Faraday Discuss. Chem. SOC., 1990, 89, 119-136 Structural Studies of High-area Zeolitic Adsorbents and Catalysts by a Combination of High-resolution X-Ray Powder Diffraction and X-Ray Absorption Spectroscopy Eric Dooryheet and G. Neville Greaves S. E. R. C. Daresbury Laboratory, Warrington WA4 4AD Andrew T. Steel, Rodney P. Townsend and Stuart W. Carr Unilever Research, Port Sunlight Laboratory, Quarry Road East, Bebbington, Merseyside L63 3JW John M. Thomas* and C. Richard A. Catlow Davy Faraday Research Laboratory, The Royal Institution, 21 Albermarle Street, London W l X 4BS We have characterized at high temperature a model uniform heterogeneous catalyst for the oligomerization of hydrocarbons (a nickel-exchanged zeolite of initial composition Na59A159Si,330384 - xH,O treated with an aqueous solution of NiC1, so as to yield a homogeneous distribution of Ni, with Si/ Ni = 7) by recording the extended X-ray absorption fine structure (EXAFS) abov? the Ni edge and also its high-resolution diffraction pattern (at A = 1.5486 A).We have obtained unique insights into the microenviron- ment of the Ni'+ ions in the as-prepared and the dehydrated as well as the reduced state of the catalyst; in particular, values of atomic coordinates, site-occupancy and bond lengths, have been obtained. High-resolution X-ray diffraction, using a specially constructed environmental cell, has enabled the precise location of sorbed Xe atoms in a sodium-exchanged zeolite X adsorbent (1.5 atoms of Xe per unit cell) under a pressure of 1.75 bar at room temperature to be determined from a Rietveld refinement of the powder diffraction pattern. Direct proof of the role of the strongly polarizing Na+ ion located at the Sn site in firmly binding the Xe to the inner wall of the supercage is obtained. X-ray absorption near-edge structures (XANES) as well as parallel EXAFS studies above the Al, Si and Ga edges on a variety of other zeolites are also reported. Combined (Cia edge) XANES and EXAFS studies show that gallosilicate networks akin to those present in crystalline (gallo) zeolite X and Y are already formed in the precursor gel and are also present in the aqueous solution in contact with the colloidal-amorphous material, which together constitute the gel. We also report one of the first applications of A1 EXAFS for the semi-quantitative 3;Jdy of the dealumination of faujasitic zeolites.Introduction Highly microporous adsorbents and catalysts, typified by zeolites, possess the convenient attribute that all, or nearly all, of their bulk atoms are, at one and the same time, surface atoms, accessible to reactant molecules of diameter up to ca. 8 A. Their surface area, which may frequently be in excess of 600m2gp', is thus internal. When, therefore, $ Also at: Davy Faraday Research Laboratory. 119120 High-area Zeolitic Adsorbents and Catalysts Fig. 1. Structure of zeolite-Y, with extraframework cation sites shown. gaseous species, enter the interior of such solids and are either adsorbed or subsequently converted catalytically into products, the fate of those bound species as well as the (internal) surface characteristics of the microporous host, may be probed by the majority of the techniques available to the solid-state chemist and physicist.’-4 Bulk powder diffraction (XRD) s t ~ d i e s , ” ~ along with bulk absorption fine structure (XAFS) studies, both of which are optimally carried out using synchrotron radiation sources, therefore yield, in combination, unique insights into the surface and catalytic properties of such high-area solids.Study A concentrates on the characterization of the microenvironment and migration of nickel ions at the internal surfaces of a zeolite Y catalyst5-’ in which some of the Na’ ions have been exchanged by Ni’+ ions. The resulting NaNi-Y catalyst has been probed both by XRD and XAFS prior to and after dehydration and also after reduction in hydrogen.Study B aims to determine the location of sorbed xenon inside a Na+ ion-exchanged zeolite X adsorbent. Study C is a quantitative study of the dealumination of a zeolite by comparing the changes in XAFS of aluminium and silicon edges. Study D is a detailed analysis of the microenvironment