The International Association of Geoanalysts (IAG) published a protocol for the certification of reference materials in close accord with the International Organisation for Standardisation (ISO) guidelines (Kane et al. 2003). This article supplements that protocol, providing additional discussion of best approaches for pre‐selecting laboratories for participation in certification projects. This discussion also makes a distinction between inter‐laboratory certifications, where n = 15 is the general standard, and expert laboratory certifications, where a much smaller number of laboratories will be deemed qualified to provide data of the quality needed for certification.

The National Bureau of Standards [NBS, now The National Institute of Standards and Technology (NIST)] certified eight glass Standard Reference Materials (SRMs) in the early 1970s. The work was undertaken jointly with Corning Glass Works and the American Society for Testing and Materials. The materials were intended to be used for calibration and control of techniques for the bulk analysis of glass only. In the absence of any CRMs designed for use in microanalysis, many laboratories are using SRMs 610‐617 as though the certification included this use. This manuscript reviews the early certification effort, with attention to those aspects of the original work that have particular relevance to current use of SRMs 610‐617 by microanalytical laboratories.

The International Association of Geoanalysts (IAG) has certified a slate sample, OU‐6, for twelve major and minor constituents, as well as thirty‐five trace elements through an interlaboratory programme conducted in close compliance with the International Organization for Standardization (ISO) Guide 35 (1989). Laboratories were qualified for participation in the certification programme, based on their performance in the prior analysis of OU‐6 as a proficiency test material. Thirty laboratories provided data for the certification, though not for each constituent that was certified. Certification criteria included a means of establishing traceability for the certification data, generally through concurrent analysis of the existing Geological Survey of Japan reference material JSl‐1, agreement of results between laboratories and methods, and a minimum of data rejection (4% rejection rate for OU‐6 data and 6% for JSl‐1 data), preferably for well‐understood technical reasons only. Information values are provided for an additional eight constituents where certification criteria were not met. Uncertainties developed in accordance with the “Guide on Uncertainty in Measurement” (Eurachem 2000) and representing the 95% confidence interval of the certified and information values are reported for all fifty five constituents. The material is currently available in 40 g units for distribution by the IAG. Supply is anticipated to last about ten years.

This certified reference material is intended to be used in the analysis of geological materials -rocks, soils and sediments -or other materials that have a silicate matrix. A unit of OU-6 includes approximately 35 g of pulverised material (98.33% of material < 63 µm and 0.03% > 125 µm) and is packaged in a plastic envelope.The certification project was carried out under the direction of the IAG certification committee following the IAG protocol for the certification of reference materials (Kane et al. 2003). Analyses were provided by thirty participating IAG laboratories, qualified by their performance in the GeoPT™ proficiency testing programme round nine (Potts et al. 2001).The certified concentrations are given in Table 1 for ten major oxides and thirty-five trace elements; information values are given for nine other oxides/elements in the same table. The values are based on measurements using two or more independent analytical techniques. In some cases, definitive methods (e.g., gravimetry for major oxides, isotope dilution-mass spectrometry (ID-MS) for a select few elements) have been used, though generally, certified values were established using comparative analytical techniques.

The results of the first international proficiency test (GeoPZ'l) of geochemical laboratories are presented. A total of 49 laboratories contributed analytical data on 694, a sample of Threlkeld microgranite which had been collected at the International Conference, Geoanalysis 94. Assigned values were derived by robust statistical analysis of contributed data for 12 major elements and 39 trace elements, of which 3 (As, Pb and Th) were provisional values. Z-scores were calculated, as required by the protocol using either 'pure geochemistry' or 'applied geochemistry' fitness for purpose criteria selected by contributing laboratories. An overview of the z-score results indicates that 28.7% of contributed data fell outside the -2