The structures of strontium tellurite and strontium telluride aluminate sodalites studied by powder neutron diffraction, EXAFS, IR and MAS NMR spectroscopies Sandra E. Dann and Mark T. Weller Department of Chemistry, The University of Southampton, Highfeld, Southampton, UK SO1 7 1BJ Aluminate sodalites containing the telluride and tellurite ions have been synthesised for the first time. The structures of Sr, [A10,],,( TeO,), and its reduced product Sr, [AlO,] 12Te, have been determined using Rietveld refinement of powder neutron diffraction data. In Sr, [A102]i2Te2 the anion occupies a position at the sodalite cage centre, whilst in Sr, [A102]12(Te03)2 the larger tellurite ion is displaced to allow reasonable coordination to the strontium ions. The reduction of the occluded TeO,,- to Te2- has also been investigated using IR spectroscopy, 27Al and 125Te MAS NMR spectroscopies and Te L,,,-edge EXAFS.The sodalites are a well known class of anion-containing framework consisting of P-cages, formed from TO4 (T =Be, Si, Ge, Al, Si, Ga, etc.) tetrahedra, directly linked through the six- membered rings and containing a centrally placed anion tetrahedrally coordinated to four cations (Fig. 1). The composi- tion of sodalite structures, with general formula [MnI8+ [{ TTO, ],I6- -[X,I2-,is very diverse;'-, many different cat- ions and anions (M=Na+, Ca2+, Kf, Ag+, Sr2+, etc.; X= C1-, Br-, S2-, Se2-, Mn04-, NO3-, etc.) can be accommo- dated in the sodalite cage through partial collapse of the framework, by tilting the tetrahedra out of their normal planes and by deviation from perfect tetrahedral geometry around T( T').The guest-host interactions of the framework structure with the occluded species are exploited in pigments where the framework stabilises unstable or transient species. Of note, as members of this class, are the well known ultramarines which contain and stabilise the polysulfide radicals S2-and S3-and the corresponding selenium analogues. Recently attention has centred on the occlusion of semiconducting particles in frame- work structures by encapsulation of S2-and Se2- and divalent cations; however, the telluride (Te2-) ion has been less well studied. Sodalite synthesis has been achieved in several ways includ- ing hydrothermal methods,'~~ high-temperature sintering5 and structural conversion.6 After phase synthesis, exchange and Fig.1 Depiction of sodalite structure showing the central anion (large sphere) tetrahedrally coordinated by the cations (dark spheres) occluded within the sodalite framework reaction of the anions and cations entrapped in the sodalite cage is also possible without decomposition of the surrounding framework, e.g. SCN- decomposition to form S3-and S2-in the synthesis of ~ltramarine.~ The group of materials where T=T'=Al are the aluminate sodalites with general formula M, [AlO,],,X, (M =Ca2+, Sr2+, Cd2+; X=S04,-, W042-, Cr04,-, MOO^^-, S2-). Many of these materials have been extensively studied by Depmeier,,-lo particularly with respect to the orientation of the tetrahedral anions located at the centre of the P-cage unit. These materials are also of note as they disobey Lowenstein's rule of aluminium avoidance'' and form a group of aluminium- rich structures along with bicchulite, Ca, [A1,Si06],(OH), .Owing to the high aluminium content, highly acidic frame- works are produced, which may have potential uses in catalysis and ion exchange. In recent studiesi2*13 it was found that sulfate- and selenite-bearing sodalites could be reduced to yield sodalites containing sulfide and selenide by reaction under hydrogen. This reaction proceeds by an intracage process, i.e. the framework remains intact during modification of the anion. In this paper we describe the synthesis of the first aluminate sodalites bearing tellurium-containing anions and their struc- ture determination using a combination of neutron diffraction, EXAFS, IR and MAS NMR spectroscopies.Experimental The parent compound Sr, [AlO,] 12( Te03), was prepared using solid-state synthesis at high temperature as explored initially by Kondo14 and Depmeier." SrO (99.95%), TeO, (99.9%) and A1203 (99.5%) were ground homogeneously using a pestle