SPE-187176-MS The Impact of Surfactant Imbibition and Adsorption for Improving Oil Recovery in the Wolfcamp and Eagle Ford Reservoirs J. O. Alvarez, I. W. R. Saputra, and D. S. Schechter, Texas A&M University Copyright 2017, Society of Petroleum Engineers This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, USA, 9-11 October 2017. This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright. Abstract Improving oil recovery from unconventional liquid reservoirs (ULR) is a major challenge and knowledge of recovery mechanisms and interaction of completion fluid additives with the rock is fundamental in tackling the problem. Fracture treatment performance and consequent oil recovery can be improved by adding surfactants to stimulation fluids to promote imbibition by wettability alteration and interfacial tension (IFT) moderate reduction. Also, the extent of surfactant adsorption on the ULR surface during imbibition of completion fluids is an important factor to take into account when designing frac jobs. The experimental work and modeling presented in this paper focuses on analyzing alteration of wetting behavior of Wolfcamp and Eagle Ford reservoir rock with the introduction of surfactants additives. We focus on effectiveness of surfactant additives for improving oil recovery as well as the extent of surfactant loss by adsorption during imbibition of surfactant-laden completion fluid. Altering the wettability with the use of surfactant additives is accompanied by alteration of the IFT as well as surfactant adsorption. We carefully evaluate these interactive variables as key constituents of imbibition capillary pressure to improve oil recovery. We assume this is a free imbibition process with no confining pressure on the rock sample. During imbibition spontaneous imbibition, as the sign of the capillary pressure changes from negative (oil wet) to positive (water wet). Original rock wettability is determined by contact angle (CA) at reservoir temperature. Then, different types of surfactants, anionic, anionic-nonionic, and cationic, at concentrations utilized in the field, are evaluated to gauge their effectiveness in altering wettability and IFT. Wettability is also studied by zeta potential to address water film stability on the shale rock surface as an indication of wetting fluid affinity and to determine the surfactant electrostatic charges. Moreover, surfactant adsorption measurements are performed using an ultraviolet–visible spectroscopy. Calibration curves for surfactants are determined by relating their concentration to light absorbance and used to calculate the amount of surfactant adsorption into the shale rock. Next, potential for improving oil recovery via surfactant additives in ultralow permeability Wolfcamp and Eagle Ford shale core is investigated by spontaneous imbibition experiments at reservoir temperatures. In order to visualize the movement of fluid as it penetrates into liquid rich shale samples, we use computed tomography (CT) methods to determine fluid imbibition in real time. In addition, oil recovery is recorded with time to compare the performance of surfactants and water alone. Finally, laboratory data are used in numerical simulation to model laboratory results and upscale these findings to the field. The results 2 SPE-187176-MS showed that aqueous solutions with surfactants altered rock wettability from oil-wet and intermediate-wet to water-wet and reduced IFT to moderately low values. In addition, cationic surfactant presented the highest adsorption capacity following a Langmuir type adsorption profile. Spontaneous imbibition results showed that aqueous solutions with surfactants had higher imbibition and were better at recovering oil from shale core compared to water without surfactants, which agrees qualitatively with wettability and IFT alteration. However, rock lithology and surfactant type play an important role in adsorption capacity and oil recovery. Our upscaling result shows that compared to a well that is not treated with surfactant, a 24% increase on the initial peak oil rate as well as a 8% increase on the 3-year cumulative oil production are observed. For the results obtained, we can conclude that the addition of surfactants to completion fluids can improve oil recovery by wettability alteration and IFT reduction, maximizing well performance after stimulation from Wolfcamp and Eagle Ford unconventional reservoirs. Introduction Unconventional liquid reservoirs (ULR) have become an important source of energy in the United States. The increasing hydrocarbon exploitation of liquid rich shale reservoirs has positioned the country as one of the biggest oil producers on the planet (Doman 2015). The two most current productive regions in the United States are the Permian Basin and the Eagle Ford with more than 300 rigs on Permian and 75 rigs on Eagle Ford (EIA 2017). The use of multiple hydraulic fracture treatments in horizontal wells in these ULR has boosted their production. Hence, the Permian Basin oil production has increased from less than 900 thousand barrels per day in 2008 to currently more than 2.5 million barrels per day whereas the Eagle Ford has increased its production from 100 thousand barrels per day in 2008 to more than 1.2 million barrels per day nowadays (EIA 2017). However, ULR petrophysical characteristics of low porosity, ultralow permeability, heterogeneity and high total organic content (TOC) create a challenge for oil exploitation. Thus, current recovery factors for ULR do not exceed more than 10% of the original oil in place (OOIP) with average values of 5 to 6% (Alharthy et al. 2015; Wang et al. 2016). Multistage hydraulic fracturing allows ULR to produce at commercial flow rates by creating effective paths for hydrocarbons to flow towards the wellbore. The effectiveness of fracture treatment in increasing recovery and consequently current low oil recovery factors may be improved if proper surfactants are added to completion fluids, thereby altering wettability, reducing interfacial tension (IFT) and consequently improving water imbibition. Imbibition is largely dependent on the capillary forces, which are defined by the Young-Laplace equation (Eq. 1) relating capillary forces (Pc) with wettability as contact angle (θ), IFT (σ), and pore radius (r). Hence, capillary forces are significant in organic liquid rich shale nanopores and complex as the contact angle and IFT varies simultaneously. (1) In petroleum systems, wettability is the affinity of either water or oil to spread onto a rock surface. The fluid that has higher affinity to the rock is then called the wetting phase and the non-wetting phase is the other fluid (Anderson 1986a). Thus, rock surface can be oil, water or intermediate-wet. Wettability can be measured quantitatively by contact angle (CA), Amott-Harvey index, and US Bureau of Mines (USBM) methods, and qualitatively by nuclear magnetic resonance (NMR), relative permeability determination and zeta-potential methods (Alvarez and Schechter 2017; Anderson 1986b; Wang et al. 2012). All these methods have been successfully used to measure wettability in conventional reservoirs; however, in unconventional reservoirs, many of these techniques have limited application. Methods such as Amott-Harvey index and USBM require acquisition of fluid saturation changes by displacement of brine by crude for the drainage cycle into the matrix. Without direct injection to cover the imbibition and drainage cycles, these rocks cannot be utilized for special core analysis in the conventional manner. The lack of direct injection necessitates more modern core-handling techniques as described in this paper. Hence, for this study we chose the CA SPE-187176-MS 3 method for being the most viable way to quantitatively determine wettability alteration in ULR. Unlike conventional reservoir rocks, shale chips provide ample, almost molecularly smooth, surfaces to measure directly the contact angle of reservoir fluids at reservoir temperature. In addition, we used zeta-potential measurements to qualitatively investigate the wetting affinity of these ULR based on the dependency of wettability on the stability of the double-layer between oil and the shale rock surface. The layer thickness and stability is determined by relative charge of the surface and the fluids interacting with the surface (Hirasaki 1991). Thus, stable solution film gives an indication of water-wet systems whereas unstable solution film is considered oil or intermediate-wet. Also, the strength and nature of the rock and the aqueous solution charge can be gauged by zeta-potential measurements. The Young-Laplace equation (Eq. 1) also considers the IFT as an important parameter in capillarity. IFT is the force that holds a phase in the pore space in which oil, water and gas coexist. Depending on its value and application, IFT can be measured by sessile drop, spinning drop and pendant drop methods. We measured IFT using the pendant drop method, which relates the deformation of a drop in another liquid phase and is very reliable for values larger than 0.2 mN/m. Finally, the last parameter affecting capillary pressure is the mean pore radius, which we determined from pore throat distribution using mercury injection capillary pressure (MICP) analysis. Due to the mixture of water-wet inorganic pores and oil-wet organic pores in ULR, wettability in these unconventional rocks is originally oil and intermediate-wet. Alvarez and Schechter (2016a) compiled the different wettability measurements for different ULR available in the literature showing that most of the results reported intermediate to oil-wet affinities originally. In addition, Alvarez and Schechter (2016b) performed several more wettability studies in the Bakken, Eagle Ford, Wolfcamp and Barnett ULR concluding that their original wettability was mostly intermediate towards oil-wet with the Wolfcamp exhibiting the greatest degree of oil-wetness. In order to let capillarity favor imbibition, capillary pressure must be positive. Thus, ULR wettability must be altered to water-wet when original wettability is oil and intermediate-wet. Also, IFT should be moderately reduced to favor water imbibition, but not to ultra-low values which would make capillary pressures negligible (Alvarez et al. 2017). Wettability