Memristive Synapses with Photoelectric Plasticity Realized in ZnO1âx

Subscriber access provided by READING UNIV

Article

Memristive synapses with photoelectric plasticity realized in ZnO1-x/AlOy heterojunction Dan-Chun Hu, Rui Yang, Li Jiang, and Xin Guo ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b01036 • Publication Date (Web): 01 Feb 2018 Downloaded from http://pubs.acs.org on February 1, 2018

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 30 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

Memristive synapses with photoelectric plasticity realized in ZnO1-x/AlOy heterojunction Dan-Chun Hu, Rui Yang*, Li Jiang, Xin Guo* Laboratory of Solid State Ionics, School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, P. R. China

________________________ * Authors to whom correspondence should be addressed. E-mail: [email protected]; [email protected]

1

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 With the end of Moore’s law in sight, new computing architectures are urgently needed to satisfy the increasing demands for big data processing. Neuromorphic architectures with photoelectric learning capability are good candidates for energy efficient computing for recognition and classification tasks. In this work, artificial synapses based on the ZnO1-x/AlOy heterojunction were fabricated and the photoelectric plasticity was investigated. Versatile synaptic functions such as photoelectric short-term/long-term plasticity, paired-pulse facilitation, neuromorphic facilitation and depression were emulated based on the inherent persistent photoconductivity and volatile resistive switching characteristics of the device. It is found that the naturally formed AlOy layer provides traps for photogenerated holes, resulting in a significant persistent photoconductivity effect. Moreover, the resistive switching can be attributed to the electron trapping/detrapping at the trapping sites in the AlOy layer.

Keywords: Photoelectric memrisitor, Artifical synapse, Heterojunction, Persistent photoconductivity, Volatile resisitve switching

2


Page 2 of 30

Page 3 of 30 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 In the last several decades, digital computers based on the Von Neumann architecture have been proved to be more powerful than the human brain in solving complex and well-structured mathematical problems.1 However, the von Neumann computing system, in which the memory and processor are physically separated, is less suitable for solving unstructured problems, and it consumes lots of energy in the course of processing large amounts of data.2 In contrast, in the human brain, the data are stored and processed simultaneously in the same place in the neural network, enabling the brain to quickly process a massive amount of information in parallel with low power consumption.3 The human brain outperforms digital computers in unstructured problems, such as recognizing various objects and visual information in complex environments. Therefore, to develop artificial neural networks that rival their biological counterpart is a new frontier in computing.4 In the human brain, there are ~1011 neurons and ~1015 synapses.5 A synapse is a conjunction of two neighboring neurons (presynaptic and postsynaptic neurons), as schematically illustrated in Figure 1a. Its weight, i.e., connecting strength, can be precisely modified according to the activities of pre- and post-synaptic neurons. This activity-dependent synaptic plasticity is the biological foundation of learning, forgetting and memory.6 In biological systems, most of the information is transferred though electric signals. However, neurons are also subject to light stimulation. For example, in the technique of photogenetics, light is employed to control genetically modified neurons to express the photosensitive ion channels.7The key reagents used in

3



the photogenetics are light-sensitive proteins. The neuron control is achieved by photo initiators, like channel rhodopsin and ancient rhodopsin. Optical recording of the neuronal activity can be mediated by calcium, vesicle release, neurotransmitters and membrane voltage.8 Many efforts have been made to emulate synaptic plasticity using micro- and nano-electronic devices.3, 9-25 In order to enable devices to emulate the highly efficient neuromorphic operations of the brain, the utilization of transistors,3, 10-11 memristors9, 12-14, 20-24

and atomic switches18-19, 25 has been proposed. Recently, memristors have

attracted intensive interests due to their intrinsically learning functions and compact structure similar to biological synapses. In the neuromorphic or synaptic devices reported in literatures, 26 the synaptic strength is altered by applying electric stimuli, but in this case, the speed of the neural operation is limited due to the bandwidth connection density trade-off. In contrast, a photonic neuromorphic system is a more favorable option to enhance the computing speed, since it has high bandwidth, low crosstalk, and low power computation.27-34 In previous works concerning photonic influence on the switching behavior, the light stimulation is employed only for multistage storage. Agnus et al.27 reported optically controlled programmable resistors based on carbon nanotubes in multiple input-output prototype circuits. Tan et al.31 demonstrated an optically writable and electrically erasable multilevel memory in ITO/CeO2-x/AlOy/Al devices. Ungureanu et al.32 achieved multilevel storage using Pd/Al2O3/SiO2/Si devices. It is noteworthy that an alumina layer usually exists in the reported optical memristive devices31-32, 35 and

