
Residual Solvents: 5 Checks Before You Trust A CoA Line
Residual solvents usually appear on a certificate of analysis as one word: conforms. Here is what the line covers, what it leaves out, and the sentence that turns it into numbers.
Table of Contents
The line usually sits near the bottom of the results table, between water content and appearance: Residual solvents — Conforms. No solvent names. No numbers. Just that.
Under time pressure it reads like a formality, and most receiving files close without anyone asking what it covers. That is a mistake, though not because the line is likely to be false. It is a mistake because “conforms” is a conclusion with the reasoning left out, and the reasoning is precisely what a QA reviewer, an auditor, or a downstream customer will ask for six months later.
This article is about that one line: what it reports, what a class of solvent changes about the question, and what to put in writing so the answer arrives before you need it.
Every synthesis runs in solvents. The reaction medium, the crystallisation solvent, the wash before drying — each leaves a residue that drying reduces but does not necessarily eliminate. Residual solvent control is the discipline of measuring what remains and holding it to a limit.
For an intermediate, the relevant solvent list is determined by the route, not by the molecule. A step running in toluene and crystallising from ethanol leaves two solvents to control. Insert a tetrahydrofuran stage and there are three. Change the crystallisation solvent and the whole list changes, together with the limits that apply to it.
This is why the question “what solvents does your material contain?” has no answer in the abstract, and why a supplier who answers it instantly, without asking about the route, is answering from a template.
The measurement itself is chromatographic. Gas chromatography, usually with headspace sampling, separates the volatile residues and reports them against the specification. Results appear as parts per million, or as a statement that the level sits below the method’s limit of quantification — a phrase worth understanding before it appears in your file.
Solvent limits do not come from the customer, the auditor, or the supplier’s preference. They come from toxicological assessment, codified in the ICH Q3C guideline, which sorts solvents into three classes according to the hazard they present and sets limits accordingly.
| Class | Examples | Typical limit | What it means for a buyer |
|---|---|---|---|
| Class 1 — to be avoided | Benzene, carbon tetrachloride, 1,2-dichloroethane, 1,1-dichloroethene | 2–8 ppm for the first four; 1,1,1-trichloroethane at 1500 ppm | Solvents with unacceptable toxicity that a route should not need. If one appears on a certificate, ask why it is in the route at all before discussing the number. |
| Class 2 — to be limited | Toluene 890 ppm, methanol 3000 ppm, dichloromethane 600 ppm, tetrahydrofuran 720 ppm, acetonitrile 410 ppm, hexane 290 ppm | Concentration limits, derived from permitted daily exposure | The working class for most routes. These are the solvents whose numbers are worth reading lot to lot. |
| Class 3 — low toxicity | Ethanol, acetone, isopropyl alcohol, ethyl acetate, methyl ethyl ketone | Usually 5000 ppm (0.5%) treated as acceptable | Normally a formality. A failure here usually points to a drying problem rather than a solvent choice. |
One detail in that table deserves a paragraph of its own. Q3C is written for drug substances and excipients, and its concentration limits assume a daily dose of 10 g or less. An intermediate is not dosed at all; it is consumed in the next step, where part of what remains is purged. A manufacturer may therefore apply the Q3C limits directly, or justify a higher value on the basis of purge in subsequent steps. Both approaches exist in commercial specifications.
What matters to a buyer is that the basis is written down, because a specification quoting 3000 ppm for a Class 2 solvent without stating its basis is a document you cannot defend in a review.
The most useful question about the line is also the simplest: which solvents are in the specification, at what limits, and what did the batch actually measure?
Three answers are possible, and they are not equivalent. The best is a table: solvent name, limit, measured result, method. The middle answer is a measured value for each controlled solvent, sometimes attached as an annex to the certificate. The weakest is the one-word statement, which tells you a comparison was made but not what was compared.
Buyers ask us, reasonably, why a certificate would stop at the one word. The honest answer is that some specifications control residual solvents as a pass-or-fail against a method rather than as a numeric result, and a certificate can only report what the specification asks for.
That is defensible as a document convention. It is not sufficient as an input to a release decision, and the gap is closed by asking, not by assuming.
There is a second reason the line stays short, and it is less flattering. A numeric result invites comparison across lots, and comparison across lots invites questions about drying consistency. A one-word statement resists that scrutiny without technically withholding anything. Most suppliers are not thinking this way. Some are.
A residual solvent result is only as informative as the method behind it. Four details determine whether two numbers can be compared at all.
The first is the method itself: gas chromatography with headspace sampling is the standard approach, but column choice, temperature programme, and sample preparation differ between laboratories. The second is the limit of quantification. A certificate reporting “not detected” for toluene means one thing if the method reaches 20 ppm and something quite different if it only reaches 300 ppm.
The third is the method edition: an in-house method revised two years ago is not the same test as the edition your specification was written against. The fourth is who ran it — the manufacturing site’s own laboratory, or a contracted laboratory whose report is referenced rather than reproduced.
None of these make a result wrong. Together they decide whether the number is comparable to the previous batch, to a competitor’s certificate, or to the specification sitting in your file. When a buyer tells us two suppliers’ certificates “do not match”, the mismatch is usually in the method, not in the material. The certificate reading discipline that covers the rest of the page applies here in full: read the method line as carefully as the result line.
A residual solvent figure does not stand alone. Four cross-checks inside the same file will tell you whether to trust it, and all four take less time than reading the method.
Water content and loss on drying come first. Solvent removal and water removal happen in the same drying step, so a batch that dried poorly usually shows both: water above its usual range and a solvent result at the upper end of its limit. One of the two moving alone is worth a question. Both moving together is a pattern.