and mortar in a glove box. The reaction mixture was then sealed in a silica tube (to prevent volatisation of TeO,, which would occur at 900 "C, below the reaction temperature of alumina) and heated to 1200 "C in a tube furnace for 48 h. The tube was then slow-cooled to room temperature. The product crystallised as a fine white powder whic! was shown to be a single cubic phase with a=9.425( 1) A by powder X-ray diffraction.Sr8[A10,],,Te2 was prepared by the reduction of Sr8[A10,],,(Te0,), in a stream of hydrogen at 850 "C for 8 h. The mass loss of 5.3% was in good agreement with the expected value of 5.45% for the reaction: Sr, [A10,],,(Te0,), +ST, [A1O,],,Te2 +6H20 The product recrystallised as a light-brown powder which was shown to be a single cubic phase by powder X-ray diffraction with a =9.379(1)A. IR spectra were obtained for both compounds on a Perkin- J. Muter. Chem., 1996,6(10), 1717-1721 1717 67.341 (a) A I EXAFS *$ 1100 900 700 500.-c c 0.33 Fig. 2 Part of the IR spectra of Sr,[AlO,],,Te, (a) and Sr8[A102]12(Te03)2(b) showing loss of the Te032- v1 stretch at 758 cm-' on reduction to Tez- sodalite in hydrogen.The aluminate cage bands are also assigned as vBs, vsl, vSz. Elmer FTIR 1710 spectrometer with a 3600 data station on pressed KBr discs. Reduction of the parent sodalite showed loss of the v1 stretch of Te03,- at 758 cm-', as seen in Fig. 2. The other absorptions in the IR spectra were assigned to the four IR-active bands expected for aluminate sodalites, which will be discussed in detail later (uide infra). Te L,,,-edge EXAFS were collected in transmission mode on station 7.1 at the SRS laboratory at Daresbury on a sample of Sr, [A102]12(Te03)2 diluted to 5% in mass with boron nitride. Data reduction and background subtraction was undertaken using PAXAS16 and data analysis using EXCURV92.I7 Fourier transform of the Te-edge EXAFS data cn the TeO,,--containing sodalite gave a single peak at 1.845 A.Best-fit parameters were achieved for three nearest neighbours at this distance and gave an R, value of 24.2%. The fit to the data is shown in Fig. 3. Solid-state 125Te and 27Al MAS NMR spectra were collected using a spin rate of 12 kHz at room temperature referenced to dimethyl tellurium and 1 mol dm-, AlCl, solution, respect- ively, on a Varian VXR300 spectrometer fitted with a 7T superconducting magnet at the University of Durham. 27Al NMR spectra were recorded using a field of 78.158 MHz and a 1 ps pulse corresponding to a 10" pulse in the reference solution of 1 mol dm-, AlCl, over 2000 repetitions.lZ5Te NMR spectra were recorded using a field of 94.45 MHz with a 1 ps pulse width and a relaxation delay of 30 s. Typically 1000 scans were performed to achieve good signal-to-noise ratios. The 125Te and 27Al NMR patterns for Sr8[A102]12Te, are shown in Fig.4. A sharp resonance is observed in the 125Te NMR pattern at 6-1372 which is characteristic of the Te2- anion. In contrast the tellurite ion resonates at 6 + 1742. Although there are very few lZ5Te data available for compari- son, this shift appears in good agreement with that of K2Te03 which has a lZ5Te resonance at 6-t 1732.18 The 27Al NMR spectra showed the characteristic shapes expected for a quadru- polar nucleus in a slightly distorted environment with incom- plete averaging of quadrupolar coupling. However, the distortion of the tetrahedral geometry in the aluminate soda- lites is relatively small and the isotropic chemical shift can be obtained by fitting the spectrum for various quadrupolar coupling constants.19 Simulation of the experimental spectra allowed calculation of the isotropic chemical shift, giving values of 6 75.7 and 75.2, respectively, for the telluride- and tellurite- containing sodalites.1718 J. Mater. Chern., 1996, 6(lo), 1717-1721 rlA Fig. 3 Fit to the Fourier transform of the Te L,,,-edge data for best fit to three oxygens at 1.845 A: (-) experimental data; (---)theoretical data 250 150 50 -50 -150 , -1000 -1400 -1800 -2200 -2600 6 Fig. 4 (a) 27Al and (b) 125Te MAS NMR spectra obtained from Sr8[A102]12Te2.Details of data collection are given in the text Structure determination Time-of-flight neutron diffraction data were