Platinum Metals Rev., 1985, 29, (2), 63
Ruthenium Melt Catalysis
Producing Chemicals from Synthesis Gas
A range of commodity chemicals and fuels can be generated directly from synthesis gas by ruthenium melt catalysis. Careful selection of the catalyst components enables a wide range of oxygenates to be produced, while other advantages of this flexible process include high productivity and the ease with which the products may be recovered and the catalyst recycled.
In platinum-metal catalysed organic synthesis, the primary purpose is to produce pure products in high yields, with good reproducibility and ease of product isolation. While both homogeneous and heterogeneous catalysis systems may fulfill these roles in laboratory or industrial processes, an alternative approach, that has many of the inherent advantages of both systems, is the use of molten salts as catalyst and/or reaction media.
Molten salts may be classified into simple ionic salts (for example the alkali and alkaline earth metal halides), simple and polymeric oxyanionic salts (such as metal nitrates, carbonates, phosphates, borates), and salts containing organic cations and/or anions (including the quaternary phosphonium and ammonium salts) (1). The high thermal stability of many of these melts, together with their low vapour pressure, good thermal and electrical conductivity, and low viscosity provides them with a range of important physical and chemical advantages over conventional homogeneous and heterogeneous catalyst systems (2). Molten salts, such as the quaternary phosphonium salts, dissolve a broad range of metals, oxides and complexes; high reaction rates and yields are often possible (for example organic syntheses in molten tetrachloroaluminate solvents (3) and fused alkali metal salts (4)) with ease of product/catalyst separation and recovery. Where chemical synthesis occurs at a melt-gas interface, as in gas reactions, the ability to present a renewable surface may avoid many of the problems of deactivation seen with conventional heterogeneous catalysts (4).
Frequently, the melt is called upon to play more than one role in a chemical process. It may, for example, simply act as a solvent for reactants and products; it may be a catalyst; it may supply one or more of the reactants, as well as aid in controlling the heat of reaction. The salts may be used singly (for example molten Na2CO3, ZnCl2 or AlCl3), or as eutectic mixtures (such as Li2CO3-Na2CO3-K2CO3, CuCl-CuCl2-KCl, and AlCl3-BuPyCl) where the lower liquidus temperatures may be advantageous.
Molten salts have already found numerous technical applications in industrial catalytic processes. Examples include the Transcat, Deacon and SO2 oxidation processes (1,2). The literature illustrating the diverse uses of melt catalysis with the platinum group metals includes applications in chlorination (2), oxidation (5) and selective hydrogenation (6) processing, as in Equation i:
In the case of synthesis gas (carbon monoxide/hydrogen) chemistry, Fischer-Tropsch production using Ir4(CO)12 dissolved in NaCl-AlCl3 melts has now been demonstrated (7); ethane is the major product fraction, see Equation ii or, under flow conditions, isobutane and propane. A further examination, by Collman and his colleagues, implicates chloromethane (or methanol) as the primary intermediate, with methane, ethane, and chloromethane the major carbon-containing products (8).
Ruthenium Melt Catalysis
In other, potentially far-reaching applications of melt catalysis by the platinum group metals, we at Texaco have demonstrated the synthesis of a range of commodity chemicals and fuels directly from synthesis gas by the use of ruthenium-containing molten salt catalysis. Products include ethylene glycol, C1−C4 alcohols, acetic acid, acetate esters, C2+ olefins and vicinal glycol esters. In its simplest form, this new class of melt catalyst comprises one or more ruthenium sources, including ruthenium carbonyls, oxides and complexes dispersed in a low-melting (m.p. <150°C) quarternary phosphonium or ammonium salt, for example tetrabutylphosphonium bromide. The key components are selected such that:
The catalyst mix is a solid under ambient conditions.
The quarternary salt melts below the temperatures of carbon monoxide hydrogenation so that, under typical operating conditions, it is a highly polar, non-volatile, liquid media that readily solubilises the ruthenium catalyst precursor and the synthesis gas components, thereby ensuring rapid rates of carbon monoxide hydrogenation to the desired aliphatic oxygenates.
Optionally, upon cooling, the ruthenium-quarternary salt may resolidify, allowing ease of separation of the liquid organic products from the solid inorganic catalyst components by decantation or distillation.
The melt catalyst enjoys, thereby, certain of the intrinsic advantages of both homogeneous and heterogeneous systems; the inherent high selectivity and reproducibility of liquid homogeneous catalysts under normal carbon monoxide hydrogenation conditions, and the ease of product-catalyst separation, typical of heterogeneous catalyst systems, once the reaction is complete.
