- Email: [email protected]
Infrared Study of Acetophenone Adsorption on Mordenite and Dealuminated Mordenite
JOURNAL OF COLLOID AND INTERFACE SCIENCE ARTICLE NO.
207, 371–378 (1998)
CS985796
Infrared Study of Acetophenone Adsorption on Mordenite and Dealuminated Mordenite Imtiaz Ahmad, James A. Anderson, Trevor J. Dines, and C. H. Rochester1 Chemistry Department, University of Dundee, Dundee DD1 4HN, United Kingdom Received May 27, 1998; accepted August 6, 1998
Infrared spectra of acetophenone adsorbed on mordenite and acid-leached mordenite are reported. The effects of dealumination were characterized by pyridine adsorption and 29Si and 27Al magic angle spinning nuclear magnetic resonance spectroscopy (MAS NMR). Adsorption of pyridine gave bands at 1634 and 1545 cm21 due to vibrations of pyridinium ions and a band at 1621 cm21 characteristic of pyridine ligated to Lewis acidic Al31 sites. 27Al NMR gave peaks at 55 ppm due to tetrahedral framework Al and at 0 ppm due to octahedral extraframework Al. 29Si NMR gave peaks at 2106 and 2112 ppm due to Si(1Al) and Si(OAl) species, respectively. For acetophenone on mordenite, spectra contained n(CAO) bands at 1685, 1671, 1650, and 1630 cm21 due to physisorbed acetophenone containing unperturbed CAO groups, acetophenone hydrogen bonded with external and internal silanol groups, acetophenone hydrogen bonded to Brønsted acidic Si(OH)Al groups, and acetophenone ligated to Lewis acidic Al31 ions, respectively. Dealumination converted 56% of the Si(1Al) species to Si(OAl), reduced the concentrations of both framework and extraframework Al, reduced the concentration of Si(OH)Al groups, and greatly reduced the population of Lewis acidic sites, but generated internal silanol groups which formed hydrogen bonds with acetophenone molecules. Residual Al31 sites and Si(OH)Al groups after dealumination could interact with adsorbed pyridine molecules but for steric reasons were unable to form adsorption complexes with acetophenone. Evidence for the protonation of acetophenone was inconclusive. © 1998 Academic Press Key Words: acetophenone adsorption; mordenite acidity.
nation had occurred. Color change alone is insufficient and measurement of UV/visible spectra is strongly recommended (2). However, infrared spectra of adsorbed pyridine have been widely used to characterize surface acidity (5). Pyridine has a pK BH1 value (4) of 5.20 (6), and therefore Brønsted sites for which H o , ca. 5.20 would be identified. If, as with UV/visible spectroscopy (2), the infrared method could be applied to a series of bases covering a wide range of pK BH1 values, then this would provide an alternative method for determining and comparing relative strengths of Brønsted acid sites on different oxides. An advantage would be that other modes of adsorption, such as ligation to Lewis acidic sites or hydrogen bonding, would also be characterized by the infrared method. Acetophenones have been used to measure acidity functions for concentrated acid solutions (7), and benzalacetophenone (pK BH1 5 25.60) has been used in UV/visible spectroscopic studies of indicator adsorption (2, 8). Benzalacetophenone was too involatile to be adsorbed from the vapor phase, but acetophenone (pK BH1 5 26.15) (7) proved suitable for study. Mordenite has been classed as a mild superacid (2) which should protonate acetophenone and was therefore chosen to test whether infrared spectroscopy could be used to assess indicator protonation and hence the Brønsted acidity of hydroxyl groups. This paper reports a study of acetophenone on mordenite and dealuminated mordenite. EXPERIMENTAL
INTRODUCTION
The unambiguous determination of the strength of Brønsted acidic groups on the surfaces of colloid particles and heterogeneous catalysts, particularly oxidic materials, is a desirable objective. One approach has been to record visible spectra of adsorbed basic indicators which undergo color change on protonation. Acid– base indicators may adsorb on oxidic surfaces in a variety of ways, and resulting color changes may not necessarily be a result of simple indicator protonation (1–3). The assessment of H o acidity function (4) values for Brønsted acidic sites therefore relies on unambiguous proof that proto1
To whom correspondence should be addressed.
The microporous aluminosilicate heterogeneous catalyst Hmordenite (EZ-321, Engelhard) with surface area 478 m2 g21, pore volume 0.77 cm23 g21, and average particle size ca. 67 mm was pressed into self-supporting discs which were suspended in a vacuum infrared cell fitted with an external furnace and fluorite optical windows. Heat treatment at 673 K for 2 h in vacuum was followed by cooling to ambient temperature (ca. 295 K) and exposure to aliquots of acetophenone vapor. Spectra were recorded with a Perkin–Elmer 1720 FTIR spectrometer at 4 cm21 resolution. To test the effects of changing populations and possibly acid strengths of Lewis and Brønsted acidic sites, similar experiments were carried out with mordenite which had been dealuminated (9) by treatment with 8 M
371
0021-9797/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
372
AHMAD ET AL.
aqueous hydrochloric acid for 24 h at 358 6 1 K. The product was repeatedly washed with deionized water until chloride free (AgNO3 test), dried overnight in air at 398 K before being pressed into disks, heated (673 K, 2 h) in the infrared cell, and exposed to acetophenone. The adsorption of pyridine on both mordenite samples was also studied. The mordenites were also characterized by 29Si and 27Al NMR spectroscopy using a Chemagnetics CMX-300 Lite MAS NMR spectrometer with a Chemagnetics 7.0 mm MAS Turbo Pencil probe. Conditions for 29Si were as follows: spinning speed, 5 kHz; pulse width, 2.00 ms (45°); pulse delay, 5 s. Conditions for 27Al were 7 kHz; 2.00 ms (45°), and 1 s. RESULTS
Characterization of Effects of Dealumination The 29Si NMR spectrum of mordenite contained two peaks at 2106 and 2112 ppm which may be ascribed to Si(1Al) and Si(OA1) species, respectively (11, 13, 14). Dealumination by acid leaching (14) considerably enhanced the latter at the expense of the former (Fig. 1). Consideration of the peak intensities showed that ca. 56% of the Si(1Al) species present for the parent mordenite had been converted to Si(OAl) by the acid extraction dealumination procedure. Maxima in the 27Al MAS NMR spectra (Fig. 2) were at 55 ppm due to tetrahedral framework aluminum and at 0 ppm due to octahedral extraframework aluminum (14). In accordance with results of Barras et al. (14) the concentrations of both framework and extraframework aluminum were considerably reduced by acid leaching. The residual peak intensities for dealuminated mordenite [Fig. 2b] were 19% of the value for the parent mordenite for framework aluminum and 7% for extraframework aluminum. Strong maxima at 1634 and 1545 cm21 for pyridine on mordenite [Fig. 3a] were due to vibrations of pyridinium cations (5, 10) and confirmed the presence of a high concentration of Brønsted acidic sites. A band at 1621 cm21 is characteristic of pyridine ligation to Lewis acidic Al31 sites, and bands at 1597 and 1446 cm21 were due to pyridine hydrogen-bonded to surface hydroxyl groups (5). The involvement of hydroxyl groups in the adsorption was shown by decreases in intensities of bands at 3740 cm21 due to external silanol groups, and 3605 cm21 due to structural Brønsted acidic hydroxyl groups, each bridging one Si and one Al atom as represented by Si(OH)Al (11). The loss of the latter is accounted for by protonation of pyridine molecules. The loss of the former is more likely to be due to hydrogen bonding. The perturbed silanol groups gave a broad band at 2826 cm21 with additional bands also appearing at 2159 and possibly 1930 (vw) cm21. These bands are reminiscent of ABC triad bands for strong but neutral hydrogen bonding interactions (12). Dealumination decreased the numbers of both Lewis acidic sites and Brønsted acidic sites in the mordenite. The latter were
FIG. 1. mordenite.
