Testing for the within-vial homogeneity

All tests must be performed in the most repeatable way one operator using one instrument in a short period of time for the test. If possible experiments should be done in parallel all extractions in parallel, if possible one calibration. The content of the 

Establish the test method uncertainty on n experiments in the most repeatable way (see 4.3.1.1)  

Take from one vial n sub-samples and measure them with the tested method Repeat the exercise on a second vial  

Where M represents the mass of the test sample, R the relative standard deviation and k the sampling constant corresponding to the mass of sample required to limit the sampling uncertainty to 1% with 68% confidence. The magnitude of k,. may be determined by estimating R from a series of measurements of samples of a mass M. Once k is determined for a given sample, the minimum mass M required for a certain standard deviation can be calculated. 

The minimum sample intake for which homogeneity becomes a significant uncertainty factor of the measurement is (often) however debatable. The CRM is intended to validate methods used for routine measurements. In daily practice of environmental monitoring, the prerequisite is to sub-sample representative samples (see Chapter 2). The problem lies more in having the possibility of handling large samples rather than extremely low sample sizes. It is another debate when forensic analyses are performed as methods may have to be validated on extremely small samples. 

---