⇦Chinese Physics Letters, 2017, Vol. 34, No. 9, Article code 097301 Characterization of Interface State Density of Ni/p-GaN Structures by Capacitance/Conductance-Voltage-Frequency Measurements * Zhi-Fu Zhu(朱志甫)1,2,3, He-Qiu Zhang(张贺秋)4**, Hong-Wei Liang(梁红伟)4, Xin-Cun Peng(彭新村)2, Ji-Jun Zou(邹继军)2**, Bin Tang(汤彬)2, Guo-Tong Du(杜国同)1,4,5 Affiliations 1School of Physics, Dalian University of Technology, Dalian 116024 2Engineering Research Center of Nuclear Technology Application (Ministry of Education), East China Institute of Technology, Nanchang 330013 3Jiangxi Province Engineering Research Center of New Energy Technology and Equipment (Ministry of Education), East China Institute of Technology, Nanchang 330013 4School of Microelectronics, Dalian University of Technology, Dalian 116024 5State Key Laboratory on Integrated Optoelectronics, School of Electronic Science and Engineering, Jilin University, Changchun 130012 Received 24 April 2017 *Supported by the Natural Science Foundation of Jiangxi Province under Grant No 20133ACB20005, the Key Program of National Natural Science Foundation of China under Grant No 41330318, the Key Program of Science and Technology Research of Ministry of Education under Grant No NRE1515, the Foundation of Training Academic and Technical Leaders for Main Majors of Jiangxi Province under Grant No 20142BCB22006, the Research Foundation of Education Bureau of Jiangxi Province under Grant No GJJ14501, and the Engineering Research Center of Nuclear Technology Application (East China Institute of Technology) Ministry of Education under Grant No HJSJYB2016-1.
**Corresponding author. Email: hqzhang@dlut.edu.cn; jjzou@ecit.cn Citation Text: Zhu Z F, Zhang H Q, Liang H W, Peng X C and Zou J J et al 2017 Chin. Phys. Lett. 34 097301 Abstract For the frequency range of 1 kHz–10 MHz, the interface state density of Ni contacts on p-GaN is studied using capacitance-voltage ($C$–$V$) and conductance-frequency-voltage ($G$–$f$–$V$) measurements at room temperature. To obtain the real capacitance and interface state density of the Ni/p-GaN structures, the effects of the series resistance ($R_{\rm s}$) on high-frequency (5 MHz) capacitance values measured at a reverse and a forward bias are investigated. The mean interface state densities obtained from the $C_{\rm HF}$–$C_{\rm LF}$ capacitance and the conductance method are $2\times10^{12}$ eV$^{-1}$cm$^{-2}$ and $0.94\times10^{12}$ eV$^{-1}$cm$^{-2}$, respectively. Furthermore, the interface state density derived from the conductance method is higher than that reported from the Ni/n-GaN in the literature, which is ascribed to a poor crystal quality and to a large defect density of the Mg-doped p-GaN. DOI:10.1088/0256-307X/34/9/097301 PACS:73.20.At, 73.40.-c, 73.50.-h © 2017 Chinese Physics Society Article Text Mg-doped p-type GaN is a critical semiconductor material for fabrication of optoelectronic devices, such as high-electron-mobility transistors, ultraviolet detectors, light-emitting diodes and microwave-operation field-effect transistors.[1-5] Among these GaN-based electronic devices, Schottky devices with relatively simple structures, a low turn-on voltage and fast switch features, have been widely applied in high-power electronic and optoelectronic devices.[6] Extensive research has focused on rectifying contacts such as Pd, Au, Pt, Ti, Al and Ni. The reported Schottky barrier heights range from 0.5 to 2.9 eV.[7-14] However, the results from Choi et al.[15] and Rickert et al.[10] indicate that the surface Fermi level of the p-GaN is pinned perfectly because of high-density deep level defects, which is essentially independent of the metal work function. Choi et al.[15] proposed that the $S$-parameter of the semi-polar p-GaN Schottky diode (SD) that contains different Schottky metal contacts was nearly zero, which indicates a perfectly pinned Fermi level and a high density of deep-level defects due to vacancy-related or Mg-induced defects on the p-GaN surface. Nguyen et al.