Second, the impurity profile. The same chromatographic work that separates solvents often reports related substances, and the impurity work the same certificate reports shares its sample preparation with the solvent test. A batch whose impurity profile is unusually clean while its solvent result sits at the limit is not impossible, but it is uncommon enough to deserve a look.
Third, the lot history. Single certificates say little; sequences say a lot. Three consecutive lots with toluene at 210, 340, and 610 ppm against an 890 ppm limit describe a drying step that is losing performance, and drift between lots shows up here first, months before anything fails. The buyer who asks for the last three certificates is not being difficult. He is reading a trend where others read a snapshot.
Fourth, the specification version and the method edition, compared against the sheet in your own file. If the certificate quotes version 3 of the specification and your file holds version 2, the solvent limits may have moved without anyone telling you. Where the limit itself comes from is a decision your specification should record, precisely so that a later certificate can be checked against it.
Four things, and knowing them prevents most of the arguments that follow a failed incoming test.
It cannot tell you the process is controlled. One conforming batch is one data point; control is a property of a series, demonstrated by trends rather than by a single pass.
It cannot tell you the solvent list is complete. A route change can introduce a solvent that the specification never controlled, and the certificate would still read “conforms” while being accurate about the solvents it covers. This is exactly the failure mode a change-notification clause is designed to catch, and it is why the list of controlled solvents belongs in the quality agreement rather than in a conversation.
It cannot tell you the material suits your downstream step. Purging factors differ between processes; a level acceptable to one customer may be marginal in another’s crystallisation. That assessment belongs to your process chemistry, not to the supplier’s certificate.
And it cannot stand in for your own testing. Your own laboratory will measure it again on arrival, and the value of that second measurement depends on both labs having used a comparable method — which is a question to settle before the first shipment, not after a discrepancy appears.
The remedies are documentary, and they cost one paragraph each.
Name the solvents. At enquiry stage, ask the supplier to declare which solvents the route uses and which ones the specification controls. A supplier who cannot produce that list from the specification has not looked at it recently.
State the basis. Q3C limits applied directly, or limits justified by downstream purge — either is workable, and the choice should appear in the specification rather than be inferred later.
Require the numbers on the certificate. One line in the specification turns “conforms” into a table: solvent, limit, result, method. Suppliers who already work this way will not blink. The others will negotiate, and the negotiation itself is information.
Name the method and its edition, so that a result from this year can be compared with a result from last year without a laboratory investigation. Then add the change-notice obligation: a new solvent in the route, a new method, or a revised limit triggers notification before the next shipment. How the solvent is chosen upstream explains why route changes happen at all — usually for yield or recovery, rarely with the customer’s specification in mind.
Sooner or later a batch fails, or two laboratories disagree. The sequence that follows decides whether the file stays defensible.
Start with the laboratory question. What an out-of-specification result looks like in a well-run system is an investigation before a conclusion: the sample re-prepared, the method checked, the instrument verified, and only then a finding. A supplier who reports an OOS result with an explanation attached has done the work. One who reports a passing retest with no explanation has done something else.
Then the material question. A genuine solvent failure is usually a drying problem, and drying problems are batch-specific: they respond to re-drying, to extended drying time, or to a review of the drying step itself. That makes rework possible in many cases — and rework, properly documented in a deviation report, is a legitimate outcome. What is not legitimate is a certificate quietly reissued with a new result and the same batch number, with no record of what happened in between. The deviation record is what separates the two.
If the disagreement is between your laboratory and the supplier’s, the tie-breaker is a third laboratory working to a method both sides agreed before the sample was sent. That agreement is worth thirty minutes at the start of a supply relationship; without it, a two-week dispute becomes a two-month one.
And if the batch is already in your warehouse and the failure is confirmed, the choice narrows to return, replacement, or a documented concession against a downstream use you can defend. The rejected batch route exists for this situation, and it is less painful when the specification was specific enough to make the rejection a fact rather than an opinion.
One closing observation. In eight years of reviewing certificates with buyers, the disagreements that cost the most time were never about chemistry. They were about documents that recorded a conclusion — conforms, passes, within limits — without recording the basis for it, leaving two parties to argue about what a word meant.
The residual solvents line is the clearest example in the whole file. It is also the easiest to fix, because the fix is a sentence in a specification and one question at the next certificate review. The three dates that decide a document’s fate matter, but they are a different problem. This one is about what the document actually says.
The reference behind the limits, for anyone who wants to read the source rather than the summary: Q3C sits inside the ICH quality guideline set, alongside the Q7 text that governs how intermediates are made.

Residual solvents usually appear on a certificate of analysis as one word: conforms. Here is what the line covers, what it leaves out, and the sentence that turns it into numbers.

A certificate of analysis does not carry its own expiry date. Batch match, retest window and spec version decide whether it still counts — three dates, five desk checks, five minutes.

A certificate of origin states where the drums were made — and nothing else. Which version your shipment needs, and the checks that keep it from stalling at customs.
Tianming Pharmaceutical is a manufacturer of high-quality pharmaceutical intermediates and fine chemicals for research, development and industrial synthesis. We support global B2B partners with process development, custom synthesis and reliable supply, backed by consistent quality documentation.
Products are supplied as pharmaceutical intermediates for research and industrial use only. They are not for human or veterinary use and are not intended for the manufacture of finished pharmaceutical products without independent regulatory authorization.