Indonesian experience in providing accuracy-based proficiency testing scheme using metrologically traceable reference values for elements mass fraction in drinking water

Abstract

Proficiency testing (PT) scheme for the mass fraction of arsenic (As), cadmium (Cd), copper (Cu), nickel (Ni) and zinc (Zn) in bottled drinking water matrices was conducted based on ISO/IEC 17043. The aim of the PT scheme was to evaluate laboratory capability in measuring minerals and contaminants in drinking water to comply with the implementation of Indonesian National Standard SNI 3553:2015 and the Regulation of Indonesia Minister of Health No. 492/2010. The PT material was a candidate of reference material, the homogeneity and stability testing on the material was performed and evaluated according to ISO 13528 and ISO Guide 35, and the material found sufficiently homogenous and stable during the transportation condition and PT period. Metrologically traceable reference values were used as the PT assigned values which were obtained using a high accuracy method, isotope dilution inductively coupled plasma mass spectrometry (ID ICP-MS) and gravimetric standard addition (GSA) with internal standard ICP-MS. The performance of the participants was evaluated mainly against the reference value based on z-scores and ζ-scores which is meaningful for assessing participants’ ability in measurement capabilities and measurement uncertainty evaluation. The PT results according to z-scores showed that more than 80 % of total 22 participants obtained satisfactory performance for Cu, Ni and Zn, while only 47 % and 58 % for As and Cd. The differences between the reference values and the robust mean of participants for As (21 %) and Cd (68 %) were found significant in the scheme which shown a big variability of measurement capability of Indonesian testing laboratories in this fields. Unsatisfactory performance of laboratories within their claimed uncertainty based on ζ-scores evaluation means that the laboratories may not assess accurately their measurement uncertainty. Therefore, re-assessment of their uncertainty estimation is needed, and the laboratories are encouraged to improve their capability in the measurement uncertainty for further enhancement.