and IFT can be modified by surfactants. Typically, we work with commercial products containing blends of which we have no precise information on the molecular structure, only the general charge of the surfactant solution utilized. As surface-active agents, they contain two groups, a lipophilic group that has a strong attraction for the solvent phase and a lipophobic group that tends to repel the solvent phase molecules. In oil-water systems, surfactants generally consist of a hydrophilic head that is an ion bearing positive charged (cationic surfactants), negative charged (anionic surfactants) or neutrally charged (nonionic surfactants), and a hydrophobic tail composed of various hydrocarbon chain lengths and isomer groups. Some surfactants consist of hydrophilic heads bearing both positive and negative charges, and they are characterized as amphoteric or zwitterionic surfactants. Wettability alteration coupled with reduction in interfacial tension at the oil-water fluid interface have been the focus of various enhanced oil recovery (EOR) studies in conventional reservoirs as they affect the imbibition profile and consequent oil recovery. In conventional reservoirs, altering the wettability of the rock to water-wet helps recover additional oil by mobilizing oil trapped in fine channels after primary recovery, which is a strong motivation behind the research of EOR techniques (Guo et al. 1998; Schechter et al. 1991; Schechter et al. 1994). However, these methods have limited application in ULR due to their unique petrophysical properties such as low porosity and permeability, mixed lithology and elevated total organic carbon (TOC) content. Hence, the use of surfactant as well as recovery mechanisms may change between conventional and unconventional reservoirs. For example, the forces that determine imbibition and drainage in porous media in both conventional and unconventional reservoirs are capillary, gravity and viscous forces. However, in conventional reservoirs, reduction of capillary forces related to viscous forces via capillary number is the primary mechanism whereas improvement of adhesion-repulsion forces at the surface thereby altering capillary forces is hypothesized to be the driving force in ULR. 4 SPE-187176-MS In addition to the role of surfactants as wettability and IFT modifiers, the effect of adsorption in ULR must be studied. Currently, there are very limited investigations on surfactant adsorption in ULR and, to our knowledge, none for the Wolfcamp and Eagle Ford formations. Using Marcellus shale outcrops, Zelenev et al. (2011) studied nonionic surfactant static adsorption by measuring surface tension of diluted solutions before and after crushed shale equilibrium. The resulting adsorption values for nonionic surfactant were close to 15 mg/g of rock at surfactant concentration of 3000 mg/L. Next, using ultraviolet-visible (UV-Vis) spectroscopy in an undisclosed crushed calcite and clay-rich ULR sample, Mirchi et al. (2014) performed static adsorption measurements on anionic surfactant, noticing very low adsorption values (0.508 mg/g at CMC of 0.03 wt.%) and Langmuir type adsorption behavior. They suggested that these results were caused by shale surface capacity for attracting predominately cations instead of anions. Zhang et al. (2016) used anionic, nonionic and blended surfactants to measure surfactant adsorption in siltstone Middle Bakken by UV-Vis spectroscopy. Their results showed adsorption capacities of 0.62 mg/g of rock for the blended surfactant and 11.91 to 33.08 mg/g for nonionic and anionic surfactants. However, the authors attributed the latter elevated values to unreliable measurements caused by solution turbidity. Finally, Alvarez et al. (2017) measured anionic and complex nanofluid (CNF) dynamic adsorption in siliceous and carbonate cores from Three Forks formation in the Bakken using UV-Vis spectrophotometry. Their results showed surfactant adsorption ranges from 6.2 to 8.9 mg/g of rock at concentration of 2 gallons per thousand gallons (gpt) with a Langmuir-type adsorption mechanism on the Bakken samples. In addition, they found that lithology and surfactant type influences adsorption capacity with anionic surfactant absorbing more on siliceous rock whereas positively charged CNF shows higher adsorption capacity in carbonate-rich rocks (Alvarez et al. 2017). Currently, there is very scant literature on the study of combined effect of wettability and IFT alteration on imbibition process on the Wolfcamp and the Eagle Ford ULR. However, the recent use of ULR as a source of liquid hydrocarbons has caught the attention of the industry in regards to wettability alteration and imbibition in ultralow permeability reservoirs. Xu and Fu (2012) used oil saturated crushed Eagle Ford sample in a packed pressure column. Surfactant solutions were flooded through this arrangement to evaluate oil recovery. The results showed that weakly emulsifying surfactant recovered more oil than nonemulsifying surfactant by reducing capillary pressure. The authors determined that rock wettability was altered by surfactants using the Washburn method, but no direct wettability measurement was performed. Nguyen et al. (2014) experimented with outcrops from Eagle Ford. They used two cationic, three nonionic, two zwitterionic, three anionic and blends of surfactants at concentration from 0.1 to 0.2 wt.% and in a 2 wt.% brine. Spontaneous imbibition experiments were performed in Amott cells at reservoir temperature. The results showed that anionic surfactants recovered 48% of the original oil in place (OOIP) and cationic surfactants 38% and 23% of OOIP. However, for the second cationic surfactant, brine alone was better in recovering oil (30% of OOIP). Also, CA was not properly measured as they were qualitatively measured just by dispensing a brine drop to the shale surface. Such measurements are not analogous to recovery of oil on an initially oil-wet surface by dislodging that oil via wettability alteration. The authors concluded that wettability alteration is the main mechanism for oil recovery because they did not find correlation with IFT and recovered oil. Finally, Alvarez and Schechter (2017) performed spontaneous imbibition experiments in siliceous core samples from the Wolfcamp using anionic and nonionic-cationic surfactants in modified Amott cells tracking oil movement by CT scan time-lapse sequences. Their results showed oil recoveries up to 33.9 % of OOIP for anionic surfactant and up to 19.7 % of OOIP for nonionic-cationic surfactant with water only recovering up to 10.5 % of the OOIP. This study aims to close the current gaps in the literature by evaluating the role of surfactants to completion fluids in improving oil recovery when fracturing Wolfcamp and Eagle Ford ULR by a correlated set of experiments. First, ULR original wettability and wettability alteration by surfactants is studied by CA and zeta-potential measurements. Then, additives capacity in decreasing oil/brine IFT is addressed by IFT measurements. Next, surfactant adsorption measurements are performed using ultraviolet-visible SPE-187176-MS 5 spectroscopy to address the extent of surfactant loss by adsorption during imbibition of completion fluids. Then, spontaneous imbibition experiments coupled with CT scan technology is used to analyze the impact of surfactants on water imbibition and consequently oil recovery, which holds the key to design chemical treatment for improving oil recovery. Finally, laboratory results are simulated using reservoir modeling to reproduce experimental set-up and findings and upscale these results to estimate field production. The uniqueness of this work is that it evaluates a holistic range of parameters the impact rock-fluid and fluid-fluid interactions by studying the effect of surfactant additives on wettability, IFT and adsorption and their implications for improving oil recovery under multiple scenarios. To our knowledge, this has not been rigorously performed in the Wolfcamp and Eagle Ford unconventional reservoirs. The results showed that surfactant are capable of altering wettability from intermediate and oil-wet to water-wet in Wolfcamp and Eagle Ford core as well as simultaneously reducing water-oil IFT. Moreover, in spontaneous imbibition experiments, aqueous solutions with surfactants imbibed further, as measured by CT scanning, and recovered more oil than water alone. These findings are consistent with CA, zeta potential, and IFT measurements. On the field-scale simulation, adding surfactant into the system is observed to increase the initial oil production rate and the cumulative oil production. From these results, we can conclude that the addition of surfactants to completion fluids alter wettability and reduce IFT, thereby improving oil recovery in Wolfcamp and Eagle Ford core. Thus, we begin to lay the foundation of the necessary tests and significant parameters that determine the optimum exchange of fluids during aqueous phase treatments in multiple reservoirs. Methodology This study investigates the interaction of completion fluids, with and without surfactants, and ULR. In addition, it addresses the effect of wettability, IFT, adsorption and imbibition on recovering hydrocarbons from liquid-rich shale cores from the Wolfcamp and Eagle Ford formations. These objectives are achieved by performing contact angle, zeta potential, adsorption and IFT measurements, as well as spontaneous imbibition monitored by computer tomography (CT) methods. Hence, this chapter describes a novel set of correlated experiments to evaluate and compare the efficiency of surfactants in altering wettability and recovering hydrocarbons from ULR core. Finally, wettability, IFT, adsorption and spontaneous imbibition results are used as inputs for a reservoirs simulation to reproduce laboratory experiment and scale the results to the field. Rock and Fluid Properties Sidewall core plugs received from the liquid-rich portion of the Wolfcamp and the Eagle Ford play were used. Wolfcamp samples were taken from well W-1 at depths from 7,850 to 7,900 ft. whereas Eagle Ford samples were from well EF-1 at depths from 13,000 to 13,050 ft. Cores are 1-inch in diameter and 1.5 to 2.5-inches in length. Wolfcamp samples show porosity from 6 to 7 %, permeability to air of 200 to 300 nD, and median pore radii of 0.005 microns all measured by mercury injection capillary pressure analysis (MICP). Similarly, Eagle Ford samples show porosity of 9 to 12 %, permeability to air of 100 to 300 nD, and median pore radii of 0.007 microns. Total organic carbon (TOC), measured on a LECO C230 Carbon Analyzer is from 5 to 6 wt.