(607a) Pore Scale Models for Imbibition of CO2 Analogue Fluids in Etched Micro-Model Junctions Using Micro-Fluidic Experiments and Direct Flow Calculations

In this paper, we investigate pore scale models for the storage of CO₂ in subterranean rock formations.  Various mechanisms for storage and trapping have been proposed, including structural, solution, mineralisation and the immobilisation of CO₂ by capillary trapping. Currently, the capillary trapping mechanism seems to be a suitable option, as it is a fast way to render the injected supercritical CO2 immobile. In addition, this trapping mechanism does not rely entirely upon the cap rock maintaining its integrity on geologic time scales. One important parameter for the success of capillary trapping is the residual saturation after CO₂ injection in the formation. The saturation is determined by the wetting properties of the rock formation. Often, and particularly in carbonate reservoirs, the rock may be wetting to CO₂, especially when injected under supercritical conditions. In this case, the mechanism of entry of CO₂ into the pores of the rock matrix is determined by spontaneous imbibition. To understand this problem at the pore scale, it is essential to have a clear view of the mechanism of capillary filling in pore junctions. In the 1980s, Lenormand and co-workers carried out pioneering work on displacement mechanisms of fluids in etched networks [1]. They observed that for drainage, the mechanism of displacement is simply determined by the Young-Laplace law, stating that the capillary pressure is proportional to surface tension and the cosine of the contact angle, and inversely proportional to the radius of the capillary. In addition, Lenormand et al. claimed that capillary filling rules for imbibition are also determined by the Young-Laplace law. The rules for capillary filling for both drainage and imbibition are currently used in pore network models. However, in recent years, several spurious results were obtained for spontaneous imbibition in pore network models. For this reason, we decided to revisit and investigate the validity of the capillary filling rules for the case of spontaneous imbibition, using a combination of micro-fluidic experiments and direct Navier-Stokes CFD flow calculations, using the lattice-Boltzmann (LB) method. The micro-fluidic cells consisted of single junctions with unequal throats, etched into PMMA. First, we validated the case of drainage by injecting DI water into the oil-wet PMMA models. Indeed we observed that, when the pore body has been filled, the subsequent displacement shows drainage into the largest capillary throat first, in agreement with the Young-Laplace law. Then we carried out a set of imbibition experiments. This was done by imbibing decane, which has shown to be an excellent analogue fluid at ambient conditions for CO₂ under supercritical conditions, into the junction. Each time, we observed that the fluid does not imbibe into the smallest throat, as predicted by the Young-Laplace law. Instead, we observed that the capillary filling rules for displacement are determined by the local geometry of the junction. To validate the experimental results, we carried out LB flow calculations in an exact representation of the 2D experimental geometry. Indeed we observed that the flow calculations reproduce the experimentally observed displacements in detail. This means that the capillary filling rules in pore network models (PNMs) will have to be overhauled. In this paper, we will describe the experimental and modeling results in detail. Also, we will discuss the implications for CO2 injection at the Darcy scale.

This work was carried out as part of the activities of the Qatar Carbonates & Carbon Storage Research Centre (QCCSRC). We gratefully acknowledge the funding of QCCSRC provided jointly by Qatar Petroleum, Shell, and the Qatar Science and Technology Park.

[1] G. Lenormand, C. Zarcone and A. Sarr,  J. Fluid Mech. 135 , 337-353 (1983).