Atomic Scale Photodetection Enabled by a Memristive Junction - ACS

25 Jun 2018 - *E-mail: [email protected]., *E-mail: [email protected]. Cite this:ACS ... View: ACS ActiveView PDF | PDF | PDF w/ Links | Full Text ...
0 downloads 0 Views 4MB Size
Download PDF
www.acsnano.org

Atomic Scale Photodetection Enabled by a Memristive Junction Alexandros Emboras,*,† Alessandro Alabastri,‡ Fabian Ducry,§ Bojun Cheng,† Yannick Salamin,† Ping Ma,† Samuel Andermatt,§ Benedikt Baeuerle,† Arne Josten,† Christian Hafner,† Mathieu Luisier,§ Peter Nordlander,∥ and Juerg Leuthold*,† †

Institute of Electromagnetic Fields (IEF) and §Computational Nanoelectronics Group, ETH Zurich, 8092 Zurich, Switzerland Department of Electrical and Computer Engineering and ∥Department of Physics and Astronomy, Rice University, 6100 Main Street, Houston, Texas 77005, United States

Downloaded via UNIV PIERRE ET MARIE CURIE on August 25, 2018 at 10:03:32 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: The optical control of atomic relocations in a metallic quantum point contact is of great interest because it addresses the fundamental limit of “CMOS scaling”. Here, by developing a platform for combined electronics and photonics on the atomic scale, we demonstrate an optically controlled electronic switch based on the relocation of atoms. It is shown through experiments and simulations how the interplay between electrical, optical, and light-induced thermal forces can reversibly relocate a few atoms and enable atomic photodetection with a digital electronic response, a high resistance extinction ratio (70 dB), and a low OFF-state current (10 pA) at room temperature. Additionally, the device introduced here displays an optically induced pinched hysteretic current (optical memristor). The photodetector has been tested in an experiment with real optical data at 0.5 Gbit/s, from which an eye diagram visualizing millions of detection cycles could be produced. This demonstrates the durability of the realized atomic scale devices and establishes them as alternatives to traditional photodetectors. KEYWORDS: atomic contacts, photodetection, quantum plasmonics, memristor, surface plasmons, local oxidation, silicon photonics, ab initio calculation

F

are not aware of any previous demonstration of an integrated, atomic scale, and memristive photodetector. Here, we demonstrate such a component relying on the motion of individual atoms rather than on optically induced electron and hole currents. The photodetection concept is based on an atomic contact that upon illumination enables the reversible hopping of a few atoms between adjacent sites, which correspond to distinct electronic quantum states. Experimental results show that the two quantum states differ by a resistance extinction ratio of 70 dB despite the small dimension. The device features a low current of only a few tens of picoamperes in the OFF-state and operates reliably at room temperature. Noticeably, the device exhibits the characteristic hysteresis loop of a memristor that is actuated optically instead of electrically with reliable operations over millions of cycles, as demonstrated by a 0.5 Gbit/s data experiment. The reported experimental behavior agrees well with atomistic calculations and numerical modeling based on finite element methods that

ifty years after the introduction of Moore’s law, the fundamental scaling limit is still one of the most relevant items on the agenda of the electronics and photonics industry. Indeed, decreasing the device dimensions is the key to reduce their energy consumption and achieve a higher density of active components. The semiconductor industry is currently at the 10 nm technology node1 with transistor gate lengths on the order of 18−20 nm, but Moore’s law keeps pushing it toward the ultimate limit, the atom.2−12 In contrast, commercial photonic devices operating in the 1550 nm telecommunication window require hundreds of μm2 if not mm2 of footprint to properly work, a reality which is stimulating research on scaling photonics to the micrometer scale.13−15 Most recently, another generation of atomic scale electro-optical switches has emerged.16−21 In these devices, the optical properties are switched by electrically relocating individual atoms.16−19 Alternatively, light can be used as a catalyst to assist the initial filament formation of resistive switching devices, in particular, memristors,20,21 or as a reset stimulus to dissociate these filaments.22,23 Whereas research on electrically and/or optically controlled atomic scale switches7,8,11,16,17,19,24−27 is rapidly gaining momentum, we © 2018 American Chemical Society

