Carryover and ghost peaks in peptide HPLC: what blank injections reveal about adsorption, memory effects, and purity figures

A chromatogram is usually read as a record of what was in the vial. For peptides, it is also a record of what was in the system before the vial: residues retained on the needle, in the rotor seal, on column frits, or on the stationary phase itself. Peptides are among the more problematic analytes in this respect. Basic residues interact with silanols and metal surfaces, hydrophobic sequences partition into polymer components, and larger molecules desorb slowly from wide-pore packings. The consequence is that a signal can appear in a run where the corresponding analyte was never injected, or can inflate a legitimate peak that happens to elute nearby.

This post reviews the main origins of carryover and ghost peaks in reversed-phase peptide chromatography, how blank injections are designed to distinguish between them, and why the distinction matters for the interpretation of a purity percentage on a certificate of analysis.

Carryover versus ghost peaks: two different problems

The terms are often used interchangeably, but they describe different phenomena with different remedies. Carryover is the appearance of analyte from a previous injection in a subsequent run. Its signature is a peak at the retention time of the preceding sample component, with an area that decreases over consecutive blank injections. Ghost peaks are signals that arise independently of the previous sample: contaminants in the mobile phase, leachables from tubing or vial closures, or compounds that accumulate on the column during equilibration and elute when the organic fraction rises.

The behavior in a gradient run offers a first diagnostic. Contaminants in the aqueous mobile phase (A) concentrate on the head of the column during the equilibration period and elute as the gradient progresses, so their peak area scales with equilibration time rather than injection volume. A blank with no injection at all, run through the same gradient, will show them. Carryover, by contrast, requires an injection event; a true no-injection blank will be clean even when an injected solvent blank is not. Comparing these two blank types is one of the simplest ways to assign an unexpected peak to a source.

Reports in the analytical literature have also described late-eluting peaks from one gradient appearing in the next, when the gradient does not reach a sufficiently high organic composition or is not held long enough to strip strongly retained material. In that case the peak is neither sample carryover in the autosampler sense nor mobile phase contamination; it is incomplete column washing, and it appears at a time determined by the new gradient rather than at the original retention time.

Mechanisms specific to peptides

Several surface interactions have been characterized as contributors to peptide carryover. Autosampler needles and needle seats, typically stainless steel or polymeric materials such as PEEK or Vespel, present both metal and polymer surfaces. Peptides containing multiple basic residues, such as arginine and lysine clusters, have been reported to adsorb strongly to such surfaces, and sequences with phosphate groups or acidic side chains can interact with exposed metal ions. Hydrophobic and amphipathic peptides adsorb to polymeric components through non-polar contact, a behavior closely related to the container-surface losses discussed in earlier posts on low-concentration adsorption.

Rotor seals in injection valves are a frequently identified site. Peptides can accumulate in the microscopic grooves of the seal, and the material may be released slowly into the flow path over multiple subsequent injections. Because the release is slow, the carryover signal may be small in any individual blank but persistent across a sequence, producing a baseline elevation or a low, broad hump at the elution position of the analyte.

On the column, the relevant factors are pore size, surface chemistry, and the strength of the strong solvent in the gradient. Peptides with high hydrophobicity can show extended desorption tails, and aggregation-prone sequences may form assemblies at the column head that dissolve only gradually. Wide-pore stationary phases reduce restricted diffusion for larger peptides but do not remove the need for an adequate high-organic wash.

Concentration is a further variable. Carryover is commonly expressed as a percentage of the preceding peak area, and for a given system that percentage is relatively stable. A high-concentration injection, however, delivers a proportionally larger absolute residue. A system that appears acceptable at a low calibration level may therefore show a visible artifact after a concentrated stock solution has been injected.

Designing blank injections that diagnose rather than reassure

A single solvent blank at the start of a sequence confirms little. More informative designs have been described in system suitability and validation guidance. A blank injected immediately after the highest-concentration standard or sample directly challenges the system. Several consecutive blanks following that injection show the decay profile: a rapid decrease points to needle or loop residue that a stronger wash can remove, while a slow, sustained decrease points to a reservoir such as the rotor seal or column.

The composition of the blank matters. A blank in the sample diluent tests the autosampler path under realistic conditions, while a blank in a mobile-phase-like solvent may fail to reveal carryover that depends on the diluent’s solvent strength. A true no-injection gradient run, as noted above, isolates mobile phase and column contributions. Running these three types in a defined order, and recording them in the sequence, gives a data set from which the origin of an artifact can be assigned rather than guessed.

Wash solvent selection is the main experimental lever. Needle wash solutions are generally chosen to be stronger than the mobile phase and to disrupt the interaction responsible for retention: higher organic content for hydrophobic adsorption, acid or ion-pairing additives for ionic interaction, and sometimes mixed solvents such as organic combinations with small amounts of acid. Multi-solvent wash sequences, alternating a strong organic wash with an aqueous acidic wash, are reported to address both interaction types. The wash conditions are method parameters and are expected to be documented, since a different instrument with a different wash configuration can yield a different carryover profile for the same method.

Acceptance criteria for carryover are typically defined relative to the limit of quantitation or the reporting threshold. In impurity methods, the response of an interfering peak in a blank is commonly compared with the response at the quantitation limit; a peak exceeding a small fraction of that response may be considered to interfere. The exact fraction is method- and laboratory-specific.

Consequences for purity figures

Chromatographic purity is calculated as the main peak area divided by the total integrated area. Any artifact that is integrated enters the denominator, and any artifact that co-elutes with the main peak enters the numerator. Ghost peaks that are integrated as impurities lower the reported purity; carryover from a previous lot of the same compound can raise the apparent area of the main peak or add small peaks at the retention times of related impurities, making two consecutive lots appear more similar than they are.

This second effect is easy to overlook. When a sequence runs several samples of the same peptide back to back, carryover from a high-concentration sample will resemble the main component and may be absorbed into its peak area without visible distortion. The practical implication is that sample order, bracketing blanks, and, where relevant, the injection of a low-concentration sample directly after a high-concentration one are all design choices that bear on whether a comparison between lots is valid.

Integration practice interacts with these artifacts. Peaks present in the blank at the same retention time are often excluded from the impurity calculation, but this exclusion requires that the blank be representative: same gradient, same equilibration, same column history. Peak purity assessment by photodiode array detection or mass spectrometry, discussed in earlier posts, can indicate whether a peak is spectrally homogeneous, but it does not distinguish between sample-derived and system-derived components that happen to share a spectrum.

Interpreting what a report states

A certificate of analysis reports a number, and the method behind it may or may not have addressed carryover explicitly. Method validation summaries that list specificity, linearity, and precision may not specify how blank interference was evaluated. Where a report refers to a validated HPLC method, the relevant documentation would typically include blank chromatograms, the wash program, and the carryover assessment. The presence of such documentation is a more useful indicator of method robustness than the purity value alone.

Carryover and ghost peaks illustrate a broader point about chromatographic data: the trace reflects the entire analytical system, and peptides, because of their surface-active chemistry, make that system visible. Understanding which part of a chromatogram belongs to the sample and which to the instrument is a prerequisite for interpreting any purity figure, and blank injections, designed deliberately, are the primary tool for drawing that line.