Gradients versus isocratic runs

Some mixtures separate at one solvent strength. Others cover such a wide range of retention that a single composition either piles the early peaks together or stretches the late peaks until they are too wide to use. The decision is which of those mixtures you have, and whether the instrument's dwell and the column's equilibration can support the programme you want. Modes themselves are outlined in chromatography methods in life-science labs.

What stays constant, and what has to change

An isocratic run holds mobile-phase composition still. Retention is then a fixed preference for the stationary phase under that one condition, which is the distribution IUPAC names as chromatography. Early peaks can be sharp. Later peaks broaden because they spend longer in the bed and spread by diffusion and slow mass transfer. If every analyte you care about elutes within a handful of column volumes at that composition, isocratic work is simpler to transfer and kinder to detectors that cannot tolerate a changing solvent.

A gradient changes solvent strength on purpose. In reversed phase the organic fraction rises. In ion exchange the salt, or sometimes the pH, rises. In normal phase the polar modifier rises. Strongly retained molecules start to move once the mobile phase reaches the strength that releases them, so late peaks stay narrower than they would have been in a weak isocratic hold. You pay for that with equilibration. After the strong end of the programme, the bed has to return to the starting strength for several column volumes before the next injection means the same thing. A standard that elutes earlier on the second injection is telling you the reset was incomplete.

Dwell volume is part of the method

The gradient you draw on the screen is the gradient at the mixer. The column sees it later, delayed by the dwell volume: tubing, mixer, and any pulse damper that sits upstream of the bed. Divide dwell by flow and you have a delay in minutes. On a large preparative pump with a bulky mixer, that delay can be a large fraction of a short analytical programme. The column then spends the opening minutes in the starting composition even though the screen shows a ramp. Peaks shift. A method moved from a low-dwell instrument to a high-dwell instrument without editing the table will not overlay.

Measure dwell the way the pump manual describes, often with a step of a ultraviolet tracer and no column, or record the value the manufacturer quotes for that configuration and then check it. Write it next to the gradient table. When you specify a new system, dwell belongs in the same sentence as flow and column volume. A fraction collector timed from the detector will also disagree with the programmed salt or organic percentage until you shift by that delay. Ion-exchange fractions labelled "eluted at 200 millimolar" are mislabelled if the label came from the screen clock and the dwell was ignored.

Criteria a specification should name

Name the retention problem first. If a scout isocratic run already spaces the peaks you care about, specify isocratic composition, flow, temperature, and column. If the scout shows a crowded front and a long tail of late peaks, specify a gradient. State the starting composition, the ending composition, the time or the column volumes spent on the ramp, and any holds. A shallow ramp spends resolution. A steep ramp spends time and can collapse neighbours into one peak. There is no universal slope. The slope that resolves your pair on your column volume is the one to write down.

State the equilibration hold in column volumes, not only in minutes, so a future change in flow does not silently shorten the reset. State the detector. Ultraviolet and mass spectrometry tolerate gradients. Refractive index generally does not, because the solvent change itself is a signal. If you need that detector, an isocratic method is the honest specification even when a gradient would have been prettier on an absorbance trace.

State how fractions will be cut. An isocratic peak can be collected by a volume window once retention is stable. A gradient peak should be collected against the detector, with the dwell in mind, and then assayed. Column volume still scales the method: a ramp of ten column volumes is a different chemical programme from a ramp of ten minutes on an unspecified bed.

Injection volume and sample solvent still apply. A gradient does not rescue a sample dissolved in a stronger solvent than the starting composition. The strong plug moves before the ramp has any say. Match the sample to the start, and measure the injection with a volumetric device used as ISO 8655-1 expects of that class of apparatus.

The dwell bracket in that drawing is the volume you forgot when two instruments disagreed. Fix the table or quote the dwell. Do not average the retention times and call the difference column ageing until the delay is ruled out.

Branches when the repeat injection moves

If every peak, standard and sample, shifts later by the same fraction of the ramp, suspect flow or a changed dwell, not a mysterious change in selectivity. Check delivered flow by collecting volume over a timed minute. If only the late peaks move, the ramp shape or the strong solvent concentration has changed. Remake the strong solvent and confirm the pump's composition with a step test. If the second injection of a standard is earlier than the first, lengthen the equilibration in column volumes and repeat the pair. Samples run in that gap are not comparable to yesterday's calibration.

A power cut mid-ramp leaves the bed at an unknown composition. After supply returns, re-equilibrate from the start for the full hold and inject a standard before any sample whose fraction you mean to keep. In a building where the supply drops on a hot afternoon, a long shallow gradient is a gamble unless the sample can be injected again. Prefer a scout you can finish, or a system on a supply that holds, when the fraction is the entire stock.

Protein identity is still a separate record. A gradient salt concentration at elution is a useful annotation next to a sequence from UniProt. It is not an identification in the sense HUPO expects of proteomics data.

Safety and the limit of the claim

Gradients mix solvents that may be flammable, acidic, or incompatible. A programme that crashes phosphate into high acetonitrile is a precipitation hazard as well as a chromatography failure. Stay inside the solubility you have checked in a beaker, and inside the column's pH and pressure window. Research traces are not diagnostic results. Waste is mixed solvent. Ventilate as your chemical assessment requires.

What to send

Say whether you need isocratic or gradient operation, the solvent pair, the column volume you plan to use, the flow, the detector, the fraction size, and any dwell you already measured on the sending system. Ask for mixer type and an achievable dwell. Hardware classes are in the scientific instruments catalogue. Put the table in the quote request. A protein gradient question can be discussed against the custom protein expression and purification reference as a method, not as a campaign already being run. ---