ROOM TEMPERATURE NANOWIRE IR, VISIBLE AND UV PHOTODETECTORS
1. A photodetector comprising:
- an array of semiconductor nanowires either standing vertically on a conducting substrate and capped by a conductor transparent to light, or lying horizontally on a non-conducting substrate and capped by conductors at both ends; and
wire leads attached to the conductors and conducting substrate for connection to an electrical circuit.
Room temperature IR and UV photodetectors are provided by electrochemical self-assembly of nanowires. The detectivity of such IR detectors is up to ten times better than the state of the art. Broad peaks are observed in the room temperature absorption spectra of 10-nm diameter nanowires of CdSe and ZnS at photon energies close to the bandgap energy, indicating that the detectors are frequency selective and preferably detect light of specific frequencies. Provided is a photodetector comprising: an aluminum substrate; a layer of insulator disposed on the aluminum substrate and comprising an array of columnar pores; a plurality of semiconductor nanowires disposed within the pores and standing vertically relative to the aluminum substrate; a layer of nickel disposed in operable communication with one or more of the semiconductor nanowires; and wire leads in operable communication with the aluminum substrate and the layer of nickel for connection with an electrical circuit.
View as Search Results
|Opticoanalytical Devices With Capacitance-Based Nanomaterial Detectors|
Patent #US 20170268926A1
Current AssigneeHalliburton Energy Services Incorporated
Sponsoring EntityHalliburton Energy Services Incorporated
|Opticoanalytical devices with capacitance-based nanomaterial detectors|
Patent #US 10,260,946 B2
Current AssigneeHalliburton Energy Services Incorporated
Sponsoring EntityHalliburton Energy Services Incorporated
|Deformable paper origami optoelectronic devices|
Patent #US 10,461,204 B2
Current AssigneeKing Abdullah University of Science and Technology
Sponsoring EntityKing Abdullah University of Science and Technology
- 1. A photodetector comprising:
an array of semiconductor nanowires either standing vertically on a conducting substrate and capped by a conductor transparent to light, or lying horizontally on a non-conducting substrate and capped by conductors at both ends; and wire leads attached to the conductors and conducting substrate for connection to an electrical circuit.
- View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
- 21. A method of making a photodetector comprising:
providing an aluminum foil with a thickness between 0-1 mm; degreasing and electropolishing the foil to reduce surface roughness; optionally rinsing the foil in distilled water, anodizing the aluminum foil in sulfuric, oxalic or phosphoric acid to create a nanoporous alumina film, with columnar pores on a surface of the aluminum foil; stripping off the alumina film using hot chromic/phosphoric acid; repeating the anodizing until an alumina film with a desired thickness is obtained; removing or reducing thickness of any alumina barrier present between the aluminum foil and the bottom of the pores using step anodization, reverse polarity etching, soaking in phosphoric acid, or combinations thereof; depositing semiconductor material into at least some of the columnar pores to obtain an array of semiconductor nanowires vertically disposed with respect to the aluminum foil; depositing a layer of optically transparent conductive material on top of the alumina film; and attaching wire leads to the layer of optically transparent conductive material and the aluminum substrate.
This application claims priority to and the benefit of the filing date of U.S. Provisional Application No. 61/611,047, filed Mar. 15, 2012, the disclosure of which is hereby incorporated by reference herein in its entirety.
1. Field of the Invention
The present invention relates to the field of photodetectors. Embodiments of the present invention provide room temperature IR, visible and UV photodetectors that incorporate wavefunction engineering and/or photomodulated electron tunneling. They are made by electrochemical self-assembly of nanowires. Photodetectors of embodiments of the invention have an infrared (wavelength averaged) detectivity exceeding 107 cm-(√Hz)/watt at room temperature.
2. Description of Related Art
A photodetector is a device that senses light by producing an electrical current under illumination. The onset of this current signals the presence of light. Infrared (IR) photodetectors are widely used in car collision avoidance systems, detection of buried mines, night vision, forensics, search for water in the moon and extra-solar planets, formation of stellar nurseries, missile defense, and analytical chemistry to name a few. Ultra-violet (UV) photodetectors are used for fire and explosion detection, UV exposure detection, chemical-related chromatography applications, such as High Performance Liquid Chromatography (HPLC), astronomy and missile defense. Visible light detectors are used in forensics and analytical chemistry.
The best photodetectors are typically made of semiconductors. They work by absorbing an incident photon that excites an electron from the valence band (Ev) to the conduction band (Ec) of the semiconductor leaving behind a hole (electron vacancy) in the valence band, as shown in
hf≧ΔE EQ. (1)
This condition is mandated by the principle of energy conservation. Because of the requirement in Equation (1), IR detectors will need semiconductors of very small bandgap ΔE because IR photons have very small energies hf (typically a few milli-electron volts, or meV).
Typical IR detectors are indeed made of small-bandgap semiconductors like InSb and HgCdTe. However, if the semiconductor has a small bandgap, then phonons (quanta of lattice vibrations) can also excite electrons from the valence to the conduction band, since phonons have energies comparable to those of IR photons, and satisfy Equation (1). Therefore, current can flow through the detector even in the dark because of the phonons exciting electrons from the valence to the conduction band. The population of phonons increases rapidly with temperature because phonons obey Bose-Einstein statistics. At room temperature, the phonon population in a detector may vastly exceed the photon population in incident IR light. In that case, phonons would have already generated a large number of electron-hole pairs in the dark and the few additional pairs generated by IR light will make little difference to the pair population or the total current. In other words, the current under illumination by IR light Ilight will only slightly exceed the dark current Idark. That will make the signal-to-noise ratio (SNR) of the detector only slightly larger than unity:
A low signal-to-noise ratio severely impairs the detection ability of the detector since there is hardly any difference between the current flowing in the dark and current flowing under illumination. Therefore, it is imperative to increase the room-temperature signal-to-noise ratio and make it considerably exceed unity by clever engineering.
The easiest way to increase the signal to noise ratio is to cool the detectors with cryogens, which will reduce the phonon population and the dark current. However, there is a strong need for semiconductor infrared (IR) photodetectors that work at room temperature without cooling and yet has large signal-to-noise ratio because of the cost and inconvenience of cryogenic or Peltier cooling. Room temperature IR detectors with adequate signal-to-noise ratio are usually fashioned out of microbolometers, thermopiles or pyroelectric detectors, but they are expensive and fragile. See P. W. Kruse and D. D. Skatrud, Uncooled Infrared Imaging Arrays and Systems, (Academic, San Diego, 1997).
Rugged room-temperature IR detectors have been derived from semiconductor structures fabricated with expensive molecular beam epitaxy. See A. Rakovska, V. Berger, X. Marcadet, B. Vinter, G. Glastre, T. Oksenhendler and D. Kaplan, Appl. Phys. Lett., 77, 397 (2000); and P. V. V. Jayaweera, S. G. Matsik, A. G. U. Perera, H. C. Liu, M. Buchanan and Z. R. Wasilewski, Appl. Phys. Lett., 93, 021105 (2008). Such fabrication methods, however, are not amenable to high throughput and therefore not commercially viable. Examples of conventional detectors include those disclosed in U.S. Pat. Nos. 8,313,972, 8,203,195, 7,989,753, 7,906,361, and 7,501,611, and in US Published Patent Application Nos. 20120138901 and 20080178924. They are typically fabricated with expensive techniques that have slow throughput.
There is a recent report that millions of IR detectors will be required across the planet to monitor the effects of global warming. Such detectors must operate without cooling and must be robust, reliable, rugged, cheap and mass-producible with high throughput. Present detectors cannot meet this requirement because they are either too fragile, or too expensive, or do not have sufficient signal-to-noise ratio. The invented detector meets this requirement since it works at room temperature and is robust, rugged, cheap and mass-producible.
