A certificate of analysis is a record of measurements made on one batch of material at one point in time. It is not a warranty, not a description of the vial in your hand, and not a single number that can be reduced to "how good is this." Read properly, a COA tells you which questions were asked, by which method, and what the instrument returned. Read carelessly, it tells you a percentage that you will then use in an arithmetic it was never designed to support.

The organizing idea of this page is that a COA answers at least three separate questions — identity, purity, and content — and that the answers are not interchangeable. The analytical strategy behind a well-characterized peptide reference standard treats identity, purity, and strength as distinct tests requiring distinct methods, with value assignment carried out as its own two-step process — a mass balance purity assignment on bulk material, which then serves as the standard for assigning a value to the vialed material — rather than read off the purity chromatogram [1]. A commercial COA is a much thinner document than a reference-standard dossier, but the same divisions apply, and most misreadings collapse two of them into one.

Identity: does this material have the structure the label names?

The identity section asks whether the substance in the batch is the sequence on the label. In practice this is usually answered by mass spectrometry, sometimes supported by NMR or by a chromatographic retention-time match against a standard. Reference-standard characterization programs use NMR, mass spectrometry, and chromatographic methods for identity testing, alongside separate HPLC and GC methods for content and impurities [1].

What an observed molecular mass establishes is narrower than it looks. A mass measurement is consistent with a composition; it does not by itself order the residues. Nor does it distinguish isomers of identical mass. Deamidation of an asparaginyl residue and isomerization of an aspartyl residue both proceed through a succinimide intermediate and yield a mixture of aspartyl and isoaspartyl products, and isoaspartic acid is an isomer of aspartic acid — same mass, backbone lengthened by one methylene unit. Distinguishing the two in peptides has been accomplished with electron capture dissociation on a Fourier transform mass spectrometer, which produces fragment ions specific to the isoaspartyl form; conventional collisionally activated dissociation methods had relied on comparison against aspartyl standards [2]. If a COA reports only an intact mass, isomeric variants of this kind are outside what it measured.

Where a COA does give a sequence confirmation — MS/MS coverage, or an amino acid composition — note that this is a materially stronger identity claim than an intact-mass match. Most COAs do not contain it.

Purity: what fraction of what was detected is the target?

The purity figure on almost every peptide COA is a chromatographic area percentage: the target peak's integrated area divided by the summed area of all integrated peaks, usually from a reversed-phase gradient with UV detection in the low-200s nm. The number is real and useful, and it is conditional on three things a careful reader should look for: the wavelength, the gradient and column, and what the integration excluded.

Peptide quality control is commonly based on a generic reversed-phase method — a C18 stationary phase and an acetonitrile gradient with 0.1% trifluoroacetic acid in the mobile phase — which performs exceptionally well across a wide variety of impurities. The limitation is specific and well described: structurally closely related impurities with similar sequences, which were not resolved during preparative purification, may easily coelute in the corresponding QC run as well [3]. A single generic gradient can therefore report a high area percentage on material that a second, orthogonal separation would resolve into more than one component. That is not a defect in the number; it is the boundary of what one separation dimension can see.

"Purity by HPLC" and "purity" are not synonyms, and a COA that names its method is worth more than one reporting a bare percentage.

Area percent is not mass percent

A UV area percentage weights each species by how strongly it absorbs, not by how much of it is present. At 214 nm, measured in the presence of acetonitrile and formic acid, the peptide bond has a molar extinction coefficient of 923 M⁻¹ cm⁻¹; tryptophan absorbs approximately thirty times more strongly than a peptide bond, and phenylalanine, tyrosine, and histidine approximately six times more [4]. Two species in the same chromatogram therefore do not convert area to moles by the same factor.

Work an example with those measured coefficients. Take a ten-residue peptide containing one tryptophan — nine peptide bonds. Assuming its other nine residues contribute negligibly — true of most amino acids, though in-chain proline absorbs approximately three times a peptide bond and methionine about the same as one — its predicted absorptivity at 214 nm is roughly (9 x 923) + (30 x 923) = 8,307 + 27,690 = 35,997 M⁻¹ cm⁻¹. Now suppose a deletion impurity is present that is missing that tryptophan: nine residues, eight peptide bonds, no aromatic. Its predicted absorptivity is 8 x 923 = 7,384 M⁻¹ cm⁻¹. If the two are present in equal molar amounts, the area percentage reports the impurity as 7,384 / (35,997 + 7,384) = 17.0% — not 50%. The arithmetic runs the other way too: an impurity that gains an aromatic residue is over-reported.

