J. MATER. CHEM., 1994, 4( 12), 1913- 1920 Investigations of Structure and Protonic Conductivity in the so-called Tin Zeolites Gary B. Hix,*aRobert C. T. Slade," Kieran C. Mollof and Bernard Ducourantc a Department of Chemistry, University of Exeter, Stocker Road, Exeter, UK EX4 4QD Department of Chemistry, University of Bath, Claverton Down, Bath, UK BA2 7AY Laboratoire des Agregats Moleculaires et Materiaux Inorganiques, URA CNRS 79, Universite Montpellier 2, Place Eugene Bataillon, 34095 Montpellier Cedex 5, France 'Tin zeolites', composites containing a zeolite and SnO,, have been produced from parent zeolites (mordenite, zeolites X, Y and A) by thermal and microwave methods. Materials characterisation employed X-ray powder diffractometry (XRD), 29Si and 27AI MAS NMR, "'Sn Mossbauer spectroscopy, and ac conductivity investigations. Both XRD and MAS NMR studies show degradation of the zeolite framework in materials prepared by the thermal route.The extent of the damage depends upon the tin salt used in synthesis and is higher for those zeolites with higher Al contents. Mossbauer spectra are dominated by a central peak arising from microdispersed SnO,, but a second Sn'" environment is also present in materials prepared thermally from zeolites X, Y and A. All the 'tin zeolites' exhibit protonic conductivities greater than those of the parent zeolites, with thermally prepared samples exhibiting conductivities up to an order of magnitude greater than corresponding materials prepared by the microwave method. The introduction of tin(1v) oxide into mordenitei,2 and zeolite Y3 has been reported in the literature.The emphasis of those investigations of 'tin zeolites' was on the ionic (H' ) conduction in the materials, with structural investigations limited to X-ray powder diffraction studies. Samples were made by melting varying quantities of tin@) salts (SnCl, or SnSO,) into the zeolite frameworks and oxidising them in situ by thermal treatment, giving dispersed SnOz with an enhanced specific surface area relative to pure SnO,. For the samples in which mordenite was used as the zeolite component, the use of small quantities of tin(11) salt in the syntheses was reported to result in tin ion exchange only (the oxidation state of the exchanged ions was not reported), and increasing the amount of tin salt was said to produce SnO,., Preparations involving zeolite Y were reported to give both SnO, and ion e~change.~ All products were therefore composite in nature.The ionic (protonic) conduction in these so-called 'tin zeolite' materials has been studied by ac and dc techniques.'-, Measurements concerning the mordenite-based materials were carried out in a water-moistened atmosphere, at 100% relative humidity (RH), and conductivity crz lop2S cm-' at 100"C. Studies using dc techniques showed that the conducting species to be the proton, the conduction being assigned to a proton-hopping mechani~m.~ Measurements on zeolite Y-derived samples at 75% RH gave 0% lop2S cm-I at 116 "C; the conduction mechanism was assigned in that case as being a polyatomic vehicle conduction of proton^.^ Since water molecules act as the 'vehicles' for the charge carriers (H'), the water content of the sample affects the conductivity directly. There is a consequent variation of conductivity with relative humidity, with higher relative humidities leading to higher conductivities.Hydrous tin@) oxide itself (Sn0,-nH,O) is formed as a white precipitate on hydrolysis of tin@) salts. It is an ion- exchange material which has been shown to be a fast protonic conductor, with a room-temperature conductivity CJ = 4 x S cm-l.' This type of material has been termed a particle hydrate,6 consisting of charged particles with partly protonated oxide-hydroxide surfaces separated by a weakly acidic aqueous region.We now report an in-depth study of Sn0,-containing composites produced by two different synthetic routes and derived from mordenite, and from zeolites A, X and Y. The first route is that of Knudsen et d.,'resulting in materials which 'have been ion-exchanged' and also contain a separate tin oxide phase. The second method employed the microwave heating and hydrolysis of an organotin compound. A compari-son of conductivity and microstructures in these materials is presented. Experimental Materials synthesized using the salt-melt procedure are desig- nated T-series materials (e.g. T-Sn-X refers to the material made by the salt melt