11:00 AM - 1:00 PM
[PCG18-P08] Elemental analysis of meteorites using laser-induced breakdown spectroscopy for future asteroid explorations
Small landers MASCOT and MINERVA landed on the asteroid Ryugu and explored its surface. The MMX rover will explore the surface of Phobos in 2027. Developing instruments for in-situ elemental analysis on asteroids is important for future asteroid lander missions. Laser-induced breakdown spectroscopy (LIBS) is a method for elemental analysis of a sample by measuring the emission spectra of plasma generated by laser pulses. LIBS was used by NASA’s Curiosity and Perseverance rovers [1, 2]. However, the intensity of plasma emission is known to decrease significantly under a high vacuum [3], potentially deteriorating its measurement capability when used in vacuum. Moreover, the analytical performance of LIBS for meteorites from asteroids has not been investigated yet. Because such meteorites, especially carbonaceous chondrites, have elemental compositions distinct from terrestrial or Martian rocks, the accuracy achieved for such asteroidal samples is not understood yet.
In this study, we used LIBS to analyze the elemental abundance of meteorites from asteroids: an ordinary chondrite Chelyabinsk, carbonaceous chondrites Allende and Murchison, and eucrite Millbillillie. We evaluate prediction accuracy by comparing our results with the elemental compositions reported in previous studies.
Each meteorite was placed in a vacuum chamber (3x10-2 Pa) 1 m away from the LIBS instrument. The samples were irradiated with 100 laser pulses at 3 spots. Pulse energy was 30 mJ. We also confirmed that this laser can be used in vacuum. A spectrometer was synchronized with the laser to acquire emission spectra (350-885 nm). Ten geological standards distributed by AIST or USGS were used as calibration samples. A multivariate spectral analysis method called Partial Least Squares Regression (PLS-R) [4] was used to construct a quantification model using these samples. The elemental compositions of the meteorites were predicted by putting their spectra into the model.
The LIBS results for the Millbillillie meteorite, for example, showed that the LIBS- and literature values differ by -0.8 wt% (SiO2), -0.4 (TiO2), +0.5 (Al2O3), +3.8 (MgO), +3.2 (CaO), +0.28 (Na2O), +2.1 (K2O) and -13 wt% (Fe2O3). When the prediction error in the PLS model is considered, the results for Si, Ti, Al, Na and K were consistent with the literature values, but the predicted elemental abundances were underestimated for Fe and overestimated for Mg and Ca. The predicted abundances of the carbonaceous chondrite Murchison were consistent for Ti and K, but underestimated for Mg and Fe, and overestimated for Si, Al, Ca and Na.
Our results confirmed that the LIBS measurements of the meteorites yield elemental abundances consistent with literature values when the elemental compositions of the meteorite falls within those of calibration sample set. In contrast, some elements showed systematic deviations from the known values. For example, Ca was overestimated and Fe was underestimated for all the meteorites; Si and Al were overestimated and Mg was underestimated for Chelyabinsk, Allende and Murchison; Na was overestimated in Allende and Murchison. Our interpretation of these results is that Millbillillie, a basaltic meteorite, was predicted relatively accurately because the calibration samples used in this study were primarily made of terrestrial igneous rocks. In contrast, carbonaceous chondrites, whose SiO2 concentration was as low as 30 wt%, were overestimated in terms of SiO2 because of its large deviation from the compositional range of the calibration samples. These results suggest that making standard samples specific to those chondrites would enable more accurate measurement of carbonaceous chondrites. Those with 25-45 wt% SiO2, 1-5 Al2O3, 20-30 MgO, 0.1-0.4 Na2O and 15-40 wt% Fe2O3 would be particularly important.
References
[1] Wiens+ 2012, SSR 170, 167-227 [2] Wiens+ 2021, SSR 217: 4
[3] Knight+ 2000, Appl. Spectrosc. 54, 331
[4] Clegg+ 2009, Spectrochim. Acta. B 64, 79
In this study, we used LIBS to analyze the elemental abundance of meteorites from asteroids: an ordinary chondrite Chelyabinsk, carbonaceous chondrites Allende and Murchison, and eucrite Millbillillie. We evaluate prediction accuracy by comparing our results with the elemental compositions reported in previous studies.
Each meteorite was placed in a vacuum chamber (3x10-2 Pa) 1 m away from the LIBS instrument. The samples were irradiated with 100 laser pulses at 3 spots. Pulse energy was 30 mJ. We also confirmed that this laser can be used in vacuum. A spectrometer was synchronized with the laser to acquire emission spectra (350-885 nm). Ten geological standards distributed by AIST or USGS were used as calibration samples. A multivariate spectral analysis method called Partial Least Squares Regression (PLS-R) [4] was used to construct a quantification model using these samples. The elemental compositions of the meteorites were predicted by putting their spectra into the model.
The LIBS results for the Millbillillie meteorite, for example, showed that the LIBS- and literature values differ by -0.8 wt% (SiO2), -0.4 (TiO2), +0.5 (Al2O3), +3.8 (MgO), +3.2 (CaO), +0.28 (Na2O), +2.1 (K2O) and -13 wt% (Fe2O3). When the prediction error in the PLS model is considered, the results for Si, Ti, Al, Na and K were consistent with the literature values, but the predicted elemental abundances were underestimated for Fe and overestimated for Mg and Ca. The predicted abundances of the carbonaceous chondrite Murchison were consistent for Ti and K, but underestimated for Mg and Fe, and overestimated for Si, Al, Ca and Na.
Our results confirmed that the LIBS measurements of the meteorites yield elemental abundances consistent with literature values when the elemental compositions of the meteorite falls within those of calibration sample set. In contrast, some elements showed systematic deviations from the known values. For example, Ca was overestimated and Fe was underestimated for all the meteorites; Si and Al were overestimated and Mg was underestimated for Chelyabinsk, Allende and Murchison; Na was overestimated in Allende and Murchison. Our interpretation of these results is that Millbillillie, a basaltic meteorite, was predicted relatively accurately because the calibration samples used in this study were primarily made of terrestrial igneous rocks. In contrast, carbonaceous chondrites, whose SiO2 concentration was as low as 30 wt%, were overestimated in terms of SiO2 because of its large deviation from the compositional range of the calibration samples. These results suggest that making standard samples specific to those chondrites would enable more accurate measurement of carbonaceous chondrites. Those with 25-45 wt% SiO2, 1-5 Al2O3, 20-30 MgO, 0.1-0.4 Na2O and 15-40 wt% Fe2O3 would be particularly important.
References
[1] Wiens+ 2012, SSR 170, 167-227 [2] Wiens+ 2021, SSR 217: 4
[3] Knight+ 2000, Appl. Spectrosc. 54, 331
[4] Clegg+ 2009, Spectrochim. Acta. B 64, 79