-
忆阻器驱动的神经形态计算芯片通过在硬件层面模拟生物突触的多维可塑性, 实现了高能效并行计算架构, 为类脑智能系统提供了新型硬件范式. 然而, 现有的大部分有机忆阻器在动态突触可塑性调控方面仍面临环境适应性不足的挑战. 本文提出了一种基于酞菁钴 (CoPc) 的双场调控忆阻器, 利用外加电场和温度协同耦合控机制, 在293—473 K的宽温域内展现出温度弹性特性, 临界电压随温度变化而自适应漂移 (3—9 V), 实现了跨尺度动态突触可塑性的高效调控. 在此基础上, 构建了由CoPc忆阻器阵列与深度学习模型集成的智能火灾预警系统, 有效保障了家用电热器的安全监测需求. 该研究不仅提出了环境自适应的忆阻器动态调控策略, 也为发展下一代鲁棒、高效的类脑神经形态计算平台奠定了物理与工程基础.Memristor-driven neuromorphic computing offers a promising path for brain-inspired intelligence by emulating the multidimensional plasticity of biological synapses, thereby achieving energy-efficient parallel computation. However, in the context of dynamically modulating synaptic plasticity, achieving strong environmental adaptability, especially in response to temperature fluctuation, remains a major challenge for organic memristors. In order to solve this problem, a bio-inspired cobalt phthalocyanine (CoPc)-based memristor is developed specifically for synergistic electric-thermal field modulation. The device utilizes the stable planar π-conjugated system of CoPc molecules and leverages dynamic oxygen vacancy (OV) migration at the CoPc/AlOx interface. A comprehensive electrical characterisation is conducted, incorporating X-ray photoelectron spectroscopy (XPS), in-situ Raman spectroscopy, and temperature-dependent electrical measurements across a wide range (293–473 K). This is supported by physical modelling (SCLC, FNT, Arrhenius) to elucidate the underlying mechanisms. Evidence indicates that the device can effectively replicate key aspects of synaptic plasiticy, including short-term potentiation/depression (STP/STD), and pairedpulse facilitation/depression (PPF/PPD), through the regulation of an electric field. The index increases to 151%, indicating a significant increase. Spike-amplitude-dependent plasticity (SADP, 45% weight increase), spike-timing-dependent plasticity (STDP, ΔW = ±90%), and learning-forgetting-relearning dynamics are revealed, unveiling cumulative memory effects linked to OV transport. The device exhibits excellent temperature resilience over the range of 293–473 K, characterised by a linear adaptive shift in its critical voltage (VCritical) from 8.7 V at 293 K to 4.5 V, with dVCritical/dT = 0.023 V/K. Physical analysis attributes this adaptive threshold and stable operation to a dual-field synergistic mechanism based on trap-assisted carrier transport. Elevated temperature thermally activates carriers, reducing the effective barrier for trap escape and OV migration activation energy (Ea = 0.073–0.312 eV), which facilitates conduction through Fowler-Nordheim tunneling (FNT) at lower electric fields. Conversely, lower temperatures require higher electric fields to enhance trap ionization efficiency through the Poole-Frenkel effect, compensating for reduced thermal energy. The validation of the linear VCritical-T relationship as a sensitive temperature transduction mechanism is achieved by developing an intelligent fire warning system. This study involves a 6 × 6 CoPc memristor array integrated into household heaters, combined with a deep learning model consisting of a fully connected network with 20 × 16 + 16 × 8 + 8 × 1 neurons. The resulting model achieves an accuracy of 96.54% in identifying high abnormal temperature. This work establishes a novel paradigm for environmentally adaptive neuromorphic devices through molecular/ interface design and synergistic multi-field modulation, providing a physical realization of temperature-elastic synaptic operation and demonstrating its practical feasibility for powerful next-generation brain-inspired computing platforms.