of a typical, tetrahedrally bonded atom (in this case gallium) in the course of formation of a (gallo) zeolite from its amorphous precursor gel to the final crystalline state. We have previously reported briefly on preliminary results obtained in study A.” Study B is an extension of earlier using laboratory X-ray sources, aimed at locating within zeolite cages, important adsorbents such as xenon 12-14 (which simulates the behaviour of methane) and methyl chloride.Study C, although complementary to solid-state NMR studies of dealumination, ‘’,16 can yield quantitative values for bond distances such as A1-0 in the dealuminated zeolite. Study D is, to our knowledge, the first report of an analysis which charts the local environment of a tetrahedrally bonded element from the pre-crystalline state to the crystalline solid. We have briefly reported” these results previously. The structure of zeolite-Y, with which we are mainly concerned in this paper, is shown in fig. 1 , which displays the customary polyhedral framework structure of the In this communication we focus upon four distinct categories of study.E. Dooryhee et al. ( a ) 121 1 6 1 2 0 8 0 4 0 2 4 6 8 1 0 Fig. 2. Background-subtracted Ni-edge EXAFS spectrum and ( b ) Fourier transform for hydrated Ni-Y.Dotted lines show calculated spectrum using structural parameters given in table 1. zeolite, each full line representing a T-0-T bond (T= Si or Al) and each vertex being a T site. S, sites are situated at the centre of the double-six rings (D6R), i.e the hexagonal prisms connecting two sodalite cages, and the S1, and Sill sites are at the walls of the supercage. S,, and SII, sites are mirror images, inside the sodalite cages, of the S, and SII sites, respectively. Experimental All our X-ray absorption and diffraction studies were carried out using the synchrotron radiation source (SRS) at the S.E.R.C. Daresbury Laboratory. For study A, three different samples were used. The first was a hydrated nickel-ex- changed zeolite-Y catalyst containing some un-exchanged sodium ( Na,7Ni21A159Si ,330384 exchanged with 0.005 mol deg-3 NiC12 solution, henceforth labelled NaNi-Y); the second was the same material dehydrated in vacuo at 300 "C for 14 h; and the third was the product of reducing the dehydrated sample in a stream of hydrogen at 450°C for 32 h.The EXAFS data for the Ni K-edge (8333 eVt) were collected in transmission using station 7.1 at the SRS Daresbury Laboratory. As noted, the measurements on the dehydrated sample were collected at room temperature in situ in a Lytle cell in which the dehydration had been effected. The hydrated and reduced samples may be safely exposed to the atmosphere and the data were obtained under normal ambient conditions.The spectra were background ~ubtracted'~ and analysed using the standard Daresbury EXAFS software package ( E X C U R V S ~ ~ ~ ) . Least-square fitting of the data was undertaken using phase shifts that had been refined by employing NiO and Ni metal as model122 High-area Zeolitic Adsorbents and Catalysts 0 8 0 4 0 0 4 Fig. 3. ( a ) Background-subtracted Ni-edge EXAFS spectrum and ( b ) Fourier transform for dehydrated Ni-Y. compounds; these yield information on backscattering from respectively surrounding 0 and Ni atoms, the latter being relevant to the reduced compounds. Calculated phase shifts for Si backscattering proved to be adequate. Background subtracted data and Fourier transforms for the three samples are shown in fig 2-4. The diffraction studies used the high-resolution Debye-Scherrer diffractometer of Station 9.1 on the hard X-ray (Wiggler) line at the SRS, Daresbury.Data were collected on the dehydrated sample, in situ in the GTP furnace. The pattern which was measured using a wavelength of 1.5486 A between angles 26, of 6 and 70°, is shown in fig. 5. The data were refined using the Rietveld profile refinement technique, employing the PDPL programme suite developed by Fitch and Murray. For study B, prior runs were carried out, using a conventional volumetric apparatus for studies of gas uptaken by porous solids, on a zeolite-X sample (Si/AI = 1.15) prepared at the Port Sunlight Laboratories of Unilever plc. An