The Effect of Dealumination on the Structure and Acidic Properties of Offretite C. FERNANDEZ, A. AUROUX, J.C. VEDRINE Institut de Recherches sur la Catalyse, Laboratoire Propre du CNRS, conventionne a l'universite Claude Bernard, Lyon I, 2 avenue Albert Einstein, 69626 Villeurbanne Cedex, Fr ance J. GROSMANGIN, G. SZABO Compagnie Franyaise de Raffinage, Centre de Recherches, BP 27, 72700 Harfleur, France Acidity of offretite zeolite dealuminated by hydrothermal treatment has been characterized by microcalorimetry and infra-red studies of NH3 adsorption. It has been observed that dealumination up to ca. 50% primarily decreased the middle acid strength of the offretite and let part of strong acid sites still present. INTRODUCTION Modifications of offretite lattice under hydrothermal steaming have not been previously studied to our knowledge although much effort has been devoted to such studies primarily for y-type and mordenite zeolites 11,21. In another paper 131 it has been shown that steaming at 870 K results in dealumination of the lattice of offretite. Holes evidenced by HREM are created during dealumination corresponding probably to the removal of some cage stackings. Aluminium compounds presumably of hydroxy-aluminic acid type formed dur ing dealumination seem to occupy some gmelinite cages or even channels for the more dealuminated samples, introducing additional physical constraints to reactant and product diffusion within the offretite channels. In the present work, we have been interested in the study of the modifications of acidity under hydrothermal treatment. EXPERIMENTAL Materials Steaming treatments were performed on an acidic offretite synthetised in the laboratory. The sample was placed in a quartz reactor and submitted to a flow of argon and water vapour. Temperature was then rised to 870 K. The hydrothermal treatment was achieved after three hours. After that treatment, the sample was treated by a solution of hydrochloric acid (N/2) in order to hopefully dissolve the aluminium compounds which were removed from the porous network and was further washed with deionized water and finally heated under air flow at 770 K. several cycles of such hydrothermal treatment were performed in order to increase the dealumination rate. Table 1 gives the chemical composition for all samples obtained. Al : Si ratios varied from 0.23 to 0.10. Dry (over Na wire) and degassed (by freeze-pump-thaw cycles) NH3 was the probe adsorbate.
Table 1 : Chemical composition of parent and dealuminated samples* Sample
Deal.r ate (%)
H-OFF H-Dl-0FF H-D2-0FF H-D3-0FF
H2.63 KO.56NSO.22(AI02) H2. 55KO. 18NaO. 03 ( Al02) Hl • 96KO. 09 NaO. 15 ( Al02) Hl.50KO.04NaO.14(Al02)
3.41 (Si0 2) 14.59 2.76 (Si0 2) 15.24 2.20 (Si0 2) 15.50 1.68(Si0 2) 16.32
0.234 0.181 0.139 0.103
19.9 35.7 50.9
*determined by atomic absorption analysis of Na,K and Al Methods Characterization by IR spectroscopy was performed using a PERKIN-ELMER 580 spectrometer. The vibrational bands of hydroxyl groups in the 4000-3400 cm- 1 domain were observed on a self-supported zeolite pellet (10 to 15 mg), placed in a pyrex cell. Samples were dehydrated under vacuum, at 720 K for two hours. Acidity of OH-groups was studied by adsorption of a basic probe. The acid strength of the different acid sites was determined by studying the recovering of the hydroxyl group band intensity after desorption at increasing temperatures. Ammonia which can easily enter all cages of offretite has been chosen as a basic probe. Differential heats of ammonia adsorption were measured using a SETARAM heat flow microcalorimeter connected to a vacuum and gas line equipped with a Barocel type pressure gauge to monitor the gas pressure within the range 0-133 Pa as described elsewhere 141. The zeolite samples (ca. 100 mg) were outgassed at 670 K in a quartz cell under 1.33 mPa vacuum overnight and then transferred to the microcalorimeter. The adsorption of NH3 was carried out by introducing successive doses of known amounts of the adsorbate onto the samples. The experimental set up was monitored by an Apple II microcomputer so that the evolved heat and the residual gas pressure could be recorded continuously following