route with zeolite X), whilst those made following microwave procedures are designated M series.Tin@) sulfate, tin(r1) chloride and triphenyltin chloride (BDH Chemicals, AnalaR grade) were used as supplied with- out purification. Zeolites (mordenite, zeolites A, X and Y) were supplied by Laporte Inorganics. Analyses (XRF) of the parent zeolites are given in Table 1. T Series Tin oxide-containing zeolites were synthesized following the methods described by Knudsen et a/.' Tin@) sulfate (1.4 g) was intimately ground with sodium zeolite (4.0 g), and then heated at 400 "C for 20 h. The products were washed for 5 x 24 h in 200 cm3 of deionised water (with filtration and fresh water daily), and then stored in an atmosphere of 50% RH (over a saturated solution of sodium hydrogensulfate).Tin(r1) chloride dihydrate (3.0 g) was intimately ground Table 1 Chemical analyses (XRF) and Si:Al ratios of the parent zeolites used in this study Si :A1 SiOz A1,0, Na,O zeolite (%) (Yo) ("/.I XRF NMR Na mordenite 81.80 10.90 6.88 6.15 6.11 H mordenite Na-A 90.90 43.08 7.31 36.23 <100ppm 8.05 9.15 1.01 8.91 1" Na-X 49.75 31.82 18.43 1.35 1.32 Na-Y 62.72 22.94 12.11 2.49 2.41 " Only one site type in the "Si NMR spectrum. with sodium zeolite (4.0 g) and heated at 100°C under dynamic vacuum for 3 h to remove any excess of water. The mixtures were heated at 200 "C for 20 h in air, oxidising the tin, and then at 400°C for 90min, to remove the tin(1v) chloride formed. The products were washed and stored as above.A preparation using tin(I1) chloride dihydrate was also carried out using hydrogen mordenite. This sample has been designated T-Sn-MorH. Where appropriate in the further discussion, the anion characteristic of the tin-containing reagent in the preparation is denoted by (C1-) or (SO,"). Thus T-Sn-X(Cl-) denotes a material prepared from zeolite X and tin@) chloride. M Series This synthesis involved hydrolysis of an organotin species to form tin(1v) oxide, following Ashcroft et aL7 A solution of triphenyltin chloride ( 1.0 g) in ethanol (30 cm3) was placed in a flask with the sodium zeolite (5.0 g). The flask was covered and then subjected to nine 1 min bursts of microwave heating (at 100% power) in a standard Samsung 600 W microwave oven.Between bursts of microwave radiation, the solutions were allowed to cool, and ethanol that had been lost by evaporation was replaced. Care was needed with this pro- cedure due to the flammability of ethanol. The materials were recovered by filtration, washed and stored as above. XRD Powder diffraction profiles were recorded on a Philips diffractometer (PW 1050/25 goniometer; Cu-Ka radiation, i= 1.54178 A). Data were collected for 10 s in steps of 0.1" of 20. Observed patterns were compared with computer-generated simulations.* MAS NMR 29Si(59.584 MHz) and 27Al(78.152 MHz) MAS NMR spectra at ambient temperature were recorded on a Varian VXR 300 spectrometer by the SERC Solid State NMR Service (Durham).Relaxation delays in recording 29Si spectra were 30s, determined as sufficiently long to avoid all saturation effects. The corresponding delay for 27Al spectra was 0.5 s. 29Si spectra were deconvoluted assuming gaussian line- shapes, thereby giving spectral parameters for the different Si sites. '19Sn Mossbauer Spectroscopy T-series samples were finely ground and suspended in an inert matrix. Once mounted, the samples were maintained at 77 K by a continuous-flow liquid-nitrogen cryostat linked to a digital temperature controller. Temperature stability was held to within kO.1 K throughout the experiment. Spectra were collected on a constant-acceleration Mossbauer spectrometer (Bath), using a Ca11gmSn03 source. The drive unit was con- trolled by a digitally generated saw-tooth waveform to pro- duce constant acceleration.Velocity calibration was based upon the spectrum of cr-Fe, with CaSnO, being used as the zero-velocity reference. Spectra of M-series materials were similarly recorded at Montpellier. Data were acquired through a multi-channel scaling analyser as two spectra related through a mirror plane. Folding these two spectra improved the counting statistics and removed the geometric effect. Experimental Mossbauer spectra were computer-simulated (with a combination of Lorentzian lines) and the calculated J. MATER. CHEM.. 1994, VOL. 4 spectrum was subsequently refined, using the 'General Mossbauer Fitting Program' written by K. Ruenbauer and T. Birchall.' Isomer shifts are expressed