Glass Transition And Eutectic Point: The Two Temperatures A Cycle Is Built Around

Freezing an aqueous solution does not freeze the solutes. Water crystallises out as pure ice, and everything dissolved in it is pushed into the shrinking liquid between the ice crystals. That residual phase becomes steadily more concentrated as freezing proceeds, which is why it is called the freeze-concentrated solution. What happens to that phase at low temperature divides formulations into two families.

If the solutes crystallise, the system reaches a eutectic point: a single temperature at which ice and solute crystals melt together to a liquid of fixed composition. Sodium chloride and water do this at -21.6 °C, measured by differential scanning calorimetry as a sharp endotherm. Below that temperature everything is solid. Above it, liquid is present.

If the solutes do not crystallise, there is no eutectic point at all. The freeze-concentrated phase keeps getting more viscous until it vitrifies into an amorphous glass. The temperature at which that happens is written Tg prime, the glass transition temperature of the maximally freeze-concentrated solution, and it appears in a calorimetry scan as a step change in heat capacity rather than a peak. Most peptide formulations containing sugars fall here.

The numbers, and how they are obtained

Two instruments do the work. Differential scanning calorimetry heats a small frozen sample at a controlled rate and records heat flow, which resolves a eutectic melt as an endothermic peak and a glass transition as a baseline shift, typically a few tenths of a joule per gram per kelvin. Freeze-dry microscopy watches a thin frozen film dry under vacuum through a microscope and records the temperature at which the drying front loses structure, which is the collapse temperature. Collapse temperature commonly sits 1 to 3 °C above Tg prime, because on the short timescale of drying a slightly rubbery matrix can still hold its shape.

These are formulation-specific measurements, not lookup values. Adding a buffer salt to a sucrose solution lowers Tg prime; the exact figure for a given composition has to be measured on that composition.

Why primary drying is slow, and why it is deliberately slow

Sublimation is endothermic, so the drying front sits colder than the shelf beneath it. Product temperature during primary drying is the outcome of shelf temperature, chamber pressure and the resistance of the already-dried layer above the front, and it rises as the dried layer thickens and that resistance grows. The design question is how close the product temperature can be run to the critical temperature without crossing it.

A common target is a product temperature 2 to 5 °C below the measured critical temperature throughout primary drying. Running colder than that is safe and wasteful: sublimation rate roughly doubles for every 5 °C the product temperature rises, so an unnecessarily conservative cycle can take twice as long as it needs to. Running warmer is fast until it is catastrophic, because the penalty is not gradual.

The failure mode: collapse above Tg prime

The characteristic failure is cake collapse. Once the freeze-concentrated phase crosses Tg prime it stops behaving as a solid; viscosity falls from the region of 10 to the twelfth pascal seconds, which is the conventional definition of the glassy state, by orders of magnitude over a few degrees. The matrix then flows into the pores that sublimation has just opened, and the pore network that vapour was escaping through closes.

What follows is self-reinforcing. A collapsed layer has much higher resistance to vapour flow, so the front below it heats further, so more of it collapses. The visible result is a shrunken, glassy or syrupy plug instead of an open porous cake, often retracted from the vial wall. In a crystalline system the equivalent event is eutectic melt-back, where the frozen mass liquefies and the cake structure is simply never formed.

The measurable consequences are consistent. Residual moisture is higher, because the escape route for water closed while water was still leaving: collapsed cakes routinely show 3% to 6% water by Karl Fischer titration where an intact cake from the same formulation shows below 1%. Time to redissolve lengthens. Specific surface area, by gas adsorption, falls sharply.

• Partial collapse is common and looks like a cake that has slumped or glazed on one face rather than a fully liquefied plug.
• Vials at the edge of a shelf receive additional radiant heat from the chamber wall, so edge vials collapse first when a cycle is run close to the limit.
• A cake that collapsed during drying and a cake that collapsed later from moisture ingress look similar in the vial and cannot be distinguished by appearance alone.

What appearance can and cannot establish

An intact, uniform, full-height cake is evidence that product temperature stayed below the critical temperature for that formulation. It is not evidence of the peptide's chemical state, because collapse and chemical degradation are separate processes with separate causes. Conversely, a collapsed cake is not by itself evidence of chemical change; it is evidence that the physical structure was lost, which raises residual moisture and therefore raises the rate of moisture-dependent reactions over time.

Where we stand on this

None of the numbers in this article are among the specifications Aurum publishes. Tg prime, eutectic temperature, collapse temperature, shelf and product temperature traces and residual moisture figures are not part of the published record for any material Aurum lists, and nothing here should be read as a claim about them. What is published is identity and purity: purity independently assayed by reverse-phase HPLC, identity assayed by mass spectrometry. Those are chemical measurements on the material, and they say nothing about the cycle that dried it. What cake appearance shows is physical structure, and that is a separate question from either.