A synthetic peptide is not assembled in a vacuum. It is assembled in litres of organic solvent, deprotected in a base dissolved in more of it, cleaved in a strong acid, crashed out into an ether, and purified in a water-acetonitrile gradient before it is ever frozen and dried.
Every one of those liquids leaves a trace. And none of them appear on the analytical panel a typical certificate of analysis carries — not because they are undetectable, but because the instruments that produce a purity percentage and a mass spectrum are structurally unable to see them.
It is an independent quality axis, alongside identity, purity, mass accounting and endotoxin — and the one most often absent entirely.
What a peptide is actually made in
The dominant chemical route, Fmoc solid-phase peptide synthesis, runs on a specific and well-known solvent inventory. N,N-dimethylformamide (DMF) — or N-methylpyrrolidone (NMP) — as the reaction medium. Piperidine dissolved in it to remove the Fmoc protecting group each cycle. Dichloromethane for resin swelling and washing. Trifluoroacetic acid with scavengers to cleave the finished chain off the resin and strip side-chain protection. Diethyl ether or tert-butyl methyl ether to precipitate the crude product. Then acetonitrile, water and more trifluoroacetic acid through a preparative reverse-phase column.
The route determines the list. A peptide past the practical limit of chemical synthesis — HGH 191AA at 191 residues, IGF-1 LR3 at 83 — is expressed in a host organism, and its residue profile is host cell proteins, host DNA and endotoxin rather than organic solvents. A compound carrying a D-amino acid, like SS-31 or FOXO4-DRI, cannot be expressed at all and therefore must have passed through the solvent list above. Same logic as synthesis routes and impurity profiles: the route decides which failure modes are possible, and a test only has value if it can detect one of them.
Three classes and a fourth table
The reference framework is ICH Q3C(R9), Impurities: Guideline for Residual Solvents, which reached Step 5 with a legal effective date of 29 April 2024. It sorts solvents by toxicological data rather than by chemistry.
Class 1 solvents should be avoided. Class 2 solvents are limited by a permitted daily exposure (PDE), which converts to a concentration limit in parts per million. Several sit directly on the peptide synthesis list:
- Acetonitrile — PDE 4.1 mg/day, 410 ppm
- N,N-dimethylformamide — 8.8 mg/day, 880 ppm
- Dichloromethane — 6.0 mg/day, 600 ppm
- N-methylpyrrolidone — 5.3 mg/day, 530 ppm
- Methanol — 30.0 mg/day, 3000 ppm
- Pyridine — 2.0 mg/day, 200 ppm
Class 3 solvents have PDEs of 50 mg or more per day — 5000 ppm (0.5%) under the guideline's Option 1 — and are limited by quality-based requirements rather than a toxicological number. Dimethyl sulfoxide, ethyl ether, ethanol, 2-propanol, acetone, ethyl acetate, tert-butylmethyl ether and triethylamine all live here. Option 1 applies the tabulated ppm figure per component; Option 2 lets the mg/day PDE be met across the whole formulation instead.
Then there is Table 4: "Solvents for which no adequate toxicological data was found." It is short. And trifluoroacetic acid is in it — alongside trichloroacetic acid, isooctane, petroleum ether and isopropyl ether.
That is worth sitting with. The single most characteristic residue of peptide chemistry, the acid used for both cleavage and chromatography and the default counterion on the finished salt, has no class, no PDE and no Option 1 ppm figure in the guideline that governs residual solvents. Piperidine is not listed anywhere in Q3C at all. This is part of why the peptide field handles residual trifluoroacetate as a counterion and mass-accounting question with product-specific limits, rather than by looking up a harmonised number — there is no harmonised number to look up.
As with every regulatory frame this library describes: Q3C applies to drug substances and drug products intended for human use. It confers nothing on an unapproved research compound. It is useful here only as a description of what a complete answer to this question looks like.
It takes a different instrument
Residual solvents are measured by USP General Chapter <467>, which specifies static headspace sampling into a gas chromatograph with flame-ionisation detection. Procedures A and B are limit tests; Procedure C is quantitative.
A gas chromatograph. Not the liquid chromatograph that generated the purity percentage.
The reason is not sensitivity. It is method design, and it is close to absurd when stated plainly: you cannot measure residual acetonitrile with a method that uses acetonitrile as the eluent. The analyte is the mobile phase.
DMF does absorb in the low-UV window where peptide purity is read at 210-220 nm — but as a small polar molecule it elutes at or near the solvent front, exactly the region whose integration is an operator choice, as covered in reading the chromatogram and the mass spectrum. Mass spectrometry fares no better: these are small volatiles far below the scan range used for a peptide, introduced in a solvent of the same kind.
