Download Hi-Res ImageDownload to MS-PowerPointCite This:ACS Appl. Electron. Mater. 2019, 1, 1, 96-104
This publication is Open Access under the license indicated. Learn More
Article
An Optically Gated Transistor Composed of Amorphous M + Ge2Se3 (M = Cu or Sn) for Accessing and Continuously Programming a MemristorClick to copy article linkArticle link copied!
- Kristy A. Campbell*Kristy A. Campbell*E-mail [email protected]Department of Electrical and Computer Engineering, Boise State University, 1910 University Dr., Boise, Idaho 83725, United StatesMore by Kristy A. Campbell
- Randall A. BassineRandall A. BassineDepartment of Electrical and Computer Engineering, Boise State University, 1910 University Dr., Boise, Idaho 83725, United StatesMore by Randall A. Bassine
- Md. Faisal KabirMd. Faisal KabirDepartment of Electrical and Computer Engineering, Boise State University, 1910 University Dr., Boise, Idaho 83725, United StatesMore by Md. Faisal Kabir
- Jeremy AstleJeremy AstleDepartment of Electrical and Computer Engineering, Boise State University, 1910 University Dr., Boise, Idaho 83725, United StatesMore by Jeremy Astle
Abstract
Click to copy section linkSection link copied!
We demonstrate that a device composed of sputtered amorphous chalcogenide Ge2Se3/M + Ge2Se3 (M = Sn or Cu) alternating layers functions as an optically gated transistor (OGT) and can be used as an access transistor for a memristor memory element. This transistor has only two electrically connected terminals (source and drain), with the gate being optically controlled, thus allowing the transistor to operate only in the presence of light (385–1200 nm). The switching speed of the OGTs is <15 μs. The OGT is demonstrated in series with a Ge2Se3 + W memristor, where we show that by altering the light intensity on the OGT gate, the memristor can be programmed to a continuous range of nonvolatile memory states using the saturation current of the OGT as a programming compliance current. By having a continuous range of nonvolatile states, one memory cell can potentially achieve 2n levels. This high density, combined with optical programmability, enables hybrid electronic/photonic memory.
Copyright © 2018 American Chemical Society
Introduction
Click to copy section linkSection link copied!
The use of resistive random access memory (ReRAM), also known as memristor devices, (1,2) in circuits such as cross-point arrays, (3) spiking neural networks, (4) or other computing applications (5−7) where a large number of these memristive memory elements need to be incorporated into the circuit has been challenging due to their resistive nature. Without proper isolation via a selector device, (8−11) each device behaves within a circuit like it is part of a resistive network. The resistive nature of these memory elements can create multiple unwanted paths of current flow in a cross-point array even when only a single element is being addressed. Because they traverse a path through nonaddressed memristor elements, these currents are termed sneak path currents. They eventually sum with the desired current through the addressed memristor, giving a false reading of the addressed memristor’s state. Circuit techniques to eliminate or reduce sneak path current have included addition of a diode in the memristor device array element. (10,12) A diode in series with each memristor prevents access to a nonaddressed element because the voltage at that node remains lower than the forward voltage of the diode, preventing current flow through the memristor element. This is a reasonable approach for the case of unipolar memristor elements that can change state with the application of a single polarity electronic input signal, such as phase-change devices. However, bipolar memristors, such as the redox-based memristors, (13) need both a positive and negative potential to increase and decrease the element’s resistance state. A circuit design option to allow for both polarities is to use two diodes with opposite direction in series with the memristor element in an array. Because this increases the array feature size, it is not the best solution in many circuit design situations. Another way that memory elements have been addressed is through the use of a traditional access transistor, the most common being a MOSFET access transistor. This type of access transistor is typically used in an array where the voltages applied to the bit and word lines are at fixed values. In this case, the MOSFET access transistor allows the memory element to be electrically accessed by application of a pulse to the MOSFET gate that turns it ON, allowing current to flow between the source and drain terminals, with some fixed transistor drain-to-source resistance, Rds. The memory element is placed in series with the drain and source, and when the MOSFET is ON, it has a resistance of Rds, resulting in a voltage divider between the memory element resistance and the transistor drain to source resistance. This is not a desirable situation for a memristor element due to the varying nature of the memristor’s resistance within a fixed circuit path, especially if multistate programming is desired. The variable resistance creates a variable voltage across the memristor due to the voltage divider and can lead to memristor device failure or a “stuck” state. Researchers have therefore been investigating alternate materials and devices to act as “selector” devices that can prevent sneak-path current in memristor arrays. (10,14) These include the ovonic threshold switch (OTS) (15,16) and other materials such as Cu-doped HfO2 (17) and Ag-based materials. (18−20) These selector materials are incorporated in series with the memristor and must be electrically driven to a state that allows the memristor to be either read (without perturbation) or programmed (higher or lower resistance). While the materials may vary, there are common themes among them. First, these selector materials must all be electrically switched. For optimum circuit design, the selector device electrical properties need to complement the memristor it will be used with. Considerations include consistent threshold voltage, switching speed, operating temperature tolerance, uni- or bidirectional operation, and number of cycles possible. Most of the selector devices have a distribution of threshold voltages. (19) In some materials, especially those that rely on Ag or Cu metal filament formation to produce the threshold switching, (17−20) the switching is not bidirectional, (19) meaning that the switching properties in the reverse voltage direction are different than in the forward voltage direction. (18) In some cases, this limits their use to unidirectional memristor elements such as phase change memory. Second, many of the selector devices, especially the Ag- and Cu-based selector devices, are sensitive to temperature. The metal can migrate into the material and saturate it with increased temperature or cycling, leading to large variations in threshold voltages (especially low voltages at higher temperatures) and eventual device failure due to metal saturation, leading to the inability to switch. Third, selectors may have low ON and OFF ratios (“selectivity”). Last, the switching speeds may be too long. Attempts to incorporate two different selector materials in series to produce an optimum hybrid selector to achieve a faster switching and higher selectivity have been performed using Ag-SiTe-based selectors and NbOx selectors. (18) While the initial results show promise, the operational timing between selector elements is complicated and may be difficult to achieve consistently.
Given the need for new selector device options, we have investigated the amorphous chalcogenide Ge2Se3, composed of alternating layers of doped/undoped material, to achieve an optically controlled selector device with current–voltage (I–V) properties similar to a light-gated transistor. (21) The photoinduced properties of GeSe alloys have long been an area of research interest and have more recently been studied as electrically driven selector materials. (15,16) GeSe materials are known for their photoinduced responses, such as photoconductivity and photoinduced defects, (22) as well as structural and morphological changes due to light exposure. (23) More recently, crystalline GeSe2 has been investigated for optoelectronic switching applications (24) and as wide band gap materials for waveguides. (25) In the work reported here, we experimentally demonstrate two different optically gated access transistors (OGTs) composed of four alternating layers of amorphous Ge2Se3/M + Ge2Se3 films (M = Sn or Cu) deposited on a thin native oxide layer on p-Si. The chalcogenide material acts as the gate, using light instead of voltage, making this a two (electrical) terminal and one optical terminal transistor. Metallic (W) source and drain electrodes are placed directly on the chalcogenide material stack.
Both OGT device types, Cu + Ge2Se3 and Sn + Ge2Se3, were tested in a circuit in series with a memristor to investigate memristor state programming through the optically gated access transistor/selector device. LED light was used to gate the OGT ON to access and allow current to flow through the memristor. I–V curves were measured for the OGTs and memristors separately and across the series combination of the two devices.
Methods
Click to copy section linkSection link copied!
Fabrication of the Optically Gated Transistor (OGT) and Memristor
The OGT and memristor devices were fabricated in the Idaho Microfabrication Laboratory at Boise State University. The OGT devices were fabricated on p-Si wafer pieces taken from 200 mm wafers (1–100 ohm·cm, Microsil, LLC Silicon Services). An AJA International ATC Orion 5 UHV magnetron sputtering system and a Ge2Se3 target from Processed Materials were used to sputter alternating layers of 100 Å Ge2Se3 (rf sputtered) and a 30 Å layer of cosputtered M + Ge2Se3, (M = Cu or Sn), for a total thickness of ∼360 Å (Figure 1a), directly onto the native oxide layer of the p-Si substrate. W electrodes (350 Å) were deposited on the surface using a shadow mask (Figure 1b). The memristors used in this work were fabricated at Boise State University but packaged in an 8-pin DIP package by Knowm, Inc., part BS-AF-W 8, and donated for this work.
Figure 1
Figure 1. Optically gated device structure and electrical measurement description. (a) Optically gated transistor structure, not drawn to scale. The active material consists of alternating layers of Ge2Se3 and cosputtered M + Ge2Se3 where M = Cu or Sn. Each Ge2Se3 layer target thickness is 100 Å, and the M + Ge2Se3 layer target thickness is 30 Å; however, the layers mix during deposition, and the material layers remain amorphous. (b) Top-down optical view of the electrodes and separation of the OGT.
Raman spectra were collected of the Ge2Se3 and M + Ge2Se3 OGT films using a Renishaw inVia Raman confocal microscope with a 1 μm beam spot size.
The UV–vis index of refraction, n, and extinction coefficient, k, data were collected using an n&k Technology 1280 broadband UV–vis spectrometer with thin films deposited on quartz substrates (Alfa Aesar, #042295 quartz microscope slide, fused, 25.4 × 25.4 × 1.0 mm3, GE type 124 grade quartz).
Electrical Testing
The OGT was electrically tested at the wafer level on a microprobe station equipped with Cascade Microtech micromanipulators and size 7B W probe tips. An HP4156A semiconductor parameter analyzer was used for I–V measurements. A 470 nm LED (C503B-BAN-CY0C0461, Cree, Inc.) was used as a light source for the I–V sweep measurements, and a 770 nm LED (MTE1077N1-R, Marktech Optoelectronics) was used for the switching time measurements. It should be noted that a broadband microscope light source (from the microprobe station) as well as LEDs with wavelengths of 385, 525, 590, 616, 770, 850, 1060, and 1200 nm also worked to effectively gate the OGT ON. The LED was connected to the microscope objective of the microprobe station, so that it was placed directly above the OGT, roughly centered between the electrodes during measurements (Figure 2). The LED driver circuit consisted of a series resistor and a variable dc voltage, supplied from a Digilent Analog Discovery 2 (AD2). The AD2 also supplied the pulses for switching response tests. The two-channel oscilloscope on the AD2 was used to collect the switching speed data from the OGT. The LED intensity was measured using a Thor Laboratories PM16-121 standard photodiode sensor with a detector area of 9.7 mm × 9.7 mm. The light intensity was measured before and after the OGT circuit measurements; however, while effort was made to place the power meter detector at the same location as the sample, it is not exact. The power readings reported are with respect to the detector area (mW/cm2) instead of with respect to the area of the OGT since the exact area of illumination or area that contributes to the current is unknown at this research stage.
Figure 2
Figure 2. Electrical testing measurement configuration.
To test the memristor in series with the OGT, the memristor was placed on a breadboard and connected to the OGT via triaxial cables connected to micromanipulator probe tips placed on the OGT W electrodes. The I–V sweep measurements were made with the HP4156A connected to one electrode of the OGT and to the bottom electrode of the memristor so that the voltage was swept across the entire circuit. All I–V sweep measurements were “double sweep” measurements, meaning that the voltage was incremented forward to the end voltage and then reversed back to the starting voltage. All voltage sweep measurements, in both the positive and negative voltage directions, started and ended at 0 V. To characterize the OGT device alone, the voltage sweep ranged from 0 to ±10 V. The memristor was measured from 0 to ±1 V using a 100 μA compliance current. The memristor–OGT combination circuit was measured using a voltage sweep of 0 to ±2 V. Pulse testing was performed with the AD2 by applying a square pulse train with a 50% duty cycle at a frequency of 100 Hz to the LED drive circuit in place of the dc supply. To measure the switching time, the two-channel AD2 oscilloscope was connected to the pulse train input on the LED drive circuit and to the top electrode of the memristor.
Results and Discussion
Click to copy section linkSection link copied!
OGT Materials Raman Spectra and Optical Band Gap Estimate
Raman spectra were used to verify the stoichiometry of the Ge2Se3 film, the incorporation of Cu and Sn into the Ge2Se3 material, and that the films were amorphous (Figure 3). Amorphous Ge2Se3 has a Raman spectrum that is distinct from the other GexSe1–x stoichiometries. (26) The incorporation of M into the Ge2Se3 film is apparent by the changes in the Raman features. The most sensitive indicator of doping an amorphous Ge2Se3 thin film is a change in the Ge–Ge homopolar bond peak around 175 cm–1 and the broad peak around 300 cm–1. (26)
Figure 3
Figure 3. Raman spectra of the OGT films: (a) Cu + Ge2Se3, (b) Sn + Ge2Se3, and (c) Ge2Se3.
The absorption coefficient, α, was calculated from the measured extinction coefficient data using the relationship
where k is the measured extinction coefficient for the film. The optical band gap was estimated using a Tauc plot (Figure 4):
This corresponds to indirect allowed transitions (27) with an optical gap estimated to be 1.87 eV for Ge2Se3 and 1.61 and 1.85 eV for the Sn and Cu doped, respectively. The undoped Ge2Se3 film optical gap of ∼1.87 eV falls within the range of the GexSe1–x evaporated films previously reported. (28) The decreased optical gap energies for Sn and Cu doped M + Ge2Se3 OGT films are expected when traps are introduced to the film by the addition of dopants. (29)
(1)
(2)
Figure 4
Figure 4. Tauc plot of each OGT material. The optical gap is estimated to be 1.87 eV for Ge2Se3, 1.61 eV for Ge2Se3 + Sn, and 1.85 eV for Ge2Se3 + Cu.
OGT I–V Curves and Photoresponse Speed
I–V curves for the OGT device types are shown in Figure 5a–c, under different light intensity conditions, using a 470 nm LED light aimed near the center region between the electrodes. It should be noted that a broadband microscope light source (from the microprobe station) as well as LEDs with wavelengths of 385, 525, 590, 616, 770, 850, 1060, and 1200 nm could effectively gate the OGT ON. When the LED was centered between the two electrodes, the positive and negative current amplitudes were similar (as shown in Figure 5). As the LED was moved toward one electrode, the current at that electrode increased, with the opposite polarity voltage sweep showing a decrease in amplitude, similar to that reported for photocurrent measurements of GeSe2 crystals. (24) As previously reported for GeSe2 crystals, this is an indication that the photoconductivity occurs primarily at the electrode. While this is an issue for a benchtop circuit measurement such as this work, this is not a limitation for integrated circuits and large-scale arrays since the integrated circuit design would be able to fix the position of the light with respect to the OGT electrodes. There is a zero current “dead zone” region around V = 0 in each OGT type which provides a block against unwanted sneak path current when operated as an access transistor for a memristor or resistive memory element. The Sn and Cu + OGT devices have a “dead zone” between approximately ±0.21 and ±0.12 V, respectively. A “dead zone” of approximately the same voltage width is predicted in the theoretical density functional theory calculations of the I–V curves for the 2D GeSe/SnS heterobilayer photovoltaic structure. (30) At voltages outside of this zone, the I–V curve has a diode-like shape until a voltage is reached where photoinduced current no longer changes with applied voltage (saturation).
