Carbon nanotube field-effect transistor (CNT FET) biosensors have demonstrated considerable advances in clinical applications, yet the absence of a unified predictive model hinders systematic optimization towards detection limits. Here, we present an experimentally calibrated physics-informed simulation framework that quantitatively elucidate the electrostatic signal transduction mechanisms in FET biosensors, facilitating a systematic exploration of the design space to achieve high sensitivity. Employing a CNT floating-gate FET biosensor configuration. We derive a noise-limited detection threshold, providing a direct link between electrical noise and the minimum resolvable molecular coverage. We then reveal that sensor sensitivity depends non-monotonically on gate dielectric thickness, with performance optimized at a specific thickness rather than at the minimum value. Moreover, our results demonstrate that at ultra-low concentrations (<10(-15) M), the specific positioning of biomolecular binding sites is critical. The tunneling barriers at the nanojunctions within carbon nanotube networks induce exponential current responses, though these effects attenuate at higher target concentrations. By establishing a robust physical model, this study provides a foundational platform for analyzing biomolecular electrostatic coupling and offers comprehensive design guidelines to push the performance boundaries of nextgeneration biosensors.

This paper presents a comprehensive power and thermal integrity analysis of a commercial IP based 7nm 3D CPU with a much larger SRAM area compared to its logic section. We systematically investigate the impact of different 3D stacking architectures—Memory-on-Logic (MoL) and Logic-on-Memory (LoM)—combined with both front-side and back-side power delivery networks (FSPDN/BSPDN). A key contribution is a novel lightweight IR drop modeling tool developed in-house, which enables supper fast and highly accurate power integrity estimation at early physical design stages—far before signoff—significantly reducing design iteration time caused by IR violations. This tool also fills a critical gap in commercial EDA support for advanced 3D integration and BSPDN evaluation. Using this tool alongside multi-physics thermal simulations, we compare four 3D design scenarios. Results show that the MoL architecture with BSPDN achieves an optimal balance between power and thermal integrity: it reduces worst-case IR drop in the logic die to just one-fourth of the 2D reference, and lowers peak temperature by over 15°C compared to a LoM counterpart. Further improvements, 50% in IR drop decrease and 14°C temperature reduction, are attainable through TSV optimization and high-thermal-conductivity material integration. This study provides essential 3D architeture, design and technology cooptimization methodologies for future high-perfermance 3D CPUs of advanced technology nodes.

While three-dimensional (3D) Memory-on-Logic integration benefits high-performance computing (HPC), it faces critical bottlenecks in power delivery and thermal management. This paper presents a comprehensive power, performance, area, and thermal (PPAT) evaluation of a 3D Memory-on-Logic CPU utilizing Frontside Power Delivery Network (FSPDN) and Backside Power Delivery Network (BSPDN). Our analysis reveals a fundamental trade-off: while BSPDN significantly improves power integrity by reducing logic IR drop by 7.7 × (vs. 3D FSPDN CPU) and 12× (vs. 2D CPU), the extreme substrate thinning required for backside connectivity severely impedes lateral heat dissipation, raising peak temperatures by ~8°C (vs. 3D FSPDN CPU) and ~12°C (vs. 2D CPU). By incorporating thermal-electrical coupling into a spatial-temperature-aware timing analysis, we demonstrate that unlike 3D FSPDN which yields negligible gains over 2D case due to through-silicon via bottlenecks, the superior power integrity of BSPDN decisively outweighs thermal penalties, achieving a net ~30% performance improvement over the 2D counterpart.

Pixel miniaturization is crucial for advancing high-resolution short-wave infrared (SWIR) imaging, particularly for remote sensing, consumer electronics, and portable platforms. However, conventional epitaxial photodiode technologies have been limited to approximately 10 µm pixels, and systematic investigations of pixel miniaturization limits in SWIR detection are lacking. Here, we report the pixel scaling behavior of carbon nanotube (CNT) film heterojunction-gated field-effect transistor (HGFET) detectors and demonstrate the feasibility and performance improvement of subwavelength pixel SWIR detectors. Specifically, the main performance indicators of the HGFET detectors consistently improve, particularly the specific detectivity and response speed as the pixel is scaled from 5.5 to 1.0 µm. A HGFET detector with a subwavelength pitch of 0.65λ (∼1.0 µm) has a record specific detectivity exceeding 1015 cm·Hz1/2·W-1 at 1300 nm and rise/fall times of 132/148 µs. Through their compatibility with silicon-based readout circuits, submicron-pixel CNT HGFET arrays offer scalable and cost-effective platforms for next-generation high-resolution SWIR imaging systems.

