Open Access
Add to BibSonomy
Abstract
In the field of diamond MESFETs, this work is what we believe to be the first to investigate the optoelectronic properties of hydrogen-terminated polycrystalline diamond MESFETs under visible and near-UV light irradiation. It is shown that the diamond MESFETs are well suited for weak light detection in the near-ultraviolet region around the wavelength of 368 nm, with a responsivity of 6.14 × 106 A/W and an external quantum efficiency of 2.1 × 107 when the incident light power at 368.7 nm is only 0.75 µW/cm2. For incident light at 275.1 nm, the device's sensitivity and EQE increase as the incident light power increases; at an incident light power of 175.32 µW/cm2 and a VGS of -1 V, the device's sensitivity is 2.9 × 105 A/W and the EQE is 1.3 × 106. For incident light in the wavelength range of 660 nm to 404 nm with an optical power of 70 µW/cm2, the device achieves an average responsivity of 1.21 × 105 A/W. This indicates that hydrogen-terminated polycrystalline diamond MESFETs are suitable for visible and near-UV light detection, especially for weak near-UV light detection. However, the transient response test of the device shows a long relaxation time of about 0.2 s, so it is not yet suitable for high-speed UV communication or detection.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The ultraviolet (UV) light in the wavelength range of 200 nm to 280 nm is known as the “solar-blind” waveband [1–3], because UV light in this waveband is almost completely absorbed by the ozone in the atmosphere. Detectors operating in the solar-blind waveband can reduce the interference of sunlight, so the sun-blind UV detector has the advantages of low background interference, high signal-to-noise ratio, and strong detection ability of small signals, etc. For day-blind detectors, the forbidden band width of the constituent detector materials needs to be greater than 4.43 eV. This means that common Si detectors or other pyroelectric detectors are not up to the task. Therefore, most of the materials used to prepare day-blind detectors are AlN, high Al-component AlGaN, gallium oxide and diamond. All three materials suffer from low mobility and difficult activation of doped impurities. For the existing material system, the improvement route of the device includes methods to improve the material crystal quality, optimize the device structure, and find suitable doping elements. Diamond is an ultra-wide band semiconductor material with high electron mobility (2000 cm2·V-1·s-1), high electron saturation rate (2 × 107 cm·s-1), high breakdown field strength (1 × 107 V·cm-1) and high thermal conductivity (2000 W·m-1·K-1), and its band width is 5.47 eV, which corresponds to the absorption edge at 225 nm. It means that diamond is a suitable material for making solar-blind ultra-violet photodetectors (UVPDs) [4], which is suitable for direct UV signal detection or other systems requiring UV detectors, such as sensors for fire alarm systems, receivers for UV fiber optic communication systems, etc. [5] However, the research on diamond UVPDs is greatly limited by the diamond doping challenges, especially the inability to achieve shallow energy level doping. In 1989, Shiomi et al. reported a planar type metal–semiconductor field effect transistor (MESFET) has been fabricated on a boron-doped diamond epitaxial film for the first time [6]. Boron impurities in diamond are difficult to ionize at room temperature, resulting in low carrier concentration and affecting device performance [7,8]. In 1982, Pate et al. illustrated for the first time the effect of hydrogen termination of surface dangling bonds on diamond. It’s found that atomic hydrogen restores the electronic structure of the (111) diamond from a 2 × 2/2 × 1 surface structure to that of the 1 × 1 surface structure [9]. In 1989, Landstrass et al. found the carrier transport properties of diamond films related to defects in the hydrogen passivation layer. In 1992, Maki et al. research found that there were at least three acceptor levels in hydrogen-terminated diamonds (H-diamonds) by Seebeck effect measurement, where the deepest energy level position and density was estimated to be 225 meV and 2 × 1020 cm-3 at room temperature. This resulted in the majority of carriers in the H-diamond being holes and the carrier concentration was estimated to reach 3 × 1017 cm-3 [10]. In 1995, Hayashi Maki et al. research found the high-conductivity layer formed in the as-deposited diamond films originates from the incorporated hydrogen in the subsurface region [11]. In order to solve the problem of low carrier concentration in diamond, H-diamonds have been used to fabricate phototransistors [12]. Compared to pin diode, Schottky diode, metal-semiconductor-metal and other two-port photodetectors, metal-semiconductor field-effect transistor structure phototransistors have higher responsivitys and lower dark currents [13]. In 2022, Pengju et al. have fabricated a MESFET phototransistor based on the exfoliated Ga2O3 microflake and graphene thin film, which shows excellent performance with a responsivity of 2.82 × 103 A/W, a rejection ratio of 5.88 × 105, and a detectivity of 2.67 × 1015 Jones under 254 nm illumination. This suggests that MESFETs made with H-diamond are also suitable for UVPDs [14]. So far, relatively few diamond-based detectors have been reported, and there is a lack of research on the response characteristics of photons whose energies are less than the band gap of diamonds. These response characteristics are mainly derived from the defect levels in diamonds [15–18].
