J. MATER. CHEM., 1991, 1(1), 43-49 43 Mechanism of n-Alkylammonium Ion Intercalation into the Layered Host a-VOP04*2H20 Michael Morris,aJohn M. Adamsband Alan Dyer*" a Department of Chemistry and Applied Chemistry, The University of Salford, Salford M5 4WT, UK English China Clays International Ltd., John Keay House, St. Austell, Cornwall PL25 4DJ, UK The mechanism of redox intercalation reactions between alkylammonium iodides and the layered host a-VOP02-2H20has been investigated by synthesis of mixed- and single-ion intercalates, by ion-exchange experiments, and by EPR study of in situ reactions. Intercalation was found to occur first at the edge of crystallites. With alkyl chains of butyl and longer, intercalation of ions was seen to proceed throughout the interlayer galleries, resulting in a bilayer arrangement, with the chains making an average angle of 39"to the host layers.Smaller ions were not intercalated to any great degree, with reaction occurring only on crystal edges and faces. Intercalation of these smaller ions at crystal edges in mixed intercalation reactions prevented larger, co-present octylammonium ions from intercalating, indicating that they represented a barrier to free interlayer diffusion of further reactant. This barrier is suggested as being the result of smaller ions intercalating parallel to the host layers at crystallite edges. Keywords: Intercalation; Alkylammonium iodide; a-Vanadyl phosphate dihydrate The intercalation of guest molecules in layered hosts has received considerable attention in recent years'-4 owing to the potential that such materials show for catalysis, molecular sieving, ion exchange, and so on.Perhaps the most studied of guest molecules are n-alkylamines (whether protonated or electrically neutral), which have a remarkable ability' to penetrate the interlamellar regions of layered substances, an activity thought to arise from the combination of the polarity of the primary amino group and the ability of these molecules to alter their molecular shape by conformational changes of the alkyl chain^.^,^ Alkylamine intercalates of layered hosts display interesting structural properties, with multilayers of intercalated guest molecules being Such multi- layers are thought to be favoured by numerous van der Waals interactions between adjacent alkyl chains bonded to opposite layers,' the energy of interactions balancing, to some extent, that required for increasing the interlayer distance.The layered phase alpha-vanadyl phosphate dihydrate, a-VOP04*2H20, has a tetragonal unit cell (Z=2) with par- ameters a=b= 6.21 A." The c parameter corresponds to the basal spacing (BS) of the phase, i.e. the repeat distance perpendicular to the layers. The structure consists of distorted vanadium(v)-oxygen octahedra which are condensed with four phosphate tetrahedra in their equatorial planes. One of the axial groups in each octahedron is a short V=O bond, while the other is a replaceable water molecule.'2 The electri- cally neutral layers are held together by hydrogen bonds between the co-ordinated water molecules and other interlayer water molecules (Fig.1). The weak interlayer bonding means that a-VOP04 2H20 readily undergoes intercalation reac-tions with the uptake of small polar molecules such as ammonia, n-alkanols,' and n-alkylamines.' 3714 These species are intercalated as neutral molecules with no change in the uncharged nature of the layers being implied. A second type of intercalation reaction involves the reduction of a fraction of the vanadium(v) to vanadium(1v) with concomitant intercalation of cations to counterbalance the induced nega- tive layer charge. This 'redox intercalation' process may conveniently be carried out using iodides. These allow both control and determination of vanadium(v) reduction.Martinez-Lara et al.' ' intercalated n-alkylammonium ions Fig. 1 Schematic represenration of the structure of a-VOP0,.2H20: . small circle, V; large circle, interlayer H,O into a-VOP0,*2H20 by this method according to RNH31+ a-VOP04*2H20-+ X(RNH3)x(V02+),(V03+)1-xP04*nH20+(2-n)H20+-122 (1) The basal spacings of the products were found'' to increase linearly with the number of carbon atoms in the alkyl- ammonium ion. Martinez-Lara et al. suggested that the alkylammonium ions are orientated perpendicular to the vanadyl phosphate layers, but commented no further on the arrangement in the interlayer region. In addition, no mention was made as to the possibility of intercalating methyl- and ethyl-ammonium ions into the host by this method.The purposes of the present study are to investigate the mechanism of these redox intercalation reactions, and to determine the arrangement of alkylammonium ions in the products. To these ends, the original intercalation reactions, carried out by Martinez-Lara et al. were repeated, and the data supplemented by an electron paramagnetic resonance (EPR) study, mixed alkylammonium ion intercalation reac- tions, and ion-exchange experiments. Experimental Alkylammonium Iodide and Host Production Solid alkylammonium iodides and a-VOPO4 2H20 were prepared as described by Martinez-Lara et al.15 The identity and composition of the host were confirmed by powder X-ray diffraction (XRD) and by wet chemical and thermogravimetric (TG) analysis.Intercalation Reactions Batch intercalation reactions with individual alkylammonium iodides were carried out by dissolving the particular iodide in acetone (0.02 mol in 30-100 cm3 depending upon solubility), and stirring a-VOP04*2H20 (0.01 mol) with this solution for 30min at room temperature. The intercalate was then obtained by centrifugation and washed iodine-free with ace- tone. The liberated iodine was titrated with standard thios- ulphate to obtain a value for the degree of vanadium(v) reduction, x. This value was found to be consistent with that obtained by wet chemical analysis. For example, an x value of 0.70, obtained by thiosulphate titration, for a butylam-monium ion intercalate, was found to compare well with 0.69 determined by redox titration of vanadium(1v) and total vanadium present in the sample.Similar study of an octylam- monium intercalate gave x values of 0.77 and 0.76 for thio- sulphate titration and chemical analysis, respectively. Chemical analysis showed also that no molecular iodine had been included by the solid product. The above method was adapted also to produce butylam- monium (C4) and octylammonium (C8)intercalates of various x values, by reacting a-VOPO4 *2H20 with 0.2-2.0 equiva-lents of the appropriate iodide. Mixed alkylammonium ion intercalation reactions were carried out by treating 5 x mol samples of the host with an alkylammonium iodide mixture (containing a total of 0.01 mol) by the same method.The iodide mixtures contained C8 iodide and one other alkylammonium iodide, with the mole fraction of each iodide in the mixture being varied from zero to one. EPR Spectroscopic Study of Intercalation Reactions EPR study of single alkylammonium iodide reactions was carried out by placing 0.12 g of lightly ground a-VOP04*2H20 in an EPR tube and wetting this with 1cm3 of acetone. The tube was then shaken to obtain a suspension, and 1 cm3 of an acetone solution of the alkylammonium iodide (containing 2 equivalents) added. Shaking was resumed for 30 s, and the tube contents were allowed to settle for 30 s (note that this settling period was considered to be part of J. MATER. CHEM.,1991, VOL. 1 ation, to be followed qualitatively.The host compound yields an EPR spectrum (due to the presence of V" impurities) consisting of numerous hyperfine linesI6 arising from interac- tions of the isolated d' electrons with "V, I=$, nuclei. Redox intercalates of a-VOPO, *2H20,such as the proton interca- lates prepared by Zazhigalov et a1.,17 give spectra that show rapid collapse of hyperfine lines into singlets with increasing x, owing to exchange interactions between d' electrons. In the present case, butylammonium ion intercalates with x values as little as 0.16, gave spectra consisting essentially of singlets with some hyperfine