Figure 1. Experimental setup: a fractured rock sample, saturated with aqueous saline solution, mounted inside the polytetrafluoroethylene column.
WATER RESOURCES RESEARCH, VOL. 38, NO. 2, 10.1029/2001WR000246, 2002
[6] Most technical details of the experimental setup and application of the NMRI technique are discussed in detail by Dijk et al. [1999]. Sections 2.1 and 2.2 provide a brief summary of the major issues, and section 2.3 describes the mineralogical analysis of the source rock.
|
|
[7] A schematic of the experimental setup is presented in Figure 1. Relatively planar, rough fractures were induced along the width of natural halite samples. Each sample has a height and width of ~18 mm and a length of ~36 mm. The mean aperture was set at ~2 mm, which is convenient for NMRI measurements. The fractured rock samples were mounted in cylindrical polytetrafluoroethylene (PTFE) columns so that the only flow passage was the fracture. Deionized water, fully saturated with the rock matrix components (sodium chloride) to prevent fracture wall dissolution, flowed through the columns driven by a peristaltic pump. The mean velocity of the saline aqueous solution in the fractures was set at ~1 mm s-1, representative of groundwater flow in many natural rock fracture systems. Pulses of undersaturated saline solution in between NMRI measurements were applied to induce dissolution of the rock fracture walls. The inflowing solute content of these pulses was used to adjust the dissolution rates. Two sealed polyvinylchloride (PVC) tubes with inner diameters of ~1.6 mm containing the fully saturated saline solution were mounted in each column to serve as a zero-velocity reference.
[8] Three different halite fracture systems were investigated. They vary in fracture orientation (horizontal and vertical), temporal mean of the applied volumetric flow rate 
Q
t, and solute content of the inflowing pulse cin/csat (cin/csat = 0 for pure water and cin/csat 
0.5 for ~50% saturated saline solution). As a result, the time intervals between NMRI measurements 
t, the number of images in the time series itot, and therefore the total dissolution times ttot = itott differ. The choice of t was such that the change in fracture walls was detectable yet minimal. The dissolution experiments ceased when the saline solution flowing through the rock fractures came into contact with the sealing material surrounding the rock samples. The orientation of the principal flow direction is horizontal for all three fracture systems. The experimental parameters for the fracture systems are given in Table 1.
[9] The overall dimensionless Damköhler number Da´, which is a measure of dissolution versus flow rate, is used to qualitatively describe the overall dissolution rates. Note that a generic definition of the Damköhler number is given by, for example, Da = kncn-1/(v/L), where kn is the kinetic reaction coefficient for an nth-order reaction, c is the solute concentration, v is the fluid flow velocity, and L is the characteristic length scale. High Da´ indicates the inflow of pure water (cin/csat = 0), while low Da´ indicates the inflow of ~50% saturated saline solution (cin/csat 0.5).
[10] The samples were inserted in a horizontal, cylindrical proton radio frequency probe inside a horizontal, cylindrical static electromagnet bore operating at a field strength of 4.7 T. The classical NMRI phase imaging approach, i.e., a three-dimensional (3-D) flow-encoded spin echo pulse sequence [e.g., Moran, 1982], was applied. This sequence allows the measurement of water densities and selected flow velocity components [Dijk et al., 1999]. By combining velocity components for three perpendicular directions, 3-D velocity vectors can be obtained. In this study, 3-D distributions, i.e., a one-dimensional (1-D) profile in the y direction at each location (x, z), of the 1-D velocity component in the principal flow (z) direction were investigated.
[11] The pulse sequence was calibrated extensively [Dijk et al., 1999]. The reproducibility of the measurements was confirmed by analyzing rapid successions of two-dimensional (2-D) images with relatively short acquisition times.
[12] The images have a matrix of 128 × 128 × 128. The spatial field of view and thus the spatial resolution vary slightly between samples. Only a subset of the matrix that contains the actual fracture volume was analyzed. The NMRI parameters for the three fracture systems are given in Table 2. The resolution of the velocities is ~0.025 mm s-1. A relatively short acquisition time of ~1 hour 12 min per image was made possible by introducing a nuclear magnetic relaxation agent, i.e., ~5 mM NiCl2(aq), to the aqueous solution. To measure the dissolution processes with timescales of minutes using NMRI with acquisition times of the order of an hour, the following procedure was employed: (1) Pure water or ~50% saturated saline solution was pumped through the fracture during time intervals t (Table 1) to invoke dissolution of the rock matrix. (2) Fully saturated saline solution was pumped through the fracture during the NMRI acquisition time to inhibit dissolution during the actual measurements. This procedure is then repeated. Thus the measured flow patterns do not refer to those during the actual dissolution process.
2.3. Mineralogical Analysis of Source Rock
[13] The rock samples were obtained from Mount Sedom, which is a salt diapir located to the west of the Dead Sea in Israel. The salt originates from marine ingression at Neogene times [Zak, 1967] that precipitated the Sedom formation. This formation is composed mainly of halite (NaCl, ~96% of the whole rock) but also contains minor amounts of calcium sulfates (anhydrite CaSO4 and gypsum CaSO4 · 2H2O), carbonates (e.g., dolomite CaMg(CO3)2 and some calcite CaCO3), potassium and magnesium salts (e.g., sylvite KCl and carnallite KMgCl3 · 6H2O), and quartz and clays [Zak, 1967]. These minor minerals may occur as discrete thin layers or dispersed within the halite [Zak, 1967]. Chemical impurities can exist in various forms, for example, as embedded or absorbed macroscopic phases, microscopic fluid inclusions, and ionic crystal substitutions.
[14] The dissolution and precipitation rates depend on, for example, the mineralogical composition of the rock matrix and the chemical composition of the aqueous solution, the rock-solution interface area, the thickness of the diffusional layer adjacent to the interface, the occurrence of surface coatings, the temperature, and the flow conditions. The dissolution rates and saturation contents (solubilities) of different constituents in water vary considerably. For example, they equal ~0.02 × 10-2 kg m-2 s-1 [Alkattan et al., 1997] and ~4 × 102 kg m-3 [Windholz, 1976], respectively, for halite and a factor of ~15 [Raines and Dewers, 1997] and ~0.005 [Ohmoto et al., 1991] as much for gypsum at ~25°C. Thus, while larger amounts of halite than gypsum can be dissolved in water, gypsum dissolves more rapidly than halite.
[15] The content of halite, calcium sulfate, and other minor elements of the water-soluble phase of the halite samples was determined by chemical analysis. The elements Na, Ca, SO4, K, Mg, and Sr were analyzed by inductively coupled plasma (ICP), and Cl was analyzed by titration. The presence of halite, anhydrite, calcium carbonate, and quartz in these samples was also examined by mineralogical analysis using X-ray diffraction (XRD). The chemical and mineralogical composition of the water-insoluble phase (e.g., clay) was not determined. The extent of the variation of the mineralogical composition was determined by examining the source rock from which the samples were cut. Several distinct types of halite were classified on the basis of the visual observation of color and texture. Chemical and mineralogical analysis was performed on samples of these rock types. In addition, chemical analysis was performed on halite from two parts of the residual halite samples after completion of each of the experiments.
Citation: Measurement and analysis of dissolution patterns in rock fractures, Water Resour. Res., 38(2), 10.1029/2001WR000246, 2002.