The escalating antimicrobial resistance (AMR) crisis demands molecular diagnostics that deliver actionable results without bacterial isolation, culture, or nucleic-acid amplification. Field-effect transistor (FET) biosensors enable fast electronic readout, but their translational impact is constrained by two coupled bottlenecks: Debye screening, which attenuates charge-based sensing, and the lack of efficient gain for trace targets. Here, we establish a sensing paradigm that addresses both limitations simultaneously by coupling a re-engineered CRISPR/Cas12a module with multivalent, proton-generating enzymatic amplification on a pH-sensitive carbon-nanotube FET. Central to the design is a trefoil-shaped reporter that spatially decouples the Cas12a cleavage site (DNA “stem”) from multiple enzyme-tagging domains (RNA “lobes”), maximizing conversion of target-induced cleavage into catalytic amplification. Together with a high-fidelity 2’-OMe-modified crRNA, single recognition event is translated into a pronounced local pH shift — inherently immune to ionic screening. This CRISPR/enzyme cascade strategy achieves culture- and PCR-free detection of the methicillin-resistance gene mecA with single-nucleotide discrimination and an attomolar detection limit of 39.8 aM. In 30 clinical Staphylococcus infections, a dual-parameter 2D classification demonstrated perfect concordance with reference methods, robustly distinguishing methicillin-resistant from -susceptible cases. Beyond mecA detection, this work establishes a generalizable signal-transduction framework for FET-based nucleic-acid biosensors, while providing a scalable route toward rapid AMR genotyping and informed antibiotic decision-making.

Electron beam-induced current (EBIC) is a vital characterization technique for promising semiconducting single-walled carbon nanotube (CNT) devices, yet its underlying imaging mechanism remains poorly understood. This study elucidates the EBIC imaging mechanism in CNTs. By simultaneously analyzing secondary electron (SE) and EBIC signals at landing energies of 1 keV and 10 keV in scanning electron microscopy (SEM), it is demonstrated that the EBIC signal is strongly correlated with SE emission intensity. This finding indicates that, unlike traditional three-dimensional semiconductor materials where EBIC imaging is dominated by built-in potential, the Pd-CNT system is governed by substrate charging polarity and electron dose. Moreover, the signal intensity distribution is determined by the resistance gradient along the CNT. This fundamental clarification of the physical origin of EBIC in CNTs provides the essential mechanistic foundation required for the reliable quantitative analysis of electrical properties at nanoscale interfaces in low-dimensional electronics.

The emerging carbon nanomaterial-based field-effect transistor (FET) biosensing technologies promise bridging the performance gap between experiments and applications in the field of point-of-care testing (POCT) due to the advantages of ultrahigh sensitivity, fast, label-free detection, and potential for integration. The specific recognition and signal transduction of these sensors are achieved by a biomolecular interface layer at the solid-liquid interface, in which the arrangement and controllable assembly of this layer are crucial for developing the sensor potential. DNA nanotechnology represents a highly programmable approach to design the biosensing interfaces that enables precisely controlling the orientation, molecular conformation, and density of surface-confined biomolecular probes at the nanoscale. In this review, we focus on using designed DNA nanostructures as bioprobes or linker molecules in combination with carbon-based FET biosensors to achieve highly sensitive detection. Specifically, we introduce the structure and principle of carbon-based FET sensors and advantages of using DNA nanostructures for interface engineering and then outline representative DNA nanostructures for biosensing probes classified according to assembly dimensions. We further summarize the latest progress of using DNA nanostructure engineered carbon-based FET biosensors in virus, biomarker, and SNP detection. Most importantly, we have thoroughly analyzed and summarized the challenges encountered in the practical promotion of FET biosensors and have drawn a technology developing roadmap. This review is expected to provide some rational design principles and inspire additional techniques to enhance the performance of carbon-based FET biosensors, thereby promoting carbon-based FET biosensors in early diagnosis and POCT applications.

Abstract Reducing power in integrated circuits (ICs) ultimately requires lowering the supply voltage, but a conventional MOSFET cannot do so indefinitely because thermionic injection fixes the room-temperature subthreshold swing (SS) near 60 mV/dec. The field has therefore pursued switches that amplify the gate response, reshape the injected carrier spectrum, or exploit controlled internal feedback. Here we review steep-slope transistors according to where steep switching originates: the gate, the source, a series element, or channel feedback, and emphasize the underlying physical logic, not only the record metrics, of each route. We compare the major device families in terms of mechanism, drive current, hysteresis, variability, thermal robustness, and integration burden, and relate device characteristics to circuit-level strategies such as subthreshold design and power gating. The central message is that sub-60 mV/dec switching is now experimentally widespread, but its technological meaning depends less on the minimum reported swing than on whether steepness survives over useful current ranges and under realistic constraints of heating, disorder, repeatability, and manufacturability.