Optionally, a second or third transition metal may be added to the ruthenium to further modify its performance and to steer the product distribution toward specific aliphatic oxygenates. These modifiers may be derivatives of rhodium, cobalt, manganese, rhenium, zirconium and titanium, in either halogen-free or halogen-containing forms.
Ethylene glycol synthesis directly from synthesis gas using ruthenium-rhodium catalysts or ruthenium alone (9).
C1−C4 Alcohol/acetate ester fuel products, which are potential octane enhancers, using ruthenium in combination with cobalt, titanium or zirconium (10).
Effect of varying rhodium, ○;
Effect of varying ruthenium ×
Acetic acid production directly from synthesis gas catalysed by halogen-promoted ruthenium-cobalt combinations (11).
Through the addition of aliphatic carboxylic acid coreactants, this chemistry may be further extended to:
Ethylene glycol commands a 4.5 × 109 lb/yr market in the United States of America, and while this large-volume petrochemical is presently generated entirely by ethylene oxide hydrolysis, there has been considerable research activity during the past decade into alternate routes to polyhydric alcohols either directly, or indirectly, from synthesis gas. We have demonstrated the preparation of ethylene glycol and its monoalkyl ether derivatives from synthesis gas (Equation iii) using ruthenium or a ruthenium-rhodium catalyst combination dispersed in a low-melting quaternary phosphonium salt such as tetrabutylphosphonium bromide. The advantages of the mixed ruthenium-rhodium system include improved glycol productivity and ability to operate at lower synthesis gas pressures. It may be noted that (see Table I):
The productivity of the ruthenium-rhodium melt catalyst is high; liquid weight gains may exceed 200 weight per cent (expt. 2) and turnover frequencies surpass 5 × 10−3/s at 220°C.
Glycol:alkanol weight ratios of 1:1.37 have been realised (expt. 6).
Both alkanol and diol products may be readily isolated by fractional distillation of the crude liquid product and the solid residual ruthenium-rhodium catalyst recycled.
|Expt.||Ruthenium-rhodium source||Quaternary salt||Melting point,b °C||Product yield, mmole||Total liquid yield, weight per centd|
Charge: ruthenium 4.0 mmole; rhodium 2.0 mmole; R4PX 15g; run conditions: 220°C; 430 atm constant pressure; CO/H2(1 : 1); 6−18 hr.
The primary requirements in selecting the quaternary Group VB salts (Table I) as suitable reaction media are that they be thermally stable under typical carbon monoxide hydrogenation conditions and that they melt below the reaction temperature (200 to 250°C) to give stable, highly polar, fluid media for solubilisation of the transition-metal component.
Figure 1 illustrates how both the ruthenium and rhodium catalyst components are beneficial to the formation of ethylene glycol, expressed here in terms of total glycol + glycol ether productivity, for the Ru(acac)3-Rh(acac)3/Bu4PBr catalyst precursor. Maximum glycol production is achieved with increasing rhodium addition, up to rhodium: ruthenium ratios of about 1:1. Very little diol is generated, however, in the absence of either the ruthenium component (see Figure 1), or the quaternary phosphonium salt (Table I, expt. 5). While the presence of ruthenium is necessary to the formation of glycol, the data of Figure 1, together with the isolation of mixed ruthenium-rhodium clusters such as Ru2Rh(CO)12 (9), point to the role of the second metal, rhodium, in providing improved diol selectivity.
A second class of platinum metal melt catalysts—comprising a ruthenium source, such as Ru3(CO)12, optionally in combination with a halogen-free cobalt, titanium, zirconium, manganese or rhenium component, dispersed in a quaternary phosphonium salt—has been found to be effective for the selective synthesis of (a) methanol, (b) ethanol and (c) their acetate esters directly from synthesis gas (see Table II). These liquid products are intended for use as octane enhancers for gasoline. The shift to unleaded and low-leaded gasoline in both the United States and Europe will likely raise the clear octane requirements by about two octane numbers. Although some of this increase could be satisfied by higher severity processing, octane improvers, particularly liquid aliphatic oxygenates, are becoming an increasingly attractive alternative.
|Expt.||Catalyst precursor||Quaternary salt||Melting point °C||Liquid product composition, weight per centc||Liquid yield, per cent|
Typical operating conditions: 220°C; 430 atm constant pressure; CO/H2 (1 : 1); catalyst charge: ruthenium 2–8 mmole, quaternary salt, 10−25 g
Here it may be seen from an inspection of the summary data in Table II that:
Ethanol may constitute near 60 per cent of the crude liquid product with 76 per cent being C1–C3 alcohol, when employing the RuO2-Bu4PI couple (expt. 7). The ruthenium/halide-free titanium/Bu4PBr combination likewise produces C1–C3 alcohols in 61 per cent selectivity of which up to 66 per cent is ethanol (expts. 10 and 11).
