Generation of Tunable Stochastic Sequences Using the Insulator–Metal TransitionClick to copy article linkArticle link copied!
- Javier del Valle*Javier del Valle*Email: [email protected]Department of Quantum Matter Physics, University of Geneva, 24 Quai Ernest-Ansermet, 1211 Geneva, SwitzerlandMore by Javier del Valle
- Pavel SalevPavel SalevDepartment of Physics and Center for Advanced Nanoscience, University of California-San Diego, La Jolla, California 92093, United StatesMore by Pavel Salev
- Stefano GariglioStefano GariglioDepartment of Quantum Matter Physics, University of Geneva, 24 Quai Ernest-Ansermet, 1211 Geneva, SwitzerlandMore by Stefano Gariglio
- Yoav KalcheimYoav KalcheimDepartment of Material Science and Engineering, Technion - Israel Institute of Technology, Haifa 32000, IsraelMore by Yoav Kalcheim
- Ivan K. SchullerIvan K. SchullerDepartment of Physics and Center for Advanced Nanoscience, University of California-San Diego, La Jolla, California 92093, United StatesMore by Ivan K. Schuller
- Jean-Marc TrisconeJean-Marc TrisconeDepartment of Quantum Matter Physics, University of Geneva, 24 Quai Ernest-Ansermet, 1211 Geneva, SwitzerlandMore by Jean-Marc Triscone
Abstract
Probabilistic computing is a paradigm in which data are not represented by stable bits, but rather by the probability of a metastable bit to be in a particular state. The development of this technology has been hindered by the availability of hardware capable of generating stochastic and tunable sequences of “1s” and “0s”. The options are currently limited to complex CMOS circuitry and, recently, magnetic tunnel junctions. Here, we demonstrate that metal–insulator transitions can also be used for this purpose. We use an electrical pump/probe protocol and take advantage of the stochastic relaxation dynamics in VO2 to induce random metallization events. A simple latch circuit converts the metallization sequence into a random stream of 1s and 0s. The resetting pulse in between probes decorrelates successive events, providing a true stochastic digital sequence.
This publication is licensed under
License Summary*
You are free to share(copy and redistribute) this article in any medium or format within the parameters below:
Creative Commons (CC): This is a Creative Commons license.
Attribution (BY): Credit must be given to the creator.
Non-Commercial (NC): Only non-commercial uses of the work are permitted.
No Derivatives (ND): Derivative works may be created for non-commercial purposes, but sharing is prohibited.
*Disclaimer
This summary highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. Carefully review the actual license before using these materials.
License Summary*
You are free to share(copy and redistribute) this article in any medium or format within the parameters below:
Creative Commons (CC): This is a Creative Commons license.
Attribution (BY): Credit must be given to the creator.
Non-Commercial (NC): Only non-commercial uses of the work are permitted.
No Derivatives (ND): Derivative works may be created for non-commercial purposes, but sharing is prohibited.
*Disclaimer
This summary highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. Carefully review the actual license before using these materials.
License Summary*
You are free to share(copy and redistribute) this article in any medium or format within the parameters below:
Creative Commons (CC): This is a Creative Commons license.
Attribution (BY): Credit must be given to the creator.
Non-Commercial (NC): Only non-commercial uses of the work are permitted.
No Derivatives (ND): Derivative works may be created for non-commercial purposes, but sharing is prohibited.
*Disclaimer
This summary highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. Carefully review the actual license before using these materials.
License Summary*
You are free to share(copy and redistribute) this article in any medium or format within the parameters below:
Creative Commons (CC): This is a Creative Commons license.
Attribution (BY): Credit must be given to the creator.
Non-Commercial (NC): Only non-commercial uses of the work are permitted.
No Derivatives (ND): Derivative works may be created for non-commercial purposes, but sharing is prohibited.
*Disclaimer
This summary highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. Carefully review the actual license before using these materials.
License Summary*
You are free to share(copy and redistribute) this article in any medium or format within the parameters below:
Creative Commons (CC): This is a Creative Commons license.
Attribution (BY): Credit must be given to the creator.
Non-Commercial (NC): Only non-commercial uses of the work are permitted.
No Derivatives (ND): Derivative works may be created for non-commercial purposes, but sharing is prohibited.
*Disclaimer
This summary highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. Carefully review the actual license before using these materials.
The increasing demand for computational power is fueling the emergence of alternative computation paradigms, beyond the standard von Neumann architecture. (1) One of them is probabilistic computing, where information is represented as the probability of a metastable system (a probabilistic bit or p-bit) to be in a particular state. (2,3) Initially proposed by von Neumann in 1956, (4) probabilistic computing gained traction through the 1960s and 1970s because of its promise of massive circuitry reduction, (5,6) but interest in it eventually faded due to the miniaturization of electronic components and the absence of practical sources of probabilistic bits. In recent years, the growing challenges of big data have motivated a resurgence in probabilistic processing approaches, such as probabilistic programming, (7) stochastic neural networks, (8,9) and Boltzmann machines. (10,11) These methods’ working principles resemble classical counterparts of quantum computing schemes, and they might offer an intermediate platform, easier to realize, between classical and quantum computing. (12)
The development of probabilistic hardware has traditionally been limited by the absence of adequate p-bit sources, that is, scalable hardware capable of producing stochastic and tunable sequences of “1s” and “0s”. CMOS-based solutions, such as linear-feedback shift registers, are large and often not truly stochastic. (13) Recently, magnetic tunnel junctions near the superparamagnetic limit have shown great promise for fast and scalable implementation of p-bits, (12,14−16) motivating the search for other systems where intrinsic fluctuations (17) could also be exploited. Natural candidates are materials featuring insulator–metal transitions (IMTs), (18) as they could, a priori, fluctuate between high and low resistance states. For instance, thermal fluctuations in NbO2 nanodevices have been shown to be useful in solving optimization problems in an analog way, (19,20) and the jitter in the self-oscillations of VO2 devices has successfully been utilized to generate nontunable random numbers. (21) But to date, no method for generating digital, tunable p-bits using the IMT has been proposed.
Here, we show that p-bits can be generated using the voltage-triggered IMT in VO2 nanodevices. We apply two trains of voltage pulses -controlled by two clocks- to our system: the “pumping” and the “probing” trains. The pumping pulses have higher amplitude and always metallize the VO2. The probing pulses have lower amplitude, sometimes triggering the transition (“1”) and sometimes not (“0”). The triggering probability depends on the probing pulse amplitude, and the stochasticity of the process is ensured due to the randomness of the relaxation path the system follows after the pumping pulse. By using a simple latch circuit, it is possible to generate a continuous string of 0s and 1s with tunable probability. Analysis of the output string indicates true stochasticity with cryptographic quality.
Our devices are fabricated using a 100 nm thick VO2 film grown by reactive sputtering on top of an R-cut sapphire substrate. Two closely spaced metallic electrodes (Pt or Ti/Au) have been patterned on top of it using electron beam lithography (see methods in the Supporting Information). Electrodes are 400 nm wide, are separated from each other by a 100–300 nm gap and form low resistance ohmic contacts with the VO2 film. Figure 1a shows resistance versus temperature in one such device. A sharp, first-order IMT is observed with a resistance drop of around 2 orders of magnitude. This transition can also be induced electrically by applying a large enough voltage. (22) The voltage-driven IMT holds much promise for application in related fields such as spiking neural networks (23−25) or optoelectronics. (26,27)Figure 1b shows current versus time when different voltage pulses are applied to a device which is initially insulating (T = 300 K in all experiments). For low pulse amplitudes, little current flows through the device as it remains in the insulating state. But once a threshold voltage (VTh) is surpassed, partial metallization takes place and the current rapidly increases. We will refer to these voltage pulses above VTh as pumping pulses, since they always trigger the IMT.
The voltage-triggered IMT is known to happen via percolation of a metallic filament between the metallic electrodes, (28,29) as schematically depicted in the top panel of Figure 1c. Once the voltage returns to zero, the device locally cools down and starts relaxing back into the insulating state, as depicted in the bottom panels of Figure 1c. The first order character of the IMT in VO2 plays a crucial role in this process: it has been recently shown that the metastability of the metallic phase leads to slow relaxation dynamics, which can be orders of magnitude slower than the cool down time. (30) This has important consequences in the transport properties. If a second pulse, this time below VTh, is sent to the device before it relaxes completely, there is a nonzero probability of triggering the IMT again. We will refer to these pulses below VTh as probing pulses, and they may or may not induce the transition. Figure 1d shows the probability that a probing pulse of amplitude VProbe induces the IMT 100 μs after it has been triggered by a pumping pulse (VPump = 3 VTh). The probability has a sigmoidal dependence on VProbe, varying smoothly from 0 to 1. Whether a specific probing pulse triggers the IMT depends on the specific arrangement of unrelaxed metallic islands within the device at the moment the pulse is applied. Since the relaxation path followed by the system after each pumping pulse is variable, (22,30) cycle-to-cycle stochasticity is expected.
The top panel in Figure 2a shows the protocol we propose for implementing p-bits using this phenomenology. Two intercalated trains of voltage pulses are applied to a VO2 nanodevice: a fixed amplitude (>VTh) pumping train and a tunable amplitude (<VTh) probing train. In our proof-of-concept experiment, we use a function generator to regulate the probing amplitude, but a gated transistor could also be employed for this task. The bottom panel in Figure 2a shows current versus time in the same device. As expected, the IMT is induced by every pumping pulse, leading to a sharp increase in current. When a probing pulse is applied, the IMT is induced in some cases (green dots), which are to be considered 1; but not in others (blue dots), which are considered 0. The first order, discontinuous character of the IMT in VO2, implies that only two well-defined device states (conducting or insulating) will be induced when the probing pulse is applied, naturally producing a digital output. Figure 2b shows three strings of 1s and 0s obtained in this way, for three different amplitudes of the probing pulse. Lower voltage pulses yield strings that are mostly 0s, while higher voltages create mostly strings of 1s. This can be better seen in Figure 2c, where the 1s to 0s ratio (the nominal value of the p-bit) is plotted as a function of VProbe. Our protocol produces tunable p-bits, whose value can be continuously varied from 0 to 1 by controlling the applied voltage.
All data shown in Figure 2 were taken keeping a period of 5 μs between probing pulses. Similar results were observed for periods of 1, 2, 10, and 20 μs, as shown in Figures S1–S10. The typical working VProbe range depends on this period with VProbe being lower for shorter periods. This is expected since for short periods there are more remaining metallic islands when the probing pulse is applied. In our experiments, we applied the probing pulse half a period after the pumping pulse for the sake of simplicity. But different time separation between pump and probe could be used with VProbe becoming lower as this time is made shorter.
In order to assess whether the sequence is truly random, we computed the autocorrelation function. It compares how similar the sequence is to itself but with the sequence elements (Xi) shifted by S positions forward in the chain. S can take any integer between 0 and N, the total sequence length. We define the autocorrelation in eq 1 where P is average of the sequence, or p-bit value
The stochasticity of our signal can be further tested. For the particular case of P = 50%, our p-bit can behave as a random number generator. We used our device to create a random sequence of 49 000 1s and 0s with P = 50.00 ± 0.01%. The sequence is long enough to be subjected to 12 of the 15 tests of the NIST Suite for Random Numbers Generators for Cryptographic Applications, (31) passing all 12 tests. The remaining three tests require sequences over 1 000 000 digits, unfeasible to record with our experimental setup. In order to obtain P as close as possible to 50%, eight independent sequences were combined using XOR logic. Details can be found in the Supporting Information.
While our protocol produces truly stochastic p-bits, the signal output is not convenient since the 1s and 0s are only readable for a short amount of time, during the probing pulse. This can be solved using a latch circuit that holds the transient voltage across the device. In our case, we add a load resistor in series with the device and use a clocked NAND latch, as depicted in Figure 3a. The NAND latch works in inverted logic: the output goes to zero voltage whenever its set input is logic 0 and the reset is logic 1. Since it is a clocked latch, the output is only changed when a voltage is simultaneously applied to the clock terminal. For operation, we apply both pumping and probing pulses to the VO2 nanodevice through the load resistor (RLoad = 500 Ω in our case). The voltage VIn between load and VO2 is applied to the set terminal of the latch, while the inverted logic value of VIn is sent to the reset terminal using a NOT gate. The probing pulses that are sent to the device are also applied to the clock terminal of the latch. If the probing pulse does not induce the IMT, VIn is high, driving set to 1, reset to 0, and the output to 0. If the IMT is triggered, the situation is reversed and the output changes to 1. Since the output can only be updated during the probing pulse, the last value is held in between pulses. This can be observed in Figure 3b for three different probing pulse amplitudes. The output voltage stays mostly around 0 V, around 5 V, or fluctuates between them depending on the probe. Figure S11 shows the probability as a function of VProbe as well as the autocorrelation functions for several cases, confirming that adding these components to the setup does not compromise the stochasticity of the p-bits.