In this work, as far as we know in the field of diamond MESFET research, we have investigated the optoelectronic properties of hydrogen-terminated polycrystalline diamond MESFETs under visible and near-ultraviolet illumination for the first time, specifically to investigate the effect of illumination in the wavelength range of 660 nm to 275 nm on the responsivity and external quantum efficiency (EQE) of the diamond MESFETs. Specifically, the effects of incident light wavelength and gate voltage on device responsivity, photo-to-dark-current ratio (PDCR) and EQE were first investigated for an optical power fixed at 175.32 µW/cm2. In this case, the device shows a trend that the responsivity and EQE increase with decreasing wavelength. Then, the relationship between incident light power and device responsivity, PCDR and EQE is investigated for incident light wavelengths of 368.7 nm and 275.1 nm, respectively. The responsivity of the device decreases with the increase of the incident optical power when the incident wavelength is 368.7 nm, but increases with the increase of the incident optical power when the incident optical power is 275.1 nm. It is worth mentioning that when the incident light wavelength is 368.7 nm and the optical power is only 0.75 µW/cm2, the responsivity of the device is as high as 6.14 × 106 A/W and the EQE is as high as 2.1 × 107, indicating that the device is well suited for near-UV weak light detection and has the potential to become a visible-UV broad spectrum detector.
2. Experiment
The H-diamond MESFETs used in experimental research were fabricated by polycrystalline diamond films, which was epitaxial grown on homoepitaxial substrates by microwave plasma chemical vapor deposition (MPCVD) system. Before the device fabrication process, the polycrystalline diamond film was placed in a hydrogen plasma treatment system to produce H-terminated diamond surfaces, the ambient temperature during processing was 900 °C, and the processing time was 15 min. This step was designed to generate two-dimensional hole gas(2DHG) near the diamond surface to increase the carrier concentration in the diamond [12,18,19]. Then, An Au film was deposited on the area where the device needs to be fabricated to form ohmic contacts. The thickness of Au film was about 80 nm, which can protect the H-diamond surface from being destroyed in the next step. Subsequently, non-conductive isolated areas are created by etching oxygen plasma and the processing time was also 15 min. Then, the gate area, source and drain pads were formed by photolithography and Au etching process. The final step, evaporating aluminum by electron beam to form a 100 nm thick gate, whose length and width of the gate metal were about 5 µm and 48 µm respectively.
As shown in the figure, some of the geometric parameters of the device gate area were standardized with the measurement tool of metallographic microscope, the spacing between the gate metal and the source drain of the device is only 1.165 µm, and the width of the gate area is 47.534 µm, so the diamond surface area that can contact the incident light is only 110.75 µm2. As shown in the Fig. 1(c), the whole device looks very flat as the preparation process of the device does not include mesa etching. The area between the source & drain and the gate is the bare H-diamond surface, which is the light absorbing region of the MESFET device. Figure 1(d) shows a schematic diagram of the cross-sectional structure of the device, the face Orientation of homogeneous epitaxy is {100} faces with a thickness of 0.5 mm, the Boron and Nitrogen concentration in the polycrystalline is less than 1 ppb and 5ppb respectively. On the surface of polycrystalline H-diamond, the element hydrogen attracts negatively charged HCO3- and OH- etc. in the air, forming a negatively charged water layer on the surface and inducing the formation of 2DHG near the surface of the device [18–20].
