2.3.2. Capillary Seals

The second phase of GRI funding addressed the origin of the seals that bound the pressure compartments. How capillary forces impede the migration of hydrocarbons was understood: Water typically wets silicate surfaces and surface tension pulls water strongly into a shale. Oil, on the other hand, must be compressed to move from the large pores of a sand into the finer openings of a shale, and, unless the required entry pressure is exceeded, it will pond below the shale [50]. How a capillary seal of this nature could impede the movement of both brine and water was not clear, and Bradley and Powley [51] were careful to distinguish the compartment "pressure seal" that impedes both brine and hydrocarbons, from a capillary seal which impedes just hydrocarbons.

As illustrated in Figure 7, experiments carried out by Jennifer Shosa [52] show that, under the right circumstances, capillary seals can impede both brine and hydrocarbons. The right circumstance is that sufficient quantities of both wetting and non-wetting fluid phases be present in the pore space. Shosa passed CO2-laden water through a 0.5 m long 12.7 mm inner diameter sand-packed tube that contained 2 to 8 fine-grained impeding layers. The impeding layers consisted of 13 to 15 mm thick intervals of 2 μm diameter sand held in place by adjacent 45 μm diameter sand layers ~11 mm thick, as illustrated in the insert. Flow was driven through the tube and through these fine impeding layers by a high-performance liquid chromatography (HPLC) piston pump. Pressure was controlled with a backpressure regulator at the discharge end, and reduced slowly with time. Until the pressure dropped below 290 psi and the CO2 began to exsolve, flow through the tube was single phase and the pressure difference between the entry and exit ends was very small (first 514 h in Figure 7). When the CO2 began to exsolve, flow through the tube was terminated and the pressure at the discharge and was reduced to atmospheric. The pressure at the entry end dropped to 163 psi over a short transition period (cross hatched band), and then remained unchanged for 3 weeks. The pore fraction of CO2 gas after this decompression was 54%.

**Figure 7.** Pressure changes as CO2-laden water is driven across a sand-packed tube with 5 fine-sand barriers as pressure is reduced. Until the pressure is reduced to the point that CO2 exsolves, the pressure difference between in inlet and outlet is small and predicted by the darcy flow equation. After CO2 exsolves, the flow of both gas and water stops despite a 163 psi pressure drop. The fine-sand barriers are illustrated in the insert, as is the nature of the sealing (red gas bubbles blocking flow into all the pores of the fine-grained layer). Figure modified from [52].

The total pressure drop is linearly related to the number of barriers. Six barriers have twice the pressure drop of 3 barriers and six times the drop of one barrier. Re-pressuring past 290 psi restores single phase flow and the single phase permeability is unchanged and predicted well by the Kozeny-Carmen relationship. The capillary pressure drop is predicted well by the Laplace relationship (insert) which equates the pressure drop across a fine-grained layer to the product of twice the CO2-water interfacial tension, σ, and the difference of the inverse radii of the fine and coarse pores on either side of the interface. The Laplace relationship does not contain permeability, and the temperature dependence of the capillary pressure drop, measured between 20 and 100 ◦C, is that of CO2-water interfacial tension. The two-phase barrier to flow in the experiment is thus clearly of a capillary nature. The capillary seals remained intact when 0.3 tube pore volumes of CO2-saturated fluid were passed through them. The flow blockage can be visualized as gas bubbles preventing flow into all the pores in the fine sand like toilet plungers (see insert). Capillary seals are remarkably easy to form in the laboratory, and remarkably durable.

The kind of seal formed in Shosa's experiments has many new properties and implications. It is not a lithologic seal. Although it may seem tied to lithology because it forms at a particular fine-grained layer or sequence of fine-grained layers, it can shift to other fine-grained layers. The top of overpressure can migrate upwards. Since sealing depends on both wetting (brine) and non-wetting (usually gas) phases being present, where either is dominant the seal does not exist. In a gas chimney, for example, gas is free to move inside the pipe and water is free to move outside the pipe. Only in the transition zone between gas and water are both fluids present in sufficient quantities to be immobilized. A proper-mixture-proportion impermeable sheath confines the pipe. A consequence is that a Shosa seal can fail completely, and release a grea<sup>t</sup> volume of gas from a pressurized compartment, and then re-heal. Finally, Shosa seals should form naturally and spontaneously in basins, and produce just the kind of pressure compartments observed. While recognizing that there is much we don't know about the kind of seals formed in Shosa's experiments, the following paragraphs discuss our current perceptions.