Carbon nanotube field-effect transistor (CNT-FET) biosensor, featured with high sensitivity, label-free operation and miniaturization, has emerged as a promising candidate platform to detect trace microRNAs (miRNAs), which are critical disease biomarkers whose aberrant expression is closely associated with cancer progression. However, the complex assembly process of probes on gates often damages the performance of CNT FETs, which makes sensitivity of the sensor unable to meet reliable detection of ultralow concentration biomarkers. Here, we report a detachable CNT FET biosensor with a modular gate chip for ultrasensitive and reproducible miRNA detection. Specifically, the sensitive gate array and the CNT FET are separately fabricated as two modules with reserved liquid metal connection ports. After biofunctionalized process, the sensitive gate modules are assembled to FET modules to form the complete biosensors. This design not only avoids damage to CNT FETs caused by harsh environments during the assembly process of biological probes, but also allows the CNT FETs to be reused repeatedly. Using miRNA-21 as a model target, the modular biosensors exhibit ultrahigh sensitivity (limit of detection of 0.36 aM) and exceptional reproducibility (a response variation below 5.1%). Clinical serum analysis of 48 samples reveals significantly elevated miRNA-21 levels in liver, lung, and breast cancer patients compared with healthy individuals, with results showing excellent agreement with qRT-PCR (R2 = 0.98). The CNT FET biosensor with a modular sensitive gate provides a robust paradigm for cost-effective, decentralized miRNA screening.

Gas sensor arrays are essential for artificial olfaction and broader intelligent sensing systems, yet on-chip recognition of multiple gas species typically necessitates an additive complex fabrication process for sensor customization or substantial computational resources for deep learning. Here, we propose and demonstrate an electrostatically reconfigurable gas-sensing unit based on a carbon nanotube field-effect transistor (CNT FET), in which gas selectivity is achieved by tuning the bottom-gate voltage to modulate the chemical potential of the top-gate sensing layer, without requiring any physical change to sensing material. The fabricated gas-sensing unit exhibits high sensitivities, with responses exceeding 1000% for 5 ppm of SO2, 2 ppm of NO2, and 20% O2 respectively under appropriate working conditions, and enables selective detection through electrostatic modulation. While it shows negligible responses to representative reducing gases, NO2 maintains a measurable response under elevated humidity (64 at 45% RH). The combined field-dependent and time-resolved behavior produces clear separation in PCA and classification accuracies exceeding 90% within 3 min. Integrating arrays of the reconfigurable gas-sensing units via back-end-of-line (BEOL) processing on silicon-based complementary metal-oxide-semiconductor (CMOS) integrated circuits provides a viable pathway toward real-time, high-precision, multitarget gas-sensor systems.

High-density, semiconducting aligned carbon nanotube arrays (A-CNTs) are essential for next-generation electronics. Dimension-limited self-alignment (DLSA) is a leading route to wafer-scale A-CNT fabrication, yet the instabilities that limit array quality remain poorly understood, hampering targeted process optimization. Here, we establish a multiscale characterization framework combining cross-polarized optical microscopy (xPOM) and scanning electron microscopy (SEM) to evaluate DLSA-prepared A-CNT wafers and identify the underlying film-formation mechanism. Notably, characterization of the entire 4-inch wafer is achieved within approximately 1 h, providing an efficient route for high-throughput macroscopic inspection. Cross-validation across multiple characterization methods reveals a hierarchical morphology comprising four spatial regions and recurring macroscopic defects. These observations, together with in situ measurements of DLSA dynamics, support a two-stage self-assembly mechanism: initial two-dimensional biphasic self-assembly behavior followed by quasi-one-dimensional triphasic self-assembly coupled to continuous transfer. The second stage, imposed by orientational confinement at the air/top-layer/CNT-dispersion interface, is proposed to drive globally consistent alignment. Specific defects are further associated with distinct kinematic processes and mechanical stress states. This mechanistic picture provides a framework for suppressing non-idealities in A-CNTs and supporting their eventual integration into CNT-based logic circuits.

