A Flexible Transparent Organic Artificial Synapse ... - ACS Publications

Subscriber access provided by UNIV OF LOUISIANA

Functional Nanostructured Materials (including low-D carbon)

A Flexible Transparent Organic Artificial Synapse based on Tungsten / Egg Albumen/Indium Tin Oxide/polyethylene Terephthalate Memristor Xiaobing Yan, Xiaoyan Li, Zhenyu Zhou, Jianhui Zhao, Hong Wang, Jingjuan Wang, Lei Zhang, Deliang Ren, Xin Zhang, Jingsheng Chen, Chao Lu, Peng Zhou, and Qi Liu ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.9b04443 • Publication Date (Web): 30 Apr 2019 Downloaded from http://pubs.acs.org on April 30, 2019

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 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 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.

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 20 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

A Flexible Transparent Organic Artificial Synapse based on Tungsten /Egg Albumen/Indium Tin Oxide/polyethylene Terephthalate Memristor

Xiaobing Yan,1*# Xiaoyan Li,1# Zhenyu Zhou,1 Jianhui Zhao, 1 Hong Wang,1Jingjuan Wang,1Lei Zhang1, Deliang Ren,1 Xin Zhang,1 Jingsheng Chen,2 Chao Lu 3, Peng Zhou,4 Qi Liu5

1College

of Electron and Information Engineering, Key Laboratory of Digital Medical Engineering of Hebei

Province, Key Laboratory of Optoelectronic Information Materials of Hebei Province, Hebei University, Baoding 071002, P. R. China 2

Department of Materials Science and Engineering, National University of Singapore, Singapore 117576,

Singapore 3

Department of Electrical and Computer Engineering, Southern Illinois University Carbondale, Carbondale, IL,

United States 47901 4

State Key Laboratory of ASIC and System, School of Microelectronics, Fudan University, Shanghai 200433, China

5

Key Laboratory of Microelectronic Devices and Integrated Technology Institute of Microelectronics Chinese

Academy of Sciences Beijing 100029, P. R. China # These

authors contributed equally to this work.

* Author to whom correspondence should be addressed: [email protected].

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 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

Abstract As artificial synapses in biomimetics, memristors have received increasing attention due to their great potential in brain-inspired neuromorphic computing. The use of biocompatible and degradable materials as the active resistive layer is promising in memristor fabrication. In this work, we select egg albumen as the resistive layer to fabricate flexible W/Egg Albumen/ITO/PET devices, which can operate normally under mechanical bending without significant performance degradation. This proposed memristor device exhibits a transparency of more than 90% under visible light with a wavelength range of 230-850nm. Moreover, by changing amplitudes of pulse voltage instead of intervals, paired-pulse facilitation (PPF) can be transmitted to paired-pulse depression (PPD), which can faithfully mimic dynamical balance of Ca2+ concentration shaped by voltagesensitive calcium channels. The device resistance can be modulated gradually by applied pulse trains to mimic certain neural bionic behaviors, including excitatory postsynaptic current (EPSC), short-term plasticity (STP) and long-term plasticity (LTP), transitions between STP and LTP. The reasons behind these behaviors are analyzed through power consumption calculation. Excellent bio-simulation characteristics have been demonstrated in this egg albumen based memristor device, which is desirable in biocompatible and dissolvable electronics for flexible artificial neuromorphic systems.


Page 2 of 20

Page 3 of 20 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 A human brain is a powerful parallel information processor with high fault tolerance and energy efficiency.

1-3

A cranial nervous system comprises numerous neurons

(1011~1012) and synapses (~1015) that are involved in memory and learning functions. 4-6

The emerging memristor, a dual-terminal non-volatile device whose resistance can

be adjusted through the application of electrical pulses or biases,

7-11

is considered as

an intelligent online training device. Similar to the human brain, a group of memristors can mimic the neuromorphic activities.2 Although the conventional approach can also fulfill these functions, it requires transistors and capacitors that are expensive, bulky, and limited by the von Neumann architecture.12-15 The features of memristors, e.g., nonvolatility, high density, scalability, quick operation, and low power consumption, are advantageous for implementing the synaptic behaviors observed in biological systems.16-18 This new type of system efficiently can perform learning, memory, and other complex processing tasks.3,5,19-21 To date, memristors have replicated several synaptic functions, e.g., spike-rate-dependent plasticity (SRDP), spike-timingdependent plasticity (STDP), excitatory postsynaptic current (EPSC), short-term plasticity (STP) and long-term plasticity (LTP). 3,5,6,22-26 Implantable biological or biomedical devices must be biocompatible and biodegradable. 3,27-29

However, most memristors reported in the literature have been constructed from

inorganic materials that are brittle and incompatible with flexible versatile functionalities.27,30-32 To establish cost-effective production routes for biocompatible and biodegradable devices, new research must employ naturally sourced compounds to satisfy these requirements. Organic materials are widely used for research and applications due to low cost, environmentally stability, and available for large-size film manufacturing.3,5,33,47 Among these materials, egg albumen (or egg white) is a natural organic substance that is biodegradable, bioabsorbable, and environmentally friendly. From the perspective of biological applicability, egg albumen is also biocompatible, implantable, and cheaply produced.