environmental chamber, capable of accommodating a flat plate on which the sample was mounted, permitted us to record high-resolution (Debye-Scherrer) diffraction data, again using Station 9.1 at the SRS Daresbury.Xenon could be admitted to the environmental cell at a variety of pressures up to 1.75 bar.? For study C, three samples of ammonium-exchanged zeolite-Y (prepared at the Unilever Port Sunlight Laboratories) were used, as well as a-A1203 and y-A1203, as references. The first of these, labelled 252, was a standard zeolite (Si/Al ratio of 2.52). The second and third, C2 (Si/Al=5.2) and D2 (Si/Al=5.6), respectively, had both been subjected to steam dealumination. XAFS data were collected at the SRS using the soft X-ray EXAFS (SOXAFS) Station 3.4. A double crystal quartz monochromator was employed, enabling the A1 K-edge (1559 eV) to be conveniently scanned.Si K-edge (1839eV) data were obtained in a separate set of experiments in which an InSbE. Dooryhee et al. 123 0 4 Fig. 4. ( a ) Background-subtracted Ni-edge EXAFS spectrum and ( b ) Fourier transform for reduced Ni-Y. 16000 Fig. 5. Diffraction pattern for dehydrated Ni-zeolite-Y (300 "C). monochromator was used. The use of soft X-rays necessitates that the samples are under a high vacuum. The EXAFS signal is measured straightforwardly by monitoring the total photo-electron yield. To prevent charging of the samples, they were mixed with graphite and supported on a copper plate. For study D, the following gallium-containing zeolites were prepared (as described in the respective cited reference): Na(GaY)," Na(GaX)," gallosodalite." Results on studies of zeolite g a l l ~ - o m e g a , ~ ~ - ~ ~ and K(GaL) 2224 will be described elsewhere.The synthesis of all these zeolites uses a two-stage process, i.e. the formation of an amorphous alkaline slurry or gel (on mixing) of gallate solution and a source of silica, followed by prolonged hydrothermal crystallization.124 High-area Zeolitic Adsorbents and Catalysts The ammonium form of the gallo-Y zeolite, NH,(GaY) was prepared by ion exchange (10 times) of Na(GaY) with 0.1 mol dm-' NH,NO,. Gallium-exchanged NH,(AIY) and NH,(GaY) were prepared from 0.1 mol dm ' aqueous solution of gallium(II1) sulphate (adjusted to a pH of 3 by addition of conc. ammonia). l o g of the zeolite NH,(AIY) or NH,(GaY) was slurried with 100 cm' of this solution; these mixtures were set aside at 25 "C for 16 h.The zeolites were recovered by filtration, washed with distilled water and dried in air. NH,(GaY) was treated with 0.1 mol dm' NH4N0, acidified to a pH of 3 with nitric acid. This sample was set aside at 25 "C for 24 h and the zeolite collected by filtration. The precursor gels were prepared as for the crystalline zeolites, except that they were separated from solution at an early stage either by filtration or by centrifuging at 4000r.p.m. for ca. 1 h. EXAFS spectra at the Ga K edge (10367 eV) were recorded on Station 7.1 of the Daresbury SRS. The spectra were collected in transmission mode. Solid samples were finely ground and supported between strips of adhesive tape. Solutions and gels were housed within heat-sealed polythene containers. Normalization and background sub- traction were undertaken and the data were analysed using the EXCURVE programme."' As well as using gallosodalite as a reference, we also used two aqueous solutions of gallium to pin point the differences between 4- and 6-coordinated gallium.(Both these solutions exhibited (see below) a single EXAFS oscillation with no 'longer range' interactions.) Sodium gallate solutions (made up of 10,8 and 82% respectively of NaOH, Ga203 and water) showed a tetrahedral Ga-0 distance of 1.80 a; gallium sulphate [made up from 0.1 mol dm-3 Ga,(SO,), buffered with 0.1 mol dm-3 NH,NO, to pH 3) had a Ga-0 distance of 1.93 A, which is characteristic of 6-coordination. Results and Discussion Study A: Microenvironment and Migration of Nickel Ions at the Internal Surfaces of NaNi-Y Catalyst From the EXAFS studies several conclusions may be drawn.The EXAFS spectrum, fig. 2, of the as-prepared NaNi-Y sample clearly indicates the presence of a hydrated Ni2+ cation. As is evident from the Fourier transform