Determination of Dust Particle Mineral Percentages by Semiquantitative Analysis Using X-Ray Powder Diffraction and Linear Spectral Unmixing

Mineral dust is the most abundant atmospheric aerosol, primarily from suspended soils in arid and semi-arid regions on earth. Dust aerosols profoundly affect the earth and its atmosphere in a wide range of areas. Mineral dust plays a crucially important role in the Earth’s system due to having a profound impact on the energy budget of the atmosphere through its direct interactions with radiation and indirectly via its influence on the formation processes of clouds and precipitation. Understanding of the mineralogy and other characteristics of particulate matter emission and dispersion from the surface is important to understand its impact on human health, its short and longwave scattering effects, its impact on earth’s temperature, wind, cloud, precipitation rate, and climate system. Small size atmospheric particulate matter, PM₁₀ and PM_2.5 (particulate matter with aerodynamic diameter less than 10 μm and less than 2.5 μm, respectively) are regularly collected on filters and analyzed for the organic and inorganic compounds, and assessed for impacts on air quality.

To characterize minerals, X-ray diffraction is often used to semi-quantitatively identify individual components in polymineralic samples. Using DIFFRAC.EVA (XRD evaluation software), helped to identify phases in a specimen by comparison with standard patterns existing in the library. Precise and accurate determination of the minerals percentages for a sample were carried out following these steps: remove the background, remove kα2, displacement, match/search, matching peaks’ height, and S-Q analysis using DIFFRAC.EVA.

The other technique to quantify endmembers is to perform linear spectral unmixing to determine the relative mineral abundances for heterogeneous sample spectra. This method decomposes the spectrum of a mixed pixel (or spectra) into a collection of constituent spectra, or endmembers, and a set of corresponding fractions. To aim this, lsqnonneg command in MATLAB was used.

For the proposed work, to trap the dust samples in Ilam, a city located western part of Iran, marble dust collectors (MDCO) were used. MDCO was found to be more efficient in collecting dust particles in desert regions (Goossens and Offer, 2000). A total of 37 of samples have been analyzed using VNIR/SWIR, LWIR, and XRD. Out of 37 samples, 5 were analyzed for mineral characterization so far. XRD semiquantitative analysis showed variable amount of quartz (19-56 %, avg. 42 %), calcite (0-54 %, avg. 21 %), clays (illite, kaolinite) (4-37 %, avg. 18 %), feldspar (albite) (10-31 %- avg. 19 %), and phyllosilicate (muscovite) (15-23 %, avg. 18 %). Concentration level indicated most minerals within the samples as the major components, except for calcite (in site1, collected on December 21^st and site2, collected on December 21^st) and Kaolinite ( in site 2, collected on December 21^st). SWIR and LWIR measurements (using ASD and Nicolet FTIR instruments, respectively) were obtained to compare minerals identified by XRD. The mineral signature was identified based on USGS spectral library version 7 (Koklay, et al. 2017). Analyzing sample SWIR spectra indicated features of calcite and illite in all samples except in S1 which did show calcite pattern in XRD where it only had diagnostic valley for illite using ASD. Quartz was identified as major components in some samples in XRD, but this common rock-forming silicate is transparent at shorter wavelengths and will not exhibit any spectral absorption features in this range. There were not any absorption features (in SWIR) attributed to kaolinite, muscovite, and albite in those samples. In the LWIR all samples spectra showed features arising from carbonate. Absorption features for Calcite, illite, and kaolinite are dominant in all samples. Albite showed absorption bands in S1 as it was identified as a major mineral in XRD with 31 percent of abundances roughly more than the other phases. Rather, muscovite was not identifiable in S4 where it has the most percentage of abundance based on the XRD analysis. It was not observed any quartz features in the longwave range due to a decrease in reflectance as the percentage of finer particles increases (Salisbury and D’Aria, 1992). Particle size decreases surface scattering increases, resulting in greater energy absorption, which translates to lower reflectance (Conel 1969).

We expect to perform spectral mixture analysis to determine the relative abundances of minerals compared to those determined by XRD. These data should support modeling of atmospheric dust loading, mass balance, and radiative forcing by different atmospheric constituents.

Keywords: dust mineralogy, IR spectroscopy, X-Ray diffraction, semiquantitative analysis, linear spectral unmixing

References:

Conel, J. E. 1969. Infrared emissivities of silicates: Experimental results and a cloudy atmosphere model of spectral emission from condensed particulate mediums. Journal of Geophysical Research 74(6):1614–1634.

Goossens, D., Offer, Z. Y. Wind tunnel and field calibration of six aeolian dust samplers. Atmospheric Environment, v. 34, n. 7, p. 1043-1057, 2000/01/01/ 2000. ISSN 1352-2310. Disponível em: < http://www.sciencedirect.com/science/article/pii/S1352231099003763 >.

Kokaly, R.F., Clark, R.N., Swayze, G.A., Livo, K.E., Hoefen, T.M., Pearson, N.C., Wise, R.A., Benzel, W.M., Lowers, H.A., Driscoll, R.L., and Klein, A.J., 2017, USGS Spectral Library Version 7: U.S. Geological Survey Data Series 1035, 61 p., https://doi.org/10.3133/ds1035.

Salisbury, J. W., and D. M. D’Aria. 1992. Infrared (8–14 micron) remote sensing of soil particle size. Remote Sensing of Environment 42:147–165.

Submit to:

Student Poster session on “Atmospheric Chemistry, Aerosols, and Air Quality”