Activity Dependent Synaptic Plasticity Mimicked on IndiumâTinâOxide

Subscriber access provided by UNIVERSITY OF THE SUNSHINE COAST

Article

Activity Dependent Synaptic Plasticity Mimicked on Indium-Tin-Oxide Electric-Double-Layer Transistor Juan Wen, Li Qiang Zhu, Yang Ming Fu, Hui Xiao, Liqiang Guo, and Qing Wan ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b13215 • Publication Date (Web): 04 Oct 2017 Downloaded from http://pubs.acs.org on October 4, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 18

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

ACS Applied Materials & Interfaces

Activity Dependent Synaptic Plasticity Mimicked on

Indium-Tin-Oxide

Electric-Double-Layer

Transistor Juan Wen1, 2, 4, Li Qiang Zhu1, 4, * Yang Ming Fu1, 4, Hui Xiao1, 4, Li Qiang Guo2, Qing Wan3, * 1) Key Laboratory of Graphene Technologies and Applications of Zhejiang Province, Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, Zhejiang, People’s Republic of China 2) Micro/Nano Science & Technology Center, Jiangsu University, Zhenjiang, 212013, Peoples Republic of China 3) School of Electronic Science & Engineering, Nanjing University, Nanjing 210093, Jiangsu, Peoples Republic of China 4) University of Chinese Academy of Sciences, Beijing 100049, Peoples Republic of China

ABSTRACT: Ion coupling has provided an additional method to modulate electric properties for solid-state materials. Here, phosphorosilicate glass (PSG)-based electrolyte gated protonic/electronic coupled indium-tin-oxide electric-double-layer (EDL) transistors are fabricated. The oxide transistor exhibits good electrical performances due to an extremely strong proton gating behavior for the electrolyte. With interfacial electrochemical doping, channel conductances of the oxide EDL transistor can be regulated to different levels, corresponding to different initial synaptic weights. Thus, activity dependent synaptic responses, such as excitatory post-synaptic current, paired-pulse facilitation and high-pass filtering, are discussed in detail. The proposed proton conductor gated oxide EDL synaptic

1 ACS Paragon Plus Environment


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

transistors with activity dependent synaptic plasticities may act as fundamental building blocks for neuromorphic system applications.

Keywords: Protonic/electronic coupling, Synaptic transistor, Synaptic activity, Initial synaptic weight, High-pass filtering


Page 2 of 18

Page 3 of 18

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


1. INTRODUCTION Conventionally, ionic contamination in Si-based integrated circuits would degrade electrical behaviors. However, the figure of merit is different for bio-chemical sensors and bionic electronics. Ions are endowed with new intention in recent advancements of condensed matters. As early as 1950s, electrolytes were used in germanium based transistors.1 In 1980s, several works were reported on electrolyte-gated transistors (EGTs).2,

3

Ion coupling has

provided an additional method to modulate electric properties for solid-state materials.4-6 Moreover, phase transformations from one to others are expected by ion transfer under external electric field, which could greatly enrich material functionality.7 In ionic/electronic hybrid electrolyte gated transistors (EGTs), the movements of ions within electrolytes can alter carrier transport characteristics of condensed materials greatly. Basically, there are two types of operation mechanism for EGTs. For EGTs operating in electrostatic modulation mode, charged ions within electrolyte will move towards the electrode, inducing an electricdouble-layer (EDL) at the interface.8-10 For EGTs operating in electrochemical modulation mode, charged ions within electrolyte will penetrate into semiconductor channel, resulting in an electrochemical doping of channel.11 Ions also act as a vital role in nervous systems. Our brain consists of ~1011 neurons and ~1015 synapses.12 Cognitive behaviors of our brain are realized through ion migrations in nervous systems, which involves physical alterations in neuronal substrates that modulate neuron activities and communications. 13-15 Such ionic fluxes related information storages and retrievals occur by re-wiring the neuronal networks. Mixed ionic-electronic conductor based resistive switching memories and atomic switches exhibit unique ionic/electronic hybrid behavior. Thus, important synaptic responses and neuromorphic computations have been demonstrated recently, including paired pulse facilitation (PPF), spiking time dependent plasticity (STDP), short-term memory to long-term memory transition, pattern recognition, etc.16-18 Field-effect transistors have also been creatively proposed for emulating biological 3 ACS Paragon Plus Environment


