2D α-In₂Se₃ under strain and ferroelectric neuromorphic computing applications

Abstract

Since the discovery of graphene, the family of 2D materials covering insulators, semiconductors, and metals has been extensively studied in the field of optoelectronics and provides an excellent platform to explore fantastic physical phenomena at the atomic level. As a subgroup, 2D piezoelectric and ferroelectric materials offer an additional degree of freedom to modulate the optoelectronic properties of the material and multifunctionalize the device with the help of ferroelectricity and piezoelectricity. In this thesis, firstly, the controllable biaxial strain is experimentally imposed on α-In2Se3 nanosheets by an electromechanical device. A red shift of Raman spectra of the nanosheets is observed under the strain. The Grüneisen parameter is calculated to analyze the strain effect on the vibrational behavior. Photoluminescence shows a blue shift which can reach up to 215 meV per 1% strain. Such tunability of optical characteristics observed from α-In2Se3 nanosheets is much higher than that of conventional semiconductors. Physical mechanism behind the observation is investigated, which is related to the variations in energy band and photoexcited carriers under piezoelectric field and laser power. Secondly, 2D ferroelectric α-In2Se3 is utilized as the channel to construct a novel artificial neuron with an easy-to-prepare structure. And both of linearly separable and nonseparable logic can be performed at the single device level rather than a network based on the principle of combinational and stateful logic, respectively. Compared to the traditional design based on CMOS, the transistor resource can be decreased to 1/6 for XOR logic gate. In addition, as artificial neurons, there is a high demand for implementing flexible and adaptable behaviors. Thus, the behaviors of multi-terminal synapses are exhibited and the ability to modulate synaptic weights is investigated. Specifically, controlling food intake as a complex nervous system-level behavior is integrated to carry out positive and negative feedback in our device for the first time. Thirdly, a novel multifunctional synaptic device based on ferroelectric α-In2Se3/GaSe vdW heterostructure is proposed to emulate the entire biological visual system. Essential synaptic behaviors were observed in response to light and electrical stimuli; additionally, the retina-like selectivity for light wavelengths and the achievement of Pavlov's dog experiment demonstrate the device's capacity for processing complex electrical and optical inputs. Beyond the optoelectronic synaptic behaviors, the device incorporates memory and logic functions analogous to those in brain's visual cortex. The results of artificial neural network simulations show that the vdW heterostructure-based device is completely capable of performing logic operations and recognizing image with a high degree of accuracy. In conclusion, our work demonstrates the ability of piezoelectric fields induced by strain to modulate the energy band structure of 2D materials, and the working mechanism is analyzed in detail. Electronic and optoelectronic synaptic devices are designed based on 2D ferroelectric α-In2Se3 which exhibits great potential for breaking the bottleneck of von Neumann computers and simplifying current artificial neural networks as well as artificial vision systems.