% for Wolfcamp and 6 to 6.5 wt.% for Eagle Ford. ULR rock sample lithology for the intervals studied is shown in Table 1. The X-Ray Diffraction (XRD) analyses for both Wolfcamp and Eagle Ford demonstrate that the samples are predominately carbonaceous. 6 SPE-187176-MS Table 1—Lithological composition of rock samples from wells W-1 and EF-1 Well W-1 EF-1 Quartz 13 17 Clays 15 35 Calcite 46 40 Dolomite 19 1 Feldspar 4 3 Pyrite 3 4 Illite/mica 74 65 Illite/Smectite 26 35 Kaolinite 0 0 Chlorite 0 0 Mineral Composition (wt.%) Relative Clay (%) Wolfcamp crude oil from well W-1 was used with density of 0.82 g/cm3 and 32.4° API at reservoir temperature of 165 °F. On the Eagle ford, crude oil from well EF-1 was used with density of 0.72 g/cm3 and 52.61° API at testing temperature of 180 °F. Oil total acid number (TAN) and total base number (TBN) was determined by titration methods in a Metrohm 905 Titrando apparatus. Wolfcamp oil TAN and TBN are 0.09 and 0.12 mg KOH/g oil, respectively whereas Eagle Ford oil TAN and TBN are 0.02 and 0.61 mg KOH/g oil, respectively. Surfactants and Brine Four different surfactants, which are commercially available and commonly used as additives in completion/ fracture fluids were used. In order to evaluate a broad range of surfactant types, we tested two anionic, one blended anionic-nonionic and one cationic surfactant at field used concentration of 2 gallons per thousand gallons (gpt). Surfactant descriptions are in Table 2. Table 2—Description of the surfactants used in the experiments Surfactant Anionic 1 Anionic-nonionic Anionic 2 Cationic Primary Components Composition (wt.%) Methyl alcohol 40-70 Proprietary sulfonate 10-30 Methyl alcohol 10-30 Sulfonate A 7-13 Sulfonate B 3-7 Ethoxylated alcohol 3-7 Isopropyl Alcohol 10-30 Citrus Terpenes 10-30 Proprietary 10-20 Isopropyl Alcohol 10-30 Ethoxylated alcohol 3-7 Quaternary Ammonia Compound 3-7 Citrus Terpenes 5-10 pH Specific Gravity 5.8-7.2 0.866 - 0.892 4.7-5.7 0.974 - 0.999 6.8-8.3 0.953 - 0.956 4.6 0.951-1.011 SPE-187176-MS 7 Aqueous solutions were prepared with distillated water and 4 wt.% potassium iodide (KI) brine solution (named as Water on this study). KI was added as a dopant to increase the contrast between oil and aqueous phases on the CT scanner, a needed procedure to observe fluid penetration during spontaneous imbibition experiments. The effect of KI on sample wettability was found to be negligible, so it is not considered for comparative analyses. Moreover, pH values remained constant for all surfactant solutions and brine in the experiments executed in this study. Contact Angle Measurements CA experiments were performed on a Dataphysics OCA 15 Pro apparatus using the captive bubble method. Wolfcamp and Eagle Ford core samples were carefully cut, to fit inside the measuring device, polished, to minimize measurement errors related to surface roughness, and cleaned in toluene and methanol to remove any impurities due to sample handling. Then, core trims were aged in Wolfcamp and Eagle Ford oil for more than 4 weeks at reservoir temperature. Wolfcamp and Eagle Ford samples were then submerged in water alone as well as surfactant aqueous solutions at concentration of 2 gpt. Wolfcamp and Eagle Ford oil is dispensed bottom up using a capillary needle. As the droplet of oil slowly attaches to the shale surface forming an angle, the shape of the oil's droplet is analyzed by enhanced video-image digitalization technique. Further details are in Alvarez et al. (2014). In order to represent as accurate as possible the liquid phases on the reservoir, oil and water phases are present and the contact angle is formed on the rock surface. Contact angles ranging from 0° to 75°, 75° to 105° and 105° to 180° are employed for water, intermediate and oil-wet as defined by Anderson (1986b). Moreover, CA measurements for Wolfcamp samples were performed at reservoir temperature of 165 °F; similarly, Eagle Ford experiments were performed at 180 °F, which is the closest experiment temperature to reservoir temperature (220 °F) that avoids bubble movement, which might affect the reliability of our results. Error bars are assigned based on the experiment confidence level with upper and lower bounds of 3 degrees. Zeta Potential Measurements Zeta potential measurements were performed on a NanoBrook ZetaPALS device using the Phase Analytical Light Scattering (PALS) method. The device measures the electrophoretic velocity of the particles in the solution and use this value to calculate the electrophoretic mobility. Measurements were performed in aqueous solution with and without surfactants, at concentrations of 2 gpt, and ULR rock samples from wells W-1 and EF-1 (Wolfcamp and Eagle Ford). Aqueous solutions were prepared and triple filtered before placing them into the vial for measurement. Moreover, core trims were finely crushed and screened in an ASTM 325 sieve of 45-μm diameter. Rock and aqueous solutions were mixed in a shaker for 1 minute. The sonicated solution was left to stabilize by letting it sit for 5-10 hours for the heavy insoluble particle to settle down. Then, the solution was placed in the zeta potential measuring device. Further details are in Alvarez and Schechter (2017). Error bars are assigned based on the experiment confidence level with upper and lower bounds of 2 mV. IFT Measurements IFT measurements were performed in a Dataphysics OCA 15 Pro apparatus using the pendant drop method. Wolfcamp and Eagle Ford oil is dispensed through the capillary needle, which is submerged into the aqueous solution with and without surfactants at concentration of 2 gpt. The experiment is recorded by a highresolution video camera and the image of when the drop is about to leave the needle is analyzed by enhanced video-image digitalization technique. IFT experiments were performed at the same temperature as the CA experiments of 165 °F and 180 °F for Wolfcamp and Eagle Ford, respectively. Finally, to calculate the brine/ oil IFT, the droplet shape profile is matched to the Laplace equation by the device software, using the density 8 SPE-187176-MS of the oil and aqueous solutions at testing temperature. Further details are in Alvarez et al. (2014). Error bars are assigned based on the experiment confidence level with upper and lower bounds of 0.2 mN/m. Surfactant Adsorption Measurements Surfactant adsorption on the Wolfcamp and Eagle Ford rock surface was measured by UV-Vis spectrophotometry. A Hitachi U-4100 UV-Vis-NIR spectrophotometer was used to calculate the concentration of the surfactant as time progresses. The device produces light with specific wavelength that is shined through the tested solution. Water without additive added was used as the reference solution and aqueous solution with surfactants are used from calibration curve and adsorption experiments. The calibration curves for each surfactant were done by correlating the amount of light adsorbed on that wavelength to the surfactant concentration in the solution. The wavelength scan range used was 190 to 300 nm. After the calibration curves were constructed, dynamic adsorption measurement was performed. To that end, Wolfcamp and Eagle Ford samples were cleaned in toluene then methanol, for 3 days and 2 days consecutively, then vacuum-dried for 3 days. Cleaned samples were crushed and passed thru a 300 μm sieve. Aqueous solution with surfactants were mixed in a 1:20 weight ratio with the rock samples. During the experiments, fluid samples were taken at different times from 10 minutes to 24 hours and filtered through a 20 μm syringe filter before measuring the light adsorption to remove rock particles and consequently stopping the adsorption reaction. Lastly, using the calibration curve for each surfactant, surfactant dynamic adsorption was calculated at each time step using Eq. 2 (Alvarez et al. 2017). (2) where θA is the amount of surfactant, ϕisurf and ϕfsurf the initial and final surfactant concentrations, respectively, Vsurf and ρsurf the surfactant volume and density, respectively, and Wrock the weight of rock. Further details are in Alvarez et al. (2017). Error bars are based on the UV-Vis spectrophotometer precision of 0.05 light absorbance for the range of wavelength utilized in the experiments. This error was used when calculating surfactant adsorption. Spontaneous Imbibition Experiments Monitored by CT Scan Methods The potential of surfactants in imbibing Wolfcamp and Eagle Ford cores and recover liquid hydrocarbons is studied by spontaneous imbibition experiments. Wolfcamp and Eagle Ford cores were aged for more than 6 months at experimental temperature to reconstitute them with the missing liquid hydrocarbons due to sample handling. Modified Amott cells were designed to allow their use on the CT scanner. The modified Amott cell consists of a temperature resistant glass base with a graduated measuring scale at the top to measure oil recovery with time. Error bars are assigned based on the measuring scale precision of 0.01 ml. Then, this error was used when calculating oil recovery as function of the OOIP. Cores were placed horizontally in a Plexiglas core base, at the bottom of the cell, to trace radial fluid imbibition toward the core center. The same surfactants used in previous experiments and shown in Table 2 as well as water without additives were tested at field-used concentration of 2 gpt inside an environmental chamber to have a constant temperature of 165 °F for the Wolfcamp samples and 180 °F for the Eagle Ford samples. Further details are in Alvarez and Schechter (2017). In order to observe fluid penetration into the ULR core in real time, a Toshiba Aquilion TSX-101A CT scanner was used as X-rays produce tomographic images of specific areas of the cores, allowing us to see inside them. Helical scans were set on 135 kV and 350 mA with a rotation time of one second and a slice thickness of 0.5 mm with intervals between each slice of 0.3 mm. Initial and final core wettability was determined