Fig. 1. (a) The photograph of the H-diamond MESFET. (b) A finer photograph of the gate area of the device.
To further understand the material properties in the gate region of the device, the region has been characterized by scanning transmission electron microscopy (STEM) and distribution of the four elements Al, C, Si and O in the region near the edge of the gate electrode has been characterized using energy disperse spectroscopy (EDS). The results are shown in Fig. 2.
Fig. 2. (a) STEM image of the device gate area profile. (b-e) The distribution of the elements Al, C, Si and O in the red box area, which is obtained by EDS-mapping.
As shown in the Fig. 2, The device is vertically split in the gate region by the focused ion beam (FIB). By comparing the different distributions of Al, C, Si and O in the same region, it can be confirmed that the gate is composed of Al metal and there is a clear distribution of O elements at the junction of the Al and C element distribution regions. According to the EDS characterization results, an extremely thin layer can be found between the gate metal and the diamond material, the main composition of which may be AlxO and some functional groups composed of C, H and O elements. This thin layer is most likely the water layer reported in other literature, which itself carries a negative charge and induces a high concentration of 2DHG at the interface of the diamond material [20].
In the experiment, the electrical and optical response characteristics of the H-diamond MESFET device were tested by a semiconductor parameter tester with a high precision module (Agilent B1500) at room temperature. A total of ten different wavelengths of incident light were used, which were all from the LED light source. Testing by spectrometer (AvaSpec ULS3648-SPU2), the central wavelengths of these ten LED light sources are 660.5 nm, 594.4 nm, 543.5 nm, 503.5 nm, 478.3 nm, 407.8 nm, 403.8 nm, 368.7 nm, and 275.1 nm, respectively. During the experiment, the light emitted from the LED is guided through the optical fiber to the microscope objective, and is focused through the objective and then directed vertically onto the sample or optical power probe with a spot diameter of 1.21 cm.
Before each experiment, the device was placed in the dark for at least four hours to allow sufficient relaxation of the captured charge in the trap. By means of an optical power probe (Thorlabs S120VC) placed in the center of the sample stage, the voltage and current values for each LED have been obtained for about 70 µW incident light power. The diaphragm diameter of the optical power probe is 7.13 mm, so the optical power density incident on the device is about 175.32 µW/cm2. For each test, the transfer characteristic curve of the device in the dark was first tested to determine the relaxation condition of the captured charge in the device. When the transfer characteristic curve of the device in the dark was not much different from the results measured in the previous experiments, conducting subsequent experiments on the effect of different wavelengths to the device.
3. Results and discussion
Since H-diamond MESFETs rely on 2DHG conductivity near the surface, the distribution of 2DHG within the diamond material is a key factor affecting the device properties. To characterize the carrier concentration profiles over depths of the device, we tested the CV characteristics of the device at room temperature with an alternating voltage (AC) signal at 5 MHz. The carrier concentration profiles can be obtained as follows [21].
As shown in Fig. 3, when the gate bias voltage is -0.5 V, the capacitance of the device is 2.55 × 10−13 F. When the gate bias voltage becomes -4 V, the capacitance of the device is 4.06 × 10−12 F. The device has a large varactor ratio. The carrier distribution inside the H-diamond can be obtained according to the CV characteristics of the device. According to the calculation results, at 495 nm from the diamond surface, 2DHG appears and the hole concentration is as high as 1.03 × 1019 cm-3, while at 514 nm from the surface, the hole concentration drops to 5.40 × 1016 cm-3, which shows that the width of 2DHG channel is only about 9 nm. The conductive channel is a highly concentrated 2DHG channel with a width of only about 9 nm located near the surface. Due to this characteristic of the device, the channel conductance of the device is almost entirely determined by the concentration of 2DHG, so the traps that can trap charge near the device surface or interface will directly affect the optoelectronic characteristics of the device [22].