[16] studied the admittance results for Ni/Au/p-GaN Schottky structures theoretically and experimentally, using thermal-admittance spectroscopy measurements from 90 K to 300 K. The electrical characteristics and the basic parameters for the metal-semiconductor junction were explained. Other research groups have used various methods to produce high-quality Ni/p-GaN Schottky diodes.[17-20] Yu et al. studied the Schottky barrier characteristics and the origins of leaky characteristics for the Ni/p-GaN Schottky barrier through metal-organic chemical vapor deposition (MOCVD).[14,21] Shiojima et al.[22] proposed that a Schottky barrier with a height of as high as 2.4 eV was obtained from the current-voltage ($I$–$V$) measurements of Ni/p-GaN contacts. Yow et al.[7] studied the Schottky barrier of Ni on p-GaN from the $I$–$V$ measurements with a thermionic field-emission model. Recently, Nagaraju et al.[8] studied the interface state density of Ti/p-GaN SD using $I$–$V$ and capacitance-voltage ($C$–$V$) methods at different annealing temperatures, and found that the interface state density was of a 10$^{12}$ eV$^{-1}$cm$^{-2}$ magnitude. The interface state density as determined from $I$–$V$ and $C$–$V$ techniques is limited by the bulk doping density and profile of the semiconductor.[23] To obtain the energy distribution of the interface state density, the mechanism of forward current flow that is extracted from the forward $I$–$V$ measurements in the metal/p-GaN interface has been established from data analysis. The experimental interface state density and surface potential are found to correlate well with the bulk dopant concentration and semiconductor profile.[24] The existence of an oxide layer and interface states at the contact metal and semiconductor interface and series resistance, and the high dislocation density could disrupt the idealized case and allow for the development of high-power electronic and optoelectronic devices.[25] Therefore, the interface state density that is obtained by the $I$–$V$ and $C$–$V$ techniques on metal/p-GaN could be incomplete. To the best of our knowledge, the detailed characterization of the interface state density of the Ni/p-GaN structures has not been studied with conductance($G$)-voltage($V$)-frequency($f$) measurements. Capacitance and conductance measurements can be used to estimate the interface trap density because the input capacitance of the equivalent circuit approximates the interface trap charge. The conductance method is based on conductance losses, which stems from the majority carrier exchange between the interface state and the majority carrier band.[26] We have studied the frequency- and voltage-dependence of the $C$–$V$ and conductance-frequency ($\frac{Gp}{\omega}$–$f$) characteristics for Ni/p-GaN structures. To obtain the real capacitance and interface state density of the Ni/p-GaN structures, the effects of series resistance ($R_{\rm s}$) on high-frequency (5 MHz) capacitance values measured under reverse and forward biases were amended. For 1 kHz–10 MHz at room temperature, the interface state density ($N_{\rm SS}$) of the Ni/p-GaN structures was estimated from the high-low frequency ($C_{\rm HF} -C_{\rm LF}$) capacitance and the conductance method.
cpl-34-9-097301-fig1.png
Fig. 1. Schematic diagram of Ni/p-GaN structures (a) and energy band diagram of metal/p-GaN Schottky diode with interface states under the zero-bias condition (b). Here ${\it \phi}_{\rm b}$ is the Schottky barrier height, $q{\it \phi}_{\rm m}$ is the metal work function, $\chi_{\rm e}$ is the electron affinity of the semiconductor, $E_{\rm g}$ is the energy gap, $V_{\rm bi}$ is the built-in potential, and $W_{\rm D}$ is the depletion width.
