01 COA & batch testing

What the COA for whey protein isolate actually contains

Technical document 01 · Reviewed 2026-11-20 · Whey Protein Isolate

On this page
  1. The certificate as a set of gates
  2. Dry basis, the nitrogen factor and the as-is figure
  3. Release limit and batch result
  4. What moves from lot to lot
  5. When an outside laboratory is worth it
  6. What to compare between suppliers

A whey protein isolate certificate reports protein by Kjeldahl with the 6.38 nitrogen factor on a dry basis, and the moisture figure that converts it to the powder as weighed; fat, lactose, ash, the metal panel, microbiology and particle size follow. Each line has to carry the method and the release limit the lot was passed against, with the measured result in its own column. The catalog limits are the gate; the COA shows a particular pack went through it.

In one line

The protein figure only means something with its nitrogen factor, its basis and the moisture beside it; the fat, lactose and ash then describe the powder behind the number.

The certificate as a set of gates

For each row the reader needs three things: what was measured, the method that produced the number, and the limit the lot was released against. A row that says only “conforms” asks the reader to accept a conclusion without evidence. On a whey isolate the protein row is the commercial one, the identity row is the one that says the powder is whey, and the fat and lactose rows are the ones the label will quote.

Two rows are often merged and should not be. Protein by Kjeldahl is a nitrogen measurement; identification is a comparison against a reference. A certificate can report 91.5% protein and still fail to show that the protein is whey protein rather than a blend with another nitrogen source.

Protein, composition and release tests · whey protein isolate
ParameterMethodRelease limit
IdentificationHPLC protein profile against a referenceConforms
Protein (dry basis)Kjeldahl, N × 6.38≥90.0%
FatSoxhlet or Weibull–Stoldt extraction≤1.0%
LactoseHPLC or enzymatic≤1.0%
AshGravimetric, muffle furnace≤3.0%
MoistureLoss on drying, vacuum oven≤5.0%
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
Particle sizeSieve analysis≥95% through 80 mesh

Dry basis, the nitrogen factor and the as-is figure

The protein result is reported on a dry basis: the laboratory measures the nitrogen in the powder as received, multiplies by 6.38, and divides by the dry matter. At the 5.0% moisture limit a 90.0% dry-basis figure is 85.5% as-is, because the water is part of the mass the label weighs. A lot at 92.0% dry basis with 3.5% moisture is 88.8% as-is, and that is the figure a serving calculation has to use.

The factor matters as much as the basis. Milk protein is conventionally calculated at 6.38, which reflects its nitrogen content; the general food factor is 6.25. The same nitrogen result read with the two factors differs by 2.1%, so the certificate should print the factor in the protein row rather than leave it to a note.

Kjeldahl measures total nitrogen, not protein, so the figure includes the non-protein nitrogen in the powder. That is why the maxima on a whey isolate certificate do not sum to 100: 85.5% protein as-is at the moisture limit, plus 1.0% fat, 1.0% lactose, 3.0% ash and 5.0% moisture, reaches 95.5%. The remainder is the gap between a nitrogen calculation and a weighed mass. A “true protein” result comes from a precipitation or an amino-acid sum and is not interchangeable with the Kjeldahl figure.

Release limit and batch result

What the result column should contain · whey protein isolate
ParameterRelease limitResult column
Protein≥90.0% dry basisA measured percentage, with the factor and the moisture
Fat and lactose≤1.0% eachTwo measured percentages, not one combined figure
Moisture≤5.0%The measured loss on drying for the lot
Heavy metalsPer elementA figure in mg/kg for each of the four elements
MicrobiologyCounts and absenceThe count and the absence statement for the lot

“Passes specification” is a conclusion. The reader should be able to reach it from the numbers: a measured protein with its basis, a measured fat and lactose, the moisture that converts them, a figure per metal, a count plus an absence statement, and the percentage through the sieve. A result column that repeats the limit is a copy of the gate, not evidence about the lot.

What moves from lot to lot

Whey isolate is made by separating protein from whey with ion exchange or microfiltration, then washing, concentrating, pasteurising and spray-drying. Each step leaves a signature. The diafiltration volume moves the lactose and the ash; the separation and the fat removal move the fat; the dryer outlet moves the moisture and the colour; the mill and the sieve move the particle size.

Colour is the least quantitative line and the most revealing. Off-white to cream is a wide window, and a shift inside it is not a failure; a powder that darkens noticeably points to a heat load or a Maillard reaction with the lactose, and the protein figure can still pass. Solubility is the second functional line: a lot that disperses more slowly or leaves a haze may have been heated longer than the one before it.

An instantized lot is not comparable on bulk density or flow with a plain 80 mesh lot. Both can meet the same protein, fat and lactose limits and still fill a sachet differently, so the fill weight should be checked against the form that was ordered.

When an outside laboratory is worth it

Order an independent test for a new source, a first shipment into a regulated market, a disputed result, or a form the supplier's certificate does not prove. The useful panel is not a copy of the certificate: it is the protein profile by HPLC, the amino-acid profile, lactose by HPLC, melamine by LC-MS/MS, the metal panel by ICP-MS and the micro count.

Two screens are worth naming. An ELISA for casein or for β-lactoglobulin shows whether the whey source is what it claims, and an amino-acid profile exposes a nitrogen spike that a Kjeldahl figure alone would report as protein. Ask for the detection limit of each screen: “not detected” without a limit is a statement about the method, not about the sample.

The sample tested should be the retained sample, and the report should identify the lot. When a result is contested, give the referee the retained sample, the certificate and the specification, and ask for the tests that separate the possibilities rather than a repeat of the same assay.

What to compare between suppliers

Put seven lines side by side: protein with its factor and basis; fat; lactose; ash; moisture; particle size and form; and the pack with its lot code. Two quotations that agree on protein but differ on basis, factor or form are not describing the same powder.

Then ask who made the lot and by which route. The catalog limits for this product sit on the whey protein isolate product page; the grade and form differences are on the grade comparison; the concentrate it is often quoted against is on the whey protein concentrate page; the separation is described in the HPLC method note; 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