Search

With the rapid progress of intelligent vision of the Internet of Things, the gap between the resultant high-level demands and conventional silicon-based hardware architectures is continuously widening. Biomimetic artificial neural networks (ANNs) have recently been proposed to solve this problem. However, they still face the predicaments in uniformity, heterogeneous integration, and hardware implementation, etc. Here we demonstrate an intelligent vision ANN with two-dimensional material (2DM)/molecular ferroelectric (MF) bilayer electronic/optoelectronic memristors. In contrast to conventional ferroelectric transistors, in-plane ferroelectric polarization was also found to enable a simpler two-terminal structure with a lateral p/n-type doping in the adjacent 2DM layer. After a demonstration of fabricated array and board-level driving circuits, an image recognition and classification task is proposed, reaching an accuracy of 85.2%, which implies great potential towards multi-modal artificial intelligent vision systems.

Abstract Despite that in-sensor processing has been proposed to remove the latency and energy consumption during the inevitable data transfer between spatial-separated sensors, memories and processors in traditional computer vision, its hardware implementation for artificial neural networks (ANNs) with all-in-one device arrays remains a challenge, especially for organic-based ANNs. With the advantages of biocompatibility, low cost, easy fabrication and flexibility, here we implement a self-powered in-sensor ANN using molecular ferroelectric (MF)-based photomemristor arrays. Tunable ferroelectric depolarization was intentionally introduced into the ANN, which enables reconfigurable conductance and photoresponse. Treating photoresponsivity as synaptic weight, the MF-based in-sensor ANN can operate analog convolutional computation, and successfully conduct perception and recognition of white-light letter images in experiments, with low processing energy consumption. Handwritten Chinese digits are also recognized and regressed by a large-scale array, demonstrating its scalability and potential for low-power processing and the applications in MF-based in-situ artificial retina.

Abstract With the burgeoning developments in artificial intelligence, hardware implementation of artificial neural network is also gaining pace. In this pursuit, ferroelectric devices (i.e., tunneling junctions and transistors) with voltage thresholds were recently proposed as suitable candidates. However, their development is hindered by the inherent integration issues of inorganic ferroelectrics, as well as poor properties of conventional organic ferroelectrics. In contrast to the conventional ferroelectric synapses, here we demonstrated a two-terminal ferroelectric synaptic device using a molecular ferroelectric (MF)/semiconductor interface. The interfacial resistance can be tuned via the polarization-controlled blocking effect of the semiconductor, owing to the high ferroelectricity and field amplification effect of the MF. Typical synaptic features including spike timing-dependent plasticity are substantiated. The introduction of the semiconductor also enables the attributes of optoelectronic synapse and in-sensor computing with high image recognition accuracies. Such interfaces may pave the way for the hardware implementation of multifunctional neuromorphic devices.

In this investigation, we experimentally demonstrated a high-speed Gbps long distance real-time visible light communication system based on non-return-to-zero on-off keying modulation by using a high-bandwidth Gallium nitride-based micro-LED with a modulation bandwidth of 230 MHz and a size of 40 μm x 40 μm. The maximum real-time data rate obtained is up to 1.3 Gbps at a 3 m free-space transmission distance with a bit-error rate (BER) of 3.4 x 10-3 and 1 Gbps at a 10 m distance with a BER of 3.2 x 10-3, both of which are underneath the forward error correction threshold of 3.8 x 10-3 required for free-error operation. In addition, when the transmission distance is increased to 16 m in free space through reflecting the emission light beam by blue reflectors mounted on the wall, a data rate of 0.87 Gbps with a BER of 3.5 x 10-3 is achieved successfully.

Transition metal dichalcogenides (TMDCs) in monolayer form have attracted a great deal of attention for electronic and optical applications. Compared to mechanical exfoliation and chemical synthesis, laser thinning is a novel and unique “on-demand” approach to fabricate monolayers or pattern desired shapes with high controllability and reproducibility. Its successful demonstration motivates a further exploration of the dynamic behaviour of this local thinning process. Here, we present an in-situ study of void formation by laser irradiation with the assistance of temporal Raman evolution. In the analysis of time-dependent Raman intensity, an empirical formula relating void size to laser power and exposure time is established. Void in thinner MoS2 flakes grows faster than in thicker ones as a result of reduced sublimation temperature in the two-dimensional (2D) materials. Our study provides useful insights into the laser-thinning dynamics of 2D TMDCs and guidelines for an effective control over the void formation.

