Light-driven gallium nitride-gold asymmetric nanotunnel devices and their fabrication methods
By designing a light-field driven gallium nitride-gold asymmetric nanotunnel junction device, and utilizing sub-10 nanometer gaps and asymmetric structures, the problems of easy damage under high potential barriers and small photoresponse current of existing devices were solved, and stable and efficient light-field driven ultrafast electron emission was achieved at low power.
Patent Information
- Application Number
- CN202310761649.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-26
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-06-26
AI Technical Summary
Existing light-driven ultrafast electron emission devices are limited by a high potential barrier, requiring a high field strength to enter the light-driven mechanism, which may cause device damage. In addition, the photoresponse current is small, which is not conducive to practical applications.
We designed a light-field driven gallium nitride-gold asymmetric nanotunnel device by fabricating a bowtie-shaped nanoantenna with a sub-ten nanometer gap on a gallium nitride substrate, optimizing the pattern using a geometric correction method to reduce the emission barrier, and combining it with an ultrafast laser-induced asymmetric structure to achieve light-field driven ultrafast electron emission.
By entering the optical field driving mechanism at lower power, the possibility of device damage is reduced, and the photoresponse current is increased from the pA level to the nA level, making it suitable for practical applications of optical frequency electronic devices.
Smart Images

Figure CN116895648B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of optical field driving technology, specifically to an optical field driven gallium nitride-gold asymmetric nanotunnel device and its fabrication method. Background Technology
[0002] In the ultrafast electron emission process driven by femtosecond pulsed lasers, the emission process transitions from photon-driven to field-driven as the optical field intensifies. The alternating electric field excited by a femtosecond pulsed laser on a metal surface is:
[0003]
[0004] Where F0 is the peak electric field, A(t) is the normalized pulse envelope, and ω is the carrier frequency. This refers to the carrier-envelope phase (CEP). In photon-driven multiphoton emission processes, specifically within a device, electrons at the metal Fermi level absorb photon energy, overcome the potential barrier, escape from the surface, and are emitted into the metal at the other end. This emission process depends on the photon energy and luminous flux density of the laser pulse. Here, the luminous flux density is the pulse intensity envelope, which is related to |F0A(t)|. 2 The response speed of this device is directly proportional to the pulse envelope of the field. At extremely high field strengths, the strong alternating electric field formed by the pulse periodically compresses the surface barrier of the metal, allowing electrons to tunnel through when the barrier is compressed. The metal emits sub-periodic electron packets in each cycle of the driving electric field, forming an ultrafast response current. In this case, the photoelectric emission process is controlled by the driving light field. This field-controlled light-matter interaction forms the basis of attosecond science and has recently expanded from gases to solid-state nanostructures. Fabricating on-chip integrated nanodevices for light-driven ultrafast electron emission has become a research hotspot. However, such ultrafast tunneling devices are limited by a high barrier (4-5 eV), requiring a high field strength to enter the ultrafast electron emission mechanism driven by the light field, which may damage the device. The photoresponse current is also relatively small (on the order of pA), requiring high current detection sensitivity, which is detrimental to the subsequent practical applications of ultrafast electronic devices.
[0005] Therefore, there is an urgent need for a new type of light field-driven emission mechanism to provide an effective solution to the shortcomings of existing technologies. Summary of the Invention
[0006] The purpose of this invention is to provide an optically driven gallium nitride-gold asymmetric nanotunnel junction device and its fabrication method, so as to solve the problems mentioned in the background art.
[0007] To achieve the above objectives, the present invention provides the following technical solution:
[0008] A light-field driven gallium nitride-gold asymmetric nanotunnel device includes a gallium nitride substrate, on the upper surface of which a bowtie-shaped nanoantenna is deposited. The bowtie-shaped nanoantenna is made of gold film material, and a gap is provided at the center of the bowtie-shaped nanoantenna. The width of the gap is on the sub-ten nanometer scale.
[0009] The gallium nitride substrate has a length and width of 1cm × 1cm, and the bowtie-shaped nanoantenna includes a pair of symmetrically distributed isosceles triangular units with a gap between the two units. The gap width of the bowtie-shaped nanoantenna is less than 15nm.
[0010] The gallium nitride substrate is an unintentionally doped intrinsic gallium nitride (~3.0μm uGaN / 650μm Sapphire).