Received: March 9, 2018 Accepted: June 25, 2018 Published: June 25, 2018 6706

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713

Article

Cite This: ACS Nano 2018, 12, 6706−6713

Article

ACS Nano

Figure 1. Optical detection with single atoms. (a) Schematic showing the concept of optical detection on atomic scale (center panel). The inset to the right shows the experimental setup used to perform the dynamic measurements (Vdc = 0.4 V, RL = 10 kΩ). The plasmonic hot spot at the tip focuses energy on the atomic quantum point contact and enables the reversible relocation of atoms between adjacent sites and thus the creation of two distinct quantum electronic states. A modulated input optical power (left panel) is converted into a digital electronic current (right panel). (b) Scanning electron microscope picture of the atomic photodetector. The inset shows an atomic force microscopy image of the device before the deposition of the metals and demonstrates a vertically oriented 3D waveguide tip.

and h is Planck’s constant. The fact that the device reversibly switches from the OFF- (1010 Ohms) to the ON-state with a quantized conductance on the order of a few G0 demonstrates the formation of an atomic scale contact.28 To enable the switching, the minimum required optical power that needs to be dissipated in the switching region is estimated to be on the order of 500 nW. The details of how this value was determined from the experiment are presented in the Supporting Information. Such an amount of optical power has been found to be sufficiently large to modulate the electronic current with a high extinction ratio of 70 dB at wavelengths around 1.55 μm. The atomic photodetector was fabricated on a 220 nm silicon-on-insulator wafer. A fabrication yield of 40% has been found; that is, 40% of the devices work as described above. A top view of the manufactured device is shown in Figure 1b. The fabrication consists of three electron-beam lithography steps on one chip. First, we patterned the silicon waveguides by using the modified local oxidation of silicon approach,29 which allows one to produce a 3D pyramid-like waveguide. An atomic force microscopy image of the 3D waveguide is given in the inset of Figure 1b. Second, a 100 nm wide Pt nanowire was evaporated and processed by means of a lift-off process. Subsequently, a 20 nm thick plasma-enhanced chemical vapor deposition α-SiO2 layer was deposited everywhere. The α-SiO2 acts as a diffusion matrix for the silver ions. It is worth noting that the thickness of the SiO2 layer was carefully chosen in order to minimize the optical loss while keeping the operation voltage in a reasonable range. Then, a 2 μm wide silver layer was added using e-beam evaporation and again a lift-off process. The Ag and Pt electrodes overlap over a 20 nm region

highlight the interplay between electrical, optical, and lightinduced thermal forces and show how these forces may relocate atoms, thereby switching an electric current between OFF- and ON-states.

RESULTS AND DISCUSSION To appreciate how the device operates, we refer to Figure 1a (center panel). The light is first propagated into a silicon waveguide and piped within 2 μm into a vertical plasmonic slot waveguide. The latter is then squeezed within 2 μm laterally and vertically to create a pyramid-like 3D plasmonic tip consisting of Ag-α-SiO2-Pt. An offset voltage (Vdc) is applied to the metallic tip and forms an atomic scale filament that short circuits the two metals.8 Photodetection then effectively happens in the tiny junction of the plasmonic tip. The detection concept can be summarized as follows: the localization of the light at the plasmonic tip creates a hot spot (both high optical field strengths and temperature) that allows for the reversible hopping of a few atoms between neighbor sites. This gives rise to two distinct electronic quantum states. Figure 1a shows how an optical stimulus that is fed into the device (left panel) and the resulting electric field can reversibly switch the resistance of the device (right panel) between these two states. Specifically, when light is turned on, the resistance dramatically increases and the device enters a tunneling regime due to the partial dissolution of the Ag nanofilament. When the light is turned off, the atomic contact is re-established, and the resistance abruptly decreases by several orders of magnitude. The corresponding conductance reaches a value that is on the order of the quantum conductance unit G0 = 2e2/h, where e is the electron charge 6707

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713

Article

ACS Nano

Figure 2. Atomic optical memristor: Memristive current as a function of the optical input power (black and red solid circles are experiments, and blue circles represent simulations). The device features memristive hysteretic behavior when the optical stimuli is varied from −25 to 10 dBm. The on-resistance is controlled by the compliance current Icc = 10 μA. The inset areas in A−D show the different phases of the switching, the applied forces (measured at the point where the junction is broken; see Supporting Information), as well as the simulated conductivity of the junction. Area (A) shows the ON-state where the filament is completely formed and the device is conductive. Area (B) shows the RE-SET state where the dissociation of the filament is initiated and the conductivity starts to decrease. The thermal diffusion becomes dominant in this state. Area (C) shows the OFF-state where the filament is completely dissociated and the conductivity is abruptly decreased. In this case, the thermal diffusion competes with the electrical and optical force, leading to current oscillations. Area (D) shows the SET-state where the filament is re-formed and the conductivity abruptly increased. In this case, the electrical force is dominating the dynamics and driving the filament to the complete reconstruction.

