The role of electrophilic species in the Fischer-Tropsch reaction.

The heterogeneously catalysed Fischer-Tropsch (FT) synthesis converts syngas (CO+H2) into long chain hydrocarbons and is a key step in the economically important transformation of natural gas, coal, or biomass into liquid fuels, such as diesel. Catalyst surface studies indicate that the FT reaction starts when CO is activated at imperfections on the surfaces of late transition metals (Fe, Ru, Co, or Rh) and at interfaces with "islands" of promoters (Lewis acid oxides such as alumina or titania). Activation involves CO cleavage to generate a surface carbide, C(ad), which is sequentially hydrogenated to CHx(ad) species (x=1-4). An overview of practical aspects of the FT synthesis is followed by a discussion of the chief mechanisms that have been proposed for the formation of 1-alkenes by polymerisation of surface C1 species. These mechanisms have traditionally postulated rather non-polar intermediates, such as CH2(ad) and CH3(ad). However, electrophiles and nucleophiles are well-known to play key roles in the reactions of organic and organometallic compounds, and also in many reactions homogeneously catalysed by soluble metal complexes, including olefin polymerisation. We have now extended these concepts to the Fischer-Tropsch reaction, and show that the polymerisation reactions at polarising surfaces, such as oxide-metal interfaces, can be understood if the reactive chain carrier is an electrophilic species, such as the cationic methylidyne, CH(delta+)(ad). It is proposed that the key coupling step in C-C bond formation involves the interaction of the electrophilic methylidyne with an alkylidene (RCH(ad), R=H, alkyl), followed by an H-transfer to generate the homologous alkylidene: CHdelta+(ad)+RCH(ad)-->RCHCH(ad) and RCHCH(ad)+H(ad)-->RCH2CH(ad). If the reactions occur on non-polarising surfaces, an alternative C-C bond forming reaction such as the alkenyl+methylene, RCH=CH(ad)+CH2(ad)-->RCH=CHCH2(ad), can take place. This approach explains important aspects of the enigmatic Fischer-Tropsch reaction, and allows new predictions.

Is it homogeneous or heterogeneous catalysis? Compelling evidence for both types of catalysts derived from [Rh(eta5-C5Me5)Cl2]2 as a function of temperature and hydrogen pressure.

Addressed herein is the 20+ year-old question of whether the true benzene and cyclohexene hydrogenation catalysts derived from the organometallic precursor [Rh(eta5-C5Me5)Cl2]2, 1, are homogeneous or heterogeneous. The methodology employed is that developed earlier (Lin, Y.; Finke, R. G. Inorg Chem. 1994, 33, 4891, "A More General Approach to Distinguishing Homogeneous from Heterogeneous Catalysis..."). The kinetic evidence especially, but also the metal product (nanoclusters plus bulk metal), Hg0 poisoning and other experiments, provide compelling evidence that Rh0 nanoclusters are the true benzene hydrogenation heterogeneous catalyst derived from [Rh(eta5-C5Me5)Cl2]2, 1, at the required more vigorous conditions of 50-100 degrees C and 50 atm H2. However, the same methods reveal that the cyclohexene hydrogenation catalyst derived from 1 at the milder conditions of 22 degrees C and 3.7 atm H2 is a nonnanocluster, homogeneous catalyst, most likely the previously identified complex, [Rh(eta5-C5Me5)(H)2(solvent)] (Gill, D. S.; White, C.; Maitlis, P. M J. C. S. Dalton Trans. 1978, 617). In short, the present results solve the two-decade-old problem of identifying the true benzene and cyclohexene hydrogenation catalysts derived from [Rh(eta5-C5Me5)Cl2]2. Perhaps most significant is the demonstration that the methodology employed has the ability to identify both heterogeneous and homogeneous catalysts from the same catalyst precursor.

Iodide effects in transition metal catalyzed reactions.

The unique properties of I(-) allow it to be involved in several different ways in reactions catalyzed by the late transition metals: in the oxidative addition, the migration, and the coupling/reductive elimination steps, as well as in substrate activation. Most steps are accelerated by I(-)(for example through an increased nucleophilicity of the metal center), but some are retarded, because a coordination site is blocked. The "soft" iodide ligand binds more strongly to soft metals (low oxidation state, electron rich, and polarizable) such as the later and heavier transition metals, than do the other halides, or N- and O-centered ligands. Hence in a catalytic cycle that includes the metal in a formally low oxidation state there will be less tendency for the metal to precipitate (and be removed from the cycle) in the presence of I(-) than most other ligands. Iodide is a good nucleophile and is also easily and reversibly oxidized to I(2). In addition, I(-) can play key roles in purely organic reactions that occur as part of a catalytic cycle. Thus to understand the function of iodide requires careful analysis, since two or sometimes more effects occur in different steps of one single cycle. Each of these topics is illustrated with examples of the influence of iodide from homogeneous catalytic reactions in the literature: methanol carbonylation to acetic acid and related reactions; CO hydrogenation; imine hydrogenation; and C-C and C-N coupling reactions. General features are summarised in the Conclusions.

Promotion of iridium-catalyzed methanol carbonylation: mechanistic studies of the cativa process.