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

synapses, including hybrid nanoparticle organic memory field-effect transistors19 and ferroelectric-gate field-effect transistor.20-22 It is interesting to note here that ion-related electrostatic and electrochemical processes in EGTs are quite similar to ion migration related information processings in synapses. Thus, EGTs can act as ideal artificial synapses.23 In fact, some important synaptic responses were demonstrated,24-29 including short term plasticity, dynamic filtering, spatiotemporal dynamic logic, etc. In nervous system, prior history of synaptic activity can significantly affect synaptic state.

30

It is a form of functional and

structural neuroplasticity that arises from cognitive functions and personal experiences. It contributes to several neural activities, including learning and memory.31 Thus, mimicking of activity dependent synaptic plasticity on solid-state ionic/electronic hybrid transistor is also meaningful for neuromorphic platforms. Here, indium-tin-oxide (ITO) electric-double-layer (EDL) synaptic transistors were made. Initial synaptic weights were set to different levels by interfacial proton electrochemical doping. Then, influences of initial synaptic weights on short-term synaptic plasticities were demonstrated. The proposed oxide synaptic transistors may act as fundamental building blocks for neuromorphic system applications.

2. EXPERIMENTAL PROCEDURE First, ~2.1µm thick nanogranular phosphorosilicate glass (PSG) films were deposited on ITO glass substrate by plasma-enhanced chemical vapor deposition (PECVD).23 Then, patterned ITO source and drain electrodes with self-assembled ITO channel were deposited on the PSG based electrolyte by sputtering. Figure 1a schematically shows the PSG gated ITO transistor. Channel length (L) and width (W) are ~80µm and ~1000µm, respectively. Figure 1b schematically shows a biological synapse. When a nerve impulse arrives at pre-synapse, voltage gated ions channel will open. Thus, ions will migrate across the channel, resulting in the release of neurotransmitters into synaptic cleft. Such neurotransmitters will combine with receptors on post-synapse, causing the movements of ions across the post-synaptic membrane. 4 ACS Paragon Plus Environment

Page 4 of 18

Page 5 of 18

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


Thus, an action potential will be triggered in post-synapse. Therefore, the migration of ions plays a critical role in adjusting synaptic weight and transporting information. Correspondingly, ITO bottom gate, ITO channel with source/drain electrodes and channel current are deemed as pre-synapse, post-synapse and synaptic weight, respectively. Thus, an artificial synapse is obtained. With a relatively high gate bias, interfacial electrochemical doping will be triggered, which alters channel conductivities. As a result, initial synaptic weight can be modulated correspondingly. Proton conductivities and protonic/electronic coupling behaviors of the PSG film were analyzed with an impedance analyzer. Electrical characterizations of the proposed oxide EDL synaptic transistor were carried out with a semiconductor parameter analyzer in air ambient with a relative humidity of 55%.

Figure 1. (a) PSG based electrolyte gated ITO EDL synaptic transistor. (b) Schematic draw of a biological synapse.

3. Results and discussions Figure 2a illustrates specific capacitance of PSG based electrolyte film as a function of frequencies. An extremely high electric-double-layer (EDL) capacitance of ~15.7 µF/cm2 is obtained at 1.0 Hz. Figure 2b shows a typical impedance spectroscopy. Impedance real value (R) is estimated to be ~632 Ω with impedance imaginary value equal to zero. Thus, proton conductivity (σ) of ~2.3×10-4S/cm is obtained. Figure 2c shows output curves of the ITO EDL transistor, exhibiting a strong saturation and a clear pinch off behavior. Figure 2d shows 5 ACS Paragon Plus Environment


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

transfer characteristics (Ids vs Vgs) with a constant Vds of 1.5 V. An anticlockwise hysteresis window of ~0.1 V is observed, attributing to the mobile protons within PSG electrolyte.32 The ITO synaptic transistor exhibits good performances, including a subthreshold swing of ~87 mV/dec, an on/off ratio of ~1.2×107 and a mobility of ~3.2 cm2/Vs.

