10:10 AM - 10:25 AM
[MGI31-04] Relationship between extracted iron content and color values of soils in the forensic geological examination
Keywords:forensic geology, soil color, dithionite-citrate extract, free iron oxides
Soil is an important trace evidence to prove the linkage between a criminal and a crime scene. In the forensic examination, color analysis is fundamental for discrimination and screening of soil samples, and elemental analysis is also informative to characterize soil. However, the relationship between the color and the elements is less applied to forensic soil examination in Japan. Therefore, we studied the relationship between color and iron content in supernatants of removal of free iron oxides obtained by pretreatments for soil examination to establish prediction model of free iron content.
Samples were collected from 39 sites around Nirasaki, Yamanashi Prefecture, Japan. First, sieved samples (less than 2 mm in diameter) were treated with hydrogen peroxide to decompose organic matter, and silt and clay fractions (“clay” fraction) were separated by wet sieving using stainless steel mesh (53 µm). Next, the fractions were treated with the dithionite-citrate system to remove free iron oxides. Finally, the suspensions were separated with supernatants and “clay” fractions by centrifugation. The 50 µL of the supernatant was dropped on a filter paper and dried. Per each supernatant, three filtered samples were prepared and examined by X-ray fluorescent analysis (Primus ZSX, Rigaku). The iron content was quantified by the fundamental parameter method, and the log ratio with sodium content (ln(Fe/Na)) was calculated. Color analysis was carried out with sieved samples (less than 2 mm in diameter), "clay" fractions after the organic decompositions, and after removal of free iron oxides using spectrophotometry (CM-700d, Konica Minolta)., and repeated five times. The correlation coefficient was calculated between ln(Fe/Na) values with chromaticity values (L*, a*, and b*) to investigate the relation between the extractable iron content and color values. In addition, the linear regression model of ln (Fe/Na) and a* (organic decomposition) were constructed, and the determination coefficient was used to evaluate the stability of the model.
The ln(Fe/Na) values negatively correlated with L* values in each treatment, and positively correlated with a* and b* values in each treatment. The strongest correlation with ln(Fe/Na) was a* of organic decomposition, and the correlation coefficient of ln(Fe/Na) and a* was 0.95. Therefore, a* after the organic decomposition is considered to be suitable for linear regression modeling of ln(Fe/Na). The determination coefficient of the regression line of ln(Fe/Na) and a* was 0.9. The high correlation implied that the iron oxide extracted by the dithionite-citrate system was originated from secondary minerals such as hematite and goethite. The quantification of color values will be helpful for predicting the content of extractable iron.
Samples were collected from 39 sites around Nirasaki, Yamanashi Prefecture, Japan. First, sieved samples (less than 2 mm in diameter) were treated with hydrogen peroxide to decompose organic matter, and silt and clay fractions (“clay” fraction) were separated by wet sieving using stainless steel mesh (53 µm). Next, the fractions were treated with the dithionite-citrate system to remove free iron oxides. Finally, the suspensions were separated with supernatants and “clay” fractions by centrifugation. The 50 µL of the supernatant was dropped on a filter paper and dried. Per each supernatant, three filtered samples were prepared and examined by X-ray fluorescent analysis (Primus ZSX, Rigaku). The iron content was quantified by the fundamental parameter method, and the log ratio with sodium content (ln(Fe/Na)) was calculated. Color analysis was carried out with sieved samples (less than 2 mm in diameter), "clay" fractions after the organic decompositions, and after removal of free iron oxides using spectrophotometry (CM-700d, Konica Minolta)., and repeated five times. The correlation coefficient was calculated between ln(Fe/Na) values with chromaticity values (L*, a*, and b*) to investigate the relation between the extractable iron content and color values. In addition, the linear regression model of ln (Fe/Na) and a* (organic decomposition) were constructed, and the determination coefficient was used to evaluate the stability of the model.
The ln(Fe/Na) values negatively correlated with L* values in each treatment, and positively correlated with a* and b* values in each treatment. The strongest correlation with ln(Fe/Na) was a* of organic decomposition, and the correlation coefficient of ln(Fe/Na) and a* was 0.95. Therefore, a* after the organic decomposition is considered to be suitable for linear regression modeling of ln(Fe/Na). The determination coefficient of the regression line of ln(Fe/Na) and a* was 0.9. The high correlation implied that the iron oxide extracted by the dithionite-citrate system was originated from secondary minerals such as hematite and goethite. The quantification of color values will be helpful for predicting the content of extractable iron.