by CA measurements, and the cores’ average initial and final CT numbers were calculated using CT scan methods. The cores were scanned periodically throughout the experiments to trace fluid SPE-187176-MS 9 imbibition. Moreover, oil production was monitored recurrently using the graduated scale on the modified Amott cell. CT data was processed using an open source software package called ImageJ to generate colorcoded relative density images, and fluid imbibition as penetration magnitude was calculated using the core average initial and average final CT numbers as determined by Eq. 3 (Alvarez et al. 2014). (3) Numerical Modeling of Experimental Results Upscaling the result from the laboratory-scale to field-scale provides a more thorough view on the effect of surfactant imbibition on the oil recovery in shale oil reservoir. In this work, we used the commercial reservoir simulator CMG®. Numerical modelling was divided into two parts, the laboratory-scale and the field-scale. The first part of the simulation was done to model the results from the spontaneous imbibition experiments. A grid model of the core plug used on the tests was constructed and the experiment was modeled by placing the core grid in the middle of a water bath with the same laboratory conditions. Then, history matching on different properties was done to obtain the oil production curve observed in the experiments. All imbibitionrelated properties applied on the core-scale model that provide the best match were applied on the fieldscale model. The field-scale system was a mechanistic model consisting of single hydraulic fracturing stage with the starting conditions that mimic the reservoir on shut-in stage after hydraulic fracturing. The change in capillarity caused by surfactants was modeled by providing two capillary pressure curves representing the two-wettability and IFT stages. For each grid block, a new capillary pressure curve was calculated by averaging the two curves with the amount of surfactant adsorbed as the weighting value. Adsorption data of each surfactant was used in the function of surfactant concentration. Shift from the initial capillary pressure curve to the final curve changed capillary equilibrium, resulting in grid water imbibition accompanied by oil release. Hence, capillary pressure curve was constructed by matching the oil recovery from the numerical simulation to the laboratory experiment result. Next, to model the heterogeneity of the core used in the spontaneous imbibition experiment, the core grid was constructed using the CT-scan rock digitalization method. Each core was scanned using the CTscan, which produces a matrix of CT numbers that can be used to reconstruct the core digitally. The CT number matrix was converted to density matrix (ρ) using Eq. 4 (Massicano et al. 2009), and by converting these matrices of CT number to matrices of density, the porosity distribution (ϕ) of the core was derived by Eq. 5, where ρr and ρfl are the density of the rock and fluid, respectively. We assumed that ρfl is equal to zero due to the rigorous cleaning process that the core undertook before CT scanning. The cleaning process consisted of performing dean stark with the core plug in toluene and methanol consecutively followed by vacuum drying for an extensive period. (4) (5) A mechanical model was used for the field-scale simulation. The model consisted of single stage hydraulic fracture on a dual porosity grid system. We assume the surfactant solution was introduced into the reservoir during the hydraulic fracturing process as completion fluid; therefore, the initial condition of the simulation was set to begin at well shut-in. Then, surfactant were added into both fracture and matrix system, where we assumed that fractures were filled with only completion fluid and matrix had a mixture of completion fluid and in-place brine resulting in the reduction of surfactant concentration in the matrix system. The simulation was run with the assumption of 30-days shut-in followed by production for three years period using the water without surfactant as the base case and surfactants to compare the oil production in term of peak oil rate, 1-year cumulative recovery, and 3-years cumulative recovery on a single well basis. 10 SPE-187176-MS Results and Discussion The results and observations from the proposed correlated set of experiments to evaluate surfactant performance in altering wettability and IFT, adsorbing on the rock surface and recovering hydrocarbons from Wolfcamp and Eagle Ford ULR are presented in this section. In addition, results are discussed to show consistency in the workflow proposed and reproduced by reservoirs modeling tools to upscale the experimental results to the field. Contact Angle Results CA measurements were performed in carbonate samples from the Wolfcamp (well W-1) and Eagle Ford (well EF-1) as described in Table 1. Original wettability for both ULR as well as the extent of wettability alteration by the use of surfactants, described in Table 2, are shown in Fig. 1. Wolfcamp initial CA shows its original wettability as oil to intermediate-wet with a CA of 121°, as shown in Fig. 1, left. Similarly, Eagle Ford initial wettability, shown in Fig. 1, right, is also oil to intermediate-wet (CA of 127°). This initial rock wettability of oil to intermediate-wet is due to the mixture of organic and inorganic matter in ULR. Wolfcamp and Eagle Ford samples have TOC of 5 to 6.5 wt.% as an indication of the presence of organic matter. This organic matter is oil-wet whereas the inorganic matter is water-wet and their combination gives ULR surface intermediate toward oil-wetting affinity. These results are consistent with other researchers who measured Wolfcamp, Bakken and Eagle Ford original wettability by NMR methods (Odusina et al. 2011) and contact angle methods (Alvarez et al. 2017; Alvarez and Schechter 2016b; Alvarez and Schechter 2017; Morsy and Sheng 2014; Nguyen et al. 2014). Figure 1—CA results for well W-1 (left) and well EF-1 (right) for four surfactants at concentration of 2 gpt. After determining ULR initial wettability, surfactants were added, at concentration on 2 gpt, to evaluate their effect on wettability alteration. The results showed that all surfactants were capable of altering Wolfcamp and Eagle Ford rock surface wettability from oil and intermediate-wet to water-wet. In addition, both reservoirs presented similar performance for different groups of surfactants. Cationic surfactant altered wettability of both Wolfcamp and Eagle Ford samples in higher amount. Bear in mind that among the surfactants tested, the cationic is the only one that has a positive charge on its head, and all others (anionic 1, anionic 2, and anionic-nonionic) are mostly anionic. We suggest that the cationic surfactant performed better than the other surfactants due to the electrostatic interactions between its positively charged heads and the negatively charged oil compounds, mostly acid compounds, attached to positively charged carbonate surfaces presented in both Wolfcamp and Eagle Ford rocks. Hence, wettability alteration takes place when the oil molecules attached to the rock surface are stripped and moved to the oil phase. Similarly, negatively charged surfactants such as anionic 1, anionic 2 and anionic-nonionic lacked these electrostatic interactions, changing CA in lesser amounts by hydrophobic interactions. In addition, the presence of nonionic surfactant in the anionic-nonionic blend improved its efficacy as compared to the anionic surfactants alone. For the SPE-187176-MS 11 Wolfcamp, the effect of electrostatic interactions can be confirmed when revising the findings reported by Alvarez and Schechter (2017). In that case, Alvarez and Schechter (2017) used siliceous cores from the Wolfcamp showing that anionic surfactants were better in altering CA than positively charged surfactants. The results presented in this study showed better performance in carbonate cores with positively charged surface-active agents. In summary, additive the cationic surfactant reduced CA more in carbonate cores from Wolfcamp and Eagle Ford. These findings suggest that lithology and surfactant type have a direct impact on surfactant performance in altering rock wettability. Next, to determine the charges of the surfactant solutions as well as the stability of the water films on the Wolfcamp and Eagle Ford rock as an indication of wettability alteration, zeta potential experiments are performed. Zeta Potential Results Zeta potential measurements were performed to further study wettability alteration in these ULR. Film aqueous solution instability, referenced as zeta potential values between −30 to + 30 mV, is an indication of intermediate or oil-wetness whereas zeta potential values more positive than +30 mV or more negative than −30 mV can be interpreted as stable indicating water-wet behavior. In addition, this technique gives us an indication of the surfactant electrostatic charges. Zeta potential results are shown in Fig. 2. Initially, zeta potential values for water alone for Wolfcamp (Fig. 2, left) and Eagle Ford (Fig. 2, right) samples showed an unstable water film as an indication of intermediate to oil-wet behaviors. These results are consistent with original wettability determined by CA. Figure 2—Zeta potential results for well W-1 (left) and well EF-1 (right) for four surfactants at concentration of 2 gpt. Conversely, when surfactants were added at a concentration of 2 gpt, aqueous film stability on the rock surface increased as a sign of water wetness. In addition, zeta potential values showed the nature of the additives evaluated. Negatively charged surfactants (anionic 1, anionic 2 and anionic-nonionic) showed negative zeta potential magnitudes whereas the cationic surfactant demonstrated its positive charges as its zeta potential values were found to be positive. These changes in zeta potential values when surfactants are added to water are consistent with CA measurements as wettability alteration, and they may favor imbibition in these Wolfcamp and Eagle Ford rocks by shifting capillary pressure signs from negative to positive. In the next section, the effect of surfactants in reducing IFT is studied. IFT Measurement Results Fluid-fluid interactions between Wolfcamp and Eagle Ford oil