conductive filament decrease and become weak, as soon as an electrical contact is formed (Figure 2A). After illumination with increasing power, Psig, the state of the device moves to area (B) on the hysteresis curve. Increasing the power of the optical signal results in the creation of a hot spot around the filament. Local heating may strongly enhance electrochemical processes and cause filament dissolution.34,35 Consequently, atoms undergo oxidation and two effects set in: a thermally induced lateral diffusion of Ag ions, which depends on an effective diffusive force FD and a rather weak optical force Fopt. Particularly, the diffusive force dominates the drift of Ag ions since the Vdc-induced electric force, Fel, is negligible in the conducting filament state: most of the applied voltage drops over the resistance put in series with the photodetector (see Supporting Information). The force fields are illustrated in Figure 2B. At the threshold of 6 dBm, the effective diffusion force FD ∼ D(T) × ∇CAg, where D(T) is the temperature-dependent diffusion constant of the Ag ions and C Ag is their concentration, dominates over the DC electrical force Fel and Fopt and pushes the Ag ions away from the filament. The optical force Fopt is aligned with the direction of the gradient of the electromagnetic field (optical tweezer effect), that is, toward the filament. Upon the relocation of the Ag ions, the filament dissociates, the electrical current abruptly drops, the electric field sets back in, and the system is moved into area

at their closest location. The exact geometry of the fabricated device at the level of the junction is given in the Supporting Information. Finally, the atomic scale quantum point contact was built up at the level of the tip by employing the electrochemical metallization30 technique that is widely used in memristive technology.31,32 The final device is operated by applying a positive voltage to the Ag electrode, whereas the Pt electrode is grounded. The induced electric field ionizes Ag atoms that become mobile in the SiO2 matrix and are converted back to Ag atoms by capturing free electrons coming from the Pt contact.31,32 Thus, a Ag-based atomic contact is progressively formed, starting from the Pt contact in the present experiments. Because the underlying formation processes are field-dependent,30 it is expected that a single atomic junction is nucleated and grows at the tip where the electric field is maximum. The operation principle is revealed by measuring the current as a function of the optical input power injected into the active area of the device (Figure 2). The plot shows the pinched characteristic of a memristive device33 triggered by an optical input power rather than an electrical signal. To achieve this feature, the device is first electrically switched to a quantum conductance of a few G0, which corresponds to a stable Ag−Pt point contact as thick as a few atoms; see area (A) on the center figure.5,28 The switch is then in a stable quantum state. The Vdc-induced electric field and force Fel across the 6708

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713

Article

ACS Nano

Figure 3. Switching dynamics of the optically induced filament dissociation showing quantized transitions. (a) (Left panel): Distinct traces of the optically induced evolution of the nanofilament resistance at room temperature. (Middle panel): Zoomed-in image of the area of abrupt resistance changes with the corresponding conductance levels. (Right panel): Enlarged zoomed-in image shows that the fastest switching speed is on the order of 10 μs. The OFF-state represents the noise level of the measurement setup. (b) Histogram of optically induced conductance quantization.

based memristors,38 and was here extended to metallic conductive filaments. The blue line in Figure 2 shows the result of the full “electric current versus optical power” simulation. It can be seen that the theoretical calculation (blue line) reproduces the experimental findings well. Of particular interest is the concentration of the Ag atoms that ultimately define the conductivity of the filament: the modeling approach considers Ag as a diffusive species which, depending on its concentration, behaves as ions (thus being pulled or pushed by electric fields) or as neutral atoms (therefore mimicking the conductive filament). This is implicitly obtained through a concentration-dependent conductivity which, by means of a threshold, modulates the response of the species to electric fields. See eqs S3 and S6 and Figure S4 for additional details. The conductivity plots are depicted as insets in Figure 2A−D (red indicates a large conductivity, blue a low one). Initially, in the device ON-state, a highly conductive (red color) filament channel is visible. Upon increasing the optical power, electromagnetic dissipation increases at the contact, leading to a temperature increase that enhances the diffusion coefficient for the Ag atoms (see Figure 2B). Upon reaching a critical temperature, the ions diffuse away from the junction and the conductive bridge is broken, resulting in a current drop (see the conductivity plot in Figure 2C). When this happens, the heat dissipation decreases and the diffusive force is reduced. At the same time, the electric field sets in again, and wherever Ag concentration is low enough, the model considers it as composed of positively charged ions, thus responding to electrical forces. Upon the optical input power being further reduced, the diffusion vanishes and the force from the electric DC field rebuilds a conductive Ag filament (Figure 2A). To further investigate the filament dissociation shown in Figure 2B,C, we performed time-resolved experiments. To reveal the dynamics, an optical signal with a nanosecond rise time was fed into the device, and the electrical resistance was measured with microsecond resolution, as illustrated in Figure

