Espectroscopia micro-Raman e machine-learning aplicados ao mineral zircão: estudo dos efeitos de radiação e de annealing

Abstract

Minerals such as apatite, zircon, epidote, and muscovite contain uranium and thorium in parts per million (ppm), making them suitable for thermochronological dating methods such as Fission Track Thermochronology (FTT). In these minerals, uranium and thorium atoms may undergo spontaneous fission or decay, producing structural modifications that interfere with diffusion kinetics, chemical etching, and annealing processes. Based on this premise, this Final Undergraduate Project expands the discussion within the FTT and Materials Science communities by investigating the effects of radiation and thermal annealing on the crystal lattice of zircon. In addition to conventional Raman spectral analysis, an algorithm was developed to process large amounts of spectral data efficiently. This new approach, combined with traditional methodologies, provided robust statistical insights. The results showed that thermal annealing was able to recover approximately 92% of the detectable structural damage, while the irradiated sample presented only minor residual defects. Analysis of the full width at half maximum (FWHM) of the ν₁(SiO₄) Raman band effectively distinguished different crystallographic states, with the natural sample (Fossil) showing greater disorder. These findings contribute significantly to the understanding of structural transformations in minerals subjected to geological and thermochronological processes. The integration of Raman spectroscopy with machine-learning algorithms proved to be an effective and reproducible method for characterizing crystal damage and recovery. Such a methodological framework offers promising applications in refining annealing models and improving the calibration of thermochronological techniques.