Search

Further reactions of ethylene oxide over oxide supports have been investigated in the context of the oxidation of ethylene to ethylene oxide using supported silver. As noted by others, high area aerosil type silicas have low activity for the isomerisation of ethylene oxide to acetaldehyde and silver supported on such materials exhibits good selectivity for ethylene epoxidation. By contrast silver supported on TiO2 and γ-Al2O3 shows zero selectivity and this correlates with the high activity of these supports for further reactions of ethylene oxide. In the case of γ-Al2O3 there is complete conversion of ethylene oxide at 239°C in the presence or absence of oxygen. Acetaldehyde is the major organic product in both cases but dioxane and small amounts of 2,3-butanedione, ethyl acetate and an unidentified compound, probably an unstable C4, are also observed. With oxygen present some carbon oxides are seen. In the absence of oxygen the amount of the unknown is greater and the gaseous mass balance is incomplete. With TiO2 the distribution amongst the major products is similar to that for γ-Al2O3 but the conversion of ethylene oxide is much less. In situ infrared measurements with TiO2 show that ethylene oxide reacts instantaneously with surface hydroxyl groups to form a glycol like species. Reaction of acetaldehyde over TiO2 and γ-Al2O3occurs at a much slower rate than that of ethylene oxide under the same conditions. The dominant product is 2-butenal. Thus the C4 products formed when using ethylene oxide do not arise by further reaction of acetaldehyde. Formation of the major organics from both ethylene oxide and acetaldehyde can be accounted for in terms of catalysis by weak Bronsted acid sites. Silicas have little activity for the further reaction of acetaldehyde but silver supported on them induce oxidation of it. It is concluded that the zero selectivity of silver supported on TiO2 and γ-Al2O3 for ethylene epoxidation is due to isomerisation of ethylene oxide to acetaldehyde on the support followed by complete oxidation on the silver.

The characteristics of the oxidative dehydrogenation of ethane and the oxidative coupling of methane have been compared over four high purity rare earth oxides--La 2 O 3 , CeO 2 , Sm 2 O 3 and Pr 6 O 11 . With each oxide ethane reacts approximately four times as fast as methane and the activity order per g is Pr 6 O 11 〉Sm 2 O 3 〉La 2 O 3 〉CeO 2 . In terms of unit area is La 2 O 3 〉Sm 2 O 3 ≥Pr 6 O 11 ≫CeO 2 . The selectivity to hydrocarbons increases with temperature for both reactions but the absolute values are not only much higher for ethane than for methane but also much less catalyst dependent. For ethane at 750°C the order in selectivity to ethylene is La 2 O 3 (74%)〉Sm 2 O 3 (68%)〉CeO 2 (57%)P˛r 6 O 11 (53%). Under identical conditions, (13% hydrocarbon, 3.5% oxygen, balance helium) the selectivity of the oxidative coupling of methane to ethane plus ethylene together is about 40% for La 2 O 3 and 33% for Sm 2 O 3 but less than 3% for CeO 2 and Pr 6 O 11 , the oxides with variable valence state. This selectivity pattern can be explained in terms of two interacting factors, the average lifetime of alkyl and alkylperoxy species in the gas phase and the tendency of the oxides to oxidise alkyl radicals. The higher oxidation power of CeO 2 and Pr 6 O 11 relative to Sm 2 O 3 and La 2 O 3 , also manifests itself in a low yield of carbon monoxide relative to carbon dioxide and a lesser ratio of hydrogen to water. The water-gas shift reaction is not at equilibrium for the former pair of oxides. The yield of minor products, propene during methane coupling and methane and butene during ethane oxidation, increases with temperature. The more selective oxides, La 2 O 3 and Sm 2 O 3 , also show a small maximum in methane production at about 650°C which may be due to carbon-carbon bond cleavage in an adsorbed ethoxy species.

The selective hydrogenation of propyne over copper powder and various supported copper catalysts has been examined in a dynamic flow system. Increased hydrogenation activities were found for supported copper catalysts, where for example, the turnover frequency over an ion-exchanged Cu/SiO2 catalyst was 3.5·10-2 s-1 at 140°C as compared to 1.1·10-3 s-1 over unsupported copper powder. The increased selectivity to propene with increasing temperature was found to be a general property of all copper catalysts. However, the exact pattern of activity and selectivity was found to depend upon the particular supported copper system. Deactivation was found to depend on the acidity of the support and the dispersion of copper. Overall, the most active and selective catalysts that deactivate least were copper on silica prepared by the ion-exchange technique and copper on magnesia.

