A paradigm shift in the construction of heterobimetallic complexes:Synthesis of group 2 & 4 metal–calix[6]arene complexes

Antonella J. Petrella,a Donald C. Craig,a Robert N. Lamb,a Colin L. Raston b andNicholas K. Roberts a

a School of Chemical Sciences, University of New South Wales, Sydney 2052, Australia.E-mail: r.lamb@unsw.edu.au

b School of Biomedical and Chemical Sciences, University of Western Australia, Crawley,Perth WA 6009, Australia. E-mail: clraston@chem.uwa.edu.au

Received 15th October 2003, Accepted 1st December 2003First published as an Advance Article on the web 11th December 2003

Deprotonation of calix[6]arenes with barium in methanol followed by the addition of [Ti(OPri)4] or [Zr(OBun)4] iseffective in the formation of novel dimeric 2 : 1 barium–titanium()/zirconium() calix[6]arene complexes. In thesecomplexes a central Ti()/Zr() coordinated in the exo-position connects the two calix[6]arenes in the 1,3-alternateconformation, each with an endo-barium sharing common phenolate groups with the titanium/zirconium centre andparticipating in cation–π interactions. A homometallic barium calix[6]arene dimer was also prepared wherein thecalix[6]arenes are in the 1,3-alternate conformation with each coordinating one endo- and one exo-barium centre. Theexo-barium cations connect the two calix[6]arenes through bridging methanol ligands. In this and the heterometalliccomplexes, cation–π complexation of the Ba2� ion within the 1,3 alternate conformation of calix[6]arene facilitatesthe formation of the dimeric complexes in methanol. In contrast, the smaller Sr2� ion did not form similar complexesin methanol, and the formation of an analogous 2 : 1 strontium–titanium calixarene complex required the use of themore sterically demanding donor alcohol, isopropanol, the resulting complex being devoid of cation–π interaction.The results show (i) that a subtle interplay of solvation strength, coordination array type and cavity/cation sizeinfluences the accessibility of heterobimetallic complexes based on calix[6]arenes, and (ii) a synergistic endo–exobinding behaviour.

IntroductionPotential applications of heterometallic alkoxide complexes inthe formation of multimetallic ceramics, either by sol-gel orMOCVD processes, has motivated considerable research intomethods for tailoring properties and tuning metal stoichi-ometries of the precursor complexes.1–5 Since the discovery ofheterometallic alkoxide complexes in 1924, various syntheticmethodologies have been established to gain access to suchcomplexes; the initial limited scope of unidentate alkoxideligands in mediating the formation of mixed metal speciesthrough bridging led to the use of multidentate ligands includ-ing bi/tri-dentate functionalised alkoxides, β-diketonates,acetate and oxo-ligands.6,7

Assembling heterometallic complexes as precursors forceramics depends on the ability of each metal to fill its co-ordination sphere with anionic or neutral oxygen donor atomsof ligands in a combination of terminal or bridging modes.Surprisingly few aryloxide ligands have been used in the syn-thesis of such complexes,8,9 especially in gaining access toheterometallic complexes of hard transition metal cations andlarge soft polarisable cations, whereby the hard metal cationscan, in principle, preorganise the aryloxide ligands, creatingaromatic pockets for π-arene coordination of the soft cations.These types of interactions have become recognised in recentyears for alkali and alkaline earth metal complexes in gener-al.10–12 To date there are very few structurally characterisedcomplexes manifesting these structural features, i.e. O-boundtransition metal centres organising aryloxide ligands such thatsoft metal cation–arene π interactions are facilitated/pro-moted.13 This may be related to the tendency of heterometalliccomplexes to undergo competing redistribution/hydrolysis reac-tions, and the difficulty of preorganising ligands into a con-formation suitable for soft metal–transition metal complexinterplay.

We have explored the use of calixarenes to stabilise cation–πinteractions in heterometallic species.14,15 Calixarenes are

comprised of a macrocylic arrangement of methylene linkedphenols which are poised to act as multidentate oxo-ligandsthrough their phenoxy groups, and often the resulting conform-ation provides shape-specific π-rich cavities. Complexes of cal-ix[4]arene, as the lowest oligomer in the series, generally adoptthe cone conformation,16,17 except for some group 13 organo-metallic systems,18 and provides two distinct cation bindingsites: either endo- or exo- to the cavity. In the endo-case, a metalion can interact with the π faces of the aromatic groups whilstsimultaneously interacting with calix[4]arene phenol groups atthe bottom of the cone whereas, in the exo-case, up to fourmetal ions can interact with the phenol groups. There areseveral examples of calixarene complexes with endo-alkalimetals in association with exo-transition metals,19–21 but there isonly one structurally authenticated example of an endo-group 2complex.22 Complexation of the heavier group 2 congeners isproblematic due to the relatively low coordination saturationafforded by the macrocycle.

The larger calix[6]arenes offer greater coordination of group2 metal ions, but this is at the expense of the increased con-formational flexibility of the calixarene, and controlling this is achallenge, as is controlling the number of metal ions complex-ing to a larger number of exo-binding sites. Whilst strategiesfor making complexes of the relatively rigid calix[4]arene aregenerally well defined, this is not the case for calix[6]arene.

Herein we report the formation of novel mixed barium–titanium, 1, mixed barium–zirconium, 2, homometallic barium,3, and mixed strontium–titanium, 4, calix[6]arene complexes.Interestingly the structural organisation of the calix[6]arenein the complexes containing barium occurs not only throughbarium–titanium/zirconium phenoxo coordination, but alsothrough barium π-complexation, while for the smaller stron-tium ion, no cation π-complexation occurs. Moreover, while theformation of the barium complexes occurs in high yield inmethanol, for the smaller strontium cation, complex form-ation can be effected only in the more sterically bulky donorsolvent isopropanol. Significantly the synthesis represents a

327

DO

I:1

0.1

03

9/ b

31

29

19

e

T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4 D a l t o n T r a n s . , 2 0 0 4 , 3 2 7 – 3 3 3

Publ

ishe

d on

11

Dec

embe

r 20

03. D

ownl

oade

d by

Uni

vers

ity o

f W

inds

or o

n 30

/10/

2014

04:

03:3

8.