obtained for Sr, [A10,],,( TeO,), and Sr, [AlO,] ,,Te2 at room temperature over the d-spacing range 3.08-0.568 A using the medium- resolution high-flux instrument, POLARIS, at the Rutherford Appleton Laboratory. Both patterns could be indexed on a body-centred cubic structure with systematic absences consist- ent with the space group 143~2,as expected for a sodalite with a single tetrahedral framework species. Rietveld refinement analysis was performed in this space group using the pro- gram GSAS.20 Starting models were taken from Brenchley and Weller13 and used as the basis for both refinements with Te on the 2a (O,O,O) site, Sr on 8c (x,x,x) x=O.24, aluminium on 12d (1/4,1/2,0) and the cage oxygen on the 24g (x,x,z) x=O.15, z= 0.47.For the tellurite the oxygen of the tellurite group was initially positioned on 24g (x,x,z) x=0.38 and z=0.51, which allowed good coordination to the cage and was chosen by Depmeier for chromite sodalite' and Brenchley and Weller', for selenite sodalite. Neutron scattering lengths were taken as Te 0.580, Sr 0.702, A1 0.3449 and 0 0.5805 x cm. Data were normalised and corrected for sample absorption. Initial stages of the refinement proceeded well in both cases including scale factor, cell constant and a five-point polynomial back- ground. For the tellurite sodalite a more multi-term complex cosine Fourier background was required to fit modulations in the background, probably associated with thermal diffuse scattering from the reorienting Te032- group.Four small excluded regions were also included in the background to remove reflections from poorly crystalline SiO,, from the quartz reaction tube, which were not apparent in the X-ray diffraction pattern. In the case of the telluride sodalite, all variable atomic positions and peak-shape parameters varied smoothly and finally the isotropic temperature factors were added. In the case of the tellurite, although the parameters defining framework geometry refined smoothly the coordinates delineating the occluded species were unstable suggesting the initial model to be incorrect.A better structural model for sodalites containing large asymmetric anions has been proposed by Mead and Weller21 for the Br0,2--containing sodalites and by Felsche and co- workers for acetate sodalite.22 In this description, unrealistically short interactions between the anions and the non-framework cation, M, are alleviated through the introduction of a second type of site for M. This effectively allows displacement of M through a six-ring into a neighbouring cage such that M coordinates only to one side of the anion, Fig. 5. This structural model involves placing a portion of the strontium atoms on (x,x,x) and the remainder on (-xl,-xl,-xl) and this model was therefore adopted in this refinement with ~~0.25 and XI 250.2.Refinement of the strontium atoms on these sites gave a vastly improved fit to the profile with significantly better R \ 'i // Fig.5 Proposed model for the cation displacement caused by the introduction of a large pyramidal anion into the sodalite structure factors. Refinement of the occupancy of the two sites gave 0.504(9) and 0.496(9), respectively. These values are somewhat further from the 0.75/0.25 ratio expected in the simple model proposed initially, where only one quarter of the strontium ions need to be displaced to accommodate the Te0,2- ion. However, the refined values indicate that further strontium ions are displaced to positions that coordinate to the oxygen side of the tellurite ion. Further improvement of the structural model involved dis- placement of the tellurium atom of the TeO, group to an (x,x,x)site with x =0.05, which produces four-fold disordering.This also produced a significant improvement in the fit and is also expected in order to allow room for oxygen atoms within the beta cage. The expected geometry of the tellurite anion is pyramidal; using the information gleaned from the tellurium-edee EXAFS, the three oxygens were expected to be located 1.84 A from the refined tellurium site. Evidence for this rigid body was then searched for in a difference Fourier map calculated using the structure factors obtained from fitting the