Processes affecting the efficiency of limestone in passive treatments for AMD: Column experiments

• Column experiments were run to study the behavior of passive treatment systems.
• X-ray microtomography was used to monitor the porosity variation in the column.
• Reactive transport modeling was used to simulate the passive treatment behavior.
• Initial input concentration and secondary phases limit the system efficiency.

Experiments using columns packed with calcite grains (size 1–2 mm) were carried out to study the behavior of passive treatment systems designed to remediate contaminated water from acid mine drainage (AMD). Two types of synthetic acidic solutions (H2SO4) were injected in the columns: (1) iron-rich with Fe(III) at pH 2, and (2) aluminum-rich at pH 2 and 3. Fluid flow was constant during the experiments with Darcy velocities ranging from 6 × 10−4 to 1 × 10−3 L m−2 s−1. The columns worked efficiently removing aqueous iron and aluminum as long as calcite dissolved and buffered the solution pH (increasing pH and promoting the precipitation of Fe(III)- or Al-oxyhydroxides). However, Ca released from dissolving calcite, along with the sulfate in solution, led to formation of gypsum coatings on the calcite grain surfaces which eventually prevented calcite dissolution. This passivation process limited the efficiency of the columns. Larger input sulfate concentrations or higher pH led to shorter passivation times. Characterization of the pore structure and composition by X-ray microtomography (mCT) and X-ray microdiffraction (mXRD) showed the precipitation of gypsum coatings on the calcite grains and secondary oxyhydroxides between the grains. This secondary mineral precipitation favored the formation of preferential flow paths, isolating regions of non-reacted limestone. An improved experimental design (mixing limestone grains and glass beads) minimized the formation of these preferential flow paths. Experimental results have been modeled with the CrunchFlow reactive transport code. Fitting of the results required a decrease in the reactive surface area of calcite, which is consistent with the passivation process.