For an institutional buyer, the most critical attachment to an order is often a single-page document: the certificate of analysis. This document is the record that establishes, by measurement, what the material in your hands actually is. Read correctly, it answers many questions up front; read superficially, it creates false expectations. This article goes through the document line by line and shows where each piece of information is to be found.
What does the certificate document?
A certificate of analysis reports the results of measurements made on a sample taken from a specific batch. The key words are a specific batch: the document belongs not to a product family but to a single production unit. The certificate of one batch therefore does not count as evidence for another.
Who issued the document also determines what it means. A declaration based on the producer's own measurements and a report issued by an independent laboratory do not carry the same weight; in the latter, the party doing the measuring is separate from the party doing the selling. The analyses for our facility's reference batch were carried out outside the facility, in an independent laboratory.
The certificate is often confused with the safety data sheet, yet the two answer different questions.
| Certificate of analysis | Safety data sheet | |
|---|---|---|
| What does it relate to? | A single batch | The product itself |
| What does it tell you? | Measured values | Hazard, handling and emergency information |
| How is it referenced? | By batch number | By version number and date |
| How often is it renewed? | For every batch | Revised when information changes |
The section layout of the form was covered in the article on the SDS. The two documents cannot replace each other; both belong in the acceptance file.
Which fields does it contain?
Although the format of certificates varies from laboratory to laboratory, the set of information to look for is largely the same.
- Identification: The name of the material, its source and species, batch number, sample form and quantity.
- Dates: The date the sample was received, the date the analysis was performed and the date the report was issued. All three may differ, and the time between them is informative.
- Methods: The method and instrument used for each measurement.
- Results: Parameter name, value found and its unit.
- Conditions: Details that affect the result, such as measurement temperature, sample concentration and dilution.
- Approval: The organization issuing the report, the responsible person and signature details.
How the sample was taken is sometimes also stated in the document. A small sample taken from a large batch does not automatically represent the batch; stating the sampling method shows what whole the result can be generalized to. If this line is missing, the report speaks only for the sample delivered.
Including species information in the identification section is particularly important. For a material of biological origin, the name of the source species is the first condition for interpreting the results. The consistency of the batch number across documents is covered in the article on traceability.
Reading the results column
A number is incomplete without two pieces of information beside it: which method produced it and in which unit it is given. The same parameter can give different results when measured by different methods; this is not an inconsistency but a sign that the methods see different things.
Protein content, for example, is calculated from a standard curve. When the equation of the curve and its goodness-of-fit measure are written in the report, the reader can judge how much confidence the result deserves; how this calculation is set up is described in the article on protein determination.
A purity statement also needs care. A purity value in a report means that no foreign substance was detected within the scope of what the applied method can see. Because not every method reveals every impurity with the same sensitivity, this statement is not an absolute claim of purity but a result limited by the method.
Some rows may contain a statement instead of a number. Entries such as not detected, out of range or not applicable do not mean that no measurement was made; they carry information about the limits of the method. A substance not being detected is not the same as its absence; it only shows that it remained below the sensitivity of the method used.
Another detail is whether the result comes from a single measurement or from the average of replicate measurements. When the number of replicates is stated, it also becomes clear how well established the value is. A single measurement is not a shortcoming in itself; it is simply necessary to know how the result was obtained.
Common reading errors
| Error | Why is it a problem? |
|---|---|
| Treating the result for one batch as a fixed property of the product | In material of biological origin, values vary from batch to batch |
| Comparing numbers without reading the method | Different methods measure different quantities |
| Skipping the measurement conditions | Temperature and dilution directly affect the result |
| Ignoring the dates | The time the sample was kept waiting may have affected the result |
| Treating the document as sufficient on its own | Species, permit and hazard information are found in other documents |
Most of these errors come from seeing the document as a stamp of approval. A certificate is not an approval but a record of measurements. It is the buyer who compares the results with expected values: the buyer determines which parameter should fall within which range for their own use and assesses the report against those criteria.
The last row sums up the acceptance process. The certificate of analysis tells you what is in the material; other documents tell you where it came from, under which permits it was produced and how it should be handled. Why the analysis is carried out outside the facility is a separate topic, covered in the article on independent laboratories.
The methods and values found in the report for our reference batch are published on the analysis page. The results are specific to that batch; measurements are repeated for each batch, and requests for documents relating to a particular batch are handled separately.
