevant In contrast when solvents are dried with
molecular sieves contact time with the sieves is for an extended period of time
On the other hand the mesh size appears to play a more significant role in the drying
of the solvent (Figure 35) The silica of mesh size 35-60 is relatively large grained
silica as the numbering of the mesh size increases the silica becomes finer As the
silica becomes finer its ability to remove water from the THF increases (Figure 35)
Higher mesh numbers (finer particles) provide greater surface area due to smaller
particle sizes as this aspect appears to be determinative of drying ability
131
Figure 35 Graph indicating decreasing water content with increasing silica mesh size
For comparative purposes with the DCM and toluene THF was also dried through a
column of activated neutral alumina (Table 315) This method proved very efficient
for the removal of water from the THF After only one passage over the alumina (10
(wv) a 94 reduction in the water content of the THF was seen This is a vast
improvement from the traditional drying methods This method also proved to be
repeatable
132
Figure 36 Graph of residual water content in THF after drying with various
desiccants under given conditions
133
In summary the results indicate that the most effective method of drying THF is by
passage over a column of neutral alumina (Figure 36) The use of 3Aring molecular
sieves successfully removes water from the solvents However this method of water
removal requires at least 48 hours to be as efficient as the alumina column method As
in the case of acetonitrile it would appear that the best method of drying this solvent
would be first passage over a column of neutral alumina and then storage over 3Aring
molecular sieves to ensure that the solvent remains dry
39 Discussion
The following chapter of this thesis (chapter 4) is concerned with the role of water in
Lewis acid catalysis in particular with metal triflates These Lewis acids are thought
to be able to form a type of Broslashnsted acidity in the presence of water in organic
solvent or in the presence of a protic solvent
The above study has shown that traditional drying methods can still leave solvents
with relatively high water content with respect to a catalyst The study showed THF
that had been dried over sodium still contained approximately 43 ppm water which
equates to 0024 mmol of water in 10 mL of the solvent If one uses a hypothetical
situation in which a typical reaction is carried out in the above mentioned 10 mL of
solvent using 100 mg of benzaldehyde and a catalyst for example Al(OTf)3 at a
loading of 10 mol (0094 mmol) the ratio of water to catalyst in the reaction
mixture would be 14 This ratio is relatively high for a solvent that would be
considered dry under normal circumstances (the THF was blue from the ketyl radical
from the benzophenone) especially if one considers that most of the metals used in
Lewis acid catalysis are capable of expanding their coordination sphere extensively
and that through this the Lewis acid is capable of forming another acidic species that
may influence the outcome of the reaction
310 Conclusions
Karl Fischer titration for the determination of water content in organic solvent proved
to be an effective analytical method By using the Karl Fischer apparatus in a glove
box errors due to atmospheric water where minimised Measurements were obtained
134
with high precision especially at higher water concentration At lower concentrations
of water it was found that the accuracy of this method increases with the amount of
sample introduced into the titration cell However when a large number of titrations
are needed this is not always a viable option
The most effective method of drying was found to be contact of the solvent with 3Aring
molecular sieves This technique worked for a broad spectrum of solvents including
the alcohols and THF which are notoriously more difficult to dry than solvents like
DCM and toluene The drawback of using the molecular sieves method however is
that it is time consuming Drying THF to a level that was acceptable for the current
study required 72 hours using 20 (wv) molecular sieves
Neutral alumina that has been activated in an oven overnight is an extremely efficient
way to dry organic solvents At just 10 (wv) loading a 94 reduction in the water
content of the THF was seen When the same method was used with acetonitrile a
96 reduction in water content was obtained This method has the advantage over
molecular sieves of been rapid Should the two methods be combined ie passage
over alumina followed by storage over molecular sieves low levels of water in the
organic solvent can be rapidly obtained and maintained
Some success was achieved by using activated silica as a drying reagent This method
was more efficient for the non polar solvents such as DCM and toluene Limited
success was realised with the drying of THF using silica A more extensive study
indicated that the pore size of the silica was irrelevant when using this method but
that mesh size played an important role As the silica becomes finer grained its
capacity for removing water from the organic solvent increases
A simple calculation highlights the importance of using dry solvents in Lewis acid
catalysis especially if the Lewis acid in question is subject to hydrolysis (eg TiCl4 or
BF3) The following chapter of this thesis investigates the role of water in Lewis acid
catalysis and highlights the importance of working in ultra dry solvents so as to
eliminate the possibility of forming a Broslashnsted acid species
135
136
38 References
1 a) Jacobs T L Truce W E Robertson G R Laboratory Practice of
Organic Chemistry Macmillan New York NY 1974 b) Roberts R M
Gilbert J C Rodewald L B Winegrove A S An Introduction to Modern
Experimental Organic Chemistry 2nd ed Holt Rineholt and Winston New
York NY 1974 c) Fieser L F Fieser M Reagent for Organic Synthesis
Wiley New York NY 1967 d) Vogel A I A Text Book of Practical
Organic Chemistry 3rd ed Longmans London 1964
2 Plesch P H Ed The Chemistry of Cationic Polymerisation Pergamon Press
Oxford 1963 682
3 a) Burfield D R Anal Chem 1976 48 2285 b) Burfield D R Smithers
R H J Org Chem 1978 43 3966 c) Burfield D R J Org Chem 1984
49 3852 d) Burfield D R Smithers R H J Org Chem 1983 48 2420
4 Bower J H J Res Nat Stand 1934 12 241
5 Trussell F Diehl H Anal Chem 1963 35 674
6 Meeker R L Critchfield F Bishop E F Anal Chem 1962 34 1510
7 Fischer K Angew Chem 1935 48 394
8 Fundamentals of the Coulometric Karl Fischer Titration with Selected
Applications Mettler Toledo Switzerland 2003
9 Perrin DD Armarego W L F Purification of Laboratory Chemicals
Pergamon Press New York 1988
10 Merck Drying Agents Catalogue 2005
11 Burfield D R Lee K H Smithers R H J Org Chem 1977 42 3060
Chapter 4 Investigations on the role of water in metal triflate catalysed reactions
41 Introduction
Another focus of this study is to contribute to the understanding of the role of water which
influences the activity of metal triflate as catalysts
The use of metal triflates as Lewis acid catalysts is now a well established practice amongst
chemists In 2002 Kobayashi1 wrote a comprehensive review containing over 400 references
on rare earth metal triflates in organic synthesis Included in this review are many
comparative studies to many other metal triflates Since this review was published the interest
in metal triflate catalysis has continued to grow2
Metal triflates are remarkable in that they are easy to handle when compared to the more
traditional Lewis acids ie many of them can be weighed using a balance open to the
atmosphere They can be used in catalytic amounts and they are recoverable and reusable
without loss of activity from reaction mixtures
Even more remarkable is the fact that metal triflates are known to act as Lewis acid catalysts
in the presence of water3a-d This is noteworthy as traditional Lewis acids such as AlCl3 and
TiCl4 hydrolyse in the presence of even the smallest amounts of moisture rendering them
ineffective as Lewis acid catalysts
On closer inspection of the literature however there seems to be some disagreement as to the
role that the metal triflate plays in catalysis Does the metal triflate itself catalyse the reaction
acting as a Lewis acid4 or does it undergo hydrolysis to form triflic acid which in turn
catalyses the reaction In the latter case the reactions could be Broslashnsted acid catalysed5
Alternatively reactions may be catalysed by a dual mechanism ie Broslashnsted assisted Lewis
acidity6
The reaction of metal compounds with water as given in standard inorganic textbooks is
often referred to as hydrolysis This is the reaction of MXn species with water such that water
is inserted into one or more of the M-X bonds (Scheme 41)
137
SnCl4 + 4H2O Sn(OH)4 + 4HCl H+ + Cl- Scheme 41 Hydrolysis of a metal
However organic chemists usually use the term hydrolysis for the cleavage of an organic
molecule or specific covalent bond with the agency of water (Scheme 42)
ROSi(CH3)3 + H2O ROH + HOSi(CH3)3 Scheme 42 Hydrolysis of organic compounds
In this chapter the term hydrolysis will refer to all reactions with water leading to the
formation of protons or their equivalent unless the context clearly refers to the type of
hydrolysis referred to in scheme 42 To some extent the two types of hydrolysis shall be
distinguished by the terms lsquoinorganic hydrolysisrsquo and lsquoorganic hydrolysisrsquo
42 The arguments for and against triflic acid as the true catalyst taken from the
literature
A great deal of work has been carried out by numerous people in an attempt to determine the
mechanism of metal triflate catalysis The question of whether or not metal triflates form
triflic acid in water containing organic solvents is not a trivial one From our own work on the
ring opening of epoxides the answer appears to be that triflic acid does not form and the
catalysis is Lewis acid driven4
When styrene oxide was ring-opened using ethanol in the presence of 00005 mol Al(OTf)3
a 94 yield of the glycol ether was obtained in 1 hour Even when 0005 mol TfOH was
used in the same reaction the product is formed in only 5 yield after 24 hours (at a catalyst
loading ten times greater than that of the Al(OTf)3) This strongly suggests that TfOH is not
the catalyst in these reactions
Similarly when Bi(OTf)3 was used in aqueous medium to catalyse the ring-opening of
epoxides7 it was far more efficient than TfOH When 10 mol of the bismuth triflate was
used in a reaction with aniline and cyclohexene oxide the expected product was afforded in
83 yield in 7 hours When the same reaction is carried out using 10 mol of TfOH the
138
product is obtained in only 28 yield in the same time period According to the authors this
suggests that the Lewis acid is involved in activating the epoxide
In 1969 Brown8 synthesised a range of sterically hindered 26-disubstituted pyridine
derivatives in order to study the effects of the substituents on the basicity of the molecules
The study showed that there was an increase in base strength as the substituent in the 26
positions became bulkier However the basicity of the pyridine series dropped dramatically
in the case of 26-di-tert-butyl pyridine (Figure 41)
Figure 41 Effect on the pKa of increasing steric requirements of the alkyl groups in the 26
positions (in 50 ethanol)
The authors suggest that the loss of basicity is due to the steric hindrance of the bulky-tert
butyl groups and the fact that they are cannot minimise strain by rotating about a C-C bound
the way for example isopropyl groups can This results in steric interactions with the
nitrogen-bound H+ ion and leads to a higher tendency to ionise (Scheme 43)
NHCH3
H3C
H3C CH3
CH3H3CN
CH3H3C
H3C CH3
CH3H3C
+ H+
Scheme 43 Ionisation of 26-di-tert-butylpyridine
139
In a further study8 they showed that boron trifluoride reacts readily with 26-lutidine and 2-
tert-butylpyridine but failed to react with 26-di-tert-butylpyridine On the other hand 26-di-
tert-butylpyridine reacts with HCl forming the dihydrochloride product (the second HCl
molecule co crystallising in the solid state) The authors suggest that the base can be used to
separate a mixture of BF3 and HCl because of its differentiation between Broslashnsted acids and
even small Lewis acids
In the discussion that follows the work carried out by Brown has been used by other workers
to distinguish between Lewis acidity and Broslashnsted acidity 26-Di-tert-butylpyridine (DTBP)
and its derivative 26-di-tert-butyl-4-methylpyridine (DTBMP) may be used as proton
scavengers in reactions where there are possibilities for both Lewis and Broslashnsted activity
Once the base has been added the reaction should then be exclusively Lewis acid driven
Ollevier et al9 carried out a study on the Mannich reaction in water using Bi(OTf)3 A model
reaction (Scheme 44) was used to determine if triflic acid played a role in the mechanism
5 mol Bi(OTf)3 H2OH
O
Ph NH2
O NH OPh
+ +
41 42 43 44
Water 7 h
Scheme 44 Direct type catalytic Mannich reaction of benzaldehyde aniline and
cyclohexanone in water
The reaction was carried out using 5 mol Bi(OTf)3 one equivalent of benzaldehyde one
equivalent of aniline and three equivalents of cyclohexanone in water at room temperature for
seven hours The desired product was afforded in an 84 yield with a stereoselectivity of
8614 antisyn when the reaction was repeated with 15 mol of the base DTBMP 44 was
obtained in a 83 yield with a stereoselectivity of 7425 antisyn However this does not
indicate that a Lewis acid is necessarily responsible for the catalysis of the reaction When the
reaction is repeated without the Bi(OTf)3 and only the 15 mol DTBMP 44 is afforded in a
76 yield with a stereo selectivity of 7921 antisyn The Bi(OTf)3 catalysed Mannich
reaction was repeated this time in the presence of the proton scavenger K2CO3 and afforded
a 44 yield of 44 (7030 antisyn) The authors interpreted this to indicate that Broslashnsted
acidity was involved in the process When the reaction was carried out using 15 mol TfOH
140
the yield obtained was 92 but with lower stereo selectivity (7723 antisyn) The reaction
shows a clear preference for the anti product the excess of which is determined to some
extent by the reaction conditions The fact that some of the antisyn ratios are so close in
many instances complicates the interpretation of this data
It should also be noted that K2CO3 which the authors used as a proton scavenger generates
KOH in water This relatively strong base could produce Bi(OH)(OTf)2 which will reduce the
Lewis acidity of the catalyst accounting for the lower yield of the reaction a point which the
authors failed to make
In their investigation on the metal triflate catalysed acylation of alcohols Dumeunier and
Markoacute10 found triflic acid to be an active catalyst in the reaction In an attempt to benzoylate
45 (Scheme 45) no desired product 46 was formed when Sc(OTf)3 was used Surprisingly
the acid sensitive TES group was clipped off resulting in the diol 47 This was ascribed to
the action of triflic acid formed by the reaction of Sc(OTf)3 with water arising from the
dehydration of 45 (a side reaction that consumed 30 of the starting material 48) In contrast
the benzoylation of 49 resulted in the smooth formation of 410 in a 96 yield
OTES OBz
PhO2S
OH OH
PhO2S
OTES OH
PhO2S
45
46
47
OH
PhO2S
48
Bz2O CH2CN
10 mol Sc(OTf)3
OH
SO2Ph
10 mol Sc(OTf)3
OBz
SO2PhBz2O CH3CN
49 410 Scheme 45 Metal triflate catalysed acylation of alcohols
141
These results suggested to the authors that triflic acid was responsible for the deprotection of
the silyl ether In order to neutralise the acid the benzoylation of 49 was repeated but 26-di-
tert-butyl-4-methyl pyridine (DTBMP) a highly hindered base thought to only interact with
protons as discussed above8 was added to the reaction mixture No benzoylation of 49
occurred even after the reaction mixture was heated to 50 degC The authors suggest that this
result strongly points to triflic acid playing a prominent role in the reaction
In order to shed more light on the active species a model reaction was carried out (Scheme
46) using a variety of triflates The reactions were carried out with and without DTBMP the
addition of DTBMP and the results are summarised in Table 41
OH O Ph
O3 eq Bz2O
M(OTf)n MeCN Scheme 46 Benzoylation of hex-2-en-1-ol using metal triflates
Table 41 Conversions () of metal triflate catalysed benzoylations
Entry Metal Triflate
(5)
Conversion
()a
DTBMP
15
Conversion
()a
1 Yb(OTf)3 99 15 11
2 Bi(OTf)3 40 15 8
3 In(OTf)3 34 15 3
4 Sc(OTf)3 63 15 2
5 Cu(OTf)2 3 - -
6 Sc(OTf)3 63 5 6
7 TfOH (8) 30 - - a Measured by capillary GC after calibration of the response for each component
Adding the hindered base dramatically suppressed the formation of the benzoylated product
in all cases The reaction was also carried out using 8 mol of triflic acid and a conversion of
