J O U R N A L O F C H E M I S T R Y Materials Tin sulfide clusters in zeolite Y, Sn4S6-Y Carol L. Bowes and GeoVrey A. Ozin* Materials Chemistry Research Group, L ash Miller Chemical L aboratories, University of T oronto, 80 St. George St., T oronto, Canada, M5S 3H6 The synthesis of a tin sulfide cluster array, denoted Sn4S6-Y, using the quantitative sequential surface anchoring and reaction of tetramethyltin with hydrogen sulfide (metal organic chemical vapor deposition, MOCVD, reagents) within the supercages of acid zeolite-Y is detailed. The tethered methyltin species and their transformation to the encapsulated tin sulfide clusters are elucidated through gravimetry, coupled with mid-IR and 119Sn Mo�ssbauer spectroscopy.Transmission electron microscopy (TEM) and Rietveld refinement of synchrotron powder X-ray diVraction (PXRD) data show that the clusters are internally confined and homogeneously dispersed in the supercages of the host zeolite, while optical spectroscopy show a cluster size dependent blue-shift of the absorption edge with respect to the bulk tin sulfide phase. A geometry is proposed for the encapusulated Sn4S6-Y cluster.Introduction Experimental Materials Various approaches to the fabrication of semiconductor quantum dots have been explored, both ‘top-down’ and ‘bottom- Acid zeolite Y was prepared by repeated ion-exchange of up’.1 The former starts with two-dimensional quantum well calcined and defect-removed Na56Y (UOP Y-52 lot no. 13076- structures formed usually by molecular beam epitaxy or chemi- 81) with ammonium nitrate.Crystallinity was confirmed by cal vapour deposition, followed by lithography, etching and powder XRD; 27Al MAS NMR indicated no occluded alumimilling techniques to reduce the lateral dimensions and form nous material and 29Si MAS NMR confirmed the Si/Al ratio.16 cylindrical dots.2 Great gains have been made in recent years Dehydrating at 430 °C resulted in deammination, leaving proand the minimum size of such structures has decreased.Dots tons as charge-balancing cations. At this stage, the resulting on the order of tens of nanometers can be created using material was very moisture sensitive, and was maintained modified STM, but not on a practical production scale, and under high vacuum (10-5 Torr) for subsequent steps. Elemental by sub-optical lithography, although the thicknesses of dots analysis (Galbraith, ICP) indicated a unit cell formula of made using this technique tend to be many times this size.3 H44Na12[(AlO2)56(SiO2)136] corresponding on average to 5.5 Variations of this form of fabrication include the construction protons and 1.5 Na+ ions per a-cage and b-cage.Tetraof nanoelectrodes within the quantum well structure so that methyltin (Aldrich, 99+%) was stored over a molecular sieve electric fields can be produced to ‘squeeze’ the electrons and and degassed by a freeze–evacuate–thaw cycle.H2S (Matheson, holes electrostatically.4 In this way coupling between the dots 99.5%) was passed through freshly dehydrated sodium Y can be controlled and adjusted via an applied gate-voltage.zeolite before use. Structures have been made in which layers are created ‘epitaxially’ such that they are not well lattice-matched, resulting in a Synthetic methods strained interface. Selective etching of the straining layer relieves the compressive stress locally and results in confine- Most reactions involving the zeolite were carried out on a selfment of the charge carriers in that area.5 supporting pressed disk of acid zeolite Y (approximately 40 mg) The ‘bottom-up’ route involves chemical synthesis in which in an in situ cell which allowed dehydration, incremental attempts are made to tailor the size, shape and distribution of Me4Sn adsorption, and had NaCl and quartz windows to semiconductor nanoclusters using self-limiting techniques.6 allow mid-IR and UV–VIS spectroscopy.17 Some experiments This was initially approached by solution-phase synthesis were performed in a small reaction cell on larger amounts of with organic capping7 or through sol–gel-type synthesis.8 acid zeolite Y to allow gravimetric measurements. In general, Alternatively, synthesis within structured media was employed the cluster was produced by dehydrating and deaminating the to constrain size.This was originally attempted in micelles, zeolite wafer at 430 °C, followed by cooling under dynamic vesicles and lipid bilayers, polymers, glasses, clays, and zeo- vacuum. Degassed tetramethyltin was introduced incremenlites. 9–12 Organic-passivated clusters are now prepared in tally by allowing the vapour above the liquid to expand into macroscopic quantities with tunable sizes of 15–100 A ° with a a small ‘titration’ volume and then allowing this volume of standard deviation of less than 4%, and can be manipulated gas to enter the main body of the closed in situ cell.After to form superlattices of close-packed nanoclusters as faceted allowing the gas to adsorb into and react with the zeolite for crystals or thin films.13 Colloidal chemistry is used to produce a brief time (ca. 15 min), the wafer was heated in the quartz layered nanocrystals, for example CdS/HgS/CdS, with epitaxial end of the sample cell to 150 °C for 2 h, after which it remained matching between the core and layers.14 It has also been found white. The sample cell was then returned to the vacuum line, that nanoclusters will self-assemble at high-strain regions when evacuated and filled to 150 Torr with dry H2S.The wafer was direct epitaxy is performed on the slopes of the valleys of a again heated for 2 h at 150 °C, whereupon it turned a pale corrugated layer. Such nanoclusters even self-align to form golden yellow colour, and the cell was evacuated. Any further quasi-ordered arrays.15 handling of the sample was performed under inert atmosphere.In this paper a self-limiting synthetic approach is described which takes advantage of the host–guest chemistry of acid Characterization zeolite Y, in which the quantitative sequential anchoring and FT-mid-IR spectra were collected on a Nicolet 20SXB (reso- reaction of tetramethyltin with hydrogen sulfide forms tin sulfide clusters within the zeolite supercages.lution 2 cm-1) by co-adding 100 interferograms. UV–VIS J. Mater. Chem., 1998, 8(5), 1281–1289 1281diVuse reflectance spectra were recorded with respect to BaSO4 standards on a Perkin-Elmer 330 spectrophotometer using an integrating sphere attachment. Data were digitized and converted to absorbance using Kubelka–Munk theory.18 Powder X-ray diVraction data collected for Rietveld structure refinement were obtained at Brookhaven National Laboratories on beam-line X7A of the synchrotron.A germanium double crystal monochromator was used to select wavelengths of about 0.7 A ° , which was detected by a Kevex solid-state detector. Samples