This method for generating p-bits is scalable, because both the VO2 devices and the few gates needed to implement a latch can be produced at the submicrometer scale. This would potentially allow for patterning very large-scale arrays of p-bits as part of a chip architecture. All p-bits could rely on the same two system clocks for creating the pumping and probing pulses, and probing voltages could be individually adjusted using a transistor. Regarding speed, here we demonstrate operation up to 1 MHz, but are constrained by device size and limits of the measuring setup, making it difficult to use voltage pulses shorter than 100 ns. Smaller devices or different geometries, such as a vertical configuration, could dramatically speed up the process to values closer to the gigahertz. Switching speeds below 1 ns have been demonstrated in materials such as VO2 and V2O3. (32,33) We must also mention possible drawbacks of our approach. The most salient might be the narrow VProbe range (∼0.1 V) that allows varying P between 0 and 1. This problem, also present in magnetic tunnel junctions, might pose a challenge for properly controlling p-bit values in large integrated circuits. Temperature changes or device-to-device variability could induce small VProbe variations that would result in large P changes. For instance, we analyzed over 10 VO2 nanodevices, finding similar phenomenology among them but a relatively large threshold voltage variability ΔVTh ≈ 0.3 V, larger than the working voltage range. However, these problems can be tackled: device-to-device variations can be greatly reduced with further optimization, and temperature sensitivity can be minimized exploring materials with a nonthermal IMT, such as GaTa4Se8. (34) Another potential source of variability might be device degradation over time due to repeated switching. Similar devices to the ones presented here showed no measurable changes in the threshold voltage after 108 switching cycles, (22) whereas other VO2 works have reported no degradation in the material properties after 1010 cycles (35) or after 100 h of continuous operation. (36) However, stability over months and years would have to be monitored before any technological implementation is possible. These considerations, however, are beyond the proof-of-concept scope of our paper and should be carefully addressed in future works.
In conclusion, we developed a method for generating probabilistic bits using materials with first order metal–insulator transitions. We induce the transition electrically and use the intrinsic stochasticity of the relaxation pathway as a source of randomness when retriggering the IMT. By combining VO2 nanodevices with a latch circuit we create tunable and truly stochastic p-bits which can be potentially downscaled below the micrometer scale. Our work presents metal–insulator transitions as a new physical system where applications for probabilistic computing can be explored, contributing to the prospects created by recent progress in the field.
Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.nanolett.1c04404.
Details on sample fabrication, randomness evaluation using the NIST suite, performance for different pulse periods (Figures S1–S10) and performance with the latch circuit (Figure S11) (PDF)
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.
Acknowledgments
We thank Suhas Kumar, Mark Stiles, and Iaroslav Gaponenko for helpful discussions. We also thank Marco Lopes, Alberto Morpurgo, and Patrycja Paruch for their support during the fabrication and measurement of these samples. This work was funded by the Swiss National Science Foundation with an Ambizione Fellowship (#PZ00P2_185848). Part of the nanofabrication was paid by the U.S. Office of Naval Research through the NICOP Grant N62909-21-1-2028 and by the Swiss National Science Foundation Project No. 200020-179155. The VO2 growth and part of the nanofabrication was supported by the Quantum Materials for Energy Efficient Neuromorphic Computing (Q-MEEN-C) Energy Frontier Research Center (EFRC), funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences under Award No. DE-SC0019273. Y.K. acknowledges funding from the Norman Seiden Fellowship for Nanotechnology and Optoelectronics and the Israel Science Foundation (grant No. 1031/21).
References
This article references 36 other publications.
- 1Beyond von Neumann. Nat. Nanotechnol 2020, 15, 507– 507 DOI: 10.1038/s41565-020-0738-x .Google ScholarThere is no corresponding record for this reference.
- 2Alaghi, A.; Qian, W.; Hayes, J. P. The Promise and Challenge of Stochastic Computing. IEEE Trans. Comput. Des. Integr. Circuits Syst. 2018, 37, 1515– 1531, DOI: 10.1109/TCAD.2017.2778107Google ScholarThere is no corresponding record for this reference.
- 3Alaghi, A.; Hayes, J. P. Survey of stochastic computing. Trans. Embed. Comput. Syst. 2013, 12, 1, DOI: 10.1145/2465787.2465794Google ScholarThere is no corresponding record for this reference.
- 4von Neumann, J. Probabilistic Logics and the Synthesis of Reliable Organisms From Unreliable Components. Autom. Stud. 1956, 34, 43– 98, DOI: 10.1515/9781400882618-003Google ScholarThere is no corresponding record for this reference.
- 5Poppelbaum, W. J.; Afuso, C.; Esch, J. W. Stochastic computing elements and systems. AFIPS’67 Proc. Fall Joint Computer Conference ; Anaheim, CA, 1967; ACM Press, 1967; Vol. 1, pp 635– 644.Google ScholarThere is no corresponding record for this reference.
- 6Gaines, B. R. Stochastic Computing Systems. Advances in Information Systems Science; Springer US, 1969; pp 37– 172.Google ScholarThere is no corresponding record for this reference.
- 7Wood, F.; Van De Meent, J. W.; Mansinghka, V. A New Approach to Probabilistic Programming Inference. Proc. 17th Int. Conf. Artif. Intell. Stat. 33, Reykjavic, Iceland (22th-25th April 2014). Preprint availale at arXiv:1507.00996Google ScholarThere is no corresponding record for this reference.
- 8Ghahramani, Z. Probabilistic machine learning and artificial intelligence. Nat. 2015 5217553 2015, 521, 452– 459, DOI: 10.1038/nature14541Google ScholarThere is no corresponding record for this reference.
- 9Bottou, L. Large-Scale Machine Learning with Stochastic Gradient Descent. Proc. COMPSTAT’2010 2010, 177– 186Google ScholarThere is no corresponding record for this reference.
- 10Salakhutdinov, R.; Hinton, G. Deep Boltzmann Machines. Proc. 12th Int. Conference Artif. Intell. Stat. 2009, 5, 448Google ScholarThere is no corresponding record for this reference.
- 11Zhang, N.; Ding, S.; Zhang, J.; Xue, Y. An overview on Restricted Boltzmann Machines. Neurocomputing 2018, 275, 1186– 1199, DOI: 10.1016/j.neucom.2017.09.065Google ScholarThere is no corresponding record for this reference.
- 12Camsari, K. Y.; Sutton, B. M.; Datta, S. p-bits for probabilistic spin logic. Appl. Phys. Rev. 2019, 6, 011305, DOI: 10.1063/1.5055860Google Scholar12https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXkvVSrsb4%253D&md5=3cdc33a79537e878a3165f4acc93e0deThe p-bits for probabilistic spin logicCamsari, Kerem Y.; Sutton, Brian M.; Datta, SupriyoApplied Physics Reviews (2019), 6 (1), 011305/1-011305/12CODEN: APRPG5; ISSN:1931-9401. (American Institute of Physics)We introduce the concept of a probabilistic or p-bit, intermediate between the std. bits of digital electronics and the emerging q-bits of quantum computing. We show that low barrier magnets or LBMs provide a natural phys. representation for p-bits and can be built either from perpendicular magnets designed to be close to the in-plane transition or from circular in-plane magnets. Magnetic tunnel junctions (MTJs) built using LBMs as free layers can be combined with std. NMOS transistors to provide three-terminal building blocks for large scale probabilistic circuits that can be designed to perform useful functions. Interestingly, this three-terminal unit looks just like the 1T/MTJ device used in embedded magnetic random access memory technol., with only one difference: the use of an LBM for the MTJ free layer. We hope that the concept of p-bits and p-circuits will help open up new application spaces for this emerging technol. However, a p-bit need not involve an MTJ; any fluctuating resistor could be combined with a transistor to implement it, while completely digital implementations using conventional CMOS technol. are also possible. The p-bit also provides a conceptual bridge between two active but disjoint fields of research, namely, stochastic machine learning and quantum computing. First, there are the applications that are based on the similarity of a p-bit to the binary stochastic neuron (BSN), a well-known concept in machine learning. Three-terminal p-bits could provide an efficient hardware accelerator for the BSN. Second, there are the applications that are based on the p-bit being like a poor man's q-bit. Initial demonstrations based on full SPICE simulations show that several optimization problems, including quantum annealing are amenable to p-bit implementations which can be scaled up at room temp. using existing technol. (c) 2019 American Institute of Physics.
- 13Matsumoto, M.; Nishimura, T. ACM Trans. Model. Comput. Simul. 1998, 8, 3– 30, DOI: 10.1145/272991.272995Google ScholarThere is no corresponding record for this reference.
- 14Borders, W. A.; Pervaiz, A. Z.; Fukami, S.; Camsari, K. Y.; Ohno, H.; Datta, S. Integer factorization using stochastic magnetic tunnel junctions. Nature 2019, 573, 390, DOI: 10.1038/s41586-019-1557-9Google Scholar14https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhvVWit7%252FE&md5=665cb85a1bf3ab1148cbe6495c94ccafInteger factorization using stochastic magnetic tunnel junctionsBorders, William A.; Pervaiz, Ahmed Z.; Fukami, Shunsuke; Camsari, Kerem Y.; Ohno, Hideo; Datta, SupriyoNature (London, United Kingdom) (2019), 573 (7774), 390-393CODEN: NATUAS; ISSN:0028-0836. (Nature Research)Conventional computers operate deterministically using strings of zeros and ones called bits to represent information in binary code. Despite the evolution of conventional computers into sophisticated machines, there are many classes of problems that they cannot efficiently address, including inference, invertible logic, sampling and optimization, leading to considerable interest in alternative computing schemes. Quantum computing, which uses qubits to represent a superposition of 0 and 1, is expected to perform these tasks efficiently1-3. However, decoherence and the current requirement for cryogenic operation4, as well as the limited many-body interactions that can be implemented, pose considerable challenges. Probabilistic computing1,5-7 is another unconventional computation scheme that shares similar concepts with quantum computing but is not limited by the above challenges. The key role is played by a probabilistic bit (a p-bit)-a robust, classical entity fluctuating in time between 0 and 1, which interacts with other p-bits in the same system using principles inspired by neural networks8. Here we present a proof-of-concept expt. for probabilistic computing using spintronics technol., and demonstrate integer factorization, an illustrative example of the optimization class of problems addressed by adiabatic9 and gated2 quantum computing. Nanoscale magnetic tunnel junctions showing stochastic behavior are developed by modifying market-ready magnetoresistive random-access memory technol.10,11 and are used to implement three-terminal p-bits that operate at room temp. The p-bits are elec. connected to form a functional asynchronous network, to which a modified adiabatic quantum computing algorithm that implements three- and four-body interactions is applied. Factorization of integers up to 945 is demonstrated with this rudimentary asynchronous probabilistic computer using eight correlated p-bits, and the results show good agreement with theor. predictions, thus providing a potentially scalable hardware approach to the difficult problems of optimization and sampling.
- 15Camsari, K. Y.; Faria, R.; Sutton, B. M.; Datta, S. Stochastic p-Bits for Invertible Logic. Phys. Rev. X. 2017, 7, 031014, DOI: 10.1103/PhysRevX.7.031014Google Scholar15https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXitV2ks7jF&md5=fce208cd8818c75e525245bc02121c6bStochastic p-bits for invertible logicCamsari, Kerem Yunus; Faria, Rafatul; Sutton, Brian M.; Datta, SupriyoPhysical Review X (2017), 7 (3), 031014/1-031014/19CODEN: PRXHAE; ISSN:2160-3308. (American Physical Society)Conventional semiconductor-based logic and nanomagnet-based memory devices are built out of stable, deterministic units such as std. metal-oxide semiconductor transistors, or nanomagnets with energy barriers in excess of ≈ 40-60 kT. In this paper, we show that unstable, stochastic units, which we call "p-bits", can be interconnected to create robust correlations that implement precise Boolean functions with impressive accuracy, comparable to std. digital circuits. At the same time, they are invertible, a unique property that is absent in std. digital circuits. When operated in the direct mode, the input is clamped, and the network provides the correct output. In the inverted mode, the output is clamped, and the network fluctuates among all possible inputs that are consistent with that output. First, here they present a detailed implementation of an invertible gate to bring out the key role of a single three-terminal transistorlike building block to enable the construction of correlated p-bit networks. The results for this specific, CMOS-assisted nanomagnet-based hardware implementation agree well with those from a universal model for p-bits, showing that p-bits need not be magnet based: any three-terminal tunable random bit generator should be suitable. They present a general algorithm for designing a Boltzmann machine (BM) with a sym. connection matrix [J] (Jij = Jji) that implements a given truth table with p-bits. The [J] matrixes are relatively sparse with a few unique wts. for convenient hardware implementation. Then show how BM full adders can be interconnected in a partially directed manner (Jij not equal to Jji) to implement large logic operations such as 32-bit binary addn. Hundreds of stochastic p-bits get precisely correlated such that the correct answer out of 233 (≈ 8 x 109) possibilities can be extd. by looking at the statistical mode or majority vote of a no. of time samples. With perfect directivity (Jji = 0) a small no. of samples is enough, while for less directed connections more samples are needed, but even in the former case logical invertibility is largely preserved. This combination of digital accuracy and logical invertibility is enabled by the hybrid design that uses bidirectional BM units to construct circuits with partially directed interunit connections. In this paper, they establish this key result with extensive examples including a 4-bit multiplier which in inverted mode functions as a factorizer.