The extinction coefficients here are measured values from a study that validated the additive prediction against several peptides and proteins [4]; the ten-residue peptide is a constructed illustration, not a real analysis. Area percent answers "what fraction of the absorbance is the target," and only approximates "what fraction of the mass is the target."

Content: how much peptide is actually in the vial?

Content — sometimes labeled peptide content, net peptide, or assay — is a mass question, and it requires a mass method. The mass balance approach assigns purity by subtracting everything that is not the target from the whole, which means every non-target component has to be measured separately.

A worked SI-traceable mass balance on a synthetic oxytocin material illustrates the categories. Structurally related peptide impurities — including deamidations, oxidations, and amino acid insertions — were identified and quantified by validated LC-HRMS, twenty-one of them, ranging from 0.05 mg/g to 15.65 mg/g. Separately, the non-peptide-related components were determined by their own techniques: water by coulometric Karl Fischer titration, non-volatile counterions by ion chromatography, inorganic elements by ICP-MS, and volatile organic compounds by headspace GC-MS. The resulting assigned value for that material was 796.5 mg/g, with an expanded uncertainty of 6.5 mg/g (k = 2) [5].

That figure says the dry solid was about 79.7% oxytocin by mass — the balance being water, counterion, elemental and solvent residues, and the structurally related peptide impurities. Purity and content answer different questions, and the gap between them is the mass that is not target peptide.

The consequence is immediate. Take a vial containing 10 mg of solid and suppose its COA reported that same content figure — a value measured on one batch of one peptide, not one to assume for material whose content was never determined: 10 mg x 0.7965 = 7.97 mg of peptide. Reconstituted in 2 mL, that is 7.97 / 2 = 3.98 mg/mL, not the 5.00 mg/mL the nominal fill mass implies — about 20% low. This is the number that propagates into every downstream molarity calculation, which is why the Echo Peptides concentration calculator notes that the label mass is nominal and that content should be taken from the lot's COA. A COA that reports a purity percentage but no content value has not given you the input that calculation needs.

Which content method was used, and does it matter?

It can. A multi-laboratory collaborative study coordinated by USP compared three approaches to quantifying the nonapeptide oxytocin — an HPLC assay, quantitative NMR, and amino acid analysis — across laboratories. In that study, the HPLC assay using the same peptide bulk material as the standard showed the lowest inter-laboratory variability; the authors are explicit that the coefficient of variation was calculated without counting the uncertainty associated with the purity assignment of the standard by mass balance. They describe qNMR as a direct measurement against an internal standard, not difficult to perform under common laboratory conditions, and conclude that its use as a primary method for peptide reference standard value assignment deserves further exploration [6]. That is a hedged conclusion about one peptide in one study, and it should be read as such.

Amino acid analysis is the third route: the peptide is hydrolyzed and the released amino acids are quantified, with peptide quantitation based on the summed amino acids, in one such method against certified isotopically labeled internal standards [7]. Whichever route a COA used, one requirement is shared. Working on oxytocin calibrators, the BIPM and NIM authors state that mass balance, amino acid analysis, qNMR, and nitrogen determination alike require a correction for structurally related peptide impurities in order to avoid biases; in their study material eighteen such impurities were identified and quantified by LC-HRMS, totaling 31.1 mg/g with an expanded uncertainty of 1.7 mg/g [8]. A content number stated without a method is a number you cannot audit.

The counterion is part of the weight

Solid-phase synthesis and purification rely on trifluoroacetic acid as a cleavage agent and ion-pairing reagent, so peptides are commonly obtained as TFA salts. Residual TFA can be detected and quantified by FT-IR, 19F-NMR, and HPLC with evaporative light-scattering detection; in one systematic study of counterion exchange, TFA removal with 10 mM HCl was determined to be the optimal condition among those tested, and no impact on peptide purity was observed at any HCl concentration examined. The measured counterion contents varied with sequence, particularly with positive charges, and the authors emphasize the importance of counterion quantification and specification for synthetic peptides used in assays [9].