of the data, there is one shell of oxygen atoms surrounding the central Ni; the least-squares fit to the data summarised in table 1 shows a coordination number of 6 and an Ni--.O bond length of 2.06 A, identical to that in the hydrated Ni cation in Ni( N03)2 - 6H20. We should note, however, that the oxygen backscattering peak is not entirely symmetrical which may indicate some partial hydrolysis of the hydrated cation. In addition we observe smaller peaks at larger distances due to backscattering from more distant shells; these almost certainly arise from framework atoms.We conclude, therefore, that Ni is present as a hydrated, and possibly partially hydrolysed cation, occupying sites in the supercage where it is loosely bonded to the framework. Fig. 3 reveals a much more complex structure after dehydration. The main feature is a peak corresponding to a Ni...O bond at 2.02 A, which is slightly less than in the hydrated systems. In addition, there is a peak corresponding to a short bond length of 1.87 A. Clearly, Ni is present in a variety of environments in this system, as shown in table 1 . Reduction of the material gives rise to additional features, in particular the peak at 2.50 A (see fig. 4) which can clearly be attributed to backscattering by surrounding Ni atoms, hence indicating the presence of metallic Ni particles on reduction.The Ni.-.Ni spacings at 4.29 and 5.09 A lie close to the second and third shells in the f.c.c. structure. Note also the disappearance of the short 1.87 8, Ni.S-0 bond length, suggesting that this site is a precursor of the metal sites in the reduced material.E. Dooryhee et al. 125 Table 1. Bond lengths between nickel and: ( i ) oxygen (ii) tetrahedral (Si or Al) atoms ( i i i ) other Ni atoms as revealed by EXAFS studies Ni-T Ni-Ni Ni-0 sample Nl R, 2af N, R, 2ai N3 R3 2 4 hydrated 6 Ni-Y dehydrated 1 Ni-Y 3 2 2 reduced 5 Ni-Y - 2.06 0.010 4 1.87 0.005 4 2.02 0.009 2 2.49 0.037 - 2.68 0.017 - 2.07 0.015 - 2 - - - - - 3.30 0.014 - - - - - 3.15 0.023 - 3.69 0.019 - - - - - - - - - - 1 2.50 0.013 3.41 0.020 1 4.20 0.016 - - 1 5.09 0.006 T = framework (Al-Si), R is the bond length, N is the coordination number and 2 a ' is the Debye- Waller factor.Table 2. Atomic positions, temperature factors and site occupancies of hydrated Ni-Y zeolite at 300 "C atom X Y Z B A' N comments Si -0.05302 0.12297 0.03562 1.49 137.0 Ni(1)-Si = 3.37 At -0.05302 0.12297 0.03562 1.49 55.0 0 1 -0.10930 0.10930 0.00000 2.01 96.0 Ni( 1)-01 = 3.77 0 2 -0.00195 -0.00195 0.14766 0.65 86.0 Ni( 1)-02 = 3.60 0 3 0.18785 0.18785 -0.02923 1.49 96.0 Ni( 1)-03 = 2.25 0 4 0.16719 0.167 198 0.3 13 15 2.75 96.0 Ni( 1)-04 = 4.80 Nil 0.23378 0.23378 0.23378 3.3 19.6 Sll. site Ni (1) 0 0 0 1.7 13.9 SI site Ni (2) 0.07058 0.07058 0.07058 5.8 1.4 S18 site Ni (3) 0.10570 0.10570 0.10570 5.8 2.4 SI.site Ni (4) 0.1973 1 0.1973 1 0.19731 5.8 2.1 SII site o w 1 0.08 155 0.03539 0.12309 5.8 35.9 sodalite cage o w 2 0.40970 - 0.283 15 -0.21 161 5.8 1 1.6 supercage Ow 3 0.173 -0.197 0.048 5.8 10.9 supercage Space group Fd3m; lattice parameters: 24.3661, 24.3661, 24.3661, 90, 90, 90; R factors: R , = 4.41, R,, = 12.14, R, = 7.77; note refinement used a variable Lorentzian peak shape. The XRD study of the dehydrated material provided important additional structural information. The diffraction pattern (fig. 5) was successfully refined using the Rietveld technique (yielding a weighted profile 'R factor' of 12.14% as shown in table 2). Evidently, as the EXAFS also show, there is a variety of sites which the Ni2+ ions occupy. The conclusions drawn from the results are that on dehydration there is migration of the Ni'+ ions from their original position in the supercage into the hexagonal prism, S , sites.This is in line with the earlier work of Olson.'' However, our refinements also show that Nil+ ions occupy two types of S , . site (see fig. 6 ) . The S , , sites are close to the so-called six-rings but project into the sodalite cage, the two S , , sites being distinguished from one another by their different distances along the (111) axes from the centre of the six-ring. Our XRD data also show that heating at 300 "C for 14 h does126 High-area Zeolitic