and aqueous solutions with and without surfactants were investigated via IFT measurements. The same surfactants used in CA and zeta potential experiments were used at a concentration of 2 gpt. IFT results are shown in Fig. 3. Original IFT between water and Wolfcamp oil had an initial value of 21.8 mN/m (Fig. 3, left) whereas Eagle Ford oil showed 12 SPE-187176-MS an initial IFT of 34.4 mN/m (Fig. 3, right). Then, as surfactants were added to the aqueous solutions, oilbrine IFT was reduced, in some case by more than one order of magnitude. This reduction in IFT is due to the alignment of surfactant molecules on the oil-brine interface. As a result of its amphiphilic nature, surfactant molecules placed themselves at the interface facing the different phases, the tail group with the hydrophobic oil phase and head group with the hydrophilic water phase. Thus, the surfactant lowered the surface energy and decreasing IFT. Figure 3—IFT results for well W-1 (left) and well EF-1 (right) for four surfactants at concentration of 2 gpt. Fig. 3 also shows that anionic and anionic-nonionic surfactants reduced IFT the most, being the surfactant anionic 1 the one with the largest reduction. This is caused by the presence of sulfonates on their formulation. Sulfonates are highly polar and negatively charged functional groups which reduce IFT by creating a strong affinity between the surfactant head and the water phase and consequently a poor arrangement in the wateroil interface. On the other hand, the cationic surfactant showed the lowest reduction in IFT, which is a typical behavior by surfactants that have quaternary ammonia compounds on their composition. Hence, the IFT results indicate that surfactant efficacy in reducing IFT depends on the surfactant nature. In addition, the oil type plays a role in fluid-fluid interactions as IFT reduction. As reported on the methodology section, Wolfcamp oil TAN and TBN are 0.09 and 0.12 mg KOH/g oil, respectively whereas Eagle Ford oil TAN and TBN are 0.02 and 0.61 mg KOH/g oil, respectively. These values suggest that Wolfcamp oil is slightly basic and Eagle Ford oil is basic in nature. These positively charged Wolfcamp and Eagle Ford oil interact better with anionic surfactants, as negatively charged compounds, favoring IFT reduction. IFT reduction by surfactants also favors wettability alteration in these ULR. Due to ULR very small pore sizes (0.005 microns for Wolfcamp and 0.007 for Eagle Ford), capillary pressures are high and the effect of decreasing IFT by surfactants aids fluid imbibition and consequently surfactant interaction with the rock surface. Once in contact with the rock, surfactants solubilize the oil attached to the rock altering wettability. This is the main reason why all surfactants used in these study, regardless of their charge, altered wettability at the concentrations tested. Thereby, a moderate IFT reduction is desired to favor wettability alteration in tight pores and reduce capillary pressures. However, contrary to conventional EOR techniques such as surfactant flooding, in ULR ultralow IFT should be avoided to prevent total elimination of capillary forces, which plays a vital role in imbibition and oil recovery in ULR. Surfactant Adsorption Measurements Surfactant dynamic adsorption was measured by evaluating the difference in surfactant concentration before and after the specified time of reaction. The concentration was measured using the UV-Vis spectrophotometer, which requires a calibration curve correlating the light adsorption on each surfactant's specific wavelength with the surfactant concentration to be constructed first. Fig. 4 shows the measured calibration curve that was used in the dynamic adsorption experiments for the four surfactants tested. In addition, Fig. 4 shows that light adsorption and surfactant concentration are related in a linear trend where SPE-187176-MS 13 the trend deflects as it approaches certain concentration, mostly between 1.5 to 2.0 gpt. We suggest that this trend change in is caused by the aggregation of the surfactant molecules to form micelles. These micelles have different light adsorptions from the individual surfactant molecule. Thus, the change in trend line was observed as surfactant critical micelle concentration (CMC) was reached. Figure 4—Calibration curve correlating the amount of light adsorbed to the concentration of the four surfactants used in the experiment. Next, surfactant adsorption on rock samples from wells W-1 (Wolfcamp) and EF-1 (Eagle Ford) is shown on Fig. 5. For well W-1, the cationic surfactant adsorbed the most, followed by anionic 1 and anionicnonionic surfactant, then surfactant anionic 2 as the least adsorbed on W-1 surface. Most of the adsorption took place in the first six hours of reaction time followed by either a plateau or more adsorption with slower rate. This trend showed that a reduction of adsorption site occurred implying that most of the surfactants were adsorbed on the rock surface in a monolayer configuration with some surfactant-surfactant interaction as the adsorption continues but at slower rate. Figure 5—Surfactant dynamic adsorption for well W-1 (left) and well EF-1 (right) for four surfactants. 14 SPE-187176-MS Surfactant adsorption highly depends on both the additive type and rock composition. The results showed that the cationic surfactant adsorbed with substantially higher quantity compared to the other surfactants tested. By correlating these results to the rock mineralogical analysis shown on Table 1 and the composition of the surfactants shown on Table 2, we suggest that the amount of clay in the sample could be the cause of higher cationic surfactant adsorption. The quaternary ammonia compound in the cationic surfactant adsorbed strongly on clay mineral, especially in those belong to illite group (Sánchez-Martín et al. 2008). On the other hand, both surfactants anionic 1 and anionic-nonionic were adsorbed more compared to surfactant anionic 2 since they both have sulfonate compounds which known to be reactive to the carbonate mineral in Wolfcamp and Eagle Ford samples (Sheng 2015). The result of surfactant dynamic adsorption on sample EF-1 are shown on Fig. 5, right. Similar adsorption results on well W-1, the cationic surfactant adsorbed the most, followed by anionic 1, anionic-nonionic, and anionic 2. The adsorption on EF-1 also showed the significance of rock and surfactant composition to surfactant adsorption. As shown on Table 1, rock sample EF-1 has more than 30 wt.% clay mineral which explains the high adsorption nature of the cationic surfactant. High carbonate content would also explain the relatively high adsorption showed by the anionic surfactants as sulfonate compound adsorbed strongly on carbonate minerals. Surfactant adsorption would cause loss of these additives in the reservoir. Different surfactant composition and rock mineralogy have a considerable influence on the amount of surfactant adsorbed on the rock. Therefore, when designing stimulation treatments using surfactants as additives, surfactant adsorption must be considered. Next, we evaluate the impact of wettability and IFT alteration as well as surfactant adsorption in recovering oil by spontaneous imbibition experiments. Spontaneous Imbibition Experiment Results Monitored by CT Scan Methods In the previous sections, we demonstrated that surfactants are capable of altering wettability, reducing brineoil IFT and adsorbing in ULR samples from the Wolfcamp and Eagle Ford. These changes in wettability and IFT modify capillary pressures. Thus, in this section, we are going to evaluate the impact of these variables in oil recovery from Wolfcamp and Eagle Ford sidewall cores by spontaneous imbibition experiments. Table 3 shows the well, core dimensions, initial properties, and type of fluid used. These values, along with initial water saturations, are used to calculate core initial oil in place (OOIP) and to relate the oil recovered as experimental time elapsed, to the amount of oil initially on each core. Initial water saturations were determined by mercury intrusion and extrusion analysis and the results showed a Wolfcamp initial water saturation (Swi) for well W-1 of 0.15 whereas Eagle Ford, well EF-1 initial water saturation of 0.1. In addition, Table 3 also exhibits the initial CA as wettability measurement for the cores used showing that all cores are initially intermediate and oil-wet due to the extended aging period. Table 3—Properties of cores used in spontaneous imbibition experiments ULR Wolfcamp Eagle Ford Well W-1 EF-1 Core Porosity (%) Diameter (in) Length (in) Initial CA (°) Type of Fluid 1 6.5 0.974 1.956 142.8 Anionic 1 2 6.8 0.990 2.037 144.8 Anionic-nonionic 3 6.5 0.974 1.887 137.3 Anionic 2 4 6.8 0.989 1.777 130.1 Cationic 5 6.5 0.990 1.451 143.1 Water 6 12 0.998 1.453 132.5 Anionic 1 7 13 0.994 1.959 126.0 Anionic-nonionic 8 12 1.001 2.214 123.4 Anionic 2 9 12 1.002 2.391 130.5 Cationic 10 12 0.995 2.585 131.1 Water SPE-187176-MS 15 Imbibition experiments for the Wolfcamp formation were performed with cores from well W-1. Cores 1 to 5 were submerged in aqueous solution with and without surfactants as indicated in Table 3. When surfactants were used, the concentration utilized was 2 gpt. Expelled oil from the Wolfcamp cores was collected at the top of the modified Amott cell with time to construct the recovery curves shown in Fig. 6. Consistent with CA, zeta potential, IFT and adsorption experiments, spontaneous imbibition results for the Wolfcamp carbonate cores clearly indicate that aqueous solution with surfactants as additives recovered more oil than water alone. This correlates with the potential of surfactants of altering wettability (Fig. 1, left) and moderately reducing IFT (Fig. 3, left). IFT reduction favored water penetration in the ULR pores shifting wettability and capillary pressure sign. Capillary pressure not only changed from negative to positive, but also its value was reduced due to IFT alteration. Hence, this combined effect in capillary pressure favored water imbibition and oil expulsion from the core in a countercurrent movement. Figure 6—Oil recovered for Wolfcamp well W-1 by spontaneous imbibition. In addition, the core submerged in positively charged additive cationic surfactant (Core 4) produced more oil than