with respect to the source.Impedance Analysis Microcrystalline samples of the modified zeolite systems and the parent zeolites were compressed in a 13 mm die at 5 tonne cm-2 to give pellets of cu. 1 mm thickness. To aid the binding of the pellets, a small quantity of water was added prior to pressing. Pellet faces were coated with conduc- tive silver paint (Acheson Electrodag 915), and copper or nickel disc electrodes were attached using silver paint as ad- hesive. Samples were then allowed to equilibrate at ambient temperatures in a desiccator over saturated aqueous sodium hydrogensulfate (RH =50%). Sample assemblies were mounted in a brass cell which held six samples simultaneously. Impedance spectra (admittance or impedance plane) were collected in the range 5 Hz-1 MHz (oscillation voltage 100 mV) using a Hewlett-Packard 4192A LF impedance analyser controlled by an IB M-compatible computer, which used software embedding EQUIVCRT mod- elling software." Temperature was controlled in the range 293 < T/K <353 by immersion of the cell in a controlled-temperature water-ethylene glycol bath, allowing at least 30 min for equilibration at each temperature (this period was found by experience to be in excess of the minimum necessary to give temporal stability of impedance spectra).RH within the cell was controlled by means of saturated aqueous sodium hydrogensulfate solution placed in a sponge at the bottom of the cell. Results and Discussion Analyses and Si :Al Ratios Analyses (XRF) of the samples, and Si :A1 ratios determined thereby, are given in Table 2.Application of Loewenstein's rule, which disallows Al- 0-A1 linkages in zeolite frame- works," also allows the Si:Al framework ratios to be calcu- lated from the relative intensity contributions of different sites in the 29Si MAS NMR spectra via eqn. (1). Si:Al= 114,mII (n1/4)1~.~ (1) m=0,4 m=0,4 where 14,mis the relative intensity of a line assigned to a Q4(mAl) silicon site (rn is the number of -OAl linkages Table 2 Chemical analyses (XRF) of the tin zeolite samples Si : A1 ratio SiOz A1,0, Na,O SnO, material (%) (Yo) (%) (Yo) XRF NMR T-Sn-A(CI -) 29.13 25.25 12.36 11.36 0.98 1.08 T-Sn-A( 24.56 21.30 11.92 16.30 0.98 1.03 T-Sn-X(C1-) 33.68 23.45 9.81 11.04 1.24 1.32 T-Sn-X(SO4'-) 30.74 21.10 5.49 18.14 1.22 1.34 T-Sn-Y (Cl- ) T-Sn-Y(SO,,-) 44.52 54.48 16.40 13.77 7.92 3.89 8.94 19.46 2.35 2.30 2.58 2.78 T-Sn-Mor( Cl -) 54.48 7.23 2.51 12.25 6.39 7.34 T-Sn-Mor( SO,'-) 50.73 7.06 3.86 18.04 6.12 7.16 T-Sn-MorH(C1F) 45.79 4.03 0.08 24.24 9.52 7.39 M-Sn-A 33.83 28.45 16.24 6.32 1.01 -~ M-Sn-X 36.11 23.10 13.38 5.82 1.32 1.39 M-Sn-Y 51.60 19.13 9.5 1 2.96 2.32 2.43 M-Sn-Mor 71.88 9.98 5.59 2.82 6.11 6.25 (Cl-) indicates samples made using SnCl,, (SO,2-) indicates samples made using SnSO,.J. MATER. CHEM., 1994, VOL. 4 around the Si atom).The values thus obtained are also given in Table 2. T Series It can be seen (Table 2) that the samples made using SnC1, contain less tin than the samples made using SnSO,. This is to be expected since the preparation using SnC1, involves a step in which SnCl, is removed as vapour by heating the samples to over 380°C. The reduction in sodium content, with respect to the parent zeolites, indicates that some ion exchange has taken place. However, the amount of tin in the samples is greater than would be required to compensate for the observed Na loss; the remainder of the tin is present as oxide. The Si : A1 ratios calculated from XRF data are closer those of the parent materials than are those obtained from NMR spectra. The values obtained from XRF account for all of the silicon and aluminium present in the whole sample, whereas the NMR method allows calculation of the %:A1 ratio for the framework only.The higher 'Si: A1 ratio' resulting from NMR data is a consequence of leaching of aluminium from the framework during processing. M Series The tin contents (Table 2) are lower than those of the corre- sponding T-series materials. The reductions in sodium contents of samples are consistent with the proportion of zeolite in the final material, i.e. there has been no ion exchange. The Si:Al ratios determined by XRF and from NMR are essentially those of the parent materials. There is no evidence for leaching of aluminium from the zeolite frameworks. XRD Profiles TSeries Data for materials prepared from the same parent zeolite differ markedly depending on whether SnC12.2H,0 or SnSO, was used as the other reagent (Fig.1-4). Materials which had SnS0, as the tin precursor exhibited sharp narrow XRD lines associated with the zeolite frame- work. These lines were, however, reduced in intensity with respect to the zeolite parents; this is due to (i) reduction in the amount of crystalline zeolite present (by degradation or fragmentation of the zeolite by acid leaching of framework aluminium), (ii) 'dilution' of the zeolite by introduction of the tin oxide phase (which gives additional broad features). The latter is predominant. 