近來由於傳統的范紐依曼結構的運算瓶頸已越發明顯，使得類神經網路運算的發展也越來越蓬勃。由於光脈衝與電脈衝對於氧化物薄膜電晶體皆能改變其電流值，進而對應到類神經網路之權重改變，故氧化物薄膜電晶體也成為了一項作為類神經網路運算硬體元件的候選。
本研究採用溶液法製備氧化鋅錫薄膜(ZTO)電晶體來模擬突觸之應用，其汲極類比為後突觸(post-synapse)，在汲極量測到之電流則類比為突觸後電流與突觸權重，並以閘極正偏壓電脈衝、閘極負偏壓電脈衝以及光脈衝(照射於通道)作為前突觸刺激，並藉此模擬突觸之多樣行為。
本研究採用了六種元件進行比較，前三種元件分別為主動層有進行微影蝕刻的一層ZTO元件並進行後退火0分鐘、5分鐘、60分鐘，元件標示為Patterned 1-ZTO Nonpostanneal、Patterned 1-ZTO Postanneal 5min、Patterned 1-ZTO Postanneal 60min，後三種元件為主動層無進行微影蝕刻的一層ZTO、十層ZTO、十層ZTO上具金奈米粒子的元件，元件標示為Nonpattern 1-ZTO、Nonpattern 10-ZTO、Nonpattern Au NPs(20mA/60s)/10-ZTO。
首先，第一部分中，本研究先利用量測元件去程以及回程的ID-VG轉換特性曲線來取得不同元件的遲滯窗口大小，遲滯窗口大小於六個元件，依序為7.7V、2.4V、1.8V、0.9V、0.1V、0.3V。
突觸操作部分，此研究中六種元件吾人皆採用相同脈衝條件進行操作。先採以單一閘極負偏壓電脈衝-5V、-10V、-15V，閘極正偏壓電脈衝+3V、+5V、+10V，且脈衝時間為500ms來觀看進行脈衝後的突觸後電流上升或下降行為，結果為閘極負偏壓電脈衝為增益型，閘極正偏壓電脈衝為抑制型刺激，且遲滯窗口大小變化對於進行電脈衝後的電流值變化行為為正相關。本研究以此探討以此為基礎機制來做的後續操作。接下來則為成對脈衝促進以及成對脈衝抑制，藉由兩刺激的時間間距變化來決定突觸後電流的促進或抑制程度，此兩種行為皆為短期突觸可塑性的展示。前者利用成對的-5V,500ms的脈衝來仿擬，後者利用成對的+5V,500ms的脈衝來仿擬，兩種操作時距皆為由100ms、200ms、300ms、500ms至1000ms。隨著時距的增加，成對脈衝促進或抑制程度在六個元件中並無規律性的變化，顯示電脈衝對元件並不具有短期突觸可塑性效應。再者，推展至長期突觸可塑性的行為展示，利用多次閘極負偏壓電脈衝-5V,500ms作為增益的條件，閘極正偏壓電脈衝+5V,500m，作為抑制的條件，藉由3次的增益刺激/抑制循環，並計算其線性程度α以及電流最高最低比值Imax/Imin，來量化其作為仿神經型態計算元件之效益。
第二部分中，本研究為了使類神經網路計算能夠有更多操作面向，引入光脈衝以及電脈衝作為兩前突觸刺激，觀看交互作用對於突觸後電流的影響，於六種元件下的條件皆為相同，光脈衝為405nm雷射，功率密度為2mW/cm2，脈衝時間為500ms；相對正偏壓電脈衝為+5V,500ms，相對負偏壓電脈衝為-5V,500ms。
當整合光脈衝與正偏壓電脈衝並調整兩脈衝的時序時，六種元件皆顯示出，隨著將光脈衝的時序提前，其電流值變化會向負方向增加，但依據各元件所進行電脈衝以及光脈衝的電流值不同，電流變化值則不相同，但皆產生了相同的趨勢，在同樣的脈衝操作條件下，僅改變兩種脈衝的時序，即能得到不同的突觸權重調控，給予元件更靈活的方式去調控突觸權重。當整合光脈衝與相對負偏壓電脈衝並調整兩脈衝的時序，元件在同時進行光脈衝與相對負偏壓電脈衝時產生了最大的電流值改變，光脈衝在前為電流值改變次之，光脈衝在後為電流值改變最小。
在同時進行光脈衝與相對負偏壓電脈衝時產生了最大的電流值改變的操作，吾人獨立分析其電流值變化，同時施加兩種脈衝時相比單獨施加分別的脈衝所提升的電流值要高，產生了協同作用，此現象使我們的元件能夠以更有效率的方式提升突觸權重，吾人利用同時施加兩種脈衝所產生的電流值變化除以個別脈衝所產生的電流值變化來檢視協同作用的增幅倍率，協同作用的倍率在六種元件下操作的結果依序為1.83、3.71、2.80、3.85、1.83、3.25，最高能夠達到3.85倍的增幅效益。

The traditional Von Neumann computer architecture experiences bottleneck in Big Data Era. Neuromorphic computing system which integrate computing and memory units has high potential to deal with such issue. Researches for such system using various devices have increased drastically in the past few years. According to literatures, amorphous oxide thin film transistor can response to electrical pulse and light pulse to change its current value which emulate synaptic weight. In this study, Zinc Tin Oxide (ZTO) thin film transistor (TFT) was fabricated using spin-coating process to simulate synaptic