德克萨斯大学奥斯汀分校

Shale oil is one of the most important oil resources in the United States, yet recovery remains challenging due to its extremely low permeability. Horizontal drilling and multi-stage hydraulic fracturing improve productivity, but depressurization recovery remains below 10% of OOIP, leaving substantial oil unrecovered and many horizontal wells on the verge of abandonment. This study systematically evaluates and compares the effectiveness of water-based chemical EOR agents-including surfactants, nanoparticles, and ketones-for enhancing shale oil recovery under reservoir conditions. A comprehensive experimental workflow was implemented, in which chemical candidates were first screened for aqueous stability and compatibility under high-temperature, highsalinity (HTHS) conditions, followed by evaluation of oil/water interfacial tension reduction and wettability alteration. Static spontaneous imbibition and dynamic Huff-n-Puff experiments were performed using Mancos shale outcrops (5% porosity, 100-200 nD permeability), supported by nuclear magnetic resonance (NMR) analysis employed to support pore-scale interpretation of recovery mechanisms. Four surfactants and two nanoparticles were selected based on aqueous stability and compatibility; one anionic and one zwitterionic surfactant were selected due to their wettability alteration capabilities. The spontaneous imbibition experiments demonstrated that the anionic surfactant achieved higher oil recovery than the zwitterionic one, while the surfactant/nanoparticle blend and the ketone exhibited slightly superior performance. Consistent recovery trends were observed during the Huff-n-Puff experiments, indicating similar chemical effectiveness under dynamic conditions. This work advances understanding of chemical EOR in shale reservoirs by providing a unified evaluation of surfactants, nanoparticles, and ketones under identical shale-relevant thermobaric conditions. The results highlight key differences in recovery behavior among these chemical agents and clarify their relative effectiveness in enhancing shale oil recovery. While laboratory-scale improvements were observed, further investigation is recommended to assess scalability and performance under field conditions.

The most common method to capture CO2 emitted from point sources is the amine-based absorption process. The regeneration of the amine uses high temperatures desorption which incurs a significant energy penalty and contributes to the faster degradation of the amine. Chemical regeneration using alkaline industrial waste offers a low-energy alternative; however, no kinetic model is currently available for this process. Here, we develop a kinetic model for amine regeneration using coal fly ash as the alkaline reactant. Monoethanolamine (MEA), N-methyldiethanolamine (MDEA), and Piperazine (PZ) solutions were used to represent primary, secondary, and tertiary amines. 15 wt% of each amine solution was CO2-saturated with a flue gas surrogate (10% CO2) and contacted with fly ash at solid-to-liquid (S/L) ratio of 350 g/L, 40 degrees C, and 1 atm. The liquid was sampled regularly, and the extent of regeneration was monitored using CO2 loading. At the end of the regeneration process, the solid product was imaged by SEM and TEM for morphological changes. Results show that fly ash mineralizes dissolved CO2 while regenerating the amines. The tertiary amine, MDEA regenerates faster because CO2 is present as bicarbonates, not carbamates. The kinetics of the amine regeneration step can be modeled by a shrinking core model with the diffusion through solid-layer as the rate-limiting step. The effective diffusion coefficient was calculated for the three amines studied. The model achieved a good agreement with experiments (R2 >= 0.85) for all the amines. This work provides the kinetic description of mineralization-mediated amine regeneration.