The iridium/iodide-catalyzed carbonylation of methanol to acetic acid is promoted by carbonyl complexes of W, Re, Ru, and Os and simple iodides of Zn, Cd, Hg, Ga, and In. Iodide salts (LiI and Bu(4)NI) are catalyst poisons. In situ IR spectroscopy shows that the catalyst resting state (at H(2)O levels > or = 5% w/w) is fac,cis-[Ir(CO)(2)I(3)Me](-), 2. The stoichiometric carbonylation of 2 into [Ir(CO)(2)I(3)(COMe)](-), 6, is accelerated by substoichiometric amounts of neutral promoter species (e.g., [Ru(CO)(3)I(2)](2), [Ru(CO)(2)I(2)](n), InI(3), GaI(3), and ZnI(2)). The rate increase is approximately proportional to promoter concentration for promoter:Ir ratios of 0-0.2. By contrast anionic Ru complexes (e.g., [Ru(CO)(3)I(3)](-), [Ru(CO)(2)I(4)](2)(-)) do not promote carbonylation of 2 and Bu(4)NI is an inhibitor. Mechanistic studies indicate that the promoters accelerate carbonylation of 2 by abstracting an iodide ligand from the Ir center, allowing coordination of CO to give [Ir(CO)(3)I(2)Me], 4, identified by high-pressure IR and NMR spectroscopy. Migratory CO insertion is ca. 700 times faster for 4 than for 2 (85 degrees C, PhCl), representing a lowering of Delta G(++) by 20 kJ mol(-1). Ab initio calculations support a more facile methyl migration in 4, the principal factor being decreased pi-back-donation to the carbonyl ligands compared to 2. The fac,cis isomer of [Ir(CO)(2)I(3)(COMe)](-), 6a (as its Ph(4)As(+) salt), was characterized by X-ray crystallography. A catalytic mechanism is proposed in which the promoter [M(CO)(m)I(n)] (M = Ru, In; m = 3, 0; n = 2, 3) binds I(-) to form [M(CO)(m)I(n+1)](-)H(3)O(+) and catalyzes the reaction HI(aq) + MeOAc --> MeI + HOAc. This moderates the concentration of HI(aq) and so facilitates catalytic turnover via neutral 4.

Investigations by (13)C NMR spectroscopy of ethene-initiated catalytic CO hydrogenation.

13C NMR spectroscopy shows that the n-alkene and n-alkane products from the catalytic hydrogenation of CO in the presence of (13)C(2)H(4) probes over Ru/150 degrees C, Co/180 degrees C, Fe/220 degrees C, or Rh/190 degrees C (1 atm, CO:H(2) 1:1, "mild conditions") contain terminal (13)CH(3)(13)CH(2)- units. This is consistent with their formation by a regiospecific polymerization of C(1) species derived from CO and initiated by (13)C(2)H(4). Although the activities toward individual products differed somewhat, similar distributions and similar product labeling patterns were obtained over all the four catalysts. 1-Butene and the higher 1-n-alkenes from all the catalysts were largely (13)CH(3)(13)CH(2)(CH(2))(n)()CH=CH(2) (n = 0-3), propene formed over Ru or Co was (13)CH(3)(13)CH=CH(2), while both (13)CH(3)(13)CH=CH(2) and (13)CH(2)=(13)CHCH(3) were formed over Fe or Rh. Comparison of the conclusions from these probe experiments with those from isotope transient experiments by other workers indicates that the ethene initiator does not significantly modify the course of the CO hydrogenation. The reaction products are largely kinetically determined, and the primary products are mainly linear 1-n-alkenes, while the n-alkanes and 2-n-alkenes largely arise via secondary processes. Since the distribution of products and the labeling in them is so similar, it is concluded that one basic primary mechanism applies over all the four metals. Several different reaction paths involving a polymerization of surface methylene, [CH(2(ad))], have been proposed. Although the predictions based on several of these mechanisms agree with many of the results, the alkenyl + [CH(2(ad))] mechanism, initiated by a surface vinyl [CH(2)=CH((ad))], most easily accommodates the experimental evidence. An alternative path involving sequential addition of surface methylidyne and hydride either to a growing alkylidene chain (alkylidene + [CH(ad) + H(ad)]) or to an alkyl chain (alkyl + [CH((ad)) + H(ad)]) has recently been proposed by van Santen and Ciobica. The [CH(2(ad))] mechanism offers an easier explanation for the formation of the various alkenes, the distribution of products, and of the initiation, while the [CH(ad) + H(ad)] mechanism can explain any n-alkanes formed as primary products and not derived from alkenes. At higher reaction temperatures over Ru and Co, considerable (13)C(1) incorporation (from natural abundance in the CO and from cleavage of the (13)C(2)H(4) probe) was found in all the hydrocarbons. Thus, at higher temperatures (13)C(1(ad)) in addition to (13)C(2(ad)) species participate in both chain growth and initiation. In summary, adsorbed CO is transformed very easily into surface C(1(ad)), probably [CH(2(ad))] in equilibrium with [CH((ad))+H(ad)], which act as the propagating species.