How to Read a Certificate of Analysis

What a certificate of analysis is, and what it is not

A certificate of analysis (COA) is a document that reports the results of analytical tests performed on a defined lot of material. It is downstream of a second document that is frequently conflated with it: the specification. In the vocabulary of ICH Q6A, a specification is a list of tests, references to the analytical procedures used, and appropriate acceptance criteria — numerical limits, ranges, or other criteria that the results must satisfy. The specification is the standard; the certificate is the record of what was measured against that standard. A certificate that reports values without showing the acceptance criteria they were judged against has given the reader half of the pair.

Several things a certificate is not. It is not a comprehensive characterization of the material. It reports what was tested, and nothing about what was not tested; the absence of a field is not evidence of the absence of the thing that field would have measured. It is not forward-looking. Every value on the document describes the sample as it existed at the moment of analysis, and the certificate makes no assertion about the material’s condition at any later date. It is also not a statement of fitness for any particular use. Suitability is a judgment the receiving laboratory makes against its own requirements, using the certificate as one input among several.

There is a further scope limitation that is easy to overlook. Results apply, strictly, to the sample that was analyzed. They apply to the lot only by inference, and the strength of that inference depends entirely on how the sample was drawn — whether it was a single grab from one container or a composite across the lot, and whether the certificate says which. Finally, format is not provenance. A certificate issued for a research-grade chemical may be laid out identically to one issued for a pharmaceutical ingredient manufactured under GMP, but the two are produced under entirely different obligations, and the visual resemblance carries no information about which regime applied.

Provenance: an in-house quality-control certificate versus an independent third-party report

The single most consequential fact about any certificate is who generated the data. A manufacturer’s own quality-control certificate is a self-report: the party that produced the material also performed the measurement and wrote the document. This is not by itself a criticism. In-house QC laboratories are the norm across the chemical industry, they usually know the material and its method better than anyone, and the vast majority of such certificates are produced in good faith by competent analysts. But the structural fact remains — the analysis and the material originate with the same party — and it is a different kind of evidence from a report produced by a laboratory with nothing at stake in the result.

An independent third-party report carries concrete, checkable markers, and a reader can usually determine in under a minute which kind of document is in front of them. Such a report names a testing laboratory as an organization distinct from the manufacturer, with its own address. It carries the testing laboratory’s own report, job, or work-order number, separate from the product lot number. It describes the sample as received, together with the date received and the client identifier the laboratory assigned. It gives an accreditation reference — most commonly ISO/IEC 17025, which sets out requirements for the competence, impartiality, and consistent operation of testing laboratories — naming the accrediting body and the accreditation certificate number. And it identifies an authorized signatory belonging to the testing laboratory. A document carrying all of these was written by someone other than the manufacturer. A document carrying none of them was not.

The ambiguous middle deserves attention, because that is where most real documents sit. A certificate printed on a manufacturer’s letterhead may reproduce a chromatogram generated elsewhere without naming the laboratory that produced it; the chromatogram is then an unattributed image. An accreditation logo without a scope reference is weaker than it appears, because accreditation is granted for a defined scope of methods and matrices — a laboratory accredited for one technique is not thereby accredited for every technique it performs, and a report may legitimately fall outside its own laboratory’s accredited scope. A PDF with no signatory, no report number, and no dates is a formatted table, whatever it is titled.

Two linkage checks apply to any certificate regardless of origin. First, the lot or batch number on the document must match the number on the container. A certificate cannot be transferred between lots, and a certificate for a different lot is not weaker evidence about the lot at hand; it is no evidence at all. Second, a third-party report describes a sample the laboratory received from a client. The laboratory can vouch for what it measured; it generally cannot vouch that the sample it received came from any particular container. That link is documentary rather than physical, and it is only as strong as the chain-of-custody information the report contains.

Appearance and identity

The appearance or description field is normally a short qualitative statement — a white to off-white lyophilized powder, a white amorphous solid, a clear colorless solution. It carries little analytical information on its own, and it is tempting to skip. It is worth reading anyway for one reason: a gross discrepancy between the described appearance and the material actually observed is a signal that the document and the container may not correspond, which is a document-control problem rather than a chemistry problem, and one the certificate cannot flag on its own. Some certificates also carry a solubility line. Where present, it records a physical-property observation made by the testing laboratory under conditions of that laboratory’s choosing, and without those conditions stated it is not interpretable. It is a measured property of the substance, not an instruction, and should not be read as one.

