Estimation of the uncertainty of metal content in a batch of waste printed circuit boards from computer motherboards

Electronic wastes are a valuable resource due to their critical and precious metal content. To include these wastes in recycling or recovery chains, it is necessary to precisely determine their metal content. Because analysing the whole sample of a batch of electronic waste is not practical, different preparation and sampling or subsampling steps are necessary. Sampling induces an error in the composition of the final sample compared to that of the initial batch, which finally leads to uncertainty in the final metal content measurement as compared to the "actual" batch metal content. The aim was to characterize the uncertainty in metal content of a batch of 372 kg of WPCB. Thirty-nine metals were analysed and thirty-two were considered: base, precious, rare-earths and critical metals. An empirical method (i.e. replicated measurement tests) was thus applied, based on statistical calculations according to Eurachem Guidelines. Uncertainty arising during the 3 different stages of the preparation process (primary, secondly and tertiary sampling steps) was calculated. For the analysed given weight (0.5 g), the shredding efficiency, which directly affects metal particle size distribution, was found to be the most important factor influencing the uncertainty. Uncertainties in base metal content, which is often concentrated in the coarsest particles, arose mainly from the last preparation step (tertiary sampling). Conversely, precious metals and rare-earths were finely ground during the 3 preparation steps, which led to low uncertainties, despite their low concentration in the waste (<337 mg/t for precious and < 35 mg/t for rare-earths).