the other cores in different surfactants. We suggest that this better performance is caused by electrostatic interactions as determined in zeta potential and adsorption experiments. Rock-fluid interactions between the negatively charged oil compounds attached to the Wolfcamp carbonate samples and positively charged cationic surfactant heads remove oil from the core surface, favoring water imbibition; thus, oil recovery. These rock-fluid interactions are also evidenced in the time in which cores began to produce oil. Conversely, the cores submerged in anionic 1 and anionic-nonionic surfactants (cores 1 and 2), which are the most negatively charged surfactants, as determined in zeta potential experiments (Fig. 2, left), recovered the least amount of oil among all surfactants. These results also confirm the impact of electrostatic interaction in surfactant efficacy where repulsive force between the negatively surfactant heads and negative charges of the oil attached to the positive core surface prevented oil on the rock to fully leave the pores. As shown in Fig. 6, the core 4 (cationic surfactant) not only recovered more oil, but also began producing oil a few hours before the other cores demonstrating that wettability alteration took place faster and more effectively. In the same way, it can be seen that core 5 (water) began to produce oil more than 18 hours after the cores in surfactants. This is because water alone is not capable of shifting wettability and reducing IFT. At the end of the experiments, Core 4 (cationic surfactant) recovered 47.3% of the OOIP, followed by core 3 (anionic 2) with 32.6%, and cores 1 and 2 (anionic 1 and anionic-nonionic) with 24.3% and 18.9% of the OOIP, respectively. Finally, aided only by gravity forces, the core submerged on water alone (core 16 SPE-187176-MS 5) marginally recovered only 7.6% of the OOIP due to water lack of potential of changing wettability or reducing IFT. Moreover, the oil recovered as a function of the OOIP from Eagle Ford ULR cores is shown in Fig. 7. Similar to the results from the Wolfcamp cores, the cores in contact with surfactants had higher oil productions than the one in water alone due to additives capability of altering wettability (Fig. 1, right) and reducing IFT (Fig. 3, right). The high carbonate content in the Eagle Ford samples favored better electrostatic interactions between the positively charged cationic surfactant and the negatively charged oil compounds attached to the carbonate surface. Thus, core 9 (cationic surfactant) recovered the highest amount of oil among all Eagle Ford cores tested. Also, core 9 began to produce in less than one hour whereas other cores began to expel oil two or more hours after. Cores 6 to 8 (anionic 1, anionic-nonionic and anionic 2) recovered less oil than core 9 due to electrostatic repulsions that retarded and repressed an effective interaction of surfactants with the rock-oil system. Therefore, wettability alteration and IFT reduction induced by surfactants aided core 9 (cationic surfactant) to have a final recovery of 9.0% of the OOIP whereas cores 6 to 8 recovered 6.5%, 4.5% and 5.8%, respectively. Lastly, consistent with Wolfcamp results, the core submerged in water alone (core 10) produced only 2.1% of the OOIP due to negative capillary pressures that prevented oil recovery by imbibition. Figure 7—Oil recovered for Eagle Ford well EF-1 by spontaneous imbibition. From the oil recovery results in both the Wolfcamp and Eagle Ford cores, we highlight two important findings. First, the potential that surfactants have when added to completion fluid to improve oil recovery by imbibition due to their capability of altering wettability and moderately reducing IFT. Second, the vital role in imbibition of rock-fluid electrostatic interactions. From spontaneous imbibition results, it can be clearly seen that the positively charged cationic surfactant recover oil faster and in higher amounts than the other chemical tested regardless of its lower IFT reduction compared to anionic surfactants. Next, CT scan technology was utilized to observe spontaneous imbibition inside the Wolfcamp and Eagle Ford cores. Also, CT scans were used to correlate oil recovery with water imbibition. To that end, the modified Amott cells were periodically scanned during the experiments to see fluid movement inside the cores. The resulting CT numbers can be related to water imbibition by knowing the CT numbers of the phases present. Thus, Wolfcamp and Eagle Ford oil CT number is close to −150 and −100 HU, respectively, and aqueous solutions CT number are approximately 800 HU due to the use of KI as dopant. The CT number difference between the oil and the aqueous solutions allow us to trace imbibition into the cores while water occupies pores originally filled with oil. Hence, in our experiments, fluid imbibition is reached when SPE-187176-MS 17 increasing CT numbers are observed. CT scan images at progressive times for Wolfcamp cores (cores 1-5) during spontaneous imbibition experiments are shown in Fig. 8. Consecutive images for Wolfcamp cores submerged in surfactant solutions (cores 1-4) showed visible changes in colors as CT numbers increased with time. Color changes from red to green and dark blue to light blue demonstrate water imbibition and consequently oil displacement. Contrarily, the core in water without surfactants (core 5) showed lesser changes as an indication of limited water penetration. These observations qualitatively agreed with oil recovery profiles showed in Fig. 6, in which cores submerged in surfactant solutions produced up to four times more oil than the core in water alone, confirming our theory that oil recovery increases when water imbibition is promoted by wettability and IFT alteration. Figure 8—CT images for Wolfcamp, well W-1 as a function of time in spontaneous imbibition experiments. Furthermore, CT scan images with time for Eagle Ford cores (cores 6-10) during spontaneous imbibition experiments are shown in Fig. 9. Same as Wolfcamp cores, consecutive CT images showed clear changes in color for cores submerged in surfactant solutions (cores 6-9) changing from red to green and dark/light blue to yellow as evidence of water imbibition. On the other hand, the Eagle Ford core submerged in water alone showed inferior water imbibition as determined by timid color changes. In addition, CT scan results for cores 6 to 10 shown in Fig. 9 correlate with the oil recovery performance in Fig. 7 where higher imbibition in cores 6 to 9 led to higher oil recovery for cores in surfactant solutions whereas lower imbibition in core 10 resulted in lower oil recovery. 18 SPE-187176-MS Figure 9—CT images for Eagle Ford, well EF-1 as a function of time in spontaneous imbibition experiments. CT scan images allow seeing fluid movement inside the cores as well as core heterogeneities such as bedding planes and natural fractures. ULR commonly present with these types of heterogeneities that dominate fluid flow and oil production in shale systems. Hence, as shown in Fig. 8 and Fig. 9, fluid flow was not radially homogeneous towards the core center. These heterogeneities present an important challenge when attempting to model laboratory results and upscaling them to the field. For that reason, our numerical modeling considers these heterogeneities by creating the grids from CT scan images. By CT scan technology, we observed that imbibing fluids moved throughout less resistant pathways inside the core, displacing hydrocarbons. In addition, during the experiments, countercurrent fluid movement was evidenced where oil was expelled from the core surface as water imbibed due to changes in capillary pressure. The results for the spontaneous imbibition experiments performed in Wolfcamp and Eagle Ford cores are summarized in Table 4. Aqueous solution imbibition was quantified as penetration magnitudes using CT scan technology and Eq. 3. For Wolfcamp cores (cores 1-5), penetration magnitude were clearly higher for cores submerged in surfactant solutions (core 1-4) compared to core in only water (core 5). Moreover, the Wolfcamp core that exhibited the highest penetration magnitude (core 4) also had the highest oil recovery confirming the hypothesis that higher imbibition leads to higher oil recovery in these ULR cores. In addition, the role of rock surface and surfactant charges as rock-fluid interactions was evidenced when comparing these carbonate core Wolfcamp results to the siliceous Wolfcamp results documented in Alvarez and Schechter (2017). Alvarez and Schechter (2017) showed that Wolfcamp siliceous cores had higher penetration magnitudes (imbibition) and oil recoveries when negatively charged surfactants (anionic surfactants) where used whereas our results showed that Wolfcamp carbonate cores exhibited higher penetration magnitudes and oil recoveries when positively charged surfactants are used (cationic surfactant). Similarly, the Eagle Ford cores exposed to surfactant additives (cores 6-9) showed higher penetration magnitudes and oil recovery than the cores in water without additives (core 10). Next, final SPE-187176-MS 19 IFT values were also measured showing that aqueous solutions with surfactants reduced IFT in almost two orders of magnitude for anionic and anionic blended surfactants and one order of magnitude for the cationic surfactant. It is important to highlight that IFTs were not reduced to ultra-low values as conceived in conventional EOR. In ULR, capillary forces play a big role in imbibition and oil recovery; hence, they should not be eliminated by reaching ultra-low IFT values. Table 4—Spontaneous imbibition experiment results Core Type of Fluid Penetration magnitude (HU) Final IFT (mN/m) Final CA (°) Initial Pc (psi) Final Pc (psi) Oil Recovery (%OOIP) Wolfcamp (Well W-1) 1 Anionic 1 28 0.4 45.6 -1007 16 24.3 2 Anionicnonionic 22 0.9 47.4 -1033 35 18.9 3 Anionic 2 35 3.9 48.7 -929 149 32.6 4 Cationic 42 8.9 38.1 -815 406 47.3 5 Water 16 21.8 89.9 -1011 2 7.6 Eagle Ford (Well EF-1) 6 Anionic 1 18 0.7 47.2 -963 20 6.5 7 Anionicnonionic 15 1.2 53.4 -838 30 4.5 8 Anionic 2 17 2.3 48.3 -785 63 5.8 9 Cationic 21 6.9 34.3 -926 236 9.0 10 Water 9 34.4 89.5 -937 12 2.1 Regarding wettability alteration determined by contact angle methods, Table 4 shows rock wettability before and after