Method and System For Generating a Photo-Response From MoS2 Schottky Junctions
Devices incorporating a single to a few-layer MoS2 channels in combination with optimized substrate, dielectric, contact and electrode materials and configurations thereof, exhibit light emission, photoelectric effect, and superconductivity, respectively.
Latest Georgetown University Patents:
- Targeted liposomes
- USP13 INHIBITORS AND METHODS OF USE THEREOF
- Methods of identifying novel proteins and antigens in cancer cells
- ENHANCING CHEMOTHERAPY IN MEDULLOBLASTOMA AND GLIOBLASTOMA WITH HIGH BASAL p53 LEVELS
- Enhancing chemotherapy in medulloblastoma and glioblastoma with high basal p53 levels
The present application is a continuation of U.S. application Ser. No. 13/661,979, filed Oct. 26, 2012, entitled “METHOD AND SYSTEM FOR GENERATING A PHOTO-RESPONSE FROM MoS2 SCHOTTKY JUNCTIONS,” which claims benefit of priority to. U.S. Provisional Patent Application Nos. 61/552,914, filed Oct. 28, 2011, entitled “ELECTRONIC AND OPTOELECTRONIC DEVICES USING SINGLE LAYER MoS2,” and 61/616,202, filed Mar. 27, 2012, entitled “METHOD AND SYSTEM FOR GENERATING A PHOTO-RESPONSE FROM MoS2 SCHOTTKY JUNCTIONS,” all of which are incorporated herein by reference.
BACKGROUND1. Field of the Embodiments
The embodiments described herein are generally directed to components incorporating molybdenum disulfide (MoS2).
2. Description of the Related Art
MoS2 is a semiconductor made of layers that are weakly coupled by van der Waals forces and can be easily separated via chemical or mechanical exfoliation to obtain few-layer or single-layer samples. A layer is made of three atomic planes that are strongly bonded to each other: two hexagonal planes of sulfur atoms separated by one hexagonal plane of molybdenum atoms, with each molybdenum atom covalently bonded to six sulfur atoms in the adjacent planes (see
Bulk MoS2 is a semiconductor with an indirect bandgap of 1.2 eV. When the thickness is reduced to a few layers, the indirect bandgap is tuned by quantum confinement and increases by 0.5 eV or more, until it eventually exceeds the energy spacing of the direct gap for single-layer thickness, at about 1.9 eV. When varying the thickness, photoluminescence experiments have shown that the quantum yield increases by a few orders of magnitude for single-layer samples, confirming the crossover from indirect to direct gap. For a single layer, the quantum yield has also been found to be higher when the layer is suspended.
These results make single-layer MoS2 a possible candidate for optoelectronics applications, as well as for other device applications where a large energy gap, a large surface area, and a high surface-to-volume ratio are crucial. Presently, the technological potential of this material still remains largely unexplored. Single-layer and few-layer transistors have been only recently realized by depositing thin flakes of MoS2 on a doped silicon substrate capped with an insulating silicon dioxide layer. The MoS2 thin flakes are first deposited using the same method that yielded the first successful single-layer graphene devices, i.e. mechanical exfoliation with scotch tape, then attachment of gold source and drain electrodes across the MoS2 layer. The number of layers can be roughly sorted by contrast imaging with an optical microscope, whereas accurate measurements are obtained either by Raman spectroscopy or by AFM measurements of the thickness, which is about 0.65 nm for a single layer. In some devices, top gates with a different material for gate dielectric are also deposited (see
There is a need in the art to advantageously implement MoS2 in various processes and devices to improvements in light emitting devices, photoelectric devices, superconductivity and the like.
SUMMARYIn a first exemplary embodiment, a solid-state device includes: a channel formed of a single-layer of MoS2 on a substrate; and a source contact formed of at least a first material and a drain contact formed of at least a second material on the substrate and in contact with first and second ends of the channel, wherein Schottky junctions are formed at an interface of at least one of the source and drain contacts and the channel exhibiting a photovoltaic effect.