34,48

Meanwhile, owing to their high flexibility,

shape diversity, and light weight, flexible materials can realize foldable, stretchable, wearable, and implantable systems. Flexible memristors also have attracted increasing attention as information storage components and brain-inspired computing.


35-38

In a


biological synapse, voltage-sensitive calcium channels is the main factor to shape the dynamical balance of the Ca2+ concentration, 39 whereas the Ca2+ dynamics naturally leads to short-term plasticity in which residual elevation of pre-synaptic (Ca2+) directly relates to the enhancement of synaptic transmission. Therefore, we intend to modulate device conduction through voltage pulse amplitudes instead of intervals. On the other hand, the metal ion dynamics of the memristor can emulate the Ca+ accumulation, which causes modification of synaptic strengths. 40 Moreover, the protein of egg white can work well as an interstitial fluid. In this work, the resistive layer and substrate of an artificial synaptic device were fabricated using egg albumen and polyethylene terephthalate (PET), respectively. Meanwhile, PET is a natural polymer with good flexibility and transparency. This W/Egg Albumen/ITO/PET memristor device exhibits thresholding switching behaviors. We systematically investigated its electronic artificial synaptic behaviors. Interestingly, the paired-pulse facilitation (PPF) can be transmitted to paired-pulse depression (PPD) by changing voltage amplitudes instead of intervals. This study paves the way for implementing biopolymer-based artificial synapses for neuromorphic computing, and helps to optimize memristive synapse devices in future research. 2. Experimental Section 2.1. Materials: All chemicals except eggs are used as received without further purification. We obtained albumen from a fresh chicken egg, which had been randomly selected from a supermarket. Ultrapure, deionized water was used in all experiments. Preparation of egg white solution: In step 1, an albumen film was prepared in a clean laboratory environment. Egg white was collected by breaking the shell of a fresh, precleaned chicken egg. In step 2, separate the egg white from the yolk. Some egg white can be filtered out by a stainless-steel mesh spoon. In step 3, place the filtered albumen on the test bench for 3 minutes to extract more egg white. In step 4, extract the sample (volume: 1 ml) from the egg white solution using a pipette. In step 5, the 1-ml albumen sample was mixed with 15 ml of deionized water to synthesize a thin dielectric film for device fabrication. Because untreated raw egg white has a high viscosity, a direct spin coating generally results in a non-uniform thickness of the albumen film. In step 6, 16 ml of the diluted albumen solution was sonicated for 1 min. The sonication step was


Page 4 of 20

Page 5 of 20 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


performed five times with a 1-min interval between successive sonication. In this manner, the albumen was thoroughly dispersed in the deionized water. In step 7, the mixture was filtered through a dust-free cloth to remove any suspended solids. In step 8, the diluted albumen liquid was spin coated onto the ITO/PET substrate for 5 s at 500 rpm, followed by spin coating for 40 s at 4000 rpm, forming a thin albumen film (thickness: ~300 nm) on the substrate. Next, the albumen film was baked at a temperature as high as 105°C for 10 min. Please note that raw egg whites are soluble, because they contain various proteins (e.g., albumins, mucoproteins, and globulins), trace minerals, fats, vitamins, and salts (ions such as Na+, K+, and Fe+, which float within egg white).41 Drying at low temperatures would cause high levels of water, resulting in unstable properties of the albumen film. 2.2. Characterization of egg white memristor: AFM (Bruker, USA) was used to obtain surface features. At room temperature and atmospheric pressure, the device I-V characteristics were measured using a Keithley 2635B semiconductor parameter analyzer. Pulse waveforms were captured with a LeCroy 62MXs-B oscilloscope. Two pulse channels of Keithley 4225 PMU provided pre- and post-spiking, respectively. 3. Results and Discussion 3.1. Flexible Switching Behaviors. Figure 1 illustrates the eight steps involved in the production of a tungsten (W)/egg albumen/indium tin oxide (ITO)/polyethylene terephthalate (PET) memristor and the schematic of the whole device. The preparation process of the egg albumen solution is shown in the experimental section. In an atomic force microscopy (AFM) examination, the root-mean-square roughness of the fabricated albumen film was 0.4777 nm as shown in Figure S1, indicating good uniformity. Finally, to minimize the impacts of the upper electrode on the device behaviors, metal W was deposited as the top electrode.



Figure 1. Proposed fabrication process and schematic of a tungsten (W)/egg albumen/indium tin oxide (ITO)/polyethylene terephthalate (PET) memristor.