View Article Online / Journal Homepage / Table of Contents for this issue

Scheme 1 S = methanol, * Ti(OPri)4 was premixed with one equivalent of TMEDA in methanol.

new methodology for the construction of heterobimetalliccomplexes based on calix[6]arene and illustrates the synergisticbinding of metals to polyphenolic systems and the crucialeffects of solvent and cation size. These complexes may reflectbonding principles underpinning some biological systems, someof which are known to feature a mixture of oxy and aromaticbinding sites.

Results and discussion

Synthesis and characterisation

Syntheses of the new complexes are outlined in Scheme 1.Complex 1 was prepared in 88% yield by the reaction of oneequivalent of barium with one equivalent of p-But-calix[6]-arene in methanol, followed by the addition of a 1 : 1 solu-tion of [Ti(OPri)4] and N,N,N�,N�-tetramethylethylenediamine(TMEDA) in methanol. It was necessary to mix the [Ti(OPri)4]with the TMEDA before addition of methanol, to avoid pre-cipitation of titanium hydroxides/methoxides. The preparationof this complex has been communicated previously.23 Complex2 precipitated as a gelatinous solid on refluxing half an equiv-alent of [Zr(OBun)4] with a suspension of the solid formedwhen calix[6]arene is partially deprotonated by barium inmethanol. This is the same preparation as that reported for 1,

Scheme 1. However, whereas 1 could be crystallised from amixture of methanol and dichloromethane, attempts to crystal-lise 2 in the same way resulted in decomposition, affording onlycrystals of the calix[6]arene starting material. The 1H NMRspectrum of the crude Ba–Zr calix[6]arene reaction product indichloromethane indicated complexation of the calix[6]arenehad indeed occurred, thus suggesting that the method ofrecrystallisation resulted in decomposition of the complex. Inorder to recrystallise the Ba–Zr calix[6]arene complex it wasnecessary to first remove methanol solvate from the crudesolid, under vacuum. The dry solid could then be recrystallisedby overlaying a dichloromethane solution with a mixture ofdimethoxyethane and acetonitrile.

The solid-state structure of 1 suggests that the 1H NMRspectrum should contain three sets of methylene doublets.Three such doublets (one slightly broad) are indeed observedwhen solid 1 is desolvated under vacuum prior to dissolution inCDCl3. In contrast, the 1H NMR spectrum of the as-obtainedmethanol solvate 1 in CDCl3 shows only one of the expecteddoublets, the others presumably being broadened into the base-line. These results suggest that the presence of methanol resultsin rapid intramolecular re-organisation on the NMR time scale.This notion is also supported by the observation that the 1HNMR solution of 2 prepared from a crystalline sample, i.e.material containing no methanol, gave three sharp sets of

328 D a l t o n T r a n s . , 2 0 0 4 , 3 2 7 – 3 3 3

Publ

ishe

d on

11

Dec

embe

r 20

03. D

ownl

oade

d by

Uni

vers

ity o

f W

inds

or o

n 30

/10/

2014

04:

03:3

8.

View Article Online

methylene doublet signals. Signals for the residual phenolic OHprotons are not observable in any spectra of 1 or 2.

The atom-to-atom connectivities in complexes 1 and 2 in thesolid state were established by single crystal diffraction studies.Both complexes crystallise in space group P21/n with the asym-metric unit comprised of half a molecule, the other half gener-ated by an inversion centre, Fig. 1.24 The calix[6]arenes are in the1,3-alternate conformation. The symmetry of the moleculeimposes a linear arrangement of the three metal centres, so thatthe cavities devoid of metal ions are pointing in opposite direc-tions. For complex 1, each such cavity contains a molecule ofdichloromethane, whereas for 2 no solvent is found within thesecavities. Interestingly the chlorinated solvent molecules fitsnugly in the cavities and are associated with C–H � � � π inter-actions, the closest CH � � � centroid distance being 2.61 Å,which is comparable to values reported previously for di-chloromethane inclusion complexes of calix[4]arene (2.46, 2.48,2.61 Å) 25 and calix[6]arene (2.49 Å).26

The central Ti(), 1, and Zr(), 2, are coordinated octa-hedrally by three contiguous phenolate groups from each calix-[6]arene, the two 1,3-arranged groups from each calixarenebridging also to its proximate barium centre. The aromaticrings of the unique non-bridging phenolates are involved inBa � � � π-arene interactions. The Ba(1)–C(12–17) distancesrange from 3.28–3.53 Å (Ba–arene centroid 3.10 Å) in 1 and3.41–3.85 Å (Ba–arene centroid 3.35 Å) in 2. The considerablylonger Ba–C distances in complex 2 may reflect the steric con-straints of the bidentate chelating dimethoxyethane ligand

Fig. 1 Molecular projections of 1 and 2.

(see below) which, on steric grounds, limits the proximity of thebarium–aromatic ring interaction. The barium centres alsobind to two other phenol moieties and, on charge balance con-siderations, one of these for each calixarene must be proton-ated, assuming that all groups around titanium/zirconiumare deprotonated (as expected for a polarising quadrivalentmetal centre). Two unidentate methanols and one bidentate di-methoxyethane ligand complete each barium environmentfor 1 and 2, respectively, thus the barium centres are seven-coordinate, counting a π-arene as occupying one coordinationsite. Unfortunately, the large scale of the structures precludedlocation of hydrogen atoms, and any attempt to assign whichoxygen centres correspond to the phenolic groups would bepurely speculative. Selected bond distances and angles for 1 and2 are listed in Table 1.