data with just the framework atoms, strontium and tellurium. Fig. 6 shows one calculated section and ,a possible site for oxygen at (0.12, -0.01, 0.125) about 2A from tellurium and similar to that used by Brenchley and Weller in their model for seo3,12313 that is (x,x,z) [equivalent to (x,z,x)] with x 250.12, z =0.05.Oxygen was introduced into the refinement of this site and a slack cqnstraint introduced setting the Te- 0 distance to 1.84( 1) A as required by the EXAFS data. Refinement of the oxygens on this site, again with four-fold disordering, produced an excellent fit to the d$a though a high temperature factor on the oxygen of 7.4A2; this high temperature factor is probably due to librations of the tellurite group. Similarly, higher temperature factors are generally observed for the oxygens of other tetrahedral and pyramidal occluded species.23 This choice of site for the oxygen of the tellurite group produces 12 possible sites with the unit cell; of these six produce reasonable Te- 0 distances and these six sites may be assigned to two orientations of the TeO, anion.Of note is the 0-Te-0 angle of 102" produced for these two choices, which is in very good agreement with data from other tellurite structures; in Na2T?O3-5H2024 the tellurite group has three oxygens at 1.86 A and an 0-Te-0 angle of 99.5". Final refined param- eters are given in Table 1 and 2 for the telluride and tellurite, 0.5 -0,5 -0.5 0 0.5 X Fig. 6 Section, z =0.125, of the neutron diffraction difference Fourier map of Sr,[A10,],,(Te0,)2 showing a possible site near 0.11, 0.00, 0.125 for the tellurite oxygen J.Mater. Chew., 1996, 6(lo), 1717-1721 1719 Table 1 Refined parameters for Sr, [A102]12Te,; e.s.d.s are given in parentheses site X Y Z B,,, 12d 24g 8c 2a 0.25 0.15627(8) 0.20665(8) 0 0.5 0.15627(8) 0.20665(8) 0 0 0.46618(7) 0.20665(8) 0 0.18 (4) 0.70(2) 0.94(2) 1.54(6) 1111 I I I I I I I 0.4 016 018 1.0 1.2 1.4 1.6 1:s TOF/ms Fig. 7 Final fit to neutron diffraction data for Sr8[A10,],,(Te0,),. atom A1 Sr Te e .-ci’-ol0.0 Fig. 8 Refined orientation of the Te03,- anion in Sr8[A102],2(Te03), showing its interaction with the strontium ions. Only one possible position for the tellurite ion is shown. Table 3 Derived bond distances and angles for Sr, [AlO,],,Te, and Sr, [A10,],,(Te03), sodalites bond distance/& The crosses represent the experimental data points and the upper bond angle/degrees Sr, [AlO,] 12Te, ,,( Sr, [AlO,] TeO,), continuous line the calculated pattern.The lower continuous line represents the difference. Tick marks for the reflection positions are also shown. respectively, and the final fit to the experimental profile is shown in Fig. 7 for the tellurite sodalite. Results and Discussion Bond distances and angles calculated from the refined atomic positions are given in Table 3. In Sr,[AlO,],,Te, the central telluride is tetrahedrally coordjnated to four strontium atoms. The bond distance of 3.355 A is qslightly shorter than that observed for bulk SrSe of 3.401 A, which has the rock-salt structure and hence octahedral coordination to the metal.This behaviour has also been observed in calcium sulfide, strontium sulfide and strontium selenide-bearing sodalites and is presum- ably related to the smaller effective ionic radius of the anion in four-fold, as opposed to the more normal six-fold, coord- ination. The strontium coordination approximates to tetra- hedral with three strong bonds to the oxide ions in the six- ring and a single bond to the central tellurium. There are also three much weaker bonds to the other three oxygens in the ring. The 3+3 coordination to the ring is caused by partial collapse of the cage around the occluded anion. The structure of the tellurite sodalite is more complex due to the large anion which has lower symmetry and is disordered.Placement of the pyramidal TeO, with the tellurium at the cage centre (O,O,O) would make it impossible to site four strontium ions within the unit cell to produce chemically reasonable Sr-0 and Sr-Te distances. For this reason the ~ Al-0 Sr-0 Sr-Te Te-0(Te) Sr-0(Te) Sr(2)-0 0-O(Te) 0-Al-0 x 2 0-Al-0 x 4 A1-0-A1 tilt angle, 4 0(Te)-Te-0(Te) ~~ ~ ~~ ~ 1.737( 1) 2.523( 1) x 3 