pubs.acs.org/JPCL
Rate-Limiting Processes Determining the Switching Time in a Ag2S Atomic Switch Alpana Nayak,*,† Takuro Tamura,‡ Tohru Tsuruoka,† Kazuya Terabe,† Sumio Hosaka,‡ Tsuyoshi Hasegawa,† and Masakazu Aono† †
WPI Center for Materials Nanoarchitectonics, National Institute for Materials Science, 1-1Namiki, Tsukuba, Ibaraki 305-0044, Japan and ‡Department of Production Science and Technology, Gunma University, 1-5-1 Tenjin-cho, Kiryu 376-8515, Japan
ABSTRACT The switching time of a Ag2S atomic switch, in which formation and annihilation of a Ag atomic bridge is controlled by a solid-electrochemical reaction in a nanogap between two electrodes, is investigated as a function of bias voltage and temperature. Increasing the bias voltage decreases the switching time exponentially, with a greater exponent for the lower range of bias than that for the higher range. Furthermore, the switching time shortens exponentially with raising temperature, following the Arrhenius relation with activation energy values of 0.58 and 1.32 eV for lower and higher bias ranges, respectively. These results indicate that there are two main processes which govern the rate of switching, first, the electrochemical reduction Agþ þ e-fAg and, second, the diffusion of Agþ ions. This investigation advances the fundamental understanding of the switching mechanism of the atomic switch, which is essential for its successful device application. SECTION Electron Transport, Optical and Electronic Devices, Hard Matter
N
room temperature, enable the development of a novel programmable logic device5 that can achieve all functions with a single chip. Because the operating mechanism of the atomic switch is very different from that of conventional semiconductor devices, a fundamental understanding of the switching mechanism is necessary prior to its use in commercial devices. Though several intriguing results, both experimental3,6-8 and theoretical,9,10 have already been obtained concerning this switch, its actual working mechanism has not been well clarified yet. Accordingly, we measured the switching time of a Ag2S atomic switch with a nanogap as a function of bias voltage and temperature. We determined, for the first time, the activation energy values for switching, which are of prime importance in understanding the mechanism. Our results suggest that, in addition to the electrochemical reaction, the diffusion of Agþ ions plays a significant role in determining the rate of switching. In this study, a Ag2S atomic switch is realized across a gap of 1 nm between a Ag2S electrode and a Pt tip of a scanning tunneling microscope (STM) as a counter electrode. Figure 1 shows a scanning electron microscope (SEM) image of the Ag2S electrode, which was used as the solid-electrolyte electrode, with crystals ranging from submicrometer to micrometer size. A schematic representation of the Ag2S atomic switch and the corresponding electric potential distribution
anoionics-based resistive switching devices have been attracting much attention in recent years to overcome the physical and economical limitations of current semiconductor technology.1 A lot of research has been aimed at finding a reliable switching mechanism that can permit ever smaller and more powerful electronics.2 Recently, we have developed a conceptually new nanodevice called an atomic switch, in which formation and annihilation of a metal atomic bridge across a nanogap between a solid-electrolyte electrode and a counter metal electrode is controlled by a solid-electrochemical reaction.3 The switching operation can be achieved by only changing the polarity of the bias voltage applied to either electrode. For instance, applying a positive bias voltage to the solid-electrolyte electrode, the metal ions in the electrode reduce to metal atoms, forming a conductive atomic bridge between the electrodes. This decreases the resistance between the two electrodes to a certain ON resistance, which means that the switch is turned ON. When the polarity of the applied voltage is reversed, the metal atoms in the conductive atomic bridge are oxidized and are incorporated back into the solid-electrolyte electrode. This annihilates the conductive bridge between the two electrodes, turning the switch OFF. Similar controlled formation and annihilation of an atomic bridge has also been achieved in an ionic conductive material sandwiched between two electrodes using the solid-electrochemical reaction.4 The ease of operation and simple structure of the atomic switch make it suitable for configuring logic gates3 and memory devices.4 In addition, its unique features, namely, low ON resistance, scalability down to nanometer size, low power consumption, and operation at
r 2010 American Chemical Society
Received Date: December 8, 2009 Accepted Date: January 6, 2010 Published on Web Date: January 11, 2010
604
DOI: 10.1021/jz900375a |J. Phys. Chem. Lett. 2010, 1, 604–608
pubs.acs.org/JPCL
Figure 1. SEM image of the Ag2S electrode showing crystals of submicrometer to micrometer size. The Ag protrusions, which grew during scanning, are indicated by arrows. Figure 3. Switching time tsw as a function of bias voltage Vsw measured at room temperature (∼25 °C). The mean value of tsw for each Vsw is shown by open circle (O) symbol. The exponential fits of the data, considering the mean values of tsw, are shown by the solid lines. The extrapolations of the exponential fits are shown by the dashed lines.
