W.K. Hall

On the strength of solid acids

The objective of the present work was to establish the relative strengths of some common solid acids and to put them on a scale familiar to chemists. A modified spectrophotometric approach involving Ho and HR indicators was developed. The red-shifts of spectra from the adsorbed neutral forms of 4-nitrotoluene and 4-nitrofluorobenzene were used to deduce the equivalent H2S04 concentrations. Thus, values of 70 to 80%, and 90 to 98% H2S04 were determined for silica-alumina and HY preparations, respectively. HM preparations were found to be mild superacids (stronger than 100% H2S04). On the other hand, H2S04/ZrO2 preparations had surface acidities about equivalent to 100% H2S04, and SbF5/silica-alumina was a superacid with − 14.5 < H0 < − 13.7. These data were correlated with the catalytic activities of the solid acids expressed either as In (reaction rate) at 370°C or as the temperature required to obtain 0.5% conversion of isobutane. The extent and nature of the secondary reactions also varied with the acidity. These results demonstrate that the intensive factor of the acidity (acid strength) dominated over the extensive factor (surface concentration of Bronsted sites) in controlling the acidity, and therefore the catalytic properties, of these solids.

The mechanism of neopentane cracking over solid acids

Abstract Neopentane was used as a probe to test whether catalyst protons can attack the CC and/or the CH σ-bonds in the cracking of alkanes. Over a variety of solid acids approximately one CH 4 molecule was formed for every neopentane reacted. Moreover, in most cases nearly as much isobutene was formed. With HY zeolite (HY) and particularly with H-mordenite (HM), however, hydride transfer leading to paraffin formation became important or dominant. Generally, the results conformed to with the t -butyl ion either decomposing to isobutene or else undergoing secondary reactions. The latter tended to increase with the intensive factor of the acidity, i.e., with the strength of the acid-base interaction. The fraction of neopentane converted to CH 4 , when plotted in the Arrhenius fashion against T −1 , produced straight lines from which apparent activation energies could be calculated. Values obtained fell between 30 kcal/mol for silica-alumina and 14 kcal/mol for HM (the most acidic and active catalyst investigated). Controversial views found in current literature are discussed in light of the present results.

The assay of acid sites on zeolites as measured by ammonia poisoning

Abstract The decomposition of neopentane was studied with and without selective site poisoning with NH3. The lethal dose, which was just sufficient to kill the catalytic activity, was estimated for each of a series of H-mordenites, HY and HZSM-5 zeolites, all with Si Al ratios higher than 7.0, i.e., over silica-rich compositions where ideally all the sites are supposedly identical. It was found that poisoning no more than 10% of the sites (counted as lattice aluminum ions) was sufficient to eliminate the activity. Some ramifications of these observations are discussed. Unpoisoned and partially poisoned zeolites were investigated using 13C-MASNMR following adsorption of isopropanol-2-13C. These findings clearly show the efficacy of adsorbed NH3 in repressing formation of polymeric surface residues.

On the surface chemistry of molybdena-alumina catalysts prepared from Mo(CO)6

Catalysts were prepared by subliming Mo(CO)6 onto partially dehydroxylated (PDA) and exhaustively dehydroxylated (DA) alumina made from the same parent preparation (American Cyanamid Aero 100PHF). The chemisorption of NO and CO on these materials was studied using volumetric, Chromatographic, and spectroscopic techniques. ESCA data indicated that metallic Mo crystals formed on Mo(CO)6DA whereas on PDA both Mo4+ and some lower valence state, Mo2+ or Mo0, were present. NO chemisorbed on both preparations at 195 K without releasing either N2O or N2. The chemisorption on the PDA preparations was over tenfold higher than that on the DA-supported catalysts under these conditions, but on raising the temperature to 300 K the difference was reduced to a factor of 2. Moreover, redox chemistry occurred at this higher temperature as evidenced by the release of N2O and N2; this was quantified. The amounts of NO actually chemisorbed under these conditions correlated well with the integrated IR band intensities. These data suggest that lower valence states are oxidized to Mo4+ at 300 K and that the observed IR bands stem from Mo4+(NO)2, irrespective of the initial catalyst. Infrared spectra from residual CO remaining on decomposition of Mo(CO)6 on DA and PDA at increasing temperatures up to 473 K showed bands which could be attributed to residual Mo(CO)6 and/or to subcarbonyl species formed during decomposition. By 573 K, no residual bands could be observed. On adding-back CO at 300 K to the PDA preparation, bands at 1989 and 2170 cmt−1 appeared, suggesting the presence of Mo4+ and residual Mo0. Spectra from similar experiments with the DA preparation demonstrated that chemisorbed Mo(CO)6 was reforming and possibly some subcarbonyl species; there was scant evidence for any higher valence states. The results are discussed in terms of the current literature.

Study of the mechanism of the cracking of small alkane molecules on HY Zeolites

Cracking of i-butane and n-pentane was studied on HY zeolites. These reactions were initiated by the protonation of C-H and C-C bonds by the Bronsted acid sites. The pentacoordinated carbonium ions thus formed decomposed into carbenium ions and the products of the initiation reactions, viz., H2, CH4, and C2H6. These carbenium ions propagated chain reactions, mainly by isomerization followed by hydride transfer. Disproportionation occurred concomitantly. For these small alkanes the contribution of /gb-scission to product formation was negligible. “Chain length” (the number of chain cycles per initiation) was defined as the ratio of reactant molecules consumed by hydride transfer to those reacted by protonation, i.e., the ratio of rates of bimolecular propagation to unimolecular initiation. Thus, chain length reflected the average lifetime of the carbenium ions. The extent of protonation was found to increase with the strength of acid sites while the chain length remained relatively constant for preparations of a similar nature. The product distribution obtained was therefore critically dependent on the steady-state population of carbonium ions. Finally the chain reactions were terminated by decomposition of carbenium ions into corresponding alkenes. Mass balances derived from these initiation, propagation, and termination steps were in agreement for both the substrates. The product distributions obtained for i-butane and n-pentane cracking were satisfactorily explained on this basis.

Reactions of benzene and alkylbenzenes with deuterium over molybdena-alumina and alumina catalysts

The reactions of D2 with aromatic systems were studied. Benzene did not hydrogenate at atmospheric pressure. On the other hand, exchange between benzene-do and D2 was facile at 70 ° and 1 atm over reduced (eMo ≈ 1.7) and at 200 ° over sulfided (eMo≈ 3.0) catalysts. The exchange reaction also occurred over the alumina support at an intermediate rate. Over alumina the exchange at 70 ° was stepwise (M = koko≈1.0), but as the temperature and hence k0 increased, multiple exchange (M ≥ 2.0) set in. The usual tests showed that this was not a result of a pore diffusion limitation. The rate of exchange of C6D6 with the alumina hydroxyl groups increased sharply as the temperature was increased to near 200 ° in a way which mimicked the increase in M. Thus, the increase in multiplicity may be ascribed to the opening of new pathways for exchange. The much faster rates and higher values of M obtained over the reduced catalysts than over the sulfided ones, coupled with the results of poisoning experiments using NO and CO2 (both of which drastically reduced the rates and returned the exchange process to M ≈ 1.0), demonstrated that with the reduced catalyst both portions of the surface act synergistically. The results may be rationalized by the supposition that exchange occurs on the alumina support and that hydrogen may spill-over on reduced catalysts, but not on sulfided ones. In the exchange reactions, alkylaromatics were found to have higher rates for ring than for side chain hydrogens.