Publications 2017 - Section for Atomic Scale Materials Modelling

Arghya Bhowmik, Heine A. Hansen and Tejs Vegge, ACS Catal., 2017, 7 (12), pp 8502–8513

Read the paper here.

We propose an innovative concept of ligand effects in oxide catalysts. Both IrO₂ and RuO₂ binds OH* and other CO2RR intermediates strongly, but the stable and miscible system Ir_xRu_(1-x)O₂ exhibits anomalous weaker binding energy in presence of CO* spectators due to Ru-Ir ligand effects. A RuO₂ surface doped with Ir move close to the top of the predicted CO2RR volcano for oxides. This leads to very low CO2RR onset potential (methanol evolution at -0.2 V-RHE). This offers an unprecedented improvement over state of the art electrocatalysts for conversion of CO₂ into methanol.

Vladimir Tripkovic, H.A. Hansen and Tejs Vegge. ASC Catal. 2017, 7, 8558–8571.

You can read the paper here.

We investigate the structural stability, catalytic activity, and electronic conductivity of pristine and doped Ni and Co oxyhydroxides ranging from bulk (3D) to single-layer (2D) catalysts. We establish the dependence of the electronic conductivity and activity on potential and find it is more energetically favorable to dope Ni than Co oxyhydroxides. We identify first-row transition and noble metals to be the most stable dopants, with Rh-doped Ni having the highest calculated activity.

Reshma R. Rao, et al. Energy Environ. Sci., 2017, 10, 2626-2637

You can read the paper here.

This article is part of the themed collection: 2017 Energy and Environmental Science HOT articles

In collaboration with Prof. Yang Shao-Horn’s group at MIT, we have combined in situ surface X-ray scattering measurements and DFT calculations to determine the surface structural changes on single-crystal RuO₂(110) as a function of potential in acidic electrolyte. At potentials relevant to the oxygen evolution reaction (OER), an –OO species on coordinatively unsaturated Ru sites (CUS) sites was detected, which was stabilized by a neighboring –OH group on the Ru CUS or bridge site. Combining potential-dependent surface structures with energetics from DFT led to a new OER pathway, where the deprotonation of the –OH group used to stabilize –OO was rate-limiting.

Arghya Bhowmik , Heine Anton Hansen, and Tejs Vegge, J. Phys. Chem. C, 2017, 121 (34), pp 18333–18343
You can read the paper here.

Employing density functional theory based computational electrocatalysis models we studied binding energy amendment due to adsorbate interaction (steric and electronic) with varying coverage of CO* spectators on the catalyst surface. Implications on the reaction pathways help us rationalize differences in experimentally observed carbonaceous product mix and suppression of the hydrogen evolution reaction (HER). We show that a moderate CO* coverage (∼50%) is necessary for obtaining methanol as a product and that higher CO* coverages leads to very low overpotential for formic acid evolution.

Qiang Fu, et al. Chem. Mater, 2017, 29, 1641-1649.

You can read the paper here.

We have analyzed Nb-doped tin dioxide (NTO) as an alternative support for Pt-based catalysts in PEMFCs. Through a combined DFT and non-equilibrium Green’s function study, we investigate the Nb segregation at Pt/NTO interfaces under operational conditions, and reveal the effects on the electronic transport and catalytic properties. We find that the Nb dopants tend to aggregate in the subsurface layers of the NTO substrate, whereas their transport across the Pt/NTO interface is hindered. This understanding will help shed light on future applications of Nb/Sb-doping in SnO₂.

A Peterson, R. Christensen, and A. Khorshidi; Phys. Chem. Chem. Phys., 2017, 10978-10985
You can read the paper here.

Machine-learning regression can emulate more expensive electronic-structure calculations of potential energies and forces. Here, neural network calculators are trained to emulate another calculator based on a set of training images. However, the uncertainty of a given machine-learning calculator is unknown outside the not always well-defined area of training. We here demonstrate how an ensemble of trained neural network calculators can be used to estimate the uncertainty in calculations with trained neural networks.

Ask Hjorth Larsen et al J. Phys.: Condens. Matter, 2017, 29, 273002

You can read the paper here.

The Atomic Simulation Environment (ASE) is an open source Python software package for initializing, performing and analyzing atomistic structure simulations. The use of Python and various libraries make the code both powerful and user-friendly. ASE interfaces uniformly to a wide range of electronic structure codes. We are part of the development community. In this overview paper of fundamentals and highlighted features of ASE, we contributed with description of genetic algorithms and Bayesian error analysis tools.

Kelsey A. Stoerzinger, et al. ACS Energy Lett., 2017, 2, 876-881.

You can read the paper here.

In collaboration with Prof. Yang Shao-Horn’s group from MIT, we have examined the OER kinetics on rutile RuO₂ in the (110), (100), (101), and (111) orientations, finding (100) the most active. We find no evidence of oxygen exchange in acidic or basic electrolytes, suggesting it is not exchanged in catalyzing OER on crystalline RuO₂ surfaces. This is supported by the correlation of activity with the number of active Ru-sites calculated by DFT, where more active facets bind oxygen more weakly. This new understanding provides a design strategy to enhance the OER activity of RuO₂.