01 COA & batch testing

What the COA for retinol oil actually contains

Technical document 01 · Reviewed 2026-11-20 · Retinol

On this page
  1. The certificate as a test report
  2. The as-is basis, the standard and the IU arithmetic
  3. The limit and the measured value
  4. What moves between batches
  5. When to order an outside test
  6. What to compare between suppliers

The certificate for a retinol lot reports appearance, identification by HPLC with UV detection, assay as retinol, peroxide value, heavy metals and microbiology. Each line needs its method and its release limit, and the measured result belongs in a column of its own. The limits below are the catalog specification; the batch COA is the document that releases a particular bottle.

In one line

An oil needs two gates on the certificate: the assay says how much retinol is present, and the peroxide value says how much oxidation the oil has already taken up.

The certificate as a test report

A certificate is only as useful as its least specific line. Every line should answer three questions: the quantity measured, the method that produced it, and the limit the lot was released against. A row that reads only “conforms” asks the reader to accept a conclusion with nothing behind it.

The two rows that decide the material are the identity row and the assay row. Retention time plus UV spectrum says the peak is retinol rather than another retinoid; the assay says how much is present. The third decisive row is the peroxide value: an oil can pass its assay while already carrying oxidation.

Identity, assay, oxidation and contaminants · retinol oil
ParameterMethodRelease limit
AppearanceVisualYellow to amber, clear to slightly viscous oil
IdentificationHPLC retention and UV spectrumConforms to the retinol reference
Assay (as-is)HPLC with UV detection≥95.0%
Peroxide valueTitrimetric≤10.0 meq O₂/kg
Heavy metalsICP-MSPb ≤3 · As ≤2 · Cd ≤1 · Hg ≤0.1 mg/kg
MicrobiologyPlate count and absence testsAPC ≤1,000 cfu/g; yeast and mould ≤100 cfu/g; E. coli and Salmonella absent

The as-is basis, the standard and the IU arithmetic

The assay is reported on the material as supplied. That sounds trivial until a buyer compares two certificates written on different bases. The catalog sets no loss on drying for this oil, so there is no dry-matter divisor to apply: 95.0% means 95.0% of the oil in the bottle. A “dry basis” result for an oil should be questioned: ask what was dried and how.

The certificate should name the standard and the wavelength. The assay is calculated against a retinol reference of declared purity, read in the ultraviolet near 325 nm. A result that does not name the reference leaves the reader unable to tell whether the standard was retinol or an ester.

The same bottle can be described as a percentage or as IU/g, and the two figures are arithmetically linked. One microgram of all-trans-retinol is defined as 3.33 IU, so 95.0% retinol is 950,000 µg/g and about 3.16 million IU/g. That IU figure is calculated from the chromatogram, not measured a second time. A certificate quoting only IU/g has hidden the percentage and the conversion, and the buyer cannot check either.

The limit and the measured value

The release limit is the boundary for the lot; the batch result is what the laboratory measured. The two belong side by side, with the method named in the same row. A result column that repeats the release limit has measured nothing.

What the result column should contain
ParameterRelease limitResult column
Assay≥95.0% as-isA measured percentage, with the reference and the wavelength
Peroxide value≤10.0 meq O₂/kgA measured figure, with the test date
Heavy metalsPer elementA figure in mg/kg for each element, not a total
MicrobiologyCounts and absenceThe count and the absence statement for the lot
AppearanceYellow to amber oilThe observed description for the lot

“Passes specification” is a conclusion, and the reader should be able to reach it from the numbers. For heavy metals the evidence is a figure per element; a single summed figure has hidden the element that matters. For microbiology the evidence is a colony count together with a statement of absence. For the peroxide value it is a number and its date, because the value climbs in storage.

What moves between batches

The peroxide value is the most variable line on this certificate. A lot released quickly can sit far below the 10.0 limit; a lot held warm, or packed with a large headspace, can arrive close to it. A value of 9.0 passes the gate and still leaves almost no margin for a year in a warm warehouse, so read the number and the date together rather than the pass alone.

Isomer content moves with heat, light and processing time. All-trans-retinol is the form the label assumes; light and heat push part of it towards cis isomers, which show a shifted and weaker ultraviolet band. The catalog does not set an isomer limit, so the chromatogram is where the shift becomes visible. Colour moves from pale yellow to amber across the same window, and a darker oil is a hint rather than a failure.

The assay can drift with the reference standard and with the dilution used to reach the detector's linear range. A trend across lots — assay sliding low, peroxide value climbing, a new shoulder in the spectrum — says more than one result at the edge of its limit.

When to order an outside test

Order an independent test for a new source, a first shipment, a disputed result, or a form claim the certificate does not prove. The panel does not need to repeat the whole certificate. A useful one combines HPLC with diode-array detection for identity and assay, peroxide value by titration, the isomer profile from the same chromatogram, and a direct method that resolves retinol from the retinyl esters.

Add ICP-MS for the metal panel, an antioxidant identification and assay if the oil is stabilised, and a residual-solvent screen if a solvent-borne dilution is suspected. Ask for the detection limit behind every screen the laboratory reports. Without a limit, “not detected” describes the method's sensitivity rather than the oil.

Chain of custody matters. The portion tested should be the portion retained, and the report should carry the lot code. When a result is contested, give the referee the retained sample and the specification, and ask for tests that separate the forms rather than a repeat assay.

What to compare between suppliers

Six lines make two quotations comparable: form and CAS number; assay as-is and the IU/g derived from it; the isomer profile; the peroxide value with its date; the declared antioxidant; and the pack with its headspace. Two offers that agree on assay but differ on antioxidant or pack are not offering the same shelf life.

The catalog specification sits on the retinol product page. The tests that separate the free alcohol from an ester are on the identity page; the market conditions are on the compliance page; the ester products are described on the vitamin A page; the chromatographic method is outlined in the analytical methods library; and the release workflow is on the quality page.

Send the specification you intend to use — grade, quantity and destination market — and we will reply with the batch COA, the technical data sheet and a quotation.

Request a quotation