In addition to the low signal-to-noise ratio at room temperature, a serious shortcoming of most semiconductor photodetectors is that they are not frequency-selective, i.e. they cannot lock on to a specific frequency of light at the exclusion of all other frequencies. Thus, they cannot satisfy the requirement of frequency-resolved analysis of radiation which is important in astronomy, forensics and molecular spectroscopy. However, nanowire based photodetectors are naturally frequency-selective since they absorb (and therefore detect) light within a narrow window of frequency. This happens because the absorption spectrum of nanowires exhibit distinct peaks at specific frequencies and these frequencies can be tailored with appropriate choice of nanowire materials. The peaks are caused by the fact that electrons and holes in nanowires are constrained to move in only one dimension (along the length of the nanowire) and cannot move sideways. This is known as quasi one-dimensional carrier confinement. One-dimensional confinement leads to frequency-selectivity of absorption and hence detection.
Synthesizing nanowires that exhibit absorption peaks (and hence frequency selectivity) is remarkably difficult since the nanowire diameter has to be smaller than the thermal DeBroglie wavelength of electrons, which may be a few nanometers. Nanowires have been made by various processes, including such processes as disclosed in U.S. Pat. Nos. 7,892,956, 7,691,201, 7,651,944, 7,635,600, 7,608,902, and 7,566,435 and in US Published Patent Application Nos. 20110108803, 20100239488, 20100147674, 20100002324, and 20070155173. Not all of them exhibit absorption peaks and hence not all of them can be fashioned into frequency selective photodetectors.
Nanowires prepared by the “porous alumina template” technique where semiconductors are selectively electrodeposited in nanometer sized pores of anodic alumina films produced by anodizing an aluminum foil in oxalic or sulfuric acid [see C. R. Martin, Science, 266 (1994) 1961; and S. Bandyopadhyay, et al., Nanotechnology 7 (1996) 360] are desirable since the fabrication process is inexpensive and have a high throughput. However, such nanowires are usually polycrystalline, have a high density of surface states (see V. Pokalyakin, et al., J. Appl. Phys. 97 (2005) 124306) and are replete with impurities absorbed during synthesis. Consequently, they do not commonly exhibit absorption peaks. Current reports in the literature reveal that CdS and CdSe nanowires produced by this technique exhibit blue shifts in their absorption spectra, but the absorption strengths increase monotonically with photon energy (see S. P. Mondal, K. Das, A. Dhar, S. K. Ray, Nanotechnology 18 (2007) 095606; and S. P. Mondal, K. Das, A. Dhar, S. K. Ray, in: Proceedings of the International Workshop on Physics of Semiconductor Devices, Mumbai, India, 2007, p. 855), instead of exhibiting peaks. Such nanowires therefore cannot be fashioned into frequency-selective detectors. The invented electrochemically synthesized nanowires of CdSe and ZnS, exhibit absorption peaks and hence frequency selectivity at room temperature.
In order to devise an IR detector that does not require cooling and yet has an adequate signal-to-noise ratio (SNR), a new approach to photodetection was adopted. It is based on the following premise. The rate of transition of electrons from one energy-level E1 to another E2 in any material, due to either photons or phonons, is given by Fermi'"'"'s Golden Rule:
where E1 is the lower energy level, E2 is the higher energy level, and ME1,E2 is the matrix element for transition between the two levels, which is a measure of the strength of coupling between electrons and either photons or phonons. Coupling induces the transition from level E1 to E2.
Matrix element for electron-photon coupling is given by:
Melec-photon∝Nphoton|∫ψ2(r)e[−iev·∇ψ1(r)]d3r|2, EQ. (4)
where Nphoton is the photon occupation number that depends on the incident light intensity, er is the unit vector in the direction of light polarization, k is the photon'"'"'s wavevector, ψ1(r) is the initial state wavefunction in level E1 and ψ2(r) is the final state wavefunction in level E2. On the other hand, the electron-phonon coupling strength is:
Melec-phonon∝Nphonon|∫ψ2(r)eikrψ1(r)d3r|2, EQ. (5)
where Nphonon is the phonon occupation number. At equilibrium, it is given by the Bose-Einstein factor
Normally, Nphonon>>Nphoton at room temperature, which makes the phonon induced transitions much stronger than the photon induced ones. Consequently, Ilight−Idark<<Idark and the SNR is only slightly larger than unity.
However, if the electron'"'"'s initial state (level E1) is a trap state where the electron is trapped and has a very localized wavefunction ψ1(r), as shown in
In order to exploit wavefunction engineering, an IR detector is made out of a wide bandgap semiconductor whose bandgap energy is much larger than the energy of IR photons. Thus, neither photons not phonons can excite electrons directly from the valence band to the conduction band and cause current by generating electron-hole pairs. However, if the semiconductor is fashioned into a nanowire embedded in an insulating matrix, then the interface defects cause shallow trap states to form close to the conduction band in energy. The energy difference between the conduction band and the trap states is less than the energy of IR photons and phonons, so that both can excite electrons from the trap states into the conduction band. The electron in traps states cannot produce current since they are not free to move, but electrons excited into the conduction band can since they become free. Since the wavefunction in the traps state is localized but the wavefunction in the conduction band is delocalized, the matrix element for electron photon coupling becomes comparable to that of electron phonon coupling despite the fact that the phonons vastly outnumber the photons. Therefore, about equal numbers of electrons are excited from trap states into the conduction band by photons and phonons at room temperature when there are many more phonons around than photons. The current produced by the photons and that produced by the phonons are very comparable despite the preponderance of phonons. This makes the ratio of the light to dark current (or the signal to noise ratio) significantly exceed unity even at room temperature and enables room temperature IR detection.
Embodiments of the present invention relate generally to semiconductor nanowires and photodetectors comprising them. Three basic figures of merit for a photodetector include: (1) Detectivity—a measure of how sensitive the detector is to light, usually determined by the minimum light intensity that the detector can detect in the presence of noise; (2) Signal-to-noise ratio—the ratio of the current under illumination to the current in the dark, where the larger this quantity is, the easier it is to detect the presence of light; and (3) Standby power dissipation—the current that flows through the detector in the dark causes power dissipation and drains the battery, thus, reducing the dark current so that the standby power dissipation is reduced is a plus.
Embodiments of the invention provide IR photodetectors that can operate at room temperature (without cooling) with high signal-to-noise ratio (i.e. large ratio of light current to dark current). This is made possible by the use of wavefunction engineering, whereby electrons from shallow trap states with localized wavefunctions are excited by IR photons (and phonons) to the conduction band with delocalized wavefunction. To implement such a “wavefunction-engineered” detector, two modifications are made. First, a semiconductor with large bandgap is used so that neither IR photons, nor phonons, can directly excite electrons from the valence to the conduction band (because Equation (1) cannot be satisfied). Second, electron traps states are introduced slightly below the conduction band level by embedding the nanowires in an insulating matrix. The material mismatch at the interface of the semiconductor nanowire and the insulator causes these trap states. The energy separation between the conduction band and the trap states is less than the energy of IR photons and phonons, so that in accordance with Equation (1), both photons and phonons should be able to excite electrons from the trap states (where electrons are trapped and cannot carry current) to the conduction band (where electrons are free and can carry current). The difference with conventional photodetectors is that instead of exciting electrons from the valence to the conduction band, here one is exciting electrons from trap states to the conduction band.
Embodiments of the invention provide photodetectors with frequency selectivity. To accomplish this, the photodetectors comprise nanowires instead of bulk semiconductor material. The absorption of photons in a semiconductor is directly proportional to a quantity called the “joint electron-hole density of states,” which is the number of allowed electron and hole states per unit energy per unit volume. In a nanowire, this quantity is non-zero only within a narrow window of frequencies whereby only photons that have energies within that window are absorbed and detected. This enables frequency selective detection.
A third modification that can further increase the signal to noise ratio is to exploit photomodulated electron tunneling. Here, the nanowires are modified to integrate a thin insulator layer at one end, as shown in
The tunneling current depends on the quantum mechanical tunneling (or transmission) probability through the insulator or tunnel barrier. The tunneling current through an arbitrary potential barrier was derived by Leo Esaki and Raphael Tsu as:
where e is the electronic charge, h is Planck'"'"'s constant, E is the tunneling electron'"'"'s kinetic energy, EF1 is the Fermi energy at one side of the potential barrier and EF2 is the Fermi energy at the other side, and |T(E)| is the transmission (or tunneling) probability of an electron with energy E to tunnel through the barrier.
