“98% pure” and “confirmed identity” sound like the same claim. They are not. Reading a supplier’s testing documentation well means keeping at least three separate test claims straight: purity, identity, and sterility. Each is measured by a different instrument, each can pass while the others are undocumented, and conflating them is the single most common misreading of a supplier’s testing documentation.

What a chromatogram is actually showing you

Reversed-phase HPLC injects a sample into a column packed with a hydrophobic stationary phase and elutes it with an increasingly organic mobile phase. Different molecules travel through the column at different rates depending on their hydrophobicity, so each distinct species exits — elutes — at a characteristic retention time and appears as a peak on the resulting trace. The instrument’s integrator then measures the area under each peak. The main peak’s area, divided by the total area of every peak on the trace, is reported as the purity percentage.

Reading a trace well means checking three things beyond the headline number: how many minor peaks are present, how close they sit to the main peak (a shoulder peak that nearly co-elutes is easy to under-integrate), and whether the baseline is flat enough that the integration software isn’t inventing peak boundaries out of noise. A trace with one large, cleanly separated peak and a flat baseline is a meaningfully stronger result than the same reported percentage produced by an integrator squeezing value out of a messy trace.

Where area-percent misleads

The purity percentage on a standard CoA is an area-percent, not a weight-percent, and the two are only the same number under an assumption that rarely gets stated: that every co-eluting impurity absorbs UV light at the detection wavelength exactly as efficiently, per unit mass, as the target peptide does. In practice, impurities that absorb weakly at 214 nm can be present at a higher mass fraction than their peak area suggests, and impurities that absorb strongly can look worse than they actually weigh. A 99% area-percent purity claim is real and useful, but it is a statement about detector response, not a direct measurement of composition by mass. Asking a supplier which detection wavelength and column their trace was run on is a fair diligence question — it changes what the number means.

Common ways a purity figure misleads

Beyond the area-percent vs. weight-percent gap, a handful of specific patterns account for most cases where a clean-looking purity number turns out to overstate what the sample actually is:

  • Under-integration of a shoulder peak. A minor impurity that elutes very close to the main peak can get folded into the main peak’s integration boundary by the software (or by an operator adjusting the integration manually), inflating the reported purity above what a stricter integration would show.
  • Non-UV-active impurities. Standard HPLC purity uses UV absorbance detection. Species that don’t absorb at the chosen wavelength — some salts, some small-molecule byproducts — are invisible to the trace entirely, regardless of how much of the sample they make up. A high UV-purity number says nothing about non-UV-active content; that requires a different detection method (evaporative light scattering or charged aerosol detection) to catch.
  • Gradient vs. isocratic method choice. A shallow, well-optimized gradient resolves closely related impurities into separate peaks; a fast, coarse method can compress several distinct species into what looks like one peak. The same sample can report meaningfully different purity depending on how much resolving power the method was built with — which is another reason the method itself, not just the number, is worth asking about.
  • Reporting only the best lot. A single strong trace from one lot, presented as representative of a product line, tells you nothing about the lot actually shipped to you. This is the specification-sheet-as-CoA problem: the fix is always to ask for the document tied to your specific lot number, not a reference trace.

A worked example

Consider a trace reporting a main peak at 99.1% area, with two minor peaks at 0.6% and 0.3%. Read on its own, 99.1% looks like a near-perfect result. Reading it properly means asking a few more questions before treating it that way. First, where do the minor peaks sit relative to the main peak’s retention time — a peak eluting well before or after the main peak (a different, unrelated species) is a different concern than one crowding right next to it (a closely related variant the integration may be underestimating). Second, is the 0.6% peak new to this lot, or does it show up consistently across the supplier’s last several lots — a recurring minor peak at a stable level suggests a known, characterized synthesis byproduct; a new peak that wasn’t on prior lots is worth a direct question to the supplier. Third, does the CoA report a mass result for the main peak at all — a 99.1% purity trace with no accompanying identity confirmation has told you about the chromatogram’s shape and nothing about what’s actually eluting.

What mass spectrometry proves that HPLC can’t

HPLC purity and chemical identity are answering different questions, and a purity trace alone cannot answer the second one. Two peptides that differ by a single deamidation, oxidation, or truncation can co-elute at the same retention time and be indistinguishable on an HPLC trace, while having different masses. Mass spectrometry — LC-MS or MALDI-TOF — measures the actual mass of the eluting species and compares it against the theoretical mass calculated from the ordered sequence. A match within a stated tolerance (commonly ≤10 ppm on high-resolution LC-MS, wider on lower-resolution MALDI) is what confirms identity: that the dominant species in the vial has the mass the sequence predicts.

Mass spec has its own blind spot, though: isobaric species and structural isomers — same mass, different arrangement — are invisible to it. A peptide with the correct sequence but the wrong disulfide bridge pairing, for example, shows an identical mass to the correctly folded version. Neither HPLC retention time nor MS mass catches that on its own; it takes an assay-specific functional check to rule it out, which is outside what any standard CoA reports.

Sterility is a third, separate claim

“Sterile” describes a manufacturing and packaging process — aseptic filling, sterile filtration during synthesis, bioburden or endotoxin testing — and it is evidenced by its own test line item, not implied by a clean purity or identity result. A vial can carry a pristine 99% HPLC trace and a confirmed mass-spec identity while having no sterility testing documented at all, because nobody ran that test. When a listing or a CoA doesn’t report a sterility or endotoxin result, the honest reading is that the claim is untested, not that it’s satisfied by proxy through the purity number.

Reading the three together

A complete testing picture for a research lot answers all three questions independently: what fraction of the detected signal is the main peak (HPLC purity), does the main peak’s mass match the ordered sequence (MS identity), and was the fill process aseptic and tested (sterility documentation). A supplier that publishes only the first and calls it comprehensive is publishing one-third of the picture. When comparing suppliers, ask for all three, ask which instrument and method produced each number, and treat a missing field as missing data — not as a passed test.

For laboratory research use only. Not for human consumption, diagnostic, or therapeutic use.