[0011] The fabrication method of light-field driven gallium nitride-gold asymmetric nanotunnel junction devices includes the following steps:
[0012] Step 1: Substrate cleaning: Select an intrinsic gallium nitride substrate that is not intentionally doped (~3.0μm uGaN / 650μm Sapphire), cut it into 1cm×1cm pieces by mechanical cutting, and ultrasonically clean it in acetone, isopropanol and deionized water for 10 minutes respectively. After cleaning, remove it and blow it dry with a nitrogen gun, and place it in a vacuum adsorption box for later use.
[0013] Step 2, Spin Coating: PMMA 950K photoresist was selected, the spin coating speed was fixed at 3000 r / min, the pre-baking temperature was 180℃, and the time was 2 min. The thickness of the photoresist was measured to be about 330 nm using a step meter.
[0014] Step 3, Electron Beam Exposure: Use L-Edit software to draw a series of bowtie-shaped patterns with equal gaps (including electrodes, where the height and base of the isosceles triangle in the bowtie-shaped pattern are both set to 5um). Using the geometric correction method, the gap range on the layout is from -10nm to +10nm, with a step size of 2nm.
[0015] Before exposure, a 10nm aluminum film was deposited by magnetron sputtering to improve the conductivity of the substrate. The electron beam dose variation range during exposure was selected as 800μC / cm2~1000μC / cm2 with a step size of 50μC / cm2. After exposure, the aluminum film was removed with an alkaline solution, then rinsed in ultrapure water and dried with a nitrogen gun.
[0016] The developing solution used is a pre-prepared developing solution (4-methyl-2-pentanone: isopropanol = 1:3) and a fixing solution (isopropanol). The developing time is adjustable from 40s to 70s with a step size of 5s, and the fixing time is fixed at 30s.
[0017] Step 4, Evaporation and Stripping: The developed sample is deposited with a 40nm gold film by electron beam evaporation. After evaporation, it is placed in acetone heated to 85°C in a water bath for stripping. After the gold film except for the bowtie structure and electrodes is removed, it is cleaned in isopropanol heated to 85°C in a water bath, dried with a nitrogen gun, and placed in a vacuum adsorption box for testing.
[0018] In step three, the gap range on the layout is set to -2nm.
[0019] In step three, the electron beam dose during exposure is selected as 900 μC / cm2.
[0020] In step three, the development time is set to 55 seconds.
[0021] In step four, no adhesion layer is added between the 40nm gold film deposited by electron beam evaporation and the gallium nitride substrate. Further,
[0022] Compared with the prior art, the beneficial effects of the present invention are:
[0023] 1. In this invention, the light-field driven gallium nitride-gold asymmetric nanotunnel junction device, through the design of a low emission barrier and the fabrication of a nanotunnel junction with an extremely short channel, facilitates the device to enter the photon-driven or light-field-driven mechanism at lower power, thereby reducing energy consumption while protecting the device from damage. This invention reduces the external field strength required to achieve the ultrafast process of light-field driving, significantly reducing the possibility of device burn-out, while simultaneously increasing the photoresponse current from the pA level to the nA level, contributing a feasible technical solution for the practical application of optical frequency electronic devices.
[0024] 2. In this invention, the fabrication method of the light-field driven gallium nitride-gold asymmetric nanotunnel junction device employs a geometric correction method. This method reserves a certain space for the influence of the proximity effect and designs the pattern in the opposite direction to compensate for the deformation that may occur due to the proximity effect, thereby obtaining a pattern with a gap of the target size. This invention uses three factors affecting the gap size—pattern gap, electron beam dose, and development time—as variables, and continuously experiments with the optimal parameter combination using the controlled variable method. Attached Figure Description
[0025] Figure 1 A schematic diagram of the structure of a light-driven gallium nitride-gold asymmetric nanotunnel junction device;
[0026] Figure 2 Scanning electron microscope image of a gallium nitride-gold asymmetric nanotunnel junction device with a sub-ten nanometer gap driven by a light field;
[0027] Figure 3The static emission behavior of the light field-driven gallium nitride-gold asymmetric nanotunnel junction device is shown in the figures: (a) IV characteristic curve (gap 7.43 nm); (b) IV characteristic curve (gap 52.9 nm); (c) FN fitting (gap 7.43 nm).
[0028] Figure 4 To illustrate the effect of asymmetric structures on photocurrent, (a) a comparison of IV characteristic curves of the device under white light and pulsed laser illumination; (b) the dependence of photoresponse current on the relative position of the light spot; (c) the effect of homogenizers on photocurrent stability; and (d) scanning electron microscope images comparing the curvature of the two tips.