2.896( 1) x 3 3.355( 1) ----119.30(8) 104.83(8) 145.09(8) 12.21(8) -~~~~ 1.723(4) 2.581 (4) x 3 2.740( 5) x 3 -1.842(2) 3.026( 6) 3 x 0.25 2.791(6) 3 x 0.25 2.794( 6) x 3 2.716(4) x 3 2.979( 5) 119.46( 12) 104.72( 5) 150.22( 12) 2.75( 12) 102.00( 2) tellurite ion is displaced off the cage centre towards one of the six-rings. Obviously a strontium ion located at (x,x,x), ~~0.2,woul4 be unreasonably close to the Te4+ centre (Sr-Te M 2.3 A).Therefore this ion is displaced into the neigh- bouring cage, where presumably it is located on the oxygen side of a TeO, group. Hence in the tellurite sodalite there are two strontium environments to be considered. The strontium centred at (0.25, 0.25, 0.25) has a very similar coordination to that observed for the selenite-containing aluminate sodalite. There are three short and three long distances to the oxides in the six-ring plus two contact distances to the oxygens of the tellurite group. The other strontium atom near (-0.2, -0.2, -0.2) has only six bonds to the ring oxygens It is noteworthy that the two strontium positions and their refined occupan- Table 2 Refined atomic parameters for Sr, [AIO,],,( TeO,),; e.s.d.s are given in parentheses ~ ~~ ~ atom site X Y Z B,,, occupancy A1 12d 0.25 0.5 0 0.80( 11) 1 0.15795( 14) 0 24gSr 8c 0.2426( 3) 0.15795( 14) 0.2426( 3) 0.4924( 5) 0.2426(3) 2.45(6) 1.09( 12) 1 0.504(9) Te 8c 0.0599(2) 0.0599( 2) 0.0599( 2) 2.45(4) 0.25 Sr 8c -0.2027(4) -0.2027(4) -0.2027( 4) 2.54( 16) 0.496( 9) O(Te) 24g 0.1228( 2) 0.1228( 2) 0.0302(3) 7.40( 44) 0.25 ~~ Q =9.42061(8)A; R,, =2.76%; Rexp=0.83%; x2 =11.18.1720 J. Muter. Chem., 1996, 6(10), 1717-1721 cies are very similar to those observed by Engelhardt and co- workers22 for the single-crystal structure determination of acetate sodalite. The cell collapse is almost totally performed by variations in the Al-0-A1 bond angle which can take values between 120 and 160" and is generally larger than that for aluminosilic- ate sodalites where bond ionicites are lower. The degree of collapse can be described by several values as investigated by De~meier.~~Of importance are a and a' which are the tetra- hedral distortion angles for the 0-Al-0 angles in the AlO, tetrahedron, which can distort considerably from the perfect 109.48".The more important term in defining cell collapse is the tilt angle, 4, which describes how much the tetrahedra have tilted out of their 4 axis. 4 can hold values between 0 and 35", where 0 equates to a fully expanded centrosymmetric sodalite with symmetry lm3rn.It can be seen that on changing the cell contents from tellurite to telluride the tilt angle increases from 2.75" to 12.21" to accommodate the smaller anion. The IR spectra of aluminosilicate sodalites have been mod- elled extensively by Creighton et These authors showed that of the fourteen expected IR-active modes of aluminate sodalites, only eight had detectable intensity and only four of these were experimentally observed over the range 300-1000 cm-' with moderate intensity. These four bands have been denoted v,,, vS1, vS2 and 6 for the asymmetric, two symmetric and deformation modes, respectively. The frequency of these bands in aluminate sodalites have been shown to linearly correlate with the cell parameter13 and the IR absorp- tion bands at 878,676,620 and 397 cm-I for the telluride and 884, 718, 622 and 395 cm-I for the tellurite are fully consistent with the previously determined data.Previous studies on the correlations between 27Al MAS NMR spectra and the coordination geometry of framework al~minates~~'~~,~~have shown a linear relationship between the quadrupolar coupling constant (C,) and the tetrahedral distortion, $, measured according to the equation =Itanlax- 109.481 X where ax is one of the six tetrahedral angles (0-A1-0) around the aluminium. Values of C, were determined as C,= 5.91 and 5.96 MHz for the telluride and tellurite sodalites, respectively, from the fitting of the spectra.