The selective hydrogenation of propyne over a Cu/SiO 2 catalyst in a dynamic flow system has been investigated. It was found that only propene and, depending on the reaction conditions, different amounts of oligomeric material, referred to as green oil or foulant, were formed. However, the selectivity to propene increased significantly at higher temperatures and it was possible to convert all the propyne selectively to propene, leaving a great potential for further development and industrial applications. Deactivation of the catalyst was found to be only a minor problem as 100% conversion and a selectivity of > 99.5% was obtained during a long-term test with an industrial raw propene feed (2 mol-%propyne/propadiene), even after one month on line. It is interesting to note that even at high propene concentrations and high temperatures no propane was detected. The conversion of propyne increased with temperature and reached 100% above 160°C. Between 90 and 150°C the selectivity was constant at about 20%, irrespective of hydrogen and propyne concentrations, the degree of conversion, the total flow-rate or the temperature. The addition of propene to the initial mixture decreased the conversion slightly but increased the selectivity markedly. It was possible to obtain 100% conversion of propyne to propene with nearly 100% selectivity at higher temperatures. The kinetic data observed showed a reaction order of zero for propyne at 100°C and a slightly positive reaction order (0.15) at 130°C. The reaction order was one for hydrogen irrespective of temperature. Propene decreased the activity slightly, corresponding to a reaction order of -0.16. The activation energy (68.6kJ/mol±5kJ/mol) was similar for the formation of propene and green oil between 90 and 150°C.

The epoxidation of ethylene over silver catalysts in the presence of added carbon monoxide and hydrogen has been studied as a possible method for directly using ethylene in streams generated by partial oxidation of methane. When tested separately over a silver sponge catalyst the oxidation of both carbon monoxide and hydrogen are more than one order of magnitude faster than that of ethylene. Measured activation energies are 44, 43 and 61 kJ/mol respectively. Under conditions of moderate ethylene conversion (ca. 40% at 239°C) the effect of including carbon monoxide or hydrogen in amounts up to 20% of the ethylene is quite small. The added material is totally oxidised in the initial section of the catalyst bed leaving the remainder available for ethylene epoxidation. Neither ethylene conversion nor ethylene oxide yield is affected significantly. The situation is somewhat different at lower conversions over chlorine moderated catalysts. Added carbon monoxide is then incompletely oxidised and the fraction emerging inhibits ethylene oxidation by direct competition for the silver surface. Conversion is reduced and the selectivity to ethylene oxide is also reduced under ethylene-rich conditions. Under similar conditions hydrogen is also incompletely oxidised but it inhibits less. A particular effect was observed when carbon monoxide was included in the feed for long periods (>10 h). Ethylene conversion fell steadily while the selectivity to ethylene oxide rose. The effect was very similar to that noted when experiments were commenced with 1,2-dichloroethane at the ppm level in the feed. It is attributed to small quantities of an impurity in the carbon monoxide since it disappeared if that stream was purified.

The effect of ethanol and dioxane as solvents for water on the hydrolysis of acrylonitrile to acrylamide in the presence of Raney copper catalysts has been investigated. Both solvents lower the rate but the reduction is somewhat greater for dioxane. The magnitude of the effect is best explained in terms of a Langmuir-Hinshelwood model in which both acrylonitrile and water are adsorbed. The added solvent inhibits reaction through competitive adsorption thus reducing the surface concentration of the two reactants. Substitution of2H2O changes the hydrolysis rate only slightly leading to the suggestion that the slow step is addition of water to acrylonitrile adsorbed by partial opening of the C–N bond. The hydroxyimine species so formed then yields acrylonitrile by rearrangement.

The influence of supports of copper catalysts produced by the ion-exchange technique on SiO 2 , Al 2 O 3 and TiO 2 has been studied for the hydrolysis of acrylonitrile to acrylamide. Cu/SiO 2 has been found to be the best catalyst, deactivating less rapidly than Cu/Al 2 O 3 . Metal-support effects were observed in Cu/TiO 2 catalysts, both for carbon monoxide adsorption and acrylonitrile hydrolysis. The best Cu/TiO 2 catalyst was less active than Cu/SiO 2 .

Studies have been made of the impregnation process used to prepare catalysts suspended on pellets of γ-alumina. Attention has been focused on the impregnation of tungsten salts, alone or in combination with nickel or cobalt. The profile of the metals distribution within the pellet has been found to be very dependent on the conditions of impregnation. Impregnation was carried out from excess solution containing known concentrations of metal salt at controlled pH. The amount of tungsten deposited on the catalyst was found to be highest at pH > 6, while the amount of cobalt and nickel deposited increased as pH increased from 6 to 8 at which point bulk hydroxides precipitated on the pellet surface. The penetration of the metal salts into the pellet was found to be a function of pH, of concentration and of soak time, with clear evidence emerging that pore blockage was limiting penetration under some conditions. The results were explained in terms of the species present in solution under different conditions and their diffusion in the pore system versus adsorption on the pore wall. Conditions were established for the preparation of catalysts with metals concentrated near the pellet exterior or spread evenly throughout the pellet.

The preparation of hydrotreating catalysts by the impregnation of alumina with solutions of heteropolyanions has been studied. These anions have the general structure [HwAxByOz]n−, where A may be Co or Ni and B may be Mo or W. The anions have advantage over conventional impregnating agents in that both catalytically active metals are in the same complex and should be deposited in the same position in the pellet. In addition, the solubility of the anion is higher than conventional impregnating agents. This allows higher concentrations of catalytic metal (and particularly of tungsten) to be deposited on or in the support. Conditions for the preparation of “shell” or “uniform” impregnated catalysts have been established. Comparisons have been made of the efficiency of such catalysts and of conventional catalysts for the hydrotreating of a model feed-stock. Heteropolyanion-based catalysts are very effective for hydrodesulphurization and hydrogenation but their relatively low content of promoter metals restricts hydrodenitrogenation activity.

