Artificial vision system with tunable photoconductivity based on organic molecule nanowire heterojunctions

Abstract

The application provides an artificial optical synapse device, which comprises an organic molecular nanowire heterojunction. The artificial vision system based on the organic molecular nanowire heterojunction provided by the embodiment of the application can imitate the visual function of human beings and has continuous negative photoconductivity or positive photoconductivity. Irradiation of different wavelengths (from solar blind to the visible light range) and power densities can stimulate the devices, and the synaptic behavior of the two different photoconductivities is effectively realized.

Description

Artificial vision system with tunable photoconductivity based on organic molecular nanowire heterostructures Technical Field

[0001] This invention relates to artificial vision systems, such as vision systems for intelligent sensory electronic devices. Background Technology

[0002] Since the emergence of artificial intelligence (AI), the development of information and communication technologies has greatly promoted the recent development of the Internet of Things (IoT), artificial vision systems, autonomous driving, and other technologies. [1-3] In fact, the human brain is one of the most complex structures in nature, capable of processing and storing massive amounts of data simultaneously with high speed and ultra-low power consumption. [4-5] Inspired by the human brain, devices and systems based on artificial synapses have made significant progress. Various intelligent sensory electronic devices, such as artificial throats, artificial skin, and artificial vision systems, have received particular attention. [7-10] It is understood that approximately 80% of information is acquired through human vision. Therefore, vision is the fundamental way to receive external signals. [11-12] (See Figure 6a). Figure 2a shows a schematic diagram of a biological synapse, in which presynaptic cells transmit information to postsynaptic cells via neurotransmitters. Furthermore, the human visual system can recognize complex images of different wavelengths and convert optical signals into biological signals for processing, which has spurred the development of artificial vision systems equipped with optoelectronic devices.

[13] (See Figure 6c). In this case, the artificial vision system could potentially act as a signal receiver for both a bionic processor and an artificial synapse.

[0003] As is well known, neuronal activity is a fundamental property of the visual system in the human brain and is related to neural plasticity.

[14] This neuroplasticity depends on the ability of neurons to process, maintain, and inherit neurons based on previous stimuli. Unlike electrical signals, optical signals in the human visual system have several advantages here, including wide bandwidth, low power consumption, and long-distance transmission.

[15] Sun et al. combined SnO2 nanowires with ionogels to fabricate electro-neural synapses that operate in the range of solar blindness.

[16] Subsequently, In-Ga-Zn-O materials were introduced together with ionogels to fabricate an artificial vision system with low retrace subthreshold swing (<60mV / 10).

[17] Besides introducing a dielectric layer, using heterojunctions is also an effective method to achieve synaptic function. Meng et al. introduced a 2D / 0D hybrid structure to fabricate artificial synapses, achieving sub-fejoule power consumption by simply integrating black phosphorus quantum dots (QDs) onto two-dimensional MoSSe (molybdenum sulfide crystal).

[18] Similarly, Huang et al. dispersed CsPbBr3QDs in C8-BTBT / PS organic thin films for synaptic transistors, exhibiting ultra-low power consumption in the ultraviolet band and effectively mimicking synaptic behavior.

[19] Furthermore, various artificial vision systems with negative photoconductivity based on two-dimensional (2D) materials and peroxides have been developed. [1,20,21] On the one hand, in order to achieve positive and negative photoconductivity at different wavelengths, heterojunctions of multiple layers of two-dimensional materials are usually used. [22,23] However, it is difficult to avoid using complex transfer processes to prepare these heterojunctions, and large-area heterojunction arrays are also difficult to achieve in practical applications. On the other hand, perovskite materials are suitable for large-area fabrication, but suffer from poor air stability. Due to the unique wide-range photoreactivity, excellent electrical properties, robust stability, and, more importantly, ease of fabrication into large-scale arrays, III-V nanowires (NW) are expected to have excellent application prospects in artificial synaptic devices. [24,25] Notably, the morphology and topological features of these one-dimensional (1D) nanowires resemble tubular axons, which are crucial for efficient connectivity in the nervous system. [26,27] In this regard, integrating III-V nanoarrays with heterojunctions of other material systems (such as organic semiconductors) to construct high-performance artificial vision systems is highly desirable. Summary of the Invention

[0004] Therefore, in one aspect, the present invention provides an artificial photosynaptic device comprising an organic molecule-nanowire heterojunction.

[0005] In some embodiments, the artificial photosynapse device further includes a first heterojunction formed by wrapping a first organic material film on a first nanowire device, or a second heterojunction formed by wrapping a second organic material film on a second nanowire device.

[0006] In some embodiments, the first nanowire device and the second nanowire device are printed nanowire arrays.

[0007] In some embodiments, the first nanowire device and the second nanowire device are III-V semiconductor nanowire arrays.

[0008] In some embodiments, the first nanowire device and the second nanowire device are InGaAs nanowire arrays.

[0009] In some embodiments, the first organic material film is different from the second organic material film, thereby enabling the artificial photosynaptic device to have adjustable photoconductivity.

[0010] In some embodiments, the first organic material membrane is C8-BTBT, and the second organic material membrane is PC. 61 BM.

[0011] In some embodiments, the InGaAs nanowire arrays are synthesized via a two-step catalytic solid-source chemical vapor deposition (CVD) process.

[0012] In some embodiments, sustained negative photoconductivity (NPC) or positive photoconductivity (PPC) is achieved in the first heterojunction and the second heterojunction.

[0013] In some embodiments, the III-V semiconductor nanowire array is p-type, n-type, or bipolar.

[0014] In some embodiments, the III-V semiconductor nanowire array is binary, ternary, or quaternary.

[0015] In some embodiments, the III-V semiconductor nanowire array is located above or below the first organic material film or the second organic material film.

[0016] In some embodiments, the dielectric layer of the artificial photosynapse device is a silicon dioxide layer thermally grown on a silicon wafer.

[0017] In another aspect of the invention, a method for manufacturing an artificial photosynaptic device is provided, comprising the steps of ultraviolet lithography, spin coating, and electron beam lithography.

[0018] The artificial vision system based on organic molecular nanowire heterojunctions provided by the embodiments of the present invention can mimic human visual function, exhibiting continuous negative or positive photoconductivity. Irradiation of different wavelengths (from the solar blind to the visible light range) and different power densities can stimulate these devices, effectively realizing synaptic behaviors with two different photoconductivities. Compared with devices that can achieve positive and negative photoconductivity at different wavelengths, complex transfer processes are avoided, and air stability is improved. Furthermore, large-area heterojunction arrays are realized through solution processing. In this regard, utilizing heterojunctions integrating III-V group nanowire arrays with organic semiconductors to construct high-performance artificial vision systems is highly desirable. Attached Figure Description

[0019] The above and further features of the present invention will become apparent from the following description of embodiments, which are provided as examples in conjunction with the accompanying drawings, wherein:

[0020] Figure 1a is a schematic diagram of a photosynaptic device according to an embodiment of the present invention.

[0021] Figure 1b is a high-resolution transmission electron microscope (HRTEM) image of a typical InGaAs NW, with the inset showing the NW along its length. <111> The lattice spacing in the direction.

[0022] Figure 1c is a scanning electron microscope (SEM) image of a printed InGaAs NW array device according to an embodiment of the present invention.

[0023] Figure 1d illustrates the transmission curves of the obtained InGaAs NWs and artificial photosynaptic devices, with and without C8-BTBT, respectively.

[0024] Figure 1e shows the laser with and without C8-BTBT (1550 nm laser, power density 3.4 mW / mm²). 2 The optical response curve of the device.

[0025] Figure 1f shows the optical response curve of the printed InGaAs NW array device (450nm laser, with a power density of 0.64mW / mm²). 2 ).

[0026] Figure 2a is a schematic diagram of a biological synapse.

[0027] Figure 2b shows the photosynaptic transfer curves of an artificial photosynaptic device according to an embodiment of the present invention, which comprises a printed InGaAs NW array wrapped with a C8-BTBT thin film and subjected to 405 nm light irradiation at different power densities.

[0028] Figure 2c shows the artificial photosynapse device at a wavelength of 450 nm and a power density of 0.16 mW / mm². 2 The behavior of EPSC (excitatory postsynaptic current) under laser light.

[0029] Figure 2d shows the artificial photosynapse device at a wavelength of 405 nm and a power density of 0.16 mW / mm². 2 EPSC behavior under laser.

[0030] Figure 2e shows the artificial photosynapse device at a wavelength of 405 nm and a power density of 0.26 mW / mm². 2 EPSC behavior under laser.

[0031] Figure 2f shows the EPSC behavior of the artificial photosynapse device under different numbers of 405nm light pulses and at a fixed frequency (1.5Hz).

[0032] Figure 2g shows the Paired Pulse Promotion (PPF) index of the photosynaptic device under two consecutive pulses with an interval of 450 nm. The inset in this figure shows the curve corresponding to the highest PPF index value.

[0033] Figure 2h shows the PPF index of the photosynaptic device under two consecutive pulses with an interval of 405 nm. The inset in this figure shows the curve corresponding to the highest PPF index value.

[0034] Figure 3a shows the photoluminescence (PL) spectra of the C8-BTBT thin film and the InGaAs / C8-BTBT heterojunction.

[0035] Figure 3b shows the corresponding time-resolved phase distances (PLs) for the C8-BTBT thin film and the InGaAs / C8-BTBT heterojunction.