structure superimposed. At x = 0.34, only a narrow singlet was observed. The intensity of these signals is directly related to the V" content of the intercalate.The spectrometer used in the present study, how- ever, yielded first-derivative spectra of the signals, making estimation of signal intensity difficult. Nevertheless, the dis- tance between the minima and maxima (or 'peak-to-peak height') of the first-derivative spectra was found to be directly related to the V" content of an intercalate, as is shown for a series of butylammonium ion intercalates in Fig. 2. A qualitat-ive method of following alkylammonium ion intercalation was therefore provided by EPR. It should be noted, however, that no calculations upon reaction kinetics were attempted owing to the relative crudity of the method. Instrumental XRD data were obtained using a Philips PW1050 diffractometer with Co-Ka radiation, and a Philips PW1730 instrument with monochromated Cu-Ka radiation.TG of samples was carried out from 298 to 1223 K using a Mettler TG 50 TG/DTG under still air. (n.b. Vanadium(1v) reoxidation was not found to be a problem under these conditions.) EPR 100 . " " " ' ' ' /€ i, the total reaction time for the purposes of this study). The 0 EPR spectrum was then recorded, and the above shaking- 0.15 0.25 0.35 0.45 0.55 0.65 settling cycle repeated until no further change was observed X in the EPR spectrum. This method allowed the reduction of Fig. 2 Relationship between EPR intensity ('peak-to-peak height')the EPR-invisible Vv to VIV (d' configuration) upon intercal- and vanadium(1v) content (x) of C4 intercalates J.MATER. CHEM., 1991, VOL. 1 spectra were recorded on a JEOL FE-FE3X spectrometer at room temperature. Results The redox intercalation reactions with single alkylammonium iodides of butylammonium (C4) and larger ions were found to proceed in a similar fashion to that described by Martinez- Lara et al.” The basal spacings (BS) of these products (Table 1) were found to be at least double that of the host itself, suggesting that the larger ions were propping the inorganic layers of the host apart. The x values of the materials confirmed also that a substantial degree of reduction had occurred, being in the range 0.7-0.8. On the other hand, the three smaller (IC3) alkylammonium iodides were found to give a low degree of reduction (x =0.1-0.2), and products with BS similar to that of the host itself.This indicates that the amount of alkylammonium ion intercalation was rather low in these cases, and contradicts the observation of Martinez- Lara et al. who found that propylammonium (C3) ions were intercalated to yield a product with a BS of 14.6 A and an x value of 0.5. Alkylammonium Intercalates We concentrate initially on the larger alkylammonium ion intercalates. A plot of the BS against the number of carbon atoms (24) in the alkyl chain of the ion yields a straight line of gradient (ABSlAn) 1.59A per carbon (Fig. 3). Since the maximum possible increment in an alkyl chain is 1.27 A per carbon atom, it is clear that a multilayer arrangement of alkylammonium ions must exist between the layers of the host.The simplest arrangement consistent with an increment of 1.59 8, would be a bilayer of straight, all-trans, alkylam- monium ions with their alkyl chains orientated at an angle, (6, of 39” to the inorganic layers. This value of (6 is extremely low. In comparable intercalates, (6 ranges from ca. 55” in a-zirconium phosphate (ZrP)”.” and KNiAs04,” to 90” (i.e. perpendicular to the layers) in silicic acidz1 and high layer- charge density clay minerals.” One of the few examples of all-trans, alkyl chains having (6 values of less than 40” is in alkylpyridinium ion-exchanged bent~nite.’