(C) of the curve. Note that the transition from the ON- to the OFF-state exhibits several conductive plateaus, which indicates that the dissociating filament may go through several metastable states before reaching a stable open-circuit configuration, as depicted schematically in Figure 2C. Simulations indicate that the temperature in the filament decreases so that thermal lateral diffusion processes are reduced (see Figure S8a−d of section 1.4 in the Supporting Information). Simultaneously, the dissolution of the conductive filament moves back the voltage drop to the memristive device and reincreases the drift force Fel caused by the electric field. The latter pushes the Ag ions toward the Pt electrode, whereas the remaining FD still prevents any filament re-formation. The optical force Fopt, if relevant at all, would push the Ag ions toward the tips of the structure, as does Fel. As the optical signal intensity is reduced, the system transits to area (D) where the driving forces for lateral thermal diffusion, FD, and the optical force, Fopt, vanish. The Ag ions are now only exposed to the external electrical force Fel (see Figure 2D). When the optical power reaches a value below −10 dBm, the Ag ions undergo reduction and a nanofilament is formed again (area (A)), which ultimately leads to a sudden electrical current increase. The resulting hysteretic current generated by the sweep of the optical input power is inherently related to the efficient optical-induced mechanical atomic relocations (atomic optomechanics). This is in sharp contrast to conventional optoelectronic devices where light interacts by means of electrons (optoelectronics). In order to more deeply analyze the SET and RE-SET dynamics of the fabricated device, we designed an integrated simulation environment based on the finite element method combining ionic transport, thermal distribution, electrical current, and electromagnetic radiation (see Supporting Information for a more detailed description). This approach is inspired by previous works,36,37 more specifically on oxide6709

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713

Article

ACS Nano

Figure 4. Atomistic ab initio calculations on the filament dissociation showing thermally induced quantized diffusion. (a) Investigated structure consisting of two metallic electrode and a nanofilament that is embedded into an α-SiO2 matrix. (b) Close-up view of the filament tip. The various plots show the positions of the three atoms (red, yellow, and blue) with time. (c) Plot of the displacement of the three atoms with time. It can be seen that the diffusion of atoms is not continuous but rather takes place in discrete steps.

Figure 5. Eye diagram and data transmission setup for testing the atomic scale detector at 0.5 Gbit/s. (a) Random bit-pattern at 0.5 Gbit/s is generated with an offline DSP and encoded by means of a commercial modulator onto a laser signal. The signal is then detected in the atomic scale photodetector. (b) Transmission experiment shows a clear eye diagram. From the sampled data, a bit-error ratio on the order of 8.9 × 10−3 has been found.

preferable atomic configuration. A histogram of the quantum conductance values for several consecutive optically induced current variations is reported in Figure 3b. The histogram spans conductance values from a fraction of G0 up to multiples. Peaks around integer multiples of G0 (G0, 2G0, 4G0, 5G0) are clearly visible, demonstrating that the fabricated Ag-αSiO2-Pt device operates at the atomic scale and involves few-atom relocations. The observed small deviations of the conductance from integer values of G0 can be explained by contact resistance effects and defects and impurities at the Ag/Pt interface, which induce scattering of electron waves and energy loss.3 To shed more light onto the underlying principle of the atomistic filament dissociation, we performed calculations based on Ehrenfest molecular dynamics (EMD) and ab initio molecular dynamics (AIMD). A detailed explanation of the simulation setup is given in the Supporting Information. The investigated structure consists of a nanofilament placed between two metallic electrodes embedded in α-SiO2 (Figure 4a). To reveal the effect of the local heating on the relocation of atoms, we use the AIMD approach. A heat source of 800 K was added to emulate the electromagnetic absorption at the tip of the structure (the motivation of using such high temperature

