Physics

Amorphous oxide semiconductors such as indium tin oxide (ITO) are promising channel materials for back-end-of-line memory and logic devices, yet their performance and reliability are hindered by self-heating at high power densities. Precise quantification of channel temperature and heat dissipation pathways into the surrounding materials is therefore essential for effective thermal management and the 3D heterogeneous integration of these devices. This work investigates heat dissipation in ultrathin ITO devices on sapphire by Raman thermometry, using an interfacial monolayer of MoS₂ as an in situ temperature sensor. Thermal boundary resistance of each interface in the device is independently measured by time-domain thermoreflectance and then used in finite-element simulations of device temperatures, thereby decoupling temperature measurements from interface characterization. Simulated temperatures incorporating measured electrical and thermal characteristics corroborate well with Raman thermometry. Our validated framework highlights the central role of interfacial resistance and heat spreading through the substrate on heat dissipation in thin-film nanoelectronics.

Acidic electrochemical CO₂ reduction reaction (CO₂RR) offers an attractive route to store intermittent renewables as valuable chemicals with high carbon efficiency but suffers from low selectivity due to predominant hydrogen evolution reaction. Utilizing concentrated alkali cations steers the acidic CO₂RR to multicarbon (C₂₊) products but leads to salt precipitation. Here we report a molecular tuning strategy to facilitate acidic CO₂RR to ethylene under a low K⁺ concentration by modifying tetraphenylporphyrin-based molecules onto a Cu surface. At 200 mA cm^-2, we achieve a record ethylene Faradaic efficiency (FE) of 53% on 5,10,15,20-tetraphenyl-21<i>H</i>,23<i>H</i>-porphine zinc functionalized Cu catalysts (a 1.2× improvement compared to the best reports at above 100 mA cm^-2 under an acidic electrolyte having a low alkali cation concentration) and a high C₂₊ FE of 85%, as well as a high CO₂ single-pass utilization of 72%. This work presents a catalyst design strategy for efficient acidic CO₂-to-ethylene electrolysis under low alkali-cation availability.

Recent elucidation of the synthesis mechanism of WS₂ nanotubes has raised fundamental questions about the origin of structural subtypes, imperfect nanotubes, and internal scroll-like morphologies. Although such 1D structures are observed in both laboratory and industrial batches, their formation pathways remain unclear. As applications of WS₂ nanotubes─particularly in optoelectronics─continue to expand, understanding these structural variations is essential for process optimization and performance control. Here, we develop an advanced imaging approach that combines stepwise sulfidation on a microelectromechanical-system chip with sequential cross-sectional imaging of individual nanostructures. This methodology enables time-resolved, atomic-scale correlation between the internal structure of tungsten oxide nanowhisker precursors and the resulting WS₂ morphologies. We identify multiple reaction pathways and establish direct links between precursor anisotropy and nanotube geometry, refining the sulfidation mechanism and clarifying the transformation from tungsten oxide to tungsten disulfide. These insights provide a framework for controlled synthesis of structurally optimized WS₂ nanotubes.

Vector vortex beams (VVBs), characterized by spin-orbit angular momentum coupling, present significant promise for advancing photonic technologies. Nonetheless, the realization of a compact, integrated platform capable of generating VVB arrays with independently addressable optical states has remained an elusive challenge. Here, we demonstrate a monolithic dielectric metasurface based on a dual meta-atom architecture in silicon carbide that overcomes this fundamental barrier. This design provides full access to the spectrum of spatial vector states encoded on higher-order Poincaré spheres. We experimentally realize multimodal VVB arrays and showcase their utility in high-capacity holographic multiplexing and secure information encryption. Our work provides a compact and versatile platform for on-chip multidimensional structured light generation in the visible domain, with promising applications in quantum information science, optical manipulation, and high-density communications.

Topological polar vortices in ferroelectric superlattices offer intriguing opportunities for nanoscale functional devices; however, achieving nonvolatile electric-field control remains a formidable challenge due to their inherent elastic recovery. Here, we demonstrate reversible nonvolatile switching of polar vortices in PbTiO₃/SrTiO₃ (PTO/STO) superlattices, enabled by a thickness-engineered mixed-phase state. Using <i>in situ</i> transmission electron microscopy, we reveal that in PTO₇/STO₇ superlattices, polar vortices structurally coexist with ferroelectric <i>a</i>-domains, forming a laterally modulated mixed-phase configuration. Under a local electric field, vortex switching proceeds via deterministic lateral propagation of vortex-<i>a</i>-domain phase boundaries, resulting in stable domain configurations upon field removal. In stark contrast, thicker PTO₁₀/STO₁₀ superlattices, which host a pure vortex phase, exhibit a volatile switching behavior that elastically relaxes to the ground state. Phase-field simulations further confirm that phase-boundary-mediated pathways provide the necessary flattened energy landscape for topological reconfiguration. These results establish mixed-phase engineering as an effective strategy for nonvolatile control of polar topological textures.

