Microfluidic observation of the onset of reactive‐infitration instability in an analog fracture

Abstract

Reactive‐infiltration instability plays an important role in many geophysical problems yet theoretical models have rarely been validated experimentally. We study the dissolution of an analog fracture in a simple microfluidic setup, with a gypsum block inserted in between two polycarbonate plates. By changing the flow rate and the distance between the plates, we are able to scan a relatively wide range of Péclet and Damkhöhler numbers, characterizing the relative magnitude of advection, diffusion, and reaction in the system. We quantify the characteristic initial wavelengths of the perturbed fronts during the onset of instability. The results agree well with theoretical predictions based on linear stability analysis, thus experimentally validating current reactive‐infiltration instability theory and opening new opportunities for experimental assessment of mineral reactivity.

The Development of Wormholes in Laboratory‐Scale Fractures: Perspectives From Three‐Dimensional Simulations

“…In Garcia-Rios et al (2015) we observe all four phases: face dissolution (Figure 6a), compact wormholes (Figures 7a and 7b), a necked wormhole (Figure 7d, XCMT images), and uniform dissolution (Figure 6d). The microfluidic experiments reported by Osselin et al (2016) show an instability in the dissolution front (Figure 2a of their paper). Fingers are being formed, but the shapes of the fully developed wormholes were not reported.…”

“…accounts for experiments (Detwiler et al, 2003;Osselin et al, 2016) with only a single reactive surface (n = 1) as well as the remaining experiments (Elkhoury et al, 2013;Garcia-Rios et al, 2015) and simulations (n = 2). This definition of the Damköhler number accounts for varying length to aperture ratios (L∕h 0 ) and for the hindrance arising from reactant diffusion across the aperture (16).…”

“…If the rock matrix is somewhat porous, reactant ions will penetrate to a depth of the order of l pm = √ D ∕ , where is the matrix porosity and is the dissolution rate of the matrix (in s −1 ). In the calcite and gypsum dissolution experiments we are comparing with (Elkhoury et al, 2013;Garcia-Rios et al, 2015;Osselin et al, 2016), l pm is in the range from 10 to 100 μm; in Detwiler et al (2003) the KDP single crystal can be viewed as impermeable. In each case, the consumption of reactant is localized to a thin region near the fracture-matrix interface and can be characterized by a surface reaction rate constant k ∼ l pm , which is the experimentally measured quantity (Colombani, 2008;Plummer & Wigley, 1976;Pokrovsky et al, 2009).…”

“…However, natural fractures have much larger aspect ratios (10 4 -10 7 ), and by using an unstructured mesh of hexahedral elements aspect ratios up to 10 5 have been simulated (Starchenko et al, 2016). Such simulations can form a bridge between laboratory-scale experiments on the scale of centimeters (Detwiler & Rajaram, 2007;Elkhoury et al, 2013;Garcia-Rios et al, 2015;Gouze et al, 2003;Noiriel et al, 2013;Osselin et al, 2016) and the field scale. However, in this paper we focus on fractures at the laboratory scale and investigate the phase diagram that can be constructed from the different dissolution patterns observed as a function of Péclet and Damköhler numbers.…”

Pore Fluid Extraction by Reactive Solitary Waves in 3‐D

“…This reactive‐infiltration mechanism (see review in De Wit ()) was used to explain patterns observed in a variety of geological processes including oil recovery (Daccord et al, ), karst formation (Ortoleva et al, ; Szymczak, ), fluid ascent in the crust (Koulakov et al, ), melt migration (Jordan & Hesse, ; Hewitt, ; Hesse et al, ; Liang et al, ; Keller & Katz, ; Rudge et al, ; Schiemenz et al, ; Weatherley & Katz, ; Spiegelman et al, ), and inner core growth (Sumita et al, ). It also led to numerous studies aiming at reproducing instabilities in the laboratory (Osselin et al, ; Pec et al, ; Vardoulakis et al, ) and developing the theoretical framework to model this process (Szymczak & Ladd, , ; Wangen, ).…”