The proposed memristor was built as a sandwich-like structure of W/egg albumen/ITO/PET. Figure 2a shows the length of the device is 10mm, and Figure 2b shows the albumen/ITO/PET memristor with a transparency of more than 90% under visible light with a wavelength range of 230–850 nm. This transparent property was assessed to be excellent. Figure 2c shows the measured I–V curve as the DC voltage was swept in the direction a → b → c → d → e → f → a. Here, a positive voltage was applied to the device (i.e., the voltage bias and ground were applied to the top and bottom electrodes, respectively). Initially, the device was in the OFF state, exhibiting high resistance. When the applied voltage exceeded the set voltage (2 V), the device state switched from HRS to LRS. The device remained in the ON state until the applied voltage reached 5 V. Then, the device regained the high-resistance state (HRS) at a higher applied voltage (7 V). Symmetric resistance changes were observed under an applied voltage with opposite polarity (i.e., 0 V → −1 V → −4 V → −8 V). By testing retention of the voltage point b and voltage point e. The resistance can be maintain although there are some fluctuation. It can be seen in Figure R2. This behavior is similar to that reported in a complementary resistive switches (CRS) and distinct from the normal resistive switching behavior. On sweeping to a moderate maximum voltage, the device is Set at a lower voltage and Reset at a higher voltage of the same polarity.50,51 In a typical CRS behavior, the set voltage and reset voltage of the device are not related


Page 6 of 20

Page 7 of 20 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


with the polarity of the applied voltage. A positive voltage and a negative voltage are applied to our devices, respectively, the set and reset show a slight asymmetry. This may be related to uneven ion distribution in the egg white.52 Figure 2d plots the measured HRS and low-resistance state (LRS) over 100 repeated cycles, implying strong robustness and reliability of the device. Moreover, the device under bending conditions shows excellent flexibility. Figure 2e shows the effect of the number of bends on the device’s resistance. Even after the memristive device has been bent 1000 times and under different bending length, the resistance of the HRS and LRS remains the appropriate ratio as shown in Figure 2e and 2f. In Figure 2e, these data were measured when the substrate is flat after bent of the device. The property have no obvious decay if measured at bended state as shown in the figure S3. In Figure 2f, when the bending degree of our device is less than 7mm, we chose about 5mm, and its high and low resistance states are about 5108 Ω and 1×106 Ω, respectively. These are basically maintained at the same level, as shown in Figure S4.The experimental results in Figures 2d–2f confirmed the stability of the proposed memristor for flexible artificial bio-synapse applications. Figure S5 shows the dissolution process of the memristor we fabricated. After immersing in DI water for 1 hour, the top W electrode began to dissolve, and the W electrode dissolved away significantly after 5 hours of invading the water (Figure S5c), and uneven dissolution of the albumen film was observed. 34 After 20 hours, the entire albumen film almost disappeared in water (Figure S5d) with only traces of residuals remaining. Therefore, memristors using albumen as a resistive layer have good degradability. This provides one idea for implantable biomedical devices and the manufacture of degradable devices.



Figure 2. Resistance switching performance and bending characteristics of the fabricated transparent, flexible W/egg albumen/ITO/PET memristor device. (a) Surface characteristics of the device under the optical microscope. (b) Transparency demonstration of the albumen/ITO/PET samples. (c) I–V characteristics of the flat device as the voltage sweeps in the direction a → b → c → d → e → f → a. (d) Repeatability test of the high-resistance-state (HRS) and low-resistance-state (LRS) values of the fabricated device. (e) Extracted device HRS and LRS vs. bending time. (f) Extracted device HRS and LRS vs. bending length.

3.2. Emulator of Artificial Electronic Synapse. Furthermore, the current responses of this device to voltage pulse stimulations were investigated. The pulse was used as a signal source in a simple circuit. This circuit was composed of a 1.1MΩ resistor in series with the devices as shown in the Figure S6. Figure 3a shows the waveform of a negative pulse sequence applied in this measurement. The applied voltage and read voltages were −5 V and −0.2 V, and the pulse width and interval were


Page 8 of 20

Page 9 of 20 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. Device responses to large voltage stimulations. (a) Pulse sequence used in our measurement. Plots of current vs. time in device measurement under successive voltage pulses: (b) 4 pulses, (c) 30 pulses, and (d) 50 pulses.

both fixed at 2 μs. The number of successive voltage pulses was varied as 4, 30, and 50. Panels (b)–(d) of Figure 3 plot the measured device currents over time. Initially, the device was in the OFF state and was turned on by the first voltage pulse. But our device does not immediately decay after the pulse stimulation removed. It might be caused by that the metal ions in egg albumen still can move forward a short distance due to the inertia effect after the pulse voltage is removed, which may result in the formation of ion-mediated EPSC. This also causes the EPSC to not decay rapidly after the voltage is removed, which is similar to other reports. 42-45 As more stimulating voltage pulses were applied, the current gradually decreased and eventually reached a steady state. Similar phenomena were also found in oxide memristors.

46

The conductance first increased

rapidly, and then slowed down with the number of pulses increased. Symmetric phenomena were observed under an applied voltage pulse with positive polarity in Supplementary Information Figure S7. To further explore the amplitude effects on current regulation in details, the device behaviors were systematically studied under different pulse amplitudes. Herein, the



pulse width and interval were both fixed to 2 μs, and the voltage amplitude was varied from −2 V to −6 V. The number of pulses was varied as 4, 30, and 50. Starting from the

Figure 4. Plots of device currents vs. different amplitudes and numbers of voltage pulses. The pulse width and interval are both fixed to 2 µs, and the amplitude ranges from −2 V to −6 V. The devices are subjected to (a) 4 pulses, (b) 30 pulses, and (c) 50 pulses. (d) Amount of energy required by the devices to switch from HRS to LRS under different voltages.

same initial state, the device current responded to the various conditions shown in Figure 4. Under pulse amplitudes of −2 V, −3 V, and −4 V, the current rose at the beginning and then remained almost saturated. This phenomenon was introduced in our other work

13.

