(118e) Comparison Between CO-PROX and CO-SMET Catalytic Processes as Final Step for Syngas CO-Clean-up

Thanks to their characteristics of high energy efficiency and low emissions, Polymer Electrolyte Fuel Cells (PEM-FCs) are the most promising candidates for electric power generator in transportation applications. At present, the most probable fuel is H₂-rich gas coming from reformed or partially oxidized hydrocarbon fuels, which downstream the Water Gas Shift step contains at least 0.5-1% of CO. Unfortunately, the performance of the Pt, the most effective electro-catalyst for H₂ oxidation in PEM-FCs, is seriously depressed from the presence of only 5 to 10 ppm of CO. There are several methods suitable to remove CO from syngas. For PEM-FCs applications however, the feasible choice may be only between the preferential CO oxidation (CO-PROX) and the selective CO methanation (CO-SMET) because of the limited available spaces and low operating pressures in such application systems. The CO-PROX process has so far been extensively tested since it is somehow reliable to remove CO down to 10 ppmv by raising the O₂ availability unlimitedly. Nonetheless, the technology requires a closely controlled low O₂ supply to keep the possibly lowest H₂ parallel oxidation and a working temperature window suitably wide for control purposes. This, while making the method costly and complex, obviously hampers its application to low-power PEM-FCs where very small oxidant rates have to be provided by use of well-refined expensive mass flow meters. Without any additional reactant required, the CO-SMET can avoid the above-mentioned shortcomings of the CO-PROX application. Furthermore, the CO methanation is less exothermic than the CO and H₂ oxidations. Thus, a CO-SMET reactor is inherently more easily controllable than a CO-PROX one. The present work deals with a comparison of catalysts performance for CO abatement via CO-PROX or CO-SMET.

Considering the CO methanation reaction (Δ°₂₉₈ = −206.2 kJ mol⁻¹), the potential chemical energy of reactants is practically conserved considering the heat developed by the reaction itself and the combustion energy of the produced CH₄. Moreover, the removal of 1 CO mole via methanation requires 3 mol of H₂. Of these, 2 mol are restored as CH4 product; the latter in the case where the FPU is integrated with the PEM-FC, as in APU systems, can be reused downstream the stack by recirculating the anode off-gas to the steam reformer burner where it is burnt for heating purposes with the main fuel rate to the burner. In a PEM-FC system, in general, the unreacted H₂ in the anode off-gas may reach 10–20% of its original feed rate. Considering H₂ consumption, for CO-SMET the actual H₂ loss should be of 3 mol per mole of removed CO, if the methanation selectivity toward CO is equal to 1 (i.e., no CO₂ methanation occurs).

For the CO-PROX there is theoretically no H₂ loss if the O₂ selectivity towards CO is equal to 1. This theoretical reaction efficiency, however, is hardly possible to be reached in practice over commercial catalysts, but also none is willing to risk the stack safety by operating at λ = O/CO = 1.0 v/v, even if the catalyst could allow, also taking into account that an actual PEM-FC system fluctuates intrinsically. Therefore, some O₂ excess allows a safer reactor management but, as a consequence, some H₂ loss, via H₂ oxidation, surely exists in practical CO-PROX reactors. This loss, based on literature reports, should be at least from 1 to 3 mol of H₂ per mole of removed CO because actual COPROX reactors run generally at λ = O/CO ratios of 2.0–4.0 v/v.

Hence, if one considers the possible heat recovery (to maintain as highest as possible the transformation efficiency of primary fuel to final H₂, FPU requires thermal energy, especially the “noble” one at high temperature), the CO-PROX reactor (assuming average λ of 3 and heat transfer efficiency of 1) allows a reaction heat recovery about 3.7 times higher compared to the CO-SMET reactor. Though, the heat from CO-PROX is at low temperature, so thermodynamically of not high significance. On the contrary, CO-SMET process, through the combustion of produced CH4, allows also obtaining, at high temperature and hence thermodynamically very useful for the FPU necessities, an amount of heat per mole of removed CO a little bit higher than the heat released from the CO-PROX reaction.

So, considering the global heat recovery (directly from the reactor and from the produced CH4 combustion), that from CO-SMET is about 1.3 times the one obtainable from CO-PROX and about 80% of this thermal energy (the one coming from CH₄ combustion) is available at a higher temperature level compared to the CO-PROX case. The availability of high temperature heat is very appreciated and useful since it positively affects the APU efficiency: APU, in fact, needs high temperature heating in the FPUs, otherwise satisfied by primary fuel burning.

On the other hand, considering the power losses from the two clean-up process final steps, using the above assumptions, CO-SMET consumes 1 H₂ mole per mole of removed CO more than CO-PROX (3 mol compared to 2 mol, respectively); assuming an FC average efficiency of 0.5 and anode stoichiometry of 1.2, the power loss for CO-SMET is independent from the stack power, but dependent on the residual CO concentration. Hence, the power loss for CO-SMET is higher compared to CO-PROX and ranges from 1.2% to 2.5% of the stack power for a CO residual inlet concentration in CO-SMET reactor of 0.5% or 1% b.v., respectively.

Thus, considering pros and cons, the CO-SMET process can be proposed as a promising route to remove CO in the reformate stream and as an alternative process to CO-PROX in FPUs for mobile and residential PEM FCs applications. Obviously, the development of catalysts able to remove the CO from percent levels (generally about 0.5–1%) down to ppmv level (10–50ppmv and lower) and characterized by high selectivity values towards the CO methanation in the presence of high CO₂ concentrations is a mandatory issue.

The present paper deals with the comparison of CO-PROX and CO-SMET catalysts for the final step of CO clean-up in FPs. The comparison has been carried out both experimentally, by developing and testing suitable catalysts for both the clean-up processes, and by modeling a FPU for stationary or mobile applications, carrying on as clean-up reatcot the CO-PROX or CO-SMET units. Comparisons between the energetic aspects of the modelled FPS will be presented in the full paper.

On the experimental pointof view, concerning the CO-PROX, Rh-based catalysts carried on Al₂O₃ and 3A zeolite were studied, whereas Ru-based catalysts carried on ZrO₂ and CeO₂ were employed for the CO-SMET. All the catalysts were prepared by a conventional impregnation method and CO removal was determined by varying the operating conditions (inlet CO/CO₂/H₂/H₂O concentrations, temperature, weight space velocity WSV). Among the prepared catalytic materials, the 1% Rh-zeolite 3A catalyst, tested at a WSV of 0.66 NL×min^-1×g_cat^-1, was found to be the most suitable one for the CO-PROX at low temperature: it reduced the inlet CO concentration below 10 ppmv within a temperature range of at least 80-120 °C without the appearance of undesirable side reactions. On the contrary, for the CO-SMET the best results were obtained with the 5% Ru/CeO₂ catalyst: the CO inlet rate was completely transformed into methane in a temperature range of 250-280 °C at WSV of 0.45 NL×min^-1×g_cat^-1, and simultaneously the parasite reactions (CO₂–MET and R-WGS) rate remained at an acceptable level.