Almost every research peptide arrives with a purity percentage attached, and that number generally comes from a single reversed-phase HPLC run: quality control for synthetic peptides is commonly based on one generic reversed-phase method [1]. It is calculated one way: the area under the main peak, divided by the summed area of every peak the software integrated, times one hundred. This is a real measurement made by a real instrument. It is also a much narrower statement than it appears to be, because it describes a chromatogram, not the contents of a vial. This guide is about the distance between those two things, and about the specific decisions — wavelength, column, gradient, integration threshold — that move the number without changing the material.

What the separation is doing

In reversed-phase chromatography the stationary phase is non-polar and the mobile phase is polar, so retention tracks hydrophobicity: the more strongly a species partitions into the packing, the later it reaches the detector. The generic quality-control configuration for synthetic peptides is a C18 stationary phase with an acetonitrile gradient and 0.1% trifluoroacetic acid in the mobile phase, and it performs well across a wide variety of impurities [1]. That last clause is doing real work. It performs well across a wide variety — not across all of them.

The impurities in question are not exotic. They span amino acid substitutions, chain cleavages, and D-/L-isomers of the target sequence [2], and a broader survey of synthetic peptide characterization groups the sources into three buckets: the starting materials, the manufacturing process, and the storage conditions afterward [5]. Some of those species differ from the target peptide by a great deal and separate easily. Others differ by the stereochemistry of one residue.

The arithmetic, and what normalization hides

Area percent is a normalized quantity. Every integrated peak contributes to the denominator, which means the reported figures always sum to one hundred regardless of what is actually in the sample. Take a hypothetical chromatogram with a main peak of 9,700,000 area counts and three impurity peaks of 180,000, 90,000 and 30,000. The total is 10,000,000, and the main peak is 9,700,000 / 10,000,000 = 97.0%.

Now suppose the 180,000-count impurity happens to respond half as strongly per unit mass as the main peptide at the chosen wavelength. Its area under-represents its mass by a factor of two. Converting each peak to mass-proportional units by dividing by its relative response: 9,700,000 / 1 = 9,700,000; 180,000 / 0.5 = 360,000; and 90,000 and 30,000 unchanged. The new total is 10,180,000, and the main component is 9,700,000 / 10,180,000 = 95.3%. Nothing about the sample changed. One assumption about detector response changed, and the number moved by nearly two percentage points.

Area percent implicitly assumes every species responds identically. That assumption is a convenience, not a finding. Impurities arising from insertion, truncation, deamidation, isomerization and oxidation may not respond equally to the main analyte under HPLC with UV detection, and applying a relative response factor of 1 by default may therefore overestimate or underestimate an impurity, depending on which way its true response factor runs [13]. In a non-peptide system — quinolone antibiotics, whose impurities differ substantially in UV absorption — quantifying impurities without matched reference substances was described as a standing challenge, and the method the authors developed, reading a diode array detector and a charged aerosol detector together, was reported from its correction-factor results to give a uniform response to impurities with different chromophores [8]. The advanced generic peptide method of Karongo and colleagues splits the flow after its second-dimension separation to a diode array detector in line with a charged aerosol detector, alongside a quadrupole time-of-flight mass spectrometer; the authors describe these as complementary detectors and state no response-factor rationale for the combination [1].

Why 220 nm and 280 nm are different questions

The amide backbone absorbs in the far ultraviolet; aromatic side chains absorb near 280 nm. That difference is not a detail of instrument tuning — it changes which molecules exist as far as the chromatogram is concerned.

At 280 nm the signal is essentially a report on aromatic residues. The molar absorptivity at 280 nm can be predicted directly from a protein sequence, but the method is only applicable to sequences that contain tryptophan or tyrosine [6]. A peptide with neither is close to invisible at that wavelength. When peak area at 280 nm is used to quantify peptides against a sequence-estimated extinction coefficient, tryptophan-containing peptides agree with expected values more closely than those carrying only tyrosine, and the authors report that analogous, previously used methods could not be applied directly at lower wavelengths [7].

Run the same vial at a far-UV wavelength and the population of visible species changes. Every peptide bond contributes, so truncated and deletion sequences that lack aromatics — close to invisible at 280 nm — now appear as peaks, enter the denominator, and lower the main peak's share. Run it at 280 nm and those same species contribute little or nothing to the calculation, which raises the reported purity. Neither number is a lie. They answer different questions, and only one of them is usually written on the document.