Figure 2. (a) Specific capacitance of the PSG electrolyte as a function of frequencies. Inset: Schematic image of a testing structure. (b) Impedance spectroscopy of the PSG electrolyte film. (c) Output and (d) Transfer characteristics of the oxide EDL transistor.

Figure 3a illustrates a dynamic channel current response recorded at a fixed Vds of 1.5 V. Initial channel current is ~30 nA. With a gate spike of (0.5 V, 10 ms), a peak channel current of ~77 nA is obtained, followed by a slow decay. Such transient characteristic is analogous to EPSC response in nervous system.33 For a biological synapse, synapse efficacy (or synaptic weight) reflect connection-strength between neurons. Here, the initial channel current reflects 6 ACS Paragon Plus Environment

Page 6 of 18

Page 7 of 18

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


the initial synaptic weight. Basically, there are two types of operation mechanism for PSG gated oxide EDL synaptic transistors. For transistor operating in electrochemical mode, protons within PSG electrolyte will migrate across the ITO channel/PSG interface and come into the ITO channel. Since protons are effective donor in ITO, an increased channel conductivity is observed. In our case, electrochemical doping is triggered by gate voltage spikes with amplitudes ranged between 3 V and 5 V.34, 35 Thus, the channel currents have been tuned to another four levels with values of ~50 nA, ~100 nA, ~200 nA and ~350 nA. Thus, five initial synaptic weights were obtained. The modulation of initial synaptic weight implies the realization of long-term plasticity. For transistor operating in electrostatic modulation mode, protons within PSG film will accumulate at the ITO channel/PSG interface when a positive gate spike is applied, forming an EDL and resulting in an increased channel current. When the spike finishes, protons will drift back to equilibrium position, inducing the decay process of EPSC. It is reported that a stretched exponential relaxation mode can be used for explaining ion hopping transport in condensed materials.36 Similarly, decayed EPSC curves can be fitted by a relation below:



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

Figure 3 (a) A dynamic channel current response of the oxide EDL transistor. Inset: Retention time as a function of initial synaptic weight. (b) Spike width-dependent and (c) Spike amplitude-dependent peak EPSC values at different initial synaptic weights.

I EPSC = ( I EPSC − Peak − I EPSC −∞ ) exp[−(

t − t0

τ

) β ] + I EPSC −∞

where τ, t0, IEPSC-∞ and β are retention time, time when the spike is finished, resting current and stretch index ranged between 0 and 1. As shown in Figure 3a, the SEF fitting curve approaches the EPSC decay curve very well. EPSC was also trigged by the same spike at different initial synaptic weight. Inset in Figure 3a illustrates τ values as a function of initial synaptic weights. It is observed that the τ value decreases from ~52 ms to ~20 ms for initial synaptic weight increasing from ~30 nA to ~350 nA. The behaviors are related to EDL effects.37 Positive gate spike results in the formation of an EDL layer at ITO 8 ACS Paragon Plus Environment