dashed lines in Figure 3. Interestingly, the exponent β is about 2 times greater for the lower bias range than that for the higher range. It should be noted that the resistance of the Ag2S electrode (RAg2S) is on the order of 100 Ω;, therefore, most of the voltage drop occurs in the vacuum gap of much higher resistance (1 MΩ-12.9 kΩ) between the Ag2S electrode and the STM tip (Figure 2). Hence, the bias voltage VAg2S effectively applied to the Ag2S electrode is much smaller than Vsw and is given by It RAg2S, where It is the tunneling current flowing in the atomic switch during the switch-ON process. Consequently, it is useful to examine tsw as a function of VAg2S. We find that the tsw decreases exponentially as well with VAg2S, consistent with our earlier report on the growth rate of Ag protrusion.12 This suggests that the rate of the solid-electrochemical reaction increases exponentially with increasing bias and that the charge-transfer coefficient1 is proportional to the exponent β. Furthermore, a smaller value of β for higher Vsw indicates that the operating mechanism might be accompanied by some additional activation barrier at higher Vsw. More insight into the switching mechanism can be obtained from the temperature dependence of tsw. Upon increasing the temperature of the Ag2S electrode, the tsw decreases exponentially following the Arrhenius relation tsw µ exp-Ea/kBT, where Ea is the activation energy for switching. The temperature (T) dependence of tsw for Vsw of 0.15 and 0.25 V are shown in Figure 4a and b, respectively. The Ea values extracted from the slope of the Arrhenius plots (cf. insets of Figure 4a and b) are 0.51 and 1.26 eV, respectively. Similarly, we have extracted the Ea values from the Arrhenius plots of tsw versus T data for various Vsw. Figure 5 shows the Ea values for Vsw in the range of 0.1-0.275 V. Interestingly, the Ea values at the lower bias range are smaller than those at the higher bias range, complementing our results from the voltage dependence measurements. The mean values of Ea are found to be 0.58 and 1.32 eV for lower and higher bias ranges, respectively. We infer from these results that, at the higher Vsw, the supply of Agþ ions might become slower than the rate of electrochemical reduction, leading to a higher value of Ea.10
Figure 2. Schematic representation of the Ag2S atomic switch and the corresponding electric potential distribution induced by applying a bias voltage Vsw for switching. The bias voltage effectively applied to the Ag2S electrode is shown by VAg2S.