Within the far-UV region the exact wavelength is itself a convention. The generic peptide quality-control method described above recorded its first-dimension chromatogram at 215 nm [1]; certificates of analysis commonly quote 214 or 220 nm instead. We could not find a primary study establishing any one of these as the correct value for peptides generally, and treat the choice as a working trade-off between backbone signal and mobile-phase absorbance rather than a measured optimum. The practical consequence stands regardless: a purity figure quoted without its detection wavelength cannot be compared with another purity figure, and two certificates recorded at 220 nm and 280 nm are not measuring the same quantity on the same sample.

Co-elution: the impurity that has no peak

Integration can only count peaks that exist. If an impurity elutes under the main peak, it is not a small peak — it is part of the main peak's area, and it inflates the purity figure rather than reducing it.

This is the failure mode the peptide chromatography literature takes most seriously. Structurally close impurities with similar sequences that were not resolved in preparative reversed-phase purification may easily co-elute in the corresponding quality-control run as well [1]. Mass spectrometry does not rescue every case: compounds with the same mass-to-charge ratio are not readily differentiated by MS and must therefore be separated chromatographically [3]. Positional isomers and epimers of the target sequence are exactly such compounds.

A concrete example. In commercial aqueous salmon calcitonin ampoules stored for up to two years, oxidation derivatives of two cysteine residues, with cleavage of the disulfide bridge, were found in every sample. Under the authors' reversed-phase separation those oxidation derivatives co-eluted with each other and showed HPLC behavior identical to the disulfide-reduced derivative — and a previous study had tentatively identified that reduced derivative as a major degradation product without mass spectrometric confirmation [10]. One retention time, several molecules, and a published misassignment that only orthogonal evidence caught.

Detector-side peak-purity checks help less than one might hope. Multivariate analysis of diode-array data — asking whether the UV spectrum across a peak is homogeneous — was shown to suffer inherent sensitivity limitations for minor components in the setting the authors assessed, where the co-eluting species have very similar UV spectra [9].. That work assessed drug compounds rather than peptides, so it does not establish the limitation for peptide analysis directly. For peptides, the serious approaches are two-dimensional: use the first dimension to detect impurities across a broad range of properties and the second to focus specifically on species that might co-elute with the target in the first [2].

Orthogonality, and what a second method finds

Because reversed-phase impurity analysis of synthetic cyclic peptides often gives inaccurate results owing to the abundance of chemically similar impurities, Yoshida and colleagues ran the same materials by hydrophilic-interaction chromatography, a technique orthogonal to reversed-phase. Across four nine-residue cyclic peptides, three polymer-based HILIC columns and seven mobile phases, the two techniques proved mutually orthogonal: most impurities with short reversed-phase retention had long HILIC retention, and vice versa [4]. For the peptides in that study, an impurity buried under the main peak in one system would have to be an unusual case to sit under it in the other as well. How far that carries to peptides outside this cyclic nine-residue class was not tested.

Column and mobile phase choice within reversed-phase matters too. A ranking of more than 300 UV and MS chromatograms from 30 column and mobile phase combinations suggested that, when developing a new method, screening four columns and four volatile mobile phases with differing characteristics should be adequate to cover the selectivity space; and that as a second-dimension separation, a C8 or C18 phase with no ionic functionality and an acetic acid / ammonium acetate mobile phase buffered at pH 5 gave good selectivity at 25 °C for peptide isomers below about 10 kDa. Notably, for the peptides studied, trifluoroacetic acid reduced the selectivity differences between columns of diverse properties — probably, the authors suggest, due to its potential to form ion pairs with peptides — even though it often improves peak shape for very large peptides [3]. The ubiquitous 0.1% TFA generic method, in other words, may be flattening exactly the selectivity differences that would expose a co-eluting impurity. It is a trade-off rather than a verdict: in the same dataset, which also contained smaller peptides, trifluoroacetic acid received the highest ranking for 40% of the column and mobile phase combinations, on better selectivity and/or peak shape [3].