ABSTRACT Interfacial charge extraction is the kinetic bottleneck governing the open‐circuit voltage ( V OC ) in perovskite solar cells (PSCs), yet it is conventionally modeled as a static process characterized by a constant extraction velocity. Here, we demonstrate that interfacial extraction is inherently dynamic and self‐limiting, driven by a negative feedback loop where accumulating charges generate a transient electric field that suppresses subsequent carrier transfer. Using interface‐sensitive transient reflectance (TR) spectroscopy combined with a field‐modified diffusion‐extraction model, we resolve the time‐dependent extraction velocity, S F ( t ), at ITO/perovskite and ITO/SAM/perovskite buried interfaces. We uncover a fundamental trade‐off: efficient interfaces (e.g., SAM‐functionalized) exhibit high initial extraction rates ( S 0 ) but suffer from rapid field saturation (short τ ), leading to a sharp decay in extraction efficacy. This self‐generated field mechanism is corroborated by Kelvin probe force microscopy and explains the performance disparity in operational devices. To quantify this dynamic behavior, we introduce the time‐integrated extraction capacity, , as a robust descriptor of sustained performance. This metric accurately reflects the superior V OC (1.18 V versus 1.13 V) and efficiency (26.47% versus 25.45%) of optimized SAM‐based devices, establishing a new paradigm for designing interfaces that minimize field‐induced losses.

ABSTRACT Metasurface coupling constitutes a fundamental yet intricate electromagnetic interaction that occurs within a lattice of artificial subwavelength unit cells. Despite its prevalence, such coupling is typically ignored in conventional metasurface design frameworks due to the high characterization complexity, leading to suboptimal device performance. Here, we reveal a distinctive long‐range coupling that exceeds an order of magnitude compared with the interaction range of evanescent waves, substantially enriching the metasurface design landscapes. This coupling exhibits pronounced graph topological features, and we design a graph neural network (GNN) to accurately abstract its inherent physics. Through strategic enhancement of the coupling effects, the discrete metasurface responses are transformed into continuous states, thereby unlocking diverse multiplexing channels. By further integrating the GNN into an inverse design agent, we tailor the multi‐channel global response of the metasurface to support simultaneous multiplexing across angle, frequency, and polarization domains. Experimentally, we demonstrate a compact metasurface multiplexer with eight independent channels, showcasing its potential for next‐generation vehicular networks. This work establishes a new paradigm for highly integrated multifunctional metasurfaces, with promising prospects for high‐density optical storage, information encryption, and high‐capacity wireless communication.

ABSTRACT Solar‐driven interfacial evaporation (SDIE) offers a promising approach for sustainable water purification, yet hydrogel evaporators often suffer dehydration‐induced network deformation, uneven wetting, and salt accumulation under extended operation. Here, inspired by leaf transpiration and vein‐guided fluid transport, we introduce a photothermally driven structural self‐optimization strategy that converts dehydration‐induced hydrogel aging into functional evolution. Rapid heating of embedded multi‐walled carbon nanotubes (MWCNTs) under illumination triggers contraction of the polyvinyl alcohol hydrogen‐bond network, generating robust multiscale water‐distribution textures and vapor‐release channels. This structural adaptation amplifies interfacial temperature‐salinity gradients, thereby enhancing coupled thermo‐solutal Marangoni circulation for efficient energy utilization and sustained uniform wetting. The optimized hydrogel evaporator delivers a high evaporation rate of 5.71 kg m −2 h −1 under one sun illumination, ranking among the highest reported values for hydrogel‐based solar interfacial evaporators. It sustains salt‐free operation for 16 h in 3.5 wt.% brine and remains stable over 25 days of continuous day–night cycling at the higher salinity of 7 wt.%, demonstrating long‐term resistance to salt crystallization. It further confirms that the evaporation–condensation process enables efficient rejection of organic dyes, endowing the system with multifunctional capability for water treatment. This work demonstrates a generalizable design strategy for developing adaptive, durable, and high‐performance SDIE systems.

