Effect of Capillary Number on the Microstructure of Residual Oil in Strongly Water-Wet Sandstones I. Chatzls,· SPE, New Mexico Inst. of Mining & Technology M.S. Kuntamukkula, SPE, New Mexico Inst. of Mining & Technology N.R. Morrow, SPE, New Mexico Inst. of Mining & Technology Summary. Changes that occur with increase in capillary number in the detailed structure of residual oil trapped in water-wet sandstone core samples have been investigated. The technique of using a nonwetting phase that can be solidified and separated from the porous medium has been applied with styrene monomer as the nonwetting phase and 2% CaCl2 brine as the wetting phase. The size distributions of residual oil blobs, obtained under various flow conditions, were measured by both image analysis and Coulter counter techniques. Specific features of blob shapes and dimensions were checked by optical and electron microscopy. The changes in size distribution and shapes of blobs provide insight into the mechanisms of trapping and mobilization of residual oil. Introduction At the conclusion of waterflooding an oil-bearing reservoir, a sig- nificant fraction of the original oil still remains in the swept region as trapped residual oil. In water-wet reservoirs, this residual oil, Sbr' may typically occupy 25 to 50% of the pore space and provides a main target for tertiary oil recovery. Trapped oil can be recovered from a core sample at Sbr by immiscible displacement if the ratio of viscous to capillary forces, expressed in this work as the capillary number Nc =kwilp/Lu, exceeds a critical value. 1,2 Changes in microscopic distribution of oil within pore spaces can still occur at capillary numbers less than critical. Above the critical capillary number, Nc(crit), oil is displaced from the core sample. In laboratory investigations, nondimensional relationships between capillary number and the ratio Sor/Sbr (residual oil saturation, Sor, normalized with respect to Sbr) have been found to be remarkably similar for a variety of sandstones. 3 In addition to the amount of trapped oil, its microscopic distri- bution within the pore spaces of a reservoir rock is important to gain a better understanding of oil-recovery mechanisms. This knowledge may also be important to the design and implementation of tertiary recovery processes. For example, in modeling the recovery of residual oil, the viscous force required for mobilization of a residual oil blob trapped under water-wet conditions is expected to be inversely proportional to blob length. 4-6 The technique of using a nonwetting phase, which after flooding to residual saturation can be solidified and then separated from the porous medium to study the microscopic structure of residual non- wetting phase, was probably first employed by Craze, 7 who re- ferred to the observed capillary structures as irregulariy shaped blobs. Blob-size distributions have been measured in the past in sandpacks with styrene monomer as the oleic phase before solidifi- cation. 8 The results of this study, 8 although released, 9 have not been made available through publication to the research community at large. A previous studylO in which styrene polymerization was used has also been cited 8 but is not available. A technique for the study of residual oil structures that involved trapEing of melted wax has been used by Morrow ll and Humphrey. 2 Since Reed and Healy12 credit the method used by Humphrey to Taber's much earlier unpublished work, it is clear that blobs prepared by this tech- nique have been examined by several investigators. Also, scanning electron micrographs (SEM's) of pore casts of blobs of residual nonwetting phase obtained through solidification of Wood's metal with hot toluene as the wetting phase have been presented by Swanson. 13 Although considerable attention has been paid to the obviously important subject of residual oil structure, the amount of experimen- 'Now at U. of Waterloo. Copyright 19B8 Society of Petroleum Engineers 902 tally determined, quantitative information on blob structure and the statistics of blob populations is very limited. To obtain such infor- mation, satisfactory techniques for preparing statistically represen- tative blob samples and measuring their size distributions must be devised. Once obtained, the experimentally determined blob-size distributions can be related to measured conditions for mobilization and compared with changes in size distribution predicted by theory. 9, 14 Experimental Fluids and Porous Media. The fluid pair used in this study was styrene monomer (the oil phase) and 2 wt% CaCl2 brine. The styrene monomer contained I wt% benzoyl peroxide as initiator for the in-situ polymerization of trapped styrene following coreflooding. An important reason for using benzoyl peroxide was that it maintained water-wetness throughout the polymerization process, whereas other initiators that were tested were either in- effective or produced gross changes in wetting behavior. The in- terfacial tension (1FT) of this fluid pair was about 29.7 mN /m [29.7 dynes/cm] at room conditions (-24°C [-75°FD. However, it was not uncommon for lower values to be observed with aging. The core samples were all Berea sandstone, except one that was a Boise sandstone. All but one sample (Sample B-BF) were unfired and had dimensions of 3.82 cm [1.5 in.] in diameter and 7 to 15 cm 3 in PV. The cores were coated with epoxy resin (Devcon-Mill Crusher Liner, liquid-type) by placing the core sample in a cylin- drical mold with an ID of 5.08 cm [2 in.] and filling the annulus between the mold wall and the core sample with resin. After the resin had set, the core with the resin coat was removed from the mold and the inlet and outlet faces of the epoxy-coated core sample were machined off with a lathe. End plates were machined from polytetrafluoroethylene. Schematics of the core holder and the coreflooding apparatus are shown in Fig. la and b. Coreflooding Procedures. Procedures used for studying the structure of residual oil as a function of capillary number are listed below. 1. The core properties, porosity, cp, air permeability, k a , and brine permeability, kw (see Table 1), were first determined with standard core analysis procedures. 2. A high initial oil saturation, Soi, was established by flooding a brine-saturated sample with styrene monomer at pressure gra- dients (!::.p/L) of about 150 kPa/cm [55 psi/in.]. About 15 to 20 PV of styrene was passed through the sample. This included flow in both directions to reduce saturation gradients caused by capillary end effects. 3. Subsequent floods fell into one of three categories. Type 1 was low-capillary-number floods to Sbr (about 1.5 PV of brine was SPE Reservoir Engineering, August 1988