induced by applying a bias voltage (Vsw) for switching are shown in Figure 2. Most of the potential drop occurs in the vacuum gap, and the potential drop in the Ag2S electrode (VAg2S) occurs mainly at the surface and the interface. We define switching time (tsw) as the time taken for the resistance between the Ag2S electrode and the STM tip to decrease from an initial OFF resistance to 12.9 kΩ after applying a Vsw. Here, 12.9 kΩ is considered as the ON resistance corresponding to a single atomic contact.11 Figure 3 shows the tsw as a function of Vsw measured at room temperature (∼25 °C) for an initial OFF resistance of 1 MΩ. Switching becomes exponentially faster with increasing Vsw.3,8 The variation in the values of tsw at a particular Vsw is acceptable, taking into account that the measurements were done at different positions on the Ag2S surface. As can be seen in Figure 1, Ag precipitation depends on the local conditions of surface structure and ionic distribution. However, considering the mean value of tsw for each Vsw, we find that the data fits well with a simple exponential function tsw µ exp-βVsw. Further, this exponential relation appears to consist of two components with different values of exponent β, indicating that more than one process participates in controlling the rate of switching. The exponential fits to the data are shown by the solid lines, and the extrapolations of the fits are shown by the
r 2010 American Chemical Society
605
DOI: 10.1021/jz900375a |J. Phys. Chem. Lett. 2010, 1, 604–608
pubs.acs.org/JPCL
Figure 6. Schematic illustration of the operating mechanism in a Ag2S atomic switch. The solid line represents the rate of precipitation of Ag atoms, and the dashed line represents the Agþ ionic density at the surface of the Ag2S electrode due to diffusion. At lower bias voltages (Vsw), the electrochemical reduction for the precipitation of Ag atoms is the rate-limiting process, whereas the diffusion of Agþ ions toward the surface of Ag2S becomes the ratelimiting process at higher Vsw. The crossover between these two rate-limiting processes occurs at around a Vsw referred as the critical point. Accordingly, the activation energy (Ea) for switching is greater at higher Vsw than that at lower Vsw.
toward the Pt tip leading to the formation of an atomic bridge. We believe, on the basis of our results, that more than one process participates in controlling the overall kinetics of the atomic switch and that the final rate-limiting step depends on the applied bias voltage for switching. Since the electric field applied during switching is very high (∼106 V/cm), the precipitated Ag atoms are most likely to grow toward the Pt tip, leading to the formation of an atomic bridge, and hence, step (iii) does not seem to be a rate-limiting process. This renders steps (i) and (ii) as the probable but competitive rate-limiting steps determining the kinetics of the switch-ON process. At lower Vsw, the time taken for switching is on the order of milliseconds or greater, providing enough time for the diffusion of sufficient Agþ ions to the Ag2S surface for the electrochemical reduction.13,14 In other words, the rate of diffusion of Agþ ions is faster than the rate of electrochemical reduction. Therefore, in this range of Vsw, the rate of electrochemical reduction for the precipitation of Ag atoms is the most probable rate-limiting step. On the other hand, at higher Vsw, the rate of precipitation of Ag atoms becomes faster due to the increased It. However, a greater value of Ea and a smaller value of β indicates that the availability of Agþ ions at the Ag2S surface might not be sufficient to accompany the fast precipitation of Ag atoms. This is because there is not enough time for the diffusion of sufficient Agþ ions toward the surface of the Ag2S electrode. Therefore, at higher Vsw, the rate of diffusion of Agþ ions in Ag2S toward its surface is the most probable rate-limiting step. On the basis of our results and the above discussion, we have schematically illustrated, in Figure 6, the mechanism for the switch-ON process of the Ag2S atomic switch for lower and higher Vsw. The solid line represents the rate of precipitation of Ag atoms with increasing Vsw, and the dashed line represents
Figure 4. Temperature (T) dependence of the switching time (tsw) for bias voltages (Vsw) of (a) 0.15 and (b) 0.25 V. The error bars correspond to the standard deviation of the measured tsw. The solid lines correspond to the exponential fits to the data. The Arrhenius plots and the extracted activation energy (Ea) values are shown in the insets of the respective figures.
Figure 5. The activation energy (Ea) values extracted from the Arrhenius plots of switching time versus temperature data for bias voltages (Vsw) in the range of 0.1-0.275 V.
