Gram to kilo: the four things that change when a route is scaled

A route that works at 10 grams and fails at 2 kilograms has usually failed for one of four reasons, and none of them is “the chemistry stopped working”.

1. Heat removal, not heat input

At gram scale a reaction flask loses heat to the room faster than an exotherm can build. At kilo scale the surface-to-volume ratio falls by an order of magnitude, and the same exotherm that was invisible becomes a runaway risk. This is why the addition rate is a scale-dependent variable: the same number of moles added over the same number of minutes can be safe in one vessel and unsafe in another.

2. Mixing time becomes a real time

In a small flask, “add the reagent” takes a second and the local concentration spike disappears immediately. In a reactor, the same instruction can produce a region of high concentration for tens of seconds. Reactions that are sensitive to local stoichiometry – deprotonations, halogen-metal exchanges, additions to aldehydes – are the ones that show it first, usually as an increase in a dimeric or over-addition impurity.

3. Work-up is no longer the cheap part

Extraction of a litre is a morning; extraction of 200 litres is a day and a drum of solvent. Phase separation that took two minutes in a sep funnel can take an hour in a reactor, and emulsions that were a nuisance become the bottleneck. This is the step where a route that was “fine” at gram scale turns out to need a different solvent.

4. The impurity you could not see becomes visible

At gram scale, a 0.5% impurity is 5 milligrams – often below the detection threshold of the method used to check the product. At kilo scale the same 0.5% is 5 grams, and it is now the thing the customer’s QA group asks about. The impurity did not appear; the scale made it measurable.

What to fix before the first kilo lot

Three things, in this order: the addition rate and the cooling capacity, the work-up and the phase separation, and the analytical method at the impurity level you will actually need to control. The first two are engineering; the third is the one people skip, and it is the one that delays release.

Custom routes are quoted in stages for this reason – see custom synthesis for how a feasibility lot is separated from a scale-up.