Vol.5 No.1 2012

Research paper : An analysis method for oxygen impurity in magnesium and its alloys (A. Tsuge et al.)−28−Synthesiology - English edition Vol.5 No.1 (2012) of Sn and magnesium oxide originally contained in the Mg sample. As shown in the profile of Fig. 3, it was expected that the oxygen could be extracted completely as carbon monoxide when heated with power of over 5,000 W. Therefore, we studied the percentage (recovery rate) of the amount of oxygen in the carbon monoxide extracted and detected against the 4 mg of oxygen predicted from the added magnesium oxide (the stoichiometric composition of oxygen in magnesium oxide is 39.7 %) when Mg was separated using the model sample where 10 mg of magnesium oxide was added to a mixture of 0.3 g Mg and 0.5 g Sn. The residue was heated at 5,000 W. In this experiment, however, the recovery rate remained at about 20 %, and much time was consumed to figure out the cause of discrepancy. Finally, we found that the Mg rebonded with the oxygen in the carbon monoxide that was produced from the residue when the crucible was heated with evaporated Mg remaining in it. We reached a simple procedure to prevent the rebonding: the residue is removed from the graphite crucible in association with the removal of the remaining Mg by opening the furnace and followed by the oxygen measurement with a new graphite crucible.Since the reactivity between the Mg vapor and the generated carbon monoxide was unexpectedly high, we were possessed by a bias that the analysis procedure had to be done under inert gas flow in the IGF-IRA method. As a consequence, much time was wasted to find the cause of discrepancy in oxygen analysis. There was a concern that the sample may become oxidized when the residue was exposed to the atmosphere during analysis, but the effect of the oxidation was ruled out by conducting the blank test, analysis without placing the sample in the crucible. 2.2 The method of samplingIn commercial transaction, analysis values measured according to a standard are considered to be representative ones of traded merchandises. Therefore, in standardizing the analysis method, it is necessary to specify a sampling method to have the analysis value representing the characteristic of the entire sample, in conjunction with basic analysis procedures. Normally, a larger amount of sample than that actually used in the analysis is taken from several parts of the object to be measured and then these samples are mixed well to homogenize.(homogenizing process) At the beginning of this research, we tried the “chip sampling method”, which is used widely in the chemical analysis of the major component and impurities of metal, allowing homogenization of the sampled material. In this method, homogenizing process was done for the chips sampled by drilling several areas of the object to be measured.Mg is, however, readily oxidized and the oxidation of the samples during the chip sampling procedure could not be avoided only with simple and practical oxygen shielding. Consequently, we examined the core drill sampling method which has only a small effect of oxidation, but lack of homogenization. The results of examination of the sampling methods and additional findings expected to contribute to the standardization will be described below.2.2.1 Chip sampling methodIn the ordinary chip sampling method, the sample will have a large surface area in the chip formation process. The surface oxidation of the chip cannot be avoided in materials with high oxygen affinity like that of Mg. Direct application of the chip sampling to Mg was expected to be quite difficult.Considering these results, we attempted to apply the chip sampling with a simple oxidation resistant treatment. Figure 5 shows the schematic diagram. Using a miniature lathe placed inside a nitrogen-purged glove box, a 10 mm diameter rod was turned and its stock removal ranged 0.2~0.6 mm. The chips formed by turning were measured with a tablet frame, taken out from the glove box through a pass box, followed by forging to form tablet samples. As shown in Fig. 6, the chip sampled under nitrogen purge showed higher oxygen concentration than that of the core of the rod. The oxidation of the chips could not be avoided, the same as those Fig. 6 Effect of surface oxidation by atmosphere during chip samplingFig. 5 Chip sampling method and creation of samples for analysis②Form the chip to size 7 mmΦ × 4 mm7 mm4 mmRod of 10 mm in diameteris turned with stock removal of 0.2 mm~0.6 mmChip①Cut to 4 mm length, leaving core of diameter 7 mmMg rod sampleTo oxygen analysisTablet frameDepth from surface (mm) purgeIn atmosphere0Oxygen concentration (wt%)0.9 - 1.50.3 - 0.90 - 0.3CoreCore1.2 - 1.50.6 - 0.80.2 - 0.40 - 0.2


page 31