Research / 2019 / Article

Review Article | Open Access

Volume 2019 |Article ID 9490413 |

Feichi Zhou, Jiewei Chen, Xiaoming Tao, Xinran Wang, Yang Chai, "2D Materials Based Optoelectronic Memory: Convergence of Electronic Memory and Optical Sensor", Research, vol. 2019, Article ID 9490413, 17 pages, 2019.

2D Materials Based Optoelectronic Memory: Convergence of Electronic Memory and Optical Sensor

Received29 Apr 2019
Accepted14 Jul 2019
Published21 Aug 2019


The continuous development of electron devices towards the trend of “More than Moore” requires functional diversification that can collect data (sensors) and store (memories) and process (computing units) information. Considering the large occupation proportion of image data in both data center and edge devices, a device integration with optical sensing and data storage and processing is highly demanded for future energy-efficient and miniaturized electronic system. Two-dimensional (2D) materials and their heterostructures have exhibited broadband photoresponse and high photoresponsivity in the configuration of optical sensors and showed fast switching speed, multi-bit data storage, and large ON/OFF ratio in memory devices. In addition, its ultrathin body thickness and transfer process at low temperature allow 2D materials to be heterogeneously integrated with other existing materials system. In this paper, we overview the state-of-the-art optoelectronic random-access memories (ORAMs) based on 2D materials, as well as ORAM synaptic devices and their applications in neural network and image processing. The ORAM devices potentially enable direct storage/processing of sensory data from external environment. We also provide perspectives on possible directions of other neuromorphic sensor design (e.g., auditory and olfactory) based on 2D materials towards the future smart electronic systems for artificial intelligence.

1. Introduction

The advance of digital technology enables data storage and processing in binary form with high speed, accuracy and density. The development of computing hardware has been driven by Moore’s law in the past few decades. However, the downward scaling trend is hitting the scaling limits according to the International Technology Roadmap for Semiconductors (ITRS) 2.0. Different from the strategy of device miniaturization involved in the digital components, a new trend of “More-than-Moore” (MtM) is highlighted in the ITRS 2.0, which is featured as a diversified function of data collection and information storage/ processing devices. The function diversification described in MtM can be considered as another form of “scaling down” for the electronic system, which includes noncomputing functions into the digital ones, enabling the interaction of digital devices with the outside world directly. Device integration of noncomputing (e.g., sensors and actuators) and digital components (e.g., memories and computing units) is one of the main approaches for function diversification of MtM. Considering the large occupation proportion of image data in both data center and edge devices, we will mainly discuss the integrations of optical sensors and digital devices (data storage and processing) in this review.

Optoelectronic random-access memories (ORAMs) offer great opportunities for multifunctional integration of optical sensing, data storage and processing into one single device. In contrast to traditional random-access memories (RAMs), optical excitation provides another method to manipulate the memory cell in ORAMs, which enables to store light information with an electronic readout at low programming voltages and is beneficial for the storage and transport of secured information. The combination of both optical stimuli and electrical readout voltage has potential for enlarging the ratio of low resistance state (LRS) and high resistance state (HRS), thus enabling multilevel data storage and improving data storage capacity. From another aspect, the optical trigger is employed in ORAM for the switching, bringing the benefits of wide bandwidth and reduction of electrical loss during sensory data transmission [16]. In addition, it is promising to perform digital processing functions with light as an extra control parameter by combining both optical and electrical inputs as input signals, such as arithmetic and logic operations and coincidence detection [4, 7]. Furthermore, the integration of optical sensors and memory devices in ORAMs provides possibilities in simplifying the image sensor circuitry.

In the conventional von Neumann computing architecture, memory and processing units are physically separated, which is inefficient in terms of energy consumption and data transmission speed. The human brain consists of a large collection of neurons connected through synapses, and deals with learning and processing complex cognition tasks (e.g., pattern recognition, and multiobject detection) with high energy efficiency. The power consumption in human brain is 5 orders magnitude lower than conventional computers [8]. Neuromorphic computing mimics the working schemes of information storage and processing like the human brain. The electronic memory synaptic devices are the most investigated and fundamental component in the hardware-based neuromorphic computing. The tunable resistance under continuous stimuli of voltage or current inputs is similar to the synaptic weight (generally determined as biological strength of the connection between two neurons). However, the conventional memory synaptic devices for neuromorphic computing are manipulated through the electrical method and are isolated from the sensors (e.g., visual, auditory or olfactory) [917]. ORAM synaptic devices integrate optical sensors and synaptic RAMs, which can respond to optical stimuli and exhibit light-tunable synaptic behaviors, including light-tunable short-term plasticity (STP), long-term plasticity (LTP), spike-timing-dependent plasticity (STDP), and spatiotemporal learning rule (STLR). The development of ORAM synaptic devices can provide potentials for bridging the gap between optical (visual) sensing and neuromorphic computing [18]. Figure 1 illustrates a comparison of the conventional Von Neumann architecture and a computing architecture based on the ORAM/ORAM synaptic devices for image storage and processing. The architecture based on ORAM/ORAM synaptic devices offer the possibilities of directly storing and processing the optical information, in contrast to the separated memory and processing unit and preliminary optical-to-electrical signal conversion for image processing in the conventional Von Neumann architecture.

Among different materials systems for ORAMs, two-dimensional (2D) materials, such as graphene, transition metal dichalcogenides (TMDs) (including MoS2, WS2, WSe2, MoSe2, MoTe2, etc.) and black phosphorus, have exhibited unprecedented electrical and optical properties for the applications of soft optoelectronics and nonvolatile memories. In terms of the integration of optical sensing, data storage and processing, 2D materials and their heterostructures also exhibit promising potentials for building ORAMs with broadband photo response, excellent photoresponsivity [1925], fast switching speed, multibit data storage, large ON/OFF ratio and high energy efficiency [2632]. More specifically, the strong light-matter interactions and the gate tunability owing to the unique physical and structural properties in the ultrathin 2D materials benefit the construction of both volatile and nonvolatile ORAMs. The large-surface-to-volume ratio in these 2D materials allows significant photo-excited charge trapping and further the ORAM design. (3) The mechanical flexibility and optical transparency of 2D materials are appropriate for the application of ORAMs in an artificial eye. The structure and physical properties of 2D materials have been overviewed in the existing reports [3335]. In this paper, we mainly focused on the review and perspective of 2D materials for memory applications.

2. 2D Materials Based Nonvolatile Electronic Memory

The 2D materials based electronic memory provides ideas for the design of 2D materials based ORAM structure. The working mechanisms of state-of-the-art ORAMs are generally based on charging trapping in three-terminal FET memory and light-tunable filament formation/rupture in two-terminal resistive random-access memories (RRAMs). In this part, we will mainly review two types of nonvolatile electronic memory based on 2D materials: floating gate memories and RRAMs.