[0029] Figure 5 Simulation of electric field distribution at an asymmetric tip;
[0030] Figure 6 A comparison of positive bias IV characteristic curves under different laser powers (0.5mW, 1.0mW, 1.5mW, 2.0mW);
[0031] Figure 7 A comparison of the IV characteristic curves for continuous laser and pulsed laser incident light;
[0032] Figure 8 The photocurrent-power dependence of devices with different gaps;
[0033] Figure 9 A schematic diagram of the two-photon emission mechanism of a gallium nitride-gold bowjunction device;
[0034] Figure 10 The photocurrent-optical power dependence diagrams are shown for incident light wavelengths of 750nm and 800nm. Detailed Implementation
[0035] Example 1: Please refer to Figures 1-10 A light-driven gallium nitride-gold asymmetric nanotunnel device, comprising a gallium nitride substrate, wherein a bowtie-shaped nanoantenna is deposited on the upper surface of the gallium nitride substrate, the bowtie-shaped nanoantenna being made of gold film material, and a gap being provided at the center of the bowtie-shaped nanoantenna, the width of the gap being on the sub-ten nanometer scale.
[0036] In this embodiment, a symmetrical bowtie-type nanoantenna was constructed on a silicon dioxide substrate. Its excellent local enhancement effect was demonstrated through photoresponse testing and field enhancement factor simulation. When the nanogap of the bowtie-type antenna is reduced to an extremely small size (sub-10 nanometers), the tip asymmetry caused by process errors will strongly affect the field distribution of the nanogap, making it possible for this device to generate directional current simply through photoexcitation. Nitride semiconductor materials themselves have significant advantages in high-frequency electronic devices due to their wide bandgap, continuously tunable bandgap, high electron saturation velocity, high breakdown voltage, and stable physicochemical properties. Currently, the excellent electrical performance of gallium nitride nanotunnel junctions has been demonstrated.
[0037] Based on the above two points, a bowtie antenna with a sub-ten nanometer gap was fabricated on a gallium nitride substrate, and a novel optical field-driven emission mechanism was proposed through optical response testing and field strength distribution simulation.
[0038] This embodiment, by designing a low emission barrier and fabricating a nanotunnel junction with an extremely short channel, facilitates the device to enter photon-driven or optical field-driven mechanisms at lower power, thereby reducing energy consumption while protecting the device from damage. A series of photoelectric response tests were conducted on the device, revealing that it exhibits both photon-driven and optical field-driven mechanisms at different incident photon energies. When the photon energy is greater than 1.7 eV, the gallium nitride two-photon absorption mechanism dominates. Even at zero bias, due to the non-uniform distribution of the laser-induced alternating electric field, a relatively high response current (on the order of nA) still exists, and this current reverses with the shift in the light spot position, which has certain guiding significance for fabricating femtosecond-level optically controlled devices. When the photon energy is less than 1.7 eV but greater than 1.1 eV, electron absorption at the Gold Fermi level, which allows photons to cross the Schottky barrier (1.1 eV), dominates.
[0039] This embodiment utilizes an ultrafast laser-induced asymmetric bowtie antenna on a gallium nitride substrate to observe and verify an ultrafast electron emission process driven solely by a light field under zero bias voltage. Leveraging the excellent physical properties of nitride semiconductors, such as tunable bandgap, other purely optically controlled ultrafast electronic devices on nitride substrates can be fabricated in practical applications, contributing to the development of ultra-high sensitivity photoelectric detectors in near-infrared and other bands. Furthermore, on-chip integrated waveguides can be used for ultrafast optical signal transmission over micrometer-scale distances. The prototype device in this embodiment converts these signals into high-frequency electrical signals, overcoming the power consumption and speed limitations of traditional electronically controlled devices, potentially breaking through the bottleneck of Moore's Law and constructing photonic chips for optoelectronic interconnection. This embodiment reduces the external field strength required to achieve the light field-driven ultrafast process, significantly reducing the possibility of device burn-out, while simultaneously increasing the photoresponse current from the pA level to the nA level, contributing a feasible technical solution for the practical application of optical frequency electronic devices.
[0040] The specific working principle of this embodiment is as follows:
[0041] 1. Rectification phenomenon during static emission of gallium nitride-gold collar junction nanotunnel devices
[0042] First, the static emission performance of the device was tested. Device gaps of 7.43 nm and 52.9 nm were selected, and their IV characteristic curves under dark conditions were measured, as shown below. Figure 3 As shown in (a) and (b), the emission current under positive and negative bias is not consistent, exhibiting obvious rectification characteristics. Furthermore, the rectification effect of the 7.43nm gap device is more pronounced than that of the 52.9nm device, and its IV characteristic curve at low voltage is similar to that of a tunneling diode. This is because the fabricated bowtie structure cannot guarantee that the geometry of the two tips is completely identical, resulting in an asymmetrical electric field distribution at both ends. This uneven electric field distribution becomes more pronounced as the gap becomes smaller. Subsequently, FN fitting was performed on the IV characteristic curve of the 7.43nm device under forward bias, as shown... Figure 3 (c) indicates that its linear relationship shows that it is the tunneling current of field emission.