4


Page 4 of 30

Page 5 of 30 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


photodetectors36-38. However, the role of the alumina layer playing in the photoelectric properties has not yet been clarified. Moreover, there are quite limit reports about mimicking neurological functions through photonic stimuli. Recently, Lee et al.39 used photonic signals solely to simulate neurological functions, representing a significant progress towards photonic synapses. However, to the best of our knowledge, the photoelectric synapse responding to both electric and photonic stimuli is yet to be achieved. Motivated by these considerations, we experimentally demonstrate the photoelectric memristive synapses based on ZnO1-x/AlOy heterojunctions. The microstructure, electric resistive switching behavior and photonic response of the present device are systemically investigated. Versatile synaptic functions are mimicked by applying electric and photonic stimuli, including photoelectric short-term/long-term plasticity, paired-pulse facilitation, neural facilitation and depression. In addition, it is found that the naturally formed alumina layer provides traps for photogenerated holes, resulting in a significant persistent photoconductivity effect.

2. Experimental SiO2/Si substrates were consecutively cleaned in acetone, isopropanol, and deionized water for 15 min each for 3 times. Then the substrates were blown dry by N2 flux and then rapidly transferred into the vacuum chamber. A 30 nm thick Al electrode was deposited by sputtering (Angstrom Engineering, Canada). After that, a ZnO layer with

5



a thickness of 50 nm was deposited by sputtering in a gas mixture of 50% Ar and 50% oxygen. Finally, ITO electrodes with a thickness of 100 nm and a diameter of 100 µm were fabricated by sputtering at 250 ℃. Since Al will be unavoidably oxidized by ZnO and ZnO will also be reduced by Al, an AlOy layer is expected at the ZnO/Al interface,31, 40-43 thus the prepared device is actually ITO/ZnO1-x/AlOy/Al, the structure of which is schematically illustrated in Figure 1b. All the electric measurements of the device were performed at room temperature in air using a semiconductor characterization system (Keithley 4200-SCS) connected with a Cascade SUMMIT 11000B semi-automatic probe station. The positive voltage was defined as the current flow through the ZnO layer from the top ITO electrode to the bottom Al electrode. The photonic responses of the device were recorded under illumination of 310 nm UV light (CEL-LEDS35).

3. Results and discussion Typical current-voltage (I-V) curves of the ZnO1-x/AlOy/Al heterojunction measured in darkness and under UV light illumination are shown in Figure 2a; the device performs similar resistive switching behaviors in darkness and under illumination. The voltage was swept in a sequence of 0 V→ -8 V → 0 V→ +5 V → 0 V at a sweep rate of 0.02 V/s, triggering the reset process from the low resistance state (LRS) to the high resistance state (HRS) at negative biases, and the set process from HRS to LRS at positive biases. We also measured the endurance of the device (Supporting Information Figure S1); negligible degradation during consecutive 1000 switching cycles was observed, demonstrating an excellent endurance. Figure 2b reports the 6


Page 6 of 30

Page 7 of 30 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


retention characteristics of HRS and LRS in darkness and under illumination; HRS is volatile and gradually restores to LRS with time, while LRS is quite stable. It is worth noting that the retention characteristics of HRS and LRS of the present device are different from usual devices with volatile resistive switching behavior, in which LRS is volatile and HRS is stable. The volatile LRS was attributed to either the fracture of the conductive filaments due to the Joule heating44 or the migration of oxygen vacancies under the concentration gradient.45-46 However, the volatile HRS observed in the present device is proposed to be due to the electrons trapping by the AlOy layer, which will be explained later. The responses to photonic stimuli of LRS and HRS are also different, as shown in Figure 2c. Figure 2d shows the typical photo-response characteristics of the present device in LRS; the current of the device increases under the UV light exposure and then gradually decays over time after stopping stimuli, showing the persistent photoconductivity (PPC) behavior.39 Photonic neuromorphic functions are realized based on the PPC effect in the present device. The conductance of the device is considered as the synaptic weight (or strength), while the optical stimuli are treated as the neuron spikes. Spikes from the presynaptic neuron can be transmitted through the synapse to the postsynaptic neuron and generate an excitatory postsynaptic current (EPSC) or an inhibitory postsynaptic current (IPSC), whose sign and magnitude are determined by the connection strength, i.e., the synaptic weight. The basic operation principle of neural networks for learning and memory is based on the plasticity of synapses, where the synaptic strength or weight can be modulated in both short-term and long-term dynamics according to the