admission of each dose. Identical doses of NH3 were repeatedly admitted onto the calorimeter cell until a final equilibrium pressure of 133 Pa was reached. The calorimeter was maintained at 420 K. After the adsorption of the first doses the residual pressure was very low which means that the adsorbate was irreversibly held on the surface at 420 K. After the adsorption of further doses a measurable equilibrium pressure built up, while further evacuation resulted in the desorption of the reversible fraction of the adsorbate. RESULTS I.R. Spectroscopy analysis Figure 1 displays the IR spectra in the hydroxyl vibration domain of the parent and the dealuminated samples, after dehydration under vacuum at 720 K. The spectrum of the protonated offretite is very similar to that published previously 15,61. The spectrum exhibits four bands which correspond to different locations of OH groups within the porous framework as previously proposed 161. On the parent offretite, the 3740 cm- 1 band did not interact with ammonia due to the very weak acidity of the corresponding Si-OH groups. The 3660 cm- 1 band was restored after desorption at 420 K which confirms its weak acid character. The other two 3610 and 3550 cm- 1 bands were restored after outgassing to 573K. Therefore it can be concluded that the latter two bands correspond to strong ac id sites. oealumination of offretite zeolite by steaming affected most of the bands (see fig.l). Only the 3660 crn- 1 band, corresponding to the OH-sites located in hexagonal prisms, remained unchanged for all samples, whereas the 3610 and 3550 crn- 1 bands intensity decreased with the dealumination rate. Therefore, it may be concluded that only the more accessible sites were affected by the hydrothermal treatment. The variations of optical density for the two 3740 and 3610 cm- 1 bands are shown on figure 2. The 3740 cm- 1 band intensity increased linearly with dealumination, while that of the 3610 cm- 1 band decreased. The slope of the 3740 cm- 1 straight line was expected to be four times that associated with the 3610 crn- 1 i f four Si-OH groups
C. Fernandez et al.
were created for one aluminum atom removed in a kind of OH-group nest as suggested earlier for y-type dealuminated zeolite 111.
cm-' Fig .1
IR spectra of hydroxyl groups of H-OFF and steamed dealuminated samples.
dealuminatlon rate (%) Fig. 2 : variations of the 3740 and 3610 cm- 1 IR band vs dealumination rate. As the slopes are similar and reverse and although quantitative measurements in I.R. are not very reliable, it may be reasonably suggested that only some OH nests were created if any while the major part of aluminium atoms extracted from the lattice by' steaming was replaced by silicon atoms stemming from another part of the material, creating locally holes as evidenced by HREM 131.
348 (IM-3-2) Dealumination is known to involve a reduction of the number of acid sites. Chemisorption and further desorption of ammonia indicated that the strength of the remaining sites was preserved since recovering of IR OH band intensity was obtained by outgassing at 570K for parent and all dealuminated samples. Microcalorimetry The calorimetry results are illustrated on figure 3 which reports the variations of the heats of adsorption as a function of the coverage of the surface by successive doses of NH3' The heat of NH3 adsorption was initially equal to 170-180 kJ per mole for all samples. For the parent H-OFF, that heat decreased slowly with the amount of NH3 adsorbed : it can thus be concluded that the strength of the corresponding Br8nsted sites is very heterogeneous in contrast to other zeolites 171 For dealuminated samples the strongest acidic sites have roughly the same strength as the reference H-OFF sample but the NH3 adsorption heat decreased rapidly with coverage (Fig.3). The most dealuminated sample (H-D3-0FF) still exhibited a small amount of residual strong acidic sites. It is thus worthwhile to note that the heat of adsorption of the strongest acid sites did not vary with the aluminium content while the number of medium acid strength sites decreased with dealumination increasing.
'b...... 0 '-.r-._ ............. H-D2-oFF tRi3-oFF 5
(mol / u.c.)
