Vol.9 No.2 2016

Research paper : Development and utilization of geochemical reference materials (T. OKAI)−68−Synthesiology - English edition Vol.9 No.2 (2016) grinding was a fist-sized to egg-sized mass of the same rock (source rock ball). By grinding through “tomozuri,” contamination was minimized. Afterwards, the mortar and pestle was considered too inefficient, and a steel jaw crusher became the mainstream tool for grinding (blade is made from manganese steel). Around this time, as a certain amount of iron and manganese were contained in samples and equipment was improved, the contamination from the jaw crusher ceased to be a problem. Currently, as shown in Fig. 6, after grinding with a steel jaw crusher, samples are pulverized in an alumina-lined ball mill with an alumina or source rock ball. When JG-1 and JB-1 were exhausted, the re-prepared samples JG-1a and JB-1a were prepared from the remaining samples collected at the time of initial reference preparation; they were ground by the current method. As JB-1b is a re-re-prepared sample of JB-1, the analytical values of major components in three samples of JB-1, 1a, and 1b were compared as an example (Table 3), and there was no sign of effects from the crusher.4.3 Standard value determination method and publication of data4.3.1 Standard value determination and publication of data during free distributionPrepared samples were analyzed first by GSJ and then distributed worldwide with these initial analytical values, while additional analytical data were gathered. Basically, it was notified publicly through reviews of academic societies etc. that the samples had been prepared,[14][15] requests for distribution were gathered, and under the condition to send back analytical data of the sample, reference materials were distributed free of charge. There were two methods of distributing reference materials. One of these methods was used by GSJ: in return for free distribution, the reporting of analytical values was required, and the standard values were decided from the collected analytical values. The other method was to sell the reference materials with standard values assigned. Currently, the latter method is the mainstream method; however, at the time, only NBS (present-day NIST) and BAS used the latter method, and a majority of the organizations distributed reference materials via the first method. Only the latter fits the original definition of reference materials, and the former should be called a common analysis sample for research. However, based on the USGS’s experience with G-1 and W-1, analytical values may change according to evolution of analytical methods. Although a majority of the elements are targets of analysis (research) of geological materials, it is difficult to assign a definite value to many of the elements from the beginning. Therefore, to collect analytical values of various elements in response to evolving analytical methods, the former method was more effective.Reported analytical results were summarized and published in a journal for the first time in 1971.[16] At that point, all analytical values (24 for JG-1 and 17 for JB-1, including analytical methods and the name of the analysts), the overall mean, standard deviation, and the mean except value that exceeds ±2 σ from the mean were published. Subsequently, after a certain amount of analytical values had been collected, the reported analytical values were calculated statistically to obtain the standard value, and this standard value was published. However, as the number of analytical values increased, not all of the results could be published, and reports were limited to the mean value for each analytical method and the range of analytical values. In short, although it is called “standard value,” its name also has changed through time. In the early stage, it was called “Consensus Mean (Value),” but afterwards, results with a sufficient number of analytical values and high reliability were called “Recommended Value.” Results with a limited number of analytical values and low reliability were initially called “preferable data,” but were later changed to “Reference Value.” In addition, because of the nature of the method by which the standard value was decided based on collected analytical values after distribution, the standard value had the potential to change. Table 4 shows changes in the major component values of the JG-1 sample, which actually changed little.4.3.2 Change in the distribution method and certified reference materialsOnce geochemical reference materials became widely used, free distribution with the requirement of reporting of analytical values became difficult for general users, and the demand for the sale of reference materials without a requirement for reporting increased. In addition, global standardization by ISO started to affect geochemical reference materials in the late 1990s, and the production of certified reference materials, which are distributed with certified values (determined (w/w %)SiO2Al2O3T-Fe2O3MnO(μg/g)CoCrCuNi13355.142538.20.1538.9914.5352.37recommendedvaluesJB-1 (1968)13956.739238.60.1489.0514.4552.41recommendedvaluesJB-1a (1984)14855.543940.30.1479.0214.3851.11initial analyticalvaluesJB-1b (1996)T-:total, JB-1&JB-1a:Imai et al. (1995)[12], JB-1b:Terashima et al. (1998)[13]Table 3. The major component contents of the JB-1 sample and the re-prepared sample


page 12