(722e) Theoretical Analysis of Formic Acid Decomposition On Transition-Metal Catalysts

Hydrogen is a promising energy carrier that can be used to generate electricity in fuel cells without pollution. To implement this fuel cell based hydrogen economy, however, practical solutions to controlled storage and release of hydrogen are essential. Previously, hydrogen storage has been achieved by compressing the gas at high pressures, liquefying it at low temperatures, and storing it in the form of chemical/metal hydrides. All these methods, however, suffer from loss of hydrogen, safety issues, and low energy densities.

Recently, formic acid has been suggested as a sustainable material for hydrogen storage. It contains 4.4 wt.% of hydrogen with a volumetric capacity of 53.4 g/l at STP. Even though its hydrogen content falls short of the milestones (5.5 wt.%) set by the US Department of Energy for 2010, its volumetric capacity surpasses that of most other storage materials used today. Hydrogen stored in formic acid can be released on demand by decomposing formic acid into H₂ and CO₂. If we can produce formic acid again by CO₂ hydrogenation, a carbon-neutral hydrogen storage cycle would be established on the globe. To develop formic acid as a hydrogen storage material, however, we first need to find catalyst materials that can actively decompose formic acid into H₂ and CO₂.

Previous studies have shown that some elements of noble metals were highly active for formic acid decomposition. For example, Solymosi and co-workers investigated formic acid decomposition over a range of monometallic noble metal catalysts (Ir, Pd, Pt, Ru or Rh supported on carbon) at 423 K. All their catalysts showed promising activity with turnover numbers of 1280-4120 hydrogen molecules per site per second. When the performance of different catalysts was compared, the catalytic activity trends of Ir ≈ Pt ≈ Pd > Ru > Rh were observed. However, these trends may not reflect the intrinsic catalytic properties of the noble metals because the metal dispersion of the catalysts was different ranging from 5.6% to 26.1%. 

For direct applications in fuel cells, formic acid needs to be decomposed via dehydrogenation (HCOOH --> H₂ + CO₂) rather than via dehydration (HCOOH --> H₂O + CO) such that virtually CO-free hydrogen is produced. Many researchers have tried to achieve this by adding a secondary metal to some noble metals. The improvements in hydrogen selectivity as well as catalytic activity were particularly significant on Pd-Au and Pd-Ag catalysts that produced CO below 100 ppm under ambient conditions. However, these catalysts are heavily based on expensive rare-earth elements. We need to find alternative materials that are not only active and selective for formic acid dehydrogenation but also cost-effective.

First principles based analysis was used to investigate formic acid as a sustainable material for hydrogen storage. The reaction mechanism and the energetics of formic acid decomposition were obtained based on the adsorption energies of the reaction species calculated over five (211) surfaces: Ag(211), Cu(211), Pd(211), Pt(211) and Rh(211). Formic acid decomposition was found to proceed via dehydrogenation rather than via dehydration on the surfaces we investigated, although the minimum energy pathway varied depending on the surface. The DFT-calculated adsorption energies decreased significantly for some species adsorbed on Cu(211) when the BEEF-vdW functional was used instead of the RPBE functional. This was interpreted as vdW forces playing an important role in the interaction of the species with the metal surface.

We then extended the study to other transition-metal surfaces by scaling the vdW-included adsorption energies of the reaction species with CO and OH adsorption energies (E_CO and E_OH). The scaling relations for individual species were independent of the surface structure, allowing us to estimate the adsorption energies of the reaction species adsorbed on other surfaces by calculating two parameters per surface: E_CO and E_OH. A micro-kinetic model was used to construct a theoretical activity volcano that mapped out the turnover frequencies for hydrogen production from formic acid as a function of E_CO and E_OH. Hydrogen and CO selectivity maps were also established as a function of E_CO and E_OH. The catalytic activity and selectivity trends obtained from our first-principles analysis were consistent with many experimental observations reported in the literature. The interpolation concept was used to quickly search for binary alloys that can actively and selectively decompose formic acid into H₂ and CO₂. As a result, Cu₃Ni alloy was newly identified to be interesting. We hope to find even more interesting catalyst materials when we add more data points to the volcanoes presented in this study.