Decreasing temperature is generally expected to attenuate or arrest product degradation; however, freezing often introduces undesirable and unexpected quality issues. Freezing is inherently a process of microscale phase separation, in which solute concentrations increase dramatically as water crystallizes. This process is accompanied by pH shifts and elevated osmotic and hydrostatic pressures (reaching up to ~2 kbar), forcing concentrated phases and living cells through the porous ice network and exposing them to cold denaturation and significant mechanical stresses.

Here, we approach freeze–thaw engineering from the microscale to the bulk scale by computationally simulating local stresses experienced by proteins and cells using computational fluid dynamics. These simulations are used to design heat-transfer geometries that better control ice-growth behavior. Case studies demonstrate reduced protein aggregation and improved post-thaw functionality of stem cells, highlighting the importance of ice-growth geometry as a critical design parameter in cryopreservation systems.