How to spot an altered or reused certificate of analysis

Most COAs are exactly what they claim to be. A few are not, and the problems tend to fall into three groups: a real report attached to the wrong lot, a real report edited after the lab issued it, and a document made to look like a lab report. This page covers what each one tends to look like on the page.

None of the signs below proves anything by itself. Labs reissue reports, change templates and fix typos. The point is to know when to ask for the original.

Reused: one report, many lots

This is the most common problem and the easiest to miss, because the report itself may be entirely genuine.

Signs:

• The lot code on the report does not match the vial, or appears only in a filename or a caption added outside the report.
• The same report number shows up for different lots or different strengths.
• Identical chromatograms across lots. Two separate runs never produce the same baseline noise. If two "different" reports have the same tiny wiggles in the same places, they came from one run.
• Purity figures identical to two decimals across lots that were made months apart.
• A report date that falls before the lot could have existed, for example before the date on the packing slip or the stated manufacture date.

Altered: a real report, edited

Edits usually leave traces where new text meets old.

• Font or size changes inside one field, such as a lot code in a slightly different typeface from the label beside it.
• Misaligned characters in a figure, or a number whose baseline sits higher or lower than its neighbors.
• A peak table that does not add up. Area percents should sum to about 100. If the main peak was edited and the others were not, the column will not total.
• A purity figure that does not match its own peak table.
• A chromatogram whose main peak area does not fit the area-percent figure. A peak the table calls 99.5% should dwarf everything else on the chart.
• Different PDF properties from the lab's other reports, such as a creation tool or date that does not fit. This is weak evidence on its own, since files get re-saved for all sorts of reasons.

Made up: a document that only looks like a report

• No lab name, or a name with no address or contact route
• No report, task or sample number
• No method details at all: no column, no wavelength, no instrument
• Only a pass/fail summary with no measured values
• A chart with no axis labels or no time scale, or a chromatogram cropped so tight that baseline and run length are gone

What to do when something looks off

1. Go to the issuing lab directly. Use the lab's own website or contact details, typed in rather than followed from a link that came with the PDF. Many labs have a lookup page where a report number brings up their stored copy.
2. Compare line by line: lot code, dates, every figure, the chromatogram image.
3. Ask the supplier for the original from the lab, or for the lab's report number if it was not printed.
4. Hold the material out of any experiment until the question is settled, and note the issue in the lab record against that lot.

A mismatch with the lab's own record is the strongest signal there is. Without it, the signs above are reasons to ask questions, not conclusions.

Why this matters for the research record

An experiment is only as traceable as its materials. If the COA in the lab file does not describe the lot that went into the work, the documentation for that work has a hole in it, whatever the material actually was.