10 20 30 40 50 60 70 80 28/degrees Fig. 1 X-Ray powder diffraction profiles of (a) T-Sn-Mor(Cl-), (b)T-Sn-Mor (c) T-Sn-MorH and (d) M-Sn-Mor AmA.-(a,)_I ~ 1 IsI'I'I'I1 ! ' l ' l ' t ' T 7 1 ' l ' i 10 20 30 40 50 60 70 80 2 @degrees Fig.2 X-Ray powder diffraction profiles of (a) T-Sn-A(Cl-), (b)T-Sn-A(SO?-) and (c) M-Sn-A I I ' I ' I ' I ' I ' I ' 1 ' I ' 10 20 30 40 50 60 70 80 2€J/degrees Fig. 3 X-Ray powder diffraction profiles of (a) T-Sn-X(C1-), (b)T-Sn-X(S042-) and (c) M-Sn-X (a) I ' I ' I ' I ' I ' I ' I ' 1 7 1 7 ' I ' I ' I ' I ' 10 20 30 40 50 60 70 80 2Wdegrees Fig. 4 X-Ray powder diffraction profiles of (u) T-Sn-Y(C1-), (b)T-Sn-Y(S0,2-) and (c) M-Sn-Y Samples derived from SnCl2.2H,O show, in contrast, much greater reductions in XRD line intensities, the extent of the reduction varying with the zeolite parent.In the most extreme examples, those of T-Sn-X and T-Sn-Y (Fig. 3 and 4, respect- ively), there are no observable zeolite lines, only very broad lines due to small particles of tin oxide being seen. Since these samples have a lower tin content than the analogous samples made with SnSO,, the loss of intensity in the zeolite lines is attributed to disruption of the framework, which is also evident in 27Al and 29Si NMR spectra (see below). Cell parameters for the zeolite framework in the samplFs giving a zeolite-related XRD profile showed very little (< 0.1 A) variation with respect to those of the parent zeolites (Table 3). M Series The XRD profiles (Fig. 1-4) show lines attributable to the zeolite frameworks, but none due to the presence of SnO,.This is due to the low percentage and small particle size of SnO, in the samples, and would also account for the non- occurrence of the intensity reductions observed for the T-series samples. These observations indicate that the struc- tural integrity of the zeolite framework has been maintained throughout the synthetic procedure. Calculated cell param- eters vary only slightly with respect to those of the pristine zeolite parents (Table 3). NMR Spectra M Series The 27Al spectra of these materials indicated exclusively tetrahedrally coordinated Al, as in the parent zeolites. Peak positions are given in Table 4. The 29Si spectra were also almost identical in all spectral parameters to those of the zeolite parents.The 29Si spectrum of M-Sn-X is shown in Fig. 5, in which deconvolution into constituent gaussians is also illustrated. The microwave preparative procedure has no Table 3 Unit-cell parameters for tin zeolites material a/A b/A CIA )T-Sn-Mor(C1~ 18.24(6) 20.77( 6) 7.58(3) T-Sn-Mor(SO4'-) 18.13( 2) 20.46( 3) 7.50(1) T-Sn-MorH (C1 -) 18.28( 3) 20.71(3) 7.52( 1) T-Sn-A( C1 -) 24.61( 1) 24.61 ( 1) 24.61(1) T-Sn-A(SO,'-) 24.59( 3) 24.59 (3) 24.59( 3) T-Sn-X( C1- )" T-Sn-X( SO,'^ ) 25.02( 13) 25.02( 13) 25.02( 13) T-Sn-Y (Cl- )" T-Sn-Y ( -) 24.94(4) 24.94( 4) 24.94( 4) M-Sn-Mor 18.11(2) 20.5 1( 1) 7.52( 1) M-Sn-A 24.61( 1) 24.61 (1) 24.61 (1) M-Sn-X 25.00( 1) 25.00( 1) 25.00( 1) M-Sn-Y 24.95( 1) 24.95( 1) 24.95( 1) (Cl-) indicates samples made using SnCl,, (SO4'-) indicates samples made using SnSO,.No zeolite lines in the XRD profile. Table 4 Observed 27Al MAS NMR peak positions for tin zeolites __~ material 6 (tetrahedral) S (octahedral) T-Sn-Mor(C1-) 54.1 0.2 T-Sn-Mor(S0,' -1 54.7 -0.4 T-Sn-MorH(C1-) 54.1 -0.4 T-%-A( C1- ) 57.5 -0.1 T-%-A( SO,' -) 54.3 -0.2 T-Sn-X( C1- ) 59.0 -2.5 T-Sn-X(SO,'-) 59.1 -3.9 T-Sn-Y (Cl-) 59.3 -3.1 T-Sn-Y(SO4'-) 58.7 -1.5 M-Sn-Mor 54.6 --M-Sn-A 58.3 M-Sn-X 59.4 -M-Sn-Y 60.0 -(C1-) indicates samples made using SnCl,, (SO,*-) indicates samples made using SnSO,. J. MATER. CHEM., 1994, VOL. 4 Fig. 5 Proton-decoupl6d high-resolution 29Si MAS NMR spectrum of M-Sn-X, recorded at ambient temperature, with 21 spin rate of ca.5 kHz and n/2 rf pulses. The recorded spectrum is giben as the dotted line (...) and fitted gaussians are shown by the solid lines. The difference between the observed and calculated