Reading a Certificate of Analysis without taking the vendor’s word for it
A practical walkthrough of every field on a real research-grade COA — what each number actually means, and which ones you can ignore.
Most certificates of analysis you see attached to peptide vials are marketing documents in lab clothing. A real COA is a record of a measurement, traceable to the instrument and the date it was made. Here’s how to tell the difference.
The header is not the data
Every COA opens with a batch identifier, manufacturing date, expiration date, and the vendor’s logo. None of this tells you anything about the molecule in the vial. The header just establishes which measurement you’re about to read. Skip past it.
Identity confirmation by mass spectrometry
The first real number on a COA should be the measured molecular weight. Mass spectrometry — usually electrospray ionization (ESI-MS) — produces a spectrum where the strongest peak is the protonated molecular ion. A confident identity check looks like this:
- Theoretical MW: 4117.5 Da
- Measured MW: 4117.6 Da
- Tolerance: ±0.5 Da
Both numbers should be present. If the COA only shows “PASS” without a measured value, treat it as decorative.
Purity by HPLC, area-percent at a fixed wavelength
High-performance liquid chromatography separates the compound from everything else in the vial — synthesis byproducts, residual solvents, degradation fragments. The detector measures absorbance at a fixed wavelength (typically 220 nm for peptide bonds), and the result is reported as a percentage of the total integrated peak area.
A serious COA appends the chromatogram itself, not just the number. Look for:
- A single dominant peak with the target retention time
- Small ancillary peaks separated and visible — not buried under a wide solvent front
- Baseline that returns to flat between peaks
If the chromatogram is missing, the purity number is unverifiable. “≥99%” without an attached chromatogram is a claim, not a measurement.
Endotoxin: EU/mg, not pass/fail
Bacterial endotoxin (LPS) is one of the most common contaminants in peptide synthesis, and it’s a real problem in cell culture work even when the assay isn’t biological. The Limulus amebocyte lysate (LAL) assay measures it in endotoxin units per milligram. A clean batch lands well under 1.0 EU/mg; a research-grade preparation should be reporting somewhere in the 0.05 – 0.5 EU/mg range.
If the COA says “endotoxin: passes,” the lab either didn’t run a quantitative LAL or doesn’t want you to see the number.
Bacterial and fungal counts in CFU/g
Plate counts after a 14-day incubation at 30–35°C (bacteria) and 20–25°C (mold/yeast). Reported as colony-forming units per gram. Zero is the expected number for sterile-filtered, aseptically lyophilized material. Anything above zero on either count means the vial isn’t sterile, which has implications well beyond cell culture.
Residual moisture
A properly lyophilized peptide cake retains less than 2% residual moisture by Karl Fischer titration. Below 1% is excellent. Above 3% means the lyophilization cycle was cut short — and the compound’s shelf life shrinks dramatically.
What’s missing tells you as much as what’s there
If a COA omits any of the above, it’s not because the lab forgot. It’s because the result wouldn’t have flattered the batch. The standard for a real research-grade COA is short and complete:
- Identity by MS — measured value with tolerance
- Purity by HPLC — area-percent with the chromatogram appended
- Endotoxin by LAL — quantitative value in EU/mg
- Bacterial & fungal counts in CFU/g
- Residual moisture by Karl Fischer
- Tested-by attribution with a date and a third-party lab name
Ask for the chromatograms and the mass spec spectra. Vendors who run their own QC are happy to send raw data. Vendors who don’t, won’t.