In a second exemplary embodiment a solid-state light emitting device includes: a channel formed of a single-layer of MoS2 on a substrate; a source contact formed of at least a first material and a drain contact formed of at least a second material on the substrate and in contact with the channel; a dielectric layer formed on the single-layer of MoS2 and the source and drain contacts; and first and second gate electrodes formed on the dielectric layer.
The following figures exemplify the embodiments described herein and are intended to be reviewed in combination with the detailed descriptions provided below.
Ambipolar MoS2 transistors with Au source-drain contacts were very recently demonstrated, but they were gated with ionic liquids (electrolytic gate) and the mechanism that leads to ambipolar transport in these samples is still unclear. By contrast, embodiments described and optimized through process discussed herein are all-solid-state devices that show either n-type or p-type transistor behavior, depending on the, choice of electrode material. It has been found that when source and drain contacts are made of palladium (Pd), MoS2 transistors always show p-type behavior. Devices made with two different contacts, Au and Pd, show asymmetric ambipolar behavior and diode characteristics in a wide range of gate voltages. These results can be explained assuming different work functions for Pd, MoS2 and Au. The sizable photovoltaic effect measured within these devices is thought to arise from the built-in potential of the space charge accumulated at the source-drain contacts.
In a first embodiment, samples are made of exfoliated multi- and few-layer flakes of MoS2 (about 50 nm thick) on a doped Si substrate capped with 300 nm SiO2. Additional substrate materials may include glass, as well as flexible materials discussed below. Source and drain electrodes are patterned by e-beam lithography and deposited by sputtering.
The reversed polarity is explained by assuming that 1) the thin adhesion layers of Cr and Nb can be neglected, since their thickness is too small to obtain a uniform film (that is, the adhesion layer is made of small disconnected islands or patches of metal) and the contact to the source and drain electrode is, mainly made by the thick Pd and Au layers and 2) (φPd>φMoS2>φAu, as shown schematically in
In the case of electron-doping Au contacts (
Although the literature values for the Au and Pd work functions are very similar, chemical interactions and charge redistribution at the interfaces change the alignment of Fermi levels for the. Au/MoS2 and the Pd/MoS2 interfaces consistently with the scenario sketched in
A Kelvin probe measurement tool (KP Technology—Wick, Scotland) uses the principle of a vibrating capacitor to determine work function of a metal, φ. The technique determines the contact potential difference (CPD) between two dissimilar metals that are electrically connected to each other. With the two metals initially unconnected but under the same ambient condition (
In the actual Kelvin probe tool, one of the metals forms a suspended probe above the second sample metal under test, effectively forming a variable capacitor. The probe is actuated at a set frequency that generates a sinusoidal variation in current response, thereby allowing the capacitor to go through a charge/discharge cycle. The counterbalancing DC voltage is ramped from a negative to positive potential, and by plotting the IV curve, a zero value for AC current can be interpolated to determine the CPD.
The Kelvin probe system is used to measure the work function of both gold and palladium (Pd) with respect to the stainless steel probe in two different states: one is to perform the measurement in air immediately after sputter deposition, while the second is after both samples have been in air for 8 days (
The effect of the electrode material on MoS2 was determined by using two Pd electrodes and one Au electrode on the same MoS2 flake, as shown in
These results may be explained considering that the built-in Schottky potential at the Au/MoS2 and the Pd/MoS2 interfaces separates the electrons and holes generated by the absorbed photons and accumulates them at the electrodes, yielding an open-circuit voltage VOC=0.1 V, as shown by the laser ON curve in
In accordance with the embodiments described in U.S. Provisional Patent Application Ser. No. 61/616,202, it is estimated that the maximum electrical power that can be extracted from the device is about 2.5% of the laser power incident on the MoS2 region between the electrodes. This percentage is a lower bound for the conversion efficiency, because only the depletion region in the MoS2 channel is photoactive. Nevertheless, this room-temperature efficiency is better than the highest value previously reported in MoS2, which was about 1% at 120 K for bulk samples.
The following variations to the devices described and illustrated herein are expected to enhance the photovoltaic effect and improve device characteristics.