29
Si MAS NMR spectra of (a) mordenite and (b) dealuminated
reduced to ca. 33% of their concentration for the parent mordenite. The percent reduction for Lewis sites could not be estimated because of overlap of the band at 1621 cm21 with other bands (Fig. 3) due to pyridinium ions or pyridine involved in hydrogen bonding interactions. Considerably more pyridine was hydrogen-bonded to silanol groups, as shown by the growth in intensities of the bands at 1597 and 1446 cm21 (Fig. 1b). The loss of Al31 from Brønsted acidic Si(OH)Al groups generated internal silanol groups (11) which for dealuminated mordenite provided an enhancement in the amount of pyridine able to adsorb via hydrogen-bonding interactions. The internal silanol groups gave, as before (11), a broad band at 3500 cm21 and an increase in absorption intensity at ca. 3700
ACETOPHENONE ADSORPTION ON MORDENITES
373
to the overall band envelope at 3700 –3740 cm21 due to SiOH groups, 3580 –3620 cm21 due to Si(OH)Al groups, and a broad tail to lower wavenumbers [Fig. 4a]. The adsorption of acetophenone had little effect on this spectral region although band growth resulted in a dominant broad maximum at 3400 cm21 (Fig. 4d). However, results in the lower spectral region suggested that several modes of adsorption occurred. Spectra of adsorbed acetophenone contained four bands due
FIG. 2. mordenite.
27
Al MAS NMR spectra of (a) mordenite and (b) dealuminated
cm21. Perturbation of the silanol groups by hydrogen bonding gave bands at 2776 and 2297 cm21 due to the AB members of an ABC triad for strong but neutral hydrogen-bonding interactions (12). The band at 3500 cm21 was the most weakened as a result of hydrogen bonding. The combined infrared and NMR data show that acid leaching led to loss of aluminum which reduced the number of Lewis acidic adsorption sites. Conversion of Si(1Al) to Si(OAl) decreased the concentration of Brønsted acidic Si(OH)Al sites but generated additional more weakly acidic silanol groups within the mordenite structure. Dealumination changed the (Si/Al)NMR ratio (14) from 9 to 22. XRD patterns for mordenite and dealuminated mordenite were the same, showing that dealumination did not induce a loss of mordenite crystallinity. Acetophenone Adsorption on Mordenite Spectra of hydroxyl groups on mordenite were rather broad and featureless although there was evidence for contributions
FIG. 3. Infrared spectra of pyridine adsorbed on (a) mordenite and (b) dealuminated mordenite.
374
FIG. 4. 295 K.
AHMAD ET AL.
Infrared spectra of mordenite after (a) evacuation at 673 K, (b– e) exposure to increasing amounts of acetophenone, and (f) 16 h further contact at
to n (CO) vibrations. Bands at 1671 and 1630 cm21 were due to hydrogen-bonded carbonyl groups and ligated carbonyl groups, respectively (Fig. 4c). At higher coverages a shoulder at ca. 1685 cm21 resembled the n (CO) band for liquid acetophenone and may be ascribed to weakly physisorbed molecules. During prolonged contact (Fig. 4f) this band was weakened, whereas the band at 1630 cm21 continued to grow in intensity. After 16 h contact there appeared to be an additional band at 1650 cm21. Other bands in the spectra at 1599, 1572, and 1452 cm21 (aromatic n (CAC) vibrations), 1421 cm21 (nas(CH3)), and 1366 cm21 (ns(CH3)) were less sensitive to adsorption and were at 1600, 1583, 1450, 1430, and 1360 cm21 for the liquid. A weak band at 1493 cm21 was scarcely detectable for the liquid. The results suggest that the dominant modes of adsorption involved retention of the acetophenone molecular
structure. However, a band which grew at 1324 cm21 was not in the liquid spectrum (15). Prolonged contact also gave a sharp maximum at 3305 cm21 (Fig. 4f), which together with the band at 1671 cm21 disappeared on evacuation at 295 K for 1 h. The first overtone of the band at 1685 cm21 for the liquid was at 3349 cm21. The band at 3305 cm21 could in part be similarly assigned for hydrogen-bonded acetophenone although this cannot be the sole attribution for such a strong band. Removal of hydrogen-bonded species resulted in growth of a band at ca. 3650 cm21 due to a n (OH) stretching vibration, probably of isolated silanol groups hydrogen-bonded to aromatic nuclei of acetophenone molecules ligated to adjacent Al31 sites. The band at 1630 cm21 due to the latter was not reduced on evacuation. Discs remained white after acetophenone adsorption for both
ACETOPHENONE ADSORPTION ON MORDENITES
FIG. 5.
