To enable the study of improved oil recovery (IOR) from carbonate rock via laboratory experiments at the pore scale, we have developed a novel microfluidic chip containing a 3D packed bed of calcite particles. The utilization of fluorescently labeled water phase enabled visualization up to 1–2 particle layers with confocal laser scanning microscopy. Porosity and residual oil saturation (ROS) in this space are quantified from image stacks in the depth direction (Z). To obtain reliable average ROS values, Z stacks are captured at various XY locations and sampled over several time-steps in the steady state. All image stacks are binarized using Otsu’s method, subsequent to automated corrections for imperfect illumination and Z-drifts of the microscope stage. Low salinity IOR was mimicked using a packed bed that was initially saturated with water and then with mineral oil. Steady state ROS values showed no significant dependence on capillary number (Ca) in the range from 6 × 10^–7 to 2 × 10^–5. In contrast, chemical modification of the pore space via adsorption of water-extracted crude oil components yielded significantly higher ROS values, in agreement with a more oil-wet porous medium. These results indicate a good potential for using packed beds on a chip as an efficient screening tool for the optimization and development of different IOR methods.

为了能够通过孔径实验通过碳酸盐岩改善采收率（IOR）的研究，我们开发了一种新型的微流体芯片，其中包含3D方解石颗粒堆积床。利用荧光标记的水相，可以通过共聚焦激光扫描显微镜观察多达1-2个颗粒层。从深度方向（Z）上的图像堆栈中量化该空间中的孔隙率和残余油饱和度（ROS ）。为了获得可靠的平均ROS值，在稳定状态下在多个XY位置捕获Z堆栈，并在多个时间步上对其进行采样。在自动校正不完美的照明和Z值之后，使用Otsu方法对所有图像堆栈进行二值化处理镜台的漂移。使用填充床模拟低盐度IOR，该填充床最初用水饱和，然后用矿物油饱和。稳态ROS值在6×10 ^–7至2×10 ^–5的范围内，对毛细管数（Ca）的依赖性不明显。相反，与更多油湿性多孔介质相一致，通过吸水提取的原油成分的吸附对孔隙空间进行化学修饰可产生更高的ROS值。这些结果表明使用芯片上的填充床作为用于优化和开发不同IOR方法的有效筛选工具的良好潜力。