However, under pulse amplitudes of −5 V and −6 V, the device was

switched on and the current reach the largest value suddenly then reaching a steady state, which is analogous to the behaviors in Figure 3. These results revealed that pulse amplitude plays an important role in the evolution of device currents, and a pulse sequence with a higher and lower amplitude voltage would result in different current variation tend in the device. To analysis the physical mechanism, we calculate its power consumption to investigate the reason why the device current increases instantaneously at a large voltage. Figure 4d shows the amount of energy required under different amplitude. Table S1 provides a comparison of the energies of the device under different numbers and amplitudes of successive voltage pulses. The amount of required energy increased at higher amplitudes of the applied voltage. At voltages of −2 V, −3 V, and


Page 10 of 20

Page 11 of 20 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 V, the energy in the device was relatively small; thus, the current flowing through the device rose slowly with the pulse number increase. At voltages above −5 V, the energy gained by the device exceeded 10−6 J, which causes the device to suddenly open during the first pulse stimulus and the current to increase instantaneously. This is similar to the PPF and the PPD in biological synapses. Our work uses the magnitude of the pulse amplitude to simulate these two synaptic effects. From Figure 4a, it can be seen that the device exhibits PPF effect, when the amplitude is -2 V, -3 V, -4 V. The device exhibits PPD effect, when the amplitude reaches -5 V, -6 V. Obviously, the PPF can be effectively transmitted to the PPD by changing the amplitude. It is also proved by the applied voltage with positive polarity as shown in Figure S8. Moreover, this mode of transformation from PPF to PPD also faithfully similar to the dynamical balance of the Ca2+ concentration in chemical synapse shaped by voltage-sensitive calcium channels. Biologically, each synapse generally has a limited amount of resources U: chemical neurotransmitters. Each spike will activate a portion of these resources (a×U (a < 1)).

The transmitted spike amplitude is a function of this portion.

The portion of neurotransmitters, which is used to transmit information, then is recovered with a feature time. Larger amplitude (i.e., larger energy) usually leads to the need for more neurotransmitters. Therefore, a depressing behavior is observed, due to lower recovery of decreasing neurotransmitters.

Figure 5. Measured excitatory postsynaptic current (EPSC) results after (a) 4 consecutive pulses and (b) 30 consecutive pulses. (c) EPSC variation varies with pulse amplitudes, where I0 and I are the initial and measured post-stimulation currents, respectively.

It is well known that short-term plasticity (STP) and long-term plasticity (LTP) exist in human memory. An STP becomes an LTP after sufficient training. These properties are also achieved by the short-term and long-term modulations of memristor resistance, which is observed in the time domain by monitoring the relaxation of the memristor



Page 12 of 20

current after voltage pulse stimulations, i.e., excitatory post-synaptic current (EPSC). Figure S7b clarifies the EPSC of the proposed device under a 5 V positive pulse with a width of 2 μs. Figures 5a and 5b plot the measured EPSC after applying 4 and 30 consecutive positive pulses, respectively. And the pulse width and interval were both fixed to 2 μs and the pulse amplitude was increased from 2 to 6 V. A small pulse (0.1 V) was applied to read the memristor current without affecting the device state. The EPSC variation was calculated by comparing the initial current I0 and the measured current I after 0.01 and 10 μs of pulse stimulations. Figure 5c plots the pulse amplitude dependence of the EPSC variation. Under relatively small pulse amplitudes, the resulting EPSC almost remained at its initial value as shown in Figure 5a, which is defined as an STP behavior. Under larger pulse amplitudes, EPSC enhancement still exists after 10 μs, which could be defined as an LTP behavior. The device showed the same phenomenon under 30 pulses of stimulation as shown in Figure 5b. The change of EPSC increased with higher pulse amplitudes, implying a transition from STP to LTP, which indicates a significant dependence on pulse amplitudes. In addition to the pulse amplitude, the pulse frequency can also have a significant effect on the synaptic properties of the device. We take the reciprocal of the pulse period as the using frequency here. To investigate the effect of spike rate on synaptic plasticity, EPSC was measured immediately after applying 50 consecutive pulses of different frequencies (333, 435, 588, and 1000 kHz) to the memristor, starting from the same initial state. The pulse voltage and width were fixed to 5 V and 1 μs, respectively. At each frequency (Figures. 6a–6d), the normalized synaptic weight drastically decreased from its initial value, followed by a gradual decrease over time. The measured data in Figures 6a–6d were fitted with a modified Kohlrausch equation that is widely accepted to represent a forgetting function in psychology 5. This equation is written as follows: It = Is + A exp(−t/τ),

