Structure-Function Relationships of PGM-Free ORR Electrocatalysts from Density Functional Theory

Abstract

Pyrolyzed platinum group metal free (PGM-free) oxygen reduction reaction (ORR) electrocatalysts are a promising class of earth-abundant materials for low-temperature polymer electrolyte fuel cell (PEFC) cathodes. Understanding of the atomic scale structure of PGM-free ORR active sites remains a key focus of research efforts with the aim of improving performance of these materials. The pyrolysis process leads to highly heterogeneous catalyst systems which complicates direct active site study. In particular, (i) understanding how these active sites give rise to ORR activity, (ii) how they interact with the environment during material degradation, (iii) their interactions with probe molecules for the purposes of active site quantification, and (iv) their spectroscopic signatures are all important aspects governed by atomic scale structure. Through the use of density functional theory (DFT), we investigate these four structure-function relationships. ORR activity, one of the greatest challenges faced by PEFCs, is explored via several binding energy parameterized descriptor models. These models include the computational hydrogen electrode (CHE) model which yields thermodynamic limiting potentials, and the linearized Gibbs energy relationship (LGER) model which yields reversible potentials for reaction steps. Additionally, kinetics of OOH bond dissociation via nudged elastic band (NEB) DFT calculations are also considered as this has been proposed by some groups to be rate determining for pathways that include binding to local C sites. Combined, these models give insight into how local arrangement of atoms affects reaction pathways and enables exploration of varied reaction pathways on and local to the active site. Durability of PGM-free electrocatalysts remains a key challenge in these materials. While a variety of degradation mechanisms have been proposed, the relation between environment and degradation mechanism has not been firmly established, complicating any mitigation strategies that might be applied. Through the use of an automated ab initio molecular dynamics (AIMD)-based model, the kinetics of bond breaking local to the active site is explored. Resulting degraded structures can then be studied via the activity models previously mentioned to show how such degradation affects calculated ORR activity descriptors. Additionally, the use of reaction rate models applied to activity loss curves can also enable some discrimination of degradation mechanism indicating either a single autocatalytic process or two mechanisms with different temporal behavior. Another issue faced by PGM-free electrocatalysts is the lack of a method for quantifying the number of ORR active sites. Unlike Pt-based systems where integration of charge transferred in the H adsorption/desorption region can give information about density of active sites, no such method has been fully established for determining the number of PGM-free active sites, a value required to deconvolute turn over frequency from ORR current densities. Promising approaches include use of probe molecules that, if bound to ORR active sites (and only ORR active sites) could provide insight into how many active sites are in a given sample. This is highly dependent on the specificity of binding for these molecules which can be explored via binding energy calculations with DFT. Finally, DFT can also be a valuable tool for understanding spectroscopic signatures of highly ORR active materials. The main issue faced in such studies is focusing on just the active sites which generally requires considering changes in the Fe states, particularly for so called atomically dispersed catalysts that exhibit the highest ORR activities to date. X-ray adsorption spectroscopy (XAS, in particular XANES and EXAFS) can give information about local structure and the state of Fe with and without probe molecules. DFT and related theoretical approaches can simulate these spectra as a function of local environment. Vibrational calculations with and without probe molecules give insight into Fe-specific nuclear resonance vibrational spectroscopy (NRVS). Calculation of energy density at the Fe nucleus and electric field gradient gives input regarding isomer shift and quadrupole splitting, giving much needed interpretation of Mössbauer spectroscopy, especially in instances where no standard exists or changes in measured parameters with ligands/probes are required. Combined, these coupled experimental/theoretical approaches give insight into the nature of active sites in PGM-free systems. In particular, in this contribution we will focus on the presence of spontaneously evolved ligands that, using DFT, are shown to be likely contributors to electrocatalyst activity in-situ and experimental signatures thereof.