 LIBS ANALYSIS OF PERCHLORATES AND CHLORIDES IN SOIL IN MARS-LIKE CONDITIONS.  S. Schröder 1,2 K. Rammelkamp 1, A. Cousin 2, D. Vogt 1, P.-Y. Meslin 2, S. Maurice 2 and H.-W. Hübers 1,3 , 1 Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Optische Sensorsysteme, Berlin, Germany. 2 Institut de Recherche en Astrophysique et Planétologie (IRAP), Toulouse, France. 3 Humboldt-Universität zu Berlin, Institut für Physik, Berlin, Germany. (Susanne.Schroeder [at] dlr.de).   Introduction:  Laser-Induced Breakdown Spectroscopy (LIBS) permits rapid in-situ multi-elemental analysis and has been evaluated for applications in planetary science in a number of studies e.g., [1;2]. The first extraterrestrially employed LIBS instrument is ChemCam on NASA's Mars Science Laboratory (MSL) [3;4], which has been successfully analyzing materials on the martian surface since the rover's landing in August 2012 [5]. NASA's upcoming Mars2020 mission with a similar rover will again apply LIBS in combination with other spectroscopy methods with the follow-up SuperCam instrument [6]. Perchlorates (ClO4 ) are among the very interesting species that have been identified on Mars in several locations such as at the Phoenix landing site [7] and by the Sample Analysis at Mars (SAM) instrument on board MSL at Gale Crater [8;9]. Perchlorates can give insights into aqueous geochemical processes, they can affect the water content of soils and of the atmosphere and due to their ability to significantly reduce the freezing point of water their presence can have broad implications for astrobiology. There is big interest in identifying perchlorates also in the ChemCam LIBS data as ChemCam is one of the most frequently used instruments of MSL that is on average applied several times per week. ChemCam data with a strong Cl(I) emission line at 837 nm was often found to also have enhanced Na, which could be indicating the presence of Na-chloride (Halite) or Na-perchlorate. Previous work has shown that it is feasible with LIBS to analyze and classify samples of similar salts such as perchlorates and chlorides in pellets of martian regolith simulant (MRS) and even in frozen water solutions [10;11]. In this work we investigate further the potential of LIBS to identify perchlorates and chlorides in martian analogue soil. We extended the sample suite and are comparing data taken of samples with Ca- Mg-, and Na-perchlorates to those comprising the chlorides with the same cation, each. Data was taken at two different LIBS setups for comparison: a setup with a highresolution Echelle-spectrometer at DLR Berlin and a ChemCam copy at IRAP, Toulouse. Experimental and Samples:  In the LIBS technique, radiation from a high-power, pulsed laser is focused onto a sample from which material is ablated and a luminous plasma is produced, e.g.[12]. Information on the elemental composition is obtained from the relaxation of excited atoms and ions due to specific transitions that produce characteristic spectral lines. A typical LIBS spectrum has tens to hundreds of emission lines ranging from ultra-violet to near infra-red wavelengths. Bremsstrahlung (free-free transition) and recombination (free-bound transition) of electrons in the plasma result in a superimposed continuous spectrum.  At DLR Berlin, an infrared Nd:YAG laser was used to generate the plasma at short stand-off distances (< 1 m): 1064 nm wavelength, 15 mJ pulse energy on the sample's surface, 8-10 ns pulse duration, 10 Hz repetition rate. The plasma emission was detected with a high resolution echelle spectrometer (14-96 nm) with a time-gated intensified CCD enabling a continuous coverage from 270 nm to 850 nm.  