Sang-Hoon Bae

The transition toward function-centric three-dimensional heterogeneous integration (3DHI) requires ultrathin photodetectors with strong photoresponse and compatibility with vertical stacking. Monolayer transition metal dichalcogenides (TMDCs) are promising active layers for such systems but suffer from intrinsically limited light absorption. Moreover, strategies for enhancing light absorption in monolayer TMDCs without compromising their electronic quality remain limited, as direct deposition can introduce defects and interfacial disorder. Here, we report deposition-free vdW lamination of an anti-reflective MgF2 nanomembrane onto monolayer MoS2. Uniform MgF2 nanomembranes are formed on an amorphous carbon monolayer (ACM) template and cleanly released at the MgF2/ACM interface, enabling integration without compromising the integrity of monolayer MoS2 while enhancing its photoresponse. Applied to mixed-dimensional 3D/2D/3D stacked MgF2/MoS2/Si photodetectors, this approach enables broadband visible-to-near-infrared detection with a maximum responsivity of 51.36 A/W and a measured-noise-based detectivity of 2.27 × 1012 Jones, demonstrating compatibility with vertically stacked 3DHI platforms and effective photoresponse enhancement.

Thermal management has emerged as a materials bottleneck for three-dimensional integrated circuits, which are being actively explored to enable dense vertical connectivity and mitigate interconnect delay and energy consumption. Unlike 2D-structure chips, stacked tiers confine dissipation within a boundary-rich back-end-of-line (BEOL) heterostructure. Heat must cross ultrathin dielectrics, porous interlayers, and numerous metal/dielectric and bonding interfaces. Consequently, interfacial resistance and thin-film size effects often determine temperature increase and reliability margins. This perspective highlights the materials physics underlying thermal limits and relates it to integration and design considerations. First, lattice heating is described by a carrier-phonon relaxation pathway in which the optical phonon bath serves as a transient energy reservoir and modulates ultrafast thermal responses. Second, heat transport in nanoscale BEOL multilayer stacks is discussed with emphasis on thickness-dependent conduction and finite thermal penetration that filters temperature transients across tiers. Then, vertical heat removal is governed by interfacial thermal boundary conductance (TBC) and via-network electrothermal coupling with parasitic Joule heating. Finally, thermal pathways are discussed, including thermally conductive insulating dielectrics and heat spreaders, TBC enhancing interlayers, low-temperature bonding and alternative metallization, and functional via architectures consistent with the BEOL thermal budget.

Two-dimensional (2D) semiconductors enable atomically thin channels and attractive electrostatics, but practical scaling increasingly hinges on gate-dielectric integration rather than channel performance. A key challenge is forming high-quality dielectrics on chemically inert, dangling-bond-free 2D surfaces while pushing equivalent oxide thickness to the sub-nanometer regime without excessive leakage, traps, or electrical breakdown. This review addresses the materials and process physics that govern dielectric formation in 2D devices, with an emphasis on atomic layer deposition nucleation, surface pretreatment and functionalization, and the use of seed and buffer layers for conformal high-κ oxides. The roles of layered insulators, such as hexagonal boron nitride, are discussed in terms of interface quality, electrostatic scaling limits, and transport limitations. The impact of dielectrics and processing on leakage mechanisms, defect generation, device-to-device variability, and reliability metrics, including time-dependent dielectric breakdown, bias-temperature instability, hysteresis, and threshold-voltage drift, is examined. Finally, we highlight van der Waals dry integration and dielectric transfer approaches that reduce process-induced damage and support wafer-scale uniformity, as well as opportunities for mixed-dimensional and 3D stacked architectures across logic, memory, and emerging functional systems.

Wearable eye-tracking technologies remain constrained by bulky optics, high power consumption, and reliance on external computation. We present a hardware–software codesigned electrooculography (EOG) interface that integrates ultrathin conformal e-skin sensors with resistive random-access memory (RRAM) crossbar, used to implement synaptic vector-matrix multiplication within a neuromorphic processing pipeline for real-time gaze decoding. Conformal e-skin sensors provide stable and continuous acquisition of both vertical and horizontal oculomotor signals, which are transformed into attention-guided spike features for classification by a lightweight spiking neural network (SNN). Implemented on an RRAM array with quantized weights after noise-aware training, the system achieves robust inference while substantially lowering latency and energy demand. The proposed framework is designed for local edge-level computation, thereby preserving user privacy and eliminating the need for cloud-based inference. By eliminating head-mounted optics, this glassless architecture enables unobtrusive and energy-efficient wearable interfaces. These results establish flexible bioelectronics with neuromorphic processors to advance immersive computing, assistive interfaces, and mobile health monitoring.