were sealed in Lindemann capillary tubes, cooled to 15 K, and rocked by about 2° while data were collected in 0.005° intervals between 2 and 50° 2h, with collection times increasing with increasing 2h for a total of 12 h.Data were analyzed using GSAS software.19 119Sn Mo�ssbauer data were collected in constant acceleration mode using a Ranger Scientific MS-1200 instrument. The radiation source was 5 mCi 119Sn in a CaSnO3 matrix. Data were collected at room temperature or at 77 K relative to SnO2.Semi-quantitative Fig. 1 Acid zeolite Y showing 13 A ° a-cage formed by 6 A ° b-cages (2.5 atom%) scanning electron microscopy with energy disper- linked by double six-rings. Oxygen atoms bridge between Si and Al sive X-ray microanalysis, SEM–EDX, was performed by tetrahedral centres located at each vertex. Cation sites within the a- Imagetek, using a Hitachi 800 analytical transmission electron cage are shown, as well as Brønsted acid sites, Ha, Hb.microscope operating at 100 kV. The probe size was 50 nm in area and penetrated greater than 1 mm in depth. A LINK X- specific crystallographic positions of the cations, considered ray microanalyser was employed for data anas, and SnS oxide coordination sites in the ‘zeolate’ coordination chemistry and SnS2 were used as standards.Independent elemental model,23,24 are indicated in Fig. 1. analysis was performed by Galbraith Laboratories, Inc., using The adsorption of tetramethyltin is observed in the mid-IR ICP for Si, Al, and Sn determinations and for C, the IR spectrum between 4000 and 1200 cm-1, using an in situ intensity of CO resulting from sample combustion.TEM reaction cell as described in the experimental section. Fig. 2(a) images of the lattice at more than 1.2×106 times magnification shows the spectrum of HY dehydrated at 430 °C. The bands were obtained also by Imagetek using a Hitachi 7000 trans- at 3640 and 3540 cm-1 have been assigned to the a and b mission electron microscope, operating at an accelerating bridging nOH modes.25,26 The small peak at 3740 cm-1 repvoltage of 100 kV.The resolution was estimated conservatively resents terminal nOH stretches, i.e. those OH existing on the at 4.5 A ° . Exposure times were limited to 2 s as electron beam external surface, or in hydroxyl nests and defect sites. Fig. 2(b) induced sample alteration was observed at times of 45 s or shows the final eVect of titration of the a-proton with numerous more.Extended Hu�ckel molecular orbital, EHMO, calculations incremental additions of thoroughly degassed tetramethyltin. were undertaken for isolated tin sulfide clusters, Sn4S6, using The IR intensity of the a-proton was used as the indicator for the ICONCL software.20 The program used the standard the titration. The a-protons reacted, evolving methane and EHMO theory wherein atomic ionization potentials were anchoring a trimethyltin moiety in the a-cage as follows: used for Hii and Hjj, and Hijs were calculated using the Z-O-H+Me4SnAZ-O-SnMe3+CH4 Wolfsberg–Helmholtz approximation, Hij=1/2K(Hii+Hjj)Sij.Overlap integrals, Sij, were calculated for all atomic orbitals Heating to 150 °C for 2 h resulted in the consumption of all rather than just nearest neighbours, using Slater-type orbit- reactive a- and b-protons, as is shown in Fig. 2(c). The resulting als. A weighted Wolfsberg–Helmholtz approximation, K= methane IR spectrum (n3 and n4 modes) was observed in the k¾+D2+D4(1-k¾) where D=(Hii-Hjj)/(Hii+Hjj) and k¾ is a atmosphere of the sealed in situ cell and its intensity was constant,21 was used in this program to account for the measured after each step.An intensity–pressure calibration diVering ‘diVuseness’ characteristic of complexes containing was made, in order to ensure that Beer’s law holds over the unoccupied high-energy basis functions. In addition, a distance pressure range studied. Acid zeolite Y is known not to adsorb dependence for K was included, according to the work of methane, unlike other zeolites or other alkanes.27,28 The further Calzaferri et al.22 Results and Discussion The internal surface of acid zeolite Y is ideal for MOCVD because the protons, which are charge balancing in the structure, can participate in surface anchoring thereby providing control over reaction stoichiometry.The reaction and regeneration of protons as well as their solvation and migration are all observable by mid-IR spectroscopy. Thus the progress of the host–guest chemistry can be monitored in situ providing control over the cluster assembly process.23 Synthesis details Anchoring of tetramethyltin.Zeolite Y has an open-framework topology comprised of a- and b-cages, roughly 13 A ° and 6 A° diameter, respectively, interconnected through 6 T-atom windows.In the illustration of the structure of zeolite Y shown Fig. 2 The titration of protons in zeolite Y with tetramethyltin. Midin Fig. 1, each vertex represents a tetrahedral (T-atom) site IR spectra of (a) dehydrated H44Na12Y, (b) spectrum following awhich is either SiO4 or AlO4-. Extra framework cations proton titration and tetramethyltin anchoring, (c) spectrum following reaction at 150 °C.balance the framework charge due to the AlIII centres, and the 1282 J. Mater. Chem., 1998, 8(5), 1281–1289reaction of the anchored methyltin species with H2S (150 Torr, The combination of restricted space in the a-cage and unfavourable b-proton reactivity limits the extent of Me4Sn 150 °C, 2 h) released essentially all of the remainder of the methyl groups bound to tin as methane, the IR intensity of loading at room temperature. It is critical to this ‘deductive’ reaction characterization that which was measured.By equating the final methane intensity measured with the the tin centres remain in the same oxidation state of +IV throughout. Confirmation that the anchored tin centre maximum of four methyl groups per tin, it was possible to conclude that, on average, 1.0±0.1 methyl groups per tin were remained in the original oxidation state of tetramethytin was ascertained by 119Sn Mo�ssbauer spectroscopy, Fig. 3(a). The released from the titration of the four a-cage protons at room temperature, and therefore that four trimethyltin species were anchored precursor and the cluster product, which will be discussed later, have isomer shifts of 1.27 and 1.25 anchored per a-cage.By the same calculations, after reaction at 150 °C, 1.5±0.1 methane molecules per tin were present in (±0.01) mm s-1, respectively. Isomer shifts of less than 1.5 define tin in oxidation state IV.31 the atmosphere of the in situ cell. This corresponds to a total of 6 protons per a-cage/b-cage unit (48 per unit cell ), in reasonable agreement with the 5.5 (44 per u.c.) determined by Hydrogen sulfide treatment. After the Me4Sn reaction and anchoring were complete, the sample was subjected to approxi- elemental