A label is a claim. A COA is supposed to be the evidence behind it. Here's how I actually read one — what matters, what's noise, and the red flags that tell you a report can't be trusted.
A Certificate of Analysis — a COA — is a laboratory report tied to one specific batch of one specific material. In the peptide and research-compound world, it's one of the most useful documents you can ask for before you trust what a label says.
A real COA doesn't just announce "tested" or "99% pure." It shows enough that you can answer: what batch was tested, what compound it was, which lab ran it, which methods they used, whether the results match the labeled identity, whether purity and contaminants were screened, and whether any of it can be traced back to a real record. The operative word is batch-specific — the report should belong to the exact lot you're holding, not a tidy example from some earlier run.
COAs can carry a lot of fields, but three carry most of the weight — and a product can pass one while failing another. Something can be extremely pure and still be the wrong molecule; it can match the right mass and still be full of impurities.
Is this actually the molecule on the label? Usually mass spectrometry.
How much of the sample is the intended compound? Usually HPLC / UPLC.
Safety screens — endotoxin and heavy metals, where relevant.
For peptides and many research compounds, identity is checked with mass spectrometry — you'll see it written as MS, LC-MS, ESI-MS, MALDI-TOF, or simply "molecular weight confirmation." Mass spec measures the mass-to-charge behavior of molecules and confirms whether the detected compound matches the expected molecular weight within a reasonable analytical tolerance.
A useful identity section lists the compound name, the expected molecular weight or formula, the observed mass, the method, a result statement ("conforms," "consistent with expected mass"), the test date, and the batch number. A strong result shows the expected and observed mass — not just the word "pass."
Purity is usually measured by High-Performance Liquid Chromatography (HPLC), or the higher-resolution UPLC. The method separates components as they move through a column and produces a chromatogram — a graph of peaks. The main compound should be the dominant peak; smaller peaks can be impurities, fragments, synthesis byproducts, degradation, solvents, or related compounds.
You'll usually see something like Purity: 98.7% or Main peak area: 98.7% — meaning the main peak accounted for that share of the integrated chromatographic area under the test conditions. When you read the chromatogram, ask: is there one clearly dominant peak? Are the secondary peaks small? Does the stated purity actually match the peak table? And is the method shown at all? A bare purity number with no chromatogram, method, or lab detail is weaker than one that shows its work.
A sample can be the right molecule and look very pure while still carrying contaminants that matter for safety or research integrity. Two categories matter most, especially for injectable, sterile-labeled, implantable, cell-culture, or any human-adjacent use.
Endotoxins are bacterial components — most often from Gram-negative bacteria — that can trigger strong inflammatory responses. Results are reported in Endotoxin Units: EU/mL, EU/mg, EU/vial, or below a stated limit such as <0.05 EU/mL. A strong section names the method (commonly an LAL-based assay), a numerical result, an acceptance limit, the batch number, and the date. "Endotoxin tested," with no number, method, or limit, tells you almost nothing.
These screen for toxic elements — lead, arsenic, cadmium, mercury, and others — that can enter through raw ingredients, catalysts, water, equipment, containers, or the environment. Look for which elements were tested, a result for each, the units (ppm, ppb, µg/g), the method (often ICP-MS or ICP-OES), and the limits. "Heavy metals: pass," with no elements, values, or limits, is the weak version.
A COA is only useful if it matches the batch in your hand. Compare the batch/lot number on the COA against the one on the vial or package, plus the manufacturing date, any retest/expiration date, the product name and strength, the vendor, and the lab. This is one of the most common COA problems: a clean-looking report that actually belongs to a different batch, concentration, date, or even compound.
A COA can come from the manufacturer, the vendor, or an independent third-party lab. Third-party testing is generally stronger because it reduces conflict of interest — but even third-party reports deserve a careful read. Look for the lab name, a contact or verification system, a report number, the batch number, the dates received and tested, the method, an analyst or authorized signature, and a QR/verification link if one exists. Missing lab name, batch, method, or date? Treat it cautiously.
Nothing links the report to the product.
A report for a different lot verifies nothing.
Purity without mass-spec identity is incomplete.
Mass alone doesn't show the impurity profile.
For sterile/injectable material, that's a critical gap.
Vague claims aren't results.
Nothing to trace it back to.
A partial image is weaker than a full report.
You can't tell when — or if — it was run.
Each compound and batch needs its own report.
A website number isn't a lab report.
A number with no method is hard to judge.
| Field | Why it matters |
|---|---|
| Product name | Confirms what was tested |
| Batch / lot number | Links the report to the product |
| Test date | Shows when analysis occurred |
| Lab name | Identifies who performed testing |
| Report number | Helps verify the record |
| Identity test + mass-spec result | Confirms the molecule |
| Purity test + chromatogram | Shows main-compound percentage and peak pattern |
| Endotoxin result | Critical for injectable or sterile contexts |
| Heavy-metal results | Screens for elemental contamination |
| Method details | Shows how the results were generated |
| Acceptance criteria | Defines what "pass" actually means |
| Signature / authorization | Supports report authenticity |
Run a COA through this the next time one lands in front of you. Tap each item you can confirm — the bar fills as the report earns your trust.
These get conflated constantly, and they answer different questions. A COA reports analytical results for a specific batch — "what did this batch test as?" An SDS (Safety Data Sheet) covers hazard, handling, and storage — "how do I handle this safely?" GMP (Good Manufacturing Practice) is a manufacturing quality system, not a test report. A product can have a COA without being made under GMP, and a COA alone doesn't prove GMP, sterility, regulatory approval, or clinical suitability. Keep them separate: COA = batch test results, GMP = manufacturing system, sterility = microbial control, endotoxin = bacterial screen, regulatory approval = legal authorization.
Certificate of Analysis — a laboratory report documenting analytical testing results for a specific material, product, batch, or lot.
Batch-specific identity confirmation. If the report doesn't confirm the compound is what the label says, the rest of the COA is limited.
It estimates how much of the detected sample appears as the main compound peak under the test conditions, usually reported as a percentage.
Not necessarily. A 99% figure is only meaningful if identity is confirmed, the COA matches the batch, the method is appropriate, and relevant safety screens are included.
Usually not on its own. It's best paired with mass spectrometry or another identity-confirming method.
Yes. It should match the exact batch or lot on the product. If it doesn't match, it may not apply to what you have.
No. It provides analytical results, but it doesn't prove safety, legality, sterility, clinical appropriateness, or regulatory approval.
Educational only. Not medical advice. Not a recommendation for use.