Flow-only and coupled flow-diffusion experiments at 95 °C and 100 bars p_CO2 carried out in micro-capillary tubes packed with forsterite mineral grains were used to constrain a coupled classical heterogeneous nucleation and crystal growth reactive transport model describing the formation of secondary magnesite. The study made use of a novel experimental setup in which one capillary tube for flow is connected via a three-way tee to a perpendicular capillary tube sealed at the distal end in which only molecular diffusion is allowed to occur—an experimental analogue of a single fracture-rock matrix system. While the high flux of CO₂ bearing fluids and their low pH did not result in the formation of secondary carbonates in the flow-dominated channels, as much as 2.7% magnesite formed in a diffusion-controlled capillary tube sample after 300 hours of reaction. The precipitation of secondary magnesite was not uniformly distributed, however, but showed a distinctive peak shaped pattern along the sample reacted as quantified by RAMAN spectroscopic analysis. About 50% of the total magnesite precipitation formed within a narrow interval of a few millimeters close to the middle of the 3 cm diffusion sample. To simulate the behavior of the diffusion–reaction column, and in particular the pronounced mm scale peak at approximately 1.8 cm, an interfacial free energy of approximately 70 mJ·m⁻² combined with a relatively high crystal growth rate for secondary magnesite was required. In agreement with the observations based on RAMAN spectroscopy, the simulations suggest that more than 50% of Mg²⁺ dissolved from the primary forsterite precipitated as secondary magnesite. The magnesite precipitation at the central peak position in the diffusion sample showed such rapid growth that it created a local minimum in pore fluids Mg²⁺ concentration close to the peak, creating a “nucleation shadow” that focused continued growth there as nearby regions had no nucleation seeds available for crystal growth. Continued crystal growth on the initial magnesite band acted to further increase the reactive surface area at this point, thus enhancing the spatially focused crystal growth rate and creating a positive feedback leading to pattern formation. This highlights the potentially critical role of an initial nucleation event in controlling mineral precipitation patterns in subsurface porous media, patterns that may determine how the pore structure subsequently evolves physically and chemically over time due to reactive flow and transport.

在 95 °C 和 100 bar p _{CO2 条件}下，在填充镁橄榄石矿物颗粒的微毛细管中进行的纯流动和耦合流动扩散实验被用来约束耦合经典异相成核和晶体生长反应输运模型，该模型描述了二次形成菱镁矿。该研究利用了一种新颖的实验装置，其中一个用于流动的毛细管通过三通三通连接到一个垂直的毛细管，该毛细管密封在远端，其中只允许发生分子扩散——一个单一的实验类似物。裂缝-岩石基质系统。而高通量的 CO ₂含流体及其低 pH 值不会导致在流动主导通道中形成次生碳酸盐，在反应 300 小时后，在扩散控制的毛细管样品中形成多达 2.7% 的菱镁矿。然而，次级菱镁矿的沉淀并不是均匀分布的，而是通过拉曼光谱分析量化反应的样品显示出独特的峰形图案。大约 50% 的菱镁矿沉淀形成在靠近 3 cm 扩散样品中间几毫米的狭窄间隔内。为了模拟扩散反应柱的行为，特别是在大约 1.8 cm 处的明显 mm 标度峰，界面自由能大约为 70 mJ·m ^-2再加上次生菱镁矿需要相对较高的晶体生长速率。与基于拉曼光谱的观察结果一致，模拟表明从原生镁橄榄石中溶解的 Mg ²⁺超过 50%作为次生菱镁矿沉淀。扩散样品中心峰位置的菱镁矿沉淀显示出如此快速的增长，以至于它在孔隙流体 Mg ^{2+ 中}产生了局部最小值浓度接近峰值，形成“成核阴影”，集中在那里继续生长，因为附近区域没有可用于晶体生长的成核种子。在初始菱镁矿带上的持续晶体生长进一步增加了此时的反应表面积，从而提高了空间聚焦的晶体生长速率并产生了导致图案形成的正反馈。这突出了初始成核事件在控制地下多孔介质中矿物沉淀模式方面的潜在关键作用，这些模式可能决定孔隙结构随后如何因反应流动和传输而随时间物理和化学演变。