spectrum is given as a solid line below the spectrum. effect (observable by NMR spectroscopy) upon the zeolite frameworks, which is fully consistent with the XRD results. TSeries The 27Al spectra of all T-series samples are dominated by the signal of tetrahedrally coordinated A1 (in the zeolite frame- work), but also indicate the presence of some octahedrally coordinated aluminium, i.e. a peak at 6~0typical of A]"' octahedrally coordinated by 0 atoms [cf. Al( H20)63+(as), which is the reference]. Peak positions are given in Table 4.The octahedral A1 is that acid-leached from the framework,I2 and is consequent on the heat treatment of SnSO, and SnCl,, producing SO, and HC1 (from the reaction of SnC1, with water), respectively. For a given zeolite host, the intensity due to octahedral A1 in the 27Al spectrum of the material made using SnCl, is greater than that in the spectrum of the corresponding material made using SnS0, (Fig. 6). This implies that, as is also evident in XRD studies (see earlier), the extent of framework damage is higher if SnC1, is used in the synthesis. Chemical shifts in the deconvoluted 29Si spectra of all samples in the T series (summarised in Table 5) are easily assigned to specific Q4(mAl) units.13 The positions of the lines vary slightly (<2.5 ppm) from those of the host zeolites (in this study and also reported in the literat~re),'~ and this is attri- buted to cation effe~tsl~,'~ brought about by tin ion exchange, to distortions in local geometry and to susceptibility differences.The damage to the zeolite frameworks is evident in 29Si MAS NMR spectra in the observation of (i) additional peaks in the spectra (except for T-Sn-Mor samples). (ii) a general upfield shift in the intensity distribution of the spectra (Si nuclei have, on average, fewer A1 nearest neighbours) and (iii) broadening of constituent lines. Where aluminium atoms are widely separated in the framework (e.g. in mordenite, Si :A1zz 5.8) their partial removal leaves the framework essen- tially intact, but results in observable changes in the relative intensities of lines in the 29Si spectra.Fig. 7 illustrates the greater framework disruption for the T-series products relative to those of the M series and also the greater disruptive effect of using SnCl, as a reagent. The 29Si spectrum for T-Sn-X( SO4,-) [prepared from SnSO,, Fig. 7(u)] is changed relative to that for the parent zeolite (to J. MATER. CHEM., 1994, VOL. 4 1917 * A\ 100 50 0 -50 * .. I‘ 100 50 0 -50 6 Fig. 6 Proton-decoupled high-resolution 27Al MAS NMR spectra of (a) T-Sn-Y(Cl-), (b)T-Sn-Y(SO,’-) and (c) M-Sn-Y (no octahedral Al), recorded at ambient temperature with spin rates of cu. 5 kHz and employing 71/6 rf pulses. *indicates a spinning side band.which that for M-Sn-X, Fig. 5, is a close approximation), but observable concentrations of Q3 sites [Si(OT),O- or silanol peak positions remain essentially unchanged. The 29Si spec- groups], with a chemical shift of ca. -100 ppm.17 Leaching trum for T-Sn-X(Cl-) [prepared from SnCl,, Fig. 7(b)],on can result in fragmentation of the crystallite, giving a variety the other hand, bears little resemblance to that for the of Q” (n=1, 2, 3 or 4) sites, all of which will exhibit different parent zeolite. chemical shifts. The additional lines observed are broad, In the composites formed from zeolite A (the parent richest indicating the presence of a variety of similar (but differently in aluminium) dealumination results in the production of distorted) fragment environments. Table 5 29Si MAS NMR parameters for tin zeolites Si(OA1) Si( 1 Al) Si(2AI) Si(3A1) Si(4A1) material -6 W/HZ I(%) -6 W/HZ I(%) -6 W/HZ I(%) -6 W/HZ I(%) -6 W/HZ I(%) T-Sn-A(SO,’-) 93.6 262 12 89.5 151 75 95.8 71 1 86.2 69 2 83.8 287 10 T-Sn-A(CI-) 100.5 74 2 95.4 210 20 90.8 238 74 98.3 82 3 85.2 91 2 T-Sn-X(SO,’-) 103.0 74 1 99.3 182 4 95.3 192 12 89.9 236 34 85.7 146 82.6 274 45 4 T-Sn-X( C1- ) 104.1 239 3 98.5 466 31 94.0 219 11 89.7 270 22 85.1 253 15 108.2 635 13 79.5 393 5 T-Sn-Y (SO,’-) 113.1 571 7 99.9 347 39 94.5 190 29 89.3 235 14 83.7 238 1 106.2 281 10 T-Sn-Y(Cl-) 114.1 588 12 100.9 300 35 95.3 234 24 90.1 283 9 80.8 283 1 106.9 353 19 T-Sn-Mor( T-Sn-Mor(Cl-) T-Sn-MorH ) 113.9 113.9 112.7 288 300 310 50 44 52 107.0 107.1 107.2 368 347 394 45 52 42 M-Sn-A - _ _ ~--- 89.4 100 M-Sn-X 103.4 170 3 99.4 181 8 94.5 183 20 89.5 174 36 85.1 110 33 M-Sn-Y 105.9 220 7 100.1 273 37 94.7 186 40 89.6 190 15 84.8 187 2 M-Sn-Mor 113.0 274 36 106.5 379 58 99.7 318 7 --.