The usual way to freeze a living cell is like trying to put a spacesuit on it: you first need to get a protective substance inside, but while it's crossing the membrane, the cell can burst from salt imbalances or get poisoned by the chemicals.

The new approach skips the ready-made antifreeze and assembles it on the spot. Harmless 'construction blocks'—small molecules that barely affect osmotic pressure and aren't toxic — enter the cell. They can slip in passively or via normal cellular uptake. Then, when freezing begins, the pressure inside a confined volume (like a vessel or the cell itself) spikes. That pressure breaks the blocks apart into individual active cryoprotectant molecules, and protection appears right when it's needed.

The key idea is that the 'disassembled' state takes up less space than the 'assembled' one. So when pressure rises, the equilibrium shifts toward disassembly.

This mechanism works both when freezing at constant volume (pressure arises on its own from ice expansion) and under external pressure in an open system.

This approach separates the delivery of the substance across the membrane from the protection itself: the cell doesn't need to stockpile active molecules in advance, which could cause harm. Instead, it gets protection in sync with cooling, without any upfront load.