spontaneous imbibition experiments. All cores from Wolfcamp (cores 1-4) and Eagle Ford (cores 6-9) submerged in surfactant solutions altered wettability from oil-wet (Table 3) to waterwet; whereas Wolfcamp (core 5) and Eagle Ford (core 10) cores submerged in water without surfactants did not change the CA significantly enough to change wettability to water-wet. Surfactants capability of reducing IFT and altering wettability allowed capillary pressures to shift from high negative values to moderate positive magnitudes as shown in Table 4 and calculated using Eq. 1. These fluid-fluid and rockfluid interactions in ULR favored changes in capillary forces for cores exposed to surfactants, favoring imbibition and improving oil recovery as demonstrated in the last column of Table 4. Conversely, the cores submerged in water alone were barely able to alter capillary pressure sign to positive, thus marginally favoring imbibition and consequently insignificantly recovering oil by the aid of fluid densities difference as gravity forces inside the modified Amott cell. In summary, the proposed set of correlated experiments used in this investigation allows us to evaluate in the laboratory the impact of surfactant imbibition and adsorption on improving oil recovery in the Wolfcamp and Eagle ford unconventional reservoirs. This study evaluated rock-fluid and fluid-fluid interactions of aqueous solution, with and without surfactants, and their impact on imbibition and oil recovery. The results showed that wettability was altered and IFT was moderately reduced when chemical additives are used in completion fluid. These wettability and IFT alterations changed capillary forces favoring water imbibition and oil explosion from ULR cores. In addition, these set of experiments give us the necessary parameters to be used in numerical simulation to model laboratory results and upscale these findings to the field as shown in the next section. 20 SPE-187176-MS Numerical Modeling of Experimental Results With all the data gathered from the correlated set of experiments showed previously, a numerical simulation to model the process of surfactant imbibition was created. The goal of the numerical modelling is to provide an estimation for oil recovery aided by fluid imbibition when using aqueous solutions with and without surfactants on a field-scale application. Upscaling was done by implementing the capillary pressure curve built from history-matching our laboratory data to a mechanistic field-scale model. Spontaneous imbibition results from Eagle Ford core 10 (water) and core 9 (cationic surfactant) were used on the numerical simulation. The objective is to compare oil productions from a core without chemical additives (core 10) to create a baseline and a core submerged in surfactants to assess improved oil recovery by imbibition. Since the cationic surfactant was observed to give the best performance in producing oil on the Eagle Ford formation (Fig. 7), the numerical modelling was done on upscaling these results, along with the water without additive results, to the field-scale. The core grid model was built using CT scan based rock digitalization method to ensure the incorporation of core heterogeneities into the model. As shown in Fig. 8 and Fig. 9, ULR cores are highly heterogeneous, thus, in order to guarantee the validity of the rock grids, rock digitalization from CT scan images was considered. From Table 3, Eagle Ford cores 9 (cationic surfactant) and 10 (water) were used on the numerical simulation. To that end, Fig. 10 shows the comparison of the grid model and the CT scan 3D reconstruction images for core 10 and core 9. Moreover, Fig. 10 shows marked heterogeneity in both cores, emphasizing the need of building the numerical simulation with the original cores used on the laboratory. By employing CT scan technology, we successfully captured core heterogeneities in the grid model. The color on the grid model represents porosity whereas the color on the CT scan image represents the CT number, which is linearly related to density. Therefore, the inverted relation between density and porosity causes the color difference between the grid and the CT image (Eq. 4 and Eq. 5). As shown on the color-scale, darker color on the CT scan image shows lower density, which means a higher porosity, represented by brighter color on the grid model. Figure 10—Grid models and CT scan images for the Eagle Ford core 10 (left) and for core 9 (right). Next, history-matching for the laboratory result was done in two steps. First, to obtain the original unaltered capillary pressure curve core 10 was used. Then, we used core 9 to obtain the cationic surfactant altered capillary pressure curve. Capillary pressure curve endpoints on both cases were calculated from the measured IFT and contact angle of Eagle Ford oil and rock samples using water and the cationic surfactant, SPE-187176-MS 21 as shown in Table 3 and Table 4. Thus, the history-match method was only used to build the curvature of the capillary pressure profile throughout the saturation change. Fig. 11 shows the best-matched capillary pressure curve and oil production curve from our laboratory experiment and numerical simulation. As expected, higher positive capillary pressures were evidenced when the cationic surfactant was used. Another prominent comparison of the two curves is the intersection point at capillary pressure equal to zero, which for the cationic surfactant is further to the right compared to the unaltered capillary pressure curve. We believe that this difference is the manifestation of wettability alteration in our numerical model, and we predict that stronger wettability alteration would move the intersection further to the right and weaker alteration would move the intersection to the left. Moreover, production curve of both cases showed a good agreement between numerical simulation and laboratory result with some difference in the middle part of the curve. We believe that these differences are due to immeasurable amount of oil stuck onto the core walls that was released in later times of the experiments by gravity forces. Figure 11—Capillary pressure curves for Eagle Ford water and cationic surfactant (left) and simulation and spontaneous imbibition experiment result for Eagle Ford water and cationic surfactant (right). Next, with the capillary pressure curve successfully constructed from the core-scale simulation, we applied the same modelling mechanism to the field-scale model. The mechanistic model used properties collected from the general average properties that are commonly observed on the Eagle Ford. To save running time, the mechanistic model consists of only one stage of hydraulic fracture and the result will be multiplied by twenty assuming the well was completed with twenty stages. The dimension, configuration and main properties of the model are on Fig. 12. Figure 12—Mechanistic model schematic and field-scale model properties. The results of the field-scale simulation are shown in Fig. 13. Well performance using completion fluids with and without cationic surfactant was simulated. The results showed that adding the cationic surfactant 22 SPE-187176-MS to the frac fluid had a positive impact on oil rate and cumulative oil production of the well. In fact, the cationic surfactant increased the peak oil rate of the well by 24% resulting in 8% addition of cumulative oil recovery for a three years production period. Figure 13—Comparison of oil cumulative and rate of production of well completed with and without cationic surfactant. By examining the change in oil saturation distribution during the whole simulation runtime, we noticed that the mechanism of production enhancement is consistent with what we observed in the laboratory experiment, imbibition of water coexisting with oil expulsion. However, this process mostly occurred during the shut-in period. At this stage, an exchange of oil and water occurred where the reservoir matrix took water and oil was expelled out to the fracture system. After the shut-in period was finished, this mechanism was overshadowed by the pressure drainage activity of the well. Therefore, although that 24% increment in peak oil rate and 8% addition of three years cumulative oil production is comparatively significant on this ultra-tight reservoir, we believe that additional increment is possible on a longer shut-in time. A more insightful study on the effect of shut-in time and different operating condition will be covered on a future work. Nevertheless, the result of the field-scale numerical simulation in this work proved our hypothesis that by modifying the capillary state of the oil, water, rock system in a ULR by altering wettability and moderately reducing IFT with chemical additives, an improvement of oil recovery can be achieved. Conclusions The results from this work have practical implications on the design of stimulating fluids to improve oil recovery in ULR. Thus, completion fluid additives should be carefully selected while taking into consideration surfactant, oil and rock type. These considerations can reduce completion costs and improve oil recovery after flowback as compared to adding an unknown chemical that may not effectively promote imbibition into the rock. Moreover, experimental and simulation results showed that soaking and flowback schedules may be beneficial when using surfactants. Lastly, field-testing is recommended to reproduce laboratory and simulation results. However, laboratory and field results may differ due to ULR heterogeneities, rock lithologies and fluids proportions. Nevertheless, the comprehensive correlated set of experiments and results reported in this study can serve as a prescreening tool as the beginning of a successful field trial. However, we must emphasize that these wettability studies encompass a tiny fraction of the heterogeneity that truly exists in the field. Subtle improvements in petrophysical properties from one SPE-187176-MS 23 core sample to the next may significantly enhance or reduce imbibition rate relative to other tests. Therefore, instead of observing a chemical phenomenon, we may simply be lucky on the petrophysical properties of one sample to the next. The unfortunate fact of not being able to measure standard values like permeability