30 was seen after 5 hours When a second portion of the acid was added (8 mol) after this
period a conversion of 82 was achieved This poorer conversion effected by triflic acid is
explained by the authors to be a result of the formation of water generated by the side
reaction (dehydration reaction of hex-2-en-1-ol) which could be responsible for attenuating
142
the acidity of the triflic acid To test the theory a 21 H2OTfOH solution was prepared in
acetonitrile and added to the reaction mixture Approximately the same yield (21) was
obtained after 20 minutes as previously The authors suggest that the results strongly support
the intervention of triflic acid as the active catalyst in these reactions and that water plays an
important role in modulating its activity
If TfOH is the causative agent in these reactions it can be questioned why its activity is not
modulated in the M(OTf)3 catalysed reactions assuming of course that these metal triflates
do hydrolyse to form triflic acid as the authors have suggested in this paper The authors fail
to address this issue in their article
The peracetylation of sugars was successfully carried out using 5 mol of In(OTf)3 in acetic
anhydride when stirred for 1 hour at zero degrees11 Triflic acid readily catalysed the
acetylation of galactose (Table 42 entry 1) However the reaction with glucose does not
proceed when 26-di-tert-butylpyridine (DTBP) is added to the reaction mixture with the
In(OTf)3 (Table 42 entry 3)
Table 42 Yield () of acetylation reactions
Entry Sugar Catalyst Time Yield
()
1 O
OHHOHO
HOOH
TfOH (005 eq) 10 min 84
2 O
OHHOHO
HO
HO
In(OTf)3 (005 eq) 1 h 99
3 O
OHHOHO
HO
HO
In(OTf)3 (005 eq)
With DTMP 24 h 11
4 O
OHH2NHO
HO
HO
In(OTf)3 (005 eq)
1 h NR
The authors concluded that these acetylation reactions involved a dual pathway namely a
reaction catalysed by triflic acid presumably formed from when In(OTf)3 binds to the sugar
and a proton is released (Scheme 47) and a reaction catalysed by the Lewis acid In(OTf)3
143
O
OHH2NHO
HO
HO
+ In(OTf)3
O
OIn(OTf)2H2N
HOHO
HO
+ HOTf
Scheme 47 Possible source of triflic acid in acetylation reaction
The reaction involving a Broslashnsted acid could be suppressed by the addition of the hindered
base DTBMP However it should be noted that the addition of the primary base glucosamine
(not sterically hindered) suppressed both reactions (Table 42 entry 4) The lack of reactivity
seen in the presence of the amino group could also be ascribed to the coordination of the
metal centre to the nitrogen atom of the amine this decreasing the Lewis acidity of the metal
During their research on the Yb(OTf)3 catalysed preparation of calyx[4]resorcinarenes carried
out by Barrett et al12 a study was conducted on the isomerisation of the benzaldehyde-
derived resorcinarene 411 and 412 (Figure 42) It has been previously reported that the all
cis isomer 11 is consistent with Broslashnsted acid catalysis1314
Ph
PhPh
Ph Ph
Ph
Ph
Ph
HO OH
HO
HO
HO OH
OH
OH
HO
HO
HO OH
HO OH
OH
OH
411 412 Figure 42 Resorcinarene isomers
Their study of isomerisation over time was interpreted to be a Broslashnsted acid mechanism of
reversible cyclisation This interpretation was based on the suppression of the reaction
(isomerisation in particular) by the addition of the hindered base DTBMP
However the possibility of a dual mechanism operating throughout the reaction with at least
equal importance for both Broslashnsted and Lewis acidity was not considered This is because the
hindered base was only considered to be a ldquoproton spongerdquo and its reaction with the protic
product or the solvent ethanol to produce phenoxide or ethoxide ions was not considered
144
This omission is particularly noteworthy because the base was needed in such a vast excess in
order to suppress isomerisation The possibility that phenoxide or ethoxide ions so generated
could also deactivate Lewis acids was never considered The possibility of a dual mechanism
for the isomerisation reaction can therefore not be excluded
Further more the authors put forward from their work with metal triflate catalysed nitration
reactions15-16 and acetylations with AcOH17 that it would appear that these Lewis acids bind
to the ligand to form a chelate structure (Figure 43) These in turn will make stronger
Broslashnsted acids than the parent nitric or acetic acid
n(H2O)YbO
ON O
H
n(H2O)YbO
OH
3+
3+
3OTf-
3OTf-
Figure 43 Enhanced Broslashnsted acid
The authors speculate that the chelates will undergo loss of triflic acid until equilibrium is
established with an associated gain of an inner sphere nitrate or acetate They indicate that
triflic acid or the enhanced Broslashnsted acidity is responsible for the promotion of the reaction
and not the Lewis acid directly
From the above discussion it is clear that there is no clear evidence in the literature as to
which moiety catalyses the reactions the Lewis acid metal triflate itself or the product of
hydrolysis (formed by the reaction with water or protic solvent) to form a protic acid
Another possibility suggested by Barrett is that another type of Broslashnsted acid is formed in
which the Lewis acid binds to the protic acid that is already present leading to the creation of
a Lewis-assisted Broslashnsted acid
The investigation that follows is an attempt to determine if the mechanism of metal triflate
catalysis is Broslashnsted acid driven Lewis acid driven or by a combination of the two and also
to what extent if any water plays a role in the catalysis
145
43 The Mukaiyama aldol Reaction
431 Introduction
The name aldol is given to a class of reactions between enolates (or enols) and carbonyl
compounds (Scheme 48) such as aldehydes and ketones18 The aldol reaction is almost
exclusively base catalysed Broslashnsted acid catalysed aldol reactions are known but are rare19
R1 R2
OH
OO
OH H H
O
Enolate
R1 H
OO
OH H
R1 H
OOH
Aldol
OH
R2 R2
Scheme 48 Base catalysed aldol reaction
The Mukaiyama aldol reaction20 is a type of aldol reaction where the ketone is treated with a
silyl enol ether (Scheme 49) The reaction is Lewis acid catalysed Originally TiCl4 was
used while other catalysts that have been successfully employed in this reaction are InCl321
SmI222 and Bi(OTf)323
R1
CHR2
OSiMe3+ R3 C R4
O
R3
C R4
OH
HCR2
CO
R1
Silyl enol ether Aldol
Lewis acid
Scheme 49 Mukaiyama aldol reaction
432 Metal salt catalysed Mukaiyama aldol reaction
In 1998 Kobayashi et al24 published a paper on work carried out on the Mukaiyama aldol
reaction using a range of Lewis acid catalysts including metal triflates chlorides and
perchlorates The objective of the investigation was to determine the correlation between the
catalytic activity of these Lewis acids in a THF-water mixture on the one hand and their
hydrolysis constants and water exchange rate constants on the other Benzaldehyde was
reacted with (Z)-1-phenyl-1-(trimethylsilyoxy) propene (Scheme 410) using 20 mol of the
Lewis acid in a THF H2O (91) mixture at room temperature for 24 hours
146
OSiMe3 O
H
O OH
+MXn 02eq
THFH2O (91)12 h rt413 414 415
Scheme 410 The Mukaiyama aldol reaction catalysed by a Lewis acid
This reaction is suitable for testing catalytic ability of the Lewis acids in aqueous media as
the silyl enol ether is sensitive to hydrolysis (of the organic type) under acidic conditions If
the Lewis acids hydrolyse in water the silyl ether will decompose and the aldol reaction will
proceed no further
This reaction has a two-fold purpose Firstly the results according to Kobayashi et al24 will
provide some indication as to the catalytic ability of the Lewis acid in aqueous media If the
Lewis acid undergoes hydrolysis slowly and conditions remain relatively non acidic the silyl
enol ether should remain stable and the aldol product should form Secondly and most
importantly to this investigation the reaction can also be used to determine if the metal
triflates hydrolyse to form Broslashnsted acid species If this is the case the silyl enol ether will be
clipped off affording little or no aldol product
The prior investigation (Table 43) shows that the rare earth metal triflates used in their
investigation all successfully catalysed the aldol reaction to afford good yields of the desired
product
That work formed the starting point of the present investigation In this work reactions were
carried out using the same conditions as previously described but non-rare earth metal
triflates were used here For comparative purposes several of the Lewis acids used by
Kobayashi24 were included as part of the present study
147
Table 43 Yield () of aldol product from different metal triflates24
Metal Triflate Yield ()
La(OTf)3 80
Ce(OTf)3 81
Pr(OTf)3 83
Nd(OTf)3 78
Sm(OTf)3 85
Eu(OTf)3 88
Gd(OTf)3 90
Tb(OTf)3 81
Dy(OTf)3 85
Ho(OTf)3 89
Er(OTf)3 86
Tm(OTf)3 85
Yb(OTf)3 92
Lu(OTf)3 84
Table 44 below shows the results of the reactions carried out in the present study Of
particular interest in this investigation was the reaction performed with Al(OTf)3 which gave
a poor yield of the aldol product and showed extensive silyl ether organic hydrolysis (79)
On the other hand Cu(OTf)2 Sc(OTf)3 Y(OTf)3 and Zn(OTf)2 afforded high yields of the
aldol product and showed relatively little silyl ether hydrolysis The 4 valent metals Hf and Zr
showed only hydrolysis of the silyl enol ether and no aldol product LiOTf failed to catalyse
the reaction and the 1H NMR spectrum of the product showed mainly starting material
remaining Triflic acid and TiCl4 also effected quantitative silyl ether hydrolysis as expected
From these results we can assume that those Lewis acids that catalysed the reaction to afford
high yields of the aldol product are good Lewis acid catalysts in aqueous media as they
appear to undergo weak inorganic hydrolysis Examples include Sc Cu and Y triflate
148
Table 44 Results of Mukaiyama aldol reaction
Lewis acid Yield ()a
aldol
Yield ()a
Hydrolysis of
Silyl ether
Yield () aStarting material
(silyl ether)
La(OTf)3 92 (80)b 8 0
InCl3 89 (68)b 3 8
CuCl2 86 (25)b 11 3
LiCl 6 (-)b 7 77
ScCl3 59 (70) 41 -
Er(OTf)3 56 (86)b 6 38
GaCl3 28 (-)b 56 16
ZnCl2 22 (10)b 0 78
AlCl3 0 (0)b 100 0
Cu(OTf)2 96 4 0
Sc(OTf)3 94 6 0
Y(OTf)3 89 11 0
Zn(OTf)2 77 2 21
Ca(OTf)2 42 28 30
In(OTf)3 41 59 0
Al(OTf)3 21 79 0
Zr(OTf)4 0 100 0
Hf(OTf)4 0 100 0
Li(OTf) 0 11 89
TiCl4 0 100 0
TfOH 0 100 0 a Yields determined by 1H NMR analysis b Yields in parenthesis refer to those of Kobayashi et al14
From Table 44 it can be seen that unlike the rare earth metal triflates several of the triflates
used in this investigation appeared quite susceptible to inorganic hydrolysis and gave rise to
in some form of Broslashnsted acidity (particularly those with a high charge density) If this was
the case then the pH of the solutions that cause silyl ether organic hydrolysis should be more
acidic relative to the pH of those that catalyse the aldol reaction In an attempt to rationalise
these results solutions were made up that mimicked the reaction mixtures (20 mol Lewis
149
acid in THF H2O) and the pH readings were taken of these solutions (Table 45) With the
clear exception of Sc(OTf)3 which is classified as a rare earth metal all of the Lewis acid
solutions with pH lt 4 caused the silyl enol ether to hydrolyse to an extent greater than 40
Table 45 Average pH readings of Lewis acids in THF H2O mixture
Lewis acid Ave pHbc Hydrolysis
TiCl4 -098 100 TfOH 034 100
Zr(OTf)4 138 100 Hf(OTf)4 160 100 GaCl3 214 56 ScCl3 254 41 AlCl3 273 100
Sc(OTf)3 275 6 In(OTf)3 277 59 Al(OTf)3 325 79 CuCl2 408 11 InCl3 432 6
Y(OTf)3 445 6 Cu(OTf)2 464 4 Zn(OTf)2 540 2
ZnCl2 549 0 Er(OTf)3 559 0 La(OTf)3 564 8 Ca(OTf)2 607 28
LiCl 830 7 LiOTf 889 11
a Readings were carried out at 25 degC (temperature-controlled water bath) b Average of 3 readings c Refer to values obtained
with a standard aqueous pH electrode
According to Kobayashi there is a correlation between the catalytic activity of the Lewis acid
in aqueous media and their WERC (water exchange rate constant) as well as their pKh (Kh =
hydrolysis constant)24 Cations with large pKh values ie gt43 hydrolyse weakly whereas
cations with pKh values lt43 are extensively hydrolysed and oxonium ions are formed25
It is under these latter conditions that we would expect to see a significant drop in the pH
values If we compare the pH values recorded in this study for the different Lewis acids with
the hydrolysis constants found listed in the literature25 (Table 46) there is a general
150
correlation between them The cations that are more readily hydrolysed show lower pH
values while the cations that hydrolyse with difficulty gave higher pH values This correlation
is more visible for the cations that are more difficult to hydrolyse
Table 46 Comparison of pH and pKh values
Lewis acid Ave pH Hydrolysis pKh25
Zr(OTf)4 138 100 022 Hf(OTf)4 160 100 025 Al(OTf)3 325 79 114
TiCl4 -098 100 230 easily hydrolysed GaCl3 214 56 260
In(OTf)3 277 59 400 InCl3 432 6 400 ScCl3 254 41 430
Sc(OTf)3 275 6 430 CuCl2 408 11 753
Cu(OTf)2 464 4 753 Y(OTf)3 445 6 770 Difficult to hydrolyse Er(OTf)3 559 0 790 La(OTf)3 564 8 850 Zn(OTf)2 540 2 896
ZnCl2 549 0 896 Ca(OTf)2 607 28 1285
LiCl 830 7 1364 LiOTf 889 11 1364 TfOH 034 100
The idea of invoking the hydrolysis constants may be useful but poses some problems if
applied in a simplistic fashion To simply write a pKh value for a given metal ion is to ignore
the potential effects of the counterion of that cation on the pKh value of that given metal
entity The pKh values cited in Table 46 are derived from previous work and were calculated
for a variety of metal salts including those of sulphates nitrates and chlorides These
constants would be useful for the particular metal with the particular counterions but not
necessarily useful for the extrapolation to other counterions
151
As a case in point salts of Zn+2 (if we accept that pKh values hold for all salts of a given
metal) are difficult to hydrolyse since they present with a pKh value of 896 This would
equally hold true of salts of Li+ which have a pKh value of 1364 (It should be remembered
at this stage that the pKh values were determined primarily from sulphates nitrates and
chloride salts of these metals) No chemist with any experience of organozinc or
organolithium compounds would support the view that such compounds are difficult to
hydrolyse Indeed the opposite view would be held
The converse would be true of certain compounds of tin(IV) SnCl4 is correctly held to
hydrolyse rapidly In contrast Bu3SnOH used as an anti-foulant in paints for marine
structures including ships is sufficiently long-lived to be a marine pollutant that disrupts the
reproductive cycles of many marine organisms26 These simple examples exemplify the
stance that pKh values should be used judiciously in the interpretation of other chemical data
In the present work the readerrsquos attention is drawn to the entries for In+3 and Sc+3 In(OTf)3
and InCl3 give pH values of 28 and 43 while leading to hydrolysis of the silyl enol ether to
the extent of 59 and 6 respectively Exactly the opposite result is obtained for Sc+3 ScCl3
gives a pH of 25 and hydrolysis of the enol ether of 41 while ScCl3 gives a pH of 28 but
shows hydrolysis of the enol ether of only 6 These results taken together with the
arguments above clearly indicate the situation to be more complex than a simple reliance
upon a single parameter purports There is an equally strong if not stronger correlation
between pH and the extent of hydrolysis of the enol ether than there is when using the pKh
values but even here there are some results that are difficult to rationalise (see for example
the entry for Ca(OTf)2 in Table 46) where even at high pH values some hydrolysis takes
place In this instance it is quite possible that the use of a combination of Broslashnsted-driven and
metal-driven mechanisms is the best approach to rationalising the outcomes
433 The possible role of water in the Mukaiyama aldol reaction
In this study and others still to be reported many metal triflates are not only tolerant to water
but at least in some cases require the presence of water to operate efficiently Kobayashi
although never expressed in this way has alluded to the importance of water in the catalytic
activity of metal triflates and has specifically ascribed the effect to the water exchange
phenomenon While not in disagreement with this view the results of the present
152
investigation suggest a more complex situation with regard to the referred to water exchange
rate constant
Kobayashi2427a indicated that metal triflates that are good Lewis acid catalysts in aqueous
media have fast WERC (water exchange rate constants) In a mixture of THFwater the metal
will preferentially bind to the water molecules present and that these will rapidly exchange
on the metal with other water molecules This phenomenon is said by Kobayashi not to occur
