Aerosols play a significant role in the climate system. Aerosol microphysics models with detailed process-level representation are highly complex and demand substantial computational resources, posing a challenge for efficient climate simulations. In this work, we propose a deep learning model designed to emulate the aerosol microphysics processes in the version 2 of the Energy Earth System Model (E3SMv2). The current version of the surrogate model is trained on a dataset comprising 10 million samples obtained from a E3SMv2 simulation. By augmenting input dimensions, the surrogate model effectively captures the intricate representations of aerosol and atmospheric state variables. This presentation will explore the feature importance of input variables and their impact on the predictive capacity of the surrogate model in relation to the E3SM outputs. Additionally, we will provide a computational comparison of online inference time when deployed on CPUs and GPUs, highlighting its efficiency and potential for rapid prediction applications in climate modeling.