Simplification Application of Two-error Regression Model in the Uncertainty Evaluation on Inductively Coupled Plasma-Mass Spectrometric Determination of Sr in Drinking Waters

This paper introduces the method of double-error regression in the course of standard curve fitting was introduced. The uncertainties of standard solutions and signal shift of the instrument were propagated into the final analytical results of Sr in drinking water by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) and by using a continuous propagation model. The fitting results obtained by one time and multi-time linear regression models did not reflect the true result and two different sources of uncertainty. The double-error regression model was applied to the standard work curve reported in this paper. The uncertainty calculation for the standard work curve was derived from the double-error regression equations, which built the relationship among detection limits, parameters for curve regression and associated uncertainties. The equation for total uncertainty was simplified by a continuous propagation model and calculation equations for the uncertainty components. The results indicate that the relative uncertainty of the measurement results is greater when the Sr concentration is lower. It also shows that the value of the detection limit increases and the sensitivity is reduced with the rougher processing for the preparation of a standard series. Both issues above prove that the test result matches the practical situation well.