HPLC vs mass spectrometry: what peptide purity and identity tests show

A certificate for a synthetic peptide usually carries two headline lines: a purity figure from high-performance liquid chromatography (HPLC) and an identity statement from mass spectrometry (MS). They answer two separate questions, and each has blind spots the other covers. This article explains how each instrument produces its number, what can hide inside it, and what to look for in the chromatogram and the spectrum themselves.

Why are purity and identity measured by different tests?

Purity asks how much of the material is the main component. Identity asks what that main component is. No single routine technique answers both well for a peptide.

The ICH guideline on validating analytical procedures, Q2(R2), describes the property an analytical method needs here as specificity or selectivity: the extent to which other substances interfere with the measurement of the analyte. It adds that where one procedure does not discriminate well enough, a combination of two or more procedures is recommended. HPLC and MS are that combination. HPLC separates components by how they behave on a column and counts them by UV absorbance. MS weighs molecules. A component that one method cannot distinguish from the product is often obvious to the other.

How does reversed-phase HPLC separate a peptide from its impurities?

Reversed-phase HPLC (RP-HPLC) pumps the dissolved sample through a column packed with particles carrying a water-repellent surface, usually C18 or C8 alkyl chains. Peptides stick to that surface through their hydrophobic side chains. The mobile phase starts mostly aqueous and gradually takes on more organic solvent, typically acetonitrile, so peptides let go of the surface in order of increasing hydrophobicity. Most peptide methods run below pH 3 with trifluoroacetic acid (TFA) in water and acetonitrile; TFA is volatile and acts as an ion-pairing agent that sharpens the separation of charged peptides, and gradients of roughly 0.5 to 2% acetonitrile per minute are common (Mant et al., 2007).

As each component leaves the column it passes a UV detector set to a low wavelength, usually around 210 to 220 nm, where the peptide bond absorbs. The detector trace, plotted against time, is the chromatogram. Each separated component appears as a peak; its retention time reflects its hydrophobicity, and its area reflects how much UV light it absorbed.

A missing hydrophobic residue, an oxidised methionine or an unremoved protecting group usually shifts a peak far enough to be resolved. Even subtle stereochemical changes can sometimes be separated: in a model α-helical peptide, analogues carrying isoleucine, allo-isoleucine, D-isoleucine or D-allo-isoleucine at one position were resolved by conventional RP-HPLC, because the substitutions changed the peptide’s secondary structure (Huang et al., 2014). That separation depended on column, temperature and sequence context, so a method that resolves one diastereomer cannot be assumed to resolve every other one.

What does “% purity by area” actually mean?

The purity figure is the area of the main peak divided by the summed area of all integrated peaks, times 100. Three things decide how close that figure sits to the true share of the product.

Detection wavelength and response factors. Area % treats every peak as if equal amounts give equal signal. They do not. Measured at 214 nm, a single peptide bond has a molar extinction coefficient of about 923 M⁻¹ cm⁻¹; tryptophan absorbs about 30 times more strongly, phenylalanine, tyrosine and histidine about six times more, and proline inside a chain about three times more, while most other residues add little (Kuipers and Gruppen, 2007). An impurity missing a tryptophan therefore gives a smaller peak per mole than the product and is under-counted; a truncated fragment rich in aromatic residues can be over-counted. Components with no chromophore at the chosen wavelength, such as many salts, are not seen at all.

Co-elution. Two components that leave the column at the same time produce one peak, and the whole area is credited to the product. A single symmetric peak is consistent with purity but does not prove it. The way to test a method for this is to show it separates the impurities expected for that sequence, or to compare it with a second method based on a different separation principle.

Integration choices. The software has to decide where each peak starts and ends, how to draw the baseline under it, how to split a shoulder from the main peak, and which small peaks fall below the reporting threshold. Peaks from the sample solvent or the blank are normally excluded. Each of these choices moves the purity figure, which is why a figure with no chromatogram and no method behind it is hard to judge.

Finally, area % is relative. It describes the peptide-like material that the method detected, not the powder in the vial, which also holds counter-ions and water. That gap is covered in the peptide-content section below.

Which impurities does the purity method need to separate?

Solid-phase peptide synthesis (SPPS) builds the chain one residue at a time on a resin, and each step leaves a small fraction of chains behind. A structured review of peptide-related impurities lists the main families: deletion sequences, caused by inefficient Fmoc deprotection; insertion sequences, from excess amino acid reagent; racemisation of residues during deprotection, giving diastereomeric impurities; peptide-protection adducts from incomplete side-chain deprotection; oxidised side chains; dimers and oligomers; and degradation products such as pyroglutamate, diketopiperazine and succinimide forms (D’Hondt et al., 2014). Truncated sequences, chains whose growth stopped and which are often deliberately capped so they cannot extend further, are also named in the EU guideline on synthetic peptides.

