Neuromorphic devices based on fluorite‐structured ferroelectrics

Abstract

A continuous exponential rise has been observed in the storage and processing of the data that may not curtail in the foreseeable future. The required data processing speed and power consumption are restricted by the buses between the logic and memory devices that are characteristic of the von Neumann computing architecture. Bio-mimicking neuromorphic computing has garnered considerable academic and industrial interest to resolve these challenges.Additionally, devices based on emerging nonvolatile memories capable of mimicking the behaviors of synapses and neurons, which are the main elements in biological computing systems (brains), are attracting significant interest from the device community. With the discovery of ferroelectricity in fluorite-structured oxides, such as HfO 2 and ZrO 2 , which are compatible with the state-of-the-art complementary-metal-oxide-semiconductor processes, ferroelectric devices have rapidly evolved as the main direction of these research and development activities. Fundamental science related to fluorite-structured ferroelectrics has been intensively studied over the last decade. At present, the focus is gradually moving to practical applications, including neuromorphic computing and advanced classical processing or memory units in the conventional von Neumann architecture. However, despite its rapid development, the wealth of recent progress in neuromorphic computing devices based on fluorite-structured ferroelectrics has not been reviewed and systemized. This progress report comprehensively reviews and systemizes the recent progress in artificial synaptic and spiking neuron devices for neuromorphic computing based on fluorite-structured ferroelectrics.

Recent Advances and Future Prospects for Memristive Materials, Devices, and Systems

“…Various types of materials such as perovskites, 2D materials, polymers, and fluorite oxides have been found to have ferroelectricity and studied for next-generation memory devices. − The memory effect of the ferroelectric materials is attributed to the switching of electric dipole alignments between an upward and downward direction, driven by electric field. Ferroelectric memories include ferroelectric random-access memory (FeRAM), ferroelectric tunnel junction (FTJ), and ferroelectric transistors.…”

Emerging Fluorite-Structured Antiferroelectrics and Their Semiconductor Applications

“…Thus, the obtained film contains a mixture of different thermodynamically stable phases, such as the t-phase, m-phase, and c-phase. Because these phases have different dielectric constants (t-phase, 24–57; o-phase, 24–29; m-phase, 16; c-phase, 29), a large field drop in the non-FE m-phase or c-phase can cause an earlier hard breakdown due to larger stress. , Comparatively, the AFE t-phase is a thermodynamically stable phase in thin films owing to its lower interfacial energy, which helps achieve a uniform phase distribution. − This can also be the reason for the lower device-to-device variation in the homogeneous phase content in each cell.…”

“…Mimicking biological neurons or synapses is currently an important topic for future computing paradigms with high power efficiency. − Nonvolatile FE memories are considered promising candidates for artificial synapses or neuron devices based on gradual polarization modulation or stochastic nucleation-dominated polarization switching. The field-driven volatile induction of polarization in AFE materials has recently attracted considerable attention from academic researchers.…”

Stretchable, skin‐conformable neuromorphic system for tactile sensory recognizing and encoding

“…Three encoding strategies, including peak (gray columns), timing (purple columns), and peak-timing-convolution (blue columns), are applied to recognize the tactile information (Figure 5D). For peak-feature extraction, the proportion is defined as the peak number in each row, and they are (2,5), (1,5), and (2, 5) for "A", "F", and "H", respectively. For the temporal-feature encoding, proportion is defined as the time differences between the last and the first handwritten entries in each row.…”

“…In biological somatosensory systems, the perception, transmission, and processing of information rely on the distributed parallel nervous networks to efficiently solve complex and unstructured real-world problems. 1,2 For instance, tactile sensation is associated with the conversion of mechanical signals to electrical signals by mechanoreceptors. 3 Then these electrical signals flow through nerve fibers to presynaptic membranes, inducing the release of neurotransmitters and firings of the postsynaptic membranes, and finally deliver them to the brain to form tactile sensations.…”