Standard solution stability: how long a prepared standard can be used

A certificate describes a material, not a preparation

Reference standard documentation is precise about the container and silent about the flask. The leaflet for a pharmacopoeial standard allows the closed container to equilibrate at ambient temperature before opening, requires the material to be used as is, and states that once the container is opened the entire content must be used immediately, because further storage and re-use are not warranted. The producer characterised a solid, or a sealed solution, and holds stability data for that form; nothing in the file covers what happens after the material meets a diluent.

Two regulators divide the work usefully. For non-quantitative solutions such as mobile phases, FDA accepts an assessment, and a literature review may be enough. For in-house solutions used in quantitative work, including standard and sample solutions in assay or impurity testing, FDA expects a formal stability study behind the expiry, a use-by date written on the container when the solution is prepared, and discarding at that date. ICH Q2(R2) points the same way: robustness testing should show the stability of sample preparations and reagents for the duration of the procedure, and its annex for separation techniques names sample and reference material preparations explicitly.

The practical consequence is that the hold time is a laboratory measurement, not a specification inherited from the supplier. The certificate settles what the powder is worth; the solution is a separate object with a separate clock.

Two baselines, and one that hides the answer

A stored-standard study has two possible reference points, and choosing the wrong one turns a failing solution into a passing report.

The third row catches laboratories. In a single-point assay the result is a ratio of sample response to standard response, so a standard and a sample that each lost 1.5 per cent produce a clean 100.0 per cent. The arithmetic below is an illustration, not data.

Against the stored standard the assay reads 101.0 per cent, the 1.5 per cent standard loss minus the 0.5 per cent sample loss; against a fresh standard it reads 99.5 per cent, the sample as it actually is. The first number is wrong because the ruler shrank, not because the instrument misbehaved. An impurity method behaves differently: percentage areas are normalised to the total, so a standard that degrades into a peak near the analyte shifts the profile rather than cancelling.

What actually changes in a vial

Six mechanisms account for most failures, each leaving a signature in the data like those a forced degradation study produces, and the signature matters more than the mechanism because it names the control to change.

The adsorption row deserves the arithmetic. One millilitre at 1 microgram per millilitre holds 1 microgram, so a 1 per cent loss is 10 nanograms; over the roughly 10 square centimetres of a 2 mL vial that is 1 nanogram per square centimetre, well below a monolayer. A loss invisible at 1 mg/mL is therefore routine at 1 microgram/mL, which is why an impurity standard and an assay standard of the same substance need separate hold times.

Designing the study

The study has to reproduce routine use: the same vial, cap, diluent lot and tray temperature that the sequence will use.

Run the stored aliquots as one sequence against standards prepared fresh at each timepoint, and inject the zero-time aliquot again at the end so instrument drift is visible. Use two preparations rather than one: a single vial cannot distinguish a preparation error from a stability effect.

Fixing the criterion before the data arrive

A criterion written after the numbers are known is not a criterion. Two parts are needed and they are not interchangeable. For the standard, the stored preparation has to give a result within a stated percentage of a fresh one, with 2 per cent a common starting point that still has to be justified against the precision the method achieves. For the chromatogram, no new peak may appear at or above the reporting threshold, and each specified impurity may change only within a set limit.

The percentage has to be smaller than the tolerance the final result must meet, or the drift eats the specification. An assay released against 95.0 to 105.0 per cent that repeats to 1 per cent cannot afford a standard allowed to drift 5 per cent; the allowance is gone before the sample is injected. For impurities the relevant comparison is the limit for that impurity, not the main component.

Report the direction as well as the size. A monotonic trend across five timepoints is evidence of instability even when every point passes, because the next point is predictable; scatter with no trend is more often injection variability than chemistry. The number on the label is the last timepoint that passes both the value and the trend, not the longest timepoint that happened to pass.

From data to a number on the label

Three constraints decide the usable period, and the shortest governs: the validated hold time, the longest routine sequence, and the expiry written on the container at preparation. The label carries the preparation date and time, the diluent, the concentration and the expiry, and the procedure states that the solution is discarded at expiry rather than re-tested and re-dated. Treating a solution that fails inside its date as a routine re-preparation removes the only signal that the hold time is wrong.

Long sequences need a further rule, because bracketing solves less than it appears to. Re-injecting a standard every ten injections corrects for instrument drift and does nothing for a standard that is degrading, since every bracketing injection uses the same ageing solution. Where a sequence must run longer than the hold time, the honest options are to split it, to prepare a second standard part way through, or to shorten the method, and the choice belongs in the procedure rather than in the analyst judgement at two in the morning.

When a hold time does not transfer

A hold time established for one preparation is data about that preparation. It stops applying when the diluent composition or pH changes, when the working concentration moves by more than a small factor, when the container or septum material changes, when detection moves into the region where extractables absorb, when freezing replaces refrigeration, or when the material is a different salt or hydrate form. Method transfer adds two more variables: ambient temperature, and the tray temperature of different equipment.

The economics are worth stating plainly. Re-running a stability study costs a few days of a chromatograph and a little standard; a hold time that was assumed costs a batch investigation, and the degradation study that defined the impurity profile will not explain a result that drifted because the standard did.

What to ask a supplier

None of this is answerable from a catalogue entry, so it belongs in the questions asked before an order. Ask for the batch validity statement current at the time of use, not at the time of purchase. Ask whether the material is a solid or a solution; for a solution, ask for the mass fraction and its expanded uncertainty, the density at a stated temperature and the solvent composition, because converting a mass fraction to a molar concentration without the density at the right temperature costs several per cent on an acidic solution standard. Ask for storage and shipping conditions, the filled weight so a unit can be taken whole, and whether the material is single-use after opening. The distinction between a reference standard and an analytical standard decides which questions apply.

Then keep three clocks separate. The material has a shelf life the producer supports, the weighing that starts the solution has its own uncertainty, and the hold time is measured in your own laboratory, on your own diluent, in your own vials. A laboratory that can show those three numbers separately has nothing to reconstruct when a result is questioned.