Identity is usually established by mass spectrometry, most often electrospray ionization (ESI-MS) or matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF). A well-constructed field reports both the observed mass and the theoretical mass calculated from the claimed sequence, and states which mass convention is in use. That last point is where certificates most often fall short. Monoisotopic mass is calculated from the most abundant isotope of each element; average mass is weighted across natural isotopic abundance. For small peptides on instruments with adequate resolution the monoisotopic value is the meaningful one; for larger peptides on lower-resolution instruments the isotope envelope is unresolved and the reported figure is an average mass. The two diverge progressively with molecular size, so a line reading “observed 3648.5, theoretical 3647.9” cannot be evaluated at all unless the reader knows which convention was intended. Related conventions matter too: whether the value is the neutral mass or an observed ion such as [M+H]⁺ or a multiply charged species, and whether the free base or the salt form was used for the calculation.

What mass spectrometry establishes, and what it does not, is worth stating plainly. A mass match confirms that the molecular mass of the principal species is consistent with the claimed composition. It does not confirm sequence. Leucine and isoleucine are exactly isobaric. Lysine and glutamine differ by roughly 0.036 Da, which is below the resolving power of many routine instruments. Any transposition of residues within a sequence is mass-neutral by definition, and D-amino acid substitutions are mass-identical to their L counterparts. Confirming sequence requires tandem mass spectrometry with fragmentation, or Edman degradation; amino acid analysis establishes composition but not order; chiral purity requires a separate determination. A certificate that reports a single intact mass under a heading reading “Identity” is therefore making a narrower claim than the heading implies, and reading the heading rather than the method is a common error.

Purity by HPLC, and what area percent actually means

The standard purity determination for a synthetic peptide is reversed-phase HPLC with ultraviolet detection, typically at 210–220 nm, where the peptide bond itself absorbs, and sometimes additionally at 280 nm, where tryptophan and tyrosine absorb. The number that appears on the certificate is almost always an area percent: the integrated area of the main peak, divided by the summed integrated area of all detected peaks, times one hundred. Nothing about that calculation involves mass. It is a ratio of detector responses.

Area percent rests on an assumption that is worth making explicit, because it is rarely stated on the document itself: that every species present produces the same detector response per unit mass at the chosen wavelength — that all relative response factors equal one. This assumption degrades whenever an impurity’s chromophore differs from that of the main compound. At 214 nm the dominant chromophore is the amide bond, so a deletion sequence missing one residue has one fewer amide bond and responds slightly less per unit mass; a short truncated fragment responds substantially less; and at 280 nm any impurity lacking an aromatic residue is simply invisible. D’Hondt and colleagues survey the impurity classes that arise in peptide synthesis and purification — deletion and truncation sequences, side-chain modifications, aggregation and oxidation products — and the analytical consequences of each. ICH Q2(R2) treats response-factor considerations as part of validating a procedure for its stated purpose, which is the framework under which a purity method is judged adequate or not.

Area percent also counts only species that both elute during the run and absorb at the detection wavelength. Several categories of material satisfy neither condition and are absent from the denominator entirely: inorganic salts, water, counterions (acetate and trifluoroacetate have only weak end absorbance and are typically eluted in or near the void), non-chromophoric scavengers and reagent residues, and any material that fails to elute at all — irreversibly retained aggregates or species too hydrophobic for the gradient used. The largest single limitation, however, is co-elution. A single-wavelength trace cannot distinguish one peak from two unresolved peaks. What addresses this is peak-purity assessment using diode-array spectral comparison across the peak, or an orthogonal separation — a different stationary phase chemistry, a different mobile-phase pH, or LC-MS on the main peak. A certificate reporting a single chromatographic condition reports purity under that condition, and the reader has no basis for extending it to any other.

Two further details determine how much a purity figure is worth. The first is method disclosure: column chemistry and dimensions, mobile phase composition and gradient, flow rate, column temperature, detection wavelength, and injection amount. The more of this the certificate prints, the more the number can be independently evaluated or reproduced. The second is the integration convention, including any disregard limit below which small peaks were excluded from the summation, and how the baseline was drawn. Integration is a judgment, and two competent analysts can obtain different area percents from the same raw data. The practical conclusion is that a figure such as “99% purity” is a statement about the composition of the UV-absorbing, eluting fraction under one set of conditions. It is not a statement that 99% of the solid’s mass is the named compound. That second quantity has its own name and its own methods.

Mass balance of a lyophilized solid: water, counterion, and net peptide content

A lyophilized peptide solid is a mixture by mass. It contains the peptide itself, water, one or more counterions, and residual salts and reagents carried through from synthesis and purification. HPLC purity addresses the internal composition of the peptide-related, chromatographically visible fraction. The remaining certificate fields address everything else in the mass, and together they are what allow a reader to reconstruct an approximate mass balance for the material.

