University of Twente Student Theses


Identifying Different Epidotes Using Infrared Imaging Spectroscopyin the Rantau Dedap Geothermal System, Indonesia

Dulzamirki, Wildan Alicondro (2023) Identifying Different Epidotes Using Infrared Imaging Spectroscopyin the Rantau Dedap Geothermal System, Indonesia.

Full text not available from this repository.

Full Text Status:Access to this publication is restricted
Abstract:Epidote is a sorosilicate mineral, which is important in geothermal systems as an index mineral for subsurface temperature. In Rantau Dedap, it is used as an indicator mineral for determining the top reservoir and casing production setting of the drilling process. A problem arose when two different types of epidotes, relict and fresh epidotes, were identified within the system. An uncertainty occurred as to which epidote was representative of the indicator. To address this issue, this study employed Infrared Imaging Spectroscopy (IRIS) as the main method and incorporated other methods, such as X-ray Diffraction (XRD), Attenuated Total Reflectance (ATR), petrography, and Electron Probe Microanalysis (EPMA). Two types of samples were used: 85 cuttings and four short pieces of drill cores. In cuttings, the IRIS results were not achieved due to the presence of the resin effect, so XRD and ATR were used to validate the findings. The results confirmed through XRD analysis indicated that pyroxene, displaying greenish brown colours, represented the relict epidotes found from the surface to shallow depths. Most of the epidotes in the shallow depth displayed yellow colours, as confirmed by XRD and ATR. In drill core samples, two different epidote episodes were found: the first was epidote + plagioclase, and the second was chlorite + epidote ± calcite. Both were confirmed by thin sections, and the second epidote association was the only one detected from IRIS. The decision tree classification revealed that all epidotes contained Fe-OH absorption features at 1544-1548 nm, 2251-2254 nm, and 2335-2340 nm. These features were also observed in the mean epidote spectra, which were compared to the background spectra. However, the mean class spectra contained a mixture of chlorite, calcite, and illite. Some issues were detected in IRIS. The presence of noise caused vague features, which classified chlorite and calcite as epidote classes. Consequently, the results showed significant differences from the thin sections. Furthermore, the noise created two nm shifts at the pixel level within an individual crystal. Therefore, any compositional differences of epidote (zoning) could not be identified. In cutting samples, the epoxy resin influenced most of the suspected epidote grains. This mixed resin was caused by the penetration depth, where the biggest grain (thickest) had the purest spectrum. Despite these problems, the comparison of mean epidote spectra versus background spectra offers valuable information on the epidote spectral features. In addition, IRIS provides more extensive spatial information compared to thin sections. Even though the resin in the cuttings had an effect, there was still a clear difference between the spectral features of epidote grains and the reference spectra of clinozoisite between 2340 and 2344 nm. This observation suggests that IRIS could differentiate epidote from clinozoisite. To improve epidote detection using a decision tree classifier, it is necessary to have mineral references consisting of a mixture of epidote, chlorite, calcite, and illite. For cutting samples, a 1.5-mm thickness is recommended to avoid a resin mixture.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Subject:38 earth sciences
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page