Page 8 of 18

Page 9 of 18

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


channel/electrolyte interface. When the spike finishes, the EDL layer will wear off due to the diffusion back of the protons originated from the concentration gradient. The positive drain bias will facilitate the process. At low initial synaptic weight, the channel resistance is relatively high. Correspondingly, contract resistance between the drain electrode and the channel is relatively high. Under a constant drain voltage (Vds), effective drain voltage will decrease. Thus, such effect gets weaken. Therefore, higher τ values are obtained at lower initial synaptic weights. We also found that retention time increases with the decreased Vds at a same gate voltage spike (results not shown here). The results provide additional evidences that the drain bias will promote such diffusion. Figure 3b illustrates spike width-dependent peak EPSC values at different initial synaptic weights. Spike amplitude is fixed at 0.5 V. For initial synaptic weight at ~30 nA, peak EPSC value increases from ~77 nA to ~1.9 µA for spike width increasing from 10ms to 200ms. With the increased initial synaptic weight, peak EPSC value increases too. Figure 3c illustrates spike amplitude-dependent peak EPSC values at different initial synaptic weights. Spike width is fixed at 10ms. For initial synaptic weight at ~30nA, peak EPSC value increases from ~77nA to ~1.5µA for spike amplitude increasing from 0.5V to 2.0V. With the increased initial synaptic weight, peak EPSC value increases too. The results here indicate that the initial synaptic weight has significant influences on synaptic plasticity of the proposed oxide EDL synaptic transistor. In neuroscience, neural facilitation, also known as paired pulse facilitation (PPF), is a dynamic enhancement of transmitter and reflects the receptivity of biological synapses for information processing.38-40 The neural facilitation is involved in several neural activities, such as learning and information processing.41 In our case, synaptic activity dependent neural facilitation can be demonstrated by applying temporally correlated spikes at different initial synaptic weights. Figure 4a illustrates a typical PPF response. The gate spike has an amplitude of 0.5V and a duration time of 20ms. The interval time (∆t) is 300 ms. Reading 9 ACS Paragon Plus Environment


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

voltage is fixed at 1.5 V. Initial synaptic weight is ~30 nA. The first (A1) and the second absolute EPSC (A2) are ~110 nA and ~130 nA, respectively. Therefore, PPF index (A2/A1×100%) is estimated to be ~118%. Figure 4b shows ∆t dependent PPF indexes at different initial synaptic weights. For initial synaptic weight at ~30nA, PPF index decreases from ~235% to ~115% for ∆t value increasing from 30 ms to 300 ms. While for initial synaptic strength at ~350 nA, PPF index decreases from ~152% to ~100% for ∆t value increasing from 30ms to 300 ms. In principle, A2 should be higher than or close to A1. However, due to the existences of experimental errors, A2/A1 will scatter at values of ~100% with larger interval time. That’s the reason why some PPF indexes are below 100% at interval time of 300ms. Interestingly, PPF Index decreases with the increased initial synaptic weight. Such behaviors can be explained as follows. Retention time reflects characteristic time required for protons to drift back to original equilibrium position, which influences effective residual proton concentration contributed to the second EPSC. If ∆t is much higher than the retention time, the protons extracted by the previous spike will have enough time to drift back. Thus, the followed EPSC value will be very close to the previous one. Therefore, the variation in PPF index at higher ∆t value will be very small. If ∆t is close to or shorter than the retention time, some protons extracted by the previous spike will have no enough time to drift back when the followed spike comes. Moreover, the number of such accumulated protons will be higher for larger retention time. As discussed above, higher retention time (τ) is observed for lower initial synaptic weight at the same spike condition, which means higher opportunity for the triggered protons to reside at channel/PSG interface. Thus, more protons will get accumulated when the followed spike comes. Therefore, higher PPF index is expected for lower initial synaptic weight.


Page 10 of 18

Page 11 of 18

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


Figure 4. (a) A typical PPF response for the oxide EDL synaptic transistor. (b) ∆t dependent PPF index with initial synaptic weight at 30nA, 50nA, 100nA, 200nA and 350nA.