So three of the standard panel's answers — purity %, intact mass, sequence coverage — are silent on this axis by construction. The same relationship holds as for endotoxin (needs LAL or a recombinant reagent), copper stoichiometry on GHK-Cu (needs ICP-MS or atomic absorption), and stereochemistry (needs a chiral method). A missing line means nobody ran that instrument.
Loss on drying is not water content
Q3C makes one explicit concession: where only Class 3 solvents are likely to be present and loss on drying is less than 0.5%, a non-specific method such as loss on drying may be used instead of a solvent-specific one.
That sentence defines the limits of a number that does appear on paperwork. Loss on drying measures everything volatile that leaves the sample — water and residual solvent together. Karl Fischer titration measures water specifically. They are not interchangeable, and the guideline permits the non-specific option only on an assumption about which solvents the process used. If a document reports both and the two disagree, the gap is volatile material that is not water.
This folds back into net peptide content: everything in the vial that is not peptide — counterion, water, salts, scavengers, solvent — is mass you weighed and did not get peptide for. Residual solvent is usually the smallest of those terms, and the only one that routinely has no line at all.
The solvent is a reaction variable, not just a residue
Here is the part that makes this worth more than a missing checkbox.
Kumar and colleagues (ChemSusChem 2020;13:5288-5294) replaced DMF with the greener solvent N-butylpyrrolidinone in Fmoc-SPPS and reported lower or equal racemisation in the amino acids most prone to it, and clearly lower aspartimide formation, than DMF gave. Their own framing: a green solvent did not hamper the synthesis and could even improve it.
Run that backwards. The solvent a peptide was built in partly determines which impurities it contains. Epimerisation produces species that are isobaric — identical mass, invisible to intact-mass MS, addressable only by a chiral method. Aspartimide formation produces beta-peptides and racemised material at Asp-Gly motifs, which is a live concern for any sequence carrying one: Epithalon ends in Asp-Gly; BPC-157 carries an aspartate-rich stretch. A residual solvent identity is therefore also a clue about which side reactions the process was running.
And that process is changing. DMF has been restricted in the EU under REACH Annex XVII entry 76 (Commission Regulation (EU) 2021/2030), applying from 12 December 2023 at a 0.3% w/w threshold with narrow derogations — a substance-level restriction rather than a pharmaceutical impurity limit, but one that has pushed manufacturers to design DMF out. The recent literature is full of the attempts: dipropyleneglycol dimethylether (Pharmaceutics 2023;15:1773); mesoporous polymer supports enabling triethyl phosphate, acetonitrile, isopropanol and DMSO mixtures (J Pept Sci 2025;31:e70038, Fresenius Kabi Ipsum co-authors); and, six days before this post, a fully automated aqueous Fmoc-SPPS platform using water and ethyl acetate with no DMF at all, applied to peptides over 30 residues including GLP-1(1-37) (RSC Adv, online 10 September 2026; Biotage-affiliated co-authors and related patent applications disclosed).
None of that is a claim about any product. It is a reason a residual solvent panel is lot-specific and route-specific — so a solvent list reproduced identically across products and years is evidence about a document template, not about material.
Four things to look for
- Is there a solvent line at all, and does it name a method (headspace GC) rather than a bare number?
- Do the named solvents match a plausible route? An organic solvent panel on a recombinant protein, or none on a D-amino-acid peptide that could only have been made chemically, is a mismatch worth noticing.
- Is loss on drying standing in for a water figure, or sitting beside one? They answer different questions.
- Is trifluoroacetate addressed as its own specification? Q3C gives it no limit, so if it is controlled, it is controlled by somebody's product-specific decision.
FAQ
Does residual solvent matter at parts-per-million in a milligram vial? In absolute mass terms it is microgram-scale, smaller than counterion or residual water. Its value on a document is mostly diagnostic: it reports whether drying and purification were controlled and finished, and which chemistry the material passed through. Some residues are also chemically active rather than inert, which is a stability question rather than a mass one.
Why isn't this on more certificates? Because it requires a gas chromatograph with headspace sampling, and a separate validated method per solvent set. It is omitted for the same reason as endotoxin, copper content and chiral purity: a different instrument answering a different question, at additional cost, on a destructive sample. Absence is not evidence of a problem — it means the question was not asked.
Does a greener synthesis mean a better peptide? Not by itself. One comparison found a DMF replacement produced less racemisation and less aspartimide, so solvent choice can move the impurity profile favourably — but that is a finding about a method in a paper, not a property conferred by the word "green." What it establishes is that solvent and impurity profile are the same story read at two ends, and neither is answered by a purity percentage.
More on analytical documentation in /quality/, and compound-by-compound profiles in /library/.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.