Why the discovery route usually fails at scale
The route a medicinal chemist uses to make a few hundred milligrams is optimised for *speed and flexibility*, not for kilograms. As a candidate advances toward preclinical studies and trials, someone must make much more active pharmaceutical ingredient reproducibly and safely. That is the job of process chemistry, and it almost always means redesigning the route, not just running the old one bigger.
What process chemists optimise for
- Safety — eliminate or contain hazardous reagents (azides, strong oxidants), and characterise every exotherm before scaling.
- Robustness — favour steps with wide tolerance to temperature, stoichiometry, and timing so the process survives a real plant.
- Purification by crystallisation, not chromatography — design intermediates that crystallise cleanly to control impurities.
- Cost and greenness — cut steps, swap precious-metal catalysts where possible, and reduce solvent and waste.
Step count matters enormously. A shorter, more convergent route with high-yielding, crystallisation-controlled steps beats a clever long one every time at scale. Process chemists will also revisit how chirality is set: an early asymmetric catalytic step that avoids a wasteful late resolution can transform the economics of a single-enantiomer drug.
Quality, impurities, and GMP
Material that goes into people must be made under [[med-good-manufacturing-practice|good manufacturing practice]], with documented, validated, reproducible steps. A central concern is the impurity profile: every reagent, by-product, and especially any genotoxic impurity must be controlled to tight limits. A discovery route that leaves a trace of a mutagenic alkylating agent is unacceptable for an API, even if the final compound is pure enough to test in a dish.