Differential heats of NH3 adsorption vs coverage at 420 K
Such a feature is also shown in Fig.4 where the dn/dQ versus Q plots represent the number of sites dn corresponding to a given strength Q, i.e. the acid strength distribution 181. So it appears on the acidity spectra that essentially the medium acid strength sites were removed by steaming. Chemisorption, rather than physisorption, involves site activation and motion of the adsorbed molecule from an active site to a stronger one. The thermogram resulting from chemisorption is therefore expected to be broad because it takes longer time when the chemisorption strength increases. This idea has been exploited in the present investigation to follow the changes of sorption thermogram with sorbate coverage. The thermokinetic parameter studied is the peak width at half thermogram height. Its variations with NH3 coverage are shown in Fig.5.
Fernandez et al.
0.5 H-OFF H-03-0FF
o !----:.~~;;;;;;;;;;;!l!!!!!!~!!!!E!E~~~__1 o 50 Fig 4
Acid strength distribution vs differential heat of adsorption of ammonia.
NH 3 coverage (cm3. g-1)
Thermokinetic parameter vs NH3 coverage
Two kinetics regimes of heat evolution were observed for the starting offretite which may arise from two differently accessible sites presumably located in the cages and channels. Only one type was observed for dealuminated samples. It may thus be suggested that dealumination removed a partiCUlar type of sites which, as shown above· exhibited a medium acidic strength. It is worth noting that thermogravimetric measurements of n-hexane and 3-methylpentane adsorption 131 indicate that possibly gmelinite cages were destroyed or blocked by non extracted aluminium compounds while wide channel pattern was preserved during dealumination.
CONCLUSION Dealumination of offretite zeolite has been achieved by successive hydrothermal treatments at 870 K, separated by acidic treatments in order to hopefully dissolve and remove out of the porous network, aluminium compounds extracted by the steaming treatment. IR spectroscopy and microcalor imetry have clear ly shown that the number of acid sites decreased with dealumination particularly for the medium acid strength sites ~hich adsorbed NH3.with a heat in a range of 70 to 110 kJ.mol- l . Both techniques have also indicated that the remaining strongest acid sites conserved their strength (170-180 kJ.mol- l ) . Investigations concerning thermokinetic parameters versus ammonia coverage showed for the starting H-OFF sample two kinetic regimes of heat evolution corresponding to two different locations of acid sites, presumably in wide channels and in gmelinite cages and only one regime for dealuminated samples. It is then concluded that dealumination destroyed gmelinite cages reSUlting in changes in hydrocarbon diffusion within the zeolite grains 131. It follows that for catalytic methanol conversion reaction at 640K a better resistance to deactivation was obtained 191 for dealuminated samples which may be correlated to the decrease of acid sites density and also to the change in porous network. REFERENCES 1. For instance J.Scherzer, Catalytic Materials, ACS Symposium series,Eds. T.E. Whyte et al, 248, 157 (1984) 2. J. Bandiera, C. Hamon and C. Naccache, Proceed. 6th Intern.Zeol.Conf., Reno, 1983, D. Olson and A. Bisio Eds, Butterworths, London, 337 (1984) 3. C. Fernandez, J.C.Vedrine, J.Grosmangin and G.Szabo, Zeolites submitted. 4. A. Auroux, J.C. Vedrine and P.C. Gravelle, in "Adsorption at the GasSolid and Liquid-Solid Interfaces". J. Rouquerol and K.S.W. Sing Eds, Elsevier, Amsterdam, Study in Surf. Sci. and Catal., 10, 305 (1982) 5. E.L.Wu, T.E.Whyte, J.R.Venuto, J. Catal., 21, 384 (1971) 6. C. Mirodatos, A. Abou kaYs, J.C. Vedrine and D. Barthomeuf, J. Chern. Soc., Faraday Trans I, 57, 1786 (1977) 7. A. Auroux, V. BolTS, P. Wierzchowski, p.C.Gravelle, J.C. Vedrine, J. Chern. SOc, Faraday Trans II, 75, 2544 (1979) 8. T. Masuda, H. Taniguchi:-K. Tsutsumi, H. Takahashi, Bull. Chem.Soc. Japan, 52 (10), 2849 (1979) Y7 Mitani, K. Tsutsumi, H. Takahashi, Bull. Chern. Soc. Japan, 56, 1917 (1983) 9. C. Fernandez, J.C. Vedrine, J. Grosmangin and G. Szabo, Appl. catal. submitted.