ABSTRACT With the rise of the “metaverse” concept, 3D display technologies that deliver strong presence and immersion, are attracting increasing attention. Among them, holographic 3D displays are regarded as one of the most promising solutions, because they directly reconstruct the complete wavefront of a 3D light field and provide all depth cues to the human eye, including vergence‐accommodation. This review systematically summarizes recent advances in liquid crystal (LC)‐based holographic 3D displays from the perspectives of materials systems and device architectures. It concentrates on LC devices that leverage the intrinsic optical anisotropy, programmable alignment and their field‐tunable properties, covering chiral structured LC devices, LC polarization elements, LC metasurfaces, and other representative LC optical components. Their roles in holographic wavefront modulation and ultra‐thin holographic optical elements (HOEs) are highlighted. This review also discusses the opportunities and challenges of cascaded multilayer LC devices for multidimensional, programmable wavefront control and high‐capacity holographic displays. Overall, this review provides a systematic reference for the design, optimization, and integration of LC‐related materials and devices, and offers insights into next‐generation holographic 3D display systems with immersive and visually comfortable performance.

ABSTRACT Upcycling waste wool into high‐performance actuators remains challenging because keratin's hierarchical architecture is readily disrupted during extraction, and the mechanistic connection between its secondary‐structure transitions and long‐range molecular order is still unclear. Herein, we present a sustainable strategy for fabricating regenerated wool fibers (RWFs) that enable tunable α‐helix‐to‐β‐sheet transitions and exhibit humidity‐triggered shape‐memory behavior. A selective disulfide‐cleavage strategy is introduced to extract intact keratin fibrils from waste fabrics, and long‐range molecular order of keratin is reconstructed through a continuous flow‐induced alignment process, yielding RWFs with hierarchical orientation and native secondary structures. Under tensile strain and hydration, RWFs undergo a reversible α‐helix‐to‐β‐sheet transition, resulting in rapid and repeatable shape recovery and humidity‐triggered actuation. By integrating machine learning with molecular‐dynamics simulations, the molecular pathway comprising helix uncoiling, β‐sheet formation, and hydrogen‐bond rearrangement is systematically elucidated. The resulting fibers can function as programmable humidity‐responsive switches, self‐tightening sutures, and adaptive compression bandages. This work establishes a viable route to hierarchically ordered, shape‐memory biomass fibers with molecular‐level precision, opening avenues for advanced bioengineering and biomedical applications.

ABSTRACT Porous functional materials have great potential in adsorption, separation, heterogeneous catalysis, and energy storage/conversion. Nanoporous poly(divinylbenzene) (PDVB), prepared via unique solvothermal and template‐free methods, is distinct in this field. Its highly crosslinked aromatic framework and hierarchical porosity provide excellent chemical/thermal stability, intrinsic superwettability (superhydrophobicity and superoleophilicity), and tunable catalytically active sites, making PDVB a versatile platform for regulating surface/interface wettability and reaction microenvironments, especially in heterogeneous catalysis. This review summarizes recent advances in designing and synthesizing PDVB‐based catalysts. Key strategies for tailoring pore structure, surface/interface chemistry, and wettability are highlighted, as well as in situ and post‐functionalization methods for integrating diverse catalytic functionalities. Representative catalytic systems (acid/base, metal‐supported, chiral, hybrid, and photocatalytic catalysts) are elaborated to illustrate rational PDVB engineering for versatile catalysis. Special emphasis is on the synergistic effects of superwettability, network swelling and hierarchical porosity in modulating reactant enrichment, mass transport, and active‐site accessibility – enhancing catalytic activity, selectivity, and recyclability. Typical applications in biomass conversion, fine chemical synthesis, and environmental remediation are systematically discussed, focusing on structure–performance relationships. Finally, current challenges and future outlooks (scalable synthesis, catalyst shaping, long‐term stability, and process‐oriented engineering) are outlined to guide the rational development of PDVB‐based catalysts toward practical and sustainable applications.