In a first variation, the device as illustrated in
In a second variation, the device as illustrated in
In a third variation, the device as illustrated in
In a fourth variation, the device as illustrated in
Returning to the single layer embodiment,
In
It is expected that light can be detected from the resulting pn junction. Forward biasing of this pn junction leads to electron-hole recombination in the junction region and emission of photons with energy matching the single-layer MoS2 energy gap, 1.9 eV. Furthermore, radiative recombination of electron-hole pairs should be favored by the strong excitonic effects predicted in single-layer MoS2. This light is in the visible portion of the electromagnetic spectrum and will be emitted from the gap separating the two top electrodes. The detector of a Raman Renishaw spectrometer may be used to measure the emitted light and its spectrum.
This type of light emitting pn junction, where the pn regions are obtained by electrostatic doping, has been already successfully used for carbon nanotubes. MoS2 can be adapted to batch fabrication techniques much more easily than carbon nanotubes, which are not at all amenable to mass production. Accordingly, although to date MoS2 devices are currently fabricated by mechanical exfoliation, it is anticipated that similar to graphene, methods to grow single layer MoS2 on wafers will be developed.
Successful detection of light from MoS2 pn junctions is extremely attractive for both applications as well as basic science. For example, shrinking the size of the light source by confining electrons and holes may lead to new types of lasers with no threshold and no need of high Q cavity. High confinement can be obtained by forming pn junctions from MoS2 nanoribbons, so this type of device can be used to build nanoscale light sources and to study possible occurrence of coherent emission due to confinement of electrons and holes. Nanoscale lasers have been recently demonstrated using surface plasmon oscillations, as a first step towards the development of on-chip optical computing. However, nanolasers achieved to date can only be pumped optically, whereas on-chip light sources for optical interconnects require electrical pumping, making MoS2 light sources extremely desirable. In addition, arrays of light sources could be easily fabricated from MoS2 layers.
There are still many open basic questions on transport properties of MoS2, including the origin of the large hysteresis observed in most devices and whether it can be controlled or reduced, the contribution to the resistance due to the contact versus the scattering in the MoS2 layers and the dependence of these contributions on the number of layers. Regarding scattering within few-layers or single-layer MoS2, it is not clear whether ballistic transport can be achieved in these systems and what is the role played by the substrate and therefore the roughness of the gate oxide. However, zero resistance due to superconductivity has been measured in MoS2 intercalated with sodium and potassium yielding critical temperatures of a few. Kelvin. Recently superconductivity has also been measured in thin MoS2 flakes doped with an electrolytic gate, with critical temperatures up to 9 K. The authors find that superconductivity occurs only in a narrow region of gate voltage where the Fermi energy is shifted within the conduction band. These measurements suggest that the onset of superconductivity is related to the increase in charge carrier density. Superconductivity due to hole-doping, however, has not been reported. The following systems and processes are implemented to determine the transport properties of MoS2 flakes, including the onset of field-effect superconductivity and its dependence on the doping, the number of layers, and quantum confinement obtained by cutting MoS2 flakes into nanoribbons.
Initially, for both single- and few-layer MoS2 flakes, electrodes are deposited with increasing spacing on the same flake to extract the contact resistance as well as the sheet resistance of the flake and its temperature dependence. This process is repeated for different electrode materials to find the material that provides the best contact to the flakes.
The influence of the substrate is determined by measuring the sheet resistance of flakes on different dielectric substrates deposited by atomic layer deposition, to investigate the effect of electron-phonon scattering and of the substrate on transport properties. Different dielectric layers are deposited by atomic layer deposition, including HfO2 and Al2O3, and results studied. More particularly, characterization of the roughness of the layer and its effect on the MoS2 transport properties is determined.
