Research ArticleGEOLOGY

In situ observation of nanolite growth in volcanic melt: A driving force for explosive eruptions

See allHide authors and affiliations

Science Advances  23 Sep 2020:
Vol. 6, no. 39, eabb0413
DOI: 10.1126/sciadv.abb0413
  • Fig. 1 Nanolites in natural and experimental samples.

    The Raman spectra in (A) represent ~1000-nm areas of the nanolite bearing glass seen in the STEM images (B to E), as well other nanolite-free areas or samples that have been experimentally melted and quenched. The broad silicate bands at ~500 and ~1000 cm−1 are characteristic of nanolite-free glass, while the sharp peak at ~670 cm−1 has been attributed to FeO-bearing nanolites (33). Samples in (B) phonotephrite from Colli Albani, (C) trachy-andesite from Tambora, and (D) a hydrous experimental Mt. Etna basalt sample all contain 4 to 5 volume % of nanolites around 10 to 20 nm in size. In (E), a natural sample collected from the 122 BCE eruption of Mt. Etna shows two sets of nanolites; one solid (on the left adjacent to a plagioclase crystal) while the others appear to be agglomerates, enlarged in (F). The latter are more typical for the 122 BCE eruption. These agglomerates represent 13 to 20 volume % of the imaged samples, although it should be noted that the individual finer 5-nm aggregated particles in (F) (corrected for the intergranular melt) would represent <5 volume %. STEM wafers are ~100-nm thick. a.u., arbitrary units.

  • Fig. 2 In situ synchrotron XRD patterns of molten Mt. Etna basalt during slow and fast cooling.

    In (A), the Mt. Etna basalt was cooled at 1°C min−1 in air starting from pure melt at 1300°C down to 1083°C. Results show the microlite crystallization for 60-s acquisitions at the temperatures indicated. The sharp Bragg peaks at 1224° and 1208°C represent a “spinel-structured” phase. Plagioclase peaks appear from 1198°C. The sharp peak observed at 2.9 Å−1 in the melt at 1224°C is the 2 2 0 Bragg diffraction peak of crystalline Pt (black tick positions), arising from incident x-ray beam impinging on the Pt-Rh10% heating wire as it contracts. The spectra in (B) are for a much faster cooling of Mt. Etna basalt from 1475°C to room temperature (in black) and then to each target temperature in the legend, in a pure Ar atmosphere. The asterisk (*) indicates the small-angle “nanolite peak,” which increases systematically with deeper undercooling. Note the absence of Bragg peaks. All data are displaced vertically. The background curve shows the XRD measurement performed with an empty cell and results from the Kapton window enclosing the wire furnace.

  • Fig. 3 Nanolite growth with time.

    The spectra collected simultaneously at the (A) SAXS and (B) WAXS detectors show nanolite crystallization of Mt. Etna basalt experiment at a constant temperature of 950°C after fast cooling from 1600°C. The SAXS-WAXS scattering curves were acquired at a sampling interval of 0.5 s from 9 to 95 s of dwell time at 950°C. The color changes from hot (red) to cold (blue) with time. (A) The development of a SAXS signal with time. The lower red SAXS spectra were collected in the first 15 s. (B) WAXS scattering curves show initially no diffraction peaks (hot colors). Afterward, two peaks appear at 2.485 and 4.255 Å−1 after 15-s dwell time. The peak intensity increases slowly between 15 and 23 s. Last, the pattern shows a fast increase of multiple peaks at 2.128, 2.316, 2.47, 2.48, 2.51, 2.554, 2.806, 3.004, 3.173, 3.634, 4.135, and 4.144 Å−1. In general, peak intensity increases with time, whereas the peak width decreases. Numbers show the position of the peak. In (C), the evolution of the nanolite radius with time for the two populations of nanolites is derived by modeling the SAXS patterns collected during nanolite crystallization (see the Supplementary Materials for details).

  • Fig. 4 The nanolite effect on viscosity.

    (A) Illustrates the measured relative viscosity (ηr = ηsuspensionoil) of analog magma made of silicon oil and spherical SiO2 nanoparticles of ~15-nm diameter at different shear rates. In (B), the relative viscosity from (A) is used to calculate the expected viscosity of nanolite-bearing melt with an initial viscosity of 200 Pa s (ηprojected = ηr × 200 Pa s), equivalent to Mt. Etna basalt at preeruptive conditions (temperature and water content). We also show the equivalent curve for microlites, using the relative viscosity of Mader et al. (26) for SiO2 microspheres, also in silicon oil at a shear rate of 1 s−1. It is clear that <10 volume % of nanoparticles is equivalent to >60 volume % of microparticles and can raise the viscosity to ~106 Pa s required for magma fragmentation (41). The traditionally calculated viscosity for <10 volume % microparticles according to the Krieger and Dougherty (75) model, as described by Mader et al. (26), would barely show above the x axis in these figures.

  • Fig. 5 Nanolite and pumice formation.

    (A) Simultaneous thermal analysis (STA) showing the heat flow (red) and weight loss (blue) during heating of a nanolite- and bubble-free Mt. Etna hydrous (H2O = 1.48 wt %) glass (experiment 1). The different thermal events are reported together with the onset of degassing that led to pumice formation in (B). (B) Picture of the recovered sample after STA measurement in (A). Note the considerable increase in the sample volume due to the development of a very high porosity. Before the measurement, the sample was a doubly polished dense nanolite- and bubble-free glass measuring 3 mm by 3 mm by 2 mm. Photo credit: Danilo Di Genova. (C) High-angle annular dark-field (HAADF)–STEM image of the pumice in (B). The rectangle shows the investigated area in (D) and (E) with STEM-EDS that suggests nanolites are Fe rich (E). (F) HAADF-STEM and (G) STEM-EDS images of the STA experiment 2 (stopped at 620°C). Note that the number density of nanolites is lower and their size is smaller than in (C).

Supplementary Materials

  • Supplementary Materials

    In situ observation of nanolite growth in volcanic melt: A driving force for explosive eruptions

    Danilo Di Genova, Richard A. Brooker, Heidy M. Mader, James W. E. Drewitt, Alessandro Longo, Joachim Deubener, Daniel R. Neuville, Sara Fanara, Olga Shebanova, Simone Anzellini, Fabio Arzilli, Emily C. Bamber, Louis Hennet, Giuseppe La Spina, Nobuyoshi Miyajima

    Download Supplement

    This PDF file includes:

    • Sample descriptions for images in Figure 1
    • In situ XRD measurements
    • In situ SAXS-WAXS measurements
    • Modelling of nanoparticle agglomeration
    • Simultaneous Thermal Analysis (STA)
    • The time dependence of the nanolite viscosity effect
    • Appendix: Relationships used in agglomeration modelling
    • Figs. S1 to S10
    • References

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article