Recent advances in electrical multiply-accumulate (MAC) operations leveraging resistive-switching materials have catalysed significant progress in optoelectronic sensing and computing technologies through the exploration of emerging materials. These innovative approaches facilitate the encoding of optical amplitude information such as retina-like functionalities. However, a critical dimensional mismatch persists between electrical and optical information, resulting in a substantial portion of high-dimensional data channels remaining unexplored in conventional MAC operation schemes. Combined with advanced device architectures and data algorithms, two-dimensional materials are considered promising candidates to realize in situ encoding and optoelectronic sensing of multi-dimensional optical information under precise control owing to their tunable physical properties. In this Review, we outline the progress of emerging two-dimensional-materials-based 'integrated sensors', and benchmark electrical inputs with optical scenarios in a framework unifying information encoding. Exciting opportunities for integrated sensors are discussed as well, highlighting the requirements and differences in the encoding of different dimensions of information and exploring the potential for integrated sensors in other fields.

Optoelectronic computing devices capable of bipolar responses offer a route to simplified architectures for processing complex tasks. However, advancing such systems towards large-scale, in-sensor computing has been constrained by the difficulty of monolithically integrating neuromorphic optoelectronic arrays with peripheral circuits, largely due to high material growth temperatures and the non-uniform performance of complex device stacks. Here we report Mo_4/3B₂T_z (T_z= O, OH, F) boridene as a low-thermal-budget platform for neuromorphic optoelectronics, enabling twelve-inch deposition below 150 °C with excellent wafer-scale uniformity. Ordered metal vacancies and interlayer registry variations generate an unusual electrical anisotropy in which through-plane conduction dominates over in-plane transport. This anisotropy enables a simplified three-terminal device architecture that exhibits intrinsic bipolar and highly linear programmable photoresponses. Correlative conductive atomic force microscopy and first-principles simulations reveal that vacancy-mediated interlayer charge transfer governs the observed behaviour. We further fabricate a 54 × 54-pixel² optoelectronic computing array with a 99.48% yield and 16 fully separable states. Using a 3k-pixel system prototype, we demonstrate the diagnosis of ophthalmic disorders. Our work establishes Mo_4/3B₂T_z boridene as a scalable nanomaterial platform that brings neuromorphic optoelectronic computing closer to practical implementations.

Albeit the speed and energy advantages offered by silicon (Si) and silicon nitride (SiN) photonics, a single material platform, with varying intrinsic limitations, can hardly meet the ever-diversifying requirements posed for photonic integrated circuits (PICs). The heterogeneous integration of functional materials to established photonic platforms has thus gained tremendous momentum, underpinning various high-performance optoelectronic applications. However, stringent lattice-matching constraints severely hinder the heteroepitaxial material quality grown on dissimilar optical substrates. Here we obviate this hurdle by exploiting freestanding single-crystalline nanomembranes enabled by advanced epitaxy and layer lift-off techniques. We present a versatile framework leveraging photonic van der Waals (vdW) integration to infuse desired functionalities to Si and SiN photonics. By transferring single-crystalline thin-film barium titanate (BTO) to Si chips, we experimentally demonstrate ultraefficient electro-optical (EO) modulation, featuring giant Pockels coefficient r_42 over 1290 pm/V and a 3 dB EO bandwidth over 23 GHz, enabled by well-defined BTO crystallographic quality and orientations. Ultracompact non-reciprocal magneto-optical (MO) isolators are further realized, via vdW-integrating cobalt ferrite (CFO) nanomembranes to Si microrings, manifesting an ultrahigh Faraday rotation coefficient θ_F exceeding 33,800 °/cm. To embody the salient capabilities of coalescing functional materials with disparate crystal structures to arbitrary photonic templates, we laterally stitch gallium arsenide and gallium nitride single-crystals atop SiN photonics to realize expansive photodetection from ultraviolet to near infrared wavelengths. Multifunctional ring-resonators for simultaneous EO and MO modulations are also demonstrated by constructing vertical CFO/BTO heterostructures, ushering a vast playground to prototype record-setting heterogeneous integrated applications and study diverse physical coupling phenomena.

Freestanding single‐crystalline nanomembranes have gained increasing attention as promising platforms for both fundamental research and advanced electronic applications. However, internal stress gradients arising from epitaxial strain within the oxide membranes often result in high crack density during fabrication, leading to unsatisfactory yield and limited reliability. Here, an elastically graded polymer (EGP) support that enables wafer‐scale crack‐free transfer of single‐crystalline oxide membranes are developed. The engineered elastic gradient within the EGP accommodates the internal strain of the oxide membrane, effectively minimizing crack formation during lift‐off. Notably, this ability to spatially control the interfacial stiffness between the polymer and the oxide film enables crack suppression under both tensile and compressive strain. This approach provides a robust and scalable route to producing high‐quality freestanding oxide membranes, paving the way not only for their integration into novel device architecture but also opening new avenues for scientific exploration of functional systems.