Elsevier

Chemical Geology

Volume 174, Issues 1–3, 1 April 2001, Pages 209-223
Chemical Geology

The viscosity of hydrous phonolites and trachytes

https://doi.org/10.1016/S0009-2541(00)00317-XGet rights and content

Abstract

The 1-atm viscosities of hydrated synthetic iron-free phonolite and trachyte melts are reported between 108.4 and 1013.1 Pa s for water contents between 0 and 5 wt.%. These show a very strong reduction with increasing water content, particularly at low contents. Empirical formulae are derived for the dependence of viscosity on temperature and water content. At magmatic temperatures of about 1100 K and water contents of 5 wt.%, viscosities are about 910 Pa s for the phonolite and 1120 Pa s for the trachyte. Comparisons with data on peralkaline rhyolites show that the effects of dissolved water are similar, even though the viscosities of the different compositions vary. This indicates that the mechanism by which water reduces the viscosities of partially depolymerized aluminosilicate liquids is independent of the alkali/alkaline earth element ratio. In addition, no major change in this pattern is observed whether water dissolves primarily as OH– or molecular H2O. This suggests that a third water-dissolution mechanism may be important in these compositions, in addition to Si–O–Si bond-breaking and dissolution of molecular water.

Introduction

The viscosity of silicate melts is an important parameter which exerts strong controls over many geological processes, including rates of melt extraction, transport mechanisms, and most spectacularly, volcanic eruptive style. In comparison with the other parameters that affect these processes, viscosity is probably both the most variable and the least well known, particularly under conditions relevant to volcanic processes. The main factors controlling the viscosity of volcanic magmas are temperature, pressure, chemical composition, volatile content, and crystal and bubble content (e.g. Dingwell et al., 1993). Of these factors, the effects of temperature, chemical composition and volatile content are the most important to be quantified, since small changes in these factors may affect the viscosity by several orders of magnitude.

There exist very few studies that have directly measured the viscosity of volatile-bearing melts at the low pressures pertinent to volcanism. Measurements near the glass transition temperature are particularly useful, as the effects are greater than at higher temperatures, in addition to which measurement of 1-atm viscosities of hydrous liquids is impossible at higher temperatures due to exsolution. Following the first such study, on a synthetic andesite Lejeune et al., 1994, Richet et al., 1996, recent work has concentrated on haplogranites (Dingwell et al., 1996) and closely related rhyolitic compositions Dingwell et al., 1998a, Dingwell et al., 1998b, Stevenson et al., 1998. These results indicate that the effects of water on less polymerized peralkaline compositions differ from those on the fully polymerized haplogranite system. In this study, we extend the measurements to hydrous phonolite and trachyte liquids, two volcanologically relevant compositions which are already partially depolymerized.

These two compositions were selected in order to compare the effect of water on compositions rich in alkaline earth elements, such as the andesite of Richet et al. (1996), and those rich in alkali elements, such as phonolites. The trachyte is intermediate between the two, and these three samples cover the widest range of alkali/alkaline earth ratios encountered in natural magmas of similar silica content (Fig. 1). An alkaline earth-free end member to the series is provided by the peralkaline rhyolite HPG8N10 of Dingwell et al. (1998a), which has a similar degree of polymerization but a significantly higher silica content. For these peralkaline liquids, Si4+, Ti4+ and Al3+ are assumed to be exclusively in tetrahedral coordination, with some cations (Mg2+, Ca2+, Na+ and K+) charge-balancing aluminium. The surplus (network-modifying) cations then terminate Si–O bonds, coordinating nonbridging oxygens. Each of these four compositions have comparable bulk melt structure, as indicated by the ratios of nonbridging oxygens to tetrahedral cations (NBO/T) which is about 0.2 in each case (Table 1). While titanium is likely to exist in coordination states other than tetrahedral in these liquids, the low quantities (less than 0.7 mol%) mean that the NBO/T values listed in Table 1 would change by about 0.02 at most.

Section snippets

Experimental methods

One problem of low-pressure viscosity measurements is that the correspondingly low temperatures may allow crystallization of the melt during the experiment. This can be exacerbated by the presence of iron. To avoid this problem and also to facilitate spectroscopic investigation of the samples, we made iron-free, trace-element-free synthetic analogues. All iron present in the original compositions, corresponding to about 4 wt.% total iron, was transferred to calcium and magnesium, preserving the

Results

Viscosities of the anhydrous liquids are reported in Table 3, Table 4, Table 5. While viscosity may be apparently Arrhenian over the restricted range of individual techniques, it is clearly non-Arrhenian for both compositions over a large temperature range, emphasizing once more that care must be taken in extrapolating high-temperature data to volcanologically relevant conditions (Fig. 2). At high temperatures the three compositions have similar viscosities, but at low temperatures the

Viscosities at magmatic temperatures

The range of viscosities accessible by the two techniques used here brackets temperatures relevant to magmatic processes. As described above, TVF equations were used to allow interpolation throughout the whole temperature ranges for the measured set of water contents. To predict viscosities at any water content, we further parameterized the data using the anhydrous viscosities as a reference. For both phonolite and trachyte, deviations from these values are well described by equations of the

Acknowledgements

This work was supported by the E.U. TMR network ERBFMRX 960063 “In situ hydrous melts”. We gratefully acknowledge the assistance of Harald Behrens, Jasper Berndt, Susanne Ohlhorst, Frank Schulze, Nathalie Tamic, Max Wilke and Tony Withers with hydration experiments, and stimulating discussions with M. Ali Bouhifd and Anne Sipp. We thank Claudia Romano and an anonymous reviewer for detailed comments, and Don Dingwell for editorial handling.

References (41)

  • P Richet et al.

    Water and the viscosity of andesite melts

    Chem. Geol.

    (1996)
  • A.C Withers et al.

    Reconciliation of experimental results on H2O speciation in rhyolitic glass using in-situ and quenching techniques

    Earth Planet. Sci. Lett.

    (1999)
  • G Adam et al.

    On the temperature dependence of relaxation phenomena in glass-forming liquids

    J. Chem. Phys.

    (1965)
  • C.A Angell

    Strong and fragile liquids

  • R Böhmer et al.

    Correlations of the nonexponentiality and state dependence of mechanical relaxations with bond connectivity in Ge–As–Se supercooled liquids

    Phys. Rev. B

    (1992)
  • H Behrens

    Determination of water solubilities in high-viscosity melts: an experimental study on NaAlSi3O8 and KAlSi3O8 melts

    Eur. J. Mineral.

    (1995)
  • Y Bottinga et al.

    The viscosity of magmatic liquids: a model for calculation

    Am. J. Sci.

    (1972)
  • D.B Dingwell et al.

    Relaxation in silicate melts

    Eur. J. Mineral.

    (1990)
  • D.B Dingwell et al.

    Magma rheology

  • D.B Dingwell et al.

    The effect of water on the viscosity of a haplogranitic melt under PTX conditions relevant to silicic volcanism

    Contrib. Mineral. Petrol.

    (1996)
  • Cited by (128)

    View all citing articles on Scopus
    View full text