9.17 : Contaminants and Errors

Effective sample preparation is crucial for accurate and reliable laboratory analysis. During this process, two significant sources of error can arise: concentration bias from improper sample splitting and contamination caused by methods used to reduce particle size, such as grinding or homogenization. Identifying and minimizing these potential errors is crucial to ensuring the validity of the analysis.

Another key consideration is determining the appropriate number of samples required to achieve a desired sampling error. If the target population follows a normal distribution, the sample size n_scan be calculated using the equation:

In this equation, t represents the t-value based on the desired confidence level, s_srepresents the sampling standard deviation, which measures how spread out the samples are from the mean, and e denotes the acceptable sampling error. This formula helps define the necessary number of samples to meet a specific error threshold, ensuring precision in the sampling process.

Furthermore, minimizing the total variance of the analysis involves addressing both the method and sampling variances. Sampling variance can be reduced by collecting an adequate number of correctly sized samples, while method variance improves when multiple analyses are conducted on each sample. By managing both variances effectively, more accurate and reliable results can be obtained.