The function ƒ(E−EF) is the Fermi-Dirac occupation probability, which is the probability of finding an electron having kinetic energy E. It is given by:
where kB is the Boltzmann constant, T is the absolute temperature and EF is the Fermi energy, which is the energy where the occupation probability is 50%.
If the potential difference across the barrier is V, then:
EF1−EF2=−eV(reverse bias) EQ. (10)
For illustration and as shown in
Substituting Equation (10) in (8), the current versus voltage relation of tunneling current is represented as:
Equation (11) shows that the tunneling current depends on the transmission probability |T(E)|.
According to the Wentzel-Kramcrs-Brillouin (WKB) approximation, the tunneling probability is given by:
where k(x) is the electron'"'"'s wavevector as a function of location x within the barrier. It is related to the electron'"'"'s kinetic energy K.E. as:
where m is the electron'"'"'s effective mass within the barrier, E is the electron'"'"'s total energy (kinetic+potential), Eb(x) is the potential energy of the barrier as a function of distance, and h=h/2π. Within the barrier itself, Eb(x)=Φ−eEx, where E is the electric field in the barrier. If there are no fixed charges within the barrier, then the electric field within the barrier is independent of x, according to the Poisson equation,
where ρ is the charge density, ∈ is the dielectric constant, and when ρ=0 (i.e., no charges present) then E is independent of x.
is then given by:
Expanding the factor
in a binomial series and retaining up to second order terms since
the result is:
Substituting this result in Equation (12), the result is:
Note that at 1 volt bias across the tunnel barrier, eEW≈1 eV, whereas Φ≈3 eV if the semiconductor is CdS and the insulator is alumina. Therefore the second term within the curly brackets is at least one order of magnitude smaller than the first term.
It is clear that if one can modulate the tunnel barrier height with light, then the current can increase dramatically under illumination since the current and transmission probability are exponentially sensitive to the barrier height Φ. This is possible since light will excite electrons from the trap levels at the semiconductor-insulator interface into the conduction band of the semiconductor. Once again, phonons do not excite electrons from trap levels into the conduction band of the semiconductor and therefore do not modulate the tunnel barrier height.
If electrons are excited into the conduction band by light, then the electron concentration in the conduction band increases. This “bends down” the conduction band since the electron concentration n(x) [number of electrons per unit length] in the conduction band of a semiconductor nanowire, at any position x, is given by:
where kF(x) is the Fermi wavevector as a function of position x within the barrier, EF is the Fermi level, which is constant at equilibrium, and Ec(x) is the conduction band edge energy in the barrier as a function of position. Note that
Clearly if n(x) increases, then EF−Ec(x) increases, which means that the conduction band “bends down” in the energy band diagram with respect to the Fermi level. However, if the conduction band in the semiconductor bends down, then that decreases the tunnel barrier height Φ. From Equation (16), this will increase the tunneling probability exponentially and hence the tunneling current exponentially with increasing n(x). This is the basis of “photomodulated electron tunneling”. Because of the exponential dependence, instead of the current increasing linearly with the number of electrons excited into the conduction band by light, it will now increase exponentially with that number. This will further increase the signal-to-noise ratio Ilight/Idark.
In accordance with embodiments of the invention, a photodetector was indeed built on this principle of wavefunction engineering and tested at room temperature. The signal-to-noise ratio of this photodetector was measured to be 2:1 at room temperature, which is adequate but not impressive. What is extremely impressive however is that the detectivity of this detector averaged over all detected IR wavelengths exceeded 107 Jones at room temperature. This is among the highest room temperature detectivity reported for IR detectors
In order to increase the signal-to-noise ratio further, a tunneling photodetector incorporating photomodulated electron tunneling was built. It exhibited a signal to noise ratio of 7:1 at 1.5 V bias, but more importantly reduced the standby power dissipation 7,500 times to only 0.12 microwatts, making it exceptionally energy-efficient.
The current invention addresses disadvantages of the prior art and overcomes a major shortcoming of room-temperature IR detectors, namely poor signal-to-noise ratio and poor detectivity. Room temperature detectors should have high detectivity (or sensitivity) to detect the faintest intensity of light. Additionally, it is desired that such detectors have a high light-to-dark contrast ratio for distinguishing and for reliability.
The inventors have discovered that the light-to-dark contrast ratio is increased by incorporating photomodulated electron tunneling. It also reduces standby power dissipation drastically. Specific objects of embodiments of the invention include IR and UV photodetectors comprising: (a) a conductive substrate, such as an aluminum substrate, with an upper surface; (b) a layer of insulator disposed on the upper surface of the aluminum substrate and comprising an array of columnar pores; (c) a plurality of semiconductor nanowires disposed within one or more of the pores and standing vertically relative to the upper surface of the aluminum substrate; (d) a layer of nickel disposed in operable communication with one or more of the semiconductor nanowires; and (e) wire leads in operable communication with the aluminum substrate and the layer of nickel for connection with an electrical circuit.
In embodiments, the photodetector can comprise: an array of semiconductor nanowires either standing vertically on a conducting substrate and capped by a conductor transparent to light, or lying horizontally on a non-conducting substrate and capped by conductors at both ends; and wire leads attached to the conductors and conducting substrate for connection to an electrical circuit. Non-conducting substrates can include quartz and similar materials. The semiconductor nanowires can comprise any direct bandgap semiconductor of bandgap energy between 0.3 eV and 4 eV or direct bandgap insulator of bandgap between 4 and 10 eV. Lithium fluoride (LiF) and magnesium oxide (MgO) are examples of direct bandgap insulators that can be used as nonowires. Preferred bandgap insulators can include semiconductors with a bandgap energy ranging for example from 4-10 eV, such as from 4 to 4.5 eV, or 5-5.5 eV, or 6-6.5 eV, or from 7-7.5 eV, or from 8-8.5 eV, or from 9-9.5, eV, or 10 eV. It is not critical how the nanowires are fabricated and methods can include self-assembly, or nanolithography, or combinations thereof.
Preferred embodiments include a photodetector comprising: a conductive substrate with an upper surface; a layer of insulator disposed on the upper surface of the conductive substrate and comprising an array of columnar pores; a plurality of semiconductor nanowires disposed within one or more of the pores and standing vertically relative to and with one end of the semiconductor nanowires in operable communication with the upper surface of the conductive substrate; a layer of conductive material disposed in operable communication with an opposing end of one or more of the semiconductor nanowires; and wire leads in operable communication with the conductive substrate and the layer of conductive material for connection with an electrical circuit. In preferred embodiments, the conductive substrate is aluminum and the conductive material is nickel.
Such photodetectors can be configured such that the semiconductor nanowires have a diameter greater than 0 and up to 100 nm. Preferred embodiments comprise nanowires with a diameter in the range of 5-50 nm. Especially preferred are photodetectors with semiconductor nanowires having a diameter of about 10 nm.
Photodetector embodiments can comprise semiconductor nanowires of at least one of CdS, CdSe, CdTe, GaAs, ZnS, ZnSe, or ZnTe, or one or more alloy thereof. Indeed, the semiconductor nanowires can comprise any binary or ternary group II-VI or III-V semiconductor compound, or alloy.
The insulator preferably comprises a layer of porous alumina (aluminum oxide). The alumina layer can be about 1-10 μm thick, but preferred embodiments comprise an alumina or insulator layer with a thickness of about 1 μm. In embodiments, the barrier layer of insulator is disposed between the columnar pores and the upper surface of the conductive substrate.
Photodetector embodiments can comprise an insulator or alumina layer with columnar pores extending lengthwise therethrough. In embodiments, the columnar pores pass completely through the layer of alumina. In other embodiments, the pores have a bottom and do not pass completely through the alumina such that there is a barrier layer of alumina disposed between the bottom of the columnar pores and the upper surface of the aluminum substrate. Preferably, when there is a barrier layer of alumina, the barrier layer has a thickness ranging from above 0 to 20 nm, and more preferably from about 1-5 nm. In preferred embodiments, some of the columnar pores are through holes through the alumina or insulator layer, while some of the pores have a bottom.