or to occur to a much slower extent with coordinating solvents such as THF Due to the rapid
exchange of the water molecule on the metal there is a chance for an aldehyde to bind to and
be activated by the metal and for the aldol reaction to occur
Against this background Kobayashi ascribed the lack of reactivity in dry THF to the slow
exchange between the THF coordinated metal and the aldehyde This suggestion cannot go
unchallenged in view of the fact that other metal coordinating species less likely to participate
in ligand exchange on metal cations (eg diamines27b and crown ethers) do not suppress the
catalytic activity of metal triflates27c Furthermore it has been demonstrated that in aqueous
solutions of Eu(OTf)3 the Eu3+ ion remains fully or nearly full coordinated to water even with
less than 5 water in THF27d The effect of low water concentration could well be associated
with the incomplete dissociation of Eu(OTf)3 the same situation should apply to other metal
triflates However an alternative explanation is tentatively afforded namely that the water is
an essential reagent to drive the reaction to completion by removing the silyl group of the
immediate product (Scheme 411)
OSiMe3
O
R R
M
+
H2O
OSiMe3
RR
OM
O O
M
RR
O OHR
R
+ H+
Scheme 411 Mukaiyama aldol reaction showing silyl ether removed by water
In the absence of water the poorly nucleophilic triflate will have to act as a nucleophile to
remove the silyl protecting group but in the process TMSOTf will be generated However
this is a very active silylating agent In this case the reaction would be at best reversible It is
therefore suggested that water (or possibly metal bound water as M-OH or M-OH2) plays the
role of the Cl- ion in the well known TiCl4ndashcatalysed or related Mukaiyama reactions
(Scheme 412) In this regard it is of interest to note that the majority of metal triflate
153
catalysed reactions which appear to be significantly enhanced by the presence of water
involves reactions of enol ether silyl enol ethers in particular1
OSiMe3
MePh
Ph
PhCHO TiCl4 CH2Cl2
OMe
Ph Ph
SiMe
MeMe
O
H Ph
TiCl3Cl -Me3SiCl
Me PhPhPh
O OTi
ClCl Cl
H2O
Me PhPhPh
O OH
+
Scheme 412 The Mukaiyama aldol reaction
Even if the role of water is not directly involved in the immediate aldol reaction it must still
play a role in the hydrolysis of the metalated aldol product in order to return the metal to the
catalytic cycle
In order to further investigate the role of water THF was dried in the present study by
passing it over a column of activated alumina (see chapter 3 for details) The water content
was determined by Karl Fischer analysis and found to be 8 ppm The Lewis acids were dried
under vacuum at elevated temperature to remove water The aldol reactions were repeated
using the same conditions as before and to ensure that there was no influence from external
water the reactions were carried out in a glove box The results both in the presence and
absence of water are summarised in Table 47
Table 47 Yield () aldol reactions in dry THF
Lewis acid Yield () Aldol
THF H2O (91)
Yield () Aldol in dry
THF
La(OTf)3 92 35 Y(OTf)3 89 83
InCl3 89 44 Zn(OTf)2 77 0 Sc(OTf)3 96 0a Cu(OTf)2 94 35
a Sc(OTf)3 polymerised the THF No desired reaction was detected
154
The question immediately arises whether this dramatic decrease in yield may be due to poor
solubility in dry THF Kobayashi made no comment in this regard In the present study it was
found that the effect could not be ascribed to low solubility since the metal triflates in
question displayed sufficient solubility in THF
(It is interesting to note that many metal triflates promote polymerisation of THF and that
such solutions cannot be kept for long periods of time This fact was included in the design of
these experiments and fresh solutions were always prepared)
Dichloromethane is a non-coordinating solvent and therefore even in the absence of water
this solvent should leave the metal open to the aldehyde for coordination and activation
(polarisation) for the aldol reaction to take place This would be in contrast to Kobayashirsquos
comments relating to the THF inhibiting such coordination in dry THF To determine if the
WERC was the only critical factor in these reactions the DCM was dried by passing it over a
column of alumina and Karl Fischer titration determined the water content to be only 2 ppm
The aldol reactions in this solvent were carried out in a glove box to prevent the ingress of
atmospheric water (Table 48)
Table 48 Yield () aldol reactions carried out in dry DCM
Lewis acid Yield () Aldol
La(OTf)3 lt1 Sc(OTf)3 lt1
InCl3 lt1 Cu(OTf)2 lt1 Zn(OTf)2 lt1
The virtual total unreactivity in this medium could not be ascribed solely to the insolubility of
the triflates in the DCM It is true that most metal triflates have limited solubility in this
solvent but it has been used successfully in related reactions28 where lower solubility resulted
in lower reactivity but not in the cessation of reactivity
From these results it is clear that it is not only the WERC that is playing a role in these
reactions although this concept canrsquot be altogether discounted as being partly determinative
of the success of the reaction at this stage When these reactions are carried out in a dry non
binding solvent DCM which for all intents and purposes contains no water the aldehyde
155
should be free to bind to the metal of the triflate However the reaction does not proceed It
would appear that water plays an important and possibly different role from that implied by
invoking the WERC concept in these reactions It has already been indicated earlier in this
thesis that it may be an essential ingredient in the reaction medium In order to ensure the
success of the reaction
To ensure that the binding of the aldehyde takes place in DCM (a prerequisite for the aldol
reaction to take place) NMR experiments were carried out The chemical shifts of
benzaldehyde were determined in CD2Cl2 (Figure 44) The respective metal triflates and
chlorides were then added to the benzaldehyde DCM mixture Any change in the chemical
shift of the signals in the spectra provides an indication of coordination The NMR samples
were made up in the glove box using dry CD2Cl2 (dried by passage over activated alumina)
and the Lewis acids were pre-dried under vacuum at elevated temperature to remove water
H
O
1001
788764755
Figure 44 Benzaldehyde showing 1H chemical shifts in CD2Cl2
Table 49 Change in chemical shift of benzaldehyde on complexation of metal triflate
Lewis Acid
Δδ H aldehyde
Δδ Ortho
protons
Δδ Para
protons
Δδ Meta
protons 1001 788 764 755
Sc(OTf)3 -0033a 0109 0108 0044
ScCl3 0054 0093 009 0029
In(OTf)3 -0015a 016 016 007
InCl3 0507 012 012 0064 aNegative values indicate an upfield shift
The results from the 1H NMR data (Table 49) not only provide direct evidence of metal
triflate solubility but also indicate that the aldehyde does in fact bind the metals in DCM In
all cases we see a shift in the 1H NMR spectra when the metal triflate or chloride is added to
the benzaldehydeDCM mixture indicating complexation of the metal This complexation
should in turn activate the aldehyde towards nucleophilic attack and the aldol reaction should
proceed (Figure45)
156
R1
CHR2
OSiMe3
H
O
M(OTf)3
δ
δ
Figure 45 Activation of benzaldehyde by metal triflate
It is of interest to note that on the addition of the metal salts to the aldehyde a single sharp
aldehyde peak (shifted from its original position as shown in Figure 46) is observed rather
than two peaks representing the complexed and uncomplexed aldehyde respectively (Figure
47)29 This situation points to the presence of only a single type of species in solution
(bound aldehyde) but may also represent the presence of rapid exchange of bound and free
aldehyde as shown pictorially in Figure 48 Low temperature NMR may have resolved this
question but was not pursued due to solubility problems (See chapter 5 for further discussions
on NMR) Any reduction in reaction rate in this solvent is unlikely to be due to exchange
phenomena
Figure 46 1H NMR of benzaldehyde complexed to ScCl3
157
Figure 47 NMR peaks showing the difference between fast intermediate and slow exchange
of ligands
Also noteworthy is the effect of the different Lewis acids on the aldehyde proton Both of the
metal triflates shift the aldehyde proton downfield the change in the shift is small Contrary
to this the metal chlorides shift the aldehyde peak upfield and the change in shift is much
greater this is possibly due to the chlorine atom interacting with the aldehyde proton this
would not occur with the oxygen atom of the triflate as it is electron deficient
44 A perspective of metal triflates in organic solvents
At this stage of the discussion it is important to give consideration to the manner in which
ionic compounds dissolve in organic solvents which vary widely in terms of polarity
dielectric constants and co-ordination abilities to metals
158
441 On solvation of the metal triflates in dry organic solvents
Judging by their high melting points and ease of dissolution in water it is reasonable to
assume that most metal triflates are ionic compounds The heat of solvation (qmx) of such
compounds in water will be given by equation 4130
qmx = umx ndash (wm ndash wx) eq 41
Where
umx = the energy required to convert the ionic lattice into separate ions
wm = the energy released on solvation of the cation
wx = the energy released on solvation of the anion
The lattice energy of umx of metal salts of the type MX2 is extremely high (in the region of
500ndash700 KCalmol-1 range) This will ensure relatively low solubility in organic solvents with
low polarity (and therefore low dielectric constants) Even here ion associations rather than
free ion pairs will be obsereved30 which decreases the potential energy of ions in solution31
In fact association between cations and anions in solution (close ion cluster formation) has
been shown to be magnitudes higher in solvents with intermediate to low dielectric constants
such as acetone (22 є) than in solvents with high dielectric constants such as
dimethylsulfoxide (472 є) (Table 410)
159
Table 410 Dielectric constants of several organic solvents
Solvent Dielectric constant (є)31
Water 79
Dimethyl sulphoxide 472
Acetonitrile 375
Methanol 315
Ethanol 242
Acetone 22
Chlorobenzene 56
Tetrahydrofuran 76
Ether 45
Benzene 23
Pentane 18
The clusters will (because of lower potential energy) have considerably less polarising ability
(compare I and II Figure 48) for activating ligands (aldehydes) than the corresponding fully
charged metal species (irrespective of which solvent molecules are associated with the cluster
or free metal cation respectively) Furthermore the metal in the cluster is shielded from the
ligand by the close association of anions30
OH
OH
M3+M OTfTfO
I Cluster (close ion pair) II Free ion
OTf M OTf M OTF
Quadruplet
3+
OTf M OTfTriplet
-
Examples of close ion pairs
Figure 48 Possible cluster formations versus free ion formation of metal triflate in organic
solvents
Even solvents with some coordinating properties (eg nitromethane and acetonitrile
preferred solvents for metal triflate catalysis) are poor solvents for most ionic compounds In
these cases there will remain a tendency to form clusters of close ion pairs (eg -+- +-+ +-+
+-+- +-+- etc)32
160
Furthermore solvents with a low dielectric constant have a low capacity for supporting
charge separation which is a necessary consequence of polarisation This is the reason for the
exceedingly large differences in rates of some SN2 reactions (Scheme 413) between neutral
species in different solvents31 This will also apply to the polarisation of aldehydes and the
likes thereof by Lewis acids
Et3N + CH3Iδ δ
Et3NMe IEt3N Me I Scheme 413 Typical SN2 reaction
Table 411 Rate of SN2 reaction in Scheme 413 in different solvents
Solvent Dielectric
constant (є) Rate (Ks )
Hexane 20 1
Chlorobenzene 56 1200
Acetonitrile 375 12000
In summation then the activation (polarisation) of aldehydes or related compounds by Lewis
acids in solvents of low dielectric constants and low coordinating ability can be expected to
be limited Secondly through the shielding effect of the counter ions of the clusters the rate
of exchange of the ligands (such as aldehydes) in the bulk solvent and the inner metal of the
clusters where interaction with the metals can take place will be slower
MSn OH
+ MSn-1
O
H
+ S eq 42
MnXm
P
OH
+
ClusterP = net charge
MnXm
PO
H
MnXm-1
P+1O
H
+ x-1
Aldehyde inbulk solvent eq 43
Scheme 414 Activation of the aldehyde by the Lewis acid in dry organic solvent
161
Thus according to Kobayashi the slow exchange seen in equation 42 (Scheme 414) accounts
for his results and the slow exchange seen in equation 43 (Scheme 414) would account for
the results seen in the present investigation However as is clear from the foregoing poorly
catalysed reactions are not necessarily due to slow exchange between the ligand and the
solvated metal The decrease in the reaction rate may not be due to a simple solvent
exchange but may in contrast be rationalised on the basis of solvent-induced cluster or tight
ion-pair formation
442 The possible role of water in organic solvents
One now has to consider the effects of the small amounts of water in the organic solutions of
the metal triflates Due to its high dielectric constant and significant coordinating ability to
metal ions it is expected that water would have a dramatic effect on the properties of the
solutions It will quickly result in the solution of the metal ions and dissociation of the
lsquoclustersrsquo into smaller units eventually to close ion pairs and then eventually to free solvated
ions
Kobayashirsquos results26 suggest that this occurs with the rare earth metal triflates at around 50
equivalents of water This assertion is based on the assumption that optimal catalytic activity
will be reached at the maximum degree of dissociation of the solute (triflate)
Along the way the Lewis acid capability of the metals is increased provided that rapid
exchange between the free aldehyde and the solvated metal can occur It is reasonable to
suggest that this rapid exchange will be possible with metals that can change their
coordination number over a wide range as is the case with the rare earth metals which have
coordination spheres of up to 12 ligands
One aspect of metal triflates which has to be addressed now is its possible source of Broslashnsted
acidity in protic solvents It is known that many higher valent metal species associate with
water or other protic solvents to generate Broslashnsted acids of varying acid strength (Scheme
415)34
162
M+n + H2O M(H2O)x
+n
M(H2O)x-1OH+n-1
+ H+
Broslashnsted acid
Kh
Scheme 415 Formation of Broslashnsted acidity through polarisation of water by a Lewis acid
These can be differentiated by different Kh values24 In a case where the Kh value is high the
question arises to what extent the metal salt will act as a Lewis acid and to what extent does
Broslashnsted acidity play a role (As has already been argued however one must exercise
caution when using an hydrolysis constant since amongst others the counter ion on the
metal plays an important role in the rate of hydrolysis) Other workers1112 claim that it is
possible to distinguish between these possibly by ascertaining the effect of an added hindered
base on the reaction rate
This attempt at rationalisation was aimed at indicating that Lewis acid acidity and catalytic
activity are complex phenomena that depend on many factors the presence of water in
particular
45 The case for 26-di-tert-butyl-4-methyl-pyridine From the results discussed earlier it is unlikely that the Mukaiyama aldol reaction is mainly
Broslashnsted acid catalysed since Broslashnsted acids so readily hydrolyse the enol ethers However
if the hindered base is added to the reaction medium one can expect that the Broslashnsted
acidityactivity will be suppressed allowing the reaction to proceed under Lewis acid
conditions
Therefore addition of the base to the reaction medium should have one of two effects 1) no
effect on the reaction where Lewis acids are almost exclusively present and 2) allow the aldol
reaction to proceed where previously the silyl ether was destroyed by Broslashnsted acidity
The Mukaiyama aldol reactions were therefore carried out in the presence of 26-di-tert-
butyl-4-methylpyridine (Table 412) Surprisingly the addition of the base generally resulted
in a rapid reduction of the reaction rate
163
Table 412 Yield () aldol reactions carried out in the presence of DTBMP
Lewis acid
Without
DTBMP
Yield ()
Aldol
15 eq
DTBMP
Yield ()
Aldol
50 eq
DTBMP
Yield ()
Aldol
La(OTf)3 92 0 0
Sc(OTf)3 96 79 50
InCl3 89 99 0
Cu(OTf)2 94 65 0
Clearly the effect of the hindered base cannot be ascribed simply to the action of a proton
scavenger As stated earlier other authors also neglected to take into account the possible
