Why the Same Method Gives Different Numbers in a Different Lab

Two laboratories running what is described as the same method routinely produce different numbers. Method transfer is the formal exercise of moving a method from one laboratory to another and demonstrating that the results still mean the same thing — and the reason it is an exercise at all is that a written method is an incomplete description of what actually happens.

What a written method leaves out

A method specifies a column, a gradient, a flow rate, a temperature, a wavelength and an injection volume. All of that transfers cleanly. What does not transfer is everything the originating laboratory knew without writing down.

• Dwell volume. The volume between where the gradient is mixed and where it reaches the column head. It differs between instrument designs by several-fold, and it delays the gradient’s arrival, shifting every retention time.
• Extra-column volume. Tubing lengths, fitting dead volumes and detector cell size all contribute band broadening that is invisible in the method text.
• Column batch. Nominally identical packings from different production lots differ measurably in selectivity, and columns from different manufacturers with the same nominal chemistry differ more.
• Mobile phase preparation. Whether the additive is measured by volume or by weight, whether the organic is added before or after, and the purity of the water all shift retention slightly.
• Integration settings. Threshold, peak width and baseline algorithm are usually left to the operator, and they move the number directly.

What the exercise actually consists of

A comparative study, run in both laboratories on the same material, with agreed acceptance criteria decided before any data is collected.

The receiving laboratory typically analyses the same samples as the originating one, in replicate. The two sets of results are compared for agreement in the reported value, in precision, and in what the method detects — because a method that reports a matching purity while resolving fewer impurity peaks has not transferred successfully.

Acceptance criteria are absolute differences agreed in advance. Deciding afterwards what counts as close enough is not a transfer; it is a retrospective justification.

The system suitability test is what makes it portable

This is the mechanism that does most of the work. A system suitability test defines what the chromatography must achieve before any sample result is accepted: a minimum resolution between a specified pair of peaks, a maximum tailing factor, a minimum plate count, and a maximum variation across replicate injections.

Those criteria are properties of the separation rather than of the instrument, so a second laboratory that meets them has demonstrated its system is performing equivalently, whatever its dwell volume happens to be. A method without a system suitability test has no portable definition of working correctly. It is described in system suitability in peptide analysis.

Why retention times are the wrong thing to match

The receiving laboratory will not reproduce the originating laboratory’s retention times, and trying to force it to by adjusting the gradient usually makes things worse.

Relative retention time — each peak’s retention divided by the main peak’s — cancels most of the system-to-system variation and is the figure that should agree. The underlying reasons are set out in retention time as identity evidence.

Where transfers commonly fail

Three patterns recur. A pair of impurities resolved on the original column merges on the receiving one, so the impurity count drops and the purity figure rises. A peak that tailed acceptably on a fresh column tails unacceptably on an older one, and integration becomes irreproducible. Or the receiving laboratory’s water or additive quality raises the baseline at 214 nm enough that small peaks disappear into it.

All three produce a higher purity figure in the receiving laboratory, which is the direction that does not prompt anyone to investigate.

The informal version

Most research-grade analysis involves no formal transfer at all. A laboratory takes a published or supplied method and runs it, and the question of equivalence never arises explicitly.

The practical substitute is a shared reference: both laboratories analysing the same retained sample and comparing. That single comparison detects most of the failures above without any of the formal apparatus, and it is what makes a retained reference sample worth keeping.

What this explains about differing certificates

When two documents on the same material report different purity figures, the gap is usually method transfer that never happened rather than an error by either party. Both laboratories may have run correctly, and their numbers still describe slightly different separations.

That is why comparing purity figures between suppliers or between laboratories has limited meaning, while comparing lots of the same compound analysed the same way has a great deal. Related reading: why certificates disagree on purity and measurement uncertainty.