The onset of superconductivity in gated MoS2 devices is determined by measuring the resistance as a function of temperature. The source-drain electrode material determined to provide the best contact to the MoS2 is utilized. In the case of Pd electrodes, fully solid state devices are achieved where hole transport can be measured. These devices are cooled to low temperature (2.5 K) and the gate is continuously scanned to shift the Fermi energy in the valence band and detect superconductivity in the hole-doping regions. This allows for differential study of the gate dependence of the superconducting properties when compared to samples with prior art electrolytic gate. This is because the electrolytic gate freezes at the low temperatures required to measure superconductivity, therefore the sample needs to be warmed up in order to effectively change the gate and the (electrically induced) doping, making a fine tuning of the gate very time consuming when using an electrolytic gate.
Additionally, superconductivity in suspended and gated MoS2 flakes is determined. In suspended structures, electrons may couple to different phonon modes, yielding a different critical temperature.
Further, the MoS2 flakes are cut in the shape of nanoribbons, to obtain a density of states with one-dimensional subbands and sharp van Hove singularities at the onset of each subband. The doping dependence of the superconducting properties is expected to be substantially different and yield higher critical temperatures. This is expected because the density of states increases dramatically when the Fermi energy is shifted into a van Hove singularity by applying a gate voltage to a nearby gate electrode. The nanoribbons are patterned by e-beam lithography and masking a narrow stripe of MoS2 with PMMA prior to dry etching.
In recent prior art investigations of superconductivity in isolated carbon nanotubes in the field effect transistor configuration, an applied gate voltage can be used to shift the Fermi energy into van Hove singularities of the density of states, yielding transport features that can be interpreted as signatures of superconductivity, with critical temperature higher than 25 K for some samples. In single carbon nanotube devices, Andreev reflection at the contacts as well as thermal and quantum fluctuations strongly affect the temperature dependence of the resistance and cause a resistive tail below the critical temperature. The transition is expected to get much sharper when multiple nanotubes are coupled together. However, since the spacing of van Hove singularities depends on the carbon nanotube diameter and the nanotube diameter cannot be easily controlled, when samples are made with multiple nanotubes it is very difficult to shift the Fermi energy into van Hove singularities for all the nanotubes and optimize the critical temperature. As a result, reported critical temperatures of samples with multiple nanotubes have always much lower than 25 K, but the superconducting transition was sharper. It has been argued that ropes are different from a system of uncoupled 1D conductors, because Cooper pairs can hop even between tubes of different helicity. In the case of few-layer MoS2 nanoribbons, each layer can be considered as a 1D conductor coupled to the adjacent layers. These layers will have the same width, therefore the same spacing of van Hove singularities. Accordingly, it is possible to shift the Fermi energy into van Hove singularities for all the layers with a common gate voltage and achieve critical temperature higher than the 9 K critical temperature recently reported for electrolytic gated MoS2.
Processes described herein include optimization of devices for achieving superconductivity in single and/or few-layer MoS2 devices through investigation of interface properties and charge scattering mechanisms, considering both electron and hole conduction. More particularly, investigative processes seek to improve electron mobilities and determine how to achieve and enhance hole conduction. In the prior art, for n-type conduction in single-layer MoS2, mobilities of less than 10 cm2V−1s−1 were originally reported. The enhancement to 200 cm2V−1s−1 in the presence of the HfO2 top gate as reported in the prior art is likely due to a reduction in impurity scattering by dielectric screening, though substrate-induced changes in the phonon properties could also reduce scattering of charge carriers.
Recent prior art density-functional-theory (DFT) calculations of the intrinsic electron-phonon scattering mechanisms in n-type single-layer MoS2 report an estimate of about 400 cm2V−1s−1 as the upper bound on the mobility. Present process embodiments assess mobility in the valence band to see whether the intrinsic limits on hole transport are more stringent. While the effective masses at the conduction band minimum and the valence band maximum in single-layer MoS2 are calculated to be similar (see
The mobility optimization in nanoscale systems contemplated the choice of contact material as well as the channel material in determining the electronic and transport properties of the system. Recent prior art computational studies of the electronic structure of single-layer MoS2 in contact with various metals suggests that the most commonly used contact material, gold, may not be optimal, and that this may explain, in part, why conduction-band mobilities have not been as large as expected. Titanium was suggested as a better choice, given that the MoS2/Ti interface has a high density of delocalized states at the Fermi energy. Optimization hereby includes determining compatibility in terms of atomic structure as well as electronic structure at the interface, and also determining Schottky barriers. An optimal contact material is identified to enhance p-type transport and to optimize the built-in potential necessary for separating charge in the photovoltaic effect.
