J.R. Anderson

Reactions on ZSM-5-type zeolite catalysts

Catalytic reactions and sorption measurements have been carried out with ZSM-5 and silicalite catalysts which are distinguished by variation in skeletal heteroatom concentration. The catalysts were used in both the hydrogen-exchanged and sodium-exchanged forms. Sorption measurements were made with the hydrocarbons n-hexane, 3-methylpentane, and 2,2-dimethylbutane, and with the bases ammonia, n-butylamine, t-butylamine, and 4-methylquinoline. Catalytic reactions were carried out on both unpoisoned and base-poisoned catalysts using methanol, propylene, and 3-methylpentane reactants. In addition, the behavior of ethylene and ethanol reactants was also explored. The ease of base sorption and hydrocarbon sorption has been assessed in terms of effective molecular size in relation to the channel size of the catalyst, and this factor is also used as a basis for explaining the effectiveness of bases for poisoning the catalytic reactions. Temperature-programmed desorption (TPD) measurements with ammonia have been used to assess the energetic distribution of sorption sites for bases, and very strong binding sites with a TPD maximum for ammonia at about 780 °K have been identified as the probable sites used in the conversion processes. The main features of the catalytic conversion process are discussed. It is concluded that sorbed C3,C4 olefinic residues are general intermediates leading to aromatic formation. Under most circumstances, ethylene was relatively unreactive, and it is inferred that a sorbed C2 residue, which is related to ethylene by sorption and desorption, is an unlikely general intermediate. A mechanism is suggested for the formation of sorbed C3,C4 olefinic residues, and for initial carbon-carbon bond formation from methanol. Catalyst self-poisoning was observed with all catalysts except hydrogen-exchanged ZSM-5.

Reactions of neopentane and neohexane on platinum/Y-zeolite and platinum/silica catalysts

Isomerization and hydrogenolysis reactions of neopentane and neohexane have been studied in the presence of excess hydrogen at 455 to 625 K in a flow reactor over various Pt/silica and Pt/Y-zeolite catalysts having dPt in the range 1 to 20 nm. Additional catalyst characterization was provided by ESCA and by hydrogen TPD. Evidence is given to show that the reaction of neopentane was entirely confined to the platinum, even with Y-zeolite carrier. However, over Pt/Y-zeolite neohexane showed features which indicate dual-function catalytic behavior. As dPt decreased there was a general trend for the isomerization selectivity for neopentane (SI) to decrease and for the activation energy (Ea) to increase. From the way in which SI and Ea varied with dPt, it is concluded that there are two reaction pathways for neopentane, one on low index crystallite facets, and a second at platinum atoms of low coordination, the probable site for the latter being a single platinum atom. Hydrogen TPD revealed an increasing proportion of higher energy binding sites at very small values of dPt, and a correlation exists between SI and the concentration of adsorbed hydrogen under reaction conditions as estimated by TPD. ESCA examination of Pt/Y-zeolite (dPt ≈ 1 nm) showed the presence of relatively electron-deficient platinum, the extent of this being greater for Pt/(La)Y-than for Pt/(Na)Y-zeolite. For the neopentane reaction, SI and Ea were somewhat greater for Pt/(Na)Y- than for Pt/(La) Y-zeolite. Kinetic pressure dependence data showed that provided pH2 was high enough, hydrogen was an inhibiting gas. Relative to the strength of hydrogen adsorption, the surface reaction intermediate from neopentane was adsorbed more strongly on Pt/silica (dPt ≈ 4 nm) than on Pt/Y-zeolite (dPt ≈ 1 nm). The nature of the adsorbed intermediates and the reaction pathways are discussed.