Water content is normally determined by Karl Fischer titration, the method introduced by Karl Fischer in 1935, which exploits the stoichiometric reaction of water with iodine and sulfur dioxide in an alcohol medium containing a base. The endpoint is determined volumetrically for higher water contents, or coulometrically, with iodine generated electrochemically, for low water contents on small samples. USP <921> describes the pharmacopeial implementations. Lyophilized peptide solids are commonly hygroscopic, and reported water contents ranging from a few percent to well above ten percent are unremarkable depending on the material and the lyophilization cycle. Because the analyte is ubiquitous in laboratory air, Karl Fischer results carry a genuine method- and handling-dependent variability, and sample handling during transfer into the titration vessel is part of the measurement rather than incidental to it. Loss on drying is a different determination and is not interchangeable: it measures all volatiles lost under defined conditions, of which water is only one, so a loss-on-drying figure is generally equal to or greater than a Karl Fischer figure on the same material.

Counterion content follows from how the peptide was purified and isolated. Peptides containing basic residues are isolated as salts, and reversed-phase purification using trifluoroacetic acid as an ion-pairing modifier leaves trifluoroacetate as the associated counterion. An acetate salt form is the result of a deliberate exchange step — ion-exchange chromatography, or repeated lyophilization from acetic acid. Roux and colleagues evaluated several such approaches for a dicationic peptide and documented that trifluoroacetate is tightly associated and that its removal is not trivially complete, which is the reason a certificate stating “acetate salt” is describing a form rather than certifying the absence of trifluoroacetate. The relevant determinations are acetate by ion chromatography or HPLC, trifluoroacetate by ion chromatography or ¹⁹F NMR, and total fluorine by combustion ion chromatography. Counterion mass is real mass: for a peptide with several basic residues it can account for a meaningful fraction of the solid. A certificate that names a salt form but reports no counterion assay has left that portion of the mass unquantified.

Net peptide content is the fraction of total solid mass that is peptide, excluding water, counterion, and residual salts. Two classical routes appear on certificates. The first is nitrogen determination: total nitrogen is measured by Kjeldahl digestion, the sulfuric-acid method Johan Kjeldahl published in 1883, or by combustion (Dumas) elemental analysis, and the result is converted to peptide mass using the nitrogen mass fraction calculated from the claimed sequence. The conversion is only as good as that assumption, and the method counts nitrogen from any source — residual ammonium salts, dimethylformamide, or piperidine all contribute nitrogen and inflate the apparent peptide content, whereas trifluoroacetate contains none and does not. The second route is quantitative amino acid analysis: complete acid hydrolysis followed by derivatization and chromatographic quantitation of the liberated amino acids against calibrated standards, along the lines described in USP <1052>. Amino acid analysis is more informative but more laborious, and it has its own well-known systematic issues with tryptophan destruction and incomplete hydrolysis of certain bonds.

This is why purity and net peptide content are different numbers, and why one should never be substituted for the other. Purity is a chromatographic ratio describing composition within the peptide-related fraction; net peptide content is a gravimetric fraction describing the whole solid. For a salt-form lyophilized peptide the two commonly differ by ten to twenty-five percentage points. A material can legitimately be 99% pure by area percent and 80% peptide by mass; neither figure is wrong, and neither can be derived from the other. The distinction matters for any quantitative record that cites a mass, because a stated milligram figure may refer to gross lyophilized solid or to net peptide, and both conventions are in circulation. Where a certificate reports a fill weight together with water content, counterion content, and net peptide content, the relationship between the two bases can be reconstructed arithmetically. Where a single figure appears with no stated basis, it is an unlabeled quantity, and every downstream calculation that cites it inherits that ambiguity.

Residual solvents and other trace-level fields

Solid-phase synthesis, resin cleavage, and chromatographic purification all introduce organic solvents and reagents that can persist in the isolated solid: dimethylformamide, N-methylpyrrolidone, dichloromethane, acetonitrile, methanol, diethyl ether, trifluoroacetic acid, bases such as piperidine and diisopropylethylamine, and scavengers such as triisopropylsilane and ethanedithiol. Residual solvent determination is conventionally performed by headspace gas chromatography with flame ionization or mass-spectrometric detection, and the analytical challenge is one of volatility and matrix: the method must partition the analyte into the headspace reproducibly from a solid that may retain it strongly. ICH Q3C sorts solvents into classes according to the toxicological risk assessment underlying the guideline — Class 1 to be avoided, Class 2 limited, Class 3 of lower concern — and specifies concentration limits accordingly; USP <467> implements the same framework pharmacopeially. Revision 9 of the guideline reached Step 4 of the ICH process on 24 January 2024.

