Rhodium Porphyrin Preferential Oxidation (PROX) Catalysts For Carbon Monoxide Conversion

Stephen DiMagno


One difficulty for the widespread implementation of hydrogen powered vehicles is the hydrogen storage problem. Since hydrogen is a gas that can only be liquefied with an energy penalty that is roughly 1/3 of the total energy content of the fuel itself, advanced hydrogen storage materials and liberation of hydrogen from liquid organic substances (such as alcohols) have been pursued as alternative, low cost strategies to cryogenic liquid hydrogen storage. Hydrogen fuel streams produced by hydrocarbon, alcohol, or biomass steam reforming are invariably contaminated with carbon monoxide (CO). Subsequent high temperature water gas shift (WGS) catalysis can reduce CO concentrations to the thermodynamic limit (20-50 ppm) in such hydrogen gas streams. However, the platinum anodes in current generation proton exchange membrane (PEM) fuel cells are poisoned by CO concentrations as low as 10 ppm. Our approach to this problem is to develop a catalyst or a catalytic reformer that can remove CO selectively at low temperature. Design requirements for such a catalyst are severe, since selective CO oxidation needs to be performed in the presence of hydrogen, carbon dioxide, and water prior to entry of the gas into the fuel cell.

Preferential oxidation of CO can be performed indirectly or directly. The advantage of the indirect approach (Scheme 1) is that CO’s reducing equivalents can be captured at a fuel cell anode; such energy recovery is appropriate for gas streams containing high CO concentrations. In contrast, direct oxidation of CO with air or O2 is an attractive option for scrubbing CO from streams produced by WGS catalysis. Because CO concentrations are already low in these gas streams, any potential energy recovery from CO oxidation is modest.


Scheme 1. Overall reaction pathway for the oxidation of carbon monoxide with rhodium porphyrins.

Group IX transition metal metalloporphyrins are known to be good catalysts for a wide variety of useful reactions, such as alkane activation and functionalization. The DiMagno group and others have studied intensively the impact of ?–fluorination on porphyrin ring and metal redox potentials. Ligand fluorination alone can shift the formal potential for the rhodium(II)/rhodium(I)- reversible redox couple by 540 mV in nonaqueous solution. This shift in potential makes these complexes excellent oxidants for a wide variety of reductants, including CO.

We discovered homogeneous, room temperature aqueous carbon monoxide PROX catalysts based upon rhodium(III) salts of 5,10,15,20-tetrakis(4-sulfonatophenyl)-2,3,7,8,12,13,17,18-octafluoroporphyrin, 1, and 5,10,15,20-tetrakis(2,6-difluoro-3-sulfonatophenyl)-2,3,7,8,12,13,17,18-octafluoroporphyrin, 2. The chemistry of these complexes was compared to the catalytically inactive, but structurally related benchmark compound 5,10,15,20-tetrakis(4-sulfonatophenyl)porphyrin 3 (Figure 1). The room activity of these complexes for carbon monoxide oxidation in water is unprecedented under such mild conditions


fig 1

Figure 1. Structures of compounds 1, 2, and 3.

To gather mechanistic information about PROX catalysis by these complexes, we studied the kinetics of individual steps in the process. The rate of reduction of 1, 2, and 3 to the corresponding RhI complexes by 1 atm of CO at 30°C under a variety of conditions and with various additives was examined. Generally, it was found that rates of CO oxidation increased with pH. For example, 2 was reduced to the corresponding RhI complex with the greatest rate observed (140 ± 13 M-1s-1) in 100 mM sodium hydroxide. The increased rate constant for 2 is consistent with the hypothesis that the most electron deficient porphyrin would increase the electrophilicity of the bound CO, which, in turn, increases the rate of nucleophilic attack. The rates of reduction of 1 and 3 are identical (within experimental error), and much slower than that of 2. Unfortunately, at high base concentration significant decomposition of the metal complexes was observed, as was evidenced by the appearance of the signal for free fluoride in the 19F NMR spectrum. We speculated that if the role of hydroxide could be filled by some nonbasic, non-nucleophilic additive, a reasonable PROX catalytic system would result.

An alternative explanation for the improvement in catalyst performance with increasing pH is that the five-coordinate hydroxo complex bearing an open coordination site for CO binding is the catalytically active species. Thus, we examined the impact of strongly-coordinating halide ions in the hope that they might also free an additional metal coordination site and thereby increase the turnover rate of CO oxidation. The association constants for the binding of halide ions (pH = 7) to 2 are shown in Table 1. The addition of halide ions led to CO oxidation rates in neutral or acidic (pH = 4) conditions that were comparable to those obtained under alkaline conditions, with the advantage that the ligand system is robust at neutral or acidic pH.

Table 1: Halide association constants with 2.

Cl- 45 ± 2 1
Br- 11,000 ± 400 3
I- 7,000 ± 200 8


To make the system catalytic, rather than stoichiometric, we sought regeneration systems compatible with O2 as the terminal oxidant. The blue dye indigo carmine was found to be suitable for the reoxidation of reduced rhodium porphyrins. Moreover, rate constants for the redox reaction of indigo carmine with the porphyrin complexes are large, ensuring that the rate limiting step remains turnover of CO. The scrubbing of CO from a gas mixture by a 10-5 M solution of 2 in water is shown in Figure 2. The figure shows the bleaching of the indigo carmine upon reduction (2H+ + 2 e-) and subsequent reoxidation of the dye by O2 to form water. Importantly, the catalyst 2 does not have any hydrogen activity; CO can be removed from hydrogen gas mixtures without oxidizing hydrogen.

fig 2

Figure 2: Reaction of 2 with CO in the presence of 100 equivalents of indigo carmine at pH = 7. a) Initial, b) 15 min. after CO purge, c) after complete bleaching of the dye, d) immediately after exposure to air.

These studies indicate that CO scrubbing by PROX catalysis is possible under exceptionally mild conditions. Current work in our laboratory is focused on converting these homogeneous catalysts to robust, supported heterogeneous catalysts.

Stephen DiMagno