-
Keywords:
- memristor /
- dual-field modulation /
- critical voltage /
- dynamic synaptic plasticity
-
图 1 CoPc忆阻器性能 (a) 器件结构示意图; (b) 酞菁钴 (CoPc)的分子结构; (c) 器件横截面SEM图像; (d) CoPc薄膜AFM图像; (e) ITO/CoPc/Al忆阻器在0→–9.5→0 V(左)和0→9.5→0 V(右)的20个电压扫描周期的I-V曲线; (f) 从(e)中提取的增强(9.5 V)和抑制(–9.5 V)特性的电导
Fig. 1. Performance of CoPc memristor: (a) Schematic diagram of the device structure; (b) molecular structure of cobalt phthalocyanine (CoPc); (c) cross-sectional SEM image of the device; (d) AFM image of the CoPc film; (e) I-V curves of the CoPc memristor during 20 voltage-sweep cycles of 0→–9.5→0 V (left) and 0→9.5→0 V (right), respectively; (f) conductance of the potentiation (9.5 V) and depression (–9.5 V) characteristics extracted from (e).
图 2 CoPc忆阻器机理分析 (a) 器件面积-电流依赖特性; (b) 扫描速率-回滞面积依赖特性; 忆阻器高阻态的I-V曲线进行SCLC (c)和FNT (d)拟合; 忆阻器低阻态的I-V曲线进行SCLC (e)和FNT (f)拟合; 在HRS (g)和LRS (h)时从ITO电极收集的O 1s XPS能谱; 在不同深度采集的(i) Co 2p和(j) O 1s的XPS能谱; (k) 9.5 V电压刺激下的原位拉曼光谱; (l) 在不同温度(293—473 K)下连续施加9.5 V尖峰电压获得的电流曲线
Fig. 2. Mechanism analysis of CoPc memristor: (a) Cell size-current dependent characteristics; (b) scan speed-hysteresis area dependent characteristics; I-V curves of the HRS of the memristor fitted by SCLC (c) and FNT (d); I-V curves of the LRS of the memristor fitted by SCLC (e) and FNT (f); O 1s XPS spectra collected from the ITO electrode at HRS (g) and LRS (h); XPS energy spectra of (i) Co 2p and (j) O 1s collected at different depths; (k) in situ Raman spectra under 9.5 V voltage stimuli; (l) current obtained by continuously applying spike voltages of 9.5 V at different temperatures (293–473 K).
图 3 CoPc忆阻器的突触可塑性 (a) 短时程增强和(b) 短时程抑制曲线(50个电压脉冲刺激, VPos = 9.5 V, VNeg = –9.5 V, VRead = 1 V, Δt = 430 ms); (c) 配对脉冲异化PPF和(d) 配对脉冲抑制 PPD指数曲线, 插图为对应的电流曲线; (e) 脉冲幅值 (7, 8, 9.5 V) 依赖可塑性曲线; (f) 脉冲频率(2.30, 1.15, 0.769, 0.577, 0.462 Hz)依赖可塑性曲线; (g) 尖峰时序依赖可塑性曲线; (h) 归一化的增强(VPos = 9.5 V)和抑制(VNeg = 5.5 V)曲线; (i) 学习-遗忘-再学习曲线
Fig. 3. Synaptic plasticity of CoPc memristor. STP (a) and STD (b) curves under 50 voltage pulses (VPos = 9.5 V, VNeg = –9.5 V, VRead = 1 V, Δt = 430 ms); PPF (c) and PPD (d) index curves, with insets showing the corresponding current curves; (e) spike amplitude (7, 8, and 9.5 V) dependent plasticity curves; (f) spike frequency (2.30, 1.15, 0.769, 0.577, 0.462 Hz) dependent plasticity curves; (g) STDP curves; (h) normalized potentiation (VPos = 9.5 V) and depression (VNeg = 5.5 V) curves; (i) learning-forgetting-relearning curves.
图 4 CoPc忆阻器的温度梯度调控特性 CoPc忆阻器依次在(a) 温度为293 K、刺激电压为9 V; (b) 333 K, 8 V; (c) 373 K, 7 V; (d) 423 K, 6 V; (e) 473 K, 5 V下的短时程可塑性; (f) 不同电压(6, 7, 8和9 V)下的LRS电阻随1000/T变化的Arrhenius模型; 在不同温度(g) 293 K, (h) 333 K, (i) 373 K, (j) 423, (k) 473 K的测试环境下器件50个临界电压的概率分布; (l) 在5个温度(293, 333, 373, 423和473 K)与5个电压(5, 6, 7, 8和9 V)下, 总计25种工作状态下的100 s刺激后的稳态电流; (m) 家用电热器存在的火灾隐患示意图; (n) 基于CoPc忆阻器的智能预警装置示意图; (o) 在100次训练后对异常状态识别正确率
Fig. 4. Temperature gradient control characteristics of CoPc memristor: STP of the CoPc memristor at (a) a temperature of 293 K and an amplitude of 9 V; (b) 333 K and 8 V; (c) 373 K and 7 V; (d) 423 K and 6 V; (e) 473 K and 5 V in sequence; (f) Arrhenius model of LRS resistance varying with 1000/T at different voltages (6, 7, 8, and 9 V); the probability distribution of 50 critical voltages of the device under the different temperatures of (g) 293 K, (h) 333 K, (i) 373 K, (j) 423 (k) and 473 K; (1) steady-state currents after 100 s of stimulation at five temperatures (293, 333, 373, 423, and 473 K) with five voltages (5, 6, 7, 8, 9 V) for a total of 25 operating states; (m) schematic diagram of fire hazards existing in household electric heaters; (n) schematic diagram of the intelligent early warning device based on CoPc memristor; (o) the accuracy rate in identifying abnormal states after 100 training epoch.
表 1 文献中报告的温度弹性忆阻器工作温度范围、动态可调性、功耗、识别准确率总结
Table 1. Summary of the operating temperature range, dynamic adjustability, power consumption and recognition accuracy of temperature-elastic memristors reported in the literature.
Active
layerWorking
temperature/KMaterial Dynamic
modulationRecognition
accuracy rate/%Power
consumptionRef CoPc 293–473 Organic Yes 96.54 1.95 nJ This work SF:Ca2+ 300 Organic No — 2.25 μJ [59] TaOx 298–418 Inorganic No 100 — [60] CIGSe 623 Inorganic No 90 — [61] TaO1.8/TaO2.7/TaO1.8 300–343 Inorganic No — — [62] AlFeO3 298–413 Inorganic No — — [63] Chitosan/PNIPAM 290–410 Organic Yes — 0.23 mJ [64] Co-TCPP 288–313 Organic No 93.95 — [65] CsPbBr2I 300–513 Inorganic No — — [66] VO2 294–315 Inorganic No 98.1 3.9 nJ [67] VOPc 303–373 Organic Yes — — [68] 
下载: 导出CSV
-
[1] Carlos E, Branquinho R, Martins R, Kiazadeh A, Fortunato E 2020 Adv. Mater. 33 2004328
[2] Hong X, Loy D J, Dananjaya P A, Tan F, Ng C, Lew W 2018 J. Mater. Sci. 53 8720
Google Scholar
[3] Liao K H, Lei P X, Tu M L, Luo S W, Jiang T, Jie W J, Hao J H 2021 ACS Appl. Mater. Interfaces 13 32606
Google Scholar
[4] Babacan Y, Kaçar F 2017 AEU - Int. J. Electron. Commun. 73 16
Google Scholar