- 16Safranski, C.; Kaiser, J.; Trouilloud, P.; Hashemi, P.; Hu, G.; Sun, J. Z. Demonstration of Nanosecond Operation in Stochastic Magnetic Tunnel Junctions. Nano Lett. 2021, 21, 2040– 2045, DOI: 10.1021/acs.nanolett.0c04652Google Scholar16https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3MXkvFyisLw%253D&md5=69b0f1b971fe41b20ae8244ff43b85d0Demonstration of Nanosecond Operation in Stochastic Magnetic Tunnel JunctionsSafranski, Christopher; Kaiser, Jan; Trouilloud, Philip; Hashemi, Pouya; Hu, Guohan; Sun, Jonathan Z.Nano Letters (2021), 21 (5), 2040-2045CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)Magnetic tunnel junctions operating in the superparamagnetic regime are promising devices in the field of probabilistic computing, which is suitable for applications like high-dimensional optimization or sampling problems. Further, random no. generation is of interest in the field of cryptog. For such applications, a device's uncorrelated fluctuation time-scale can det. the effective system speed. It has been theor. proposed that a magnetic tunnel junction designed to have only easy-plane anisotropy provides fluctuation rates detd. by its easy-plane anisotropy field and can perform on a nanosecond or faster time-scale as measured by its magnetoresistance's autocorrelation in time. Here, we provide exptl. evidence of nanosecond scale fluctuations in a circular-shaped easy-plane magnetic tunnel junction, consistent with finite-temp. coupled macrospin simulation results and prior theor. expectations. We further assess the degree of stochasticity of such a signal.
- 17Chen, L.; Zhou, P.; Kalcheim, Y.; Schuller, I. K.; Natelson, D. Percolation and nanosecond fluctuators in V2O3films within the metal-insulator transition. APL Mater. 2020, 8, 101103, DOI: 10.1063/5.0023475Google Scholar17https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhvFyqurbK&md5=1fa2db014ea7bd67ecae1c70391382dfPercolation and nanosecond fluctuators in V2O3 films within the metal-insulator transitionChen, Liyang; Zhou, Panpan; Kalcheim, Yoav; Schuller, Ivan K.; Natelson, DouglasAPL Materials (2020), 8 (10), 101103CODEN: AMPADS; ISSN:2166-532X. (American Institute of Physics)Vanadium sesquioxide (V2O3) exhibits a metal-insulator transition (MIT) at 160 K between a low temp., monoclinic, antiferromagnetic Mott insulator and a high temp., rhombohedral, paramagnetic, metallic phase. In thin films, a percolative transition takes place over a finite temp. range of phase coexistence. We study the fluctuating dynamics of this percolative MIT by measuring voltage noise spectra at both low frequencies (up to 100 kHz) and radio frequencies (between 10 MHz and 1 GHz). Noise intensity quadratic in bias is obsd. in the MIT region, as expected for resistive fluctuations probed nonperturbatively by the current. The low frequency noise resembles flicker-type 1/fβ noise, often taking on the form of Lorentzian noise dominated by a small no. of fluctuators as the vol. fraction of the insulating phase dominates. Radio frequency noise intensity also quadratic in the bias current allows the identification of resistance fluctuations with lifetimes below 1 ns, approaching timescales seen in non-equil. pump-probe studies of the transition. We find quant. consistency with a model for fluctuations in the percolative fraction. The thermodn. of the MIT suggests that dominant fluctuations are ones that alter small vols. affecting the connectivity of domain boundaries. This noise serves as a sensitive and nonperturbative probe for the dynamics of switching phenomena in this system. (c) 2020 American Institute of Physics.
- 18Imada, M.; Fujimori, A.; Tokura, Y. Metal-insulator transitions. Rev. Mod. Phys. 1998, 70, 1039– 1263, DOI: 10.1103/RevModPhys.70.1039Google Scholar18https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaK1cXns1OhtLk%253D&md5=ff8bff3d0414933473d17808b96562b0Metal-insulator transitionsImada, Masatoshi; Fujimori, Atsushi; Tokura, YoshinoriReviews of Modern Physics (1998), 70 (4, Pt. 1), 1039-1263CODEN: RMPHAT; ISSN:0034-6861. (American Physical Society)A review with a large no. of refs. Metal-insulator transitions are accompanied by huge resistivity changes, even over tens of orders of magnitude, and are widely obsd. in condensed-matter systems. This article presents the observations and current understanding of the metal-insulator transition with a pedagogical introduction to the subject. Esp. important are the transitions driven by correlation effects assocd. with the electron-electron interaction. The insulating phase caused by the correlation effects is categorized as the Mott Insulator. Near the transition point the metallic state shows fluctuations and orderings in the spin, charge, and orbital degrees of freedom. The properties of these metals are frequently quite different from those of ordinary metals, as measured by transport, optical, and magnetic probes. The review 1st describes theor. approaches to the unusual metallic states and to the metal-insulator transition. The Fermi-liq. theory treats the correlations that can be adiabatically connected with the noninteracting picture. Strong-coupling models that do not require Fermi-liq. behavior have also been developed. Much work has also been done on the scaling theory of the transition. A central issue for this review is the evaluation of these approaches in simple theor. systems such as the Hubbard model and t-J models. Another key issue is strong competition among various orderings as in the interplay of spin and orbital fluctuations. Exptl., the unusual properties of the metallic state near the insulating transition have been most extensively studied in d-electron systems. In particular, there is revived interest in transition-metal oxides, motivated by the epoch-making findings of high-temp. supercond. in cuprates and colossal magnetoresistance in manganites. The article reviews the rich phenomena of anomalous metallicity, taking as examples Ti, V, Cr, Mn, Fe, Co, Ni, Cu, and Ru compds. The diverse phenomena include strong spin and orbital fluctuations, mass renormalization effects, incoherence of charge dynamics, and phase transitions under control of key parameters such as band filling, bandwidth, and dimensionality. These parameters are exptl. varied by doping, pressure, chem. compn., and magnetic fields. Much of the obsd. behavior can be described by the current theory. Open questions and future problems are also extd. from comparison between exptl. results and theor. achievements.
- 19Kumar, S.; Strachan, J. P.; Williams, R. S. Chaotic dynamics in nanoscale NbO 2 Mott memristors for analogue computing. Nature 2017, 548, 318– 321, DOI: 10.1038/nature23307Google Scholar19https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXhtlSrsLzJ&md5=59224eec4fd1185459b875bb2828a44aChaotic dynamics in nanoscale NbO2 Mott memristors for analogue computingKumar, Suhas; Strachan, John Paul; Williams, R. StanleyNature (London, United Kingdom) (2017), 548 (7667), 318-321CODEN: NATUAS; ISSN:0028-0836. (Nature Publishing Group)At present, machine learning systems use simplified neuron models that lack the rich nonlinear phenomena obsd. in biol. systems, which display spatio-temporal cooperative dynamics. There is evidence that neurons operate in a regime called the edge of chaos that may be central to complexity, learning efficiency, adaptability and analog (non-Boolean) computation in brains. Neural networks have exhibited enhanced computational complexity when operated at the edge of chaos, and networks of chaotic elements have been proposed for solving combinatorial or global optimization problems. Thus, a source of controllable chaotic behavior that can be incorporated into a neural-inspired circuit may be an essential component of future computational systems. Such chaotic elements have been simulated using elaborate transistor circuits that simulate known equations of chaos, but an exptl. realization of chaotic dynamics from a single scalable electronic device has been lacking. Here we describe niobium dioxide (NbO2) Mott memristors each less than 100 nm across that exhibit both a nonlinear-transport-driven current-controlled neg. differential resistance and a Mott-transition-driven temp.-controlled neg. differential resistance. Mott materials have a temp.-dependent metal-insulator transition that acts as an electronic switch, which introduces a history-dependent resistance into the device. We incorporate these memristors into a relaxation oscillator and observe a tunable range of periodic and chaotic self-oscillations. We show that the nonlinear current transport coupled with thermal fluctuations at the nanoscale generates chaotic oscillations. Such memristors could be useful in certain types of neural-inspired computation by introducing a pseudo-random signal that prevents global synchronization and could also assist in finding a global min. during a constrained search. We specifically demonstrate that incorporating such memristors into the hardware of a Hopfield computing network can greatly improve the efficiency and accuracy of converging to a soln. for computationally difficult problems.
- 20Kumar, S.; Williams, R. S.; Wang, Z. Third-order nanocircuit elements for neuromorphic engineering. Nature 2020, 585, 518– 523, DOI: 10.1038/s41586-020-2735-5Google Scholar20https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhvFegu7nL&md5=847286203f18d8c1bd5c85a737735508Third-order nanocircuit elements for neuromorphic engineeringKumar, Suhas; Williams, R. Stanley; Wang, ZiwenNature (London, United Kingdom) (2020), 585 (7826), 518-523CODEN: NATUAS; ISSN:0028-0836. (Nature Research)Abstr.: Current hardware approaches to biomimetic or neuromorphic artificial intelligence rely on elaborate transistor circuits to simulate biol. functions. However, these can instead be more faithfully emulated by higher-order circuit elements that naturally express neuromorphic nonlinear dynamics1-4. Generating neuromorphic action potentials in a circuit element theor. requires a min. of third-order complexity (for example, three dynamical electrophys. processes)5, but there have been few examples of second-order neuromorphic elements, and no previous demonstration of any isolated third-order element6-8. Using both expts. and modeling, here we show how multiple electrophys. processes-including Mott transition dynamics-form a nanoscale third-order circuit element. We demonstrate simple transistorless networks of third-order elements that perform Boolean operations and find analog solns. to a computationally hard graph-partitioning problem. This work paves a way towards very compact and densely functional neuromorphic computing primitives, and energy-efficient validation of neuroscientific models.
- 21Jerry, M.; Ni, K.; Parihar, A.; Raychowdhury, A.; Datta, S. Stochastic Insulator-to-Metal Phase Transition-Based True Random Number Generator. IEEE Electron Device Lett. 2018, 39, 139– 142, DOI: 10.1109/LED.2017.2771812Google Scholar21https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXitFGksbrO&md5=d7761d622597a7361acf3d99505e9f40Stochastic insulator-to-metal phase transition-based true random number generatorJerry, Matthew; Ni, Kai; Parihar, Abhinav; Raychowdhury, Arijit; Datta, SumanIEEE Electron Device Letters (2018), 39 (1), 139-142CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)An oscillator-based true random no. generator (TRNG) is exptl. demonstrated by exploiting inherently stochastic threshold switching in the insulatorto-metal transition (IMT) in vanadium dioxide. Through experimentation and modeling, we show that the origin of stochasticity arises from small perturbations in the nanoscale domain structure, which are then subsequently amplified through a pos. feedback process. Within a 1T1R oscillator, the stochastic cycle-to-cycle variations in the IMT trigger voltage result in random timing jitter, which is harnessed for a TRNG. The randomness of the IMT TRNG output is validated using the NIST SP800-22 statistical test.
- 22del Valle, J.; Vargas, N. M.; Rocco, R.; Salev, P.; Kalcheim, Y.; Lapa, P. N.; Adda, C.; Lee, M.-H.; Wang, P. Y.; Fratino, L.; Rozenberg, M. J.; Schuller, I. K. Spatiotemporal characterization of the field-induced insulator-to-metal transition. Science (80-.) 2021, 373, 907, DOI: 10.1126/science.abd9088Google ScholarThere is no corresponding record for this reference.
- 23Pickett, M. D.; Medeiros-Ribeiro, G.; Williams, R. S. A scalable neuristor built with Mott memristors. Nat. Mater. 2013, 12, 114– 117, DOI: 10.1038/nmat3510Google Scholar23https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC38XhvVersr%252FP&md5=2fcede63aebca82666a458a870dc418cA scalable neuristor built with Mott memristorsPickett, Matthew D.; Medeiros-Ribeiro, Gilberto; Williams, R. StanleyNature Materials (2013), 12 (2), 114-117CODEN: NMAACR; ISSN:1476-1122. (Nature Publishing Group)The Hodgkin-Huxley model for action potential generation in biol. axons is central for understanding the computational capability of the nervous system and emulating its functionality. Owing to the historical success of silicon complementary metal-oxide-semiconductors, spike-based computing is primarily confined to software simulations and specialized analog metal-oxide-semiconductor field-effect transistor circuits. However, there is interest in constructing phys. systems that emulate biol. functionality more directly, with the goal of improving efficiency and scale. The neuristor was proposed as an electronic device with properties similar to the Hodgkin-Huxley axon, but previous implementations were not scalable. Here we demonstrate a neuristor built using two nanoscale Mott memristors, dynamical devices that exhibit transient memory and neg. differential resistance arising from an insulating-to-conducting phase transition driven by Joule heating. This neuristor exhibits the important neural functions of all-or-nothing spiking with signal gain and diverse periodic spiking, using materials and structures that are amenable to extremely high-d. integration with or without silicon transistors.