2D materials and their heterostructures have offered great practical advantages in the design of floating gate memory device with a large ON/OFF ratio and long retention time. The application of 2D materials as the floating gate can effectively reduce the capacitive interference between the floating gate and drain electrode for high-capacity data storage. In early 2013, Bertolazzi et al. designed a nonvolatile floating gate memory with MoS2 as channel material, multilayer graphene as floating gate and charge trapping layer, and HfO2 as control gate (Figure 2(a)). As a positive control gate voltage is applied (+18 V), electrons tunnel from MoS2 channel to multilayer graphene floating gate through the HfO2 tunneling barrier. The electrons are stored in the floating gate, programming the device. A negative control gate voltage (-18 V) sweeps the electrons in the floating gate back to the channel material, erasing the device (Figure 2(b)). With this structure, the device exhibits nonvolatile switching characteristics with ON/OFF ratio of 104 [36]. Later, Choi et al. demonstrated a similar floating gate memory with MoS2/h-BN/graphene heterostructure with h-BN as a tunneling barrier, showing a comparable ON/OFF ratio of 104 [37]. By further tuning the h-BN thickness appropriately, two-terminal operation of the MoS2/h-BN/graphene floating gate memory can be achieved (Figure 2(c)), which allows for higher integration density. A source-drain voltage of -6 V programs the device to LRS state by driving the electrons to tunnel through the h-BN layer and be accumulated at graphene floating gate. A positive source-drain voltage of 6 V then is employed for erasing the device (Figure 2(d)). The device exhibits an extremely low OFF current of 10−14 A, large ON/OFF ratio of up to 109, and long retention time (104 s) [38]. As the operation speed in floating gate memories is usually relatively slow, a semi-floating gate memory based on van der Waals heterostructures was further developed as an effective device structure to achieve ultrahigh speed and long refresh time (Figure 2(e)). In the semi-floating gate memory, WSe2 is the channel material. HfS2 and h-BN serve as a floating gate and tunneling layer (drain region), respectively. The WSe2/MoS2 (source region) works as a p-n junction switch. For the programming operation ( = − 5 V, = 0 V), the p–n junction is switched on. The electrons in the HfS2 are transferred to the channel material through the h-BN tunneling layer (flash path) or the WSe2/MoS2 p–n junction (switch path). For the erasing operation ( = 5 V, = 0 V), as the p–n junction is switched off, electrons in the channel are transferred to the floating gate and combined with holes in the HfS2 layer. The fast writing speed and long refresh time in the device design originate from the fast charge injection through switch path and floating gate charge storage, respectively [28].

Although excellent nonvolatile properties can be achieved in the 2D nonvolatile floating gate memory, its three-terminal structure limits the integration density and requires large programming voltage. Thus, the two-terminal RRAMs are considered as promising candidates towards nonvolatile memory applications for the high integration density and low programming voltage. Two-terminal RRAMs based on 2D TMD monolayers have also been rationally designed for low operation voltage, including MoS2, MoSe2, WS2 and WSe2 [32]. The good thermal stability and well-defined interfaces of 2D heterostructures provide possibilities in overcoming the thermal stability issue in the conventional RRAMs. Wang et al. reported a robust RRAM based on graphene/Mo/graphene heterostructure with excellent endurance up to 107 switching cycles and high operating temperature of 340°C (Figures 3(a) and 3(c)) [29]. The high-temperature endurance of the device originates from the atomically sharp interfaces even at elevated temperatures, avoiding the undesirable ion migrations (Figure 3(b)). The 2D materials RRAMs also show promising potentials in constructing artificial synapses with low power consumption for energy-efficient neuromorphic computing. Shi et al. reported a 2D material RRAM synaptic device with metal/h-BN/metal structure (Figure 3(d)) based on the switching mechanism of formation/rupture of conductive filament (~15 nm). The RRAM synaptic device exhibits both nonvolatile and volatile switching behaviors, as well as the tunable synaptic behaviors (Figures 3(e) and 3(f)). The low power consumption (0.1 fW in standby and 600 pW per transition) and fast switching speed (10 ns) implicate the potential applications for brain-like computing [39].

3. 2D Materials Based Nonvolatile ORAM

The conventional 2D materials based RRAM is usually manipulated by electrical triggers. As 2D materials have exhibited a broad response from ultraviolet to infrared light with high responsibility or fast response speed, it is possible to integrate the optical sensing and data storage together by rationally designing device structures, which is potential for the various applications, such as image sensors for artificial vision, optical demodulators and logic gates for optoelectronic circuits, and synaptic devices for neuromorphic systems. The constructions of ORAMs can be mainly divided into two types according to the device structure: ORAM based on 2D materials and ORAM based on 2D heterostructures, which will be discussed in the Sections 3.2 and 3.3, respectively.

3.1. Nonvolatile ORAM versus Photodetectors

Both ORAMs and traditional photodetectors convert the optical signal to an electronic one. However, the design of ORAMs is in sharp contrast to the photodetectors, in which the photo-generated charge carriers recombine immediately in the photodetectors after the removal of light stimuli, corresponding to an immediate current drop from (photo current) to (dark current) (Figure 4(a)). In ORAM, the optical information can be stored and retained for a long period even after the removal of light stimuli, corresponding to a stored and retained photocurrent (Figure 4(b)). The responsivity R in photodetector, which represents the capability of optical-to-electrical conversion, can be expressed according to where is the light power density and is the device area. As the photocurrent usually suffers from a slight degradation (after the removal of light stimuli) until reaching a stable current storage current of , here we can define a specific nonvolatile responsivity (NR) for ORAM in a similar way, which can be expressed as where is the storage current.

3.2. 2D Materials Based ORAM

The working mechanism of ORAMs based on 2D materials mainly lies in providing charge trapping centers for photo-generated charge carriers (electrons/holes), which store the photo-generated electrons/holes even after the removal of light stimuli. In this case, the optical information can be stored in the ORAM. Because the light intensity determines the amount of photo-generated charge carriers, which is generally related to the possibility of charge trapping or interfacial physics. Thus, tuning the light intensity provides an approach to achieve multiple resistance states. Additionally, light stimuli with different wavelengths correspond to different optical excitation energies and different amount of light absorption, which may also induce to different charge trapping/detrapping physics and multiple resistance states.