At IRAP Toulouse LIBS spectra of the same samples were taken with the ChemCam replica (1067 nm, up to 14 mJ, 4.5 ns pulse duration, 3 Hz). The pure salts (mostly as hydrates) were mixed with JSC Mars-1A Martian Regolith Simulant (MRS) each and compressed into pellets containing about 50 wt% of the pure salt. 1 g of the soil mixture was pressed at 5 tons to pellets with a diameter of 14 mm. The pellets were stored in low pressure environments before the LIBS analysis to prevent the highly hygroscopic salts from getting more humid. Measurements were performed simulating a martian environment with an appropriate gas mixture composed of mainly CO2 at 7 mbar. Data Analysis and Results:  In order to explore the data qualitatitvely and to identify the spectral features and emission lines that could be used to differentiate LIBS data from chlorides from data of perchlorates we applied the multivariate analysis method Principal Component Analysis (PCA). Here, the spectra are sorted due to their similarities and variabilities and uncorrelated variables called principal components (PCs) replace a bigger number of correlated variables in the data. Similar spectra cluster in the space defined by the PCs.  The data from DLR Berlin was recorded with a time delay of 300 ns to exclude the continuous emission of the early stages of plasma emission but to not lose too much of the ionic emission. The LIBS data from IRAP includes the total plasma emission with an integration time of 3 ms and was analyzed two times for comparison: (A) with and (B) without the continuum removed from the data.  300 400 500 600 700 800 900 -0.15 0.00 0.15 0.30 -0.18 0.00 0.18 0.36 300 400 500 600 700 800 900 -0.12 0.00 0.12 0.24   loading PC1  loading PC2 wavelength (nm) loading PC3 Fig. 1:  PCA loadings for the first three principal components (PCs) of the ChemCam-like LIBS data with no continuum removal applied. The most intense lines are from the cations (Ca, Na, Mg).       The first three PCA loadings of the data which include the continuous emission (B) are shown in Fig. 1, the corresponding PCA scores plots are shown in Fig. 2. While PC1 and PC2 mostly comprise the strong emission lines of the cations of the salts (Ca, Na, Mg) and explain together already 68% + 28% of the data, the PC3 (2%) separates the chlorides from the perchlorates with only one exception of a Na-perchlorate spectrum. Along PC3, the chlorides cluster on the positive side with stronger cation emission, while the perchlorates are characterized by a stronger emission of oxygen from the 777 nm triplet and at 844 nm. This can be explained by the higher number of oxygen atoms in the perchlorates. The usually strongest chlorine emission in ChemCam data at 837 nm is challenging to detect [13;14] and was not found to significantly contribute in the first three loadings. The highest hydrogen emission at 656 nm is mostly observed in the data of the Mg containing salts and seems to be of little importance for the discrimination between chlorides and perchlorates.      In the IRAP LIBS data with the continuum not removed, the clusters of the salts with the same cation become slightly better defined. A better improvement, however, can be obtained when only specific wavelength channels with important emission lines are selected for the PCA. The PCA of the DLR data shows mostly similar results. -0.02 -0.01 0.00 0.01 0.02 0.03 0.04 -0.020 -0.015 -0.010 -0.005 0.000 0.005 0.010 0.015 0.020 -0.02 -0.01 0.00 0.01 0.02 0.03 0.04 -0.004 -0.002 0.000 0.002 0.004 0.006 0.008 MgCl2 Mg(ClO4)2 CaCl2 Ca(ClO4)2 NaCl NaClO4 PC 2 ( 28 %) PC1 (68%) PC 3 ( 2% )PC1 (68%) Fig. 2:  PCA scores plots of ChemCam-like LIBS data. With PC3 the LIBS data of the perchlorates can be distinguished from data of chlorides.       Conclusion:  Our results show that with multivariate methods and optimized data processing it is possible to differentiate between chlorides and perchlorates in LIBS data. Further experiments will be done with different compositions, studying also the effect of the water content in the samples.  References:  [1] Knight et al. (2000) Appl. Spectrosc. 54, 331-340. [2] Del Bianco et al. (2006) Int. Planet. Probe Workshop, 116-123. [3] Maurice S. et al. (2012) Space Sci. Rev. 170, 95 166. [4] Wiens R.C. et al. (2012). Space Sci. Rev. 170, 167227. [5] Maurice S. et al. (2016) J. Anal. At. Spectrom. 31 4, 863-889. [6] Clegg S. et al. (2015) LPSC #1832. [7] Hecht et al. (2009) Science, 325:64-67. [8] Leshin L.A. et al. (2013) Science 341 1238937. [9] Ming D.W. et al. (2014) Science 343 1245267. [10] Schröder S. et al. (2013) Icarus, 223, 6173. [11] Schröder S. et al. (2011) LPSC #1608. [12] Cremers & Radziemski (2006) Wiley. [13] Forni O. (2014) LPSC #1328. [14] Cousin A. et al. (2015) EMSLIBS P37. 