A balanced mix of C1–C3 alcohol and acetate esters may be realised in high yields using a variety of ruthenium-cobalt bimetallic catalysts as illustrated here by expts. 8 and 9.
Methanol will be the predominant product fraction when the melt comprises ruthenium plus a halogen-free manganese or rhenium component. With RuO2-Re2(CO)10/Bu4PBr (expt. 12), methanol selectivity in the liquid product is 85 per cent and the turnover frequency is 24 mol MeOH/g atom ruthenium/hr.
The addition of certain solvents may lead to further changes in product composition. The presence of added p-dioxane, for example (expt. 14), makes ethyl acetate the predominant product, with alkyl acetates providing 67 per cent of the organic liquid fraction.
Numerous factors must, of course, be considered in the selection of organic oxygenates as blending components in gasoline, particularly cost, octane blending value, volatility, water and gasoline compatibility, raw material availability, engine performance and corrosion, as well as environmental problems (10). In our work, we have focused on two C1–C3 alcohol/ester blends produced by carbon monoxide hydrogenation (Equation iv) using the Texaco melt technology. The first blend (see Table III, product mix “A”), rich in methanol and ethanol, is similar in composition to those prepared by the procedure of Table II, expt. 7. The second blend (“B”) contains sizeable quantities of methyl and ethyl acetates, and more closely resembles the products of Table II, expt. 9, generated using a ruthenium-cobalt couple. Both blends are, of course, the products of numerous runs under standard operating conditions; they have been isolated with the minimum of fractionation.
|Component||Volume per cent|
|Product mix “A” C1-C3 alcohols||Product mix “B” C1-C3 alcohol/esters|
Preliminary performance data for the two blends indicates:
The octane improvement properties of the C1–C3 alcohol mix, “A”, appear to be good, and on this basis it may show promise as a future octane blending agent. The improvements in motor octane number and research octane number values, about 1.4 and 3.4, respectively, for mix “A”, are noticeably higher than those for composite alcohol-ester mix, “B”, but the increase in volatility is greater for the alcohol/ester gasoline blend (14).
The oxidative stability and gum-forming tendencies of the gasoline-containing alcohol/ester mix appear poorer than for gasoline containing mix “A”. Stability of mix “A” is equivalent to that of gasoline (14).
The alcohol composition in gasoline exhibits superior corrosion properties, when compared to “B” blend, for each of the metals commonly used in automotive engine surfaces.
Acetic acid is another commodity chemical with a large market requirement, of about 2.9 × 109 lb/yr in the U.S., that has been targeted as being potentially made directly from synthesis gas (Equation vi) (15), rather than by methanol carbonylation or through oxidation chemistry. We have established that acetic acid may be generated directly from synthesis gas as the predominant oxygenate product in 68 weight per cent selectivity (11) using Ru3(CO)12–Co2(CO)8/Bu4PBr melt catalyst, preferentially where the cobalt: ruthenium and carbon monoxide: hydrogen ratios exceed unity (for example see Figure 2).
Short-Chain Fatty Acids
Platinum group metals dispersed as melt catalysts have also been utilised by Texaco in a novel process for making short-chain (C5-C10) fatty acids (SFAs). Now generated from coconut oil and through oxo chemistry, these SFAs find expanding applications in the synthetic lubricant and plasticiser markets.
We have demonstrated that ligand-stabilised palladium and platinum salts dispersed in quaternary Group VB salts of trichlorostannate(II) are effective catalysts for SFA manufacture by the regioselective carbonylation of α -olefins (Eq. ix). Preferred catalyst formulations include bis(triphenylphosphine)-palladium(II) chloride and its para-tolyl analogue, PdCl2[(p-CH3.X6H4)3]2, dispersed in tetraethylammonium trichlorostannate(II). In the case of nonanoate ester synthesis, yields consistently exceed 80 mole per cent for both systems (16), and carboxylic acid linearity is greater than 80 per cent. A variety of alkene and alkyne structures may be carbonylated in the presence of different alkanol coreactants (see Table IV).
|Expt.||Alkene/alkyne||Alkanol||Conversion mole per cent||Identity||Selectivity, per centb||Yield, mole per cent|
|22||4-Methylcyclohexene||Ethanol||80||Ethyl 4-methylcyclohexane carboxylated||80.0||42|
The palladium catalyst, PdCl2[P(C6H5)3]2-10[(C2H5)4NISnCl3], exists under carbonylation conditions (Table IV) as a clear, homogeneous, phase with the α -olefin and alkanol coreactant. However, this particular quaternary salt melts at 78°C, so, following carbonylation, the reactant mixture may be cooled to reprecipitate the inorganic catalyst components, thereby once again ensuring ease of separation of the crude product acid esters from the now resolidified palladium catalyst.