interaction of the base with the protic solvent (Scheme 416) The base (nucleophile)
generated in this equilibrium will undoubtedly deactivate the Lewis acid present
N+ ROH
NH
+ RO
Scheme 416 Interaction of DTBMP with protic solvent to form nucleophile
Solutions of THFH2O (91) and metal triflates (in the same ratio as was used in the aldol
reactions) were made up and the pH measurements taken after which 15 equivalents of the
hindered base DTBMP were added to and the pH measurements were again recorded The
results are summarised in Table 413 The aqueous solutions of THF and triflate exhibit
higher pH values when large excesses of the base are added (15 equivalents)
Table 413 pH measurements of Lewis acids in THFH2O with and without DTBMP
Lewis acid pH in
THFH2O (91)
pH in THFH2O
(91) and 15 eq
DTBMP
La(OTf)3 629 678 Al(OTf)3 294 357 Sc(OTf)3 212 327 Cu(OTf)2 443 506
164
These results (Table 413) indicate that because of the lower basicity of this sterically
hindered base (see Figure 41) it is difficult to suppress the Broslashnsted acidity completely using
15 equivalents thereof This apparently explains why Barrett and others were required to use
such large excesses of the base (up to 1000 equivalents in some cases) However the
possibility at these large excesses of base that the solution will become alkaline particularly
in the case of metals with low Kh values was not considered
46 Summary
So far the investigation strongly suggested that some of the metal triflates formed a type of
Broslashnsted acidity in the presence of water in organic solvents These triflates cannot be used in
the aldol reaction because of their rapid destruction of the silyl ether Other metal triflates
with relatively low Kh values however effectively catalyse the aldol reaction On the other
hand these metal triflates are not only water tolerant but are particularly water dependent for
their successful catalysis Specifically the present study on the Mukaiyama aldol reaction
showed that without a certain amount of water present which has been shown to form an
amount of Broslashnsted acid the reactions do not proceed in both coordinating (THF) and non-
coordinating (DCM) solvents
47 Carbocation formation
471 Carbocation formation in wet and dry solvents
The formation of retinylic carbocations in a number of solvents was studied by Blatz and
Pippert35 By using low temperatures (between -35 degC and -50 degC) and rapid handling they
were able to detect the carbocations of retinyl acetate and retinol in a number of solvent-
Broslashnsted acid systems Treatment of retinyl acetate or retinol with a Broslashnsted acid results in a
carbocation being formed this carbocation is a characteristic blue colour and can be easily
measured using UVVis spectrophotometry (Scheme 417) They found the λmax of the cation
to be solvent dependent
165
Retinyl acetate
O
O
H+O
OH
CH2+ HO
O
Acetic acidCarbocation
Scheme 417 Carbocation formation using retinyl acetate
Barrett et al12 then used the same protocol in his study of resorcinarenes In their study they
used retinol as the probe and found that [Yb(H2O)9(OTf)3] produced carbocations in THF
when AcOH and PhCO2H were added but not when resorcinol was added TfOH gave
carbocations and surprisingly so did [Yb(H2O)9(OTf)3] in MeNO2 in the absence of any
additives The conclusion drawn from the study was that the action of [Yb(H2O)9(OTf)3] on
the resorcinarene was the result of Broslashnsted acidity
To further investigate the role of water in metal triflate catalysed reactions and the possible
formation of Broslashnsted type acidity in the present study the reaction of retinyl acetate and
acid was explored This reaction was used in order to determine if solutions of the metal
triflates in organic solvents form a type of Broslashnsted acidity in the presence of water
In order to establish a working protocol experiments were initially carried out using Broslashnsted
acids A 22 x 10-5 M solution of retinyl acetate in ether was prepared and cooled to -50 degC in
an acetone dry ice bath to mimic Blatz and Pippertrsquos conditions35 The reactions are
performed at this temperature in order to prolong the lifetime of the carbocations35 At room
temperature the lifespan of the carbocation is so fleeting that spectrophotometric
measurements would be impossible35 In the present instance nitrogen was continuously
flushed across the optics of the spectrophotometer and the cells to eliminate condensation
The retinyl acetate was added first to the quartz cuvette and then the acid under investigation
was added A UVVis spectrum of the untreated solution of retinyl acetate shows its
characteristic absorption peak at around 289 nm (Figure 49) The Broslashnsted acids were then
added to fresh retinyl acetate solutions and their spectra were recorded The UVVis scans
were run over a period of time to ensure that the whole life-span of the carbocation was
166
recorded Carbocations were successfully formed using methanesulfonic acid triflic acid and
p-toluenesulfonic acid (Table 414)
Figure 49 UVVis scan of retinyl acetate and its corresponding carbocation showing the
typical wavelengths and colours of the solutions
Table 414 Variation of λmax and absorptivity of the retinyl carbocation
Acida Solvent λ max Absorption
Methanesulfonic acid Ether 604 0205
Triflic acid Ether 600 0491
p-Toluenesulfonic acid Ether 599 0151 a 50 uL of the acid were added to the 3 mL of retinyl acetate solution
The results show the characteristic wavelength of the retinylic carbocation which absorbs at
around 600 nm (Figure 411) Blatz35 showed the wavelength of the maximum absorption
(λmax) to be a slight function of the solvent it did not deviate more than a few nm to either
side of this wavelength
To determine if the metal triflates would yield carbocations on addition of the retinyl acetate
they had to be dissolved in a non protic solvent in order to eliminate any proton source The
metal triflates were found to be relatively soluble in nitrobenzene This was established after
much trial and error involving a large number of solvents
167
As a visual test Al(OTf)3 dissolved in nitrobenzene was added to a solution of retinyl acetate
which immediately turned blue Interestingly the carbocation was sustainable at room
temperature for several minutes This was in contrast to previous work which indicated that
low temperatures are imperative to the longevity of the carbocation This is possibly due to
the cation being stabilised by the solvating effects of the nitro groups (Figure 410)
N
NO
O
O
O
Figure 410 Solvation of carbocation by nitrobenzene
Encouraged by these results the same visual tests were carried out using In(OTf)3 Hf(OTf)4
Yb(OTf)3 and Sc(OTf)3 All of these metal triflates tested formed sustainable carbocations in
nitrobenzene from retinyl acetate at room temperature
In order to carry out the investigation in a more quantitative manner Al(OTf)3 was chosen
because of previous successes that had been realised in the present study with this metal
triflate
Karl Fischer titration showed that the nitrobenzene used thus far from the bottle (Aldrich
product) contained 352 ppm of water The aim of the experiments was to determine whether
Al(OTf)3 formed the carbocation through the formation of Broslashnsted acidity The experiments
were to be repeated in the presence of the sterically hindered base DTBMP as a test for
Broslashnsted acidity Additionally the solvent would be dried as much as possible in an attempt
to prevent the formation of water promoted Broslashnsted-type acidity by the presence of water
Stock solutions of the retinyl acetate Al(OTf)3 DTBMP and triflic acid were made up in
nitrobenzene 15 mL of the retinyl acetate solution were added to the quartz cuvette To this
was added the solution either of the Lewis acid or Broslashnsted acid (Table 415)
168
Table 415 Results of carbocation formation in nitrobenzene
Solution Additive C+ formation Absorption
Retinyl acetatea Al(OTf)3d Positive 0314
Retinyl acetateb TfOHe Positive 0810
Retinyl acetate +
DTBMPc TfOH Negative 0
Retinyl acetate +
DTBMPc Al(OTf)3 Negative 0
a Retinyl acetate solution 10 M b Retinyl acetate solution 0001 M c DTBMP solution 3 molar equivalents per acid added
dAl(OTf)3 solution 01M e TfOH solution 001 M
Carbocations were formed when the Al(OTf)3 or the triflic acid solutions were added to the
retinyl acetate stock solutions (Table 415) The intensity of the carbocation formed with the
triflic acid was much greater than that formed with Al(OTf)3 especially considering that the
solutions used for the triflic acid experiments are far more dilute than those used in the
Al(OTf)3 experiments (0001M versus 10 M retinyl acetate respectively) This is to be
expected if the formation of the carbocation from the Al(OTf)3 is due to Broslashnsted acidity by
hydrolysis
In the next series of experiments the sterically hindered base was added to the reaction and
mixtures no carbocations are formed with either the triflic acid or the Al(OTf)3 This strongly
suggests that the metal triflate is forming a Broslashnsted-type acid in the presence of water
Pleasingly unlike previous work that had been carried out using this hindered base12 only
three equivalents of DTBMP had to be added before the reaction was quenched
However to be absolutely sure that the carbocations were being formed due to Broslashnsted
acidity the next step was to remove the water and thus the source of the protons from the
solvent The nitrobenzene was dried by passing it through a column of activated alumina and
the water content was determined by Karl Fischer titration to be 3 ppm The stock solution of
retinyl acetate was made up to the same dilutions as before However it was found that the
Al(OTf)3 was now insoluble in the dried nitrobenzene Even after vigorous stirring and mild
heating the Al(OTf)3 powder remained at the bottom of the volumetric flask This was not the
only instance of solubility problems with metal triflates in non-polar in particular in dry
non- polar solvents
169
Little information could be found on the solubilising effect of small amounts of water on
metal triflates in non-polar organic solvents It is reasonable to assume in view of earlier
discussions on the solvation of ionic compounds in organic solvents that small amounts of
water through solvation of ions will increase the solubility significantly In this regard it
may be of importance to note that the following general observation was made in all of the
relevant experiments in this study addition of the functionalised substrate such as the
aldehyde to the non-polar organic solvents resulted in the dramatic increase in the solubility
of the metal triflates It is suggested that the solvationligation of ions (cations in particular)
by functionalised substrates may be responsible for this phenomenon
A series of other dry non protic solvents was evaluated for the purpose of carrying out this
reaction with Al(OTf)3 The metal triflate has a very limited solubility in non-coordinating
solvents making the choice of solvent very difficult However ionic liquids with a non-
coordinating counter ion were considered to be a potential solution to this problem For this
1-butyl-3-methylimidazolium triflate ([bmim][OTf]) was prepared (Figure 411)
N+ N
-OTf416
Figure 411 1-butyl-3-methylimidazolium triflate ([bmim][OTf])
The ionic liquid was prepared in the following way 1-chlorobutane and N-methylimidazole
were heated at 80 degC for 48 hours The resultant ionic liquid was then washed with ethyl
acetate to remove any unreacted starting material (the ionic liquid is immiscible with ethyl
acetate) The residual solvent in the ionic liquid was then removed under vacuum to yield 1-
butyl-3-methylimidazolium chloride ([bmim][Cl]) an ionic liquid An excess of LiOTf was
then added to this ionic liquid of [bmim][Cl] in water and the solution was allowed to stir for
24 hours in order for an ion exchange reaction to take place between the -Cl and the -OTf The
mixture was extracted with ethyl acetate and the residual solvent was removed under vacuum
Water and an excess of NaOTf was then added resulting in a biphasic system namely the
ionic liquid [bmim][OTf] and an aqueous solution of NaCl and excess NaOTf After 12
hours the aqueous layer was separated from the ionic liquid which was then dried under
vacuum at 80 degC for 72 hours
170
Karl Fischer titration of the [bmim][OTf] determined the water content to be 845 ppm water
The Al(OTf)3 readily dissolved in the ionic liquid However unexpectedly all attempts to
form carbocations in the ionic liquid failed The failure to generate Broslashnsted acidity in this
wet solvent may be due to the common ion effect in this case the triflate counter ions of the
ionic liquid (Scheme 418) which may suppress Broslashnsted acid formation by competing with
water molecules for coordination
Al(OTf)3 + H2O Al(OTf)2(OH) + OTf- + H+
BA OTf-
Al(OTf)3
BA =Al(OTf)3K
OTf-
α1
OTf-
K =
Scheme 418 Common ion effect on Al(OTf)3 in [bmim][OTf]
Al(OTf)3 was found to have some solubility in DCM A mixture of DCM and Al(OTf)3 was
allowed to stir at 35 degC overnight to generate a saturated solution after which it was allowed
to cool and the undissolved triflate settled to the bottom of the volumetric flask An aliquot of
the supernatant (5 mL) was measured out and the solvent removed under vacuum The
Al(OTf)3 that remained was weighed and it was found that 25 mg of Al(OTf)3 was soluble in
5 mL of DCM
Using this information stock solutions of Al(OTf)3 retinyl acetate DTBMP and triflic acid
were made up in DCM Karl Fischer titration determined the water content of the DCM from
the bottle to be 24 ppm UVVis spectrophotometry experiments were carried out as before
(Table 416)
171
Table 416 Results of carbocation formation in DCM
Solution Additive C+ Formation Absorption
Retinyl acetatea Al(OTf)3d Positive 174
Retinyl acetateb TfOHe Positive 317
Retinyl acetate +
DTBMPc TfOH Negative 0
Retinyl acetate +
DTBMPc Al(OTf)3 Negative 0
a Retinyl acetate solution 001 M b Retinyl acetate solution 0001 M c DTBMP solution 3 molar equivalents as per acid
added d Al(OTf)3 solution 001M e TfOH solution 0001 M
Carbocations were formed when Al(OTf)3 or triflic acid were added to the retinyl acetate
solutions As was the case with the nitrobenzene solutions the intensity of the cation formed
with the triflic acid was greater than that formed with the Al(OTf)3 (Figure 412) When
DTBMP was added to the solutions no carbocation formation is seen in either case
The DCM was dried by passing it over a column of alumina that had been activated in an
oven at 250 degC for 24 hours Karl Fischer titration was then carried out on the DCM and the
water content was found to be 2 ppm The corresponding stock solutions as previously were
made up To ensure that no atmospheric water found its way into the samples all work was
carried out in the glove box
172
Figure 412 UVVis scan showing the different intensities of carbocation formation with
triflic acid and Al(OTf)3 in DCM
The interesting shifts that can be seen in the λmax of the above UVVis scans may be the result
of the different counter ions formed in the reactions ie -OTf and Al(OTf)4-
Carbocations were formed at a similar intensity as before when the experiments are carried
out using triflic acid in the dry DCM When attempts were made to form carbocations in dry
DCM with Al(OTf)3 the solution turned a very faint blue a slight absorption peak can be
seen on the UVVis spectrum (Figure 413)
Figure 413 UVVis scan of Al(OTf)3 and retinyl acetate in dry DCM
173
At such a low concentration of water this result was unexpected as most of the water and
therefore also the source of Broslashnsted acidity had been removed from the system However a
DSC (differential scanning calorimetry) analysis of the Al(OTf)3 showed that the salt
contains a relatively large amount of water (Figure 414) The sample of Al(OTf)3 for that
analysis was made up in an inert atmosphere (glove box) and the scan was conducted under a
blanket of nitrogen The results of the scan showed one endotherm peak at a temperature of
170 degC and another at 260 degC (Figure 414) The lower temperature peak was assumed to
belong to lsquoloosely boundrsquo water and the higher temperature peak to that of water bound
directly to the metal centre
Figure 414 DSC scan of standard Al(OTf)3
This water along with the small amount of water left in the DCM may have been the source
of the Broslashnsted acidity that was promoting the weak carbocation formation that was seen in
the previous experiments A sample of the same Al(OTf)3 was then dried under reduced
pressure at 120 degC for 48 hours and the DSC scan was repeated Both of the endotherm
peaks had disappeared (Figure 415)
174
Figure 415 DSC scan of dried Al(OTf)3
To determine if it was in fact water that had been removed from the Al(OTf)3 sample and not