Adsorbents and Catalysts Fig. 6. Ni and Na sites in zeolite-\(. Number convention is as in table 2. Fig. 7. Configuration of the proposed partial hydrated Ni S , , site.not fully remove the water from the hydration shell of all the Ni2+ ions, with Nil+ ions at S,, sites persisting. Comparing tables 1 and 2, there is an apparent contradiction between our EXAFS and XRD results, especially in regard to Ni-0 bond lengths for the dehydrate! samples, which is, at first, puzzling. The N i - 0 bond length for the S , site of 2.25 A , derived from the diffraction data, is significantly longer than the value of 2.02 A obtained from analysis of the EXAFS results. This apparent discrepancy may be rationalised by noting that only about three-quarters of the S , sites are occupied. In an unyccupied S, site the distance of the surrounding oxygen atoms from the centre is 2.6 A , where as in the occupied sites the 0 atoms relax inwards.EXAFS will measure the relaxed Ni.q.0 bond length, whereas diffraction will see an average 0 position. I t is reasonable to postulate, therefore, that the diffraction bond length is an aveEage of the relaxed and unrelaxed bond lengths. If we use the EXAFS value of 2.02 A for the former and the value of 2.6 A referred to above for the latter togtther with an occupation number of 0.75 for the S , site we obtain an estimate of 2.17 A for the average Ni...O bond length, which is in acceptable agreement with the diffraction value of 2.25 A. An alternative explanation that may be advanced postulates asymmetric relaxations of the Ni” ions away from the centres of the hexagonal prism. However, we would expect such relaxations to lead to enhanced temperature factors which are not observed.The explanation in terms of relaxation of the surrounding 0 atoms seems therefore to be the more plausible.E. Dooryhee et al. i I 8 0 127 1 i 4 ’ I I I I I 1 I 6 1 0 1 4 18 22 26 30 34 201’ Fig. 8. Diffraction pattern of unloaded zeolite-)< (outgassed at 31 “C). There is good evidence that the non-S, Ni ions are all partially hydrated species. Of the two S,, sites, the first is bonded to the framework oxygens but with additional coordinating water molecules as illustated in fig. 7; the second, which is further towards the centre of the sodalite cage, is largely hydrated. The positions of the oxygen atoms (of the water molecules) located in the sodalite cage are consistent with these models for hydration of the S,, Ni2+.The Ni in the SI, sites is also directly bonded to the framework oxygen, and it is also likely that these cations are partially hydrated. Study B: Location of Xenon in the Cavities of Na+-exchanged Zeolite-X Previous studies of xenon sorbed within the intrazeolite cavities of zeolite-p were carried out by Gameson et alx using laboratory X-ray sources. These studies, along with comparable ones9 involving sorption of methyl chloride by zeolite-p, were carried out at low temperatures ( 10-200 K). Although valuable structural data pertaining to the location and dynamics of physically adsorbed Xe were obtained, complementing data derived from solid-state NMR,”.“ it is clear that the flux and monochromaticity of synchrotron sources yield quantitatively improved data at higher temperatures. This is especially important in obtaining basic information about guest species sorbed at ambient temperatures and around atmospheric pressures, the conditions most frequently used in gas separations.We also note that high-quality structural data on guest species sorbed within zeolites may be derived from neutron powder diffraction studies, as exemplified by the work of Fitch et a1.26 on the benzene-zeolite-Y system, and the work of Wright and coworkers” on Xe in zeolite-p, and Wright and coworkers’x on pyridine in zeolite-L, Williams” and Newsam3(’ on benzene in zeolite-L. From our raw data (see fig. 8 for the ‘unloaded’ and fig. 9 for the ‘loaded’ zeolite-)< samples) it is immediately apparent that the diffraction patterns of the two ‘loaded’ samples are significantly different from one another and from that of the ‘unloaded’ sample. Using the Rietveld-profile-refinement procedure, we find that for both loaded samples there is one predominant