375
Infrared spectra of dealuminated mordenite after (a) evacuation at 673 K and (b–f ) exposure to increasing amounts of acetophenone.
mordenite and dealuminated mordenite although this was probably not significant in the context of whether protonation had occurred or not because lmax for protonated acetophenone is at 290 nm (7), well into the UV region. Acetophenone Adsorption on Dealuminated Mordenite Spectra of acetophenone adsorbed on dealuminated mordenite were dominated by the effects of hydrogen bonding (Fig. 5). Loss of the band at 3745 cm21 due to isolated silanol groups paralleled the growth of a broad band at 3205 cm21 due to perturbed hydroxyl groups and a band at 1665 cm21 due to perturbed carbonyl groups. The latter had a shoulder at 1671 cm21 at high coverages. At low coverages some molecules were ligated to Al31 sites as shown by a weak band at 1630
cm21 (Fig. 5c). Under these conditions two of the bands at 1599 and 1572 cm21 due to aromatic n (CAC) vibrations were as for acetophenone on the parent mordenite. However, at higher coverages, where the hydrogen-bonded species dominated, the bands were in positions identical to those for the liquid at 1600 and 1583 cm21 (15). The shift of the latter band to 1572 cm21 for ligated acetophenone is consistent with results for acetophenone in metal complexes (16). Contact between dealuminated mordenite and acetophenone for long times (.16 h) at 295 K had little effect on the spectrum, which contrasted with the result for the parent mordenite for which slow replacement of a band at 1685 cm21 by a band at 1630 cm21 took place. Also, evacuation at 295 K for 24 h failed to desorb any acetophenone, unlike the result for
376
AHMAD ET AL.
groups and possibly resulting from enolisation reactions (17). In accordance with the expected band shift associated with the replacement of an OH-group by an OD group (18), a weak band appearing as a shoulder at ca. 2630 cm21 can be ascribed to the n (OD) mode of surface deuteroxyl groups confirming that some exchange has occurred. A broader, more intense maximum at 2466 cm21 could be the OD equivalent of the n (OH) band at 3400 cm21. However, this interpretation was confused by a similar broad band at 2420 cm21 in spectra of acetophenone on mordenite where there were no deuteriumcontaining species present. The latter band was accompanied by a second maximum of about the same shape and intensity at 2890 cm21. The combination of these two maxima closely resembled a spectrum of CH3CN on H-ZSM-5 (12), suggesting they were the AB members of an ABC triad for a hydrogenbonded complex of medium-high strength (12). The bands were particularly clear when adsorption was followed by evacuation for .1 h at 295 K, showing that this mode of adsorption was sufficiently strong to resist disruption under these conditions. Spectra in the 1750 –1300 cm21 region for the adsorption of acetophenone–methyl-d3 are not shown because their implications were wholly consistent with the corresponding results for acetophenone. DISCUSSION
FIG. 6. Infrared spectra of mordenite after (a) evacuation at 673 K, (b) exposure to acetophenone–methyl-d3 at 295 K, and (c) 16 h further contact at 295 K.
mordenite, for which the band at 1671 cm21 was considerably weakened as evacuation took place. The band at 1324 cm21 for mordenite was absent for dealuminated mordenite. Adsorption of Acetophenone–Methyl-d3 on Mordenite Deuterated acetophenone was adsorbed (Fig. 6) in order to test for H/D isotopic exchange involving surface hydroxyl
The behavior of mordenite after acid leaching is heavily dependent on the considerable dealumination which occurs. The NMR data confirm the extensive loss of aluminum from the present mordenite, and the results for pyridine adsorption confirm the loss of Lewis acidic Al31 adsorption sites. The accompanying loss of Brønsted acidic Si(OH)Al sites (11, 13, 14) generated internal silanol groups (11) which were additional to silanol groups on the external catalyst surface. The results for acetophenone adsorption paralleled those for pyridine in that dealumination decreased the extent of ligation to Lewis acidic sites but favoured hydrogen bonding interactions with silanol groups. The growth for mordenite of the band at 1630 cm21 due to ligated acetophenone at the expense of the band at 1685 cm21 due to weakly physisorbed acetophenone molecules containing unperturbed CO-groups is attributable to slow diffusion of molecules to Lewis acidic sites in the mordenite pore structure. External silanol groups on mordenite (11) formed hydrogen bonds with acetophenone as in complex (I) in Scheme 1, which gave n (OH) and n (CO) bands at 3305 and 1671 cm21. This constitutes a weak mode of adsorption and hence evacuation caused desorption. However, at least some silanol groups were sufficiently close to ligated acetophenone molecules as in (II) to generate complex (III) for which n (OH) and n (CAO) were at 3650 cm21 (a typical band position for silanol groups perturbed by benzene rings (19)) and 1630 cm21. Complex (III) was not detected for dealuminated mordenite because Al31 sites were far fewer than for mordenite.
ACETOPHENONE ADSORPTION ON MORDENITES
377
SCHEME 1
The enhancement of internal silanol groups made hydrogen bonding the dominant mode of interaction for dealuminated mordenite. The n (OH) and n (CO) bands at 3205 and 1665 cm21 corresponded to larger shifts from the band positions for unperturbed groups than for hydrogen bonding involving external silanol groups. This is indicative of a stronger form of hydrogen bonding, which is borne out by the resistance of acetophenone to desorption from internal silanol groups. The shoulder at 1671 cm21 [Fig. 5e] showed that external silanol groups on dealuminated mordenite also formed hydrogen bonds which, as for untreated mordenite, were broken by evacuation. Mordenite in the presence of pyridine gave infrared bands at 2826, 2159, and 1930 cm21 which conform with an ABC triad characteristic of medium-high strength hydrogen bonds (12). Dealuminated mordenite behaved similarly. This may be ascribed to the presence of external and internal silanol groups which act as hydrogen bond donors. However, protonation of pyridine also occurred, particularly for mordenite which had not been dealuminated, implying the presence of hydroxy groups with Brønsted acidity which was greater than that of internal or external silanol groups. Bridging Si(OH)Al groups (11, 13, 14, 20) fulfill this criterion. Protonation of acetophenone could not be unambiguously confirmed from the spectra. However, the spectra for mordenite, but not for dealuminated mordenite, exhibited a band at 1650 cm21 which represents a larger shift in the n(CO) band position from that for unperturbed carbonyl groups than the shift for hydrogen bonding involving silanol groups which gave a band at 1671 cm21. The implied enhanced hydrogen bond strength would be consistent with the involvement of Si(OH)Al groups. Brønsted acid sites having a range of acid strengths exist in mordenites and not all the sites will be accessible to acetophenone molecules (10). The Si(OH)Al sites in mordenite which interacted with acetophenone via hydrogen bonding were destroyed by dealumination. This parallels the elimination by dealumination of nearly all the Al31 sites in mordenite which were capable of interacting via ligation with acetophenone. The numbers of Si(OH)Al and Al31 sites available for interaction with pyridine were
reduced on dealumination but not eliminated. Thus the residual sites after dealumination must be accessible to interaction with pyridine, but are unable for steric reasons to generate adsorption complexes (H-bonding or ligation) involving acetophenone. This may result either from inability of acetophenone to diffuse to the sites, or because of steric impedance of the required acetophenone orientation for complex formation. The expectation that acidic hydroxy groups in mordenite should protonate acetophenone (pK BH1 5 26.15) (7) is based on visible spectra which showed that both of the two mordenites which were studied protonated benzalacetophenone (pK BH1 5 25.60) and contained sites which protonated 4-nitrotoluene (pK BH1 5 211.3) and 4-nitrofluorobenzene (pK BH1 5 212.4) (2). We carried out Hartree–Fock ab initio calculations at the SCF level (3-21G basis set) which, after appropriate scaling, predict n (OH) and n (CO) band positions of 3427 and 1317 cm21 for acetophenone protonated on the carbonyl oxygen atom. The present spectra for mordenite contained a sharp band at 1324 cm21 (Fig. 4f) which was absent for dealuminated mordenite (Fig. 5e), and in one instance (Fig. 4f) a band at 3305 cm21 which was also not observed for dealuminated mordenite (Fig. 5f). The assignment of these bands to protonated acetophenone would fit the predictions and also the considerable decrease caused by dealumination in the number of Brønsted sites detected by pyridine adsorption. However, this explanation is suspect, particularly with respect to the band at 3305 cm21. The latter must contain a contribution from the n (CO) band overtone for adsorbed acetophenone. The appearance of the band only after prolonged contact between mordenite and acetophenone and its rapid removal (,5 min) during subsequent evacuation at 295 K was at complete variance with the growth of the band at 1324 cm21 at all surface coverages and the retention of this band after evacuation. The two bands must have been due to different surface species. The band at 3305 cm21 cannot be due to protonated acetophenone, which could, however, be responsible for the band at 1324 cm21. An alternative attribution of the latter would be to an aromatic in-plane CH deformation vibration perturbed by ligation of acetophenone to Lewis acidic Al31 sites. This form of
378
AHMAD ET AL.
