The impact of pore structure and surface roughness on capillary trapping for 2‐D and 3‐D porous media: Comparison with percolation theory

Abstract

We study the impact of pore structure and surface roughness on capillary trapping of nonwetting gas phase during imbibition with water for capillary numbers between 10−7 and 5 × 10−5, within glass beads, natural sands, glass beads monolayers, and 2‐D micromodels. The materials exhibit different roughness of the pore‐solid interface. We found that glass beads and natural sands, which exhibit nearly the same grain size distribution, pore size distribution, and connectivity, showed a significant difference of the trapped gas phase of about 15%. This difference can be explained by the microstructure of the pore‐solid interface. Based on the visualization of the trapping dynamics within glass beads monolayers and 2‐D micromodels, we could show that bypass trapping controls the trapping process in glass beads monolayers, while snap‐off trapping controls the trapping process in 2‐D micromodels. We conclude that these different trapping processes are the reason for the different trapping efficiency, when comparing glass beads packs with natural sand packs. Moreover, for small capillary numbers of 10−6, we found that the cluster size distribution of trapped gas clusters of all 2‐D and 3‐D porous media can be described by a universal power law behavior predicted from percolation theory. This cannot be expected a priori for 2‐D porous media, because bicontinuity of the two bulk phases is violated. Obviously, bicontinuity holds for the thin‐film water phase and the bulk gas phase. The snap‐off trapping process leads to ordinary bond percolation in front of the advancing bulk water phase and is the reason for the observed universal power law behavior in 2‐D micromodels with rough surfaces.

Microfluidic Model Porous Media: Fabrication and Applications

“…[21c,86] In drainage, the wetting defending phase is mainly trapped by the bypass mechanism; in imbibition, however, the nonwetting defending phase can be trapped by either bypass or snap‐off, or a combination of both mechanisms, depending on the local topology of the pore structure, the surface roughness, and the Ca. [21a,c,87] In general, surface roughness could enhance the snap‐off trapping; bypass is more likely to be dominant at smaller Ca. [21c,88] Notably, using confocal microscopy on 3D index‐matching micromodels one can directly visualize and characterize the full 3D structure of the trapped ganglia, as shown in Figure c.…”

New Capillary Number Definition for Micromodels: The Impact of Pore Microstructure

“…It is the dependence of the new N ca on pore geometry as described in the bracketed term in equation that shifts the separate CDC's in Figure to converge to Figure . The large values of G (Table ) explain the reason why the value of N ca needed to mobilize the nonwetting phase in micromodels using the conventional definition is much less than that in rock (Buchgraber et al, ; Geistlinger et al, ). The conventional definition is misleading, when using micromodel results to interpret displacements in porous rock.…”

“…Thus, one is especially likely to misinterpret mobilization in micromodels, based on the value of N ca defined in equation . For instance, capillary numbers around 10 −5 as defined by equation correspond to considerable mobilization in micromodels, for example, S nw of about 10–15% (Buchgraber et al, ; Geistlinger et al, ). This greatly overestimates mobilization implied by similar capillary numbers in geological porous media.…”

Wettability Effects on Primary Drainage Mechanisms and NAPL Distribution: A Pore‐Scale Study

“…It should be noted that the capillary numbers for WW and IW are outside the range of applicability for percolation theory (Ca < 10 −6 ) (Geistlinger et al, ). As a result, quantitative predictions from percolation theory (e.g., power‐law predictions of residual ganglia size) are not applicable to this study.…”