Processing of X-ray microtomography images to investigate crystallized volcanic rock textures

Fabio Arzilli1,2, Margherita Polacci1, Don R Baker3, Patrizia Landi1, Daniele Giordano4 and Lucia Mancini2, (1)Istituto Nazionale di Geofisica e Vulcanologia, Pisa, Italy, (2)Elettra - Sincrotrone Trieste, Basovizza (Trieste), Italy, (3)McGill University, Earth and Planetary Sciences, Montreal, QC, Canada, (4)Università di Torino, Dipartimento di Scienze della Terra, Torino, Italy

Contact First Author: Fabio Arzilli; arzilli.fabio@gmail.com

Abstract ID#: 35286

 

English Abstract:
X-ray computed microtomography has become a fundamental tool in geosciences to study rock textures directly in 3D. This technique is useful to study the vesiculation of pumices [1], but only recently there has been more interest to study the crystallization process using volumetric datasets. Crystal and silicate glasses often have similar density and absorption contrast, therefore, image processing requires more sophisticated procedures to separate them.

Feldspars are one of the most abundant crystalline phases of igneous rocks, hence their crystallization conditions and abundance are intensely investigated by geologists. Moreover, feldspar is one of the most difficult phases to segment during image analysis, making the separation of feldspars from the silicate glass a challenge.

Here we show that the crystallinity of natural pumice of Stromboli and synthetic trachyte from Campi Flegrei magma can be studied using microtomographic datasets. Data were collected using the propagation-based synchrotron X-ray phase contrast microtomography technique at the SYRMEP beamline of the Elettra-Sincrotrone Trieste Laboratory (Basovizza (Trieste), Italy). Although X-ray phase contrast imaging can improve the visualization of weakly absorbing features, making visible the edge of feldspars, the segmentation of these crystals cannot be directly obtained from the raw phase contrast X-ray images. Phase-retrieval methods are needed for extracting phases of interest. A single distance phase-retrieval algorithm [2] was applied to the dataset to improve the contrast between feldspar and glass. Through this technique we are able to isolate feldspars both in natural samples and in synthetic ones, and to perform quantitative analysis of these phases using volumetric datasets. Our preliminary results have demonstrated that phase retrieval processing will be an invaluable tool for geologists to study rock textures.

[1] Polacci M. et al. (2009) J. Geophys. Res. 114, B01206. [2] Paganin D. et al. (2002) J. of Microscopy 206, 33-40.