Other fields appear on certificates with varying frequency, and each is a distinct determination with its own method and its own acceptance criterion: elemental impurities by ICP-MS, framed by ICH Q3D; bacterial endotoxin by LAL or recombinant factor C assay; bioburden or sterility; chiral purity or D-amino acid content, typically by chiral GC or LC after hydrolysis; sequence confirmation by tandem MS or Edman degradation; and specific optical rotation. The general rule that applies to all of them is the same rule that applies to the core fields: a result is interpretable only in the company of the method that produced it and the criterion it was judged against. And the absence of any of these fields means only that the test was not reported. It is not a negative result, and it should not be read as one.

The conclusion statement, and reading the document as a whole

Most certificates end with a disposition line — “Conforms to specification,” “Passes,” “Meets in-house specification,” or similar. It is worth being precise about what this sentence asserts. It asserts that the measured values agreed with a set of pre-established acceptance criteria. It asserts nothing else: not that the material is well characterized, not that the tests performed were the right ones, and not that the criteria were demanding. Its informational content is therefore exactly as large as the specification standing behind it, and that specification is frequently not printed on the certificate. “Conforms” against an undisclosed in-house specification approaches zero information. “Conforms” printed alongside each test, its method, its acceptance criterion, and its measured result is fully checkable, and lets the reader see not just whether each result passed but by how much margin — a result sitting at the edge of its limit and one sitting comfortably inside it are reported identically by the word “conforms.”

A workable reading order puts provenance first and numbers last, because the numbers cannot be weighed until their source is known. Who performed the testing, and is that entity distinct from the manufacturer. Does the lot number on the document match the container. Is there a report number, a date of analysis, a date of issue, and a named signatory. Is each result accompanied by the method that produced it. Are acceptance criteria shown next to results. For any mass figure, is the basis — gross solid or net peptide — stated explicitly. Is there an accreditation reference, and does it name both the accrediting body and a scope or certificate number. One further question separates strong documents from weak ones: whether raw data are appended. A certificate with the actual chromatograms and mass spectra attached permits an independent reader to evaluate baseline placement, integration, peak shape, and spectral quality. A certificate with only a table of numbers requires the reader to accept those judgments unseen.

The underlying posture is that a certificate is a piece of evidence, and evidence is weighed rather than accepted or rejected. Its weight depends on who produced it, how specific it is about method, and how much of the underlying data it exposes. The reader’s task is not to decide whether the document is trustworthy in the abstract but to determine precisely what it asserts, who is asserting it, and — usually the more informative question — what it leaves entirely unaddressed. Where a laboratory’s own work depends on a property the certificate does not pin down, no certificate substitutes for determining that property in-house.

References

1. International Council for Harmonisation. ICH Q6A: Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and New Drug Products: Chemical Substances. Step 4 version, 6 October 1999.
2. ISO/IEC 17025:2017. General requirements for the competence of testing and calibration laboratories. 3rd edition. International Organization for Standardization / International Electrotechnical Commission, Geneva, 2017.
3. D’Hondt, M.; Bracke, N.; Taevernier, L.; Gevaert, B.; Verbeke, F.; Wynendaele, E.; De Spiegeleer, B. Related impurities in peptide medicines. Journal of Pharmaceutical and Biomedical Analysis, 2014, 101, 2–30. DOI: 10.1016/j.jpba.2014.06.012; PMID: 25044089.
4. International Council for Harmonisation. ICH Q2(R2): Validation of Analytical Procedures. Step 4 version adopted 1 November 2023.
5. Fischer, K. Neues Verfahren zur maßanalytischen Bestimmung des Wassergehaltes von Flüssigkeiten und festen Körpern. Angewandte Chemie, 1935, 48(26), 394–396. DOI: 10.1002/ange.19350482605.
6. United States Pharmacopeia. General Chapter <921> Water Determination. USP–NF, United States Pharmacopeial Convention, Rockville, MD.
7. Roux, S.; Zékri, E.; Rousseau, B.; Paternostre, M.; Cintrat, J.-C.; Fay, N. Elimination and exchange of trifluoroacetate counter-ion from cationic peptides: a critical evaluation of different approaches. Journal of Peptide Science, 2008, 14(3), 354–359. DOI: 10.1002/psc.951; PMID: 18035848.
8. Kjeldahl, J. Neue Methode zur Bestimmung des Stickstoffs in organischen Körpern. Zeitschrift für analytische Chemie, 1883, 22, 366–382. DOI: 10.1007/BF01338151.
9. United States Pharmacopeia. General Chapter <1052> Biotechnology-Derived Articles — Amino Acid Analysis. USP–NF, United States Pharmacopeial Convention, Rockville, MD.
10. International Council for Harmonisation. ICH Q3C(R9): Impurities: Guideline for Residual Solvents. Step 4 version, 24 January 2024.
11. United States Pharmacopeia. General Chapter <467> Residual Solvents. USP–NF, United States Pharmacopeial Convention, Rockville, MD.