Over the past decades, significant effort has been dedicated to the development of polymer electrolyte membrane fuel cells (PEMFC) with improved durability. To improve the durability of PEM fuel cells, the scientific focus has been to understand the factors that determine a PEM fuel cell’s lifetime, so that the lifetime can be increased without losing performance or increasing cost. Specifically, attention has been devoted to the stability of low Platinum cathode catalyst layers. Degradation of Platinum nanoparticles over carbon support is critical in consequence of several mechanisms, among which Platinum dissolution. As proposed in the literature, a possible strategy to mitigate Platinum dissolution consists in the development of catalyst layers with gradient properties [1]. In the present work, a theoretical analysis of Platinum dissolution is applied to the interpretation of degradation data recorded on catalyst-coated membranes (CCMs) fabricated by reactive spray deposition technique with 25 cm² active area and low Pt loading 0.10 mg cm^-2 for the cathode and 0.05 mg cm^-2 for anode. The standard protocol defined by U.S. Department of Energy is used to subject the cells to accelerated stress test (AST) with triangle sweep cycles between 0.6 V and 1.0 V at 50 mVs⁻¹ scan rate under inert atmosphere. The effect of aging on cell performance was monitored by measuring the electrocatalyst active surface (ECSA), hydrogen crossover, polarization curves and Electrochemical Impedance Spectroscopy (EIS) after 5000, 10,000, and 30,000 cycles. In order to get insight into the degradation process of cathode catalyst layer under accelerated durability testing, a 1D transient model referenced to the literature [2] has been calibrated and validated on a set of experimental data. The model solves the mass conservation equation for Platinum dissolved in the ionomer phase and a simple model for Platinum dissolution is proposed. The validity of the proposed model is discussed versus the experimental data that include in situ measurements collected on samples with different Platinum nanoparticle size (2nm, 3nm, 5nm) and ex situ data obtained from TEM observations. The evolution of catalyst active surface as predicted by the simulations reveals good consistency between model predictions and measurements, as reported in Fig. 1. It is concluded that the model captures the effect of particle size on nanoparticle stability and is thus used as a tool to analyze PEMFC durability in accelerated degradation tests. The model predicts the formation of a catalyst depleted zone, in consequence of the Platinum flux from the catalyst layer to the polymer membrane, in agreement with TEM observations. The model is later applied to the analysis of samples with gradient properties. Two sets of gradient catalyst layers have been analysed: materials with gradient in catalyst loading across the catalyst layer thickness; materials with gradient in Platinum particle size across the catalyst layer. In both cases the model can predict the trend of measured ECSA. The model predicts that the gradient materials show improved durability if they limit Platinum depletion in the region close to the polymer membrane. According to the model predictions, gradient samples with improved durability show higher content of Platinum catalyst retained in the region next to the membrane. Consistency with TEM observations obtained on gradient samples is discussed and some discrepancies are highlighted as a starting point to improve the model.