site for the sorbed Xe as shown in fig. 10.The latter is situated close to Naf ions in the SI1 site, and it is clear that the highly polarizable Xe atom is bound to the polarizing Nat cation. Study C: XAFS of Dealuminated Zeolites Fig. 1 l(a)-(c) show the Al edge data collected for the three zeolite-\( samples; corre- sponding Si edge data are shown in fig. 12(a)-(c). In addition, data were obtained for128 High-area Zeolitic Adsorbents and Catalysts - ( a ) 20000 15000 i ' 3 10000 4500L 3000 - 1 5 0 0 . I 1 1 1 I 1 I 1 2 14 16 1 8 20 6 8 1 0 2 0 / O Fig.9. Diffraction pattern of the Xe loaded zeolite-)< (at 31 OC): ( a ) data for 1 atm; ( b ) data for 1.75 atm. Xe atom Na+ ions Na'(4) ion Fig. 10. Site occupied by Xe in zeolite-><.E. Dooryhee et al. 129 1 5 5 0 ‘ 5 7 0 *, 590 :61C 1630 1550 energy/eV 1 0 0 8 0 2 0 0 1590 :510 :530 : 550 : 550 1570 energy/eV 0 0 I I , I ‘550 ‘ 5 7 3 : 59c ‘ 5 ’ 0 ‘530 ‘ 6 5 3 energy/eV Fig. 1 1 . A1 XANES for three ammonium-exchanged zeolite-\( samples. ( a ) sample 252 (before dealumination) ( h ) sample ‘C2’ (after dealumination) ( c ) sample ‘D2’ (after dealumination).130 1 0 0 8 0 6 E! =r 0 4 0 2 0 0 High-area Zeolitic Adsorbents and Catalysts . ( a ) I .- i I I - I I 1 I I I 0 6 O C 0 3 : 8 3 5 l o r ( c ) ‘885 1 3 9 5 ‘ 8 4 5 1 8 5 5 1 8 6 5 : 8 7 5 energy/eV Fig.12. Si XANES of zeolite-Y samples ( a ) before (sample 252) and ( h ) and ( c ) after dealumina- tion. ( b ) sample C2, ( c ) sample D2.E. Dooryhee et al. 131 Table 3. First coordination shell parameter obtained from analysis of EXAFS data for A1 edge compound A l m a -0 bond length/ A coordination number Debye- Waller factor/A2 252 D2 c 2 1.7 (6) 1.8 (8) 1.8 ( 3 ) 5 7 6 0.02 ( I ) 0.03 (8) 0.01 (3) several Al- and Si-containing model compounds, including cu-A1203. There are clear qualitative differences between the A1 spectra of the dealuminated samples in fig. 11 ( b ) and (c) compared with the spectrum in fig. l l ( a ) . In contrast no important change is observed in the Si edge data in fig. 12(a)-(c). These observations confirm the known fact that there are substantial structural changes on dealumination and that aluminium jettisoned from the framework ends up mainly in the intrazeolitic cavities.Quantitative analysis of the data, showed that, in all samples, there was little well defined structure beyond the first shell. The results for the first shell for the A1 data are, however, of considerable interest and are show! in table 3. It is clear that there is a marked change in the AI...O bond length (ca. 0.1 A); moreover, this is accompanied by an increase in the analysed coordination number of the Al. Neither effect is observed in the silicon environment. The results on the bond lengths which indicate a change in coordination are of greater quantitative significance, as it is known that the extraction of precise coordination numbers from EXAFS studies is difficult owing to the high correlation between the coordination number and the Debye- Waller factor.The change in the Al-0 bond lengths indicates a change in the coordination of the Al. The values for the C2 and De samples are towards the range of those observed for octahedrally coordinated AI(AI2O3 , Y = 1.91). The results therefore provide strong corroborating evidence that steam dealumination results in the deposition of some octahedrally coordinated A1 (probably in the form of a hydroxy species) in the channels of the zeolite. The results also show the general value of SOXAFS studies of aluminosilicate systems. EXAFS studies of light elements such as A1 and Si have not been extensively pursued.Our results present some of the first quantitative studies of the EXAFS above the A1 edge. Study D: Microenvironment of Ga in Proceeding from the Precursor, Dispersed State to the Crystal line Gall-zeolite The principal aim here was to ascertain the resemblance between the local environment surrounding G a atoms in the precursor gels and the gallo-framework of the final crystalling zeolite. Fig. 13-16 show a representative selection of XAFS data and the respective Fourier transforms for the precursor gels and crystalline gallo-zeolites-)< and -Y. The results are also