Short-term synaptic plasticity endows biological synapse possessing ability of signal decoding and calculation.42 Short-term synaptic depression contribute to low-pass temporal filtering, while short-term synaptic facilitation contribute to high-pass temporal filtering. In our case, a high-pass temporal filtering behavior is expected for the proposed oxide EDL synaptic transistor because of the PPF behaviors. Figure 5a illustrates a typical EPSC response triggered with sequential spikes (0.5 V, 20 ms) at different frequencies. The initial synaptic weight is ~30nA. Because of short-term facilitations, EPSC amplitude increases with the increased number of stimulus. Moreover, EPSC amplitude increases with the increased stimulus frequency. Inset in Figure 5a illustrates the amplified EPSC response triggered at 10 Hz. The first (A1) and the fifth absolute EPSC (A5) are ~100nA and ~260nA, respectively. Therefore, an EPSC gain (A5/A1) of ~2.6 is obtained. Figure 5b illustrates frequency dependent EPSC gains at different initial synaptic weights. With the increased frequency, the EPSC gain increases correspondingly. For initial synaptic weight at ~30 nA, EPSC gain value increases from ~1.1 to ~7.3 for stimulus frequency increasing from 2 Hz to 33 Hz. With the increased initial synaptic weight, the EPSC gain decreases correspondingly. To establish a high-pass filter, the oxide EDL synaptic transistor can be connected to a neuron circuit which sets a threshold. When EPSC gain is above the threshold, the neuron circuit would let signals 11 ACS Paragon Plus Environment


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

Page 12 of 18

pass. On the contrary, the neuron circuit would not let signals pass. Thus, a cutoff frequency (fcut) can be determined by a certain threshold. Interestingly, the fcut values can be regulated by tuning the initial synaptic weight. As shown in Figure 5b, if the threshold is set at 2.0, the fcut value is estimated to be ~6 Hz for initial synaptic weight at ~30 nA. While for initial synaptic weight at ~350 nA, the fcut value is estimated to be ~20 Hz.

Figure 5. (a) A typical EPSC response triggered with sequential spikes (0.5 V, 20 ms) at different frequencies. Inset: Amplified EPSC current triggered at 10 Hz. (b) Frequency dependent EPSC gains at different initial synaptic weights.

In nervous system, prior history of synaptic activity can significantly affect synaptic state.30 In another word, synaptic weight is governed by activity dependent plasticity behaviors regarded as essential mechanisms for information processing. It is reported that the major

forms

of

activity

dependent

synaptic

plasticity

include

short-term

potentiation/depression, long-term potentiation/depression, spiking-time-dependent plasticity, etc.43 Moreover, activity dependent synaptic plasticity may also contribute to classical condition44. Thus, mimicking activity dependent synaptic plasticity on a single solid-state synaptic device is interesting for neuromorphic platforms. The proposed oxide EDL synaptic transistors may act as potential building blocks for neuromorphic system applications.


Page 13 of 18

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


4. CONCLUSION In summary, phosphorosilicate glass (PSG)-based electrolyte gated protonic/electronic coupled indium-tin-oxide electric-double-layer transistors were fabricated. The oxide electricdouble-layer (EDL) transistor exhibits good electrical performances. Furthermore, the oxide EDL transistor can be electrochemically regulated to different initial synaptic weights. Thus, activity dependent synaptic responses were discussed in detail, such as excitatory postsynaptic current, paired-pulse facilitation and high-pass filtering. Moreover, cutoff frequency of the proposed high-pass filter can be modulated by setting different initial synaptic weight. The proposed oxide synaptic transistors may act as potential building blocks for neuromorphic system applications.

AUTHOR INFORMATION

Corresponding Author *E-mail address: [email protected], [email protected]. Notes The authors declare no competing financial interest.

ACKNOLEDGEMENTS This work was supported by Zhejiang Provincial Natural Science Foundation of China,

Ningbo Science and Technology Innovation Team (2016B10005), Youth Innovation Promotion Association CAS (2014259) and Key Research Program of Frontier Sciences, Chinese Academy of Sciences (No.QYZDB-SSW-JSC047).