In 2014, Lei et al. demonstrated multilayered CuIn7Se11 based ORAM array for image sensing, and further extended to other atomically thin 2D materials (InSe and MoS2) [40]. The design concept of the ORAM is to build potential well with an energy band bending between 2D material channel and metal electrodes, trapping photo-generated electrons. Although the ORAM exhibits good photo-response, its short retention time (less than 50 s) and low ON/OFF ratio (less than 10) limit the applications for multilevel nonvolatile memories. Further, Lee et al. introduced artificially charge traps at the MoS2/SiO2 interface through O2 plasma treatment of SiO2 before MoS2 monolayer growth (Figure 5(a)). The band alignments of the initial state, upon light exposure and reading process for ON and OFF states are illustrated in Figure 5(b). At the initial state, a potential well is formed by metal-semiconductor Schottky barrier with the applied gate voltage. A positive gate voltage pulse first programs the device to OFF state by driving the electrons in the channel to the interfacial trap states, leading to a low OFF current of 4 pA (=3V). Upon light illumination (450 nm laser), photo-generated electrons are accumulated in the potential well, while holes escape from both electrodes through upward energy band bending. After the removal of light illumination, the readout process (readout bias is applied) eliminates one of the Schottky barriers and releases the stored electrons. By intentionally introducing charge traps, the MoS2 based ORAM device exhibits improved ON/OFF ratio of 4700 and retention time of 6000 s. However, the photo-current still suffers from relatively large degradation after the removal of light exposure [41].

Towards the design of ORAMs with high photo-responsivity, large ON/OFF ratio, low dark current and good retention properties, researchers adopted other charge trapping methods to fabricate multilevel and highly responsive nonvolatile 2D ORAM. Cho group successfully reported a multibit MoS2 ORAM with high ON/OFF ratio of 107, long retention time over 104 s, and high nonvolatile responsivity exceeding 8000 A/W at an illumination power of 0.1 μW [42]. They employed a layer of Au nanoparticles as the charge trapping layer and floating gate between the blocking and tunneling dielectrics to store photo-generated charge carriers in the MoS2 ORAM (Figure 5(a)). The MoS2 ORAM can be programmed in both electrical and optical manner, which enables versatile applications, such as performing logic operations and coincidence detection [3]. The schematic band diagrams of electrical and optical programming are illustrated in Figures 5(d) and 5(e), respectively. For the electrical programming, a negative gate voltage induces the transfer of electrons in Au nanoparticle layer to the conduction band of MoS2 layer through Fowler-Nordheim tunneling; for the optical programming, the electrons can be easily transferred from Au nanoparticle layer to MoS2 valence band with the light illumination (655 nm laser), which prevents the recombination of photo-generated electrons and holes in the MoS2 layer. The programmed resistance states can be erased by applying a positive gate voltage. The MoS2 ORAM exhibits light intensity-dependent multilevel resistance states, integrating both optical sensing and nonvolatile data storage (Figure 5(f)).

In addition to the typical FET ORAMs, ferroelectric ORAM based on monolayer MoS2 has also exhibited unique characteristics, including nondestructive data readout and low operating voltage [43] (Figures 6(a) and 6(b)). The optical illumination can induce the domain re-arrangement in the MoS2-PZT ORAM, which further causes the change in conductivity. The ferroelectric ORAM can be globally optically SET and electrically RESET or electrically SET and optically RESET, depending on the initial polarization states of PZT (Figure 6(c)). Wang et al. fabricated a two-terminal MoS2 ORAM with a structure of Au/MoS2 nanospheres/Au [44]. With a poling voltage of 6 V, upon the light illumination the device can be switched from HRS to LRS at -2.9 V with an ON/OFF ratio of 10. A positive voltage of 4.2 V can switch the device back to HRS. By combining different poling voltage and light illumination, multilevel resistance states can be obtained in the two-terminal MoS2 ORAM device, which provides possibilities to the design of two-terminal ORAM based on 2D materials.

3.3. 2D Heterostructure Based Nonvolatile ORAM

The 2D material based ORAMs are generally sensitive to environmental factors because most of the memory functions in ORAMs are dependent on the charge trapping carriers in original or artificial defects at the materials interface. The charge trapping mechanism usually leads to relatively short retention time in the memory device. By combining the 2D materials with other semiconductors, the ORAMs based on 2D heterostructure can widen ORAM design possibilities and improve the ORAM performance, including the light absorption range, environmental stability and responsivity, nonvolatile characteristics and data storage capability.

Roy et al. reported an ORAM based on graphene/MoS2 vertical heterostructure, which demonstrates a high photo-responsivity of 5×108 A/W at room temperature. However, the switching ON/OFF ratio is less than 2, owing to the high electrical conductivity of graphene and the large dark current of the device. The low ON/OFF ratio restricts the application for multibit ORAM [45]. Later, Du et al. designed a broadband (ultraviolet to visible region) ORAM based on WSe2/h-BN heterostructure (Figure 7(a)), which exhibits a large ON/OFF ratio of 1.1×106, over 128 distinct resistance states and remarkable retention time over 4.5×104 s [46]. Upon the light illumination (405 nm laser), the photo-generated electrons in BN conduction band can be transferred to WSe2 conduction band under a negative gate voltage of -20 V (programming process). The photo-generated holes are left and localized in the middle of BN bandgap even after the removal of light stimulation and electric field, which indicates the nonvolatile switching characteristics. A positive gate voltage with light illumination erases the device and turns the device back to the initial states by driving holes from BN to WSe2 (Figure 7(b)). Figure 7(c) illustrates 130 resistance states obtained by applying an increased number of light pulses. The WSe2/h-BN can respond to the lights with a wide range of wavelength from 410 nm to 750 nm. A color image sensor array integrated with sensing and storage capability exhibits promising potentials for the future design of smart image sensor (Figure 7(c) inset). The heterostructure configuration can also be extended to other materials, such as BP/ BN based ORAM, which exhibits ultrahigh nonvolatile responsivity of 1.2×107 A/W [46]. However, the three-terminal FET ORAM usually involves with high programming gate voltage, relatively low integration level and complex circuitry for integration. Two-terminal ORAM is highly desirable for the design of the device with simple structure, high integration level and low power consumption.

To further decrease the programming voltage and power consumption in ORAM device, Tran et al. demonstrated a nonvolatile multibit ORAM based on MoS2/BN/graphene van der Waals heterostructure, showing a programming voltage (source-drain voltage) of -10 V, extremely low dark-current of 10−14 A and high nonvolatile responsivity of 2×104 A/W (@458 nm, 2nW) (Figure 7(c)) [47]. A negative source-drain voltage of -10 V induces the electrons tunneling through BN to graphene floating gate. The accumulated electrons in the graphene layer serve as a negative gate bias, thus programs the device with extremely low dark current (LRS). The visible light pulses then induce the photo-generated holes in MoS2 layer, tunneling through the low hole barrier of MoS2/BN and combining with the stored electrons in the graphene layer, and RESET the device to HRS (Figure 7(d)). This operation scheme leads to a high ON/OFF ratio of 106 and over 16 distinct storage levels by tuning the light intensities [47]. Hu et al. reported a two-terminal and vertical ORAM based on BP/ZnO nanoparticles heterojunction, exhibiting broadband absorption from ultraviolet (380 nm) to near infrared (785 nm) and low SET voltage < 4V (Figure 7(e)).The optically-controlled resistive switching originates from fast electron/hole separation in the heterojunction, which is further involved in the manipulation of the formation/rupture of oxygen vacancy filament [48].