A COA that names the salt form is therefore telling you something about the mass you weighed out. And purity and counterion load are separable: exchanging the counterion changed the salt without changing purity in that study [9] — another instance of the rule that these fields are independent.

Degradation hides in small mass differences

Peptides degrade by chemistry that shifts mass very little. In a stability characterization of recombinant human PTH (1-34) stored at 25 ± 2 °C, LC-MS confirmed deamidated (+1 Da) and succinimide (-17 Da) variants alongside three oxidized species — the sulfoxide of Met8, the sulfoxide of Met18, and a doubly oxidized variant — and low-molecular-weight species were confirmed by N-terminal sequencing and LC-MS to be fragments arising from backbone cleavage. The authors concluded that under those conditions the material degraded not only by oxidation but that deamidation and peptide bond cleavage were also prominent modes [10].

For COA reading this matters twice. A +1 Da or -17 Da species sits close enough to the parent that a low-resolution intact-mass identity check may not separate it, and the corresponding chromatographic peaks are exactly the "structurally closely related" impurities that a single generic gradient may fail to resolve [3]. And the COA describes the batch on its date of analysis. Degradation chemistry does not stop because a document was signed.

Microbiological results are not one result

Where a COA carries microbiological data, note that sterility and endotoxin are separate determinations that can disagree. In an evaluation of products purchased from illegal online sellers, the lyophilized samples were devoid of viable microorganisms at the time of testing, yet endotoxin was detected in all samples, at levels between 2.1645 and 8.9511 EU/mg [11]. Absence of viable organisms is not absence of endotoxin, and a document reporting one has not reported the other.

What a missing batch identifier means

A COA without a batch or lot identifier that matches the vial is not a weak COA. It is a document about an unspecified quantity of material, and there is no procedure that recovers the link. Every number discussed above — purity, content, counterion, water, impurity profile — is a property of a specific synthesis and purification campaign, not of a compound name. Two batches of the same sequence from the same manufacturer are different materials analytically.

The literature on out-of-quality-system peptide products is the empirical case for taking this seriously. A systematic screening of the ten most frequently encountered falsified peptide products on the Belgian market, obtained from three suspected illegal internet pharmacies, examined the active ingredient, ingredient-related impurities, small-molecule contaminants, elemental impurities, and residual solvents — a design premised on the fact that no single analysis characterizes such material [12]. In the online-seller study cited above, measured purity ranged from 7.7% to 14.37% against a 99% purity claimed on the product labels, while measured content exceeded the labeled amounts by 28.56% to 38.69% [11]. Those are the two numbers this page has been separating — one measured above the label, one measured far below the claim, on the same samples.

What a COA does not establish, stated plainly

  • It does not establish that your vial is the batch tested. That link is administrative — identifier on the document matching identifier on the vial — and nothing analytical replaces it.
  • It does not establish current condition. The measurements have a date. Storage and handling after that date are outside the document.
  • It does not establish anything the methods listed did not measure. A COA with RP-HPLC and MS reports chromatographic purity and a mass; it is silent on water content, counterion, elemental residues, and residual solvents unless those tests appear on it.
  • It does not establish behavior in an assay. A COA is a chemistry document.

How much of this is convention

Considerably more than the format suggests. The pharmacopeial and metrological literature on peptide characterization is detailed, and where a COA follows it — a named method, a stated wavelength, a content value with an assignment route, an uncertainty — you can trace what was done. But the layout of a commercial COA, the choice to report area percent rather than mass fraction, and the decision about which impurities to integrate and which to treat as solvent front are conventions of practice, not standards with primary literature behind each choice. Treat the fields you can trace as evidence and the rest as a claim of the same rank as any other statement from the supplier — and note that a supplier's COA and an independent laboratory's COA on the same lot are two different kinds of document.

A workable reading order: find the batch identifier and match it; find the date; read which methods were run and at what wavelength; separate the identity result from the purity result from the content result; check whether a content value exists at all before using the nominal fill mass in a calculation; and note what is absent. What a COA does not say is part of what it says.