Figure 5
Figure 5. OGT I–V curves under 470 nm LED variable light intensity illumination. (a) Cu + Ge2Se3 OGT device. (b) Sn + Ge2Se3 OGT device. (c) Undoped Ge2Se3OGT device. An example of I–V curves for two different memristors is given in (d). A compliance current of 100 μA (the flat line in the I–V curve at 100 μA) was applied only during memristor measurements to protect the device from damage during the dc voltage sweep. Saturation current values in (a–c) correspond to light illumination ranging from a detector reading of 15.23 mW/cm2 at the highest saturation current reading to 28 μW/cm2 for the lowest current reading. The “dark” trace appears near 0 A (pink trace in (a) to (c)).
The memristors used in the OGT–memristor circuit measurements are composed of a set of alternating Ge2Se3/W + Ge2Se3 material layers and selected due to the their compatibility for future OGT–memristor device microfabrication integration. An example of I–V curves for two different memristor devices used in the OGT–memristor circuit tests is shown in Figure 5d. A compliance current of 100 μA was used during the I–V measurements of the memristor device (not during OGT–memristor circuit measurement) to limit the current through the memristor during I–V sweep testing to prevent damage. Given the amorphous nature of the active layer material, the memristors will have a slight variation in threshold voltage and I–V curve shape as a function of device operating conditions and history. This is clear when comparing the two different memristor device I–V curves in Figure 5d. The memristor write threshold voltage is typically in the range 0.21 to 0.35 V. The erase threshold voltage is typically in the range −50 to −200 mV. The dead zone of each OGT device allows the memristors to switch when the programming voltage is applied and the OGT is illuminated, without the OGT hindering access during a programming operation, yet blocking current flow when a memristor is not being accessed (since those devices will be under dark conditions). The variation in memristor threshold voltage is not an important factor when using the OGT to program a memristor since the OGT is operated at a potential placing it in its saturation regime, which is above any memristor threshold voltage range variation. Therefore, the OGT enables a consistent single circuit operating voltage for programming every memristor device.
The background (“dark”) measurements for the OGT measurements in Figure 5a–c were collected with the LED and room lights off, but not enclosed in a dark box. For purposes of determining photocurrent, the background incident power is considered to correspond to the “dark” current, and the resulting photocurrent (Iph) is the difference between the measured current and I“dark”. For each OGT, the resultant photocurrent is linear with respect to light intensity when measured in the saturation region at 2 V (Figure 6a). This linear response is desirable for a phototransistor device. Also shown in Figure 6a is the photocurrent measured in the OGT–memristor circuit, which aligns with the photocurrent of the OGT devices alone, as expected.
Figure 6
Figure 6. Photocurrent generated by 470 nm illumination during. (a) Photocurrent, Iph = Imeas – I“dark” measured at 2 V, as a function of light intensity. Black traces denote the Cu + OGT alone (solid line, open circle symbols) and in series with a memristor (dashed, + symbols); red traces denote the Sn + OGT alone (solid, open square symbols) and in series with a memristor (dashed, × symbols). (b) Iph as a function of V at a light power density of 14.17 mW/cm2 for the OGT devices alone. The measurement is a double sweep (forward and reverse voltage sweep) where the green arrows denote the reverse voltage return path. Regions denoted by 1 correspond to thermally generated carrier dominated current. The region marked by 2 denotes the onset of trap filled conduction, and that marked by 3 denotes the trap-filled limit voltage, where all traps are filled and conduction is space charge limited. Red trace (solid line) corresponds to Cu + OGT; black trace (dashed line) corresponds to Sn + OGT.
Current saturation was not reported for GeSe2 nanocrystals (24) but has been reported in organic crystals where it results from injection limited current at an electrode. (31) In Figure 6b, a plot of the log Iph versus log V for the Cu and Sn + OGTs shows three distinct conduction mechanism regions. (32) The first region, labeled “1” in Figure 6b, is an ohmic conduction region. The voltage at the onset of the second region is denoted by arrows labeled “2” and indicates the voltage at which the transition from ohmic to a trap-filled limited current begins. The voltage at which all traps are completely filled and the conduction becomes space-charge limited (obeying Child’s law) is denoted labeled “3” in Figure 6b. In the OGT devices, the saturation current is due to charge injection filling all available trap states at a given light intensity (Figure 6b). In this scenario, an increase in current with light intensity is likely due to an increase in the concentration of trap states due to photoinduced defect formation. In the memristor–OGT circuit, the saturation current is advantageous since it provides built-in current limiter, variable by light intensity, which is similar to variable state programming of a memristor through compliance current limiting (33−37) and is key to multilevel memristor programming.
The response time of the OGT devices in series with a 1 kohm load resistor using a 770 nm LED light pulse is shown in Figure 7a for a full pulse response and in Figure 7b for an expanded view of the ON to OFF transition. The voltage signal is measured across the 1 kohm load and converted to a current value to show the current through the OGT during the test. When there is no light on the OGT, it is OFF, and the resistor has only background “dark” current through it. The 770 nm LED switching time is in the nanoseconds range, faster than the response speed of the OGTs in the measurement configuration, which is approximately 11 and 15 μs for the Cu + OGT and Sn + OGT devices, respectively. The OFF to ON transition is at least an order of magnitude faster (Figure 7a) than the ON to OFF transition. In contrast, previous studies on GeSe single crystal nanosheet phototransitors have measured switching speeds on the order of 100 ms (OFF to ON transition) and >10 s (ON to OFF) under 808 nm illumination. (38) The OGT speeds are at least 3 orders of magnitude faster than the other light-gated memristor circuits which show ON to OFF transitions of ∼40 s for a CeO2–x/AlOy/Al multilayer structure (39) and milliseconds times for a light-gated memristor composed of CH3NH3PbI3. (40)
Figure 7
Figure 7. Photoswitching response for OGT devices to a 770 nm LED pulse at 7.4 mW/cm2. (a) Response to pulse train. (b) Expanded view of the ON to OFF transition. Sn + OGT, black open circles; Cu + OGT, red open squares. The OGTs are in series with a 1 kohm resistor and biased at 6 V on the drain electrode, with the source electrode connected to the resistor, and the other resistor terminal connected to ground.
Accessing and Programming the Memristor with the OGT Selector
The quasi-static I–V curves for the coupled OGT–memristor circuits (no series resistance added to the measurement circuit) are given in Figure 8. Note that in these measurements the light incident on the OGT is kept the same for both samples, including the same background light. The memristor used in this work operates under a similar mechanism as the self-directed channel memristor, (34) except that it contains a sandwiched layer of cosputtered W–Ge2Se3 between two active Ge2Se3 layers which allows it to operate more readily as a continuum-state memristor, with a mixed ion-conducting/phase-change mechanism. The memristor-only I–V curve (dashed), for each memristor used in the two different circuits, is superimposed on the circuit I–V curves for comparison. The “loop” in the I–V curves of the positive voltage sweeps of the OGT–memristor circuits (seen in the region before saturation) show the gating effect on memristor programming. As the voltage is incremented to higher values in the positive direction, the memristor will write to a lower resistance when its threshold voltage is exceeded and the current through the OGT is high enough to latch the device. The saturation current of the transistor, which is set by the light intensity, determines the memristor ON state resistance. The backward voltage sweep, which completes the “loop” in the positive voltage quadrant, shows that the memristor has been written to a lower resistance state. The negative voltage erase sweeps show the memristor in a low resistance state until the memristor erase threshold voltage and current are reached. It is not uncommon for this memristor type to exhibit two erased states, which can be seen in the I–V curves for the Cu + OGT/memristor circuit (Figure 8a). This can occur in the W–Ge2Se3 mixed ion-conductor/phase-change memristor device when erase potentials higher than −1 V are applied, such as in this measurement.
Figure 8
Figure 8. I–V curves for OGT–memristor circuits. (a) Cu + OGT. (b) Sn + OGT. An I–V curve for the memristor used in each OGT circuit is placed on each graph for comparison (∗). (c) Programmed resistance was determined by measuring the resistance of a line fit to the erase curve. (d) OGT–memristor circuit resistance as a function of LED light intensity. Sn + OGT, black open circles; Cu + OGT, red open squares. Illumination source was a 470 nm LED.
Memristors can be programmed to intermediate states through the process of current limiting. (33−37) Because the incident light intensity on the OGT determines its saturation current level, the light intensity can be used to program intermediate states in a memristor device (Figure 8d). The programmed resistance states of the OGT–memristor circuit were determined from the data shown in Figure 8c using the slope of the rising edge of the erase sweep peak.
The OGT–memristor circuit for the Sn + OGT–memristor was also pulse tested (Figure 9). As in the quasi-static I–V measurements (Figure 8), no series resistance was added to the pulsed test OGT–memristor measurement circuit. Because of the large capacitance of the experimental setup, long pulse widths (0.5–1 s) were used to reduce capacitive effects. However, it is noted that the memristor has switching speeds in the 10–9 s regime, (34) and the OGT has 10–6 s switching speeds (Figure 7)—much faster than used for the OGT–memristor circuit measurements. Figure 9a demonstrates that the OGT–memristor circuit cycles the memristor between high and low resistance states. In these tests, the OGT–memristor circuit applies a pulse to the OGT to bias it within one of the operation regions of the OGT (Figure 6b). OGT bias pulses correspond to the black traces (right graph axes) in Figure 9a–c. A “read” pulse biases the OGT to a voltage low enough not to perturb the memristor state; this is within the thermally generated current region (region 1 in Figure 6b) of OGT operation. A “write” OGT pulse biases the OGT in the saturation region (region 3 in Figure 6b) in the positive potential direction, thus programming the memristor to a low resistance value. An “erase” OGT pulse also biases the OGT in the saturation region, but in the negative potential direction, thus programming the memristor to a higher resistance value. The voltage across the memristor (left axis, red trace, Figure 9a–c) measures the memristor state change since no series resistor was added to the test circuit. A lower memristor resistance state corresponds to a smaller voltage measurement across the memristor.
Figure 9
Figure 9. Sn + OGT–memristor circuit response to programming pulses. (a) OGT–memristor circuit cycling. (b) Consecutive application of write pulses. (c) Consecutive application of erase pulses. The OGT pulse (black trace, right vertical axis) corresponds to the pulse that biases the OGT into a saturated state. The memristor voltage corresponds to the voltage measured across the memristor. The OGT is driven by a 770 nm LED.
The pulse testing data in Figure 9b,c demonstrate another approach (besides methods of compliance current dc or pulsed programming) of continuously programming the memristor state via continuous resistance changes through either repeating Write (Figure 9b) or Erase (Figure 9c) pulses. The blue curve in each graph highlights the change in memristor resistance state as a function of consecutive pulses. It is clear from these data that the memristance is capable of multistate programming into a continuous range of write and erase states.
Conclusion
Click to copy section linkSection link copied!
A multipurpose optically gated transistor has been demonstrated that enables nonvolatile photonic memory and (1) functions as an access transistor for a memristor device, (2) enables uniform voltage memristor programming, (3) allows a memristor to be programmed within a continuous range of states by using light intensity applied to the OGT gate, and (4) includes built-in compliance current limiting via the OGT saturation current. When light is applied, the transistor turns ON, analogous to the application of a gate voltage turning on an enhancement mode MOSFET. The applied gate light intensity controls the saturation current level of the OGT. This is an advantage of amorphous materials and their large concentration of trap states: the series resistance in the saturation state acts as a current limiter for the memristor, in the same role as a compliance current limit. Compliance current limiting has also been demonstrated in an amorphous Cu/Ge2Sb2Te5-based memory device with an Ag/TiO2-based selector device. (20) The memory device programming current in that example is self-limiting (similar to a compliance current) likely due to the formation of a CuTe phase within the memory cell which limits growth of the switching Cu filament. In contrast, the OGT in this work also provides compliance current limiting to the memristor but does not rely on the formation and dissolution of a CuTe phase and Cu filament growth control, both of which will have associated randomness due to the amorphous materials disorder. In the OGT, the magnitude of the current that is allowed through the memristor sets the device resistance state; the light intensity therefore directly programs the memristor state. Above a minimum light level, variable light intensity corresponds linearly to the change in memristance. The OGT also shows a reasonable photoresponse time (<15 μs) compared to other light-gated memristors, making it promising for application in a large memristor array. In addition to the light-intensity-based compliance current option for continuous memristor state programming, the application of consecutive pulses to the OGT–memristor circuit can also be used for continuous write and erase state programming.
There are advantages to specifically using amorphous chalcogenide materials in the OGT device, in addition to the advantages stated for amorphous materials in general. Chalcogenides are known for their optical activity and have been used in many optical applications, for example, as photodetectors, phase-change memory devices, waveguides, and photoresist, and their properties have been well-investigated. In addition, the Ge2Se3 OGT devices developed here are compatible with ion-conducting Ge2Se3-based memristors, (34) including the W-Ge2Se3 cosputtered memristor used in this work, making them ideal for incorporation into an existing fabrication process flow. Our ongoing research has shown that, in addition to Sn and Cu, other dopants such as Al, C, and Ti, to name a few, can be used in the Ge2Se3 glass to alter the access device properties, including threshold voltages, saturation current levels, photoresponse time, and light wavelength sensitivity. This research direction may show that OGT devices can be tuned to specific applications. Other research directions include investigation of the effects of oxygen in the OGT. Because the devices tested in this work were not encapsulated, it is possible that oxygen has entered the films over time; however, no effects on the OGT operation have been observed over the past year. Additionally, annealing, which can alter the trap density, and aging are both topics that need to be studied with respect to the OGT operation and are part of the ongoing characterization efforts.
Supporting Information
Click to copy section linkSection link copied!
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsaelm.8b00034.
I–V curves for the undoped Ge2Se3–memristor circuit in Figures S1 and S2 (DOCX)
Terms & Conditions
Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.
Author Information
Click to copy section linkSection link copied!
- Kristy A. Campbell - Department of Electrical and Computer Engineering, Boise State University, 1910 University Dr., Boise, Idaho 83725, United States;
http://orcid.org/0000-0003-3734-3616;
The manuscript was written through contributions of all authors. K.A.C. developed the OGT and materials, fabrication methods, and electrical testing procedures, supervised all fabrication and measurements, and prepared the manuscript. R.A.B., M.F.K., and J.A. performed OGT fabrication and electrical testing and assisted with manuscript preparation. All authors have given approval to the final version of the manuscript.
