Evaluation of analytical uncertainty in quantitative determination of elements - The case of boron

The uncertainty including accuracy and precision is the most vital factor that determines the overall quality of quantitative analysis. The uncertainty has been, however, evaluated relatively within the same analytical technique. Given this background, the present study evaluates the uncertainty on quantitative elemental analysis with a quasi-absolute approach. The objectives of this study are (1) to investigate the analytical uncertainty of prompt gamma-ray analysis (PGA), a chemical interference-free method in principle, on the quantitative analysis of boron and (2) to evaluate the applicability of inductively coupled plasma optical emission spectrometry (ICP-OES), a common technique for quantitative elemental analysis including boron.

PGA provided analytical quantity that is equivalent to the true quantity. The quantity values determined for a series of boron-containing samples are all well above the detection limit of the PGA system, the quantity resolution of which is also much smaller than the minimum difference in quantity among the samples. These facts confirm that the evaluation of analytical uncertainty with the present PGA system is statistically meaningful. The analytical uncertainty in both methods was adequately evaluated by comparing the results from PGA and ICP-OES for a series of boron-containing materials with different physical/chemical properties (i.e. CrB₂, B₄C, and solidified products of stainless steel-B₄C melt) and the major sources of uncertainty in both methods are specified. The conditions for sample preparation/pretreatment were optimized to lower the uncertainty.

This study proposes a new concept to perform the quasi-absolute evaluation of analytical uncertainty by employing a chemical interference-free technique. The proposed concept is not limited to the combination of PGA and ICP-OES as demonstrated in this study, but it is applicable to any combination of any analytical methods for any element. Hence, the concept demonstrated in this study could be beneficial to a wide range of analytical chemistry.