The complex porous structure of the PEM fuel cell catalyst layer (CL) necessitates the use of multiscale modeling strategies such as the agglomerate approach. In this study a 2D steady-state model for the cathode CL is developed using the spherical agglomerate approach. A new, more accurate, method is introduced to determine the effective agglomerate surface area, which plays a key role in estimating diffusion losses in the CL. Specifically, the reduction in the effective surface area due to overlapping particles is modeled geometrically based on a sphere-packing approach. In addition, the equations for the agglomerate model are reformulated to correctly account for the agglomerate surface area reduction due to overlapping particles. The importance of an accurate geometric model for the effective surface area is demonstrated by investigating the effect of CL composition on performance, and the results show that the new method provides more realistic predictions than the existing approach. Results from the new approach for optimal Pt loading, ionomer loading, and Pt|C ratio show good agreement with experimental results. Great strides have been made in proton exchange membrane (PEM) fuel cell research since the introduction of the technology in the early 1960s. Computational models represent a major tool for the design and development of PEM fuel cells because they decrease the dependence on expensive and time-consuming experimental procedures. Accurate computational models can also provide key insights into the fundamental processes that govern fuel cell performance. The overall accuracy of computational models depends strongly on the description of the catalyst layer (CL) as all the critical electrochemical reactions, and heat and mass transport processes occur within the CL. Therefore, developing better modeling strategies for the CL has been an important research objective to obtain a realistic representation of PEM fuel cell behavior.Catalyst layer modeling approaches, which have grown in complexity in the past decades, were described in detail in. 1 Here, we present only a brief review. Among the models reported, the interface model 2-4 is the simplest approach as it represents all of the CL activity as boundary conditions applied at the interface between the membrane and the gas diffusion layer (GDL). The macro-homogenous model [4][5][6][7] is a more advanced approach that utilizes the effective medium theory and approximates the CL as a porous matrix of gas pores, platinum, carbon, and electrolyte. The most comprehensive device-level model to date is the agglomerate approach [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22] as it represents CL activity as a multiscale problem. Overall, the CL is assumed to consist of gas pores and aggregates of C|Pt particles that are covered by an electrolyte film. The agglomerate approach attempts to model the C|Pt aggregates (shown schematically in Fig. 1a) that are typically seen in microscopic images of the CL. For simplicity, the aggregate microstructure, de...