What an HPLC purity number does not cover

What is the number, mechanically?

The compendial definition: under the normalization procedure in the chromatography chapter, and provided linearity of the peaks has been demonstrated, the percentage content of a component is calculated by taking the area of its peak as a percentage of the total area of all the peaks, excluding those due to solvents or reagents, those arising from the mobile phase or the sample matrix, and those at or below the disregard limit or reporting threshold.

Read that sentence twice: the denominator is not the sample. The denominator is the set of peaks the method produced and the integration kept.

The threshold that sets the floor: the same chapter states the reporting threshold, the limit above which a peak is reported, is generally 0.05 percent. Everything below it is discarded before the arithmetic starts.

The exclusions in the same sentence: peaks from solvents and reagents, and peaks arising from the mobile phase or the sample matrix, are disregarded by design. That is correct practice, and it means a purity figure is silent about that whole category of content.

Why is area not mass?

The compendial term: detector sensitivity is the signal output per unit concentration or unit mass of a substance, also called the response factor, and the relative response factor expresses that sensitivity against a standard substance. Correction factors are applied in a related-substances test only where a monograph indicates them.

Which means: in the absence of a monograph correction factor, an area percent is being read as though every peak responds identically. Nothing in the chemistry makes that true.

The numbers behind that: measured at 214nm, the peptide bond has a molar extinction coefficient of 923 M-1 cm-1. Tryptophan absorbs roughly 30 times more strongly than a peptide bond, while phenylalanine, tyrosine and histidine absorb roughly 6 times more, and most other residues absorb far less (Kuipers and Gruppen, 2007).

The consequence for an impurity: two impurities at the same mass, one carrying a strongly absorbing group (a chromophore) and one not, report as different areas. The area percent is a weighted view of the sample, weighted by something that has nothing to do with how much of each is present.

What a single gradient can hide

The compendial admission: the chromatography chapter instructs that integration of the peak area of any impurity not completely separated from the principal peak is preferably performed by tangential skim. The chapter supplies a procedure for the case where separation failed. It does not say how often that case arises, and neither do we.

Why peptides make it worse: a deletion sequence differs from the intended molecule by one residue, and a racemized residue produces a diastereomer that is otherwise identical. Both classes are documented products of solid-phase synthesis (D'Hondt, 2014). A diastereomer carries the same formula and the same mass as the intended peptide, and a deletion sequence differs from it by a single residue, so one column and one gradient is not guaranteed to resolve either of them from the main peak.

The standard answer: orthogonal methods. On the sameness question, FDA's draft guidance asks applicants for more than one analytical method to characterize primary sequence and physicochemical properties, secondary structure, oligomer and aggregation states, and biological activities. One system, one number, is the weakest form of the evidence.

What that means for a certificate: a single purity figure carries only what one separation and one detector resolved. The guidance asks for more than one method because one does not settle the characterization. That recommendation is written for applicants establishing that a generic peptide is the same as an approved one, not for vendor certificates, so it describes what a thorough characterization looks like rather than a standard any certificate is obliged to meet.

The list of what a purity number does not cover

How to read this: each row is a question people believe the purity figure answers, with the test that answers it instead.

The pattern across the table: the purity figure answers one narrow question well. Every other question on a buyer's mind is answered by a different test on a different sample.

What does 99 percent purity mean?

It means the main peak accounted for 99 percent of the integrated area of the peaks that eluted and were detected, after the excluded categories were removed. It is a statement about a chromatogram produced by one method on one day, not a statement about the contents of a vial.

The reframe: the useful reading is not the headline. It is the impurity profile underneath. In the FDA guidance cited below, which kind of application a synthetic peptide needs turns largely on the data required to evaluate the difference in impurities against the previously approved product, because peptide-related impurities may create the potential for differences in immunogenicity.

Can something be 99 percent pure and still be mostly not peptide?

