HR: 09:15h
AN: V41I-06 INVITED [Abstracts]
TI: A Non-Arrhenian Viscosity Model for Natural Silicate Melts with Applications to Volcanology
AU: * Russell, J K
EM: krussell@eos.ubc.ca
AF: Volcanology & Petrology Laboratory, Earth & Ocean Sciences
University of British Columbia, Vancouver, BC V6T 1Z4
Canada
AU: Giordano, D
EM: dgiordano@eos.ubc.ca
AF: Volcanology & Petrology Laboratory, Earth & Ocean Sciences
University of British Columbia, Vancouver, BC V6T 1Z4
Canada
AU: Dingwell, D B
EM: dingwell@lmu.de
AF: Earth & Environment, University of Munich, Munich, Bav 80333
Germany
AB:
Silicate melt viscosity is the most important physical property in volcanic systems. It governs styles and rates of flow,
velocity distributions in flowing magma, rates of vesiculation, and, ultimately, sets limits on coherent(vs. fragmented or
disrupted) flow. The prediction of melt viscosity over the range of conditions found on terrestrial planets remains a
challenge. However, the extraordinary increase in number and quality of published measurements of melt viscosity suggests the
possibility of new models. Here we review the attributes of previous models for silicate melt viscosity and, then, present a
new predictive model natural silicate melts. The importance of silicate melt viscosity was recognized early [1] and
culminated in 2 models for predicting silicate melt viscosity [2,3]. These models used an Arrhenian T-dependence; they were
limited by a limited experimental database dominated by high-T measurements. Subsequent models have aimed to: i) extend the
compositional range of Arrhenian T-dependent models [4,5]; ii) to develop non-Arrhenian models for limited ranges of
composition [6,7,8], iii) to develop new strategies for modelling the composition and T-dependence of viscosity [9,10,11],
and, finally, to create chemical models for the non-Arrhenian T-dependence of natural melts [12]. We present a multicomponent
model for the compositional and T dependence of silicate melt viscosity based on data spanning a wide range of anhydrous
melt compositions. The experimental data include micropenetration and concentric cylinder viscometry measurements covering a
viscosity range of 10-1 to 1012 Pa s and a T-range from 700 to 1650°C. These published data provide a high-
quality database comprising ~ 800 experimental data on 44 well-characterized melt compositions. Our model uses the
Adam-Gibbs equation to capture T-dependence:
log η = A + B/[T · log (T/C)]
where A, B, and C are adjustable parameters that vary for different melt compositions. We assume that all silicate melts converge to a common, but
unknown, high-T limit (e.g., A) and that all compositional dependence is accommodated for by B and C. We adopt a linear compositional dependence for B and C:
B = σi=1..n [xi βi]
C = σi=1..n [xi γi]
where xi's are the mole fractions of oxide components (n=8) and βi and γi are
adjustable parameters. The model, therefore, comprises 2 · n+1 adjustable parameters which are optimized for against
the experimental database including a common value of A and compositional coefficeints for B and C. The new model reproduces
the original database to within experimental uncertainty and can predict the viscosity of silicate melts across the full
range of conditions found in Nature. References Cited: [1] Friedman et al., 1963. J Geophys Res 68, 6523-6535. [2] Bottinga Y
& Weill D 1972. Am J Sci 272, 438- 475. [3] Shaw HR 1972. Am J Sci 272, 438- 475. [4] Persikov ES 1991. Adv Phys Geochem 9,
1-40. [5] Prusevich AA 1988. Geol Geofiz 29, 67-69. [6] Baker DR 1996. Am Min 81, 126-134. [7] Hess KU & Dingwell DB 1996.
Am Min 81, 1297- 1300. [8] Zhang, et al. 2003. Am min 88, 1741- 1752. [9] Russell et al. 2002. Eur J Min 14, 417-428. [10]
Russell et al. 2003. Am Min 8, 1390- 1394. [11] Russell JK & Giordano D In Press. Geochim Cosmochim Acta. [12] Giordano D &
Dingwell DB 2003. Earth Planet. Sci. Lett. 208, 337-349.
UR: http://www.eos.ubc.ca/~krussell/
DE: 8410 Geochemical modeling (1009, 3610)
DE: 8411 Thermodynamics (0766, 1011, 3611)
DE: 8445 Experimental volcanism
SC: Volcanology, Geochemistry, Petrology [V]
MN: Fall Meeting 2005