functions. With different experimental parameters, six devices with different electrical properties are fabricated and further discussed in this study. In the first part of the study, our devices shows clockwise hysteresis in transfer characteristic curve. Excitatory post synaptic current (EPSC) and inhibitory post synaptic current (IPSC) are shown by applying negative gate electrical pulse and positive gate electrical pulse to devices, respectively. Potentiation and Depression cycle is also mimicked and analyzed by linearity and dynamic range. In the second part of the study, 405nm light serves as a second pre-synaptic input. Dual operation mode are exhibited by altering temporal order of positive gate electrical pulse and light pulse, temporal order of negative gate electrical pulse and light pulse. PSC varies as temporal order changes which give us a more dynamic way of synaptic weight change. Synergistic effect occurs by simultaneously applying negative gate electrical pulse and light pulse. A more efficient way of reaching higher synaptic state is demonstrated.

1. Wan, Q.; Sharbati, M. T.; Erickson, J. R.; Du. Y.; Xiong, F. Emerging Artificial Synaptic Devices for Neuromorphic Computing. Adv. Mater. Technol. 2019, 4, 1900037. doi:10.1002/admt.201900037
2. T. Kamiya, H. Hosono Handbook of Visual Display Technology, J. Chen, W. Cranton, M. Fihn (eds.), Springer, 2016.
3. Thomas, S. R.; Pattanasattayavong, P.; Anthopoulos, T. D. Solution-processable metal oxide semiconductors for thin-film transistor applications. Chemical Society Reviews, 2013, 42(16), 6910-6923. doi:10.1039/C3CS35402D
4. Ahn, S. E.; Song, I.; Jeon, S.; Jeon, Y. W.; Kim, Y.; Kim, C.; Ryu, B.; Lee, J. H.; Nathan, A.; Lee, S.; Kim, G. T.; Chung, U. Metal oxide thin film phototransistor for remote touch interactive displays. Advanced materials, 2012, 24(19), 2631-2636. doi:10.1002/adma.201200293
5. Jiang, J.; Guo, J.; Wan, X.; Yang,Y.; Xie, H.; Niu, D.; Yang, J.; He, J.; Gao, Y.; Wan, Q. 2D MoS2 Neuromorphic Devices for Brain-Like Computational Systems. small. 2017, 13, 1700933. doi:10.1002/smll.201700933
6. John, R. A.; Liu F.; Chien, N. A.; Kulkarni, M. R.; Zhu, C.; Fu, Q.; Basu, A.; Liu, Z.; Mathews, N. Synergistic Gating of Electro-Iono-Photoactive 2D Chalcogenide Neuristors:Coexistence of Hebbian and Homeostatic Synaptic Metaplasticity. Adv. Mater. 2018, 30, 1800220. doi:10.1002/adma.201800220
