A 20 mg research batch and a 10 g batch of the same peptide normally start from the same chemistry, the same resin and the same coupling reagents. Whether they can stay identical is something the larger batch has to demonstrate, because resin, reagents, solvent volumes or reaction times may all need to change as scale increases. What differs is that at scale the mixture starts heating itself, the chromatography stops being a formality, and every percentage point of crude purity turns into real money. The useful question is not whether a sequence can be scaled, but which stage will become the bottleneck and what can be settled in advance to avoid it.
Chain assembly scales close to linearly, because nearly every reagent volume in a solid-phase protocol is specified per gram of resin: solvent for swelling and washing, deprotection solution, coupling solution. Doubling the resin roughly doubles these quantities, and this first stage is where scaling comes closest to arithmetic. Even here it is only a starting basis: mixing, heat transfer, resin swelling and reagent distribution all change with vessel size, so coupling efficiency has to be re-checked at the new scale rather than assumed. The three stages after it behave differently.
Loading sets how many peptide chains sit per gram of support, and therefore how much product a given vessel can make. For routine sequences, substitution in the range of about 0.5–0.8 mmol/g works well; for longer or difficult sequences the standard advice is to drop to roughly 0.1–0.2 mmol/g, because crowded chains on a highly substituted resin aggregate and couple incompletely1.
At scale this becomes a genuine conflict rather than a preference. Process efficiency rises with loading — more product per vessel, per litre of solvent, per hour of operator time — while on-resin precipitation falls as loading drops. Industrial practice treats the optimum as something to be found by experiment for each sequence and protocol, typically somewhere between 0.2 and 0.9 mmol/g2. Choosing high loading to save a vessel and then losing half the batch to deletion sequences is the most expensive mistake available at this stage, and it is made before a single residue is coupled.
Rule of thumb: loading is a throughput lever at the start of a synthesis and a purity penalty at the end of it. If the sequence has already shown signs of aggregation at research scale — sluggish Fmoc removal, deletion peaks in the crude — the larger batch should be made at lower loading, not the same loading in a bigger reactor.
Cleavage cocktail is dosed per gram of resin, usually around 10–15 mL/g, freshly mixed immediately before use3. That part does scale. What does not scale is heat removal. A flask holding a few hundred milligrams of resin loses heat easily; a vessel holding several hundred grams has far less surface per unit volume, and the acidolysis is exothermic. Process development work on the manufacture of a therapeutic GLP-1 receptor agonist peptide recorded a temperature rise of 5–8 °C on cleavage, noted that this promotes side products and depresses yield, and investigated pre-swelling the peptide-resin in an organic co-solvent specifically to blunt the exotherm4. A self-heating strong-acid mixture containing volatile thiol scavengers is also a safety question long before it is a yield question.
The same work is a useful corrective on cocktail composition. Scavenger loading is not a case of more being safer: increasing the 1,2-ethanedithiol content significantly reduced the yield, and omitting several components altogether increased it. Cocktail volume could be cut by up to 30% with no effect on yield, though larger reductions did cost yield. A simpler cocktail was preferred at production scale for the practical reasons — less analysis, lower cost, easier handling4.
The operational conclusion is unglamorous and widely recommended: test the chosen cocktail and timing on 50 mg or less before committing a large batch to it3. A cleavage that is 10% suboptimal on a test scale is an annoyance; the same 10% on 300 g of peptide-resin is a repeat synthesis. For the chemistry behind cocktail choice, see our page on cleavage cocktails and scavengers.
Reversed-phase chromatography scales by a single well-defined rule. Keep the stationary phase, particle size and bed length constant, and increase flow rate and sample load in proportion to the column's cross-sectional area; the separation then transfers approximately unchanged5. A 20 mm i.d. column has about nineteen times the cross-sectional area of a 4.6 mm column, so flow and load scale by roughly nineteen5. The same area ratio is what sets loading, flow and eluent consumption at production scale6.
The number that makes this concrete is loading capacity. In elution-mode preparative RP-HPLC, synthetic peptides typically load at about 1–2 mg per mL of packed column volume — on the order of 0.1–0.2% of the column volume7. That is a ceiling set by the separation, not by the pump, and it is where the arithmetic becomes unforgiving.
| Target | Crude purity | Crude to inject | Column work |
|---|---|---|---|
| 5 g purified | 60% | ~8 g (before recovery losses) | Baseline |
| 5 g purified | 40% | ~13 g | About 1.5× baseline |
| 5 g purified | 20% | ~25 g | About 3× baseline |
The column accepts roughly the same mass of crude regardless of what fraction of that mass is the product, so a poor crude multiplies injections, solvent volume, fraction analysis, pooling and drying time all at once. This is why effort spent on assembly — lower loading, pseudoproline or backbone-protected building blocks, better coupling on a known problem residue — is usually cheaper than the chromatography it saves, though the balance is process-specific. It is also why a quoted price for a gram-scale peptide depends far more on the sequence than on the mass.
Expectations set by short peptides do not survive contact with long ones. A published synthesis of an 82-residue chimeric peptide on a low-loading PEG resin, reduced from 0.60 to 0.20 mmol/g precisely to limit aggregation, obtained a crude yield of 9.98% by standard synthesis and 13.06% with microwave assistance, with recoveries after purification of 9.16% and 12.67% respectively; the microwave route also cut total synthesis time from 162 to 37.8 hours8. Those are competent results for a sequence of that length, not poor ones.
The practical reading is that overall yield falls with length and with sequence difficulty, and it falls before purification rather than during it. A gram of an 82-mer is a different proposition from a gram of a 12-mer, and the resin charge required differs by far more than the length ratio suggests. See long peptide synthesis for the strategies that apply above roughly 50 residues.
| Scales linearly | Reagent and solvent volumes, dosed per gram of resin |
| Needs scale-specific design | Heat removal during cleavage; chromatographic load per injection |
| Capacity-limited | Lyophilisation, by shelf area and condenser duty |
| Typical prep loading | ~1–2 mg crude per mL of packed column volume |
| Decides the cost | Crude purity, which is set during assembly |
LifeTein synthesises peptides from milligram research quantities up to multi-gram scale at its New Jersey facility, with HPLC and mass spectrometry characterisation on every batch. If you are moving a sequence from a pilot experiment to a larger study, send us the sequence and the amount you need in hand — we will tell you what the crude is likely to look like and what that means for the price, before you commit.