REFERENCES

(1) Brattain, W. H.; Garrett, C. G. B., Electrical Properties of the Interface Between a Germanium Single Crystal and an Electrolyte. Phys. Rev. 1954, 94, 750. 13 ACS Paragon Plus Environment


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

(2) White, H. S.; Kittlesen, G. P.; Wrighton, M. S. Chemical Derivatization of an Array of Three Gold Microelectrodes with Polypyrrole: Fabrication of a Molecule-Based Transistor. J. Am. Chem. Soc. 1984, 106, 5375-5377. (3) Tardella, A.; Chazalviel, J. N. Highly Accumulated Electron Layer at a Semiconductor Electrolyte Interface. Phys. Rev. B 1985, 22, 2439-2448. (4) Ueno, K.; Nakamura, S.; Shimotani, H.; Ohtomo, A.; Kimura, N.; Nojima, T.; Aoki, H.; Iwasa, Y.; Kawasaki, M. Electric-Field-Induced Superconductivity in an Insulator. Nat. Mater. 2008, 7, 855-858. (5) Ueno, K.; Nakamura, S.; Shimotani, H.; Yuan, H. T.; Kimura, N.; Nojima, T.; Aoki, H.; Iwasa, Y.; Kawasaki, M. Discovery of Superconductivity in KTaO3 by Electrostatic Carrier Doping. Nat. Nanotechnol. 2011, 6, 408-411. (6) Li, L. J.; O’Farrell, E. C. T.; Loh, K. P.; Eda, G.; Ozyilmaz, B.; Castro Neto, A. H. Controlling Many-Body States by the Electric-Field Effect in a Two-Dimensional Material. Nature 2016, 529, 185-189. (7) Lu, N. P.; Zhang, P. F.; Zhang, Q. H.; Qiao, R. M.; He, Q.; Li, H. B.; Wang, Y. J.; Guo, J. W.; Zhang, D.; Duan, Z.; Li, Z. L.; Wang, M.; Yang, S. Z.; Yan, M. Z.; Arenholz, E.; Zhou, S. Y.; Yang, W. L.; Gu, L.; Nan, C. W.; Wu, J.; Tokura, Y.; Yu, P.; Electric-Field Control of Tri-State Phase Transformation with a Selective Dual-Ion Switch, Nature, 2017, 546, 124-128. (8) Said, E.; Crispin, X.; Herlogsson, L.; Elhag, S.; Robinson, N. D.; Berggren, M. Polymer Eield-Effect Transistor Gated via a Poly (Styrenesulfonic Acid) Thin Film. Appl. Phys. Lett. 2006, 89, 143507. (9) Shimotani, H.; Asanuma, H.; Tsukazaki, A.; Ohtomo, A.; Kawasaki M.; Iwasa, Y. Insulator-to-Metal Transition in ZnO by Electric Double Layer Gating. Appl. Phys. Lett. 2007, 91, 082106.


Page 14 of 18

Page 15 of 18

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


(10) Yuan, H. T.; Shimotani, H.; Ye, J. T.; Yoon, S.; Aliah, H.; Tsukazaki, A.; Kawasaki, M.; Iwasa, Y. Electrostatic and Electrochemical Nature of Liquid-Gated Electric-Double-Layer Transistors Based on Oxide Semiconductors. J. Am. Chem. Soc. 2010, 132, 18402-18407. (11) Kim, S. H.; Hong, K.; Xie, W.; Lee, K. H.; Zhang, S.; Lodge, T. P.; Frisbie, C. D. Electrolyte-Gated Transistors for Organic and Printed Electronics. Adv. Mater. 2013, 25, 1822-1846. (12) Drachman, D. A. Do We Have Brain To Spare? Neurology 2005, 64, 2004-2005. (13) Bi, G. Q.; Poo, M. M. Synaptic Modifications in Cultured Hippocampal Neurons: Dependence on Spike Timing, Synaptic Strength, and Postsynaptic Cell Type. J. Neurosci. 1998, 18, 10464-10472. (14) Zucker, R. S.; Regehr, W. G. Short-Term Synaptic Plasticity. Annu. Rev. Physiol. 2002, 64, 355-405. (15) Voglis, G.; Tavernarakis, N. The Role of Synaptic Ion Channels in Synaptic Plasticity. EMBO Rep. 2006, 7, 1104-1110. (16) Hu, S. G.; Liu, Y.; Chen, T. P.; Liu, Z.; Yu, Q.; Deng, L. J.; Yin, Y.; Hosaka, S. Emulation the Paired-Pulse Facilitation of a Biological Synapse with a NiOx-Based Memristor, Appl. Phys. Lett., 2013, 102, 183510; (17) Ohno, T.; Hasegawa, T.; Tsuruoka, T.; Terabe, K.; Gimzewski, J. K.; Ano, M. ShortTerm Plasticity and Long-term Potentiation Mimicked in Single Inorganic Synapses. Nat. Mater., 2011, 10, 591-595; (18) Prezioso, M.; Merrikh-Bayat, F.; Hoskins, B. D.; Adam, G. C.; Likharev, K. K.; Strukov, D. B. Training and Operation of an Integrated Neuromorphic Network Based on Metal-Oxide Memristors, Nature, 2015, 521, 61-64 (19) Alibart, F.; Pleutin, S.; Guerin, D.; Novembre, C.; Lenfant, S.; Lmimouni, K.; Gamrat, C.; Vuillaume, D. An Organic Nanoparticle Transistor Behaving as a Biological Spiking Synapse. Adv. Funct. Mater. 2010, 20, 330-337. 15 ACS Paragon Plus Environment