The photodetectors can be configured to be capable of detecting IR, UV, or visible light. In preferred embodiments, the photodetectors are configured to selectively detect only IR or UV light and only certain wavelengths of IR or UV light.
In embodiments, the wire leads are gold wire, but can be any conductive material.
The semiconductor nanowires can have a bandgap ranging between 1-4 eV.
The photodetectors can be capable of a signal-to-noise ratio of from 2 to 10, such as about 3, 4, 5, 6, 7, 8, or 9. Additionally or alternatively, the photodetectors can be capable of a detectivity at room temperature ranging from 1×105 Jones to 1×108 Jones.
In embodiments, when exposed to IR or UV light, electrons in the semiconductor nanowires of the photodetectors are excited from trap states to the conduction band.
The photodetector can have a layer of nickel that is up to 50 nm thick, such as from about 5-45 nm thick, or about 10-30 nm thick, or even about 20-25 nm thick. In preferred embodiments, the layer of conductive material (such as nickel) is 50 nm thick. The top contact does not have to be nickel and can be any transparent conductor. Indium tin oxide can also be used as the conductor or conductive material.
Semiconductor material can be disposed between the insulator layer and the layer of nickel. In embodiments, this is achieved by overfilling the pores with semiconductor material. This layer of additional semiconductor material is thus in operable communication with the semiconductor nanowires within the pores and the layer of nickel. Again, when referring to the layer of nickel, this can be any conductor material.
The photodetectors and nanowires can be made by any applicable process known in the art. Provided in this specification are representative methods that can also be used. Any one or more of the method steps can be omitted to fit a desired application or result. Likewise, any one or more of the method steps can be performed in any order, if applicable. Other method steps can also be added. One such method for making nanowires comprises: (a) providing 99% purity or greater aluminum foil with a thickness between 0-1 mm; (b) degreasing and electropolishing the aluminum foil in a solution of perchloric acid, butyl cellusolve, ethanol and distilled water to reduce surface roughness; (c) optionally rinsing the aluminum foil in distilled water; (d) anodizing the aluminum foil using sulfuric acid or oxalic acid to create an alumina film with columnar pores on a surface of the aluminum foil; (e) stripping off the alumina film using hot chromic/phosphoric acid; (f) repeating the anodizing and stripping until an alumina film with a desired thickness and pore diameter is obtained; (g) removing or reducing thickness of any alumina barrier present between the aluminum foil and the pores by reverse polarity etching using phosphoric acid at room temperature; and (h) depositing semiconductor material into at least some of the columnar pores to obtain an array of semiconductor nanowires vertically disposed with respect to the aluminum foil.
A preferred method includes depositing the semiconductor material using dc electrodeposition and overfilling at least some of the pores such that semiconductor material is disposed on top of the alumina film.
Methods of making a photodetector according to embodiments of the invention can comprise: (a) providing 99% purity or greater aluminum foil with a thickness between 0-1 mm; (b) degreasing and electropolishing the aluminum foil in perchloric acid, butyl cellusolve, ethanol and distilled water to reduce surface roughness and obtain an aluminum substrate; (c) optionally rinsing the aluminum foil in distilled water; (d) anodizing the aluminum foil using sulfuric acid or oxalic acid to create an alumina film with columnar pores on a surface of the aluminum foil; (e) stripping off the alumina film using hot chromic/phosphoric acid; (f) repeating the anodizing and stripping until an alumina film with a desired thickness and pore diameter is obtained; (g) removing or reducing thickness of any alumina barrier present between the aluminum foil and the pores by reverse polarity etching using phosphoric acid at room temperature; (h) depositing semiconductor material into at least some of the columnar pores to obtain an array of semiconductor nanowires vertically disposed with respect to the aluminum foil; (i) depositing a layer of nickel on top of the alumina film; and (j) attaching wire leads to the layer of nickel and the aluminum substrate.
A more general method of making a photodetector can comprise: providing a conductive substrate with 99% purity or greater with a thickness between 0-1 mm; degreasing and electropolishing the conductive substrate to reduce surface roughness; optionally rinsing the conductive substrate in distilled water; anodizing the conductive substrate to create an insulative film, which comprises an oxide of the conductive substrate, with columnar pores on a surface of the conductive substrate; stripping off the film; repeating the anodizing and stripping until a film with a desired thickness and pore diameter is obtained; removing or reducing thickness of any oxide barrier present between the conductive substrate and the pores; depositing semiconductor material into at least some of the columnar pores to obtain an array of semiconductor nanowires vertically disposed with respect to the conductive substrate; depositing a layer of conductive material on top of the insulative film; and attaching wire leads to the layer of conductive material and the conductive substrate.
Methods also include a method of making nanowires comprising: providing a conductive substrate with 99% purity or greater and with a thickness between 0-1 mm; degreasing and electropolishing conductive substrate to reduce surface roughness; optionally rinsing the conductive substrate in distilled water; anodizing the conductive substrate to create an insulative film, which comprises an oxide of the conductive substrate, with columnar pores on a surface of the conductive substrate; stripping off the insulative film; repeating the anodizing and stripping until an insulative film with a desired thickness and pore diameter is obtained; removing or reducing thickness of any oxide barrier present between the conductive substrate and the pores; and depositing semiconductor material into at least some of the columnar pores to obtain an array of semiconductor nanowires vertically disposed with respect to the conductive substrate.
Detectors provided by embodiments of the invention can detect the entire electromagnetic spectrum from infrared (IR) to ultraviolet (UV) at room temperature. This is made possible by quantum mechanical wavefunction engineering, which selectively suppresses phonon-mediated excitation of electrons into the conduction band, while allowing photon-mediated excitation. This increases the room-temperature signal-to-noise ratio (SNR) which is the ratio of current under illumination to the current in the dark. This ratio is further enhanced by integrating the detector with a tunnel barrier where light modulates the barrier height and hence the tunneling current.
An additional feature of the inventive nanowire IR photodetectors is an SNR of approximately 2 and a detectivity exceeding 107 cm-(√Hz)/watt at room temperature. This detectivity is ten times better than that reported in the state-of-the-an as described by Jayaweera, et al., Applied Physics Letters, 93, 021105 (2008). Furthermore, because this detector is fabricated with electrochemical self assembly in a beaker, it is much cheaper than other photodetectors and is robust, reliable and mass-producible as well due to the high throughput capability of its production process.
Additional desired features of such detectors are that they have low standby power dissipation for energy efficiency. The tunnel detector of embodiments of the present invention has very low dark current and hence very low standby dissipation. Further, the capacitive detector is non-leaky and ideally does not dissipate any energy.
Yet another object of embodiments of the present invention is that the detectors are light frequency selective, i.e. preferentially detect light within a band of frequencies thereby enabling frequency-resolved detection. Such a feature is not possible with bulk semiconductor photodetectors. The inventors have made this possible by one-dimensional density of states in 10 nm diameter nanowires. Specific embodiments of the invention provide detectors capable of detecting both IR and UV light, as well as detectors capable of selectively detecting only IR light or selectively detecting only UV light.
The nanowire material of detectors of the invention is a semiconductor with a large bandgap far exceeding IR photon or phonon energies, which disallows excitation of electrons from the valence to the conduction band by either photons or phonons. By embedding the nanowires in an insulating matrix (alumina), one can introduce trap states in the semiconductor'"'"'s bandgap which form as a result of the material mismatch between the semiconductor and the insulator. These trap states have been characterized in the past and were found to be filled with electrons at room temperature. Capacitance-voltage measurements had shown that the nanowires are n-type, which means that the Fermi level must be much closer to the conduction band than the valence band at equilibrium and all but the shallowest traps will be filled with electrons.