analysis.After reaction at 150 °C, there is some mately 150 Torr H2S at room temperature followed by 150 °C small, remaining b-proton intensity which was thought to be for 2 h.Fig. 4( b) displays the eVect of room temperature H2S either inaccessible or weakly acidic defect proton sites, lying treatment, while Fig. 4(c), (d) show the results of heating at within the b-proton bandwidth. The reaction of essentially all 150 °C and evacuating at room temperature, respectively. protons was taken to indicate the ‘homogeneous dispersion’ Features T, A and B are familiar; the terminal and b-proton of the reactants throughout the zeolite.Although the protons are essentially unaVected while all remnants of the a-proton are to some extent mobile, they cannot delocalize from the have disappeared. Vestiges of the methyl groups remain also, unit cell entirely because it would involve creating chargefeature E. In experiments in which the hydrogen sulfide was separated regions in the zeolite lattice which is expected to be introduced in small, incremental volumes, features C and D energetically unfavourable.Therefore, the disappearance of all grew in together with increasing amounts of H2S. Features C protons implies the homogeneity of the anchored precursor. and D, centred at about 3000 and 2370 cm-1, have been The reaction pathway for the tetramethyltin adsorption may assigned to the solvated a-protons and its solvators, respect- be summarized as follows ively: ZMOMHa,(SH2)n.Earlier work32 documented similar broad and intense peaks resulting from solvation (hydrogen Ha32Hb12Na12Y+32Me4Sn bonding) of zeolite Y a-cage protons by anhydrous hydrogen CA RT (Me3Sn)32Hb12Na12Y+32CH4 halides, of which the vibration was similarly bathochromically shifted.In the earlier work, the degree of shifting of the vibration was found to depend upon the electronegativity of (Me3Sn)32Hb12Na12Y the halide and the number of solvating molecules. In all cases a second broad band was observed, corresponding to band D CA 150 °C (Me3Sn)20(Me2Sn)12Na12Y+12CH4 in this experiment, which also shifted with electronegativity, but to a lesser extent.It was bathochromically shifted from It is interesting to consider the factors contributing to the the expected frequency of the free molecule nHX and in this reaction path described above. One factor is the relative reactivity of the methroups on the tin centre. It is clear that the first methyl group reacts with a proton with ease at room temperature, while the second is less readily removed.It is expected that the nucleophilicity of the second methyl carbon will be less than that of the first because an electron releasing methyl group of Me4Sn has been replaced by an electron withdrawing oxide-type ligand of ZO-SnMe3. Therefore the reactivity towards an electrophile (proton) will incrementally decrease as first, second or third methyl groups react.Also, electrostatics may contribute to the diminished reactivity. For a second methyl group to react, the proton must eVectively interact with a Me3Sn+ centre bearing a formal 1+ charge. For a third, it must approach a Me2Sn2+ centre bearing a formal 2+ charge. Electrostatics are expected to be an important factor, based on decationization and dehydrohalogenation kinetic experiments in zeolite Y.29 A second factor is the steric limitation imposed by the size of the a-cage.It was observed, through IR spectroscopy of the sample pellet and sample-cell atmosphere, that Me4Sn added after completion of the titration of the a-protons remained in the atmosphere of the cell (outside the zeolite), despite the presence of further unreacted b-protons.This information should be coupled with the idea that there are expected to be thermodynamic and kinetic barriers to the reaction of the b-protons. It must be so, or the initial, room temperature uptake of tetramethyltin would aVect a- and bprotons equally which is not observed (RT decrease of the bproton intensity corresponds only to 0.15 protons).Because Me4Sn is too large to enter the b-cages, the b-protons must leave the b-cage in order to react; plus, b-protons are less Fig. 3 Mo�ssbauer spectra of (a) the precursor material (CH3)2Sn-Y acidic and less mobile than their a-counterparts, as shown and (b) the product material Sn4S6-Y. Isomer shifts are reported with respect to SnO2. through adsorption and desorption energies of given bases.30 J. Mater.Chem., 1998, 8(5), 1281–1289 1283hydrosulfurization–deanchoring (proton regeneration): Me2Sn-Y+H2SAMe2(SH)Sn-Y+H-Y Me2(SH)Sn-Y+H2SAMe2(SH)2Sn+H-Y dehydrosulfurization–condensation (cluster assembly): (SH)4Sn+(SH)3Sn-YA(SH)3Sn-S-Sn-(SH)2-Y+H2S The alternative reaction for the last step is an intramolecular self-dehydrosulfurization, in which a SnNS species is formed following H2S elimination.While this step could not be ruled out a priori, it is not thought to contribute in light of the remainder of the data, as will be discussed. Thus the proposed synthesis in which four anchored methyltin species per a-cage are demethylated, hydrosulfurized, and self-assembled via dehydrosulfurization–condensation reactions accounts adequately for the formation of a tin sulfide cluster product, except for the question of cluster stoichiometry.Several methods were employed for the determination of the Fig. 4 Adsorption and reaction of hydrogen sulfide in methyltintin to sulfur ratio in the nanocluster product. The first was a loaded zeolite Y. (a) Mid-IR spectrum of anchored dimethyl- and quantitative gravimetric experiment in which the entire syn- trimethyltin in zeolite Y, (b) spectrum following room temperature thesis was carried out on a large sample (approximately introduction of 150 Torr H2S, (c) the product of 150 °C reaction, (d) spectrum following room temperature evacuation of the reaction 300 mg) in a small cell (approximately 60 g) which was accucell. rately weighed after each step in the reaction process.The second was semiquantitative SEM–EDX using SnS and SnS2 standards. A conservative estimate of the error is ±2.5 atom% for ratio measurements with standards and within the thin crystal limit for micro-X-ray fluorescence.34 The third was case the vibration may be assigned also to the nSH stretch for independent elemental analysis (Galbraith) using ICP and with solvating hydrogen sulfide.standards to establish errors. The results were as follows: Finally, the peaks F and G at about 2562 cm-1 have been assigned also to nSH vibrations. In the literature,33 the n3 gravimetric: S/Sn=1.63±0.05 STEM–EDX: S/Sn=1.42±0.04 Galbraith: S/Sn=1.50±0.15 mode of molecularly adsorbed hydrogen sulfide as well as the nSH stretching mode of a hydrosulfide group resulting from dissociatively adsorbed H2S in zeolites, have been reported at giving an average result of 1.5 for the S/Sn