The p-GaN with a Mg doping concentration of $1.0\times10^{19}$ cm$^{-3}$ and a thickness of 4.5 μm, and undoped 2-μm-thick GaN were grown consecutively on a $c$-plane sapphire substrate through MOCVD with a $3\times2$ in the Aixtron close-coupled shower head system. Before fabrication of the Schottky and Ohmic contacts, the p-GaN samples were cleaned with a standard RCA clean (a recipe developed by the RCA Company in America). Samples were soaked in boiling aqua regia (HCl:HNO$_{3}$ = 3:1) for 15 min to remove native oxides, and films were rinsed with deionized water. Samples were desiccated using high-purity nitrogen gas (99.999%). Ni/Au (15 nm/20 nm) Ohmic contacts and Ni (15 nm) Schottky contacts were formed by a standard photolithographic technique and metal deposition by an e-beam evaporator. Films were treated at an annealing temperature of 500$^{\circ}\!$C for 15 min in a quartz-tube furnace in air to form an Ohmic contact before deposition of the Schottky contacts. A schematic diagram and the energy band diagram under the zero bias of the Ni Schottky diode as prepared on p-GaN is shown in Fig. 1. The Schottky contact diameter is 125 μm and a 100-μm gap exists between the Schottky and ohmic contacts, whose width and outer diameter are 165 μm and 665 μm, respectively. The carrier concentration and mobility for p-GaN were characterized by the Hall-effect measurements at room temperature. The Hall-effect measurements showed that the carrier concentration is $1.05\times10^{17}$ cm$^{-3}$ and the mobility is 6.9 cm$^{2}$V$^{-1}$s$^{-1}$ in the p-GaN layer. To obtain detailed information on the interface state density, Ni/p-GaN structures were characterized by $C$–$V$ measurements from 1 kHz to 10 MHz at room temperature using a 4200-SCS semiconductor parameter analyzer (Keithley Instruments Inc., Cleveland, OH, USA) and a Lake Shore Model TTPX cryogenic probe station. Samples were measured from $2\times10^{-6}$ to $8\times10^{-6}$ Torr. A small sinusoidal signal was applied to the device to meet the requirements.
cpl-34-9-097301-fig2.png
Fig. 2. Experimental forward and reverse bias $I$–$V$ characteristic curves of the Ni/p-GaN structures at 300 K.
Figure 2 shows the experimental forward and reverse bias $I$–$V$ characteristic curves of the Ni/p-GaN structures. A nonlinear $I$–$V$ curve is obtained for the Ni/p-GaN structures, which exhibits a strong rectifying behavior. Devices were reverse biased in turn by applying a negative potential to the Ohmic contact (Ni/Au) and by grounding the Schottky contact (Ni). When a reverse bias is applied on the Ni/p-GaN structures, the basic equivalent circuits including interface state effect are shown in Fig. 3.[27] In Fig. 3, $C_{\rm ox}$ and $C_{\rm m}$ are the insulator capacitance and the semiconductor depletion layer capacitance, respectively, $C_{\rm it}$ and $R_{\rm it}$ are the capacitance and resistance associated with the interface trap, and $R_{\rm s}$ is the series resistance from the metal contact and the bulk resistance of p-GaN.
cpl-34-9-097301-fig3.png
Fig. 3. Measurement equivalent circuits model: (a) the capacitance method including interface trap effects, (b) the conductance method including interface trap effects, (c) low frequency, and (d) high frequency.
cpl-34-9-097301-fig4.png
Fig. 4. (a) Measured $C$–$V$ plots of Ni/p-GaN structures with different frequencies at room temperature. (b) Variation of series resistance as a function of voltage for various frequencies at room temperature.