Field-effect transistor (FET)-based biosensors exhibit exceptional sensitivity and are readily adaptable to portable electronics. However, detecting electrically neutral small molecules-such as aflatoxin B1 (AFB1)-remains challenging, as their lack of intrinsic charges prevents effective modulation of the FET channel's electric field, leading to weak or undetectable signal output. To address this issue, we develop a smart DNA hydrogel-FET biosensing platform that converts molecular recognition of AFB1 into a robust electrical signal via pH-mediated transduction. The urease encapsulated in the DNA hydrogel is released upon AFB1 binding to catalyze urea hydrolysis, generating ammonia that shifts pH and modulates FET conductivity. This biosensor achieves an outstanding detection limit of 41.69 fg/mL, excellent specificity against structural analogs, and high reproducibility. It enables accurate quantification of AFB1 in various real-world food and herbal samples, with strong agreement to ELISA results (R-2 > 0.98). Notably, the sensing chip is readily regenerable through a simple rinse-and-dry procedure, allowing for multiple reuse cycles without loss of performance. This work presents a generalizable signal conversion strategy to overcome the limitations of FET-based detection for uncharged small molecules, providing a cost-effective, portable, and highly sensitive solution for on-site food safety monitoring and environmental toxin screening.

Developing complementary metal oxide semiconductor (CMOS) integrated circuits (ICs) combining high flexibility and ultrastrong radiation tolerance features will expand conventional chips into ever-increasing extreme applications. Here, we develop a technology to fabricate flexible CMOS field-effect transistors (FETs) and ICs with ultrastrong radiation tolerance based on a semiconducting carbon nanotube (CNT) film via a system technology co-optimization strategy encompassing materials, fabrication process, device structure, circuit architecture, and passivation/encapsulation. The fabricated CNT CMOS FETs exhibit high and symmetric performances, excellent flexibility, and especially recorded radiation tolerance to total ionizing doses up to 24 Mrad (Si), and then flexible and strong radiation-tolerant ICs including inverters, ring oscillators, and static random-access memory cells have been demonstrated. Notably, high-energy irradiation introduces two competing effects where it causes damage but also reduces gate interface state density to improve the performance of the ICs. These findings position flexible CNT CMOS technology as a promising candidate for use in electronics in extreme environments.

In-sensor optoelectronic computing offers a compelling route toward low-latency and energy-efficient machine vision, yet its expansion into the infrared spectrum remains fundamentally constrained by the absence of scalable and highly sensitive infrared kernel primitives (IKPs), which limits the application particularly under ultralow-light and complex conditions. Here we report highly efficient IKPs enabled by an ambipolar photoresponsive infrared photodetector consisting of an ambipolar field-effect transistor (AFET) and a heterojunction according to an optoelectrically decoupled architecture. Taking advantage of the symmetric ambipolar transport in Schottky-barrier (SB) carbon nanotube FET, the heterojunction-gated ambipolar FET (HGAFET) converts a directed photovoltage from the heterojunction gate into a positive photocurrent (PPC) or negative photocurrent (NPC) through adjusting the bottom gate voltage. The IKPs exhibit symmetric and sensitive ambipolar photoresponses, with both specific detectivities exceeding 10 13 Jones, to broad spectrum light from visible-to-shortwave infrared (vis-SWIR), and then have been demonstrated in-sensor multiply-accumulate operations and analog-domain interframe difference computations. The proposed in-sensor computing system based on these IKPs enables efficient feature recognition and motion awareness, achieving recognition accuracies above 90% under low-light conditions while reducing inference latency and computational cost by factors of 16.2× and 73.0×, respectively, and can act as a hardware-level strategy for all-weather infrared edge vision.

Sensitive photodetection covering UV, visible, and short-wave infrared (SWIR) lights will greatly promote applications in all-weather surveillance, remote sensing, and non-destructive inspection, but remains challenging in terms of bandwidth or dark noise based on either conventional semiconductors or emerging low-dimensional materials. Here, we take full advantage of the excellent designability and compatibility of the heterojunction-gated field-effect transistor (HGFET) phototransistor, and extend the SWIR detection upper limit from 1400 to 1700 nm through optimizing the lead sulfide (PbS) colloidal quantum dots (CQDs) based diode on the gate. Specifically, the mean diameter of CQDs is increased from 3.8 to 6.0 nm to enable efficient long-wavelength (1700 nm) absorption, and a hybrid ligand passivation strategy is used to significantly suppress defect states on the nonpolar (100) facets, thereby enhancing heterojunction photovoltage. The resulting HGFETs exhibit a broadband radiation detection from 350 to 1700 nm with a room-temperature detectivity of up to 5.7 × 1013 cm Hz1/2 W-1 and a minimum detectable power density of 6.4 nW cm-2 at 1650 nm. The hybrid-passivated CQD HGFETs provide a possible route toward next-generation, highly sensitive, and broadband infrared photodetectors from UV to short-wave infrared (beyond 1700 nm) light.