$ was calculated as 0.672 and 0.685 for the telluride and tellurite sodalites from the structure refinements. These results are in very good agreement with values expected by interpolation of the pre- viously determined re1ati0nships.l~ In addition a linear corre- lation between the 27Al isotropic chemical shift (hiso) and the A1-0-A1 bond angle has been determined" and the values reported in this work [S 75.7, 145.1" (Te2-); 6 75.2, 150.2" (Te032-)] fit this behaviour well. Conclusions The tellurite ions has been incorporated into the sodalite structure for the first time. Previous reports of oxotellurium anions in sodalite structures are limited to the compound Ca2Na, [AlSiO,],( TeO,),, a hauyne derivative reported by Neurgaonkar and H~mmel.~~ However, as the free Te042- ion is not known it is likely that this compound was incorrectly described and is in fact a tellurite.Indeed, our attempts to produce sodalites containing the Te042- ion have been unsuc- cessful, producing tellurites even under very strongly oxidising conditions. Structure determination of sodalites containing complex anions requires the application of a variety of techniques in order to define the species present. In this work a combination of MAS NMR, EXAFS and neutron diffraction was required in order to define the nature and position of the oxoanion. Reduction of tellurite ion to telluride can be accomplished inside the sodalite cage, showing that intracage reactions are accomplished easily.Attempts to oxidise tellurite to tellurate by heating in a stream of oxygen at various temperatures were unsuccessful. As the sodalite structure often stabilises tetra- hedral species, e.g. permanganate is stable to 500°C inside a sodalite cage,30 it might be expected that the tellurate ion could be generated inside the aluminate sodalite. However, the framework of Sr, [A102]12(Te03)2 is rather expanded (Al-0-A1 angle 150.2", tilt angle 2.75") and further expansion to accommodate the tellurate ion is impossible. We thank the EPSRC for a grant in support of this work, David Apperley at the solid-state NMR service at Durham for running the 27Al and lzsTe NMR spectra and Dr. Richard Oldroyd for collection and refinement of the Te Lm-edge data.References 1 R. M. Barrer and J. F. Cole J. Chem. SOC. A, 1970, 1516. 2 D. Taylor and C. M. B. Henderson, Phys. Chem. Miner., 1978, 2, 325. 3 M. T. Weller and G. Wong, J. Chem. SOC., Chem. Commun., 1988, 1103. 4 S. E. Dann and M. T. Weller, Inorg. Chem., 1996,35, 555. 5 J. S. Prener and R. J. Ward, J. Am. Chem. Soc., 1988,72,2780. 6 I. Chang, J. Electrochem. Soc., 1974, 121, 815. 7 F. Hund, Z. Anorg. Allg. Chem., 1984,511,225. 8 W. Depmeier, Acta Crystallogr., Sect. B, 1988,44,201. 9 W. Depmeier, Acta Crystallogr., Sect. C, 1987,43,2251. 10 W. Depmeier, Acta Crystallogr., Sect. C, 1984,40,226. 11 W. Lowenstein, Am. Mineral, 1954,39,92. 12 M. E. Brenchley and M. T. Weller, J. Muter. Chem., 1992,2, 1003.13 M. E. Brenchley and M. T. Weller, Chem. Muter., 1993,5,970. 14 R. Kondo, Yogyo Kyokai Shi., 1965,71, 1. 15 W. Depmeier, Kristall. Technik., 1972, 7, 229. 16 N. Binsted, PAXAS: EXAFS Analysis Program, 1988. 17 N. Binsted, S. J. Gurman and I. Ross, J. Phys. C.: Solid State Phys., 1984, 17, 143. 18 R. K. Harris and B. E. Mann, NMR and the Periodic Table, Academic Press, London, 1978. 19 M. T. Weller, M. E. Brenchley, D. C. Apperley and N. A. Davies, Solid State Nucl. Magn. Reson., 1994,3, xxx. 20 A. C. Larson and R. B. Von Dreele, GSAS: Generalised Structural Analysis System, Los Alamos, NM, 1990. 21 P. J. Mead and M. T. Weller, Zeolites, 1995, 15, 561. 22 P. Sieger, A. M. Scheider, M. Wiebcke, P. Behrens, J. Felsche and G. Englehardt, Chem. Muter., 1995,7, 163. 23 M. E. Brenchley and M. T. Weller, Zeolites, 1994, 14, 682. 24 E. Philippot, M. Maurin and J. Moret, Acta Crystallogr, Sect. B, 1979,35,1337. 25 W. Depmeier, Acta Crystallogr., Sect. B, 1984,40, 185. 26 J. A. Creighton, H. W. Deckman and J. M. Newsam, J. Phys. Chem., 1991,95,2099. 27 G. Engelhardt, H. Koeller, P. Sieger, W. Depmeier and A. Samoson, Solid State. Nucl. Magn. Reson., 1992, 1, 127. 28 G. Engelhardt and W. J. Veeman, J. Chem. SOC., Chem. Commun., 1993,622. 29 R. R. Neurgaonkar and F. A. Hummel, Muter. Res. Bull., 1976, 11,61. 30 M. T. Weller and K. E. Haworth, J. Chem. SOC.,Chem. Commun., 1991,734. Paper 6/03606F; Received 23rd May, 1996 J. Mater. Chem., 1996,6(lo), 1717-1721 1721