~ In these materials, however, the low value of (6 seems to be determined by the interaction of the aromatic ring of the intercalated ion with the aluminosilicate layers.The angle which alkylammonium ions make with a layer depends on the ratio of the available interlayer surface area per unit charge to the cross-sectional area of the ions. In the present case, the surface area is 77.1 8,’ per unit cell and 38.6 per vanadium site (2=2). As the intercalates possess an x value of ca. 0.7 (Table l), the surface area available to each intercalated ion is 55A2,which can be compared with the cross-sectional area for an all-trans alkyl chain of only 22-24 It seems likely that the alkylammonium ions would be Table 1 The x [fraction of vanadium(v) reduced] values and major basal spacings of intercalates of different alkylammonium ions R in RNH,’ X basal spacing/A 0.18 7.19 0.09 7.40 0.18 7.19 0.70 15.55 0.72 17.40 0.7 1 18.66 0.74 20.74 0.77 2 1.83 22.51 4 5 6 7 8 number of carbon atoms in ion Fig.3 Plot of basal spacing of intercalates against number of carbon atoms in intercalated ion present as gauche-block structures, i.e. the alkyl chains contain isolated kinks.7 The loose packing of such conformations is known to allow the formation of the maximum possible number of van der Waals interactions and therefore be energetically favourable. Such structures have generally been found where alkyl chains have packing densities lower than 33 8,’ per chain.’ gauche blocks also enable the amine groups to improve their interaction with ‘ro~gh”~ or puckered layers such as are found in the present case.Alkylammonium Ion Mixtures As expected from the results obtained with single alkyl- ammonium iodides, the degree of reduction was high (ca. 0.7) when C4-C7 ions were used in mixtures with octylammonium iodide. Of special interest is the fact that XRD traces showed single basal reflections of comparable width to that of a pure, single-ion intercalate (e.g. Fig. 4). This suggested that the two types of ion were randomly distributed in each interlayer gallery. If interstratification, i.e. segregation of different ions into different interlayer galleries, had occurred the intercalate would possess more than one repeat distance parallel to its 001 plane, and a broad XRD peak would be expected.That random intercalation of ions occurred was supported by the observations that both the BS of the products and their total weight losses from TG increased linearly with the fraction of octylammonium iodide in the mixture (Fig. 5 and 6). The mixed-ion intercalation reactions with the smaller C1-C3 and C8 iodides were found to result in a considerably different type of material. The obtained products displayed x values of 0.15-0.30, suggesting low degrees of intercalation. It is apparent (Table 2) from the XRD results that the materials produced fell into two classes. The first was that produced by reactions with mixtures of mole fraction C8 50.6 (Cl) and J. MATER. CHEM., 1991, VOL. 1 10 5 28J" Fig.4 XRD traces (Cu-Ka) of (a) C8 and (b) mixed C6/C8 (mole fraction C8 iodide 0.8) intercalates 22 --21 -20 5 9) .-04 19--3 f2 18-17-0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 mole fraction of octylammonium iodide Fig.5 Plot of basal spacings of mixed C4/C8 intercalates against mole fractions of C8 iodide in the initial reaction mixture 39 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 mole fraction of octylammonium iodide Fig. 6 Plot of total TG weight losses of CS/C8 mixed intercalates against mole fraction of C8 iodide in the initial reaction mixture C8 I0.4 (C2,C3). These materials gave XRD traces with reflections of BS equal to, or slightly less than, that of the host. This low degree of intercalation was similar to that observed in the single, smaller ion intercalation reactions.The second type of material produced in Cl-C3/C8 mixed intercal- ation reactions was found when mixtures containing higher mole fractions of C8 were used (Table 2). As in the case of the first class of materials, their XRD traces were dominated (Fig. 7) by peaks with a BS of ca. 7 A from the unreacted host. For both classes of mixed intercalate this suggests that intercalation of ions is limited to crystallite faces and edges, with the bulk vanadium sites remaining unreacted. In the second class, however, low-intensity