Transition metal dichalcogenides (TMDCs) demonstrate great potential in numerous applications. However, these applications require a precise control of layer thickness at the atomic scale. In this work, we present an in-situ study of the self-limiting oxidation process in MoTe2 by ozone (O3) treatment. A precise layer-by-layer control of MoTe2 flakes can be achieved via multiple cycles of oxidation and wet etching. The thinned MoTe2 flakes exhibit comparable optical properties and film quality to the pristine exfoliated ones. Besides, an additional p-type doping is observed after O3 oxidation. Such a p-doping effect converts the device properties of MoTe2 from electron-dominated to hole-dominated ambipolar characteristics.

Photoinduced space-charges in organic optoelectronic devices, which are usually caused by poor mobility and charge injection imbalance, always limit the device performance. Here we demonstrate that photoinduced space-charge layers, accumulated at organic semiconductor-insulator interfaces, can also play a role for photocurrent generation. Photocurrent transients from organic devices, with insulator-semiconductor interfaces, were systematically studied by using the double-layer model with an equivalent circuit. Results indicated that the electric fields in photoinduced space-charge layers can be utilized for charge generation and can even induce a photovoltage reversal. Such an operational process of light harvesting would be promising for photoelectric conversion in organic devices.

Dopant-segregated source/drain contacts in a p-channel Schottky-barrier metal-oxide semiconductor field-effect transistor (SB-MOSFET) require further hole Schottky barrier height (SBH) regulation toward sub-0.1 eV levels to improve their competitiveness with conventional field-effect transistors. Because of the solubility limits of dopants in silicon, the requirements for effective hole SBH reduction with dopant segregation cannot be satisfied using mono-implantation. In this study, we demonstrate a potential solution for further SBH tuning by implementing the dual implantation of boron (B) and aluminum (Al) in combination with microwave annealing (MWA). By using such a method, not only has the lowest hole SBH ever with 0.07 eV in NiSi/n-Si contacts been realized, but also the annealing duration of MWA was sharply reduced to 60 s. Moreover, we investigated the SBH tuning mechanisms of the dual-implanted diodes with microwave annealing, including the dopant segregation, activation effect, and dual-barrier tuning effect of Al. With the selection of appropriate implantation conditions, the dual implantation of B and Al combined with the MWA technique shows promise for the fabrication of future p-channel SB-MOSFETs with a lower thermal budget.

The effects of UV-ozone (UVO) treatment on the sensing behaviours of extended-gate field-effect transistors (EGFETs) that use Al2O3 as the sensing film have been investigated. The Al2O3 sensing films are UVO-treated with various duration times and the corresponding EGFET sensing behaviours, such as sensitivity, hysteresis, and long-term stability, are electrically evaluated under various measurement conditions. Physical analysis is also performed to characterize the surface conditions of the UVO-treated sensing films using X-ray photoelectron spectroscopy and atomic force microscopy. It is found that UVO treatment effectively reduces the buried sites in the Al2O3 sensing film and subsequently results in reduced hysteresis and improved long-term stability of EGFET. Meanwhile, the observed slightly smoother Al2O3 film surface post UVO treatment corresponds to decreased surface sites and slightly reduced pH sensitivity of the Al2O3 film. The sensitivity degradation is found to be monotonically correlated with the UVO treatment time. A treatment time of 10 min is found to yield an excellent performance trade-off: clearly improved long-term stability and reduced hysteresis at the cost of negligible sensitivity reduction. These results suggest that UVO treatment is a simple and facile method to improve the overall sensing performance of the EGFETs with an Al2O3 sensing film.

In this paper, capacitive-type humidity sensors were prepared by sequentially drop-coating the aqueous suspensions of zinc oxide (ZnO) nanopowders and polyvinyl pyrrolidone–reduced graphene oxide (PVP-RGO) nanocomposites onto interdigitated electrodes. Significant improvements in both sensitivity and linearity were achieved for the ZnO/PVP-RGO sensors compared with the PVP-RGO/ZnO, PVP-RGO, and ZnO counterparts. Moreover, the produced ZnO/PVP-RGO sensors exhibited rather small hysteresis, fast response-recovery time, and long-term stability. Based on morphological and structural analyses, it can be inferred that the excellent humidity sensing properties of the ZnO/PVP-RGO sensors may be attributed to the high surface-to-volume ratio of the multilayer structure and the supporting roles of the PVP-RGO nanocomposites. The results in this work hence provide adequate guidelines for designing high-performance humidity sensors that make use of the multilayer structure of semiconductor oxide materials and PVP-RGO nanocomposites.