7



history of the stimulation. 47 By changing the frequency of stimulation, we can obtain short-term plasticity (STP), long-term plasticity (LTP) and the transition from STP to LTP. As shown in Figure 3a, the device performs STP, in which the synaptic weight sharply increases under illumination and then gradually decays to the original state after stimuli, when the input light pulses are at a low frequency of 0.1 Hz. However, when the frequency of the input light pulses increases to 1 Hz, the device exhibits LTP with the high-current state persisting for a long time. This transition from STP to LTP is similar to the consolidation process observed in biological systems.48 In addition to the photonic stimulus frequency, the energy density of the photonic pulses has also effects on the response of the device, as shown in Figure 3b. The transition from STP to LTP can also be achieved through increasing the energy density of the stimuli. Other stimulus parameters, including width and number, also affect the device response; details are given in the Supporting Information Figure S2. The device current accumulatively increases upon application of photonic stimuli, as shown in the left panel of Figure 4a. This photonic response of the device are quite similar to the potentiation effect observed in biological excitatory synapses.49 In addition, the depression effect in biological inhibitory synapses can be emulated using electric stimuli in the present device (see the right panel of Figure 4a); the current of the device gradually decreases by applying negative electric pulses. Moreover, the paired pulse facilitation (PPF), an important short-term plasticity observed in both biological excitatory and inhibitory synapses,

50-53

can be vividly emulated using the

present device, as reported in Figure 4b and 4c. PPF depicts the phenomenon that the

8


Page 8 of 30

Page 9 of 30 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


EPSC (IPSC) generated by the second pulse is larger than that of the first pulse and is determined by the time interval between the two pulses, when two consecutive presynaptic pulses are applied. A smaller interval usually leads to a larger EPSC (IPSC) amplitude enhancement. PPF can be manifested by the ratio between the EPSC amplitude measured immediately after the first pulse (A1) and the second pulse (A2), that is, the PPF index is A2/A1. The dependence of the PPF index on the pulse interval follows a double-exponential function: PPF = 1 + − ⁄ + − ⁄

(1)

where t is the pulse interval, C1 and C2 are the initial facilitation magnitudes of the respective phases, and τ1 and τ2 are the characteristic relaxation time of the respective phases. As shown in Figure 4b, the amplitude of the EPSC generated by the second presynaptic pulse is larger than that of the first one. Moreover, the dependence of the PPF index on the interval time (∆t) can be fitted very well by the double exponential function. The depression PPF induced by the electric stimuli are also in line with that observed in biological systems, as shown in Figure 4c. These results clearly indicate that the present device is an ideal synapse emulator with the photonic potentiation and electric depression effects. In biology, PPF is observed if the first action potential cannot release the transmitters at most publishing sites.50 Similar to that, it is suggested that the emulated PPF effect in the present device is achieved because a portion of holes excited by the first light are trapped in the AlOy layer. To understand the mechanisms of the PPC effect and the resistive switching behavior, the cross-section of the device was investigated by TEM and EDX, as

9



shown in Figure 1b and Figure 5. An interfacial layer ~20 nm thick is found at the interface between Al and ZnO. The EDX results clearly indicate that the interfacial layer is composed of Al and O. To disclose the valence distribution of Al in this layer, the XPS depth analysis of the interface was performed. The XPS data of Al 2p, O 1s and Zn 2p at various etching time are shown in Figure 6a; the contents of O and Zn decrease with increasing etching time, while the content of Al increases. Figure 6b shows the XPS depth profiling of the ZnO/Al interface. When the etching time is 100 s, Al appears and coexists with Zn, and the percentage of Al gradually increases. When the etching time is 300 s, only Al and O exist, indicating that Al captures O from ZnO to form AlOy, as a result, ZnO is reduced to ZnO1-x.38 The deconvolution fitting of the Al 2p signal at an etching time of 300 s clearly indicate that Al is oxidized at the ZnO/Al interface, as shown in Figure 6c. This is consistent with the TEM investigation. Furthermore, the variation of the relative contents of Al and Al3+ with etching time shown in Figure 6d, indicates that the formed interfacial layer is no-stoichiometric AlOy rather than stoichiometric Al2O3. There exist lots of defects in the AlOy layer; these defects, working as trapping sites, play a key role in the resistive switching and the PPC behavior.32, 35,55-56 From the TEM image, we can see that the thickness of the AlOy layer is relatively large, so that it is difficult for electrons to tunnel through. According to the energy levels of zinc oxide and alumina, the valence band offset between ZnO1-x/AlOy is less than the conduction band offset.54

35

The mechanism of the PPC effect of the

present device is explained in Figure 7a. In the initial equilibrium state, a built-in

