CoChat
Sign Up
Cited by 29Advanced Memory and Neural ComputingFerroelectric and Negative Capacitance DevicesSemiconductor materials and devices

The DOI matches its Crossref record; 19 of the 19 references it cites were found in a registry or an index.

Scalable Memdiodes Exhibiting Rectification and Hysteresis for Neuromorphic Computing

Open access
Shank, Joshua C., Tellekamp, M. Brooks, Wahila, Matthew J., Howard, Sebastian, Weidenbach, Alex S., Zivasatienraj, Bill, Piper, Louis F. J., Doolittle, W. Alan·2018·Scientific Reports
DOI 10.1038/s41598-018-30727-9PMC PMC6113211
Open this paper in CoChat — free accountOpen PDF

Abstract

Metal-Nb 2 O 5−x-metal memdiodes exhibiting rectification, hysteresis, and capacitance are demonstrated for applications in neuromorphic circuitry. These devices do not require any post-fabrication treatments such as filament creation by electroforming that would impede circuit scalability. Instead these devices operate due to Poole-Frenkel defect controlled transport where the high defect density is inherent to the Nb 2 O 5−x deposition rather than post-fabrication treatments. Temperature dependent measurements reveal that the dominant trap energy is 0.22 eV suggesting it results from the oxygen deficiencies in the amorphous Nb 2 O 5−x. Rectification occurs due to a transition from thermionic emission to tunneling current and is present even in thick devices (>100 nm) due to charge trapping which controls the tunneling distance. The turn-on voltage is linearly proportional to the Schottky barrier height and, in contrast to traditional metal-insulator-metal diodes, is logarithmically proportional to the device thickness. Hysteresis in the I–V curve occurs due to the current limited filling of traps.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41598-018-30727-9

