This report presents the outcome of an inter-laboratory comparison exercise (ILC) on the determination of selected metals in acetic acid solutions and the determination of migration of elements from ceramic and glass tableware. The exercise was organised by the European Union Reference Laboratory for Food Contact Materials (EURL-FCM) to evaluate the enforceability of measures under discussion for the future revision(s) of ceramics Directive 84/500/EEC and the implementation of provisions laid down in the Commission Regulation 10/2011 for plastics.
National Reference Laboratories (NRLs) requested the EURL-FCM to organise an ILC with the aim to check (i) the analytical abilities of participating laboratories to quantify Cu, Fe, Zn and Sb in a solution of acetic acid 3% and Ba, Co, Mn, Pb, Cd and Al in a solution of acetic acid 4%; (ii) to test the laboratories performance to carry out the migration test on ceramic and glass articles and (iii) to derive precision criteria, including repeatability and reproducibility for the release of elements from tableware.
Participation in this ILC was mandatory for the nominated NRLs, and open to Official Control Laboratories (OCLs) and other invited laboratories. A total of 53 participants from 27 countries registered to the exercise. Twenty-nine NRLs from 27 countries participated in this ILC and all of them reported results.
Laboratory results were rated using z-score in accordance with ISO 13528:2015 . The target standard deviation for the ILC assessment (σ), for all measurands was calculated based on previous ILC03/04 2014 exercises for spiked solutions and using the robust reproducibility standard deviation for the migration exercise on tableware. Repeatability and reproducibility standard deviations for the quantification of elements in acetic acid 3% and acetic acid 4% (spiked solutions and leachates) were calculated using robust approaches .
The outcome of this exercise was satisfactory. The rate of success was almost always higher than 80% for all elements in all samples. No difference between the performance of NRLs and OCLs was observed. Since the exercise required the uncertainty estimation, an additional assessment was provided to each laboratory, indicating how reasonable their measurement uncertainty estimation was. lζ-scoresl were systematically higher than z-scores, that underlines the need of an improvement in estimation of the measurement uncertainties.