[0036] Figure 3c shows the X-ray photoelectron spectra (XPS) of the C8-BTBT thin film and the InGaAs / C8-BTBT heterostructure.

[0037] Figure 3d shows the EPSC behavior of the artificial photosynaptic device under single light pulses of different wavelengths.

[0038] Figure 3e shows the band structure of the InGaAs / C8-BTBT heterojunction before irradiation.

[0039] Figure 3f shows the band structure of the InGaAs / C8-BTBT heterojunction after irradiation.

[0040] Figure 3g illustrates the long-term duration enhancement (LTP) characteristics of the artificial photosynaptic device under light pulses of different wavelengths (1.5 Hz).

[0041] Figure 3h shows the LTP characteristics of the artificial photosynaptic device under 450 nm light pulses of different power densities.

[0042] Figure 3i shows the LTP characteristics of the artificial photosynaptic device under 405 nm light pulses of different power densities.

[0043] Figure 4a shows the LTP characteristics of the artificial photosynapse device under 405nm light pulses at different power densities (1.5Hz). The artificial photosynapse is configured with a printed InGaAs NW array and is mounted on a PC. 61 BM film wrapping.

[0044] Figure 4b illustrates the EPSC behavior of the synaptic device under 405nm light pulses of different pulse widths.

[0045] Figure 4c shows the EPSC behavior of the synaptic device under different numbers of 405nm light pulses and at a fixed frequency (1.5Hz).

[0046] Figure 4d shows the LTP characteristics of the artificial photosynaptic device under 450nm light pulses at different power densities (1.5Hz).

[0047] Figure 4e illustrates the EPSC behavior of the synaptic device under 450nm light pulses of different pulse widths.

[0048] Figure 4f shows the EPSC behavior of the synaptic device under different numbers of 450nm light pulses and at a fixed frequency (1.5Hz).

[0049] Figure 5a shows the synergistic optoelectronic modulation for long-term memory using InGaAs / C8-BTBT heterojunction devices, where the optical memory function is realized using an array of artificial photosynaptic devices.

[0050] Figure 5b shows the corresponding relaxation time fitted using the Kohlrausch stretching exponential function.

[0051] Figure 5c shows different memory states configured using the NPC phenomenon.

[0052] Figure 5d shows an optical image of a 4×4 array based on InGaAs / C8-BTBT heterojunction devices, in which a human hair is used to block laser irradiation.

[0053] Figure 5e shows the EPSC in the initial state under a single 405nm pulse irradiation.

[0054] Figure 5f shows the percentage change in EPSC after 10 seconds of exposure to a single 405nm pulse.

[0055] Figure 5g shows the percentage change in EPSC after 30 seconds of irradiation with a single 405nm pulse.

[0056] Figure 5h shows the percentage change in EPSC after 60 seconds of irradiation with a single 405nm pulse.

[0057] Figure 5i shows the percentage change in EPSC after 120 seconds of exposure to a single 405nm pulse.

[0058] Figure 6a is a schematic diagram of the human visual system.

[0059] Figure 6b is a demonstration of a kernel array network that simulates the visual processing and recognition functions of the human eye in a hardware kernel based on NPC and PPC devices.

[0060] Figure 6c shows the capabilities of the hardware kernel-based vision processing.

[0061] Figure 6d shows the experimental and simulation results of hardware kernel operation.

[0062] Figure 6e shows a schematic diagram of a neural network used for classification and recognition.

[0063] Figure 6f shows the recognition rate of the neural network evaluated with and without a hardware kernel.

[0064] Figure 7 is a schematic diagram of the manufacturing steps of an optical synaptic device array according to an embodiment of the present invention.

[0065] Figure 8a shows a schematic diagram of the CVD apparatus used for NW synthesis.

[0066] Figure 8b shows a SEM image of the grown InGaAs NWs, with a scale bar of 5 μm.

[0067] Figure 8c shows a SEM image of the grown InGaAs NWs, with a scale bar of 1 μm.

[0068] Figure 9a shows a transmission electron microscope (TEM) image of NW.

[0069] Figure 9b shows the energy-dispersive X-ray spectra (EDS) of the corresponding InGaAs NWs.

[0070] Figures 10a-10d show the elemental mappings of In, As, and Ga in representative InGaAs NWs, respectively.

[0071] Figure 11a shows the molecular structure of C8-BTBT.

[0072] Figure 11b shows the output curve of the device using C8-BTBT / InGaAs.

[0073] Figure 11c illustrates the NPC phenomenon in devices with C8-BTBT / InGaAs.

[0074] Figure 12a shows a SEM image (left) of InGaAs NWs coated with a C8-BTBT film and EDS (right) of C and S elements as components of C8-BTBT.

[0075] Figure 12b is an atomic force microscopy (AFM) image of C8-BTBT thin film covering InGaAs NWs.

[0076] Figures 13a-13c are EDS diagrams of devices with C8-BTBT / InGaAs.

[0077] Figure 13d is a SEM image of the corresponding device.

[0078] Figure 14 is an optical image of a device with C8-BTBT / InGaAs.

[0079] Figure 15a illustrates the photoreaction of printed InGaAs NWs under a 405 nm laser.

[0080] Figures 15b and 15c illustrate EPSCs of photosynaptic devices using 450nm light at different power densities.

[0081] Figure 16a shows the peak interval (ΔTime = 3 seconds) and 0.32 mW / mm 2 PPF behavior of C8-BTBT / InGaAs synaptic devices at optical power densities.

[0082] Figure 16b illustrates the effect at a fixed peak interval (ΔTime = 3 seconds) and 0.96 mW / mm 2 PPF behavior of C8-BTBT / InGaAs synaptic devices at optical power densities.

[0083] Figure 16c illustrates the effect at a fixed peak interval (ΔTime = 3 seconds) and 1.3 mW / mm 2 PPF behavior of C8-BTBT / InGaAs synaptic devices at optical power densities.

[0084] Figure 16d shows the peak interval (ΔTime = 3 seconds) and 1.8 mW / mm 2 PPF behavior of C8-BTBT / InGaAs synaptic devices at optical power densities.

[0085] Figure 16e shows the peak interval (ΔTime = 3 seconds) and 2.1 mW / mm 2 PPF behavior of C8-BTBT / InGaAs synaptic devices at optical power densities.

[0086] Figure 16f illustrates the effect at a fixed peak interval (ΔTime = 3 seconds) and 2.4 mW / mm 2 PPF behavior of C8-BTBT / InGaAs synaptic devices at optical power densities.

[0087] Figures 17a-17f show the results at a fixed power density (and a wavelength of 450 nm, with a power density of 0.96 mW / mm). 2 PPF behavior of C8-BTBT / InGaAs synaptic devices under different spike intervals (Δ time).

[0088] Figures 18a-18g show the results at a fixed power density (and a wavelength of 405 nm, with a power density of 0.17 mW / mm). 2PPF behavior of C8-BTBT / InGaAs synaptic devices under different spike intervals (Δ time).

[0089] Figure 19a shows the band gap of the PL spectrum of C8-BTBT.

[0090] Figure 19b shows the absorption characteristics of the C8-BTBT thin film.

[0091] Figure 20a shows the absorption of InGaAs NWs.

[0092] Figure 20b shows the corresponding (αhν) of InGaAs NWs. 2 -hν curve.

[0093] Figure 21 shows the ultraviolet photoelectron spectroscopy (UPS) of InGaAs NWs.

[0094] Figure 22a shows the transmission curves of the C8-BTBT / InGaAs device after illumination with 450nm and 1550nm light.

[0095] Figure 22b shows the IT curve of the C8-BTBT / InGaAs device when two different wavelengths of light are applied simultaneously after an infrared pulse.

[0096] Figure 23a shows the LTP characteristics of the C8-BTBT / GaSb-based artificial photosynapse device under a 405nm light pulse.

[0097] Figure 23b shows the LTP characteristics of the C8-BTBT / GaSb-based artificial photosynaptic device under a 450nm light pulse.

[0098] Figures 24a-24c show the PC 61 EDS plot of BM / InGaAs devices.

[0099] Figure 24d is a SEM image of the corresponding device.

[0100] Figure 25a shows PC 61 The molecular structure of BM.

[0101] Figure 25b shows a PC. 61 Optical images of BM / InGaAs devices.

[0102] Figure 25c shows the PC 61 AFM image of InGaAs NWs covered by BM thin film.

[0103] Figure 26a shows PC 61 Absorption of BM thin film.

[0104] Figure 26b shows the PC 61 Band structure of BM / InGaAs.

[0105] Figure 27a shows the PC 61 PL of BM / InGaAs devices.

[0106] Figure 27b shows the PC. 61 Time-resolved pulse (PL) of BM / InGaAs devices.

[0107] Figures 28a-28h show the PC at a fixed power density (wavelength 405 nm) and different Δ times. 61 PPF behavior of BM / InGaAs synaptic devices.

[0108] Figures 29a-29h show the PC at a fixed power density (wavelength 450 nm) and different Δ times. 61 PPF behavior of BM / InGaAs synaptic devices.

[0109] Figure 30a shows the LTP characteristics of the artificial photosynaptic device under a 261 nm light pulse.

[0110] Figure 30b shows the LTP characteristics of the artificial photosynaptic device under a 532 nm light pulse.

[0111] Figure 30c shows the EPSC behavior of the synaptic device under different numbers of 532nm light pulses and at a fixed frequency (1.5Hz).

[0112] Figure 31a shows the air stability of the photosynaptic device based on C8-BTBT / InGaAs NWs.