3. The experimental setup used to perform such pulsed measurement is illustrated in the inset of Figure 1a. Two transient regimes could be identified, a slow and a fast one, indicating that the optically induced reset process depends on two distinct effects. The slow transient regime occurring on the millisecond scale is the time to initiate the device conductance changes and is marked as t1 in Figure 3 (left panel). The faster time constant t2 is on the order of microseconds (Figure 3, right panel) and is attributed to the lateral migration of Ag ions away from the filament tip and into the surrounding SiO2 matrix (i.e., Ag ions preferably diffuse toward less Ag-populated areas). This effect, as shown also by our calculations, leads to a dissolution of the Ag filament that finally ruptures the conductive channel.39−41 The experiment further confirms the atomic scale operations by showing the discrete conductance levels,42,43 as can be seen in the middle panel of Figure 3a. It can be observed that the device resistance passes through many metastable steps during an optical pulse injection. Clear plateaus and abrupt changes with discrete values that are close to an integer or multiple integers of G0 can be identified (Figure 3a, middle panel). Such conductance plateaus are inherently related to step-by-step atom dissociations,3,44 and each step corresponds to an energetically 6710

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713

Article

ACS Nano

optimization of such a technology platform will allow digital photodetectors to be built and may find applications in systems where optical signals should be directly converted into digital ones (with states “0” and “1”) without the need for decision circuits or analog-to-digital converters.

is explained in the Supporting Information). It can be seen that the resulting displacement of atoms as a function of the time are not continuous but rather discrete (Figure 4b), similar to the experimental measurements in Figure 3a (middle panel). In particular, the atoms are diffusing and get an energetically favorable position on the picosecond time scale when they receive a specific portion of quantized energy. To investigate the effect of the optical forces (optical tweezer effect), the EMD approach is used. Our calculations show that the forces directly caused by the optical field are weaker than the thermal forces that exist at room temperature. Thus, it appears that the optical forces in the present device are not sufficiently high to rearrange the configuration of the atoms at the tip and to induce the observed resistance changes. Finally, the reliability of the atomic quantum point photodetector was confirmed by conducting data modulation experiments as in Figure 5. The eye diagram demonstrates the durability of the device over ∼240000 consecutive cycles. For the experiment, we generated a 0.5 Gbit/s electrical signal comprising a random bit sequence by means of an offline digital signal processing (DSP) unit and a high-speed digitalto-analog converter. The electrical data stream was amplified with an RF amplifier and fed into a commercial intensity modulator which converted the continuous-wave laser at 1310 nm to an optical data stream. The optical data were then detected by the atomic scale photodetector, and the electrical signal was sampled in a real-time oscilloscope. We measured a bit-error ratio of 8.9 × 10−3 after applying standard telecom equalization techniques. This error ratio is well below the soft forward error correction limit. This suggests possible error-free transmissions with a 20% overhead.45 It is worth mentioning that in this experiment our device was operated in the ONstate with an atomic filament of 7−8 G0. The optical input power in the waveguide was detected before the device was experimentally tested with reference waveguides and found to be on the order of 0 ± 1 dBm. Based on simulations of the fabricated structure, the power of the plasmonic mode at the tip can be estimated to reach 400 nW with an uncertainty of 300 nW that comes from fabrication intolerances, surface roughness, or alignment issues. The experiment has been performed at 1310 nm because the light was coupled from the fiber to the waveguide through grating couplers optimized for 1310 nm. However, the device is expected to properly operate at 1550 nm, too.

METHODS Measurements. The DC electrical characterizations were performed using an Agilent B2912A precision source/measurement unit. Two different Agilent LASER sources (8164B) emitting from 1260−1370 and 1460−1640 nm were used for the electro-optical testing. For the data experiment, a DSP unit and a high-speed digitalto-analog converter was used. The detected data were measured with a Digital storage osciloscope (DSO-X-96204Q). The Fib/scanning electron microscope images of the devices were acquired by a Helios Nanolab 450S. A high-resolution scanning electron microscope was also used (SU8230). Atomistic Calculations. Two different approaches were used to shed light into the various physical properties of the atomic photodetector. The first approach comprised ground-state Born− Oppenheimer molecular dynamics coupled with the nonequilibrium Green’s function formalism, and the second approach was EMD simulations. More details are given in the Supporting Information.

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.8b01811. Modeling of the filament dynamics of the optical memristor (introduction, 2D electromagnetic, and 2D axisymmetric electric simulations, diffusive, electrical and thermal simulations, additional results from EEOC and FDC cluster calculations, 3D electromagnetic model for propagation losses); ab initio atomistics calculation (electronic simulation overview), methods (description of the system and the setup), results (plasmonic forces, thermal effects, electron transport), conclusion; additional electro-optical experimental results (electrical current−voltage characteristics); geometrical parameters of the fabricated devices (PDF)

AUTHOR INFORMATION Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected].

