01 COA & batch testing

What the COA for riboflavin actually contains

Technical document 01 · Reviewed 2026-11-18 · Vitamin B2

On this page
  1. The certificate as a test report
  2. What the dried basis and the HPLC line measure
  3. Release limit versus batch result
  4. What moves between batches, and what the colour shows
  5. When to order an outside test
  6. What to compare between suppliers

The certificate for a riboflavin lot reports appearance, identification by IR and UV, assay by HPLC on the dried basis, loss on drying, residue on ignition, particle size, heavy metals and microbiology. Each row should carry the method and the release limit. The ranges here are the catalog specification; the batch COA confirms the particular lot, and the assay only means something when its basis and the identity lines are read with it.

In one line

The assay says how much riboflavin is in the lot; the dried basis, the identity lines and the colour say whether that number belongs to riboflavin.

The certificate as a test report

A certificate is useful when it can be audited line by line. Each parameter needs three things: what was measured, which method produced the number, and which limit the lot was released against. The tables below separate identity and purity from contaminants and microbiology, because different people read them and they fail in different ways.

Identity, assay and purity · riboflavin
ParameterMethodRelease limit
AppearanceVisualYellow to orange-yellow crystalline powder
Identification AIR spectroscopySpectrum conforms to the reference
Identification BUV spectroscopyMaxima and absorbance ratios conform
Assay (dried basis)HPLC with UV detection98.0–102.0%
Loss on dryingGravimetric≤1.5%
Residue on ignitionGravimetric≤0.3%
Particle sizeSieve analysis≥95% through 80 mesh

The identification block is not decorative. IR compares the fingerprint region with a reference spectrum; UV compares the absorption maxima and their ratios. Either can pass while the assay is low, or fail while the assay is inside the window, which is why the specification carries all three. The appearance line is part of identity too: a colour that has drifted is a reason to check the storage history before the assay.

Contaminants and microbiology · riboflavin
ParameterMethodRelease limit
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 heavy-metal panel is reported element by element, not as one “heavy metals” number, and the method should be named so its detection limits can be judged. The microbiological lines are release parameters as well; a clean assay does not excuse a failing plate count.

What the dried basis and the HPLC line measure

The assay is reported on the dried basis, so the result is corrected for the water measured by loss on drying. That matters because the specification caps loss on drying at 1.5%, and a lot at the top of that window carries more water than one at the bottom. The as-is figure is the dried-basis figure multiplied by one minus the loss on drying; the certificate should state the basis rather than leave the reader to convert.

The 98.0–102.0% window is a pharmacopoeial-style purity window, not a statement that the powder can contain 102% riboflavin. The upper bound allows for the method, the reference standard and measurement uncertainty. A figure at the top of the window is not a more concentrated vitamin; it is the same material measured near the top of the same range.

IR is the fingerprint test; UV measures the conjugated isoalloxazine chromophore, with maxima in the ultraviolet and visible range and characteristic absorbance ratios. The exact maxima belong in the method, not in a marketing line. HPLC separates riboflavin from related substances and from the phosphate ester, and the assay is calculated against a riboflavin reference standard. A certificate that names “HPLC” without the column, the detection wavelength and the basis has given a method name, not a method.

Release limit versus batch result

A release limit is the boundary for the lot; a batch result is what the laboratory measured. Both belong on the certificate, in separate columns, with the method beside them. A result column filled with the specification range is a copy of the gate, not a result. The same rule applies to appearance: “conforms” is a conclusion, and the descriptor it conforms to should be printed next to it.

Appearance and identification are pass/fail, loss on drying and residue are ceilings, and the assay is a window. A result close to a limit is read with the trend across lots, not on its own.

What moves between batches, and what the colour shows

Riboflavin is made by fermentation or by synthesis, and both routes end in the same molecule. What moves between lots is less the assay, held in a narrow window, than the physical picture: crystal size and shape, shade of yellow, loss on drying, residue and trace metals. Milling and sieving set the particle size; drying and packing set the moisture.

Colour is the most visible variable and the easiest to misread. Yellow to orange-yellow is the normal range, and the shade varies with crystal size and with the light the powder has seen. A dull or brown tint is not a normal shade; repeat the assay and check how the lot was stored and packed. Under 365 nm light, a thin layer should show a strong yellow-green glow; a weak glow points to dilution, degradation or a different material.

When to order an outside test

Order an independent test for a new source, a first shipment into a regulated market, a form claim the certificate does not prove, or a disputed result. The useful panel is not a full repeat: HPLC assay with peak purity, IR and UV identity, loss on drying and residue, ICP-MS for the heavy-metal panel, and the microbiological set, all on one submitted sample. If the phosphate form is possible, add a phosphate assay or a second HPLC method that separates the ester.

Ask the laboratory to report its detection limit for each contaminant. A “not detected” line without a limit is evidence about the method's sensitivity, not about the sample. Chain of custody matters as well: the sample tested should be the sample retained, and the lot code should travel with it.

What to compare between suppliers

Put six lines side by side: form and CAS number; assay method and basis; loss on drying; residue on ignition; particle size; and the colour and fluorescence description. Two quotations that match on assay but diverge on the other five are not describing the same powder.

The second layer is lot traceability. The lot code should match on the certificate, the pouch or drum and the retained sample. The certificate should say who manufactured the lot, which route was used and when it was packed; a correct certificate can still belong to a different drum. Ask for the COA of the lot being shipped, not a representative certificate from a previous campaign.

The catalog specification sits on the riboflavin product page. The two forms are compared on the grade page, the tests that separate them on the identity page, and the release workflow on the quality page. The HPLC method note describes the assay, and the loss-on-drying entry defines the basis arithmetic.

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