10


Page 10 of 30

Page 11 of 30 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


electric field is formed at the interface of ZnO1-x/AlOy due to the difference in work function (denoted by a purple arrow).57 When the light stimulation is applied to the device, photons excite electrons to the conduction band to participate in conduction, leaving holes in the valence band. Therefore, the conductance of the device increases. Photo-generated holes accumulate at the interface and are trapped in the AlOy layer under the built-in electric field, as shown in the left part of Figure 7a. When the light source is removed, photo-generated holes move back across the interfical barrier under thermal excitation, which results in the PPC effect. The resistive switching process is schematically shown in Figure 7b. When a large negative voltage is applied to the device (the blue arrow direction), electrons injected from the ZnO1-x layer are trapped in the deep trapping sites in the AlOy layer, while electrons in the shallow trap near the Al electrode will be diverted into the Al electrode to participate in the conduction. This leads to the formation of an electron depletion layer in the AlOy layer, resulting in the high-resistance state (HRS) of the device. When the negative electric field is removed, free electrons in the Al electrode will be injected into the AlOy layer under thermal excitation, narrowing the depletion layer and resulting in the low-resistance state (LRS). Therefore, HRS is unstable and is easily lost to LRS, as shown in Figure 2b. Different responses to the light illumination for HRS and LRS can also be expained by the above proposed resistive switching mechanism. In LRS, photo-generated holes move to the AlOy layer and are trapped by the trapping sites in the AlOy layer. These photo-generated holes will move back to the original state after

11



removing the light. Thus, the current increases upon light illumination and then decreases after removing the light illumination in LRS. In contrast, no obvious current change is observed by turning on/off the light in HRS. This is because the current of HRS spontaneously increases after removing the electric filed due to the volatility of HRS. The photo-induced current modification is attenuated by the spontaneous current increase in HRS, because photo-generated electrons and the electrons from the Al electrode have the same drift direction. Therefore, the effect of light is different in LRS and HRS, as shown in Figure 2c. To verify the above mechanism, a sample of ITO/ZnO/Pt without the AlOy layer was fabricated for comparison. As shown in Figure 7c, the PPC effect of ITO/ZnO1-x/AlOy/Al device is much more obvious than that of the ITO/ZnO/Pt device, therefore, the AlOy layer really plays a key role in the PPC effect. The weak PPC effect observed in the ITO/ZnO/Pt device is related to the ionization of oxygen vacancies in the ZnO layer.57-60 In addition, the ITO/ZnO/Pt device also performs electric resistive switching behaviors, details are given in Supporting Information Figure S3. However, both LRS and HRS are volatile in the ITO/ZnO/Pt device (Figure 7d), which also demonstrates the importance of the AlOy layer in the volatile HRS of the ITO/ZnO1-x/AlOy/Al device.

4. Conclusions Based on the PPC effect and the resistive switching behavior in the ZnO1-x/AlOy heterojunction, we successfully emulated versatile synaptic functions including STP/LTP, PPF, neural facilitation and depression using both photonic and electric 12


Page 12 of 30

Page 13 of 30 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


stimuli. The AlOy layer at the ZnO/Al interface works as the trapping sites for holes and play a key role in PPC and the resistive switching. The present optoelectronic device with multifunctions has a very promising prospective in the applications of multidimensional neural networks, biological brain simulations and high-performance information storage and computing.

Associated content Supporting Information. Endurance of the ITO/ZnO1-x/AlOy/Al device; EPSC induced by presynaptic pulses with different (a) widths and (b) numbers; Typical I-V curves obtained for the ITO/ZnO/Pt and ITO/ZnO1-x/AlOy/Al devices.

Acknowledgements This work is supported by the National Natural Science Foundation of China (Grant No. 51772112, 51302095 and 51372094) and the Fundamental Research Funds for the Central Universities (HUST: 2016YXZD058 ).

13



References 1.

Von Neumann, J.; Morgenstern, O., Theory of Games and Economic Behavior.

Bull. Amer. Math. Soc 1945, 51 (7), 498-504. 2.

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, I.; Esser, S. K.; Appuswamy, R.; Taba, B.; Amir, A.; Flickner, M. D.; Risk, W. P.; Manohar, R.; Modha, D. S., A Million Spiking-neuron Integrated Circuit with a Scalable Communication Network and Interface. Science 2014, 345 (6197), 668-673. 3.

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. 4.

Machens, C. K., Building the Human Brain. Science 2012, 338 (6111),

1156-1157. 5.

Ananthanarayanan, R.; Esser, S. K.; Simon, H. D.; Modha, D. S. In The Cat is out

of the Bag: Cortical Simulations with 109 Neurons, 1013 Synapses, High Performance Computing Networking, Storage and Analysis, Proceedings of the Conference on, IEEE: 2009, pp 1-12. 6.