The chromatogram can manufacture its own impurities

A purity run is not a passive observation. Baumans and colleagues investigated oxidation of methionine-containing peptides occurring during the reversed-phase separation itself. On two of three stationary phases they observed an additional broad peak corresponding to an on-column oxidized species. On one phase the on-column oxidation level rose with the age of the analytical column and with gradient length, reaching 56% when a one-year-old column set was used with a 180-minute method; the same measurement gave 0% on a second phase and 18% on a third. The proportion decreased as the sample load applied to the column increased, which the authors say suggests the presence of a discrete number of oxidation sites within the stationary phase of the analytical column [11].

Read that as a purity result and the sample looks badly degraded. It was the column. The reverse error is equally available: a peak that appears in one laboratory's chromatogram and not another's may be a genuine impurity, an artifact of column age, or an artifact of gradient length, and area percent alone cannot distinguish them.

What is in the vial that the chromatogram never reports

Purity is a proportion of what eluted and absorbed. Several things in a lyophilized vial are neither.

The counterion is the clearest case. Solid-phase synthesis relies on trifluoroacetic acid as both cleavage agent and ion-pairing reagent, so peptides are routinely obtained as TFA salts; quantifying the residual counterion requires separate methods — FT-IR, fluorine-19 NMR, or HPLC with an evaporative light-scattering detector. When Erckes and colleagues exchanged TFA for chloride by freeze-drying from hydrochloric acid, they found 10 mM HCl optimal and observed no impact on peptide purity at any HCl concentration tested [12]. That last point is the useful one for reading a certificate: the salt form of the material changed substantially and the chromatographic purity did not move. Purity and peptide content are independent measurements, and only one of them tells you how much peptide a vial of a given nominal mass actually holds.

Bound water is the second case, and residual inorganic salts a third. Neither is reported by a UV chromatogram. Beyond those, any species that fails to elute under the gradient used — very strongly retained material — or that elutes unretained in the void with the solvent front will not be integrated as an impurity in a normal purity run. How much material this represents in typical commercial lots is not something we can quantify from the literature reviewed here; mass balance is not routinely published on certificates.

Reading a purity figure without over-reading it

A purity number is interpretable in proportion to how much of its method travels with it. Useful questions, roughly in order of how much they change the answer:

  • At what wavelength? A far-UV number and a 280 nm number on the same vial are different quantities, and for a peptide lacking tryptophan and tyrosine the sequence-based 280 nm method does not apply at all [6].
  • On what column and gradient? Selectivity is a property of the column and mobile phase combination, and the choice materially changes which isomers resolve [3].
  • Was peak identity confirmed? Chromatography answers how much; identity is a separate question requiring mass spectrometry, and even MS cannot separate same-mass isomers without chromatography doing it first [3].
  • Was any orthogonal method run? A single reversed-phase result is one projection of the sample. HILIC or a second-dimension separation is what tests whether the main peak is one compound [2, 4].
  • Is peptide content reported separately? If not, the vial's mass and its peptide mass are different numbers by an unstated amount [12].
  • When was it run? Storage conditions are one of the recognized sources of peptide modification [5], so a purity figure is a statement about the lot on the date of analysis, not a property the material retains indefinitely. Handling after the run is yours; see our reconstitution and handling guide.

What is not settled

Several things in this area are convention rather than measured fact, and it is worth being explicit about which. The specific far-UV wavelength used for peptide purity is one. Integration thresholds — the area below which a peak is not counted at all — are another; they directly determine the denominator and are essentially never published alongside the percentage. The assumption of equal response factors is a third. Alternatives exist and are documented — the advanced generic peptide method reads a charged aerosol detector in line with its diode array detector [1], a review of that detector class describes its response as not depending on the spectral properties of the analyte [14], and in a non-peptide system a diode array and a charged aerosol detector read together gave a uniform response to impurities with different chromophores [8] — but we do not see them reflected on the certificates that accompany research material.

What the literature does establish clearly is narrower and more useful: reversed-phase area percent is prone to missing structurally similar impurities [1, 4], the number moves with detection wavelength for reasons intrinsic to peptide chemistry [6, 7], co-elution has produced at least one published misassignment that only orthogonal evidence caught [10], and the separation itself can generate species that were not in the vial [11]. A high purity figure is evidence. It is evidence about a chromatogram, produced by one method, on one day.