Most of the absorption spectrum in ORAM device currently is limited in the ultraviolet and visible region. As infrared radiation is the communication medium for data transfer in wireless networks, night vision, image sensing, medical diagnosis, etc., the extension of absorption to the infrared region in ORAM is highly desirable, which not only enables the conversion and storage of infrared data, but also benefits for the optical data communication in the photonic circuit and artificial visual intelligence. The application of 2D heterostructures provides possibilities in designing infrared ORAMs by combining 2D materials with small bandgap semiconductors or up-conversion nanoparticles. Wang et al. reported a nonvolatile infrared ORAM by adopting a MoS2/PbS van der Waals heterostructure, which can respond to a wide wavelength range from 880 nm to 1940 nm (Figure 7(f)) [49]. The optical switching originates from the localized photo-generated holes in PbS near the interface and the injection of photo-generated electrons from PbS to MoS2, thus modulating the MoS2 conductivity. A retention time of over 104 s and multi-level storage states can be obtained by applying 1940 nm infrared light pulses. The nonvolatile responsivity reaches up to 1.2×104 A/W. The up-conversion nanoparticles based on lanthanide ions-doped NaYF4 with narrow absorption band at 980 nm also provide possibilities for 2D materials based infrared ORAM. The up-conversion nanoparticles/MoS2 serve as charge trapping and separation centers upon the infrared light illumination, which is responsible for the optical assisted switching process [50]. The up-conversion nanoparticles/MoS2 based infrared ORAM enables to sense and store the 980 nm infrared light with different light intensities, corresponding to different resistance states.

Table 1 compares the optical switching characteristics of nonvolatile ORAMs based on 2D materials and their heterostructures with other material systems. The nonvolatile ORAMs based on 2D materials have exhibited high responsivity, low optical power consumption, large ON/OFF ratio, good endurance, and broadband absorption compared with that based on other materials (e.g., oxides, hybrid organic-inorganic perovskite, organic materials), however, the high programming voltage in most of 2D material ORAM is required to be further lowered to achieve a low power consumption in the future.

Active layerThree /two terminalControl gate/floating gateSwitching mechanismLight wavelength (nm)Light intensity (μW cm−2)Light pulse width (ms)Non-volatile responsivity (A/W)Max On/off Retention (s)Programming voltage (V)Multi-levelEnduran-ceRef.

graphene/MoS2ThreeSiO2/--Charge trapping635 nm, white light37.630-100--<210550Yes--[45]

MoS2ThreeFunctionalized SiO2/--Charge trapping450, 650 nm0.019 μW1-3000--470010480Yes--[41]

CuIn7Se11ThreeSiO2/--Charge trapping543 nm1.5×10550-500--<105080----[40]

MoS2/PVP/AuNPsThreeSiO2/AuNPsCharge trapping655 nm0.1-1000 μW1000800010710410Yes200[42]

MoS2/PZTThreeSiO2/--Ferroelectrichalogen bulb150 W----221046--500[43]

WSe2/BNThreeSiO2/--Charge trapping410-750 nm2 nW500--1064.5×10420Yes (128)200[46]

BP/BNThreeSiO2/--Charge trapping410-500 nm2 nW5001.2×1074154×10415Yes200[46]

MoS2/h-BN/grapheneTwo--/grapheneCharge trapping458,638, 725, 811 nm2nW-160nW10-20002×104 @458 nm@2nW1064.5×10410Yes (18)104[47]

BP/ZnOTwo--Oxygen filament 380, 532, 633, 785 nm20-500----106--5----[48]

MoS2/PbSThreeSiO2/--Charge trapping850, 1310, 1550 nm2710001.2×104@1940 nm150104No programming voltage requiredYes2000[49]

MoS2/NaYF4:Yb3+, Er3+Two--Charge trapping980 nm0-5×105----1041040.5Yes200[50]

CH3NH3Two--Charge trappingWhite light0-320010001×1031041040.1Yes400[3]

Ce /AlTwo--Charge trapping499, 560, 638 nm40020000130104No programming voltage requiredYes30[4]

ZnO nanorods/Nb-doped SrTiO3Two--Charge trapping365 nm2.5×106--1.31031042----[51]

ROT300/VOTc/ZnO nanparticlesThreeAl/ZnO nanparticlesCharge trapping850 nm80-105001000--104106No programming voltage requiredYes--[52]

BBTNDTThreeSiO2/--Charge trappingWhite light4000010000433106106120----[53]

4. 2D Materials Based ORAM Synaptic Device

4.1. ORAM Synaptic Device versus Photodetector

The integration of optical sensing and data storage in ORAM enables direct storage of optical information from the outside world without optical-to-electrical conversion, however, these functions are restricted in the digital applications. Towards the future low power neuromorphic computing, ORAM synaptic devices are desirable that provide possibilities to bridge the sensing and neuromorphic computing and the direct processing of optical information (e.g., image processing) [18]. In comparison to conventional photodetectors and nonvolatile ORAMs, ORAM synaptic devices exhibit both optically-controlled nonvolatile and volatile switching characteristics. It can respond to optical stimuli and exhibit light-tunable synaptic behaviors, such as STP, LTP and STDP. Figure 8 illustrates the light modulating schemes in photodetectors (a) and ORAM synaptic devices (b). In sharp contrast to the conventional photodetector or image sensors, ORAM synaptic devices sense information in ways like neural signals with light-tunable and time-dependent plasticity, implicating the potential application in the construction of neuromorphic visual systems.