Nanopores widely developed in shale reservoirs can alter confined fluid phase behavior and affect CO2 based recovery and storage processes. Existing studies have mainly emphasized pore size effects, whereas the role of pore wall wettability is often treated qualitatively or indirectly. In this study, a modified Peng Robinson equation of state framework is developed to evaluate the coupled influence of pore size and representative pore wall wettability on confined phase behavior and miscibility. A dimensionless confinement factor is introduced by combining normalized pore radius with a contact angle dependent term, and a semi empirical critical property shift correlation is incorporated into the PR-EOS parameter correction. The confined phase equilibrium calculation is further coupled with the multiple mixing cell method to estimate minimum miscibility pressure. The model is compared with published saturation pressure data for ethane, propane, and butane in mesoporous materials, showing reasonable agreement within the investigated pore size and wettability range. Model predictions suggest that smaller pores and lower contact angles lead to larger apparent shifts in critical temperature and pressure, while the corrections weaken as pore size increases or the surface becomes less wetting. Under the present correction framework, CH4 exhibits larger predicted critical property deviations than CO2 under the same pore size and contact angle conditions. For a representative Eagle Ford condensate composition, pore size and representative pore wall wettability may affect vapor liquid composition redistribution, including the CH4 fraction in the vapor phase and the heavy component fraction in the liquid phase. The predicted MMP increases with contact angle and decreases with decreasing pore size, while higher CO2 content in the injection gas reduces both MMP and its sensitivity to wettability. This work provides an engineering thermodynamic framework for assessing model predicted confinement and wettability effects in shale nanopores, with potential relevance to CO2 enhanced recovery and storage applications.

Against the backdrop of global carbon mitigation and efficient energy development, carbon capture, utilization, and storage (CCUS) has become a key pathway for large-scale CO 2 emission reduction and low-carbon fossilenergy utilization. CO 2 -enhanced shale stimulation still faces several unresolved challenges, including the unclear coupling between fluid-rock interaction and mechanical weakening, the difficulty of predicting fracture initiation and propagation in heterogeneous shale, and the lack of multiscale criteria for selecting prepad fluids that can simultaneously enhance reservoir stimulation and promote CO 2 migration and retention. To address these problems, a multiscale framework linking micromechanical properties, interfacial friction, and hydraulic fracture propagation was established for the continental shale oil reservoir in the Yingxiongling area of the Qaidam Basin. Microindentation, friction tests, surface morphology characterization, and three-dimensional hydraulic-fracturing simulations were integrated to evaluate the effects of slickwater, gel, acid, and supercritical CO 2 (SC-CO 2 ) on shale mechanical behavior, frictional response, and fracture evolution. The results show that acid causes the strongest weakening of shale, significantly reducing hardness, elastic modulus, and fracture toughness. Slickwater and gel mainly weaken interfacial strength through clay hydration, with slickwater having the stronger effect. In contrast, SC-CO 2 promotes localized brittle damage while maintaining relatively high matrix stiffness, thereby lowering the resistance to fracture initiation and propagation. Frictional behavior is jointly controlled by fluid type and mineral composition: with increasing illite content, the friction coefficient first increases and then decreases, whereas higher carbonate content increases interfacial frictional resistance, and acid markedly weakens carbonate-rich interfaces. Simulations further show that prepad fluid type strongly controls pressure response, fracture geometry, leakoff, and stimulation efficiency. Compared with slickwater, gel increased fracture volume by 48.14 % and reduced leakoff by 24.07 %. The SC-CO 2 prepad yielded the largest fracture area (18,516 m 2 ), the largest fracture volume (1,153 m 3 ), the lowest leakoff volume (1,028 m 3 ), and the highest effective fracturing ratio (52.87 %), demonstrating the best overall stimulation performance and greater potential for CO 2 migration, retention, and long-term storage.

Reasoning models have made rapid progress on many benchmarks involving math, code, and science. Yet, there are still many open questions about the best train- ing recipes for reasoning since state-of-the-art models often rely on proprietary datasets with little to no public information available. To address this, the goal of the OpenThoughts project is to create open-source datasets for training reasoning models. Our OpenThoughts2-1M dataset led to OpenThinker2-32B, the first model trained on public reasoning data to match DeepSeek-R1-Distill-32B on standard reasoning benchmarks such as AIME and LiveCodeBench. We then improve our dataset further by systematically investigating each step of our data genera- tion pipeline with 1,000+ controlled experiments, which led to OpenThoughts3. Scaling the pipeline to 1.2M examples and using QwQ-32B as teacher yields our OpenThinker3-7B model, which achieves state-of-the-art results: 53% on AIME 2025, 51% on LiveCodeBench 06/24-01/25, and 54% on GPQA Dia- mond – improvements of 15.3, 17.2, and 20.5 percentage points compared to the DeepSeek-R1-Distill-Qwen-7B. All of our datasets and models are available on ANONYMIZED.