summarized in table 4. Not unexpectedly, EXAFS spectra for certain pairs of solids, e.g. Na(GaY) and Na(GaX) are almost identical, the domin?ting influence of the first-shell Ga-0 backscattering, at the tetrahedral distance of 1.8 A, being apparent. The refined outer-$hell distances (Ga-Si) of these two crystalline gallozeolites fall within 3.10 * 0.05 A, the variation being considerably less than expected from the semi-empirical calculations based on the correlations between NMR "Si chemical shifts and T-0-T angles.The XANES spectra of a number of crystalline gallozeolites (see fig. 15) show some distinct differences, reflecting the variety of the local stereochemistry rather than the coordination number of the gallium atom in the framework. in general we find that the EXAFS spectrum of the G a at the gel stage is very similar to that of the Ga in the crystalline zeolite, indicating that the gallo-silicate networks are132 High-area Zeolitic Adsorbents and Catalysts 4 6 a 1 0 1 2 k / A I I 1 I I 1 3 5 9 R I A Fig.13. Ga-edge EXAFS data and ( b ) corresponding Fourier transform for crystalline Na(GaX) zeolite. comparable in the respective states. We also find that the EXAFS fingerprint of the fourfold coordinated Ga in the separated aqueous phase and the separated, semi- colloidal, suspended solid phase of the zeolite precursor gel is almost identical. There are, however, some discernible differences in the XANES fingerprints for the correspond- ing samples. Evidently, the local environment of the Ga changes slightly between the supernatant liquid and the dried gel; but, by and large, similar local environments exist in the gallosilicate networks present in solution and in the amorphous solid phase. Upon crystallization, however, the Ga XANES and EXAFS show small but distinct differences from those of the dispersed aqueous or amorphous suspended phases. More details of similar studies on other gallosilicate zeolite precursors and crystalline phases, along with quantitative analyses of the various associated XAFS spectra, will be pub- lished elsewhere.3'E.Dooryhee et al. 133 0 4 - 0 Y, * - 4 - 8 . 1 2 2 A c.l .- E l c) C .- 0 I I I I I 4 6 a 1 0 1 2 k/ A I I I I 1 1 3 5 7 9 R I A Fig. 14. ( a ) Ga-edge EXAFS data and ( b ) corresponding Fourier transform for crystalline Na(GaY) and its precursor gel (broken line). Na(Ga X) NaNMe(Ga-Omega) 0 Gallo sodalite a 1 - 0 Na(Ga Y) I 1 I I I I 0 20 4 0 6 0 80 100 1 ----- 4 0 2 0 energy / e V Fig. 15. Ga-edge XANES data for a range of crystalline gallium-containing zeolite K(GaL), Na NMe,(G-IR) and gallosodalite shown for comparison.134 High-area Zeolitic Adsorbents and Catalysts 4 6 8 1 0 1 2 k l A I I I I I 1 3 5 7 9 R I A Fig.16. ( a ) Ga-edge EXAFS data and ( b ) corresponding Fourier transform for the supernatant liquid (broken line) and the dried gel of zeolite-Y. Conclusions Study A In an as-prepared nickel-exchange zeolite--Y catalyst, hydrated Ni'+ ions are shown (by EXAFS) to be situated within the supercages either as Ni(H,O)i+ species or as solvated ions attached to the walls of the supercage via the oxygen atoms of the framework. On heating to 300°C, water is progressively removed from the Ni2+ ions, which migrate out of the supercages into the double-six rings (S,-sites).There is an apparent discrepancy between the Ni-0 bond distances obtained from EXAFS and XRD. But this is resolved in terms of the partial occupancy of the SI sites. Some 30-40% of the Ni'+ ions in the dehydrated state are situated, still partially hydrated, in sodalite cages where they are accessible to attack by hydrogen, thereby yielding small crystallites of metallic nickel.E. Dooryhee et al. Table 4. EXAFS-derived gallium-oxygen distances 135 type number radius/A U 2 / P GaA Gallate (as) Na[GaY] Na[GaX] whole gel dried gel supernatant Ga-exchanged [ Al YI Ga-exchanged [Gay1 0 0 0 Si Na Si/ Na 0 Si Na Si/ Na 0 Si Na Si/ Na 0 Si Na Si/ Na 0 Si Na Si/ Na 0 0 Si Na Si/ Na 0 0 Si 6 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 2 4 4 4 4 2 4 1.92 1.80 1.78 3.14 3.40 4.34 1.78 3.14 3 -40 4.32 1.79 3.1 1 3.41 4.36 1.79 3.1 1 3.40 4.35 1.80 3.1 1 3.41 4.37 1 .80 1.99 3.18 3.41 4.3 1 1.83 1.98 3.27 (0.009) (0.005) 