This work was partially supported through sponsorship of M. Alexander Nugent Consulting under U.S. Air Force Research Laboratory Contract FA8750-16-C-0183.
Acknowledgments
Click to copy section linkSection link copied!
The authors acknowledge support from the Department of Electrical and Computer Engineering at Boise State University for this work and Knowm, Inc., for donating the memristors used in this work.
| I–V | current–voltage |
| MOSFET | metal oxide semiconductor field-effect transistor |
| OGT | optically gated transistor |
| RAM | random access memory |
| ReRAM | resistive random access memory |
| UV–vis | ultraviolet–visible. |
References
Click to copy section linkSection link copied!
This article references 40 other publications.
- 1Chua, L. O. Memristor-The Missing Circuit Element. IEEE Trans. Circuit Theory 1971, 18, 507, DOI: 10.1109/TCT.1971.1083337Google ScholarThere is no corresponding record for this reference.
- 2Chua, L. Everything You Wish to Know About Memristors but are Afraid to Ask. Radioengineering 2015, 24, 319– 368, DOI: 10.13164/re.2015.0319Google ScholarThere is no corresponding record for this reference.
- 3Teimoori, A.; Amirsoleimani, A.; Ahmadi, A.; Ahmadi, M. A 2M1M Crossbar Architecture: Memory. IEEE Trans. VLSI Syst. 2018, 26, 2608– 2618, DOI: 10.1109/TVLSI.2018.2799951Google ScholarThere is no corresponding record for this reference.
- 4Lashkare, S.; Panwar, N.; Kumbhare, P.; Das, B.; Ganguly, U. PCMO-Based RRAM and NPN Bipolar Selector as Synapse for Energy Efficient STDP. IEEE Electron Device Lett. 2017, 38, 1212– 1215, DOI: 10.1109/LED.2017.2723503Google Scholar4PCMO-based RRAM and NPN bipolar selector as synapse for energy efficient STDPLashkare, S.; Panwar, N.; Kumbhare, P.; Das, B.; Ganguly, U.IEEE Electron Device Letters (2017), 38 (9), 1212-1215CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)Resistance random access memories (RRAMs) are widely explored to show spike time dependent plasticity (STDP) as a learning rule to show biol. synaptic behavior, as these devices possess analog conductance change. To implement STDP, pre- and post-neuronal waveforms are superposed. Only the peak voltage changes the conductance of memory. But due to the remaining part of the waveform (which don't affect the conductance change), there is a significant amt. of inadvertent current flow leading to unnecessary energy consumption. In this letter, we exptl. demonstrate that, the PCMO-based RRAM, a memristor (1M) in series with the NPN selector can be used as a synapse to reduce the undesirable energy consumption. First, we propose the pre- and post-neuronal waveform engineering required for PCMO-based memristor + Selector (1S1M) to reduce energy consumption. Second, we demonstrate exptl. that 1S1M synapse gives >4700× redn. in energy consumption compared with 1M synapse for a single neuronal waveform. Third, we implement and compare anti-STDP for 1M and 1S1M synapse to show no loss of generality. Thus, we exptl. demonstrate 1S1M-based energy efficient synapse for brain inspired computing.
- 5Nugent, M. A.; Molter, T. W. AHaH Computing–From Metastable Switches to Attractors to Machine Learning. PLoS One 2014, 9, e85175 DOI: 10.1371/journal.pone.0085175Google Scholar5AHaH computing-from metastable switches to attractors to machine learningNugent, Michael Alexander; Molter, Timothy WesleyPLoS One (2014), 9 (2), e85175/1-e85175/29, 29 pp.CODEN: POLNCL; ISSN:1932-6203. (Public Library of Science)Modern computing architecture based on the sepn. of memory and processing leads to a well known problem called the von Neumann bottleneck, a restrictive limit on the data bandwidth between CPU and RAM. This paper introduces a new approach to computing we call AHaH computing where memory and processing are combined. The idea is based on the attractor dynamics of volatile dissipative electronics inspired by biol. systems, presenting an attractive alternative architecture that is able to adapt, self-repair, and learn from interactions with the environment. We envision that both von Neumann and AHaH computing architectures will operate together on the same machine, but that the AHaH computing processor may reduce the power consumption and processing time for certain adaptive learning tasks by orders of magnitude. The paper begins by drawing a connection between the properties of volatility, thermodn., and Anti-Hebbian and Hebbian (AHaH) plasticity. We show how AHaH synaptic plasticity leads to attractor states that ext. the independent components of applied data streams and how they form a computationally complete set of logic functions. After introducing a general memristive device model based on collections of metastable switches, we show how adaptive synaptic wts. can be formed from differential pairs of incremental memristors. We also disclose how arrays of synaptic wts. can be used to build a neural node circuit operating AHaH plasticity. By configuring the attractor states of the AHaH node in different ways, high level machine learning functions are demonstrated. This includes unsupervised clustering, supervised and unsupervised classification, complex signal prediction, unsupervised robotic actuation and combinatorial optimization of procedures-all key capabilities of biol. nervous systems and modern machine learning algorithms with real world application.
- 6Burr, G. W.; Shelby, R. M.; Sebastian, A.; Kim, S.; Kim, S.; Sidler, S.; Virwani, K.; Ishii, M.; Narayanan, P.; Fumarola, A.; Sanches, L. L.; Boybat, I.; Le Gallo, M.; Moon, K.; Woo, J.; Hwang, H.; Leblebici, Y. Neuromorphic Computing Using Non-Volatile Memory. Advances in Physics: X 2017, 2, 89– 124, DOI: 10.1080/23746149.2016.1259585Google ScholarThere is no corresponding record for this reference.
- 7Chen, Z.; Schoeny, C.; Dolecek, L. Hamming Distance Computation in Unreliable Resistive Memory. IEEE. Trans. On Comm. 2018, 66, 5013, DOI: 10.1109/TCOMM.2018.2840717Google ScholarThere is no corresponding record for this reference.
- 8Song, B.; Xu, H.; Liu, S.; Liu, H.; Li, Q. Threshold Switching Behavior of Ag-SiTe-Based Selector Device and Annealing Effect on its Characteristics. IEEE J. Electron Devices Soc. 2018, 6, 674– 679, DOI: 10.1109/JEDS.2018.2836400Google Scholar8Threshold switching behavior of ag-site-based selector device and annealing effect on its characteristicsSong, Bing; Xu, Hui; Liu, Sen; Liu, Haijun; Li, QingjiangIEEE Journal of the Electron Devices Society (2018), 6 (1), 674-679CODEN: IJEDAC; ISSN:2168-6734. (Institute of Electrical and Electronics Engineers)Programmable metalization cell is one of important threshold switching selectors. We first performed a study on the selector based on amorphous chalcogenide material (Si0.4Te0.6) because of the rigid structure applied in ovonic threshold switch. In the meantime, annealing process is implemented to improve the performance. Results show that devices without annealing process demonstrate a minor threshold switching characteristic, revealing the potential as selector for cross-point memristor array. After implementing annealing process, threshold voltage (Vth), selectivity and endurance of selectors improve. Meanwhile, requirements of high current and a low holding voltage (Vh) for an ideal selector are fulfilled. Using the Ag filament formed during motion of Ag ions, a steep-slope (1.7 mV/dec) for threshold switching with high selectivity (∼ 104) could be achieved. Owing to the faster diffusivity of Ag atoms in solid-electrolytes, the resulting Ag filament easily dissolved under low current regime. It is deduced that performance improvement is due to the defect redn. within annealing process. Finally, time characteristics of selector devices are tested to verify fast switching and recovery speed for practical applicability.
- 9Koo, Y.; Lee, S.; Park, S.; Yang, M.; Hwang, H. Simple Binary Ovonic Threshold Switching Material SiTe and its Excellent Selector Performance for High-Density Memory Array Application. IEEE Electron Device Lett. 2017, 38, 568, DOI: 10.1109/LED.2017.2685435Google Scholar9Simple binary ovonic threshold switching material SiTe and its excellent selector performance for high-density memory array applicationKoo, Yunmo; Lee, Sangmin; Park, Seonggeon; Yang, Minkyu; Hwang, HyunsangIEEE Electron Device Letters (2017), 38 (5), 568-571CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)In this letter, simple binary Ovonic threshold switching (OTS) material with outstanding selector device performance has been demonstrated. Even with its simple material compn. and easy fabrication process, the selector device with the binary OTS material showed excellent selector performance such as high-OFF resistance ( > 1 G Ω at 0.1 V), low-ON resistance (< 1 k Ω at 2.0 V), extremely sharp switching slope (< 1 mV/dec), fast operating speed (ttransition < 2 ns, tdelay < 7 ns), high endurance (>10♂ cyclesof 150 ns pulse),high elec. stability (>1ks at 1.2 V), and high thermal stability (> 400 °C / 30 min). Furthermore, conduction mechanism of the OTS has been explained by Poole-Frenkel-based anal. modeling.
- 10Burr, G. W.; Shenoy, R. S.; Hwang, H. Select Device Concepts for Crossbar Arrays. In Resistive Switching: From Fundamentals of Nanoionic Redox Processes to Memristive Device Applications, 1st ed.; Ielmini, D., Waser, R., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2016.Google ScholarThere is no corresponding record for this reference.
- 11Liu, M.; Wang, W. Application of Nanojunction-Based RRAM to Reconfigurable IC. Micro Nano Lett. 2008, 3, 101– 105, DOI: 10.1049/mnl:20080029Google ScholarThere is no corresponding record for this reference.
- 12Potteiger, T.; Robinson, W. H. A One Zener Diode, One Memristor Crossbar Architecture for a Write-Time-Based PUF. IEEE 58th International Midwest Symposium On Circuits And Systems , 2015; pp 1– 4.Google ScholarThere is no corresponding record for this reference.
- 13Zidan, M. A.; Chen, A.; Indiveri, G.; Lu, W. D. Memristive Computing Devices and Applications. J. Electroceram. 2017, 39, 4– 20, DOI: 10.1007/s10832-017-0103-0Google Scholar13Memristive computing devices and applicationsZidan, Mohammed A.; Chen, An; Indiveri, Giacomo; Lu, Wei D.Journal of Electroceramics (2017), 39 (1-4), 4-20CODEN: JOELFJ; ISSN:1385-3449. (Springer)A review. Advances in electronics have revolutionized the way people work, play and communicate with each other. Historically, these advances were mainly driven by CMOS transistor scaling following Moore's law, where new generations of devices are smaller, faster, and cheaper, leading to more powerful circuits and systems. However, conventional scaling is now facing major tech. challenges and fundamental limits. New materials, devices, and architectures are being aggressively pursued to meet present and future computing needs, where tight integration of memory and logic, and parallel processing are highly desired. To this end, one class of emerging devices, termed memristors or memristive devices, have attracted broad interest as a promising candidate for future memory and computing applications. Besides tremendous appeal in data storage applications, memristors offer the potential to enable efficient hardware realization of neuromorphic and analog computing architectures that differ radically from conventional von Neumann computing architectures. In this review, we analyze representative memristor devices and their applications including mixed signal analog-digital neuromorphic computing architectures, and highlight the potential and challenges of applying such devices and architectures in different computing applications.
- 14Woo, J.; Peng, X.; Yu, S. Design Considerations of Selector Device in Cross-Point RRAM Array for Neuromorphic Computing. 2018 IEEE International Symposium on Circuits and Systems (ISCAS), Florence , 2018; pp 1– 4.Google ScholarThere is no corresponding record for this reference.
- 15Cyrille, M. C.; Verdy, A.; Navarro, G.; Bourgeois, G.; Garrione, J.; Bernard, M.; Sabbione, C.; Noé, P.; Nowak, E. OTS Selector Devices: Material Engineering for Switching Performance. 2018 International Conference on IC Design & Technology (ICICDT), Otranto , 2018; pp 113– 116.Google ScholarThere is no corresponding record for this reference.
- 16Avasarala, N. S.; Govoreanu, B.; Opsomer, K.; Devulder, W.; Clima, S.; Detavernier, C.; van der Veen, M.; Van Houdt, J.; Henys, M.; Goux, L.; Kar, G. S. Doped GeSe Materials for Selector Applications. 2017 47th European Solid-State Device Research Conference (ESSDERC), Leuven , 2017; pp 168– 171.Google ScholarThere is no corresponding record for this reference.
- 17Luo, Q.; Xu, X.; Lv, H.; Gong, T.; Long, S.; Liu, Q.; Li, L.; Liu, M. Endurance Characterization of the Cu-Dope HfO2 Based Selection Device With One Transistor-One Selector Structure. 2017 IEEE Electron Devices Technology and Manufacturing Conference (EDTM), Toyama , 2017; pp 178– 179.Google ScholarThere is no corresponding record for this reference.
- 18Park, J.; Yoo, J.; Song, J.; Sung, C.; Hwang, H. Hybrid Selector With Excellent Selectrivity and Fast Switching Speed for X-Point Memory Array. IEEE Electron Device Lett. 2018, 39, 1171– 1174, DOI: 10.1109/LED.2018.2845878Google Scholar18Hybrid selector with excellent selectivity and fast switching speed for X-point memory arrayPark, Jaehyuk; Yoo, Jongmyung; Song, Jeonghwan; Sung, Changhyuck; Hwang, HyunsangIEEE Electron Device Letters (2018), 39 (8), 1171-1174CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)In this letter, we demonstrate a new concept of threshold selector by combining Ag-based selector with NbOx insulator metal transition (IMT) selector. To overcome intrinsic limits of slow turn-off speed (>1μs) of the Ag-based selector, we adopted NbO x IMT selector with fast turn-off speed. The hybrid device exhibits excellent selector characteristics, such as extremely low off-current (<1 pA) and fast turn-off speed (<1 μs). It shows excellent readout margin (>75% at >210 word lines) in a fast sequential operation of cross point memory array.
- 19Song, B.; Xu, H.; Liu, S.; Liu, H.; Li, Q. Threshold Switching Behavior of Ag-SiTe-Based Selector Device and Annealing Effect on Its Characteristics. IEEE J. Electron Devices Soc. 2018, 6, 674– 679, DOI: 10.1109/JEDS.2018.2836400Google Scholar19Threshold switching behavior of ag-site-based selector device and annealing effect on its characteristicsSong, Bing; Xu, Hui; Liu, Sen; Liu, Haijun; Li, QingjiangIEEE Journal of the Electron Devices Society (2018), 6 (1), 674-679CODEN: IJEDAC; ISSN:2168-6734. (Institute of Electrical and Electronics Engineers)Programmable metalization cell is one of important threshold switching selectors. We first performed a study on the selector based on amorphous chalcogenide material (Si0.4Te0.6) because of the rigid structure applied in ovonic threshold switch. In the meantime, annealing process is implemented to improve the performance. Results show that devices without annealing process demonstrate a minor threshold switching characteristic, revealing the potential as selector for cross-point memristor array. After implementing annealing process, threshold voltage (Vth), selectivity and endurance of selectors improve. Meanwhile, requirements of high current and a low holding voltage (Vh) for an ideal selector are fulfilled. Using the Ag filament formed during motion of Ag ions, a steep-slope (1.7 mV/dec) for threshold switching with high selectivity (∼ 104) could be achieved. Owing to the faster diffusivity of Ag atoms in solid-electrolytes, the resulting Ag filament easily dissolved under low current regime. It is deduced that performance improvement is due to the defect redn. within annealing process. Finally, time characteristics of selector devices are tested to verify fast switching and recovery speed for practical applicability.