ratio and therefore about 2560 cm-1.While the former may be present (some H2S a likely cluster stoichiometry of Sn4S6, with an overall charge molecules not involved in hydrogen bonding to a-cage protons of 4+ based on SnIV and S-II oxidation states, Fig. 3(b), see but rather physisorbed to the oxygen framework), the latter is Mo�ssbauer discussion above. almost certainly present, as will be considered in the discussion The absence of any significant contribution from tin sulfide of the reaction path below. material aggregating on the external surface of the zeolite Heating to 150 °C has little eVect on the IR spectrum, except crystals was determined by TEM lattice imaging, Fig. 5. for a decrease in the D-band and a shift to higher frequency Aggregations of the heavier guest species generally appear as of the C-band, both associated with a decrease in the number darker spots on the crystal surface which was not observed in of solvating H2S molecules. Also, the intensity of and ratio any TEM images obtained from many of these materials.In between bands F and G have changed, possibly indicating the addition, a general impression of the homogeneity of the reaction of, for example, H2S to form HS- species. The sample interior structure can be obtained which suggests that the tin appearance, at this stage, changed from the previously white sulfide clusters are encapsulated and homogeneously distribcolour to a pale golden colour.uted. In fact, because not all protons are regenerated in the After 150 °C reaction, the excess hydrogen sulfide and methtin sulfide cluster formation process, it is expected that the ane reaction product were pumped out of the in situ cell at clusters are anchored to the lattice, performing some frameroom temperature, leaving a material whose IR spectrum is work charge-balancing function, in accord with the expected shown in Fig. 4(d). The most noticeable aspect of this spectrum charge on Sn4S64+. is that protons, both a- and b-cage, were regenerated. It is not Finally, the regeneration of unsolvated protons has importclear at what point in the reaction scheme the protons are ant ramifications with respect to cluster growth within the regenerated, because in the presence of H2S the a-protons are zeolite. Preliminary molecular graphics structural models indisolvated and their number cannot be easily quantified from cated that a Sn4S6 cluster would not fill the a-cage void space their IR intensity.Nevertheless, in repeated experiments, on and this idea is supported in that there exist a-protons average 40% and up to 50% of the original intensity of these unsolvated by the cluster in the final product.In principle, it bands has returned after evacuation. Band D disappeared should be possible to re-titrate the regenerated protons and in completely, while some portion of the protons remain solvated, a second, controlled loading step add tin and sulfur to the indicating that all of the H2S was removed but that the tin existing clusters in order to increase their size.Doping might sulfide product species solvates the neighbouring protons. also be achieved in this way. Finally, the bands F and G disappeared, suggesting that no Growing the clusters was accomplished as shown in Fig. 6, hydrosulfide species remained. Reactions of the following type, where each IR spectrum represents a complete Me4Sn titration in which Y represents the zeolite framework, are thought to and anchoring/H2S adsorption and reaction/evacuation cycle.contribute to the process observed in the IR: Important aspects of these spectra are the decrease of both hydrosulfurization–demethylation (methane evolution): solvated and lvated protons and the lack of increase of methyl or hydrosulfide group intensity.The former indicated Me3Sn-Y+H2SAMe2(SH)Sn-Y+CH4 that more of the cation sites are occupied ‘capping’ the clusters, the latter suggests that methyl groups are decreasingly required Me2Sn-Y+H2SAMe(SH)2Sn-Y+CH4 1284 J. Mater. Chem., 1998, 8(5), 1281–1289most significant. Having control over the process, the ability to stop, and change the reaction conditions at any point in the synthesis is what is most unusual, and most useful about this method.Therefore it is interesting to consider what species are present at various points; in short, to consider the reaction pathway. The reaction of methyltin moieties with protons is fairly straightforward, as was described above, the first loss of methane occurring easily at room temperature, and the second with a higher activation energy due to the change in nucleophilicity and electrostatics. Thus the anchoring of Me4Sn in HY finished with a mixture of anchored trimethyl- and dimethyltin, there being 5.5–6.0 reactive protons and 4 tin centres per a-cage.At room temperature, upon H2S addition, methane is released corresponding to almost one equivalent per tin (0.85±0.10) so that 2.4±0.1 of the four methyls per tin have been released.This is curious, since the reacting species were mixed di- and trimethyltin species. No argument based on relative SnMC bond strengths will easily account for this observation. The outcome of the reaction must depend upon which other factor was dominant.First, the degree to which the HMS bond of coordinating H2S is polarized by the tin centre will aVect the reactivity of the proton. The replacement of methyl groups by hydrosulfide groups would increase the ability of tin to polarize that bond. Second, the susceptibility of the methyl carbon to electrophilic attack will be decreased by the introduction of hydrosulfide groups about the tin centre that are more electron withdrawing, having the eVect of attracting electron density away from carbon and decreasing its nucleophilicity.Fig. 5 Typical high resolution TEM lattice image at 1.2×106 magnifi- If the reactivity of the H2S proton were the limiting step, cation of zeolite Y with encapsulated tin sulfide clusters, showing lack tin nuclei with fewer electron donating methyl groups and of external tin sulfide and homogeneity of sample more electron withdrawing hydrosulfide groups would be more reactive.After a dimethyltin species reacted, it would become even more reactive, and one equivalent of methane could result from those centres alone. Alternatively, if the reactivity of the carbon were limiting, then the result might be that a centre, having gained one hydrosulfide group, would be invulnerable to further attack at room temperature, so that methane produced represents one methane from each centre.Neither scheme accounts for interference by de-anchoring proton regeneration reactions which may compete for hydrogen sulfide protons. Described in reaction (2) above, the dissociative deanchoring must occur in order to form a tin sulfide cluster. The IR band corresponding to solvated protons in