CONCLUSIONS In summary, we experimentally demonstrated a photodetection concept based on an optical-induced relocation of individual atoms in a metallic quantum point contact at room temperature. The efficient interaction of photons with atom transport results in a pinched hysteretic electronic current from which an optical memristor behavior is derived. Inspired by the optical memristance concept, we showed a practical photodetector device application. Experimental results indicate a high resistance extinction ratio (70 dB) that emanates from the interplay of atomic and electronic transport. The device features a low OFF-state current of tens of picoamperes and operates reliably for weeks. The detector was also tested with a 0.5 Gbit/s data signal. Open eyes with low bit-error ratios are reported. These results have been enabled by a technology platform for atomic scale electronics and photonics that relies on the coexistence of an optical hot spot and an atomic filament at a 3D vertically oriented plasmonic tip. The

ORCID

Alexandros Emboras: 0000-0002-8843-5118 Alessandro Alabastri: 0000-0001-6180-8052 Bojun Cheng: 0000-0001-5551-4242 Yannick Salamin: 0000-0003-1797-2968 Ping Ma: 0000-0003-0674-8530 Peter Nordlander: 0000-0002-1633-2937 Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS The Volkswagen Stiftung, the Werner Siemens Stiftung Centre for Single-Atom Electronics and Photonics, and ETH Zurich under Grant No. 3515-2 are acknowledged for partial funding. The work was carried out at the Binnig and Rohrer Nanotechnology Center, Switzerland. It used computational resources from the Swiss National Supercomputing Centre 6711

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713

Article

ACS Nano

(21) Hino, T.; Tanaka, H.; Hasegawa, T.; Aono, M.; Ogawa, T. Photoassisted Formation of an Atomic Switch. Small 2010, 6, 1745− 1748. (22) Zhu, X.; Lee, J.; Lu, W. D. Iodine Vacancy Redistribution in Organic−Inorganic Halide Perovskite Films and Resistive Switching Effects. Adv. Mater. 2017, 29, 1700527. (23) Kawashima, T.; Zhou, Y.; Yew, K. S.; Ang, D. S. Optical Reset Modulation in the Sio2/Cu Conductive-Bridge Resistive Memory Stack. Appl. Phys. Lett. 2017, 111, 113505. (24) Natterer, F. D.; Yang, K.; Paul, W.; Willke, P.; Choi, T.; Greber, T.; Heinrich, A. J.; Lutz, C. P. Reading and Writing Single-Atom Magnets. Nature 2017, 543, 226−228. (25) Savage, K. J.; Hawkeye, M. M.; Esteban, R.; Borisov, A. G.; Aizpurua, J.; Baumberg, J. J. Revealing the Quantum Regime in Tunnelling Plasmonics. Nature 2012, 491, 574−577. (26) Loth, S.; Baumann, S.; Lutz, C. P.; Eigler, D. M.; Heinrich, A. J. Bistability in Atomic-Scale Antiferromagnets. Science 2012, 335, 196− 199. (27) Liu, X.; Kang, J.