Under illumination, the shallow traps are continuously emptied by light as electrons are excited out of them into the conduction band. Once in the conduction band, these electrons flow out into the contact and contribute to the photocurrent. However, the emptied traps are quickly refilled by thermal (multiple phonon-mediated) transitions from the valence band as shown in
Additional information about the nanowires and photodetectors included within the invention may be found literature authored by the inventors. See Saumil Bandyopadhyay, Jennette N. Mateo, Supriyo Bandyopadhyay, and Garry Glaspell, “Signature of quasi one-dimensionality in the absorption spectra of electrochemically self-assembled nanowires,” Physica E, 43, 1255-1261 (2011) and see Saumil Bandyopadhyay, Pratik Agnihotri and Supriyo Bandyopadhyay, “A self assembled room temperature nanowire infrared photodetector based on quantum-mechanical wavefunction engineering,” Physica E, 44, 1478 (2012).
The accompanying drawings illustrate certain aspects of some embodiments of the present invention, and should not be used to limit or define the invention. Together with the written description the drawings serve to explain certain principles of the invention.
Reference will now be made in detail to various exemplary embodiments of the invention. It is to be understood that the following discussion of exemplary embodiments is not intended as a limitation on the invention. Rather, the following discussion is provided to give the reader a more detailed understanding of certain aspects and features of the invention.
Nanowire IR detectors according to embodiments of the invention can be electrochemically self assembled using a porous alumina template technique. In a representative fabrication method, the starting material is commercially available aluminum foil, such as 99.999% pure aluminum foil, with a thickness in the range of about 0.1 to 1 mm thick. A preferred thicknesses for example can be about 100 μm, 200 μm, 300 μm thick and so on.
In embodiments, the aluminum foil is preferably chemically degreased and electropolished. A solution of perchloric acid, butyl cellusolve, ethanol and distilled water can be used for degreasing and electropolishing in a commercial Buehler Electromet electropolisher for about 30-60 s at 60 V dc to reduce the surface roughness to about 3 nm. See R. E. Ricker, A. E. Miller, D-F Yue, G. Banerjee and S. Bandyopadhyay, J. Elec. Mat., 25, 1585 (1996). A preferred degreasing and electropolishing solution can comprise 62 cc perchloric acid, 700 cc ethanol, 100 cc butyl cellusolve and 137 cc distilled water. The aluminum foil treated in this manner is characterized by a mirror-like surface. To prevent overheating of the solution, the electropolishing can be carried out in intervals of 5 s. This results in reducing the root-mean-square surface roughness of the foil to about 3 nm as verified by atomic force microscopy. See R. E. Ricker, et al., J. Electron. Mater. 25 (1996) 1585.
The foil is then rinsed in distilled water and anodized in a Perkin-Elmer flat cell using the aluminum as anode and a platinum mesh as cathode. The electrolyte is either 15% sulfuric acid or 0.3M oxalic acid (3% oxalic acid). The anodization is preferably performed for a time ranging from about a few minutes to about 15 minutes and at a voltage of about 5-60 V, such as about 15 V. The process of anodization creates a porous alumina film on the surface of the foil, which is stripped off in hot chromic/phosphoric acid (0.2M H2CiO4/0.4M H3PO4) and the anodization can be repeated for 1 h to produce a thicker porous film. This film too is stripped off and the final anodization is carried out for 15 min to produce a porous film on the foil.
In embodiments, the porous film can be approximately 1 μm thick and preferably comprises a well ordered array of nanopores having a diameter of about 1-100 nm, such as from about 5-95 nm, or from about 10-15 nm, or from about 20-30 nm, or from about 25-35 nm, or from about 40-50 nm, or from about 45-55 nm, or from about 60-70 nm, such as from about 65-75 nm, or from about 80-90 nm, and so on. Preferably, the nanopores have a diameter in the range of about 10-50 nm, such as from about 5-60 nm, including 11 nm, 12 nm, 13 nm, 14 nm, 15 nm, 16 nm, 17 nm, 18 nm, 19 nm, 20 nm, 21 nm, 22 nm, 23 nm, 24 nm, 25 nm, 26 nm, 27 nm, 28 nm, 29 nm, 30 nm, 31 nm, 32 nm, 33 nm, 34 nm, 35 nm, 36 nm, 37 nm, 38 nm, 39 m, 40 nm, 41 nm, 42 nm, 43 nm, 44 nm, 45 nm, 46 nm, 47 nm, 48 nm 49 nm, 50 nm, 51 nm, 52 nm, 53 nm, 54 nm, 55 nm, 56 nm, 57 nm, 58 nm, 59 nm, or 60 nm. For some applications, even larger diameter nanowires may be desired, such as with a diameter ranging from 150-500 nm, including from about 200-400 nm, or from 250-35 nm, such as about 300 nm. A field emission scanning electron micrograph (SEM) image of a nanoporous film prepared using an electrolyte of oxalic acid with pores of about 50 nm in diameter is shown in
At the bottom of the pores, there is a thin barrier layer of alumina (typically about 20 nm thick) separating the bottom of the pores from the aluminum substrate. This is called a “barrier layer” in the literature since photocurrent will always flow along the length of the nanowires, and hence must tunnel through this barrier. Before carrying out dc electrodeposition to selectively deposit the semiconductor within the nanopores, the thin alumina barrier layer that forms at the bottom of the pores can be removed by a method known as “reverse polarity etching.” For this purpose, the anodized foil is contacted with 0.05 M (5%) phosphoric acid at room temperature in a flat cell and a voltage of 7 V dc is applied between the foil and platinum mesh using the foil as the cathode. This can be performed for about 30 s to about 1 hr, such as from about 45 s to about 45 minutes, to remove the barrier layer, or at least reduce its width, which barrier layer would have blocked dc current flow along the pore'"'"'s length and prevented de electrodeposition. The final barrier width typically varies between 2 and 20 nm. Alternatively, the barrier layer can be removed by soaking the film in 5% phosphoric acid for up to about 60 minutes. This process is known to etch the barrier layer much faster than the pore walls, so that the increase in the diameter of the pores during the soaking process is negligible. There are many other methods of removing the barrier layer that can also be used, none of which is any better controlled. This method has been found to be the simplest.
Next, the pores are selectively filled with the semiconductor of interest by de electrodeposition using selected electrolytes. This process results in the formation of vertically standing semiconductor nanowires within the pores. The bottom ends are in contact with the underlying aluminum if the barrier layer had been removed in the preceding step, or in embodiments the barrier layer can be reduced in thickness. Semiconductors that can be deposited in this manner include binary and ternary II-VI and III-V compounds (including but not limited to one or more of cadmium sulfide (CdS), cadmium selenide (CdSe), cadmium telluride (CdTe), zinc oxide (ZnO), zinc sulfide (ZnS), zinc selenide (ZnSe), zinc telluride (ZnTe), gallium arsenide (GaAs), and alloys of such compounds including cadmium zinc telluride (CdZnTe), otherwise referred to as CZT). Indeed, any semiconductor compound can be used with preferable compounds having a bandgap in the range of about 1 to 4 eV at 25° C., such as a semiconductors with a bandgap of 1-2 eV, such as 1.1 eV, 1.2 eV, 1.3 eV, 1.4 eV, 1.49 eV, 1.5 eV, 1.6 eV, 1.7 eV, 1.73 eV, 1.8 eV, 1.9 eV; or semiconductors with a bandgap ranging from 2-3 eV, such as 2.1 eV, 2.2 eV, 2.25 eV, 2.3 eV, 2.4 eV, 2.42 eV, 2.5 eV, 2.6 eV, 2.7 eV, 2.8 eV, or 2.9 eV; or semiconductors with a bandgap ranging from 3-4 eV, such as 3.1 eV, 3.2 eV, 3.3 eV, 3.37 eV, 3.4 eV, 3.5 eV, 3.54 eV, 3.6 eV, 3.7 eV, 3.8 eV, 3.84 eV, 3.9 eV, or 3.91 eV.
