HPLC for peptides — what the buyer needs to look for

High-performance liquid chromatography is the analytical method that produces the single most-cited number on a peptide Certificate of Analysis: the purity figure. Understanding what HPLC actually measures, what a credible method specifies, and where the technique can mislead is the difference between trusting a 99% purity claim and being able to evaluate it.

This article walks through reversed-phase HPLC as applied to peptides — the workhorse mode that dominates analytical lab work in this space — and gives a peptide buyer the vocabulary to read a chromatogram, evaluate a method, and recognise the patterns that mark an unrigorous report.

What HPLC actually is

High-performance liquid chromatography (HPLC) separates a mixture of compounds based on differences in how strongly each compound interacts with a stationary phase packed inside a column, versus a mobile phase pumped through it. A small volume of sample (typically 1–20 microlitres) is injected into a stream of flowing solvent. As the solvent carries the sample through the column, compounds that interact more strongly with the stationary phase travel more slowly; those that interact less travel faster. They emerge from the column at different times and are detected as they pass through a detector at the column's outlet.

The output is a chromatogram: a plot of detector signal versus time, with each distinct compound appearing as a peak. The position of a peak along the time axis (its retention time) identifies it; the area under the peak quantifies it. For a peptide sample, the chromatogram shows the target peptide as a main peak, with smaller peaks for synthesis impurities, degradation products, and other artefacts.

This is the foundation. Everything else — purity numbers, identity confirmations, method validation — builds on the same separation principle.

Reversed-phase HPLC for peptides

Peptides are almost always analysed by reversed-phase HPLC. In this mode, the stationary phase is non-polar (typically an octadecyl-bonded silica known as C18) and the mobile phase is polar (water with a small amount of organic modifier, usually acetonitrile). Peptides interact with the C18 surface through hydrophobic contact between non-polar amino acid side chains and the alkyl chains bonded to the silica. More hydrophobic peptides retain more strongly; more hydrophilic ones elute earlier.

To achieve separation, the mobile phase composition is changed during the run — a gradient. A typical peptide gradient starts at high water content (e.g. 95% water / 5% acetonitrile), then progressively increases the acetonitrile percentage over 15–40 minutes. As acetonitrile rises, peptides desorb from the C18 surface in order of hydrophobicity. The result is a chromatogram where peptide peaks are well-spread across the run.

Two further wrinkles matter:

Ion-pairing reagent. Pure water/acetonitrile gradients separate peptides poorly because peptide charges (positive at low pH) cause non-specific interactions with residual silanol groups on the silica. A small amount of acid is added — most commonly 0.1% trifluoroacetic acid (TFA), sometimes 0.1% formic acid — to keep peptides protonated and to ion-pair with the C18 surface. The choice of acid changes peak shape and retention. TFA gives sharp peaks but suppresses electrospray ionisation; formic acid is gentler on mass-spec coupling but produces broader peaks.
Column dimensions and particle size. Modern peptide HPLC uses sub-2-micron particle columns at higher pressures (UHPLC) for narrower peaks and faster runs. A 100 mm long, 2.1 mm internal diameter column packed with 1.7 micron C18 particles is a contemporary default. Older 5-micron columns at 4.6 mm diameter still appear but give lower resolution.

A peptide HPLC method that does not specify these choices is incomplete. The principles formalised in USP General Chapter <621> on chromatography are the foundation for evaluating any peptide separation method against expected practice.

Reading a chromatogram

Three numbers tell you most of what a chromatogram reveals.

Retention time. Where the main peak appears along the time axis. For a given method, the retention time of a particular peptide is reproducible to within seconds across runs. A chromatogram with a retention time noted but unstated method conditions is half a document; the retention time only means something against the method.

Peak area as a percentage of total area. This is the source of the "purity" figure. The integration software sums the area under all peaks in the run and reports each peak's area as a percentage of that total. A peak listed as 99.0% means it accounts for 99% of all the UV-absorbing material that passed through the detector during the run. Crucially, this is a relative measure — it does not say how much absolute material is in the vial, only how uniform what was analysed turned out to be.

Peak shape. A symmetric Gaussian-like peak is ideal. Peaks that tail (asymmetric with a long trailing edge) or front (asymmetric with a long leading edge) usually indicate problems: silanol interactions, column overload, sample preparation issues, or column degradation. The asymmetry is quantified as a tailing factor (or asymmetry factor); USP requires tailing factors below 2.0 for system suitability in most chromatographic monographs.

Two related concepts complete the picture:

Resolution between peaks. A peptide impurity that elutes very close to the main peak — say within 0.2 minutes — may not be cleanly separated. Their peak areas blend together. Resolution (R) quantifies the separation; values above 1.5 are considered baseline-resolved.
Coelution. When an impurity is so chromatographically similar to the target peptide that it elutes inside the main peak. Coelution is the most common cause of overstated purity — the impurity is included in the "main peak" area and inflates the apparent purity. Method optimisation is partly the work of finding conditions where likely impurities are pushed away from the main peak.

A chromatogram with three or four well-separated peaks, a symmetric main peak, retention times noted, and the percent areas listed is a readable document. One with a single fat peak and no surrounding detail is, at best, ambiguous.

