Reference Standards, and Why Two Purity Figures Are Not Comparable
Ninety-nine point four one percent, against what?
That is not a rhetorical question and it does not have a comfortable answer. A purity figure is not a property that a peptide possesses. It is the output of a comparison — this peak against those peaks, on this column, with this detector, calibrated against some material that somebody decided was the reference. Change any term in that comparison and the number moves. Two certificates from two suppliers, each reading 99.4%, are two different comparisons that happen to have landed on the same digits.
This article is about what sits underneath that number, what the international metrology system has learned by testing it directly, and what a research-grade certificate can and cannot honestly claim as a result.
All material discussed here is supplied for research use only and is not intended for human use.
What a reference standard actually is
The vocabulary here is precise and worth using precisely, because the marketing use of “reference standard” is much looser than the metrological one.
The National Institute of Standards and Technology defines a reference material as a material sufficiently homogeneous and stable with respect to specified properties to be fit for its intended use in a measurement process. A certified reference material is a narrower thing: one “characterized by a metrologically valid procedure for one or more specified properties, accompanied by a certificate that provides the value of the specified property, its associated uncertainty, and a statement of metrological traceability.” A NIST Standard Reference Material is a certified reference material delivering at least one certified property value traceable to internationally recognized higher-order reference systems.
Three requirements are doing the work in that definition: a stated value, a stated uncertainty, and a stated traceability chain. A number without an uncertainty is not a certified value in this system. It is an assertion.
What happened when NIST tried this on peptides
The instructive case is a NIST reference material built from three short synthetic peptides: a 14-residue sequence at roughly 1,562 g/mol, a 26-residue sequence at roughly 2,949 g/mol, and an 11-residue sequence at roughly 1,290 g/mol (Reference Material 8327, report of investigation, National Institute of Standards and Technology).
NIST issued no certified values for purity on that material. The purity figures are reference values, which the document itself defines as “a noncertified value that is the best estimate of the true value; however, the value does not meet NIST criteria for certification and is provided with associated uncertainties that may reflect only measurement precision and may not include all sources of uncertainty.”
The uncertainties are the part to sit with. Total ultraviolet area-percent purity carries expanded uncertainties of ±3.6% to ±12.6% at a coverage factor of 2.776. Percent peptide mass purity — determined by amino acid analysis and ultraviolet spectrophotometry at 280 nm — carries ±11% to ±16% at a coverage factor of 4.303. Experimental molecular mass, by contrast, is tight: ±0.37 to ±1.7 g/mol.
Read that ordering carefully. Identity, the national metrology institute can pin down to a fraction of a dalton. Purity, on peptides of eleven to twenty-six residues, it could not certify at all, and the mass-purity interval it published is wider than ten percentage points.
Now set that beside a commercial certificate reporting 99.41% to two decimal places with no uncertainty statement of any kind. The commercial figure is not necessarily wrong. It is a different kind of statement, made under different rules, and it should not be read as though it were the other kind.
When expert laboratories measure the same batch
The international key comparison program run through the CCQM Protein Analysis Working Group tests exactly this by sending aliquots of one homogeneous batch to national metrology institutes and comparing what comes back.
On synthetic oxytocin, nine laboratories participated. The key comparison reference value for oxytocin mass fraction was 787.2 mg/g with an expanded uncertainty of 12.9 mg/g. The values the participants reported ran from 766.3 mg/g to 987 mg/g (CCQM-K115.b, key comparison study on peptide purity, synthetic oxytocin, Bureau International des Poids et Mesures final report). That is a spread of roughly 220 mg/g — about 28% of the reference value — among national measurement institutes, on subsamples of a single batch, using declared and documented methods.
The follow-on pilot study on the HbA0 hexapeptide VHLTPE was tighter in absolute terms and worse in relative agreement. Participant values ran 608.7 to 641.2 mg/g against an assigned 613 ± 20 mg/g, and the report notes “less agreement of participants’ results” than the earlier comparisons in the series (CCQM-P55.2.2018, pilot study on peptide purity, hexapeptide of HbA0, BIPM final report).
The diagnosis in both reports lands on the same place, and it is not the main peak. It is the impurities. In the hexapeptide study, every pilot participant failed to identify the major depsipeptide impurity, which systematically pushed total impurities down and the peptide value up. In the oxytocin study, misidentification and misquantification ran the other way for some laboratories, inflating the impurity total and deflating the assigned oxytocin content.
That is the mechanism behind interlaboratory disagreement on peptide purity, and it is worth stating plainly: laboratories do not disagree much about the main peak. They disagree about what is in the rest of the chromatogram, and a purity percentage is defined by the rest of the chromatogram.
The comparison that looks reassuring until you read the caveat
The United States Pharmacopeia ran a multi-laboratory collaborative study on the same question, comparing an HPLC assay, quantitative NMR and amino acid analysis for quantitation of oxytocin (Li, Bhavaraju, Melanson, Blomgren, Rundlöf, Kilpatrick, Rudd, Aubin and colleagues, survey of peptide quantification methods and comparison of their reproducibility, a case study using oxytocin, Journal of Pharmaceutical and Biomedical Analysis 2019, 166:105–112, DOI 10.1016/j.jpba.2018.12.028).
The headline result is that the HPLC assay — run against the same peptide bulk material as the standard — showed the lowest inter-laboratory variability. That finding is frequently quoted, and quoting it without the next sentence inverts its meaning. The authors state that the coefficient of variation “was calculated without counting the uncertainty associated with the purity assignment of the standard with mass balance.”