[0113] Figure 31b shows a PC-based... 61 Air stability of BM / InGaAs NWs photosynaptic devices.

[0114] Figure 32a shows the application of 261 nm laser stimulation to C8-BTBT / InGaAs and PC, respectively. 61 BM / InGaAs photosynaptic devices to achieve long-term memory functionality.

[0115] Figure 32b shows the application of 532nm laser stimulation to C8-BTBT / InGaAs and PC, respectively. 61 BM / InGaAs photosynaptic devices to achieve long-term memory functionality.

[0116] Figures 33a-33c show the photoresponse of NPC devices with different optical power densities and wavelengths.

[0117] Figures 34a-34c show the photoresponse of PPC devices with different optical power densities and wavelengths.

[0118] Figure 35 shows the experimental details of visual processing. Detailed Implementation

[0119] The visual system is one of the most critical units in the human sensory system, combining multi-wavelength signal detection and data processing capabilities. Embodiments of the present invention, described below, provide a large-scale array of artificial synaptic devices based on organic molecular nanowire heterojunctions, exhibiting tunable photoconductivity.

[0120] Referring to FIG1a, a first embodiment of the present invention is an artificial photosynaptic device 20 comprising an organic molecular nanowire heterojunction. The device 20 is configured with a printed InGaAs NW array 22 and a top-coated organic semiconductor thin film 24. The device 20 also includes source / drain (S / D) metal contact electrodes 30 made of nickel (Ni). The InGaAs NW array 22 is parallel to each other and is fabricated on top of a SiO2 layer 26, wherein a Si layer 28 serves as the substrate and gate electrode. In a specific fabrication process, the Si layer 28 is a highly boron-doped p-type silicon wafer. The SiO2 layer 26 is a 50 nm thick thermally grown gate oxide, and the metal contact electrode 30 is thermally deposited on the SiO2 layer 26 with a thickness of 50 nm.

[0121] As understood by those skilled in the art, InGaAs is an alloy of gallium arsenide and indium arsenide. More generally, it belongs to the InGaAsP quaternary system, which consists of alloys of indium arsenide (InAs), gallium arsenide (GaAs), indium phosphide (InP), and gallium phosphide (GaP). Since gallium and indium belong to Group III of the periodic table, while arsenic and phosphorus belong to Group V, these binary materials and their alloys are all group III-V compound semiconductors.

[0122] Organic semiconductor thin film 24 is p-type 2,7-dioctyl[1]benzothiophene[3,2-b][1]benzothiophene (C8-BTBT) or n-type phenyl-C 61 methyl butyrate (PC) 61 Organic thin films (BM) are used to encapsulate InGaAs NW arrays 22 to configure two different type I heterojunctions (i.e., pn C8-BTBT / InGaAs and nn PC). 61 BM / InGaAs). These two type I heterojunctions (i.e., InGaAs / C8-BTBT and InGaAs / PC) 61 Based on their different photoconductivities, heterojunctions (BMs) can effectively simulate basic synaptic visual behaviors such as PPF and EPSC. Due to the different carrier injections, these heterojunctions can achieve sustained NPC or PPC.

[0123] Now let's explore the working principle of the artificial photosynaptic device 20. Irradiation with different wavelengths (from the solar blind to the visible light range) and power densities can stimulate the photosynaptic device 20, effectively mimicking the synaptic behaviors of two different photoconductivities. Long-term memory (LTM) and polymorphic optical memory are also achieved through synergistic photoelectric modulation. Notably, the InGaAs NW array 22, along with NPC and PPC phenomena, is employed to construct the hardware core of the vision system. Due to its tunable photoconductivity and response to multiple wavelengths, the neural network employing the hardware core achieves 100% recognition rate with low complexity and power consumption. As will be demonstrated later, the artificial photosynaptic device 20 exhibits retinal-like behavior and the ability to be integrated over large areas, revealing its enormous potential in artificial vision systems.

[0124] Due to two different organic materials (i.e., p-type C8-BTBT and n-type PC) 61 The primary charge carriers in C8-BTBT differ from those in BM, allowing for tunable photoconductivity under irradiation from ultraviolet to visible light. Studies have shown that under light stimulation, the hole carriers or PCs in C8-BTBT... 61 Electron carriers from the BM are injected into InGaAs NWs, thereby inducing either NPC or PPC accordingly. Due to the high surface-to-volume ratio of InGaAs NWs, these two distinct heterojunctions can readily achieve well-defined synaptic functions. [28,29] Notably, synaptic behaviors such as short-term plasticity (STP), long-term synaptic time (LTM), excisional postsynaptic current (EPSC), inhibitory postsynaptic current (IPSC), and long-term potential / inhibition (LTP / LTD) can be reliably visualized using visible and ultraviolet lasers, including in the solar-blind range. [30,31] Compared to single NW-based devices, the integration of printed NW arrays and solution-processed organic materials demonstrates their significant potential as active materials for large-area neural visual networks. A prototype of a 4×4 artificial photosynaptic device array was tested with a gate bias voltage (V0). G It was proven to have optical memory function under certain conditions. Simultaneously, two photosynthetic devices with tunable photoconductivity (i.e., InGaAs / C8-BTBT for NPC and InGaAs / PC for PPC) were employed. 61Two photosynaptic devices (BM) were constructed to simulate the visual processing and recognition of the human eye. The structures of these two devices are similar to the artificial photosynaptic device 20 shown in Figure 1a. The different hardware kernels can detect various features in the image, such as those perceived by the human retina. These features are then fed into a neural network to classify letters of six different colors. Impressively, the classification accuracy reached 100%, compared to only 51% without the hardware kernel. All these results clearly demonstrate the great potential of these photosynaptic devices based on organic molecular nanowire heterostructures for practical applications in artificial vision systems.

[0125] Referring to Figures 7 and 8a, according to one embodiment of the present invention, a method for manufacturing InGaAs NWs and artificial photosynaptic devices is provided. A schematic diagram of a CVD setup for NW synthesis is shown in Figure 8a. Furthermore, Figure 7 illustrates a process flow for manufacturing InGaAs NWs and then the artificial photosynaptic device. As described above, Figure 1a depicts a schematic diagram of an artificial photosynaptic device 20 configured with a printed InGaAs NW array and a top-coated organic semiconductor thin film; the artificial photosynaptic device 20 of Figure 1a is used as an example in Figure 7 to illustrate the process flow. It should be emphasized that the artificial photosynaptic device 20 of Figure 1a can also be manufactured using any other suitable method, without being limited to the methods shown in Figures 7 and 8a.

[0126] In this method, InGaAs nanowires are obtained through existing technology. [28,32]The InGaAs nanowires were synthesized using a two-step catalytic solid-source CVD method proposed in [the original text]. Specifically, as shown in Figures 7 and 8a, the InGaAs nanowires were synthesized on SiO2 / Si wafers 26 and 28 (including a 50 nm thick thermally grown oxide layer, i.e., the aforementioned SiO2 layer 26) using a chemical vapor transport method in a dual-zone horizontal tube furnace 42 (see Figure 7). A mixed InAs and GaAs powder 44 (weight ratio 1:9) was loaded into a boron nitride crucible 46, which was placed in the upstream region 48a of the furnace 42. A growth substrate pre-deposited with a 0.5 nm thick (nominal thickness) gold film (not shown) as a catalyst was placed in the downstream region 48b of the furnace 42. The temperature of the downstream region 48b was first raised to 800 °C and held for 10 minutes to anneal the gold catalyst. Then, the temperature of the downstream region 48b was directly cooled to the growth temperature (660 °C) for the first step of growth, while the source temperature began to rise. The first nucleation step begins when the source temperature reaches the specified value (820°C). After 1-2 minutes, downstream heating is stopped, and the temperature is then cooled to the second-step growth temperature (580°C). Finally, the second-step growth lasts for 40 minutes. Throughout the growth process, hydrogen (99.9995%) is used as the carrier gas, and the flow rate is maintained at 100 sccm. Arrow 50 in Figures 7 and 8a shows the flow direction of hydrogen as the carrier gas in furnace 42.

[0127] As shown in Figure 7, after the InGaAs NWs are synthesized, they proceed to the fabrication of photosynthetic devices. A mature contact printing technique is used to fabricate a parallel array 22 of NWs on highly boron-doped p-type Si substrates 26 and 28, on which a 50nm thick thermally grown gate oxide layer, i.e., a SiO2 layer 26, is placed. [67,68] Photolithography was used to define the source and drain regions, while a 50nm thick Ni layer was thermally deposited as the top electrode 30, followed by removal of excess metal. C8-BTBT and PC 61 BM was dissolved in chlorobenzene at concentrations of 20 mg / ml and 25 mg / ml, respectively. Top-coated C8-BTBT films (not shown in Figure 7) were fabricated using spin coating at 2000 rpm for 30 seconds. After spin coating, the films were annealed at 100°C for 5 minutes. PC 61 The BM organic film (not shown in Figure 7) was prepared using the same spin-coating method at 1800 revolutions per 30 seconds, followed by annealing at 60°C for 5 minutes. At this point, the artificial photosynaptic device 20 described in Figure 1 has been successfully manufactured.