- 24del Valle, J.; Salev, P.; Kalcheim, Y.; Schuller, I. K. A caloritronics-based Mott neuristor. Sci. Rep. 2020, 10, 1– 10, DOI: 10.1038/s41598-020-61176-yGoogle ScholarThere is no corresponding record for this reference.
- 25Bohaichuk, S. M.; Kumar, S.; Pitner, G.; McClellan, C. J.; Jeong, J.; Samant, M. G.; Wong, H. S. P.; Parkin, S. S. P.; Williams, R. S.; Pop, E. Fast Spiking of a Mott VO2-Carbon Nanotube Composite Device. Nano Lett. 2019, 19, 6751– 6755, DOI: 10.1021/acs.nanolett.9b01554Google Scholar25https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhs1aju73I&md5=6ecc4bbd2b4f2fce96a56fd2db1a7758Fast Spiking of a Mott VO2-Carbon Nanotube Composite DevicesBohaichuk, Stephanie M.; Kumar, Suhas; Pitner, Greg; McClellan, Connor J.; Jeong, Jaewoo; Samant, Mahesh G.; Wong, H-.S. Philip; Parkin, Stuart S. P.; Williams, R. Stanley; Pop, EricNano Letters (2019), 19 (10), 6751-6755CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)The recent surge of interest in brain-inspired computing and power-efficient electronics has dramatically bolstered development of computation and communication using neuron-like spiking signals. Devices that can produce rapid and energy-efficient spiking could significantly advance these applications. Here we demonstrate d.c. or voltage-driven periodic spiking with sub-20 ns pulse widths from a single device composed of a thin VO2 film with a metallic carbon nanotube as a nanoscale heater, without using an external capacitor. Compared with VO2-only devices, adding the nanotube heater dramatically decreases the transient duration and pulse energy, and increases the spiking frequency, by up to 3 orders of magnitude. This is caused by heating and cooling of the VO2 across its insulator-metal transition being localized to a nanoscale conduction channel in an otherwise bulk medium. This result provides an important component of energy-efficient neuromorphic computing systems and a lithog.-free technique for energy-scaling of electronic devices that operate via bulk mechanisms.
- 26Lee, D.; Lee, J.; Song, K.; Xue, F.; Choi, S. Y.; Ma, Y.; Podkaminer, J.; Liu, D.; Liu, S. C.; Chung, B.; Fan, W.; Cho, S. J.; Zhou, W.; Lee, J.; Chen, L. Q.; Oh, S. H.; Ma, Z.; Eom, C. B. Sharpened VO2 Phase Transition via Controlled Release of Epitaxial Strain. Nano Lett. 2017, 17, 5614– 5619, DOI: 10.1021/acs.nanolett.7b02482Google Scholar26https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXht1elu7vM&md5=eedbbe736545e23f5232881dce40bb67Sharpened VO2 phase transition via controlled release of epitaxial strainLee, Daesu; Lee, Jaeseong; Song, Kyung; Xue, Fei; Choi, Si-Young; Ma, Yanjun; Podkaminer, Jacob; Liu, Dong; Liu, Shih-Chia; Chung, Bongwook; Fan, Wenjuan; Cho, Sang June; Zhou, Weidong; Lee, Jaichan; Chen, Long-Qing; Oh, Sang Ho; Ma, Zhenqiang; Eom, Chang-BeomNano Letters (2017), 17 (9), 5614-5619CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)Phase transitions in correlated materials can be manipulated at the nanoscale to yield emergent functional properties, promising new paradigms for nanoelectronics and nanophotonics. Vanadium dioxide (VO2), an archetypal correlated material, exhibits a metal-insulator transition (MIT) above room temp. At the thicknesses required for heterostructure applications, such as an optical modulator discussed, the strain state of VO2 largely dets. the MIT dynamics crit. to the device performance. The authors develop an approach to control the MIT dynamics in epitaxial VO2 films by employing an intermediate template layer with large lattice mismatch to relieve the interfacial lattice constraints, contrary to conventional thin film epitaxy that favors lattice match between the substrate and the growing film. A combination of phase-field simulation, in situ real-time nanoscale imaging, and elec. measurements reveals robust undisturbed MIT dynamics even at preexisting structural domain boundaries and significantly sharpened MIT in the templated VO2 films. Utilizing the sharp MIT, the authors demonstrate a fast, elec. switchable optical waveguide. This study offers unconventional design principles for heteroepitaxial correlated materials, as well as novel insight into their nanoscale phase transitions.
- 27Butakov, N. A.; Knight, M. W.; Lewi, T.; Iyer, P. P.; Higgs, D.; Chorsi, H. T.; Trastoy, J.; Del Valle Granda, J.; Valmianski, I.; Urban, C.; Kalcheim, Y.; Wang, P. Y.; Hon, P. W. C.; Schuller, I. K.; Schuller, J. A. Broadband Electrically Tunable Dielectric Resonators Using Metal-Insulator Transitions. ACS Photonics 2018, 5, 4056– 4060, DOI: 10.1021/acsphotonics.8b00699Google Scholar27https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhs1KhurzE&md5=b0625158c361d56b16ae833b3ce9e270Broadband Electrically Tunable Dielectric Resonators Using Metal-Insulator TransitionsButakov, Nikita A.; Knight, Mark W.; Lewi, Tomer; Iyer, Prasad P.; Higgs, David; Chorsi, Hamid T.; Trastoy, Juan; Del Valle Granda, Javier; Valmianski, Ilya; Urban, Christian; Kalcheim, Yoav; Wang, Paul Y.; Hon, Philip W. C.; Schuller, Ivan K.; Schuller, Jon A.ACS Photonics (2018), 5 (10), 4056-4060CODEN: APCHD5; ISSN:2330-4022. (American Chemical Society)Dielec.-resonator-based nanophotonic devices show promise owing to their low intrinsic losses, support of multipolar resonances, and efficient operation in both reflection and transmission configurations. A key challenge is to make such devices dynamically switchable, such that optical behavior can be instantaneously reconfigured. Large, broadband, and continuous elec. tuning of reflection resonances in hybrid dielec.-VO2 devices is exptl. demonstrated. The calcns., in strong agreement with exptl. reflectance measurements, also indicate large transmission and absorption modulation. Addnl. independent modulation of both reflection amplitude and phase at Fabry-P´erot antinodes and nodes, resp., a key requirement for metasurface design, are demonstrated. Rapid electronic modulation rates of ∼3 kHz are achieved, substantially faster than other recent approaches. These findings greatly expand the potential of designing nanophotonic devices that exploit the tunable behavior of hybrid dielec.-VO2 resonators.
- 28Kumar, S.; Pickett, M. D.; Strachan, J. P.; Gibson, G.; Nishi, Y.; Williams, R. S. Local Temperature Redistribution and Structural Transition During Joule-Heating-Driven Conductance Switching in VO 2. Adv. Mater. 2013, 25, 6128– 6132, DOI: 10.1002/adma.201302046Google Scholar28https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXhtFelsL%252FF&md5=2bf6037c787384b1e02b9df9129decc2Local Temperature Redistribution and Structural Transition During Joule-Heating-Driven Conductance Switching in VO2Kumar, Suhas; Pickett, Matthew D.; Strachan, John Paul; Gibson, Gary; Nishi, Yoshio; Williams, R. StanleyAdvanced Materials (Weinheim, Germany) (2013), 25 (42), 6128-6132CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)The authors have utilized a complementary set of techniques to study the elec. and thermal behavior, and electronic structure of fabricated VO2 devices. Blackbody spectromicroscopy provided a comprehensive description of abrupt redistributions in local temp. during Joule-heating-driven insulator metal transition IMT. Filamentary conduction was accounted for in numerical simulations using a simplified electrothermal model. Using STXM maps we obsd. that the structure phase transition SPT occurred beyond the same current threshold that induced the IMT, regardless of whether or not they occur at precisely the same temp. The occurrence of IMT and SPT at the threshold current was justified by the jump in local temp. upon on-switching. The techniques employed in this study provide a spatially-resolved, direct and unambiguous means to study local temp. and local structure correlated with electronic transport. These results present a clear understanding of the driving mechanism in metal-insulator transition devices, which have gained recent interest due to their promising applications.
- 29Shabalin, A. G.; Valle, J.; Hua, N.; Cherukara, M. J.; Holt, M. V.; Schuller, I. K.; Shpyrko, O. G. Nanoscale Imaging and Control of Volatile and Non-Volatile Resistive Switching in VO 2. Small 2020, 16, 2005439, DOI: 10.1002/smll.202005439Google Scholar29https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXisVGlt7fJ&md5=2954b2f54c8b24a2d275441ab95577cfNanoscale Imaging and Control of Volatile and Non-Volatile Resistive Switching in VO2Shabalin, Anatoly G.; del Valle, Javier; Hua, Nelson; Cherukara, Mathew J.; Holt, Martin V.; Schuller, Ivan K.; Shpyrko, Oleg G.Small (2020), 16 (50), 2005439CODEN: SMALBC; ISSN:1613-6810. (Wiley-VCH Verlag GmbH & Co. KGaA)Control of the metal-insulator phase transition is vital for emerging neuromorphic and memristive technologies. The ability to alter the elec. driven transition between volatile and non-volatile states is particularly important for quantum-materials-based emulation of neurons and synapses. The major challenge of this implementation is to understand and control the nanoscale mechanisms behind these two fundamental switching modalities. Here, in situ X-ray nanoimaging is used to follow the evolution of the nanostructure and disorder in the archetypal Mott insulator VO2 during an elec. driven transition. Our findings demonstrate selective and reversible stabilization of either the insulating or metallic phases achieved by manipulating the defect concn. This mechanism enables us to alter the local switching response between volatile and persistent regimes and demonstrates a new possibility for nanoscale control of the resistive switching in Mott materials.
- 30del Valle, J.; Salev, P.; Tesler, F.; Vargas, N. M.; Kalcheim, Y.; Wang, P.; Trastoy, J.; Lee, M. H.; Kassabian, G.; Ramírez, J. G.; Rozenberg, M. J.; Schuller, I. K. Subthreshold firing in Mott nanodevices. Nature 2019, 569, 388– 392, DOI: 10.1038/s41586-019-1159-6Google Scholar30https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXovFeisL0%253D&md5=4536dbb8211e48cde67d820463186624Subthreshold firing in Mott nanodevicesdel Valle, Javier; Salev, Pavel; Tesler, Federico; Vargas, Nicolas M.; Kalcheim, Yoav; Wang, Paul; Trastoy, Juan; Lee, Min-Han; Kassabian, George; Ramirez, Juan Gabriel; Rozenberg, Marcelo J.; Schuller, Ivan K.Nature (London, United Kingdom) (2019), 569 (7756), 388-392CODEN: NATUAS; ISSN:0028-0836. (Nature Research)Resistive switching, a phenomenon in which the resistance of a device can be modified by applying an elec. field1-5, is at the core of emerging technologies such as neuromorphic computing and resistive memories6-9. Among the different types of resistive switching, threshold firing10-14 is one of the most promising, as it may enable the implementation of artificial spiking neurons7,13,14. Threshold firing is obsd. in Mott insulators featuring an insulator-to-metal transition15,16, which can be triggered by applying an external voltage: the material becomes conducting ('fires') if a threshold voltage is exceeded7,10-12. The dynamics of this induced transition have been thoroughly studied, and its underlying mechanism and characteristic time are well documented10,12,17,18. By contrast, there is little knowledge regarding the opposite transition: the process by which the system returns to the insulating state after the voltage is removed. Here we show that Mott nanodevices retain a memory of previous resistive switching events long after the insulating resistance has recovered. We demonstrate that, although the device returns to its insulating state within 50 to 150 ns, it is possible to re-trigger the insulator-to-metal transition by using subthreshold voltages for a much longer time (up to several milliseconds). We find that the intrinsic metastability of first-order phase transitions is the origin of this phenomenon, and so it is potentially present in all Mott systems. This effect constitutes a new type of volatile memory in Mott-based devices, with potential applications in resistive memories, solid-state frequency discriminators and neuromorphic circuits.
- 31Bassham, L. E.; Rukhin, A. L.; Soto, J.; Nechvatal, J. R.; Smid, M. E.; Barker, E. B.; Leigh, S. D.; Levenson, M.; Vangel, M.; Banks, D. L.; Heckert, N. A.; Dray, J. F.; Vo, S. A statistical test suite for random and pseudorandom number generators for cryptographic applications. NIST SP 800-22r1a; National Institute of Standards and Technology: Gaithersburg, MD, 2010.Google ScholarThere is no corresponding record for this reference.