In a first example, CdS is selectively electrodeposited within the pores from a solution of cadmium perchlorate, dimethyl sulfoxide (DMSO) and sulfur powder. The electrodeposition is carried out in a Perkin Elmer flat cell using an ac voltage of 12 V rms and 200 Hz frequency at a temperature of 100° C. During the negative ac cycle, the Cd++ ions in the solution is reduced to zero-valent material and is selectively deposited within the pores because they offer the least impedance path for the ac current to flow. The Cd metal then reacts with sulfur in the solution at the high temperature and forms CdS nanowires within the pores. In embodiments, the nanowires can have a length corresponding to the thickness of the porous film, or can be less than the thickness of the film if the pores are underfilled or slightly more than the thickness of the film if the pores are overfilled during the fabrication process. The length of the nanowires can range between 0-10 μm. Preferably, the nanowires have a length ranging from about 1-5 μm, such as 2 μm, 3 μm, or 4 μm, or from about 6-9 μm. In especially preferred embodiments, the nanowires have a length of about 1 μm. Here, the CdS nanowires were fabricated to have a diameter of about 50 μm and a length of about 1 μm.
The diameter of the nanowires also corresponds to the diameter of the pores of the film. In embodiments, the nanowires can have a diameter of about 1-100 nm, such as from about 5-95 nm, or from about 10-15 nm, or from about 20-30 nm, or from about 25-35 nm, or from about 40-50 nm, or from about 45-55 nm, or from about 60-70 nm, such as from about 65-75 nm, or from about 80-90 nm, and so on. Preferably, the nanopores have a diameter in the range of about 2-50 nm, such as from about 1-5 nm, or from 2-10 nm, or from 3-4 nm, or 5-60 nm, including 11 nm, 12 nm, 13 nm, 14 nm, 15 nm, 16 nm, 17 nm, 18 nm, 19 nm, 20 nm, 21 nm, 22 nm, 23 nm, 24 nm, 25 nm, 26 nm, 27 nm, 28 nm, 29 nm, 30 nm, 31 nm, 32 nm, 33 nm, 34 nm, 35 nm, 36 nm, 37 nm, 38 nm, 39 nm, 40 nm, 41 nm, 42 nm, 43 nm, 44 nm, 45 nm, 46 nm, 47 nm, 48 nm, 49 nm, 50 nm, 51 nm, 52 nm, 53 nm, 54 nm, 55 nm, 56 nm, 57 nm, 58 nm, 59 nm, or 60 nm. For some applications, even larger diameter nanowires may be desired, such as with a diameter ranging from 150-500 nm, including from about 200-400 nm, or from 250-35 nm, such as about 300 nm. In preferred embodiments, the nanowires have a length of about 1 μm and a diameter of about 10-50 nm. Especially preferred are nanowires having a length of about 1 μm and a diameter of about 5-10 nm
When CdS is electrodeposited in porous alumina films, numerous traps states form at the interface between CdS and alumina. These traps play the critical role in room temperature IR detection with large signal-to-noise ratio, as explained earlier. Energy dispersive analysis of x-ray (EDAX) was carried out to verify that the wires formed inside the pores comprise CdS. The results are shown in
Once the wires are formed within the pores, 50 nm of nickel or indium tin oxide (ITO) or any other conductor which is semi-transparent to IR is evaporated through a mask on the sample surface for contact formation using an electron beam evaporator. In embodiments, the pores are slightly overfilled during dc electrodeposition so that they not only get filled to the brim, but the semiconductor spills out into the surface forming a thin surface layer. A representative process for making photodetectors is illustrated in
Since pore filling is non-uniform, it is expected that only certain regions of the samples will contain overfilled pores. The thickness of the spilled layer is difficult to measure, but since the underlying pores through this layer can been seen with a 2 keV electron beam, the layer cannot be more than a few tens of nm thick. An electrical contact is then delineated on the top surface of each sample by e-beam evaporating 50 nm of nickel through a mask. The contact area is large enough (0.5 cm×0.5 cm) to cover many regions that have the surface layer caused by overfilling. This ensures that several parallel nanowires are electrically connected at both ends by the nickel contact and the bottom aluminum contact, so that good electrical characteristics can be expected. Finally gold wires are attached to the top contact (e.g., the front of the nickel film) and the bottom contact (e.g., the back of the aluminum substrate) with silver paste to complete the photodetector fabrication. See
In a second example, in order to electrodeposit CdSe selectively within the pores, the samples are rinsed thoroughly in distilled water and brought into contact with a solution of dimethylsulfoxide (DMSO) comprising 50 mM cadmium perchlorate, 10 mM lithium perchlorate (to increase conductivity of the solution) and 10 mM selenium powder, preheated to 200° C., in a flat cell. A dc voltage of 8 V is applied between the foil and the platinum mesh, using the former as the cathode. This drives the cation Cd++ selectively into the pores since the latter offer the least resistance path for the current to flow. Once inside the pores, the Cd++ ion is reduced to zero-valent Cd atoms by absorbing electrons and reacting with Se in the solution to form CdSe. In the end, a dense array of vertically standing CdSe nanowires of diameter 10 nm dispersed uniformly in alumina is produced. The areal density of this wire is about 5×1011 g·cm2. In order to fabricate ZnS nanowires, zinc perchlorate can be substituted for cadmium perchlorate and sulfur powder for selenium powder with the same process. Likewise, to prepare CdTe or ZnTe nanowires similar substitutions using cadmium perchlorate and tellurium powder or zinc perchlorate and tellurium powder can be used. If desired, the composition of one or more of the nanowires can be changed during the fabrication process by introducing a different composition after filling of the pores has begun. For example, a first composition of CdS can be grown within one or more of the pores, then a composition of CdTe could be grown on top of the CdS, then CdS can be grown on top of the CdTe layer.
Similar samples have been extensively characterized by the inventors in the past using cross-section scanning electron microscopy to verify pore filling. Here, energy dispersive x-ray analysis verified the composition of the material inside the pores. For example,
In order to test the photodetectors, i.e. to verify that a current response is produced when IR or UV light is incident on the detector, the current versus voltage characteristic of the photodetector can be measured in the dark under IR or UV illumination. A Hewlett-Packard semiconductor parameter analyzer can be used for this purpose. Alternatively, or in addition, infrared absorption can be measured with a Fourier Transform Infrared (FTIR) spectrometer. For testing UV absorption, an ultra-violet laser (such as an He—Ne laser or an Nd:YAG laser) and a monochromator with wavelength resolution of 1 nm can be used. The IR or UV absorption spectra of samples with unfilled pores (no electrodeposition) can be used for background subtraction in order to eliminate spurious absorption.
For testing IR absorption, the IR source can be a 250 Watt IR lamp radiating within n radians in the near- to mid-infrared. The bulb'"'"'s surface can be placed 4 cm away from the sample for the measurements. If the bulb is brought closer to the sample, the incident light power will increase and the light current will increase, whereas if the bulb is retracted farther, the incident light power decreases and the light current decreases. If it is found that a single sample does not produce enough absorption, three samples can be stacked together. The absorption scans are preferably repeated up to about 64 times to achieve good signal-to-noise ratio. The samples can be mounted on a large metal chuck that immediately conducts away any heat generated by the IR lamp to eliminate spurious thermal effects and ensure that any change in the current when illuminated will be due to the IR light alone.
In embodiments, for measurement of the IR or UV absorption spectra of the nanowires, the samples should be disposed on a semi-transparent substrate by removing the metallic aluminum substrate from the bottom. This is accomplished by first applying a thick organic varnish on the surface of the sample and then soaking the film overnight in a solution of HgCl2 which selectively etches away the aluminum substrate (which is opaque to IR and UV light). The purpose of the varnish is to provide mechanical stability to the approximately 1 μm thin film during this process. Finally, the sample is immersed in ethanol overnight to dissolve out the varnish. The released thin film floats to the surface of the ethanol solution since its density is less than that of ethanol. The floating film is then carefully captured on a quartz slide that is transparent to ultra-violet radiation. The film is nitrogen dried and ready for absorption measurement. Representative preparation steps are shown in
The measured IR absorption spectrum of 50-nm diameter CdS and CdSe nanowires, respectively, is shown in
Absorption decreases with increasing photon wavenumber (or energy). This is characteristic of nanowires which possess a quasi one-dimensional density of electron and hole states that ideally varies with photocarrier energy E as 1/√E in each subband. Since the absorption is non-zero, there must be shallow traps in the bandgap from where infrared photons are exciting electrons into the conduction band.