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

(20) Nishitani, Y.; Kaneko, Y.; Ueda, M.; Morie, T.; Fuji, E. Three-Terminal Ferroelectric Synapse Device with Concurrent Learning Function for Artificial Neural Networks. J. Appl. Phys. 2012, 111, 124108. (21) Kaneko, Y.; Nishitani, Y.; Ueda, M. Ferroelectric Artificial Synapses for Recognition of a Multishaded Image. IEEE Trans. Electron Devices 2014, 61, 2827-2833. (22) Shi, J.; Ha, S. D.; Zhou, Y.; Schoofs, F.; Ramanathan, S. A Correlated Nickelate Synaptic Transistor. Nat. Commun. 2013, 4, 2676. (23)Guo, L. Q. ; Wen, J.; Zhu, L.Q.; Fu, Y.M.; Xiao, H. Humidity-Dependent Synaptic Plasticity for Proton Gated Oxide Synaptic Transistor. IEEE Electron Device Lett, 2017, 38, 1248-1251. (24) Kim, K.; Chen, C. L.; Truong, Q.; Shen, A. M.; Chen, Y. A Carbon Nanotube Synapse with Dynamic Logic and Learning. Adv. Mater. 2013, 25, 1693-1698. (25) Zhu, L. Q.; Wan, C. J.; Guo, L. Q.; Shi, Y.; Wan, Q. Artificial Synapse Network on Inorganic Proton Conductor for Neuromorphic Systems. Nat. Commun. 2014, 5, 3158. (26) Gkoupidenis, P.; Schaefer, N.; Garlan, B.; Malliaras, G. G. Neuromorphic Functions in PEDOT:PSS Organic Electrochemical Transistors. Adv. Mater. 2015, 27, 7176-7180. (27) Wan, C. J.; Zhu, L. Q.; Zhou, J. M.; Shi, Y.; Wan, Q. Memory and Learning Behaviors Mimicked in Nanogranular SiO2-Based Proton Conductor Gated Oxide-Based Synaptic Transistors. Nanoscale 2013, 5, 10194-10199. (28) Lai, Q. X.; Zhang, L.; Li, Z. Y.; Stickle, W. F.; Williams, R. S.; Chen, Y. Ionic/Electronic Hybrid Materials Integrated in a Synaptic Transistor with Signal Processing and Learning Functions. Adv. Mater. 2010, 22, 2448-2453. (29) Gkoupidenis, P.; Koutsouras, D. A.; Malliaras, G. G. Neuromorphic Device Arthitectures with Global Connectivity Through Electrolyte Gating, Nat. Commun., 2017, 8, 15448. (30) Huang, Y. Y.; Cion, A.; Selig, D. K.; Malenka, R. C. The Influence of Prior Synaptic Activity on the Induction of Long-Term Potentiation. Science 1992, 255, 730-733. 16 ACS Paragon Plus Environment