Significantly 2 represents the first heterobimetallic complexfor zirconium based on the Zr(OR)6

2� motif, with an additionalnovelty being the simplicity of its synthesis involving aprotic reaction solvent. Typically heterobimetallic complexes ofzirconium are based on the confacial bis-octahedral metallo-ligand motif Zr2(OR)9

�, which is capable of chelating/bridgingto its heterometallic partner in a uni-, bi-, tri- or quadri-dentatefashion, although most commonly quadri-dentate ligationoccurs through four alkoxy oxygen atoms (two terminal andtwo bridging).4 The stable Zr2(OR)9

� anion forms where R =Me, Et, Prn, Pri, Bun, Bus, but for R = But the Zr2(OR)9

� anionis destabilised due to steric hindrance,27 and alternative anionsform, such as Zr(OBut)5

� and Zr(OBut)62�.3 In contrast, hetero-

metallic complexes involving titanium are typically based onthe trigonal bipyramidal Ti(OR)5

�, octahedral Ti(OR)62� and,

to a lesser extent, confacial bi-octahedral Ti2(OR)9� anions.4

The observed differences relate to the larger size of thezirconium cation, whose ionic radius is 0.72 Å (Zr4�), cf. 0.61 Å(Ti4�).

In order to rationalise the formation of 2, we considered thatthe type of anion formed by zirconium would be a function ofthe reaction solvent, heterometallic partner and the solubilityof possible complexes of the two metals. Since the largeralkaline earth metals generally prefer coordination numbers> 6, including solvation of Ca2�, Sr2� and Ba2� by methanol,it is desirable that any pairing of the metals should resultin >6 coordinate alkaline earth centres. In 2 each bariumatom is seven coordinate, counting the Ba2�–π-arene inter-action as the metal bound to one sterically demanding ligand.The octahedral coordination of zirconium is also favourableand therefore the formation of the 2 : 1 complex is possible.Precipitation of the dimer from methanol presumably drives itsformation at the expense of any other soluble alternatives.Nevertheless, this compound decomposes when dissolved indichloromethane–methanol, although establishing the natureof the resulting species remains elusive. In the absence ofprotic reagents the complex is stable, as evidenced by ourability to recrystallise it from a mixture of dichloromethane,dimethoxyethane and acetonitrile.

We observed that the reaction of one equivalent of bariummetal with a suspension of calix[6]arene in hot methanolafforded an intermediate species followed by a new crystallineprecipitate that could not be redissolved in methanol. The samebehaviour was observed for other group 2 cations. In contrastthe reactions of Group 1 metals with suspensions of calix-[6]arene in methanol formed only clear solutions, indicating thepresence of more labile species, possibly as solvent-separatedion pairs. Unfortunately we were unable to structurally charac-terise the 1 : 1 Ba calix[6]arene reaction product. All attempts toform X-ray quality crystals either by slow crystallisation uponits formation, or by adding methanol or dimethoxyethane todichloromethane solutions of the product were unsuccessful.

As evident from the 2 : 1 Ba–(Ti/Zr()) calix[6]arene com-plexes, the additional coordination of either titanium or zir-conium stabilises the endo-complexation of barium by fixing

329D a l t o n T r a n s . , 2 0 0 4 , 3 2 7 – 3 3 3

Publ

ishe

d on

11

Dec

embe

r 20

03. D

ownl

oade

d by

Uni

vers

ity o

f W

inds

or o

n 30

/10/

2014

04:

03:3

8.

View Article Online

the calix[6]arene conformation, however solutions of thezirconium complex are stable only in the absence of proticreagents. An interesting complementarity of coordinationmodes is displayed in the homometallic barium calix[6]arenecomplex 3, where the endo-complexation of barium by calix-[6]arene is also stabilised by additional exo-coordination ofbarium to calix[6]arene. Complex 3 was prepared in 90% yieldby the addition of 2.5 equivalents of barium metal to a suspen-sion of calix[6]arene in methanol (Scheme 1), and crystals weregrown from a dichloromethane–methanol mixture.

Complex 3 crystallises in the space group P1̄, with half thedimer of two calix[6]arene molecules as the asymmetric unit,the other being generated by an inversion centre. The calix-[6]arenes are in the 1,3-alternate conformation with each co-ordinating one endo- and one exo-barium centre (Fig. 2). Theexo-barium cations connect the two calix[6]arenes throughbridging methanol ligands, Fig. 2,24 such that cavities devoid ofmetal centres point away from each other, as found in 1 and 2.Located within each of these cavities is a dichloromethanemolecule which has two C–H � � � π interactions at 2.813 and2.819 Å. These are somewhat longer than those previouslyreported for dichloromethane inclusion complexes of calix-[4]arene (2.46, 2.48, 2.61 Å) 25 and calix[6]arene (2.49, 2.61 Å).26

Each endo-bound barium centre is complexed by twomethanol molecules, and four calixarene phenol oxygens, twoof which also form bridges to the exo-binding barium atoms.Additionally, each endo-barium forms cation–π interactionswith the calixarene phenol that has its O-centre uniquely co-ordinating the exo-barium cation. The Ba(1)–C(12–17) dis-tances range from 3.34 to 3.46 Å, with the Ba–arene centroiddistance at 3.09 Å, these distances being consistent with thosefound in 1.

