Statistics – Computation
Scientific paper
Dec 2008
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2008agufm.p53a1429t&link_type=abstract
American Geophysical Union, Fall Meeting 2008, abstract #P53A-1429
Statistics
Computation
1042 Mineral And Crystal Chemistry (3620), 1060 Planetary Geochemistry (5405, 5410, 5704, 5709, 6005, 6008), 3255 Spectral Analysis (3205, 3280), 5494 Instruments And Techniques, 6225 Mars
Scientific paper
The ChemCam instrument on the Mars Science Laboratory will use laser-induced breakdown spectroscopy (LIBS) to quickly and accurately determine elemental compositions of surface materials up to 9 meters away from the rover. Some key mineral groups may occur on Mars in sample sizes large enough that single phases can be probed by the LIBS instrument with a laser spot size of ~0.5mm. Of these, phyllosilicates are significant because spectral features attributed to vibrations of OH groups bound to Al, Fe, and Mg in phyllosilicates have recently been identified on Mars by OMEGA and the MRO CRISM instrument. Furthermore, to maximize science return from the ChemCam LIBS data, it is necessary to understand chemical matrix effects, which arise when chemical and physical interactions affect spectral intensities in unpredictable ways. This work explores matrix effects and calibration issues for a group of phyllosilicate minerals. Because they share the same crystal structure, some aspects of matrix effects should be mitigated. LIBS spectra of 17 phyllosilicate samples were recorded at the Los Alamos National Laboratory in ChemCam- and Mars-analogue conditions at a stand-off distance of 9 meters in the spectral range of 224-927 nm. Three analysis techniques are employed to determine elemental concentrations: univariate analysis, which correlates peak intensity with elemental concentration; peak area stepwise regression analysis (PASRA), which correlates multiple peak areas with elemental concentration; and partial least squares analysis (PLS), which correlates the entire spectrum with elemental concentration. The relative effectiveness of each technique's ability to predict known concentrations was assessed and compared. Results show that all three methods give reasonable results for most elements. The conventional univariate analysis is the least effective in predicting chemical compositions accurately. Both PLS and PASRA appear to compensate for or exploit the chemical matrix effects, though PLS is superior in predicting compositions of unknowns in this data set because it utilizes all the channels in a spectrum. The PASRA technique is far less computationally intensive, but may be more instructive with regard to the physics contributing to the chemical matrix effects because it identifies only a few key peaks that are diagnostic of specific elements. However, PASRA will require a larger data to set achieve accuracy comparable to PLS.
Barefield James E.
Bishop Janice L.
Clegg Samuel M.
Darby Dyar M.
Schaefer Martha W.
No associations
LandOfFree
Laser-induced Breakdown Spectroscopy of Phyllosilicates for ChemCam Calibration does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Laser-induced Breakdown Spectroscopy of Phyllosilicates for ChemCam Calibration, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Laser-induced Breakdown Spectroscopy of Phyllosilicates for ChemCam Calibration will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1239309