Cites

19 of 19 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Current-Voltage and Current-Time characteristics of metal-Nb2O5−x-metal memdiodes. (a) Current-Voltage curve for a metal-Nb2O5−x-metal device showing rectification and hysteresis. (b) Current transients at several voltage steps in (a) normalized to the first data point showing a decreasing current below the turn-on voltage of 1 volt and increasing current above the turn-on voltage. (c) An extended current transient above the turn-on voltage showing that the time scale to reach steady state is hundreds of seconds.
Figure 1 Current-Voltage and Current-Time characteristics of metal-Nb2O5−x-metal memdiodes. (a) Current-Voltage curve for a metal-Nb2O5−x-metal device showing rectification and hysteresis. (b) Current transients at several voltage steps in (a) normalized to the first data point showing a decreasing current below the turn-on voltage of 1 volt and increasing current above the turn-on voltage. (c) An extended current transient above the turn-on voltage showing that the time scale to reach steady state is hundreds of seconds.CC BY · PubMed Central
Figure 2 Capacitance characteristics of metal-Nb2O5−x-metal memdiodes. (a) The capacitance and conductance as a function of frequency. The capacitance is highly dependent on frequency with a relative dielectric constant greater than 180 at neurological frequencies (<100 Hz). (b) The capacitance and conductance as a function of DC bias voltage measured at 100 Hz. The capacitance is relatively unchanging below the turn-on voltage, increasing by only 6%. The slight increase in capacitance is likely due to image-force lowering decreasing the effective insulator thickness. Above the turn-on voltage the capacitance decreases and the conductance exhibits rectification.
Figure 2 Capacitance characteristics of metal-Nb2O5−x-metal memdiodes. (a) The capacitance and conductance as a function of frequency. The capacitance is highly dependent on frequency with a relative dielectric constant greater than 180 at neurological frequencies (<100 Hz). (b) The capacitance and conductance as a function of DC bias voltage measured at 100 Hz. The capacitance is relatively unchanging below the turn-on voltage, increasing by only 6%. The slight increase in capacitance is likely due to image-force lowering decreasing the effective insulator thickness. Above the turn-on voltage the capacitance decreases and the conductance exhibits rectification.CC BY · PubMed Central
Figure 3 Linear and logarithmic scaling of the memdiode transition voltage. (a) Semi-log plot of the current-voltage sweep defining a transition voltage between the off and on states. (b) The transition voltage is linearly proportional to the ideal metal work function indicating that the transition is tied to the Schottky barrier height. (c) The transition voltage is logarithmically related to the oxide thickness in contrast to traditional MIM diodes.
Figure 3 Linear and logarithmic scaling of the memdiode transition voltage. (a) Semi-log plot of the current-voltage sweep defining a transition voltage between the off and on states. (b) The transition voltage is linearly proportional to the ideal metal work function indicating that the transition is tied to the Schottky barrier height. (c) The transition voltage is logarithmically related to the oxide thickness in contrast to traditional MIM diodes.CC BY · PubMed Central
Figure 4 Charge density maps and band diagrams below and above turn-on. The charge density and electron potential of the proposed device model both (a) below and (b) above the turn-on voltage. Removal of electrons from traps in Region 1 leaves fixed positive charge that supports a large electric field (106 V/cm). The steady state width of Region 1 is determined by balancing the injection of electrons by thermionic emission below the turn-on voltage or tunneling current above the turn-on voltage with the removal of electrons by Poole-Frenkel current through Region 2. (a) Below the turn-on voltage a small electric field (103 V/cm) is required in Region 2 to remove the electrons injected by thermionic emission. (b) Above the turn-on voltage a larger electric field (105 V/cm) is required in Region 2 to remove electrons injected by tunneling.
Figure 4 Charge density maps and band diagrams below and above turn-on. The charge density and electron potential of the proposed device model both (a) below and (b) above the turn-on voltage. Removal of electrons from traps in Region 1 leaves fixed positive charge that supports a large electric field (106 V/cm). The steady state width of Region 1 is determined by balancing the injection of electrons by thermionic emission below the turn-on voltage or tunneling current above the turn-on voltage with the removal of electrons by Poole-Frenkel current through Region 2. (a) Below the turn-on voltage a small electric field (103 V/cm) is required in Region 2 to remove the electrons injected by thermionic emission. (b) Above the turn-on voltage a larger electric field (105 V/cm) is required in Region 2 to remove electrons injected by tunneling.CC BY · PubMed Central
Figure 5 Poole-Frenkel current characterization. A Poole-Frenkel plot exhibiting a linear region above the turn-on voltage indicating that the Poole-Frenkel mechanism controls the current flow within this voltage range. The coefficient of determination (R2) is 0.998 indicating good agreement between the experimental measurements and the linear fit.
Figure 5 Poole-Frenkel current characterization. A Poole-Frenkel plot exhibiting a linear region above the turn-on voltage indicating that the Poole-Frenkel mechanism controls the current flow within this voltage range. The coefficient of determination (R2) is 0.998 indicating good agreement between the experimental measurements and the linear fit.CC BY · PubMed Central
Figure 6 Temperature dependent Current-Voltage characteristics show a 0.22 eV trap activation energy. (a) Temperature dependent measurements of the device current showing rectification with temperature. The current was measured at 1.25 V, which is above the turn-on voltage. At low temperatures the Poole-Frenkel current controls the current flow with a low activation energy while at high temperatures thermionic emission over a higher energy barrier controls current flow. (b) Plotting the Poole-Frenkel activation energy at a variety of applied voltages reveals a defect trap depth of 0.22 eV ± 0.01 eV matching the previously reported trap depth for oxygen vacancies in Nb2O5.
Figure 6 Temperature dependent Current-Voltage characteristics show a 0.22 eV trap activation energy. (a) Temperature dependent measurements of the device current showing rectification with temperature. The current was measured at 1.25 V, which is above the turn-on voltage. At low temperatures the Poole-Frenkel current controls the current flow with a low activation energy while at high temperatures thermionic emission over a higher energy barrier controls current flow. (b) Plotting the Poole-Frenkel activation energy at a variety of applied voltages reveals a defect trap depth of 0.22 eV ± 0.01 eV matching the previously reported trap depth for oxygen vacancies in Nb2O5.CC BY · PubMed Central
Figure 7 Average film compositions from XPS. (a) An example O 1 s core region from XPS showing the two peaks used during peak fitting as a part of the compositional analysis. (b) Nb:O film ratios for 30, 40, and 50 W depositions determined from XPS measurements, both including and excluding surface contamination related oxygen in the ratio.
Figure 7 Average film compositions from XPS. (a) An example O 1 s core region from XPS showing the two peaks used during peak fitting as a part of the compositional analysis. (b) Nb:O film ratios for 30, 40, and 50 W depositions determined from XPS measurements, both including and excluding surface contamination related oxygen in the ratio.CC BY · PubMed Central
Figure 8 The role of oxygen stoichiometry in mid-gap state creation. HAXPES spectra of the core levels and valence band show an increase in mid-gap states at the Fermi Level with decreases in bulk oxygen content. The inset shows the change in mid-gap states, with the estimated valence band maxima (gray dotted line) shown near 3 eV indicating the mid-gap states exist throughout the ~3.15 eV band gap.
Figure 8 The role of oxygen stoichiometry in mid-gap state creation. HAXPES spectra of the core levels and valence band show an increase in mid-gap states at the Fermi Level with decreases in bulk oxygen content. The inset shows the change in mid-gap states, with the estimated valence band maxima (gray dotted line) shown near 3 eV indicating the mid-gap states exist throughout the ~3.15 eV band gap.CC BY · PubMed Central
Figure 9 The role of oxygen stoichiometry in device performance. Films with higher oxygen stoichiometry turn on in a less defined manner, and the current above turn-on is approximately 10× lower than the lower oxygen stoichiometry films analyzed in the preceding sections.
Figure 9 The role of oxygen stoichiometry in device performance. Films with higher oxygen stoichiometry turn on in a less defined manner, and the current above turn-on is approximately 10× lower than the lower oxygen stoichiometry films analyzed in the preceding sections.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 19 references

Cited by

Sorted by citations

Related

See all

Metadata

Authors
Shank, J. C., Tellekamp, M. B., Wahila, M. J., Howard, S., Weidenbach, A. S., Zivasatienraj, B., Piper, L. F. J., & Doolittle, W. A.
Affiliations
  • Georgia Institute of Technology (US)
  • Binghamton University (US)
Corresponding author
Doolittle, W. Alan
ORCID
Venue
Scientific Reports
Publisher
Springer Science and Business Media LLC
Published
2018-08-28
Volume
8
Issue
1
Pages
12935-12935
ISSN
2045-2322 (ISSN-L)
DOI
10.1038/s41598-018-30727-9
PMC
PMC6113211
OpenAlex
W2888632952
Cited by
29
References
19
Citations by year
2026: 3 · 2025: 5 · 2024: 2 · 2023: 4 · 2022: 6 · 2021: 3 · 2019: 6
FWCI
0.86
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, other-oa)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • Diamond Light Source
  • Defense Threat Reduction Agency (HDTRA1-12-1-0031)
  • Air Force Office of Scientific Research (FA9550-, FA9550-18–1-0024)
Topics
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Ferroelectric and Negative Capacitance Devices
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Semiconductor materials and devices
Keywords
memristors, Nb2O5, metal-insulator-metal devices, rectification, hysteresis, charge trapping, Poole-Frenkel effect, oxygen vacancies, thermionic emission, tunneling current, Schottky barrier height, neuromorphic computing, amorphous metal oxides

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review