<p>A detailed microphysical model is developed using a Lagrangian-particle-based direct numerical simulation framework to simulate ice growth in a turbulent mixed-phase environment. The Lagrangian particle method is employed to track the interactions between ice, droplets, and turbulence at the native scales. The investigation reveals for the first time the mixed-phase processes at the sub-meter length scales using direct numerical simulation.</p> <p>This paper examines the conditions that favor effective ice growth in the cloud top generating cells. Investigations over a range of environmental (macrophysical and turbulent) and microphysical conditions (ice number concentrations) that distinguish generating cells from their surrounding cloudy air were conducted. Results show that high liquid water content (LWC) or high relative humidity (RH) is critical to maintaining effective ice growth and mixed-phase. As a result, generating cells with high LWC and high RH provide favorable conditions for rapid ice growth. Sensitivity studies on ice number concentrations show that when the ice number concentration is below 1 <em>cm</em><sup>-3</sup>, a typical range in the mixed-phase clouds, a high LWC is needed for efficient formation of big ice particles. The study also found that supersaturation fluctuations due to small-scale turbulent mixing have a negligible effect on the particle mean radius but substantially broaden the size spectra which can affect the subsequent collection process.</p>