An experimental and numerical study of surface tension-driven melt flow

R. A. Parsons, F. Nimmo, J. W. Hustoft, B. K. Holtzman, D. L. Kohlstedt

Research output: Contribution to journalArticlepeer-review

16 Scopus citations


To determine the role of surface tension-driven melt migration in planetary bodies, we investigated the effect of static annealing on the evolution of melt-rich bands in partially molten samples. In shear deformation experiments, deviatoric stress causes melt to segregate; when the stress is removed, surface tension causes the melt to relax back to a homogeneous distribution. Samples composed of 76 vol.% olivine + 20 vol.% chromite + 4 vol.% MORB were deformed to shear strains of ~ 3.5 at 1523 K, 300 MPa and shear stresses of 20 to 55 MPa. After deformation, the samples were statically annealed for 0, 10, or 100 h. During annealing, melt transport driven by surface tension occurs, but takes place much more slowly than flow driven by deviatoric stress. Finite difference numerical simulations were performed of surface tension-driven melt flow resisted by viscous deformation of the olivine matrix. These models best reproduce the distribution of melt in the annealed samples when the solid viscosity ηs = 1.7 ± 0.5 × 1012 Pa s with n = 2.4 ± 0.3 and b = 9000 ± 1900 in the expression for permeability κ = φ{symbol}nd2 / b where d is grain size. The large value of b compared with estimates from geometrical models is probably due to clogging of the melt tubes by the secondary solid phase (chromite). Redistribution of melt by surface tension is likely to be the dominant process in small (~ 10 km radius) planetesimals in the absence of convection or impact-induced deformation. However, this redistribution process is sufficiently slow that large bodies of localized melt (magma chambers) are likely to develop.

Original languageEnglish (US)
Pages (from-to)548-557
Number of pages10
JournalEarth and Planetary Science Letters
Issue number3-4
StatePublished - Mar 30 2008

Bibliographical note

Funding Information:
We would like to acknowledge the National Science Foundation (OCE-0648020) and NASA (PG&G NNX07AP68G) for funding the experimental portion of this research project. Funding for the modeling portion was provided by NASA (MFR NNG06GH13G).


  • melt migration
  • permeability
  • planetesimals
  • surface tension


Dive into the research topics of 'An experimental and numerical study of surface tension-driven melt flow'. Together they form a unique fingerprint.

Cite this