Shelf Temperature Is Not Product Temperature

Sublimation is endothermic. Converting ice directly to vapour absorbs roughly 2830 J per gram at temperatures near 0 °C, and that energy has to come from somewhere. In a vial on a shelf it comes through the glass bottom by conduction and through the surrounding gas, and the ice front sits at whatever temperature balances the heat arriving against the heat leaving as vapour.

So the shelf can be held at -10 °C while the product sits at -32 °C, and the product will stay there for as long as ice remains. The moment the ice is gone the sink disappears and product temperature climbs toward the shelf within minutes. That climb is one of the signals used to decide primary drying has ended.

This is why a freeze drying cycle is written as a set of shelf setpoints and chamber pressures, but judged against product temperature. The setpoint is an input. The product temperature is the constraint.

What the gap looks like across a cycle

Two things in that table do the work. First, the gap is largest exactly when it matters, during primary drying, because that is when the material is closest to its collapse temperature. Second, the gap closes in secondary drying, because there is no ice left to sublime and therefore no cooling.

Why the gap is a constraint and not a curiosity

An amorphous frozen system has a glass transition temperature of the maximally freeze-concentrated solution, written Tg prime. Above it the concentrated phase softens and the porous structure loses the rigidity holding it up. Collapse temperature measured by freeze drying microscopy typically sits 1 to 3 °C above Tg prime for small-molecule excipient systems. A crystalline system has a eutectic temperature instead, measured by differential scanning calorimetry or by electrical resistance during freezing.

Product temperature has to stay below that limit for the whole of primary drying. Shelf temperature can sit well above it, and routinely does, because the sublimation cooling keeps the product down. Raising the shelf setpoint shortens the cycle right up to the point where the ice front crosses the limit, and past that point the structure gives way.

How product temperature is measured, and what each method costs

Wireless probes and thermocouples both report a single vial. Manometric measurement and duct-based methods report the batch. They disagree, and the disagreement is informative rather than an error in either.

The failure mode: the monitored vial is not the batch

Placing a thermocouple in a vial changes that vial. The probe provides a nucleation site, so ice forms at a higher temperature than in the uninstrumented vials around it, producing larger ice crystals, a more open pore structure and a faster sublimation rate. The instrumented vial therefore finishes primary drying early and reads colder for longer than its neighbours.

Position compounds the problem. Vials at the edge of a shelf receive radiant heat from chamber walls and the door in addition to conduction from the shelf, so their sublimation rate runs higher and their product temperature runs warmer, by several °C in published comparisons of edge and centre vials. If the probes are in centre vials, the warmest material in the load is the material nobody is watching. A run that looks compliant on the trace can contain a ring of collapsed vials at the perimeter.

• Instrumented vials read colder than the batch because the probe promotes nucleation.
• Edge vials read warmer than centre vials because of radiation from walls and door.
• The two biases push in opposite directions, so they do not cancel in any predictable way.
• A batch-level method such as manometric temperature measurement is a cross-check on both, not a replacement for either.

How it shows up afterwards is in appearance, not in the trace: shrunken, glassy or retracted cakes concentrated at one part of the shelf, with residual moisture by coulometric Karl Fischer titration running higher in those same positions.

Where our own record stops

Freeze dryer is not among the specifications Aurum publishes. Product temperature traces is not among the specifications Aurum publishes, shelf setpoint records, collapse temperature determinations or residual moisture results for the material we sell, and any of them is not among the specifications Aurum publishes. What can be assessed from outside the process is the finished cake: its appearance, its structure and whether the seal held.

That is a limit worth stating in the same breath as the physics. A well-described process is not the same as a documented one, and we are describing the first.

Common questions

Why not just set the shelf to the collapse temperature?

Because the shelf is not where the material is. Setting the shelf at the collapse limit would leave the product tens of °C colder and the cycle far longer than necessary. The shelf is deliberately run above the limit and the sublimation cooling keeps the product below it.

Does product temperature stay constant during primary drying?

It drifts upward. As the dried layer above the ice front thickens, resistance to vapour flow rises, the sublimation rate falls, less heat is removed and the ice front warms. A cycle that starts with margin can lose it near the end of primary drying.

Can shelf and product temperature ever be equal?

Effectively yes, in secondary drying and during the freezing hold, because in neither case is large-scale sublimation removing heat. During primary drying they are equal only if nothing is subliming.

Is a colder shelf always safer?

It lowers product temperature, at the cost of a much longer cycle and a longer time at low moisture-removal rates. It is a trade rather than a free margin, and an overlong primary drying phase has its own consequences for throughput and for the total time the material spends in the machine.

Does a thicker fill change the gap?

Yes. A deeper frozen layer means a longer vapour path, higher resistance and a warmer ice front for the same shelf setpoint and pressure. Fill depth is part of the cycle specification for that reason.