XRD peaks at ca. 17 8, (together with their second-order reflections at 8.5 A) were also found. Although it is not possible to assign these reflec- tions to a particular arrangement of alkylammonium ions with any certainty, the fact that the magnitude of the BS does not change with the size of the other ion in the mixture suggests that these regions may contain monolayers of C8 ions. The low degree of interaction observed with both single- and mixed-ion intercalation reactions of smaller ions should be contrasted with the results obtained (Table 3) from reacting the host with various amounts of C8 iodide (similar results were also obtained for the C4 salt).Here a correlation was found between the amounts of intercalate and unreacted host present in a material, and the number of equivalents of C8 iodide used in the reaction. The XRD trace (Fig. 7) of the product obtained by reaction of the host with 0.4equivalents of C8 iodide illustrated the point well.The material had an x value of 0.34, similar to that of the above-mentioned mixed intercalates, yet a peak with a BS of 21.76A can be seen in addition to that due to the unreacted host at 7.08 A. The BS of this material is therefore almost identical to that of the J. MATER. CHEM., 1991, VOL. 1 47 Table 2 Major basal spacings of mixed alkylammonium ion intercalates other alkylammonium iodide in mixture mole fraction of C8 iodide= 0.0 0.2 basal spacings/A 0.4 0.6 0.8 1.o methyl ethyl ProPYl 7.06 7.22 7.23 7.12 7.21 7.27 7.14 7.22 7.22 7.08 7.13" 7.13" 7.40" 7.18" 7.18" 21.83 21.83 21.83 butyl pentyl hexyl heptyl 15.16 17.26 18.70 20.7 1 16.80 18.11 19.21 20.94 18.80 18.96 19.27 21.34 20.74 20.36 20.74 21.55 21.14 2 1.04 20.27 21.87 21.83 21.83 21.83 21.83 a Small peaks at 17 and 8.5 8, also observed in XRD traces of these materials ;!.L I I 1 10 5 2eio Fig. 7 XRD traces (Cu-Ka) of (a)C8 intercalate produced with 0.4 equivalents of C8 iodide and (b) mixed C3/C8 intercalate, mole fraction C8 iodide in initial reaction mixture = 0.8 Table 3 x [fraction of vanadium(v) reduced] values and basal spacings of C8 intercalates produced by reaction of host with varying amounts of C8 iodide equivalents of C8 iodide x XRD peaks (relative intensity) 0.2 0.16 21.98(7) 7.38(17) 7.14(76) 0.4 0.34 21.76( 19) 7.08(81) 0.6 0.44 21.66(71) 7.08(3) 6.71(26) 0.8 0.53 21.87(81) 7.03(7) 6.69(12) 1.o 0.56 21.55(92) 6.99(2) 6.68(6) 2.0 0.76 21.55(97) 6.61(3) fully reduced x = 0.76 intercalate, suggesting that the arrange- ment of alkyl chains is the same in both compounds.The most likely explanation for this is that intercalation of C8 ions occurred initially at the edge of the crystallite and spread towards the centre of the interlayer gallery as reaction pro- ceeded. The alternative explanation is that individual crystal- lites consisted either of unreacted host or of fully reduced intercalate. This seems unlikely given the relative weakness of interlayer bonding in the host. Ion-exchange Experiments The ion-exchange experiments further demonstrated the difference in the properties of large and small alkylammonium ions with respect to intercalation in cc-VOPO4.2H20.Treat-ment of an octylammonium ion intercalate with butylam- monium ions (Fig.8) caused the BS of the intercalate to fall, suggesting exchange of the larger ions (this was confirmed by the total TG weight losses of intercalates decreasing as exchange occurred). Conversely, treatment of a butyl-ammonium intercalate with ethylammonium (C2)ions caused no change in BS. The C4/C8 exchange indicated that diffusion of the smaller ions was facile, suggesting that this should not be a problem for C2/C4, and other explanations are, therefore, required to explain why ion exchange did not occur. 519.0-.-18.5--(d 3 18.0-D -17.5 17.0--16.5 15 n .v.