Fig. 3. (a) The C–V curves of the H-diamond MESFET. (b) The carrier concentration profiles over depths of the H-diamond.
For optoelectronic devices, the responsivity (R), PDCR, detectivity (D*) and the corresponding EQE can be expressed as follows [23].
where P is the incident light power density, S is the area of light absorption area, q is the elementary electric charge, h is the Plank’s constant, c is the light velocity and λ is the wavelength of incident light.The spectrum of the LEDs used in the experiments were obtained by the spectrometer and the normalized test results have been shown in Fig. 2.
As shown in Fig. 4, there are ten different central wavelengths of LED light sources in total, which vary in light intensity and half-height width. The voltage or current applied to the LED can be adjusted so that different wavelengths of light irradiate the sample with the same optical power. The wavelength of the incident light used for the experiments, the optical power incident into the optical power probe, and the optical power density irradiated to the surface of the device are shown in the following Table 1.
Table 1. Parameters of incident lights used in the experiments
First, the transfer characteristic curves and trans-conductivity of the devices were tested in dark environment and under different wavelengths of light. During the test, the power of the incident light at all wavelengths was 70 µW except for the incident light at 275.1 nm, and the power of the incident light at 275.1 nm was 10 µW as shown in Fig. 5.
In order to investigate the effect of incident light wavelength on device characteristics at the same optical power density, the incident optical power density setting was shown in condition 1-7, 14 and 21 in Table 1, VDS was set to be -8 V, VGS was stepped from 1 V to -1 V with a step interval of 50 mV. The response characteristics of the device were obtained as shown in the Fig. 6.
Fig. 5. The (a) transfer characteristic curves and (b) transconductance of the device at different wavelengths, except for the 275.1 nm light at which the power of other incident light is 70 µW.
Fig. 6. (a)The relationship between the responsivity and VGS of the device at different incident wavelengths, the illustration is in semi-logarithmic coordinate. (b) The maximum value of the responsivity and the corresponding VGS at different incident wavelengths. The size of the marker is proportional to the value of the responsivity.
As shown in the Fig. 6(a), when the gate voltage is less than 0 V, the responsivity of the device increases first and then decreases as the absolute value of the VGS increases under the 660.5 nm, 594.4 nm, 543.5 nm, 503.5 nm, 478.3 nm, 407.8 nm, or 403.8 nm light illuminations. For the 368.7 nm and 275.1 nm incident light illuminations, the responsivity of the device increases with the absolute value of the VGS when the VGS is in the [-1 V, 0 V] range. Because when the VGS is negative, the absolute value of the ID increases by about 11.23 µA for every 50 mV increase in the absolute value of the VGS, the responsivity of the device decreases when the absolute rate of increase of the photocurrent is lower than that of the absolute rate of increase of the dark current.
As shown in the Fig. 6(b), for the 660.5 nm, 594.4 nm, 543.5 nm, 503.5 nm, 478.3 nm, 407.8 nm or 403.8 nm light illuminations, the VGS is in the [-0.8 V, -0.5 V] range when the responsivity of the device reaches its maximum value. Therefore, if the device is used for visible light detection, the gate bias voltage should be set in the interval [-0.8 V, -0.5 V]. In this application, the mean value of the device's responsivity is about 1.21 × 105 A/W.
As shown in the Fig. 7(a), in the VGS region where the responsivity is high, the PDCR value of the device is low, not greater than 5, mainly because the MESFET channel is in the open state and the dark current value is large. As shown in the Fig. 7(b), when the gate voltage is negative and the absolute value is greater than 0.5 V, the EQE of the device is greater than 1.6 × 106 A/W at the UV wavelength of 275.1 nm. For the 594.4 nm, 543.5 nm, 503.5 nm, 478.3 nm, 407.8 nm, or 403.8 nm light illuminations, the EQE is greater than 2.0 × 105.