4.2. ORRAM Synaptic Device

Qin et al. first reported an ORAM synaptic device based on graphene/carbon nanotube (CNT) heterostructure, demonstrating light-tunable synaptic behaviors, such as, light-tunable STP, LTP, STLR and STDP (Figure 9(a)) [54]. The channel conductance can be tuned through optical pulse and gate pulse. The multiple state conductance can be regarded as the biological synaptic weight. Figure 9(b) shows the STP stimulated with an optical pulse (405 nm, 50 μW, 5 ms), where the conductance increases abruptly and relaxes within seconds. The STP is triggered owing to the electron transfer from CNT to graphene upon the light illumination, tuning the built-in field formed at graphene/CNT interface. Optically tuned paired-pulse facilitation (PPF) function is an important function for temporal visual information encoding, as illustrated in Figure 9(c). The plasticity can be further modulated to LTP by combining a gate voltage and an optical pulse (Figure 9(d)). The LTP originates from the trapping of photo-generated holes from graphene or SWCNT at the SiO2 surface owing to the defects. A STDP learning rule (symmetric STDP) is also demonstrated in this work (Figure 9(f)), which is related to associative learning and enables to the application of coincidence detector for spiking activity. Both the non-Hebbian learning rule (e.g., STLR) and Hebbian learning rule (e.g., STDP) are mimicked (Figures 9(e) and 7(f)), respectively. The implementations of STLR and STDP are associated with the spatiotemporal information processing in the neural network. Similar STLR and STDP learning rules can also be performed in ion-gated 2D ORAMs [55, 56].

Other types of 2D materials heterostructures have also been utilized for ORAM synaptic devices, such as MoS2/perylene-3,4,9,10-tetracarboxylic dianhydride (PTCDA) heterojunction, exhibiting large synaptic weight change and PPF ratio [22]. To further broaden the absorption spectrum and reduce the power consumption, a hybrid structure of 2D WSe2/boron-doped Si nanocrystals was employed for ORAM synaptic device with wide absorption spectrum from ultraviolet to near-infrared and low power consumption of 75 fJ, which points the way to the development of low-power artificial visual intelligence beyond human eyes [57]. Table 2 compares the basic performances of ORAM synaptic devices based on 2D materials with other different active layers.

Active layerThree /two terminalOperation mechanismSpike wavelength Spike energy (pJ)Ref.

graphene/CNTThreeCharge trapping405, 532 nm250[54]
Si-NC/WSe2ThreeCharge trapping375, 532 and 1342 nm0.000075[57]
MoS2/PTCDAThreeCharge trapping532 nm5.38 [22]
2D Perovskite/grapheneThree--520 nm0.592[58]
CH3NH3PbI3TwoCharge trapping500, 635 nm and halogen bulb--[2]
CdS/CNTTwoCharge trapping405 and 532 nm--[59]
InGaZn/Al2O3ThreeCharge trapping365 nm2.4×103[60]
Pentacene/PMMA/CsPbBr3 quantum dotsThreeCharge trapping 365, 450, 520 and 660 nm0.76[61]
ITO/Nb:SrTiO3TwoCharge trapping459, 528 and 630 nm1178[62]
C8-BTBTThreeCharge trapping 360 nm2160[63]

4.3. Applications in Neural Network and Image Processing

The multiple synaptic functions demonstrated in the ORAM synaptic device facilitate the coupling of optical sensing and neuromorphic computing, and enable the applications in the optical neural network for image processing (e.g., pattern recognition). Generally, for the conventional image processing based on neuromorphic computing, the optical signals in the images are first converted to electronic signals through image sensors. The electronic signals are then transferred to the synaptic arrays in the neural network for further processing. In contrast to the conventional memory synaptic devices operated by electrical method, the ORAM synaptic devices respond to the optical stimuli directly and exhibit light-evoked synaptic behaviors. The constructed optic-neural network through ORAM synaptic devices can potentially simplify the circuitry for neuromorphic image processing and reduce the power consumption during the data transmission. Additionally, the high integration level in ORAM synaptic device provides future miniaturized and power-efficient artificial eye design [18]. Tian et al. first developed an optic-neural network (ONN) with optical synapses based on graphene/2D perovskite, for which the pattern recognition rate in this ONN can reach up to 80 % [58]. Seo et al. further demonstrated an ONN for mimicking color-mixed pattern recognition in human vision. The ORAM synaptic device is integrated with a WSe2/h-BN photodetector and a synaptic device based on WSe2/weight control layer (WCL)/h-BN (Figure 10). The optical synapse can respond to the lights with different wavelengths (red, green and blue), showing different synaptic weight change for long-term potentiation/long-term depression (Figure 10(b)). According to the wavelength-dependent synaptic weight change, an ONN (Figure 10(c)) is developed for the task of color-mixed pattern recognition, which differentiates from the conventional neural network (NN) without wavelength-dependent synaptic weight control. The recognition rate for mixed-color digits (Figure 10(d)) in ONN is around 60% higher than that in NN [64].

5. Conclusions and Perspectives: 2D Materials in Designing Neuromorphic Devices

Owing to the strong light-matter interaction, unique physical properties, electrostatic tunability and downward scaling potential, 2D materials and their heterostructures have been widely employed for optoelectronics and memory applications in recent years. Towards the trend of functional diversification in “MtM”, the integration of optical sensing, data storage and processing drives the design of ORAM and optoelectronic synaptic devices, which can potentially simplify the circuit complexity and reduce the power consumption for the photonic circuit and artificial vision. In contrast to the conventional image sensors with light-intensity sensing, the ORAM device with light tunable plasticity (light-dosage sensing) can further be applied for the construction of ONN for pattern recognition, potentially bridging the sensory data and neural network in the neuromorphic visual system. The nonvolatile ORAMs based on 2D materials heterostructures have exhibited high photoresponsivity, stable multilevel data storage and good retention and endurance properties. Although ORAM devices based on 2D materials have exhibited various advantages over other materials, the high programming voltage is still a main issue, which causes high power consumption. Additionally, the three-terminal structure in most of the 2D material ORAMs restricts the integration density in the integrated circuits. To develop appropriate 2D material ORAMs with two-terminal structure can be a potential approach for low-power and high-integration-density devices in the future.

The ORAM synaptic device that integrates optical sensing and synaptic functions have exhibits great potentials in neuromorphic computing and visual information processing. For the future artificial intelligence, to mimic the sensing and processing of auditory or olfactory information like the human brain is also demanded for energy-efficient and miniaturized processing system. As the conventional visual, olfactory and auditory sensors generate a large amount of redundant data, neuromorphic sensory devices potentially allow for the real-time storage and processing of the sensory data with reduced power consumption. Similar to the extension from optical sensors to ORAM synaptic devices, olfactory sensors (e.g., gas sensors) and the auditory sensors (e.g., sound detectors) can also be potentially developed as neuromorphic sensors with synaptic behaviors through appropriate structure design. For both sound detectors and gas sensors, 2D materials have also exhibited their potentials in achieving high sensitivity. More specifically, 2D materials (TMDs, BP, silicene, MXene, etc.) are adopted as promising candidates for high-sensitive gas sensors because of the large surface-to-volume ratio and tunable functionality of the surface through decorating transition metals or functional groups [6569]. The excellent mechanical properties and piezoelectric properties in 2D materials also implicate the applications of sound detectors [7072]. The outstanding performances in 2D based gas sensors and sound detectors suggest the possibilities for future designs of neuromorphic olfactory sensors and auditory sensors, which could contribute to important applications in the development of edge computing and Internet of Things.