- J. MATER. CHEM., 1994, VOL. 4 Table 6 '19Sn Mossbauer parameters for tin leolites diso/ AEq/ FWI-lM/ relative material" mm s-lb mm s-lb mm s-l intensity M-Sn-A 0.16 0.55 1.04 1.o M-Sn-X 0.17 0.55 1.09 1.o M-Sn-Y 0.16 0.6 1 0.07 1.o M-Sn-Mor 0.14 0.5 1 1.27 1.o T-Sn-Mor( SO4,-) T-Sn-Mor( C1- ) T-Sn-MorH (Cl- ) 0.04 0.06 0.03 3.25 0.59 0.60 0.67 1.80 1.03 1.02 1.21 1.09 1.oo 1.oo 0.94 0.06 T-Sn-A(SO4,-) 0.05 0.14 0.59 1.34 1.04 0.88 0.91 0.09 T-Sn-A(C1-) 0.13 0.19 0.57 1.39 110 0.97 0.7 1 0.29 -50 1 ~ ~ -70 ~ ' 1 -90 ~ ~ " ~ ~"' -1 20 ~ ~I ' -1 30 ' ~ ~ ' "T-Sn-X(SO4'-) ~ ~ ~ ~0.05 0.11 ' ~ 0.55 1.35 ' 1 00 0 85 0.88 0.12 T-Sn-X( C1- ) 0.04 0.07 0.51 1.36 1 05 0 98 0.68 0.32 T-Sn-Y(SO,'-) 0.04 0.11 0.55 1.30 0 46 0.91 0.82 0.16 4.05' - 0 94 0.02 T-Sn-Y (C1 -) 0.06 0.09 0.53 1.33 0 99 0 98 0.85 0.15 ~ a indicates samples made using SnSO,, rCl-) indicates samples made using SnC1,.'diso values f0.03 mm s-', AEq values k0.06 mm s-', Fitted as a singlet.appearance similar to Fig. 8(a)] showed a single peak close to I zero velocity and similar to that attributed to tin(rv) ~xide.'~,'~ As in the related work of Berry et al. (on similar reactions with Laponite)," the peak was fitted as an unresolved doublet -60 -80 -1 00 -1 30 arising from a small quadrupole splitting (AE, ~0.5mm s-'). 6 These spectra would therefore be consistent with the presence of tin exclusively in finely dispersed SnO,.Fig. 7 Proton-decoupled high-resolution 29Si MAS NMR spectra of (a) T-Sn-X(SO,'-) and (b) T-Sn-X(C1-), both recorded at ambient temperature with spin rates of ca. 5 kHz and 42 rf pulses. Recorded T Series spectra are given as dotted lines (...) and fitted gaussians are shown For all the T-Sn-Mor samples [prepared from Na-mordenite by the solid lines.The differences between the observed and calculated and either tin@) chloride or tin(rr) sulfate] the spectra were spectra are given as solid lines below the spectra. as for the M series, arising from Sn exclusively in dispersed oxide. In the case of T-Sn-MorH(Cl-) [prepared from H-'I9Sn Mossbauer Spectroscopy mordenite and tin@) chloride] an additional doublet (AE,= 1.50mm s-') of low intensity (6%) was detected in the Spectral parameters resulting from computer-fitting of experi- spectrum [Fig. 8(b)]. The isomer shift (6,,=3.3 mm s-') for mental data are given in Table 6, with three typical spectra that doublet is consistent with residual tin@) species present illustrated in Fig. 8. as an impurity." The Mossbauer data for this component are different from either anhydrous (aiso=4.10, AEq=0.56 mm M Series s-l) or hydrated SnC1, (6,,,=3.68, AE,= 1.24 mm s-'), but For these samples (prepared by microwave-assisted hydrolysis closely resemble data for a frozen solution of SnCl, in of organotin chloride) the 'I9Sn Mossbauer spectra [with methanol (dis0 =3.36, AEq=1.76 mm s-'),~~suggesting that 100-h$ 96-v C .-0 .-2 92 6 2 aa -a4 1 -6 -2 2 6 -6 -2 2 velocity/mm s-' Fig.8 'I9Sn Mossbauer spectra of (a) T-Sn-Mor(SO4'-) and (b) T-Sn-MorH(C1-) and (c) T-Sn-X(C1-). Dashed lines indicate individual contributions to the observed spectra (see text), with the final fits being represented by solid lines. Spectral parameters are listed in Table 6.J. MATER. CHEM., 1994, VOL. 4 the Sn" species in T-Sn-MorH(C1- ) occupies an environment including several oxygen donors. The spectra of other T-series samples, made from tin(I1) sulfate or tin@) chloride and zeolite X, Y or A, also contained a peak at close to zero velocity. For those materials, satisfac- tory fits could not be obtained on the basis of a single unresolved doublet. The spectra were, however, satisfactorily simulated as the sum of two doublets, the first being the 'unresolved doublet' arising from Sn in dispersed oxide (AEq20.5 mm s-l) and the second (lower-intensity) doublet having a larger quadrupole splitting (AEqz 1.3 mm s-I). -1 -2 -3 0----_ 0 -1 -J ' 6.. Fig. 9 Temperature-dependent conductivities for (a) T-Sn-Mor and M-Sn-Mor samples: 0, T-Sn-Mor (Cl-); 0,T-Sn-Mor (SO:-); 0, T-Sn(H)-Mor; X, M-Sn-Mor; (b) T-Sn-A and M-Sn-A samples: '3, T-Sn-A (Cl-); 0,T-Sn-A (SO,'-); 0, M-Sn-A; (c) T-Sn-X and M-Sn-X samples: 0, T-Sn-X (Cl-); 0,T-Sn-X (SO:-); 0,M-Sn-X; and (d)T-Sn-Y and M-Sn-Y samples: 0,T-Sn-Y (Cl-); 0,T-Sn-Y 0, M-Sn-Y.Lines shown are the linear regressions and correspond to the activation energies given in Table 7. Fig. 8(c) shows the spectrum of T-Sn-X(C1-) [a sample