HPLC came out most reproducible because the hard part had been removed from the tally. Every participating laboratory was comparing against the same bulk material, so they agreed with each other about a quantity whose own uncertainty was excluded from the calculation. Reproducibility between laboratories using a shared standard is a real and useful property. It is not accuracy, and it says nothing about how that shared standard’s value was established. The authors’ own conclusion points the other way — that qNMR, as a direct measurement against an internal calibrant, “deserves further exploration” as a primary method for value assignment.
USP’s published work on peptide reference standards makes the same point from the other end: across their value-assignment program, the mass balance approach yielded the least inter-laboratory variability, and a worked desmopressin acetate assignment ended at 1.75 mg per vial on an as-is basis with a 95% confidence interval of 1.654 to 1.836 after outlying laboratory data were removed (McCarthy, Han, Carrick, Schmidt, Workman, Matejtschuk, Duru and Atouf, reference standards supporting quality of synthetic peptide therapeutics, Pharmaceutical Research 2023, 40(6):1317–1328, DOI 10.1007/s11095-023-03493-1).
A confidence interval on a reference standard’s content is the thing research-grade certificates in this market never print.
What this means for a research peptide certificate
The BPC-157 10 mg batch certificate — report DBAV-BPC-10-051226, received 19 May 2026, tested 23 May, reported 25 May — states an expected content of 10.00 mg, an actual mass of 10.29 mg, a fill accuracy of 102.9% and a purity of 99.41% by HPLC-UV at 214 nm on a white lyophilized powder. We read it in full this week, and it contains a sample identification block and that purity-and-content section, and nothing else.
Here is the honest description of what the 99.41% is traceable to. It is the integration of one chromatogram, on one instrument, at one contract laboratory, on one date, using a method the certificate names and a wavelength it names. There is no certified reference material for BPC-157 — none exists, and given the economics of certification for a compound with no pharmacopeial monograph, none is likely to. So the figure cannot be traceable to the SI in the sense the NIST definition requires. It is traceable to a laboratory’s procedure.
That is the convention across the entire United States research peptide market, not a peculiarity of one supplier, and a domestic operation shipping same-day from Santa Barbara is working under the same constraint as every other. What follows from it is specific and practical:
A purity number is not comparable between suppliers. Two vendors reporting 99.4% have not been measured against a common reference, because no common reference exists. Ranking suppliers by the digits on their certificates is comparing procedures, not materials.
What is comparable is the certificate’s structure. Whether it names the analysing laboratory. Whether it names a report number, a lot, and the dates it was received, tested and reported. Whether it states the method and the detection wavelength. Whether it gives an expected mass alongside the measured one, so the reader can see fill accuracy rather than compute it. Whether it is specific to the batch in the box or is a generic document reused across shipments. Those are verifiable facts about a document, and unlike the purity figure they mean the same thing on every certificate you compare.
A certificate with no lot identifier cannot be tied to a batch, which is the single function a batch certificate performs. Where a lot field is present and populated, it is doing real work; where it is blank, the document has become an assertion about a product line rather than a measurement of a shipment. This is worth checking on any certificate, including ours.
None of this makes chromatographic purity useless. It is a genuine constraint on how bad a batch can be, it catches gross synthesis failures, and it is the reason a certificate is better than no certificate. It is simply a narrower claim than the two-decimal presentation implies, and the metrology literature is unusually clear about exactly how much narrower.
All material referenced on this site is supplied for research purposes only. It is not for human use, not for diagnostic use, and not for therapeutic application. The comparison studies described above were conducted on reference and laboratory materials by national metrology institutes and pharmacopeial bodies, and are reported here as published methodology, not as a claim about any product.
References
- National Institute of Standards and Technology — Standard Reference Materials Program definitions: reference material, certified reference material, Standard Reference Material, certified value, reference value.
- National Institute of Standards and Technology — Reference Material 8327, three synthetic peptides, report of investigation.
- Bureau International des Poids et Mesures — CCQM-K115.b, key comparison study on peptide purity, synthetic oxytocin, final report, CCQM Protein Analysis Working Group.
- Bureau International des Poids et Mesures — CCQM-P55.2.2018, pilot study on peptide purity, hexapeptide of HbA0, final report.
- Li, Bhavaraju, Melanson, Blomgren, Rundlöf, Kilpatrick, Rudd, Aubin and colleagues — survey of peptide quantification methods and comparison of their reproducibility, a case study using oxytocin, Journal of Pharmaceutical and Biomedical Analysis 2019, 166:105–112, DOI 10.1016/j.jpba.2018.12.028.
- McCarthy, Han, Carrick, Schmidt, Workman, Matejtschuk, Duru and Atouf — reference standards supporting quality of synthetic peptide therapeutics, Pharmaceutical Research 2023, 40(6):1317–1328, DOI 10.1007/s11095-023-03493-1.
- Melanson, Thibeault, Stocks, Leek, McRae and Meija — purity assignment for peptide certified reference materials combining qNMR and LC-MS/MS amino acid analysis, applied to angiotensin II, Analytical and Bioanalytical Chemistry 2018, 410(26):6719–6731, DOI 10.1007/s00216-018-1272-7.
- Testides Analytical — certificate of analysis DBAV-BPC-10-051226, BPC-157 10 mg, reported 25 May 2026.