Technical Market Support » General
The key objectives of this ACARP funded study (for arsenic, boron, cadmium, chlorine, chromium, fluorine, mercury, lead, molybdenum and selenium) were to:
· Determine if Australian laboratories are meeting expected accuracy and precision requirements of ISO standards;
· Determine if precision and accuracy levels being achieved meets the requirements of the Australian coal industry, taking into account trends in environmental and trading restrictions; and
· Ascertain if the range of (certified) reference materials (RMs) for these elements in coal matrices meets the quality assurance needs of the Australian coal industry and thereby builds confidence in the accuracy of the analysis of these trace elements.
An industry survey associated with this project indicated that the published levels of precision in relevant Australian and ISO standard test methods typically meet the needs of Australian coal industry.
The ability of Australian laboratories to meet the accuracy and precision requirements of ISO standards, and whether there are an adequate range of RMs, to build confidence in the accuracy of analyses depends very much on the parameter in question. This aspect was assessed by the conduct of an interlaboratory study (ILS) in which a range of RMs were submitted blind to participating Australian coal testing laboratories and an associated laboratory survey. The number of participants was parameter dependant ranging from four to ten and resulted in insufficient data to draw firm, statistically sound conclusions for parameters other than chlorine. This was identified at the proposal stage as a project risk. The results from this ILS, and its associated laboratory survey, indicate a high level of confidence in the accuracy and precision of chlorine analysis. For the remaining parameters where the number of participants is less than ideal, results are considered interim and can be summarised as follows:
· Between laboratory precision and accuracy of mercury analyses were of concern. This was heavily influenced by one laboratory whose mercury results were biased high. This laboratory has been notified of the problem;
· Likewise for boron, analyses appear to be meeting expected precision limits with the exception of one laboratory that uses an accredited, although non-standard, test method and was biased high. This will be bought to the attention of the laboratory;
· For lead, molybdenum and chromium assessment of between laboratory precision and accuracy was inconclusive. This is attributed to the low number of participating laboratories. There however were no particular flags for concern with lead and chromium analyses; molybdenum results were less precise;
· Between laboratory precision was good for cadmium, however the accuracy of analyses presents significant concerns. All laboratories were biased very low on one RM (NIST 1635a). Further, there were a number of instances where the difference between duplicates reported was much wider than indicated as acceptable by method repeatability limits;
· For selenium and arsenic, there were some concerns regarding between laboratory precision and accuracy of analyses;
· For fluorine, a known low bias on the bomb combustion method (ASTM D3761) was confirmed. However, the results conducted by pyrohydrolysis techniques reveal a higher than expected number of flags for concerns raising the possibility that the precision and accuracy of this technique is not as good as could be expected.
Although interim, any potential identification of bias or reduced precision should be of concern to the Australian coal industry and worthy of further monitoring to establish if any identified trends are due to inherent bias or are within the normal fluctuations of method precision. For arsenic, cadmium, fluorine, mercury and selenium where industry concern is highest further rounds of testing need to be conducted to determine if identified problems have been remedied and provide more robust conclusions to support the findings of this study. Key risks to achieving this objective are the availability of suitable samples combined with the possible variability in laboratory precision over time. Ultimately this work should provide confidence that the analysis of these elements by Australian laboratories is meeting required levels of precision. Alternatively, it may identify where analytical methods need to be reviewed, either within laboratories or at the national level. Furthermore, the current project is initially focusing on the precision and bias of the analytical methods and is not addressing the likely errors that might results from sampling and sample preparation.
In summary, this project has provided a detailed evaluation of the status of trace element testing at a point in time. There is a high level of confidence in the analysis of chlorine in coals. For other trace elements, research outcomes are interim and more variable. This suggests that further rounds of testing would be of value in improving confidence in the accuracy and precision of trace element analyses. The establishment of a continuous proficiency test program is likely to be considered necessary to maintain confidence in the trace element testing results from Australian laboratories. This work, combined with the production of a range of trace element RMs in Australian coals would assist in the continuous improvement of trace element analysis and provide valuable marketing information for those elements where trade or environmental restrictions apply.