-I....,....0 1 2 3 4 5 6 7 8 9 10 equivalents of exchanging ion Fig.8'Plot of basal spacings of (a) C8 intercalate treated with C4 iodide and (b)a C4 intercalate treated with C2 iodide 48 EPR Measurements EPR study of intercalation reactions was found to be imposs- ible for Cl-C3 iodides as the presence of acetone solvent prevented the recording of the rather weak spectra of the resulting materials. The reactions of the larger ions could, however, be followed and these were found to have fast rates, being complete in minutes. C5 and larger ions gave almost immediate reaction, with maximum EPR intensity being found in the first recorded spectrum after 1 min (the reaction of C8 together with that for C4 is shown in Fig.9). The reaction with C4 iodide was found, however, to have a lag of 2min before a rapid increase to maximum vanadium(~v ) content after 3 min.C4 iodide, in addition to being the smallest alkylammonium iodide to yield a stable intercalate, is therefore unique in that its intercalation reaction with a-VOPO, 2H20 involves a short period when the degree of vanadium(v) reduction is low before a rapid reaction to yield the final intercalate. Discussion All the aata presented show that a-VOP04*2H20 discrimi- nates as an intercalation host between large (2C,) and small (5C,) n-alkylammonium ions. The larger ions are readily intercalated and form a stable bilayer between the inorganic layers of the host. Alternatively, the three smallest ions are not intercalated to any great degree and indeed inhibit the intercalation of C8 ions when mixtures of iodides are used for intercalation reactions.An explanation as to why the smaller ions have this effect is suggested by the XRD traces of the resulting materials (Table 1, 2 etc.), which indicate that intercalation has occurred at crystallite edges and faces only. 110 V 100 90 Q,-5 80-(d .-.c 5 70 23 Ln 60 I 0,.-2 50 Y Q,Q 4-6 40 2i Q,n c 30 n LLI 20 10 0 I 50 100 150 200 250 300 350 400 450 500 time/s Fig.9 Reaction of (a) C8 iodide and (b) C4 iodide with cr-VOP0,.2H20 (as followed by EPR) J. MATER. CHEM., 1991, VOL. 1 Results for C8 intercalates (Fig. 7) suggest that intercalation occurs initially at the edge of interlammellar galleries and spreads inwards towards crystallite centres. Intercalation of smaller ions at crystallite edges, however, prevents reaction with the bulk of the present vanadium(v) sites.It is interesting to speculate how these smaller ions could have this ‘poisoning’ effect on further intercalation. One possibility is that the small ions could take up a parallel orientation to the host layers when intercalated. Strong layer-ion-layer electrostatic inter- actions would then cause the layers to be held tightly together, and to effectively block the diffusion of additional reactant to the centre of interlammellar galleries. This ‘poisoning’ effect could account for the failure of both single- and mixed-ion intercalation reactions involving the smaller alkylammonium ions.The observed peaks at ca. 17A for certain mixed intercalates suggest, however, that the effect of the smaller ions may be modified somewhat if sufficient of a larger ion is used in the reaction mixture. EPR study of intercalation reactions demonstrated well how C4 ions represented the borderline between ions that successfully intercalated into the host, and those that ‘poi- soned’ subsequent intercalation reaction. The short time lag observed in the reaction of this iodide may be explainable by C4 ions being initially intercalated parallel to the host layers at the crystallite edges, thereby blocking the diffusion of further reactant. This situation would, however, be energeti- cally unstable with respect to a gauche-block arrangement of C4 ions as is present in the final intercalate, and a transform- ation (perhaps aided by unfavourable steric interactions between neighbouring parallel C4 ions) to gauche-block would occur.Diffusion of further reactant between the loose-packed C4 ions would then be facile, accounting for the fast reaction observed after the initial time lag. The failure of C2 ions to be introduced into a C4 intercalate by ion exchange may also be accountable for by a similar ‘poisoning’ effect of C2 ions at crystallite edges. XRD revealed, however, that the products