New Insights Into Impact Glass Formation and Evolution Using Machine Learning and Aerodynamic Levitation Laser Heating Experiments

dc.contributor.advisorMacris, Catherine
dc.contributor.authorMarrs, Ian James
dc.contributor.otherBarth, Andrew
dc.contributor.otherDruschel, Gregory
dc.date.accessioned2022-10-04T13:55:36Z
dc.date.available2022-10-04T13:55:36Z
dc.date.issued2022-09
dc.degree.date2022en_US
dc.degree.disciplineDepartment of Earth Scienceen
dc.degree.grantorIndiana Universityen_US
dc.degree.levelM.S.en_US
dc.descriptionIndiana University-Purdue University Indianapolis (IUPUI)en_US
dc.description.abstractImpact processes, where a meteor strikes a planetary body’s surface, are ubiquitous in the Solar System. These highly energetic events require study by both computational methods and experimental investigation. An impact process of particular interest to our study is the impact plume, a collection of vaporized rock and superheated gases that is produced during an impact event. Tektites are silica rich (roughly rhyolitic), extremely dry, and often contain both lechatelierite inclusions (amorphous SiO2) and flow textures (schlieren) and are an impact product of particular interest to this study. Tektites likely form either very early in the impact process or within the impact plume itself as condensates, and therefore offer a unique insight into the early stages of the impact cratering process. Here, we present both the results of the statistical analysis of published tektite geochemistry and the geochemical analysis of a variety of glasses produced in an aerodynamic levitation laser furnace. The major findings of the statistical analysis are that the variance of tektite geochemistry is broadly controlled by MgO, CaO, K2O, and Na2O, that the Australasian strewn field (an extensive region of tektite distribution) is best subdivided into five geochemical subgroups, and that random forest classification models can predict the strewn field or geochemical subgroup of an unknown tektite with >94% accuracy. In terms of our heating experiments, in nearly all cases, Na2O and K2O are rapidly lost from the melt due to evaporation, while Al2O3, CaO, and TiO2 become progressively enriched. Volatility is far more dependent on peak heating temperature than on heating time. Additionally, the chemical constituents of basalts are less readily volatilized than those of rhyolites or loess, with few exceptions. We also find that the volatility of the chemical constituents of non-standard samples is far more variable than for standard samples and that oxygen fugacity has a strong influence over elemental volatility in the aerodynamic levitation laser furnace. Changes in oxygen fugacity can either result in variable, exaggerated, or even opposite volatility trends depending on the material and oxide in question.en_US
dc.identifier.urihttps://hdl.handle.net/1805/30177
dc.identifier.urihttp://dx.doi.org/10.7912/C2/3033
dc.language.isoen_USen_US
dc.rightsAttribution-ShareAlike 4.0 International*
dc.rights.urihttp://creativecommons.org/licenses/by-sa/4.0/*
dc.subjectTektitesen_US
dc.subjectMachine learningen_US
dc.subjectHigh temperature geochemistryen_US
dc.subjectGeologic glass formationen_US
dc.titleNew Insights Into Impact Glass Formation and Evolution Using Machine Learning and Aerodynamic Levitation Laser Heating Experimentsen_US
dc.typeThesisen
Files
Original bundle
Now showing 1 - 5 of 5
No Thumbnail Available
Name:
IJM-MSc_Thesis-Appendix_A-TektiteDatabase.xlsx
Size:
381.05 KB
Format:
Microsoft Excel XML
Description:
Appendix A - Tektite database
Loading...
Thumbnail Image
Name:
IJM-MSc_Thesis-Appendix_F-SEMImages.pdf
Size:
9.79 MB
Format:
Adobe Portable Document Format
Description:
Appendix F - SEM images
No Thumbnail Available
Name:
IJM-MSc_Thesis-Appendix_B-SuppTables.docx
Size:
47.42 KB
Format:
Microsoft Word XML
Description:
Appendix B - Supplemental Tables
No Thumbnail Available
Name:
IJM-MSc_Thesis-Appendix_E-SuppFigs.docx
Size:
1.08 GB
Format:
Microsoft Word XML
Description:
Appendix E - Supplemental Figures
Loading...
Thumbnail Image
Name:
IJM-MSc_Thesis-FINAL.pdf
Size:
9.16 MB
Format:
Adobe Portable Document Format
Description:
Thesis text
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.99 KB
Format:
Item-specific license agreed upon to submission
Description: