Scaling Considerations
Scaling a freeze drying process from laboratory to production requires careful consideration of several factors:
- Heat transfer: Lab units typically use radiation heating, while production units use conduction or convection. The heat transfer coefficient changes significantly.
- Chamber pressure: Maintaining uniform pressure across a large chamber is more challenging than in small lab units.
- Ice thickness: Production units must accommodate thicker ice buildup on condenser coils.
Recommended Approach
We recommend a stepwise scale-up: lab (0.4m²) → pilot (4m²) → production (40-100m²). Each step should be validated with process parameters (temperature, pressure, time) adjusted based on the equipment characteristics.
What Scales and What Does Not
Drying is limited by two transports: heat reaching the frozen layer, and water vapour leaving it. In a lab unit the sample is thin and surrounded by radiant heat, so both paths are short. In production the product sits on a shelf, heat arrives mainly through the tray, and the drying front has to travel outward through a layer that is already dry. Loading density — kilograms per square metre of shelf — is therefore the meaningful scale factor, not chamber volume. Two chambers of the same volume with different shelf areas hold different loads and release water vapour at different rates.
Validating Each Step
Formulation, unit fill weight and freezing profile are settled before scale-up and should not be changed between steps: freezing rate determines ice crystal size, and with it both drying behaviour and final texture. What is re-tuned at each step is the shelf temperature ramp, the chamber pressure set point and the cycle time. Record product temperature rather than shelf temperature, with the probe in the slowest-drying position; check the wettest samples rather than the average; and watch when the condenser begins to lose vapour removal capacity as ice accumulates on the coils. A single successful run is a starting point for the next scale, not a proven process.
Common Pitfalls
- Scaling by ratio: Extrapolating cycle time by area or volume ratio assumes heat and vapour transport stay proportional. They do not.
- Unrepresentative samples: A small sample on an empty shelf behaves differently from a full load, where vapour from the middle of the load has further to travel.
- Ignoring the condenser: The later part of a production cycle runs against a frosted coil, so vapour removal is slower than the lab unit ever showed.
- Changing two things at once: If the freezing profile and the shelf ramp are altered in the same trial, neither result can be attributed.