-H.; Yuan, H.; Park, J.; Kim, S. J.; Cui, Y.; Hwang, H. Y.; Brongersma, M. L. Electrical Tuning of a Quantum Plasmonic Resonance. Nat. Nanotechnol. 2017, 12, 866−870. (28) Krishnan, K.; Muruganathan, M.; Tsuruoka, T.; Mizuta, H.; Aono, M. Highly Reproducible and Regulated Conductance Quantization in a Polymer-Based Atomic Switch. Adv. Funct. Mater. 2017, 27, 1605104. (29) Goykhman, I.; Desiatov, B.; Khurgin, J.; Shappir, J.; Levy, U. Locally Oxidized Silicon Surface-Plasmon Schottky Detector for Telecom Regime. Nano Lett. 2011, 11, 2219−2224. (30) Waser, R.; Aono, M. Nanoionics-Based Resistive Switching Memories. Nat. Mater. 2007, 6, 833−840. (31) Yang, Y.; Gao, P.; Gaba, S.; Chang, T.; Pan, X.; Lu, W. Observation of Conducting Filament Growth in Nanoscale Resistive Memories. Nat. Commun. 2012, 3, 732. (32) Yang, Y.; Gao, P.; Li, L.; Pan, X.; Tappertzhofen, S.; Choi, S.; Waser, R.; Valov, I.; Lu, W. D. Electrochemical Dynamics of Nanoscale Metallic Inclusions in Dielectrics. Nat. Commun. 2014, 5, 4232. (33) Chua, L. O. The Fourth Element. Proc. IEEE 2012, 100, 1920− 1927. (34) Waser, R. Nanoelectronics and Information Technology; John Wiley & Sons: New York, 2012. (35) Guo, X.; Schindler, C.; Menzel, S.; Waser, R. Understanding the Switching-Off Mechanism in Ag+ Migration Based Resistively Switching Model Systems. Appl. Phys. Lett. 2007, 91, 133513. (36) Larentis, S.; Nardi, F.; Balatti, S.; Gilmer, D. C.; Ielmini, D. Resistive Switching by Voltage-Driven Ion Migration in Bipolar Rram—Part Ii: Modeling. IEEE Trans. Electron Devices 2012, 59, 2468−2475. (37) Schindler, C.; Staikov, G.; Waser, R. Electrode Kinetics of Cu− Sio2-Based Resistive Switching Cells: Overcoming the Voltage-Time Dilemma of Electrochemical Metallization Memories. Appl. Phys. Lett. 2009, 94, 072109. (38) Kim, S.; Choi, S.; Lu, W. Comprehensive Physical Model of Dynamic Resistive Switching in an Oxide Memristor. ACS Nano 2014, 8, 2369−2376. (39) Valov, I.; Sapezanskaia, I.; Nayak, A.; Tsuruoka, T.; Bredow, T.; Hasegawa, T.; Staikov, G.; Aono, M.; Waser, R. Atomically Controlled Electrochemical Nucleation at Superionic Solid Electrolyte Surfaces. Nat. Mater. 2012, 11, 530−535. (40) Nayak, A.; Tamura, T.; Tsuruoka, T.; Terabe, K.; Hosaka, S.; Hasegawa, T.; Aono, M. Rate-Limiting Processes Determining the Switching Time in a Ag2s Atomic Switch. J. Phys. Chem. Lett. 2010, 1, 604−608. (41) Strukov, D. B.; Alibart, F.; Stanley Williams, R. Thermophoresis/Diffusion as a Plausible Mechanism for Unipolar Resistive Switching in Metal−Oxide−Metal Memristors. Appl. Phys. A: Mater. Sci. Process. 2012, 107, 509−518.