Fig. 7. Variation of (a) PDCR and (b) EQE of the device with the VGS at different incident wavelengths.
In purpose of studying the effect of incident optical power density on the device, 368.7 nm and 275.1 nm light were chosen as the light sources, and the specific experimental conditions are shown in condition 8-21 in Table 1.
As shown in the Fig. 8(a), the responsivity of the device at VGS = -1 V decreases with increasing incident optical power under light illumination at a wavelength of 368.7 nm. When the incident light power is 0.75 µW/cm2, the detector achieves a responsivity of 6.14 × 106 A/W for 368.7 nm light, indicating that the device is suitable for weak light detection in this wavelength band. In contrast, Fig. 8(b) shows that the responsivity of the device increases with increasing incident optical power at 275.1 nm. As the optical power irradiated to the device increases, the responsivity increases or decreases depending on the photogenerated carrier generation rate. If the incident optical power is not greater than 175.32 µW/cm2, and the rate of increase of incident optical power is kept constant, the increase of photogenerated carrier generation rate will gradually become larger per unit time in the device under 275.1 nm wavelength illumination. Similarly, when the device is illuminated at 368.7 nm and the incident optical power is in the range of [100 µW/cm2, 500 µW/cm2], the increase of the photogenerated carrier generation rate per unit time will gradually decrease if the incident optical power growth rate is kept constant.
Fig. 8. The relationship between device responsivity and VGS under different optical power illumination for (a) 368.7 nm and (b) 275.1 nm wavelengths. The illustration shows the device responsivity at different incident light powers for a gate voltage of -1 V.
As shown in the Fig. 9, the PDCR increases with the increase of optical power when the device is irradiated at 368.7 nm and the incident optical power is less than 500 µW/cm2. Meanwhile, the gate voltage has a large effect on the PDCR value. When VGS is about 0.4 V, the PDCR value of the device is maximum, but currently the device channel is not yet open and the responsivity is low. Similarly, when the incident light wavelength is 275.1 nm and the incident light power is not greater than 175 µW/cm2, the enhancement of the incident light power can increase the value of PDCR and its maximum value occurs at the gate voltage of 0.4 V.
Fig. 9. The relationship between PDCR and VGS under different optical power illumination for (a) 368.7 nm and (b) 275.1 nm wavelengths.
As shown in the Fig. 10, the EQE value of the device decreases with increasing optical power at 368.7 nm illumination and increases with increasing optical power at 275.1 nm illumination. When the power of 368.7 nm light is µW/cm2, the EQE of the device is as high as 2.1 × 107. This performance indicates that the device is suitable for weak light detection near the wavelength of 368 nm. For 275.1 nm light, the EQE of the device will also be greater than 1.1 × 106 if the incident optical power exceeds 125 µW/cm2.
Fig. 10. The relationship between EQE and VGS under different optical power illumination for (a) 368.7 nm and (b) 275.1 nm wavelengths. The illustration shows the EQE at different optic powers with VGS = -1 V.
In order to understand the dynamic characteristics of the device, the response time of the device was tested by using a 213 nm deep ultra violet laser (Brolight MCC-213-1004) with a 700 ps pulse width and a 1 kHz repetition rate as the excitation source. The variation of the device drain current with time at room temperature was recorded by a high-speed oscilloscope (Keysight DSOS604A). Due to the limitations of the test conditions, the gate of the device was not actively loaded with potential during this experiment and the gate was in a floating state.
As shown in the Fig. 11, in the recorded data, three cycles of data were intercepted. During each cycle, the current value decreases from the peak at the end of the light. Due to the presence of traps, there is a process of capturing and releasing charge from the traps during the current drop, resulting in a slower drop in current. Since 2DHG is induced by the negative charge of the water layer, traps on the diamond surface or at the interface between the diamond and the gate metal trap photoelectrons when illuminated, causing the 2DHG concentration to increase and the drain current to increase. After illumination, the traps at the surface or interface that capture electrons begin to release them, causing the 2DHG concentration to decrease and the drain current to decrease. The time it takes for the trap to release electrons is called the trap relaxation time, which can be fitted by the following equation [22,24].