Fig. 4(a) seems too large to be entirely due to a few ‘leftover’ protons, yet it is not possible to quantify the amount, as the intensity is dependant both on the number of solvated protons and on their extent of solvation, as described above. However, it is likely that there is a suYciently large excess of hydrogen sulfide so that the methane production and de-anchoring can proceed without mutual interference. Fig. 6 Mid-IR spectra of (a) Sn4S6 clusters in zeolite Y, (b) product The second process is that of cluster formation self-assembly, following a second loading of tetramethyltin and hydrogen sulfide, described in reaction (3). After thermal treatment, Fig. 4(c), a and (c) following third loading strong nSH vibration remained (feature G), along with some significant number of solvated protons (C).This is taken to indicate that many, if not all, species are in the hydrosulfide to satisfy tin coordination sites and that hydrosulfide condensation is occurring. Possibly, the new material is incorporated form, and only condense and assemble when the excess hydrogen sulfide is pumped from the cell. to form larger clusters rather that forming new, smaller clusters which would leave many tin atoms coordinatively unsaturated.Using the methods of methane quantification described earlier, Proposed cluster geometry of Sn4S6-Y. There exist in the literature a number of tin sulfide clusters of various stoichio- the reactions of second and third Me4Sn loadings were shown to produce as much as 118% as much methane as the first, metries and cluster charges having specific geometries.35 However, the clean chemistry leading to the cluster and various suggesting eight Sn atoms or more per a-cage.The UV–VIS spectra of the materials illustrate the cluster generation and pieces of evidence have led us to propose a single Sn4S6 cluster moiety.Firstly, the only example of a cluster of this stoichi- growth nicely, and are discussed in a later section. ometry in the literature is the recurring adamantanoid geometry with a 4+ charge. Because the clusters in the zeolite Reaction pathway. Although the reactions presented above are thought to represent those participating in cluster forma- are partially charge balancing with respect to the anionic framework, creating larger clusters would involve separation tion, it is the step-wise nature of this process which is perhaps J.Mater. Chem., 1998, 8(5), 1281–1289 1285of charge between the anionic lattice and cationic clusters Spectroscopic characterization of precursors and products which would be energetically unfavourable. Further evidence IR Spectroscopy of anchored methyltin reagents in zeolite Y.was obtained from a Rietveld refinement of the structural A key tool in determining the product of the first step anchoring model. Rietveld PXRD refinement is the most widely used reaction has been in situ mid-IR spectroscopy. Yet, it was not method for obtaining crystal structures of materials when it is possible to identify the products directly from their spectra, not possible to obtain crystals of suYcient size for single crystal both because of the occurrence of mixed species, and because structure determination, provided a good starting model is the spectra of the various candidates are quite similar.Fig. 7 available. In the case of tin sulfide in zeolite Y the quality of shows the spectra in question.Trace (a) is the nCH stretching the refinement was insuYcient to completely define the cluster, region of Me4Sn chemisorbed in Na56Y, any excess physisorbed due to significant symmetry-related disorder intrinsic in the Me4Sn having been pumped away. This represents a saturation problem, and possibly dynamic disorder, therefore the inforloading of four Me4Sn per a-cage, in which the Me4Sn moieties mation gleaned must be taken cautiously due to the known are most likely adsorbed to the Na+ cations with interactions pitfalls of locating extraframework species in zeolites.36 between the methyl hydrogens and the oxygen framework Nevertheless some further evidence for its structure and sites.Trace (b) of Fig. 7 is the nCH IR spectrum of the location was suggested.Using the low temperature (15 K) corresponding situation in HY whereby trimethyltin acts as synchrotron PXRD data, a preliminary Rietveld structure an extraframework cation, probably situated above the four refinement of the zeolite framework only was made in the tetrahedrally arranged, puckered oxygen six-ring sites, with up space group Fd39m, in order to locate electron density on a to three of the adjacent oxygens coordinating the tin centre.Fourier diVerence map. All electron density was located in the Trace (c) of Fig. 7 is the corresponding nCH IR spectrum of a-cage, the b-cage being completely empty, as was expected the product of thermal treatment of the sample in (b), expected due to the size of the precursor species.This confirms the ‘ato be roughly a 50–50 mixture of trimethyltin and dimethyltin cage specific’ nature of this kind of intrazeolite MOCVD anchored species (due to 5.5–6 total protons reacting with 4 synthesis. The most significant concentration of electron dentetramethyltin species). sity was located over the site II positions (see Fig. 1), centred Tetramethyltin is a regular tetrahedral molecule belonging above the three prominent oxygens of the six-ring in the ato the point group Td.The set of twelve displacement vectors cavity, at a distance of about 2.4 A ° . Introducing a tin atom at representing the CMH stretches generate an irreducible repthis location significantly reduced the refinement residuals and resentation Cvib=A1+E+T1+2T2. Of these, only the T2 are x2.The occupancy of this site, determined to be slightly less IR active. The predicted two major T2 nCH bands of Me4Sn than one quarter, is indicative of a single species symmetry in NaY were observed at 2975 and 2902 cm-1. These values disordered over the four crystallographically inequivalent sites, correspond to the reported nasym and nsym modes of liquid rather than multiple species each occupying one of the sites Me4Sn observed at 2982 and 2914 cm-1 respectively.37 In that which would lead to occupancies of one half or higher.No work it was assumed that the vibrational coupling between significant electron density was observed at any of the other the nCH methyl group and SnMC skeletal modes was small. common cation sites of the a-cages of zeolite Y.It was possible On the basis of this argument, it might be expected that the to introduce a second distinct tin at a location of high electron diVerence in nCH vibrational frequencies between Me4Sn, density which, due to the threefold symmetry of the site, formed Me3Sn and Me2Sn species anchored in the a-cage of zeolite Y a tetrahedron with the first with tin–tin distances of