Yang, R.; Terabe, K.; Yao, Y.; Tsuruoka, T.; Hasegawa, T.; Gimzewski, J. K.;

Aono, M., Synaptic Plasticity and Memory Functions Achieved in a WO3-x-Based Nanoionics Device by Using the Principle of Atomic Switch Operation. Nanotechnology 2013, 24 (38), 384003. 7.

Mancuso, J. J.; Kim, J.; Lee, S.; Tsuda, S.; Chow, N. B.; Augustine, G. J.,

14


Page 14 of 30

Page 15 of 30 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


Optogenetic Probing of Functional Brain Circuitry. Exp Physiol. 2011, 96 (1), 26-33. 8.

Treger, J. S.; Priest, M. F.; Bezanilla, F., Single-Molecule Fluorimetry and Gating

Currents Inspire an Improved Optical Voltage Indicator. Elife 2015, 4, e10482. 9. Yang, R.; Huang, H.-M.; Hong, Q.-H.; Yin, X.-B.; Tan, Z.-H.; Shi, T.; Zhou, Y.-X.; Miao, X.-S.; Wang, X.-P.; Mi, S.-B.; Jia, C.-L.; Guo, X., Synaptic Suppression Triplet-STDP Learning Rule Realized in Second-Order Memristors. Adv. Funct. Mater. 2017, 1704455 10. Shi, J.; Ha, S. D.; Zhou, Y.; Schoofs, F.; Ramanathan, S., A Correlated Nickelate Synaptic Transistor. Nat. Commun. 2013, 4,2676. 11. 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 (12), 1693-1698. 12. Tuma, T.; Pantazi, A.; Le Gallo, M.; Sebastian, A.; Eleftheriou, E., Stochastic Phase-Change Neurons. Nature Nanotech. 2016, 11 (8), 693-699. 13. Eryilmaz S B, Kuzum D, Jeyasingh R, et al. Brain-like Associative Learning Using a Nanoscale Non-Volatile Phase Change Synaptic Device Array. Front. Neuroinformatics, 2014, 8,205. 14. Ambrogio, S.; Ciocchini, N.; Laudato, M.; Milo, V.; Pirovano, A.; Fantini, P.; Ielmini, D., Unsupervised Learning by Spike Timing Dependent Plasticity in Phase Change Memory (PCM) Synapses. Front. Neuroinformatics 2016, 10.25. 15. Prezioso, M.; Merrikh-Bayat, F.; Hoskins, B.; Adam, G.; Likharev, K. K.; Strukov, D. B., Training and Operation of an Integrated Neuromorphic Network Based on Metal-Oxide Memristors. Nature 2015, 521 (7550), 61-64.

15



16. 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 (3), 2203-2211. 17. Alibart, F.; Zamanidoost, E.; Strukov, D. B., Pattern Classification by Memristive Crossbar Circuits Using Ex situ and in Situ Training. Nat. Commun. 2013, 4,2072. 18. Ohno, T.; Hasegawa, T.; Tsuruoka, T.; Terabe, K.; Gimzewski, J. K.; Aono, M., Short-Term Plasticity and Long-Term Potentiation Mimicked in Single Inorganic Synapses. Nature Mater. 2011, 10 (8), 591-595. 19. La Barbera, S.; Vuillaume, D.; Alibart, F., Filamentary Switching: Synaptic Plasticity through Device Volatility. ACS Nano 2015, 9 (1), 941-949. 20. Alibart, F.; Pleutin, S.; Bichler, O.; Gamrat, C.; Serrano-Gotarredona, T.; Linares-Barranco, B.; Vuillaume, D., A Memristive Nanoparticle/Organic Hybrid Synapstor for Neuroinspired Computing. Adv. Funct. Mater. 2012, 22 (3), 609-616. 21. 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 (13), 2759-2765. 22. Krzysteczko, P.; Munchenberger, J.; Schafers, M.; Reiss, G.; Thomas, A., The Memristive Magnetic Tunnel Junction as a Nanoscopic Synapse-Neuron System. Adv. Mater. 2012, 24 (6), 762-6. 23. Jo, S. H.; Chang, T.; Ebong, I.; Bhadviya, B. B.; Mazumder, P.; Lu, W., Nanoscale Memristor Device as Synapse in Neuromorphic Systems. Nano Lett. 2010, 10 (4), 1297-301.

