A chemical threshold controls nanocrystallization and degassing behaviour in basalt magmas

Abstract

An increasing number of studies are being presented demonstrating that volcanic glasses can be heterogeneous at the nanoscale. These nano-heterogeneities can develop both during viscosity measurements in the laboratory and during magma eruptions. Our multifaceted study identifies here total transition metal oxide content as a crucial compositional factor governing the tendency of basalt melts and glasses towards nanolitization: at both anhydrous and hydrous conditions, an undercooled trachybasalt melt from Mt. Etna readily develops nanocrystals whose formation also hampers viscosity measurements, while a similar but FeO- and TiO2-poorer basalt melt from Stromboli proves far more stable at similar conditions. We therefore outline a procedure to reliably derive pure liquid viscosity without the effect of nanocrystals, additionally discussing how subtle compositional differences may contribute to the different eruptive styles of Mt. Etna and Stromboli.

Viscosity of anhydrous and hydrous peridotite melts

“…and Tpeak; when such deviations were observed, Raman spectra were acquired to document that the sample changed during the measurement (i.e., crystallization and/or degassing). We indeed demonstrated in several recent works that Raman spectroscopy can reveal the precipitation of nanosized Fe-Ti-oxides (as expected during calorimetric measurements of magmatic melts at deep undercooling, where transition metals are poorly soluble and homogeneous crystal nucleation prevails over crystal growth), with sensitivity extending even to the early stages of amorphous phase separation anticipating the occurrence of actual crystals (Di Genova et al, 2020b, 2020a, 2017a, 2017bScarani et al, 2022;Zandona et al, 2022Zandona et al, , 2021Zandona et al, , 2019. Using this approach, we concluded that sample S38F5W1 may be reliably measured for qh ranging between 100 and 30000 K s -1 (see the expected linear dependence on qh in Figure 6); we obtained Tonset = 1055 ± 3 K and Tpeak = 1090 ± 1 K at qh = 1000 K s -1 .…”

“…The experimental procedure applied in this work has been presented, tested and optimized in previous studies (Cassetta et al, 2021;Di Genova et al, 2020b;Scarani et al, 2022;Stabile et al, 2021); its cornerstones are briefly summarized here. We have demonstrated that the combination of calorimetry (DSC) and Brillouin spectroscopy (BLS) enables the modelling of melt viscosity as a function of temperature [η(T)] (Cassetta et al, 2021).…”

“…The melt fragility m was derived spectroscopically from BLS measurements performed on glasses at room temperature (Cassetta et al, 2021) because this approach avoids the need to acquire viscosity data over an extended temperature range, which is challenging or impossible for poor glass-forming liquids (Di Genova et al, 2020b;Kleest et al, 2020;Liebske et al, 2003b;Richet et al, 1996;Scarani et al, 2022). Fragility m was obtained as (Eq.…”

“…Using this approach, we concluded that sample S38F5W1 may be reliably measured for qh ranging between 100 and 30000 K s -1 (see the expected linear dependence on qh in Figure 6); we obtained Tonset = 1055 ± 3 K and Tpeak = 1090 ± 1 K at qh = 1000 K s -1 . For qh ≥ 5,000 K s -1 we limited the matching upscans to temperatures ≤ Tpeak, as the exposure of the melt to higher temperatures induced partial crystallization (Di Genova et al, 2020b;Scarani et al, 2022), inferred from the opacification of the glass and the appearance of the sharp characteristic features (e.g., at ~670 cm -1 ) of Fe-Ti-oxide crystals in its Raman spectrum (Di Genova et al, 2020a, 2020b, 2017b. Tonset values obtained at 1000 K min -1 from all hydrous peridotite glasses, plotted as a function of the water molar content: the higher the water content, the lower Tonset.…”

“…Unfortunately, there are undeniable experimental challenges related to the viscosity measurements of volcanic melts (or their synthetic analogs), namely: (i) their undercooling-driven tendency to crystallize, which typically limits concentric-cylinder low-viscosity (~10 4 -10 -2 Pa s) determinations (e.g., Kolzenburg et al, 2018) to superliquidus temperatures and only mild undercooling; (ii) the prompt degassing of volatile-bearing melts at ambient pressure, which can be only overcome through falling-sphere experiments and the application of confining pressure (e.g., Liebske et al, 2005), with significant technical difficulties; (iii) the limited glass-forming ability of volcanic compositions, since glassy samples are necessary for high-viscosity (10 12 -10 6 Pa s) measurements by micropenetration (e.g., Di Genova et al, 2020b), parallel-plate deformation (e.g., Whittington et al, 2009), beam bending (e.g., Hagy, 1963) or fiber elongation (e.g., Taniguchi, 1992) viscometry. Moreover, even when the required glassy samples can be synthesized, their limited stability during high-temperature measurements has been repeatedly demonstrated by postmortem measurements using Raman spectroscopy and high-resolution imaging (Di Genova et al, 2020b, 2017aKleest et al, 2020;Liebske et al, 2003a;Scarani et al, 2022): nanocrystallization of Fe-Ti-oxides can affect sample homogeneity and alter the composition of the residual melt, preventing derivation of the crystal-free viscosity. As a consequence, currently available viscosity models (Giordano et al, 2008;Hui and Zhang, 2007;Langhammer et al, 2022) have been trained in a relatively restricted volcanological domain of temperature and chemistry, with questionable applicability to compositional extremes such as ultrabasic melts and very H2O-rich matrices.…”