Figure 4(a) shows the experimental $C$–$V$ characteristics of the Ni/p-GaN structures as a function of frequency at room temperature under the applied bias from $-$2 V to +2 V. Because the impacts of interface capacitance at a high 5-MHz frequency were small, changes in the depletion layer capacitance and diffusion capacitance as the voltage changes were mainly considered. At a high 5-MHz frequency, the capacitance increases, reaches a maximum and then decreases when it is under forward bias, whereas the capacitance decreases with an increase in voltage when it is under reverse bias. However, at a frequency of less than or equal to 1 MHz, the capacitance decreases with an increase in voltage when it exists under forward bias. The capacitance increases with the voltage, then reaches an extreme value and decreases with the increase of the voltage when it exists under reverse bias. The frequency-dependent capacitance characteristics are very important based on the result accuracy and reliability.[15,16,21-23] The reverse bias $C$–$V$ in depletion-region plots exhibits a kink, and the peak value for the capacitance decreases sharply as the frequency increases. Demirezen et al.[28] and Doǧan et al.[29] demonstrated that the observed peak in the forward bias $C$–$V$ curves of an n-GaN Schottky diode is caused by series resistance. It is known that the peak value capacitance of the Schottky diodes varies with the interface state density and the series resistance. Therefore, the series resistance is critical in the electrical characteristics of the Schottky barrier diodes. The effect of interface traps has been observed from the $C$–$V$ curves that stretch out in the voltage direction, which occurs because extra charges have to fill the traps, and thus more charge or an applied voltage is required to reach the same surface potential.[24] The series resistance $R_{\rm s}$ of the Schottky structures can be extracted from the measured capacitance $C_{\rm m}$ and conductance $G_{\rm m}$ in reverse and forward biases at a high frequency, as presented by Nicollian et al.[26] The series resistance $R_{\rm s}$ is given by $$\begin{align} R_{\rm s} =\frac{G_{\rm m}}{G^2_{\rm m} +(\omega C_{\rm m})^2},~~ \tag {1} \end{align} $$ where $\omega =2\pi f$ is the angular frequency. The voltage-dependent series resistance profile of the Ni/p-GaN structure is calculated from the measured $C$–$V$ and $G$–$V$ measurements according to Eq. (1) and the results with 5 MHz are shown in Fig. 4(b). Figure 4(b) shows that the series resistance of the Ni/p-GaN structures is nearly zero in the bias voltage range from $-$1.0 V to 1.0 V. However, the series resistance shows an exponential increase with the reverse bias because of an increase in the depletion-layer width. Hence, the capacitance and conductance are uncorrected with a low voltage in the following analysis. In the equivalent circuit of our Ni/p-GaN structures, the interface state capacitance $C_{\rm it}$ is a paralleled space-charge capacitance $C_{\rm sc}$ because the interface states do not follow the alternating-current signal at a high frequency, and thus do not contribute to the total capacitance.[18] The space-charge capacitance $C_{\rm sc}$ is equal to the measured high-frequency capacitance. A subtraction of the depletion-layer capacitance (that is extracted from the measured high-frequency capacitance $C_{\rm HF}$) can be used to obtain the interface state capacitance. The interface state capacitance can be obtained by subtracting the interface state capacitance (which is extracted from the measured low-frequency capacitance, $C_{\rm LF}$). We selected a low frequency of 1 kHz and a high frequency of 5 MHz to obtain the detailed interface state density $N_{\rm ss}$ of the Ni/p-GaN structures by a high-low frequency method. Here $C_{\rm LF}$ at a low frequency can be determined by $$\begin{align} C_{\rm LF} =C_{\rm it} +C_{\rm sc}.~~ \tag {2} \end{align} $$ Thus the equivalent capacitance becomes the parallel connection of interface state capacitance $C_{\rm it}$ and space charge capacitance $C_{\rm SC}$, $$\begin{align} C_{\rm sc} =C_{\rm HF}.~~ \tag {3} \end{align} $$ Substituting Eq. (3) into Eq. (2) gives the interface state capacitance ($C_{\rm it}$) in terms of the measured $lf$ and the $hf(C-V)$ curves as $$\begin{align} C_{\rm it} =C_{\rm LF} -C_{\rm HF}.~~ \tag {4} \end{align} $$ The interface state density $N_{\rm ss}$ can be obtained from Eqs. (2)-(4), and yields $$\begin{align} N_{\rm SS} =\frac{C_{\rm it}}{qA}=\frac{1}{qA}(C_{\rm LF} -C_{\rm HF}),~~ \tag {5} \end{align} $$ where $q$ is equal to 1.60$\times$10$^{-19}$ C, and $A=0.012$ cm$^{2}$ is the device area.
cpl-34-9-097301-fig5.png
Fig. 5. Voltage dependence of corrected interface state density distribution $N_{\rm ss}$ for Ni/p-GaN structures at room temperature.
