The replacement of Pt group metal (PGM) electrocatalysts with earth-abundant PGM-free materials in polymer electrolyte fuel cell (PEFC) cathodes faces several technological hurdles, including the need for improved long-term durability. In PGM systems, catalyst durability is achieved by maintaining Pt catalyst surface area that is in electrical contact with the electrode (electrochemical surface area, ESA).1,2 This ESA can be probed directly by electrochemical methods. Loss of electrical contact (via carbon corrosion or Pt particle detachment) as well as Pt surface area loss (via Pt dissolution into ionomer, particle agglomeration, surface blocking, or Ostwald ripening) lead to a loss in oxygen reduction reaction (ORR) activity in aged PGM cathodes. State-of-the-art PGM-free catalysts are highly heterogeneous systems due to pyrolysis processing and contain a great diversity of atom-scale structures. As such, there is still much debate regarding the atomic scale structure of the ORR active site. It has been proposed that a number of sites in the materials are ORR active, with varying degrees of activity and durability, further complicating study. Unlike their PGM counterparts, PGM-free materials do not have an easily electrochemically probed property such as ESA that can be determined and correlated to measured activity loss. These complexities necessitate different approaches to identifying and predicting durability properties and degradation mechanisms in PGM-free systems. One proposed approach for understanding PGM-free catalyst durability3 is combining quantum chemical modeling with a TEM beam-damage model4 to identify how local atomic structure affects the kinetics of atom removal. Calculation of the knock-on displacement threshold energy (KODTE) could serve as a durability descriptor for a given atomic structure, similar to how thermodynamic limiting potential serves as a computationally-accessible activity descriptor of ORR active site structures. This descriptor would go beyond relative thermodynamic stability of active site structures and actively probe the kinetics of bond breaking in the system. As part of the DOE Electrocatalysis Consortium (ElectroCat), an automated workflow for calculating the KODTE for a given atomic scale structure using ab initio molecular dynamics has been developed and applied to a variety of possible PGM-free ORR active site structures as well as C-host materials. These methods have identified that N are most susceptible to removal and that edge-hosted structures are more susceptible than bulk-hosted structures. Utilized methodology, as well as identified trends from these calculations and comparisons to available experiments will be discussed.