The two exo-bound barium centres each coordinate threecontiguous phenolate groups from each calixarene and fourmethanol ligands, two of which bridge the barium centres atdistances of 2.887(3) and 3.057(3) Å. Notably, the dimericassembly also appears to be stabilised by hydrogen bondinginteractions between the calixarene oxygen (O2) and themethanol solvates (O1Me2�) and (O1Me3�) which coordinatethe exo-barium associated with the other calixarene (O2 � � �O1Me2�, O1Me3�, 2.604(5), 2.651(5) Å). The comparativelyshort calixarene oxygen, (Oc), exo-barium, (Baexo) distances vs.methanol oxygen, (Om), Baexo distances (Baexo–Oc = 2.582(3),2.719(3), 2.560(3) vs. Baexo–Om = 2.820(4), 2.887(3), 2.739(4),2.808(4) Å) suggests that all three calixarene phenols that co-ordinate the exo-bariums are deprotonated. This is also consist-ent with the higher acidity of phenol groups compared tomethanol. However on charge balance considerations thecollective 8� charge of the metal centres requires two moredonors to be deprotonated. Since the large nature of the

Fig. 2 Molecular projection of 3.

Tab

le 1

Sele

cted

bon

d di

stan

ces

(Å)

and

angl

es (

�) fo

r 1

and

2; C

* is

the

cen

troi

d of

the

are

ne r

ing

inte

ract

ing

wit

h ba

rium

Ti1

–O(1

, 2, 3

)1.

930(

4), 1

.940

(4),

1.9

44(4

)Z

r1–O

(1, 2

, 3)

2.07

2(2)

, 2.0

70(2

), 2

.057

(2)

Ba1

–O(1

, 3, 4

, 6, 1

Me1

, 1M

e3)

2.72

5(4)

, 2.6

74(4

), 2

.847

(4),

2.8

05(4

), 2

.752

(5),

2.7

15 (

5)B

a1–O

(1, 3

, 4, 6

, 1D

1, 2

D1)

2.75

8(2)

, 2.7

86(2

), 2

.794

(2),

2.8

18(2

), 2

.741

(2),

2.7

45(2

)B

a1–C

*3.

10B

a1–C

*3.

35O

1–T

i(1)

–O(1

� , 2, 2

� , 3, 3

� )18

0.0,

89.

8(2)

, 90.

2(2)

, 84.

0(2)

, 96.

0(2)

O1–

Zr1

–O(1

�, 2

, 2�,

3, 3

�)18

0.0,

87.

83(8

), 9

2.17

(8),

84.

31(7

), 9

5.69

(7)

O2–

Ti(

1)–O

(2� , 3

, 3i )

180.

0, 8

9.7(

2), 9

0.3(

2)O

2–Z

r1–O

(2�,

3, 3

�)18

0.0,

86.

97(8

), 9

3.03

(8)

O3–

Ti(

1)–O

(3� )

180.

0O

3–Z

r1–O

(3�)

180.

0O

1–B

a–O

(3, 4

, 6, 1

Me1

, 1M

e3)

57.4

(1),

121

.5(1

), 9

1.3(

1), 1

63.5

(1),

116

.5(2

)O

1–B

a1–O

(3, 4

, 6, 1

D1,

2D

1)59

.97(

6), 1

27.4

4(6)

, 89.

96(6

), 1

19.2

7(6)

, 160

.75(

7)O

3–B

a–O

(4, 6

, 1M

e1, 1

Me3

)88

.6(1

), 1

24.1

(1),

109

.6(1

), 1

63.9

(1)

O3–

Ba1

–O(4

, 6, 1

D1,

2D

1)92

.65(

6), 1

27.1

1(6)

, 167

.62(

7), 1

18.6

6(7)

O4–

Ba–

O(6

, 1M

e1, 1

Me3

)68

.8(1

), 6

4.2(

1), 1

06.3

(1)

O6–

Ba–

O(1

Me1

, 1M

e3)

105.

0(1)

, 68.

4(1)

O4–

Ba1

–O(6

, 1D

1, 2

D2)

71.3

9(6)

, 96.

47(7

), 7

0.80

(7)

O1M

e1–B

a–1M

e373

.0(2

)O

6–B

a1–O

(1D

1, 2

D1)

64.1

7(7)

, 103

.37(

7)B

a1–O

(1, 3

)–T

i110

7.5(

2), 1

09.0

(2)

O1D

1–B

a1–O

2D1

57.3

8(8)

C*–

Ba1

–O(1

, 3, 4

, 6, 1

Me1

, 1M

e3)

78.6

, 78.

5, 1

44.7

, 144

.6, 8

9.2,

85.

8.B

a1–O

(1, 3

)–Z

r110

7.60

(7),

107

.05(

7)

C

*–B

a1–O

(1, 3

, 4, 6

, O1D

1, O

2D1)

77.8

, 76.

4,14

2.6,

142

.8, 9

1.3,

75.

7

330 D a l t o n T r a n s . , 2 0 0 4 , 3 2 7 – 3 3 3

Publ

ishe

d on

11

Dec

embe

r 20

03. D

ownl

oade

d by

Uni

vers

ity o

f W

inds

or o

n 30

/10/

2014

04:

03:3

8.