ABSTRACT Continuous blood pressure (BP) monitoring remains a clinical imperative with the global prevalence of cardiovascular diseases. Despite the development of flexible electronics for on‐skin sensing of hemodynamic signals, the reliability of continuous BP tracking is still limited by the poor skin‐device coupling at anatomical concavities (e.g., wrist), which leads to motion artifacts and compromised long‐term wearability. Herein, we designed a breathable, full‐module shape‐adaptive pulse sensing wristband leveraging the foam structure of the shape memory polymer (SMP). The novel device demonstrates superior pressure sensing performance, achieving ultrahigh sensitivity (&gt;2289.31 kPa −1 ) within a wide range (0–700 kPa), as well as ultralow detection limit (0.16 Pa) and response/recovery time (1.64 ms/3.42 ms), which enables the recording of arterial pulses in high fidelity. Through processing the pulse waveform with a machine learning‐based algorithm, continuous BP monitoring was achieved, with 97%/90% accuracy for systolic/diastolic BP (SBP/DBP) prediction, surpassing the clinical standard for BP measurement. Moreover, the wristband's full‐module breathability and the stress‐free skin‐sensor interface ensure long‐term wearability without skin irritation for up to 24 h. Significantly, the incorporation of SMPs dramatically optimize the bioelectronic interface for epidermal systems. And the as‐developed pulse sensing wristband holds great potential for population‐scale cardiovascular health management.

ABSTRACT High‐performance cathode materials are essential for the commercialization of sodium‐ion batteries (SIBs) in large‐scale energy storage. However, the NASICON‐type Na 3.12 Fe 2.44 (P 2 O 7 ) 2 cathode with promising structural stability and electrochemical properties has been plagued by sluggish electronic and ionic kinetics. To address this, we develop a series of superior rate and ultra‐stable Na 3.12‐2x Fe 2.44+x (P 2 O 7 ) 2 (NFPO) cathodes through a deliberate bulk defect‐engineering strategy that enables precise control of iron vacancies (V Fe ) and oxygen vacancies (V O ). Structural characterizations confirm that the material retains phase‐pure triclinic P ‐1 framework and the introduction of V Fe –V O divacancies, which induce an upshift of the valence band. Remarkably, density functional theory (DFT) calculations elucidate a dual‐functional mechanism, as V Fe and V O not only narrow the band gap but also drastically reduce the Na + migration barrier by 1.317 eV. Benefiting from this dual kinetic enhancement, the optimized NFPO@C (x = −0.1) cathode exhibits superior rate capability with a reversible capacity of 129.6 mAh g −1 at 1C, and exceptional cycling stability, retaining 99.6% of its capacity (100.8 mAh g − 1 ) after 12 000 cycles at 20C. This study establishes rational Fe and O divacancy engineering as an effective strategy to synergistically boost electronic/ionic transport, offering a generalizable paradigm for advanced polyanionic compounds in next‐generation energy storage.

ABSTRACT The practical deployment of aqueous zinc–iodine batteries (ZIBs) is hindered by the polyiodide shuttle effect and the intrinsically sluggish charge transfer within conventional host materials. To address these challenges, we designed a dual‐engineered polyimide cathode through a rational “backbone–side chain” strategy. Specifically, π‐conjugation engineering of the backbone established a carbonyl‐bridged polyimide (BTPI) constructed from 3,3′,4,4′‐benzophenonetetracarboxylic dianhydride, which uniquely combined strong chemisorption (binding energy: −2.54 eV) with a narrow bandgap (1.43 eV) arising from enhanced π‐orbital delocalization, thereby enabling enhanced charge‐transfer capability. Building on this optimized conjugated host, side‐chain engineering was employed to obtain quaternary ammonium‐functionalized BTPI (QA‐BTPI), providing potent electrostatic confinement. This synergistic dual‐anchoring network significantly enhanced the overall polyiodide binding energy to −4.47 eV. Consequently, the QA‐BTPI cathode achieved a high reversible capacity (150.6 mAh g −1 at 1 C) and exceptional cycling stability, retaining 92.8% capacity after 10000 cycles at 20 C. This work demonstrates a new design paradigm for energy storage materials via the concerted optimization of π‐conjugation‐regulated charge transfer and multi‐mode confinement.