ligation was enormously reduced for dealuminated mordenite and, therefore, the absence of the band at 1324 cm21 is also explicable. In conclusion, there is no unambiguous evidence in the infrared spectra for acetophenone protonation. The strongest Brønsted acid sites may be in narrow channels and hence, although acetophenone molecules are small enough to diffuse through these channels, there will be a steric imposition to attaining a favorable orientation of acetophenone at Brønsted acid sites for proton transfer to occur and give an ionic hydrogen-bonded complex of the type discussed by Zecchina et al. (12). However, visible spectroscopy has detected strong Brønsted sites using indicator molecules at least as large as acetophenone (2). The infrared method enables modes of adsorption involving hydrogen bonding and ligation to be distinguished and characterized but, compared with visible spectroscopy, is unfortunately inadequate to identify Brønsted acidic sites by acetophenone adsorption. Acetone on the pure rutile form of titanium dioxide undergoes slow enolization and condensation to give adsorbed mesityl oxide (17). In the present study of mordenite the spectra slowly changed with time for two possible reasons: first, diffusion of molecules into the zeolite pore structure, and second, the occurrence of chemical reactions. The product of condensation would be dypnone [Ph(Me)CACHCOPh]. Evidence for H/D isotopic exchange suggested that enolization occurred, probably via an acid-catalyzed mechanism (21). However, condensation to dypnone in the pore structure was unlikely for steric reasons. Furthermore, disks remained white. Dypnone has the benzalacetophenone structure minus a Me group and therefore should not only be protonated by mordenite but should also turn yellow (2, 8). Acetophenone adsorption on titania gives dypnone, which leads to a pale yellow color (22). CONCLUSIONS
(a) Weak physisorption of acetophenone on mordenite is accompanied by three other modes of adsorption involving hydrogen bonding to internal and external silanol groups, hydrogen bonding to Brønsted acidic Si(OH)Al groups, and ligation to Lewis acidic Al31 ions. (b) Dealumination of mordenite by acid leaching converted a proportion of the Si(1Al) species to Si(OAl), reduced the concentrations of framework and extraframework Al and of Si(OH)Al groups, greatly reduced the population of Lewis acidic sites, and generated additional silanol groups.
(c) Residual Si(OH)Al groups and Al31 sites in dealuminated mordenite could interact with adsorbed pyridine molecules, but for steric reasons could not form complexes with acetophenone. (d) Inconclusive evidence for the protonation of acetophenone suggested that the study of acetophenone adsorption by infrared spectroscopy does not provide a good method for the detection and characterisation of Brønsted acidic sites on the surface of oxidic materials. ACKNOWLEDGMENT We thank the High Commission for Pakistan for a research studentship.
REFERENCES 1. Rochester, C. H., Discuss. Faraday Soc. 52, 285 (1971). 2. Umansky, B. S., and Hall, W. K., J. Catal. 124, 97 (1990). 3. Ahmad, I., Anderson, J. A., Dines, T. J., and Rochester, C. H., J. Chem. Soc., Faraday Trans. 92, 3225 (1996). 4. Rochester, C. H., “Acidity Functions.” Academic Press, London, 1970. 5. Parry, E. P., J. Catal. 2, 371 (1963). 6. Christensen, J. J., Hansen, L. D., and Izatt, R. M., “Handbook of Proton Ionisation Heats and Related Thermodynamic Quantities.” Wiley–Interscience, New York, 1976. 7. Stewart, R., and Yates, K., J. Am. Chem. Soc. 80, 6355 (1958). 8. Drushel, H. V., and Sommers, A. L., Anal. Chem. 38, 1723 (1966). 9. Sawa, M., Niwa, M., and Murakami, Y., Zeolites 12, 175 (1992). 10. Banko´s, I., Valyon, J., Kapustin, G. I., Kallo´, D., Klyachko, A. L., and Brueva, T. R., Zeolites 8, 189 (1988). 11. Wu, P., Komatsu, T., and Yashima, T., J. Phys. Chem. 99, 10923 (1995). 12. Zecchina, A., Bordiga, S., Spoto, G., Scarano, D., Spano`, G., and Geobaldo, F., J. Chem. Soc., Faraday Trans. 92, 4863 (1996). 13. Hays, G. R., van Erp, W. A., Alma, N. C. M., Couperus, P. A., Huis, R., and Wilson, A. E., Zeolites 4, 377 (1984). 14. Barras, J., Klinowski, J., and McComb, D. W., J. Chem. Soc., Faraday Trans. 90, 3719 (1994). 15. Ahmad, I., Dines, T. J., Anderson, J. A., and Rochester, C. H., J. Colloid Interface Sci. 195, 216 (1997). 16. Driessen, W. L., and Groeneveld, W. L., Rec. Trav. Chim. 90, 258 (1971). 17. Griffiths, D. M., and Rochester, C. H., J. Chem. Soc., Faraday Trans. 1 74, 403 (1978). 18. McDonald, R. S., J. Am. Chem. Soc. 79, 850 (1957). 19. Rochester, C. H., and Trebilco, D.-A., J. Chem. Soc., Faraday Trans. 1 74, 1125 (1978). 20. Shen, J.-P., Sun, T., Yang, X.-W., Jiang, D.-Z., and Min, E.-Z., J. Phys. Chem. 99, 12332 (1995). 21. Bell, R. P., “The Proton in Chemistry,” Chap. 8. Chapman and Hall, London, 1973. 22. Ahmad, I., Ph.D. Thesis, Dundee University, 1996.
