http://hdl.handle.net/10203/17 KAIST Dept. of Chemical and Biomolecular Engineering Tue, 14 Apr 2026 21:26:13 GMT 2026-04-14T21:26:13Z http://hdl.handle.net/10203/326419 Title: Efficient extraction of hydrogen fluoride using hollow fiber membrane contactors with the aid of active-learning Authors: Park, Jimin; Cheon, Mujin; Kim, David Inhyuk; Park, Daeseon; Lee, Jay H.; Koh, Dong-Yeun Abstract: This study presents a sustainable approach to extracting hydrogen fluoride (HF) from wastewater using hollow fiber membrane contactors. HF, a widely used yet hazardous chemical, requires efficient separation techniques due to its environmental and health concerns. Our research compared two operational modes, vacuum mode and liquid-liquid extraction mode, revealing the latter as more efficient for HF separation. Notably, this study introduces a novel aspect by employing a data-driven decision-making method, Bayesian optimization (BO) for process optimization of the liquid-liquid extraction mode, aiming to maximize HF removal efficiency at low experimental costs. Subsequent validation through a 96-h experimental run confirms the suitability of the optimized conditions for industrial applications. This study not only demonstrates an efficient HF separation process using hollow fiber membrane contactors but also establishes a new standard for complex industrial process optimization. http://hdl.handle.net/10203/326419 http://hdl.handle.net/10203/338996 Title: Light oils from methane homologation over hierarchically porous TiO2/ SiO2 hybrid catalyst for a non-thermal plasma Authors: Ban, Minkyeong; Kim, Juchan; Im, Hyeonae; Kim, Seongseop; Nguyen, Hoang Phuong; Yun, Yongju; Han, Jeong Woo; Ha, Kyoung-Su; Lee, Jinwoo Abstract: Significant effort has been dedicated to the conversion of methane (CH4) into more valuable chemicals, but its direct conversion into longer-chain hydrocarbons remains challenging. Herein, we develop hierarchically porous TiO2/SiO2 (HP-TiO2/SiO2) catalysts with adjustable Ti/Si molar ratios for plasma-assisted CH4 conversion to light oils including gasoline-range and diesel-range hydrocarbons in the C5-C16 range. By exploiting interfacial interactions between block copolymer and homopolymer constituents, precise control over macro- and mesoporous structures is achieved. The hierarchical porosity promotes plasma penetration, CH4 activation and intermediate conversion. The Ti/Si ratio influences chain-growth probability, increasing selectivity for gasolineand diesel-range hydrocarbons, reminiscent of Fischer-Tropsch synthesis but without high-temperature CH4 reforming. Density functional theory calculation demonstrates the coke-resistant properties of HP-TiO2 catalysts by lowering CH3 desorption energy and increasing the activation barrier for dehydrogenation. This work provides insights into non-thermal plasma-activated catalyst optimizations for CH4 conversion, and offers prospects for direct synthetic fuel and chemical production. Mon, 01 Jun 2026 00:00:00 GMT http://hdl.handle.net/10203/338996 2026-06-01T00:00:00Z http://hdl.handle.net/10203/339086 Title: N-Doped CNT interlayer synthesized from CO2 for enhancing lithium-sulfur battery performance Authors: Kim, Hyein; Lee, Dayeon; Yun, Won Chan; Kim, Youngchan; Lee, Jae Woo Abstract: Lithium-sulfur batteries are considered promising energy storage devices due to their high theoretical capacity and energy density, yet practical use is hindered by the shuttle effect, where soluble lithium polysulfide intermediates migrate through the separator, causing active-material loss and capacity fade. This study modifies a polypropylene separator by introducing an interlayer of entangled carbon nanotubes grown on cobalt oxide catalysts via CO2 conversion. The CNT network enhances electrical conductivity and facilitates Li-ion diffusion. Melamine serves as a nitrogen precursor to produce nitrogen-doped CO2-derived CNTs (NCCNTs) with edge-site nitrogen that strengthens interactions with lithium polysulfides, accelerates liquid-solid conversion, and suppresses shuttling. With the NCCNT interlayer, cells deliver a maximum discharge capacity of 1214 mAh g-1 at 0.2 C and 727 mAh g-1 at 1.0 C and achieve 3.28 mAh cm-2 at 4.5 mg cm-2 sulfur loading. These results demonstrate a practical route linking CO2 utilization with high-performance lithium-sulfur batteries. Mon, 01 Jun 2026 00:00:00 GMT http://hdl.handle.net/10203/339086 2026-06-01T00:00:00Z http://hdl.handle.net/10203/339662 Title: A scalable microbead-based cell culture platform presenting curvature cues for the enrichment of cancer stem cell-like phenotypes beyond 3D spheroid models Authors: Yeun, Jemin; Kim, Jeongyeon; Kim, Minkyung; Park, Seonghyeon; Yoon, Sung Hyun; Sun, Sang Yu; Jeong, Booseok; Im, Sung Gap; Baek, Jieung Abstract: Cancer stem cells (CSCs) have been implicated as potential contributors to tumor recurrence, therapeutic resistance, and metastatic behavior. While traditional 3D spheroid models have advanced CSC research, their multilayered architecture introduces cellular heterogeneity and limits reproducibility. Here, we present a 3D microbead-based culture platform that enables spatially uniform induction of CSC-like phenotypes in cancer cells and scalable expansion of malignancy-enriched cells. By leveraging omnidirectional curvature cues provided from suspended microbeads, we achieved enhanced mechanotransductive stimulation across the entire ovarian cancer cell surface. To guide preferential adhesion to the beads and minimize nonspecific substrate attachment, we employed initiated chemical vapor deposition (iCVD) to coat the bottom surface with hydrophobic polymers, where poly(cyclohexyl methacrylate) (pCHMA) promoted microbead-specific adhesion effectively. Cancer cells cultured on this microbead-based system exhibited upregulation of the genes associated with tumor aggressiveness and invasive phenotypes exceeding those observed in conventional 2D monolayer and 3D spheroid models. Mechanistically, these effects were closely associated with curvature-induced RhoA signaling and cytoskeletal remodeling. Furthermore, this platform supported large-scale, high-throughput-compatible expansion of aggressive cancer cells, offering a robust tool for CSC-focused studies and drug screening. Our findings highlight the utility of curvature-mediated mechanobiology in engineering more physiologically relevant and scalable in vitro cancer models. Mon, 01 Jun 2026 00:00:00 GMT http://hdl.handle.net/10203/339662 2026-06-01T00:00:00Z