approxishould be small.This is the case in practice. mately 3.5 A ° , consistent with the sulfur-bridged tins of the The nine internal coordinates representing CMH stretches proposed adamantanoid geometry. This also resulted in sigfor a C3v pyramidal, anchored ZO-Sn(CH3)3 moiety generate nificant reduction in residuals. Beyond this point, at which an irreducible representation Cvib=2A1+A2+3E.Of these A1 x2=2.4, Rp=9.1 and weighted residual wRp=12.6 (Table 1), no further improvement in the refinement could be made, despite the observation of electron density on the Fourier diVerence map at locations suitable for sulfur positions. Consequently, no sulfur atoms are included in the model. Although the Rietveld refinement of the data was not able to unambiguously define the structure of the tin sulfide cluster, the information that was obtained was consistent with the proposed adamantane geometry and location exclusively inside the zeolite a-cages.Table 1 Atom positions and parameters of the best model for a single Sn4S64+ cluster in zeolite Y x y z Uiso/A ° 2 Si/Al(1/2) -0.0549(3) 0.1328(2) 0.0356(3) 0.0039 O(3) 0 0.1111(6) -0.1111(6) 0.0366 O(4) -0.0023(5) -0.0023(5) 0.1459(6) 0.0050 O(5) 0.0752(4) 0.0752(4) -0.0317(7) 0.0176 O(6) 0.0713(6) 0.0713(6) 0.3224(8) 0.0068 Sn(7) 0.2421(7) 0.2421(7) 0.2421(7) 0.1377 Sn(8) 0.411(13) 0.411(13) 0.205(15) 0.1292 statistics of refinement x2=2.4 wRp=12.6 Rp=9.1 lattice parameter, a/A ° 24.6537(5) r[Sn(7)MO(4)]/A ° 2.398(20) r[Sn(7)MSn(8)]/A° 3.5(3) r[Sn(8)MSn(8)]/A ° 4.7(8) Fig. 7 Mid-IR spectra of nCH region for (a) tetramethyltin impreg- [Sn(8)MSn(7)MSn(8)]/° 83(6) nated NaY, (b) anchored trimethyltin in acid zeolite Y, (c) anchored [O(4)MSn(7)MO(4)]/° 101.3(9) dimethyltin and trimethyltin in acid zeolite Y 1286 J.Mater. Chem., 1998, 8(5), 1281–1289and E are IR active and so five modes are expected. The two to values such as 3.31 mm s-1 (ref. 40) {Me2Sn[N,N¾-(2- hydroxytrimethylene)bis(salicylaldamine)]}, agreeing best with major bands observed are blue-shifted from those of chemisorbed (CH3)4Sn,NaY by roughly 15 cm-1 for the symmetric the splitting determined here. On going to more symmetric, trigonal bipyramidal structures with methyl groups both equa- stretch and 4 cm-1 for the asymmetric stretch. These nCH bands, however, remain within the linewidths of the tetra- torial {Me2Sn[N-(2-hydroxyphenyl)salicylaldimine]}, the quadrupole splitting drops to 3.04.41 This suggests that the methyltin vibrations.It is possible, through consideration of the characteristic motions which compose the nCH normal dimethyltin species, anchored over the three prominant oxygens of site II, probably has a structure intermediate modes of ZO-Sn(CH3)3, to separate the five normal nCH modes into two sets having their origins in the symmetric and between the latter two examples, a distorted ‘octahedron’ with one missing ligand, or a distorted trigonal bipyramid, with asymmetric nCH modes of uncoupled CH3 groups.The result is that the lower frequency band may be assigned as A1+E one methyl group equatorial and one axial. It is also interesting to note in this data the possiblity of a (A1, CH3), while the higher frequency major band includes the components A1+2E (E, CH3), the A2 contribution being Goldanskii–Karyagin eVect, reflected in the diVerence in areas of the two peaks of the quadrupole doublet while similar peak IR inactive. Upon heating the ZO-Sn(CH3)3 samples containing b- widths (FWHH) are maintained.42 This eVect arises when there is an anisotropy in the Mo�ssbauer recoil-free fraction, i.e., the protons to 150 °C, for 2 h, some ZO-Sn(CH3)2 species are formed.The spectrum shows a further shift of the band maxima mean-square vibrational amplitude is not cubically symmetric. This is quite reasonable in a situation in which the recoiling to higher frequency, and a change in the relative intensities.The lower frequency band shifts by 3 cm-1 and the higher by species is eVectively anchored to a surface site (i.e. the zeolite framework) such that the vibrational amplitude perpendicular 5 cm-1 with respect to the anchored ZO-Sn(CH3)3 species. As above, under C2v, the irreducible representation Cvib= to the ‘wall’ will likely be very diVerent from that parallel to it.Further work would be required to prove such a point, but 2A1+A2+2B1+B2 can be split into two groups, A1+B1 (A1, CH3) for the lower frequency band and A1+B1+B2 (E, CH3) it remains an interesting possibility. Fig. 3(b) shows the Mo�ssbauer spectrum of Sn4S6-Y. The for the higher frequency band. In these considerations, a rigid model for the methyl groups isomer shift of 1.25 mm s-1 indicates that the oxidation state of tin is IV, as mentioned earlier, and the quadrupole splitting has been assumed.While it is also possible to predict the active modes using a non-rigid Bunker-type model in which the is 2.28 mm s-1, considerably smaller than that of the precursor material. However, octahedral or tetrahedral symmetries have methyl groups are free to rotate, in the sterically hindered environment of the a-cage containing a total of four ZO- no intrinsic quadrupole splitting, and no alkyl groups are expected to contribute, as only sulfide and oxide ligands are Sn(CH3)3 and ZO-Sn(CH3)2 species, it is unlikely the methyl groups would truly be free rotors, and therefore the rigid expected.Therefore, an unsymmetric distribution of the diVerent ligands about a tetrahedral or octahedral tin(IV) centre will oscillator approach was preferred. There are very weak satellite bands or shoulders in the nCH spectra which are, in every probably describe the structure, consistent with the proposed adamantane structure. The isomer shift of the encapsulated case, split oV from the major bands by about 35 to 130 cm-1, precluding the simple assignment of these bands as components cluster, 1.25 mm s-1, agrees very closely with that measured for the methyl-capped cluster Me4Sn4S6, 1.28 mm s-1.of the nCH ZO-Sn(CH3)3/ZO-Sn(CH3)2 modes.38 Such splittings are not readily accounted for through arguments of However, the quadrupole splitting of the encapsulated cluster is considerably bigger than that of Me4Sn4S6 (2.28 vs.multiple anchoring site eVects, splitting of degeneracy through symmetry eVects, or even correlation coupling eVects, 1.34 mm s-1), probably because the latter has a more rigid symmetry, and alkyl groups more closely match the electrone- which are all typically 20 cm-1 or less. No Raman modes of Me4Sn occur at appropriate frequencies such that a change of gativity of sulfide ligands than zeolite oxygens do.The observation of a single kind of tin(IV) Mo� ssbauer site is consistent selection rules due to site symmetry eVects could result in IRactive modes at these positions. The best interpretation is that with the Rietveld PXRD