The absorption is stronger at smaller wavelengths because of the quasi one-dimensional density of states. The absorption strength is proportional to the joint electron-hole density of states. Hence, the observed spectrum is very characteristic of nanowires. Furthermore, since the absorption drops off with increasing photon energy, the IR absorption is mostly due to the nanowires and not the thin layer on the surface (due to overfilling) which will behave either as a two- or a three-dimensional layer that has either energy-independent density of states or density of states that increases with energy as √E. The spectrum matches the 1/√E dependence only approximately; an exact match would have required the bandstructure of the nanowires to be perfectly parabolic, which is extremely unlikely. This data also shows that there is most likely a dominant trap level, no deeper than 245 meV [k=2000 cm−1] from the lowest subband bottom in the conduction band of the nanowires, whence most of the electrons are excited into the conduction band by IR photons. It is noted that the equipment used does not allow for measuring absorption for wavenumbers less than 2000 cm−1.
The traps may have a varied origin. They could be caused by dislocations associated with relaxation of strain in the nanowires due to lattice mismatch between the semiconductor and alumina, or interface states. The origin however is not important for this purpose since it plays no role in photodetection.
The current versus voltage characteristic of the CdS tunnel detector was measured in the dark and under IR illumination. These results are shown in
Although time resolved measurements can be made with a pulsed IR source, it is expected that the response time will be roughly the time needed to excite electrons from the trap states into the conduction band. The discharging time of traps in some of the nanowires was measured in the past and was found to be on the order of 2.5 μs. Accordingly, it is expected that the response time under illumination will be of the same order, i.e., a few μs. Most applications of IR detectors are however in the steady-state, for which the response time is immaterial.
The fact that these wide-bandgap nanowires absorb IR light, and the conductivity increases upon absorption, confirms there must be shallow filled trap states in the bandgap from which electrons are being excited into the conduction band by IR photons. Although IR absorption can be due to free-carrier absorption as well, such absorption would not have increased the conductivity since it does not create additional carriers in the conduction band. Moreover, free carrier absorption in nanowires is weak owing to the k-selection rule. It is also possible that IR absorption may be due to excitation of electrons from the valence band into empty trap states close to the valence band, and the holes created in the valence band increased the conductivity. If that were the case, then the Fermi level in the semiconductor would have been below the empty traps, close to the valence band, and the nanowires would have behaved as p-type. Since that is not the case, this possibility can be eliminated and it can be asserted that IR absorption is due to excitation of electrons from filled shallow trap levels in the bandgap into the conduction band. This is explained in more detail below.
The UV absorption spectrum does not show a monotonic trend, but a broad peak around the bandgap frequency due to quasi one-dimensional density of states in 10 nm diameter nanowires. This is the first observation of quasi one-dimensionality in the optical absorption of self-assembled nanowires. More particularly, in
Another feature demonstrated by
On the other hand, strain can produce either a blue-shift or a red-shift depending on the a and b coefficients of the material and whether the strain is compressive or tensile. Thus, there are many possible causes of the observed red-shift in sample 2: (1) strain, (2) bandgap states and band-tailing due to the polycrystalline nature, and (3) Stark effect which reduces the effective bandgap and has been observed before in these types of nanowires. Bandgap states are most likely the cause of the red-shift since the FWHM is extremely large and can only be explained by absorption into numerous bandgap states. Note that here there is no blue-shift expected from quantum confinement (however slight), but instead a red-shift. Most likely, the red-shift is caused by the preponderance of bandgap states. This reinforces the notion that frequency shifts are unreliable indicators of quantum confinement. The feature that is most interesting in
To determine how narrow a nanowire has to be to behave as a quasi-one-dimensional structure, anodic alumina films with 50-nm diameter pores were prepared by anodizing aluminum in oxalic acid instead of sulfuric acid. Nanowires of nominally 50-nm diameter were then produced by electrodepositing semiconductors in these pores. These CdTe nanowires do not exhibit peaks in their absorption spectra, as shown in
In yet other examples, photoluminescence spectra of 10-nm diameter CdS and 10-nm diameter ZnSe nanowires are taken with a He—Ne laser. The absorption spectra are shown respectively in
The energy band diagram for the conduction band of the heterostructure comprising the semiconductor CdS and the insulator alumina is shown in
From Equation (16), it can be inferred that the transmission probabilities under forward bias conditions (i.e. when EF1−EF2=eV>0) in the dark and under illumination are, respectively,
Note that the inequality Edf>Elf should make the dark current larder than the light current, but then Φdf>Φlf, which should make the light current larger. The latter effect is much stronger since the first term within the curly bracket is more than one order of magnitude larger than the second term. Under forward bias, the light current is much larger than the dark current.
When the barrier is reverse-biased (i.e. when EF1−EF2=eV<0), and electrons flow from right to left in the right panel of
where the superscript “r” refers to reverse-bias conditions. Note that since Edr<Elr, the light current should be larger than the dark current, but this is the minor effect. The major effect is due to the barrier height. Since under reverse bias Φdr=Φlr, the light current and dark current will be about equal. Thus, the light current will be considerably larger than the dark current in forward bias, leading to a large signal-to-noise ratio, but the two currents will be approximately equal under reverse bias. This is confirmed in the current-voltage characteristic of
Thus embodiments of the invention provide an IR photodetector whose signal-to-noise ratio at room temperature has been significantly enhanced by exploiting two quantum mechanical principles, namely, wavefunction engineering via trap states and modulation of tunneling probability with light. Such detectors (e.g. 50-nm diameter CdS nanowires) work at room temperature with a signal-to-noise ratio of up to 7:1 under a forward bias of 1.5 V.
In order to calculate the effective detectivity D* of the photodetector at room temperature, proceed as follows: From
During illumination, the IR lamp is placed at a distance r=4 cm from the sample surface. Therefore, the power incident on the sample'"'"'s surface per unit area is 250 W/(πr2)=4.97 W/cm2 since the lamp radiates within a solid angle of π radians. Even if all surface reflection is neglected and it is assumed that all incident power is absorbed by the detector, the power absorbed is no more than 1.24 Watts since the detector'"'"'s surface area is 0.5 cm×0.5 cm. Since this absorbed power changes the current in the photodetector by about 0.8 mA at a bias voltage of 1 V, the responsivity R of the detector is 0.8 mA/1.24 W=6.45×10−4 A/W at 1 V bias. Therefore, the minimum IR power that the detector can detect in the presence of thermal and Shot noise at room temperature is In/R=43.4 nW/√Hz when biased at 1 volt. This is the ‘noise equivalent power’ (NEP).
The minimum IR radiation power that the detector can detect in the presence of noise is called the noise equivalent power (NEP). The smaller it is, the more sensitive is the detector. The unnormalized detectivity is the reciprocal of the noise equivalent power. The larger it is, the more sensitive is the detector.
Effective detectivity D* at room temperature is evaluated from formula D*=R(√(A))/In, where A is area of the detector in cm2. This yields detectivity in Jones (cm-(√Hz)/W). Using the values of R and In, D*=1.18×107 Jones at 1 volt bias. This is over an order of magnitude better than as reported in the literature. See P. V. V. Jayaweera, S. G. Matsik, A. G. U. Perera, H. C. Liu, M. Buchanan and Z. R. Wasilewski, Appl. Phys. Lett., 93, 021105 (2008).
While the detectivity is impressive, the SNR and the standby power dissipation are not. The latter quantity is the product of the dark current at a given bias and the bias itself. At 1 V, this quantity is 0.9 mW. In order to improve the SNR and standby power dissipation, slight design change was made whereby the barrier layer separating the nanowires from the underlying aluminum was incompletely removed during the fabrication step. As a result, the photocurrent through the nanowires will have to tunnel through a thin insulating layer. This improved both the SNR and the standby power dissipation—the latter by more than a factor of 1500—while the normalized detectivity decreased by a factor of about 36.