207, 371–378 (1998)
CS985796
Infrared Study of Acetophenone Adsorption on Mordenite and Dealuminated Mordenite Imtiaz Ahmad, James A. Anderson, Trevor J. Dines, and C. H. Rochester1 Chemistry Department, University of Dundee, Dundee DD1 4HN, United Kingdom Received May 27, 1998; accepted August 6, 1998
Infrared spectra of acetophenone adsorbed on mordenite and acid-leached mordenite are reported. The effects of dealumination were characterized by pyridine adsorption and 29Si and 27Al magic angle spinning nuclear magnetic resonance spectroscopy (MAS NMR). Adsorption of pyridine gave bands at 1634 and 1545 cm21 due to vibrations of pyridinium ions and a band at 1621 cm21 characteristic of pyridine ligated to Lewis acidic Al31 sites. 27Al NMR gave peaks at 55 ppm due to tetrahedral framework Al and at 0 ppm due to octahedral extraframework Al. 29Si NMR gave peaks at 2106 and 2112 ppm due to Si(1Al) and Si(OAl) species, respectively. For acetophenone on mordenite, spectra contained n(CAO) bands at 1685, 1671, 1650, and 1630 cm21 due to physisorbed acetophenone containing unperturbed CAO groups, acetophenone hydrogen bonded with external and internal silanol groups, acetophenone hydrogen bonded to Brønsted acidic Si(OH)Al groups, and acetophenone ligated to Lewis acidic Al31 ions, respectively. Dealumination converted 56% of the Si(1Al) species to Si(OAl), reduced the concentrations of both framework and extraframework Al, reduced the concentration of Si(OH)Al groups, and greatly reduced the population of Lewis acidic sites, but generated internal silanol groups which formed hydrogen bonds with acetophenone molecules. Residual Al31 sites and Si(OH)Al groups after dealumination could interact with adsorbed pyridine molecules but for steric reasons were unable to form adsorption complexes with acetophenone. Evidence for the protonation of acetophenone was inconclusive. © 1998 Academic Press Key Words: acetophenone adsorption; mordenite acidity.
nation had occurred. Color change alone is insufficient and measurement of UV/visible spectra is strongly recommended (2). However, infrared spectra of adsorbed pyridine have been widely used to characterize surface acidity (5). Pyridine has a pK BH1 value (4) of 5.20 (6), and therefore Brønsted sites for which H o , ca. 5.20 would be identified. If, as with UV/visible spectroscopy (2), the infrared method could be applied to a series of bases covering a wide range of pK BH1 values, then this would provide an alternative method for determining and comparing relative strengths of Brønsted acid sites on different oxides. An advantage would be that other modes of adsorption, such as ligation to Lewis acidic sites or hydrogen bonding, would also be characterized by the infrared method. Acetophenones have been used to measure acidity functions for concentrated acid solutions (7), and benzalacetophenone (pK BH1 5 25.60) has been used in UV/visible spectroscopic studies of indicator adsorption (2, 8). Benzalacetophenone was too involatile to be adsorbed from the vapor phase, but acetophenone (pK BH1 5 26.15) (7) proved suitable for study. Mordenite has been classed as a mild superacid (2) which should protonate acetophenone and was therefore chosen to test whether infrared spectroscopy could be used to assess indicator protonation and hence the Brønsted acidity of hydroxyl groups. This paper reports a study of acetophenone on mordenite and dealuminated mordenite. EXPERIMENTAL
INTRODUCTION
The unambiguous determination of the strength of Brønsted acidic groups on the surfaces of colloid particles and heterogeneous catalysts, particularly oxidic materials, is a desirable objective. One approach has been to record visible spectra of adsorbed basic indicators which undergo color change on protonation. Acid– base indicators may adsorb on oxidic surfaces in a variety of ways, and resulting color changes may not necessarily be a result of simple indicator protonation (1–3). The assessment of H o acidity function (4) values for Brønsted acidic sites therefore relies on unambiguous proof that proto1
To whom correspondence should be addressed.
The microporous aluminosilicate heterogeneous catalyst Hmordenite (EZ-321, Engelhard) with surface area 478 m2 g21, pore volume 0.77 cm23 g21, and average particle size ca. 67 mm was pressed into self-supporting discs which were suspended in a vacuum infrared cell fitted with an external furnace and fluorite optical windows. Heat treatment at 673 K for 2 h in vacuum was followed by cooling to ambient temperature (ca. 295 K) and exposure to aliquots of acetophenone vapor. Spectra were recorded with a Perkin–Elmer 1720 FTIR spectrometer at 4 cm21 resolution. To test the effects of changing populations and possibly acid strengths of Lewis and Brønsted acidic sites, similar experiments were carried out with mordenite which had been dealuminated (9) by treatment with 8 M
371
0021-9797/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
372
AHMAD ET AL.
aqueous hydrochloric acid for 24 h at 358 6 1 K. The product was repeatedly washed with deionized water until chloride free (AgNO3 test), dried overnight in air at 398 K before being pressed into disks, heated (673 K, 2 h) in the infrared cell, and exposed to acetophenone. The adsorption of pyridine on both mordenite samples was also studied. The mordenites were also characterized by 29Si and 27Al NMR spectroscopy using a Chemagnetics CMX-300 Lite MAS NMR spectrometer with a Chemagnetics 7.0 mm MAS Turbo Pencil probe. Conditions for 29Si were as follows: spinning speed, 5 kHz; pulse width, 2.00 ms (45°); pulse delay, 5 s. Conditions for 27Al were 7 kHz; 2.00 ms (45°), and 1 s. RESULTS
Characterization of Effects of Dealumination The 29Si NMR spectrum of mordenite contained two peaks at 2106 and 2112 ppm which may be ascribed to Si(1Al) and Si(OA1) species, respectively (11, 13, 14). Dealumination by acid leaching (14) considerably enhanced the latter at the expense of the former (Fig. 1). Consideration of the peak intensities showed that ca. 56% of the Si(1Al) species present for the parent mordenite had been converted to Si(OAl) by the acid extraction dealumination procedure. Maxima in the 27Al MAS NMR spectra (Fig. 2) were at 55 ppm due to tetrahedral framework aluminum and at 0 ppm due to octahedral extraframework aluminum (14). In accordance with results of Barras et al. (14) the concentrations of both framework and extraframework aluminum were considerably reduced by acid leaching. The residual peak intensities for dealuminated mordenite [Fig. 2b] were 19% of the value for the parent mordenite for framework aluminum and 7% for extraframework aluminum. Strong maxima at 1634 and 1545 cm21 for pyridine on mordenite [Fig. 3a] were due to vibrations of pyridinium cations (5, 10) and confirmed the presence of a high concentration of Brønsted acidic sites. A band at 1621 cm21 is characteristic of pyridine ligation to Lewis acidic Al31 sites, and bands at 1597 and 1446 cm21 were due to pyridine hydrogen-bonded to surface hydroxyl groups (5). The involvement of hydroxyl groups in the adsorption was shown by decreases in intensities of bands at 3740 cm21 due to external silanol groups, and 3605 cm21 due to structural Brønsted acidic hydroxyl groups, each bridging one Si and one Al atom as represented by Si(OH)Al (11). The loss of the latter is accounted for by protonation of pyridine molecules. The loss of the former is more likely to be due to hydrogen bonding. The perturbed silanol groups gave a broad band at 2826 cm21 with additional bands also appearing at 2159 and possibly 1930 (vw) cm21. These bands are reminiscent of ABC triad bands for strong but neutral hydrogen bonding interactions (12). Dealumination decreased the numbers of both Lewis acidic sites and Brønsted acidic sites in the mordenite. The latter were
FIG. 1. mordenite.