Two families deserve a closer look. Among oxidisable residues, methionine is the most reactive, converting to the sulfoxide or sulfone; but forced-oxidation studies have also detected oxidation products on tryptophan in somatostatin and on tyrosine and proline in cetrorelix, peptides that contain no methionine (Datola et al., 2023). And aspartimide formation, a base-driven side reaction at aspartic acid residues during Fmoc synthesis, remains a persistent problem despite a range of protecting-group strategies developed to suppress it (Kong et al., 2025). The aspartimide ring can reopen to give the normal α-aspartyl peptide, a β-aspartyl isomer or a racemised form, and those isomers have exactly the same mass as the product.

The table below gives the mass shift each family produces, which is how MS recognises them, and whether RP-HPLC typically separates them.

Impurity type	Mass difference from the target	Usually separated by RP-HPLC?
Deletion sequence	Minus one residue (e.g. −57.02 Da for Gly, −113.08 Da for Leu or Ile)	Often, sometimes eluting close to the main peak
Insertion sequence	Plus one residue	Often
Truncated, acetyl-capped sequence	Much lower mass; the cap adds +42.01 Da	Usually
Remaining tert-butyl protecting group	+56.06 Da	Usually; more hydrophobic, so later
Oxidised methionine (sulfoxide)	+15.99 Da	Usually; more polar, so earlier
Aspartimide	−18.01 Da	Often
β-aspartyl isomer, D-residue	0 Da	Sometimes; may co-elute
Covalent dimer	About twice the mass	Usually

A same-mass isomer is invisible to MS and can be invisible to HPLC; a co-eluting impurity is invisible to HPLC but usually obvious in the mass spectrum.

How does mass spectrometry confirm identity?

A mass spectrometer measures mass-to-charge ratio (m/z), so the peptide first has to be turned into gas-phase ions. Two soft ionisation methods dominate peptide work.

Electrospray ionisation (ESI) sprays the solution through a charged needle and produces intact ions from large, fragile molecules. Its spectra show a coherent series of multiply charged ions, each differing from its neighbour by one charge (Fenn et al., 1989). A peptide of neutral monoisotopic mass M appears at m/z = (M + z × 1.00728) / z for z added protons, so the neutral mass is recovered as M = z × (m/z − 1.00728).

Matrix-assisted laser desorption/ionisation (MALDI) mixes the sample with a light-absorbing matrix and desorbs it with a laser pulse; it was shown in 1988 to ionise intact proteins above 10,000 Da (Karas and Hillenkamp, 1988). Peptide MALDI spectra are usually dominated by singly protonated ions, [M+H]⁺, which makes them simpler to read but gives fewer independent checks on the mass.

Three features then confirm identity:

• Observed versus theoretical mass. The calculated mass of the sequence is compared with the measured one. The tolerance depends on the instrument: a low-resolution instrument may agree to within a few tenths of a dalton, a high-resolution one to a few parts per million.
• Charge-state consistency. In ESI, every ion in the charge series should give the same neutral mass.
• Isotopic pattern. Natural ¹³C, ¹⁵N and other heavy isotopes spread each ion into a cluster of peaks spaced 1/z apart, so a 2+ ion shows peaks half a mass unit apart. The spacing confirms the charge state, and the shape of the cluster should match the elemental formula. For larger molecules the lightest (monoisotopic) peak becomes too small to pick out, and its position is inferred from the whole isotope distribution (Senko et al., 1995).

Tandem MS (MS/MS) goes further: it fragments a selected ion and reads sequence information from the fragments. In a study of the model peptide angiotensin I, liquid-chromatography high-resolution MS/MS detected and structurally characterised degradation fragments such as angiotensin II and III in a single run, identifying them by comparing measured with calculated masses (Stoppacher et al., 2013).

Why can’t mass spectrometry measure purity, or HPLC prove identity?

MS cannot measure purity reliably because its signal is not proportional to amount. Molecules ionise with different efficiency, and components that enter the source together compete for charge, so a minor impurity that ionises well can look larger than a major one that ionises poorly. Same-mass impurities, such as D-residue and β-aspartyl forms, do not show up as separate masses at all. A spectrum that shows only the expected mass is good evidence of identity and weak evidence of purity.

HPLC cannot prove identity because retention time is not unique. Different sequences, including a sequence with two residues swapped, can elute at similar times, and a retention time only means something when it is compared with a reference standard of the same peptide run under the same conditions.

What does LC-MS add?