[0043] 2. Two-photon emission mechanism of gallium nitride-gold collar junction nanotunnel devices
[0044] The photoresponse of a device with a 7.43 nm gap was tested. First, the bowtie-shaped device was illuminated at its center of symmetry using white light and a 700 nm wavelength, 1 mW femtosecond pulsed laser. A bias voltage was applied, and its IV characteristic curve was measured. Figure 4 As shown in (a), both exhibit relatively obvious rectification characteristics, and it was observed that the photoresponse current is not zero when the applied bias voltage is 0V. If the tips are perfectly symmetrical and the electric field distribution around the gap is also perfectly symmetrical, then a DC signal cannot exist macroscopically without an applied electric field. Considering the limitations of the experimental equipment, there is no absolutely zero bias voltage; the dark current at 0V was measured to be approximately 10. -14 A. The influence of the source meter's built-in voltage was eliminated. Therefore, the photocurrent generated under zero bias was attributed to ultrafast laser induction. This was confirmed by observing the scanning electron microscope image of the device (...). Figure 4 (b) It can be seen that the radius of curvature of the upper tip is significantly smaller than that of the lower tip. This suggests that the asymmetric structure generates an inhomogeneous strong field under the induction of ultrafast laser, thereby driving electron transport between nanotunnel junctions. Figure 4 In this study, the influence of asymmetric structures on photocurrent is shown in the following figures: (a) Comparison of IV characteristic curves of the device under white light and pulsed laser illumination; (b) Dependence of photoresponse current on the relative position of the light spot; (c) Influence of homogenizer on photocurrent stability; and (d) Scanning electron microscope images comparing the curvature of the two tips.
[0045] Without a beam-shaping homogenizer in the optical path, the instability of the photocurrent increases significantly, a situation not observed in previous tests of the photoresponse of single-crystal gold nanotunnel junction devices. The It curves with and without a beam homogenizer in the optical path were compared, as shown below. Figure 4 As shown in (c). Without a homogenizer, the slight jitter of the optical platform will change the relative position of the light spot and the center of symmetry of the bowtie on the nanoscale, thereby changing the electric field distribution between them. This indicates that the device is very sensitive to the intensity distribution of the electric field. Combined with the IV characteristic curve of the non-tunneling behavior in 4(a), it reflects from the side that its electron emission mechanism is fundamentally different from that of single-crystal gold nanotunnel devices.
[0046] The homogenizer in the optical path was then removed, and the illumination position of the light spot was adjusted. The dependence of the photoresponse current under zero bias voltage on the displacement of the light spot center relative to the device's center of symmetry was recorded. Figure 4 As shown in (d), the current direction is positive from end B to end A. This design aims to roughly adjust the illumination intensity at the two tips using the Gaussian distribution of the laser beam intensity. Experimental results show that the current is positive when the laser beam is irradiated at the center; as the center of the beam moves from one end of the bowtie antenna to the other, the direction of the photoresponse current changes. Since the beam diameter is 2.5 μm and the relative gap is 7.43 nm, which is much larger, it can be assumed that when the beam center coincides with the center of symmetry of the bowtie antenna (i.e., the relative displacement is zero), the illumination intensity at the two tips is near the extreme value of the Gaussian distribution function, with minimal difference. At this point, the asymmetric distribution of the electric field at both ends is mainly caused by the asymmetric structure. The photoresponse current flows from end B to end A, indicating that the electric field around end A is much stronger than that around end B due to the higher field enhancement factor. Therefore, under time integration, the number of electrons transported from end A to end B is much greater than that transported from end B to end A, macroscopically manifesting as a DC signal from end B to end A. When the center of the light spot is relatively displaced from the center of the bowtie antenna, the illumination intensity at both ends decreases significantly, and the intensity difference gradually increases. The asymmetric distribution of the electric field at both ends is mainly due to the non-uniformity of the illumination intensity. When the light spot moves towards end B, the electric field distribution near end B gradually becomes stronger than that at end A, thus reversing the current direction. Simultaneously, when the center of the light spot moves the same distance in the opposite direction, the absolute value of the resulting photoresponse current is larger when the light spot is biased towards end A, which has a smaller radius of curvature. This further demonstrates that the asymmetric structure generates an asymmetric electric field distribution under ultrafast laser induction.
