61st Annual Report on Research 2016

Executive Summary. Steam methane reforming is the most widely-used industrial process for hydrogen and synthesis gas production from fossil fuels, and currently accounts for over 95% of the hydrogen produced in the United States. In our ACS-PRF program, we are exploring and establishing a new approach to tailored design of catalysts for steam methane reforming based on the use of metal catalysts integrated with ferroelectric oxide support layers, in which dynamic switching of the polarization of the support layer is used in real time to modify surface properties, and hence promote the desired chemical reactions, on the metal catalyst surface. The approach is based on polarization modulation of a ferroelectric oxide support to reduce key reaction barriers for molecular adsorption, dissociation, association, and release, and hence promote desired pathways. Using first-principles quantum mechanical modeling methods based on density functional theory, the goals of this program are to (i) identify the mechanisms and rate limiting steps for steam methane reforming reactions on nickel catalysts, (ii) design integrated ferroelectric oxide/metal systems in which the polarization tunes the electronic structure and surface chemistry of the metal, and (iii) optimize the design and assess the ultimate potential of the newly proposed mechanism. Alongside the research, the program will train undergraduate and graduate students in the atomistic modeling and simulation of chemical processes relevant to the petroleum field.

Introduction. Steam methane reforming is currently the most widely used process for hydrogen production in the United States. It is a highly endothermic reaction (requiring an external energy supply) that uses high-temperature steam to convert two stable molecules (methane to hydrogen) to the more reactive carbon monoxide gas:

CH₄ + H₂O ˆ CO + 3H₂ (DH = -206 kJ/mol). (1)

Widescale adoption of this process is currently hindered by challenges in capital costs and operation and maintenance costs of the steam methane reforming process. As their costs preclude the usage of highly-active, stable noble metal catalysts on the industrial scale, nickel is the most widely-used catalyst today. However, conversion efficiencies of nickel catalysts are not as high as their noble metal counterparts, and their performance is severely hindered by coking, ultimately leading to encapsulation and deactivation of the catalyst. There are unexplored opportunities in the design of oxide-supported metal catalysts for steam methane reforming. Amongst the class of oxides considered as catalytic support layers, ferroelectric oxides are currently receiving attention on account of their switchable polarization, which may offer routes to dynamic control over surface chemistry and chemical reactions. If successful, the development of high-efficiency catalysts based on this approach will provide a strong incentive to develop on-site infrastructures to capture and convert natural gas into products useful for society.

Methods & Results. In our second year of this program, we have finalized the analysis of the steam methane reforming reaction taking place on exposed nickel surfaces supported by lead titanate ferroelectric oxide substrates. We have considered both positive and negative poled ferroelectrics, and compared the reaction profiles to those of unsupported (conventional) nickel catalysts. Our work has revealed several exciting highlights, and we think the approach that we are proposing, although still at an early stage, exhibits many promising aspects:

• As shown in Figure 1, the Gibbs free energy profile of the steam methane reaction occurring on nickel catalysts supported by a polarized lead titanate ferroelectric substrate exhibits a remarkable sensitivity to the polarization of the ferroelectric. This suggests that the proposed use of ferroelectric supports to modify a reaction profile may in fact be effective.
• There is hardly any reconstruction or distortion of the Ni surface due to adsorption of reaction intermediates. The reaction is simulated at a total coverage of 0.5 monolayers. The surface geometries look largely the same for unsupported Ni as well as the +/- P systems.
• Despite the similarity in the adsorbate geometries, the energy barrier for the most endothermic steps of the reaction sequence, which are the hydrogen removal steps, become greatly reduced and in some cases even become slightly exothermic in the presence of the ferroelectric material. Interestingly, we find this to occur for both polarizations of the ferroelectric.
• In several cases, the difference between the energies of the positive and negative polarizations is substantial. This suggests the possibility of dynamically switching the polarization, to overcome rate limiting reaction barriers and promote the desired reaction sequence.
• The observed trends in the reaction profile can be largely understood from the chemistry of the nickel/ferroelectric interface. The stronger interfacial bond in the case of negatively polarized systems leads to a larger interfacial dipole, which induces a greater interaction with the surface adsorbates, and hence, a larger variation in the SMR Gibbs free energy profile.

Plans for Continuation. During our second year, we completed the analysis of the full steam methane reforming reaction using these integrated ferroelectric/Ni systems, for the cases of both positively and negatively polarized ferroelectrics. The chemical reaction pathways and barriers, compared to those obtained for isolated nickel catalyst layers, exhibit a remarkable sensitivity to the substrate polarization. Our future goals lie along two axes: (i) we wish to identify experimental collaborators or collaborations with industry to assess and validate our predictions, and (ii) we wish to continue to carry out computational analysis to further establish the nature of the observed trends. This includes an analysis of finite size effects of ultra-thin nickel catalysis, and the incorporation of chemical modifications at the interface to further tailor and design the reaction free energy profile.

Fig. 1. The interfacial bonding between lead titanate and Ni for positive and negative polarizations (a): the nickel, lead, titanium, and oxygen atoms are represented by pink, grey, blue, and red spheres respectively. The interfacial atoms of the lead titanate are equidistant from three nickel atoms of the first layer, in the lowest energy registry. The stronger interfacial bond in the case of negatively polarized systems leads to a larger interfacial dipole, which induces a greater interaction with surface adsorbates, and hence, a larger variation in the Gibbs free energy profile shown in (b).