A sweetener COA arrives as a table, and most buyers read the assay first. That is the line least likely to explain why two lots taste different. This note sets out the six lines that decide the grade, in the order a quality reviewer should read them, and the questions to ask when a line is missing.
Start with identity, not purity
Purity is a quantity. Identity is a statement about what the material is. A sucralose certificate that says 99.2% is telling you how much of the named peak was measured, not that the peak is sucralose, that the crystal is the right form, or that the lot came from the supplier's own plant. The identity line is the one that names the reference standard, the secondary confirmation and the chromatogram.
For a steviol glycoside, identity includes the glycoside profile: Reb A, stevioside and Reb C. For a polyol, it includes the other-polyol profile. For a fibre, it includes the degree of polymerisation. A certificate that reports only the headline percentage has hidden the identity question inside an assay number.
Read the assay basis before the assay
The same lot can read 98.0% as-is, 99.2% on a dried basis and 99.5% on an anhydrous basis. None of those numbers is wrong; they answer different questions. A supplier comparing its 99% dried-basis result with a competitor's 98% as-is result is not making a fair comparison, even if both certificates are accurate.
The basis is especially important for hydrates. Sodium saccharin dihydrate carries water in the crystal; trehalose dihydrate carries about 10% water. A 98% assay on a dried basis and a 98% assay as-is can describe the same material, or two different forms. Read the basis with the moisture method and the specification, not as a standalone figure.
The related-substance or isomer profile is the taste signal
Two lots can both meet a 98% assay and taste different because the remaining 2% is different. In stevia, the stevioside and Reb C fractions contribute to the linger and the bitterness. In a polyol, a sorbitol fraction in an erythritol lot changes the hygroscopicity and the cooling. In a rare sugar, a fructose fraction changes the browning and the sweetness.
The related-substance line is also the adulteration signal. A blend built to hit the assay will show a profile that does not match the reference. Ask for the profile, not only the sum. If the certificate reports 'related substances ≤2%' without the individual peaks, the number is a limit, not a description.
Moisture and particle size decide the process
Moisture changes the usable solids and the storage behaviour. A powder at 0.2% moisture and a powder at 0.8% can pass the same assay and flow differently, cake differently and weigh differently per scoop. For a hydrate, the water is part of the specification; for an anhydrous grade, it is a limit.
Particle size and bulk density decide how the powder dissolves, dusts, fills and blends. A 60 mesh granule and an 80 mesh powder can share an assay and behave nothing alike in a sachet line or a tablet press. The COA should report both, or the TDS should state the grade and the COA should confirm the lot.
Contaminants, microbiology and the lot number
The heavy-metal and microbiology blocks are release decisions, not decoration. Lead, arsenic, cadmium and mercury should be reported individually, with the limit of quantification. The microbiology block should name the sample size, not just say 'absent'. A result below the detection limit is only useful if the limit is stated.
Finally, the lot number. The COA is a record for one lot, and the lot number on the certificate should match the drum and the retained sample. A certificate without a lot number, or with a lot number that does not match the packing list, is not batch-specific evidence, however good the assay looks.
The six-line summary
If a certificate is missing one of those lines, the answer is not necessarily a rejection. It is a question. A supplier who can supply the missing line and explain the method is easier to qualify than one who cannot. The point of reading the COA closely is to make the conversation specific.
- Identity: reference standard, chromatogram and the secondary test.
- Assay basis: as-is, dried or anhydrous, with the moisture method.
- Profile: related substances, other polyols or glycoside distribution.
- Physical: moisture, particle size and bulk density.
- Contaminants: individual metals, limits of quantification and microbiology.
- Traceability: lot number, manufacturing date and pack.
What to take away
- Read identity before assay, and the assay basis before the number.
- Ask for the profile behind the percentage; it explains taste and catches substitution.
- Moisture and particle size are process parameters, not footnotes.
- The lot number is part of the evidence, not an administrative detail.
Go deeper
COA and batch testing
How the primary assay is built, what it proves and what the chromatogram adds.
Read the referenceCite this note
Rainwood Sweeteners. “How to read a sweetener COA: the six lines that decide the grade.” Published 20 November 2026. https://sweeteners.rainwoodbio.com/blog/how-to-read-a-sweetener-coa/