System suitability testing: what the injections before your sample are meant to prove

Every chromatographic purity figure rests on a chain of assumptions, and the weakest link is rarely the chemistry. A method that separated cleanly during validation is being run months or years later, on a different column from the same batch, by a different analyst, on an instrument whose pump seals have worn and whose detector lamp has aged. Nothing about the validation report guarantees that the separation still works today. System suitability testing is the mechanism by which that gap is closed: a short set of injections, run immediately before and often interleaved with the sample sequence, whose sole purpose is to demonstrate that the instrument, column, mobile phase, and operator constitute a system currently capable of producing the result the method was validated to produce.

The distinction matters because system suitability is frequently confused with method validation. Validation is a one-time exercise establishing that a method is fit for purpose. System suitability is a recurring, per-run exercise establishing that the method is working right now. A validated method run on an unsuitable system produces numbers that look identical to good data and are not. Nothing on the face of a certificate of analysis distinguishes the two.

What the parameters actually measure

The conventional system suitability set for a reversed-phase peptide purity method comprises four measurements, each guarding against a distinct failure mode.

Resolution between the main peak and its nearest neighbour tests whether the separation is still separating. Resolution degrades as a column ages, as void volume develops at the head of the bed, and as mobile phase composition drifts through evaporation of the organic component. A resolution requirement is typically written as not less than 2.0 between a specified critical pair, though the harmonized text of USP General Chapter <621> frames the requirement as applying where a significant impurity elutes near the analyte or where a critical pair must be distinguished, rather than as a blanket rule for every method. For peptides this is nearly always the operative case, because the impurities that matter most — single-residue deletions, D-amino acid epimers, oxidation products differing by sixteen mass units — are precisely the ones that elute closest to the parent.

Tailing factor, conventionally required to be not more than 2.0, tests peak symmetry. Asymmetry in peptide chromatography usually indicates secondary interactions with residual silanols, ion-pairing agent depletion, or column overload. Its practical significance for purity work is that integration of an asymmetric peak is ambiguous: a tailing main peak buries co-eluting late-shoulder impurities in its own tail, and the software’s choice of where to drop the baseline becomes a meaningful contributor to the reported percentage. Two analysts integrating the same tailing chromatogram can differ by tenths of a percent, which is the entire margin between a 99.0% specification and a 98.7% result.

Repeatability of replicate injections, expressed as relative standard deviation of peak area or peak response, tests the injector and the pump. This is the parameter most often quoted as a fixed 2.0%, and that quotation is an oversimplification. USP <621> contains a table relating the permitted RSD to both the number of replicate injections and the width of the assay acceptance range — where the range is narrow, the permitted RSD tightens substantially, falling well below one percent for a small number of injections against a ±2% range. The logic is straightforward: if the specification window is narrow, injection scatter has to be small enough that a passing lot cannot be pushed outside the window by instrument noise alone. Methods that quote a flat 2.0% RSD are usually operating against a wider range than a tight purity specification implies.

Signal-to-noise at or near the reporting threshold tests the detector. For a related-substances method this is the parameter that determines whether the impurity threshold printed on the certificate is real. A method reporting individual impurities down to 0.05% must demonstrate that a 0.05% peak is distinguishable from baseline on the day of the run — conventionally a ratio of at least ten to one for quantitation and three to one for detection. Detector lamps age, and an aged lamp raises the noise floor without changing anything visible about the main peak. A purity result generated on a system with a degraded lamp is not wrong about the main peak; it is silently blind to the small ones.

The peptide-specific problem: gradients

Almost every synthetic peptide purity method is a gradient method, and gradients complicate system suitability in ways that isocratic assays do not.

USP <621> permits a defined set of adjustments to chromatographic conditions without triggering revalidation — mobile phase pH by ±0.2 units, column temperature by ±10 °C, column length and particle size within a constant L/dp ratio of −25% to +50%, and, for isocratic separations, flow rate by ±50%. The latitude granted to gradient separations is materially narrower, and for good reason. In a gradient method the composition arriving at the column head at any moment depends on the instrument’s dwell volume — the volume between the mixing point and the column inlet. Dwell volume varies by an order of magnitude between low-pressure and high-pressure mixing systems, and it is not a property of the method, it is a property of the instrument. Transfer a validated peptide gradient from a system with a 1.5 mL dwell volume to one with 150 µL and early-eluting peaks shift, resolution in the front of the chromatogram changes, and the critical pair the method was designed around may no longer be critical in the same way.

This is why a system suitability failure on a transferred peptide method is so often a dwell volume problem misdiagnosed as a column problem. The column gets replaced, the failure persists, and the actual fix is an isocratic hold at the start of the gradient or a compensating adjustment to the initial composition. It is also why the method variables that govern a peptide separation — gradient slope, ion-pairing agent, temperature, column chemistry — need to be treated as a coupled set rather than as independently adjustable knobs.

Bracketing, and the question of when the system was suitable

A system suitability injection at the start of a sequence establishes that the system was working at the start of the sequence. It says nothing about hour six. Peptide sequences are frequently long — a lot-release run might include a blank, a diluent, a standard, replicate suitability injections, several sample preparations in duplicate, and a set of stability samples — and column performance can degrade measurably within a single sequence when the sample matrix contains particulates, strongly retained material, or residual synthesis reagents.

Bracketing addresses this by repeating the suitability check at intervals through the run and at its end, so that a failure can be localized to a segment rather than invalidating everything. Where bracketing is not built into the method, a purity number carries an unstated assumption about instrument stability across the run that no one has tested. This is one of several reasons a certificate showing only a final purity percentage conveys less than it appears to; the chromatographic conditions and the suitability record are what make the number interpretable.

A second and less-discussed issue is what the suitability standard actually is. Resolution and tailing can be assessed on the sample itself if it contains a suitable impurity pair, but repeatability and signal-to-noise generally require a qualified reference standard — and the traceability of that standard determines what the resulting numbers mean. A suitability check run against an in-house standard characterized by the same laboratory using the same method is circular in a way that is easy to overlook.

What suitability cannot establish

System suitability is a check on the instrument, not on the chemistry. It cannot detect a co-eluting impurity that was never separated during validation, because a peak that has never been resolved does not appear as a resolution failure — it appears as a clean, symmetrical main peak. A method whose critical pair was chosen wrongly will pass suitability indefinitely while under-reporting impurity content.

This is the structural argument for orthogonality. A reversed-phase method separating by hydrophobicity, passing every suitability criterion, remains blind to any species with the same hydrophobicity as the parent. Separating the same sample by charge, or by electrophoretic mobility, or confirming identity by mass, tests a different property and can reveal what a single mode conceals. Suitability testing guarantees that a method is performing as it was validated to perform; it makes no claim about whether the validation asked the right question.

Read in that light, the injections preceding a sample are not administrative overhead. They are the only evidence that the separation producing a purity figure was, on that day and on that instrument, the separation the method describes. Their absence from a certificate does not mean they were not run — but it does mean the reader is being asked to take the result on trust rather than on record.

Further reading

Research use only. This post is for educational and reference purposes on peptide synthetic and analytical chemistry. It does not constitute medical, veterinary, or dosing guidance.