residual TfOH a small portion of the dried Al(OTf)3 powder was exposed to the atmosphere
for 15 minutes A DSC scan of this sample was then recorded The endotherm peaks reappear
at both 170 degC and 260 degC This strongly suggests that the endotherm peaks are as a result of
water bound to the Al(OTf)3
The carbocation formation experiment was repeated using the dried Al(OTf)3 in dried DCM
Stock solutions were made up in the glove box As before solubility was a problem and the
solution had to be heated to 35 degC before the Al(OTf)3 became completely soluble in the
solvent When the Al(OTf)3 solution was added to the retinyl acetate solution the solution did
not turn blue Nevertheless after some time Al(OTf)3 could be seen accumulating slowly on
the bottom of the cuvette Around the fine powder a blue colour could be seen forming on the
interface of the powder and the solvent
A possible explanation of this phenomenon is the irreversible hydrolysis on the crystal faces
of the Al(OTf)3 that occurs on exposure to moisture to yield amphoteric patches of
aluminium oxide on the surface33 This observation has been made for certain types of
alumina surfaces and may account for the present phenomena
175
472 The proton and the sterically hindered base ndash X-ray crystallography
In(OTf)3 and DTBMP were dissolved in DCM The DCM was then allowed to evaporate
slowly allowing crystals to form The crystals were then analysed using X-ray
crystallography (Figure 416)
Figure 416 Crystal structure of protonated DTBMP with OTf- counterion (417)
Table 417 Crystal data of protonated 26-di-tertbutyl-4-methyl pyridine
C20H20F3N2O3S Dx = 1607 Mg mminus3
Mr = 42544 F000 = 884
Orthorhombic Pna21 Mo Kα radiation λ = 071073 Aring
a = 228420 (16) Aring Cell parameters from 3551 reflections
b = 90680 (6) Aring θ = 24ndash280deg
c = 84873 (6) Aring micro = 024 mmminus1
V = 17580 (2) Aring3 T = 296 (2) K
Z = 4 041 times 022 times 019 mm
The crystal structure shows a pyridium ion with no metal found in the crystal structure and
presumably In(OTf)2(OH) is formed in the process This is consistent with a previous finding
176
of this investigation (see section 45 The case for 26-di-tert-butyl-4-methyl-pyridine) The
crystals formed in the presence of In(OTf)3 are identical to those formed when the same
experiment is carried out using triflic acid The latter experiment also generated crystals
identical to those described in Figure 416 and Table 417 above
48 Friedel-Crafts alkenylation reactions of arenes
481 Optimising the reaction
So far in the investigation it has been established that the metal triflates can form Broslashnsted
type acidity to varying degrees in the presence of water in organic solvents In the case of the
Mukaiyama aldol reaction this results in the hydrolysis of the silyl enol ether Furthermore
Broslashnsted acidity has been shown to be causative in the formation of carbocations using
retinyl acetate and a metal triflate An X-ray structure determination on crystals formed upon
the reaction of In(OTf)3 with the sterically hindered base DTBMP showed that a proton binds
to the base and that triflate is the counter ion
Since a metal triflate may exhibit both kinds of activity (Lewis and Broslashnsted acidity) it needs
to be established whether the Broslashnsted acid or the Lewis acid drives the reaction or whether it
is a combination of the two Alternatively the question may be asked as whether such a metal
triflate can act purely as a Lewis acid in the absence of water or protic solvent The Friedel-
Crafts alkenylation (Scheme 418) reaction of arenes was chosen for this part of the
investigation as it is a proton-neutral reaction Once the water is removed from the reaction
there is no other source of protons available for the generation of Broslashnsted acidity In this
way the extent of Lewis acid catalysis can possibly be determined
The reaction between p-xylene and phenylacetylene (Scheme 419) is known to be catalysed
by In Sc and Zr triflates36 This served as a starting point for the current investigation Using
the same experimental procedure set out in the 2000 communication36 a range of metal
triflates (20 mol) was used in the reaction between p-xylene and phenylacetylene
177
Ph HM(OTf)n 20 mol85 oC
Ph
H
H
+ +
Ph H
H
418 419 420 Scheme 419 Friedel-Crafts reaction of p-xylene with phenylacetylene
The reactions were carried out at 85 degC for 24 hours after which the yields were determined
by 1H NMR spectroscopy (Table 418) This was done by integration of the remaining
acetylene proton signal against the signal of the vinylic hydrogen in the product The yields
of the products were mostly poor many of the metal triflates failed to catalyse the reaction at
all (Table 418) but this may be due to solubility problems in the non-polar reaction medium
The problem was somewhat overcome by the addition of nitromethane to the p-xylene The
reactions were then repeated in this solvent mixture Several of the reactions were repeated
(Table 418) The yields of the products were generally if sometimes only slightly so
improved from the previous run In an attempt to try to further optimise the reactions those
metal triflates that had performed best were used in reactions where the amount of p-xylene
was systematically reduced (Table 419)
Table 418 Yield () of Friedel-Crafts alkenylation reactions catalysed by various M(OTf)x
Lewis acid
Reaction Yield ()a
Reactions +200 uL
nitromethaneYield ()a
Zr(OTf)4 53 58 Al(OTf)3 50 86 Cu(OTf)2 0 - Ca(OTf)2 0 - Hf(OTf)4 63 64 Zn(OTf)2 0 0 La(OTf)3 0 - Sc(OTf)3 68 100 Sm(OTf)3 0 0 Y(OTf)3 0 0
ScCl3 0 21 InCl3 50 53 TfOH 31 31
a Yields determined by 1H NMR spectroscopy
178
By decreasing the volume of p-xylene used in the reaction mixture the yields of the product
were greatly improved The results are summarised in Table 419
Table 419 Yield () of Friedel-Crafts alkenylation reactions in various amounts of
p-xylenea
Metal triflate
Yield ()b 8 mL p-xylene
Yield ()b
4 mL p-xylene
Yield ()b
2 mL p-xylene
Zr(OTf)4 53 68 100 Al(OTf)3 100 100 100 Sc(OTf)3 100 100 100 Hf(OTf)4 76 100 100
a p-xylene phenylacetylene (100 uL) nitromethane (200 uL) 20 mol M(OTf)x b Yield determined by 1H NMR
spectroscopy
The application of metal triflates in the Friedel-Crafts alkenylation reaction is expected to
have a wide application For example the study also showed that phenyl acetylene could be
successfully reacted with a wide range of aromatic systems including toluene anisole etc
using the same metal triflates (Table 420)
Table 420 Yield () of Friedel-Crafts alkenylation reactions with alternative aromatic
systemsa
Lewis Acid 10 mol
Yield ()b
Cumene 16 mL 48 h
Yield ()b
Anisole 16 mL 24 h
Yield ()b
Toluene 16 mL 48 h
Zr(OTf)4 71 gt 95 66 Al(OTf)3 47 gt95 77
a p-xylene phenylacetylene (100 uL) nitromethane (200 uL) 20 mol M(OTf)x b Yields determined by 1H NMR
spectroscopy products not isolated
482 Reactions in dry solvent
Once the optimal reaction conditions had been established the p-xylene and nitromethane
were dried Karl Fischer titration was carried out on the solvents to determine their water
content before and after drying Nitromethane from the bottle was found to contain 325 ppm
water Working in the glove box the solvent was passed through a column of activated
179
alumina and the dry nitromethane was found to contain 22 ppm water The p-xylene was
dried for 24 hours over 3Aring molecular sieves that had been activated in an oven at 250 degC The
dried p-xylene was found to contain 1 ppm water When the solvents were mixed in the same
ratio as they were used in the previous reaction mixture the Karl Fischer titration was
repeated on the solvent mixture and the water content was found to be 5 ppm This mixture
was then used for the reactions
The metal triflates were dried under high vacuum at 120 degC for 48 hours to remove all traces
of water DSC scans were carried out to ensure and confirm that the all of the metal triflates
were dry Additionally all preparation work took place in a glove box The scans showed no
endotherm peaks that are characteristic of the presence of water
The Friedel-Crafts alkenylation reactions were then repeated (Table 421) using the dry
solvents in order to determine to what extent Broslashnsted acidity plays a role in these reactions
Since for all intents and purposes the water had been removed from these reactions the
possibility of generating Broslashnsted acidity had also been eliminated
Table 421 Friedel-Crafts alkenylation reaction in dry solventa
Metal Triflate
Yield ()b Solvent from
bottle
Yield ()b Dry solvent
Zr(OTf)4 68 24 Al(OTf)3 100 100 Sc(OTf)3 100 74 Hf(OTf)4 100 35
TfOH 31 21 a 4 mL p-xylene 20 mol M(OTf)x 85 degC 24 h b Yields determined by 1H NMR spectroscopy
Table 421 shows that yield of the reactions decreases moderately to significantly when they
were carried out in dry medium except in the case of Al(OTf)3 The results indicate that the
reactions can be sustained in a thoroughly dried solvent and are in this case very probably
Lewis acid promoted However the higher activity in slightly wetter solvents could be due to
several effects including increased solubility andor solvation of ions resulting in improved
ionic dissociation and exchange of the metal triflates (solvation effects) 1H and 13C NMR
spectroscopy of phenyl acetylene in deuterated DCM suggests that Al(OTf)3 does bind to the
triple bond of the phenyl acetylene Complexation results in a clear downfield shift of the
180
acetylic hydrogen and triple bond carbons (from 531 ppm to 528 ppm in the proton
spectrum and from 838 ppm to 839 ppm in the 13C spectrum) The possibility of increased
activity due to the formation of a protic acid from water binding to the metal triflate is a
realistic possibility The phenomenon of increased catalytic activity of metal triflates in the
presence of water has been observed throughout this investigation
It is clear that Al(OTf)3 is a very active catalyst for the Friedel-Crafts alkenylation reaction
under investigation Reactions were performed under dry conditions using smaller amounts of
catalyst Only at a catalyst loading of 5 mol was a decrease in reactivity observed (ie 10
mol catalyst led to quantitative conversion to product) In this case the yield of the reaction
was 60
Despite the generally lower yields obtained in the Friedel-Crafts alkenylation reaction in dry
organic medium it appears as if this particular reaction is indeed primarily Lewis acid
catalysed in the case of Al(OTf)3 (and possibly for the other metal triflates used in this study-
although a large contribution from a Broslashnsted-acid catalysed mechanism may be the force
with those metal triflates that were severely affected by the drying ie Zr(OTf)4 and
Hf(OTf)4)
The effect of the lower water content on the triflic acid can be explained in terms of
diminished dissociation in a solvent with lower dielectric constant and poor solvating
properties There seems to be no simplistic trend with regards to water on the metal triflates
This may be due to the dual mechanism and the unpredictable reactivities and quantities of
the given Lewis acid and Broslashnsted acid that forms
In cases where metal triflates were not completely soluble in the reaction medium but some
portion remained as solid particles the contribution of a heterogeneous component to the
reaction cannot be excluded This possibility has not been investigated but should command
attention
181
49 Conclusions
Summation of results described in publications and new results outlined in this investigation
led to the conclusion that the presence of water (or other protic molecules) in organic solvents
can affect the catalytic activity of the metal triflates in different ways Not only can it
increase solubility but catalytic activity can be increased by solvation water complexation
while results in the formation of Broslashnsted acid activity The effect of water and other protic
solvents will generally not be easy to determine to predict or be ascribed to a specific factor
The dramatic effect of small amounts of water on the catalytic ability of metal triflates raises
the question of the effect of water on Lewis acid activity in general and as to the nature of the
nature of the active catalyst In the minds of most practising chemists Lewis acid catalysis
appears to play out as the simple activation of a substrate by a metal centre This study has
amply demonstrated that this is not the case Instead the reality appears to be one in which
water plays a critical if sometimes determinative role in the successful outcome of the
reaction In all likelihood many (if not most) reactions that are held to be purely Lewis acid
catalysed are either Broslashnsted acid catalysed (by complex Broslashnsted acids of the type MXnmdash
OH2) or co-catalysed by Broslashnsted-Lewis synergism in which hydrogen bonding and metal
bonding where a MmdashOH2 moiety lead to favourable transition states
Throughout all of the work of the present study water has shown to play a critical role In
only one case was this not so The study clearly points to the complexity faced when
considering Lewis acid catalysis at a fundamental level as has been done here It is quite
likely given the manifold reactions investigated here and the complex interplay between
Lewis and Broslashnsted acidity (the latter being almost ubiquitous in the presence of Lewis
acids) that the fundamental way in which chemists think of such activators should be
modified
Comparing the catalytic activity of metal triflates becomes particularly problematic when
water (or a protic solvent) is a potential reagent (eg in the Mukaiyama aldol reaction) Such
comparisons should preferably be carried out with model reactions that are inert to water
The results described suggest that metal triflates in water-containing solvents often catalyse
reactions by a dual mechanism (Lewis andor Broslashnsted acid mediated) and that the relative
182
importance of these two mechanisms differ from metal to metal The results further indicate
that the interpretation of the effect of the addition of the sterically hindered base to a reaction
medium should be interpreted with caution particularly where large excesses of the base are
added
An observation of particular importance is that some metal triflates are not only tolerant to
water but require water for their catalytic activity The sometimes dramatic effect of drying
the organic solvent on the metal triflate catalytic activity highlights the role of small amounts
of water in organic reactions in general
In turn this point focuses attention as to what is meant by using what organic chemists usually
term dry solvents The previous chapter highlighted the difficulties in drying organic solvents
and serves as a relief for the present work
183
410 References
1 Kobayashi S Sugiura M Kitagawa H Lam W W L Chem Rev 2002 102
2227
2 Scifinder Scholar search of ldquoMetal Triflate Catalysisrdquo 2002-2009 ndash 307 hits
3 a) Kobayashi S Synlett 1994 9 689 b) Kobayashi S Chem Lett 1991 12 2187
c) Kobayashi S Ogawa C Chem Eur J 2006 12 5954 d) Keller E Feringa B
L Tetrahedron Lett 1996 37 1879
4 Williams D B G Lawton M Org Biomol Chem 2005 3 3269
5 Bizier P N Atkins S R Helland L C Colvin S F Twitchell J R Cloninger
M J Carb Res 2008 343 2814
6 Noji M Ohno T Fuji K Futaba N Tajima H Ishii K J Org Chem 2003
68 9340
7 Ollevier T Lavie-Compin G Tetrahedron Lett 2004 45 49
8 Brown H C Kanner B J Am Chem Soc 1966 88 986
9 Ollevier T Nadeau E Guay-Beacutegin A-A Tetrahedron Lett 2006 47 5351
10 Dumeunier R Markoacute I E Tetrahedron Lett 2004 45 825
11 Bizier P N Atkins S R Helland L C Colvin S F Twitchell J R Cloninger
M J Carb Res 2008 343 2814
12 Barrett A G M Braddock D C Henschke J P Walker E R J Chem Soc
Perkin Trans 1999 873
13 Curtis A D M Tetrahedron Lett 1997 38 4295
14 Pieroni O L Rodriquez N M Vuano B M Cabaleiro M C J Chem Res (S)
1994 188
15 Waller F J Barrett A G M Braddock D C Ramprasad D Tetrahedron Lett
1998 39 1641
16 Waller F J Barrett A G M Braddock D C Ramprasad D Chem Commun
1997 613
17 Barrett A G M Braddock D C Chem Commun 1997 351
18 Claydon J Greeves N Warren S Wothers P Organic Chemistry Oxford
University Press New York 2001
19 Smith M B March J Advanced Organic Chemistry Reactions Mechanisms and
Structure 5th ed Wiley New York 2001
20 Mukaiyama T Pure Appl Chem 1983 55 1749
184
185
21 Loh T-P Pei J Cao G-Q Chem Commun 1996 1819
22 Van de weghe P Collin J Tetrahedron Lett 1993 34 3881
23 Hollis T K Bosnich B J Am Chem Soc 1995 117
24 Kobayashi S Nagayama S Busujima T J Am Chem Soc 1998 120 8287
25 Baes C F Jr Mesmer R The Hydrolysis of Cations Wiley New York 1976
26 Hagger J A Depledge M H Galloway T S Marine Pollution Bulletin 2005 51
811
27 a) Kobayashi S Synlett 1994 9 689 b) Ding R Katebzadeh K Roman L
Bergquist K E Lindstrm U M J Org Chem 2006 71 352 c) Kobayashi S
Manabe K Acc Chem Res 2002 35 209 d) Dissanayake P Allen M J J Am
