Finding the Previous Sample in Your Own Chromatogram

Send one batch to two laboratories and you may get back two chromatograms that differ by a single small peak. The instinct is to decide which laboratory made a mistake. Often neither did. The peak may belong to whatever was injected before it, left behind in the needle, the loop or the tubing that every sample shares, and reappearing in the run that followed. Carryover is common, it is thoroughly understood, and on a peptide method it is quite capable of manufacturing an impurity that was never in the vial.

A coincidence that is not one

The signature is a modest peak arriving at precisely the retention time of something from an earlier acquisition. That exactness is what gives it away. There is no reason a genuine impurity in this sample should elute at the identical moment the previous sample’s main component did; there is every reason for a leftover trace of that component to do so.

Two different things go wrong as a result. Where purity is the object, the phantom is tallied as an impurity and the material is reported as less pure than it is. Where the object is identity or trace-level analysis, the same phantom argues for the presence of a compound that is simply absent.

Surfaces that hold on to peptide

The underlying cause is adsorption. Basic residues find residual silanols and bare steel; hydrophobic stretches take to seals and polymer tubing. Consider what that means for a strongly hydrophobic sequence: whatever makes it difficult to release from the column under gradient conditions makes it equally difficult to flush out of the plumbing, since it is the same chemistry acting in both places.

Relative concentration then decides whether the residue is noticeable. Put a concentrated standard directly in front of a dilute sample and the leftover material is substantial compared with the sample it contaminates. That ordering is the textbook way to produce a carryover peak.

Controls that settle the question

Injecting pure solvent immediately after a sample is the basic test: the resulting trace ought to be flat wherever the sample’s peaks appeared. Blanks placed on either side of a run constitute the usual control, and a method that omits them has a hole in it. Separately, nearly all autosamplers rinse the needle between injections, but the rinse must be matched to what is being analyzed – a predominantly aqueous wash will not budge a hydrophobic peptide, and a wash tuned to one analyte may be useless for the one after it. Sequencing helps too: running dilute samples ahead of concentrated ones, and inserting blanks wherever unrelated compounds sit next to each other, removes most of the opportunity for carryover to matter.

These checks belong with the other pre-run controls described in system suitability, the exercise that demonstrates an instrument was in a fit state to give a meaningful answer on the day.

What it means for comparing reports

Back to the two chromatograms. An unexplained minor peak present on one and absent on the other can reflect a real difference between the samples, or it can reflect a difference between the sequences they were run in. Both laboratories may be reporting honestly; the same material was simply analyzed in different company. Add this to the list of explanations for disagreeing reports, which already includes wavelength, column, gradient and integration convention. The full set appears in why certificates of analysis disagree, and the item that shifts the number most frequently is treated in peak integration.

Reading a published trace

Look for an impurity peak with nothing matching it in the impurity table, or one the report declines to assign. Neither observation demonstrates carryover; both justify asking. And whether blanks were included is an entirely reasonable thing to put to any laboratory.

All material is supplied for laboratory research use only. It is not a drug, not a supplement, and not for use in humans or animals.