Page 16 of 18

Page 17 of 18

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


(31) Ganguly, K.; Poo, M. M. Activity-Dependent Neural Plasticity from Bench to Bedside. Neuron 2013, 80, 729-741. (32) Yuan, H. T.; Shimotani, H.; Tsukazaki, A.; Ohtomo, A.; Kawasaki, M.; Iwasa, Y. Hydrogenation-Induced Surface Polarity Recognition and Proton Memory Behavior at ProticIonic-Liquid/Oxide Electric-Double-Layer Interfaces. J. Am. Chem. Soc. 2010, 132, 66726678. (33) Smith, A. J.; Owens, S.; Forsythe, I. D. Characterisation of Inhibitory and Excitatory Postsynaptic Currents of the Rat Medial Superior Olive. J. Physiol 2000, 529, 681-698. (34) Kamiya, T.; Nomura, K.; Hosono, H. Origins of High Mobility and Low Operation Voltage of Amorphous Oxide TFTs: Electronic Structure, Electron Transport, Defects and Doping. J. Disp. Technol. 2009, 5, 273-288. (35) Kawamura, T.; Uchiyama, H.; Saito, S.; Wakana, H.; Mine, T.; Hatano, M. Analysis of Subthreshold Slope of Fully Depleted Amorphous In-Ga-Zn-O Thin Film Transistors. Appl. Phys. Lett. 2015, 106, 013504. (36) Sturman, B.; Podivilov, E.; Gorkunov, M. Origin of Stretched Exponential Relaxation for Hopping-Transport Models. Phys. Rev. Lett. 2003, 91, 176602. (37) Liu, N.; Zhu, L.Q.; Xiao, H.; Wan, C.J.; Liu, Y.H.; Chao, J.Y. Transient Characteristics for Proton Gating in Laterally Coupled Indium-Zinc-Oxide Transistors. ACS Appl. Mater.Interfaces, 2015, 7, 6205-6210. (38) Bornschein, G.; Arendt, O.; Hallermann, S.; Brachtendorf, S.; Eilers, J.; Schmidt, H. Paired-Pulse Facilitation at Recurrent Purkinje Neuron Synapses is Independent of Calbindin and Parvalbumin During High-Frequency Activation. J. Physiol. 2013, 591, 3355-3370. (39) Buonomano, D. V.; Maass, W. State-Dependent Computations: Spatiotemporal Processing in Cortical Networks. Nat. Rev. Neurosci. 2009, 10, 113-698. (40) Atluri, P. P.; Regehr, W. G. Determinants of the Time Course of Facilitation at the Granule Cell to Purkinje Cell Synapse. J. Neurosci. 1996, 16, 5661-5571. 17 ACS Paragon Plus Environment


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

(41) Fortune, S. E.; Rose, G. J. Short-Term Synaptic Plasticity as a Temporal Filter. Trends Neurosci. 2001, 24, 381-385. (42) Abbott, L. F.; Regehr, W. G. Synaptic Computation. Nature 2004, 431, 796-803. (43) Bertrand, S. S.; Cazalets, J. R. Activity-Dependent Synaptic Plasticity and Metaplasticity in Spinal Motor Networks. Curr. Pharm. Des. 2013, 19, 4498-4508. (44) Hawkins, R. D.; Abrams, T. W.; Carew, T. J.; Kandel, E. R. A Cellular Mechanism of Classical Conditioning in Aplysia: Activity-Dependent Amplification of Presynaptic Facilitation. Science 1983, 219, 400-405.

Table of Contents (TOC) graphic


Page 18 of 18

Activity Dependent Synaptic Plasticity Mimicked on IndiumâTinâOxide

Recommend Documents

Activity Dependent Synaptic Plasticity Mimicked on IndiumâTinâOxide