[0047] To investigate how asymmetric structures affect the electric field distribution, simulations were performed using COMSOL Multiphysics software. The radius of curvature at the tip was set to 10 nm (A) and 5 nm (B), the excitation light was set to 700 nm, and the gap range was adjusted (10 nm to 100 nm, with a step size of 10 nm). The device's geometric parameters and material properties were consistent with actual measurements. The gold-gallium nitride interface was selected, and the electric field distribution and field enhancement factor were analyzed. The results are as follows: Figure 5 As shown, the field enhancement factor around the tip with a small radius of curvature is significantly higher than that around the tip with a large radius of curvature, and the gap between the two gradually decreases as the gap increases, further verifying the strong influence of the asymmetric structure on the field distribution at extremely small gaps. Through the above tests, it can be determined that the asymmetric nanotunnel junction undergoes a directional electron transport process induced by ultrafast laser.
[0048] Next, the electron generation mechanism was verified. The positive bias I-V characteristic curves under different laser powers (0.5mW, 1.0mW, 1.5mW, 2.0mW) were tested and compared, as shown in Figure 6(a). The results showed that as the laser power increased, ① the zero bias photoresponse current continuously increased; ② the I-V characteristic curve gradually tended towards linear correlation. This indicates that electron generation or transport is related to laser power, and the electron emission process is not a tunneling process. When the laser power is low, the excited alternating electric field is weak, and the space charge effect is obvious, significantly restricting electron transport. As the laser power increases, the strong alternating electric field can transport electrons into the metal in a timely manner, the space charge effect weakens, and the macroscopically, the I-V characteristic curve gradually tends towards linearity. Subsequently, with a step size of 10nm, the wavelength dependence curves of the photoresponse current in the range of 680–800nm were tested (applied bias voltages of 0V, 2V, 4V, and 8V, laser power of 0.1mW), as shown in Figure 6(a). Figure 6 As shown in (b), a sharp drop in photoresponse current occurs in the laser wavelength range of 720nm–740nm under different bias voltages, proving that the photoresponse current of this device is related to the incident photon energy. The band gap E of gallium nitride... gap The energy of a photon at a wavelength of 732 nm is approximately 3.4 eV, while the energy of a photon at 732 nm is approximately 1.7 eV, which is half the band gap of gallium nitride. gap / 2 The results match, suggesting that the electrons originate from the gallium nitride valence band, absorb two photons, and transition to the conduction band, where they are driven by a strong alternating electric field induced by ultrafast laser light. The dependence of the photoresponse current on the optical power at zero bias was then tested, such as... Figure 6 As shown in (c), it is clear that the response current and optical power exhibit a quadratic power-law dependence when the light wavelength is 700 nm and 720 nm, further verifying the above conjecture that electrons absorb two photons and emit them. Figure 6In the following sections, the verification of the two-photon emission mechanism is shown: (a) IV characteristic curves under different power laser irradiation; (b) photoresponse current-wavelength dependence under different bias voltages; (c) photoresponse current-optical power dependence under different wavelength laser irradiation; and (d) polarization dependence of the photoresponse current.
[0049] Next, the polarization dependence of the photoresponse current under zero bias was tested, and the results are as follows: Figure 6 As shown in (d), 0° corresponds to the light polarization direction being perpendicular to the bowtie antenna axis, and 90° corresponds to the light polarization direction being parallel to the bowtie antenna axis. It was observed that the photoresponse current of the device exhibits a strong polarization dependence under zero bias. The photocurrent when the polarization direction is parallel to the bowtie antenna axis is 30 times larger than the photocurrent when it is perpendicular to the bowtie antenna axis, indicating that the laser with parallel polarization has a very significant driving effect on the electron transport of the bowtie antenna. Furthermore, the power dependence and polarization dependence of the photoresponse current under zero bias also indicate that the main contribution of the current is the driving force of the light field, while other processes such as thermionic emission contribute very little.
[0050] Subsequently, the device center was illuminated with both femtosecond pulsed lasers and continuous lasers at the same power density, and its IV characteristic curves were tested. Since the spot sizes of the two lasers were different, the power density was obtained by dividing the measured power value by the spot area to ensure consistent power density during testing. It was found that the photoresponse current of the femtosecond pulsed laser was about two orders of magnitude larger than that of the continuous laser. Figure 7 As shown. This is because the average optical power density of the two can only be controlled to be the same by an optical power meter. Pulsed lasers output an extremely strong pulse in a very short time, and their peak power is much higher than that of continuous lasers. Therefore, the driving field formed by pulsed lasers is much stronger than that of continuous lasers. This experiment initially shows that pulsed lasers drive an ultrafast emission process, rather than a simple optical excitation process. Therefore, the response speed of the photoresponse current is theoretically comparable to the optical period, i.e., the "optical frequency".