Chem Soc 2008 131 6342
28 Chaminade X Chiba Shunsuke C Narasaka K Duntildeach E Tetrahedron Lett
2008 49 2384
29 Drago R S Physical Methods in Chemistry Saunders 1976
30 Gould E S Mechanism and Structure in Organic Chemistry Holt Reinhart and
Winston 1959
31 Purcell K F Kotz J C Inorganic Chemistry Sauders 1977
32 Gladstone S Textbook of Physical Chemistry Macmillen 1953
33 Isaacs N S Physical Organic Chemistry Longman 1987
34 Smith M B Organic Synthesis McGraw-Hill Singapore 1994
35 Blatz P E Pippert D L J Am Chem Soc 1967 90 1296
36 Tsuchimoto T Maeda T Shirakawa E Kawakami Y Chem Commun 2000
1573
37 Carruthers W Coldham I Modern Methods Inorganic Synthesis Cambridge
University Press UK 2004
Chapter 5
Ranking of Lewis acids
51 Introduction
The final aim of this investigation was to rank the metal triflates according to their Lewis
acid strength using spectroscopic methods This type of ranking had proved marginally
successful for other more traditional types of Lewis acids such as the metal halogens
However to our knowledge a study like this has not been undertaken for the metal
triflates
Despite attempts by many researchers the quantitative measurements of Lewis acid
strength across a broad range does not exist The quantitative measurement of Lewis
acidity appears to be one of the persistent problems of the acid-base theory Lewis
himself pointed out that relative acidity (or basicity) would depend on the choice of
reference base (or acid)
The most reliable method for determining the strength of a Lewis acid would be the
determination of the enthalpy change accompanying the formation of the acid-base
adduct in the gas phase1 This method also has its drawbacks it is not available to a wide
range of compounds and although it tells us the acidity in the gas phase the question
arises as to whether the data could be extrapolated to the solvent phase
The hard-soft acid-base (HSAB) concept was introduced in 1963 by Pearson2 and can
explain affinities between acids and bases that do not depend on electronegativity and
other related properties3 According to this principle hard acids prefer to bond to hard
bases and soft acids prefer to bond to soft bases Electrostatic interaction is presumed to
be the dominant source of stabilisation in the hard acid-hard base complex In the case of
soft acid-soft base complexes electron delocalisation between the frontier orbitals has
been thought to be the principal interaction4
186
The HSAB principles give us a good qualitative indication upon which to work
However we are unable to determine anything about the inherent strength of the acid or
base
Nevertheless there have been many successes in correlating relative Lewis acid strength
using an array of techniques (for a full review see Chapter 1) Childs carried out a study
using NMR spectroscopy by examining the shifts of complexed bases versus the
uncomplexed bases4 He was able to determine the Lewis acidity of a variety of acids
Other studies have been carried out using UVVis spectrophotometry to determine Lewis
acidity eg by the difference in the wavelength of complexed and uncomplexed carbonyl
groups Often spectroscopic information is applied in calculating equilibrium constants
which then forms the basis for ranking of the Lewis acids
The aim of the present investigation was to establish a ranking for a variety of metal
triflates with respect to Lewis acidity using NMR IR and UVVis spectroscopy and to
determine if the ranking found by these methods correlated with each other
52 Lewis acidity from NMR resonance shifts
Crotonaldehyde
The ranking of Lewis acids by NMR spectroscopy is based on the assumption that when
the Lewis acid (electron acceptor) binds to the electron donor (Lewis base) there is a
reduction of electron density on the Lewis base This reduction of electron density results
in a downfield shift in the NMR signals of the basic compound The stronger the Lewis
acid the more dramatic the shift on the NMR spectra These shifts can then be compared
to one another and a ranking obtained
One of the most comprehensive investigations carried out on the ranking of Lewis acids
using NMR spectroscopy was done by Childs et al4 In this investigation metal halides
were used as the Lewis acids The most successful probe (base) used in the investigation
187
was crotonaldehyde although others were also employed (this study is outlined more fully
in Chapter 1 section 132 of this thesis)
Childsrsquos study formed the starting point of the current investigation While investigating
the role of water in metal triflate catalysis (Chapter 4) it was found that the triflates had
some solubility in DCM and that this solubility increased when an aldehyde or other
functionalised organic compound was added to the mixture In light of these findings the
current NMR investigation was carried in deuterated DCM using (asymp 01 M)
crotonaldehyde (for numbering see Figures 51 and 53) as the probe An excess of 12
equivalents (with respect to the aldehyde) of the Lewis acid was used to ensure that all of
the aldehyde was coordinated to the metal The 1H and 13C NMR spectra in all instances
showed only one set of signals indicating complete (within the limits of NMR sensitivity)
coordination to the metal The 1H and 13C NMR results with respect to several Lewis
acids are recorded in Table 51 and 52 respectively
H
H3C
OH
H3
12
Figure 51 Proton numbering used on crotonaldehyde
Table 51 1H NMR chemical shift differences (Δδ) of crotonaldehyde on complexation
with various Lewis acids
NMR
signala
Croton
aldehyde
δ
Δ δ on
addition
of
Al(OTf)3
Δ δ on
addition
of
AlCl3
Δ δ on
addition
of
Sc(OTf)3
Δ δ on
addition
of
ScCl3
Δ δ on
addition
of
In(OTf)3
Δ δ on
addition
of
InCl3
H-1 947 -026 -016 -025 003 -006 -007
H-2 610 054 013 044 045 030 027
H-3 687 093 023 093 071 050 043
CH3 201 029 012 028 021 016 013 a Negative values indicate an upfield shift
188
The results show significant shifts for both H-3 and H-2 resonance in all cases (Table
51) Furthermore Al(OTf)3 and Sc(OTf)3 effect the greatest of these shifts particularly
with regard to H-3 The same trend is seen by the CH3 group However the shifts are of
smaller magnitude In every case the shifts are mutually consistent and can readily be
interpreted in terms of the relative Lewis acidity of the metals The following ranking is
therefore suggested Al(OTf)3 gt Sc(OTf)3 gt In(OTf)3 And for the chloride series ScCl3 gt
InCl3 gt AlCl3 (H-1 shifts did not correlate with the above suggested Lewis acid ranking
this may be due to anisotropic shielding induced by the oxygen-metal bond on the nearby
H-1 hydrogen) The order found for the chloride series appears to be anomalous with
respect to AlCl3 Other workers have found that AlCl3 gt InCl356 However probes
(bases) used in these studies were different In one case ethyl acetate was used5 and in
the other 9-fluorenone6 which could be a possible reason for the difference in the
rankings Childs4 did not use AlCl3 or InCl3 in his study
Figure 52 1H NMR chemical shift differences of crotonaldehyde versus the various
Lewis acids
189
H
O
3
12
4 Figure 53 Carbon numbering used on crotonaldehyde
Table 52 13C NMR chemical shift differences (Δδ) of crotonaldehyde on complexation
with various Lewis acids
NMR
signala
Croton
aldehyde
δ
Δ δ on
addition
of
Al(OTf)3
Δ δ on
addition
of
AlCl3
Δ δ on
addition
of
Sc(OTf)3
Δ δ on
addition
of
ScCl3
Δ δ on
addition
of
In(OTf)3
Δ δ on
addition
of
InCl3
C-1 1941 111 04 116 175 04 29
C-2 1348 -20 05 -14 -15 03 -02
C-3 1544 219 02 221 1752 04 45
C-4 187 26 06 27 24 05 09 a Negative values indicate an upfield shift
The significant shifts seen in C-1 and C-3 following the same reasoning as before these
shifts appear to support Lewis acid acidity ranking Al(OTf)3 asymp Sc(OTf)3 gt In(OTf)3 The
chloride series remains the same as before
The measure of consistency found in this method strongly suggests that this may be a
valuable method for the ranking of Lewis acids particularly since the large number of
aldehydes available that will allow a great measure of fine tuning This is a subject of an
ongoing study in our laboratory
190
Figure 54 13C chemical shift differences of crotonaldehyde versus the various Lewis
acids
Ionic liquids as a solvent for NMR spectroscopy
Work carried out previously in this investigation showed that the metal triflates were
soluble in ionic liquids in particular [bmim][OTf] (Chapter 4) In order to extend the
current study on the ranking of the metal triflates using NMR spectroscopy it was
decided to employ the use of ionic liquids as the solvent To do this a 10 mm NMR tube
was used in which the ionic liquid along with the aldehyde and the metal triflate was
placed A coaxial tube filled with deuterated benzene was then inserted into the 10 mm
NMR tube containing the ionic liquid (deuterated benzene was used due to its high
deuterium content when other deuterated solvents were used for these experiments for
example CDCl3 or CD3OD a lock could not be obtained on the NMR spectrometer)
Trans-cinnamaldehyde was used as the probe in these experiments (Figure 55 and Table
53)
191
C3C2 C1
H1
OH3
H2 Figure 55 Atom numbering on trans-cinnamaldehyde
Table 53 1H and13C chemical shift differences (Δδ) of trans-cinnamaldehyde in
[bmim][OTf] on complexation with various metal triflates
NMR signala
Trans-
cinnamaldehyde
δ ppm
H-1
969
H-2
669
C-1
1948
C-2
1291
C-3
1537
Lewis acid Δδ H-1 Δδ H-2 Δδ C-1 Δδ C-2 Δδ C-3
Hf(OTf)4 003 005 13 -02 12
Sc(OTf)3 -016 -002 24 01 25
In(OTf)3 -0171 0047 17 00 16
Ca(OTf)2 -0074 0014 04 -04 02
Zn(OTf)2 -005 0038 29 03 26
Y(OTf)3 -0048 0145 31 03 33
Zr(OTf)4 -0107 004 23 -07 24
Al(OTf)3 -0076 0032 09 00 09
LiOTf -0017 0055 10 01 07 a Negative values indicate an upfield shift
The outcomes of these experiments show a strong correlation between the shifts in the
resonance of C-1 and C-3 (Figure 56) The shifts observed for C-2 appear to be random
There is no relationship between the data obtained for H-1 and H-2 and the results could
also not be linked to the resonance shifts observed in the carbon spectra No shifts were
seen for H-3 Due to the lack of correlations any ranking obtained from this method
would be inconclusive at best
192
Figure 56 13C chemical shift differences of trans-cinnamaldehyde versus the various
Lewis acids
Phosphorus NMR
Methoxycarbonylation reactions using phosphorus ligands are one of the focuses of
work in our laboratories Recently it was discovered that a metal triflate can co-catalyse
these reactions where previously they had been exclusively Broslashnsted acid catalysed6 In
light of these findings and in order to find out more about how the metal triflates worked
in these reactions phosphorus probes were used in an NMR study in an attempt to rank
the Lewis acidity of the metal triflates
Spencer et al7 conducted a study into the ranking of Lewis acids using 31P NMR
spectroscopy in which triphenylphosphine oxide was used as the probe along with
calorimetric techniques to determine the enthalpy changes and the equilibrium constants
The Lewis acids under investigation were trimethylchlorosilane -germane and -stannane
193
The workers found little correlation between the 31P shifts recorded and the
thermodynamic data
In the current study the NMR experiments were initially carried out by a colleague using
deuterated methanol This solvent was chosen as it mimics the reaction conditions of the
methoxycarbonylation reaction which was the reaction under investigation The probe
used was triphenylphosphine and a variety of metal triflates were used 8
The results showed that there was no shift in the 31P NMR chemical shifts on addition of
any of the Lewis acids which may be due to two factors Firstly the metal triflates are
relatively hard Lewis acids and the probe being used ie triphenylphosphine is a
comparatively soft Lewis base which would account for the absence of interaction
between the two Secondly the deuterated solvent is methanol which is itself a
coordinating solvent The metal ions of the triflate salts may preferentially coordinate to
the harder oxygen atom of the alcohol over the softer phosphorus atom
To establish if any coordination with phosphorus was possible the 31P NMR experiments
were repeated with several of the metal triflates in deuterated DCM a non-coordinating
solvent The phosphorus probes were also extended to include triphenylphosphine oxide
and diphenylphosphinobenzaldehyde The results are summarised in Table 54
Table 54 31P NMR chemical shift differences (Δδ) of phosphorus compounds on
complexation with various metal triflates
Phosphorus compound
δ uncomplexeda
Δ δ on addition
of Al(OTf)3
Δ δ on addition
of Sc(OTf)3
Δ δ on addition
of In(OTf)3
Triphenyl phosphine -511 071 1072 1006
Triphenyl phosphine oxide 2778 1195 1939 1749
Diphenylphosphino benzaldehyde -1044 4399 4399 4398
a Referenced to 85 phosphoric acid in water using a coaxial tube insert
194
Gratifyingly complexation of the metal triflates to the phosphorus centre was seen in each
case causing large downfield shifts in the resonance of the 31P signals A linear
relationship exists between the shifts found for triphenylphosphine and those of
triphenylphosphine oxide ie Sc(OTf)3 gt In(OTf)3 gt Al(OTf)3 The order of this series
differs from the order found with crotonaldehyde
Interestingly when diphenylphosphinobenzaldehyde is used as the probe the change seen
in the resonance of the 31P NMR signals are identical for the three metal triflates used
(ie 439 ppm) No precedence for these phenomena could be found in the literature
However a possible explanation could be the formation of a chelate (Figure 57) between
the phosphorus atom and the oxygen atom on the aldehyde to the metal of the triflate
This chelate would form a stable six membered ring and in so doing the phosphorus
would take on a formal positive charge Presumably the primary binding in the structure
is between the harder oxygen atom and the hard metal centres The secondary binding to
the phosphorus atom is rendered advantageous by virtue of the chelate structure and
leads to the observed similarities This would account for the fact that all of the metal
triflates used in the study appear to withdraw electrons at the same rate from this Lewis
base The strong downfield chemical shift is typical of metal bound P (III) atoms
P H
O
(OTf)3M
Figure 57 Diphenylphosphinobenzaldehyde chelated to a metal triflate
What this work highlights is the need to exercise caution when using this method of
ranking It should be recommended that this method leads to relative ranking of the
Lewis acids that is relative to the probe (base) being used at the time rather than an
absolute method of ranking This is because each base has its own unique electronic
characteristics and will be affected by the Lewis acid in different ways This is unlike the
scale used for Broslashnsted acidity which uses proton acidity as a common feature
195
Equilibrium constants by NMR
NMR spectroscopy has been applied to determine the equilibrium constants (Keq) of
Lewis acid-base adducts and from this information a ranking of Lewis acidities should be
possible A 11 stoichiometry of the acid-base is generally assumed1 This method can of
course only be applied when there is slow exchange between the bound and unbound
ligand In this case it is assumed that the resonance shifts on the 1H NMR spectra of
unbound-base versus bound-base can be integrated and the respective concentrations
determined Here Keq = [acid-base complex][acid][base] This method assumes that the 1H NMR integral for the signal of a CH proton of a complexed molecule of the base in
question is directly proportional to the mole fraction of that species and may be related as
a proportional mole fraction to the integral on the same CH signal of the free base A
successful study of this kind was carried out by Branch et al1 where 9-fluorenone was
used as the probe and it was found that the ranking obtained from the NMR study could
be correlated to some other thermodynamic data obtained
In the current study crotonaldehyde was used as a probe in deuterated DCM Mixtures of
a 11 ratio of the aldehyde to metal triflate were carefully weighed out on a five decimal
balance Each solution for NMR was made up in 08 mL of deuterated DCM The
mixtures were stirred for 30 minutes at room temperature before the NMR spectra was
taken All spectra were recorded at 25 degC several metal triflates were investigates giving
comparable results therefore Table 56 contains results of only two of these triflates
namely Sc(OTf)3 and In(OTf)3
The results (Table 55) of the 11 12 14 etc ratio of metal to ligand were surprising in
that none of the spectra showed unbound crotonaldehyde Therefore clearly rapid
equilibrium between the bound and unbound aldehyde is established resulting in
weighted averaging of the signals9
In this case the following equation applies