under Project No. s714. P.N. and A.A. acknowledge support from the Robert A. Welch Foundation under Grant C-1222 and the U.S. National Science Foundation under Grant ECCS1610229.

REFERENCES (1) Courtland, R. Moore’s Law’s Next Step: 10 Nanometers. IEEE Spectrum 2017, 54, 52−53. (2) Eigler, D. M.; Schweizer, E. K. Positioning Single Atoms with a Scanning Tunnelling Microscope. Nature 1990, 344, 524−526. (3) Agraït, N.; Yeyati, A. L.; van Ruitenbeek, J. M. Quantum Properties of Atomic-Sized Conductors. Phys. Rep. 2003, 377, 81− 279. (4) Reed, M. A.; Zhou, C.; Muller, C. J.; Burgin, T. P.; Tour, J. M. Conductance of a Molecular Junction. Science 1997, 278, 252−254. (5) Xie, F. Q.; Nittler, L.; Obermair, C.; Schimmel, T. GateControlled Atomic Quantum Switch. Phys. Rev. Lett. 2004, 93, 128303. (6) Terabe, K.; Hasegawa, T.; Nakayama, T.; Aono, M. Quantized Conductance Atomic Switch. Nature 2005, 433, 47−50. (7) Strukov, D. B.; Snider, G. S.; Stewart, D. R.; Williams, R. S. The Missing Memristor Found. Nature 2008, 453, 80−83. (8) Waser, R.; Dittmann, R.; Staikov, G.; Szot, K. Redox-Based Resistive Switching Memories − Nanoionic Mechanisms, Prospects, and Challenges. Adv. Mater. 2009, 21, 2632−2663. (9) Pires, D.; Hedrick, J. L.; De Silva, A.; Frommer, J.; Gotsmann, B.; Wolf, H.; Despont, M.; Duerig, U.; Knoll, A. W. Nanoscale ThreeDimensional Patterning of Molecular Resists by Scanning Probes. Science 2010, 328, 732−735. (10) Obermair, C.; Xie, F.-Q.; Schimmel, T. The Single-Atom Transistor: Perspectives for Quantum Electronics on the AtomicScale. Europhys. News 2010, 41, 25−28. (11) Fuechsle, M.; Miwa, J. A.; Mahapatra, S.; Ryu, H.; Lee, S.; Warschkow, O.; Hollenberg, L. C. L.; Klimeck, G.; Simmons, M. Y. A Single-Atom Transistor. Nat. Nanotechnol. 2012, 7, 242−246. (12) Sangwan, V. K.; Jariwala, D.; Kim, I. S.; Chen, K.-S.; Marks, T. J.; Lauhon, L. J.; Hersam, M. C. Gate-Tunable Memristive Phenomena Mediated by Grain Boundaries in Single-Layer Mos2. Nat. Nanotechnol. 2015, 10, 403. (13) Xu, Q.; Schmidt, B.; Pradhan, S.; Lipson, M. Micrometre-Scale Silicon Electro-Optic Modulator. Nature 2005, 435, 325−327. (14) Timurdogan, E.; Sorace-Agaskar, C. M.; Sun, J.; Shah Hosseini, E.; Biberman, A.; Watts, M. R. An Ultralow Power Athermal Silicon Modulator. Nat. Commun. 2014, 5, 4008. (15) Haffner, C.; Heni, W.; Fedoryshyn, Y.; Niegemann, J.; Melikyan, A.; Elder, D. L.; Baeuerle, B.; Salamin, Y.; Josten, A.; Koch, U.; Hoessbacher, C.; Ducry, F.; Juchli, L.; Emboras, A.; Hillerkuss, D.; Kohl, M.; Dalton, L. R.; Hafner, C.; Leuthold, J. AllPlasmonic Mach−Zehnder Modulator Enabling Optical High-Speed Communication at the Microscale. Nat. Photonics 2015, 9, 525−528. (16) Schoen, D. T.; Holsteen, A. L.; Brongersma, M. L. Probing the Electrical Switching of a Memristive Optical Antenna by Stem Eels. Nat. Commun. 2016, 7, 12162. (17) Di Martino, G.; Tappertzhofen, S.; Hofmann, S.; Baumberg, J. Nanoscale Plasmon-Enhanced Spectroscopy in Memristive Switches. Small 2016, 12, 1334−1341. (18) Emboras, A.; Cheng, B.; Ma, P.; Salamin, Y.; Haffner, C.; Luisier, M.; Hafner, C.; Leuthold, J. In Atomic Photodetection, Conference on Lasers and Electro-Optics, San Jose, California, June 5, 2016; Optical Society of America: San Jose, CA, 2016; p FF1B.1. (19) Emboras, A.; Niegemann, J.; Ma, P.; Haffner, C.; Pedersen, A.; Luisier, M.; Hafner, C.; Schimmel, T.; Leuthold, J. Atomic Scale Plasmonic Switch. Nano Lett. 2016, 16, 709−714. (20) Hino, T.; Hasegawa, T.; Tanaka, H.; Tsuruoka, T.; Terabe, K.; Ogawa, T.; Aono, M. Volatile and Nonvolatile Selective Switching of a Photo-Assisted Initialized Atomic Switch. Nanotechnology 2013, 24, 384006. 6712

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713

Article

ACS Nano (42) Onofrio, N.; Guzman, D.; Strachan, A. Atomic Origin of Ultrafast Resistance Switching in Nanoscale Electrometallization Cells. Nat. Mater. 2015, 14, 440. (43) Tappertzhofen, S.; Valov, I.; Waser, R. Quantum Conductance and Switching Kinetics of Agi-Based Microcrossbar Cells. Nanotechnology 2012, 23, 145703. (44) Ducry, F.; Emboras, A.; Andermatt, S.; Bani-Hashemian, M. H.; Cheng, B.; Leuthold, J.; Luisier, M. Ab Initio Modeling of Cbram Cells: From Ballistic Transport Properties to Electro-Thermal Effects. In 63rd IEEE International Electron Devices Meeting, IEDM 2017, San Francisco, CA, USA, December 2−6, 2017; IEEE, 2017. (45) Yuan, S.; Igarashi, K.; Tsuritani, T.; Morita, I. Bit-Error Ratio Performance Improvement Using Iterative Decoding for PolybinaryShaped Super-Nyquist Wavelength Division Multiplexed Signals. J. Lightwave Technol. 2017, 35, 4605−4612.

6713

DOI: 10.1021/acsnano.8b01811 ACS Nano 2018, 12, 6706−6713