- 20Song, J.; Woo, J.; Lim, S.; Chekol, S. A.; Hwang, H. Self-Limited CBRAM With Threshold Selector for 1S1R Crossbar Array Applications. IEEE Electron Device Lett. 2017, 38, 1532– 1535, DOI: 10.1109/LED.2017.2757493Google Scholar20Self-limited CBRAM with threshold selector for 1S1R crossbar array applicationsSong, Jeonghwan; Woo, Jiyong; Lim, Seokjae; Chekol, Solomon Amsalu; Hwang, HyunsangIEEE Electron Device Letters (2017), 38 (11), 1532-1535CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)In this letter, we demonstrate a self-limited conductive-bridging random accessmemory (CBRAM) that removes the necessity for external current compliance in a one selector-one resistor (1S1R) architecture. The std. Ge2Sb2Te5 (GST) is used as a CBRAM switching layer. In addn., Te-rich GST is also considered. Their performance is then compared. Both samples exhibit selflimited on-current characteristics, and the on-currents of the std. GST and Te-rich GST are ∼300 and ∼20μA, resp. The obsd. self-limited characteristics are caused by the Te in the GST layer because in the presence of Te, Cu tends to form a more stable CuTe phase that restrict Cu filament growth. Furthermore, to confirm the feasibility of crossbar array applications, the 1S1R device is evaluated using a Ag/TiO2-based threshold selector device reported in our previous work. Hence, we confirm leakage current redn., a uniform resistance distribution, and stable retention characteristics in the 1S1R configuration with no external current compliance.
- 21Zhai, Y.; Yang, J.-Q.; Zhou, Y.; Mao, J.-Y.; Ren, Y.; Roy, V. A. L.; Han, S.-T. Toward Non-Volatile Photonic Memory: Concept, Material and Design. Mater. Horiz. 2018, 5, 641– 654, DOI: 10.1039/C8MH00110CGoogle Scholar21Toward non-volatile photonic memory: concept, material and designZhai, Yongbiao; Yang, Jia-Qin; Zhou, Ye; Mao, Jing-Yu; Ren, Yi; Roy, Vellaisamy A. L.; Han, Su-TingMaterials Horizons (2018), 5 (4), 641-654CODEN: MHAOBM; ISSN:2051-6355. (Royal Society of Chemistry)Digital technol. is one of the greatest modern breakthroughs, allowing sounds, words and images to be stored in binary form. However, there is a huge gap between the amt. of data created daily and the capacities of existing storage media. Developing multibit memory in which 2n levels, typically represented by distinguishable current levels, can be achieved in a single cell is a crit. specification for achieving high-d. memory devices. Compared with elec. operated memory, photonic memory-in which elec. read-out is orthogonal to the photo-programming operation-promises high differentiation among different data levels. From another aspect, benefiting from its high d., multifunctionality, low power consumption, and multilevel data storage, photonic memory devices hold future promise for built-in, non-volatile memory and reconstructed logic operation and are expected to bridge this capacity gap. Thus, we present a review on the development of photonic memory, with a view towards inspiring more intriguing ideas on the elegant selection of materials and design of novel device structures that may finally induce major progress in the manuf. and application of photonic memory.
- 22El Gharras, Z.; Bourahla, A.; Vautier, C. Role of Photoinduced Defects in Amorphous GexSe1-xPhotoconductivity. J. Non-Cryst. Solids 1993, 155, 171– 179, DOI: 10.1016/0022-3093(93)91322-TGoogle Scholar22Role of photoinduced defects in amorphous germanium-selenium (GexSe1-x) photoconductivityEl Gharras, Z.; Bourahla, A.; Vautier, C.Journal of Non-Crystalline Solids (1993), 155 (2), 171-9CODEN: JNCSBJ; ISSN:0022-3093.The spectral dependence of photocond. of amorphous pure Se and Ge-Se alloy (4 and 8 at.% Ge) evapd. thin films was measured for the spectral range between 250 and 800 nm. A comparison of the Se and Ge-Se spectra shows that the magnitude of photocurrent decreases as the Ge content increases. This result indicates a correlation between the magnitude of the photocurrent and the d. of photoinduced defects. The contribution of these defects to the enhancement of photocond. in a specific region of the spectrum as well as the influence of temp. on their d. are also discussed.
- 23Tanaka, K.; Shimakawa, K. Amorphous Chalcogenide Semiconductors and Related Materials; Springer: New York, 2011.Google ScholarThere is no corresponding record for this reference.
- 24Mukherjee, B.; Tok, E. S.; Sow, C. H. Photocurrent Characteristics of Individual GeSe2 Nanobelt With Schottky Effects. J. Appl. Phys. 2013, 114, 134302, DOI: 10.1063/1.4823779Google Scholar24Photocurrent characteristics of individual GeSe2 nanobelt with Schottky effectsMukherjee, Bablu; Tok, Eng Soon; Sow, Chorng HaurJournal of Applied Physics (Melville, NY, United States) (2013), 114 (13), 134302/1-134302/10CODEN: JAPIAU; ISSN:0021-8979. (American Institute of Physics)Single crystal GeSe2 nanobelts (NBs) were successfully grown using chem. vapor deposition techniques. The morphol. and structure of the nanostructures were characterized using SEM, transmission electron microscopy, X-ray diffractometry, and Raman spectroscopy. Electronic transport properties, photoconductive characteristics, and temp.-dependent electronic characteristics were examd. on devices made of individual GeSe2 nanobelt. The current increased by 3 orders of magnitude upon laser irradn. (wavelength 532 nm and intensity ∼6.8 mW/cm2) with responsivity of ∼2764 A/W at fixed 4 V bias. Localized photocond. study shows that the large photoresponse of the device primarily occurs at the metal-NB contact regions. In addn., the elec. Schottky nature of nanobelt/Au contact and p-type cond. nature of GeSe2 nanobelt are extd. from the current-voltage characteristics and spatially resolved photocurrent measurements. The high sensitivity and quick photoresponse in the visible wavelength range indicate potential applications of individual GeSe2 nanobelt devices in realizing optoelectronic switches. (c) 2013 American Institute of Physics.
- 25Liu, W. C.; Hoffman, G.; Zhou, W.; Reano, R. M.; Boolchand, P.; Sooryakumar, R. Slab Waveguides and Nanoscale Patterning of Pulsed Laser-Deposited Ge0.2Se0.8 Chalcogenide Films. Appl. Phys. Lett. 2008, 93, 041107, DOI: 10.1063/1.2965124Google Scholar25Slab waveguides and nanoscale patterning of pulsed laser-deposited Ge0.2Se0.8 chalcogenide filmsLiu, W. C.; Hoffman, G.; Zhou, W.; Reano, R. M.; Boolchand, P.; Sooryakumar, R.Applied Physics Letters (2008), 93 (4), 041107/1-041107/3CODEN: APPLAB; ISSN:0003-6951. (American Institute of Physics)Planar slab waveguides were fabricated by pulsed laser deposition from GexSe1-x glass compds. (x ∼ 0.2) that lies very close to the floppy to rigid stiffness transition. These high quality active structures, which were deposited on SiO2 cladding layers above Si substrates, support several transverse-elec. (TE) modes, and a loss of 0.24 dB/cm for the TE0 mode was measured at 632.8 nm wavelength. The ability to exploit electron beam writing at these special Ge in Se compns. to create nanoscale surface motifs are promising advances to create unique miniature optical processing devices. (c) 2008 American Institute of Physics.
- 26Edwards, T. G.; Sen, S. Structure and Relaxation in Germanium Selenide Glasses and Supercooled Liquids: A Raman Spectroscopic Study. J. Phys. Chem. B 2011, 115, 4307– 4314, DOI: 10.1021/jp202174xGoogle Scholar26Structure and Relaxation in Germanium Selenide Glasses and Supercooled Liquids: A Raman Spectroscopic StudyEdwards, T. G.; Sen, S.Journal of Physical Chemistry B (2011), 115 (15), 4307-4314CODEN: JPCBFK; ISSN:1520-5207. (American Chemical Society)The structure of the tetrahedral backbone and the nature and time scale of the temp.-dependent structural changes in binary Ge-Se glasses and supercooled liqs. with ≤ 33.33 atom % Ge have been investigated using ambient and high-temp. Raman spectroscopy. The compn. dependence of the relative fractions of edge- and corner-shared GeSe4 tetrahedra and that of the characteristic mean vibrational frequencies of these structural units are shown to be consistent with a structural model for these glasses based on a random interconnection between GeSe4 tetrahedra and Se-Se chain fragments. The most prominent temp.-dependent structural change in the Ge20Se80 glass and supercooled liq. involves progressive conversion of the edge-shared GeSe4 tetrahedra into corner-shared tetrahedra, upon lowering of temp. The time scale of this tetrahedral conversion "reaction" corresponds well with those of enthalpy and shear relaxation near glass transition. Moreover, the temp. dependence of this GeSe4 tetrahedral speciation is shown to be the most important source for the prodn. of configurational entropy in this supercooled liq. near the glass-transition range, signifying a direct link between structural relaxation, configurational entropy, and viscous flow.
- 27Mott, N. F.; Davis, E. A. Electronic Processes in Non-crystalline Materials; Oxford University Press: London, 1971; p 238.Google ScholarThere is no corresponding record for this reference.
- 28El-Shair, H. T.; Fouad, S. S. Optical Properties of Thin GexSe1-x Amorphous Films. Vacuum 1991, 42, 463– 467, DOI: 10.1016/0042-207X(91)90017-DGoogle Scholar28Optical properties of thin germanium-selenium (GexSe1-x) amorphous filmsEl-Shair, H. T.; Fouad, S. S.Vacuum (1991), 42 (7), 463-7CODEN: VACUAV; ISSN:0042-207X.The optical consts. of GexSe1-x amorphous thin films of different thicknesses (23-335 nm) and of different compns. (0.05 ≤ x ≤ 0.30) were detd. in the wavelength range 400-2000 nm. Both the refractive index n and the absorption index (k) are independent of the film thickness for a given soln., while both on the Ge content. The spectra indicate the existence of indirect optical transitions. The corresponding forbidden band gaps were detd. and the energy gap of GexSe7-x amorphous thin films obeys the relation EGeSe = YEGe + (1-Y)ESe, where Y is the vol. fraction of Ge.
- 29Feltz, A. Amorphous Inorganic Materials and Glasses; VCH Publishers: New York, 1993; p 238.Google ScholarThere is no corresponding record for this reference.
- 30Xia, C.; Du, J.; Xiong, W.; Jia, Y.; Wei, Z.; Li, J. A type-II GeSe/SnS Heterobilayer With a Suitable Direct Gap, Superior Optical Absorption and Broad Spectrum for Photovoltaic Applications. J. Mater. Chem. A 2017, 5, 13400, DOI: 10.1039/C7TA02109GGoogle Scholar30A type-II GeSe/SnS heterobilayer with a suitable direct gap, superior optical absorption and broad spectrum for photovoltaic applicationsXia, Congxin; Du, Juan; Xiong, Wenqi; Jia, Yu; Wei, Zhongming; Li, JingboJournal of Materials Chemistry A: Materials for Energy and Sustainability (2017), 5 (26), 13400-13410CODEN: JMCAET; ISSN:2050-7496. (Royal Society of Chemistry)Van der Waals (vdW) heterobilayers are emerging as unique structures for next-generation electronic and optoelectronic devices. The authors predict that the GeSe/SnS heterobilayer has a direct band structure with a gap value of ∼1.519 eV and typical type-II band alignment. Moreover, it possesses the characteristics of superior optical absorption (∼105) and a broad absorption spectrum from the visible light to the near UV region. In addn., the GeSe/SnS heterobilayer also exhibits obviously anisotropic electronic transport and optical properties with larger current and stronger optical absorption along the zigzag direction. Meanwhile, interlayer coupling and applying an external elec. field are identified to be effective methods to modify its electronic and optical properties. Thus, these predicted results indicate that the GeSe/SnS heterobilayer will have promising applications in photovoltaic devices.
- 31Mark, P.; Helfrich, W. Space-Charge-Limited Currents in Organic Crystals. J. Appl. Phys. 1962, 33, 205– 215, DOI: 10.1063/1.1728487Google Scholar31Space-charge-limited currents in organic crystalsMark, Peter; Helfrich, WolfgangJournal of Applied Physics (1962), 33 (), 205-15CODEN: JAPIAU; ISSN:0021-8979.Elec.-cond. measurements were performed with thin (50-μ) single crystals of p-terphenyl, p-quaterphenyl, and anthracene supplied with aq. electrodes, one of which was an iodineiodide soln. (acceptor electrode), and the other an iodide soln. The results strongly indicate that the acceptor electrode can form an ohmic contact for hole injection into these crystals and that space-charge-limited currents can be drawn through them. The crystals contained holetrapping states, the location-in-energy of which can be approximated by a decreasing exponential distribution above the valence band. The measurements showed that the hole mobility in p-terphenyl is ∼3 × 10-2 cm.2/v. sec., is independent of the field at least up to about 4 ×104 v./ cm., and that the hole-trap concn. is at least 1013/cc. The acceptor electrode used does not form an ohmic contact to crystals of naphthalene and biphenyl; an explanation for this is proposed. Some theoretical aspects of ohmic contact formation to org. crystals and space-charge-limited current flow in insulators are also discussed.
- 32Chiu, F.-C. A Review on Conduction Mechanisms in Dielectric Films. Adv. Mater. Sci. Eng. 2014, 2014, 578168, DOI: 10.1155/2014/578168Google ScholarThere is no corresponding record for this reference.