- 32Zhou, Y.; Chen, X.; Ko, C.; Yang, Z.; Mouli, C.; Ramanathan, S. Voltage-triggered ultrafast phase transition in vanadium dioxide switches. IEEE Electron Device Lett. 2013, 34, 220– 222, DOI: 10.1109/LED.2012.2229457Google Scholar32https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXkt1GksLY%253D&md5=62b1d97cc25ce189e4f809603bc7d92fVoltage-triggered ultrafast phase transition in vanadium dioxide switchesZhou, You; Chen, Xiaonan; Ko, Changhyun; Yang, Zheng; Mouli, Chandra; Ramanathan, ShriramIEEE Electron Device Letters (2013), 34 (2), 220-222CODEN: EDLEDZ; ISSN:0741-3106. (Institute of Electrical and Electronics Engineers)Elec. driven metal-insulator transition (MIT) in vanadium dioxide (VO2) is of interest in emerging memory devices, neural computation, and high-speed electronics. We report on the fabrication of out-of-plane VO2 metal-insulator-metal structures and reproducible high-speed switching measurements in these two-terminal devices. We have obsd. a clear correlation between the elec. driven ON/OFF current ratio and the thermally induced resistance change during MIT. It is also found that sharp MIT could be triggered by the external voltage pulses within 2 ns at room temp. and the achieved ON/OFF ratio is greater than two orders of magnitude with good endurance.
- 33Brockman, J. S.; Gao, L.; Hughes, B.; Rettner, C. T.; Samant, M. G.; Roche, K. P.; Parkin, S. S. P. Subnanosecond incubation times for electric-field-induced metallization of a correlated electron oxide. Nat. Nanotechnol. 2014, 9, 453– 458, DOI: 10.1038/nnano.2014.71Google Scholar33https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXmsVKms70%253D&md5=e857866a2a3dba2fa7d6fd981b4e7658Subnanosecond incubation times for electric-field-induced metallization of a correlated electron oxideBrockman, Justin S.; Gao, Li; Hughes, Brian; Rettner, Charles T.; Samant, Mahesh G.; Roche, Kevin P.; Parkin, Stuart S. P.Nature Nanotechnology (2014), 9 (6), 453-458CODEN: NNAABX; ISSN:1748-3387. (Nature Publishing Group)Strong interactions, or correlations, between the d or f electrons in transition-metal oxides lead to various types of metal-insulator transitions that can be triggered by external parameters such as temp., pressure, doping, magnetic fields and elec. fields. Elec.-field-induced metalization of such materials from their insulating states could enable a new class of ultrafast electronic switches and latches. However, significant questions remain about the detailed nature of the switching process. Here, we show, in the canonical metal-to-insulator transition system V2O3, that ultrafast voltage pulses result in its metalization only after an incubation time that ranges from ∼150 ps to many nanoseconds, depending on the elec. field strength. We show that these incubation times can be accounted for by purely thermal effects and that intrinsic electronic-switching mechanisms may only be revealed using larger elec. fields at even shorter timescales.
- 34Cario, L.; Vaju, C.; Corraze, B.; Guiot, V.; Janod, E. Electric-field-induced resistive switching in a family of mott insulators: Towards a new class of RRAM memories. Adv. Mater. 2010, 22, 5193, DOI: 10.1002/adma.201002521Google Scholar34https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXhsFSgs7nL&md5=a350b8919397d8b9a4593f2908212076Electric-field-induced resistive switching in a family of Mott insulators: towards a new class of RRAM memoriesCario, Laurent; Vaju, Cristian; Corraze, Benoit; Guiot, Vincent; Janod, EtienneAdvanced Materials (Weinheim, Germany) (2010), 22 (45), 5193-5197CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)We have discovered that the fragile Mott insulator compds. AM4X8 exhibit both a volatile and a non-volatile unipolar RS. All our indicate that this RS is related neither to a chem. nor to an amorphous-cryst. phase change as obsd. in all other systems where a RS occurs. Conversly we have demonstrated that the volatile RS found in the Mott insulators AM4X8 is related to an elec.-field-driven IMT. These compds. could, therefore, offer the possibility to explore a new type of RRAM based on a new mechanism likely related to the Mott physics.
- 35Radu, I. P.; Govoreanu, B.; Mertens, S.; Shi, X.; Cantoro, M.; Schaekers, M.; Jurczak, M.; De Gendt, S.; Stesmans, A.; Kittl, J. A.; Heyns, M.; Martens, K. Switching mechanism in two-terminal vanadium dioxide devices. Nanotechnology 2015, 26, 165202, DOI: 10.1088/0957-4484/26/16/165202Google Scholar35https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BC2MnosFGqsA%253D%253D&md5=0ddb03ca089be8f620f467a3825d5a3fSwitching mechanism in two-terminal vanadium dioxide devicesRadu Iuliana P; Govoreanu B; Mertens S; Shi X; Cantoro M; Schaekers M; Jurczak M; De Gendt S; Stesmans A; Kittl J A; Heyns M; Martens KNanotechnology (2015), 26 (16), 165202 ISSN:.Two-terminal thin film VO2 devices show an abrupt decrease of resistance when the current or voltage applied exceeds a threshold value. This phenomenon is often described as a field-induced metal-insulator transition. We fabricate nano-scale devices with different electrode separations down to 100 nm and study how the dc switching voltage and current depend on device size and temperature. Our observations are consistent with a Joule heating mechanism governing the switching. Pulsed measurements show a switching time to the high resistance state of the order of one hundred nanoseconds, consistent with heat dissipation time. In spite of the Joule heating mechanism which is expected to induce device degradation, devices can be switched for more than 10(10) cycles making VO2 a promising material for nanoelectronic applications.
- 36Seo, G.; Kim, B. J.; Ko, C.; Cui, Y.; Lee, Y. W.; Shin, J. H.; Ramanathan, S.; Kim, H. T. Voltage-pulse-induced switching dynamics in VO2 thin-film devices on silicon. IEEE Electron Device Lett. 2011, 32, 1582, DOI: 10.1109/LED.2011.2163922Google Scholar36https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3MXhs1emu73F&md5=2589e0094dfa797314b669e1a3162f7eVoltage-pulse-induced switching dynamics in VO2 thin-film devices on siliconSeo, Giwan; Kim, Bong-Jun; Ko, Changhyun; Cui, Yanjie; Lee, Yong Wook; Shin, Jun-Hwan; Ramanathan, Shriram; Kim, Hyun-TakIEEE Electron Device Letters (2011), 32 (11), 1582-1584CODEN: EDLEDZ; ISSN:0741-3106. (Institute of Electrical and Electronics Engineers)We demonstrate the characteristics of voltage-pulse-induced out-of-plane switching driven by metal-insulator transition (MIT jump) with VO2 thin-film devices fabricated on silicon. As the peak of a triangular pulse increases, the delay time from the insulating state to the metallic state linearly decreases and is independent of change in external resistance. The intrinsic rising time is less than 170 ns. An endurance test with continuous voltage pulse shows reliability without breakdown for more than 110 h. This work contributes to correlated oxide electronics utilizing phase transition layers.
Cited By
This article is cited by 8 publications.
- Sofie J. Studholme, Simon A. Brown. Integer Factorization with Stochastic Spiking in Percolating Networks of Nanoparticles. ACS Nano 2024, 18
(41)
, 28060-28069. https://doi.org/10.1021/acsnano.4c07200
- Tian-Yue Chen, Haowen Ren, Nareg Ghazikhanian, Ralph El Hage, Dayne Y. Sasaki, Pavel Salev, Yayoi Takamura, Ivan K. Schuller, Andrew D. Kent. Electrical Control of Magnetic Resonance in Phase Change Materials. Nano Letters 2024, 24
(37)
, 11476-11481. https://doi.org/10.1021/acs.nanolett.4c02697
- Noémie Bidoul, Nicolas Roisin, Denis Flandre. Tuning the Intrinsic Stochasticity of Resistive Switching in VO2 Microresistors. Nano Letters 2024, 24
(21)
, 6201-6209. https://doi.org/10.1021/acs.nanolett.4c00184
- Tae Joon Park, Kemal Selcuk, Hai-Tian Zhang, Sukriti Manna, Rohit Batra, Qi Wang, Haoming Yu, Navid Anjum Aadit, Subramanian K. R. S. Sankaranarayanan, Hua Zhou, Kerem Y. Camsari, Shriram Ramanathan. Efficient Probabilistic Computing with Stochastic Perovskite Nickelates. Nano Letters 2022, 22
(21)
, 8654-8661. https://doi.org/10.1021/acs.nanolett.2c03223
- Lucia Varbaro, Bernat Mundet, Claribel Domínguez, Jennifer Fowlie, Alexandru B. Georgescu, Lukas Korosec, Duncan T.L. Alexander, Jean‐Marc Triscone. Electronic Coupling of Metal‐to‐Insulator Transitions in Nickelate‐Based Heterostructures. Advanced Electronic Materials 2023, 19 , 2201291. https://doi.org/10.1002/aelm.202201291
- Theodor Luibrand, Adrien Bercher, Rodolfo Rocco, Farnaz Tahouni-Bonab, Lucia Varbaro, Carl Willem Rischau, Claribel Domínguez, Yixi Zhou, Weiwei Luo, Soumen Bag, Lorenzo Fratino, Reinhold Kleiner, Stefano Gariglio, Dieter Koelle, Jean-Marc Triscone, Marcelo J. Rozenberg, Alexey B. Kuzmenko, Stefan Guénon, Javier del Valle. Characteristic length scales of the electrically induced insulator-to-metal transition. Physical Review Research 2023, 5
(1)
https://doi.org/10.1103/PhysRevResearch.5.013108
- Parker Schofield, Adelaide Bradicich, Rebeca M. Gurrola, Yuwei Zhang, Timothy D. Brown, Matt Pharr, Patrick J. Shamberger, Sarbajit Banerjee. Harnessing the Metal–Insulator Transition of VO
2
in Neuromorphic Computing. Advanced Materials 2022, 11 , 2205294. https://doi.org/10.1002/adma.202205294
- Mohammad Samizadeh Nikoo, Reza Soleimanzadeh, Anna Krammer, Guilherme Migliato Marega, Yunkyu Park, Junwoo Son, Andreas Schueler, Andras Kis, Philip J. W. Moll, Elison Matioli. Electrical control of glass-like dynamics in vanadium dioxide for data storage and processing. Nature Electronics 2022, 5
(9)
, 596-603. https://doi.org/10.1038/s41928-022-00812-z
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
References
This article references 36 other publications.
- 1Beyond von Neumann. Nat. Nanotechnol 2020, 15, 507– 507 DOI: 10.1038/s41565-020-0738-x .There is no corresponding record for this reference.
- 2Alaghi, A.; Qian, W.; Hayes, J. P. The Promise and Challenge of Stochastic Computing. IEEE Trans. Comput. Des. Integr. Circuits Syst. 2018, 37, 1515– 1531, DOI: 10.1109/TCAD.2017.2778107There is no corresponding record for this reference.
- 3Alaghi, A.; Hayes, J. P. Survey of stochastic computing. Trans. Embed. Comput. Syst. 2013, 12, 1, DOI: 10.1145/2465787.2465794There is no corresponding record for this reference.
- 4von Neumann, J. Probabilistic Logics and the Synthesis of Reliable Organisms From Unreliable Components. Autom. Stud. 1956, 34, 43– 98, DOI: 10.1515/9781400882618-003There is no corresponding record for this reference.
- 5Poppelbaum, W. J.; Afuso, C.; Esch, J. W. Stochastic computing elements and systems. AFIPS’67 Proc. Fall Joint Computer Conference ; Anaheim, CA, 1967; ACM Press, 1967; Vol. 1, pp 635– 644.There is no corresponding record for this reference.
- 6Gaines, B. R. Stochastic Computing Systems. Advances in Information Systems Science; Springer US, 1969; pp 37– 172.There is no corresponding record for this reference.
- 7Wood, F.; Van De Meent, J. W.; Mansinghka, V. A New Approach to Probabilistic Programming Inference. Proc. 17th Int. Conf. Artif. Intell. Stat. 33, Reykjavic, Iceland (22th-25th April 2014). Preprint availale at arXiv:1507.00996There is no corresponding record for this reference.
- 8Ghahramani, Z. Probabilistic machine learning and artificial intelligence. Nat. 2015 5217553 2015, 521, 452– 459, DOI: 10.1038/nature14541There is no corresponding record for this reference.
- 9Bottou, L. Large-Scale Machine Learning with Stochastic Gradient Descent. Proc. COMPSTAT’2010 2010, 177– 186There is no corresponding record for this reference.
- 10Salakhutdinov, R.; Hinton, G. Deep Boltzmann Machines. Proc. 12th Int. Conference Artif. Intell. Stat. 2009, 5, 448There is no corresponding record for this reference.
- 11Zhang, N.; Ding, S.; Zhang, J.; Xue, Y. An overview on Restricted Boltzmann Machines. Neurocomputing 2018, 275, 1186– 1199, DOI: 10.1016/j.neucom.2017.09.065There is no corresponding record for this reference.