29
Si MAS NMR spectra of (a) mordenite and (b) dealuminated
reduced to ca. 33% of their concentration for the parent mordenite. The percent reduction for Lewis sites could not be estimated because of overlap of the band at 1621 cm21 with other bands (Fig. 3) due to pyridinium ions or pyridine involved in hydrogen bonding interactions. Considerably more pyridine was hydrogen-bonded to silanol groups, as shown by the growth in intensities of the bands at 1597 and 1446 cm21 (Fig. 1b). The loss of Al31 from Brønsted acidic Si(OH)Al groups generated internal silanol groups (11) which for dealuminated mordenite provided an enhancement in the amount of pyridine able to adsorb via hydrogen-bonding interactions. The internal silanol groups gave, as before (11), a broad band at 3500 cm21 and an increase in absorption intensity at ca. 3700
ACETOPHENONE ADSORPTION ON MORDENITES
373
to the overall band envelope at 3700 –3740 cm21 due to SiOH groups, 3580 –3620 cm21 due to Si(OH)Al groups, and a broad tail to lower wavenumbers [Fig. 4a]. The adsorption of acetophenone had little effect on this spectral region although band growth resulted in a dominant broad maximum at 3400 cm21 (Fig. 4d). However, results in the lower spectral region suggested that several modes of adsorption occurred. Spectra of adsorbed acetophenone contained four bands due
FIG. 2. mordenite.
27
Al MAS NMR spectra of (a) mordenite and (b) dealuminated
cm21. Perturbation of the silanol groups by hydrogen bonding gave bands at 2776 and 2297 cm21 due to the AB members of an ABC triad for strong but neutral hydrogen-bonding interactions (12). The band at 3500 cm21 was the most weakened as a result of hydrogen bonding. The combined infrared and NMR data show that acid leaching led to loss of aluminum which reduced the number of Lewis acidic adsorption sites. Conversion of Si(1Al) to Si(OAl) decreased the concentration of Brønsted acidic Si(OH)Al sites but generated additional more weakly acidic silanol groups within the mordenite structure. Dealumination changed the (Si/Al)NMR ratio (14) from 9 to 22. XRD patterns for mordenite and dealuminated mordenite were the same, showing that dealumination did not induce a loss of mordenite crystallinity. Acetophenone Adsorption on Mordenite Spectra of hydroxyl groups on mordenite were rather broad and featureless although there was evidence for contributions
FIG. 3. Infrared spectra of pyridine adsorbed on (a) mordenite and (b) dealuminated mordenite.
374
FIG. 4. 295 K.
AHMAD ET AL.
Infrared spectra of mordenite after (a) evacuation at 673 K, (b– e) exposure to increasing amounts of acetophenone, and (f) 16 h further contact at
to n (CO) vibrations. Bands at 1671 and 1630 cm21 were due to hydrogen-bonded carbonyl groups and ligated carbonyl groups, respectively (Fig. 4c). At higher coverages a shoulder at ca. 1685 cm21 resembled the n (CO) band for liquid acetophenone and may be ascribed to weakly physisorbed molecules. During prolonged contact (Fig. 4f) this band was weakened, whereas the band at 1630 cm21 continued to grow in intensity. After 16 h contact there appeared to be an additional band at 1650 cm21. Other bands in the spectra at 1599, 1572, and 1452 cm21 (aromatic n (CAC) vibrations), 1421 cm21 (nas(CH3)), and 1366 cm21 (ns(CH3)) were less sensitive to adsorption and were at 1600, 1583, 1450, 1430, and 1360 cm21 for the liquid. A weak band at 1493 cm21 was scarcely detectable for the liquid. The results suggest that the dominant modes of adsorption involved retention of the acetophenone molecular
structure. However, a band which grew at 1324 cm21 was not in the liquid spectrum (15). Prolonged contact also gave a sharp maximum at 3305 cm21 (Fig. 4f), which together with the band at 1671 cm21 disappeared on evacuation at 295 K for 1 h. The first overtone of the band at 1685 cm21 for the liquid was at 3349 cm21. The band at 3305 cm21 could in part be similarly assigned for hydrogen-bonded acetophenone although this cannot be the sole attribution for such a strong band. Removal of hydrogen-bonded species resulted in growth of a band at ca. 3650 cm21 due to a n (OH) stretching vibration, probably of isolated silanol groups hydrogen-bonded to aromatic nuclei of acetophenone molecules ligated to adjacent Al31 sites. The band at 1630 cm21 due to the latter was not reduced on evacuation. Discs remained white after acetophenone adsorption for both
ACETOPHENONE ADSORPTION ON MORDENITES
FIG. 5.
375
Infrared spectra of dealuminated mordenite after (a) evacuation at 673 K and (b–f ) exposure to increasing amounts of acetophenone.
mordenite and dealuminated mordenite although this was probably not significant in the context of whether protonation had occurred or not because lmax for protonated acetophenone is at 290 nm (7), well into the UV region. Acetophenone Adsorption on Dealuminated Mordenite Spectra of acetophenone adsorbed on dealuminated mordenite were dominated by the effects of hydrogen bonding (Fig. 5). Loss of the band at 3745 cm21 due to isolated silanol groups paralleled the growth of a broad band at 3205 cm21 due to perturbed hydroxyl groups and a band at 1665 cm21 due to perturbed carbonyl groups. The latter had a shoulder at 1671 cm21 at high coverages. At low coverages some molecules were ligated to Al31 sites as shown by a weak band at 1630
cm21 (Fig. 5c). Under these conditions two of the bands at 1599 and 1572 cm21 due to aromatic n (CAC) vibrations were as for acetophenone on the parent mordenite. However, at higher coverages, where the hydrogen-bonded species dominated, the bands were in positions identical to those for the liquid at 1600 and 1583 cm21 (15). The shift of the latter band to 1572 cm21 for ligated acetophenone is consistent with results for acetophenone in metal complexes (16). Contact between dealuminated mordenite and acetophenone for long times (.16 h) at 295 K had little effect on the spectrum, which contrasted with the result for the parent mordenite for which slow replacement of a band at 1685 cm21 by a band at 1630 cm21 took place. Also, evacuation at 295 K for 24 h failed to desorb any acetophenone, unlike the result for
376
AHMAD ET AL.