- 33Korolev, D. S.; Belov, A. I.; Okulich, E. V.; Okulich, V. I.; Guseinov, D. V.; Sidorenko, K. V.; Shuisky, R. A.; Antonov, I. N.; Gryaznov, E. G.; Gorshkov, O. N.; Tetelbaum, D. I.; Mikhaylov, A. N. Manipulation of Resistive State of Silicon Oxide Memristor by Means of Current Limitation During Electroforming. Superlattices Microstruct. 2018, 122, 371, DOI: 10.1016/j.spmi.2018.07.006Google Scholar33Manipulation of resistive state of silicon oxide memristor by means of current limitation during electroformingKorolev, D. S.; Belov, A. I.; Okulich, E. V.; Okulich, V. I.; Guseinov, D. V.; Sidorenko, K. V.; Shuisky, R. A.; Antonov, I. N.; Gryaznov, E. G.; Gorshkov, O. N.; Tetelbaum, D. I.; Mikhaylov, A. N.Superlattices and Microstructures (2018), 122 (), 371-376CODEN: SUMIEK; ISSN:0749-6036. (Elsevier Ltd.)Resistive switching and adaptive behavior of resistive state in response to elec. stimulation has been studied for the silicon oxide based memristive devices subjected to electroforming in the conditions of current compliance in comparison with the analogous memristive devices after electroforming without any current limitation. The limitation of current and temp. during electroforming affects the parameters of growing conductive filament ensembles and redn.-oxidn. reactions resulting in a gradual character and a wide dynamic range of resistance change important for neuromorphic applications.
- 34Campbell, K. A. Self-Directed Channel Memristor for High Temperature Operation. Microelectron. J. 2017, 59, 10– 14, DOI: 10.1016/j.mejo.2016.11.006Google Scholar34Self-directed channel memristor for high temperature operationCampbell, Kristy A.Microelectronics Journal (2017), 59 (), 10-14CODEN: MICEB9; ISSN:0026-2692. (Elsevier Ltd.)Ion-conducting memristors comprised of the layered chalcogenide materials Ge2Se3/SnSe/Ag are described. The memristor, termed a self-directed channel (SDC) device, can be classified as a generic memristor and can tolerate continuous high temp. operation (at least 150 °C). Unlike other chalcogenide-based ion conducting device types, the SDC device does not require complicated fabrication steps, such as photodoping or thermal annealing, making these devices faster and more reliable to fabricate. Device pulsed response shows fast state switching in the 10-9 s range. Device cycling at both room temp. and 140 °C show write endurance of at least 1 billion.
- 35Liu, D.; Cheng, H.; Zhu, X.; Wang, G.; Wang, N. Analog Memristors Based on Thickening/Thinning of Ag Nanofilaments in Amorphous Manganite Thin Films. ACS Appl. Mater. Interfaces 2013, 5, 11258– 64, DOI: 10.1021/am403497yGoogle Scholar35Analog Memristors Based on Thickening/Thinning of Ag Nanofilaments in Amorphous Manganite Thin FilmsLiu, Dongqing; Cheng, Haifeng; Zhu, Xuan; Wang, Guang; Wang, NannanACS Applied Materials & Interfaces (2013), 5 (21), 11258-11264CODEN: AAMICK; ISSN:1944-8244. (American Chemical Society)The authors developed an analog memristor based on the thickening/thinning of Ag nanofilaments in amorphous La1-xSrxMnO3 (a-LSMO) thin films. The Ag/a-LSMO/Pt memristor exhibited excellent pinched hysteresis loops under high-excitation frequency, and the areas enclosed by the pinched hysteresis loops shrank with increasing excitation frequency, which is a characteristic typical of a memristor. The memristor also showed continuously tunable synapselike resistance and stable endurance. The a-LSMO thin films in the memristor acted as a solid electrolyte for Ag+ cations, and only the Ag/a-LSMO/Pt memristor electroformed with a larger current compliance easily exhibited high-frequency pinched hysteresis loops. From the electrochem. metalization (ECM) theory and elec. transport models of quantum wires and nanowires, the memristance is ultimately detd. by the amt. of charge supplied by the external current. The state equations of the memristor were established, and charge was the state variable. This study provides a new analog memristor based on metal nanofilaments thickening/thinning in ECM cells, which can be extended to other resistive switching materials. The new memristor may enable the development of beyond von Neumann computers.
- 36Lee, T.-W.; Nickel, J. H. Memristor Resistance Modulation for Analog Applications. IEEE Electron Device Lett. 2012, 33, 1456– 1458, DOI: 10.1109/LED.2012.2207429Google Scholar36Memristor resistance modulation for analog applicationsLee, Tsung-Wen; Nickel, Janice H.IEEE Electron Device Letters (2012), 33 (10), 1456-1458CODEN: EDLEDZ; ISSN:0741-3106. (Institute of Electrical and Electronics Engineers)The resistance modulation (RM) of TaOx-based memristors can be precisely controlled by the SET switching compliance current. After electroforming, switching occurs in a rebuilt oxide between the electroformed conductive filament and the Pt electrode. RM is independent of initial oxide thickness. The switching mechanism is postulated as dielec. breakdown at the sidewall of the conductive channel created within the rebuilt oxide: During a SET operation to a lower resistance state, the conductive channel increases in size, conducting a larger current until limited by the external compliance; during a RESET operation to a higher resistance state, the tip of the oxygen-satd. tantalum conductive channel is oxidized, reforming the rebuilt oxide. The conduction of the rebuilt oxide follows a power law function of voltage, in parallel with the modulated channel conductance. The linear resistance can be randomly programmed with accuracy and reproducibility. Analog circuits of tunable memristive low-pass and high-pass filters demonstrate frequency tuning by RM.
- 37Kim, K. M.; Kim, G. H.; Song, S. J.; Seok, J. Y.; Lee, M. H.; Yoon, J. H.; Hwang, C. S. Electrically Configurable Electroforming and Bipolar Resistive Switching in Pt /TiO2 /Pt Structures. Nanotechnology 2010, 21, 305203, DOI: 10.1088/0957-4484/21/30/305203Google Scholar37Electrically configurable electroforming and bipolar resistive switching in Pt/TiO2/Pt structuresKim, Kyung Min; Kim, Gun Hwan; Song, Seul Ji; Seok, Jun Yeong; Lee, Min Hwan; Yoon, Jeong Ho; Hwang, Cheol SeongNanotechnology (2010), 21 (30), 305203/1-305203/7CODEN: NNOTER; ISSN:1361-6528. (Institute of Physics Publishing)This study examd. the effects of elec. forming methods on the bipolar resistance switching (BRS) behavior in Pt/TiO2/Pt sandwich structures. The BRS is confined to a region near the ruptured end of conducting nanofilaments, which are composed of a TinO2n-1 Magneli phase formed by electroforming. The intermediate phase with an oxygen vacancy concn. between the insulating TiO2 and the residual conducting filament that formed at the interface region was considered to be the switching layer (SL). The change in filament shape caused by a variation in the compliance current during filament formation resulted in a different filament rupture location and SL configuration. Precise control of the filament formation and rupture process resulted in SLs connected in an anti-parallel configuration. It was possible to reconfigure the SLs in the same fashion without any restraints, which allowed an unlimited memristive operation to be achieved. This paper presents a new technique in voltage sweep mode that applies a compliance current as a tool to achieve a memristor with unlimited operation.
- 38Mukherjee, B.; Cai, Y.; Tan, H. R.; Feng, Y. P.; Tok, E. S.; Sow, C. H. NIR Schottky Photodetectors Based on Individual Single-Crystalline GeSe Nanosheet. ACS Appl. Mater. Interfaces 2013, 5, 9594, DOI: 10.1021/am402550sGoogle Scholar38NIR Schottky photodetectors based on individual single-crystalline GeSe nanosheetMukherjee, Bablu; Cai, Yongqing; Tan, Hui Ru; Feng, Yuan Ping; Tok, Eng Soon; Sow, Chorng HaurACS Applied Materials & Interfaces (2013), 5 (19), 9594-9604CODEN: AAMICK; ISSN:1944-8244. (American Chemical Society)The authors have synthesized high-quality, micrometer-sized, single-crystal GeSe nanosheets using vapor transport and deposition techniques. Photoresponse is investigated based on mech. exfoliated GeSe nanosheet combined with Au contacts under a global laser irradn. scheme. The nonlinear, asym., and unsatd. characteristics of the I-V curves reveal that two uneven back-to-back Schottky contacts are formed. First-principles calcns. indicate that the occurrence of defects-induced in-gap defective states, which are responsible for the slow decay of the current in the OFF state and for the weak light intensity dependence of photocurrent. The Schottky photodetector exhibits a marked photoresponse to NIR light illumination (max. photoconductive gain ∼5.3 × 102 % at 4 V) at a wavelength of 808 nm. The significant photoresponse and good responsitivity (∼3.5 A W-1) suggests its potential applications as photodetectors.
- 39Tan, H.; Liu, G.; Yang, H.; Yi, X.; Pan, L.; Shang, J.; Long, S.; Liu, M.; Wu, Y.; Li, R.-W. Light-Gated Memristor with Integrated Logic and Memory Functions. ACS Nano 2017, 11, 11298, DOI: 10.1021/acsnano.7b05762Google Scholar39Light-Gated Memristor with Integrated Logic and Memory FunctionsTan, Hongwei; Liu, Gang; Yang, Huali; Yi, Xiaohui; Pan, Liang; Shang, Jie; Long, Shibing; Liu, Ming; Wu, Yihong; Li, Run-WeiACS Nano (2017), 11 (11), 11298-11305CODEN: ANCAC3; ISSN:1936-0851. (American Chemical Society)Memristive devices are able to store and process information, which offers several key advantages over the transistor-based architectures. However, most of the two-terminal memristive devices have fixed functions once made and cannot be reconfigured for other situations. Here, we propose and demonstrate a memristive device "memlogic" (memory logic) as a nonvolatile switch of logic operations integrated with memory function in a single light-gated memristor. Based on nonvolatile light-modulated memristive switching behavior, a single memlogic cell is able to achieve optical and elec. mixed basic Boolean logic of reconfigurable "AND", "OR", and "NOT" operations. Furthermore, the single memlogic cell is also capable of functioning as an optical adder and digital-to-analog converter. All the memlogic outputs are memristive for in situ data storage due to the nonvolatile resistive switching and persistent photocond. effects. Thus, as a memdevice, the memlogic has potential for not only simplifying the programmable logic circuits but also building memristive multifunctional optoelectronics.
- 40Zhu, X.; Lu, W. D. Optogenetics-Inspired Tunable Synaptic Functions in Memristors. ACS Nano 2018, 12, 1242– 1249, DOI: 10.1021/acsnano.7b07317Google Scholar40Optogenetics-Inspired Tunable Synaptic Functions in MemristorsZhu, Xiaojian; Lu, Wei D.ACS Nano (2018), 12 (2), 1242-1249CODEN: ANCAC3; ISSN:1936-0851. (American Chemical Society)Two-terminal memristors with internal Ca2+-like dynamics can be used to faithfully emulate biol. synaptic functions and have been intensively studied for neural network implementations. Inspired by the optogenetic technique that uses light to tune the Ca2+ dynamics and subsequently the synaptic plasticity, the authors develop a CH3NH3PbI3 (MAPbI3)-based memristor that exhibits light-tunable synaptic behaviors. Specifically, by increasing the formation energy of iodine vacancy (V·I·I/V×I×I), light illumination can be used to control the V·I·I/V×I generation and annihilation dynamics, resembling light-controlled Ca2+ influx in biol. synapses. The memory formation and memory loss behaviors in the memristors can be modified by controlling the intensity and the wavelength of the illuminated light. Coincidence detection of elec. and light stimulations is also implemented in the memristive device with real-time (≤20 ms) response to light illumination. These results open options to modify the synaptic plasticity effects in memristor-based neuromorphic systems and can lead to the development of electronic systems that can faithfully emulate diverse biol. processes.
Cited By
Click to copy section linkSection link copied!
Citation Statements
beta
Smart citations by scite.ai include citation statements extracted from the full text of the citing article. The number of the statements may be higher than the number of citations provided by ACS Publications if one paper cites another multiple times or lower if scite has not yet processed some of the citing articles.
Explore this article's citation statements on scite.ai
This article is cited by 11 publications.
- Su-yeon Cho, Somnath S. Kundale, Junoh Shim, Sojeong Park, Hangil Kwak, Anh Vo Hoang, Eui-Tae Kim, Sunkook Kim, Jun Hong Park. Enhanced Carrier Dynamics and Excitation of Optically Stimulated Artificial Synapse Using van der Waals Passivation Layers. ACS Applied Materials & Interfaces 2025, 17
(4)
, 6460-6472. https://doi.org/10.1021/acsami.4c12694
- Karl Griffin, Gareth Redmond. A Purely Electronic Analog Memristive Device Based on a Squaraine Nanowire Mesh. ACS Applied Electronic Materials 2024, 6
(1)
, 261-275. https://doi.org/10.1021/acsaelm.3c01238
- Sharmila B, Priyanka Dwivedi. Effect of Temperature on TiO
2
/P‐Si‐Based Memristor Devices as Optical/Electrical Synapse. physica status solidi (a) 2025, 222
(3)
https://doi.org/10.1002/pssa.202400542
- Md Faisal Kabir, Kristy A. Campbell. Effects of Group IVA Elements on the Electrical Response of a Ge2Se3-Based Optically Gated Transistor. Micromachines 2024, 15
(8)
, 1000. https://doi.org/10.3390/mi15081000
- Lu Wang, Jiazhuang Li, Wantao Su, Dianzhong Wen. Photoelectric biomemristors for artificial visual perception systems. Applied Materials Today 2023, 35 , 101964. https://doi.org/10.1016/j.apmt.2023.101964
- Yuling Yan, Niannian Yu, Ziyan Yu, Yupeng Su, Jiawei Chen, Tao Xiang, Yuenan Han, Jiafu Wang. Optoelectronic Synaptic Memtransistor Based on 2D SnSe/MoS
2
van der Waals Heterostructure under UV–Ozone Treatment. Small Methods 2023, 7
(6)
https://doi.org/10.1002/smtd.202201679
- Hee Won Suh, Dong Su Kim, Ji Hoon Choi, Hak Hyeon Lee, Kun Woong Lee, Sung Hyeon Jung, Won Seok Yang, Jeong Jae Kim, Ji Sook Yang, Ho Seong Lee, Hyung Koun Cho. Multifunctional double active layers formed with electrochemically controlled nanoparticle dispersion for resistive switching memory arrays. Applied Surface Science 2023, 608 , 155206. https://doi.org/10.1016/j.apsusc.2022.155206
- Syed Ghazi Sarwat, Timoleon Moraitis, C. David Wright, Harish Bhaskaran. Chalcogenide optomemristors for multi-factor neuromorphic computation. Nature Communications 2022, 13
(1)
https://doi.org/10.1038/s41467-022-29870-9
- Nazek El-Atab. MemSor: Emergence of the In‐Memory Sensing Technology for the Digital Transformation. physica status solidi (a) 2022, 219
(2)
https://doi.org/10.1002/pssa.202100528
- Wuhong Xue, Wenjuan Ci, Xiao-Hong Xu, Gang Liu. Optoelectronic memristor for neuromorphic computing*. Chinese Physics B 2020, 29
(4)
, 048401. https://doi.org/10.1088/1674-1056/ab75da
- Jorge Gomez, Ioannis Vourkas, Angel Abusleme, Georgios Ch. Sirakoulis, Antonio Rubio. Voltage Divider for Self-Limited Analog State Programing of Memristors. IEEE Transactions on Circuits and Systems II: Express Briefs 2020, 67
(4)
, 620-624. https://doi.org/10.1109/TCSII.2019.2923716
Learn about these metrics
Article Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.
Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.
The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.
Recommended Articles
Abstract
Figure 1
Figure 1. Optically gated device structure and electrical measurement description. (a) Optically gated transistor structure, not drawn to scale. The active material consists of alternating layers of Ge2Se3 and cosputtered M + Ge2Se3 where M = Cu or Sn. Each Ge2Se3 layer target thickness is 100 Å, and the M + Ge2Se3 layer target thickness is 30 Å; however, the layers mix during deposition, and the material layers remain amorphous. (b) Top-down optical view of the electrodes and separation of the OGT.
Figure 2
Figure 2. Electrical testing measurement configuration.
Figure 3
Figure 3. Raman spectra of the OGT films: (a) Cu + Ge2Se3, (b) Sn + Ge2Se3, and (c) Ge2Se3.
Figure 4
Figure 4. Tauc plot of each OGT material. The optical gap is estimated to be 1.87 eV for Ge2Se3, 1.61 eV for Ge2Se3 + Sn, and 1.85 eV for Ge2Se3 + Cu.
Figure 5
Figure 5. OGT I–V curves under 470 nm LED variable light intensity illumination. (a) Cu + Ge2Se3 OGT device. (b) Sn + Ge2Se3 OGT device. (c) Undoped Ge2Se3OGT device. An example of I–V curves for two different memristors is given in (d). A compliance current of 100 μA (the flat line in the I–V curve at 100 μA) was applied only during memristor measurements to protect the device from damage during the dc voltage sweep. Saturation current values in (a–c) correspond to light illumination ranging from a detector reading of 15.23 mW/cm2 at the highest saturation current reading to 28 μW/cm2 for the lowest current reading. The “dark” trace appears near 0 A (pink trace in (a) to (c)).
Figure 6
Figure 6. Photocurrent generated by 470 nm illumination during. (a) Photocurrent, Iph = Imeas – I“dark” measured at 2 V, as a function of light intensity. Black traces denote the Cu + OGT alone (solid line, open circle symbols) and in series with a memristor (dashed, + symbols); red traces denote the Sn + OGT alone (solid, open square symbols) and in series with a memristor (dashed, × symbols). (b) Iph as a function of V at a light power density of 14.17 mW/cm2 for the OGT devices alone. The measurement is a double sweep (forward and reverse voltage sweep) where the green arrows denote the reverse voltage return path. Regions denoted by 1 correspond to thermally generated carrier dominated current. The region marked by 2 denotes the onset of trap filled conduction, and that marked by 3 denotes the trap-filled limit voltage, where all traps are filled and conduction is space charge limited. Red trace (solid line) corresponds to Cu + OGT; black trace (dashed line) corresponds to Sn + OGT.
Figure 7
Figure 7. Photoswitching response for OGT devices to a 770 nm LED pulse at 7.4 mW/cm2. (a) Response to pulse train. (b) Expanded view of the ON to OFF transition. Sn + OGT, black open circles; Cu + OGT, red open squares. The OGTs are in series with a 1 kohm resistor and biased at 6 V on the drain electrode, with the source electrode connected to the resistor, and the other resistor terminal connected to ground.
Figure 8
Figure 8. I–V curves for OGT–memristor circuits. (a) Cu + OGT. (b) Sn + OGT. An I–V curve for the memristor used in each OGT circuit is placed on each graph for comparison (∗). (c) Programmed resistance was determined by measuring the resistance of a line fit to the erase curve. (d) OGT–memristor circuit resistance as a function of LED light intensity. Sn + OGT, black open circles; Cu + OGT, red open squares. Illumination source was a 470 nm LED.
Figure 9
Figure 9. Sn + OGT–memristor circuit response to programming pulses. (a) OGT–memristor circuit cycling. (b) Consecutive application of write pulses. (c) Consecutive application of erase pulses. The OGT pulse (black trace, right vertical axis) corresponds to the pulse that biases the OGT into a saturated state. The memristor voltage corresponds to the voltage measured across the memristor. The OGT is driven by a 770 nm LED.
References
This article references 40 other publications.
- 1Chua, L. O. Memristor-The Missing Circuit Element. IEEE Trans. Circuit Theory 1971, 18, 507, DOI: 10.1109/TCT.1971.1083337There is no corresponding record for this reference.
- 2Chua, L. Everything You Wish to Know About Memristors but are Afraid to Ask. Radioengineering 2015, 24, 319– 368, DOI: 10.13164/re.2015.0319There is no corresponding record for this reference.
- 3Teimoori, A.; Amirsoleimani, A.; Ahmadi, A.; Ahmadi, M. A 2M1M Crossbar Architecture: Memory. IEEE Trans. VLSI Syst. 2018, 26, 2608– 2618, DOI: 10.1109/TVLSI.2018.2799951There is no corresponding record for this reference.
- 4Lashkare, S.; Panwar, N.; Kumbhare, P.; Das, B.; Ganguly, U. PCMO-Based RRAM and NPN Bipolar Selector as Synapse for Energy Efficient STDP. IEEE Electron Device Lett. 2017, 38, 1212– 1215, DOI: 10.1109/LED.2017.27235034PCMO-based RRAM and NPN bipolar selector as synapse for energy efficient STDPLashkare, S.; Panwar, N.; Kumbhare, P.; Das, B.; Ganguly, U.IEEE Electron Device Letters (2017), 38 (9), 1212-1215CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)Resistance random access memories (RRAMs) are widely explored to show spike time dependent plasticity (STDP) as a learning rule to show biol. synaptic behavior, as these devices possess analog conductance change. To implement STDP, pre- and post-neuronal waveforms are superposed. Only the peak voltage changes the conductance of memory. But due to the remaining part of the waveform (which don't affect the conductance change), there is a significant amt. of inadvertent current flow leading to unnecessary energy consumption. In this letter, we exptl. demonstrate that, the PCMO-based RRAM, a memristor (1M) in series with the NPN selector can be used as a synapse to reduce the undesirable energy consumption. First, we propose the pre- and post-neuronal waveform engineering required for PCMO-based memristor + Selector (1S1M) to reduce energy consumption. Second, we demonstrate exptl. that 1S1M synapse gives >4700× redn. in energy consumption compared with 1M synapse for a single neuronal waveform. Third, we implement and compare anti-STDP for 1M and 1S1M synapse to show no loss of generality. Thus, we exptl. demonstrate 1S1M-based energy efficient synapse for brain inspired computing.
- 5Nugent, M. A.; Molter, T. W. AHaH Computing–From Metastable Switches to Attractors to Machine Learning. PLoS One 2014, 9, e85175 DOI: 10.1371/journal.pone.00851755AHaH computing-from metastable switches to attractors to machine learningNugent, Michael Alexander; Molter, Timothy WesleyPLoS One (2014), 9 (2), e85175/1-e85175/29, 29 pp.CODEN: POLNCL; ISSN:1932-6203. (Public Library of Science)Modern computing architecture based on the sepn. of memory and processing leads to a well known problem called the von Neumann bottleneck, a restrictive limit on the data bandwidth between CPU and RAM. This paper introduces a new approach to computing we call AHaH computing where memory and processing are combined. The idea is based on the attractor dynamics of volatile dissipative electronics inspired by biol. systems, presenting an attractive alternative architecture that is able to adapt, self-repair, and learn from interactions with the environment. We envision that both von Neumann and AHaH computing architectures will operate together on the same machine, but that the AHaH computing processor may reduce the power consumption and processing time for certain adaptive learning tasks by orders of magnitude. The paper begins by drawing a connection between the properties of volatility, thermodn., and Anti-Hebbian and Hebbian (AHaH) plasticity. We show how AHaH synaptic plasticity leads to attractor states that ext. the independent components of applied data streams and how they form a computationally complete set of logic functions. After introducing a general memristive device model based on collections of metastable switches, we show how adaptive synaptic wts. can be formed from differential pairs of incremental memristors. We also disclose how arrays of synaptic wts. can be used to build a neural node circuit operating AHaH plasticity. By configuring the attractor states of the AHaH node in different ways, high level machine learning functions are demonstrated. This includes unsupervised clustering, supervised and unsupervised classification, complex signal prediction, unsupervised robotic actuation and combinatorial optimization of procedures-all key capabilities of biol. nervous systems and modern machine learning algorithms with real world application.
- 6Burr, G. W.; Shelby, R. M.; Sebastian, A.; Kim, S.; Kim, S.; Sidler, S.; Virwani, K.; Ishii, M.; Narayanan, P.; Fumarola, A.; Sanches, L. L.; Boybat, I.; Le Gallo, M.; Moon, K.; Woo, J.; Hwang, H.; Leblebici, Y. Neuromorphic Computing Using Non-Volatile Memory. Advances in Physics: X 2017, 2, 89– 124, DOI: 10.1080/23746149.2016.1259585There is no corresponding record for this reference.
- 7Chen, Z.; Schoeny, C.; Dolecek, L. Hamming Distance Computation in Unreliable Resistive Memory. IEEE. Trans. On Comm. 2018, 66, 5013, DOI: 10.1109/TCOMM.2018.2840717There is no corresponding record for this reference.
- 8Song, B.; Xu, H.; Liu, S.; Liu, H.; Li, Q. Threshold Switching Behavior of Ag-SiTe-Based Selector Device and Annealing Effect on its Characteristics. IEEE J. Electron Devices Soc. 2018, 6, 674– 679, DOI: 10.1109/JEDS.2018.28364008Threshold switching behavior of ag-site-based selector device and annealing effect on its characteristicsSong, Bing; Xu, Hui; Liu, Sen; Liu, Haijun; Li, QingjiangIEEE Journal of the Electron Devices Society (2018), 6 (1), 674-679CODEN: IJEDAC; ISSN:2168-6734. (Institute of Electrical and Electronics Engineers)Programmable metalization cell is one of important threshold switching selectors. We first performed a study on the selector based on amorphous chalcogenide material (Si0.4Te0.6) because of the rigid structure applied in ovonic threshold switch. In the meantime, annealing process is implemented to improve the performance. Results show that devices without annealing process demonstrate a minor threshold switching characteristic, revealing the potential as selector for cross-point memristor array. After implementing annealing process, threshold voltage (Vth), selectivity and endurance of selectors improve. Meanwhile, requirements of high current and a low holding voltage (Vh) for an ideal selector are fulfilled. Using the Ag filament formed during motion of Ag ions, a steep-slope (1.7 mV/dec) for threshold switching with high selectivity (∼ 104) could be achieved. Owing to the faster diffusivity of Ag atoms in solid-electrolytes, the resulting Ag filament easily dissolved under low current regime. It is deduced that performance improvement is due to the defect redn. within annealing process. Finally, time characteristics of selector devices are tested to verify fast switching and recovery speed for practical applicability.
- 9Koo, Y.; Lee, S.; Park, S.; Yang, M.; Hwang, H. Simple Binary Ovonic Threshold Switching Material SiTe and its Excellent Selector Performance for High-Density Memory Array Application. IEEE Electron Device Lett. 2017, 38, 568, DOI: 10.1109/LED.2017.26854359Simple binary ovonic threshold switching material SiTe and its excellent selector performance for high-density memory array applicationKoo, Yunmo; Lee, Sangmin; Park, Seonggeon; Yang, Minkyu; Hwang, HyunsangIEEE Electron Device Letters (2017), 38 (5), 568-571CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)In this letter, simple binary Ovonic threshold switching (OTS) material with outstanding selector device performance has been demonstrated. Even with its simple material compn. and easy fabrication process, the selector device with the binary OTS material showed excellent selector performance such as high-OFF resistance ( > 1 G Ω at 0.1 V), low-ON resistance (< 1 k Ω at 2.0 V), extremely sharp switching slope (< 1 mV/dec), fast operating speed (ttransition < 2 ns, tdelay < 7 ns), high endurance (>10♂ cyclesof 150 ns pulse),high elec. stability (>1ks at 1.2 V), and high thermal stability (> 400 °C / 30 min). Furthermore, conduction mechanism of the OTS has been explained by Poole-Frenkel-based anal. modeling.
- 10Burr, G. W.; Shenoy, R. S.; Hwang, H. Select Device Concepts for Crossbar Arrays. In Resistive Switching: From Fundamentals of Nanoionic Redox Processes to Memristive Device Applications, 1st ed.; Ielmini, D., Waser, R., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2016.There is no corresponding record for this reference.
- 11Liu, M.; Wang, W. Application of Nanojunction-Based RRAM to Reconfigurable IC. Micro Nano Lett. 2008, 3, 101– 105, DOI: 10.1049/mnl:20080029There is no corresponding record for this reference.
- 12Potteiger, T.; Robinson, W. H. A One Zener Diode, One Memristor Crossbar Architecture for a Write-Time-Based PUF. IEEE 58th International Midwest Symposium On Circuits And Systems , 2015; pp 1– 4.There is no corresponding record for this reference.
- 13Zidan, M. A.; Chen, A.; Indiveri, G.; Lu, W. D. Memristive Computing Devices and Applications. J. Electroceram. 2017, 39, 4– 20, DOI: 10.1007/s10832-017-0103-013Memristive computing devices and applicationsZidan, Mohammed A.; Chen, An; Indiveri, Giacomo; Lu, Wei D.Journal of Electroceramics (2017), 39 (1-4), 4-20CODEN: JOELFJ; ISSN:1385-3449. (Springer)A review. Advances in electronics have revolutionized the way people work, play and communicate with each other. Historically, these advances were mainly driven by CMOS transistor scaling following Moore's law, where new generations of devices are smaller, faster, and cheaper, leading to more powerful circuits and systems. However, conventional scaling is now facing major tech. challenges and fundamental limits. New materials, devices, and architectures are being aggressively pursued to meet present and future computing needs, where tight integration of memory and logic, and parallel processing are highly desired. To this end, one class of emerging devices, termed memristors or memristive devices, have attracted broad interest as a promising candidate for future memory and computing applications. Besides tremendous appeal in data storage applications, memristors offer the potential to enable efficient hardware realization of neuromorphic and analog computing architectures that differ radically from conventional von Neumann computing architectures. In this review, we analyze representative memristor devices and their applications including mixed signal analog-digital neuromorphic computing architectures, and highlight the potential and challenges of applying such devices and architectures in different computing applications.