Fig. 11. The impulse response of the device under a 213 nm deep ultra violet laser with a 700 ps pulse width and a 1 kHz repetition rate.
Table 2. Fitting parameters of the detector impulse response model
As shown in Table 2, the fit between the model used and the test data was good, with R-square greater than 0.95 for all groups. The fitted parameters show that there are two main types of traps in the material that capture photoelectrons, one of which has a low percentage and a short relaxation time of about 6.514 × 10−5 s, and the other has a high percentage but a long relaxation time of about 0.216 s. After the end of illumination, the traps with short relaxation time release electrons rapidly, leading to a rapid decrease in drain current, but the traps with short relaxation time have lower density. Therefore, the latter is the main role of the trap with longer relaxation time, which shows a long and slow decrease of the drain current. The response time of the diamond device is about 0.2 s, which is not suitable for high-speed detection, in terms of the time to release electrons from the trap, i.e., the relaxation time.
When the energy of the incident photon is less than the forbidden band width of diamond, although the electron in the valence band cannot be excited to enter the conduction band, it can be excited to enter the energy level corresponding to the trap in the forbidden band, resulting in the capture of the electron by the trap in the forbidden band. When the trap position is located on the surface of diamond material or near the interface with the gate metal, the trap of captured photoelectrons can cause an increase in the concentration of 2DHG, leading to an increase in the drain current.
In Table 3, the comparison of the key parameters of diamond-based photodetectors in recent years is compared.
Table 3. Comparison of key parameters of diamond-based photodetectors
As shown in Table 3, the detector prepared in this work achieves a detection rate of 6.14 × 106 A/W at a gate voltage of only -1 V, which is the highest among similar detectors, compared with similar detectors. However, the response time of the device is long, and the device structure needs to be optimized to improve the response rate.
4. Conclusion
In conclusion, hydrogen-terminated polycrystalline diamond MESFETs have the potential to become visible and UV detectors, especially for near-UV weak light detection, the devices have very high responsivity and external quantum efficiency. In this experiment, we found that the 2DHG in H-diamond is located about 0.5 µm below the surface and is only about 9 nm wide, with a concentration of 1.03 × 1019 cm-3. Since traps capable of trapping electrons are present on the H-diamond surface or at the interface with the gate, photons with energy less than the forbidden band width of the diamond can excite valence band electrons to the trap energy level and be trapped when they are incident. At this point, the negative charge concentration at the diamond surface increases and induces more 2DHG, which reduces the channel resistance of the device, allowing the device to respond to the incident photons with energy less than the forbidden band width. For weak light incidence at a wavelength of 368.7 nm, the device has a high responsivity and EQE. When the incident light intensity is only 0.75 µW/cm2, the device has a responsivity of 6.14 × 106 A/W and an EQE of 2.1 × 107. For the incident light of 275.1 nm, the responsivity and EQE of the device increase with the increase of the incident light power, when the incident light power is 175.32 µW/cm2 and VGS is -1 V, the responsivity of the device is 2.9 × 105 A/W, the EQE is 1.3 × 106. For incident light in the wavelength range of 660 nm to 404 nm with an optical power of 70 µW/cm2, the device achieves an average responsivity of 1.21 × 105 A/W. The low detectivity of the device is mainly since the absorbed light area of the device is only 110.75 µm2. Transient response testing of the device confirms the existence of at least two traps in H-diamond for capturing photoelectrons with relaxation times of 6.514 × 10−5 s and 0.216 s, respectively. In the subsequent work, the ITO transparent gate will be introduced by the process to increase the absorbed light area and improve the characteristics of the device.