- 12Camsari, K. Y.; Sutton, B. M.; Datta, S. p-bits for probabilistic spin logic. Appl. Phys. Rev. 2019, 6, 011305, DOI: 10.1063/1.505586012https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXkvVSrsb4%253D&md5=3cdc33a79537e878a3165f4acc93e0deThe p-bits for probabilistic spin logicCamsari, Kerem Y.; Sutton, Brian M.; Datta, SupriyoApplied Physics Reviews (2019), 6 (1), 011305/1-011305/12CODEN: APRPG5; ISSN:1931-9401. (American Institute of Physics)We introduce the concept of a probabilistic or p-bit, intermediate between the std. bits of digital electronics and the emerging q-bits of quantum computing. We show that low barrier magnets or LBMs provide a natural phys. representation for p-bits and can be built either from perpendicular magnets designed to be close to the in-plane transition or from circular in-plane magnets. Magnetic tunnel junctions (MTJs) built using LBMs as free layers can be combined with std. NMOS transistors to provide three-terminal building blocks for large scale probabilistic circuits that can be designed to perform useful functions. Interestingly, this three-terminal unit looks just like the 1T/MTJ device used in embedded magnetic random access memory technol., with only one difference: the use of an LBM for the MTJ free layer. We hope that the concept of p-bits and p-circuits will help open up new application spaces for this emerging technol. However, a p-bit need not involve an MTJ; any fluctuating resistor could be combined with a transistor to implement it, while completely digital implementations using conventional CMOS technol. are also possible. The p-bit also provides a conceptual bridge between two active but disjoint fields of research, namely, stochastic machine learning and quantum computing. First, there are the applications that are based on the similarity of a p-bit to the binary stochastic neuron (BSN), a well-known concept in machine learning. Three-terminal p-bits could provide an efficient hardware accelerator for the BSN. Second, there are the applications that are based on the p-bit being like a poor man's q-bit. Initial demonstrations based on full SPICE simulations show that several optimization problems, including quantum annealing are amenable to p-bit implementations which can be scaled up at room temp. using existing technol. (c) 2019 American Institute of Physics.
- 13Matsumoto, M.; Nishimura, T. ACM Trans. Model. Comput. Simul. 1998, 8, 3– 30, DOI: 10.1145/272991.272995There is no corresponding record for this reference.
- 14Borders, W. A.; Pervaiz, A. Z.; Fukami, S.; Camsari, K. Y.; Ohno, H.; Datta, S. Integer factorization using stochastic magnetic tunnel junctions. Nature 2019, 573, 390, DOI: 10.1038/s41586-019-1557-914https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhvVWit7%252FE&md5=665cb85a1bf3ab1148cbe6495c94ccafInteger factorization using stochastic magnetic tunnel junctionsBorders, William A.; Pervaiz, Ahmed Z.; Fukami, Shunsuke; Camsari, Kerem Y.; Ohno, Hideo; Datta, SupriyoNature (London, United Kingdom) (2019), 573 (7774), 390-393CODEN: NATUAS; ISSN:0028-0836. (Nature Research)Conventional computers operate deterministically using strings of zeros and ones called bits to represent information in binary code. Despite the evolution of conventional computers into sophisticated machines, there are many classes of problems that they cannot efficiently address, including inference, invertible logic, sampling and optimization, leading to considerable interest in alternative computing schemes. Quantum computing, which uses qubits to represent a superposition of 0 and 1, is expected to perform these tasks efficiently1-3. However, decoherence and the current requirement for cryogenic operation4, as well as the limited many-body interactions that can be implemented, pose considerable challenges. Probabilistic computing1,5-7 is another unconventional computation scheme that shares similar concepts with quantum computing but is not limited by the above challenges. The key role is played by a probabilistic bit (a p-bit)-a robust, classical entity fluctuating in time between 0 and 1, which interacts with other p-bits in the same system using principles inspired by neural networks8. Here we present a proof-of-concept expt. for probabilistic computing using spintronics technol., and demonstrate integer factorization, an illustrative example of the optimization class of problems addressed by adiabatic9 and gated2 quantum computing. Nanoscale magnetic tunnel junctions showing stochastic behavior are developed by modifying market-ready magnetoresistive random-access memory technol.10,11 and are used to implement three-terminal p-bits that operate at room temp. The p-bits are elec. connected to form a functional asynchronous network, to which a modified adiabatic quantum computing algorithm that implements three- and four-body interactions is applied. Factorization of integers up to 945 is demonstrated with this rudimentary asynchronous probabilistic computer using eight correlated p-bits, and the results show good agreement with theor. predictions, thus providing a potentially scalable hardware approach to the difficult problems of optimization and sampling.
- 15Camsari, K. Y.; Faria, R.; Sutton, B. M.; Datta, S. Stochastic p-Bits for Invertible Logic. Phys. Rev. X. 2017, 7, 031014, DOI: 10.1103/PhysRevX.7.03101415https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXitV2ks7jF&md5=fce208cd8818c75e525245bc02121c6bStochastic p-bits for invertible logicCamsari, Kerem Yunus; Faria, Rafatul; Sutton, Brian M.; Datta, SupriyoPhysical Review X (2017), 7 (3), 031014/1-031014/19CODEN: PRXHAE; ISSN:2160-3308. (American Physical Society)Conventional semiconductor-based logic and nanomagnet-based memory devices are built out of stable, deterministic units such as std. metal-oxide semiconductor transistors, or nanomagnets with energy barriers in excess of ≈ 40-60 kT. In this paper, we show that unstable, stochastic units, which we call "p-bits", can be interconnected to create robust correlations that implement precise Boolean functions with impressive accuracy, comparable to std. digital circuits. At the same time, they are invertible, a unique property that is absent in std. digital circuits. When operated in the direct mode, the input is clamped, and the network provides the correct output. In the inverted mode, the output is clamped, and the network fluctuates among all possible inputs that are consistent with that output. First, here they present a detailed implementation of an invertible gate to bring out the key role of a single three-terminal transistorlike building block to enable the construction of correlated p-bit networks. The results for this specific, CMOS-assisted nanomagnet-based hardware implementation agree well with those from a universal model for p-bits, showing that p-bits need not be magnet based: any three-terminal tunable random bit generator should be suitable. They present a general algorithm for designing a Boltzmann machine (BM) with a sym. connection matrix [J] (Jij = Jji) that implements a given truth table with p-bits. The [J] matrixes are relatively sparse with a few unique wts. for convenient hardware implementation. Then show how BM full adders can be interconnected in a partially directed manner (Jij not equal to Jji) to implement large logic operations such as 32-bit binary addn. Hundreds of stochastic p-bits get precisely correlated such that the correct answer out of 233 (≈ 8 x 109) possibilities can be extd. by looking at the statistical mode or majority vote of a no. of time samples. With perfect directivity (Jji = 0) a small no. of samples is enough, while for less directed connections more samples are needed, but even in the former case logical invertibility is largely preserved. This combination of digital accuracy and logical invertibility is enabled by the hybrid design that uses bidirectional BM units to construct circuits with partially directed interunit connections. In this paper, they establish this key result with extensive examples including a 4-bit multiplier which in inverted mode functions as a factorizer.
- 16Safranski, C.; Kaiser, J.; Trouilloud, P.; Hashemi, P.; Hu, G.; Sun, J. Z. Demonstration of Nanosecond Operation in Stochastic Magnetic Tunnel Junctions. Nano Lett. 2021, 21, 2040– 2045, DOI: 10.1021/acs.nanolett.0c0465216https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3MXkvFyisLw%253D&md5=69b0f1b971fe41b20ae8244ff43b85d0Demonstration of Nanosecond Operation in Stochastic Magnetic Tunnel JunctionsSafranski, Christopher; Kaiser, Jan; Trouilloud, Philip; Hashemi, Pouya; Hu, Guohan; Sun, Jonathan Z.Nano Letters (2021), 21 (5), 2040-2045CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)Magnetic tunnel junctions operating in the superparamagnetic regime are promising devices in the field of probabilistic computing, which is suitable for applications like high-dimensional optimization or sampling problems. Further, random no. generation is of interest in the field of cryptog. For such applications, a device's uncorrelated fluctuation time-scale can det. the effective system speed. It has been theor. proposed that a magnetic tunnel junction designed to have only easy-plane anisotropy provides fluctuation rates detd. by its easy-plane anisotropy field and can perform on a nanosecond or faster time-scale as measured by its magnetoresistance's autocorrelation in time. Here, we provide exptl. evidence of nanosecond scale fluctuations in a circular-shaped easy-plane magnetic tunnel junction, consistent with finite-temp. coupled macrospin simulation results and prior theor. expectations. We further assess the degree of stochasticity of such a signal.
- 17Chen, L.; Zhou, P.; Kalcheim, Y.; Schuller, I. K.; Natelson, D. Percolation and nanosecond fluctuators in V2O3films within the metal-insulator transition. APL Mater. 2020, 8, 101103, DOI: 10.1063/5.002347517https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhvFyqurbK&md5=1fa2db014ea7bd67ecae1c70391382dfPercolation and nanosecond fluctuators in V2O3 films within the metal-insulator transitionChen, Liyang; Zhou, Panpan; Kalcheim, Yoav; Schuller, Ivan K.; Natelson, DouglasAPL Materials (2020), 8 (10), 101103CODEN: AMPADS; ISSN:2166-532X. (American Institute of Physics)Vanadium sesquioxide (V2O3) exhibits a metal-insulator transition (MIT) at 160 K between a low temp., monoclinic, antiferromagnetic Mott insulator and a high temp., rhombohedral, paramagnetic, metallic phase. In thin films, a percolative transition takes place over a finite temp. range of phase coexistence. We study the fluctuating dynamics of this percolative MIT by measuring voltage noise spectra at both low frequencies (up to 100 kHz) and radio frequencies (between 10 MHz and 1 GHz). Noise intensity quadratic in bias is obsd. in the MIT region, as expected for resistive fluctuations probed nonperturbatively by the current. The low frequency noise resembles flicker-type 1/fβ noise, often taking on the form of Lorentzian noise dominated by a small no. of fluctuators as the vol. fraction of the insulating phase dominates. Radio frequency noise intensity also quadratic in the bias current allows the identification of resistance fluctuations with lifetimes below 1 ns, approaching timescales seen in non-equil. pump-probe studies of the transition. We find quant. consistency with a model for fluctuations in the percolative fraction. The thermodn. of the MIT suggests that dominant fluctuations are ones that alter small vols. affecting the connectivity of domain boundaries. This noise serves as a sensitive and nonperturbative probe for the dynamics of switching phenomena in this system. (c) 2020 American Institute of Physics.
- 18Imada, M.; Fujimori, A.; Tokura, Y. Metal-insulator transitions. Rev. Mod. Phys. 1998, 70, 1039– 1263, DOI: 10.1103/RevModPhys.70.103918https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaK1cXns1OhtLk%253D&md5=ff8bff3d0414933473d17808b96562b0Metal-insulator transitionsImada, Masatoshi; Fujimori, Atsushi; Tokura, YoshinoriReviews of Modern Physics (1998), 70 (4, Pt. 1), 1039-1263CODEN: RMPHAT; ISSN:0034-6861. (American Physical Society)A review with a large no. of refs. Metal-insulator transitions are accompanied by huge resistivity changes, even over tens of orders of magnitude, and are widely obsd. in condensed-matter systems. This article presents the observations and current understanding of the metal-insulator transition with a pedagogical introduction to the subject. Esp. important are the transitions driven by correlation effects assocd. with the electron-electron interaction. The insulating phase caused by the correlation effects is categorized as the Mott Insulator. Near the transition point the metallic state shows fluctuations and orderings in the spin, charge, and orbital degrees of freedom. The properties of these metals are frequently quite different from those of ordinary metals, as measured by transport, optical, and magnetic probes. The review 1st describes theor. approaches to the unusual metallic states and to the metal-insulator transition. The Fermi-liq. theory treats the correlations that can be adiabatically connected with the noninteracting picture. Strong-coupling models that do not require Fermi-liq. behavior have also been developed. Much work has also been done on the scaling theory of the transition. A central issue for this review is the evaluation of these approaches in simple theor. systems such as the Hubbard model and t-J models. Another key issue is strong competition among various orderings as in the interplay of spin and orbital fluctuations. Exptl., the unusual properties of the metallic state near the insulating transition have been most extensively studied in d-electron systems. In particular, there is revived interest in transition-metal oxides, motivated by the epoch-making findings of high-temp. supercond. in cuprates and colossal magnetoresistance in manganites. The article reviews the rich phenomena of anomalous metallicity, taking as examples Ti, V, Cr, Mn, Fe, Co, Ni, Cu, and Ru compds. The diverse phenomena include strong spin and orbital fluctuations, mass renormalization effects, incoherence of charge dynamics, and phase transitions under control of key parameters such as band filling, bandwidth, and dimensionality. These parameters are exptl. varied by doping, pressure, chem. compn., and magnetic fields. Much of the obsd. behavior can be described by the current theory. Open questions and future problems are also extd. from comparison between exptl. results and theor. achievements.