structural conclusion, that the Sn4S64+ cluster is four-fold positionally disordered over the these shoulders and weak bands be assigned as overtones and combinations with low frequency fundamental modes, consist- four site II oxygen six-ring positions in the a-cage of zeolite Y.Variable temperature Mo�ssbauer studies (10–300 K) will be ent with the assignment of similar shoulders in the spectrum of Me4Sn.37 required to distinguish static from dynamic Sn4S64+ disorder in this system.Mo�ssbauer spectroscopy. The isomer shift, 1.27 mm s-1, measured in the Mo�ssbauer spectrum of dimethyltin anchored in Optical spectroscopy. In Fig. 8 the optical reflectance data for four forms of tin sulfide are presented. Qualitatively, the zeolite Y, Fig. 3(a), indicates that the oxidation state of tin is IV. The large quadrupole splitting, 3.44 mm s-1, speaks of the spectrum of the encapsulated clusters, Sn4S6-Y, is considerably blue-shifted as compared to the bulk semiconductor berndtite, low symmetry surrounding the tin(IV) centres.The sample of pure dimethyltin in zeolite Y was prepared by titrating the SnS2. This is as expected for confinement of electrons within a quantum dot. However, it is quite similar to the spectrum of protons while the sample was held at 150 °C, so that the reactivity was similar for the a- and b-protons.Thus every tin the molecular species, (CH3)4Sn4S6, indicating that the excitonic confinement regime was overshot. This species is within centre reacted with two protons immediately, and therefore about three tin centres, uniformly dimethyl species, were the very strong confinement size regime and is essentially molecular.Without good information about the eVective anchored in each a-cage. Quadrupole splittings of this size (3.44 mm s-1) are usually indicative of low symmetry, such masses of the electrons and holes in SnS2 it is not possible to calculate the optimum cluster size for excitonic confinement, as trigonal bipyramidal stereochemistry, for compounds (R2SnX2)n or (R3SnX)n where X is F-, Cl-, Br-, I-, or O2- but it is clear that the size obtained was too small.An EHMO calculation for the isolated cluster was consistent with the donor.31 Among alkyltin species with oxygen containing ligands, the size of the quadrupole splitting is dominated by qualitative molecular orbital diagram shown in Fig. 9, and indicated that the HOMO and LUMO were composed mainly the number and position of the alkyl ligands. For example, for dimethylated amino acid and SchiV base derivatives with of sulfur 3p and tin 5s and 5p orbitals, respectively, and give rise to a ligand-to-metal charge-transfer electronic transition octahedral structure and methyl groups trans, quadrupole splittings of up to 4.02 mm s-1 have been reported39 as seen in the optical spectrum, Fig. 8.43 In considering the possibility of intercluster coupling, the EHMO method was [Me2Sn(acac)]. Distorting the octahedra such that the CMSnMC bond angle is 160° reduces the quadrupole splitting used to determine that clusters would need to have centre-to- J. Mater. Chem., 1998, 8(5), 1281–1289 1287Fig. 10 UV–VIS diVuse reflectance spectra of the products of (a) the Fig. 8 UV–VIS absorbance spectra of four forms of tin sulfide, first loading of Sn4S6 clusters in zeolite Y, (b) a second loading, (a) molecular (CH3)4Sn4S6, (b) zeolite encapsulated Sn4S6-Y, titrating regenerated protons in order to add more tin sulfide material (c) [(CH3)4N]2Sn3S7 denoted TMA-SnS-1, (d) bulk SnS2, berndtite to each cluster, and (c) a third loading, increasing the cluster again, (d) bulk SnS2 tion edge on going from Sn4S6 clusters, (a), to clusters resulting from a second loading, (b), and from a third loading, (c).The red-shift is probably indicative of less confined charge carriers, due to an increase in the cluster size.47 This supports the idea that the titration of regenerated protons and formation of more tin sulfide material in each a-cage does increase the nuclearity of the initial tin sulfide clusters.On a final note, one would expect that adamantanoid Sn4S6-Y might be grown from intrazeolite Me4Sn and H2S reagents, as it is well known that the solution phase reaction of MeSnCl3 with H2S yields adamantane clusters, Me4Sn4S6, with terminal methyl-capping groups. The four nucleophilic methyl groups of Me4Sn4S6 are simply substituted in the zeolite encapsulated analogue by charge-balancing framework oxygens to give (ZO)4Sn4S6 according to the zeolate model of bonding.23 Conclusions Fig. 9 Extended Hu� ckel molecular orbital, EHMO, energy diagram representing a zeolite-encapsulated Sn4S6 cluster The application of size-limiting intrazeolitic MOCVD techniques has been demonstrated to be a method of choice for the synthesis of arrays of zeolite encapsulated IV–VI semi- centre distances of less than 9 A° before any modification of their electronic structure was calculated.While the clusters conductor clusters. Step-wise, molecule-by-molecule control over the intrazeolite reaction processes is possible by employing were not expected to occupy the centres of the a-cages, which are 13 A ° apart, there would nevertheless be an average inter- non-intrusive, quantitative, in situ methods of observing them, as shown in this work.Charge-balancing tin sulfide clusters of cluster distance of 13 A ° over the whole material, which puts it outside this range for potential intercluster coupling. This of average Sn4S64+ stoichiometry have been synthesised in the acages of zeolite Y, possibly with an adamantanoid geometry.course does not account for any through-bond coupling, possible via the zeolite lattice. In comparison, in a microporous It is expected that the methodologies presented in this paper will be generally applicable to other semiconductor clusters layered tin(IV) sulfide having the formula R2Sn3S7 with Sn4S3, broken cube clusters linked by double sulfur bridges,44 the within other media in syntheses that are equally controllable.It should be noted that, at the time this work was done, clusters are less than 8 A° apart and are anticipated to be significantly coupled through the double sulfur bridges. clusters in zeolites had an upper limit of 13 A ° imposed by the restricted choice of large-cage zeolites.Since then the world The intermediate value of the absorption edge of R2Sn3S7, Fig. 8, with respect to those of the bulk and molecular cluster of size-controlled large-aperture mesoporous silicas, beginning with MCM-41,48 has opened up, and ordered channels of forms of tin sulfide is noteworthy both in terms of connectivity arguments that have been put forward regarding the dimen- 20–100 A ° are available as hosts for this type of CVD topotaxy.49 sionality of such materials,45 and in the context of semiconductor quantum confinement and cluster-framework topological complements.46 The authors are indebted to Dr Robert Broach and Dr Robert L.Bedard for valuable discussion regarding the Rietveld As was discussed above, further growth of the tin sulfide clusters was possible and resulted in red-shifting of the absorp- analysis.The PXRD measurements were carried out at the National Synchrotron Light Source, Brookhaven National tion maximum, but the material was insuYciently homogeneous for good characterization. Nevertheless, Fig. 10 shows Laboratories which is supported by the US Department of Energy, Division of Materials Sciences and the Division of the results of cluster growth: a definite red-shift of the absorp- 1288 J.Mater. Chem., 1998, 8(5), 1281–128921 J. , H.