(1)

where Is is the current in the stabilized state, A is a prefactor, t represents time, and τ is the characteristic relaxation time constant, which represents the forgetting rate. As shown in Figure 6e, the relaxation time increases with the increasing frequency of the stimulation pulse. A larger τ indicates a longer decay time of the memristor current. In our case, this may be due to the spontaneous diffusion of ions from the conductive filaments between two pulse, resulting in the decaying of EPSC 19. The next pulse needs


Page 13 of 20 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


to fill the ion diffusion caused by the pulse interval, so the interval duration influences the speed of rupture and formation of conductive channels. Note that a higher pulse frequency corresponds to slower attenuation (i.e., slower forgetting). As can be seen from the figure 6(e), the higher the frequency of the applied pulse, the slower the forgetting speed of the device. We fit τ as formula: 𝜏𝑓 = 𝜏0 +𝐴exp(𝑓 × 𝐵), where 𝜏0 is the initial value of 𝜏𝑓, A and B are prefactor, 𝑓 represents frequency. The EPSC is largely affected not only by the pulse frequency but also by the number of pulses. Panels (a)–(c) of Figure S5 plot the device’s EPSC and their fitting curves (determined

Figure 6. Effect of stimulation rate on the EPSC of the memristor. EPSC was measured immediately after 50 consecutive pulses (amplitude: 5 V; pulse width: 1 µs). The EPSC and fitting curves under pulse frequencies of (a) 333 kHz, (b) 435 kHz, (c) 588 kHz, and (d) 1000 kHz. (e) A plot of relaxation time vs. pulse frequency.

using Eq. (1)) after 4, 30, and 50 stimulation pulses, respectively, and Figure S5d shows how the retention time τ varies with the number of stimulation pulses N. The relaxation time τ increased with the number of applied stimulation pulses. 3.3. Switching Mechanism. Figure 7 illustrates the possible resistance change mechanisms of the W/Egg Albumen/ITO/PET device. Most proteins in egg white are



globular proteins, which are linked together by a number of weak chemical bonds. When these bonds are formed (break), they absorb (release) molecules, electrons, and hydrogen atoms. The resistance transition occurs by the drift of a limited number of metallic elements between the ITO bottom electrode layer and the W top electrode. Apart from proteins, egg white contains trace amounts of metal ions, e.g., iron, sodium, and potassium. The switching behavior has been attributed to metal atoms, such as iron oxidation (Fe → Fe3+ + 3e−) and reduction (Fe3+ + 3e− → Fe).34 The thresholding phenomenon in our device differs from that of bipolar resistive memories because the bottom ITO electrode is inert, whereas in that work the Mg is also can occur electrochemical reaction in Ref.34. Therefore, the metal ions in the resistive layer cannot be replenished over time and are responsible for the thresholding switching behaviors. When no bias voltage is applied, most ions in the egg membrane are distributed, as shown in Figure 7a. When the upper electrode receives a positive voltage, metallic atoms (e.g., Fe) are oxidized to ions (e.g., Fe3+). Under an electric field, these ions will drift toward the ITO layer, depleting the metallic atoms at the bottom electrode. At this time, the metallic atoms and non-reduced ions in the albumen film accumulate into filaments near the ITO layer (Figure 7b). As the bias voltage increases, the metal ion conduction paths continue to grow until they reach the bottom ITO electrode (Figure 7c), thereby establishing the LRS. If the bias voltage further increases, insufficient metal ions in the albumen have been consumed, and the iron atoms in filaments near the upper electrode may be oxidized to iron ions moving downwards, causing the filament to break to form a high-impedance state (Figure. 7d). And the thresholding switching phenomenon is thus obtained.49

Figure 7. Mechanism of the filamentary model. (a) Distribution of the trace elements in albumen without bias voltage. (b) Under a positive applied voltage, the ions in the albumen layer move along the electric field lines. (c) Under a higher bias voltage, a filament and connection are built between


Page 14 of 20

Page 15 of 20 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 top and bottom electrodes. The device is in the LRS. (d) As the bias voltage increases further, the filaments are ruptured near the top electrode and the device gradually enters the HRS.

In order to further explore the effect of W electrode on device performance, we replace electrode W with the conventional high-diffusion active electrode Ag, the device became a bipolar resistive switch, as shown in Figure S10. But in this work, the W-electrode device obtained a stable threshold switching effect. So we may be that the W diffusion ability is not as strong as the Ag electrode. 4. Conclusions This study presents a highly-transparent and flexible memristor device, which is made of egg white. This device performs synaptic functions by adjusting applied pulse voltages. It is observed that a PPF-to-PPD transition can implement by changing pulse amplitudes for emulating dynamical balance of Ca2+ concentration in chemical synapse shaped by voltage-sensitive calcium channels. The forgetting speed of this device is stimulated by controlling pulse frequencies and the number of pulses. Due to the use of dissolvable a protein membrane of egg white as a material. This device can be dissolved in deionized water within 1 day. Experimental results show that this proposed albumenbased memristor device is attractive in biocompatible and biodegradable electronics, which adopt cheap and natural organic materials to achieve intelligent artificial synaptic systems. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (No.61674050), Top‐notch Youth Project of University in Hebei Province (No. BJ2014008), Outstanding Youth Project of Hebei Province (No. F2016201220), Outstanding Youth Cultivation Project of Hebei University (No. 2015JQY01), Project of Science and Technology activities for Overseas Researcher (No. CL201602), Institute of Baoding Nanyang Research - New Material Technology Platform (17H03), 2018 School level Innovation Program of Hebei University (hbu2018ss04), Project of distinguished young of Hebei province (No. A2018201231). References



