Process Development and Scale-Up
What the domain tests
Process Development and Scale-Up tests whether you can move a process from bench to pilot to commercial without losing product quality. The exam emphasizes Quality by Design (QbD): linking material attributes and process parameters to critical quality attributes (CQAs), defining a design space under ICH Q8, and justifying scale-up rules rather than copying them.
Core rules
- Establish CQAs from the target product profile, then map critical process parameters (CPPs) with risk assessment and design of experiments (DoE). A design space is a multidimensional combination of input variables proven to provide assurance of quality; movement within it is not a regulatory change, but leaving it is.
- Scale-up is rarely one rule. For turbulent mixing, constant power per volume (P/V) preserves dispersion; constant tip speed preserves shear; constant Reynolds number preserves flow regime. Choose the rule that matches the limiting mechanism.
- Geometric similarity is an assumption, not a guarantee. Heat-transfer area per volume falls as scale increases, so a bench exotherm may become a commercial bottleneck.
- Tech transfer requires a documented package: process description, CPPs and proven acceptable ranges, analytical methods, equipment requirements, and risk assessment.
Worked example
A 2 L bioreactor with a 6 cm impeller runs at 300 rpm. Scale to 2,000 L with geometric similarity (length scale factor 10). For turbulent flow, P/V ∝ N³D². Constant P/V gives N₂ = N₁(D₁/D₂)^(2/3) = 300 × (0.1)^(2/3) = 64.6 rpm. Tip speed at bench = π × 0.06 × 5 = 0.94 m/s; at commercial = π × 0.6 × 1.08 = 2.03 m/s. So constant P/V more than doubles tip speed, which may damage shear-sensitive cells. If shear is limiting, use constant tip speed instead, accepting lower P/V and possibly poor bulk mixing. The exam wants you to state the limiting mechanism, compute both, and justify the choice.
Traps
- Scaling by volume ratio alone.
- Reporting a single CPP setpoint without a proven acceptable range.
- Ignoring heat removal, feed addition time, or mixing time, which scale differently from P/V.
- Treating a design space as a single point.