View Article Online

Table 2 Selected bond distances (Å) and angles (�) for 3; C* is the centroid of the arene ring interacting with barium

Ba1–O(1, 2, 3, 1Me1, 1Me2, 1Me2�, 1Me3, 1Me4) 2.582(3), 2.719(3), 2.560(3), 2.820(4), 2.887(3), 3.057(3), 2.739(4),2.808(4)

Ba2–O(1, 3, 4, 6, 1Me5, 2Me6) 2.621(3), 2.619(3), 2.951(3), 3.181(3), 2.757(4), 2.773(4)Ba2–C* 3.09O1–Ba1–O(2, 3, 1Me1, 1Me2, 1Me2�, 1Me3, 1Me4) 77.09(9), 74.4(1), 84.1(1), 154.2(1), 107.62(9), 140.1(1), 67.8(1)O2–Ba1–O(3, 1Me1, 1Me2, 1Me2�, 1Me3, 1Me4) 75.5(1), 154.1(1), 84.5(1), 53.19(9), 131.5(1), 99.0(1)O3–Ba1–O(1Me1, 1Me2, 1Me2�, 1Me3, 1Me4) 82.5(1), 83.6(1), 125.0(1), 132.8(1), 142.1(1)O1Me1–Ba1–O(1Me2, 1Me2�, 1Me3, 1Me4) 106.6(1), 151.8(1), 73.9(1), 89.9(1)1Me2 Ba1–O(1Me2�, 1Me3, 1Me4) 74.1(1), 65.6(1), 133.8(1)1Me2� Ba1–O(1Me3, 1Me4) 81.3(1), 72.0(1)1Me3–Ba1–O1Me4 79.0(1)O1–Ba2–O(3, 4, 6, 1Me5, 2Me6) 72.8(1), 123.2(1), 78.08(9), 101.6(1), 168.1(1)O3–Ba2–O(4, 6, 1Me5, 2Me6) 86.66(9), 119.46(9), 173.9(1), 113.9(1)O4–Ba2–O(6, 1Me5, 2Me6) 67.14(8), 98.6(1), 68.0(1)O6–Ba2–O(1Me5, 2Me6) 60.5(1), 105.26(9)1Me5–Ba2–O(2Me6) 71.2(1)C*–Ba2–O1, O3, O4, O6, O1Me5, O1Me6 81.8, 82.6, 148.3, 143.2, 94.4, 89.3

structure precluded the location of hydrogen atoms, it is unclearwhat other ligands are deprotonated. On one hand, by consider-ing the comparatively long Baendo–Oc distances (2.951(3),3.181(3) Å) it is tempting to assume that these calixarene oxy-gens remain protonated and that two methanol ligands may bedeprotonated. However, based on the higher acidity of phenolscompared to methanol, and rationalising that the relativelylong Baendo–Oc distances could simply be a manifestation ofsteric strain, this assumption remains questionable. Selectedbond distances and angles for 3 are included in Table 2.

The mixed barium–(titanium/zirconium) or homometallicbarium calix[6]arene complexes show that barium can be com-plexed by calix[6]arene even in the presence of methanol, whichis a good solvent for group 2 metal cations. In contrast, ourattempts to synthesise the analogous strontium complexes frommethanol were unsuccessful. Although the reaction of stron-tium metal with calix[6]arene in methanol resulted in the form-ation of a new solid product, addition of titanium isopropoxidedid not afford any isolable mixed metal complexes. This sug-gests that either the structural rearrangements necessary forbinding of titanium are restricted, or that, once titanium iscomplexed, binding of the labile strontium is unfavourable. Forthe larger more polarisable barium cation the structuralrearrangements resulting from titanium binding do notdestabilise the chelation of barium, nor does the binding ofbarium restrict the complexation of titanium. To the contrary,it appears that titanium (or zirconium) binding organizes thecalix[6]arene structure, so that barium � � � π arene solvation isfavoured, suggesting a synergistic binding relationship.

The endo-complexation of barium (but not strontium) withinthe cavity of 1,3-alternate conformation of calix[6]arene frommethanol, presumably relates to the smaller size of Sr2� com-pared to Ba2� (ionic radius 1.18 vs. 1.35 Å). In particular thebetter size complementary of the cavity with the bariumcation allows barium to interact significantly with one aromaticring of the calixarene, thereby achieving a seven coordinatecomplexation environment. For the smaller strontium cationa significant cation aromatic interaction is less likely. This,coupled with its higher polarising strength, could explain whyno mixed strontium–titanium calix[6]arene complexation prod-ucts can be obtained from methanol, as methanol itself couldprovide a preferable complexation environment. Therefore, werationalised that a successful synthesis of mixed metal com-plexes of strontium and titanium would necessitate the attenu-ation of the competitive solvation of strontium by the solvent.For this purpose, we chose a more sterically demanding donoralcohol, isopropanol.

The reaction of strontium metal with calix[6]arene in iso-propanol followed by the addition of [Ti(OPri)4] gave 4, whichhas a structure similar to the Ba–Zr, Ba–Ti calix[6]arenecomplexes, except that three isopropanol ligands and onewater molecule complete the coordination sphere of the two

strontiums, Fig. 3.24 Given the structural similarity of 4 to 1 and2, the structural discussion relating to 4 will be limited to itsunique features. The structure determination of 4 is of limitedprecision due to weakly diffracting crystals, but neverthelessthe atom-to-atom connectivity and spatial arrangement of thecomponents is established. The complex crystallises in the spacegroup P21/n with the molecule comprising the asymmetricunit. Unlike the case for the complexes containing barium, thearomatic rings of the unique non-bridging phenolates arenot involved in Sr � � � π-arene interactions, since the Sr–C[Sr–C(12–17)((A),(B)) = 3.32–4.33 Å, 3.74–4.51 Å] distancesare longer than expected for significant Sr–C interaction (i.e.>3.30 Å). Hence the strontium cations are just six-coordinate.The comparatively low coordination afforded to strontium inthe complex suggests why it could not be prepared in methanol.Selected bond distances and angles for 4 are included in Table 3.

The pale yellow–orange solid of 4 turns deep orange rapidlyupon exposure to water or moist air. This colour change can bereversed by exposing the deep orange solid (dry) to isopropanolliquid. This cycle of colour changes can be reversed at least fivetimes with no apparent onset of irreversibility. In contrast, theorange solid of 1 (the analogous Ba–Ti calix[6]arene complex)did not change colour in air.