7. Byrne, J. H.; Byrne, J.; Roberts, J. L. (Eds.). From molecules to networks. Burlington: Academic Press, 2009.
8. Ghaffarian, N.; Mesgari, M.; Cerina, M.; Göbel, K.; Budde, T.; Speckmann, E. J.; Meuth, S, G.; Gorji, A. Thalamocortical-auditory network alterations following cuprizone‐induced demyelination. Journal of neuroinflammation, 2016, 13(1), 160. doi:10.1186/s12974-016-0629-0
9. Wang, S.; Chen, C.; Yu, Z.; He, Y.; Chen, X.; Wan, Q.; Shi, Y.; Zhang, D. W. Zhou, H.; Wang, X.; Zhou P. A MoS2-PTCDA Hybrid Heterojunction Synapse with Efficient Photoelectric Dual Modulation and Versatility. Adv. Mater. 2019, 31, 1806227.doi: 10.1002/adma.201806227
10. Yu, F.; Zhu, L. Q.; Xiao, H.; Gao, W. T.; Guo Y. B. Restickable Oxide Neuromorphic Transistors with Spike-Timing-Dependent Plasticity and Pavlovian Associative Learning Activities. Adv. Funct. Mater. 2018, 28, 1804025. doi:10.1002/adfm.201804025
11. Hu, W.; Jiang, J.; Xie, D.; Liu, B.; Yang J.; He J. Proton–electron-coupled MoS2 synaptic transistors with a natural renewable biopolymer neurotransmitter for brain-inspired neuromorphic learning. J. Mater. Chem. C, 2019, 7, 682—691. doi:10.1039/C8TC04740E
12. Bi, G. Q.; Poo, M. M. Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type. Journal of neuroscience, 1998, 18(24), 10464-10472. doi:10.1523/JNEUROSCI
13. Indiveri, G.; Liu, S. C. Memory and Information Processing in Neuromorphic Systems. IEEE. 2015, 103, 1379-1397. doi:10.1109/JPROC.2015.2444094
14. Duan, N.; Li, Y.; Chiang, H. C.; Huang, S. P.; Yin, K. S.; Chen, J.; Yang, C. I.; Chang, T. C.; Miao, X. S. Gate Modulation of Excitatory and Inhibitory Synaptic Plasticity in a Low-Temperature Polysilicon Thin Film Synaptic Transistor. ACS Appl. Electron. Mater. 2019, 1, 132−140. doi:10.1021/acsaelm.8b00060
15. Trinh, T. T.; Ryu, K.; Jang, K.; Lee, W.; Baek, S.; Raja, J.; Yi, J. Improvement in the performance of an InGaZnO thin-film transistor by controlling interface trap densities between the insulator and active layer. Semiconductor science and technology. 2011, 26(8), 085012. doi:10.1088/0268-1242/26/8/085012
16. Jang, H.; Lee, S. J.; Porte, Y.; Myoung, J. M. Selective metallization of amorphous-indium–gallium–zinc-oxide thin-film transistor by using helium plasma treatment. Semiconductor Science and Technology. 2018, 33(3), 035011. doi:10.1088/1361-6641/aaa9e7
17. Du, J. Y.; Ge, C.; Riahi, H.; Guo, E. J.; He, M.; Wang, C.; Yang, G, Z.; Jin, K. J. Dual‐Gated MoS2 Transistors for Synaptic and Programmable Logic Functions. Advanced Electronic Materials. 2020, 1901408. doi:10.1002/aelm.201901408
18. Li, J.; Jiang, D.; Yang, Y.; Zhou, Y.; Chen, Q.; Zhang, J. Li‐Ion Doping as a Strategy to Modulate the Electrical‐Double‐Layer for Improved Memory and Learning Behavior of Synapse Transistor Based on Fully Aqueous‐Solution‐Processed In2O3/AlLiO Film. Advanced Electronic Materials. 2020, 6(4), 1901363. doi: 10.1002/aelm.201901363
19. Guo, Y. B.; Zhu, L. Q.; Long, T. Y.; Wan, D. Y.; Ren, Z. Y. Bio-polysaccharide electrolyte gated photoelectric synergic coupled oxide neuromorphic transistor with Pavlovian activities. Journal of Materials Chemistry C. 2020, 8(8), 2780-2789. doi: 10.1039/c9tc06749c