[5] Campbell K A, Drake K T, Barney Smith E H 2016 Front. Bioeng. Biotechnol. 4 97
[6] Daddinounou S, Vatajelu E I 2024 Front. Neurosci. 18 1387339
Google Scholar
[7] Go S X, Lim K G, Lee T H, Loke D K 2024 Small Sci. 4 2300139
Google Scholar
[8] Li Z Y, Li Z S, Tang W, Yao J P, Dou Z P, Gong J J, Li Y F, Zhang B N, Dong Y X, Xia J, Sun L, Jiang P, Cao X, Yang R, Miao X S, Yang R G 2024 Nat. Commun. 15 7275
Google Scholar
[9] Lu J L, Sun F, Zhou G D, Duan S K, Hu X F 2024 IEEE Sens. J. 24 2967
Google Scholar
[10] Zhang J L, Li X J, Xiao P D, Wei Z M, Hong Q H 2024 IEEE Trans. Circuits Syst. I Regular Papers 71 3228
Google Scholar
[11] Huang Y F, Hopkins R, Janosky D, Chen Y C, Chang Y F, Lee J C 2022 IEEE Trans. Electron Devices 69 6102
Google Scholar
[12] Li J Y, Qian Y Z, Li W, Yu S C, Ke Y X, Qian H W, Lin Y H, Hou C H, Shyue J J, Zhou J, Chen Y, Xu J P, Zhu J T, Yi M D, Huang W 2023 Adv. Mater. 35 2209728
Google Scholar
[13] Naqi M, Yu Y, Cho Y, Kang S, Khine M T, Lee M, Kim S 2024 Mater. Today Nano 27 100491
Google Scholar
[14] Fan Z Y, Tang Z H, Fang J L, Jiang Y P, Liu Q X, Tang X G, Zhou Y C, Gao J 2024 Nanomaterials 14 583
Google Scholar
[15] He Y, Farmakidis N, Aggarwal S, Dong B, Lee J S, Wang M, Zhang Y, Parmigiani F, Bhaskaran H 2024 Nano Lett. 24 16325
Google Scholar
[16] Zhu Y B, Wu C X, Xu Z W, Liu Y, Hu H L, Guo T L, Kim T W, Chai Y, Li F S 2021 Nano Lett. 21 6087
Google Scholar
[17] Haghshenas Gorgabi F, Morant-Miñana M C, Zafarkish H, Abbaszadeh D, Asadi K 2023 J. Mater. Chem. C 11 1690
Google Scholar
[18] Hajtó D, Rák Á, Cserey G 2019 Materials 12 3573
Google Scholar
[19] Li H Z, Gao Q, Gao J, Huang J S, Geng X L, Wang G X, Liang B, Li X H, Wang M, Xiao Z S, Chu P K, Huang A P 2023 ACS Appl. Mater. Interfaces 15 46449
Google Scholar
[20] Sun Y M, Li B X, Liu M, Zhang Z K 2024 Mater. Today Adv. 23 100515
Google Scholar
[21] Tian L, Wang Y Y, Shi L P, Zhao R 2020 ACS Appl. Electron. Mater. 2 3633
Google Scholar
[22] 刘东青, 程海峰, 朱玄, 王楠楠, 张朝阳 2008 物理学报 63 187301
Liu D Q, Cheng H F, Zhu X, Wang N N, Zhang C Y 2014 Acta Phys. Sin. 63 187301
[23] Baranowski M, Sachser R, Marinković B P, Ivanović S D, Huth M 2022 Nanomaterials 12 4145
Google Scholar
[24] Mayer S F, Mitsioni M, van den Heuvel L, Robin P, Ronceray N, Marcaida M J, Abriata L A, Krapp L F, Anton J S, Soussou S, Jeanneret-Grosjean J, Fulciniti A, Moller A, Vacle S, Feletti L, Brinkerhoff H, Laszlo A H, Gundlach J H, Emmerich T, Dal Peraro M, Radenovic A 2025 BioRxiv 26 615172
[25] Rodriguez N, Maldonado D, Romero F J, Alonso F J, Aguilera A M, Godoy A, Jimenez-Molinos F, Ruiz F G, Roldan J B 2019 Materials 12 3734