0.006 0.017 0.018 0.023 0.005 0.017 0.017 0.024 0.006 0.025 0.020 0.03 1 0.006 0.023 0.020 0.029 0.006 0.028 0.02 1 0.029 0.012 0.020 0.027 0.016 0.03 1 0.015 0.016 0.023 Study B At room temperatute and under an equilibrium pressure of 1-1.75 bar, atoms of xenon are firmly bound in the intrazeolitic supercage of a Na'-exchanged zeolite-)<.These atoms of Xe, of which there are 1.5 per cent cell on average, are closely associated with the extraframework Na+ cations situated at the S,, site. Study C We report one of the first applications of A1 EXAFS, and in particular its use in elucidating the products of dealumination of a faujasitic zeolite.The average Al-0 bond length of a dealuminated zeolite has been determined demonstrating the creation of octahedral sites.136 High-area Zeolitic Adsorbents and Catalysts Study D Combined XANES and EXAFS beyond the Ga edge of a number of precursor aqueous colloidal and crystalline gallozeolites have been carried out. It is demonstrated that, both in the aqueous solution and amorphous gel, gallosilicate networks very similar to those that exist in the final crystalline gallozeolite are already pre-formed; and there is no evidence of islands or rafts of gallo-containing species. We are grateful for the support of S.E.R.C. and unilever plc and for the cooperation of many of our colleagues, notably Carol Williams, Andrew Fitch, Peter Maddox, John Couves, Richard Jones, Robert Cernik, Steven Pickett, David h, Sheehy. References 1 J.M. Thomas, Proc. 8th Int. Symp. Catal., Berlin, July, 1984, (Verlag Chemie 2 K. Klier, Langmuir, 1988, 4, 13. 3 C. R. A. Catlow and J. M. Thomas, Prog. Inorg. Chem., 1987, 35, 1. 4 J. M. Thomas, Angew. Chemie Intl. Edn. Eng., 1988, 27, 1673. 5 P. J. Maddox. J. Stachurski and J. M. Thomas, Catal. Lett., 1988, I , 191. adill and Michael 1 vol 1. 6 T. Rayment, J. M. Thomas and C. Williams, J. Chem. Soc. Faraday Trans. 1, 1988, 84, 2915. 7 J. W. Couves, R. H. Jones, B. J. Smith and J. M. Thomas, Adu. Materials, 1990, 2, 181. 8 I . Gameson, T. Rayment, J. M. Thomas and P. A. Wright, Chem. Phys. Lett., 1986, 123, 145. 9 I . Gameson, J. M.Thomas and P. A. Wright, J. Phys. Chem., 1988, 92, 988. 10 C. R. A. Catlow, J. W. Couves, E. Dooryhee, G. N. Greaves, P. J. Maddox, A. T. Steel, J. M. Thomas and R. P. Townsend, in preparation. 11 C . R. A. Catlow, E. Dooryhee, G . N. Greaves, A. T. Steel, J. M. Thomas and R. P. Townsend, Int. Zeolite Association Conf: Specialist Research Reports. 12 J. Fraissard and T. Ito Zeoliter, 1988 8, 350. 13 S. Ramdas, J. M. Thomas and P. A. Wright, J. Chem. Soc., Chem. Commun., 1984, 1338. 14 A. K. Cheetham, A. K. Nowak, J. van der Ouden, B. Petersen, S. D. Pickett and M. M. F. Post, J. 15 C. A. Fyfe, G. C. Gabbi, J. Klinowski and J. M. Thomas, Nature 1982, 296, 533. 16 J. Klinowski and J. M. Thomas, Adv. Catal., 1985, 33, 197. 17 S. W. C a n , C. R. A. Catlow, E. Dooryhee, G. N. Greaves, A. T. Steel, J. M. Thomas, and R. P. 18 S. W. Lytle, J. H . Sinfelt and G. H . Via, S~nchrotron Radiation Rerearch, ed. H . Winick and S. Doniach, 19 S. K. Harbron e t a l . , Inorg. Chem., 1986, 25, 1789. 20 N. Binstead, S. J. Gurman and I . Ross, J. Phjx C., 1984, 17, 143. 21 E. Oldfield and H. K. C. Timkin, J. Am. Chem. Soc., 1987, 109, 7699. 22 J. D. Jorgensen and J . M. Newsam, Zeolites, 1987, 7, 569. 23 J. M. Thomas and Xing Sheng Lui, J. Phys. Chem., 1986, 90, 4843. 24 Xing Sheng Lui, Ph.D. Theyis, (University of Cambridge, 1985). 25 D. H. Olson, J. Phys. Chem. 1968, 72, 652. 26 A. N. Fitch, J. Jobic and A. Renouprez, J. Chem. Soc., Chem. Commun. 1985, 284. 27 A. K. Cheetham, S. Ramdas, J. M. Thomas and P. A. Wright, J. Chem. Soc., Chem. Commun.., 1984, 1338. 28 A. K. Cheetham, A. K. Nowak, J. M. Thomas and P. A. Wright, Nature, 1985, 318, 611. 29 C. Williams, Ph.D. Thesis (University of Cambridge, 1987). 30 J. M. Newsam, in press. 31 S. W. Carr, C. R. A. Catlow, E. Dooryhee, G. N. Greaves, A. T. Steel, J. M. Thomas and R. P. Townsend, Phys. Chem., 1990, 94, 1233. Townsend, 8th Int. Zeolite Auociation Donf: (Amsterdam, 19891, Specialist Research Reports. (Plenum Press, New York, 19801, p.401. in preparation. Paper 0/00350F; Received 23rd January, 1990