- 19Kumar, S.; Strachan, J. P.; Williams, R. S. Chaotic dynamics in nanoscale NbO 2 Mott memristors for analogue computing. Nature 2017, 548, 318– 321, DOI: 10.1038/nature2330719https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXhtlSrsLzJ&md5=59224eec4fd1185459b875bb2828a44aChaotic dynamics in nanoscale NbO2 Mott memristors for analogue computingKumar, Suhas; Strachan, John Paul; Williams, R. StanleyNature (London, United Kingdom) (2017), 548 (7667), 318-321CODEN: NATUAS; ISSN:0028-0836. (Nature Publishing Group)At present, machine learning systems use simplified neuron models that lack the rich nonlinear phenomena obsd. in biol. systems, which display spatio-temporal cooperative dynamics. There is evidence that neurons operate in a regime called the edge of chaos that may be central to complexity, learning efficiency, adaptability and analog (non-Boolean) computation in brains. Neural networks have exhibited enhanced computational complexity when operated at the edge of chaos, and networks of chaotic elements have been proposed for solving combinatorial or global optimization problems. Thus, a source of controllable chaotic behavior that can be incorporated into a neural-inspired circuit may be an essential component of future computational systems. Such chaotic elements have been simulated using elaborate transistor circuits that simulate known equations of chaos, but an exptl. realization of chaotic dynamics from a single scalable electronic device has been lacking. Here we describe niobium dioxide (NbO2) Mott memristors each less than 100 nm across that exhibit both a nonlinear-transport-driven current-controlled neg. differential resistance and a Mott-transition-driven temp.-controlled neg. differential resistance. Mott materials have a temp.-dependent metal-insulator transition that acts as an electronic switch, which introduces a history-dependent resistance into the device. We incorporate these memristors into a relaxation oscillator and observe a tunable range of periodic and chaotic self-oscillations. We show that the nonlinear current transport coupled with thermal fluctuations at the nanoscale generates chaotic oscillations. Such memristors could be useful in certain types of neural-inspired computation by introducing a pseudo-random signal that prevents global synchronization and could also assist in finding a global min. during a constrained search. We specifically demonstrate that incorporating such memristors into the hardware of a Hopfield computing network can greatly improve the efficiency and accuracy of converging to a soln. for computationally difficult problems.
- 20Kumar, S.; Williams, R. S.; Wang, Z. Third-order nanocircuit elements for neuromorphic engineering. Nature 2020, 585, 518– 523, DOI: 10.1038/s41586-020-2735-520https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhvFegu7nL&md5=847286203f18d8c1bd5c85a737735508Third-order nanocircuit elements for neuromorphic engineeringKumar, Suhas; Williams, R. Stanley; Wang, ZiwenNature (London, United Kingdom) (2020), 585 (7826), 518-523CODEN: NATUAS; ISSN:0028-0836. (Nature Research)Abstr.: Current hardware approaches to biomimetic or neuromorphic artificial intelligence rely on elaborate transistor circuits to simulate biol. functions. However, these can instead be more faithfully emulated by higher-order circuit elements that naturally express neuromorphic nonlinear dynamics1-4. Generating neuromorphic action potentials in a circuit element theor. requires a min. of third-order complexity (for example, three dynamical electrophys. processes)5, but there have been few examples of second-order neuromorphic elements, and no previous demonstration of any isolated third-order element6-8. Using both expts. and modeling, here we show how multiple electrophys. processes-including Mott transition dynamics-form a nanoscale third-order circuit element. We demonstrate simple transistorless networks of third-order elements that perform Boolean operations and find analog solns. to a computationally hard graph-partitioning problem. This work paves a way towards very compact and densely functional neuromorphic computing primitives, and energy-efficient validation of neuroscientific models.
- 21Jerry, M.; Ni, K.; Parihar, A.; Raychowdhury, A.; Datta, S. Stochastic Insulator-to-Metal Phase Transition-Based True Random Number Generator. IEEE Electron Device Lett. 2018, 39, 139– 142, DOI: 10.1109/LED.2017.277181221https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXitFGksbrO&md5=d7761d622597a7361acf3d99505e9f40Stochastic insulator-to-metal phase transition-based true random number generatorJerry, Matthew; Ni, Kai; Parihar, Abhinav; Raychowdhury, Arijit; Datta, SumanIEEE Electron Device Letters (2018), 39 (1), 139-142CODEN: EDLEDZ; ISSN:1558-0563. (Institute of Electrical and Electronics Engineers)An oscillator-based true random no. generator (TRNG) is exptl. demonstrated by exploiting inherently stochastic threshold switching in the insulatorto-metal transition (IMT) in vanadium dioxide. Through experimentation and modeling, we show that the origin of stochasticity arises from small perturbations in the nanoscale domain structure, which are then subsequently amplified through a pos. feedback process. Within a 1T1R oscillator, the stochastic cycle-to-cycle variations in the IMT trigger voltage result in random timing jitter, which is harnessed for a TRNG. The randomness of the IMT TRNG output is validated using the NIST SP800-22 statistical test.
- 22del Valle, J.; Vargas, N. M.; Rocco, R.; Salev, P.; Kalcheim, Y.; Lapa, P. N.; Adda, C.; Lee, M.-H.; Wang, P. Y.; Fratino, L.; Rozenberg, M. J.; Schuller, I. K. Spatiotemporal characterization of the field-induced insulator-to-metal transition. Science (80-.) 2021, 373, 907, DOI: 10.1126/science.abd9088There is no corresponding record for this reference.
- 23Pickett, M. D.; Medeiros-Ribeiro, G.; Williams, R. S. A scalable neuristor built with Mott memristors. Nat. Mater. 2013, 12, 114– 117, DOI: 10.1038/nmat351023https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC38XhvVersr%252FP&md5=2fcede63aebca82666a458a870dc418cA scalable neuristor built with Mott memristorsPickett, Matthew D.; Medeiros-Ribeiro, Gilberto; Williams, R. StanleyNature Materials (2013), 12 (2), 114-117CODEN: NMAACR; ISSN:1476-1122. (Nature Publishing Group)The Hodgkin-Huxley model for action potential generation in biol. axons is central for understanding the computational capability of the nervous system and emulating its functionality. Owing to the historical success of silicon complementary metal-oxide-semiconductors, spike-based computing is primarily confined to software simulations and specialized analog metal-oxide-semiconductor field-effect transistor circuits. However, there is interest in constructing phys. systems that emulate biol. functionality more directly, with the goal of improving efficiency and scale. The neuristor was proposed as an electronic device with properties similar to the Hodgkin-Huxley axon, but previous implementations were not scalable. Here we demonstrate a neuristor built using two nanoscale Mott memristors, dynamical devices that exhibit transient memory and neg. differential resistance arising from an insulating-to-conducting phase transition driven by Joule heating. This neuristor exhibits the important neural functions of all-or-nothing spiking with signal gain and diverse periodic spiking, using materials and structures that are amenable to extremely high-d. integration with or without silicon transistors.
- 24del Valle, J.; Salev, P.; Kalcheim, Y.; Schuller, I. K. A caloritronics-based Mott neuristor. Sci. Rep. 2020, 10, 1– 10, DOI: 10.1038/s41598-020-61176-yThere is no corresponding record for this reference.
- 25Bohaichuk, S. M.; Kumar, S.; Pitner, G.; McClellan, C. J.; Jeong, J.; Samant, M. G.; Wong, H. S. P.; Parkin, S. S. P.; Williams, R. S.; Pop, E. Fast Spiking of a Mott VO2-Carbon Nanotube Composite Device. Nano Lett. 2019, 19, 6751– 6755, DOI: 10.1021/acs.nanolett.9b0155425https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhs1aju73I&md5=6ecc4bbd2b4f2fce96a56fd2db1a7758Fast Spiking of a Mott VO2-Carbon Nanotube Composite DevicesBohaichuk, Stephanie M.; Kumar, Suhas; Pitner, Greg; McClellan, Connor J.; Jeong, Jaewoo; Samant, Mahesh G.; Wong, H-.S. Philip; Parkin, Stuart S. P.; Williams, R. Stanley; Pop, EricNano Letters (2019), 19 (10), 6751-6755CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)The recent surge of interest in brain-inspired computing and power-efficient electronics has dramatically bolstered development of computation and communication using neuron-like spiking signals. Devices that can produce rapid and energy-efficient spiking could significantly advance these applications. Here we demonstrate d.c. or voltage-driven periodic spiking with sub-20 ns pulse widths from a single device composed of a thin VO2 film with a metallic carbon nanotube as a nanoscale heater, without using an external capacitor. Compared with VO2-only devices, adding the nanotube heater dramatically decreases the transient duration and pulse energy, and increases the spiking frequency, by up to 3 orders of magnitude. This is caused by heating and cooling of the VO2 across its insulator-metal transition being localized to a nanoscale conduction channel in an otherwise bulk medium. This result provides an important component of energy-efficient neuromorphic computing systems and a lithog.-free technique for energy-scaling of electronic devices that operate via bulk mechanisms.
- 26Lee, D.; Lee, J.; Song, K.; Xue, F.; Choi, S. Y.; Ma, Y.; Podkaminer, J.; Liu, D.; Liu, S. C.; Chung, B.; Fan, W.; Cho, S. J.; Zhou, W.; Lee, J.; Chen, L. Q.; Oh, S. H.; Ma, Z.; Eom, C. B. Sharpened VO2 Phase Transition via Controlled Release of Epitaxial Strain. Nano Lett. 2017, 17, 5614– 5619, DOI: 10.1021/acs.nanolett.7b0248226https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXht1elu7vM&md5=eedbbe736545e23f5232881dce40bb67Sharpened VO2 phase transition via controlled release of epitaxial strainLee, Daesu; Lee, Jaeseong; Song, Kyung; Xue, Fei; Choi, Si-Young; Ma, Yanjun; Podkaminer, Jacob; Liu, Dong; Liu, Shih-Chia; Chung, Bongwook; Fan, Wenjuan; Cho, Sang June; Zhou, Weidong; Lee, Jaichan; Chen, Long-Qing; Oh, Sang Ho; Ma, Zhenqiang; Eom, Chang-BeomNano Letters (2017), 17 (9), 5614-5619CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)Phase transitions in correlated materials can be manipulated at the nanoscale to yield emergent functional properties, promising new paradigms for nanoelectronics and nanophotonics. Vanadium dioxide (VO2), an archetypal correlated material, exhibits a metal-insulator transition (MIT) above room temp. At the thicknesses required for heterostructure applications, such as an optical modulator discussed, the strain state of VO2 largely dets. the MIT dynamics crit. to the device performance. The authors develop an approach to control the MIT dynamics in epitaxial VO2 films by employing an intermediate template layer with large lattice mismatch to relieve the interfacial lattice constraints, contrary to conventional thin film epitaxy that favors lattice match between the substrate and the growing film. A combination of phase-field simulation, in situ real-time nanoscale imaging, and elec. measurements reveals robust undisturbed MIT dynamics even at preexisting structural domain boundaries and significantly sharpened MIT in the templated VO2 films. Utilizing the sharp MIT, the authors demonstrate a fast, elec. switchable optical waveguide. This study offers unconventional design principles for heteroepitaxial correlated materials, as well as novel insight into their nanoscale phase transitions.
- 27Butakov, N. A.; Knight, M. W.; Lewi, T.; Iyer, P. P.; Higgs, D.; Chorsi, H. T.; Trastoy, J.; Del Valle Granda, J.; Valmianski, I.; Urban, C.; Kalcheim, Y.; Wang, P. Y.; Hon, P. W. C.; Schuller, I. K.; Schuller, J. A. Broadband Electrically Tunable Dielectric Resonators Using Metal-Insulator Transitions. ACS Photonics 2018, 5, 4056– 4060, DOI: 10.1021/acsphotonics.8b0069927https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhs1KhurzE&md5=b0625158c361d56b16ae833b3ce9e270Broadband Electrically Tunable Dielectric Resonators Using Metal-Insulator TransitionsButakov, Nikita A.; Knight, Mark W.; Lewi, Tomer; Iyer, Prasad P.; Higgs, David; Chorsi, Hamid T.; Trastoy, Juan; Del Valle Granda, Javier; Valmianski, Ilya; Urban, Christian; Kalcheim, Yoav; Wang, Paul Y.; Hon, Philip W. C.; Schuller, Ivan K.; Schuller, Jon A.ACS Photonics (2018), 5 (10), 4056-4060CODEN: APCHD5; ISSN:2330-4022. (American Chemical Society)Dielec.-resonator-based nanophotonic devices show promise owing to their low intrinsic losses, support of multipolar resonances, and efficient operation in both reflection and transmission configurations. A key challenge is to make such devices dynamically switchable, such that optical behavior can be instantaneously reconfigured. Large, broadband, and continuous elec. tuning of reflection resonances in hybrid dielec.-VO2 devices is exptl. demonstrated. The calcns., in strong agreement with exptl. reflectance measurements, also indicate large transmission and absorption modulation. Addnl. independent modulation of both reflection amplitude and phase at Fabry-P´erot antinodes and nodes, resp., a key requirement for metasurface design, are demonstrated. Rapid electronic modulation rates of ∼3 kHz are achieved, substantially faster than other recent approaches. These findings greatly expand the potential of designing nanophotonic devices that exploit the tunable behavior of hybrid dielec.-VO2 resonators.
