Why A Lyophilized Cake Collapses

Freeze-drying separates a frozen solution into two departures. Ice sublimes away during primary drying, leaving a void network. Water still bound to the solute leaves later, during secondary drying, by desorption. The cake keeps its shape only because the freeze-concentrated solute between the ice crystals is stiff enough to stand while the ice is being removed from around it.

That stiffness is temperature-dependent and the dependence is sharp. Above a characteristic temperature, the concentrated phase behaves as a viscous liquid rather than a solid, and viscosity falls by orders of magnitude across a few degrees. Once it flows, the walls between the voids cannot bear the structure and the cake collapses. For an amorphous material the threshold is the glass transition temperature of the maximally freeze-concentrated solution, written Tg prime. For a crystalline one it is the eutectic melt temperature, Teu, and the equivalent failure is called melt-back.

Collapse is therefore a thermal event, not a contamination event or a sign that the molecule itself has changed. It is also permanent: a collapsed cake does not recover on further drying.

The temperatures that set the limit

The last row matters for cycle design. Collapse onset observed under a microscope is usually a little above the calorimetric Tg prime, which is why a cycle is normally designed to hold product temperature 2 to 5 °C below the measured collapse temperature rather than exactly at it. Product temperature, not shelf temperature, is the controlled variable, and the two are not the same number during sublimation.

The failure mode: chamber pressure drifting up mid-run

The single most common route to collapse in practice is not a shelf setpoint error. It is loss of pressure control during primary drying. Sublimation cools the product surface, and that evaporative cooling is what keeps product temperature well below the shelf. If chamber pressure rises, the driving force for sublimation falls, the cooling falls with it, and product temperature climbs toward the shelf temperature within minutes.

Causes are mundane: a condenser approaching its capacity, a partially iced vapour duct, a leak in the chamber, or a vacuum pump losing performance. The signature in the data is a pressure trace stepping up while the product thermocouples rise and the pressure difference between a Pirani and a capacitance manometer narrows earlier than expected. A cycle that has been running correctly for hours can lose the whole batch in the last third of primary drying.

Collapse is also distinct from a cake broken in transit. A cake that has fractured into pieces but whose pieces are still white, matte and porous was mechanically damaged, not thermally collapsed.

What collapse changes about the material

The consequences are physical and measurable rather than speculative. A collapsed cake has lost most of its internal surface area, and internal surface area is what makes secondary drying efficient, so residual moisture in collapsed cakes is routinely several times higher than in intact ones from the same formulation. Values of 4 to 8% w/w by Karl Fischer where the intact cake sits near 1 to 2% w/w are reported in the freeze-drying literature.

• Residual moisture rises, because the pore network that carried water vapour out has closed.
• Dissolution slows, because the solvent front has to penetrate a dense mass rather than a foam.
• Appearance fails visual inspection: shrinkage, retraction from the wall, glassiness, or a puddle at the base.
• Between-vial uniformity degrades, because edge vials receive more radiant heat than centre vials and collapse first.

The edge-vial point is worth stating plainly. Vials at the perimeter of a shelf receive radiation from the chamber wall and door in addition to conduction from the shelf, and their sublimation rate can run 10 to 30% higher than centre vials in the same run. A batch can therefore contain intact centre vials and collapsed edge vials at the same time, which is why appearance is checked across the load rather than on a single sample.

What Aurum measures here, and what it does not

Aurum does not run the freeze-drying cycles for the material we sell and does not hold the cycle records: shelf temperature profiles, chamber pressure traces and product thermocouple data sit with whoever performed the lyophilization. We therefore cannot state the collapse temperature of any formulation we list, nor confirm how far below it a given run was held.

We also do not routinely determine residual moisture by Karl Fischer on incoming lots, which is the measurement that would quantify the consequence described above. What can be reported is appearance, which is observable, and chromatographic purity, which is measured. Neither of those substitutes for a moisture figure, and where a research question depends on residual moisture, our documentation does not answer it.

Common questions

Does a collapsed cake mean the peptide has degraded?

Not by itself. Collapse is a structural event. Whether the chemistry has changed is a separate question answered by chromatography and mass measurement, not by appearance. The link between the two is indirect: higher residual moisture raises the rate of hydrolysis and deamidation over time.

Can a collapsed cake be re-dried?

The lost structure does not come back. Further drying can lower moisture somewhat, but it does so slowly against a closed pore network, and the appearance defect is permanent.

Why do some materials collapse and others do not at the same shelf temperature?

Because the critical temperature is a property of the formulation, not of the machine. An amorphous sugar system with a Tg prime near -32 °C has far less thermal headroom than a crystalline mannitol system melting near -1.5 °C.

Does adding a bulking agent prevent collapse?

It changes the critical temperature and the mechanical strength of the scaffold, which can raise the headroom available. A crystallising bulking agent such as mannitol gives a stronger cake structure than an amorphous one at the same solids loading.

Is a shrunken cake always collapse?

No. Low total solids, commonly below about 2% w/v, produce thin, fragile cakes that shrink or crack without any thermal event. Volume and solids loading have to be known before appearance can be interpreted.