Preparation of some supported metallic catalysts from metallic cluster carbonyls

An examination was made of the adsorption of some metallic cluster carbonyls (MCCs), Co2Rh2(CO)12, Co3Rh(CO)12, Co4(CO)12, Ir4(CO)12, Rh6(CO)16, and Ru3(CO)12, from nonaqueous solution onto two typical catalyst supports, γ-alumina and Aerosil silica. With two MCCs, Co2Rh2(CO)12 and Ir4(CO)12, dispersed metallic catalysts were generated, and a study was made of how the main experimental conditions affected the metallic dispersion. MCC adsorption was more facile on γ-alumina than on silica and was often assisted by the presence of oxygen. An ir study showed that initial adsorption of Co2Rh2(CO)12 on γ-alumina occurred with the loss of bridging carbonyls, the remaining carbonyls being progressively lost at temperatures >300 K, while adsorption of Ir4(CO)2 on γ-alumina resulted in progressive carbonyl loss at 320–620 K. Strong adsorption involves carbonyl loss, probably by ligand exchange with a surface anion, and the effect of oxygen is probably oxidative decarbonylation. Catalysts prepared from Co2Rh2(CO)12 or Ir4(CO)12 were relatively highly dispersed (D ≈ 0.4-1 depending on conditions), and Co2Rh2(CO)12 gave a much higher dispersion than was obtained by conventional impregnation using aqueous salt solutions. MCC adsorption in the presence of oxygen favored higher dispersions.

Mechanism of some conversions over ZSM-5 catalyst

Abstract Various carbon-carbon bond formation reactions over ZSM-5 catalyst have been studied with the help of deuterium labeling. The reactions studied were: benzene methylation (benzene/methanol), benzene ethylation (benzene/ethanol and benzene/ethylene), and the conversion of propylene and of ethylene to higher molecular weight products, together with some ancillary exchange reactions. It is shown that, in the methylation of benzene, the methyl group remains intact throughout and does not undergo exchange with surface hydrogen or aryl hydrogen. All the carbon-carbon bond formation reactions are interpreted as Bronsted acid-catalyzed electrophilic alkylations. From these data it is suggested that the ZSM-5-catalyzed conversion of methanol to hydrocarbons involves electrophilic methylation of olefinic intermediates as the main propagation reaction for carbon-carbon bond formation.

The isomerization of aliphatic hydrocarbons over evaporated films of platinum and palladium

Saturated hydrocarbons in the presence of excess hydrogen undergo catalytic skeletal isomerization and hydrocracking at a platinum surface. For mechanistic purposes, a study has been made of these reactions (i) with ethane, n -butane, isobutane, neopentane, and isopentane over unoriented evaporated films of platinum and palladium, (ii) with n -butane and isobutane over (111) and (100) oriented films of platinum, and (iii) with n -butane-1-C 13 over unoriented films of platinum. Platinum films exposing (111) and (100) surfaces were prepared by deposition on mica and on an evaporated sodium chloride layer, respectively. Over palladium, most of the reaction was hydrocracking: In no case did isomerization contribute more than about 3% to the initial reaction products. Over platinum, the proportion of isomerization product was substantial. Isobutane isomerized more readily than n -butane and the isomerization of the former, but not of the latter, was markedly increased by using a (111) platinum surface. Reactions with both n -and isobutane over (100) platinum gave product distributions only marginally different from unoriented platinum. The main features of the distributions of hydrocracking products could be approximately accounted for by assuming that the residence of a molecule on the surface resulted in the rupture of not more than one carbon-carbon bond. Initial product distributions were substantially independent of temperature. With reaction mixtures having p H 2 / p HC The isobutane formed from n -butane-1-C 13 was labeled only in the 2-position (i.e., peripherally). No scrambling of the C 13 occurred. The isomerization was thus entirely intramolecular. At the same time n -butane-2-C 13 was produced in a constant proportion to the isobutane. These products were thus probably formed concurrently from a common surface intermediate. From arguments based on the influence of hydrocarbon geometry on the reaction, it was concluded that the surface intermediate for isomerization and hydrocracking was 1–3 diadsorbed, except in the case of ethane which was 1–2 diadsorbed. In agreement with this, a 1–3 diadsorbed intermediate provided a reasonably satisfactory quantitative account of the proportion of n -butane-2-C 13 relative to isobutane in the reaction products. The exceptionally high degree of isomerization of isobutane on (111) platinum was tentatively ascribed to a symmetrical triadsorbed surface intermediate.