[0128] In the experimental setup described above, the electrical performance of the FET was then characterized using a standard electrical probe station and an Agilent 4155C semiconductor analyzer (supplied by Agilent Technologies, Santa Clara, California, USA). Scanning electron microscopy (SEM, Quanta) was also employed. TM450FEG, FEI and HRTEM, Thermo Scientific TM Talos TM The morphology and EDS images of the prepared NWs were examined using an F200X. The morphology of the organic thin films was determined using an atomic force microscope (Bruker). Icon TM The evaluation was performed using AFM. The band structure of the nanowires was investigated using a UPS (ThermoFisher, ESCALAB 250Xi). The charge transfer process was characterized using a PL and a time-resolved PL (FLS980).

[0129] Next, the performance of the artificial photosynaptic device 20 fabricated using the above method will be discussed, and its experimental data will be presented. First, the thickness, length, density, and surface condition of the InGaAs NWs obtained via a two-step CVD method will be evaluated using SEM (see Figures 8b and 8c). Based on the TEM images, spherical catalytic seeds can be clearly observed at the tips of the NWs (Figure 9a), an observation consistent with the generally known vapor-liquid-solid (VLS) growth mechanism of InGaAs NWs. [33,34] HRTEM revealed distinct lattice fringes, with a spacing of 0.34 nm between adjacent lattice planes, corresponding to the InGaAs NWs. <111> The dominant direction (Figure 1b). Then, EDS was applied to evaluate the composition of InGaAs NWs, confirming the dominant direction of In. x Ga 1-x The composition of As was determined, with x approximately 0.49 (Fig. 9b). Corresponding elemental dispersion images were also obtained, demonstrating the uniform distribution of In, As, and Ga along the NW volume (see Figs. 10a-10d). Following material characterization, the various steps of the aforementioned fabrication method were applied. In particular, a large-scale InGaAs NW array was fabricated using an NW printing process.

[29] Then, photolithography was used to define a channel region with a length of 2 μm, and nickel was used as the source / drain (S / D) metal contact electrode, as shown in Figure 1c. C8-BTBT is an excellent p-type organic semiconductor widely used in high-performance field-effect transistors, artificial synapses, and optotransistors.

[35] C8-BTBT has a relatively wide band gap (~3.8 eV), exhibiting good ultraviolet response characteristics and potential applications with continuous photoconductivity. [36-38]The molecular structure of C8-BTBT is shown in Figure 11a. An artificial photosynaptic device was constructed by spin-coating the top of a C8-BTBT thin film onto an InGaAs NW array. Due to the band structure of the type I pn heterojunction, the conduction band (CB) and valence band (VB) of InGaAs are completely covered by the LOMO (lowest unoccupied molecular orbital) and HOMO (highest occupied molecular orbital) levels of C8-BTBT. Figure 1d shows the transport curves of devices containing and without C8-BTBT. Source / drain current (I0) ds The reduction in voltage (from 51 μA to 38 μA) and threshold voltage (V) th The shift in voltage (from -3.5V to -2V) indicates that holes injected from the C8-BTBT, acting as the dominant charge carriers, will recombine with electrons in the InGaAs NWs. Figure 11b shows the output curve of the InGaAs / C8-BTBT device, illustrating the effective V... G Modulation and ohmic contact. The photoresponse of devices with and without C8-BTBT is shown in Figure 1e, which again shows that the photocurrent of InGaAs / C8-BTBT devices is reduced, consistent with the results of the transmission curve. Typically, InGaAs NWs exhibit strong light absorption in the infrared band, but C8-BTBT films show no photoresponse in the same band. Therefore, due to carrier recombination within InGaAs NWs, both the dark current and photocurrent of devices with organic molecular nanowire heterojunctions decrease. Notably, many intriguing optoelectronic properties of one-dimensional nanowires have been observed, among which NPC is one of the most widely studied phenomena. [39,40] Due to the narrow band gap of InGaAs, when high-energy, short-wavelength lasers are applied, induced photothermal electrons are trapped in the surface oxide, which acts as scattering centers for charge carriers. Therefore, depending on the thickness and density of the surface oxide layer, InGaAs NWs devices will randomly exhibit two different photoconductivities. [41,42] Figure 1f shows the positive photoconductivity of the printed InGaAs NW array at 450 nm. However, after integration with a p-type organic semiconductor of C8-BTBT to construct an I-type pn heterojunction, the photoconductivity of the device is unified to exhibit consistent NPC characteristics. Figure 11c illustrates the NPC of the device with an organic molecular nanowire heterojunction, and the corresponding NPC mechanism will be discussed later.

[0130] In biological synapses, neurotransmitters are released when the presynaptic terminal receives an external signal. The electrical potential at the postsynaptic terminal changes according to the stimulation of the neurotransmitter.

[43] Here, the stable and controllable persistent NPC phenomenon based on the type I pn organic molecular nanowire heterostructure demonstrates its application potential in artificial vision systems. Specifically, Figure 2a shows a schematic diagram of the biological synapse. The morphology of the C8-BTBT film covering the nanowires is described using SEM and AFM (see Figures 12a-12b). In the InGaAs / C8-BTBT overlap region, S and C elements are uniformly distributed, and the surface roughness is small, indicating that the deposited organic film has high density and smoothness. In addition, Figures 13a-13d and 14 show the SEM and optical images of the photosynaptic device, respectively, further demonstrating the good quality of the obtained C8-BTBT film. The transmission curves of the device with NPC characteristics under 405 nm laser at different power densities are shown in Figure 2b. It is clear that the decrease in carrier concentration is amplified with the increase of irradiation intensity. Figures 2c-e show the peak intensity-dependent plasticity (SIDP) and peak time interval-dependent plasticity of the photosynaptic device under 450 nm and 405 nm lasers, respectively. It can be seen that the IPSC value is highly dependent on the wavelength, intensity, and duration of laser irradiation. The two basic synaptic functions are divided into short-term and long-term memory, demonstrating the ability to inherit and retain synaptic weights.

[44] Whether increasing exposure time or laser intensity, the photosynaptic device exhibits a transition from short-term memory to long-term memory (S / LTM). Figures 15a-15c show the photoresponse of the printed InGaAs NW array under a 405 nm laser, and the IPSC behavior of the photosynaptic InGaAs / C8-BTBT device under different power densities under a 450 nm laser. Furthermore, different numbers of 405 nm laser pulses were applied to the photosynaptic device to demonstrate its ability to inherit neural weights (Figure 2f). Undoubtedly, as the number of pulses increases, the LTM phenomenon gradually becomes dominant, and the photosynaptic device exhibits spike number-dependent plasticity (SNDP). Next, PPF, one of the fundamental synaptic behaviors for decoding temporal information in neurobiology, is simulated here using two different laser wavelengths.

[14] In this scenario, two consecutive presynaptic lights with different time intervals (ΔTime) are used to simulate the PPF behavior in the photosynaptic device. Furthermore, the PPF index can be defined by the ratio (A2 / A1) of two distinct EPSC peaks.

[45] For an irradiation time of 0.5 seconds and a light power density of 0.96 mW / mm², -2For a 450 nm light pulse (see Figure 2g), the maximum PPF exponent value of 174% was found when the Δ time was 4.5 seconds. Figures 16a-16f show the PPF under 450 nm light with different power densities and a fixed Δ time (3 seconds). Subsequently, similar to the 450 nm light, the PPF behavior was successfully mimicked by a 405 nm light pulse (see Figure 2h). The maximum PPF exponent value of 178% was obtained at ΔTime = 1.67 seconds. The pulse interval-dependent decay of the PPF exponent under two different light pulses can be represented using a double exponential decay function, as shown in the inset of Figure 2h.

[46] The y0 parameter is the PPF exponent when the pulse interval is close to infinity. The t1 and t2 parameters are the decay constants for the fast and slow phases, respectively. The A1 and A2 parameters are the initial boost amplitudes for the corresponding phases. The t1 and t2 extracted from the pulse equations for 450 nm and 405 nm light are 1.96 s and 2.39 s (see Fig. 2g), and 0.67 s and 0.83 s, respectively (see Fig. 2h). These results demonstrate that the organic molecular nanowire photosynaptic device can successfully mimic the synaptic behavior of PPF. Furthermore, data on the PPF exponents for 405 nm and 450 nm light are shown in Figs. 17a-17f and Figs. 18a-18g.

[0131] Stable and persistent photogenerated carrier transfer (NPC) phenomena were achieved through an InGaAs / C8-BTBT heterojunction. Here, the synaptic behavior of photopulses can also be mimicked by photosynaptic devices. To further investigate the NPC mechanism induced by the organic molecular nanowire heterojunction, photoluminescence (PL) spectroscopy was used to evaluate carrier transfer in the heterojunction (see Figure 3a). C8-BTBT films were fabricated on a printed nanowire array using the same process via spin coating. The PL spectra of the C8-BTBT films and the InGaAs / C8-BTBT heterojunction were measured at the same power under 320 nm laser excitation. Clearly, the emission peak intensity at approximately 372 nm decreased by a factor of 6 compared to the former, which is similar to heterostructures exhibiting photogenerated carrier transfer processes.

[47] After analyzing the absorption characteristics of the C8-BTBT thin film (see Figures 19a-19b), the wide bandgap of the C8-BTBT thin film was confirmed. Then, the time-resolved pulse distance (PL) was evaluated using data fitted with a double exponential function, revealing a relatively short lifetime for the InGaAs / C8-BTBT heterojunction (see Figure 3b).

