Identity and purity are two different questions

The claim that a percentage cannot support

Consider a certificate reporting “purity 99.2% by HPLC” for a named peptide. Taken literally, it says that 99.2% of the detected chromatographic signal appeared in one peak. It does not say which molecule produced that peak.

This is not a technicality. A sample of an entirely different peptide, cleanly synthesised and well purified, would produce exactly the same figure. So would the correct sequence with the wrong C-terminal chemistry, the wrong salt form, or a D-amino acid substituted for an L. Purity measures homogeneity — how much of what is present is one thing. Identity measures which thing.

What each test can establish

Both techniques are excellent at what they do. The failure is almost always in the inference drawn from them, not in the instruments.

Why the order matters

Read identity first. A purity figure for an unidentified substance is an interesting number about an unknown material; once identity is established, the same figure becomes informative. Reversing the order makes the purity number feel like confirmation of something it never addressed — which is exactly the effect a headline percentage on a marketing page relies on.

In practice this means looking for, in order: the claimed sequence and expected mass; an observed mass to compare it against; then the purity result together with the method that produced it. If the first two are absent, the third is uninterpretable no matter how high it is.

Where both tests stop

Even a report with both results has boundaries worth naming:

• Equal masses are not equal molecules. Leucine and isoleucine are isobaric. D- and L-substitutions do not change mass at all. Disulfide-scrambled isomers weigh the same as the correctly folded form. Mass confirmation is consistency, not proof — resolving these requires MS/MS, chiral analysis or NMR.
• Signal is not substance. A UV-based purity figure omits everything that does not absorb at the chosen wavelength: counter-ions, inorganic salts, residual water, many solvents. Those are separate tests, and their absence from a certificate means unmeasured rather than zero.
• The result belongs to a sample. Extending it to a batch depends on sampling and on documentation, which is a traceability question rather than an analytical one.

Reading practice

Three certificate fragments, and what each supports:

1. “Purity: 99.1% (HPLC). Identity: conforms.” — a purity result with an identity assertion that has no numbers behind it. Supports homogeneity; supports nothing about identity.
2. “MS: [M+H]⁺ observed 1234.6, expected 1234.5. HPLC: 96.4% at 220 nm, gradient 5–65% B over 20 min.” — both questions answered, with the method stated. The lower purity figure is the more informative document.
3. “≥99% purity, third-party tested.” — a specification and an unattributed claim. Neither a result nor an identity.

The second example is the one to want. A slightly lower number with a method attached carries more information than a higher number without one — which is the through-line of everything else on this site.