Conjugated Polymer‐Wrapped Single‐Wall Carbon Nanotubes for High‐Mobility Photonic/Electrical Fully Modulated Synaptic Transistor

Abstract

Artificial synaptic thin-film transistor (TFT) devices have recently attracted considerable attention because they can emulate biological synapses, which enables the devices to process and store information like the human brain [1] wherein computation and storage are executed simultaneously-thereby enabling spatiotemporal information processing through timedependent neuron interconnections. Synapses play a major role linking signals transmitted between pre-and postneurons, thereby enabling the brain to function. Neuronal connectivity strength, also known as "plasticity," determines the brain's cognitive intelligence. Adjusting neuronal plasticity enables the brain to concurrently process and store information. [2] Inspired by this collective and adaptive property of human memory, researchers have dedicated considerable effort to developing and implementing neuromorphic computational systems to emulate the brain's strategy of parallel information processing and storage. [3] Such a device would be highly promising for rapidly advancing artificial intelligence [4] because replacing inefficient and high-power-consuming traditional von Neumann-architecture-based computers with separate processor and memory units is feasible. [2c,e,5] To date, synaptic TFT devices have demonstrated their suitability for many practical applications including artificial tactile sensory organs, neurological electronic skin, wearable intelligent electronics, and neuromorphic vision systems. [4a,6] However, fully implementing current devices is limited by numerous challenges. First, most reported devices exhibit sophisticated architectures involving complex fabrication procedures, [1a,5b] which can render such devices expensive and ineffective. In addition, because most reported synaptic devices are electrically operated, they may be limited by the bandwidth/connection-density tradeoff and may exhibit lowspeed signal transmission. [5b,7] Photonic-modulated synaptic devices, on the other hand, have rarely been reported. However, such devices operate based on focused light stimuli exhibiting Although synaptic devices have already demonstrated their operability through electric or photonic signals or a combination thereof, current challenges include developing a single hardware synaptic device that is independently fully operational through either photonic or electric signals to improve device versatility. Additionally, most previously reported devices are fabricated using multiple technical processes-which impede device implementation-while the low-output current triggered in most such devices limits the possible integration of auxiliary gadgets. Therefore, by spontaneously wrapping a conjugated block copolymer around single-walled carbon nanotubes (SWCNTs), a thin-film transistor memory device comprising single-layered poly(9,9-dioctylfluorene)-b-polyisoprene (PF-b-PI)-wrapped-SWCNTs-which function as both a semiconductor and an electret layer-to simplify the device structure and fabrication is designed. Owing to the robust SWCNT ...

A brief review on device operations and working mechanisms of organic transistor photomemories

“…In our recent work on photosynaptic transistors, the use of single-walled carbon nanotubes (SWCNTs) wrapped single-layered poly(9,9-dioctylfluorene)- b -polyisoprene (PF- b -PI) which serves as both semiconductor and electret layer has demonstrated. 11 Owing to the high charge carrier mobility of SWCNTs (∼11.3 cm 2 V −1 s −1 ) and inherently photoactive conjugated polymer of PF- b -PI, this hybrid material endows the device with high output current (10 −4 to 10 −3 A) and finely-modulated synaptic behavior against either light or electrical signal. Besides possessing a large memory window (>70 V), the LTP to LTD could be emulated under 21 light pulses of blue light (395–415 nm at 25 mW cm −2 ) for 10 s each at the intervals of 30 s and continually monitoring the PSC in the dark for 1000 s, and the synaptic PPF behavior by the application of pair of presynaptic light pulses for 10 s each at intervals of 300 s. The electrical triggered synaptic properties resulted in the PPF index (a time interval of the synaptic device, which follows a biexponential function to identify and decode information) 20 that matches well with the biological counterpart.…”

“…As mentioned in the introduction section, many promising applications in the rapid growth of IoT 4 and AI 5,6 have been realized by adopting the concept of device operations and mechanisms of transistor photomemory. [7][8][9][10][11][12][13][14][15][16][17] Among the plethora of applications, in this section, the discussion will focus on two crucial applications, there are multilevel storage memory cells 7,8 and artificial synapses 10,11 due to the similarity of working principle utilization; 91 for other fascinating and feasible applications based on photonic memory devices can be found in the recent reports and literature reviews. 20,30,44,50,[92][93][94][95][96][97]…”

“…18–20 In general, the transistor memories could be classified into three different types based on their polarization/charge storage materials, 21 as follows: (1) ferroelectric, 18,19 (2) floating-gate, 20,22–24 and (3) polymer electret transistor memory. 7,11,13,14 Among them, the polymer electret one can behave as a semi-permanent electric field and/or dipole moment that is formed by trapping electrostatic charges. 25,26 Besides this extraordinary function, the fabrication of transistor memory using polymer electrets is the most facile choice under solution processes and can be readily applied and adapted to many other processing methods from research scale to industrial applications.…”

“…As information technology has rapidly progressed, nowadays, many innovations in electronic components have also been developed specifically in response to the fields of the Internet of Things (IoT) and artificial intelligence (AI). [4][5][6] For instance, a high-density data storage, 7,8 a pulse monitoring sensor, 9 an artificial synapse, 10,11 a smart artificial retina, 12 a photorecorder, 13,14 pattern/image identification, 15,16 and an ultrasensitive photodetector; 17 all have been built-up based on a memory transistor framework. The clear difference between a memory transistor and a conventional one is in at least one additional polarization/charge storage layer that is sandwiched within the semiconductor and the gate contact (see Fig.…”

Inverter and Ternary Content‐Addressable Memory Based on Carbon Nanotube Transistors Using Chemical Doping Strategy

“…[1][2][3][4][5][6][7] Traditional silicon CMOS is in urgent need of improving performance in terms of power consumption and space occupancy, and has no way to meet the requirements of generality of memory computing in the era of big data. Therefore, in the face of huge computing and memory requirements, it is urgent to studied in multiple application areas, including several basic logic and storage devices, [47][48][49][50][51][52][53] synaptic devices, [54][55][56][57] healthcare devices, [58,59] radio frequency technology, [60,61] biosensing and detection technology, [62][63][64][65] communication engineering, [66,67] electrical applications, such as compressor, [68] ultralow power device, [69] AC amplifier, [70] and display electronics, [71] and so on. In the case of TCAM, as far as we know, the hybrid method of making TCAM has not been tried.…”