[48] For the C8-BTBT thin film, the lifetimes of the fast and slow components were 2.65 ns and 27.05 ns, respectively, while the lifetimes of the organic molecular nanowire heterostructure decreased accordingly to 1.35 ns and 20.19 ns. This significant decrease in lifetime is attributed to the transfer of photogenerated carriers from C8-BTBT to InGaAsNWs. XPS was then used to investigate the carrier transfer characteristics of the InGaAs / C8-BTBT heterostructure (see Figure 3c). Compared to the original C8-BTBT thin film, the S2p peak of the heterostructure shifted slightly to a lower binding energy by 0.2 eV, confirming the presence of charge carrier transfer at the organic molecular nanowire interface.

[49] Figure 3d shows the IPSC behavior under light pulses of different wavelengths. Interestingly, even at relatively weak light power (i.e., 261 nm), short-wavelength light pulses can produce a strong IPSC effect. To explain the charge carrier transfer and NPC phenomenon, the banded structure of the organic molecular nanowire heterostructure before and after the application of light pulses is summarized in Figures 3e and 3f. In this work, UPS and absorption spectroscopy were used to evaluate In 0.48 Ga 0.52 The band structure of AsNWs is shown in Figures 20a-20b and 21. The band gap and Fermi level extracted from Figures 20a-20b and 21 are 0.84 eV and -4.17 eV, respectively. Thus, when In 0.48 Ga 0.52 When AsNWs and C8-BTBTs are in contact, a type I heterojunction can be confirmed. Due to the difference in Fermi level, an intrinsic electric field is generated at the InGaAs-C8-BTBT interface, resulting in band bending without the need for an optical pulse (see Figure 3e). This prediction is consistent with the XPS results. [49,50] Furthermore, the primary charge carrier in p-type C8-BTBT films is holes, which will transfer to InGaAs NWs to suppress current, as shown in Figures 1d and 1e. Moreover, due to the relatively large gap and energy barrier between the LUMO level (-1.8 eV) of C8-BTBT and the CB of InGaAs, electron transfer from C8-BTBT to InGaAs NWs is more difficult than hole transfer.

[51] Meanwhile, C8-BTBT exhibits strong absorption in the ultraviolet range (see Figures 19a-19b). When shorter wavelength light pulses irradiate the heterojunction, due to the built-in electric field at the InGaAs-C8-BTBT interface, more photogenerated holes in C8-BTBT are injected into the NWs (see Figure 3f). The injected holes will recombine with electrons in the NWs, inducing the NPC phenomenon. [22,52]Since InGaAs NWs exhibit photoreactivity in the infrared range, carrier recombination in the heterojunction was investigated using 1550 nm light after a visible light pulse. Figure 22a shows the transfer curves after irradiation with 450 nm and 1550 nm light. At 450 nm light (10 s, 2.2 mW / mm²), the carrier recombination was... -2 After irradiation, the decrease in current indicates the recombination of electrons into the NWs. Subsequently, infrared irradiation induces photocarriers in the NWs without affecting C8-BTBT, restoring the transport curve to its initial state. In Figure 22b, under an infrared pulse (2.7 mW / mm²), -2 After 5 seconds, two different wavelengths of light simultaneously act on the heterojunction. Because infrared excitation and carrier recombination occur simultaneously, different photocurrents can be controlled by the power density of the 405 nm light. As mentioned above, the persistent IPSC of the artificial photosynaptic device is highly dependent on the light power and wavelength, exhibiting a function similar to that of the human eye. Figures 3g-3i show the LTP characteristics induced by persistent IPSC in the artificial photosynaptic organic molecular nanowire heterojunction device. Different wavelengths of light pulses are used to induce different degrees of LTP characteristics (see Figure 3g). Notably, strong LTP characteristics can be achieved even under very weak light, such as solar blind light (261 nm). The pulse power in the ultraviolet and visible light ranges can also modulate the weighted changes of LTP characteristics (see Figures 3h and 3i). Therefore, the number of photogenerated holes in C8-BTBT varies with laser power, wavelength range, and pulse number. When a weak light pulse with a short irradiation time is used, only a small number of photogenerated holes can enter the nanowire, which will result in the STM characteristics. As laser intensity and duration increase, more photogenerated holes enter the nanowire, leading to increased recombination and inducing LTM (Light Transformation Mechanism) properties. Furthermore, to further verify the mechanism of the NPC (Potentially Differentially Interfering) phenomenon, gallium antimonide (GaSb) was used as the p-type nanowire to fabricate a p-type heterojunction with C8-BTBT. The printed GaSb array and C8-BTBT thin film were also fabricated using the same process. Interestingly, this p-type heterojunction did not exhibit the NPC phenomenon. Instead, as shown in Figures 23a-23b, strong forward photoconductivity was achieved under laser pulses at 405 nm and 450 nm. This result indicates that injecting the same type of charge carriers does not induce NPC.

[0132] In addition to injecting heterocarriers into the InGaAs / C8-BTBT heterojunction to achieve sustained NPC phenomenon, another popular organic material widely used in solar cells, PC, was also introduced. 61 BM, to construct type I inInGaAs / PC 61 BM heterojunction [53,54]Figures 24a-24d show SEM images and EDS analysis of the fabricated devices. Figures 25a-25c summarize the PC covering the printed InGaAs NW array. 61 Molecular structure, optical images, and AFM analysis of BM. It is clear that the deposited PC... 61 The BM film is uniform and dense, without any obvious pores, and its constituent elements are evenly distributed. This observation demonstrates the high quality of the deposited film, which is crucial for constructing high-quality heterostructures. Furthermore, InGaAs / PC... 61 The BM heterojunction can successfully realize the PPC phenomenon, demonstrating the potential for establishing an artificial vision system with positive photocurrent. Figures 4a and 4b show the LTP characteristics and EPSC behavior under 405nm light pulse illumination. Artificial photosynaptic devices based on these BM heterojunctions exhibit good weight inheritance and LTP characteristic preservation behavior under continuous illumination of 405nm and 450nm light at different power densities (see Figures 4a and 4d). Figures 4b and 4e investigate the 405nm nanometer (0.13mW / mm²) light pulse. -2 ) and 450nm (0.14mW / mm -2 The EPSC behavior under different durations of illumination with 405nm (0.13mW / mm) light was studied. -2 ) and 450nm (0.14mW / mm -2 The SNDP characteristics of the laser with 3, 5, and 10 pulses are shown in Figures 4c and 4f. The transition from STM to LTM can be achieved simply by adjusting the irradiation time and pulse number. Similar to the InGaAs / C8-BTBT heterojunction, using nInGaAs / PC 61 Artificial photosynthesis in BM heterojunctions can also successfully simulate synaptic behavior. Furthermore, EPSC behavior is highly sensitive to irradiation wavelength, power density, and pulse number. Shorter wavelength laser pulses more readily induce higher EPSC, showing potential for visual and color recognition. Figures 26a-26b illustrate PC... 61 Absorption spectrum of BM and InGaAs / PC 61 The band structure of the BM heterojunction illustrates the different mechanisms of PPC here. [55,56] To be specific, under laser pulses, PC 61 Photogenerated electrons in the BM will first be injected into the InGaAs NWs. Then, the type I heterojunction will prevent the electrons from flowing back. Furthermore, due to PC... 61 There is a significant difference between the HOMO level of BM (-6.1 eV) and the VB of InGaAs, and holes originate from PC. 61Transferring photogenerated electrons (BMs) to InGaAs NWs is difficult. Therefore, more photogenerated electrons than corresponding holes are injected into the NWs, inducing the PPC phenomenon. InGaAs / PC experiments were conducted. 61 The photoluminescence (PL) and time-resolved photoluminescence (PL) spectra of the BM heterojunction were used to evaluate and confirm charge carrier transfer under laser irradiation (see Figures 27a-27b). The photopolymer-peptide (PPF) effect, a fundamental synaptic characteristic, was also observed in the photosynaptic InGaAs / PC. 61 The BM device was successfully mimicked. A 405nm process was used with a fixed duration and power density (0.33 s, 0.14 mW / mm²). -2 ) and 450nm (0.33 s, 0.17 mW / mm -2 Two consecutive pulses of light were applied to the laser, and the responses are shown in Figures 28a-28h and 29a-29h, respectively. For the 405nm laser, when the Δ time is 1 second, the maximum PPF exponential value was found to be 130%, and the t1 and t2 values ​​extracted from the double exponential decay function in Figure 2h are 0.88 seconds and 3.11 seconds, respectively (see Figure 28a).

[46] Subsequently, similar to the 405nm laser, the 450nm laser pulse achieved a maximum PPF of 133% with a Δ time of 1 second. The t1 and t2 values ​​obtained from the same double exponential decay function were 0.39 seconds and 2.69 seconds, respectively (see Figure 29a). Furthermore, due to PC... 61 BM has a relatively wide absorption range, and lasers with wavelengths of 261nm and 532nm are used to induce PC. 61 The LTP characteristics of the BM / InGaAs photosynaptic device are shown in Figures 30a-30b. Furthermore, number-dependent spike plasticity (SNDP) was also achieved using different numbers of 532nm laser pulses, as shown in Figure 30c. In this case, the heterojunction device effectively mimics a range of synaptic behaviors, revealing its significant potential for constructing high-performance artificial vision systems composed of positive and negative photoconductors. To demonstrate its potential for practical applications, the air stability of two photosynaptic devices was investigated, as shown in Figures 31a-31b. These devices were placed in a dehumidifying chamber with a fixed humidity of 40% in the air environment for LTP characteristic testing. As shown in Figure 31a, for the C8-BTBT / InGaAs NWs-based device, the PSC decreased by 5.6% (3 days) and 18.5% (7 days) respectively under a 405nm laser pulse. Furthermore, Figure 31b shows the results with PC... 61 Air stability of BM / InGaAs NWs photosynaptic devices. Under 60 blue light pulses, the PSC decreased by 18.1% (3 days) and 33.3% (7 days) respectively.

