How to Spot a Fabricated Certificate of Analysis
Forged peptide COAs are common and most of them share the same handful of tells. What fabricated documents get wrong about chromatograms, mass spectra, batch numbers and the language of a real laboratory report.
A Certificate of Analysis is a PDF. PDFs are easy to edit, letterheads are easy to copy, and the number of buyers in a position to verify one is small. The economics of fabrication are favourable and the practice is correspondingly common.
The good news is that most fabricated certificates are not sophisticated. They are made by people who understand marketing rather than analytical chemistry, and they share a recognisable set of errors.
Tell 1: no batch number, or a batch number that never changes
The purpose of a Certificate of Analysis is to describe one specific batch. The batch number is what connects the document to the vial.
Warning signs:
- No batch or lot number anywhere on the document.
- A batch number on the certificate that does not appear on the vial label.
- The same certificate served for every order over months — meaning either every vial came from one enormous batch, or the document is a product brochure.
- A batch number in an obviously decorative format:
BATCH-001,LOT-2024.
Real batch numbers are usually unlovely, because they encode manufacturing information rather than being chosen to look convincing.
This is the strongest single tell, because it defeats the document's only purpose. Everything downstream is moot.
Tell 2: the chromatogram is wrong in a specific way
Fabricated chromatograms are usually drawn by someone who has seen one but never read one.
A perfectly flat baseline. Real detectors produce noise. Real gradient runs produce drift. A baseline that is a mathematically straight line at exactly zero is a drawing.
One peak and nothing else. Real 99% pure material still shows small impurity peaks. The 1% is somewhere. A trace with a single peak and an otherwise empty field describes material that does not exist.
No axis labels or units. Real chromatography software prints retention time, absorbance units and the detection wavelength automatically. Their absence usually means the image did not come from an instrument.
A perfectly symmetric peak. Some tailing is normal in peptide reversed-phase work. A flawless Gaussian is a drawing tool's output.
A retention time that doesn't fit the run. A main peak at 2 minutes into a 30-minute gradient, with a run that continues for 28 minutes showing nothing at all, is not how a method would be designed.
The same chromatogram on different products. Worth checking if you order more than one compound — overlay the images. Identical traces for different peptides means one image is being reused.
See reading an HPLC chromatogram for what a real one looks like.
Tell 3: the mass spectrum shows a single clean peak
This one is diagnostic, because it requires actual familiarity with the technique to get right.
Peptides are ionised by electrospray, which adds multiple protons. A large peptide is observed as a series of multiply charged ions at different mass-to-charge ratios, not as one peak at its molecular weight. A 4,963 Da peptide appears at m/z ≈ 1,655 (3+), ≈ 1,242 (4+), ≈ 993 (5+) and so on, and software deconvolutes the series to a molecular mass.
A "mass spectrum" showing a single clean spike at exactly the molecular weight is what a person would draw if they thought a mass spectrometer prints the molecular weight. It is not what the instrument produces.
Related tell: only one mass figure given. A real result states theoretical mass and observed mass, because the comparison is the result. A single figure is a claim with the evidence removed.
Tell 4: numbers that are too good, or too round
Purity of exactly 99.9%, or 100%. Purification is a process with limits, and integration has uncertainty. Real certificates carry values like 98.7% or 99.24% because that is what the software calculated.
Every batch identical. If successive batches all report exactly 99.5%, the number is not being measured.
Observed mass identical to theoretical mass to more decimal places than the instrument can deliver. If the certificate names a low-resolution instrument and then reports a match to four decimal places, the two halves of the document disagree.
Tell 5: tests that do not apply
A strong indicator of a template being filled in without understanding.
- A purity percentage on bacteriostatic water. It is a solvent. There is no peptide to assay and no purity figure exists. See bacteriostatic water.
- A single purity figure for a blend. Two components produce two main peaks; one number cannot describe both.
- "Sterility: pass" on a lyophilised research peptide that was never sterility tested — a test with a defined method and a defined duration that does not happen incidentally.
- Endotoxin results with no method or units. Real endotoxin results are reported in EU/mg against a stated method.
Tell 6: the language is wrong
Analytical reports are written in a particular register — flat, specific, hedged where hedging is warranted. Marketing language inside a laboratory document is out of place, and its presence usually means the document was written by the same person who wrote the product page.
Phrases that do not belong on a certificate: "premium quality", "highest purity available", "pharmaceutical grade" (a term with no fixed meaning in this context), "research grade guaranteed", "lab tested" with no laboratory named and no method given.
Real certificates also carry mundane detail that fabrications omit because it seems inessential: instrument model, column dimensions and chemistry, mobile phase composition, gradient conditions, injection volume, analyst identity, date of analysis distinct from date of manufacture.
Tell 7: the claims around the document
Sometimes the certificate is fine and the claims attached to it are not.
"Third-party tested" with no laboratory, method or date. Frequently this describes in-house testing relabelled. See what third-party testing actually means.
"ISO 17025 certified" with no accreditation number. Accreditation bodies maintain public registers, and the number is the whole point of holding it. The claim appears constantly in this market and almost never with the number that would make it checkable.
"Every batch independently verified" where the certificate on offer is clearly in-house work. This particular overstatement does two things at once: it overstates coverage, and it reassigns the vendor's own chromatography to somebody else's laboratory.
What you can actually do
Ask which laboratory performed the analysis. A supplier confident in their documentation will answer. Evasion is itself the answer.
Ask for the raw data. Instrument-generated PDFs contain metadata, method parameters and axis labels that a drawn image does not.
Compare across orders. Different batches should produce visibly different chromatograms. Identical traces mean one image is doing all the work.
Ask for the accreditation number if accreditation is claimed, and check the register.
Commission your own analysis where the work justifies it. It is the only approach that does not depend on trusting the supplier at all, and for anything that will be published it is worth the cost.
The underlying point
A certificate is a claim about material, not proof of it. Its value depends entirely on whether it can be traced to a real analysis of a real batch — which is why the batch number matters more than the purity figure, and why a supplier's willingness to answer questions about their documentation is more informative than the documentation itself.
Related: how to read a peptide COA and choosing a peptide supplier in Australia.