(1) Pakkenberg, B.; Pelvig, D.; Marner, L.; Bundgaard, M. J.; Gundersen, H. J. G.; Nyengaard, J. R.; Regeur, L. Aging and the human neocortex. Exp. Gerontol. 2003, 38, 95-99. (2) Kuzum, D.; Yu, S.; Wong, H. P. Synaptic electronics: materials, devices and applications. Nanotechnology. 2013, 24, 382001. (3) Park, Y.; Lee, J. S. Artificial synapses with short-and long-term memory for spiking neural networks based on renewable materials. ACS Nano. 2017, 11, 8962-8969. (4) Drachman, D. A. Do we have brain to spare?. Neurology. 2005, 64, 2004-2005. (5) Li, S.; Zeng, F.; Chen, C.; Liu, H.; Tang, G.; Gao, S.; Song, C.; Lin, Y.; Pan, F.; Guo, D. Synaptic plasticity and learning behaviours mimicked through Ag interface movement in an Ag/conducting polymer/Ta memristive system. Journal of Materials Chemistry C. 2013, 1, 5292-5298. (6) Yan, X. B.; Li, K.; Yin, J.; Xia, Y. D.; Guo, H. X.; Chen, L.; Liu, Z. G. The resistive switching mechanism of Ag/SrTiO3/Pt memory cells. Electrochem. Solid-State Lett. 2010, 13, H87-H89. (7) Chua, L. O. Memristor-the missing circuit element. IEEE Trans. Circuit Theory. 1971, 18, 507519. (8) Yu, S.; Wu, Y.; Jeyasingh, R.; Kuzum, D.; Wong, H. P. An electronic synapse device based on metal oxide resistive switching memory for neuromorphic computation. IEEE Trans. Electron Devices. 2011, 58, 2729-2737. (9) Wang, Z. W.; Yin, M. H.; Zhang, T.; Cai, Y. M.; Wang, Y. Y.; Yang, Y. C.; Huang, R. Engineering incremental resistive switching in TaOx based memristors for brain-inspired computing. Nanoscale. 2016, 8, 14015-14022. (10) 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. (11) Chua, L. O.; Kang, S. M. Memristive devices and systems. Proc. IEEE 1976, 64, 209–223. (12) Merolla, P. A.; Arthur, J. V.; Alvarez-Icaza, R.; Cassidy, A. S.; Sawada, J.; Akopyan, F.; Jackson, B. L.; Imam, N.; Guo, C.; Nakamura, Y.; Brezzo, B.; Vo, L.; Esser, S. K.; Appuswamy, R.; Taba, B.; Amir, A.; Flickner, M. D.; Risk, W. P.;

R. Manohar, R.; Modha, D. S. A million

spiking-neuron integrated circuit with a scalable communication network and interface, Science. 2014, 345, 668. (13) Yan, X.; Zhao, J.; Liu, S.; Zhou, Z.; Liu, Q.; Chen, J.; Liu, X. Y. Memristor with Ag‐Cluster‐Doped TiO2 Films as Artificial Synapse for Neuroinspired Computing. Adv. Funct. Mater. 2018, 28, 1705320. (14) 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. (15) Jeong, D. S.; Kim, K. M.; Kim, S.; Choi, B. J.; Hwang, C. S. Memristors for energy‐efficient new computing paradigms. Adv. Electron. Mater. 2016, 2, 1600090. (16) Yang, J. J.; Strukov, D. B.; Stewart, D. R.; Memristive devices for computing. Nat.