[0133] According to some embodiments of the present invention, NPC and PPC phenomena can be observed by controllably injecting different photogenerated charge carriers into two different heterojunctions. For InGaAs / C8-BTBT heterojunction photosynthetic devices, holes injected from C8-BTBT under laser irradiation will lead to carrier recombination in NWs. Furthermore, VN can be introduced... G Bias voltage is used to study the memory characteristics of these artificial photosynaptic devices. When a large positive V... G When (40V) is applied to the gate, a large number of electrons accumulate in the NW channel. [21,52,57] Once the 405nm laser (1.1mW / mm) is applied... -2 When the device is irradiated (5 seconds), the PSC value will decrease as expected. However, since electrons have already accumulated in the channel before irradiation, the generation of a large number of additional electrons in the channel is not possible after recombination. Thus, the current of the synaptic device can remain at a low level for a relatively long time. Figure 5a shows the long-term memory (>10 4 Coordinated photoelectric modulation (seconds). The relaxation time fitted with the Kohlrausch stretching exponential function shows that the memory time here is also consistent with that in biological systems (Fig. 5b). [12,14,58] Different memory states were also tested and are shown in Figure 5c. Under synergistic photoelectric modulation, a weak 405nm laser (40μW / mm²) was used. -2 More than eight memory states can be realized in optosynthetic InGaAs / C8-BTBT devices, suggesting the potential for memory functionality exceeding 3 bits (8 distinguishable levels). [59,60] Furthermore, 261nm and 532nm laser stimulation were applied to InGaAs / C8-BTBT and InGaAs / PC, respectively. 61 BM heterojunction devices achieve LTM functionality through synergistic optoelectronic modulation (see Figures 32a-32b). For the human visual system, millions of nerve fibers transmit data to the brain for various processing tasks. In this context, the integrability and scalability of devices are essential for artificial visual systems. The combination of printable NW arrays with solution-processable organic thin films is suitable for such large-area fabrication, overcoming the difficulty of large-scale integration of heterojunction devices. Furthermore, photolithography techniques, such as photolithography, can be used to fabricate high-density device arrays. Therefore, this study fabricated a 4×4 array based on photosynaptic InGaAs / C8-BTBT devices and used human hair to create patterns of laser illumination on the device array (see Figure 5d).

[61] A 405nm laser pulse was then applied, and the percentage change in output current is shown in Figures 5e-5i. Due to the light blocking effect, the covered device experienced faster current recovery. As shown in Figure 5i, the location covered by hair could still be determined 120 seconds after irradiation.

[0134] More importantly, two different photosynthetic devices with negative and positive photoconductivity (i.e., InGaAs / C8-BTBT and InGaAs / PC) 61 BM) is used to build the array to define the hardware core. As shown in Figures 33a-33c and 34a-34c, the visual processing and recognition functions of the human eye are simulated by utilizing the light response characteristics of these synaptic devices to different illumination power densities and wavelengths (i.e., red, green, and blue time periods).

[62] The key role of the receptive cortex in the human retina is to accelerate visual perception in the brain by extracting its key features, thereby processing visual information as early as possible (see Figure 6a). Utilizing this principle, key features of an image can be extracted using a 3×3 hardware kernel, where different kernels have the function of extracting different important features.

[63] In this configuration, nine PPC and nine NPC devices each form their own array. Furthermore, one PPC device corresponds to one NPC device, and the conductivity of these devices can be modulated with variations in irradiation power density and wavelength. Each pair of photosynaptic devices is connected via a corresponding voltage input.

[64] According to Kirchhoff's laws, the final output current ( Ioutput ) is the sum of the currents of the PPC device and the NPC device. This allows for the successful formation of a hardware convolutional kernel array.

[65] In this array, a pair of positive and negative photoconductors are considered a complete device. Therefore, both the "on" and "off" responses can exist within a single kernel array, as shown in Figure 6b. Additionally, Figure 6c illustrates the functionality of hardware kernel-based vision processing, including extracting key features from the image shown in Figure 35 and experimental details. As a demonstration, three different hardware kernels are used here for vision processing. These three different image processing methods can then be implemented by modulating specific conductance values ​​for each device with different irradiation wavelengths. In this case, the inversion, edge, and embossing of the convolution kernel (3×3) can be mapped onto the hardware kernel array. The experimental results of the hardware kernel operation are in good agreement with the software simulation results, reproducing the image features as shown in Figure 6d. It should be noted that the experimental results are from InGaAs / PC... 61 Extracted from the photoresponse data of BM and InGaAs / C8-BTBT devices. Furthermore, experimental results on edges and embossing show that the image features can be more richly detailed due to the different conductivity amplitudes between the two different devices in the hardware core.

[0135] In another experimental setup, features detected by a hardware kernel (built from a device with adjustable conductivity) were fed into a neural network for classification and recognition to evaluate the network's performance. The dataset consisted of six colored images (3×3 pixels) containing the letters "V" (red), "G" (green), and "B" (blue), respectively.

[66] In this work, the image input signals are defined as 0.8V to 1V (for R), 0.4V to 0.6V (for G), and 0.1V to 0.3V (for B).

[23] In this way, images can be randomly generated according to this definition. These 250 images were arbitrarily generated from each of the six color numbers. Therefore, 1200 images were used as the training set for learning, while 300 images were used as the test set to evaluate classification accuracy. Figure 6e shows the key feature values ​​of the images as input signals to a neural network with nine hardware kernels, consisting of three red, three green, and three blue optical guide kernels. Furthermore, all hardware kernels possess inverted, edge, and embossed features. Therefore, the single-layer perceptron neural network consists of nine input and six output neurons. The recognition rates of the neural networks with and without hardware kernels were then evaluated separately (see Figure 6f). Impressively, the recognition rate with hardware kernels reached 100%; in contrast, the recognition rate without hardware kernels was only about 51%, verifying that the hardware kernels are capable of detecting and classifying target features of different colors. Generally, recognizing the color of one image pixel requires three neurons; here, a 9-pixel image requires 27 input neurons. Colors can be directly distinguished by the hardware kernels of this invention. The number of input neurons required for the neural network has been reduced to nine. Notably, when constructing artificial hardware neural networks, the hardware kernel can significantly reduce the number of input neurons and synaptic weights, thereby reducing the complexity and power consumption of the neural network. All these results demonstrate the effectiveness of using two different types of devices (i.e., InGaAs / C8-BTBT and InGaAs / PC). 61 The applicability of the hardware kernel built by BM to artificial vision systems.

[0136] In conclusion, according to some embodiments of the present invention, InGaAs NW arrays were successfully coupled with C8-BTBT or PC. 61 BM's organic semiconductor thin film top layer encapsulates arrays of artificial photosynaptic devices exhibiting persistent NPC or PPC phenomena, respectively. Two different organic-inorganic type I heterojunctions are fabricated together with parallel arrays of InGaAs NWs to achieve tunable persistent NPC or PPC. These two type I heterojunctions (i.e., InGaAs / C8-BTBT and InGaAs / PC) 61Based on their different photoconductivities, artificial photosynaptic devices (ANSIDs) can effectively mimic basic synaptic visual behaviors such as PPF and EPSC. To demonstrate the large-scale fabrication capability of printable NW arrays and solution-processed organic thin films, the inventors fabricated a 4×4 array of artificial photosynaptic devices that exhibited reliable optical memory functionality when different gate voltages were applied. More importantly, utilizing the negative and positive photoconductivities of these two different photosynaptic devices, a hardware core simulating human visual processing and recognition can be constructed. Due to the tunable photoconductivity and response to different illumination wavelengths, a highly efficient neural network achieved 100% classification accuracy. All these results clearly demonstrate the great potential of artificial photosynaptic devices based on organic molecular nanowire heterojunctions for practical applications in artificial vision systems. Using a device array with multi-wavelength NPC and PPC phenomena to construct the hardware core of a vision system achieved 100% recognition rate with low complexity and power consumption.

[0137] As can be seen, the embodiments of the present invention can provide an innovative method for designing multifunctional device structures with high mobility III-V group semiconductor nanowires and organic semiconductor materials, thereby not only realizing human visual functions with negative / positive photoconductivity, but also simplifying the manufacturing process and realizing large-area integration and arraying.

[0138] Artificial vision systems based on organic molecular nanowire heterojunctions can mimic the NPC or PPC of human visual function. Irradiation with different wavelengths (from solar blind to visible light range) and power densities can stimulate these devices, effectively achieving synaptic behaviors with two different photoconductivities. Compared to devices that can achieve positive and negative photoconductivities at different wavelengths, this avoids complex transfer processes and improves air stability. Furthermore, large-area heterojunction arrays are realized through solution processing. In this regard, utilizing heterojunctions integrating III-V group nanowire arrays with organic semiconductors to construct high-performance artificial vision systems is highly desirable.

[0139] As a demonstration and proof of concept, this device structure, using group III-V semiconductor nanowires and organic semiconductor materials, enabled the hybrid device to exhibit two distinct photoconductivities. Notably, the synaptic behaviors of STP, LTM, EPSC, and LTP / LTD were reliably demonstrated under both visible and ultraviolet laser light, including the range of solar blindness. Furthermore, the integration of printed NW arrays and solution-processed organic materials illustrates their significant potential as active materials for large-area neural visual networks.