Google Scholar
[26] Wang S X, Dong X Q, Xiong Y X, Sha J, Cao Y G, Wu Y P, Li W, Yin Y, Wang Y C 2021 Adv. Electron. Mater. 7 2100014
Google Scholar
[27] Xu G H, Zhang M L, Mei T T, Liu W C, Wang L, Xiao K 2024 ACS Nano 18 19423
[28] Li J Y, Qian Y Z, Ke Y X, Li W, Huang W, Yi M D 2023 ACS Appl. Electron. Mater. 5 6813
Google Scholar
[29] Zhou J, Li W, Chen Y, Lin Y H, Yi M D, Li J Y, Qian Y Z, Guo Y, Cao K Y, Xie L H, Ling H F, Ren Z J, Xu J P, Zhu J T, Yan S K, Huang W 2020 Adv. Mater. 33 2006201
[30] Wang Z Y, Wang L Y, Wu Y M, Bian L Y, Nagai M, Jv R, Xie L H, Ling H F, Li Q, Bian H Y, Yi M D, Shi N E, Liu X G, Huang W 2021 Adv. Mater. 33 2104370
Google Scholar
[31] Pasquini C, D’Amario L, Zaharieva I, Dau H 2020 J. Chem. Phys. 152 19
[32] Bian L Y, Xie M, Chong H, Zhang Z W, Liu G Y, Han Q S, Ge J Y, Liu Z, Lei Y, Zhang G W, Xie L H 2022 Chin. J. Chem. 40 2451
Google Scholar
[33] Chen J B, Jia S J, Gao L Y, Xu J W, Yang C Y, Guo T T, Zhang P, Chen J T, Wang J, Zhao Y, Zhang X Q, Li Y 2024 Colloids Surf. A 689 133673
Google Scholar
[34] Kim H, Kim M, Lee A, Park H L, Jang J, Bae J H, Kang I M, Kim E S, Lee S H 2023 Adv. Sci. 10 2300659
Google Scholar
[35] Qian Y Z, Li J Y, Li W, Hou C H, Feng Z Y, Shi W, Yi M 2024 J. Mater. Chem. C 12 9669
Google Scholar
[36] Shakib M A, Gao Z, Lamuta C 2023 ACS Appl. Electron. Mater. 5 4875
Google Scholar
[37] Zeng J M, Chen X H, Liu S Z, Chen Q L, Liu G 2023 Nanomaterials 13 803
Google Scholar
[38] 邵楠, 张盛兵, 邵舒渊 2019 物理学报 68 214202
Google Scholar
Shao N, Zhang S B, Shao S Y 2019 Acta Phys. Sin. 68 214202
Google Scholar
[39] 贡以纯, 明建宇, 吴思齐, 仪明东, 解令海, 黄维, 凌海峰 2024 物理学报 73 207302
Google Scholar
Gong Y C, Ming J Y, Wu S Q, Yi M D, Xie L H, Huang W, Ling H F 2024 Acta Phys. Sin. 73 207302
Google Scholar
[40] Hu W J, Fan Z, Mo L Y, Lin H P, Li M X, Li W J, Ou J, Tao R Q, Tian G, Qin M H, Zeng M, Lu X B, Zhou G F, Gao X S, Liu J-M 2025 ACS Appl. Mater. Interfaces 17 9595
Google Scholar
[41] Ki S, Chen M Z, Liang X G 2023 IEEE Nanotechnol. Mag. 17 24
Google Scholar
[42] Yun S, Mahata C, Kim M H, Kim S 2022 Appl. Surf. Sci. 579 152164
Google Scholar
[43] Cantley K D, Subramaniam A, Stiegler H J, Chapman R A, Vogel E M 2012 IEEE Trans. Neural Netw. Learn. Syst. 23 565
Google Scholar
[44] Chen Z J, Zhang J D, Wen S C, Li Y, Hong Q H 2021 IEEE Trans. Very Large Scale Integr. Syst. 29 1095
Google Scholar
[45] Lee Y P 2017 IEEE Trans. Syst. Man Cybern. Syst. 47 3386
Google Scholar
