Moisture and Weighing: Build a Defensible Sample Chain

In pharmaceutical quality control the measurement is rarely the weak link. The weak link is the chain that leads to it: the sample that was identified, the mass that was recorded, the portion that was actually tested, and the record that ties the three together. A defensible moisture result is a chain, not a number.

Four links in the chain

Identity: the sample tested is the sample specified. Mass: the quantity weighed is the quantity recorded, at a resolution that matters against the tolerance. Portion: the material measured represents what was received. Result: the number reported can be traced to raw data without an intermediate transcription.

What the balance has to carry

A balance is defensible when four conditions hold: it is levelled and the level is checked, it is calibrated across the working range rather than only at capacity, it is used above its minimum weight for the required tolerance, and it stands where air movement does not reach it. The fourth is the one most often lost to a convenient bench position.

Minimum weight is a calculation, not a label. It depends on the readability and on the tolerance the method needs — a 0.1 mg balance used for a 20 mg portion with a tight tolerance carries a different risk from the same balance used for a 2 g portion.

Moisture: two methods, two claims

Loss on drying reports what leaves the sample at a defined temperature and time. Karl Fischer reports water. A specification that says “water content” and a method that says “loss on drying” are making different claims, and the difference surfaces at the audit rather than at the bench.

For hygroscopic material the handling matters more than the instrument: pre-dried vessels, a desiccator between steps, a short exposure time, and weighing by difference rather than transferring material between containers.

Where the chain breaks

One tolerance, two instruments

Moisture and mass share a tolerance chain. A moisture result expressed as a percentage inherits the uncertainty of the weighing. If the method requires three significant figures, the balance resolution and the portion size have to support three figures — and that requirement belongs in the method, not in the analyst’s judgement on the day.

The record

What a reviewer needs is short: which sample, which balance, which programme, which portion, which raw values, which calculation, who performed it and who checked it. Everything else is commentary. A workflow that produces those items without anyone reconstructing them afterwards is a workflow that survives inspection.