Calibration transforms an instrument’s raw signal into a meaningful concentration value. The fundamental principle is that, over a defined range, the response y is proportional to the concentration x: y = m x + b. The calibration function is established by measuring standard solutions of known concentration and fitting a mathematical model to the data. Without proper calibration, even the most sophisticated instrument produces data of limited value.
External standard calibration is the simplest and most widely used approach. A series of standards containing the analyte at known concentrations is prepared in a matrix that matches the sample as closely as possible. The instrument responses are plotted against concentration, and the resulting regression equation is used to calculate unknown concentrations from their measured responses. This method assumes that the standards and samples behave identically in the instrument, which requires careful matrix matching.
The internal standard method compensates for instrument drift, injection volume variations, and matrix effects. A known amount of a compound (the internal standard) — chemically similar to the analyte but not present in the sample — is added to every standard and sample. The ratio of the analyte response to the internal standard response is used for calibration. This ratio effectively cancels out systematic errors that affect both species equally. Internal standardization is standard practice in chromatography and ICP-MS.
Standard addition is employed when matrix effects are severe and cannot be replicated in calibration standards. Known quantities of the analyte are added directly to aliquots of the sample itself. The instrument response is plotted against the added concentration, and the x-intercept of the regression line gives the original analyte concentration (as a negative value). Standard addition inherently accounts for matrix effects because all measurements are made in the same sample matrix. Its main drawback is increased sample throughput time.
The linear dynamic range is the concentration interval over which the response-concentration relationship remains linear within acceptable tolerance. Working outside this range risks biased results from saturation (at high concentrations) or poor signal-to-noise (at low concentrations). Multipoint calibration (5–8 standards) is preferred over single-point calibration because it allows assessment of linearity through the correlation coefficient and residual plots. Weighted regression (1/x or 1/x²) is recommended when the variance is not constant across the concentration range (heteroscedasticity). Calibration must be verified regularly through independent quality control standards, blank measurements, and continuing calibration checks interspersed with sample analyses.