Intrinsic advantages of this case of carbonylation catalyst over its solvent-solubilised counterparts (17) would be in:
The ease of product acid ester separation.
Improved stability of the palladium catalyst during multicycling.
Potentially higher concentrations of active palladium within the reaction mix.
The long-term life of these palladium-dispersed molten salt catalysts, particularly PdCl2[P(C6H5)3]2-[(C2H5)4NISnCl3], has in fact been confirmed in multicycling experiments, with regeneration of spent palladium catalyst (particularly reoxidation of inactive Pd(O) species) by treatments with mineral acids, peroxides, chlorination and oxychlorination techniques. Chlorination in the presence of added chlorinated solvent (16), in particular, does allow multicycling of the palladium melt catalyst, over 30 to 40 cycles, with little overall loss in specific activity.
Clearly the myriad potential applications of the platinum group metals in melt catalysis has only just begun to be explored. The development of new classes of catalysts to produce aliphatic oxygenates from synthesis gas, beyond the generation of C1-C4 alkanols/diols/acids to more complex molecules, and their use with N-, P- and S-containing coreactants, are challenges for the future. The extension of this technology to oxidation, halogenation and refinery processes, particularly hydrocracking and isomerisation, has yet to be fully exploited.
The commercialisation of the novel syngas-to-aliphatic oxygenate processes outlined in this article will hinge primarily upon future economic projections, particularly the relative feedstock values of petroleum, versus coal- or natural gas-derived synthesis gas.
- 1 B. W. Hatt and D. H. Kerridge, Chem. Br., 1979, 15, (2), 78
- 2 C.N. Kenney, Catal.Rev.Sci.Eng., 1975, 11, 197
- 3 H. L. Jones, and R. A. Osteryoung,, “ Advances in Molten Salt Chemistry, ” Vol. 3, ed. J. Braunstein,, G. Mamantov, and G. P. Smith, Plenum Press, New York, 1975, Chapter 3
- 4 J. E. Gordon, “ Techniques and Methods of Organic and Organometallic Chemistry, ” Vol. 1, ed. D. B. Denney, Marcel Dekker, New York, 1969, Chapter 3
- 5 P.R. Rony, Ann. N.Y. Acad.Sci., 1970, 172, 238
- 6 G.W. Parshall, J.Am.Chem.Soc., 1972, 94, 8716
- 7 G. C. Demitras and E. L. Muetterties, J. Am. Chem. Soc., 1977, 99, (8), 2796
- 8 J. P. Collman,, J. I. Brauman,, G. Tustin, and G. S. Wann, J. Am. Chem. Soc., 1983, 105, 3913
- 9 J. F. Knifton, J. Am. Chem. Soc., 1981, 103, (13), 3959 ; J. Chem. Soc., Chem. Commun., 1983, 729
- 10 J. F. Knifton,, R. A. Grigsby, and S. Herbstman, Hydrocarbon Process., 1984, 63, (1), 111
- 11 J. F. Knifton,, R. A. Grigsby, and J. J. Lin, Organometallics, 1984, 3, (1), 62
- 12 J. F. Knifton, J. Catal., 1982, 76, (1), 101 ; J. Chem. Soc., Chem. Commun., 1981, (4), 188
- 13 J. F. Knifton, J. Catal., 1983, 79, (1), 147
- 14 J. F. Knifton,, R. A. Grigsby,, and S. Herbstman, in “ Catalytic Conversions of Synthesis Gas and Alcohols to Chemicals, ” ed. R. H. Herman, Plenum Press, New York, 1984, p. 81
- 15 J. Haggin, Chem. Eng. News, 1981, 59, (23), 23 ; 1981, 59, (46), 57
- 16 J. F. Knifton, J. Am. Oil Chem. Soc., 1978, 55, (5), 496
- 17 J. F. Knifton, J. Org. Chem., 1976, 41, (5), 793 ; J. Org. Chem., 1976, 41, (17), 2885
- 18 “ Fatty Acids and Derivatives, ” Chem Systems Report#78–2 (January 1979); A. G. Johanson, J. Am. Oil Chem. Soc., 1977, 54, (11), 848A