Research Library · Proof
How to read a certificate of analysis
A certificate of analysis (COA) answers three questions about a vial: which molecule it holds, how clean that molecule is, and how much of it is actually there. Those are different questions, and a document that answers only one is not a certificate — it is an advertisement with a table.
Below is every line of the 14-Point Quality Framework, read the way an analyst reads it: what the line looks like, what belongs on it, and what the fake version does instead. The certificate shown is a made-up one in the usual layout — not any single supplier’s, ours included — so it reads the same whoever your vial came from.
| Certificate of analysis · Lot EXAMPLE-0000 | |
|---|---|
| Appearance | White lyophilised powder |
| Identity by mass spectrometry | Conforms |
| Purity by HPLC (area %) | 99.2 % |
| Peptide content | 8.20 mg / vial |
| Assay (% of label claim) | 82.0 % |
| Net peptide content | 82.0 % |
| Water content (Karl Fischer) | 6.1 % |
| Counterion (trifluoroacetate) | 11.4 % |
| Single largest impurity | 0.31 % |
Any one of the three teal lines gives you the peptide mass. Peptide content in milligrams is the most direct — use it as-is. Assay as a percentage of label claim, and net peptide content as a percentage of the solids, both multiply the labelled amount. In this example all three agree at 8.20 mg out of a 10 mg vial.
The amber line does not. HPLC purity is an area percentage of the chromatogram — it says how much of what is in the vial is the right molecule, not how many milligrams are there. A vial can be 99.2 % pure and still hold only 8.2 mg of peptide, because the rest of the mass is counterion and water and neither shows up as an impurity peak. Purity, quantity and potency are three separate measurements.
Before you read a single number: who hosts this document?
Look at the address bar first. A real certificate lives on the testing laboratory’s own domain, where the seller cannot touch it. A PDF served from the seller’s website is a document the seller controls. A screenshot in a chat is not a document at all. A QR code the seller printed that points at the seller’s own site proves only that the seller owns a printer.
Until you know who hosts the file, every number on it is a claim, not a record. Our certificates are hosted by the laboratory that produced them.
Any of ours can be checked at Verify a Certificate.
The fourteen lines, in certificate order
Eleven of these run on every lot without exception. Three — aggregation, potency, and secondary structure — are qualification-tier: run whenever the synthesis route, the container, or a supplier changes, with the certificate stating which qualification run covers your lot.
Every instrument below is explained in Analytical Methods.
1. Identity — high-resolution LC-MS/MS
The line: “Identity by high-resolution LC-MS/MS” — liquid chromatography feeding a mass spectrometer — with a measured mass beside the theoretical mass.
Expect: Two masses agreeing to several decimal places before any “conforms,” and full sequence confirmation on the first lot from a new manufacturing site.
Red flag: “Conforms” with no mass shown, or identity claimed from an HPLC retention time — a retention time is a habit, not a fingerprint. And the central trap: a chromatogram can show one flawless peak and still be the wrong molecule. A purity instrument measures how much of one thing, never which thing. Purity cannot rescue identity.
2. Purity — RP-HPLC with full-spectrum detection
The line: “Purity by RP-HPLC (area %), PDA detection.” The instrument separates the vial’s contents into peaks; purity is the main peak’s share of the total area. A full-spectrum (PDA) detector watches every wavelength at once, so an impurity invisible at one shows at another.
Expect: An area percentage to one decimal, the method named, and the single largest impurity reported separately — 99.2% and 0.31% on the example lot above.
Red flag: A round “99%” with no decimal, no method, and no impurity line. Instruments do not produce round numbers; people do.
3. Net peptide content — qNMR
The line: “Net peptide content by qNMR” — quantitative nuclear magnetic resonance, cross-checked by amino acid analysis. The most commercially consequential number on the page.
Expect: Milligrams of peptide per vial and a percent of label claim. On the example lot, a vial labeled 10 mg carries 8.20 mg of peptide — 99.2% pure and 82.0% of label at the same time, both true. Purity says how clean. Content says how much. Purity is not content, and the label is not this number.
Red flag: No content line at all — the market’s favourite omission — or content “calculated from purity,” which is arithmetic posing as measurement.
4. Water content — Karl Fischer
The line: “Water content by Karl Fischer titration” — a chemical method that counts water directly instead of drying and weighing.
Expect: An honest percentage. Lyophilized (freeze-dried) powder always holds some water; the example lot reads 6.1%. It is measured because water is weight that is not peptide.
Red flag: “0.0%,” a number nobody measured, or no water line on material sold by gross weight.
5. Counterion content — ion chromatography
The line: “Counterion by ion chromatography,” naming the salt the peptide carried out of synthesis — trifluoroacetate (TFA) or acetate.
Expect: The salt named, with its share of the mass — 11.4% TFA on the example lot. Salt is part of what the vial weighs, so a milligram means nothing until you know the salt form.
