Scaling Up an Organic Synthesis: Where Linear Stoichiometry Stops Working

A medicinal chemist scales a 1 mmol Suzuki coupling that worked cleanly in a vial to 50 mmol in a round-bottom flask. The reagent table multiplies through fine: aryl bromide 50 mmol, boronic acid 60 mmol, Pd(PPh₃)₄ 2.5 mmol, K₂CO₃ 100 mmol, THF/H₂O 4:1. The arithmetic is linear. The reaction isn’t. Conversion stalls at 40% and the chromatography looks nothing like the 1 mmol run. This post walks the stoichiometric math that does scale linearly, then names the physical effects that don’t — the failure modes that show up when scale changes by more than ~10×.

What scales linearly: the reagent table

Stoichiometry itself is linear. Equivalents are dimensionless ratios; mmol scales with the limiting reagent; mass and volume scale with mmol. If a 1 mmol reaction uses 0.345 g of substrate, then a 50 mmol reaction uses 17.25 g of the same substrate. There is no nonlinearity in the math.

Catalyst loadings are reported in mol% precisely because mol% is scale-invariant. 5 mol% Pd at 1 mmol is 0.05 mmol of catalyst; at 50 mmol it’s 2.5 mmol. The percentage doesn’t change. So if you build the reagent table in a stoichiometry calculator at the new scale and the equivalents stay the same as the 1 mmol procedure, the math is already done.

What doesn’t scale linearly: the physics around the molecules

The four effects below show up reliably at 10–100× scale-up and are responsible for most "the reaction stopped working at 50 mmol" failures. None of them are stoichiometric problems — the reagent table is correct — but all of them change what the molecules experience at the larger scale.

1. Surface-area-to-volume ratio drops

A 1 mmol reaction in a 4 mL vial has high surface area relative to volume. Heat dissipates fast, gas evolves freely, and the headspace mixes with the liquid easily. A 50 mmol reaction in a 250 mL round-bottom has dramatically lower surface-area-to-volume — roughly proportional to V^1/3/V, or about 3.7× lower. Exotherms that vented as warmth at 1 mmol can run away at 50 mmol. Reactions that needed brief reflux at small scale may need extended reflux at larger scale because heat input through the flask wall is also surface-limited.

2. Mixing time grows

Magnetic stirring at 1 mmol with a 1 cm stir bar gives near-instantaneous mixing — the reaction is effectively well-mixed. At 50 mmol in 100 mL of solvent, a 2 cm stir bar may take seconds to fully mix a reagent addition, and biphasic reactions (like the aqueous/organic Suzuki) need vigorous stirring or mechanical agitation. Slow mixing creates transient concentration gradients: a reagent added dropwise sees high local concentration before it mixes, which can favor side products. For sensitive additions (BuLi, NaH-mediated deprotonations), assume a 1 mmol "all at once" is not equivalent to a 50 mmol "all at once" — switch to dropwise addition.

3. Concentration drift when you don’t scale solvent linearly

Reaction concentration matters. A 1 mmol coupling in 2 mL of solvent is 0.5 M. The natural temptation at 50 mmol is to use 50 mL ("the same ratio") — which is correct — but practitioners often choose round-number volumes (75 mL, 100 mL) for glassware fit. That changes the concentration. Dilution can starve bimolecular reactions; concentration can favor oligomerization. If you change scale, hold reaction concentration constant unless you have a deliberate reason to change it.

4. Catalyst-poisoning thresholds shift

Pd-catalyzed reactions are sensitive to trace contaminants — oxygen, water above tolerance, sulfur impurities in solvents. At 1 mmol with 5 mol% Pd, you have 0.05 mmol of catalyst. A small impurity burden may consume some catalyst but leave enough working catalyst to complete the reaction. At 50 mmol with the same 5 mol% (2.5 mmol Pd), a proportionally larger impurity load comes in with the larger reagent masses and solvent volumes. The impurity-to-catalyst ratio can shift. Practitioners running first-time scale-ups often increase catalyst loading to 6–8 mol% and degas more rigorously, then dial back on subsequent runs.

A worked example: 1 mmol to 50 mmol Suzuki coupling

Where this gets harder: scaling past 100 mmol

The shifts above are manageable in academic-scale flasks (50–500 mL). Beyond that — multimolar runs, multi-liter solvent volumes, anything that needs a reactor rather than glassware — the territory shifts from "synthetic chemist with good technique" to process chemistry. Heat-transfer modeling, mass-transfer coefficients, residence-time distributions for flow systems, and dose-rate control for exothermic additions all enter the picture. The reagent table still scales; the equipment, controls, and safety analysis don’t.

For most medicinal-chemistry and academic synthesis work, you’ll stop at 50–200 mmol per batch and split larger campaigns into multiple batches at known-good scale. Splitting is almost always faster than re-optimizing a new scale-up.

Edge cases worth naming

• Highly exothermic additions (BuLi, LDA generation, Grignard formation): at >10 mmol scale, switch from "all at once" to controlled dropwise via addition funnel. The exotherm that warmed a vial can boil out 100 mL of THF.
• Air-sensitive reactions with hygroscopic reagents: KOH, NaOH, K₃PO₄ all gain water on the balance over minutes to hours of weighing time. At 50 mmol you’re weighing grams of these, which extends weighing time, and the actual mol delivered may be 5–15% lower than nominal. Weigh under inert atmosphere or apply a purity correction based on known water content.
• Macrocyclization and other reactions deliberately run dilute (typically 0.001–0.01 M): scaling these linearly means scaling the solvent volume too, which becomes physically awkward fast. A 50 mmol macrocyclization at 0.005 M would need 10 L of solvent. These reactions usually don’t scale in single batches — high-dilution syringe-pump addition is the standard approach.
• Concentration-dependent selectivity: any reaction where the product/byproduct ratio depends on concentration (oligomerization, dimerization, intramolecular vs. intermolecular pathways) requires careful concentration preservation. Don’t round solvent volume to fit glassware — choose glassware to fit the concentration.

The practical scale-up checklist

Before running a scaled-up reaction:

1. Build the reagent table at the new scale. Confirm every mass and volume against the original procedure ratio.
2. Hold reaction concentration constant. Calculate the solvent volume from scale and target concentration, not from glassware fit.
3. Plan addition order and rates. Anything exothermic or anything that generates gas needs controlled addition at >10 mmol scale.
4. Choose glassware that accommodates the volume plus 50% headspace. Reactions evolve gas, foam, and reflux.
5. Consider catalyst bump (e.g., 5 mol% → 6–7 mol%) for the first run; dial back once the procedure is proven at scale.
6. Increase monitoring frequency. The first half-life of the reaction may shift; you want to see when conversion stalls, not how long it takes overall.

The stoichiometry calculator is the place to do step 1 reliably. Steps 2–6 are practitioner judgment that no calculator captures — but if the reagent table is wrong, the rest doesn’t matter.