ABSTRACT Hydrogen sulfide (H 2 S), a toxic gas and key biomarker, is crucial to environmental safety and human health, yet existing sensors struggle with real‐time, portable detection at trace levels. Here, we report a cost‐effective, flexible, self‐powered H 2 S sensor with outstanding performance at room temperature, constructed using a galvanic cell structure and leather hydrogel as the electrolyte. The degradable, tough leather hydrogel confers flexibility, mechanical robustness, and eco‐friendliness on the sensor. Upon gas adsorption, monitoring the electrode potential change induced by interfacial charge transfer enables a pronounced response even to trace concentrations. The sensor delivers exceptional sensitivity, with a detection limit down to 0.109 ppb—the lowest among reported electrochemical H 2 S sensors—while also exhibiting excellent selectivity, stability, and repeatability. Performance can be further enhanced by serially connecting, and its self‐powered nature enables visual detection of H 2 S concentration. Furthermore, a wireless sensing system based on Bluetooth and cloud technology is developed, validating the sensor's practical application potential in periodontitis grading diagnosis, meat spoilage monitoring, and H 2 S leak detection. This work provides a novel strategy for the development of high‐performance, cost‐effective, portable H 2 S sensors and holds great promise for applications in environmental monitoring, health diagnostics, and smart sensing systems.

ABSTRACT Natural actuation systems achieve adaptive motion by activating internal stress and stiffness only when required, rather than maintaining permanent structural anisotropy. In contrast, most artificial soft actuators rely on embedded fibers or layered architectures that lock anisotropy and generate motion primarily through passive swelling constraints. Here, a reversibly ion‐activated soft actuator that dynamically switches between isotropic and anisotropic mechanical states is reported. The system consists of shear‐aligned stiff chitosan (∼2.4 wt.%) chains dispersed in a poly(acrylic acid) hydrogel, where molecular alignment acts as a latent template that is mechanically inactive in the absence of ions. Upon cation exposure, the aligned chitosan directs the formation of dense ionic bridges orthogonal to the alignment, producing inverse anisotropy with higher strength perpendicular to the alignment (6.0 MPa) than parallel direction (3.0 MPa). This ion‐templated reinforcement drives robust actuation perpendicular to the alignment (67°) while suppressing deformation along the alignment direction (3°), arising from active ionic interactions rather than passive swelling. Ion removal restores mechanical isotropy, enabling repeatable and cyclable transitions. Integration with extrusion‐based 4D bioprinting further enables programmable, cross‐axis shape transformations guided by finite‐element simulations, offering a versatile platform for bio‐inspired soft actuators.

ABSTRACT The diesel oxidation catalysts (DOC) have been challenged by the high cost of noble metal and limited reactivity at low temperature. Herein, we report PtSmMnCoNb high‐entropy oxide nanoparticles supported single‐atom Pt catalyst, which achieves 90% conversion for CO at 100°C, for C 3 H 6 at 208°C, and 50% conversion for NO x at 203°C. This catalyst also achieves a thermodynamically stabilized configuration, which exhibits a long‐term durability benefited from high‐entropy effects. A combined experimental and theoretical analysis reveals the electron‐rich character of Pt δ+ sites derived by the strong interaction involved in abundant Pt‐O‐Metal interfaces, which accounts for the effective activation for reactants, leading to a transition in the reaction mechanism from Mars‐van Krevelen to Eley‐Rideal. This work provides a design paradigm of advanced catalysts for the integrated removal of multi‐pollutants.

Abstract Gallium Nitride (GaN) high-electron-mobility transistors (HEMTs) are constrained by localized selfheating, which compromises reliability and limits power scaling. While diamond integration offers a superior thermal management solution, minimizing the thermal resistance of the near-junction passivation layers remains a critical challenge. This work demonstrates an effective top-side thermal management strategy by fabricating a GaN/AlN/diamond heterostructure, replacing the conventional low-thermalconductivity SiN passivation. Enabled by a thin AlN interlayer, this architecture achieves an ultra-low thermal boundary resistance of 6.5 m²•K/GW. Consequently, under a high power density of 24 W/mm, transient thermal analysis reveals a significant peak junction temperature reduction of ~30°C compared to standard SiN-passivated devices. Cross-sectional TEM and EDS confirm atomically abrupt interfaces without interdiffusion, validating the structural integrity. These results underscore the efficacy of eliminating the thermal barrier in the near-junction region, establishing a viable pathway for realizing the full potential of diamond-enhanced GaN electronics.

