Sampling particulate matter for analysis - Controlling uncertainty and bias using the theory of sampling

Area of Science:

• Engineering and Environmental Science
• Analytical Chemistry
• Sampling Theory

Background:

• Particulate matter sampling is vital but prone to errors from heterogeneity.
• Sampling errors introduce bias and variance, affecting concentration estimates.
• Existing methods like grab sampling (GS) lack error control.

Purpose of the Study:

• To quantify bias, reproducibility, and representativeness of probabilistic sampling (PS).
• To compare PS against grab sampling (GS) with sieve screening.
• To assess sampling performance for physical and chemical analytes.

Main Methods:

• Probabilistic sampling (PS) adhering to the Theory of Sampling (TOS).
• Grab sampling (GS) with and without sieve screening.
• Analysis of steel microspheres and copper sulfate at varying concentrations.

Main Results:

• Sampling method significantly impacts bias, variance, and representativeness.
• PS improved measurement representativeness by over two orders of magnitude.
• Physical analytes (microspheres) showed higher bias and variance than chemical analytes (copper sulfate).
• Sieve screening in GS introduced significant bias and reduced variability.

Conclusions:

• Probabilistic sampling (PS) is superior to grab sampling (GS) for accurate particulate matter analysis.
• PS effectively minimizes sampling errors, enhancing measurement reliability.
• Sieve screening in GS negatively affects sample representativeness and introduces bias.