groups and possibly resulting from enolisation reactions (17). In accordance with the expected band shift associated with the replacement of an OH-group by an OD group (18), a weak band appearing as a shoulder at ca. 2630 cm21 can be ascribed to the n (OD) mode of surface deuteroxyl groups confirming that some exchange has occurred. A broader, more intense maximum at 2466 cm21 could be the OD equivalent of the n (OH) band at 3400 cm21. However, this interpretation was confused by a similar broad band at 2420 cm21 in spectra of acetophenone on mordenite where there were no deuteriumcontaining species present. The latter band was accompanied by a second maximum of about the same shape and intensity at 2890 cm21. The combination of these two maxima closely resembled a spectrum of CH3CN on H-ZSM-5 (12), suggesting they were the AB members of an ABC triad for a hydrogenbonded complex of medium-high strength (12). The bands were particularly clear when adsorption was followed by evacuation for .1 h at 295 K, showing that this mode of adsorption was sufficiently strong to resist disruption under these conditions. Spectra in the 1750 –1300 cm21 region for the adsorption of acetophenone–methyl-d3 are not shown because their implications were wholly consistent with the corresponding results for acetophenone. DISCUSSION
FIG. 6. Infrared spectra of mordenite after (a) evacuation at 673 K, (b) exposure to acetophenone–methyl-d3 at 295 K, and (c) 16 h further contact at 295 K.
mordenite, for which the band at 1671 cm21 was considerably weakened as evacuation took place. The band at 1324 cm21 for mordenite was absent for dealuminated mordenite. Adsorption of Acetophenone–Methyl-d3 on Mordenite Deuterated acetophenone was adsorbed (Fig. 6) in order to test for H/D isotopic exchange involving surface hydroxyl
The behavior of mordenite after acid leaching is heavily dependent on the considerable dealumination which occurs. The NMR data confirm the extensive loss of aluminum from the present mordenite, and the results for pyridine adsorption confirm the loss of Lewis acidic Al31 adsorption sites. The accompanying loss of Brønsted acidic Si(OH)Al sites (11, 13, 14) generated internal silanol groups (11) which were additional to silanol groups on the external catalyst surface. The results for acetophenone adsorption paralleled those for pyridine in that dealumination decreased the extent of ligation to Lewis acidic sites but favoured hydrogen bonding interactions with silanol groups. The growth for mordenite of the band at 1630 cm21 due to ligated acetophenone at the expense of the band at 1685 cm21 due to weakly physisorbed acetophenone molecules containing unperturbed CO-groups is attributable to slow diffusion of molecules to Lewis acidic sites in the mordenite pore structure. External silanol groups on mordenite (11) formed hydrogen bonds with acetophenone as in complex (I) in Scheme 1, which gave n (OH) and n (CO) bands at 3305 and 1671 cm21. This constitutes a weak mode of adsorption and hence evacuation caused desorption. However, at least some silanol groups were sufficiently close to ligated acetophenone molecules as in (II) to generate complex (III) for which n (OH) and n (CAO) were at 3650 cm21 (a typical band position for silanol groups perturbed by benzene rings (19)) and 1630 cm21. Complex (III) was not detected for dealuminated mordenite because Al31 sites were far fewer than for mordenite.
ACETOPHENONE ADSORPTION ON MORDENITES
377
SCHEME 1
The enhancement of internal silanol groups made hydrogen bonding the dominant mode of interaction for dealuminated mordenite. The n (OH) and n (CO) bands at 3205 and 1665 cm21 corresponded to larger shifts from the band positions for unperturbed groups than for hydrogen bonding involving external silanol groups. This is indicative of a stronger form of hydrogen bonding, which is borne out by the resistance of acetophenone to desorption from internal silanol groups. The shoulder at 1671 cm21 [Fig. 5e] showed that external silanol groups on dealuminated mordenite also formed hydrogen bonds which, as for untreated mordenite, were broken by evacuation. Mordenite in the presence of pyridine gave infrared bands at 2826, 2159, and 1930 cm21 which conform with an ABC triad characteristic of medium-high strength hydrogen bonds (12). Dealuminated mordenite behaved similarly. This may be ascribed to the presence of external and internal silanol groups which act as hydrogen bond donors. However, protonation of pyridine also occurred, particularly for mordenite which had not been dealuminated, implying the presence of hydroxy groups with Brønsted acidity which was greater than that of internal or external silanol groups. Bridging Si(OH)Al groups (11, 13, 14, 20) fulfill this criterion. Protonation of acetophenone could not be unambiguously confirmed from the spectra. However, the spectra for mordenite, but not for dealuminated mordenite, exhibited a band at 1650 cm21 which represents a larger shift in the n(CO) band position from that for unperturbed carbonyl groups than the shift for hydrogen bonding involving silanol groups which gave a band at 1671 cm21. The implied enhanced hydrogen bond strength would be consistent with the involvement of Si(OH)Al groups. Brønsted acid sites having a range of acid strengths exist in mordenites and not all the sites will be accessible to acetophenone molecules (10). The Si(OH)Al sites in mordenite which interacted with acetophenone via hydrogen bonding were destroyed by dealumination. This parallels the elimination by dealumination of nearly all the Al31 sites in mordenite which were capable of interacting via ligation with acetophenone. The numbers of Si(OH)Al and Al31 sites available for interaction with pyridine were
reduced on dealumination but not eliminated. Thus the residual sites after dealumination must be accessible to interaction with pyridine, but are unable for steric reasons to generate adsorption complexes (H-bonding or ligation) involving acetophenone. This may result either from inability of acetophenone to diffuse to the sites, or because of steric impedance of the required acetophenone orientation for complex formation. The expectation that acidic hydroxy groups in mordenite should protonate acetophenone (pK BH1 5 26.15) (7) is based on visible spectra which showed that both of the two mordenites which were studied protonated benzalacetophenone (pK BH1 5 25.60) and contained sites which protonated 4-nitrotoluene (pK BH1 5 211.3) and 4-nitrofluorobenzene (pK BH1 5 212.4) (2). We carried out Hartree–Fock ab initio calculations at the SCF level (3-21G basis set) which, after appropriate scaling, predict n (OH) and n (CO) band positions of 3427 and 1317 cm21 for acetophenone protonated on the carbonyl oxygen atom. The present spectra for mordenite contained a sharp band at 1324 cm21 (Fig. 4f) which was absent for dealuminated mordenite (Fig. 5e), and in one instance (Fig. 4f) a band at 3305 cm21 which was also not observed for dealuminated mordenite (Fig. 5f). The assignment of these bands to protonated acetophenone would fit the predictions and also the considerable decrease caused by dealumination in the number of Brønsted sites detected by pyridine adsorption. However, this explanation is suspect, particularly with respect to the band at 3305 cm21. The latter must contain a contribution from the n (CO) band overtone for adsorbed acetophenone. The appearance of the band only after prolonged contact between mordenite and acetophenone and its rapid removal (,5 min) during subsequent evacuation at 295 K was at complete variance with the growth of the band at 1324 cm21 at all surface coverages and the retention of this band after evacuation. The two bands must have been due to different surface species. The band at 3305 cm21 cannot be due to protonated acetophenone, which could, however, be responsible for the band at 1324 cm21. An alternative attribution of the latter would be to an aromatic in-plane CH deformation vibration perturbed by ligation of acetophenone to Lewis acidic Al31 sites. This form of
378
AHMAD ET AL.