The switch-ON process of the atomic switch at a sufficient Vsw includes the following steps: (i) a charge-transfer process that involves reduction of Agþ ions at the Ag2S surface leading to the precipitation of Ag atoms and the counter reaction representing the oxidation and dissolution of Agþ ions into the Ag2S electrolyte (Figure 2), (ii) diffusion of Agþ ions across the Ag2S electrolyte under the action of an applied electric field, and (iii) preferential growth of the precipitated Ag atoms
r 2010 American Chemical Society
606
DOI: 10.1021/jz900375a |J. Phys. Chem. Lett. 2010, 1, 604–608
pubs.acs.org/JPCL
the Agþ ionic density at the Ag2S surface due to diffusion. At lower Vsw, the electrochemical reduction for the precipitation of Ag atoms is the rate-limiting process, whereas the diffusion of Agþ ions toward the surface of Ag2S becomes the ratelimiting process at higher Vsw. The crossover between these two rate-limiting processes occurs at around a Vsw of 0.2 V, referred as the critical point in Figure 6. Further investigation to clarify the existence of such a critical point is in progress. Moreover, there are possibilities of other rate-determining processes; for instance, the electron migration process to trap defects on the Ag2S surface for the formation of a Ag cluster could also play a significant role. Investigating such processes would require a precise knowledge of the distribution of defects on the surface of the Ag2S electrode together with an atomistic study based on first-principles calculations. Thus, in order to exploit the potential of atomic switch to the limits, a considerable research effort is still needed with respect to a deeper understanding of the mechanism that governs the rate of switching. The Ag2S electrode used in this study was prepared by sulfurizing a Ag plate (8 5 0.2 mm3) at 150 °C for 2 h with sulfur vapor. Under this sulfurization condition, the Ag2S electrode is expected to exhibit the acanthite R-phase characterized by a monoclinic unit cell with semiconducting properties and is stable at room temperature.15-17 The temperature dependence measurements were performed by an indirect resistive heating of the Ag2S electrode. During all of the measurements, the STM was operated under high vacuum conditions (∼1 10-4 Pa). We measured the switching time by the following procedure: (i) The Pt tip was fixed with a tunneling resistance of 1 MΩ (V = -20 mV, I = 20 nA) as the initial OFF resistance. Consequently, a gap of about 1 nm was maintained between the Pt tip and the Ag2S electrode. (ii) The feedback system of the STM was disabled, and a positive Vsw was applied to the Ag2S electrode. (iii) The current flowing between the tip and the Ag2S surface was recorded. Since the resistance of the Ag2S electrode was less than 100 Ω, the ON resistance was calculated from the current flowing between the tip and the Ag2S electrode. The Ag protrusion formed on the surface of Ag2S during each switch-ON process was completely eliminated by applying a negative Vsw for a few minutes to the Ag2S electrode before the next measurement. In conclusion, the effects of bias voltage and temperature on the switching time of a Ag2S atomic switch were investigated. Switching becomes exponentially faster with increasing bias voltage. However, the exponential relation exhibits a greater exponent for the lower bias range than that for the higher range, suggesting an additional activation barrier for switching at higher bias voltages. Raising the temperature shortens the switching time exponentially, following the Arrhenius relation. The activation energy values are determined to be 0.58 and 1.32 eV for the lower and higher bias ranges, respectively. This is the first experimental determination of the activation energy for switching in an atomic switch and is of prime importance in understanding the switching behavior. Our results suggest that the electrochemical reduction for the precipitation of Ag atoms is the rate-limiting process at lower
r 2010 American Chemical Society
bias voltages, whereas the diffusion of Agþ ions toward the surface of Ag2S becomes the rate-limiting process at higher bias voltages. These findings provide physical insight into the operating mechanism of the atomic switch.
AUTHOR INFORMATION Corresponding Author: *To whom correspondence should be addressed. E-mail: NAYAK.
[email protected].
ACKNOWLEDGMENT Part of this work was conducted under the Key-Technology Research Project, “Atomic Switch Programmed Device”, supported by the MEXT and the Strategic JapaneseGerman Cooperative Program supported by JST.