Liquid chromatography-mass spectrometry (LC-MS) sends the column effluent into the mass spectrometer, usually with the UV detector still in line. Every peak in the chromatogram then has a mass spectrum attached. The main peak can be shown to carry the expected mass, each impurity peak can be assigned a likely identity from its mass shift (the table above), and a peak that looks clean in UV but contains two masses reveals a co-elution. A review of synthetic peptide characterisation describes RP-HPLC with UV and high-resolution MS detection as widely used in the pharmaceutical industry, alongside other separation modes such as size-exclusion and ion-exchange chromatography (Sharma et al., 2022).

One practical detail: TFA, the usual ion-pairing agent for UV purity methods, suppresses the electrospray signal, so LC-MS methods often use formic acid instead. A peptide’s retention, and so the look of its chromatogram, can differ between a certificate’s UV purity run and its LC-MS identity run.

Where does peptide content fit in?

Purity by area and MS identity together still do not say how much peptide is in a weighed amount of powder. That is net peptide content, and it is measured differently. The common approach is amino acid analysis: the peptide is fully hydrolysed, typically in 6 M hydrochloric acid, and the released amino acids are quantified, for example by LC-MS against isotope-labelled internal standards, with the peptide amount calculated from the known sequence (Qasrawi et al., 2023).

Amino acid analysis counts amino acids, not intact peptide, so it also counts peptide-related impurities. In the angiotensin I study, structurally related impurities amounted to 10.4 mg per gram of a candidate reference material, and failing to correct for them would have introduced a 1% error into concentrations determined by amino acid analysis (Stoppacher et al., 2013). The same study describes the mass-balance approach used for reference materials: estimate the mass fraction of all impurities present with a range of techniques, and assign the difference from the theoretical total to the main peptide. A certificate that reports HPLC purity, MS identity and peptide content side by side is giving three separate pieces of that balance.

How do you read the chromatogram on a certificate?

• Axes. Time (minutes) on the horizontal axis, UV absorbance (often mAU) on the vertical. Check the stated wavelength, ideally in the 210 to 220 nm range.
• Peak table. Each integrated peak should be listed with retention time, area and area %. The main peak’s area % is the purity figure; the other rows are the impurity profile.
• Baseline and run length. A flat baseline and a run that continues after the main peak (late-eluting, more hydrophobic impurities such as protected species) are signs of a complete method. A trace cropped tightly around the main peak hides what came after.
• Peak shape. Shoulders, a tail or a broad base on the main peak can mean a partially co-eluting impurity.
• Method details. Column, mobile phases and gradient should be stated. Without them the figure cannot be reproduced.

How do you read the mass spectrum on a certificate?

Take angiotensin I (DRVYIHPFHL), the model peptide in the study above, as a worked example. Its calculated monoisotopic mass is 1295.68 Da (average mass 1296.49 Da).

• Find the charge states. An ESI spectrum might show peaks at m/z 1296.69 ([M+H]⁺), 648.85 ([M+2H]²⁺) and 432.90 ([M+3H]³⁺). Each gives the same neutral mass: for example, 2 × (648.85 − 1.007) = 1295.68 Da.
• Check the isotope spacing. Around m/z 648.85 the cluster should show peaks about 0.5 apart, confirming 2+. For this formula (C₆₂H₈₉N₁₇O₁₄) the second isotope peak is roughly three-quarters the height of the first.
• Know which mass is quoted. Monoisotopic and average masses differ by about 0.8 Da here and by more for larger peptides. A certificate should say which it reports.
• Explain the other peaks. Sodium and potassium adducts sit 21.98 and 37.96 Da above the protonated ion (divided by the charge). Peaks at +15.99, −18.01 or +56.06 Da, or a residue mass below the target, point to the impurity families in the table. An unexplained major peak deserves a question to the supplier.

Reading a few real certificates against these two checklists, for example the batch certificates published in the EuroPeps COA library, is the quickest way to see what complete and incomplete reports look like.

Which standards describe these tests?

ICH Q6B, often quoted in discussions of peptide specifications, sets out identity, purity and quantity as separate attributes, but its scope covers proteins and polypeptides from recombinant or cell-culture expression systems and explicitly excludes synthetic peptides. In the EU, synthetic peptides now have their own guideline: the EMA guideline on the development and manufacture of synthetic peptides, adopted as EMA/CHMP/CVMP/QWP/367182/2025 and applying from 1 June 2026, covers manufacturing, characterisation, specifications and analytical control, including deletion, truncated and insertion sequences and stereoisomers. ICH Q2(R2) supplies the validation vocabulary, including specificity, that tells you whether a purity or identity method was shown to do its job. These documents govern medicinal products, not research-grade materials, but a careful certificate follows the same logic.