The interface state density $N_{\rm ss}$ that is extracted from the measured high- and low-frequency capacitances in line with the applied reverse bias is plotted in Fig. 5 for the Ni/p-GaN structures. The interface state density $N_{\rm ss}$ shows clearly a maximum value at a reverse voltage of $-$0.6 V in the depletion regions in Fig. 5. When a reverse bias is applied, the depletion-layer capacitance increases and subsequently reaches its peak. Thus the interface state density $N_{\rm ss}$ increases with the reverse bias and finally reaches a maximum value in the depletion region. However, the interface state density $N_{\rm ss}$ decreases with increasing the bias in the inversion regions, which leads to a difference between the low- and high-frequency capacitances. This can be ascribed to the fact that the depletion region widens and the total capacitance continues to decrease with further increase in reverse bias. Thus the interface state density $N_{\rm ss}$ decreases with the increase of the reverse bias in the inversion region. This change also appears in the high- and low-frequency $C$–$V$ curves in Fig. 4(a). The mean value of $N_{\rm ss}$ is about $2\times10^{12}$ eV$^{-1}$cm$^{-2}$. The conductance method is supposed to be most complete for probing the interface state density $N_{\rm ss}$ because of a high sensitivity and accuracy, and because it is able to detect trap densities as low as 10$^{8}$ eV$^{-1}$cm$^{-2}$.[24] The conductance method is measured by the device bias in the depletion region and the weak inversion region. According to the $C$–$V$ analysis, the reverse bias of the depletion region of the Ni/p-GaN structure occurs at $-$0.6 V. Therefore, the reverse bias of $C$–$V$–$f$ measurements is applied in the depletion region. Figure 6 shows the experimental capacitance and conductance for a frequency in the range from 1 kHz to 10 MHz and biases from $-$0.6 V to 0 V with 0.1 V increments. Figure 6(a) presents the measured capacitance decreases with the increase of the frequency. Figure 6(b) shows that when the frequency is lower than 1 MHz, the conductance almost keeps constant. However, at frequencies above 1 MHz, the conductance increases logarithmically with frequency. A large frequency dispersion in the depletion regions results for the admittance measurements. The interface trap model can be used to explain the reverse behavior. The density distribution of the interface states can be determined from the experimental $G$–$f$ measurements of the Ni/p-GaN.
cpl-34-9-097301-fig6.png
Fig. 6. (a) Experimental $C$–$f$ characteristics of Ni/p-GaN structures for different reverse bias voltages at room temperature. (b) Experimental conductance versus frequency characteristics of Ni/p-GaN structures. (c) Equivalent parallel conductance $\frac{G_{\rm p}}{\omega}$ versus $\log f$ at various reverse biases for Ni/p-GaN structures at room temperature.
To obtain the detailed interface state density $N_{\rm ss}$ of the Ni/p-GaN structures by the conduction method, the interface state conductance for Ni/p-GaN structures can be described as[30] $$\begin{align} \frac{G_{\rm p}}{\omega}=\frac{qAN_{\rm SS}}{2\omega \tau _{\rm p}}\ln[1+(\omega \tau _{\rm p})^2].~~ \tag {6} \end{align} $$ Figure 6(c) shows the parallel conductance $\frac{G_{\rm p}}{\omega}$ as functions of frequency and bias voltage. With the increase of the frequency, $\frac{G_{\rm p}}{\omega}$ increases until it reaches a peak and then it decreases rapidly. This phenomenon is not unique to the Ni/p-GaN interface. Similar features result for Au/Ni/AlGaN/AlN/GaN and Au/Ni/n-GaN. It has been demonstrated that $\frac{G_{\rm p}}{\omega}$ goes through a maximum value when $\omega \tau _{\rm it} =1$, which gives $\omega \tau _{\rm it}$. The maximum value of $\frac{G_{\rm p}}{\omega}$ is $C_{\rm it} /2$. From Eq. (5), a maximum results at $\omega \tau _{\rm p} =1.98$. Figure 6(c) shows that $({\frac{G_{\rm p}}{\omega}})_{\max}$ occurs at 4 MHz. Thus the interface state density $N_{\rm SS}$ is given by $$\begin{align} N_{\rm SS} =\frac{2.5}{qA}\Big({\frac{G_{\rm p}}{{\rm \omega}}}\Big)_{\max}.~~ \tag {7} \end{align} $$ Figure 7 shows the interface state density versus the reverse bias voltage plots for Ni/p-GaN structures at room temperature. It can be seen from Fig. 7 that the interface state density of the Ni/p-GaN structures decreases with the increase of the reverse bias. When reverse bias is applied to the Ni/p-GaN structures, the valence-band edge bends upward near the surface and is closer to the Fermi level.[27] When the applied reverse bias increases for the Ni/p-GaN structure in the depletion region, the depletion capacitance and the equivalent parallel conductance increase with the reverse bias. Therefore, the interface state density $N_{\rm ss}$ of the Ni/p-GaN structure increases as the reverse bias increases and its maximum is $\sim$0.94 10$^{12}$ eV$^{-1}$cm$^{-2}$.