REFERENCES (1)
Waser, R.; Dittmann, R.; Staikov, G.; Szot, K. Redox-Based Resistive Switching Memories ; Nanoionic Mechanisms, Prospects, and Challenges. Adv. Mater. 2009, 21, 2632– 2663. (2) Waser, R.; Aono, M. Nanoionics-Based Resistive Switching Memories. Nat. Mater. 2007, 6, 833–840. (3) Terabe, K.; Hasegawa, T.; Nakayama, T.; Aono, M. Quantized Conductance Atomic Switch. Nature 2005, 433, 47–50. (4) Sakamoto, T.; Sunamura, H.; Kawaura, H.; Hasegawa, T.; Nakayama, T.; Aono, M. Nanometer-Scale Switches Using Copper Sulfide. Appl. Phys. Lett. 2003, 82, 3032–3034. (5) Kaeriyama, S.; Sakamoto, T.; Sunamura, H.; Mizuno, M.; Kawaura, H.; Hasegawa, T.; Terabe, K.; Nakayama, T.; Aono, M. A Nonvolatile Programmable Solid-Electrolyte Nanometer Switch. IEEE J. Solid-State Circuits 2005, 40, 168–176. (6) Terabe, K.; Nakayama, T.; Hasegawa, T.; Aono, M. Ionic/ Electronic Mixed Conductor Tip of a Scanning Tunneling Microscope as a Metal Atom Source for Nanostructuring. Appl. Phys. Lett. 2002, 80, 4009–4011. (7) Liang, C.; Terabe, K.; Hasegawa, T.; Negishi, R.; Tamura, T.; Aono, M. Ionic-Electronic Conductor Nanostructures: Template-Confined Growth and Nonlinear Electrical Transport. Small 2005, 1, 971–975. (8) Tamura, T.; Hasegawa, T.; Terabe, K.; Nakayama, T.; Sakamoto, T.; Sunamura, H.; Kawaura, H.; Hosaka, S.; Aono, M. Switching Property of Atomic Switch Controlled by Solid Electrochemical Reaction. Jpn. J. Appl. Phys. 2006, 45, L364–L366. (9) Wang, Z.; Kadohira, T.; Tada, T.; Watanabe, S. Nonequilibrium Quantum Transport Properties of a Silver Atomic Switch. Nano Lett. 2007, 7, 2688–2692. (10) Wang, Z.; Gu, T.; Kadohira, T.; Tada, T.; Watanabe, S. Migration of Ag in Low-Temperature Ag2S from First Principles. J. Chem. Phys. 2008, 128, 014704. (11) Ohnishi, H.; Kondo, Y.; Takayanagi, K. Quantized Conductance Through Individual Rows of Suspended Gold Atoms. Nature 1998, 395, 780–783. (12) Terabe, K.; Nakayama, T.; Hasegawa, T.; Aono, M. Formation and Disappearance of a Nanoscale Silver Cluster Realized by Solid Electrochemical Reaction. J. Appl. Phys. 2002, 91, 10110–10114. (13) Allen, R. L.; Moore, W. J. Diffusion of Silver in Silver Sulfide. J. Phys. Chem. 1959, 63, 223–226.
607
DOI: 10.1021/jz900375a |J. Phys. Chem. Lett. 2010, 1, 604–608
pubs.acs.org/JPCL
(14)
(15)
(16) (17)
Bartkowicz, I.; Mrowec, S. Ionic Conductance of Silver Sulphide and Diffusion Mechanism of Silver Ions in R-Ag2S. Phys. Status Solidi B 1972, 49, 101–105. Kundu, M.; Terabe, K.; Hasegawa, T.; Aono, M. Effect of Sulfurization Conditions and Post-Deposition Annealing Treatment on the Structural and Electrical Properties of Silver Sulfide Films. J. Appl. Phys. 2006, 99, 103501. Lehovec, K. On the Temperature Dependence of the Optical Absorption of β-Ag2S. J. Chem. Phys. 1953, 21, 54–57. Wagner, C. Investigations on Silver Sulfide. J. Chem. Phys. 1953, 21, 1819–1827.
r 2010 American Chemical Society
608
DOI: 10.1021/jz900375a |J. Phys. Chem. Lett. 2010, 1, 604–608