Hydrocarbon reactions on supported iridium catalysts

The reactions of ethane, n -butane, isobutane, neopentane, neohexane, 2,3-dimethylbutane, and methylcyclopentane have been studied over a range of dispersed iridium catalysts in which the iridium was supported on γ -alumina or Aerosil silica, and d Ir lay in the range ≤ 1 to 20 nm. The catalysts were characterized by electron microscopy, hydrogen adsorption, and temperature-programmed desorption of hydrogen. Hydrocarbon hydrogenolysis was the sole reaction pathway, except in the reaction of neopentane over catalysts with d Ir of 7 and 20 nm, where some isomerization to isopentane was observed, and in the case of isobutane, where there was some inferential evidence for isomerization to n -butane prior to hydrogenolysis. The hydrocarbons fell into one of two classes depending upon the activation energy and the type of hydrogenolysis reaction occurring. Reaction in a C 2 -unit mode (ethane the archetypal hydrocarbon) occurred with an activation energy in the region of 175 kJ mol −1 , and was the reaction mode for C I -C I , C I -C II , and C II -C II bond types. Reaction in an iso-unit mode (neopentane the archetypal hydrocarbon) occurred with an activation energy in the region of 235 kJ mol −1 , and was the reaction mode for C I -C IV , C I -C III , and C II -C III bond types. (C I indicates a primary carbon, etc.) The activation energy and the frequency factor (expressed as rate per surface iridium atom) for reaction in the C 2 -unit or the iso-unit mode was independent of d Ir . Provided the hydrogen pressure was kept sufficiently high, all reaction rates were proportional to P 1 HC P −3 H 2 . The reaction mechanism is discussed and is compared with the behavior of the corresponding reactions over platinum catalysts.

On the reduction of supported iron catalysts studied by Mössbauer spectroscopy

Abstract The reducibility of ferrous ions exchanged onto the surface of zeolite Y, amorphous aluminosilicate, and silica gel has been studied by Mossbauer spectroscopy. The iron content in these catalysts ranged from 0.17 to 13.8 wt%. Reduction was in general carried out in about 10–20 kPa of hydrogen at 673 K for several hours. Ferric ions were also introduced, but at the expense of some loss of zeolite crystallinity. Although such ferric ions could be reduced to ferrous, further reduction to metallic iron was, except possibly in the special circumstances indicated below, not observed. This lack of reducibility included a Fe 2+ -Y zeolite sample in which 0.8 wt% of platinum had been incorporated by ion-exchange. By heating in oxygen at 673 K, ferrous ions were oxidized to the ferric state; the oxidation-reduction process was reversible. Under similar reducing conditions, bulk iron oxides could easily be reduced to metallic α-iron. When an amorphous Fe 3+ -Y sample was heated in air at 1273 K, the Mossbauer spectrum showed, in addition to a central doublet, a six-line pattern due to α-Fe 2 O 3 . The magnetic hyperfine pattern due to α-Fe was, however, not observed when this sample was reduced: it is concluded that the iron was probably in the metallic state with particle size smaller than the ferromagnetic domain size. Thermodynamic considerations suggest that while both bulk iron oxides and nickel ions in zeolites can be reduced to the metallic state by hydrogen at 673 K, much severer conditions would be needed to reduce ferrous ions on the catalyst support, in agreement with observation.