Copper orl silica catalysts have been prepared using ion-exchange of [Cu(NH3)4]]2+ with silanol sites under basic conditions (normally pH 11). Equilibrium saturation was reached within minutes. Upon washing in distilled water or dilute ammonia at pH 7–9, dissolved Cu2+-ions, present in the interstitial particle volumes, were precipitated as Cu(OH)2 and retained in the catalyst. The solid then contained two different copper species, i.e., ion-exchanged Cu2+ and precipitated Cu(OH)2, the proportions of which varied depending on the preparation conditions. The relative amounts of ion-exchanged copper were high if preparation was based on low Cu2+ concentrations in the preparation solutions and extensive washing in dilute ammonia at pH 11 (copper loadings below 1 wt%). Copper contents between 2 and 6 wt%, obtained by impregnation at pH 11 and washing in distilled water, resulted in the presence of ion exchanged and bulk copper species. At higher loadings the ion exchanged species were a negligible fraction of the exposed copper, possibly due to coverage by bulk copper deposits. The precipitated species is concentrated in very small clusters (

Methanol may be produced by carbonylating alcohols in the presence of alkali metal to produce alkyl formates which are then hydrogenated to produce the original alcohol and methanol. As part of a study of methanol production by this route, the present investigation has been concerned with the hydrogenolysis of alkyl formates over copper chromite catalysts. Hydrogenolysis of a series of alkyl formates has been found to proceed at moderate temperatures and pressures but the methanol produced reacts with the original alkyl formate to produce methyl formate and the original alcohol. The latter transesterification reaction exhibits a rapid equilibrium, the position of which is little influenced by temperature. Preliminary studies of the effect of formate structure on hydrogenolysis and transesterification reactions have been carried out.

The deactivation of a copper on silica catalyst prepared by the ion-exchange technique for the hydrolysis of acrylonitrile (AN) to acrylamide (AA) has been studied at temperatures below 350 K. Initial deactivation is attributed to a polymer network coating the outer particle surface and originating from the product. Due to the nature of the polymer and the fine catalyst support pore structure, a major part of the active copper sites remains accessible and the catalyst retains its activity over a long time and broad range of reactant and product concentrations, but undergoes very slow and constant deactivation by copper oxidation from oxygen dissolved in the reactant mixture. The level of almost constant activity is distinctly higher than for other copper-based catalysts and is described by the power law rate equation r =1.80 exp (−5910/T) C 0.80 AN C −0.12 AA

Copper chromite catalysts have been characterised by total and metal surface areas. X-ray powder diffraction, thermal gravimetric analysis and selective extraction of copper oxide in order to investigate the nature of the active catalyst for methanol dehydrogenation. After reduction in hydrogen at atmospheric pressure, the catalyst surface has been shown to consist essentially of copper metal crystallites supported on cuprous chromite. The mechanism of catalyst reduction has been identified.

Aluminas, intermediate in the transformation of γ to α-phase, have been tested as supports for vanadia in the oxidation of benzene to maleic anhydride. Treatment of γ-alumina in air or steam results in an initial rapid loss of half the original surface area corresponding to closing off of small pores with no development of bulk α-phase. However,27Al NMR measurements, and determinations of surface Lewis acidity using the dehydration of methanol to dimethyl ether as a test reaction, each show that conversion to the α arrangement is well advanced at the local and surface levels. Further reduction in area parallels formation of crystalline α-phase. Aluminas with surface areas of 20 to 40 m2 g−1 contain 30–60% α-phase and have very low surface acidity. Catalyst comprising 9% V2O5 on such supports are able to convert benzene to maleic anhydride with selectivities of up to 83%, in contrast to zero selectivity for comparison catalysts using γ-alumina supports. The activity and selectivity patterns with the partially transformed aluminas as supports can be rationalised in terms of an optimal average number of V2O5 layers of three to four. With less than this, some support is exposed and selectivity and activity per unit area of catalyst is degraded. Thicker layers lead to lower than optimal activity per unit weight of V2O5 and some loss of the favourable effect of an α-alumina surface on the selectivity of overlying V2O5 layers.

The liquid phase hydrogenolysis of methyl formate has been studied in a three phase slurry reactor using a Girdler G-89 copper chromite catalyst. Measurements at 446 K show a very high selectivity to methanol. The reaction is first order in methyl formate, first order in hydrogen for pressures of 1.7 to 4.5 MPa and has an apparent activation energy of 62 kJ mol−1. Unlike the corresponding vapour phase process at atmospheric pressure, the liquid phase reaction shows no evidence of catalyst deactivation at 450 K and to conversions beyond 70%. However, inhibition, which is at least partially reversible, becomes apparent at 473 K when carbon monoxide is detectable as a by-product. Inhibition can be observed at lower temperatures if carbon monoxide is deliberately introduced at pressures above 0.1 MPa. The apparent reaction order in carbon monoxide is then −0.32.