-B. Burgi, J. C. Thibeault and R. HoVman, J. Am. Chemical Sciences. We acknowledge the Natural Sciences and Chem. Soc., 1978, 100, 3686. Engineering Research Council of Canada (NSERC) as well as 22 G.Clazaferri, L. Forss and I. Kamber, J. Phys. Chem., 1989, 93, the Canadian Space Agency (CSA) and UOP for financial 5366. support of this endeavour. In addition, C.B. thanks NSERC 23 G. A. Ozin, C. L. Bowes and M. R. Steele, Mater. Res. Soc. Symp. and the University of Toronto for financial support during her Proc., 1992, 277, 105. 24 G. A.Ozin and S. Ozkar, Chem. Mater., 1992, 4, 51; G. A. Ozin, graduate work. A. Kuperman and A. Stein, Angew. Chem., Int. Ed. Engl., 1989, 101, 373; G. A. Ozin, S. Ozkar and R. A. Prokopowicz, Acc. Chem. Res., 1992, 25, 553; G. A. Ozin, Adv. Mater., 1992, 4, 612; G. A. Ozin, Adv.Mater, 1992, 4, 11. References 25 D. H. Olsen and E. Dempsey, J. Catal., 1969, 13, 221. 26 J. W. Ward, J. Catal., 1967, 9, 396. 1 P. Day, Chem. Br., 1996, July, 29. 27 G. A. Ozin, S. Ozkar and L. McMurray, J. Phys. Chem., 1990, 2 M. A. Reed, J. N. Randall, R. J. Aggarwal, R. J. Metyi, T. M. Moore 94, 8297. and A. E.Wetsel, Phys. Rev. L ett., 1988, 60, 535. 28 T. Yamazaki, I. Watanuki, S. Ozawa and Y. Ogino, L angmuir, 3 R. F. Pease, in Nanostructures and Mesoscopic Systems, ed. 1988, 4, 433. W.P. Kirk and M. A. Reed, Academic Press, Toronto, 1992. 29 G. A. Ozin, S. Ozkar and L. McMurray, J. Phys. Chem., 1990, 4 F. R. Waugh, M. J. Berry, D. J. Mar, R. M. Westervelt, 94, 8289. K. L. Campman and A. C. Gossard, Phys. Rev. L ett., 1995, 75, 705. 30 L. Quanzi, A. Ruiming and X. Zhiyuan, in New Developments in 5 K. Kash, D. D. Mahoney, B. P. Van der Gaag, A. S. Gozdz, Zeolite Science and T echnology, ed.Y. Murakami, A. Iijima and J. P. Harbinson and L. T. Florenz, J. Vac. Sci. T echnol. B, 1992, J. W. Ward, Elsevier, New York, 1986, p. 487; J. J. Lunsford, 10, 2030. P. N. Tutunjian, P. Chu, E. B. Yeh and D. J. Zalewski, J. Chem. 6 G. A. Ozin, Adv.Mater., 1992, 4, 612. Phys., 1989, 93, 2590 and references therein. 7 N. Herron, Y. Wang and H. Eckert, J.Am. Chem. Soc., 1990, 31 R. V. Parish, in Mo�ssbauer Spectroscopy Applied to Inorganic 112, 1322. Chemistry, ed. G. J. Long, Plenum Press, N.Y., 1984, vol. 1, ch. 16. 8 M. A. Matchett, A. M. Viano, S. L. Adolphi, R. D. Stoddard, 32 G. A. Ozin, S. Ozkar and G. D. Stucky, J. Phys. Chem., 1990, W. E. Buhro, M. S. Conradi and P. C. Gibbons, Chem. Mater., 94, 7562. 1992, 4, 508. 33 J.Howard and Z. A. Kadir, Spectrochim. Acta, Part A, 1985, 41, 9 M. L. Steigerwald, A. P. Alivisatos, J. M. Gibson, T. D. Harris, 825 and references therein. R. Korton, A. J. Muller, A. M. Thayer, T. M. Duncan, 34 A. K. Cheetham and A. J. Skarnulis, Anal. Chem., 1981, 53, 1060. D. C. Douglass and L. E. Brus, J. Am. Chem. Soc., 1988, 110, 3046; 35 B. Krebs, Angew. Chem., Int. Ed.Engl., 1983, 22, 113. Y-M. Tricot and J. H. Fendler, J. Phys. Chem., 1986, 90, 3369; 36 X. Hu and W. Depmeier, Z. Kristallogr., 1992, 201, 99. X. K. Zhao, S. Baral, R. Rolandi and J. H. Fendler, J. Am. Chem. 37 W. F. Edgell and C. H. Ward, J. Am. Chem. Soc., 1955, 77, 6486. 38 P. R. Bunker, Molecular Symmetry and Spectroscopy, Academic Soc., 1988, 110, 1012; J. H. Fendler and F. C.Meldrum, Adv. Press, New York, 1977; R. L. Flurry, Symmetry Groups, T heory Mater., 1995, 7, 607. and Chemical Applications, Prentice Hall, NJ, 1980. 10 Y. Wang, A. Suna, W. Mahler and R. Kasouuwski, J. Chem. Phys., 39 T. K. Sham and G. M. Bancroft, Inorg. Chem., 1975, 14, 2281. 1987, 87, 7315; E. S. Smotkin, R. M. Brown, Jr., L. K. Rabenberg, 40 J. N. R. Ruddick and J. R. Sams, J. Organomet. Chem., 1973, K. Salomon, A. J. Bard, A. M. A. Campion, M. A. Fox, T. Mallouk, 60, 233. S. E.Webber and J. M. White, J. Phys. Chem., 1990, 94, 7543. 41 J. N. R. Ruddick and J. R. Sams, J. Chem. Soc., Dalton T rans., 11 F. Yan and J. M. Parker, J. Non-Cryst. Solids, 1989, 112, 277; 1974, 470. R. D. Stramel, T. Nakamura and J. K. Thomas, J. Chem. Soc., 42 V. I. Goldanskii, G. M. Gorodinskii, W. V. Karyagin, Faraday T rans., 1988, 84, 1287. L. A. Korytko, L. M. Krizhanski, E. F. Makarov, I. P. Suzdalev 12 N. Herron, Y. Wang, M. M. Eddy, G. D. Stucky, D. E. Cox, and V. V. Khrapov, Dokl. Akad. Nauk SSSR, 1962, 147, 127; K. Moller and T. Bein, J. Am. Chem. Soc., 1989, 111, 530; S. V. Karyagin, Dokl. Akad. Nauk SSSR, 1963, 148, 1102. J. E. MacDougall, H. Eckert, G. D. Stucky, N. Herron, Y. Wang, 43 C. L. Bowes, MSc Thesis, University of Toronto, 1991. K. Moller and T. Bein, J. Am. Chem. Soc., 1989, 111, 8006. 44 C. L. Bowes, S. Petrov, G. Vovk, D. Young, G. A. Ozin and 13 C. B. Murray, C. R. Kagan and M. G. Bawendi, Science, 1995, R. L. Bedard, J.Mater. Chem., 1998, 8, 711. 270, 1335. 45 E. A. Axtell III, J-H. Liao, Z. Pikramenou, Y. Park and 14 A. P. Alivisatos, Science, 1996, 271, 933. M. G. Kanatzidis, J. Am. Chem. Soc., 1993, 115, 12191; E. A. Axtell 15 P. M. PetroV and G. Madeiros-Ribeiro,MRS Bull., 1996, 21, 50. III, J-H. Liao, Z. Pikramenou and M. G. Kanatzidis, Chem. Eur. 16 C. A. Fyfe, J. M. Thomas, J. Klinowski and G. C. Gobbi, Angew. J., 1996, 2, 656. Chem., 1983, 22, 259. 46 C. L. Bowes and G. A. Ozin, Mater. Res. Soc. Symp. Proc., 1992, 17 G. A. Ozin and J. Godber, J. Phys. Chem., 1988, 92, 4980; 2841. 286, 93. 18 R. W. Frei and J. D. MacNeil, DiVuse Reflectance Spectroscopy in 47 L. E. Brus, J. Chem. Phys., 1984, 80, 4403; Y. Kayanuma, Phys. Environmental Problem Solving, CRC Press, Cleveland, OH, 1973. Rev. B, 1988, 38, 9797. 19 A. C. Larson and R. B. Von Dreele, Generalized Structure Analysis 48 C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and System, LAUR 86-748, Los Alamos National Laboratory, Los J. S. Beck, Nature, 1992, 359, 710. Alamos, NM 87545, USA. 49 C. L. Bowes, A. Malek, G. A. Ozin, Chem. Vap. Deposit. (Adv. Mater.), 1996, 2, 97. 20 ICONCL software: OS/2 Version, MS-Fortran 5.0, 30 November 1989, Modified ICON8 from QCPE 344, Calzaferri Group, Institute for Inorganic Chemistry, University of Bern, Switzerland. Paper 7/08093J; Received 11th November, 1997 J. Mater. Chem., 1998, 8(5), 1281–1289 12