Page 16 of 20

Page 17 of 20 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


Nanotechnol. 2012, 8, 13. (17) Pan, F.; Gao, S.; Chen, C.; Song, C.; Zeng, F. Recent progress in resistive random access memories: materials, switching mechanisms, and performance. Mater. Sci. Eng. R. Rep. 2014, 83, 1-59. (18) Rajendran, B.; Liu, Y.; Seo, J. S.; Gopalakrishnan, K.; Chang, L.; Friedman, D. J.; Ritter, M. B. Specifications of nanoscale devices and circuits for neuromorphic computational systems. IEEE Trans. Electron Devices. 2013, 60, 246-253. (19) Chang, T.; Jo, S. H.; Lu, W.; Short-term memory to long-term memory transition in a nanoscale memristor. ACS Nano. 2011, 5, 7669-7676. (20) Hu, S. G.; Liu, Y.; Chen, T. P. ; Liu, Z.; Yu, Q.; Deng, L. J.; Yin, Y.; Hosaka, S. Emulating the Ebbinghaus forgetting curve of the human brain with a NiO-based memristor. Appl. Phys. Lett. 2013, 103, 133701. (21) Li, J.; Duan, Q.; Zhang, T.; Yin, M.; Sun, X.; Cai, Y.; Li, L.; Yang, Y.; Huang, R.; Tuning analog resistive switching and plasticity in bilayer transition metal oxide based memristive synapses. RSC Adv. 2017, 7, 43132-43140. (22) 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. (23) Kuzum, D.; Jeyasingh, R. G. D.; Lee, B.; Wong, H. S. P. Nanoelectronic programmable synapses based on phase change materials for brain-inspired computing. Nano letters. Nano Lett. 2012, 12(5): 2179–2186. (24) Yan, X. B.; Zhou, Z. Y.; Ding, B. F. ; Zhao, J. H.; Zhang, Y. Y. Superior resistive switching memory and biological synapse properties based on a simple TiN/SiO2/p-Si tunneling junction structure. J. Mater. Chem. C. 2017, 5, 2259-2267. (25) Wang, Z. Q.; Xu, H. Y.; Li, X. H.; Yu, H.; Liu, Y. C.; Zhu, X. J. Synaptic learning and memory functions achieved using oxygen ion migration/diffusion in an amorphous InGaZnO memristor. Adv. Funct. Mater. 2012, 22, 2759-2765. (26) Kim, S.; Kim, H.; Hwang, S.; Kin, M. H.; Chang, Y. F.; Park, B. G. Analog synaptic behavior of a silicon nitride memristor. ACS Appl. Mater. Interfaces. 2017, 40420-40427. (27) Tan, Z. H.; Yang, R.; Terabe, K.; Yin, X. B.; Zhang, X. D.; Guo, X.; Synaptic metaplasticity realized in oxide memristive devices. Adv. Mater. 2016, 28, 377-384. (28) Kim, S.; Du, C.; Sheridan, P.; Ma, W.; Choi, S.; Lu, W. D. Experimental demonstration of a second-order memristor and its ability to biorealistically implement synaptic plasticity. Nano Lett. 2015, 15, 2203-2211. (29) Irimia-Vladu, M.; “Green” electronics: biodegradable and biocompatible materials and devices for sustainable future. Chem. Soc. Rev. 2014, 43, 588-610. (30) Hwang, S. W.; Tao, H.; Kim, D. H.; Cheng, H.; Song, J. K.; Rill, E.; Brenckle, M. A.; Panilaitis, B.; Won, S. M.; Kim, Y. S.; Song, Y. M.; Yu, K. J.; Ameen, A.; Li, R. Su, Y. Yang, M. Kaplan, D.



L.;

Zakin, M. R.; Slepian, M. J.; Huang, Y.; Omenetto, F. G.; Rogers, J. A. A physically transient

form of silicon electronics. Science. 2012, 337, 1640-1644. (31) Dagdeviren, C.; Hwang, S. W.; Su, Y.; Kim, S.; Cheng, H.; Gur, O.; Haney, R.; Omenetto, F. G.; Huang, Y.; Rogers, J. A. Transient, biocompatible electronics and energy harvesters based on ZnO. Small. 2013, 9, 3398-3404. (32) Hwang, S. W.; Kim, D. H.; Tao, H.; Kim, T. I.; Kim, S.; Yu, K. J.; Panilaitis, B.; Jeong, J. W.; Song, J. K.; Omenetto, F. G.; Rogers, J. A. Materials and fabrication processes for transient and bioresorbable high‐performance electronics. Adv. Funct. Mater. 2013, 23, 4087. (33) Ji, Y.; Cho, B.; Song, S.; Kim, T. W.; Choe, M.; Kahng, Y. H.; Lee, T. Stable switching characteristics of organic nonvolatile memory on a bent flexible substrate. Advanced Materials. Adv. Mater. 2010, 22, 3071-3075. (34) He, X.; Zhang, J.; Wang, W.; Xuan, W.; Wang, X.; Zhang, Q.; Smith, C. G.; Luo, J. Transient resistive switching devices made from egg albumen dielectrics and dissolvable electrodes. ACS Appl. Mater. Interfaces, 2016, 8, 10954-10960. (35) Wang, X.; Lu, X.; Liu, B.; Chen, D.; Tong, Y.; Shen, G. Flexible Energy‐Storage Devices: Design Consideration and Recent Progress. Adv. Mater. 2014, 26, 4763-4782. (36) Hosseini, N. R.; Lee, J. S. Biocompatible and Flexible Chitosan‐Based Resistive Switching Memory with Magnesium Electrodes. Adv. Funct. Mater. 2015, 25, 5586-5592. (37) Hu, Y.; Zhang, S.; Miao, X.; Su, L.; Bai, F.; Qiu, T.; Liu, J.; Yuan, G. Ultrathin Cs3Bi2I9 Nanosheets as an Electronic Memory Material for Flexible Memristors. Adv. Mater. Interfaces. 2017, 4, 1700131. (38) Yan, X.; Zhou, Z.; J, Zhao.; Liu, Q.; Wang, H.; Yuan, G.; Chen, J. Flexible memristors as electronic synapses for neuro-inspired computation based on scotch tape-exfoliated mica substrates. Nano Res. 2018, 11, 1183-1192. (39) Burgoyne, R. D. Neuronal calcium sensor proteins: generating diversity in neuronal Ca2+ signalling. Nature Reviews Neuroscience. 2007, 8, 182. (40) Clapham, D. E. Calcium signaling. Cell. 2007, 131, 1047-1058. (41) Puskas, J. E. Introduction to polymer chemistry: A biobased approach. DEStech Publications, Inc, 2013. (42) Wang, Z. Q.; Xu, H. Y.; Li, X. H.; Yu, H.; Liu, Y. C.; Zhu, X. J. Synaptic learning and memory functions achieved using oxygen ion migration/diffusion in an amorphous InGaZnO memristor. Advanced Functional Materials. 2012, 22, 2759-2765. (43) Cheng, W.; Liang, R.; Tian, H.; Sun, C.; Jiang, C.; Wang, X.; Wang, J.; Ren, T.; Xu, J. Proton conductor gated Synaptic transistor based on transparent IGZO for realizing electrical and UV light stimulus. IEEE Journal of the Electron Devices Society. 2018. (44) Yin, J.; Zeng, F.; Wan, Q.; Li, F.; Sun, Y.; Hu, Y.; Liu, J.; Li, G.; Pan, F. Adaptive Crystallite Kinetics in Homogenous Bilayer Oxide Memristor for Emulating Diverse Synaptic Plasticity. Advanced Functional Materials. 2018, 28, 1706927.


