All mercury analyzers need to be calibrated in order to ensure correct results.
The method of calibration will depend on the type of instrument and the application.

For atomic absorption and atomic fluorescence analyzers used for measuring mercury in liquids, the calibration is achieved by preparing a series of solutions containing concentrations of mercury which will cover the expected range of mercury concentration in the samples. These standards are analysed using the instrument at the same settings as for the sample and a graph is constructed (automatically in the software) of concentration against detector signal strength. All samples are then analyzed and compared to this calibration graph.
Alternatively a single standard solution may be prepared and the sample compared to this standard. This method although seemingly simpler needs the standard to be very close to the sample in mercury concentration for the procedure to have any accuracy. This method is generally more accurate for atomic fluorescence instruments with their greater linearity than for atomic absorption instruments with their lack of linearity causing high curvature graphs.

For mercury in gas analysis the calibration is usually achieved by introducing measured volumes of mercury vapour directly into the instrument by syringe from a container of mercury vapour at equilibrium in air at a known temperature. By introducing varying volumes of mercury vapour again a calibration graph is obtained.

For automatic online mercury analyzers for gases using atomic fluorescence detection the calibration is achieved from a module which generates known concentrations of mercury vapour from a mixture of vapours from a source of mercury vapour and mercury free air which are mixed in controlled conditions at a fixed temperature
The mixtures are measured to cover the range of expected values of the mercury concentrations of the samples and can be programmed to me measured at fixed time intervals.

This unit can also be programmed to send exact mercury concentrations to the sampling point which can be used to validate the method of analysis.