will always limit the efficacy of these tests. However, creation of new techniques like the ones described in this manuscript and given enough benchmarking from different wettability scenarios will doubtless lead to providing useful methods for industry to screen various chemicals additives in the near future. In this paper, we have concluded: • • • • • • • • • • Original wettability, measured by CA methods, for the Wolfcamp and Eagle Ford cores showed intermediate to oil-wet affinity due to ULR mixture of oil-wet organic pores and water-wet inorganic pores. All surfactants tested, at concentrations on 2 gpt, altered the wettability of the shale samples from intermediate and oil-wet to water-wet, but wettability alteration strongly depended on rock mineral composition and surfactant type. Consistent with wettability measurements by CA, zeta potential results showed lower double layer stability for rock in contact with aqueous solutions without surfactants. Conversely, more stable water films on rock surface, and consequently more water-wetness, were evidenced in aqueous solution with surfactants. All surfactants added to completion fluids moderately reduced IFT. Anionic surfactants decreased IFT further than cationic surfactants. The relation between rock mineralogy and surfactant charges determines surfactant adsorption on the rock. This information is crucial in determining the selection of the additive on different reservoir. The cationic surfactant was found to be adsorbed the most compared to other chemicals tested on both Eagle Ford and Wolfcamp rock samples. Spontaneous imbibition results showed higher imbibition and improved oil recoveries for cores submerged in surfactants compared to water alone. This better performance was achieved by wettability and IFT alteration, which changed capillary forces to favoring water imbibition and releasing trapped hydrocarbons in the rock pores. Spontaneous imbibition using surfactants occurs on impressively rapid time scales despite immeasurable permeability. Changes in wettability and IFT reductions induced by surfactants in completion fluids improve matrix penetration favoring imbibition and increasing oil recovery from the Wolfcamp and Eagle Ford ULR. In spontaneous imbibition experiments, the majority of the oil produced by imbibition was within 3-5 days from the start of experiments. This suggest the possibility of designing and optimizing treatment duration and flowback schedules. Upscaling using numerical simulation provides an insight on the field-scale impact of the surfactant imbibition method. Incorporating surfactant into the stimulation fluid was found to be beneficial as shown by the increment of both initial peak oil production rate and 3-year cumulative oil production. The results from this study give valuable insights on designing a chemically compatible and better performing stimulating fluid at affordable costs, which can recover additional oil in unconventional liquid reservoirs. Acknowledgements The authors would like to thank the Department of Petroleum Engineering, Texas Engineering Experimental Station (TEES), and Crisman Institute for Petroleum Research at Texas A&M University for funding this work. Also Rodolfo Marquez, John Maldonado and Don Coleen for their collaboration on the experimental work. 24 SPE-187176-MS Nomenclature cm = cm3 = CT = CTinitial = CTfinal = ft = HU = gpt = gr = mV = mm = mN/m = Nm = ϕm = ϕf = Pi = Swi = μD = km = kf = khf, int = khf, mod = mg/g = mg/L = μm = ULR = xf = wf = wf, mod = Centimeter Cubic centimeter Computer tomography Average CT number of the core before spontaneous imbibition experiments Average CT number of the at the end of spontaneous imbibition experiments Feet Hounsfield unit Gallons per thousand gallons Grams Millivolts Millimeters Millinewton per meter Nanometers Matrix porosity Fracture porosity Initial reservoir pressure Initial water saturation Micro Darcie Matrix permeability Fracture permeability Hydraulic fracture intrinsic permeability Hydraulic fracture modeled permeability Milligrams per grams Milligrams per liter Microns Unconventional liquid reservoir Hydraulic fracture half-length Hydraulic fracture intrinsic width Hydraulic fracture modeled width Superscripts ° = degrees References Alharthy, N., Teklu, T., Kazemi, H.et al. 2015. Enhanced Oil Recovery in Liquid-Rich Shale Reservoirs: Laboratory to Field. Paper presented at the SPE Annual Technical Conference and Exhibition, 28-30 September, Houston, Texas, USA. Society of Petroleum Engineers. doi: 10.2118/175034-MS. Alvarez, J.O., Neog, A., Jais, A.et al. 2014. Impact of Surfactants for Wettability Alteration in Stimulation Fluids and the Potential for Surfactant Eor in Unconventional Liquid Reservoirs. Paper presented at the SPE Unconventional Resources Conference, 1-3 April, The Woodlands, Texas, USA. Society of Petroleum Engineers. doi: 10.2118/169001ms. Alvarez, J.O., Saputra, I.W.R., and Schechter, D.S. 2017. Potential of Improving Oil Recovery with Surfactant Additives to Completion Fluids for the Bakken. Energy & Fuels 31 (6): 5982–5994. doi: 10.1021/acs.energyfuels.7b00573 Alvarez, J.O. and Schechter, D.S. 2016a. Application of Wettability Alteration in the Exploitation of Unconventional Liquid Resources. Petroleum Exploration and Development 43 (5): 832–840. doi: 10.1016/S1876-3804(16)30099-4 Alvarez, J.O. and Schechter, D.S. 2016b. Wettability, Oil and Rock Characterization of the Most Important Unconventional Liquid Reservoirs in the United States and the Impact on Oil Recovery. Paper presented at the Unconventional SPE-187176-MS 25 Resources Technology Conference, 1-3 August, San Antonio, Texas, USA. Unconventional Resources Technology Conference. doi: 10.15530-urtec-2016-2461651. Alvarez, J.O. and Schechter, D.S. 2017. Wettability Alteration and Spontaneous Imbibition in Unconventional Liquid Reservoirs by Surfactant Additives. SPE Reservoir Evaluation & Engineering 20 (01). doi: 10.2118/177057-PA Anderson, W.G. 1986a. Wettability Literature Survey- Part 1: Rock/Oil/Brine Interactions and the Effects of Core Handling on Wettability. Journal of Petroleum Technology 38 (10): 1125–1144. doi: 10.2118/13932-pa Anderson, W.G. 1986b. Wettability Literature Survey- Part 2: Wettability Measurement. Journal of Petroleum Technology 38 (11): 1246–1262. doi: 10.2118/13933-pa Doman, L. 2015. U.S. Remained World's Largest Producer of Petroleum and Natural Gas Hydrocarbons in 2014. U.S Energy Information Administration. EIA. 2017. Drilling Productivity Report. U.S. Energy Information Administration, April 2017. Guo, B., Schechter, D.S., and Baker, R.O. 1998. An Integrated Study of Imbibition Waterflooding in the Naturally Fractured Spraberry. Society of Petroleum Engineers. doi: 10.2118/39801-MS. Hirasaki, G.J. 1991. Wettability: Fundamentals and Surface Forces. SPE Formation Evaluation 6 (02). doi: 10.2118/17367-PA Massicano, F., Massicano, A.V.F., Silva, N.G.d. et al. 2009. Analysis of Ct and Pet/Spect Images for Dosimetry Calculation. Paper presented at the 2009 International Nuclear Atlantic Conference, Brazil. ASSOCIAÇÃO BRASILEIRA DE ENERGIA NUCLEAR. doi: 978-85-99141-03-8. Mirchi, V., Saraji, S., Goual, L. et al. 2014. Experimental Investigation of Surfactant Flooding in Shale Oil Reservoirs: Dynamic Interfacial Tension, Adsorption, and Wettability. Society of Petroleum Engineers. doi: 10.15530/ urtec-2014-1913287. Morsy, S. and Sheng, J.J. 2014. Surfactant Preflood to Improve Waterflooding Performance in Shale Formations. Paper presented at the Unconventional Resources Technology Conference, 25-27 August, Denver, Colorado, USA. Society of Petroleum Engineers. doi: 10.2118/169519-MS. Nguyen, D., Wang, D., Oladapo, A. et al. 2014. Evaluation of Surfactants for Oil Recovery Potential in Shale Reservoirs. Paper presented at the SPE Improved Oil Recovery Symposium, 12-16 April, Tulsa, Oklahoma, USA. Society of Petroleum Engineers. doi: 10.2118/169085-ms. Odusina, E.O., Sondergeld, C.H., and Rai, C.S. 2011. Nmr Study of Shale Wettability. Paper presented at the Canadian Unconventional Resources Conference, 15-17 November, Calgary, Alberta, Canada. Society of Petroleum Engineers. doi: 10.2118/147371-MS. Sánchez-Martín, M.J., Dorado, M.C., del Hoyo, C. et al. 2008. Influence of Clay Mineral Structure and Surfactant Nature on the Adsorption Capacity of Surfactants by Clays. Journal of Hazardous Materials 150 (1): 115–123. doi: https:// doi.org/10.1016/j.jhazmat.2007.04.093 Schechter, D.S., Denqen, Z., and Orr, F.M., Jr. 1991. Capillary Imbibition and Gravity Segregation in Low Ift Systems. Paper presented at the SPE Annual Technical Conference and Exhibition, 6-9 October, Dallas, Texas, USA. Society of Petroleum Engineers. doi: 10.2118/22594-MS. Schechter, D.S., Zhou, D., and OrrF.M.Jr, 1994. Low Ift Drainage and Imbibition. Journal of Petroleum Science and Engineering 11 (4): 283–300. doi: http://dx.doi.org/10.1016/0920-4105(94)90047-7 Sheng, J.J. 2015. Status of Surfactant Eor Technology. Petroleum 1 (2): 97–105. doi: https://doi.org/10.1016/ j.petlm.2015.07.003 Wang, D., Butler, R., Zhang, J. et al. 2012. Wettability Survey in Bakken Shale with Surfactant-Formulation Imbibition. SPE Reservoir Evaluation & Engineering 15 (06). doi: 10.2118/153853-pa Wang, D., Zhang, J., Butler, R. et al. 2016. Scaling Laboratory-Data Surfactant-Imbibition Rates to the Field in FracturedShale Formations. SPE Reservoir Evaluation & Engineering 19 (03). doi: 10.2118/178489-PA Xu, L. and Fu, Q. 2012. Ensuring Better Well Stimulation in Unconventional Oil and Gas Formations by Optimizing Surfactant Additives. Paper presented at the SPE Western Regional Meeting, 21-23 March, Bakersfield, California, USA. Society of Petroleum Engineers. doi: 10.2118/154242-MS. Zelenev, A.S., Champagne, L.M., and Hamilton, M. 2011. Investigation of Interactions of Diluted Microemulsions with Shale Rock and Sand by Adsorption and Wettability Measurements. Colloids and Surfaces A: Physicochemical and Engineering Aspects 391 (1–3): 201–207. doi: https://doi.org/10.1016/j.colsurfa.2011.07.007 Zhang, J., Wang, D., and Olatunji, K. 2016. Surfactant Adsorption Investigation in Ultra-Lower Permeable Rocks. Paper presented at the SPE Low Perm Symposium, 5-6 May, Denver, Colorado, USA. Society of Petroleum Engineers. doi: 10.2118/180214-MS.