Funding
Guangzhou Science and Technology Project Fund (202201011247, 202201011290, 2023A04J2041).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. J. Xu, W. Zheng, and F. Huang, “Gallium oxide solar-blind ultraviolet photodetectors: a review,” J. Mater. Chem. C 7(29), 8753–8770 (2019). [CrossRef]
2. P. Schreiber, T. Dang, T. Pickenpaugh, G. A. Smith, P. Gehred, and C. W. Litton, “Solar-blind UV region and UV detector development objectives,” Proc. SPIE 3629 – Photodetectors (1999), p. 230. [CrossRef]
3. F. Xie, Y. Gu, Z. Hu, B. Yu, and G. Yang, “Ultra-low dark current back-illuminated AlGaN-based solar-blind ultraviolet photodetectors with broad spectral response,” Opt. Express 30(13), 23756 (2022). [CrossRef]
4. C.-N. Lin, Y.-J. Lu, X. Yang, Y.-Z. Tian, C.-J. Gao, J.-L. Sun, L. Dong, F. Zhong, W.-D. Hu, and C.-X. Shan, “Diamond-Based All-Carbon Photodetectors for Solar-Blind Imaging,” Adv. Opt. Mater. 6(15), 1800068 (2018). [CrossRef]
5. R. Min, B. Ortega, and C. Marques, “Fabrication of tunable chirped mPOF Bragg gratings using a uniform phase mask,” Opt. Express 26(4), 4411 (2018). [CrossRef]
6. H. Shiomi, Y. Nishibayashi, and N. Fujimori, “Field-Effect Transistors using Boron-Doped Diamond Epitaxial Films,” Jpn. J. Appl. Phys. 28(12A), L2153 (1989). [CrossRef]
7. R. F. Mamin and T. Inushima, “Conductivity in boron-doped diamond,” Phys. Rev. B 63(3), 033201 (2001). [CrossRef]
8. K. Muzyka, J. Sun, T. H. Fereja, Y. Lan, W. Zhang, and G. Xu, “Boron-doped diamond: current progress and challenges in view of electroanalytical applications,” Anal. Methods 11(4), 397–414 (2019). [CrossRef]
9. B. B. Pate, M. H. Hecht, C. Binns, I. Lindau, and W. E. Spicer, “Photoemission and photon-stimulated ion desorption studies of diamond(111): Hydrogen,” J. Vac. Sci. Technol. 21(2), 364–367 (1982). [CrossRef]
10. T. Maki, S. Shikama, M. Komori, Y. Sakaguchi, K. S. K. Sakuta, and T. K. T. Kobayashi, “Hydrogenating effect of single-crystal diamond surface,” Jpn. J. Appl. Phys. 31(10A), L1446 (1992). [CrossRef]
11. K. Hayashi, S. Yamanaka, H. Okushi, and K. Kajimura, “Study of the effect of hydrogen on transport properties in chemical vapor deposited diamond films by Hall measurements,” Appl. Phys. Lett. 68(3), 376–378 (1996). [CrossRef]
12. H. Kawarada, “Hydrogen-terminated diamond surfaces and interfaces,” Surf. Sci. Rep. 26(7), 205–206 (1996). [CrossRef]
13. W. Kim and K. S. Chu, “ZnO nanowire field-effect transistor as a UV photodetector; optimization for maximum sensitivity,” Phys. Status Solidi A 206(1), 179–182 (2009). [CrossRef]
14. P. Tan, Y. Zou, X. Zhao, X. Hou, Z. Zhang, M. Ding, S. Yu, X. Ma, G. Xu, and Q. Hu, “Hysteresis-free Ga2O3 solar-blind phototransistor modulated from photoconduction to photogating effect,” Appl. Phys. Lett. 120(7), 071106 (2022). [CrossRef]
15. J. P. Goss, P. R. Briddon, V. Hill, R. Jones, and M. J. Rayson, “Identification of the structure of the 3107 cm −1 H-related defect in diamond,” J. Phys.: Condens. Matter 26(14), 145801 (2014). [CrossRef]