Detection: UV at 215 and 280 nm

The detector on a peptide HPLC system is almost always a UV-visible diode array. Two wavelengths are routinely monitored:

215 nm captures the absorbance of the peptide bond itself — the n→π* transition of the amide chromophore. Every peptide bond absorbs here, so this wavelength is universal: any peptide gives a strong signal at 215 nm regardless of side-chain composition.
280 nm captures the absorbance of aromatic side chains: tryptophan strongly, tyrosine moderately, phenylalanine weakly. A peptide without aromatic residues gives almost no signal at 280 nm; a tryptophan-rich peptide gives a strong one.

A complete HPLC report records both wavelengths. The 215 nm trace is the primary quantitative measurement (universal sensitivity); the 280 nm trace cross-checks: if a peak at 215 nm has no corresponding signal at 280 nm, it cannot contain Trp, Tyr, or Phe — which constrains what it might be.

Two alternatives to UV deserve mention. Evaporative light scattering detection (ELSD) detects any non-volatile compound, including those that do not absorb UV. ELSD trades sensitivity for universality. Charged aerosol detection (CAD) is conceptually similar but with better linearity. Both are useful for detecting residual reagents, counterions, and non-chromogenic impurities that UV misses entirely.

What a credible HPLC method specifies

A method described as "HPLC" is not a method. A credible peptide HPLC method specifies, at minimum:

Column. Manufacturer, particle chemistry (C18, C8, etc.), particle size (in microns), pore size (typically 100 Å for peptides, 300 Å for larger), column length and internal diameter.
Mobile phase. Composition of solvent A (typically water with 0.1% TFA) and solvent B (typically acetonitrile with 0.1% TFA), expressed as v/v percentages.
Gradient profile. Time-versus-%B table, e.g. "0 min 5%, 30 min 60%, 31 min 95%, 35 min 95%, 36 min 5%, 45 min 5%".
Flow rate. In millilitres per minute, appropriate to the column dimensions (a 2.1 mm column typically runs at 0.3–0.5 mL/min; a 4.6 mm column at 1.0 mL/min).
Injection volume. Typically 1–20 microlitres.
Column temperature. Often 30–60 °C, controlled to within 1 °C; matters for retention reproducibility.
Detection. Wavelengths and bandwidth, e.g. "DAD, 215 nm and 280 nm, bandwidth 4 nm".
System suitability. A short test run (typically the reference standard injected three times) demonstrates that the system is working: tailing factor below 2.0, RSD of retention time below 1%, RSD of peak area below 2%. These criteria should be stated and the actual results recorded.

A peptide COA whose HPLC section reads "Method: Reversed-phase HPLC, 215 nm, 99.2% area" without any of the above is a one-line assertion, not a method.

Common pitfalls

Several artefacts distort HPLC results in ways the buyer should know about.

Coelution. Already mentioned: an impurity hiding inside the main peak. Detected by changing the gradient to spread peaks differently, or by orthogonal techniques like mass spectrometry that distinguish coeluting compounds by mass. A vendor that reports purity from a single chromatographic method without orthogonal confirmation is offering one perspective on the sample.

TFA-induced peak distortion. TFA suppresses ionisation in mass spectrometry, which is why TFA-free methods (formic acid instead) are used when LC-MS coupling is the priority. TFA can also cause subtle peak distortion at lower pH levels; some peptides exhibit unusual peak shapes that disappear when the acid is changed.

Sample preparation artefacts. Peptides reconstituted in incompatible solvents can precipitate at the column head, broaden peaks, or generate degradation artefacts. The reconstitution solvent should be the same composition as the initial mobile phase or weaker (more water). Reconstituting in pure acetonitrile and injecting onto a water-rich initial mobile phase causes characteristic peak splitting.

Column ageing. Silica-based C18 columns degrade with use. Hydrolysis of bonded alkyl chains and loss of stationary phase produce gradually broadening peaks and shifting retention times. Reputable labs replace columns on a schedule and document column age on the method record.

Method validation in brief

For HPLC results to be trusted across laboratories and across time, the method must be validated. The International Council for Harmonisation's Q2(R2) guideline on validation of analytical procedures is the global reference. It specifies the parameters that a validation package must address: specificity (does the method distinguish the analyte from impurities?), linearity (does signal scale predictably with concentration?), range (over what concentrations does linearity hold?), accuracy and precision (how close to true values are the results, and how reproducible?), detection limit and quantitation limit, and robustness (how stable is the method under deliberate small variations?).

A laboratory that has validated its peptide HPLC method to Q2(R2) is operating at a higher technical standard than one running a generic method. The difference shows up in the method documentation: validated methods come with validation summaries listing each parameter and its measured value.

Red flags

Patterns that mark an unrigorous HPLC report:

Single chromatogram without method details. A trace alone, with no column, gradient, or wavelength noted, cannot be interpreted in any quantitative way.
Purity stated without integration parameters. Modern integration software has many adjustable settings — slope sensitivity, area threshold, peak rejection — and small differences produce different "purity" numbers from the same trace. Methods that do not state these are not fully reported.
Single-wavelength detection only. Reporting at 215 nm without 280 nm leaves a substantial cross-check unexploited.
Run times under five minutes for a complex sample. Fast generic methods often coelute impurities. A peptide with multiple synthesis byproducts requires a gradient long enough to resolve them.
Tailing factors not reported or above 2.0. Indicates either column problems or method that has not been optimised.
"System suitability passed" with no values. Pass/fail without numerical results is the same problem as elsewhere on a COA: a claim, not data.