This phenomenon may arise from reversible exchange ofpart, or all, of the coordinated isopropanol by water ligand(s),which is not unreasonable considering the X-ray crystal struc-ture of 4 reveals that one strontium coordination site is alreadyoccupied by water. Such coordinative displacement may be

Fig. 3 Molecular projection of 4.

331D a l t o n T r a n s . , 2 0 0 4 , 3 2 7 – 3 3 3

Publ

ishe

d on

11

Dec

embe

r 20

03. D

ownl

oade

d by

Uni

vers

ity o

f W

inds

or o

n 30

/10/

2014

04:

03:3

8.

View Article Online

Table 3 Selected bond distances (Å) and angles (�) for 4

Ti1–O(1A, 2A, 3A, 1B, 2B, 3B) 1.950(23), 1.921(24), 1.999(23), 2.034(23), 1.903(25), 1.972(22)Sr1–O(1A, 3A, 4A, 6A, 1Ip1, 1Ip6) 2.508(23), 2.476(23), 2.615(24), 2.633(22), 2.591(24), 2.509(35)Sr2–O(1B, 3B, 4B, 6B, 1Ip2, W1) 2.498(22), 2.497(23), 2.598(21), 2.626(23), 2.571(22), 2.497(22)O1A–Ti1–O(2A, 3A, 1B, 2B, 3B) 90.9(10), 77.5(9), 161.8(10), 106.4(10), 87.0(9)O2A–Ti1–O(3A, 1B, 2B, 3B) 84.6(10), 82.4(10), 156.4(9), 108.2(10)O3A–Ti1–O(1B, 2B, 3B) 118.4(9), 83.8(10), 160.2(10)O1B–Ti1–O(2B, 3B) 85.3(10), 79.2(9)O2B–Ti1–O3B 89.0(10)O1A–Sr1–O(3A, 4A, 6A, 1Ip1, 1Ip6) 59.5(7), 132.6(8), 96.8(7), 107.3(8), 136.4(14)O3A–Sr1–O(4A, 6A, 1Ip1, 1Ip6) 95.1(8), 136.8(8), 146.0(9), 90.0(13)O4A–Sr1–O(6A, 1Ip1, 1Ip6) 74.5(7), 113.3(9), 75.4(13)O6A–Sr1–O(1Ip1, 1Ip6) 71.9(8), 125.5(14)O1Ip1–Sr1–O1Ip6 80.0(12)O1B–Sr2–O(3B, 4B, 6B, 1Ip2, W1) 61.5(7), 135.6(7), 93.0(7), 146.2(8), 93.3(8)O3B–Sr2–O(4B, 6B, 1Ip2, W1) 96.6(7), 133.6(8), 106.6(8), 138.3(9)O4B–Sr2–O(6B, 1Ip2, W1) 74.6(7), 74.1(7), 122.8(9)O6B–Sr2–O(1Ip2, W1) 113.9(8), 75.6(8)O1Ip2–Sr2–OW1 75.3(8)Sr1–O(1A, 3A)–Ti1 111.6(9), 111.2(9)Sr2–O(1B, 3B)–Ti1 108.4(9), 110.6(9)

favoured by the absence of excess isopropanol, and the form-ation of shorter Sr–O bonds, resulting in better solvation ofstrontium. Also the bound water ligands have less steric hin-drance compared to isopropanol, especially since the strontiumatom is set deeply in the calix[6]arene cone. Such reasoningsuggests that the isopropanol ligand(s) are bound weakly andcould be displaced easily by water ligand(s). Additional evi-dence supporting this rationale is that solutions of 4 are air-sensitive, with dichloromethane solutions unstable in air relativeto solutions of 1, affording a precipitate of native calix[6]areneover ∼1 day. Solutions of 4 in isopropanol were more stable inair, but eventually decomposed, also with precipitation of thenative calix[6]arene.

The 1H NMR spectrum of 4 in CDCl3 gives further insightinto the solution state behaviour described above. In particular,the presence of 12 doublets of varying intensity from methyleneprotons (δ 5.49, 5.33, 4.75, 4.58, 4.48, 4.29, 4.04, 3.43, 3.28,3.16, 3.07, 2.54 ppm) indicates that 4 decomposes in solution.Presumably this relates to the absence of excess isopropanol,necessary to stabilise 4 via the coordinative saturation ofstrontium. For what appear to be similar reasons, complex 3also decomposes in CDCl3, evidenced by the precipitation ofmaterials from solution shortly after the NMR solution wasprepared. In this case, decomposition is probably related tothe absence of excess methanol, required to stabilise 3 by co-ordinative saturation of the barium centres. Notably, a stableNMR solution of 3 could be prepared in a mixture of CD3ODand CDCl3, however the 1H NMR spectrum obtained gaveonly a broad But signal and aryl H signal (1.18 and 7.01 ppm,respectively). This, coupled with the absence of detectablemethylene signals (presumably these are broadened into thebaseline), indicate that the solution state structure of 3 is highlyfluxional. Thus for complexes 3 and 4, no meaningful 1H NMRspectra could be obtained.

ConclusionsIn this paper we have demonstrated that a subtle interplay ofsolvation strength, coordination array type and cavity sizeinfluences the successful preparation of heterobimetallic com-plexes based on calix[6]arenes. Our simple synthetic approachgives access to novel heterobimetallic complexes and thisapproach may lend itself to other combinations of metals,which could include those of other transition metals and group2 ions, or even combinations of transition metals and lan-thanides. Moreover, the presence of protonated O-centres foreach calixarene suggests the possibility of coordinating othermetals and thus building up complexes of higher nuclearity.