16


Page 16 of 30

Page 17 of 30 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


24. Zhang, C.; Tai, Y.-T.; Shang, J.; Liu, G.; Wang, K.-L.; Hsu, C.; Yi, X.; Yang, X.; Xue, W.; Tan, H.; Guo, S.; Pan, L.; Li, R.-W., Synaptic Plasticity and Learning Behaviours in Flexible Artificial Synapse Based on Polymer/Viologen System. Journal of Materials Chemistry C 2016, 4 (15), 3217-3223. 25. F Pan, S Gao, C Chen, C Song, F Zeng, Recent Progress in Resistive Random Access Memories: Materials, Switching Mechanisms, and Performance. Materials Science and Engineering: R: Reports, 2014, 83: 1-59. 26. Benner, A. F.; Ignatowski, M.; Kash, J. A.; Kuchta, D. M.; Ritter, M. B., Exploitation of Optical Interconnects in Future Server Architectures. IBM Journal of Research and Development 2005, 49 (4.5), 755-775. 27. Agnus, G.; Zhao, W.; Derycke, V.; Filoramo, A.; Lhuillier, Y.; Lenfant, S.; Vuillaume, D.; Gamrat, C.; Bourgoin, J. P., Two-Terminal Carbon Nanotube Programmable Devices for Adaptive Architectures. Adv. Mater. 2010, 22 (6), 702-706. 28. Bera, A.; Peng, H.; Lourembam, J.; Shen, Y.; Sun, X. W.; Wu, T., A Versatile Light-Switchable Nanorod Memory: Wurtzite ZnO on Perovskite SrTiO3. Adv. Funct. Mater. 2013, 23 (39), 4977-4984. 29. Park, J.; Lee, S.; Yong, K., Photo-Stimulated Resistive Switching of ZnO Nanorods. Nanotechnology 2012, 23 (38), 385707. 30. Sun, B.; Li, X.; Liang, D.; Chen, P., Effect of Visible-Light Illumination on Resistive Switching Characteristics in Ag/Ce2W3O12/FTO Devices. Chem. Phys. Lett. 2016, 643, 66-70.

17



Page 18 of 30

31. Tan, H.; Liu, G.; Zhu, X.; Yang, H.; Chen, B.; Chen, X.; Shang, J.; Lu, W. D.; Wu, Y.; Li, R. W., An Optoelectronic Resistive Switching Memory with Integrated Demodulating and Arithmetic Functions. Adv. Mater. 2015, 27 (17), 2797-2803. 32. Ungureanu, M.; Zazpe, R.; Golmar, F.; Stoliar, P.; Llopis, R.; Casanova, F.; Hueso, L. E., A light-Controlled Resistive Switching Memory. Adv. Mater. 2012, 24 (18), 2496-2500. 33. Park, J.; Lee, S.; Lee, J.; Yong, K., A light Incident Angle Switchable ZnO Nanorod Memristor: Reversible Switching Behavior Between Two Non-Volatile Memory Devices. Adv. Mater. 2013, 25 (44), 6423-6429. 34. Gao, S.; Song, C.; Chen, C.; Zeng, F.; Pan, F., Dynamic Processes of Resistive Switching in Metallic Filament-Based Organic Memory Devices. The Journal of Physical Chemistry C 2012, 116 (33), 17955-17959. 35. Li, H. K.; Chen, T. P.; Liu, P.; Hu, S. G.; Liu, Y.; Zhang, Q.; Lee, P. S., A Light-Stimulated Synaptic Transistor with Synaptic Plasticity and Memory Functions Based on InGaZnOx–Al2O3 Thin Film Structure. J. Appl. Phys. 2016, 119 (24), 244505. 36. Chen, C. Y.; Lin, C. A.; Chen, M. J.; Lin, G. R.; He, J. H., ZnO/Al2O3 Core-Shell Nanorod Arrays: Growth, Structural Characterization, and Luminescent Properties. Nanotechnology 2009, 20 (18), 185605. 37. Richters, J. P.; Voss, T.; Kim, D. S.; Scholz, R.; Zacharias, M., Enhanced Surface-Excitonic

Emission

in

ZnO/Al(2)O(3)

Nanotechnology 2008, 19 (30), 305202.

18


Core-Shell

Nanowires.