Conflicts of Interest

The authors declare that they have no conflicts of interest.


This work was supported by Research Grant of Council of Hong Kong (PolyU 152053/18E), the Hong Kong Polytechnic University (G-YBPS and G-SB79), and National Natural Science Foundation of China (61851402 and 61861166001).


  1. T. Leydecker, M. Herder, E. Pavlica et al., “Flexible non-volatile optical memory thin-film transistor device with over 256 distinct levels based on an organic bicomponent blend,” Nature Nanotechnology, vol. 11, no. 9, pp. 769–775, 2016. View at: Publisher Site | Google Scholar
  2. X. Zhu and W. D. Lu, “Optogenetics-inspired tunable synaptic functions in memristors,” ACS Nano, vol. 12, no. 2, pp. 1242–1249, 2018. View at: Publisher Site | Google Scholar
  3. F. Zhou, Y. Liu, X. Shen, M. Wang, F. Yuan, and Y. Chai, “Low‐voltage, optoelectronic CH3NH3PbI3− xClx memory with integrated sensing and logic operations,” Advanced Functional Materials, vol. 28, no. 15, Article ID 1800080, 2018. View at: Google Scholar
  4. H. Tan, G. Liu, X. Zhu et al., “An optoelectronic resistive switching memory with integrated demodulating and arithmetic functions,” Advanced Materials, vol. 27, no. 17, pp. 2797–2803, 2015. View at: Publisher Site | Google Scholar
  5. J. Shim, S. Jo, M. Kim, Y. J. Song, J. Kim, and J. Park, “Light-triggered ternary device and inverter based on heterojunction of van der waals materials,” ACS Nano, vol. 11, no. 6, pp. 6319–6327, 2017. View at: Publisher Site | Google Scholar
  6. Z. Cheng, C. Ríos, W. H. Pernice, C. D. Wright, and H. Bhaskaran, “On-chip photonic synapse,” Science Advances, vol. 3, no. 9, Article ID e1700160, 2017. View at: Google Scholar
  7. H. Tan, G. Liu, H. Yang et al., “Light-gated memristor with integrated logic and memory functions,” ACS Nano, vol. 11, no. 11, pp. 11298–11305, 2017. View at: Publisher Site | Google Scholar
  8. X. Hong, D. J. Loy, P. A. Dananjaya, F. Tan, C. Ng, and W. Lew, “Oxide-based RRAM materials for neuromorphic computing,” Journal of Materials Science, vol. 53, no. 12, pp. 8720–8746, 2018. View at: Publisher Site | Google Scholar
  9. C. Li, M. Hu, Y. Li et al., “Analogue signal and image processing with large memristor crossbars,” Nature Electronics, vol. 1, no. 1, pp. 52–59, 2018. View at: Publisher Site | Google Scholar
  10. M. Kimura, T. Shima, and T. Yamashita, “Artificial retina using thin-film photodiode and thin-film transistor,” ECS Transactions, vol. 3, no. 8, pp. 325–331, 2006. View at: Google Scholar
  11. S. Choi, S. H. Tan, Z. Li et al., “SiGe epitaxial memory for neuromorphic computing with reproducible high performance based on engineered dislocations,” Nature Materials, vol. 17, no. 4, pp. 335–340, 2018. View at: Publisher Site | Google Scholar
  12. J. Kang, Z. Yu, L. Wu et al., “Time-dependent variability in RRAM-based analog neuromorphic system for pattern recognition,” Electron Devices Meeting (IEDM), pp. 6.4.1–6.4.4, 2017. View at: Google Scholar
  13. Y. Kim, A. Chortos, W. Xu et al., “A bioinspired flexible organic artificial afferent nerve,” Science, vol. 360, no. 6392, pp. 998–1003, 2018. View at: Publisher Site | Google Scholar
  14. G. J. Lee, C. Choi, D. Kim, and Y. M. Song, “Bioinspired artificial eyes: optic components, digital cameras, and visual prostheses,” Advanced Functional Materials, vol. 28, no. 24, Article ID 1705202, 2017. View at: Publisher Site | Google Scholar
  15. Y. M. Song, Y. Xie, V. Malyarchuk et al., “Digital cameras with designs inspired by the arthropod eye,” Nature, vol. 497, no. 7447, pp. 95–99, 2013. View at: Publisher Site | Google Scholar
  16. K.-H. Jeong, J. Kim, and L. P. Lee, “Biologically inspired artificial compound eyes,” Science, vol. 312, no. 5773, pp. 557–561, 2006. View at: Google Scholar
  17. H. C. Ko, M. P. Stoykovich, J. Song et al., “A hemispherical electronic eye camera based on compressible silicon optoelectronics,” Nature, vol. 454, no. 7205, pp. 748–753, 2008. View at: Publisher Site | Google Scholar
  18. F. Zhou, Z. Zhou, J. Chen et al., “Optoelectronic resistive random access memory for neuromorphic vision sensors,” Nature Nanotechnology, vol. 14, pp. 776–782, 2019. View at: Google Scholar
  19. F. Yan, Z. Wei, X. Wei, Q. Lv, W. Zhu, and K. Wang, “Toward high-performance photodetectors based on 2D materials: strategy on methods,” Small Methods, vol. 2, no. 5, Article ID 1700349, 2018. View at: Publisher Site | Google Scholar
  20. N. Huo and G. Konstantatos, “Recent progress and future prospects of 2D-based photodetectors,” Advanced Materials, vol. 30, no. 51, Article ID e1801164, 2018. View at: Publisher Site | Google Scholar
  21. H. Qiao, J. Yuan, Z. Xu et al., “Broadband photodetectors based on graphene–Bi2Te3 heterostructure,” ACS Nano, vol. 9, no. 2, pp. 1886–1894, 2015. View at: Publisher Site | Google Scholar
  22. S. Wang, C. Chen, Z. Yu et al., “A MoS2/PTCDA hybrid heterojunction synapse with efficient photoelectric dual modulation and versatility,” Advanced Materials, vol. 31, no. 3, Article ID e1806227, 2018. View at: Publisher Site | Google Scholar
  23. Q. Guo, A. Pospischil, M. Bhuiyan et al., “Black phosphorus mid-infrared photodetectors with high gain,” Nano Letters, vol. 16, no. 7, pp. 4648–4655, 2016. View at: Publisher Site | Google Scholar
  24. M. Liu, X. Yin, E. Ulin-Avila et al., “A graphene-based broadband optical modulator,” Nature, vol. 474, no. 7349, pp. 64–67, 2011. View at: Publisher Site | Google Scholar