ExperimentalAll reactions were carried out in air. [Ti(OPri)4] (97%, HOPri),[Zr(OBun)4] (80%, 1-butanol), barium and strontium metal wereobtained from Aldrich and used as received. Reagent grademethanol and isopropanol were used, and p-tert-butyl-calix-[6]arene was synthesised according to the literature.28

Crystal data

Data for 1–4 were collected at 150(2) K on a Bruker-AXSSMART 1000 CCD diffractometer with Mo Kα radiation,and the structures were solved by direct methods (SIR92)and refined with full matrix least-squares refinements on F(RAELS).29

Synthesis of 1

To a suspension of p-But-L[OH]6 (1.0 g, 1.03 mmol) inmethanol (100 ml) was added barium metal (0.16 g, 1.17mmol). The mixture was stirred, affording a white suspensionon complete consumption of the barium. [Ti(OPri)4] (0.2 ml,97%, 0.65 mmol) and TMEDA (0.15 ml, 1 mmol) were mixedtogether and then diluted in ∼10 ml methanol and this mixturewas added to the barium calix[6]arene complex. Complex 1formed as a pale orange crystalline solid, yield 1.2 g, 88%.Crystals of 1 suitable for X-ray structure determination weregrown from a mixture of dichloromethane and methanol.

Synthesis of 2

To a suspension of p-But-L[OH]6 (1.0 g, 1.03 mmol) in meth-anol (100 ml) was added barium metal (0.16 g, 1.17 mmol). Themixture was stirred affording a white suspension on completeconsumption of the barium. [Zr(OBun)4] (0.3 ml, 0.66 mmol)was added to the barium calix[6]arene complex. Complex 2formed as a gelatinous white solid, yield 0.7 g, 62%. Afterthorough removal of methanol in vacuo, crystals of 2 suitablefor X-ray structure determination were grown from a dichloro-methane solution overlaid with acetonitrile and a small amountof dimethoxyethane.

Synthesis of 3

To a suspension of p-But-L[OH]6 (1.0 g, 1.03 mmol) inmethanol (100 ml) was added barium metal (0.35 g, 2.55mmol). The mixture was stirred affording a white suspensionon complete consumption of the barium. The mixture waswarmed and dichloromethane was added until the solid wasdissolved completely. On slow cooling, colourless prisms of 3were deposited (1.4 g, 90%).

332 D a l t o n T r a n s . , 2 0 0 4 , 3 2 7 – 3 3 3

Publ

ishe

d on

11

Dec

embe

r 20

03. D

ownl

oade

d by

Uni

vers

ity o

f W

inds

or o

n 30

/10/

2014

04:

03:3

8.

View Article Online

Synthesis of 4

To a suspension of p-But-L[OH]6 (1.0 g, 1.03 mmol) in isopro-panol (100 ml) was added strontium metal (0.1 g, 1.03 mmol).The solution was refluxed until the strontium metal dissolvedcompletely. The addition of [Ti(OPri)4] (0.2 ml, 97%, 0.65mmol) gave an orange solution which was filtered and evapor-ated to dryness in vacuo. The solid was redissolved in the mini-mum amount of dichloromethane, filtered and evaporated todryness to yield 4 (1.0 g, 75%). Crystals suitable for X-raystructure determination were obtained from a mixture ofdichloromethane and isopropanol.

1H NMR (CDCl3, 300 MHz, 298 K): 1 (sample dried invacuo) δ 7.15 (s, 4 H, aryl), 7.02 (s, 8 H, aryl), 7.01, (s, 4 H, aryl),6.89 (s, 4 H, aryl), 6.54, (s, 4 H, aryl), 4.95 (b, 4 H, CH2), 4.36(d, 4 H, J = 15 Hz, CH2), 3.97 (d, 4 H, J = 15 Hz, CH2), 3.28(d, 4 H, J = 15 Hz, CH2), 3.19 (d, 4 H, J = 15 Hz, CH2), 3.06(b, 4 H, CH2), 1.25 (s, 54 H, But), 1.20 (s, 18 H, But), 1.15(s, 36 H, But). 2 δ 7.16 (d, 4 H, J = 2.3 Hz, aryl), 7.12 (s, 4 H,aryl), 7.02, (d, 4 H, J = 2.3 Hz, aryl), 7.00 (s, 4 H, aryl), 6.84,(d, 4 H, J = 2.3 Hz, aryl), 6.47 (d, 4 H, J = 2.3 Hz, CH2), 5.18(d, 4 H, J = 15 Hz, CH2), 4.04 (d, 4 H, J = 15 Hz, CH2), 3.96(d, 4 H, J = 15 Hz, CH2), 3.56 (s, 12 H, CH2, DME), 3.35(d, 4 H, J = 15 Hz, CH2), 3.28 (d, 4 H, J = 15 Hz, CH2), 2.96(s, CH3, 18 H, DME), 2.94 (d, 4 H, J = 15 Hz, CH2), 1.23(s, 72 H, But), 1.09 (s, 36 H, But).

Crystal data

1, 0.5(C132H160Ba2TiO12)�1.25(CH2Cl2)�3(CH3OH): M = 1332.9,monoclinic, a = 14.674(3), b = 14.360(3), c = 33.624(7) Å,β = 99.936(4)�, U = 6979(4) Å3, T = 150 K, space group P21/n,Z = 4, µ(Mo Kα) = 0.767 mm�1, 69102 reflections measured,16640 unique (Rint = 0.040). The final R = 0.048, and wR = 0.060(observed data).

2, 0.5(C132H156Ba2O12Zr)�3(C4H10O2): M = 1399.0, mono-clinic, a = 15.237(5), b = 14.258(5), c = 33.827(11) Å, β =101.442(6)�, U = 7203(7) Å3, T = 150(2) K, space group P21/n,Z = 4, µ(Mo Kα) = 0.658 mm�1, 69660 reflections collected,17224 unique (Rint = 0.038). The final R = 0.044, and wR = 0.060(observed data).