Red flag: No salt form stated anywhere, or “TFA-free” claimed with no ion chromatography behind it.
6. Structural integrity — class-specific
The line: For bridged peptides, “disulfide connectivity by peptide mapping”; where loose sulfur ends must not exist, “free thiol by Ellman’s assay.”
Expect: On a molecule held in shape by disulfide bridges, evidence the right pairs are connected; on a molecule without them, an honest “not applicable.”
Red flag: “N/A” on a peptide whose structure plainly contains bridges. That is not a skipped test; it is a skipped question.
7. Residual solvents — GC-MS headspace
The line: “Residual solvents by GC-MS headspace” — the instrument samples the vapor above the material and identifies what synthesis chemicals remain.
Expect: Solvents named one by one, results in parts per million against stated limits, and “not detected” entries that still show the detection limit.
Red flag: The single word “Pass” with nothing named and nothing counted. Pass against what?
8. Elemental impurities — ICP-MS
The line: “Elemental impurities by ICP-MS” — a plasma instrument that counts metal atoms, including palladium, a catalyst some syntheses rely on.
Expect: Each metal on its own row with its own figure and limit.
Red flag: “Heavy metals: pass” as one line. Metals have names; a real report uses them.
9. Endotoxin and bioburden — USP <85>, <61>/<62>
The line: “Endotoxin (USP <85>)” and “Bioburden (USP <61>/<62>).” Endotoxin is residue from bacterial cell walls that outlasts the bacteria; bioburden counts microorganisms still alive.
Expect: Endotoxin as a number in endotoxin units per vial against a stated limit, and bioburden as a count, with the USP chapters cited — they define how the test is run.
Red flag: The word “sterile” with no chapter, no unit, and no number behind it.
10. Content uniformity — USP <905>
The line: “Content uniformity (USP <905>)” — vials pulled from across the lot and measured individually.
Expect: How many vials were sampled and the computed result against the compendial limit.
Red flag: The line is missing. A certificate can be perfectly true for the one vial tested and useless for the one you received; uniformity is what ties a lot together.
11. Aggregation state — SEC-MALS, qualification tier
The line: “Aggregation by SEC-MALS” — size-exclusion chromatography with light scattering, which sees whether peptide molecules have clumped.
Expect: A result, or a dated reference to the qualification run covering this lot. Clumping can hide from a purity column, so a clean area percentage does not answer it.
Red flag: No line and no reference — the question was never asked.
12. Relative potency — SPR or functional assay, qualification tier
The line: “Relative potency versus reference standard” — does the molecule bind its target the way the reference standard does, measured in an instrument.
Expect: A percentage relative to the reference, with the assay named.
Red flag: Potency talk in the marketing and no potency line on the certificate. What is worth claiming is worth measuring.
13. Secondary structure — CD or FTIR, qualification tier
The line: “Secondary structure by CD or FTIR” — spectroscopy that reads a peptide’s fold, for the molecules that have one.
Expect: A spectrum compared against reference material, or an honest “not applicable” for a peptide with no defined fold.
Red flag: Folding language with no spectroscopy named anywhere.
14. Container closure integrity — vacuum decay or headspace
The line: “Container closure integrity by vacuum decay or headspace analysis” — a test of the sealed vial itself, not of the peptide.
Expect: The method named and evidence the stopper and crimp keep moisture out for the vial’s whole life.
Red flag: Never mentioned. A perfect powder in a leaking vial is a temporary situation.
Certificates that look real
Fakes rarely look fake. Learn the practices, not the logos.
- Recycled batch numbers: the same lot number on certificates months apart, or on two different products — lot numbers are supposed to die with their lot.
- One purity number across a whole catalog: every product 99-point-something; different molecules do not share one result.
- No method named: “Purity: 99%” with no instrument, no detector, no laboratory — a number with no method is a sentence with no verb.
- No mass spectrum shown: identity, the question that matters most, asserted with nothing to check.
- And one tell in the other direction: a public record with no failures in it. A laboratory that tests honestly fails a lot sometimes.
Ours are published on the failed lots page.
Read one for real
The fastest way to learn is on a live certificate. Open a certificate, follow the fourteen lines in order, and check the address bar as you go. The specification each line is judged against was published before the lot existed — the only order of events that means anything.
The one mistake to never make
Entering an HPLC purity figure as though it were peptide content is the single most common way these numbers go wrong. A 99 % pure vial is not a 99 % full vial. If your certificate only reports purity, it does not establish how much peptide is present — ask the supplier for an assay or a peptide-content result before relying on a quantity.
Put a real certificate’s number to work in the Unequivocal Calculator, confirm a lot with Verify a COA, or see what every lot is tested for in The Testing Standard.
This guide explains certificate figures only. It does not describe the effect of any article in a human body and offers no dosing or treatment guidance. Every article is supplied under the vial caution: “CAUTION: Contains a new drug for investigational use only in laboratory research animals or for tests in vitro. Not for use in humans.”