Abstract The memristor, garnering considerable attention due to its potential applications in various fields, such as neuromorphic computing, analog and digital circuits, memory devices, chaotic circuits, etc., is an exceptionally promising technological breakthrough. Numerous hysteresis phenomena, each exhibiting diverse properties, are widely documented within the realm of memristors. It is widely assumed that memristors can be categorized into different types depending on the presence of three distinct key signatures. The memristor exhibits promise as a building block for synaptic devices that demand analog behavior, memory devices that require digital-type hysteresis, and applications that require butterfly-like hysteresis. The currently available models of memristors need to be more adequate in elucidating every facet of their current-voltage behavior. A novel set of window functions is methodically developed&amp;#xD;in this work, emphasizing the customization of the analog, digital, asymmetrical and butterfly-like features. The experimentally observed characteristics and behavior is explainable by these models using a variety&amp;#xD;of control parameters, including voltage range, frequency, ramp-rate, and amplitude variation. These generalized models will play a crucial role in the advancement of memristor-based circuitry and the continued expansion of the field.&amp;#xD;

Abstract Intensity-modulated proton therapy (IMPT) employs proton radiation rather than conventional X-rays to treat cancerous tumors. This approach offers significant advantages by minimizing the radiation exposure of surrounding healthy tissue, leading to improved patient outcomes and reduced side effects compared to traditional X-ray therapy. To ensure patient safety, each treatment plan must be experimentally validated before clinical implementation. However, current dosimetry devices face limitations in performing angled beam measurements and obtaining multi-depth assessments, both of which are essential for verifying IMPT treatment plans. In this study, the performance of a β-Ga 2 O 3 -based metalsemiconductor-metal (MSM) detector with a low-noise amplifier has been studied and evaluated under various proton radiation doses and energy levels delivered by a MEVION S250i proton accelerator. The detector's performance was also compared with that of an ionization chamber. The β-Ga 2 O 3 detector exhibited a linear response with proton dose for single-spot irradiations, and its response to varying proton energies closely matched both the ion chamber data and simulated dose distributions. These findings highlight the potential of β-Ga 2 O 3 -based detectors as robust dosimetry devices for IMPT applications.

Abstract We report an in situ RHEED investigation of the early-stage capping dynamics in selfassembled InAs/GaAs quantum dots (QDs) grown by molecular beam epitaxy (MBE) and encapsulated with AlₓGa₁₋ₓAs ternary alloys. By varying the aluminum molar fraction (x = 0.0-1.0), we systematically investigate the onset of QD encapsulation, with emphasis on surface planarization, morphological preservation, and interface evolution. Real-time RHEED monitoring of the transmission spot intensity was employed during both selfassembly and capping. Two capping regimes were explored: one with variable AlGaAs growth rate and another with a fixed rate (~0.75 μm/h). Regardless of growth rate, it was found that higher Al content significantly delays surface flattening and preserves faceting features up to ~6.5 nm of capping. This behavior is attributed to the lower surface mobility and the reduced intermixing of Al adatoms. Under fixed growth rate conditions, compositiondependent transitions could be clearly identified within the first few nanometers, allowing us to decouple kinetic and compositional effects. Complementary derivative analysis of the RHEED signal revealed changes in surface morphology within the first 2 nm of overgrowth. These findings demonstrate that compositional tuning of the capping layer is a powerful tool for controlling strain relaxation and morphology evolution in stacked QD heterostructures.

Abstract As CMOS technology scales, Static Random-Access Memory (SRAM) stability is increasingly jeopardized by process variability. While the impact of Gate Line Width Roughness (LWR) on spatial variability has been widely documented, the distinct role of Active Area (AA) LWR—particularly its correlation with temporal Low-Frequency Noise (LFN)—remains poorly understood. This study bridges this gap by presenting a systematic experimental investigation using a production 55 nm CMOS hard-mask (HM) process to decouple the effects of AA and gate LWR. Contrary to conventional wisdom that focuses solely on dimensional uniformity, we discover a novel physical mechanism: reducing AA LWR mitigates boundary scattering at the shallow trench isolation (STI), which directly suppresses the intrinsic LFN of the device by approximately 1 dB. This suppression of temporal noise is critical for low-voltage operation. Furthermore, we establish a dual-regime yield theory: experimental data prove that low-voltage yield is predominantly limited by LFN, whereas nominal-voltage yield is governed by spatial variability (device mismatch). Quantitative analysis reveals that a 1 nm reduction in AA LWR decreases temporal Read Static Noise Margin (RSNM) fluctuation by 37.6%, resulting in a 13.1% increase in normalized low-voltage yield. In contrast, gate LWR reduction primarily enhances spatial matching, improving parametric yield by 10.2% at nominal voltage. These findings provide a physics-guided co-optimization strategy for advanced patterning processes, offering specific design rules to manage spatiotemporal fluctuations beyond simple process tweaks.