[0140] Therefore, exemplary embodiments of the present invention have been fully described above. Although specific embodiments have been mentioned in the description, it will be clear to those skilled in the art that the invention can be practiced with variations in these specific details. Therefore, the invention should not be construed as being limited to the embodiments described herein.

[0141] While embodiments of the invention have been detailed and described in the accompanying drawings and foregoing description, they should be considered illustrative rather than restrictive. It should be understood that only exemplary embodiments are shown and described, and the scope of the invention is not limited in any way. It is understood that any feature described herein can be used in any embodiment. The illustrative embodiments do not exclude each other or other embodiments not mentioned herein. Therefore, the invention also provides embodiments comprising combinations of one or more of the foregoing illustrative embodiments. Modifications and variations can be made to the invention without departing from its spirit and scope.

[0142] In a variation of a preferred embodiment, an artificial vision system with tunable photoconductivity based on an organic molecular nanowire heterojunction is provided to achieve various visual neural functions, including color recognition, PPF / EPSC, and long-term and multi-state optical memory. The constituent III-V semiconductor nanowires and organic semiconductor materials are configured as device channels. An integrated structure of III-V semiconductor nanowires and organic semiconductor layered materials is provided. Furthermore, a process for fabricating this nanowire / organic semiconductor heterojunction device is also provided.

[0143] In some variations of the preferred embodiments, the III-V semiconductor nanowires are p-type, n-type, or bipolar.

[0144] In some variations of the preferred embodiments, the III-V semiconductor nanowires are binary, ternary, or quaternary.

[0145] In some variations of the preferred embodiments, the III-V semiconductor nanowires may be on or beneath an organic semiconductor material.

[0146] In some variations of the preferred embodiments, the dielectric layer of the device is a thermally grown SiO2 layer on a silicon wafer.

[0147] In some variations of the preferred embodiments, the electrode patterning method for nanowires and organic semiconductor materials is a combination of ultraviolet lithography, spin coating, and electron beam lithography.

[0148] References

[0149] Each of the following references (and related appendices and / or supplements) is explicitly included in this reference in its entirety:

[0150] [1]Z.Zhang,S.Wang,C.Liu,R.Xie,W.Hu,P.Zhou,Nat.Nanotechnol.2021,17,27.

[0151] [2]I.Sanchez Esqueda,X.Yan,C.Rutherglen,A.Kane,T.Cain,P.Marsh,Q.Liu,K.Galatsis,H.Wang,C.Zhou,ACS Nano 2018,12,7352.

[0152] [3]M.-K.Kim,I.-J.Kim,J.-S.Lee,Sci.Adv.2021,7,1341.

[0153] [4]K.Roy,A.Jaiswal,P.Panda,Nature 2019,575,607.

[0154] [5]LFAbbott,WGRegehr,Nature 2004,431,796.

[0155] [6]Q.Liu,S.Gao,L.Xu,W.Yue,C.Zhang,H.Kan,Y.Li,G.Shen,Chem.Soc.Rev.2022,51,3341.

[0156] [7]L.Gu,S.Poddar,Y.Lin,Z.Long,D.Zhang,Q.Zhang,L.Shu,X.Qiu,M.Kam,A.Javey,Z.Fan,Nature.2020,581,278.

[0157] [8]VKSangwan,MCHersam,Nat.Nanotechnol.2020,15,517.

[0158] [9]L.-Q.Tao,H.Tian,Y.Liu,Z.-Y.Ju,Y.Pang,Y.-Q.Chen,D.-Y.Wang,X.-G.Tian,J.-C.Yan,N.-Q.Deng,Y.Yang,T.-L.Ren,Nat.7,8.2017.

[0159]

[10] M.Wang,Y.Luo,T.Wang,C.Wan,L.Pan,S.Pan,K.He,A.Neo,X.Chen,Adv.Mater.2021,33,2003014.

[0160]

[11] J.Xue,Z.Zhu,X.Xu,Y.Gu,S.Wang,L.Xu,Y.Zou,J.Song,H.Zeng,Q.Chen,NanoLett.2018,18,7628.

[0161]

[12] S.Gao,G.Liu,H.Yang,C.Hu,Q.Chen,G.Gong,W.Xue,X.Yi,J.Shang,R.-W.Li,ACS Nano 2019,13,2634.

[0162]

[13] F.Guo,M.Song,M.-C.Wong,R.Ding,W.F.Io,S.-Y.Pang,W.Jie,J.Hao,Adv.Funct.Mater.2022,32,2108014.

[0163]

[14] L.Q.Zhu,C.J.Wan,L.Q.Guo,Y.Shi,Q.Wan,Nat.Commun.2014,5,3158.

[0164]

[15] S.Lee,R.Peng,C.Wu,M.Li,Nat.Commun.2022,13,1485.

[0165]

[16] Y.Chen,W.Qiu,X.Wang,W.Liu,J.Wang,G.Dai,Y.Yuan,Y.Gao,J.Sun,Nano.Energy 2019,62,393.

[0166]

[17] W.Qiu,Y.Huang,L.-A.Kong,Y.Chen,W.Liu,Z.Wang,J.Sun,Q.Wan,J.H.Cho,J.Yang,Y.Gao,Adv.Funct.Mater.2020,30,2002325.

[0167]

[18] J.-L.Meng,T.-Y.Wang,L.Chen,Q.-Q.Sun,H.Zhu,L.Ji,S.-J.Ding,W.-Z.Bao,P.Zhou,D.W.Zhang,Nano Energy 2021,83,105815.

[0168]

[19] Q.Shi,D.Liu,D.Hao,J.Zhang,L.Tian,L.Xiong,J.Huang,Nano Energy2021,87,106197.

[0169]

[20] J.-Y.Wu,Y.T.Chun,S.Li,T.Zhang,J.Wang,P.K.Shrestha,D.Chu,Adv.Mater.2018,30,1705880.

[0170]

[21] X.Huang,Q.Li,W.Shi,K.Liu,Y.Zhang,Y.Liu,X.Wei,Z.Zhao,Y.Guo,Y.Liu,Small 2021,2102820.

[0171]

[22] Y.Wang,E.Liu,A.Gao,T.Cao,M.Long,C.Pan,L.Zhang,J.Zeng,C.Wang,W.Hu,S.-J.Liang,F.Miao,ACS Nano 2018,12,9513.

[0172]

[23] S.Seo,S.-H.Jo,S.Kim,J.Shim,S.Oh,J.-H.Kim,K.Heo,J.-W.Choi,C.Choi,S.Oh,D.Kuzum,H.S.P.Wong,J.-H.Park,Nat.Commun.2018,9,5106.

[0173]

[24] Z.-x.Yang,N.Han,M.Fang,H.Lin,H.-Y.Cheung,S.Yip,E.-J.Wang,T.Hung,C.-Y.Wong,J.C.Ho,Nat.Commun.2014,5,5249.

[0174]

[25] D.Li,C.Lan,A.Manikandan,S.Yip,Z.Zhou,X.Liang,L.Shu,Y.-L.Chueh,N.Han,J.C.Ho,Nat.Commun.2019,10,1664.

[0175]

[26] S.Ham,M.Kang,S.Jang,J.Jang,S.Choi,T.-W.Kim,G.Wang,Sci.Adv.2020,6,1178.

[0176]

[27] P.Xie,Y.Huang,W.Wang,Y.Meng,Z.Lai,F.Wang,S.Yip,X.Bu,W.Wang,D.Li,J.Sun,JCHo,Nano Energy 2022,91,106654.

[0177]

[28] JJHou,N.Han,F.Wang,F.Xiu,S.Yip,ATHui,T.Hung,JCHo,ACSNano2012,6,3624.

[0178]

[29] D.Li,S.Yip,F.Li,H.Zhang,Y.Meng,X.Bu,X.Kang,C.Lan,J.Liu,JCHo,Adv.Opt.Mater.2020,2001201.

[0179]

[30] C.-M.Yang,T.-C.Chen,D.Verma,L.-J.Li,B.Liu,W.-H.Chang,C.-S.Lai,Adv.Funct.Mater.2020,30,2001598.

[0180]

[31] Y.Zhu,J.Wu,Z.Xu,Y.Liu,H.Hu,T.Guo,TWKim,Y.Chai,F.Li,NanoLett.2021,21,6087.

[0181]

[32] JJHou,F.Wang,N.Han,F.Xiu,S.Yip,M.Fang,H.Lin,TFHung,JCHo,ACSNano 2012,6,9320.

[0182]

[33] Y.Meng,C.Lan,F.Li,S.Yip,R.Wei,X.Kang,X.Bu,R.Dong,H.Zhang,JCHo,ACS Nano 2019,13,6060.

[0183]

[34] W.Wang,S.Yip,Y.Meng,W.Wang,F.Wang,X.Bu,Z.Lai,X.Kang,P.Xie,Q.Quan,C.Liu,JCHo,Adv.Opt.Mater.2021,9,2101289.

[0184]

[35] P.Xie,T.Liu,J.Sun,J.Yang,Adv.Funct.Mater.2022,2200843.

[0185]

[36] X.Li,P.Xie,X.Mo,Y.Xiang,Z.Xiao,T.Xiongsong,H.Huang,J.Sun,G.Dai,J.Yang,Phys.Status.solidi-R 2021,15,2000479.