With regard to attaining superconductivity in MoS2, an initial step is examination of the phonon spectrum calculated for a single layer of MoS2 as shown in
Next, the process looks to determine whether the strong coupling, of the A1g mode to electrons in the conduction band can also explain the superconductivity observed in electrolytically gated devices made of thin MoS2 flakes.
Current results show strong contributions to the electron-phonon coupling parameter from intervalley phonon wave vectors Q=M and Q′=K. While the optical A1g mode gives a large contribution at the zone center, corresponding to intrasheet coupling (Q=Γ), for Q=M and K, acoustic modes involving in-plane Mo motion and in-plane or out-of-plane S motion provide the largest contributions to the coupling constant. Although the coupling constant λQ is large at and near these wave vectors, the phase space over which the coupling is non-zero is limited by the small size of the Fermi sheets. Nevertheless, for electron doping of 0.06 electrons per formula unit, we estimate a total coupling constant of λ˜0.9. This is consistent with a superconducting Tc on the order of a few K. For comparison, the, potassium intercalated bulk material KxMoS2 has a peak Tc of about 7 K when x=0.4, while electron-doped liquid-gated few-layer MoS2 devices have been reported to show superconductivity around 9 K. The carrier density in the latter case is still unclear.
Since the location and character of the hole sheets of the Fermi surface are very different from those of the electron sheets, it is expected that different phonon modes will be important for superconductivity in the hole-doped case. In both the hole-doped and electron-doped cases, optimization includes the examination of the effect of carrier concentration, strain, number of layers, interlayer interactions, and further reduction of dimensionality (i.e., nanoribbons) on electronic structure, vibrational properties, electron-phonon coupling, and superconductivity. For example, in the electron-doped case, when the doping level is sufficient to start occupying the ½ (ΓK) valleys, there should be a significant enhancement in superconductivity. Implementing methods, to preferentially increase the density of states in these valleys, and to increase the size of the associated Fermi sheets, enlarges the phase space of phonons that couple these sheets. To make contact with the proposed nanoribbon, structures that are neither perfect armchair nor perfect zigzag ribbons may be considered.
One skilled in the art recognizes that there are variations to the configurations and optimization processes discussed herein. Those variations are considered to be within the skill in the art and the scope, of the invention.
Claims
1. A process for forming a solid-state device including at least one channel capable of exhibiting a photovoltaic effect comprising:
- forming a channel comprising a single-layer of MoS2 on a substrate; and
- forming source contact comprised of at least a first material and a drain contact formed of at least a second material on the substrate and in contact with first and second ends of the channel, wherein Schottky junctions are formed at an interface of at least one of the source and drain contacts and the channel exhibiting a photovoltaic effect and further wherein the at least a first material and the at least a second material are different materials.
2. The process of claim 1, wherein forming the source contact further includes forming a first adhesion layer on the substrate and depositing the at least a first material on the first adhesion layer.
3. The process of claim 2, wherein forming the drain contact further includes forming a second adhesion layer on the substrate and depositing the at least one second material on the second adhesion layer.
4. The process of claim 2, wherein the first material is palladium and the first adhesion layer is niobium.
5. The process of claim 3, wherein the second material is gold and the second adhesion layer is chromium or titanium.
6. The process of claim 1, wherein the substrate is formed of a flexible polymer.
7. The process of claim 1, further comprising forming a reflective material on the substrate prior to forming the single-layer of MoS2 on the substrate, wherein the reflective material is formed between the single-layer of MoS2 and the substrate.
8. The process of claim 1, further comprising forming the source contact and the drain contact such that a distance therebetween is shorter than an electron-hole recombination length of the device.
