Fig. 1. (Color online) Summary of the review. According to the working mechanisms, SiC-based synaptic devices can be categorized into two types: electrically and optically stimulated synaptic devices. Commonly used materials types include amorphous SiC thin film^[58], single-crystal SiC thin film^{[44, 57, 59]}, and SiC nano wires^{[43, 60]}. Several application scenarios for neuromorphic computing include logic functions, wireless transmission, high-temperature image learning and memory, as well as high-temperature color quantization.

Fig. 2. (Color online) (a) Schematic of the Ag/SiC/Pt structure. The device mimic (b) STDP, and (c) PPF. (d) Diagram of switching dynamics in Ag/SiC/Pt devices^[57]. (e) Schematic of the Cu/SiC/W structure. (f) Diagram of the formation of Cu conductive filament in Cu/SiC/W devices. The device mimic (g) SRDP, (h) SVDP, and (i) SDDP^[58].

Fig. 3. (Color online) (a) Schematic of the 4H-SiC/PVK/P3HT synaptic transistor. (b) Energy band diagram of 4H-SiC, PVK, and P3HT. The device mimic (c) PPF, (d) SDDP, (e) SNDP, (f) SRDP, and (g) learning-forgetting-relearning behavior. (h) EPSC of the device triggered by 400 optical spikes, which didn’t decay completely even 10⁴ s after the stimulus stopped^[59].

Fig. 4. (Color online) (a) Schematic diagram of the ITO/PMMA/3C-SiC nano wire/ITO synaptic device and a typical biological synapse. (b) Electron transport of the 3C-SiC nano wire device with and without light illumination. The device mimic (c) SDDP, (d) SNDP, and (e) classical conditioning of Pavlov’s dog^[60].

Fig. 5. (Color online) (a) Schematic of a bionic human visual system, the optoelectronic memristor array, and a single synaptic device. (b) Energy band diagram of 3C-SiC and NiO. The device mimic (c) LTP/LTD, (d) SNDP, and (e) learning-forgetting-relearning behavior^[43].

Fig. 6. (Color online) (a) Schematic of the 4H-SiC synaptic device. (b) Working mechanism of the 4H-SiC device. The device mimic (c) PPF, (d) SNDP, and (e) SRDP at 327 °C^[44].

Fig. 7. (Color online) (a) Schematic diagram of the information integration in the synaptic device with multi-terminal inputs. (b) A spiking logic response by dual modulatory input at 0.1 and 0.5 V read voltage for achievement of "AND" and "OR" logic, respectively. (c) Histograms of the post-synaptic current for "AND" and "OR" logic at 0.1 and 0.5 V read voltage, respectively^[60].

Fig. 8. (Color online) Encodement of the International Morse code of (a) "hello" and (b) "world" with the 365 nm light-stimulated EPSC of 3C-SiC nano wire synaptic device. Correlation between EPSC values and the International Morse code of English letter. Each letter is linearly correlated with (c) the sum and (d) the end of EPSC amplitude peak values^[60].

Fig. 9. (Color online) The conductance response of the synapse device stimulated by different number of light pulses after decay time of (a) 0, (b) 5, and (c) 10 s. Conductance response images were obtained in the device array after applying (d) 1, (e) 5, and (f) 10 light pulses^[43].

Fig. 10. (Color online) (a) Schematic of the array-based self-organizing map neural network. (b) Dependence of the quantization error on the number of competitive nodes. (c) Visual results of the color quantization^[44].