Page 19 of 30 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


38. Law, M.; Greene, L. E.; Radenovic, A.; Kuykendall, T.; Liphardt, J.; Yang, P., ZnO-Al2O3 and ZnO-TiO2 Core-Shell Nanowire Dye-Sensitized Solar Cells. J. Phys. Chem. B 2006, 110 (45), 22652-22663. 39. Lee, M.; Lee, W.; Choi, S.; Jo, J. W.; Kim, J.; Park, S. K.; Kim, Y. H., Brain-Inspired Photonic Neuromorphic Devices Using Photodynamic Amorphous Oxide Semiconductors and their Persistent Photoconductivity. Adv. Mater. 2017, 1700951. 40. Yang, R.; Li, X. M.; Yu, W. D.; Gao, X. D.; Liu, X. J.; Cao, X.; Wang, Q.; Chen, L. D., Stable Bipolar Resistance Switching Behaviour Induced by a Soft Breakdown Process at the Al/La0.7Ca0.3MnO3 Interface. J. Phys. D: Appl. Phys. 2009, 42 (17), 175408. 41. Yang, R.; Li, X. M.; Yu, W. D.; Liu, X. J.; Cao, X.; Wang, Q.; Chen, L. D., Multiform Resistance Switching Effects in the Al/La0.7/Ca0.3/MnO3/Pt Structure. Electrochem. Solid-State Lett. 2009, 12 (7), H281. 42. Yang, R.; Li, X. M.; Yu, W. D.; Liu, X. J.; Gao, X. D.; Wang, Q.; Chen, L. D., Resistance Switching Properties of La0.67Ca0.33MnO3 Thin Films with Ag–Al Alloy Top Electrodes. Appl. Phys. A 2009, 97 (1), 85-90. 43. Yang, R.; Li, X. M.; Yu, W. D.; Gao, X. D.; Shang, D. S.; Chen, L. D., Endurance Improvement of Resistance Switching Behaviors in the La0.7Ca0.3MnO3 Film Based Devices with Ag–Al Alloy Top Electrodes. J. Appl. Phys. 2010, 107 (6), 063703. 44. Van den Hurk, J.; Linn, E.; Zhang, H.; Waser, R.; Valov, I., Volatile Resistance States in Electrochemical Metallization Cells enabling Non-Destructive Readout of

19



Complementary Resistive Switches. Nanotechnology 2014, 25 (42), 425202. 45. Shi, T.; Yin, X.-B.; Yang, R.; Guo, X., Pt/WO3/FTO Memristive Devices with Recoverable Pseudo-Electroforming for Time-Delay Switches in Neuromorphic Computing. Phys. Chem. Chem. Phys. 2016, 18 (14), 9338-9343. 46. 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 (2), 377-84. 47. 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 (13), 133701. 48. Izquierdo, I.; McGaugh, J., Behavioural Pharmacology and its Contribution to the Molecular Basis of Memory Consolidation. Behav. Pharmacol. 2000, 11 (7-8), 517-534. 49. Salin, P. A.; Scanziani, M.; Malenka, R. C.; Nicoll, R. A., Distinct Short-Term Plasticity at Two Excitatory Synapses in the Hippocampus. Proc. Natl. Acad. Sci. 1996, 93 (23), 13304-13309. 50. Debanne, D.; Guerineau, N. C.; Gähwiler, B.; Thompson, S. M., Paired-Pulse Facilitation and Depression at Unitary Synapses in Rat Hippocampus: Quantal Fluctuation Affects Subsequent Release. J. Physiol. 1996, 491 (1), 163-176. 51. Hu, S. G.; Liu, Y.; Chen, T. P.; Liu, Z.; Yu, Q.; Deng, L. J.; Yin, Y.; Hosaka, S., Emulating the Paired-Pulse Facilitation of a Biological Synapse with a NiOx-Based Memristor. ApPhL 2013, 102 (18), 183510.

20


Page 20 of 30

Page 21 of 30 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


52. Zucker, R. S.; Regehr, W. G., Short-Term Synaptic Plasticity. Annu. Rev. Physiol. 2002, 64, 355-405. 53. Liu, Y. H.; Zhu, L. Q.; Feng, P.; Shi, Y.; Wan, Q., Freestanding Artificial Synapses Based on Laterally Proton-Coupled Transistors on Chitosan Membranes. Adv. Mater. 2015, 27 (37), 5599-604. 54. Kerber, A.; Cartier, E.; Degraeve, R.; Pantisano, L.; Roussel, P.; Groeseneken, G. In Strong Correlation between Dielectric Reliability and Charge Trapping in SiO2/Al2O3 Gate Stacks with TiN Electrodes, VLSI Technology, 2002. Digest of Technical Papers. 2002 Symposium on, IEEE: 2002; pp 76-77. 55. Specht, M.; Reisinger, H.; Hofmann, F.; Schulz, T.; Landgraf, E.; Luyken, R. J.; Rösner, W.; Grieb, M.; Risch, L., Charge Trapping Memory Structures with Al2O3 Trapping Dielectric for High-Temperature Applications. Solid-State Electron. 2005, 49 (5), 716-720. 56. Wang, J.; Wang, Z.; Huang, B.; Ma, Y.; Liu, Y.; Qin, X.; Zhang, X.; Dai, Y., Oxygen Vacancy Induced Band-Gap Narrowing and Enhanced Visible Light Photocatalytic Activity of ZnO. ACS Appl. Mat. Interfaces 2012, 4 (8), 4024-4030. 57. Mundle, R.; Carvajal, C.; Pradhan, A. K., ZnO/Al: ZnO Transparent Resistive Switching Devices Grown by Atomic Layer Deposition for Memristor Applications. Langmuir 2016, 32 (19), 4983-4995. 58. Ahn, S. E.; Song, I.; Jeon, S.; Jeon, Y. W.; Kim, Y.; Kim, C.; Ryu, B.; Lee, J. H.; Nathan, A.; Lee, S.; Kim, G. T.; Chung, U. I., Metal Oxide Thin Film Phototransistor for Remote Touch Interactive Displays. Adv. Mater. 2012, 24 (19), 2631-2636.