Page 18 of 20

Page 19 of 20 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


(45) Midya, R.; Wang, Z.; Zhang, J.; Savel'ev, S. E.; Li, C.; Rao, M.; Jang, M.; Joshi, S.; Jiang, H.; Lin, P.; Norris, K.; Ge, N.; Wu, Q.; Barnell, M.; Li, Z.; Xin, H.; Williams, S.; Xia, Q. Anatomy of Ag/Hafnia‐based selectors with 1010 nonlinearity. Advanced Materials. 2017, 29(12), 1604457. (46) Zhu, X.; Du, C.; Jeong, Y.; Lu, W. D. Emulation of synaptic metaplasticity in memristors. Nanoscale. 2017, 9, 45-51. (47) Liu, G.; Wang, C.; Zhang, W.; Pan, L.; Zhang, C.; Yang, X.; Fan, F.; Chen, Y.;

Li, R. W.

Organic biomimicking memristor for information storage and processing applications. Advanced Electronic Materials. 2016, 2, 1500298. (48) Chen, Y. C.; Yu, H. C.; Huang, C. Y.; Chung, W. L.; Wu, S. L.; Su, Y. K. Nonvolatile biomemristor fabricated with egg albumen film. Scientific reports. 2015, 5, 10022. (49) Wang, Z.; Joshi, S.; Savel’ev S, E.; Jiang, H.; Midya, R.; Lin, P.; Hu, M.; Ge, N.; Strachan, J. P.; Li, Z.; Wu, Q.; Barnell, M.; Li, G. L.; Xin, H. L.; Williams, R. S.; Xia, Q.; Yang, J. J. Memristors with diffusive dynamics as synaptic emulators for neuromorphic computing. Nature materials. 2017, 16, 101. (50) Linn, E.; Rosezin, R.; Kügeler, C.; Waser, R. Complementary resistive switches for passive nanocrossbar memories. Nature materials. 2010, 9, 403. (51) Wedig, A.; Luebben, M.; Cho, D. Y.; Moors, M.; Skaja, K.; Rana, V.; Hasegawa, T.; Adepalli, K, K.; Yildiz, B.; Waser. R.; Valov, I. Nanoscale cation motion in TaOx, HfOx and TiOx memristive systems. Nature nanotechnology. 2016, 11, 67 (52) Nardi, F.; Balatti, S.; Larentis, S.; Ielmini, D. Complementary switching in metal oxides: Toward diode-less crossbar RRAMs. In Electron Devices Meeting (IEDM), 2011 IEEE International. IEEE. 2011, 31.1.1-31.1.4



The Table of Contents The memristive device with the organic material egg albumen as the resistive layer exhibits excellent electrochemical properties and biomimetic properties. The material has the advantages of flexibility, high light transparency and degradability, and it can well simulate the important synaptic behavior of the nervous system. The formation and fracture of metal conductive channels are responsible for the memristor mechanism according to the conduction transport analysis. Keyword: memristor; flexible; egg albumen; synapse; dissolvable Mr. Xiaobing Yan*#, Xiaoyan Li,# Zhenyu Zhou, Jianhui Zhao, Hong Wang, Jingjuan Wang, Lei Zhang, Deliang Ren, Xin Zhang, Jingsheng Chen, Chao Lu, Qi Liu A Flexible Transparent Organic Artificial Synapse based on Tungsten /Egg Albumen/Indium Tin Oxide/polyethylene Terephthalate Memristor

Yan et al. ToC Figure


Page 20 of 20

A Flexible Transparent Organic Artificial Synapse ... - ACS Publications

Recommend Documents