  25. A. Di Bartolomeo, G. Luongo, F. Giubileo et al., “Hybrid graphene/silicon Schottky photodiode with intrinsic gating effect,” 2D Materials, vol. 4, no. 2, Article ID 025075, 2017. View at: Google Scholar
  26. T. Mueller, F. Xia, M. Freitag, J. Tsang, and P. Avouris, “Role of contacts in graphene transistors: A scanning photocurrent study,” Physical Review B: Condensed Matter and Materials Physics, vol. 79, no. 24, Article ID 245430, 2009. View at: Google Scholar
  27. Y. Xu, J. Yuan, L. Fei et al., “Selenium-doped black phosphorus for high-responsivity 2D photodetectors,” Small, vol. 12, no. 36, pp. 5000–5007, 2016. View at: Publisher Site | Google Scholar
  28. C. Liu, X. Yan, X. Song, S. Ding, D. W. Zhang, and P. Zhou, “A semi-floating gate memory based on van der Waals heterostructures for quasi-non-volatile applications,” Nature Nanotechnology, vol. 13, no. 5, pp. 404–410, 2018. View at: Publisher Site | Google Scholar
  29. M. Wang, S. Cai, C. Pan et al., “Robust memristors based on layered two-dimensional materials,” Nature Electronics, vol. 1, no. 2, pp. 130–136, 2018. View at: Publisher Site | Google Scholar
  30. S. Bertolazzi, P. Bondavalli, S. Roche et al., “Nonvolatile memories based on graphene and related 2D materials,” Advanced Materials, vol. 31, no. 10, Article ID e1806663, 2019. View at: Publisher Site | Google Scholar
  31. X. Wu, R. Ge, P. A. Chen et al., “Thinnest nonvolatile memory based on monolayer h-BN,” Advanced Materials, vol. 31, no. 15, Article ID e1806790, 2019. View at: Google Scholar
  32. R. Ge, X. Wu, M. Kim et al., “Atomristor: nonvolatile resistance switching in atomic sheets of transition metal dichalcogenides,” Nano Letters, vol. 18, no. 1, pp. 434–441, 2017. View at: Publisher Site | Google Scholar
  33. Y. Zhao, Y. Xie, Z. Liu, X. Wang, Y. Chai, and F. Yan, “Two-dimensional material membranes: An emerging platform for controllable mass transport applications,” Small, vol. 10, no. 22, pp. 4521–4542, 2014. View at: Publisher Site | Google Scholar
  34. Y. Zhao, K. Xu, F. Pan, C. Zhou, F. Zhou, and Y. Chai, “Doping, contact and interface engineering of two-dimensional layered transition metal dichalcogenides transistors,” Advanced Functional Materials, vol. 27, no. 19, Article ID 1603484, 2017. View at: Publisher Site | Google Scholar
  35. J. Cheng, C. Wang, X. Zou, and L. Liao, “Recent advances in optoelectronic devices based on 2d materials and their heterostructures,” Advanced Optical Materials, vol. 7, no. 1, Article ID 1800441, 2019. View at: Google Scholar
  36. S. Bertolazzi, D. Krasnozhon, and A. Kis, “Nonvolatile memory cells based on MoS2/graphene heterostructures,” ACS Nano, vol. 7, no. 4, pp. 3246–3252, 2013. View at: Publisher Site | Google Scholar
  37. M. S. Choi, G.-H. Lee, Y.-J. Yu et al., “Controlled charge trapping by molybdenum disulphide and graphene in ultrathin heterostructured memory devices,” Nature Communications, vol. 4, no. 1624, 2013. View at: Google Scholar
  38. Q. A. Vu, Y. S. Shin, Y. R. Kim et al., “Two-terminal floating-gate memory with van der Waals heterostructures for ultrahigh on/off ratio,” Nature Communications, vol. 7, no. 1, article 12725, 2016. View at: Publisher Site | Google Scholar
  39. Y. Shi, X. Liang, B. Yuan et al., “Electronic synapses made of layered two-dimensional materials,” Nature Electronics, vol. 1, no. 8, pp. 458–465, 2018. View at: Publisher Site | Google Scholar
  40. S. Lei, F. Wen, B. Li et al., “Optoelectronic memory using two-dimensional materials,” Nano Letters, vol. 15, no. 1, pp. 259–265, 2014. View at: Publisher Site | Google Scholar
  41. J. Lee, S. Pak, Y. Lee et al., “Monolayer optical memory cells based on artificial trap-mediated charge storage and release,” Nature Communications, vol. 8, no. 1, article 14734, 2017. View at: Publisher Site | Google Scholar
  42. D. Lee, E. Hwang, Y. Lee et al., “Multibit MoS2 photoelectronic memory with ultrahigh sensitivity,” Advanced Materials, vol. 28, no. 41, pp. 9196–9202, 2016. View at: Publisher Site | Google Scholar
  43. A. Lipatov, P. Sharma, A. Gruverman, and A. Sinitskii, “Optoelectrical Molybdenum Disulfide (MoS2) - Ferroelectric Memories,” ACS Nano, vol. 9, no. 8, pp. 8089–8098, 2015. View at: Publisher Site | Google Scholar
  44. W. Wang, G. N. Panin, X. Fu et al., “MoS2 memristor with photoresistive switching,” Scientific Reports, vol. 6, no. 31224, 2016. View at: Google Scholar
  45. K. Roy, M. Padmanabhan, S. Goswami et al., “Graphene–MoS2 hybrid structures for multifunctional photoresponsive memory devices,” Nature Nanotechnology, vol. 8, no. 11, pp. 826–830, 2013. View at: Publisher Site | Google Scholar
  46. D. Xiang, T. Liu, J. Xu et al., “Two-dimensional multibit optoelectronic memory with broadband spectrum distinction,” Nature Communications, vol. 9, no. 1, article 2966, 2018. View at: Publisher Site | Google Scholar
  47. M. D. Tran, H. Kim, J. S. Kim et al., “Two-terminal multibit optical memory via van der waals heterostructure,” Advanced Materials, vol. 31, no. 7, Article ID 1807075, 2019. View at: Publisher Site | Google Scholar
  48. L. Hu, J. Yuan, Y. Ren et al., “Phosphorene/ZnO nano-heterojunctions for broadband photonic nonvolatile memory applications,” Advanced Materials, vol. 30, no. 30, Article ID 1801232, 2018. View at: Publisher Site | Google Scholar
  49. Q. Wang, Y. Wen, K. Cai et al., “Nonvolatile infrared memory in MoS2/PbS van der Waals heterostructures,” Science Advances, vol. 4, no. 4, Article ID eaap7916, 2018. View at: Google Scholar
  50. Y. Zhai, X. Yang, F. Wang et al., “Infrared-sensitive memory based on direct-grown MoS2-upconversion-nanoparticle heterostructure,” Advanced Materials, vol. 30, no. 49, Article ID 1803563, 2018. View at: Google Scholar