- 14Woo, J.; Peng, X.; Yu, S. Design Considerations of Selector Device in Cross-Point RRAM Array for Neuromorphic Computing. 2018 IEEE International Symposium on Circuits and Systems (ISCAS), Florence , 2018; pp 1– 4.There is no corresponding record for this reference.
- 15Cyrille, M. C.; Verdy, A.; Navarro, G.; Bourgeois, G.; Garrione, J.; Bernard, M.; Sabbione, C.; Noé, P.; Nowak, E. OTS Selector Devices: Material Engineering for Switching Performance. 2018 International Conference on IC Design & Technology (ICICDT), Otranto , 2018; pp 113– 116.There is no corresponding record for this reference.
- 16Avasarala, N. S.; Govoreanu, B.; Opsomer, K.; Devulder, W.; Clima, S.; Detavernier, C.; van der Veen, M.; Van Houdt, J.; Henys, M.; Goux, L.; Kar, G. S. Doped GeSe Materials for Selector Applications. 2017 47th European Solid-State Device Research Conference (ESSDERC), Leuven , 2017; pp 168– 171.There is no corresponding record for this reference.
- 17Luo, Q.; Xu, X.; Lv, H.; Gong, T.; Long, S.; Liu, Q.; Li, L.; Liu, M. Endurance Characterization of the Cu-Dope HfO2 Based Selection Device With One Transistor-One Selector Structure. 2017 IEEE Electron Devices Technology and Manufacturing Conference (EDTM), Toyama , 2017; pp 178– 179.There is no corresponding record for this reference.
- 18Park, J.; Yoo, J.; Song, J.; Sung, C.; Hwang, H. Hybrid Selector With Excellent Selectrivity and Fast Switching Speed for X-Point Memory Array. IEEE Electron Device Lett. 2018, 39, 1171– 1174, DOI: 10.1109/LED.2018.284587818Hybrid selector with excellent selectivity and fast switching speed for X-point memory arrayPark, Jaehyuk; Yoo, Jongmyung; Song, Jeonghwan; Sung, Changhyuck; Hwang, HyunsangIEEE Electron Device Letters (2018), 39 (8), 1171-1174CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)In this letter, we demonstrate a new concept of threshold selector by combining Ag-based selector with NbOx insulator metal transition (IMT) selector. To overcome intrinsic limits of slow turn-off speed (>1μs) of the Ag-based selector, we adopted NbO x IMT selector with fast turn-off speed. The hybrid device exhibits excellent selector characteristics, such as extremely low off-current (<1 pA) and fast turn-off speed (<1 μs). It shows excellent readout margin (>75% at >210 word lines) in a fast sequential operation of cross point memory array.
- 19Song, B.; Xu, H.; Liu, S.; Liu, H.; Li, Q. Threshold Switching Behavior of Ag-SiTe-Based Selector Device and Annealing Effect on Its Characteristics. IEEE J. Electron Devices Soc. 2018, 6, 674– 679, DOI: 10.1109/JEDS.2018.283640019Threshold switching behavior of ag-site-based selector device and annealing effect on its characteristicsSong, Bing; Xu, Hui; Liu, Sen; Liu, Haijun; Li, QingjiangIEEE Journal of the Electron Devices Society (2018), 6 (1), 674-679CODEN: IJEDAC; ISSN:2168-6734. (Institute of Electrical and Electronics Engineers)Programmable metalization cell is one of important threshold switching selectors. We first performed a study on the selector based on amorphous chalcogenide material (Si0.4Te0.6) because of the rigid structure applied in ovonic threshold switch. In the meantime, annealing process is implemented to improve the performance. Results show that devices without annealing process demonstrate a minor threshold switching characteristic, revealing the potential as selector for cross-point memristor array. After implementing annealing process, threshold voltage (Vth), selectivity and endurance of selectors improve. Meanwhile, requirements of high current and a low holding voltage (Vh) for an ideal selector are fulfilled. Using the Ag filament formed during motion of Ag ions, a steep-slope (1.7 mV/dec) for threshold switching with high selectivity (∼ 104) could be achieved. Owing to the faster diffusivity of Ag atoms in solid-electrolytes, the resulting Ag filament easily dissolved under low current regime. It is deduced that performance improvement is due to the defect redn. within annealing process. Finally, time characteristics of selector devices are tested to verify fast switching and recovery speed for practical applicability.
- 20Song, J.; Woo, J.; Lim, S.; Chekol, S. A.; Hwang, H. Self-Limited CBRAM With Threshold Selector for 1S1R Crossbar Array Applications. IEEE Electron Device Lett. 2017, 38, 1532– 1535, DOI: 10.1109/LED.2017.275749320Self-limited CBRAM with threshold selector for 1S1R crossbar array applicationsSong, Jeonghwan; Woo, Jiyong; Lim, Seokjae; Chekol, Solomon Amsalu; Hwang, HyunsangIEEE Electron Device Letters (2017), 38 (11), 1532-1535CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)In this letter, we demonstrate a self-limited conductive-bridging random accessmemory (CBRAM) that removes the necessity for external current compliance in a one selector-one resistor (1S1R) architecture. The std. Ge2Sb2Te5 (GST) is used as a CBRAM switching layer. In addn., Te-rich GST is also considered. Their performance is then compared. Both samples exhibit selflimited on-current characteristics, and the on-currents of the std. GST and Te-rich GST are ∼300 and ∼20μA, resp. The obsd. self-limited characteristics are caused by the Te in the GST layer because in the presence of Te, Cu tends to form a more stable CuTe phase that restrict Cu filament growth. Furthermore, to confirm the feasibility of crossbar array applications, the 1S1R device is evaluated using a Ag/TiO2-based threshold selector device reported in our previous work. Hence, we confirm leakage current redn., a uniform resistance distribution, and stable retention characteristics in the 1S1R configuration with no external current compliance.
- 21Zhai, Y.; Yang, J.-Q.; Zhou, Y.; Mao, J.-Y.; Ren, Y.; Roy, V. A. L.; Han, S.-T. Toward Non-Volatile Photonic Memory: Concept, Material and Design. Mater. Horiz. 2018, 5, 641– 654, DOI: 10.1039/C8MH00110C21Toward non-volatile photonic memory: concept, material and designZhai, Yongbiao; Yang, Jia-Qin; Zhou, Ye; Mao, Jing-Yu; Ren, Yi; Roy, Vellaisamy A. L.; Han, Su-TingMaterials Horizons (2018), 5 (4), 641-654CODEN: MHAOBM; ISSN:2051-6355. (Royal Society of Chemistry)Digital technol. is one of the greatest modern breakthroughs, allowing sounds, words and images to be stored in binary form. However, there is a huge gap between the amt. of data created daily and the capacities of existing storage media. Developing multibit memory in which 2n levels, typically represented by distinguishable current levels, can be achieved in a single cell is a crit. specification for achieving high-d. memory devices. Compared with elec. operated memory, photonic memory-in which elec. read-out is orthogonal to the photo-programming operation-promises high differentiation among different data levels. From another aspect, benefiting from its high d., multifunctionality, low power consumption, and multilevel data storage, photonic memory devices hold future promise for built-in, non-volatile memory and reconstructed logic operation and are expected to bridge this capacity gap. Thus, we present a review on the development of photonic memory, with a view towards inspiring more intriguing ideas on the elegant selection of materials and design of novel device structures that may finally induce major progress in the manuf. and application of photonic memory.
- 22El Gharras, Z.; Bourahla, A.; Vautier, C. Role of Photoinduced Defects in Amorphous GexSe1-xPhotoconductivity. J. Non-Cryst. Solids 1993, 155, 171– 179, DOI: 10.1016/0022-3093(93)91322-T22Role of photoinduced defects in amorphous germanium-selenium (GexSe1-x) photoconductivityEl Gharras, Z.; Bourahla, A.; Vautier, C.Journal of Non-Crystalline Solids (1993), 155 (2), 171-9CODEN: JNCSBJ; ISSN:0022-3093.The spectral dependence of photocond. of amorphous pure Se and Ge-Se alloy (4 and 8 at.% Ge) evapd. thin films was measured for the spectral range between 250 and 800 nm. A comparison of the Se and Ge-Se spectra shows that the magnitude of photocurrent decreases as the Ge content increases. This result indicates a correlation between the magnitude of the photocurrent and the d. of photoinduced defects. The contribution of these defects to the enhancement of photocond. in a specific region of the spectrum as well as the influence of temp. on their d. are also discussed.
- 23Tanaka, K.; Shimakawa, K. Amorphous Chalcogenide Semiconductors and Related Materials; Springer: New York, 2011.There is no corresponding record for this reference.
- 24Mukherjee, B.; Tok, E. S.; Sow, C. H. Photocurrent Characteristics of Individual GeSe2 Nanobelt With Schottky Effects. J. Appl. Phys. 2013, 114, 134302, DOI: 10.1063/1.482377924Photocurrent characteristics of individual GeSe2 nanobelt with Schottky effectsMukherjee, Bablu; Tok, Eng Soon; Sow, Chorng HaurJournal of Applied Physics (Melville, NY, United States) (2013), 114 (13), 134302/1-134302/10CODEN: JAPIAU; ISSN:0021-8979. (American Institute of Physics)Single crystal GeSe2 nanobelts (NBs) were successfully grown using chem. vapor deposition techniques. The morphol. and structure of the nanostructures were characterized using SEM, transmission electron microscopy, X-ray diffractometry, and Raman spectroscopy. Electronic transport properties, photoconductive characteristics, and temp.-dependent electronic characteristics were examd. on devices made of individual GeSe2 nanobelt. The current increased by 3 orders of magnitude upon laser irradn. (wavelength 532 nm and intensity ∼6.8 mW/cm2) with responsivity of ∼2764 A/W at fixed 4 V bias. Localized photocond. study shows that the large photoresponse of the device primarily occurs at the metal-NB contact regions. In addn., the elec. Schottky nature of nanobelt/Au contact and p-type cond. nature of GeSe2 nanobelt are extd. from the current-voltage characteristics and spatially resolved photocurrent measurements. The high sensitivity and quick photoresponse in the visible wavelength range indicate potential applications of individual GeSe2 nanobelt devices in realizing optoelectronic switches. (c) 2013 American Institute of Physics.
- 25Liu, W. C.; Hoffman, G.; Zhou, W.; Reano, R. M.; Boolchand, P.; Sooryakumar, R. Slab Waveguides and Nanoscale Patterning of Pulsed Laser-Deposited Ge0.2Se0.8 Chalcogenide Films. Appl. Phys. Lett. 2008, 93, 041107, DOI: 10.1063/1.296512425Slab waveguides and nanoscale patterning of pulsed laser-deposited Ge0.2Se0.8 chalcogenide filmsLiu, W. C.; Hoffman, G.; Zhou, W.; Reano, R. M.; Boolchand, P.; Sooryakumar, R.Applied Physics Letters (2008), 93 (4), 041107/1-041107/3CODEN: APPLAB; ISSN:0003-6951. (American Institute of Physics)Planar slab waveguides were fabricated by pulsed laser deposition from GexSe1-x glass compds. (x ∼ 0.2) that lies very close to the floppy to rigid stiffness transition. These high quality active structures, which were deposited on SiO2 cladding layers above Si substrates, support several transverse-elec. (TE) modes, and a loss of 0.24 dB/cm for the TE0 mode was measured at 632.8 nm wavelength. The ability to exploit electron beam writing at these special Ge in Se compns. to create nanoscale surface motifs are promising advances to create unique miniature optical processing devices. (c) 2008 American Institute of Physics.
- 26Edwards, T. G.; Sen, S. Structure and Relaxation in Germanium Selenide Glasses and Supercooled Liquids: A Raman Spectroscopic Study. J. Phys. Chem. B 2011, 115, 4307– 4314, DOI: 10.1021/jp202174x26Structure and Relaxation in Germanium Selenide Glasses and Supercooled Liquids: A Raman Spectroscopic StudyEdwards, T. G.; Sen, S.Journal of Physical Chemistry B (2011), 115 (15), 4307-4314CODEN: JPCBFK; ISSN:1520-5207. (American Chemical Society)The structure of the tetrahedral backbone and the nature and time scale of the temp.-dependent structural changes in binary Ge-Se glasses and supercooled liqs. with ≤ 33.33 atom % Ge have been investigated using ambient and high-temp. Raman spectroscopy. The compn. dependence of the relative fractions of edge- and corner-shared GeSe4 tetrahedra and that of the characteristic mean vibrational frequencies of these structural units are shown to be consistent with a structural model for these glasses based on a random interconnection between GeSe4 tetrahedra and Se-Se chain fragments. The most prominent temp.-dependent structural change in the Ge20Se80 glass and supercooled liq. involves progressive conversion of the edge-shared GeSe4 tetrahedra into corner-shared tetrahedra, upon lowering of temp. The time scale of this tetrahedral conversion "reaction" corresponds well with those of enthalpy and shear relaxation near glass transition. Moreover, the temp. dependence of this GeSe4 tetrahedral speciation is shown to be the most important source for the prodn. of configurational entropy in this supercooled liq. near the glass-transition range, signifying a direct link between structural relaxation, configurational entropy, and viscous flow.
- 27Mott, N. F.; Davis, E. A. Electronic Processes in Non-crystalline Materials; Oxford University Press: London, 1971; p 238.There is no corresponding record for this reference.
- 28El-Shair, H. T.; Fouad, S. S. Optical Properties of Thin GexSe1-x Amorphous Films. Vacuum 1991, 42, 463– 467, DOI: 10.1016/0042-207X(91)90017-D28Optical properties of thin germanium-selenium (GexSe1-x) amorphous filmsEl-Shair, H. T.; Fouad, S. S.Vacuum (1991), 42 (7), 463-7CODEN: VACUAV; ISSN:0042-207X.The optical consts. of GexSe1-x amorphous thin films of different thicknesses (23-335 nm) and of different compns. (0.05 ≤ x ≤ 0.30) were detd. in the wavelength range 400-2000 nm. Both the refractive index n and the absorption index (k) are independent of the film thickness for a given soln., while both on the Ge content. The spectra indicate the existence of indirect optical transitions. The corresponding forbidden band gaps were detd. and the energy gap of GexSe7-x amorphous thin films obeys the relation EGeSe = YEGe + (1-Y)ESe, where Y is the vol. fraction of Ge.
- 29Feltz, A. Amorphous Inorganic Materials and Glasses; VCH Publishers: New York, 1993; p 238.There is no corresponding record for this reference.