cpl-34-9-097301-fig7.png
Fig. 7. Interface state density versus reverse bias voltage plots for Ni/p-GaN structures at room temperature from the conductance method.
As can be seen from Figs. 5 and 7, the calculated $N_{\rm ss}$ values according to the conductance method are lower than those according to the capacitance and are of approximately one order from Figs. 5 and 7. The difference could be attributed to failure of a 1 MHz $C$–$V$ curve to approximate a high frequency $C$–$V$ curve, which is considered by the high-low frequency capacitance. Another reason could be that a simplification is possible by limiting the $C$–$V$ measurements to high- and low-frequency ranges. Nagaraju et al. have calculated the interface state density $N_{\rm ss}$ of the Ti/p-GaN SDs by the $I$–$V$ and $C$–$V$ methods to be 10$^{13}$ eV$^{-1}$cm$^{-2}$. Compared with the results from Nagaraju et al., our experimental value is smaller, which stems from the fact that the mechanism of forward-current flow is undetermined, which gives rise to interpretations of $I$–$V$ characteristics. Turut et al.[25] have calculated $6\times10^{11}$ eV$^{-1}$cm$^{-2}$ for Au/Ni/n-GaN structures by the $C/G$–$f$ method at 300 K with a 0 V bias. The high $N_{\rm ss}$ values obtained from the conductance method of p-GaN could be attributed to the poor crystal quality and the large defect density of the p-GaN. In summary, $C$–$V$ and $G$–$f$–$V$ characteristics of the Ni/p-GaN structures have been studied at room temperature. The relationship between the interface state density distribution profile and the reverse bias is determined according to the high-low frequency capacitance and conductance methods. The effects of series resistance ($R_{\rm s}$) on the high-frequency (5-MHz) capacitance values as measured under reverse and forward biases are amended to obtain the real capacitance and interface state density of the Ni/p-GaN structures. Compared with the $I$–$V$ and $C$–$V$ methods for the Ti/p-GaN, the experimental results illustrate that the interface state density determined by both the methods is remarkably low. In addition, the interface state density from the conductance method is higher than the reported values for Ni/n-GaN. References 231–261nm AlGaN deep-ultraviolet light-emitting diodes fabricated on AlN multilayer buffers grown by ammonia pulse-flow method on sapphireMicrowave performance of a 0.25 μm gate AlGaN/GaN heterostructure field effect transistorCandela‐class high‐brightness InGaN/AlGaN double‐heterostructure blue‐light‐emitting diodesHigh‐responsivity photoconductive ultraviolet sensors based on insulating single‐crystal GaN epilayersGaN based transistors for high power applicationsApplication of the thermionic field emission model in the study of a Schottky barrier of Ni on p-GaN from current–voltage measurementsAnnealing effects on the electrical, structural and morphological properties of Ti/p-GaN/Ni/Au Schottky diodeSchottky Barrier Height and S -Parameter of Ti, Cu, Pd, and Pt Contacts on p-Type GaNX-ray photoemission determination of the Schottky barrier height of metal contacts to n –GaN and p –GaNElectrical properties and carrier transport mechanism in V/p-GaN Schottky diode at high temperature rangeSchottky barrier height of boride-based rectifying contacts to p-GaNAnnealing and