ligation was enormously reduced for dealuminated mordenite and, therefore, the absence of the band at 1324 cm21 is also explicable. In conclusion, there is no unambiguous evidence in the infrared spectra for acetophenone protonation. The strongest Brønsted acid sites may be in narrow channels and hence, although acetophenone molecules are small enough to diffuse through these channels, there will be a steric imposition to attaining a favorable orientation of acetophenone at Brønsted acid sites for proton transfer to occur and give an ionic hydrogen-bonded complex of the type discussed by Zecchina et al. (12). However, visible spectroscopy has detected strong Brønsted sites using indicator molecules at least as large as acetophenone (2). The infrared method enables modes of adsorption involving hydrogen bonding and ligation to be distinguished and characterized but, compared with visible spectroscopy, is unfortunately inadequate to identify Brønsted acidic sites by acetophenone adsorption. Acetone on the pure rutile form of titanium dioxide undergoes slow enolization and condensation to give adsorbed mesityl oxide (17). In the present study of mordenite the spectra slowly changed with time for two possible reasons: first, diffusion of molecules into the zeolite pore structure, and second, the occurrence of chemical reactions. The product of condensation would be dypnone [Ph(Me)CACHCOPh]. Evidence for H/D isotopic exchange suggested that enolization occurred, probably via an acid-catalyzed mechanism (21). However, condensation to dypnone in the pore structure was unlikely for steric reasons. Furthermore, disks remained white. Dypnone has the benzalacetophenone structure minus a Me group and therefore should not only be protonated by mordenite but should also turn yellow (2, 8). Acetophenone adsorption on titania gives dypnone, which leads to a pale yellow color (22). CONCLUSIONS
(a) Weak physisorption of acetophenone on mordenite is accompanied by three other modes of adsorption involving hydrogen bonding to internal and external silanol groups, hydrogen bonding to Brønsted acidic Si(OH)Al groups, and ligation to Lewis acidic Al31 ions. (b) Dealumination of mordenite by acid leaching converted a proportion of the Si(1Al) species to Si(OAl), reduced the concentrations of framework and extraframework Al and of Si(OH)Al groups, greatly reduced the population of Lewis acidic sites, and generated additional silanol groups.
(c) Residual Si(OH)Al groups and Al31 sites in dealuminated mordenite could interact with adsorbed pyridine molecules, but for steric reasons could not form complexes with acetophenone. (d) Inconclusive evidence for the protonation of acetophenone suggested that the study of acetophenone adsorption by infrared spectroscopy does not provide a good method for the detection and characterisation of Brønsted acidic sites on the surface of oxidic materials. ACKNOWLEDGMENT We thank the High Commission for Pakistan for a research studentship.
REFERENCES 1. Rochester, C. H., Discuss. Faraday Soc. 52, 285 (1971). 2. Umansky, B. S., and Hall, W. K., J. Catal. 124, 97 (1990). 3. Ahmad, I., Anderson, J. A., Dines, T. J., and Rochester, C. H., J. Chem. Soc., Faraday Trans. 92, 3225 (1996). 4. Rochester, C. H., “Acidity Functions.” Academic Press, London, 1970. 5. Parry, E. P., J. Catal. 2, 371 (1963). 6. Christensen, J. J., Hansen, L. D., and Izatt, R. M., “Handbook of Proton Ionisation Heats and Related Thermodynamic Quantities.” Wiley–Interscience, New York, 1976. 7. Stewart, R., and Yates, K., J. Am. Chem. Soc. 80, 6355 (1958). 8. Drushel, H. V., and Sommers, A. L., Anal. Chem. 38, 1723 (1966). 9. Sawa, M., Niwa, M., and Murakami, Y., Zeolites 12, 175 (1992). 10. Banko´s, I., Valyon, J., Kapustin, G. I., Kallo´, D., Klyachko, A. L., and Brueva, T. R., Zeolites 8, 189 (1988). 11. Wu, P., Komatsu, T., and Yashima, T., J. Phys. Chem. 99, 10923 (1995). 12. Zecchina, A., Bordiga, S., Spoto, G., Scarano, D., Spano`, G., and Geobaldo, F., J. Chem. Soc., Faraday Trans. 92, 4863 (1996). 13. Hays, G. R., van Erp, W. A., Alma, N. C. M., Couperus, P. A., Huis, R., and Wilson, A. E., Zeolites 4, 377 (1984). 14. Barras, J., Klinowski, J., and McComb, D. W., J. Chem. Soc., Faraday Trans. 90, 3719 (1994). 15. Ahmad, I., Dines, T. J., Anderson, J. A., and Rochester, C. H., J. Colloid Interface Sci. 195, 216 (1997). 16. Driessen, W. L., and Groeneveld, W. L., Rec. Trav. Chim. 90, 258 (1971). 17. Griffiths, D. M., and Rochester, C. H., J. Chem. Soc., Faraday Trans. 1 74, 403 (1978). 18. McDonald, R. S., J. Am. Chem. Soc. 79, 850 (1957). 19. Rochester, C. H., and Trebilco, D.-A., J. Chem. Soc., Faraday Trans. 1 74, 1125 (1978). 20. Shen, J.-P., Sun, T., Yang, X.-W., Jiang, D.-Z., and Min, E.-Z., J. Phys. Chem. 99, 12332 (1995). 21. Bell, R. P., “The Proton in Chemistry,” Chap. 8. Chapman and Hall, London, 1973. 22. Ahmad, I., Ph.D. Thesis, Dundee University, 1996.