- 30Xia, C.; Du, J.; Xiong, W.; Jia, Y.; Wei, Z.; Li, J. A type-II GeSe/SnS Heterobilayer With a Suitable Direct Gap, Superior Optical Absorption and Broad Spectrum for Photovoltaic Applications. J. Mater. Chem. A 2017, 5, 13400, DOI: 10.1039/C7TA02109G30A type-II GeSe/SnS heterobilayer with a suitable direct gap, superior optical absorption and broad spectrum for photovoltaic applicationsXia, Congxin; Du, Juan; Xiong, Wenqi; Jia, Yu; Wei, Zhongming; Li, JingboJournal of Materials Chemistry A: Materials for Energy and Sustainability (2017), 5 (26), 13400-13410CODEN: JMCAET; ISSN:2050-7496. (Royal Society of Chemistry)Van der Waals (vdW) heterobilayers are emerging as unique structures for next-generation electronic and optoelectronic devices. The authors predict that the GeSe/SnS heterobilayer has a direct band structure with a gap value of ∼1.519 eV and typical type-II band alignment. Moreover, it possesses the characteristics of superior optical absorption (∼105) and a broad absorption spectrum from the visible light to the near UV region. In addn., the GeSe/SnS heterobilayer also exhibits obviously anisotropic electronic transport and optical properties with larger current and stronger optical absorption along the zigzag direction. Meanwhile, interlayer coupling and applying an external elec. field are identified to be effective methods to modify its electronic and optical properties. Thus, these predicted results indicate that the GeSe/SnS heterobilayer will have promising applications in photovoltaic devices.
- 31Mark, P.; Helfrich, W. Space-Charge-Limited Currents in Organic Crystals. J. Appl. Phys. 1962, 33, 205– 215, DOI: 10.1063/1.172848731Space-charge-limited currents in organic crystalsMark, Peter; Helfrich, WolfgangJournal of Applied Physics (1962), 33 (), 205-15CODEN: JAPIAU; ISSN:0021-8979.Elec.-cond. measurements were performed with thin (50-μ) single crystals of p-terphenyl, p-quaterphenyl, and anthracene supplied with aq. electrodes, one of which was an iodineiodide soln. (acceptor electrode), and the other an iodide soln. The results strongly indicate that the acceptor electrode can form an ohmic contact for hole injection into these crystals and that space-charge-limited currents can be drawn through them. The crystals contained holetrapping states, the location-in-energy of which can be approximated by a decreasing exponential distribution above the valence band. The measurements showed that the hole mobility in p-terphenyl is ∼3 × 10-2 cm.2/v. sec., is independent of the field at least up to about 4 ×104 v./ cm., and that the hole-trap concn. is at least 1013/cc. The acceptor electrode used does not form an ohmic contact to crystals of naphthalene and biphenyl; an explanation for this is proposed. Some theoretical aspects of ohmic contact formation to org. crystals and space-charge-limited current flow in insulators are also discussed.
- 32Chiu, F.-C. A Review on Conduction Mechanisms in Dielectric Films. Adv. Mater. Sci. Eng. 2014, 2014, 578168, DOI: 10.1155/2014/578168There is no corresponding record for this reference.
- 33Korolev, D. S.; Belov, A. I.; Okulich, E. V.; Okulich, V. I.; Guseinov, D. V.; Sidorenko, K. V.; Shuisky, R. A.; Antonov, I. N.; Gryaznov, E. G.; Gorshkov, O. N.; Tetelbaum, D. I.; Mikhaylov, A. N. Manipulation of Resistive State of Silicon Oxide Memristor by Means of Current Limitation During Electroforming. Superlattices Microstruct. 2018, 122, 371, DOI: 10.1016/j.spmi.2018.07.00633Manipulation of resistive state of silicon oxide memristor by means of current limitation during electroformingKorolev, D. S.; Belov, A. I.; Okulich, E. V.; Okulich, V. I.; Guseinov, D. V.; Sidorenko, K. V.; Shuisky, R. A.; Antonov, I. N.; Gryaznov, E. G.; Gorshkov, O. N.; Tetelbaum, D. I.; Mikhaylov, A. N.Superlattices and Microstructures (2018), 122 (), 371-376CODEN: SUMIEK; ISSN:0749-6036. (Elsevier Ltd.)Resistive switching and adaptive behavior of resistive state in response to elec. stimulation has been studied for the silicon oxide based memristive devices subjected to electroforming in the conditions of current compliance in comparison with the analogous memristive devices after electroforming without any current limitation. The limitation of current and temp. during electroforming affects the parameters of growing conductive filament ensembles and redn.-oxidn. reactions resulting in a gradual character and a wide dynamic range of resistance change important for neuromorphic applications.
- 34Campbell, K. A. Self-Directed Channel Memristor for High Temperature Operation. Microelectron. J. 2017, 59, 10– 14, DOI: 10.1016/j.mejo.2016.11.00634Self-directed channel memristor for high temperature operationCampbell, Kristy A.Microelectronics Journal (2017), 59 (), 10-14CODEN: MICEB9; ISSN:0026-2692. (Elsevier Ltd.)Ion-conducting memristors comprised of the layered chalcogenide materials Ge2Se3/SnSe/Ag are described. The memristor, termed a self-directed channel (SDC) device, can be classified as a generic memristor and can tolerate continuous high temp. operation (at least 150 °C). Unlike other chalcogenide-based ion conducting device types, the SDC device does not require complicated fabrication steps, such as photodoping or thermal annealing, making these devices faster and more reliable to fabricate. Device pulsed response shows fast state switching in the 10-9 s range. Device cycling at both room temp. and 140 °C show write endurance of at least 1 billion.
- 35Liu, D.; Cheng, H.; Zhu, X.; Wang, G.; Wang, N. Analog Memristors Based on Thickening/Thinning of Ag Nanofilaments in Amorphous Manganite Thin Films. ACS Appl. Mater. Interfaces 2013, 5, 11258– 64, DOI: 10.1021/am403497y35Analog Memristors Based on Thickening/Thinning of Ag Nanofilaments in Amorphous Manganite Thin FilmsLiu, Dongqing; Cheng, Haifeng; Zhu, Xuan; Wang, Guang; Wang, NannanACS Applied Materials & Interfaces (2013), 5 (21), 11258-11264CODEN: AAMICK; ISSN:1944-8244. (American Chemical Society)The authors developed an analog memristor based on the thickening/thinning of Ag nanofilaments in amorphous La1-xSrxMnO3 (a-LSMO) thin films. The Ag/a-LSMO/Pt memristor exhibited excellent pinched hysteresis loops under high-excitation frequency, and the areas enclosed by the pinched hysteresis loops shrank with increasing excitation frequency, which is a characteristic typical of a memristor. The memristor also showed continuously tunable synapselike resistance and stable endurance. The a-LSMO thin films in the memristor acted as a solid electrolyte for Ag+ cations, and only the Ag/a-LSMO/Pt memristor electroformed with a larger current compliance easily exhibited high-frequency pinched hysteresis loops. From the electrochem. metalization (ECM) theory and elec. transport models of quantum wires and nanowires, the memristance is ultimately detd. by the amt. of charge supplied by the external current. The state equations of the memristor were established, and charge was the state variable. This study provides a new analog memristor based on metal nanofilaments thickening/thinning in ECM cells, which can be extended to other resistive switching materials. The new memristor may enable the development of beyond von Neumann computers.
- 36Lee, T.-W.; Nickel, J. H. Memristor Resistance Modulation for Analog Applications. IEEE Electron Device Lett. 2012, 33, 1456– 1458, DOI: 10.1109/LED.2012.220742936Memristor resistance modulation for analog applicationsLee, Tsung-Wen; Nickel, Janice H.IEEE Electron Device Letters (2012), 33 (10), 1456-1458CODEN: EDLEDZ; ISSN:0741-3106. (Institute of Electrical and Electronics Engineers)The resistance modulation (RM) of TaOx-based memristors can be precisely controlled by the SET switching compliance current. After electroforming, switching occurs in a rebuilt oxide between the electroformed conductive filament and the Pt electrode. RM is independent of initial oxide thickness. The switching mechanism is postulated as dielec. breakdown at the sidewall of the conductive channel created within the rebuilt oxide: During a SET operation to a lower resistance state, the conductive channel increases in size, conducting a larger current until limited by the external compliance; during a RESET operation to a higher resistance state, the tip of the oxygen-satd. tantalum conductive channel is oxidized, reforming the rebuilt oxide. The conduction of the rebuilt oxide follows a power law function of voltage, in parallel with the modulated channel conductance. The linear resistance can be randomly programmed with accuracy and reproducibility. Analog circuits of tunable memristive low-pass and high-pass filters demonstrate frequency tuning by RM.
- 37Kim, K. M.; Kim, G. H.; Song, S. J.; Seok, J. Y.; Lee, M. H.; Yoon, J. H.; Hwang, C. S. Electrically Configurable Electroforming and Bipolar Resistive Switching in Pt /TiO2 /Pt Structures. Nanotechnology 2010, 21, 305203, DOI: 10.1088/0957-4484/21/30/30520337Electrically configurable electroforming and bipolar resistive switching in Pt/TiO2/Pt structuresKim, Kyung Min; Kim, Gun Hwan; Song, Seul Ji; Seok, Jun Yeong; Lee, Min Hwan; Yoon, Jeong Ho; Hwang, Cheol SeongNanotechnology (2010), 21 (30), 305203/1-305203/7CODEN: NNOTER; ISSN:1361-6528. (Institute of Physics Publishing)This study examd. the effects of elec. forming methods on the bipolar resistance switching (BRS) behavior in Pt/TiO2/Pt sandwich structures. The BRS is confined to a region near the ruptured end of conducting nanofilaments, which are composed of a TinO2n-1 Magneli phase formed by electroforming. The intermediate phase with an oxygen vacancy concn. between the insulating TiO2 and the residual conducting filament that formed at the interface region was considered to be the switching layer (SL). The change in filament shape caused by a variation in the compliance current during filament formation resulted in a different filament rupture location and SL configuration. Precise control of the filament formation and rupture process resulted in SLs connected in an anti-parallel configuration. It was possible to reconfigure the SLs in the same fashion without any restraints, which allowed an unlimited memristive operation to be achieved. This paper presents a new technique in voltage sweep mode that applies a compliance current as a tool to achieve a memristor with unlimited operation.
- 38Mukherjee, B.; Cai, Y.; Tan, H. R.; Feng, Y. P.; Tok, E. S.; Sow, C. H. NIR Schottky Photodetectors Based on Individual Single-Crystalline GeSe Nanosheet. ACS Appl. Mater. Interfaces 2013, 5, 9594, DOI: 10.1021/am402550s38NIR Schottky photodetectors based on individual single-crystalline GeSe nanosheetMukherjee, Bablu; Cai, Yongqing; Tan, Hui Ru; Feng, Yuan Ping; Tok, Eng Soon; Sow, Chorng HaurACS Applied Materials & Interfaces (2013), 5 (19), 9594-9604CODEN: AAMICK; ISSN:1944-8244. (American Chemical Society)The authors have synthesized high-quality, micrometer-sized, single-crystal GeSe nanosheets using vapor transport and deposition techniques. Photoresponse is investigated based on mech. exfoliated GeSe nanosheet combined with Au contacts under a global laser irradn. scheme. The nonlinear, asym., and unsatd. characteristics of the I-V curves reveal that two uneven back-to-back Schottky contacts are formed. First-principles calcns. indicate that the occurrence of defects-induced in-gap defective states, which are responsible for the slow decay of the current in the OFF state and for the weak light intensity dependence of photocurrent. The Schottky photodetector exhibits a marked photoresponse to NIR light illumination (max. photoconductive gain ∼5.3 × 102 % at 4 V) at a wavelength of 808 nm. The significant photoresponse and good responsitivity (∼3.5 A W-1) suggests its potential applications as photodetectors.
- 39Tan, H.; Liu, G.; Yang, H.; Yi, X.; Pan, L.; Shang, J.; Long, S.; Liu, M.; Wu, Y.; Li, R.-W. Light-Gated Memristor with Integrated Logic and Memory Functions. ACS Nano 2017, 11, 11298, DOI: 10.1021/acsnano.7b0576239Light-Gated Memristor with Integrated Logic and Memory FunctionsTan, Hongwei; Liu, Gang; Yang, Huali; Yi, Xiaohui; Pan, Liang; Shang, Jie; Long, Shibing; Liu, Ming; Wu, Yihong; Li, Run-WeiACS Nano (2017), 11 (11), 11298-11305CODEN: ANCAC3; ISSN:1936-0851. (American Chemical Society)Memristive devices are able to store and process information, which offers several key advantages over the transistor-based architectures. However, most of the two-terminal memristive devices have fixed functions once made and cannot be reconfigured for other situations. Here, we propose and demonstrate a memristive device "memlogic" (memory logic) as a nonvolatile switch of logic operations integrated with memory function in a single light-gated memristor. Based on nonvolatile light-modulated memristive switching behavior, a single memlogic cell is able to achieve optical and elec. mixed basic Boolean logic of reconfigurable "AND", "OR", and "NOT" operations. Furthermore, the single memlogic cell is also capable of functioning as an optical adder and digital-to-analog converter. All the memlogic outputs are memristive for in situ data storage due to the nonvolatile resistive switching and persistent photocond. effects. Thus, as a memdevice, the memlogic has potential for not only simplifying the programmable logic circuits but also building memristive multifunctional optoelectronics.
- 40Zhu, X.; Lu, W. D. Optogenetics-Inspired Tunable Synaptic Functions in Memristors. ACS Nano 2018, 12, 1242– 1249, DOI: 10.1021/acsnano.7b0731740Optogenetics-Inspired Tunable Synaptic Functions in MemristorsZhu, Xiaojian; Lu, Wei D.ACS Nano (2018), 12 (2), 1242-1249CODEN: ANCAC3; ISSN:1936-0851. (American Chemical Society)Two-terminal memristors with internal Ca2+-like dynamics can be used to faithfully emulate biol. synaptic functions and have been intensively studied for neural network implementations. Inspired by the optogenetic technique that uses light to tune the Ca2+ dynamics and subsequently the synaptic plasticity, the authors develop a CH3NH3PbI3 (MAPbI3)-based memristor that exhibits light-tunable synaptic behaviors. Specifically, by increasing the formation energy of iodine vacancy (V·I·I/V×I×I), light illumination can be used to control the V·I·I/V×I generation and annihilation dynamics, resembling light-controlled Ca2+ influx in biol. synapses. The memory formation and memory loss behaviors in the memristors can be modified by controlling the intensity and the wavelength of the illuminated light. Coincidence detection of elec. and light stimulations is also implemented in the memristive device with real-time (≤20 ms) response to light illumination. These results open options to modify the synaptic plasticity effects in memristor-based neuromorphic systems and can lead to the development of electronic systems that can faithfully emulate diverse biol. processes.
Supporting Information
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsaelm.8b00034.
I–V curves for the undoped Ge2Se3–memristor circuit in Figures S1 and S2 (DOCX)
Terms & Conditions
Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.