A generalized formulation of interfacial tension driven fluid migration with dissolution/precipitation Yasuko Takei a, ⁎, Saswata Hier-Majumder b a Earthquake Research Institute, University of Tokyo, Tokyo, 113-0032, Japan b Department of Geology, University of Maryland, 237 Regents Drive, College Park, MD-20742, USA abstract article info Article history: Received 11 January 2009 Received in revised form 21 August 2009 Accepted 8 September 2009 Available online 13 October 2009 Editor: L. Stixrude Keywords: interfacial tension dissolution/precipitation fluid infiltration two-phase flow We present an extended formulation for the interfacial tension driven melt migration by taking into account dissolution/precipitation and diffusive matter transport through the liquid phase. Our results indicate that the melt migration is caused by two mechanisms. In the first mechanism, a change in melt fraction is accommodated by compaction/decompaction of solid matrix, and in the second mechanism, a change in melt fraction is accommodated by dissolution/precipitation. The latter mechanism is newly introduced in this study. As spatial scale decreases, the dominant mechanism changes from compaction/decompaction to dissolution/precipitation, and when the second mechanism is dominant, evolution of melt fraction is governed by a nonlinear diffusion equation. Therefore, critical scale of this transition is called “diffusion length δ d ” below which melt fraction evolves primarily by diffusion. Diffusion length δ d is usually smaller than the compaction length δ c . Important roles of the new mechanism are discussed on the basis of existing experimental data on melt infiltration, shear-induced melt segregation into melt-rich bands, and rehomogenization of melt in the bands due to static annealing. Melt distribution in the mantle is briefly discussed on the basis of the new model. © 2009 Elsevier B.V. All rights reserved. 1. Introduction Interfacial tension plays an important role in determining the grain scale geometry of melt and aqueous fluid (e.g., von Bargen and Waff, 1986; Watson and Brenan, 1987; Hier-Majumder, 2008). Also, interfacial tension driven fluid migration, or fluid infiltration, can drive melt and aqueous fluid over distances much greater than grain scale (e.g., Watson, 1982). Although interfacial tension is important at smaller scales than buoyancy force, it affects the small scale porosity structure, and hence affects the permeability of the two-phase system. Therefore, interfacial tension can significantly affect the buoyancy- driven flow. Experimental studies on the interfacial tension driven fluid migration constrain physical and chemical properties relevant to this process (e.g., Watson, 1982; Riley et al., 1990; Riley and Kohlstedt, 1991; Nakamura and Watson, 2001). Interfacial tension driven fluid migration was first formulated by Stevenson (1986) within the framework of two-phase flow. It was predicted that fluid distribution tends to be homogenized or localized depending on whether the dihedral angle is smaller or larger, re- spectively, than 60°. The qualitative result of his model explains well the experimental result that liquid phase with small dihedral angles (b 60°) infiltrates into dry rocks, while liquid phase with large di- hedral angles (N 60°) does not (e.g., Watson, 1982; Nakamura and Watson, 2001). Another model for the interfacial tension driven fluid migration was proposed by Riley and Kohlstedt (1991). They focused on the role of compaction/decompaction of the solid matrix, which was omitted in the model of Stevenson (1986). While the rate-controlling factors in the model of Stevenson (1986) are given only by melt transport properties (permeability and melt viscosity), the rate-controlling factors in the model of Riley and Kohlstedt (1991) are given not only by the melt transport properties but also by the viscosity of the solid matrix. Models similar to Riley and Kohlstedt (1991) were also derived by Bercovici and Ricard (2003) and Hier-Majumder et al. (2006). Riley and Kohlstedt (1991) conducted a melt infiltration experiment and reported that the best agreement between the melt migration profile and the simulation generated on the basis of their theory is obtained when permeability has a melt fraction exponent of 1. Because the melt fraction exponent expected for the melt existing as grain-edge tubules is about 2, there still remains a difficulty to understand the result of this analysis consistently with the micro- structural observation of equilibrium melt geometry. The purpose of this study is to extend the existing models of interfacial tension driven fluid migration by taking into account dissolution/precipitation and diffusive matter transport through the liquid phase. Previous studies on reactive, buoyancy-driven flow in porous media demonstrated that dissolution/precipitation causes channeling instability (e.g., Aharonov et al., 1995; Spiegelman et al., Earth and Planetary Science Letters 288 (2009) 138–148 ⁎ Corresponding author. E-mail address: ytakei@eri.u-tokyo.ac.jp (Y. Takei). 0012-821X/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2009.09.016 Contents lists available at ScienceDirect Earth and Planetary Science Letters journal homepage: www.elsevier.com/locate/epsl