[46] Li M, Hong Q H, Wang X P 2021 Neural Comput. Appl. 34 319
[47] Patel M, Gosai J, Patel P, Roy M, Solanki A 2024 ACS Omega 9 46841
Google Scholar
[48] Bing Z S, Baumann I, Jiang Z Y, Huang K, Cai C X, Knoll A 2019 Front. Neurorobot. 13 18
Google Scholar
[49] Pedroni B U, Joshi S, Deiss S R, Sheik S, Detorakis G, Paul S, Augustine C, Neftci E O, Cauwenberghs G 2019 Front. Neurosci. 13 357
Google Scholar
[50] Quintana F M, Perez-Peña F, Galindo P L 2022 Neural Comput. Appl. 34 15649
Google Scholar
[51] Mikaitis M, Pineda García G, Knight J C, Furber S B 2018 Front. Neurosci. 12 105
Google Scholar
[52] 王美丽, 王俊松 2015 物理学报 64 108701
Google Scholar
Wang M L, Wang J S 2015 Acta Phys. Sin. 64 108701
Google Scholar
[53] Wang Z X, Yu N G, Liao Y S 2023 Electronics 12 3992
Google Scholar
[54] Kim S I, Lee Y, Park M H, Go G T, Kim Y H, XU W T, Lee H D, Kim H, Seo D G, Lee W, Lee T W 2019 Adv. Electron. Mater. 5 1900008
Google Scholar
[55] Frenkel J 1938 Phys. Rev. 54 647
Google Scholar
[56] He D W, Qiao J S, Zhang L L, Wang J Y, Lan T, Qian J, Li Y, Shi Y, Chai Y, Lan W, Ono L K, Qi Y B, Xu J B, Ji W, Wang X R 2017 Sci. Adv. 3 e1701186
Google Scholar
[57] Takagi K, Nagase T, Kobayashi T, Naito H 2016 Org. Electron. 32 65
Google Scholar
[58] Sturman B, Podivilov E, Gorkunov M 2003 Phys. Rev. Lett. 91 176602
Google Scholar
[59] Xie Y L, Kundu S C, Fan S, Zhang Y P 2024 Sci. China Mater. 67 3675
Google Scholar
[60] Cao L J, Luo Y H, Yao J P, Ge X, Luo M Y, Li J Q, Cheng X M, Yang R, Miao X S 2024 J. Mater. Chem. C 12 19555
Google Scholar
[61] Guo T, Ge J W, Jiao Y X, Teng Y C, Sun B, Huang W, Asgarimoghaddam H, Musselman K P, Fang Y, Zhou Y N, Wu Y A 2023 Mater. Horiz. 10 1030
Google Scholar
[62] Li J C, Liu Z C, Xia Y H, Liu X, Yang H X, Ma Y X, Wang Y L 2025 Adv. Funct. Mater. 35 2416635
Google Scholar
[63] Ganaie M M, Kumar A, Shringi A K, Sahu S, Saliba M, Kumar M 2024 Adv. Funct. Mater. 34 2405080
Google Scholar
[64] Sun Y M, Liu M, Li B X 2024 Small 20 2404177
Google Scholar
[65] Ouyang G, Wang Y L, Su J, Ren M C, Zhang M H, Cao M H 2025 Nano Energy 137 110778
Google Scholar
[66] Liu Z H, Cheng P P, Li Y F, Kang R Y, Zhang Z Q, Zuo Z Y, Zhao J 2021 ACS Appl. Mater. Interfaces 13 58885
Google Scholar
[67] Li Z Y, Li Z S, Tang W, Yao J P, Dou Z P, Gong J J, Li Y F, Zhang B N, Dong Y X, Xia J, Sun L, Jiang P, Cao X, Yang R, Miao X S, Yang R G 2024 Nat. Commun. 15 7275
Google Scholar
[68] Shen D H, Zhou J, Chen Y, Kong L J, Li W, Shi W, Yi M D 2025 J. Phys. D: Appl. Phys. 58 245107
Google Scholar
下载:
计量
- 文章访问数: 299
- PDF下载量: 6
- 被引次数: 0