of this ion exchange were still monophasic with no evidence of the regions of low BS that would be anticipated if edge-blocking occurred.Another possibility is that exchange did not occur because the product would be less thermodynamically stable than the original intercalate. C2 ions have been previously to form an all-trans bilayer between the layers of ZrP; however the layer charge density (i.e. the area per unit charge) of ZrP is greater than the diameter of the C2 ion. An all-trans bilayer of C2 ions is therefore dictated on stoichiometric grounds in ZrP.” The layer charge density of the present ions was considerably lower than that of ZrP, far exceeding the diam- eter of the C2 ion. To enjoy significant stabilisation uia van der Waals interactions therefore, an alkylammonium ion would need to contain kinks and form a loose-packed, gauche- block structure.C2 ions are not large enough to contain kinks and to form gauche-blocks. The mechanism of alkylammonium ion intercalation into a-VOPO, 2H20 was a considerably more complicated pro- cess than originally envisaged. The edge-first intercalation reaction observed here has been reported previously for ZrP by Alberti,25 and seems to be a general mechanism for most layered hosts (although not for the important clay mineral kaolin’). Indeed, Clearfield and TindwaIg observed that propylamine and butylamine were intercalated parallel to the layers of ZrP at low amine loadings. It was found also that the intercalated organic ions were forced into a more upright arrangement as further intercalation proceeded. The present study is believed, however, to be the first to detail an instance where an initial intercalation of alkylammonium ions at the edge of an interlayer gallery blocks further intercalation reactions.J. MATER. CHEM., 1991, VOL. 1 One of the authors (M.M.) would like to thank Laporte Inorganics, Widnes, and the Science and Engineering Research Council for their financial support of this work via a CASE grant. References 1 Intercalation Chemistry, ed. M. S. Whittingham and A. J. Jacob- son, Academic Press, New York, 1982. 2 G. Lagaly, Philos. Trans. R. SOC. London, Ser. A, 1984, 311, 315. 3 T. J. Pinnavaia, Science, 1983, 220, 365. 4 G. Alberti, in Recent Developments in Ion Exchange, ed. P. A. Williams and M. J. Hudson, Elsevier Applied Science, London, 1987, pp.233-248. 5 G. Lagaly, Solid State lonics, 1986, 22, 43. 6 G. Lagaly and A. Weiss, Angew. Chem., Int. Ed. Engl., 1971, 10, 558. 7 G. Lagaly, Angew. Chem., Int. Ed. Engl., 1976, 15, 575. 8 F. R. Gamble, J. H. Osiecki, M. Cais, R. Pisharody, F. J. Disalvo and T. H. Geballe, Science, 1971, 174, 493. 9 C. Rosner and G. Lagaly, J. Solid State Chem., 1984, 53, 92. 10 G. Alberti and U. Costantino, in Intercalation Chemistry, ed. M. S. Whittingham and A. J. Jacobson, Academic Press, New York, 1982, pp. 147- 1 80. 11 A. W. Hewat, M. Tachez, F. Theobold and J. Bernard, Rev. Chim. Miner., 1982, 19, 291. 12 J. W. Johnson, A. J. Jacobson, J. F. Brody and S. M. Rich, Inorg. Chem., 1982, 21, 3820. 13 G. Ladwig, 2.Anorg. Allg. Chem., 1965, 338, 266. 14 K. Beneke and G. Lagaly, Inorg. Chem., 1983, 22, 1503. 15 M. Martinez-Lara, A. Jimenez-Lopez, L. Moreno-Real, S. Brusque and E. Ruiz-Hitzky, Muter. Res. Bull., 1985, 20, 549. 16 D. Ballutaud, E. Bordes and P. Courtine, Muter. Res. Bull,, 1982, 17, 519. 17 V. A. Zazhigalov, A. I. Pyanitskaya, 1. V. Bachemkova, G. A. Komashko, G. Ladwig and V. M. Belousov, React. Kinet. Catal. Lett., 1983, 23, 119. 18 E. Michel and A. Weiss, Z. Nuturforsch, Teil B, 1965, 20, 1307. 19 A. Clearfield and R. M. Tindwa, J. Inorg. Nucl. Chern., 1979, 41, 871. 20 K. Beneke and G. Lagaly, Clay Miner., 1982, 17, 175. 21 G. Lagaly, Colloid Interface Sci., 1979, 11, 105. 22 G. Lagaly, Clay Miner., 1981, 16, 1. 23 G. Lagaly, H. Stange and A. Weiss, Proc. Int. Clay Conf. 1972, 1973, 693. 24 A. Grandin, M. M. Bore1 and B. J. Raveau, J. Solid State Chem., 1985, 60, 366. 25 G. Alberti, Acc. Chem. Rex, 1978, 11, 163. Paper 0/02933E; Received 29th June, 1990