3, C66H80Ba2O6�7(CH3OH)�0.5(H2O)�0.5(CH2Cl2): M =1519.8, triclinic, a = 14.741(3), b = 16.691(3), c = 17.982(3) Å,α = 79.799(3) β = 74.419(3), γ = 64.967(3)�, U = 3851(2)Å3, T =150(2) K, space group P1̄, Z = 2, µ(Mo Kα) = 1.105 mm�1,32786 reflections collected, 17600 unique (Rint = 0.050). Thefinal R = 0.044, and wR = 0.070 (observed data).

4, C132H160O12Sr2Ti�7(C3H7O)�0.5(CH2Cl2)�H2O: M = 2636.0,monoclinic, a = 21.131(7), b = 36.571(11), c = 22.514(7) Å,β = 112.358(6)�, U = 16090(14) Å3, T = 150(2) K, space groupP21/n, Z = 4, µ(Mo Kα) = 0.759 mm�1, 91816 reflectionscollected, 27751 unique (Rint = 0.085). The final R = 0.121, andwR = 0.184 (observed data).

CCDC reference numbers 205214–221944.See http://www.rsc.org/suppdata/dt/b3/b312919e/ for crystal-

lographic data in CIF or other electronic format.

Acknowledgements

We thank Dr Peter Turner for his assistance with crystallo-graphy.

References1 V. G. Kessler, L. G. Hubert-Pfalzgraf, S. Daniele and A. Gleizes,

Chem. Mater., 1994, 6, 2336.2 M. Veith, S. Mathur and V. Huch, J. Am. Chem. Soc., 1996, 118,

903.3 M. Veith, S. Mathur, V. Huch and T. Decker, Eur. J. Inorg. Chem.,

1998, 1327.4 M. Veith, S. Mathur and C. Mathur, Polyhedron, 1998, 17, 1005.5 M. Veith, S. Mathur, N. Lecerf, V. Huch, T. Decker, H. P. Beck,

W. Eiser and R. Haberkorn, J. Sol-Gel Sci. Technol., 2000, 17,145.

6 L. G. Hubert-Pfalzgraf, Inorg. Chem. Commun., 2003, 6, 102.7 L. G. Hubert-Pfalzgraf, Coord. Chem. Rev., 1998, 178–180, 967.8 S. K. Park and S. M. Koo, Bull. Korean Chem. Soc., 1999, 20, 1115.9 V. W. Day, T. A. Eberspacher, W. G. Klemperer and S. Liang, Chem.

Mater., 1995, 7, 1607.10 G. W. Gokel, L. J. Barbour, S. L. De Wall and E. S. Meadows, Coord.

Chem. Rev., 2001, 222, 127.11 G. W. Gokel, L. J. Barbour, R. Ferdani and J. Hu, Acc. Chem. Res.,

2002, 35, 878.12 G. B. Deacon, C. M. Forsyth and P. C. Junk, J. Organomet. Chem.,

2000, 607, 112.13 L. D. Durfee, P. E. Fanwick and I. P. Rothwell, Angew. Chem., 1988,

100, 1218.14 A. J. Petrella, N. K. Roberts, D. C. Craig, C. L. Raston and

R. N. Lamb, Chem. Commun., 2003, 1728.15 A. J. Petrella, N. K. Roberts, D. C. Craig, C. L. Raston and

R. N. Lamb, Dalton Trans., 2003, 4590.16 Calixarenes 2001, ed. Z. Asfari, V. Bohmer, J. Harrowfield and

J. Vicens, Kluwer, Dordrecht, 2001.17 C. Wieser, C. B. Dieleman and D. Matt, Coord. Chem. Rev., 1997,

165, 93–136.18 J. L. Atwood, M. G. Gardiner, C. Jones, C. L. Raston, B. W. Skelton

and A. H. White, Chem. Commun., 1996, 2487–2489.19 V. C. Gibson, C. Redshaw, W. Clegg and M. R. J. Elsegood, Chem.

Commun., 1997, 1605.20 A. Zanotti-Gerosa, E. Solari, L. Giannini, C. Floriani, A. Chiesi-

Villa and C. Rizzoli, Chem. Commun., 1997, 2, 183–184.21 F. A. Cotton, E. V. Dikarev, C. A. Murillo and M. A. Petrukhina,

Inorg. Chim. Acta, 2002, 332, 41–46.22 G. Guillemot, E. Solari, C. Rizzoli and C. Floriani, Chem. Eur. J.,

2002, 8(9), 2072–2080.23 A. J. Petrella, N. K. Roberts, D. C. Craig, C. L. Raston and

R. N. Lamb, Chem. Commun., 2003, 2288.24 W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics, 1996,

14.1, 33–38, http://www.povray.org.25 M. G. Gardiner, G. A. Koutsantonis, S. M. Lawrence, P. J. Nichols

and C. L. Raston, Chem. Commun., 1996, 17, 2035–2036; J. L.Atwood, S. G. Bott, C. Jones and C. L. Raston, Chem. Commun.,1992, 18, 1349–51.

26 A. J. Petrella, N. K. Roberts, D. C. Craig, C. L. Raston andR. N. Lamb, Chem. Commun., 2003, 1014.

27 W. G. Bartley and W. Wardlaw, J. Chem. Soc., Abstr., 1958, 422.28 C. D. Gutsche, B. Dhawan, M. Leonis and D. Stewart, Org. Synth.,

1990, 68, 238.29 A. D. Rae, RAELS. A Comprehensive Constrained Least Squares

Refinement Program, University of New South Wales, 1989.

333D a l t o n T r a n s . , 2 0 0 4 , 3 2 7 – 3 3 3

Publ

ishe

d on

11

Dec

embe

r 20

03. D

ownl

oade

d by

Uni

vers

ity o

f W

inds

or o

n 30

/10/

2014

04:

03:3

8.

View Article Online