The growth, morphology, and life cycle of two marine myxobacterial isolates were studied. Results suggest that the halotolerant forms are the result of degenerative adaptation of soil myxobacteria to the marine environment, while the halophilic myxobacteria form a different evolutionary group that is indigenous to the ocean.

It has been found that in the low-frequency regime, the experimentally determined thermal conductivity of thin dielectric films deviates significantly from the theoretical expectation based on Cahill's 3ω thermal model. It is shown in the present work that this deviation is mainly caused by the heat transport in the heater strip, which is neglected in Cahill's one-dimensional model. By taking this mechanism into account, an analytical model is developed to simultaneously consider the heat transport in the heater strip and the heat flow into the underlying substrate. The validity of this two-dimensional model is confirmed by experiments using specially designed test structures as well as by numerical simulation. The results show that the heat transport along the heater strip, originated from a nonuniform temperature, becomes comparable to that in the substrate at low frequencies. This effect of a nonuniform temperature distribution can also be exploited for extraction of the thermal conductivity of the metallic strip itself. [ABSTRACT FROM AUTHOR]

Transition metal dichalcogenides (TMDCs) in monolayer form have attracted a great deal of attention for electronic and optical applications. Compared to mechanical exfoliation and chemical synthesis, laser thinning is a novel and unique “on-demand” approach to fabricate monolayers or pattern desired shapes with high controllability and reproducibility. Its successful demonstration motivates a further exploration of the dynamic behaviour of this local thinning process. Here, we present an in-situ study of void formation by laser irradiation with the assistance of temporal Raman evolution. In the analysis of time-dependent Raman intensity, an empirical formula relating void size to laser power and exposure time is established. Void in thinner MoS2 flakes grows faster than in thicker ones as a result of reduced sublimation temperature in the two-dimensional (2D) materials. Our study provides useful insights into the laser-thinning dynamics of 2D TMDCs and guidelines for an effective control over the void formation. [ABSTRACT FROM AUTHOR]

The effects of UV-ozone (UVO) treatment on the sensing behaviours of extended-gate field-effect transistors (EGFETs) that use Al2O3 as the sensing film have been investigated. The Al2O3 sensing films are UVO-treated with various duration times and the corresponding EGFET sensing behaviours, such as sensitivity, hysteresis and long-term stability, are electrically evaluated under various measurement conditions. Physical analysis is also performed to characterize the surface conditions of the UVO-treated sensing films using X-ray photoelectron spectroscopy and atomic force microscopy. It is found that UVO treatment effectively reduces the buried sites in the Al2O3 sensing film and subsequently results in reduced hysteresis and improved long-term stability of EGFET. Meanwhile, the observed slightly smoother Al2O3 film surface post UVO treatment corresponds to decreased surface sites and slightly reduced pH sensitivity of the Al2O3 film. The sensitivity degradation is found to be monotonically correlated with the UVO treatment time. A treatment time of 10 min is found to yield an excellent performance trade-off: clearly improved long-term stability and reduced hysteresis at the cost of negligible sensitivity reduction. These results suggest that UVO treatment is a simple and facile method to improve the overall sensing performance of the EGFETs with an Al2O3 sensing film. [ABSTRACT FROM AUTHOR]

The presence of carbon at the interface between NiSi and Si has been found to participate in the process of modification of effective Schottky barrier heights using the dopant segregation (DS) method. Carbon alone results in an increased Φbn from 0.7 to above 0.9 eV. Boron diffusion in NiSi is inhibited by carbon, and no B-DS at the NiSi/Si interface occurs below 600 °C. Above this temperature, B-DS at this interface is evident thus keeping Φbn high. The presence of interfacial carbon leads to an increased interfacial As concentration resulting in beneficial effects in tuning Φbp above 1.0 eV by As-DS. [ABSTRACT FROM AUTHOR]