[0186]

[37] H.Kobayashi,N.Kobayashi,S.Hosoi,N.Koshitani,D.Murakami,R.Shirasawa,Y.Kudo,D.Hobara,Y.Tokita,M.Itabashi,J.Chem.Phys.2013,139,014707.

[0187]

[38] H.Xia,S.Tong,C.Zhang,C.Wang,J.Sun,J.He,J.Zhang,Y.Gao,J.Yang,Appl.Phys.Lett.2018,112,233301.

[0188]

[39] L.Shen,S.Yip,C.Lan,L.Shu,D.Li,Z.Zhou,C.-Y.Wong,EYBPun,JCHo,Adv.Mater.Interfaces 2018,5,1701104.

[0189]

[40] J.Chen,CHYang,RAWilson,MJYang,Appl.Phys.Lett.1992,60,2113.

[0190]

[41] N.Guo,W.Hu,L.Liao,S.Yip,JCHo,J.Miao,Z.Zhang,J.Zou,T.Jiang,S.Wu,X.Chen,W.Lu,Adv.Mater.2014,26,8203.

[0191]

[42] Y.Han,M.Fu,Z.Tang,X.Zheng,X.Ji,X.Wang,W.Lin,T.Yang,Q.Chen,ACSAppl.Mater.Interfaces 2017,9,2867.

[0192]

[43] M.Zhang,Z.Tang,X.Liu,J.Van der Spiegel,Nat.Electron.2020,3,191.

[0193]

[44] H.-L.Park,Y.Lee,N.Kim,D.-G.Seo,G.-T.Go,T.-W.Lee,Adv.Mater.2020,32,1903558.

[0194]

[45] S.Dai,X.Wu,D.Liu,Y.Chu,K.Wang,B.Yang,J.Huang,ACSAppl.Mater.Interfaces 2018,10,21472.

[0195]

[46] C.Han,X.Han,J.Han,M.He,S.Peng,C.Zhang,X.Liu,J.Gou,J.Wang,Adv.Funct.Mater.2022,2113053.

[0196]

[47] C.Fan,X.Xu,K.Yang,F.Jiang,S.Wang,Q.Zhang,Adv.Mater.2018,30,1804707.

[0197]

[48] W.Eaton Samuel,M.Lai,A.Gibson Natalie,B.Wong Andrew,L.Dou,J.Ma,L.-W.Wang,R.Leone Stephen,P.Yang,P.Natl.A.Sci 2016,113,1993.

[0198]

[49] Y.Zhao,X.Liu,G.Feng,L.Lyu,L.Li,S.Wang,J.Jiang,Y.Li,D.Niu,H.Xie,Y.Gao,Results.Phys 2020,19,103590.

[0199]

[50] H.Xie,D.Niu,Y.Zhao,S.Wang,B.Liu,Y.Liu,H.Huang,P.Wang,D.Wu,Y.Gao,Synthetic.Met 2020,260,116261.

[0200]

[51] D.Yang,X.Zhang,K.Wang,C.Wu,R.Yang,Y.Hou,Y.Jiang,S.Liu,S.Priya,Nano Lett.2019,19,3313.

[0201]

[52] J.Han,J.Wang,M.Yang,X.Kong,X.Chen,Z.Huang,H.Guo,J.Gou,S.Tao,Z.Liu,Z.Wu,Y.Jiang,X.Wang,Adv.Mater.2018,30,1804020.

[0202]

[53] C.-H.Chiang,C.-G.Wu,Nat.Photonics 2016,10,196.

[0203]

[54] S.Foster,F.Deledalle,A.Mitani,T.Kimura,K.-B.Kim,T.Okachi,T.Kirchartz,J.Oguma,K.Miyake,JRDurrant,S.Doi,J.Nelson,Adv.Energy.Mater2014,4,1.1400

[0204]

[55] Z.Ahmad,M.Awais,MANajeeb,RAShakoor,F.Touati,SpringerInternationalPublishing Cham 2017,95.

[0205]

[56] ASShikoh,Z.Ahmad,F.Touati,RAShakoor,J.Bhadra,NJAl-Thani,RSCAdv.2017,7,35445.

[0206]

[57] F.-C.Wu,P.-R.Li,B.-R.Lin,R.-J.Wu,H.-L.Cheng,W.-Y.Chou,ACS.Appl.Mater.Interfaces 2021,13,45822.

[0207]

[58] T.Chang,S.-H.Jo,W.Lu,ACS Nano 2011,5,7669.

[0208]

[59] L.Wu,A.Wang,J.Shi,J.Yan,Z.Zhou,C.Bian,J.Ma,R.Ma,H.Liu,J.Chen,Y.Huang,W.Zhou,L.Bao,M.Ouyang,S.J.Pennycook,S.T.Pantelides,H.-J.Gao,Nat.Nanotechnol.2021,16,882.

[0209]

[60] J.Meng,T.Wang,Z.He,Q.Li,H.Zhu,L.Ji,L.Chen,Q.Sun,D.W.Zhang,NanoRes.2021,15,2472.

[0210]

[61] Y.Meng,F.Li,C.Lan,X.Bu,X.Kang,R.Wei,S.Yip,D.Li,F.Wang,T.Takahashi,T.Hosomi,K.Nagashima,T.Yanagida,J.C.Ho,Sci.Adv.2020,6,6389.

[0211]

[62] S.Hong,H.Cho,B.H.Kang,K.Park,D.Akinwande,H.J.Kim,S.Kim,ACSNano2021,15,15362.

[0212]

[63] C.-Y.Wang,S.-J.Liang,S.Wang,P.Wang,Z.a.Li,Z.Wang,A.Gao,C.Pan,C.Liu,J.Liu,H.Yang,X.Liu,W.Song,C.Wang,B.Cheng,X.Wang,K.Chen,Z.Wang,K.Watanabe,T.Taniguchi,J.J.Yang,F.Miao,Sci.Adv.2020,6,6173.

[0213]

[64] P.Yao,H.Wu,B.Gao,J.Tang,Q.Zhang,W.Zhang,J.J.Yang,H.Qian,Nature2020,577,641.

[0214]

[65] H.Jang,C.Liu,H.Hinton,M.-H.Lee,H.Kim,M.Seol,H.-J.Shin,S.Park,D.Ham,Adv.Mater.2020,32,2002431.

[0215]

[66] M.Prezioso,F.Merrikh-Bayat,BDHoskins,GCAdam,KKLikharev,DBStrukov,Nature 2015,521,61.

[0216]

[67] Z.Fan,JCHo,ZAJacobson,R.Yerushalmi,RLAlley,H.Razavi,A.Javey,Nano Lett.2008,8,20.

[0217]

[68] T.Takahashi,K.Takei,JCHo,Y.-L.Chueh,Z.Fan,A.Javey,J.Am.Chem.Soc.2009,131,2102

Claims

1. An artificial photosynaptic device, comprising an organic molecule-nanowire heterojunction, the organic molecule-nanowire heterojunction including a first heterojunction having a first organic material film encapsulated on a first nanowire device; wherein, The organic molecule-nanowire heterostructure is C8-BTBT / InGaAs or PC. 61 BM / InGaAs.

2. The artificial photosynapse device according to claim 1 further comprises a second organic molecule-nanowire heterojunction; wherein, The second organic molecule-nanowire heterostructure is C8-BTBT / InGaAs or PC. 61 BM / InGaAs.

3. The artificial photosynapse device according to claim 2, wherein the second organic molecule-nanowire heterojunction has a second organic material film wrapped around the second nanowire device.

4. The artificial photosynaptic device according to claim 3, wherein the first nanowire device and the second nanowire device are printed nanowire arrays.

5. The artificial photosynapse device according to claim 3, wherein the first nanowire device and the second nanowire device are InGaAs nanowire arrays.

6. The artificial photosynapse device according to claim 5, wherein the InGaAs nanowire array is synthesized by a two-step catalytic solid-source CVD method.

7. The artificial photosynaptic device according to claim 5, wherein the InGaAs nanowire array is p-type, n-type, or bipolar.

8. The artificial photosynaptic device according to claim 5, wherein the InGaAs nanowire array is binary, ternary, or quaternary.

9. The artificial photosynapse device according to claim 5, wherein the InGaAs nanowire array is located above or below the first organic material film or the second organic material film.

10. The artificial photosynaptic device according to claim 3, wherein the first organic material film is different from the second organic material film, thereby giving the artificial photosynaptic device adjustable photoconductivity.

11. The artificial photosynaptic device of claim 3, wherein sustained negative photoconductivity (NPC) or positive photoconductivity (PPC) is achieved in the first heterojunction and the second organic molecule-nanowire heterojunction.

12. The artificial photosynapse device according to claim 1, wherein the dielectric layer of the artificial photosynapse device is a SiO2 layer thermally grown on a silicon wafer.

13. A method for manufacturing an artificial photosynaptic device as claimed in claim 1, comprising the steps of ultraviolet lithography, spin coating, and electron beam lithography.

14. An artificial photosynaptic device, comprising an organic molecule-nanowire heterojunction, further comprising a first heterojunction having a first organic material film encapsulated on a first nanowire device, and a second heterojunction having a second organic material film encapsulated on a second nanowire device; wherein the first organic material film is different from the second organic material film, thereby enabling the artificial photosynaptic device to have adjustable photoconductivity; and wherein the first organic material film is C8-BTBT, and the second organic material film is PC. 61 BM.

Images