9. The process of claim 1, further comprising forming multiple individual photovoltaic devices on the substrate, each of the multiple individual photovoltaic devices including a single-layer of MoS2 and a single source contact and a single drain contact.
10. The process of claim 9, further comprising connecting the multiple individual photovoltaic devices in series.
11. The process of claim 9, further comprising connecting the multiple individual photovoltaic devices in parallel.
12. The process of claim 9, further comprising connecting a first subgroup of the multiple individual photovoltaic devices in parallel and a second subgroup of the multiple individual photovoltaic devices in series.
13. The process of claim 1, wherein forming the channel comprises exfoliating flakes of MoS2 onto doped silicon (Si).
14. A process for forming solid-state light emitting device comprising:
- forming a channel, comprising a single-layer of MoS2 on a substrate;
- forming a source contact of at least a first material and a drain contact of at least a second material on the substrate and in contact with the channel, wherein the first material and the second material are different materials;
- forming a dielectric layer on the single-layer of MoS2 and the source and drain contacts; and
- forming first and second gate electrodes on the dielectric layer.
15. The process of claim 14, wherein forming the source contact further includes forming a first adhesion layer on the substrate and depositing the at least a first material on the first adhesion layer.
16. The process of claim 15, wherein forming the drain contact further includes forming a second adhesion layer on the substrate and depositing the at least one second material on the second adhesion layer.
17. The process of claim 15, wherein the first material is palladium and the first adhesion layer is niobium.
18. The process of claim 16, wherein the second material is gold and the second adhesion layer is chromium or titanium.
19. The process of claim 14, wherein the dielectric material is formed to a thickness in the range of 5 Angstroms to 100 nanometers.
20. The process of claim 14, wherein forming the channel comprises exfoliating flakes of MoS2 onto doped silicon (Si).
21. The process of claim 14, further comprising biasing the first and second gate electrodes with opposite voltages.
22. A nanoscale light source comprising, on a substrate:
- a single layer of MoS2;
- a source electrode formed of a first material and located at a first end of the single layer of MoS2;
- a drain electrode formed of a second material, different from the first material, and located at a second end of the single layer of MoS2, wherein a gap is formed in the single layer of MoS2 between the source electrode and the drain electrode;
- a dielectric layer formed on the single layer of MoS2, the source electrode and the drain electrode; and
- first and second gate electrodes formed on the dielectric layer and biased with voltages of opposite sign, wherein a pn junction is formed in the gap and light is emitted therefrom in response to biasing the pn junction.
23. The light source of claim 22, wherein the source contact further includes a first adhesion layer.
24. The light source of claim 23, wherein the drain contact further includes a second adhesion layer.
25. The light source of claim 24, wherein the first material is palladium and the first adhesion layer is niobium.
26. The light source of claim 25, wherein the second material is gold and the second adhesion layer is chromium or titanium.
27. The light source of claim 26, wherein the dielectric material has a thickness in the range of 5 Angstroms to 100 nanometers.
28. The device of claim 27, wherein the substrate is doped silicon.
29. A nanoscale light source comprising, on a substrate:
- a nanoribbon formed of a single layer of MoS2; and
- a pn junction formed in the nanoribbon via electrostatic doping, wherein light is emitted therefrom in response to biasing the pn junction.
30. The nanoscale light source of claim 29, wherein the pn junction is formed in a gap between a source electrode and a drain electrode formed on the nanoribbon.
31. The nanoscale light source of claim 30, wherein the source electrode and the drain electrode are formed of different materials.
32. The nanoscale light source of claim 31, wherein the different materials are selected from the group consisting of gold, palladium and titanium.
Type: Application
Filed: Jun 4, 2014
Publication Date: Feb 19, 2015
Patent Grant number: 9431573
Applicant: Georgetown University (Washington, DC)
Inventors: Makarand PARANJAPE (Silver Spring, MD), Paola BARBARA (Washington, DC), Amy LIU (Alexandria, VA), Marcio FONTANA (Salvador)
Application Number: 14/295,592
International Classification: H01L 33/00 (20060101); H01L 31/07 (20060101); H01L 31/18 (20060101);