[0051] The photocurrent-power dependence of devices with different gaps was subsequently tested, such as... Figure 8 As shown, their power-law dependence is between 1 and 2. This is mainly because the field enhancement factor decreases sharply when the gap is large, resulting in a weaker driving electric field that cannot transport electrons that have transitioned to the conduction band to the other side in time. At the same time, the space charge effect also reduces the photocurrent. Therefore, the power law of the IP curve is less than 2.
[0052] In summary, a clear physical image of this embodiment has been obtained, as shown in the schematic diagram. Figure 9As shown, when the photon energy of the incident light is greater than 1.7 eV (i.e., the wavelength is less than 732 nm), electrons in the gallium nitride valence band only need to absorb two photons to transition to the conduction band. At this time, an asymmetrically distributed strong alternating electric field is formed between the two ends of the bowtie-shaped gold antenna by the pulsed laser. Electrons in the conduction band are emitted from one end to the other under the influence of the strong field. Through time integration, this macroscopically manifests as a DC signal, which is the main contributor to its photoresponse current. The electrons in the conduction band have higher energy levels and faster movement speeds, so a response current on the order of nA can be obtained under zero-bias, 1 mW pulsed laser incident light. At the same time, the photocurrent response time generated based on this mechanism is extremely short, theoretically comparable to the optical period.
[0053] 3. Other emission mechanisms of gallium nitride-gold collar junction nanotunnel devices
[0054] When testing the wavelength-dependent characteristics of the photoresponse current, it was found that when the photon energy is less than 1.7 eV (i.e., when the wavelength is greater than 732 nm), there is still a weak photoresponse current (on the order of pA). At this time, the electrons in the valence band cannot cross the band gap of 3.4 eV after absorbing two photons, indicating that there may be other emission mechanisms.
[0055] First, the photocurrent-optical power dependence of the device with a gap of 7.43 nm was tested at incident light wavelengths of 750 nm and 800 nm. Figure 10 As shown, they all exhibit similar two-stage power-law dependencies: the first stage is a linear power-law dependency, and the second stage is a quadratic power-law dependency, differing only in the inflection point. The linear power-law dependency in the first stage is relatively easy to understand. Based on the work function of gold (5.1 eV) and the electron affinity of intrinsic gallium nitride (4.0 eV), the Schottky barrier height between gold and gallium nitride can be calculated to be 1.1 eV (Φ = W(Au) - χ(GaN) = 5.1 eV - 4.0 eV = 1.1 eV). When the photon energy of the incident light is greater than 1.1 eV, an electron at the gold Fermi level can absorb a photon and cross the Schottky barrier to emit light. It is noteworthy that both linear relationships were tested with an applied bias voltage of 1V; without a bias voltage, the photocurrent is very weak (10). -13 A) Therefore, electrons crossing the Schottky barrier are primarily transported under the influence of an applied electric field. The second power-law dependence may be due to the sufficiently high peak field strength of the pulsed laser compressing the band gap of gallium nitride, allowing valence band electrons to absorb two photons and emit them into the conduction band of gallium nitride, where they are then transported under the influence of an applied electric field. The two IP curves also show that the inflection point at 750nm is earlier than at 800nm, indicating that a lower optical field strength is required to enter the two-photon emission mechanism when the incident photon energy is high. This, to some extent, corroborates the previous hypothesis about strong-field-induced band gap compression.
[0056] Example 2: Please refer to Figures 1-10 The difference between the light field-driven gallium nitride-gold asymmetric nanotunnel device and Example 1 is that the gallium nitride substrate has a length and width of 1cm × 1cm, and the bowtie-shaped nanoantenna includes a pair of symmetrically distributed isosceles triangular units with a gap between the two units. The gap width of the bowtie-shaped nanoantenna is less than 15nm.
[0057] Example 3: Please refer to Figures 1-10 The light field-driven gallium nitride-gold asymmetric nanotunnel junction device differs from Example 1 in that the gallium nitride substrate is unintentionally doped intrinsic gallium nitride (~3.0μm uGaN / 650μm Sapphire).
[0058] In this embodiment, the gallium nitride substrate is intrinsic gallium nitride that is not intentionally doped, which ensures almost no leakage current while maintaining its crystal quality and avoiding defects introduced by doping.
