Systematic comparison of delta34S measurements by multicollector inductively coupled plasma mass spectrometry and evaluation of full uncertainty budget using two different metrological approaches.

Abstract

A systematic comparison of delta34S measurements by multicollector inductively coupled plasma mass spectrometry (MCICPMS) using two different standardisation approaches and two different reference materials as working standards is reported for the first time. Full uncertainty budgets have been calculated using the Kragten approach. Internal standardisation by measuring Si ratios and external standardisation by bracketing samples with either NIST RM8553 or NIST RM8554 have been compared to correct for the effects of mass bias. The delta34S value and the associated uncertainty ( per thousand) were found to be slightly different when using different approaches. Corrected 34S:32S ratios for NIST RM8553 and NIST RM8554 differed from those previously reported by 0.1 to 0.5% when using external standardisation. 34S:32S ratios for NIST RM8553 and NIST RM8554 differed from those previously reported by 0.06 to 0.15% when using Si internal standardisation. This indicated that in order to minimise deviation from true ratio values, Si internal standardisation would be a more appropriate option when using a Neptune MCICPMS instrument. delta(34)S values were obtained for four different methionine samples and expanded uncertainties (k = 2) expressed in delta34S ( per thousand) ranged from 0.7 per thousand to 1.6 per thousand. Regardless of the approach used for the mass bias correction, three parameters provided the major contributions to the standard uncertainty of the delta34S(V-CDT) value. These were the measured 34S:32S in the sample, the measured 34S:32S for the working standard and the known delta34S(V-CDT) value of the working standard. Results using both approaches are compared and the Si internal standardisation approach is used to provide results for an inter-laboratory comparison.