The barrier layer at the bottom of the pores is roughly 20 nm thick and it is ascertained that it takes about 60 minutes of soaking in 5% phosphoric acid to dissolve it completely. Assuming a constant dissolution rate, 45 minutes of soaking will leave behind an approximately 5 nm thick barrier through which electrons can tunnel. Therefore, the samples were prepared as before, except the soaking time was reduced to 45 minutes. This definitely leaves behind a thin barrier layer since the dc electrodeposition time increased considerably at the same dc voltage (minutes to hours). This happened because the dc deposition current reduced considerably. Cross-section SEM characterization can be used to determine definitively the presence and the width of the tunnel barrier with an SEM that provides 5-nm resolution at the low voltages needed to avoid charging up the insulator. The measured current-voltage characteristic, however, changed from linear to non-linear (superlinear) which happens only if carriers either tunnel through the barrier or thermionically emit over it. Since the barrier is obviously too high for thermionic emission, it seems that carriers must be tunneling through it.
The tunnel barrier will reduce both the dark current and the light current at a given bias since it increases the resistance in the path of the current. This will reduce the standby power dissipation. It is expected that it will also increase the SNR because of the following: Without the tunnel barrier, the SNR is approximately:
where nphoton is the number of electrons excited into the conduction band by photons and nphoton is the number excited by phonons.
Since the measured SNR in the samples without the tunnel barrier is about 2, nphoton≈nphonon. In other words, the additional electrons excited into the conduction band by light is roughly equal in number to the electron population excited there in the dark by phonons.
The SNR of a photodetector with a tunnel barrier however will not be given by Equation (20) because the current will no longer be proportional to the number of electrons in the conduction band. Photo-excitation of electrons from trap states into the conduction band causes the latter to ‘bend’ because charges accumulate at the interface between the semiconductor and the insulator (tunnel barrier). Under an electrical bias, the energy band diagram in the direction of current flow looks as in
The tunneling current through a barrier with a potential drop V across it is given by the Tsu-Esaki formula:
where EF1 and EF2 are the Fermi levels on two sides of the barrier. Conservatively, it is assumed that the voltage applied across the sample is dropped entirely across the barrier which, in any case, is the most resistive region. In that event, EF1−EF2=qV, where V is the voltage across the sample as shown in
According to the Wentzel-Kramers-Brillouin (WKB) approximation:
where κ(x) is the imaginary part of the position-dependent electron wave vector in the barrier and is given by κ(x). (√[2m*(Ec(x)−E)])/h, with m* being the electron'"'"'s effective mass in the barrier. In the barrier, Ec(x)−E=Φb−qEx, where Φb is the height of the tunnel barrier encountered by the electron and E is the electric field in the barrier. Therefore, the integral ∫x1x2 κ(x)dx is given by:
where W=x2−x1, Ed is the electric field in the barrier in the dark and E1 is the electric field in the barrier under illumination. The electric field in the barrier will be spatially invariant as long as there are no fixed charges within the barrier. If there are fixed charges, then Ed and Et should be interpreted as the spatial averages of the electric field within the barrier.
Expanding the factor
in a binomial series and retaining up to second order terms since
the result is:
Note that the second term within the curly brackets is about one order of magnitude smaller than the first term when the bias across the sample is 1.5 V. Even if all of this voltage was dropped across the tunnel barrier, qEW will be 1.5 eV while Φ≈3 eV.
Equations (21)-(24) clearly show that the light current will be considerably different from the dark current because Φ′<Φ.
Again referring to
The second feature, namely the non-linearity of the current-voltage characteristic is due to the fact that transport is due to tunneling. In this case, it is expected that the current will increase superlinearly with voltage due to the exponential sensitivity of the tunneling probability to barrier height. The third feature, namely the larger SNR was expected because of the foregoing discussion and is indeed observed in forward bias.
All that remains to be explained is the first feature, namely the difference between the SNRs in the forward and reverse bias characteristics. This is demonstrated with the help of
where the superscript ‘f’ refers to forward bias and ‘r’ to reverse bias.
The responsivity of the detector is the light current produced for a given incident light power. The change in the current under illumination is 3.4×10−7 A under a radiation power of 1.24 Watts. Therefore, the responsivity for R is 0.274 μA/W in this embodiment. Responsivity of photodetectors of the invention can range from 100-400 nA/W, such as from about 150 nA/W to about 200 nA/W, or from about 250 nA/W to about 300 nA/W, or from about 350 nA/W to 400 nA/W. In preferred embodiments, the responsivity ranges from 210 nA/W to 290 nA/W, or from 220 nA/W to 280 nA/W, or from 230 nA/W to 270 nA/W, such as from about 240 nA/W to 250 nA/W.
The noise equivalent power (NEP) is In/R=1.54 μW. Therefore, the detectivity is (√A)/NEP=3.3×105 Jones. Thus, the tunnel detector has a 36 times lower detectivity at room temperature, despite the higher SNR. In embodiments, the detectivity or normalized detectivity of the photodetectors of the invention can range from 1×104 Jones to 1×108 Jones, such as from 1×105 Jones to 1×107 Jones, or from 1×105 Jones to 1×106 Jones. In preferred embodiments, the detectivity is in the range of 2×105 Jones to 3×107 Jones. In embodiments, the NEP can be up to 2 μW, such as up to 10 nW, or 20 nW, or 30 nW, or 50 nW, or 100 nW, or 200 nW, or 300 nW, or 400 nW, or 500 nW, or 600 nW, or 700 nW, or 800 nW, or 900 nW, or up to 1 μW, or up to 1.25 μW, or up to 1.5 μW, or up to 1.75 μW, and so on. In the context of this specification, the term “up to” also includes the specified endpoint.
The major advantage of the tunnel detector is however in the standby power dissipation. This quantity is 4×10−7 A×1.5V=0.6 μW. Therefore, the standby power dissipation has been reduced from 0.9 mW to 0.6 μW for the same contact area, resulting in a factor of 1500 saving. In preferred embodiments, the standby power dissipation of photodetectors of the invention ranges from 100 nW to 1,000 nW, such as up to 110 nW, or up to 120 nW, or up to 150 nW, or up to 175 nW, or from 200-300 nW, or from 400-500 nW, or from 600-700 nW, or from 800-900 nW and so on.
The figures of merit for various embodiments of the invention can be any combination of parameters or ranges of values specified in this specification. Table I below provides the figures of merit for another photodetector comprising 50-nm diameter CdS nanowires when measured at 1.5 V forward bias.
As demonstrated, the inventors provide novel nanowire-based room-temperature infrared photodetectors that rely on excitation of electrons from shallow trap states in the bandgap into the conduction band. This mechanism results in an acceptable signal-to-noise ratio of greater than 1:1, such as up to 10:1. The measured detectivity of this detector exceeds that of other comparable room-temperature detectors by a factor of about 10. The signal to noise ratio and the standby power dissipation can be improved . . . the latter vastly . . . by integrating the nanowires with a tunnel detector, while sacrificing some of the high detectivity. These detectors are inexpensive and rugged and are mass producible since they are fabricated with electrochemical self assembly.
The present invention has been described with reference to particular embodiments having various features. One skilled in the art will recognize that these features may be used singularly or in any combination based on the requirements and specifications of a given application or design. For example, with respect to particular methods disclosed herein it is understood that any one or more of the method steps can be omitted, added to another method, or performed in a different order than disclosed. Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention. Where a range of values is provided in this specification, each value between the upper and lower limits of that range is also specifically disclosed. The upper and lower limits of these ranges may independently be included or excluded in the range as well, unless otherwise indicated. All numbers and ranges disclosed above may vary by some amount. As used in this specification, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. While compositions and methods are described in terms of “comprising,” “containing,” or “including” various components or steps, the compositions and methods can also “consist essentially of” or “consist of” the various components and steps.
The references cited in this disclosure are each hereby incorporated by reference herein in their entireties, especially for information that is well known in the art for example with respect to semiconductors. If there is any conflict in the usages of a word or term in this specification and one or more patent or other documents that may be incorporated herein by reference, definitions consistent with this specification should be adopted.