Adsorption and temperature-programmed desorption of hydrogen with dispersed platinum and platinum-gold catalysts

Abstract Hydrogen adsorption has been studied by temperature-programmed desorption (TPD) and by static volumetric measurement (SVM) on a range of supported dispersed platinum and platinum-gold catalysts. For compositions in the range Pt 98, Au 2 to Pt 15, Au 85 mole%, a comparison of hydrogen adsorption data with predictions from surface enrichment theory combined with electron microscopic particle size measurement, led to the conclusion that after suitable thermal treatment, equilibrium surface enrichment by gold was achieved, although equilibrium was most difficult to achieve for the Pt 15, Au 85 composition. It was concluded from the nature of the hydrogen adsorption isotherms that, at higher pressures, some hydrogen could be adsorbed on surface gold atoms, probably as a result of spillover from the platinum component. The general shape of the TPD profiles on platinum-gold was independent of gold content; to a first approximation the gold acted as an inert diluent. The implications of this result for the mode of hydrogen chemisorptive bonding are discussed. With gold-free platinum catalysts there was a trend towards an extension of the high temperature tail of the TPD profile with decreasing d Pt in the range ⩽4 nm, due to the presence of hydrogen binding states of higher energy. Possible reasons for this are discussed. Illustrations are given of the technique of TPD-profile dissection using increased starting temperatures. A TPD peak with a maximum in the region 700–750 K has been assigned to desorption from the support, this binding state being populated via hydrogen spillover from the platinum.

The mechanism of isomerization of aliphatic hydrocarbons at a platinum surface

Abstract The present paper discusses a process by which C 4 and C 5 alkane isomerization may occur at a platinum surface. The model proposes that the surface precursor is 1,3-diadsorbed with a double bond to the surface at one carbon atom [e.g., structure (C), Fig. 2], and that isomerization occurs by the transformation of this to a bridged structure [e.g., structure (D), Fig. 2]. The likelihood of isomerization is discussed in terms of the energy difference between the precursor and the corresponding bridged intermediate. The energies were calculated by a simple Huckel MO method. The effect of methyl group hyperconjugation on facilitating the isomerization is considered explicitly, while the effect of differing electronegativities is discussed qualitatively. It is shown that the energy criterion for isomerization is favorable where there is partial electron transfer from the hydrocarbon residue to the surface metal atom, and this remains possible for finite values of overlap in forming the double bond with the surface metal atom. The model provides a more detailed interpretation for the results obtained previously using C 13 -labeled n -butane in the isomerization reaction.

Reactions of methyl chloride and of methylene chloride at metal surfaces: II. Reactions over evaporated films of titanium and other metals

Abstract The reaction of methyl chloride and hydrogen, and of methylene chloride and hydrogen has been studied at the surface of a number of evaporated metal films in a static system. Adsorption of alkyl chloride resulted in rupture of all carbon-chlorine bonds before any carbon-hydrogen bonds were broken; thus in the reaction with deuterium, virtually no deuteroparent was formed. By the use of 13 C and 35 Cl it was shown that in the adsorption of methyl chloride on titanium, carbon-chlorine bond rupture was totally irreversible. Over nickel, tungsten, copper, platinum, cobalt, manganese, aluminum and silver, methane was the only hydrocarbon product. Over palladium and titanium higher hydrocarbons were formed. In the reaction of methylene chloride over titanium, polymer was also formed, and olefins were the main C 2 and C 3 products. The reaction of methyl chloride on titanium was shown to be first order in both methyl chloride and hydrogen pressures. Adsorption measurements with the alkyl chlorides on the above metals gave surface hydrogen/carbon ratios which, in most cases, suggested the average surface hydrocarbon residue to be more extensively dehydrogenated than the group in the original alkyl chloride: the extent of dehydrogenation increased with increasing temperature. The exception to this was titanium where adsorption of methyl chloride and methylene chloride gave approximately CH 3 and CH 2 surfaces residues, respectively. From the influence of hydrogen on the reaction product distribution, and from the results from reactions with deuterium it has been concluded that on palladium, chemisorbed hydrogen was directly involved in the process of product desorption, but on titanium this was not so, and inter-residue hydrogen transfer occurred in desorption. Evidence has been adduced that carbon-carbon bond formation probably occurred by reaction of an alkyl chloride molecule with a surface CH 2 group. Activation energies and frequency factors have been evaluated for the reactions of both alkyl chlorides on titanium.