Yes, and that is the single most useful thing to understand about the number. Area percent is normalized over the peaks the method produced and the integration kept; water is measured separately under the water determination chapter and the counterion by its own dedicated method, and neither enters the calculation. A solid can therefore be almost entirely one species by chromatographic area while a share of its weighed mass is water and salt.

Which line closes that gap: peptide content, sometimes called net peptide content, reported against a weighed sample. Without it, nothing on the certificate states how much of the labeled milligram figure is peptide, which is a gap in what was reported rather than evidence that no measurement exists.

Why this matters more than it sounds: every downstream calculation about concentration starts from the assumption that the labeled mass is peptide. If that assumption is wrong, the arithmetic is wrong by the same factor, silently.

Does an HPLC purity number say anything about sterility or endotoxin?

Nothing at all. Sterility is established under USP General Chapter 71, a microbiological test, and bacterial endotoxin under General Chapter 85. Neither result can be inferred from a chromatogram, because a chromatogram is a chemical separation run on a dissolved sample and neither organisms nor endotoxin appear in it as peaks.

The asymmetry worth naming: a purity run is a chemical separation on a dissolved sample. Sterility carries its own sampling plan and incubation period under General Chapter 71, and the endotoxin test its own preparation under General Chapter 85. Those two are the results that bear on whether a preparation could be injected safely, and a purity figure is silent about both.

Why do two purity numbers for the same material differ?

Because a purity figure is method-dependent by construction. Change the column, the gradient, the buffer or the detection wavelength and both the separation and the relative responses change, which moves the area percent without anything in the vial changing.

The compendial layer on top: peptide specifications are barely harmonized between pharmacopoeias, with large differences between the European and United States compendia, and monographs are not always consistent within a single pharmacopoeia (Vergote, 2009). Two defensible methods can be built to two different standards.

The practical test: a purity figure is comparable to another only when the method, the wavelength and the system suitability criteria are stated and match. Without those, two numbers are two different measurements sharing a unit.

What stays unsettled

The open items even with a purity figure in hand:

• Detection wavelength. It changes which impurities register and how strongly, so a figure printed without it cannot be compared with another.
• Separation quality. Resolution and system suitability decide whether close impurities were separated or skimmed off the main peak.
• Everything below the threshold. Peaks at or under the reporting threshold are excluded before the calculation.
• Mass balance. Without water content, counterion content and peptide content, the fraction of the solid that is peptide is unknown.
• The vial. A result belongs to a tested sample, and connecting it to a specific container is a chain-of-custody question, not an analytical one.

How this page is sourced

Definitions from the definition source: the normalization procedure, the reporting threshold, the response factor language and the tangential skim instruction were read in the compendial chromatography chapter itself. Every compendial chapter named on this page is linked to its official USP-NF record and DOI.

One citation was replaced on 29 July 2026: an earlier FDA guidance on synthetic peptide applications no longer resolves at its FDA locator, so the two claims that rested on it were re-sourced to a current FDA guidance that does resolve and that carries the same language.

What changed on 30 July 2026: a citation audit found four sentences reaching past the documents underneath them, and no reference was added or removed to fix any of them. The tangential-skim instruction no longer carries an inference about how common coelution is, because the chapter gives an integration procedure and no frequency. The orthogonal-methods paragraph now states what the guidance asks for instead of concluding what a certificate has failed to exclude. The endotoxin row no longer asserts anything about heat, only that endotoxin is not a chromatographic peak. And the peptide content paragraph now says the certificate is silent on the labeled mass rather than that the mass was never measured. On 31 July 2026 a fifth sentence was rescoped the same way: the deletion-and-diastereomer paragraph no longer asserts that both classes travel close to the main peak, a claim D'Hondt's abstract does not carry and whose full text is paywalled, and instead states what the chemistry establishes, that one column and one gradient is not guaranteed to resolve them.

Seven sources: three papers with DOIs and PubMed IDs, one FDA guidance, two compendial entries naming the chapters by number, and one ICH entry covering the two guidances named where they apply.

The standard: the full sourcing and citation-verification standard for this site, including who checks a page before it ships, lives on the methodology page.