[0059] Example 4: A method for fabricating a light-field driven gallium nitride-gold asymmetric nanotunnel junction device, comprising the following steps:
[0060] Step 1: Substrate cleaning: Select an intrinsic gallium nitride substrate that is not intentionally doped (~3.0μm uGaN / 650μm Sapphire), cut it into 1cm×1cm pieces by mechanical cutting, and ultrasonically clean it in acetone, isopropanol and deionized water for 10 minutes respectively. After cleaning, remove it and blow it dry with a nitrogen gun, and place it in a vacuum adsorption box for later use.
[0061] Step 2, Spin Coating: PMMA 950K photoresist was selected, the spin coating speed was fixed at 3000 r / min, the pre-baking temperature was 180℃, and the time was 2 min. The thickness of the photoresist was measured to be about 330 nm using a step meter.
[0062] Step 3, Electron Beam Exposure: Use L-Edit software to draw a series of bowtie-shaped patterns with equal gaps (including electrodes, where the height and base of the isosceles triangle in the bowtie-shaped pattern are both set to 5um). Using the geometric correction method, the gap range on the layout is from -10nm to +10nm, with a step size of 2nm.
[0063] Before exposure, a 10nm aluminum film was deposited by magnetron sputtering to improve the conductivity of the substrate. The electron beam dose variation range during exposure was selected as 800μC / cm2~1000μC / cm2 with a step size of 50μC / cm2. After exposure, the aluminum film was removed with an alkaline solution, then rinsed in ultrapure water and dried with a nitrogen gun.
[0064] The developing solution used is a pre-prepared developing solution (4-methyl-2-pentanone: isopropanol = 1:3) and a fixing solution (isopropanol). The developing time is adjustable from 40s to 70s with a step size of 5s, and the fixing time is fixed at 30s.
[0065] Step 4, Evaporation and Stripping: The developed sample is deposited with a 40nm gold film by electron beam evaporation. After evaporation, it is placed in acetone heated to 85°C in a water bath for stripping. After the gold film except for the bowtie structure and electrodes is removed, it is cleaned in isopropanol heated to 85°C in a water bath, dried with a nitrogen gun, and placed in a vacuum adsorption box for testing.
[0066] The technical principle of this embodiment:
[0067] The primary factor limiting resolution at pattern gaps is the proximity effect of electron beam exposure, which originates from electron scattering within the sample. The lateral spread of scattered electrons is much larger than the electron beam spot diameter (up to several micrometers). The radiation energy absorbed at each point in the resist is the sum of direct and surrounding radiation energy. When the pattern linewidth and gaps are small enough to match the size of the scattering diffusion range, scattered electrons will severely impact the exposure of neighboring patterns. To address the low contrast and resolution degradation caused by the proximity effect, the main correction methods include dose correction and geometric correction.
[0068] This embodiment employs a geometric correction method. The principle is to reserve a certain amount of space for the effects of proximity, designing a pattern in the opposite direction to compensate for potential deformation caused by proximity effects. Specifically, in the exposure of a bowtie antenna, this means that the actual gap size of the fabricated device is usually slightly larger than the gap size designed in the layout. This embodiment compensates for this difference by designing a negative distance (i.e., overlapping of the leading edges of the two tips) gap in the layout, thereby obtaining a pattern with the target gap size. In the design of the limit gap process, this embodiment uses three factors affecting the gap size—layout gap, electron beam dose, and development time—as variables, continuously experimenting with the optimal parameter combination using the controlled variable method.
[0069] Example 5: A method for fabricating a light-field driven gallium nitride-gold asymmetric nanotunnel junction device, which differs from Example 4 in that, in step three, the gap range on the layout is set to -2nm.
[0070] In this embodiment, through continuous experimentation using the controlled variable method, the optimal value for the gap on the printing plate was found to be -2nm.
[0071] Example 6: A method for fabricating a light-field driven gallium nitride-gold asymmetric nanotunnel junction device, which differs from Example 4 in that, in step three, the electron beam dose during exposure is selected as 900 μC / cm2.
[0072] This embodiment, through continuous experimentation using the controlled variable method, determined that the optimal electron beam dose during exposure is 900 μC / cm2.
[0073] Example 7: A method for fabricating a light-field driven gallium nitride-gold asymmetric nanotunnel junction device, which differs from Example 4 in that the development time in step three is set to 55s.
[0074] This embodiment, through continuous experimentation using the controlled variable method, determined that the optimal development time is 55 seconds.