Measurement Temperature Dependence of W2B- and W2B5-Based Rectifying Contacts to p-GaNStudy of Schottky barrier of Ni on p -GaNInvestigation of Fermi level pinning at semipolar (11–22) p-type GaN surfacesThermal admittance spectroscopy of Mg-doped GaN Schottky diodesDepletion Capacitance and Diffusion Potential of Gallium Phosphide Schottky‐Barrier DiodesAn investigation of surface states at a silicon/silicon oxide interface employing metal-oxide-silicon diodesInterface states in MOS structures with 20–40 Å thick SiO2 films on nondegenerate SiDescription of the SiO2Si interface properties by means of very low frequency MOS capacitance measurementsThe origins of leaky characteristics of schottky diodes on p-GaNLarge Schottky barriers for Ni/p-GaN contactsSurface states at steam-grown silicon-silicon dioxide interfacesThe interface state density characterization by temperature-dependent capacitance–conductance–frequency measurements in Au/Ni/ n -GaN structuresInterface States in AlGaN/GaN Metal-Insulator-Semiconductor High Electron Mobility TransistorsOn the temperature dependent profile of interface states and series resistance characteristics in (Ni/Au)/Al0.22Ga0.78N/AlN/GaN heterostructuresCapacitance-conductance-frequency characteristics of Au/Ni/n-GaN/undoped GaN StructuresAnalysis of Capacitance-Voltage-Temperature Characteristics of GaN High-Electron-Mobility Transistors
[1] Hirayama H et al 2007 Appl. Phys. Lett. 91 071901
[2] Asif Khan M et al 1994 Appl. Phys. Lett. 65 1121
[3] Nakamura S et al 1994 Appl. Phys. Lett. 64 1687
[4] Khan M A et al 1992 Appl. Phys. Lett. 60 2917
[5] Shur M S 1998 Solid-State Electron. 42 2131
[6]Sharma B L 1984 Metal-Semiconductor Schottky Barrier Junctions and Their Applications (New York: Springer)
[7] Lin Y J 2005 Appl. Phys. Lett. 86 122109
[8] Nagaraju G et al 2015 Appl. Phys. A 121 131
[9] Park Y et al 2012 Jpn. J. Appl. Phys. 51 09MK01
[10] Rickert K A, Ellis A B, Kim J K et al 2002 J. Appl. Phys. 92 6671
[11] Siva Pratap Reddy M, Bengi A, Rajagopal Reddy V et al 2015 Superlattices Microstruct. 86 157
[12] Stafford L, Voss L F, Pearton S J et al 2006 Appl. Phys. Lett. 89 10
[13] Voss L F, Stafford L, Thaler G T et al 2007 J. Electron. Mater. 36 384
[14] Yu L S, Qiao D, Jia L et al 2001 Appl. Phys. Lett. 79 4536
[15] Choi Y Y, Kim S, Oh M et al 2015 Superlattices Microstruct. 77 76
[16] Nguyen N D, Germain M, Schmeits M et al 2001 J. Appl. Phys. 90 985
[17] Cowley A M 1966 J. Appl. Phys. 37 3024
[18] Terman L M 1962 Solid-State Electron. 5 285
[19] Kar S and Dahlke W E 1972 Solid-State Electron. 15 221
[20] Castagne R and vapaille A 1971 Surf. Sci. 28 157
[21] Yu L S, Jia L, Qiao D et al 2003 IEEE Trans. Electron Devices 50 292
[22] Shiojima K, Sugahara T and Sakai S 1999 Appl. Phys. Lett. 74 1936
[23] Berglund C N 1966 IEEE Trans. Electron Devices 13 701
[24]Sze S M and Kwok K 2006 Physics of Semiconductor Devices (Hoboken NJ: Wiley)
[25] Turut A, Doğan H and Yıldırım N 2015 Mater. Res. Express 2 096304
[26]Nicollian E H and Brews J R 1982 Metal Oxide Semiconductor Physics and Technology (New York: Wiley)
[27] Qian F, Kai D, Yu K L et al 2013 Chin. Phys. Lett. 30 127302
[28] Demirezen S Ã 2010 Physica B 405 1130
[29] Doǧan H, Yildirim N, Orak I et al 2015 Physica B 457 48
[30] Zhao M and Liu X Y 2015 Chin. Phys. Lett. 32 048501