Chemical vapor deposition synthesis of semiconducting transition metal dichalcogenides (TMDs) offers a new route to build next-generation semiconductor devices. But realization of continuous and uniform multilayer (ML) TMD films is still limited by their specific growth kinetics, such as the competition between surface and interfacial energy. In this work, a layer-by-layer vacuum stacking transfer method is applied to obtain uniform and non-destructive ML-MoS2 films. Back-gated field effect transistor (FET) arrays of 1L- and 2L-MoS2 are fabricated on the same wafer, and their electrical performances are compared. We observe a significant increase of field-effect mobility for 2L-MoS2 FETs, up to 32.5 cm2 V−1 s−1, which is seven times higher than that of 1L-MoS2 (4.5 cm2 V−1 s−1). Then we also fabricated 1L-, 2L-, 3L-, and 4L-MoS2 FETs to further investigate the thickness-dependent characteristics of transferred ML-MoS2. Measurement results show a higher mobility but a smaller current on/off ratio as the layer number increases, suggesting that a balance between mobility and current on/off ratio can be achieved in 2L- and 3L-MoS2 FETs. Dual-gated structure is also investigated to demonstrate an improved electrostatic control of the ML-MoS2 channel. [ABSTRACT FROM AUTHOR]

The large-signal modulation characteristics of a GaN-based micro-LED have been studied for Gbps visible-light communication. With an increasing signal modulation depth the modulation bandwidth decreases, which matches up with the increase in the sum of the signal rise time and fall time. By simulating the band diagram and the carrier recombination rate of the micro-LED under large-signal modulation, carrier recombination and the carrier sweep-out effect are analyzed and found to be the dominant mechanisms behind the variation of modulation bandwidth. These results give further insight into improving the modulation bandwidth for high-speed visible-light communication. [ABSTRACT FROM AUTHOR]

Graphene-based optoelectronic devices, including reduced graphene oxide (RGO) devices, commonly exhibit a large dark current and low on/off ratio with high dark-power consumption and a small responsivity. In this study, semi-RGO-based infrared photodetectors were directly “written” by using lasers, which can exhibit a small dark current of approximately 12 µA/cm2 with a high responsivity of approximately 0.18 A/W at 1,550 nm. Both the dark current and response speed can be tuned with GO as the gate dielectric in a field-effect transistor structure. These findings suggest the possibility of the three-dimensional “writing” of a micro-optoelectronic device or system with a low cost and high performance. [ABSTRACT FROM AUTHOR]

GaN-based light emitting diodes (LEDs) have been fabricated on sapphire substrates with different thicknesses of GaN buffer layer grown by a combination of hydride vapor phase epitaxy and metalorganic chemical vapor deposition. We analyzed the LED efficiency and modulation characteristics with buffer thicknesses of 12 μm and 30 μm. With the buffer thickness increase, cathodoluminescence hyperspectral imaging shows that the dislocation density in the buffer layer decreases from  ∼1.3  ×  108 cm−2 to  ∼1.0  ×  108 cm−2, and Raman spectra suggest that the compressive stress in the quantum wells is partly relaxed, which leads to a large blue shift in the peak emission wavelength of the photoluminescence and electroluminescent spectra. The combined effects of the low dislocation density and stress relaxation lead to improvements in the efficiency of LEDs with the 30 μm GaN buffer, but the electrical-to-optical modulation bandwidth is higher for the LEDs with the 12 μm GaN buffer. A rate equation analysis suggests that defect-related nonradiative recombination can help increase the modulation bandwidth but reduce the LED efficiency at low currents, suggesting that a compromise should be made in the choice of defect density. [ABSTRACT FROM AUTHOR]

This work has a focus on NiSi as a possible metallic contact for aggressively scaled complementary metal oxide semiconductor devices. As the bulk work function of NiSi lies close to the middle of Si bandgap, the Schottky barrier height (SBH) of NiSi is rather large for both electron (∼0.65 eV) and hole (∼0.45 eV). Different approaches have therefore been intensively investigated in the literature aiming at reducing the effective SBH: dopant segregation (DS), surface passivation (SP), and alloying, in order to improve the carrier injection into the conduction channel of a field-effect transistor. The present work explores DS using B and As for the NiSi/Si contact system. The effects of C and N implantation into Si substrate prior to the NiSi formation are examined, and it is found that the presence of C yields positive effects in helping reduce the effective SBH to 0.1–0.2 eV for both conduction polarities. A combined use of DS or SP with alloying could be considered for more effective control of effective SBH, but an examination of undesired compound formation and its probable consequences is necessary. Furthermore, an analysis of the metal silicides that have a small “intrinsic” SBH reveals that only a very small number of them are of practical interest as most of the silicides require either a high formation temperature or possess a high specific resistivity. [ABSTRACT FROM AUTHOR]