δiave = (1-Nc) δic + Nc x δif
196
Where
δiave = observed signal for nucleus i in NMR spectrum
δif = signal of nucleus i of free base (ligand)
δic = signal of nucleus i of complexed base (ligand)
Nc = mol fraction of complexed base (ligand)
Table 55 1H chemical shift differences (Δδ) of crotonaldehyde on complexation with
various metal triflates
NMR signal
Croton aldehyde
δ
Δ δ on addition
of Sc(OTf)3
11
Δ δ on addition
of Sc(OTf)3
12
Δ δ on addition
of Sc(OTf)3
14
Δ δ on addition
of Sc(OTf)3
18 aH-1 947 -027 -022 -015 -008
H-2 687 045 014 026 014
H-3 610 100 083 057 024
CH3 201 028 023 014 005
NMR signal
Croton aldehyde
δ
Δ δ on addition
of In(OTf)3
11
Δ δ on addition
of In(OTf)3
12
Δ δ on addition
of In(OTf)3
14
Δ δ on addition
of InOTf)3
18 aH-1 947 -009 -008 -007 -006
H-2 687 027 019 010 008
H-3 610 049 037 021 017
CH3 201 014 007 004 003 a negative values indicate an upfield shift
In this case in principle the equilibrium constants should still be obtainable by calculation
provided that the equilibrium is not so for to the right ie to the complex that the amount
of free metal and of ligand cannot be accurately measured The linear relationship
between the signals H3 and CH3 of the complexed crotonaldehyde strongly suggest
almost quantitative complexation of both Sc(OTf)3 and In(OTf)3 in all cases The similar
results with other triflates therefore does not allow the ranking of the Lewis acidity in
197
this solvent A ranking using the NMR method would require the use of a more polar and
coordinating solvent to decrease the affinity of the ligand for the metal but then again the
ranking will be critically dependent on the chose of the solvent and will change from
solvent to solvent
Equilibrium constants UVVis
The calculation of equilibrium constants of acid-base adducts has been carried out using
data obtained by UVVis measurements10 In spectroscopic methods concentration is
directly proportional to absorptivity according to Beerrsquos law (Equation 1)
A = εbc 1
Where A = absorbance
ε = molar absorptivity
b = cell length (cm)
c = concentration (molL)
It has been found that the addition of the Lewis acid to specific types of Lewis bases
leads to a decreases in intensity of the absorption band of the free base and a new band
characteristic of the adduct usually appearing at a wavelength longer than that of the base
(Figure 58)9
198
Figure 58 UVVis spectrum of adduct formation between Lewis base and Lewis acid
Thus by using known concentrations of base and adding known concentrations of the
Lewis acid it is possible to determine the strength of a Lewis acid This is done by
calculating the Keq in the following way Keq = [acid-base complex][acid][base] The
concentrations are calculated from the absorptivity taken from the UVVis spectra9
In an attempt to rank metal triflates using this method a dilute solution (10-5 M) of 4-
methyl-3-nitroaniline was prepared in DCM and solutions of metal triflates of equal
molarity were also prepared The UVVis absorption profile of the free base was recorded
(Figure 59)
The absorption maximum of the base is seen at 424 nm When the solution of Al(OTf)3
was added no new absorption band was seen in fact no change in the spectrum was seen
at all (Figure 59) This behaviour was also observed when Sc(OTf)3 and In(OTf)3
solutions were added to the probe
199
Figure 59 UVVis spectrum of 4-methyl-3-aniline in DCM
Dilute solutions of 2-nitrodiphenylamine and 4-nitrodiphenylamine in DCM were
investigated as alternative probes The UVVis scans were carried out as before
However no complexation of the Lewis acid to the probe could be detected The lack of
coordination seen in these experiments may be due to the metal triflates forming close ion
pairs in the DCM and thus not been available to bind to the nitrogen This would be
particularly true due to the very low concentration of the probe (low in order to record a
UVVis spectrum of this chromophore) which would favour dissociation of a weak
complex
In an attempt to circumvent this problem dilute solutions of the same probes and the
metal triflates were made up in THF This solvent has better solvating capabilities than
DCM which may make metal ions more available for coordination to the nitrogen The
UVVis scans were then repeated These experiments showed a slight attenuation of the
original band (Figure 510 and Figure 511) when the Lewis acid solutions were added to
the probes
Satchell and Wardell10 observed the same phenomena in their work where on addition of
the Lewis acid the absorption band attributed to the base was simply reduced and no new
band characteristic of an acid-base adduct appeared According to these authors this was
200
ascribed to protonation of the aniline and the resultant anilinium ion absorbing at a much
shorter wavelength (bathochromic shift end absorption) than the parent base11
In the current study many other probes were evaluated (Table 56) and similar results
were obtained in each case ie attenuation of the original band
Figure 510 UVVis spectrum of Figure 511 UVVis spectrum of 4-
4-nitrodiphenylaniline in THF nitrodiphenylaniline coordinated to
Lewis Acid
Table 56 Δλ of probe on addition of a Lewis acid
Lewis base used Solvent Result
Pyridine THF Attenuation of original peak
2-Bromoanline THF Attenuation of original peak
4-Chloroaniline Ether No observable result
Diphenylaniline THF Attenuation of original peak
4-Nitroacetophenone Ether Attenuation of original peak
As a result equilibrium constants could not be calculated The possibility of using the
extent of attenuation of Lewis acid acidity is presently under investigation in our
laboratory
201
53 Infrared Spectroscopy (IR)
Given that only marginal success was found using NMR-based and UVVis approaches
to the ranking of the metal triflates according to equilibrium constants altogether another
method was sought
The ranking of Lewis acidity using IR spectroscopy has been reported5 When a Lewis
acid binds to a Lewis base such as the oxygen of the carbonyl group perturbation of this
bond occurs The strength of the donor-acceptor bond is reflected in the extent of the
weakening of the C=O bond This can be measured by the change in the bond stretching
frequency (Δν) on IR Lappert5 carried out a study in which ethyl acetate was used as a
probe and boron halides as the Lewis acids He found good correlation of his results with
other studies (as described in detail in Chapter 1 section 132)
Pyridine was used as an infrared probe by Yang and Kou11 to determine the Lewis acidity
of ionic liquids According to the study the presence of a band near 1450 cm-1 indicates
Lewis acidity whilst a band near 1540 cm-1 is indicative of Broslashnsted acidity With respect
to the current investigation this approach may have a two-fold advantage The metal
triflates are soluble in ionic liquids which should enable an IR study to be carried out
aimed at a possible ranking of the metal triflates In addition further information
regarding the induced Broslashnsted acidity arising due to the presence of the metal triflates
may be obtained
Yang and Kou studied CuCl2 FeCl3 and ZnCl2 in [bmim][Cl] In our study [bmim][OTf]
was the ionic liquid and the corresponding metal triflates were used We also repeated
Yangrsquos study of the chlorides in [bmim][OTf] and obtained the same ranking namely
ZnCl2 gt FeCl3 gt CuCl2 The reason behind the use of ndashOTf counter ion is so that the
ionic liquid would provide the same counter ions as the those provided by the metal
triflates guaranteeing the integrity of those species
202
In the current investigation metal triflates (11 equivalents) were added to pyridine in 03
mL of [bmim][OTf] The mixture was stirred at room temperature until complete
dissolution of the metal triflate was obtained Infrared spectra were taken of the
complexes using KBr pellets The results of the experiments are summarised in Table 57
The results indicate that the Lewis acids form a complex with the pyridine this can be
seen by a shift in the peak at around 1440 cm-1 which according to Yang11 indicates
Lewis acidity Closer inspection of Table 57 shows clear difference between the mono
and divalent metals on the one hand and the trivalent metals on the other
Table 57 Δν (cm-1) of pyridine on complexation with various Lewis acids in
[bmim][OTf]
Lewis acid cm-1 Δν (cm-1)Broslashnsted
acid coordination
Pyridine (original peaks) 14400 15893 NaOTf 14410 10 -a
LiOTf 14422 22 -a
Ca(OTf)2 14424 24 -a
Zn(OTf)2 14520 120 -a
Cu(OTf)2 14530 130 -a
Ba(OTf)2 14610 210 -a
LiCl 14860 460 -a
ScCl3 14870 470 -a
In(OTf)3 14880 480 15417 Al(OTf)3 14880 480 15412 Sc(OTf)3 14890 490 15403 Hf(OTf)4 14890 490 15412 Y(OTf)3 14890 490 15403 Er(OTf)4 14890 490 15396 Zr(OTf)4 14890 490 15431 Nd(OTf)3 14890 490 15415 Sm(OTf)3 14890 490 15461 La(OTf)3 14890 490 15439
InCl3 14890 490 -a
GaCl3 14890 490 15437 a ndash indicates no Broslashnsted acid coordination observed
203
With regard to the former and assuming that the extent of the shift in pyridine peak at
1440 cm-1 is indicative of Lewis acid acidity the following ranking of the Lewis acids
can be deduced LiCl gt Ba(OTf)3 gt Cu(OTf)2 gt Zn(OTf)2 gt Ca(OTf)2 gt LiOTf gt NaOTf
Interestingly pyridine seems to have a limit as a probe It would appear that it is only
effective for the softer Lewis acids where we see a range of shifts When we move to the
harder Lewis acids we see a maximum in the shift no matter which Lewis acid is added
These observations are similar to those observed in the 31P NMR study when
diphenylphosphinobenzaldehyde was used as a probe
It is therefore suggested that the interaction between pyridine and the Lewis acid results
in quantitative bond formation and placing a full positive charge on the nitrogen The
bond orders in all of these complexes are essentially the same resulting in similar
absorption frequencies and therefore a lack of discrimination between the different
metals This was not observed by Yang et al because they used a limited number of
metal salts
According to Yang11 a peak in the region of 1540 cm-1 indicates Broslashnsted acidity The
results of the current study (Table 57) indicate that a number of the metal triflates form a
type of Broslashnsted acidity in the ionic liquid These finding are in line with those found in
chapter 4 in which it was found that the metal triflates form Broslashnsted acidity in organic
solvents in the presence of water It should be noted here that although the [bmim][OTF]
was left under high vacuum at high temperature such organic liquids are notoriously
difficult to dry because of the ionic environment The remaining water in the ionic liquid
is most likely the cause of the Broslashnsted acidity seen in this study A noteworthy point
here is that none of the softer Lewis acids used in this study showed Broslashnsted acid
activity
In a further attempt to find a probe that could be used to rank the harder Lewis acids it
was thought that by making the probe more electron poor binding through the lone pair
204
on the nitrogen would be inhibited and in this way the probe may differentiate between
the harder Lewis acids
For the purpose of this aspect of the study several of the harder metal triflates were used
namely Al(OTf)3 Sc(OTf)3 and In(OTf)3 in order to determine if these probes could
discriminate between their electron withdrawing abilities The results are summarised in
Table 58
Table 58 Δν (cm-1) of electron-poor pyridine derivatives on complexation with various
Lewis acids in [bmim][OTf]
Probe and Peak of interest (cm-1)
ν (cm-1) on addition of Al(OTf)3
ν (cm-1) on addition of Sc(OTf)3
ν (cm-1) on addition of In(OTf)3
2 ndash Chloro ndash 6 ndash Methoxypyridine
14694 14694 14694 14694
2 ndash Chloro ndash 5 ndash Nitropyridine
144428 14463 14463 14463
26 - Difluoropyridine 14484 14484 14484 14484 25 ndash Dichloropyridine 14439 14439 14439 14439
From the outcome of these experiments is appears that by withdrawing electron density
from the nitrogen of the pyridine ring it the nitrogen has become a softer base by virtue of
electron density delocalisation and thus will no longer bond with the triflates which on
the whole are hard Lewis acids possibly accounting for the lack of complexation by the
metal triflates seen in the above results (Table 58)
The next logical step in the study was to add electron density onto the ring in an attempt
to make the nitrogen of the pyridine moiety harder and in this way it may discriminate
between the harder Lewis acids Infrared experiments were carried out in the same way
using pyridine derivatives that bearing electron donating moieties using three metal
triflates as before (Table 59)
205
Table 59 Δν (cm-1) of electron rich pyridine derivatives on complexation with various
Lewis acids in [bmim]][OTf]
Probe and Peak of interest (cm-1)
ν (cm-1) on addition of Al(OTf)3
ν (cm-1) on addition of Sc(OTf)3
ν (cm-1) on addition of In(OTf)3
246-Trimethylpyridine 16112 16400 16400 16400
23-Lutidine 15880 Suppression of signal
Suppression of signal
Suppression of signal
23-Lutidine 15581 Suppression of signal
Suppression of signal
Suppression of signal
22rsquo-Dipyridine 14557 14834 14801 14791
The results show that the metal triflates do in fact bind to the electron rich pyridine
derivatives The signals of 23-lutidine and 23- lutidine are suppressed on the addition of
the Lewis acid and no conclusive results could be drawn from the spectra At best it
would appear that the 22rsquo-dipyridine base shows a ranking of the metal triflates of
Al(OTf)3 gt Sc(OTf)3 gt In(OTf)3 However the probe is not sufficiently sensitive to
distinguish Lewis acidity with ease
Numerous other probes for the infrared were also evaluated in this study but are not
discussed exhaustively Various compounds containing carbonyl groups were used but it
was found that instead of seeing a shift in the peak of interest the intensity of the peak
diminished on complexation of the metal triflate Phosphines such as triphenyl phosphine
and triphenylphosphine oxide were also studied little discrimination was seen between
the Lewis acids
It is possible that the use of harder bases possibility even anions is called for For
example it may be that sodium aryl amides (NaNRRrsquo) would be useful or even
phenoxides Due to time constraints these aspects could not be pursued but are put
forward as a possibility for future study in this area
206
55 Conclusions
What the above discussion highlights is that the acidity of Lewis acids should be looked
at in relative terms This stems from the fact that the measurements be it by NMR
spectroscopy IR etc not only measure the extent to which the acid accepts the electrons
from the base but also the extent to which the base is donating the electrons As was seen
in the NMR study different rankings can be obtained when different Lewis bases are
used Ideally one base should be used to rank all Lewis acids But as was shown in the
infrared study the softer Lewis acids could be ranked using pyridine but this probe was
unable to discriminate between the harder Lewis acids
As was mentioned in the beginning of this chapter the quantitative measurements of
Lewis acids is a perennial problem of the Lewis acid-base theory and becomes more
complicated when carried out with metal triflates Amongst others their solubility is
limited to very few solvents making spectroscopic studies difficult
In cases where Lewis acids form strong 11 complexes with ligands induced chemical
shifts appear to be a valuable method for ranking of Lewis acid acidity
207
208
References
1 Branch C S Bott S G Barron A R J Organomet Chem 2003 666 23
2 Pearson RG J Am Chem Soc 1963 85 3533
3 Corma A Garcia H Chem Rev 2003 103 4307
4 Childs R F Mulholland D L Nixon A J Can Chem 1982 60 801
5 Lappert M F J Chem Soc 1962 103 542
6 Williams D B G Shaw M L Green M J Holzapfel C W Angew Chem
Int Ed 2008 47 560
7 Spencer J N Barton S C Cader B M Corsico C D Harrison L E
Mankuta M E Yoder C H Organometallics 1985 4 394
8 Shaw M L Unpublished data University of Johannesburg 2009
9 Drago R S Physical Methods in Chemistry Saunders 1976
10 Satchell D P N Wardell J L J Chem Soc 1964 4134
11 Yang Y Kou Y Chem Commun 2004 226
Summary of conclusions and suggested future research
Main conclusions
The application of Al(OTf)3 and other metal triflates as Lewis acid catalysts for organic
transformations has now been expanded to include reactions not previously investigated
with this compound as facilitator While the mechanism of activation in extremely dry
solvents may be through Lewis acidity (coordination of functionalised substrates to metal
cations) a dual mechanism (ie including Broslashnsted acidity) appears to be operative in
aqueous systems The relative importance of the two mechanisms differs from metal to
metal
The role of water in all systems is yet to be established with certainty and may include
