A temperature-dependent in situ micro-FTIR and micro-Raman spectroscopic investigation was performed on powdered (FTIR, Raman) and single-crystal (Raman) lizardite-1T samples between room temperature and 819 °C. Between room temperature and 665 °C, the OH stretching bands shift to lower wavenumbers, demonstrating a weak expansion of the O3-H3 … O2 interlayer distance. Band deconvolution of FTIR and Raman spectra at room temperature show differences in the number of bands in the OH stretching region with respect to group theory: four (FTIR) and six (Raman) OH stretching bands, respectively. This number reduces to four (Raman) at non-ambient temperatures either caused by LO-TO splitting, the presence of non-structural OH species or the presence of different lizardite polytypes and/or serpentine polymorphs as well as defects. During dehydroxylation, the evolution of the integrated intensity of the OH bands suggests a transport of hydrogen and oxygen as individual ions/molecule or OH -. A significant change in Raman spectra occurs between 639 and 665 °C with most lizardite peaks disappearing contemporaneously with the appearance of forsterite-related features and new, non-forsterite bands at 183, 350, and 670 cm -1 . A further band appears at 1000 cm -1 at 690 °C. The long stability of Si-O-related bands indicates a delayed decomposition of the tetrahedral sheet with respect to the dehydroxylation of the octahedral sheet. Moreover, evidence for a small change in the ditrigonal distortion angle during heating is given. In general, all appearing non-forsterite-related frequencies are similar to Raman data of talc and this indicates the presence of a talc-like intermediate.

One of the most important mechanisms releasing water in subducting slabs of oceanic crust is connected to the dehydration of serpentinized oceanic rocks. This study reports on a detailed investigation of the transition from chrysotile-an important serpentine mineral-to forsterite through the release of water.The dehydroxylation of natural chrysotile and the subsequent phase change to forsterite was studied by in situ micro-Raman and micro-FTIR spectroscopy in the temperature range of 21 to 871 C. Comparisons were made with previously published data of lizardite-1T. Micro-Raman spectra obtained in the low-frequency (100-1200 cm -1 ) and high-frequency ranges (3500-3800 cm -1 ) were complemented by micro-FTIR measurements between 2500 and 4000 cm -1 to study changes in the chrysotile structure as a function of dehydroxylation progress. In general, room-temperature chrysotile bands lie at higher wavenumbers than equivalent bands of lizardite-1T except of three bands positioned at 301.7, 317.5, and 345.2 cm -1 . Different band assignments of chrysotile and lizardite-1T Raman spectra from literature are compared. The most striking assignments concern the three aforementioned Raman bands and those lying between 620 and 635 cm -1 . The present data support a chrysotile-or at least curved TO layer related origin of the latter. Deconvolution of overlapping OH stretching bands at room temperature revealed the presence of five (FTIR) and four (Raman) bands, respectively. A slight change in the ditrigonal distortion angle during heating and the effects of a radius-dependent dehydroxylation progress can be shown. Furthermore, it was possible to identify a quenchable talclike phase immediately after the onset of the dehydroxylation at 459 C. Main bands of this phase are positioned at 184.7, 359.2, and 669.1 cm -1 and a single OH band at 3677 cm -1, and are thus quite similar to those reported for dehydroxylating lizardite-1T. Their appearance coincides with the formation of forsterite. A maximum in the integral intensity of the talc-like intermediate is reached at 716 C. At higher temperatures, the intermediate phase breaks down and supports the accelerated growth of forsterite. The lack of OH bands with the concomitant appearance of broad chrysotile-related modes in the low-frequency range after heating the sample to 871 C indicates the presence of a heavily disordered phase still resembling chrysotile. However, there are no spectral evidences for further Si-and/ or Mg-rich amorphous phases during the dehydroxylation and no indications for a relationship between the breakdown of the talc-like phase and the growth of enstatite as previously reported in literature.

Abstract:The thermally induced dehydroxylation of lizardite and its phase transformation to forsterite were studied by hightemperature X-ray diffraction (HT-XRD), thermogravimetry (TGA) and Fourier-transformed infrared spectroscopy (FTIR). Primary sample characteristics like chemical composition and crystallographical structure were determined by combined (HR)TEM-EDX, electron-microprobe analyses (EMPA) as well as conventional X-ray diffraction (XRD). Isothermal HT-XRD and non-isothermal TGA data were treated with the classical Avrami-Erofe'ev method and more advanced isoconversional methods in order to obtain kinetic data of a multi-step decomposition reaction. A highly precise activation energy E a versus reaction progress (a) dependency based on non-isothermal TGA data of lizardite is provided and associated mechanisms are discussed. Here, the main focus is on recently published ab initio calculations from the phase transformation of other phyllosilicates. Moreover, the calculated overall apparent activation energy is compared with discrepant data from the literature and discussed. Especially, the usability of overall activation energies of multi-step decomposition reactions is critically discussed. The presented and discussed reaction steps of water formation from different hydroxyl species in lizardite can be used to improve ab initio calculations, especially the pre-selection of reacting hydroxyl species in hydrous sheet like minerals.

a much higher apparent E a characterised by an initial stage of around 290 kJ/mol. Afterwards, the apparent E a comes down to around 250 kJ/mol at α ~ 65 % before rising up to around 400 kJ/mol. The delivered kinetic data have been investigated by the z(α) master plot and generalised time master plot methods in order to discriminate the reaction mechanism. Resulting data verify the multi-step reaction scenarios (reactions governed by more than one rate-determining step) already visible in E a versus α plots.Keywords Serpentine dehydroxylation · Generalised master plot · Chrysotile · Brucite · Thermogravimetry · In situ high-temperature X-ray powder diffraction