- 28Kumar, S.; Pickett, M. D.; Strachan, J. P.; Gibson, G.; Nishi, Y.; Williams, R. S. Local Temperature Redistribution and Structural Transition During Joule-Heating-Driven Conductance Switching in VO 2. Adv. Mater. 2013, 25, 6128– 6132, DOI: 10.1002/adma.20130204628https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXhtFelsL%252FF&md5=2bf6037c787384b1e02b9df9129decc2Local Temperature Redistribution and Structural Transition During Joule-Heating-Driven Conductance Switching in VO2Kumar, Suhas; Pickett, Matthew D.; Strachan, John Paul; Gibson, Gary; Nishi, Yoshio; Williams, R. StanleyAdvanced Materials (Weinheim, Germany) (2013), 25 (42), 6128-6132CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)The authors have utilized a complementary set of techniques to study the elec. and thermal behavior, and electronic structure of fabricated VO2 devices. Blackbody spectromicroscopy provided a comprehensive description of abrupt redistributions in local temp. during Joule-heating-driven insulator metal transition IMT. Filamentary conduction was accounted for in numerical simulations using a simplified electrothermal model. Using STXM maps we obsd. that the structure phase transition SPT occurred beyond the same current threshold that induced the IMT, regardless of whether or not they occur at precisely the same temp. The occurrence of IMT and SPT at the threshold current was justified by the jump in local temp. upon on-switching. The techniques employed in this study provide a spatially-resolved, direct and unambiguous means to study local temp. and local structure correlated with electronic transport. These results present a clear understanding of the driving mechanism in metal-insulator transition devices, which have gained recent interest due to their promising applications.
- 29Shabalin, A. G.; Valle, J.; Hua, N.; Cherukara, M. J.; Holt, M. V.; Schuller, I. K.; Shpyrko, O. G. Nanoscale Imaging and Control of Volatile and Non-Volatile Resistive Switching in VO 2. Small 2020, 16, 2005439, DOI: 10.1002/smll.20200543929https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXisVGlt7fJ&md5=2954b2f54c8b24a2d275441ab95577cfNanoscale Imaging and Control of Volatile and Non-Volatile Resistive Switching in VO2Shabalin, Anatoly G.; del Valle, Javier; Hua, Nelson; Cherukara, Mathew J.; Holt, Martin V.; Schuller, Ivan K.; Shpyrko, Oleg G.Small (2020), 16 (50), 2005439CODEN: SMALBC; ISSN:1613-6810. (Wiley-VCH Verlag GmbH & Co. KGaA)Control of the metal-insulator phase transition is vital for emerging neuromorphic and memristive technologies. The ability to alter the elec. driven transition between volatile and non-volatile states is particularly important for quantum-materials-based emulation of neurons and synapses. The major challenge of this implementation is to understand and control the nanoscale mechanisms behind these two fundamental switching modalities. Here, in situ X-ray nanoimaging is used to follow the evolution of the nanostructure and disorder in the archetypal Mott insulator VO2 during an elec. driven transition. Our findings demonstrate selective and reversible stabilization of either the insulating or metallic phases achieved by manipulating the defect concn. This mechanism enables us to alter the local switching response between volatile and persistent regimes and demonstrates a new possibility for nanoscale control of the resistive switching in Mott materials.
- 30del Valle, J.; Salev, P.; Tesler, F.; Vargas, N. M.; Kalcheim, Y.; Wang, P.; Trastoy, J.; Lee, M. H.; Kassabian, G.; Ramírez, J. G.; Rozenberg, M. J.; Schuller, I. K. Subthreshold firing in Mott nanodevices. Nature 2019, 569, 388– 392, DOI: 10.1038/s41586-019-1159-630https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXovFeisL0%253D&md5=4536dbb8211e48cde67d820463186624Subthreshold firing in Mott nanodevicesdel Valle, Javier; Salev, Pavel; Tesler, Federico; Vargas, Nicolas M.; Kalcheim, Yoav; Wang, Paul; Trastoy, Juan; Lee, Min-Han; Kassabian, George; Ramirez, Juan Gabriel; Rozenberg, Marcelo J.; Schuller, Ivan K.Nature (London, United Kingdom) (2019), 569 (7756), 388-392CODEN: NATUAS; ISSN:0028-0836. (Nature Research)Resistive switching, a phenomenon in which the resistance of a device can be modified by applying an elec. field1-5, is at the core of emerging technologies such as neuromorphic computing and resistive memories6-9. Among the different types of resistive switching, threshold firing10-14 is one of the most promising, as it may enable the implementation of artificial spiking neurons7,13,14. Threshold firing is obsd. in Mott insulators featuring an insulator-to-metal transition15,16, which can be triggered by applying an external voltage: the material becomes conducting ('fires') if a threshold voltage is exceeded7,10-12. The dynamics of this induced transition have been thoroughly studied, and its underlying mechanism and characteristic time are well documented10,12,17,18. By contrast, there is little knowledge regarding the opposite transition: the process by which the system returns to the insulating state after the voltage is removed. Here we show that Mott nanodevices retain a memory of previous resistive switching events long after the insulating resistance has recovered. We demonstrate that, although the device returns to its insulating state within 50 to 150 ns, it is possible to re-trigger the insulator-to-metal transition by using subthreshold voltages for a much longer time (up to several milliseconds). We find that the intrinsic metastability of first-order phase transitions is the origin of this phenomenon, and so it is potentially present in all Mott systems. This effect constitutes a new type of volatile memory in Mott-based devices, with potential applications in resistive memories, solid-state frequency discriminators and neuromorphic circuits.
- 31Bassham, L. E.; Rukhin, A. L.; Soto, J.; Nechvatal, J. R.; Smid, M. E.; Barker, E. B.; Leigh, S. D.; Levenson, M.; Vangel, M.; Banks, D. L.; Heckert, N. A.; Dray, J. F.; Vo, S. A statistical test suite for random and pseudorandom number generators for cryptographic applications. NIST SP 800-22r1a; National Institute of Standards and Technology: Gaithersburg, MD, 2010.There is no corresponding record for this reference.
- 32Zhou, Y.; Chen, X.; Ko, C.; Yang, Z.; Mouli, C.; Ramanathan, S. Voltage-triggered ultrafast phase transition in vanadium dioxide switches. IEEE Electron Device Lett. 2013, 34, 220– 222, DOI: 10.1109/LED.2012.222945732https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXkt1GksLY%253D&md5=62b1d97cc25ce189e4f809603bc7d92fVoltage-triggered ultrafast phase transition in vanadium dioxide switchesZhou, You; Chen, Xiaonan; Ko, Changhyun; Yang, Zheng; Mouli, Chandra; Ramanathan, ShriramIEEE Electron Device Letters (2013), 34 (2), 220-222CODEN: EDLEDZ; ISSN:0741-3106. (Institute of Electrical and Electronics Engineers)Elec. driven metal-insulator transition (MIT) in vanadium dioxide (VO2) is of interest in emerging memory devices, neural computation, and high-speed electronics. We report on the fabrication of out-of-plane VO2 metal-insulator-metal structures and reproducible high-speed switching measurements in these two-terminal devices. We have obsd. a clear correlation between the elec. driven ON/OFF current ratio and the thermally induced resistance change during MIT. It is also found that sharp MIT could be triggered by the external voltage pulses within 2 ns at room temp. and the achieved ON/OFF ratio is greater than two orders of magnitude with good endurance.
- 33Brockman, J. S.; Gao, L.; Hughes, B.; Rettner, C. T.; Samant, M. G.; Roche, K. P.; Parkin, S. S. P. Subnanosecond incubation times for electric-field-induced metallization of a correlated electron oxide. Nat. Nanotechnol. 2014, 9, 453– 458, DOI: 10.1038/nnano.2014.7133https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXmsVKms70%253D&md5=e857866a2a3dba2fa7d6fd981b4e7658Subnanosecond incubation times for electric-field-induced metallization of a correlated electron oxideBrockman, Justin S.; Gao, Li; Hughes, Brian; Rettner, Charles T.; Samant, Mahesh G.; Roche, Kevin P.; Parkin, Stuart S. P.Nature Nanotechnology (2014), 9 (6), 453-458CODEN: NNAABX; ISSN:1748-3387. (Nature Publishing Group)Strong interactions, or correlations, between the d or f electrons in transition-metal oxides lead to various types of metal-insulator transitions that can be triggered by external parameters such as temp., pressure, doping, magnetic fields and elec. fields. Elec.-field-induced metalization of such materials from their insulating states could enable a new class of ultrafast electronic switches and latches. However, significant questions remain about the detailed nature of the switching process. Here, we show, in the canonical metal-to-insulator transition system V2O3, that ultrafast voltage pulses result in its metalization only after an incubation time that ranges from ∼150 ps to many nanoseconds, depending on the elec. field strength. We show that these incubation times can be accounted for by purely thermal effects and that intrinsic electronic-switching mechanisms may only be revealed using larger elec. fields at even shorter timescales.
- 34Cario, L.; Vaju, C.; Corraze, B.; Guiot, V.; Janod, E. Electric-field-induced resistive switching in a family of mott insulators: Towards a new class of RRAM memories. Adv. Mater. 2010, 22, 5193, DOI: 10.1002/adma.20100252134https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXhsFSgs7nL&md5=a350b8919397d8b9a4593f2908212076Electric-field-induced resistive switching in a family of Mott insulators: towards a new class of RRAM memoriesCario, Laurent; Vaju, Cristian; Corraze, Benoit; Guiot, Vincent; Janod, EtienneAdvanced Materials (Weinheim, Germany) (2010), 22 (45), 5193-5197CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)We have discovered that the fragile Mott insulator compds. AM4X8 exhibit both a volatile and a non-volatile unipolar RS. All our indicate that this RS is related neither to a chem. nor to an amorphous-cryst. phase change as obsd. in all other systems where a RS occurs. Conversly we have demonstrated that the volatile RS found in the Mott insulators AM4X8 is related to an elec.-field-driven IMT. These compds. could, therefore, offer the possibility to explore a new type of RRAM based on a new mechanism likely related to the Mott physics.
- 35Radu, I. P.; Govoreanu, B.; Mertens, S.; Shi, X.; Cantoro, M.; Schaekers, M.; Jurczak, M.; De Gendt, S.; Stesmans, A.; Kittl, J. A.; Heyns, M.; Martens, K. Switching mechanism in two-terminal vanadium dioxide devices. Nanotechnology 2015, 26, 165202, DOI: 10.1088/0957-4484/26/16/16520235https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BC2MnosFGqsA%253D%253D&md5=0ddb03ca089be8f620f467a3825d5a3fSwitching mechanism in two-terminal vanadium dioxide devicesRadu Iuliana P; Govoreanu B; Mertens S; Shi X; Cantoro M; Schaekers M; Jurczak M; De Gendt S; Stesmans A; Kittl J A; Heyns M; Martens KNanotechnology (2015), 26 (16), 165202 ISSN:.Two-terminal thin film VO2 devices show an abrupt decrease of resistance when the current or voltage applied exceeds a threshold value. This phenomenon is often described as a field-induced metal-insulator transition. We fabricate nano-scale devices with different electrode separations down to 100 nm and study how the dc switching voltage and current depend on device size and temperature. Our observations are consistent with a Joule heating mechanism governing the switching. Pulsed measurements show a switching time to the high resistance state of the order of one hundred nanoseconds, consistent with heat dissipation time. In spite of the Joule heating mechanism which is expected to induce device degradation, devices can be switched for more than 10(10) cycles making VO2 a promising material for nanoelectronic applications.
- 36Seo, G.; Kim, B. J.; Ko, C.; Cui, Y.; Lee, Y. W.; Shin, J. H.; Ramanathan, S.; Kim, H. T. Voltage-pulse-induced switching dynamics in VO2 thin-film devices on silicon. IEEE Electron Device Lett. 2011, 32, 1582, DOI: 10.1109/LED.2011.216392236https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3MXhs1emu73F&md5=2589e0094dfa797314b669e1a3162f7eVoltage-pulse-induced switching dynamics in VO2 thin-film devices on siliconSeo, Giwan; Kim, Bong-Jun; Ko, Changhyun; Cui, Yanjie; Lee, Yong Wook; Shin, Jun-Hwan; Ramanathan, Shriram; Kim, Hyun-TakIEEE Electron Device Letters (2011), 32 (11), 1582-1584CODEN: EDLEDZ; ISSN:0741-3106. (Institute of Electrical and Electronics Engineers)We demonstrate the characteristics of voltage-pulse-induced out-of-plane switching driven by metal-insulator transition (MIT jump) with VO2 thin-film devices fabricated on silicon. As the peak of a triangular pulse increases, the delay time from the insulating state to the metallic state linearly decreases and is independent of change in external resistance. The intrinsic rising time is less than 170 ns. An endurance test with continuous voltage pulse shows reliability without breakdown for more than 110 h. This work contributes to correlated oxide electronics utilizing phase transition layers.
Supporting Information
Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.nanolett.1c04404.
Details on sample fabrication, randomness evaluation using the NIST suite, performance for different pulse periods (Figures S1–S10) and performance with the latch circuit (Figure S11) (PDF)
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.