Aerosol particles significantly influence the Earth’s climate system, but accurately modeling their complex microphysical processes in global climate models is still challenging. The Energy Exascale Earth System Model (E3SM) includes a comprehensive aerosol representation using the four-mode version of the Modal Aerosol Module (MAM4). Aerosol microphysics processes considered in MAM4 include condensation/evaporation, nucleation, coagulation, aging, and aerosol water uptake. In this study, we train Deep Neural Networks (DNNs) to emulate these microphysical processes simulated by MAM4 in E3SMv2, building on the earlier work by Harder et al. (2022). The original approach is adapted to address the differences between the M7 microphysics model used in that earlier study and the MAM4 model used here, such as different soluble and insoluble modes/species and separate treatments for clear-sky and cloudy portions of the grid box in MAM4. In addition, we also evaluate sensitivities of the emulation results to some modifications to the DNNs.