  51. A. Bera, H. Peng, J. Lourembam, Y. Shen, X. W. Sun, and T. Wu, “A versatile light-switchable nanorod memory: Wurtzite ZnO on perovskite SrTiO3,” Advanced Functional Materials, vol. 23, no. 39, pp. 4977–4984, 2013. View at: Publisher Site | Google Scholar
  52. H. Wang, H. Liu, Q. Zhao et al., “A retina-like dual band organic photosensor array for filter-free near-infrared-to-memory operations,” Advanced Materials, vol. 29, no. 32, Article ID 1701772, 2017. View at: Google Scholar
  53. K. Pei, X. Ren, Z. Zhou, Z. Zhang, X. Ji, and P. K. Chan, “A high-performance optical memory array based on inhomogeneity of organic semiconductors,” Advanced Materials, vol. 30, no. 13, Article ID 1706647, 2018. View at: Publisher Site | Google Scholar
  54. S. Qin, F. Wang, Y. Liu et al., “A light-stimulated synaptic device based on graphene hybrid phototransistor,” 2D Materials, vol. 4, no. 3, Article ID 035022, 2017. View at: Google Scholar
  55. J. Jiang, W. Hu, D. Xie et al., “2D electric-double-layer phototransistor for photoelectronic and spatiotemporal hybrid neuromorphic integration,” Nanoscale, vol. 11, no. 3, pp. 1360–1369, 2019. View at: Publisher Site | Google Scholar
  56. R. A. John, F. Liu, N. A. Chien et al., “Synergistic gating of electro-iono-photoactive 2D chalcogenide neuristors: coexistence of hebbian and homeostatic synaptic metaplasticity,” Advanced Materials, vol. 30, no. 25, Article ID 1800220, 2018. View at: Google Scholar
  57. Z. Ni, Y. Wang, L. Liu et al., “Hybrid structure of silicon nanocrystals and 2D WSe 2 for broadband optoelectronic synaptic devices,” IEEE International Electron Devices Meeting (IEDM), pp. 38.5.1–38.5.4, 2018. View at: Google Scholar
  58. H. Tian, X. Wang, F. Wu, Y. Yang, and T.-L. Ren, “High performance 2D perovskite/graphene optical synapses as artificial eyes,” IEEE International Electron Devices Meeting (IEDM), pp. 38.6.1–38.6.4, 2018. View at: Google Scholar
  59. K. Pilarczyk, A. Podborska, M. Lis et al., “Synaptic behavior in an optoelectronic device based on semiconductor‐nanotube hybrid,” Advanced Electronic Materials, vol. 2, no. 6, Article ID 1500471, 2016. View at: Google Scholar
  60. H. K. Li, T. Chen, P. Liu et al., “A light-stimulated synaptic transistor with synaptic plasticity and memory functions based on InGaZnOx–Al2O3 thin film structure,” Journal of Applied Physics, vol. 119, no. 24, Article ID 244505, 2016. View at: Google Scholar
  61. Y. Wang, Z. Lv, J. Chen et al., “Photonic synapses based on inorganic perovskite quantum dots for neuromorphic computing,” Advanced Materials, vol. 30, no. 38, Article ID 1802883, 2018. View at: Google Scholar
  62. S. Gao, G. Liu, H. Yang et al., “An oxide Schottky junction artificial optoelectronic synapse,” ACS Nano, vol. 13, no. 2, pp. 2634–2642, 2019. View at: Google Scholar
  63. S. Dai, X. Wu, D. Liu et al., “Light-stimulated synaptic devices utilizing interfacial effect of organic field-effect transistors,” ACS Applied Materials & Interfaces, vol. 10, no. 25, pp. 21472–21480, 2018. View at: Publisher Site | Google Scholar
  64. S. Seo, S.-H. Jo, S. Kim et al., “Artificial optic-neural synapse for colored and color-mixed pattern recognition,” Nature Communications, vol. 9, no. 1, article 5106, 2018. View at: Publisher Site | Google Scholar
  65. J. Park, J. Mun, J.-S. Shin, and S.-W. Kang, “Highly sensitive two-dimensional MoS2 gas sensor decorated with Pt nanoparticles,” Royal Society Open Science, vol. 5, no. 12, Article ID 181462, 2018. View at: Publisher Site | Google Scholar
  66. S. Cui, H. Pu, and S. A. Wells, “Ultrahigh sensitivity and layer-dependent sensing performance of phosphorene-based gas sensors,” Nature Communications, vol. article, no. 8632, 2015. View at: Google Scholar
  67. X.-P. Chen, L.-M. Wang, X. Sun et al., “Sulfur dioxide and nitrogen dioxide gas sensor based on arsenene: a first-principle study,” IEEE Electron Device Letters, vol. 38, no. 5, pp. 661–664, 2017. View at: Publisher Site | Google Scholar
  68. V. Q. Bui, T. Pham, D. A. Le, C. M. Thi, and H. M. Le, “A first-principles investigation of various gas (CO, H2O, NO, and O2) absorptions on a WS2 monolayer: stability and electronic properties,” Journal of Physics: Condensed Matter, vol. 27, no. 30, Article ID 305005, 2015. View at: Publisher Site | Google Scholar
  69. A. N. Abbas, B. Liu, L. Chen et al., “Black phosphorus gas sensors,” ACS Nano, vol. 9, no. 5, pp. 5618–5624, 2015. View at: Publisher Site | Google Scholar
  70. L.-Q. Tao, H. Tian, Y. Liu et al., “An intelligent artificial throat with sound-sensing ability based on laser induced graphene,” Nature Communications, vol. 8, no. 1, article 14579, 2017. View at: Publisher Site | Google Scholar
  71. R. Hinchet, U. Khan, C. Falconi, and S.-W. Kim, “Piezoelectric properties in two-dimensional materials: Simulations and experiments,” Materials Today, vol. 21, no. 6, pp. 611–630, 2018. View at: Publisher Site | Google Scholar
  72. C. Cui, F. Xue, W. Hu, and L. Li, “Two-dimensional materials with piezoelectric and ferroelectric functionalities,” npj 2D Materials and Applications, vol. 2, no. 1, article 18, 2018. View at: Publisher Site | Google Scholar

Copyright © 2019 Feichi Zhou et al. Exclusive licensee Science and Technology Review Publishing House. Distributed under a Creative Commons Attribution License (CC BY 4.0).

3990 Views | 1749 Downloads | 9 Citations
 PDF Download Citation Citation

Altmetric Attention Score

Find out more
 Sign up for content alerts