Microfluidic chips for oil recovery from carbonate reservoirs at elevated temperatures
Anh Duy Le is a PhD student in the department Physics of Complex Fluids. (Co)Supervisors are prof.dr. F.G. Mugele, prof.dr. J.G.E. Gardeniers and dr. M.H.G. Duits from the faculty of Science & Technology.
Low salinity waterflooding is a technique that has been reported to improve oil recovery in carbonate reservoirs in core flooding tests. It is generally accepted that the wettability alteration towards more water wet that occurs upon low salinity injection in carbonate reservoir results in additional oil extraction. However, the underlying mechanisms and associated fluid-fluid, fluid-solid interactions during this process are still unclear and scattered due to the lack of observations of multiphase flow in core plugs and the lack of sufficient time and spatial resolution. To develop efficient methods to economically recover additional oil, a comprehensive understanding of these phenomena at pore scale is highly required. Therefore, artificial microfluidic models with similar physicochemical properties to carbonate reservoir conditions attracts a great interest from researchers in the oil recovery community. This thesis focuses on development of microchannels that mimick carbonate reservoir relevant conditions to study the oil recovery process.
Chapter 2 presents the previous development of microfluidic microchannels in the past and its application for improved oil recovery study. The geometry and geochemistry of the micromodels, as well as the temperature, pressure, and aging conditions of waterflooding experiments performed in these micromodels are summarized. In addition, a typical experimental setup for improved oil recovery study and visualization techniques of multiphase fluid flow at pore scale are also described. This chapter provides a fundamental background for the next four chapters in the thesis.
Chapter 3 addresses the challenges of fabricating a 3D microfluidic microchannel to study oil recovery for carbonate reservoirs. A novel and straightforward fabrication method to create a chemical representation of the carbonate reservoir combined with corresponding 3D features of the porous network was presented. This was achieved by the assembly of micro-sized synthesized calcite particles into a PDMS microfluidic channel. The fluorescence microscope was used to visualize calcite particles, calcite, and mineral oil along with automated image analysis to determine residual oil saturation at various locations across the channel close to the glass transparent wall. While this method uses calcite particles, similar to material in carbonate reservoirs, the optical access to the bulk of the calcite packed bed microchannel is limited due mismatch in refractive indexes between calcite and fluids (brine and oil). Therefore, we cannot visualize the two-phase displacement in the bulk and cannot know the difference in the oil recovery process for pores near boundaries and pores in the bulk. In addition, the PDMS microchannel is also not compatible with crude oil and flooding at elevated temperature conditions.
Chapter 4 addresses the lack of more realistic physicochemical conditions and a more realistic pore networks in a microfluidic microchannels. This chapter presents the effect of fluids aging and reservoir temperature on waterflooding in 2.5D glass micromodels. The 2.5D micromodel with two distinct depths offers better representation of strong variation in pore depth compared to standard 2D microchannel with a single depth. In addition, physicochemical behavior of brine/crude oil at elevated temperature conditions is also incorporated. We found that glass microchannels that have been aged in formation water (FW) and subsequently exposed to crude oil (CRO) retain significantly more oil compared to those exposed to only crude oil after waterflooding with high salinity water at room temperature. This can be explained by a stronger adsorption of hydrophilic parts of CRO components (e.g. asphaltenes, resins) stimulated by divalent cations from FW that makes the glass surface more oilwet. In addition, ROS in FW/CRO aging protocol also decreases with increasing waterflooding temperature. Wettability alteration of the glass surface toward more water-wetness at elevated temperature was also observed with microscopic ‘dewetting’ phenomena that demonstrate the role of fluid-solid interaction at the pore scale. These results highlight the significant role of geometry (dual depth) of the microchannel to study IOR in addition to chemical aspects. While Chapter 4 already contributes to further development of the microfluidic platform to optimize IOR conditions, realistic sandstone reservoir conditions could be replicated with more complicated microchannels.
Chapter 5 extends the scope from 2.5D glass microfluidic system developed in Chapter 4 by combining specific conditions of IOR (elevated temperature, crude oil, aging, surface chemistry and pore heterogeneity) to further mimic reservoir conditions with a focus on carbonate reservoirs. Our adapted method could partially coat the inner glass surface with calcite particles on the complex geometry of 2.5D microchannels. It was found that ROS values appear to be higher for the calcite coated microchannel in comparison to that in glass microchannels with the same pore network geometry and flooding conditions. The overall ROS trends as a function of flooding temperature were also found to be similar between these two types of microchannel. History and flooding conditions ultimately determine the final ROS value. We also observed formation and growth of large (10-100 μm) ‘blisters’ of water drops at 90 ◦C for longer time (days to a week) exposure to HSW. These blisters were possibly caused by osmotic effect due to a gradient in water chemical potential. This indicates the important role of fluid-fluid and fluid-solid interactions, particularly water-in-oil micro-dispersions and osmosis phenomena at elevated temperature that was hardly reported at pore scale in micromodel so far. Separate analysis of ROS trends reveals significantly more oil retention at the pore edges compared to the average pore space, suggesting that in calcite-coated micromodels, wall roughness is a stronger determinant of oil retention than surface chemistry. Furthermore, our novel interpolation and complementary image analysis method improve the precise ROS calculation in cases of coexistence of brine/oil and optical heterogeneities of calcite particles compare to standard binarization methods.
Chapter 6 briefly summarizes the main questions related to the development of microchannels to study improved oil recovery for carbonate reservoirs. I describe what answers this thesis provides and how it contributes to the current development of microfluidics in the petroleum oil community. Some critical reflections on the results in the thesis, fabrication techniques, visualization, measurement protocols and data analysis are also discussed. Finally, I describe remaining open questions and present some recommendations for future development that might improve our understanding of the problems.