T116 Mineralogy, Petrology, and Geochemistry: New Approaches to Harnessing the Multidimensionality of Complex Systems I

Wednesday, 28 October 2020
10:00 AM - 12:00 PM
Meeting Rooms

CHARACTERIZATION OF CLAY MINERALS USING DATA FUSION OF RAMAN AND LASER-INDUCED BREAKDOWN SPECTROSCOPIES

Paper number: T116-6
Program: Technical Sessions
Time: 11:10 AM – 11:25 AM

Authors

Abstract

The type and amount of constituent minerals influences the behaviour and properties of rocks and rock assemblages. Further, characterization of bulk mineralogy yields information about source rocks, depositional setting, and diagenetic processes. A detailed mineralogical analysis is thus an essential step in several applied and academic research efforts in the geological, geochemical, and geomechanical fields.

In a proof-of-concept study, we assessed the feasibility of using data fusion to enhance the accuracy of identifying clay minerals in polymineralic rocks using spectroscopy and multivariate analysis. Data fusion is a generic data science term to describe algorithms that combine data from different sources. Here, we merged spectra from Raman (RS) and laser-induced breakdown spectroscopies (LIBS) using simple concatenation. LIBS and RS offer advantages such as no sample preparation, real-time analysis, and minimal sample destruction. In addition, they provide complementary information: LIBS reveals the elemental profile, while RS reveals molecular structures and bonding environments. Given that mineralogy is a function of both the composition and the structure of a specimen, we hypothesized that fusing LIBS and Raman spectra into a single model would allow for better mapping of the data space, resulting in more accurate mineral identification.

Our results showed that the Raman model was not as effective (86% accuracy) as the LIBS model (96% accuracy) in classifying the specimens by their modal clay mineralogy. However, the Raman spectra provided discrimination power complementary to that of LIBS spectra, allowing the fused data model to attain 100% classification accuracy.

Our results indicate that data fusion strategies improve the identification of clay-rich specimens by their modal mineralogy using techniques that are amenable for remote, real-time analyses. We suggest that this data science method not only improves mineral estimates, but also streamlines the analytical workflow. Considering that the benefits of data fusion are derived from the inclusion of complementary data that offer a more comprehensive description of the target, and not the specific experimental parameters themselves, our promising results may reasonably be extended to other contexts and/or conditions not directly investigated in this study. Future research in our lab will assess the application of more complex data fusion algorithms as well as the possibility of achieving reliable quantitative mineral estimates from fused spectra.

Geological Society of America Abstracts with Programs. Vol. 52, No. 6

doi: 10.1130/abs/2020AM-354759

© Copyright 2020 The Geological Society of America (GSA), all rights reserved. Permission is hereby granted to the author(s) of this abstract to reproduce and distribute it freely, for noncommercial purposes. Permission is hereby granted to any individual scientist to download a single copy of this electronic file and reproduce up to 20 paper copies for noncommercial purposes advancing science and education, including classroom use, providing all reproductions include the complete content shown here, including the author information. All other forms of reproduction and/or transmittal are prohibited without written permission from GSA Copyright Permissions.