increasing the solubility of metal triflates solvation of metal cations or a source of
Broslashnsted acidity (through activation of water by coordination to metal cations) A starting
point for such investigation requires the availability of extremely dry organic solvents A
very successful method for the rapid drying of several organic solvents has been
identified
With respect to the identification of the Broslashnsted acid component of a possible dual
mechanism the formation of cations from retinyl acetate appears to be a proton specific
reaction However the assumed discrimination between Lewis and Broslashnsted acidity on
the basis of the effect of an added hindered pyridine should be interpreted with caution
particularly in cases where the base is added in a large excess
Comparing the catalytic activity is particularly problematic when water (or a protic
solvent) is a (potential) reagent (eg Mukaiyama aldol reaction) Such comparisons
should preferably be carried out with model reactions which are inert to water
The observation that some metal triflates are not only tolerant of water but require water
for their catalytic activity poses the question of the role of small amounts of water in
209
organic reactions in general This question is particularly relevant in light of the
difficulties experienced in drying organic solvents
Several methods which have been suggested for the comparison of Lewis acidity were
evaluated with respect to metal triflates While none of the methods proved to be ideal
with the view to establishing ranking of Lewis acidity induced NMR chemical shifts of
selected probes appeared to hold the most promise
Future research
There is a real need to establish and compare the solubility of metal triflates in different
organic solvents and to use this information to evaluate their catalytic activities on a
comparative molar basis Too many reactions in the literature proceed in low yield
without mention of whether activity was possibly precluded due to insolubility of the
metal triflates which was identified in the present study as particularly problematic in dry
solvents or when larger amounts of metal triflates are to be dissolved in given solvents
The effects of small amounts of water on metal triflate catalysed reactions in organic
solvents would constitute a useful area of research Similarly there is merit in studies
directed in establishing the role of water in metal triflate catalysed reactions carried out in
water In this regard model reactions should be selected that are completely inert to water
or at least such that water should not be a potential participant in the reaction (as is the
case for the hydrolysis of one of the reactants in the Mukaiyama aldol reaction)
A need exists for obtaining more information on the exact mode of activation of
substrates by metal triflates in non-polar non-coordinating organic solvents eg the
nature of exchange phenomena operating in these situations where non solvated partly
dissociated metal triflates are involved as against dissociated and solvated species in
aqueous or other protic media
210
211
The role of the counterion should be studied in more detail not only with a view to
obtaining better understanding but to suggest cheaper alternatives to the relatively
expensive triflates
In view of its green credentials the application of metal triflates Al(OTf)3 in particular
the identification of new opportunities in synthesis offers real rewards It will be
particularly useful to find applications where the more traditional catalysts fail due either
to a lack of (selective) activity or due to extreme sensitivity of the substrate eg the
extreme sensitivity of pyrroles and indoles to protic acids
Chapter 6
Experimental data and characterisation
61 Standard experimental techniques
611 Chromatography
Thin-layer chromatography (TLC) was conducted on Merck GF254 pre-coated silica
gel aluminium backed plates (025 mm layer) Various solvent mixtures were used to
elute the chromatograms with a mixture of hexane and EtOAc usually being the
eluent of choice Compounds were visualised either by their fluorescence under UV
light (254 nm) or after spraying the TLC plate with a chromic acid solution and then
heating it over an open flame
Flash column chromatography (FCC) refers to column chromatography under
nitrogen pressure (ca 50 kPa) The columns were loaded with Merck Kieselgel 60
(230-400 mesh) and eluted with the appropriate solvent mixtures
612 Anhydrous solvents and reagents
Toluene was dried by passing it over activated alumina under nitrogen pressure (ca
50 kPa) The toluene was then heated over sodium-benzophenone under a nitrogen
atmosphere until the solution turned a deep blue colour The solvent was freshly
distilled before use Dichloromethane dichloroethane and 12-dimethoxyethane were
respectively heated over CaH2 under N2 with subsequent distillation Ethyl acetate
was distilled from K2CO3 using a Vigreux distillation column Hexanes were distilled
prior to use
62 Spectroscopical and spectrometrical methods
621 Nuclear Magnetic Resonance Spectroscopy (NMR)
NMR spectra were recorded using a Varian Gemini 2000 300 MHz spectrometer
The samples were made up in CDCl3 unless otherwise indicated The 1H NMR data
212
are listed in order chemical shift (δ reported in ppm and referenced to the residual
solvent peak of CDCl3 [δ = 724 ppm] or in the case of aromatic compounds to TMS
[δ = 000 ppm]) the multiplicity (s = singlet d = doublet q = quartet br s = broad
singlet dd = doublet of doublets dt = doublet of triplets dq = doublet of quartetsddd
= doublet of doublets od doublets ddt = doublet of doublets of triplets p = pentet sx
= sextet sp = septet) the number of integrated protons the coupling constant J
expressed in Hz and finally the specific hydrogen allocation Spin decoupling
experiments assisted with the determination of the coupling constants and hydrogen
allocation 13C NMR data are listed in the order chemical shift (δ reported in ppm
and referenced to the residual solvent peak of CDCl3 [δ = 770 ppm] and the specific
carbon atom allocation In some cases HSQC HMBC and COSY spectroscopy were
used to assist in the allocation of the spectra
622 Mass spectroscopy (mz)
Mass spectrometry was performed on the Thermo Double Focusing sector High
Resolution mass spectrometer Techniques included EIMS and CIMS
623 Infrared spectroscopy (IR)
A Tensor 27 spectrophotometer was used to record IR spectra using an ATR fitting
The data are listed with characteristic peaks indicated in wavenumber (cm-1)
63 Melting Points
Melting points were determined using a Gallencamp oil immersion apparatus and are
uncorrected
213
64 Chemical methods
Chapter 2
Metal triflates in protection group chemistry
641 General procedure for acetal formation
The aldehyde or ketone (125 mmol) was added to a mixture of the anhydrous alcohol
and 1 mol (0059 g) of the Al(OTf)3 The orthoester (273 mL 25 mmol) was slowly
added to the mixture and the reaction was stirred at room temperature for one hour
The reactions were quenched by passing the reaction mixture through a plug of
neutral alumina to remove the Al(OTf)3 The volatiles were then removed under
vacuum If further purification was necessary bulb-to-bulb vacuum distillation was
used
1-(Dimethoxymethyl)-4-methoxy benzene1 (11)
H
MeO OMe
H3CO
Yield 97 yellow oil
IR νmax (ATR diamond crystal neat) 1464 1301 1246 1049 785 cm-1 1H NMR (300 MHz CDCl3) δH 734 (d 2H J = 87 Hz H2 and H6) 686 (d
2H J = 90 Hz H3 and H5) 532 (s 1H acetal) 377 (s 3H OCH3)
328 (s 6H 2x OCH3) 13C NMR (75 MHz CDCl3) δC 1596 (C4) 1303 (C1) 1278 (C2 and C6) 1134
(C3 and C4) 1030 (acetal) 551 (OCH3) 525 (2 x OCH3)
HRMS (mz) Calculated [M ndash OMe]+ C9H11O2 = 1510765
Obtained = 1510753
214
1-(diethoxymethyl)-4-methoxy benzene2 (12)
H
EtO OEt
H3CO
Yield 90 yellow oil
IR νmax (ATR diamond crystal neat) 2974 1511 1246 1034 772cm-1 1H NMR (300 MHz CDCl3) δH 736 (d 2H J = 93 and J= 06 Hz H2 H6)
685 (d 2H J = 87 Hz H3 H5) 543 (s 1H acetal) 376 (s 3H
OCH3) 376 ndash 345 (m 4H 2 x OCH2) 120 ( t 6H J = 71 Hz CH3) 13C NMR (75 MHz CDCl3) δC 19063 (COC3H) 1318 (ipso) 1280 (ortho)
1133 (meta) 1013 (acetal) 607 (2 x OCH2) 150 (2 x CH3)
HRMS (mz) Calculated [M ndash OEt]+ C10H13O2 = 165091
Obtained = 1650910
1-Chloro-4-(dimethoxymethyl) benzene3 (13)
MeO
H
OMe
Cl
Yield gt98 colourless oil
IR νmax (ATR diamond crystal neat) 2937 2830 1088 1052 808 cm-1 1H NMR (300 MHz CDCl3) δH 736 (d 2H J = 75 Hz H2 H6) 731 (d J =
66 Hz H3 H5) 535 (s 1H acetal) 329 (s 6H 2 x OCH3) 13C NMR (75 MHz CDCl3) δC 1366 (ipso) 1342 (para) 1283 (meta) 1283
(ortho) 1022 (acetal) 525 (2 x OCH3)
HRMS (mz) Calculated [M ndash OMe]+ C9H11ClO = 1560258
Obtained = 1569867
215
1-chloro-4-(diethoxymethyl)benzene2 (14)
EtO
H
OEt
Cl
Yield gt98 colourless oil
IR νmax (ATR diamond crystal neat) 2975 2881 1087 1051 1015 cm-1 1H NMR (300 MHz CDCl3) δH 729 (d 2H J = 51 Hz H2 H6) 729 (d 2H J
= 75 Hz H3 H5) 545 (s 1H acetal) 360ndash356 (m 4H 2 x
CH2CH3) 120 (t 6H J = 71 Hz 2 x CH2CH3) 13C NMR (75 MHz CDCl3) δC 1376 (ipso) 1339 (para) 1282 (meta) 1230
(ortho) 1006 (acetal) 608 (2 x CH2CH3) 150 (2 x CH2CH3)
HRMS (mz) Calculated [M ndash OEt]+ C9H10ClO = 1690415
Obtained = 1690416
4-Nitroacetophenone dimethyl acetal (15)
MeO OMe
O2N
Yield 97 yellow solid
mp 612 ndash 634 degC
IR νmax (ATR diamond crystal neat) 2945 1520 1350 1086 1034 cm-1 1H NMR (300 MHz CDCl3) δH 818 (d 2H J = 87 Hz H2 and H6) 765 (d
2H J = 93 Hz H3 H5) 317 (s 6H 2 x OCH3) 151 (s 3H CH3) 13C NMR (75 MHz CDCl3) δC (1501 (CNO2) 1475 (ipso) 1274 (meta) 1234
(ortho) 491 (2 x OCH3) 258 (CH3)
HRMS (mz) Calculated [M ndash CH3]+ C9H10NO4 = 1960604
Obtained = 1960593
216
4-Nitroacetophenone dimethyl acetal (16)
EtO OEt
O2N
Yield 92 yellow oil
IR νmax (ATR diamond crystal neat) 12976 1520 1347 1045 857 cm- 1H NMR (300 MHz CDCl3) δH 810 (d 2H J = 78 Hz H2 and H6) 762 (d
2H J = 78 Hz H3 H5) 345 ndash 338 (m 2H CH2ACH3) 330 ndash 320
(m 2H CH2BCH3) 147 (s 3H CH3) 14 (t J = 63 Hz 6H 2 x
OCH2CH3) 13C NMR (75 MHz CDCl3) δC (1511 (ipso NO2) 1473 (ipso acetal) 1272 (C3
and C5) 1232 (C2 and C6) 1006 (Cα) 569 (2 x OCH2CH3) 268
(CH3) 151 (OCH3)
HRMS (mz) Calculated [M ndash OEt]+ C9H10NO4 = 1940812
Obtained = 1940811
o-Nitrobenzaldehyde dimethyl acetal4 (17)
H
MeO OMe
NO2
Yield 95 light yellow oil
IR νmax (ATR diamond crystal neat) 2937 1529 1359 1108 1055 cm-1 1H NMR (300 MHz CDCl3) δH 777 (d 2H J = 66 and 12 Hz H3 H6) 774
(d 1H J = 63 and 15 Hz H4) 756 (t 1H J = 78 and 11 Hz H5)
589 (s 1H acetal) 336 (s 6H 2 x OCH3) 13C NMR (75 MHz CDCl3) δC 148 (CNO2) 1325 (ipso) 1325 (C6) 1293
(C3) 1280 (C5) 1241 (C4) 997 (acetal) 544 (2 x OCH3)
HRMS (mz) Calculated [M ndash OMe]+ C8H8NO3 = 1660499
Obtained = 1660498
217
o-Nitrobenzaldehyde diethyl acetal4 (18)
H
EtO OEt
NO2
Yield gt98 light yellow oil
IR νmax (ATR diamond crystal neat) 2977 1529 1360 1108 1055 cm-1 1H NMR (300 MHz CDCl3) δH782 (d 1H J = 63 Hz H3) 777 (d 1H J =
78 Hz H6) 757 (dt 1H J = 60 and 13 Hz H5) 743 (dt 1H J = 78
and 15 Hz H4) 371 ndash 361 (m 2H 2 x OCH2ACH3) 360 ndash 350(m
2H 2 x OCH2BCH3) 122 (t 6H J = 72 Hz CH2CH3) 13C NMR (75 MHz CDCl3) δC 1336 (CNO2) 1324 (C6) 1291 (C3) 1280
1241 (C4) 983 (acetal) 634 (2 x OCH2) 150 (2 x CH3)
HRMS (mz) Calculated [M ndash OEt]+ C9H10NO4 = 1800666
Obtained = 1800655
(33-Dimethoxy-1-propen-1-yl)-benzene3 (19)
Ph OMe
OMe
Yield gt98 yellow oil
IR νmax (ATR diamond crystal neat) 2932 1449 1190 1049 772cm-1 1H NMR (300 MHz CDCl3) δH 744 ndash 736 (m 2H ortho) 735 -726 (m 3H
meta para) 674 (d 1H J = 159 Hz PhCH=CH) 617 (dd 1H J =
161 and 50 Hz PhCH=CH) 497 (d 1H J = 11 and 45 Hz acetal)
338 (s 6H OCH3) 13C NMR (75 MHz CDCl3) δC 1360 (ipso) 1335 (para) 1285 (meta) 1280
(PhCH=CH) 1266 (para) 1256 (PhCH=CH) 1028 (acetal) 526 (2
x OCH3)
HRMS (mz) Calculated [M ndash OMe]+ C10H11O = 1470804
Obtained = 1470805
218
(33-Diethoxy-1-propen-1-yl)-benzene (110)
Ph OEt
OEt
Yield gt98 light yellow oil
IR νmax (ATR diamond crystal neat) 2975 1679 1120 1049 969 cm-1 1H NMR (300 MHz CDCl3) δH 740 (d 2H J = 78 Hz ortho) 730 ndash 723 (m
3H meta para) 670 (d 1H J = 162 Hz PhCH=CH) 620 (dd 1H J
= 162 and 51 Hz PhCH=CH) 372 ndash 367 (m 2H OCH2ACH3) 361
ndash 350 (m 2H OCH2BCH3) 124 (t 6H J = 70 Hz 2 x CH2CH3) 13C NMR (75 MHz CDCl3) δC 1361 (ipso) 1329 (PhCH=CH) 1285 (meta)
1280 (para) 1267 (ortho) 1266 (PhCH=CH) 1014 (acetal) 610 (2
x OCH2CH3) 152 (2 x OCH2CH3)
HRMS (mz) Calculated [M ndash OEt]+ C11H13O = 1610961
Obtained = 1610960
11rsquo-Dimethoxy-decane5 (111)
MeO
OMe
H
Yield gt98 colourless oil
IR νmax (ATR diamond crystal neat) 2923 1219 1122 1055 769cm-1 1H NMR (300 MHz CDCl3) δH 430 (t 1H J = 59 Hz acetal) 325 (s 3H 2 x
OCH3) 152 (m 2H CH2CH3) 123 ndash 121 (m 14H CH3(CH2)7CH2)
082 (t 3H J = 65 Hz CH2CH3) 13C NMR (75 MHz CDCl3) δC 1045 (CH) 524 (OCH3) 324 (CH2CH) 318
(CH3CH2CH2) 295 ndash 293 (CHCH2(CH2)5) 246 (CH3CH2) 226
(CH2CH2CH) 140 (CH3)
HRMS (mz) Calculated [M ndash OMe]+ C11H23O = 1711743
Obtained = 1711741
219
11rsquo-Diethoxy-decane5 (112)
EtO
OEt
H
Yield gt98 colourless oil
IR νmax (CHCl3) 2943 1222 1234 1064 765 cm-1 1H NMR (300 MHz CDCl3) δH 444 (t 3H J = 47 Hz acetal) 360 (m 2H
CH2A) 346 (m 2H CH2B) 155 (m 2H CH2CH) 124ndash198 (m
12H CH3CH2(CH2)6CH2) 175 (t 6H J = 75 Hz 2 x OCH3) 084 (t
3H J = 66 Hz CH3) 13C NMR (75 MHz CDCl3) δC 1030 (acetal) 607 (2 x OCH2CH3) 336
(CH2CH) 318 (CH3CH2CH2) 294ndash293 (CHCH2(CH2)5) 247
(CH3CH2) 226 (CHCH2CH2) 153 (2 x OCH2CH3) 141 (CH3)
HRMS (mz) Calculated [M ndash OEt]+ C12H25O = 1851900
Obtained = 1851902
Dimethoxymethyl-cyclohexane (113)
OMe
OMe
Yield gt98 colourless oil
IR νmax (ATR diamond crystal neat) 2929 1215 1083 1048 754 cm-1 1H NMR (300 MHz CDCl3) δH 395 (d 1H J = 69 Hz acetal) 330 (s 6H 2 x
OCH3) 175ndash167 (m 4H H2A H3A H5A H6A) 163ndash150 (m 2H
H4) 121ndash110 (m 2H H2B H6B) 107ndash091 (m 2H H3B H5B) 13C NMR (75 MHz CDCl3) δC 1085 (acetal) 534 (2 x OCH3) 400 (C1) 280
(C3 and C5) 263 (C4) 257 (C4 and C6)
HRMS (mz) Calculated [M ndash OMe]+ C8H15O = 1271117
Obtained = 1271099
220
Diethoxymethyl cyclohexane (214)
OEt
OEt
Yield gt98 light yellow oil
IR νmax (ATR diamond crystal neat) 2925 2853 1130 1080 1056 cm-1 1H NMR (300 MHz CDCl3) δH 408 (d 1H J = 72 Hz acetal) 365 ndash 358 (m
2H OCH2ACH3) 355 ndash 340 (m 2H OCH2BCH3) 177 -167 (m 4H
H2A H3A H5A H6A) 162 -119 (m 2H H4) 114 (t 6H J = 72 Hz
CH2CH3) 106 ndash 087 (m 4H H2B H3B H5B H6B) 13C NMR (75 MHz CDCl3) δC 1067 (acetal) 615 (2 x OCH2) 407 (C1) 281
(C3 and C5) 264 (C4) 258 (C2 and C6) 153 (2 x OCH2CH3) 1HRMS (mz) Calculated [M ndash OEt]+ C8H15O = 1411274
Obtained = 1411275
55-dimethyl-2-phenyl-13-dioxane (215)
O
O
Yield 98White solid
mp 443 ndash 474 degC
IR νmax (ATR diamond crystal neat) 2960 1456 1392 1106 770 cm-1 1H NMR (300 MHz CDCl3) δH 751ndash748 (m 2H ortho) 738ndash734 (m 3H
meta para) 538 (s 1H acetal) 373 (d 2H J = 111 Hz OCH2A)
370 (d 2H J = 111 Hz OCH2B) 129 (s 3H CH3) 078 (s 3H CH3) 13C NMR (75 MHz CDCl3) δC 1384 (ipso) 1288 (para) 1283 (meta) 1261
(ortho) 1018 (acetal) 776 (2 x OCH2) 230 (CH3) 219 (CH3)
HRMS (mz) Calculated [M]+ C12H16O2 = 1921150
Obtained = 1921147
221
Dimethoxymethyl-benzene3 (216)
H
MeO OMe
Yield 98 light yellow oil
IR νmax (ATR diamond crystal neat) 2975 1338 1094 1050 700 cm-1 1H NMR (300 MHz CDCl3) δH 745 - 750 (m 2H H-aromatic) 732 - 7393
(m 3H H-aromatic) 541 (s 1H CH(OCH3)2) 333 (s 6H 2 x CH3) 13C NMR (75 MHz CDCl3) δC 1381 (ipso) 1284 (para) 1282 (meta) 1267
(ortho) 1031 (CH) 526 (OCH3)
HRMS (mz) Calculated [M ndash OMe]+ C8H9O = 1210648
Obtained = 1210445
Dimethoxymethyl-benzene2 (217)
H
EtO OEt
Yield 92 dark yellow oil
IR νmax (ATR diamond crystal neat) 2963 1324 1089 1047 745 cm-1
H NMR (300 MHz CDCl3) δH 750ndash752 (m 2H H-aromatic) 730ndash740 (m
3H H-aromatic) 553 (s 1H CH(OCH3)2) 350ndash366(m 4H 2 x
OCH2) 126 (t 6H J = 705 Hz 2 x CH3) 13C NMR (75 MHz CDCl3) δC 1389 (ipso) 1280 (para) 1280 (meta) 1264
(ortho) 1013 (CH) 607 (OCH2) 150 (CH2CH3) 1HRMS (mz) Calculated [M ndash OEt]+ C9H11O = 1350804
Obtained = 1350807
222
(11-Dimethoxyethyl)-benzene4 (218)
MeO OMe
Yield gt98 light yellow oil
IR νmax (ATR diamond crystal neat) 2929 1215 1083 1048 754 cm-1 1H NMR (300 MHz CDCl3) δH 753 (d 2H J = 75 Hz ortho) 740 ndash 728 (m
3H meta para) 320 (s 6H 2 x OCH3) 160 (s 3H CH3) 13C NMR (75 MHz CDCl3) δC 1420 (ipso) 1280 (ortho) 1274 (para) 1261
(meta) 1016 (acetal) 488 (2 x OCH3) 260 (CH3)
HRMS (mz) Calculated [M ndash CH3]+ C9H11O2 = 1510754
Obtained = 1510755
(11-Diethoxyethyl)-benzene4 (219)
EtO OEt
Yield 82 light yellow oil
IR νmax (ATR diamond crystal neat) 2974 1219 1119 1049 772cm-1 1H NMR (300 MHz CDCl3) δH 756 (d 2H J = 156 Hz ortho) 740 ndash 726 (m
3H meta para) 355 ndash 345 (m 2H CH2ACH3) 343 ndash 333 (m 2H
CH2BCH3) 160 (s 3H CH3) 123 (t 6H J = 71 Hz 2 x CH2CH3) 13C NMR (75 MHz CDCl3) δC 1438 (ipso) 1280 (ortho) 1272 (para) 1272
(para) 1261 (meta) 1011 (acetal) 566 (2 x OCH2CH3) 271
(OCH2CH3)
HRMS (mz) Calculated [M ndash OEt]+ C11H15O2 = 1791067
Obtained = 1791067
223