This letter presents a laboratory-level, closed-loop, on-wafer predictive design methodology for emerging-device monolithic microwave integrated circuits (MMICs), enabling rapid prototyping under material/process variability without a commercial process design kit (PDK). The flow integrates broadband nonlinear device modeling, test-structure-based electromagnetic (EM) back-calibration of the passive structure, and EM/circuit co-simulation to predict impedance matching and harmonic generation. A compact zero-bias 30/60 GHz frequency doubler using aligned carbon nanotube (A-CNT) Schottky diodes on high-resistivity silicon (HR-Si) validates the approach. The prototype occupies 1.3 mm(2) and achieves a minimum conversion loss (CL) of 15 dB with -1.9 dBm output power at 60 GHz, representing a clear performance improvement compared with previously reported CNT doublers. The close agreement between simulation and measurement across the intended band validates the method for mm-wave circuit development on emerging platforms.

Single-walled carbon nanotubes (SWCNTs) are an interesting material for investigating strong light-matter coupling in the near-infrared and at room temperature due to their large exciton binding energies and stable emissions. In this work, using thin films of monochiral (6,5) SWCNTs as emitters, we study the strong light-matter coupling in three types of well-designed Fabry-Pérot microcavities with a gradual increase in the quality factor (Q factor) from ∼20 to ∼1000. We observe sharp polariton emissions in the near-infrared with a full width at half-maximum down to ∼1 meV. In the structure, exciton-like subradiant states resulting from the strong coupling manifest themselves through the relaxation dynamics of the exciton reservoir (ER). Our time-resolved photoluminescence (PL) measurements indicate that the coherence of these states can be tuned by the Q factor, which enables a high ratio of bright excitons above ∼90% relative to that of the intrinsically dark excitons in SWCNTs. With increasing Q factor, we also show that the population transfer from the ER to the lower polaritons (LPs) can be systematically enhanced. Furthermore, our angle-resolved PL spectra show a narrow distribution of the polariton emission centered around the LP ground state, which is necessary to realize the polariton condensation. These results broaden our understanding of the photophysics of both the polaritonic and subradiant states in the strongly coupled SWCNT microcavity, which will be critical for further studies on the polariton condensation and the engineering of polaritonic devices based on SWCNTs.

Abstract Carbon nanotube (CNT) field-effect transistors (FETs), considered as a promising building block for radio-frequency/terahertz electronics, can be further enhanced in performance through adopting a substrate with high thermal conductivity, low parasitic effect, and acceptable cost. In this work, we develop a technology to fabricate aligned CNT array-based FETs on a polycrystalline diamond (PD) substrate to achieve excellent heat dissipation and low substrate loss for RF applications. Specifically, a spin-on glass (SOG) layer is deposited on the PD substrate to significantly reduce the surface roughness, which is favorable for the formation of high-quality channel, contact, and gate stack in FETs. The fabricated CNT RF FETs on the SOG-treated PD substrate present cut-off frequencies (ft/fmax) of 191/211 GHz, a 35% improvement compared to those on pristine PD, while basically maintaining its excellent thermal conduction.

The rapid expansion of hydrogen economy and safety monitoring of high-energy-density lithium-ion batteries demand hydrogen (H2) sensors that reconcile two conflicting requirements: ultralow limit of detection (LOD) and wide dynamic range. Conventional sensors invariably compromise one for other, sacrificing trace-level sensitivity for linearity, or vice versa. To address this dilemma, discrete and multiple sensor elements are combined into sensor arrays; however, these are characterized by large footprint and low integration densities. Consequently, such arrays are restricted in applications, where miniaturization and high integration are prerequisites. Here, we report a monolithically integrated multi-mechanism sensor chip (IMSC) based on a scalable carbon nanotube (CNT) platform that resolves this fundamental dichotomy. By synergizing three distinct transduction architectures: metal-insulator-semiconductor (MIS), metal-semiconductor (MS), and resistive (Res)-on a single 4inch silicon wafer, the IMSC achieves a dynamic range-to-LOD ratio (DLR) spanning six orders of magnitude (106), from a LOD of 1 part per million (ppm) to pure hydrogen (100 vol %). The device leverages high transductive gain of CNT for trace detection and phase-stability of palladium-silver (Pd-Ag) alloys to ensure linearity at high concentrations. Packaged within a compact 2 mm & times; 3 mm footprint, the IMSC utilizes an adaptive calibration algorithm (ACA) to seamlessly fuse data streams. In real-time lithium-ion battery thermal runaway tests, the sensor successfully distinguished between initial leakage and catastrophic hydrogen evolution, providing a continuous quantitative profile from early warning to explosive limits. This work establishes a scalable, complementary metal oxide semiconductor (CMOS)-compatible route for next-generation multimodal sensing systems.