Proficiency Test for Chemical Analysis of Iron Ore Materials

Precise and accurate chemical analysis of iron ore is required for contracts of trade and referee analysis, to assure equitable payment for mineral exports and reduce the incidence of disputes and penalty payments. In addition, reliable in-house analysis and improved quality control has the benefit of better resource utilisation and metallurgical accounting.

Since 1992, CSIRO Mineral Resources has coordinated a quality assurance program for the iron ore industry in the form of an international proficiency test for chemical analysis of iron ore materials. The program provides a means for participants to assess their precision and accuracy relative to other industry laboratories.

Two samples are distributed every two months; participants may analyse any (or all) of the following elements using their everyday in-house method (XRF, ICP, titration etc):

Fe, SiO2, CaO, Mn, Al2O3, TiO2, MgO, P, S, K2O, Sn, V, Cr, Co, Ni, Cu, Zn, As, Pb, Ba, Sr, Zr, LOI

Results are collated and summarized by CSIRO.  All data is subjected to a statistical test in accordance with ISO 5725-2 (Accuracy (trueness and precision) of measurement methods and results –  Part 2), using the Cochran and Grubbs’ tests to exclude data on the basis of high within lab error or large laboratory bias from the mean of all labs.  The mean, overall within-laboratory error and overall between-laboratory error are then calculated, and the graphs plotted as deviations from the mean value (after exclusions).  The summarised results are plotted and sent back to participants so that they can monitor their performance relative to other analytical laboratories.

There are currently more than 80 laboratories participating from 21 countries with participants remaining anonymous with only a code to identify their results.

​​​​​​​CSIRO welcomes new participants to join the program. Find out more 

 

Mandel;s between-laboratory consistency statistic – Sample 271 – FeXRF

Mandel’s between-laboratory consistency statistic – FeXRF