made using tin@) chloride], along with the contributions of both doublets. In the case of T-Sn-Y(S02-) only, a third signal of very low intensity (2%) was attributable to residual tin@) impurity (6,,,=4.05 mm SKI).No attempt was made to identify this species, and the site was simulated by a single component in the fit (Table 6).For comparison, Mossbauer data for SnSO, are diso=4.00, AEq=1.OO mm s-.21 An unresolved question is the nature of the Sn environment giving rise to the second doublet in T-series samples prepared from zeolites X, Y and A. The small isomer shift (6,,, ~0.1mm sK1) is consistent with SnIV bonded to oxygen, but the quadrupole splitting AEq z 1.3 mm s-l corresponds to an environment considerably more distorted than the octahedral site in SnO,, a site which is itself not perfectly regular. One possibility is that hydrolysed Sn" ions introduced by ion exchange are associated with/bonded to the S6Rs (single six- rings) in zeolites X, Y and A (mordenite does not have S6Rs), with Sn coordinated both to the zeolite framework and to hydroxy/water groups with distinct SnIV -0 bond kngths.It must, however, also be borne in mind that the fr'imework disruption in these samples could well result in SnIV in regions of variable tin coordination geometry, such that the two- doublet fit merely mimics a variety of closely similar Sn" environments. ac Conductivity Studies Impedance spectra consisted of a high-frequency arc and a sloping rise of reactance with respect to resistance at lower frequency. Such spectra are typical of ionic conductors when blocking electrodes are used. The conductance of the sample was extracted as the intercept of the low-frequency line with the resistance (real) axis.The temperature dependences of the conductivities showed Arrhenius-like behaviour over the experimental temperature range (see Fig. 9). Empir tcal acti- vation energies E, (corresponding to linear regressior, fits) for protonic conduction are given in Table 7. E, values can only be regarded as empirical parameters, as there mill be a variation of the water content (and number of charge carriers) of the samples with temperature. All samples were, however, studied in the same environmental conditions, enabling facile comparisons. T Series The observed conductivities of the materials are in the range 3 x (at 293 K, for T-Sn-Mor made from SnSO,) to 3 x S cm-' (at 353 K, for T-Sn-A made from SnCl,), with significant enhancement with respect to the untreated Table 7 Activation energies for protonic conduction in 'tin zeolites' material EJkJ mol-' T-Sn-Mor(SO,'-) 70&4 T-Sn-Mor(C1-) 65k7 T-Sn-MorH (CI-) 72k3 T-Sn-A(SO,'-) 36f3 T-Sn-A(C1-) 54f3 T-Sn-X(SO4'-) 37k2 T-Sn-X(C1-) 14+2 T-Sn-Y(SO,'-) 14*4 T-Sn-Y (C1- ) 18f8 M-Sn-A 50+3 M-Sn-X 41&5 M-Sn-Y 12f 1 M-Sn-Mor 39k2 (C1-) indicates samples made using SnCI,, ( SO,'-) indicates samples made using SnSO,.parent zeolite^.'^^*^^-^^ Th e conductivities of the materials based on mordenite are lower than those reported by Knudsen et a!.' This is likely to be due to the lower relative humidity at which the measurements were made in this study. A lower water content in the sample will result in fewer charge carriers (protons) being available and hence a lower conductivity.Samples made using SnC1, exhibit higher conductivities than those in which SnSO, was used. These samples have more extensively damaged frameworks, but generally contain less tin. It is therefore unlikely that the enhancement in conductivity is due solely to the introduction of SnO, as had been suggested by Knudsen et al.' The increased conductivity of the T-Sn-Mor samples (with respect to the parent zeolites) can be attributed to the effects of damage to the zeolite framework. In T-Sn-MorH the damage to the zeolite is more significant and the conductivity is further enhanced. The T-Sn-Y materials exhibit conductivities of the same order of magnitude as those reported by Krogh Andersen et al.( lop2 S cm-' at 116 3C).3 M Series The conductivities of these samples are lower than those of the related T-series samples. The activation energies are also lower than those of the corresponding T-series samples, and Mossbauer spectra of these samples showed that tin is present only in SnO, (see earlier). The lower conductivities for the M-series samples are consequent on very little disruption of the zeolite frameworks and also on low tin contents. Conclusions All of the T-series samples exhibit degradation of the zeolite framework. This is evident from both XRD and MAS NMR studies. The extent of the damage depends upon the tin salt used in synthesis and is higher for those zeolites with higher A1 contents.More damage was incurred by a given zeolite when SnC1,-2H2O was used in the synthesis than when SnS0, was used. In their papers on 'tin mordenites' Knudsen et a[.'