-
Memristors stand out as the most promising candidates for non-volatile memory and neuromorphic computing due to their unique properties. A crucial strategy for optimizing memristor performance lies in voltage modulation, which is essential for achieving ultra-low power consumption in the nanowatt range and ultra-low energy operation below the femtojoule level. This capability is pivotal in overcoming the power consumption barrier and addressing the computational bottlenecks anticipated in the post-Moore era. However, for brain-inspired computing architectures utilizing high-density integrated memristor arrays, key device stability parameters must be considered, including the on/off ratio, high-speed response, retention time, and durability. Achieving efficient and stable ion/electron transport under low electric fields to develop low-voltage, high-performance memristors operating below 1 V is critical for advancing energy-efficient neuromorphic computing systems. This review provides a comprehensive overview of recent advancements in low-voltage memristors for neuromorphic computing. Firstly, it elucidates the mechanisms that control the operation of low-voltage memristor, such as electrochemical metallization and anion migration. These mechanisms play a pivotal role in determining the overall performance and reliability of memristors under low-voltage conditions. Secondly, the review then systematically examines the advantages of various material systems employed in low-voltage memristors, including transition metal oxides, two-dimensional materials, and organic materials. Each material system has distinct benefits, such as low ion activation energy, and appropriate defect density, which are critical for optimizing memristor performance at low operating voltages. Thirdly, the review consolidates the strategies for implementing low-voltage memristors through advanced materials engineering, doping engineering, and interface engineering. Moreover, the potential applications of low-voltage memristors in neuromorphic function simulation and neuromorphic computing are discussed. Finally, the current problems of low-voltage memristors are discussed, especially the stability issues and limited application scenarios. Future research directions are proposed, focusing on exploring new material systems and physical mechanisms that could be integrated into device design to achieve higher-performance low-voltage memristors.
-
Keywords:
- memristor /
- low-voltage /
- neuromorphic computing
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] -
器件
结构工作机制 开关电压 开关比 开关
速度/ns保留
时间耐久性
(循环)功耗/
能耗应用 文献 Pt/HfAlOx/TaN VCM BRS:
+1/–1 V— 50 — — 4.28 aJ 手写数字识别 [171] Ta/Ta2O5:Ag/Ru ECM BRS:
+0.7 V/–0.7 V— 100 ≈5×104 s 5×107 — — [42] Pt/YSZ/Zr VCM BRS:
+0.7 V/–0.7 V— 2 104 s 108 — — [172] Ag/SnOx/SnSe ECM BRS: +0.4/–0.1 V >103 — 105 s 4000 — — [157] EGaIn/MACsPbI/
PEDOT: PSS/ITOVCM BRS:
+0.6/–0.41 V>105 — 105 s 104 3.8 mW — [114] ITO/FA1–yMAyPbI3–xClx/
(PEA)2PbI4/AuVCM BRS:
+1.0/–0.5 V— — — 200 1 fJ 突触功能模拟 [49] Ag/PMMA/MAPbI3:
Ag/AuECM TS:±0.22 V — 40 — 2500 10 μW 伤害传感器 [59] Ag/CsPbI3/Ag ECM TS:100 mV — — 100 ms — 2 nW 储备池计算 [111] PET-ITO/MAPbI3/
PEAI/AuVCM BRS:
+1/–1 V— — — 50 13.5 aJ 神经元积分-
发放功能[152] Ag/MoOx/
CsI (CsBr)/AgECM BRS:
–0.16/+0.07 V>1010 <200 >106 s >105 <3.31 pW 模拟手写数字分类 [62] Pt/CuI/Cu ECM BRS:
+0.64/–0.19 V103 — 17 h 125 8.73 µW 图像硬件加密和解密 [10] Ag/PMMA/
Cs2AgBiBr6/ITOECM BRS:
+0.6/–0.6 V>10 — — — 188 pJ 手写数字识别 [58] Pt/MoS2/Ti VCM BRS:
+0.65/–0.90 V160 — 10 years 1×107 — — [117] Au/HfSe2/Au VCM BRS:
+0.742/–0.817 V102 — — 500 0.82 pJ 矩阵计算 [80] Ag/BNOx/Graphene ECM BRS: 0.6/0.1 V 100-1000 — — 100 — — [75] Ag/Protein nanowires/Ag ECM TS:60 ± 4 mV — — — 104 — 神经元-突触
联立积分发放[126] Au/PBFCL10/Ag ECM BRS:
+0.2/–0.2 V— 21 >106 s — 2.35 μW HNN [131] ITO/PEDOT:PSS/
pTPD/CsPbBr3NCs/AgECM TS:<1 V 103 — 105 s TS:2×106BRS:5.6×103 — 储备池计算 [154] Au/MSFP/Au VCM BRS:
+1.0/–1.0 V— — 104 s 100 — 图像处理 [106] ITO/PVK:TCNQ/Ag ECM BRS:
+0.69/–0.52 V
TS:0.21 V≈103 — 104 s 104 15.2 μW 突触、神经元
功能模拟[109] Au/TPPS/Au VCM BRS:
–0.1/+0.3 V— — — — 16.25 pW—
2.06 nW突触模拟 [105] W(Ag)/PI/Pt/Ti ECM TS:0.56 V ≈103 — 0.44 ms 300 80 nW 图像处理 [107] Pt/CuZnS/Ag ECM Vset=0.089 V ≈106 — >1000 s 100 0.1 nW 模式识别 [132] Pt/DDP-CuNPs/Au VCM TS:4 mV — — — 100 — SNN [145] Ag/c-YY NW/Ag ECM TS:≤0.1 V 106 — — — 750 fJ SNN [124] Ag/Ag-IPS/Au ECM BRS:
+0.43/–0.21 V108 100 105 s 900 18.5 fJ 图像处理 [138] Ag/PMMA/MAPbI3:
Ag/AuECM TS:≈0.2 V — 40 — 2500 — 伤害感受器 [59] Al/Ti3C2:Ag/Pt VCM BRS:
+2.0/–2.0 V— — — 106 0.35 pJ 突触模拟 [137] Ag/TiO2:Ag/Pt ECM BRS:
+0.1/–0.1 V— — — — 26.0 pJ 突触模拟 [140] ITO/NiSAs/
N-C/PVP/AuVCM BRS:
+0.7/–1.1 V103 100 >106 s 500 — 全加器 [99] Au/silk: AgNO3/Ag ECM TS:0.17 V 3 × 106 — 103 s 100 — 突触模拟 [147] Ag/MXene/Pt ECM BRS:
+1.33/–0.94 V>105 — 104 s 103 1~10 fJ ANN [149] Ag/a-COx/ta-C/Pt ECM BRS:1.5 V/–1.0 V — — 100 s — 6 nW — [155] Au/h-BN/Au VCM TS:0.1 V 107 40 >20000 s 500 — 逻辑门 [158] Ag/GeTe/MoTe2/Pt ECM BRS:
+0.15/–0.14 V102 — 104 s 105 ≈30 nJ 突触模拟 [162] Ag/SnS/Pt ECM BRS:
+0.2/–0.1 V108 1.5 105 s 104 100 fJ 图像分类 [73] 
DownLoad: CSV
-
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182]
DownLoad:
Catalog
Metrics
- Abstract views: 1631
- PDF Downloads: 139
- Cited By: 0