What mass spectrometry confirms that HPLC cannot

Reversed-phase HPLC answers one question: of everything the detector saw in this run, what fraction was a single species? It is a statement about homogeneity, and it is silent on what that dominant species actually is. Mass spectrometry answers a different question — what does the species weigh — and mass is the closest routine proxy analytical chemistry has for molecular identity. That is why the two methods are reported together, and why a certificate carrying only one of them is answering half a question. But the pairing has a limit that is easy to overstate past. A measured mass matching a calculated mass narrows the field of candidate structures enormously; it does not reduce that field to one. Isomers weigh the same. Some near-isobaric substitutions differ by less than the resolving power of a routine instrument. And tandem experiments that do recover sequence information return a statistical match, not a proof. Knowing which parts of an identity claim rest on measurement and which rest on inference is most of the skill in reading one.

What does a mass spectrometer actually measure?

Not mass. A mass spectrometer measures mass-to-charge ratio, m/z, for ions in the gas phase, and mass is recovered from that by working out the charge state.

Getting an intact peptide into the gas phase without destroying it is the problem that had to be solved first. Fenn and colleagues demonstrated electrospray ionisation for large biomolecules, in which a solution is dispersed into charged droplets that evaporate to leave multiply charged ions — a molecule of 10,000 daltons carrying ten charges appears at an m/z near 1,000, well inside the range of ordinary analysers (Science, 1989;246(4926):64–71; PMID 2675315, DOI 10.1126/science.2675315). That multiple-charging behaviour is itself informative: a series of peaks from the same molecule at successive charge states is a self-consistency check, because all of them must resolve to one neutral mass.

Aebersold and Mann's survey of mass-spectrometry-based proteomics sets out the general architecture that follows — ionisation, mass analysis, and optional fragmentation of a selected precursor — and the fact that sequence information comes only from that last step (Nature, 2003;422(6928):198–207; PMID 12634793, DOI 10.1038/nature01511). An intact-mass measurement alone stops short of it.

Why is a matching mass not a confirmed identity?

Because mass is a sum, and sums are degenerate. Several distinct structures share one.

Isomers are exactly equal. Leucine and isoleucine have the same molecular formula and therefore the same mass at any resolving power whatsoever. No mass measurement can separate them. They are distinguishable only by fragmentation methods that cleave the side chain, and those are specialist techniques rather than routine ones: Edwards and colleagues applied charge transfer dissociation to model and wild-type peptides, obtaining full sequence coverage and reliable side-chain diagnostic ions for leucine/isoleucine positions in more than 80% of cases, and noted that earlier tandem approaches to the same problem "all have certain limitations" (Rapid Commun Mass Spectrom, 2022;36(5):e9246; PMID 34927767, DOI 10.1002/rcm.9246).

Stereoisomers are exactly equal. A peptide containing one D-amino acid is an epimer of the all-L sequence, identical in formula and mass. Tao and colleagues opened their work on the problem with the plain statement that the presence of a single D-amino acid in a peptide "is very difficult to detect", and showed that radical-directed dissociation discriminated D-serine, D-alanine and D-aspartic acid across eight peptides better than conventional collision-induced dissociation did for all but one (Anal Chem, 2012;84(15):6814–20; PMID 22812429, DOI 10.1021/ac3013434). Chen and colleagues later evaluated higher-energy collisional dissociation on the same class of problem, using liraglutide and its D-amino-acid-containing variants as the test case, and localised the substituted positions from differences in fragment ion intensity (Int J Mol Sci, 2024;25(3):1379; PMID 38338662, DOI 10.3390/ijms25031379). Both papers exist because the routine methods do not resolve this.

Some substitutions are near-isobaric. Glutamine and lysine residues differ by roughly 0.036 Da. Distinguishing them requires resolving power and mass accuracy that a low-resolution instrument does not have, and a certificate that reports a nominal mass rather than an accurate one has not made the distinction.

What does an intact mass reliably reveal?

Modifications that change the formula. Oxidation adds about 15.995 Da. Deamidation of asparagine or glutamine adds about 0.984 Da — a small shift, but a real one, and detectable on an instrument with adequate mass accuracy.

Deamidation is worth dwelling on, because the conditions that produce it are laboratory-routine rather than exotic. Erckes and colleagues used tandem mass spectrometry with both collision-induced and electron-transfer dissociation to monitor deamidation and isoaspartate formation, and reported that conventional chromatographic methods and standard mass spectrometric analyses "often fail to distinguish structurally similar peptides with nearly identical physicochemical properties and masses". They confirmed isoaspartate formation under mildly basic conditions such as phosphate-buffered saline, and found that exposure to acidic conditions — specifically the trifluoroacetic acid routinely used during HPLC purification — produced substantial direct deamidation by hydrolysis, with C-terminal amides markedly more susceptible (RSC Med Chem, 2026;17(2):1144–1154; PMID 41541711, DOI 10.1039/d5md01025j).

Two things follow. First, the purification step can generate the modification the analysis is looking for. Second, aspartate and isoaspartate are isomers, so the isoaspartate route is mass-silent: the material has changed and the intact mass has not moved.

Can tandem MS close the gap?

Partly, and with a caveat about what a match means.

Fragmenting a selected precursor and reading the resulting ion series recovers sequence information, and the papers above show it resolving isomeric cases that intact mass cannot touch. But an identification made by matching a spectrum against a database is a statistical inference with an error rate, not a deterministic readout. Elias and Gygi's target-decoy strategy — searching a reversed or shuffled sequence set alongside the real one to estimate how often a match of a given score arises by chance — became standard practice precisely because scores needed an error rate attached to be interpretable (Nat Methods, 2007;4(3):207–14; PMID 17327847, DOI 10.1038/nmeth1019). That work addresses large-scale protein identification rather than single-compound confirmation, and the arithmetic does not transfer directly to a one-compound certificate. The underlying point does: a spectral match is evidence with a confidence attached, and the confidence is part of the result.

What do the certificates on this site report?

An HPLC purity figure, a batch number, a test date and a named issuing laboratory. Five are currently public: GHK-Cu batch 2026-03 and Retatrutide batch 2026-03, tested 11 March 2026; MOTS-c batch 2026-03 and NAD+ batch 2026-03, tested 20 March 2026; and BPC-157 batch 2026-03, tested 22 January 2026. All five were analysed by Janoshik Analytical, whose certificates carry a verification key checkable directly with the laboratory.

Those are chromatographic measurements. None of them constitutes the identity determination described above, and none should be read as one. Stating that is more useful than blurring it, and it is the same reason the test methodology page separates what a purity figure measures from what it does not.

Veridian Research supplies these materials strictly for in-vitro laboratory research. They are not drugs and are not approved for human or veterinary use.

The short version

HPLC says one species dominated the chromatogram. Mass spectrometry says what that species weighs, which excludes most wrong answers and cannot exclude isomers — leucine for isoleucine, aspartate for isoaspartate, a D residue for an L one. Tandem fragmentation reaches some of those cases, at the cost of turning the answer into a scored match rather than a measurement. An identity claim is worth exactly as much as the method that produced it, and the method is a field a certificate either carries or does not.