The Analytical Column

Every calibration curve rests on a column of numbers nobody looks at twice: the concentrations of the standards. They are typed in once, usually straight off the bottle (10, 50, 100 mg/L), and from then on every result the instrument reports is measured against them. If those numbers are a little wrong, the curve is a little wrong in exactly the same way, at every level, and nothing in the fit will tell you. A curve can only be as true as the standards behind it. This piece is about making those standards true: weighing what you put in, correcting for what it really is, and calculating the concentration you actually made instead of the one you meant to make. (Here the solid or neat material is weighed and the solution is made up to volume; preparing the whole solution by mass is covered further down.)

The one-line idea

A standard's concentration is not a label. It is a calculation from three measured things: how much you weighed, how pure it was, and how much solution you made.

Every term carries its own uncertainty, and the worked example in the Eurachem/CITAC guide shows how they combine. For a cadmium standard made from 100.28 mg of high-purity metal (standard uncertainty 0.05 mg), of purity 0.9999 (u = 0.000058), made up to 100.0 mL, which is 0.1000 L (u = 0.07 mL), the result is 100.28 mg × 0.9999 / 0.1000 L = 1002.7 mg/L, with a combined standard uncertainty of 0.9 mg/L as the guide gives it.¹ Two things to notice. The true value is 1002.7, not the 1000 on the label. And in that example, a high-purity metal, purity barely matters: the relative uncertainties are about 0.07% for the volume, 0.05% for the mass and 0.006% for the purity. For a 98%-pure organic compound with a purity uncertainty of half a percent, purity would dominate instead.