21



59. Jang, J. T.; Park, J.; Ahn, B. D.; Kim, D. M.; Choi, S. J.; Kim, H. S.; Kim, D. H., Study on the Photoresponse of Amorphous In-Ga-Zn-O and Zinc Oxynitride Semiconductor Devices by the Extraction of Sub-Gap-State Distribution and Device Simulation. ACS Appl. Mat. Interfaces 2015, 7 (28), 15570-15577. 60. Jeon, S.; Ahn, S. E.; Song, I.; Kim, C. J.; Chung, U. I.; Lee, E.; Yoo, I.; Nathan, A.; Lee, S.; Robertson, J.; Kim, K., Gated Three-Terminal Device Architecture to Eliminate Persistent Photoconductivity in Oxide Semiconductor Photosensor Arrays. Nat Mater. 2012, 11 (4), 301-305.

22


Page 22 of 30

Page 23 of 30 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 1. Memristive synapse based on the ZnO1-x/AlOy heterojunction. (a) Schematic illustration of biological neurons and synapses. (b) Schematic illustration of the ITO/ZnO1-x/AlOy/Al heterojunction. Inset is the cross-section TEM image of the device.

23



Figure 2. Resistive switching and photo response of the ZnO1-x/AlOy junction. (a) I-V curves. The black curve was measured in darkness, while the magenta curve was measured under UV-light illumination. (b) Retention characteristics of HRS and LRS in darkness and under UV-light illumination. (c) UV-light influence on the retention characteristics of HRS and LRS. (d) PPC effect. The read voltage is 0.1 V.

24


Page 24 of 30

Page 25 of 30 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. Photonic neural functions. The transition from LTP to STP realized by controlling (a) photonic pulse frequency and (b) power. The read voltage is 0.1 V.

25



Figure 4. Photonic and electrical synaptic functions. (a) Synaptic potentiation triggered by photonic stimuli (red curve) and synaptic inhabitation triggered by electric stimuli (black curve). Insets are the waveforms of the optical and electric stimuli. (b) Photonic PPF function. The inset shows the paired-pulse facilitation and a pair of pre-synaptic spikes and the triggered EPSC under an inter-spike interval of 5 s. PPF index, defined as the ratio of A2/A1, plotted as a function of inter-spike interval, ∆t, between two consecutive light spikes. (c) Electrical PPF function. The inset shows paired-pulse depression phenomenon. The read voltage is 0.1 V.

26


Page 26 of 30

Page 27 of 30 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 5. TEM image and EDX line-scan results for the cross-section of the device.

27



Figure 6. XPS investigation of the ZnO1-x/AlOy interface. (a) Al 2p, O 1s, Zn 2p. (b) Depth profiles of Al, O and Zn elements. (c) Al 2p XPS data taken from the ZnO1-x/AlOy interface. The peak at 73 eV (blue) is for metallic Al, the peaks at 74.4 eV (yellow) and 76 eV (green) are for Al3+. (d) Depth profiles of Al and Al3+.

28


Page 28 of 30

Page 29 of 30 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 7. Mechanisms of PPC and resistive swithing. Schematic illustrations of the band diagram of the ZnO1-x/AlOy junction (a) under illumination (left) and in darkness (right), (b) under negative electric bias (left) and without negative electric bias (right). Comparison of (c) the PPC effect and (d) the retention characteristics of the ITO/ZnO/Pt and ITO/ZnO1-x/AlOy/Al devices.

29



Table of Contents

30


Page 30 of 30

Memristive Synapses with Photoelectric Plasticity Realized in ZnO1âx

Recommend Documents

Memristive Synapses with Photoelectric Plasticity Realized in ZnO1âx