[0075] Example 8: A method for fabricating a light-field driven gallium nitride-gold asymmetric nanotunnel junction device, which differs from Example 4 in that, in step four, no adhesion layer is added between the 40nm gold film deposited by electron beam evaporation and the gallium nitride substrate;
[0076] There are two reasons for not adding an adhesive layer in this embodiment: firstly, the gold film has good adhesion to the gallium nitride substrate, and not adding an adhesive layer reduces the difficulty of peeling; secondly, it makes the overall structure of the device as simple as possible, which facilitates subsequent testing and analysis.
Claims
1. A light-field driven gallium nitride-gold asymmetric nanotunnel junction device, characterized in that, The invention includes an intrinsic gallium nitride substrate that is not intentionally doped, on the upper surface of which a bowtie-shaped nanoantenna is deposited. The bowtie-shaped nanoantenna is made of gold film material, and no adhesion layer is added between the gold film and the intrinsic gallium nitride substrate. The bowtie-shaped nanoantenna includes a pair of symmetrically distributed isosceles triangular units, with a gap between the opposite tips of the pair of isosceles triangular units. The width of the gap is on the sub-ten nanometer scale, and the curvature of the tips of the pair of isosceles triangular units is asymmetrical.
2. The light-field driven gallium nitride-gold asymmetric nanotunnel device according to claim 1, characterized in that: The intrinsic gallium nitride substrate has a length and width of 1cm × 1cm, the height and base of the isosceles triangular monomer are both 5μm, and the gap width is less than 10nm.
3. The light-field driven gallium nitride-gold asymmetric nanotunnel device according to claim 1, characterized in that: The intrinsic gallium nitride substrate consists of a 3.0 μm thick unintentionally doped intrinsic gallium nitride epitaxial layer on a 650 μm thick sapphire substrate.
4. A method for fabricating a light-field driven gallium nitride-gold asymmetric nanotunnel junction device, characterized in that, Includes the following steps: Step 1: Substrate cleaning: Select a substrate with an unintentionally doped intrinsic gallium nitride epitaxial layer of 3.0 μm thickness on a sapphire substrate of 650 μm thickness. Cut it into 1 cm × 1 cm pieces using mechanical cutting method. Place them in acetone, isopropanol, and deionized water for ultrasonic cleaning for 10 minutes respectively. After cleaning, remove them and blow them dry with a nitrogen gun. Place them in a vacuum adsorption box for later use. Step 2, Spin Coating: PMMA 950K photoresist was selected, the spin coating speed was fixed at 3000 r / min, the pre-baking temperature was 180℃, and the time was 2 min. The thickness of the photoresist was measured to be 330 nm using a step meter. Step 3, Electron Beam Exposure: Use L-Edit software to draw a series of bowtie-shaped patterns and electrodes with equal gaps. The height and base of the isosceles triangle in the bowtie-shaped pattern are both set to 5µm. Using the geometric correction method, the gap range on the layout is from -10nm to +10nm, with a step size of 2nm. Before exposure, a 10nm aluminum film was deposited by magnetron sputtering to improve the conductivity of the substrate. The electron beam dose variation range during exposure was selected as 800μC / cm2~1000μC / cm2 with a step size of 50μC / cm2. After exposure, the aluminum film was removed with an alkaline solution, then rinsed in ultrapure water and dried with a nitrogen gun. The developing solution used is a pre-prepared developing solution, which is a mixture of 4-methyl-2-pentanone and isopropanol in a volume ratio of 1:
3. The fixing solution is isopropanol. The developing time is adjusted in the range of 40s to 70s with a step size of 5s, and the fixing time is fixed at 30s. Step 4, Evaporation and Stripping: The developed sample is deposited with a 40nm gold film by electron beam evaporation. After evaporation, it is placed in acetone heated to 85°C in a water bath for stripping. After the gold film except for the bowtie structure and electrodes is removed, it is cleaned in isopropanol heated to 85°C in a water bath, dried with a nitrogen gun, and placed in a vacuum adsorption box for testing.
5. The method for fabricating the light-field driven gallium nitride-gold asymmetric nanotunnel junction device according to claim 4, characterized in that: In step three, the gap range on the layout is set to -2nm.
6. The method for fabricating the light-field driven gallium nitride-gold asymmetric nanotunnel junction device according to claim 4, characterized in that: In step three, the electron beam dose during exposure is selected as 900 μC / cm2.
7. The method for fabricating the light-field driven gallium nitride-gold asymmetric nanotunnel junction device according to claim 4, characterized in that: In step three, the development time is set to 55 seconds.
Citation Information
Patent Citations
Crack structures, tunneling junctions using crack structures and methods of making same
CN108700542A