16. S. Dannefaer, “Defects in diamond,” Phys. Status Solidi (c) 4(10), 3605–3613 (2007). [CrossRef]
17. S. Mohapatra, P. K. Sahu, S. Rath, P. K. Sahoo, S. Varma, and N. V. L. Narasimha Murty, “Impact of nitrogen induced defect dynamics on UV response of diamond photodetectors,” Superlattices Microstruct. 142, 106504 (2020). [CrossRef]
18. M. Kasu, K. Ueda, H. Kageshima, and Y. Yamauchi, “Gate interfacial layer in hydrogen-terminated diamond field-effect transistors,” Diamond Relat. Mater. 17(4-5), 741–744 (2008). [CrossRef]
19. A. Kubovic, A. Denisenko, W. Ebert, M. Kasu, I. Kallfass, and E. Kohn, “Electronic surface barrier characteristics of H-terminated and surface conductive diamond,” Diamond Relat. Mater. 13(4-8), 755–760 (2004). [CrossRef]
20. F. Maier, M. Riedel, B. Mantel, J. Ristein, and L. Ley, “Origin of Surface Conductivity in Diamond,” Phys. Rev. Lett. 85(16), 3472–3475 (2000). [CrossRef]
21. H. Kroemer, W.-Y. Chien, J. S. Harris, and D. D. Edwall, “Measurement of isotype heterojunction barriers by C - V profiling,” Appl. Phys. Lett. 36(4), 295–297 (1980). [CrossRef]
22. Y.-B. Liu, J.-Y. Shen, J.-Y. Xing, W.-Q. Yao, H.-H. Liu, Y.-Q. Dai, L.-K. Yang, F.-G. Wang, Y. Ren, M.-J. Zhang, Z.-S. Wu, Y. Liu, and B.-J. Zhang, “Abnormal phenomenon of source-drain current of AlGaN/GaN heterostructure device under UV/visible light irradiation*,” Chin. Phys. B 30(11), 117302 (2021). [CrossRef]
23. H. Zhang, F. Liang, K. Song, C. Xing, D. Wang, H. Yu, C. Huang, Y. Sun, L. Yang, X. Zhao, H. Sun, and S. Long, “Demonstration of AlGaN/GaN-based ultraviolet phototransistor with a record high responsivity over 3.6 × 10 7 A/W,” Appl. Phys. Lett. 118(24), 242105 (2021). [CrossRef]
24. Y.-C. Chang, “Effects of illumination on the excess carrier dynamics and variations of the surface states in an AlGaN/GaN heterostructure,” J. Appl. Phys. 107(3), 033706 (2010). [CrossRef]
25. C. Lin, Y. Lu, Y. Tian, C. Gao, M. Fan, X. Yang, L. Dong, and C. Shan, “Diamond based photodetectors for solar-blind communication,” Opt. Express 27(21), 29962 (2019). [CrossRef]
26. Z. Liu, D. Zhao, T. Min, J. Wang, G. Chen, and H.-X. Wang, “Photovoltaic Three-Dimensional Diamond UV Photodetector With Low Dark Current and Fast Response Speed Fabricated by Bottom-Up Method,” IEEE Electron Device Lett. 40(7), 1186–1189 (2019). [CrossRef]
27. A. BenMoussa, U. Schühle, K. Haenen, M. Nesládek, S. Koizumi, and J.-F. Hochedez, “PIN diamond detector development for LYRA, the solar VUV radiometer on board PROBA II,” phys. stat. sol. (a) 201(11), 2536–2541 (2004). [CrossRef]
28. M. Liao, Y. Koide, and J. Alvarez, “Single Schottky-barrier photodiode with interdigitated-finger geometry: Application to diamond,” Appl. Phys. Lett. 90(12), 123507 (2007). [CrossRef]
29. L. Ge, Y. Peng, B. Li, X. Chen, M. Xu, X. Wang, Y. Cui, D. Wang, J. Han, K. Y. Cheong, P. Tanner, M. Zhao, and X. Xu, “A Highly Responsive Hydrogen-Terminated Diamond-Based Phototransistor,” IEEE Electron Device Lett. 43(8), 1271–1274 (2022). [CrossRef]