%, did not report any such degradation of the host zeolite. The synthesis of the M-series samples caused no apparent damage to the zeolite frameworks. This is expected since the amount of HCl generated during the hydrolysis of triphenyltin chloride will be small compared with that evolved on hydroly- sis of SnCl, (present when SnCl, is used in thermal syntheses). Evidence from Mossbauer spectroscopy indicates clearly that these materials are composites. All spectra contain a dominant contribution from a central peak (an unresolved doublet) arising from dispersed SnO,.In samples prepared thermally from zeolites X, Y and A, a second doublet attribu- table to a second, unidentified, SnIV site is observed. Comparison of the conductivities of corresponding mate- rials from the two series shows there is an increase in the observed conductivity when the zeolite framework has been disrupted. We thank Dr. D. J. Jones of 1'Universite Montpellier 2 for useful discussions on the fitting of Mossbauer spectra, and J. MATER. CHEM., 1994, VOL. 4 G. Edwards for assistance with synthetic aspects. We thank Laporte Inorganics for provision of the zeolite samples. We thank the SERC National Solid State NMR service (Durham) for recording NMR spectra and for subsequent deconvol- utions. We thank SERC for a studentship for G.B.H.We thank NATO for a travel grant enabling joint studies at Exeter and Montpellier. The authors thank the referees for constructive criticism of this paper. References 1 N. Knudsen, E. Krogh Andersen. 1. G. Krogh Andersen and E. Skou, Solid State Ionics, 1988,28-30, 627. 2 N. Knudsen, E. Krogh Andersen. 1. G. Krogh Andersen and E. Skou, Solid State Ionics, 1989,35, 51. 3 E. Krogh Andersen, I. G. Krogh Andersen, N Knudsen and E. Skou, Solid State Ionics, 1991,46, 89. 4 A. Ono, J. Mater. Sci., 1984, 19, 2691. 5 L. Glasser, Chem. Rev., 1975,7521. 6 W. A. England, M. G. Cross, A. Hamnett, P. J Wiseman and J. B. Goodenough, Solid State Ionics, 1980,1, 231 7 R. C. Ashcroft, S. P. Bond, M. S. Beevers, M. A M. Lawrence, A. Gelder, W.R. McWhinnie and F. J. Berry, Polyhedron, 1992. 11, 1001. 8 R. von Ballmoos, Collection of' Simulated XRD Powder Patterns for Zeolites, Butterworth, Guildford, 1984. 9 K. Ruenbauer and T. Birchall, Hyperfine Interactions, 1979,7, 125. 10 B. A. Boukamp, Solid State Ionics, 1986,18 & 19. 136; 1986,20, 31. 11 G. Engelardt and D. Michel, in High Resolution Solid State NMR of Silicates and Zeolites, John Wiley, New York, 1987, p. 150. 12 C. A. Fyfe, G. C. Gobbi, W. J. Murphy, R. S. Ozubuko and D. A. Slack, J. Am. Chem. Soc., 1982,86,3061. 13 M. Magi, E. Lippmaa, A. Samoson, M. Tarkamand and G. Engelhardt, J. Phys. Chem., 1984,88, 1518. 14 G. R. Hays, W. V. van Erp, N. C. M. Alma, P. A. Couperus, R. Huis and A. E. Wilson, Zeolites, 1984,4, 377. 15 G. Engelardt and D. Michel, in High Resolution Solid State NMR of Silicates and Zeolites, John Wiley, New York, 1987, pp. 256-257. 16 M. Magi, E. Lippmaa, A. Samoson, G. Engelhardt and A- R. Grimmer, J. Phys. Chem., 1984, SS, 1518. 17 J. M. Thomas, C. A. Fyfe, S. Ramdas, J. Klinowski and G. C. Gobbi, J. Am. Chem. SOC., 1982,86,3061. 18 N. N. Greenwood and T. C. Gibb, in Miisshauer Spectroscopji, Chapman and Hall, London, 1972, p. 374. 19 F. J. Berry and A. G. Maddock, Radiochim. Acta. 1977,24, 32. 20 F. J. Berry, R. C. Ashcroft, M. S. Beevers, S. P. Bond, A. Gelder and W. R. McWhinnie, Hyperfine Interactions, 1991,68,261. 21 J. N. R. Ruddick, Ret.. Silicon, Germanium, Tin and Lead Compounds, 1976,2,115. 22 D. N. Stamires, J. Chem. Phys., 1962,36, 3174. 23 M. Lal, C. M. Johnson and A. T. Howe, Solid Slate Ionics, 1981, 5,451. 24 E. Krogh Andersen, I. G. Krogh Andersen, K. E. Simonsen and E. Skou, in Solid State Protonic Condirctors II, ed. J. B. Goodenough, J. Jensen and M. Kleitz, Odense University Press, 1983, pp. 155-160. 25 S. Y. de-Andersen, E. Skou, I. G. Krogh Andersen and E. Krogh Andersen, in Solid State Protonic Conductors 111, ed. J. B. Goodenough, J. Jensen and A. Potier, Odense University Press, 1985, pp. 247-257. Paper 4/02972K; Received I 8th Muy, 1994