Analytical Sciences

Abstract − Analytical Sciences, 26(2), 277 (2010).

X-ray Fluorescence Analysis of Cr6+ Component in Mixtures of Cr2O3 and K2CrO4
Tatsunori TOCHIO,*1 Shusuke SAKAKURA,*2 Hirofumi OOHASHI,*3 Hirohisa MIZOTA,*2 Yanhui ZOU,*2 Yoshiaki ITO,*2 Sei FUKUSHIMA,*4 Shigeo TANUMA,*4 Takashi SHOJI,*5 Hajime FUJIMURA,*5 and Michiru YAMASHITA*6
*1 Department of Physics, Faculty of Science, Kobe University, 1-1 Rokkodai, Nada, Kobe 657-8501, Japan
*2 Institute for Chemical Research, Kyoto University, Uji, Kyoto 611-0011, Japan
*3 Harima Office, National Institute for Materials Science, Sayo, Hyogo 679-5148, Japan
*4 National Institute for Materials Science, 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan
*5 X-ray Research Institute, Rigaku Industrial Corp., 14-8 Akaoji, Takatsuki, Osaka 569-1146, Japan
*6 Hyogo Prefectural Institute of Technology, 3-1-12 Yukihira, Suma, Hyogo 654-0037, Japan
X-ray fluorescence analysis using Cr Kα spectra was applied to the determination of the mixing ratio of Cr6+ to (Cr6+ + Cr3+) in several mixtures of K2CrO4 and Cr2O3. Because the powder of K2CrO4 contained large particles that were more than 50 μm in diameter, it was ground between a pestle and a mortar for about 8 h. The coarse particles still remaining were removed by using a sieve with 325-mesh (44 μm) in order to reduce the difference in absorption effects between emissions from Cr6+ and those from Cr3+. The mixing ratio, K2CrO4/(K2CrO4 + Cr2O3), of the five mixtures investigated is 0.50, 0.40, 0.20, 0.10, and 0.05 in weight, respectively. Each spectrum obtained was analyzed by decomposing it into two reference spectra, those of the two pure materials, K2CrO4 and Cr2O3, with a constant background. The results for the mixtures containing K2CrO4 of more than 20 wt% are that the relative deviation from the true value is less than ∼5%. On the other hand, when the content of K2CrO4 decreases to less than 10 wt%, the relative deviation gets so large as 20 – 25%. The error coming from a peak separation of spectrum involved in our results were estimated by applying our method to five sets of data for each mixture computationally generated, taking into account the uncertainty in total counts of real measurements.