Last reviewed: 29 July 2026.

1. 1

   United States Pharmacopeia. General Chapter 621, Chromatography. Quantification section, normalization procedure; Other Considerations, detector response and measurement of peaks; reporting threshold; tangential-skim instruction for an impurity not completely separated from the principal peak. Official USP-NF record accessed 29 July 2026. Re-checked 31 July 2026: the DOI resolves, HTTP 200, but doi.usp.org serves a preview behind a subscription wall carrying only the chapter introduction, about 1,100 characters of visible text, so the quantification and reporting-threshold passages this page draws on cannot be re-extracted by an automated request. USP-NF is a subscription standard: that is a paywall, not a missing document. doi:10.31003/USPNF_M99380_10_01.

2. 2

   Kuipers BJ, Gruppen H. Prediction of molar extinction coefficients of proteins and peptides using UV absorption of the constituent amino acids at 214 nm to enable quantitative reverse phase high-performance liquid chromatography-mass spectrometry analysis. J Agric Food Chem. 2007;55(14):5445-5451. doi:10.1021/jf070337l. PMID 17539659.

3. 3

   D'Hondt M, Bracke N, Taevernier L, et al. Related impurities in peptide medicines. J Pharm Biomed Anal. 2014;101:2-30. doi:10.1016/j.jpba.2014.06.012. PMID 25044089.

4. 4

   Vergote V, Burvenich C, Van de Wiele C, De Spiegeleer B. Quality specifications for peptide drugs: a regulatory-pharmaceutical approach. J Pept Sci. 2009;15(11):697-710. doi:10.1002/psc.1167. PMID 19750489.

5. 5

   U.S. Food and Drug Administration, Center for Drug Evaluation and Research. Sameness Evaluations in an ANDA, Active Ingredients. Draft guidance for industry, November 2022. Synthetic peptides section, lines 128 to 142 on the orthogonal analytical methods recommendation, and footnote 28 on peptide-related impurities, immunogenicity and the impurity-difference test. Accessed 29 July 2026. The line numbers are the ones PRINTED IN THE MARGIN of the guidance, which is how FDA publishes a draft for comment, not a count of extracted text lines. Verified 31 July 2026 by downloading the PDF, HTTP 200, and converting it with layout preserved: printed line 125 is the heading Synthetic Peptides, printed line 137 carries the sentence recommending orthogonal analytical methods, and printed lines 139 to 142 are the four bulleted properties. An extraction that renumbers the lines will disagree, and one has twice. fda.gov, Sameness Evaluations in an ANDA, Active Ingredients.

6. 6

   United States Pharmacopeia, USP-NF general chapters named as the methods that answer the questions area percent cannot, each cited by its official USP-NF record and DOI. All accessed 29 July 2026. 71 Sterility Tests, doi:10.31003/USPNF_M98810_01_01; 85 Bacterial Endotoxins Test, doi:10.31003/USPNF_M98830_02_01; 233 Elemental Impurities, Procedures, doi:10.31003/USPNF_M5193_03_01; 921 Water Determination, doi:10.31003/USPNF_M99710_05_01.

7. 7

   International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. Q3C(R8), Impurities: Guideline for Residual Solvents, Step 4 version dated 22 April 2021; Q3D(R2), Guideline for Elemental Impurities, adopted 26 April 2022. Both accessed 29 July 2026.

Related pages

Around this page: five neighboring pages on certificates, vials and the arithmetic that assumes the label is right.

• How to read a COA: the whole document, line by line, and the compendial tests no chemical line on it can answer.
• What a lyophilized peptide is: where the water in a vial comes from, and why it is measured separately.
• What a vial cake means: the visual check that precedes every one of these tests.
• Concentration against amount: the calculation that breaks if the labeled mass is not peptide.
• Why the charts disagree: what uncertainty in the vial does downstream to every number.

Next in this run: the rest of the certificate series arrives in The Decadewise briefing as each page clears its gate.