New Developments of Gallium Nitride and the ... - Semantic Scholar
New Developments in Gallium Nitride and the Impact on Power Electronics M.A. Khan, G. Simin, S.G. Pytel, A. Monti, E. Santi and J.L. Hudgins* Department of Electrical Engineering University of South Carolina 301 South Main Street Columbia, South Carolina 29208
*Department of Electrical Engineering 209N WSEC, PO Box 880511 University of Nebraska Lincoln, NE 68588 [email protected]
I. INTRODUCTION The interest in III-Nitride semiconductors dates back to 1970s. It was initially stimulated by pioneering works by Pankove, Akasaki, Nakamura, Khan and many others showing tremendous potential of these materials for optoelectronics. Nearly at the same time, field-effect transistors have been recognized as the most promising electronic devices, which has been a topic of intensive investigations since the first report in 1991 [1]. The major
advantages of nitride-based devices that make them extremely promising for high-power, high-temperature applications are high electron mobility and saturation velocity, high sheet carrier concentration at heterojunction interfaces, high breakdown field, and low thermal impedance when grown over SiC substrates. The chemical inertness and excellent temperature stability of nitrides are the key properties to provide high reliability. Owing to these unique properties, group III-Nitride semiconductors provide great promise for overcoming the fundamental limitations associated with silicon and gallium arsenide semiconductors.
1000
600 400 200 0 1970
1975
1980
1985
2D electron gas
800
1990
Blue LED
1200
Breakthrough in growth
Number of publications
Abstract: Wide bandgap III-Nitride semiconductor materials possess superior material properties as compared to silicon, GaAs and other III-V compound materials. Recent achievements in Gallium Nitride (GaN) technology for optoelectronics have resulted in ultra-bright blue light emitting diodes and lasers, ultraviolet emitters, and solar-blind optical detectors. In the electronic area, drastic improvement of microwave device performance has been achieved, yielding record high power densities of 20 - 30 W/mm. Novel applications of these materials in high-power electronics for switching, energy conversion and control are just emerging. This paper provides an overview of the state-of-the-art IIINitride wide bandgap technologies and it explores power electronic applications while illustrating the enormous potential that GaN based devices have for overcoming the major challenges of power electronics in the 21st Century. The paper discusses the unique material and device properties of GaN-based semiconductors that make them promising for high-power, high-temperature applications. These include high electron mobility and saturation velocity, high sheet carrier concentration at heterojunction interfaces, high breakdown voltages, and low thermal impedance (when grown over SiC or bulk AlN substrates). The chemical inertness and radiation hardness of nitrides are other key properties. As applied to power electronics, the III-Nitride technology allows for high-power switching with sub-microsecond and nano-second switching times. The paper will present the innovations that further improve the performance of highpower DC-DC converters, switches and other building blocks. These include novel insulated gate HFET design that significantly expands the allowable input voltage amplitude, further increases the device peak currents, and most importantly, tremendously improves the large-signal stability and reliability. Insulated gate switching devices have been shown to operate at up to 300C with no noticeable parameter degradation. Novel monolithic integrated circuits of high-power switches and DC-DC converters and their performance parameters will be presented. The paper also discusses the major challenges associated with modern GaN technology and work in progress to overcome them.
1995
2000
Year Figure 1. Progress in the number of publications per year in group IIINitride material and device research (after [2]).
Figure 1 shows the number of publications per year related to the III-Nitride semiconductors during past thirty years. As seen, after 1980s - 1990s when significant progress in MOCVD epitaxial growth was achieved, and the first electronic and optoelectronic devices were successfully demonstrated, the interest in the field increased tremendously. In the past ten years, the research in the IIINitride device applications has mainly focused on optoelectronics and microwave electronics. This resulted in several novel unique device types, which include blue and ultraviolet light emitters and lasers, visible- and solar-blind photodetectors and high-power microwave field-effect transistors. The device improvement goes alongside with material growth development, which led to tremendous improvement in the material quality and significant cost reduction. Following the tremendous progress in the material quality and device processing, a worldwide market for nitride-based devices has developed. One of the highest profile markets for nitride-based LEDs is in mobile phones,
The bandgap energy values for the most important semiconductor materials are shown in Figure 3. As one can see the bandgap energy of GaN, EG 3.4 eV, is more than three times higher than that of Si (1.1 eV) and more than two times higher than that of GaAs (1.41 eV). Since the probability of ionization and many other processes depend exponentially on the bandgap energy, a large bandgap is the key factor for high temperature operation, chemical inertness, and high breakdown voltage of GaN devices. The bandgap energy of AlN is even higher, ~ 6.2 eV. The bandgap value of the third important III-Nitride material, InN, is yet somewhat unclear, between 0.95 and 1.9 eV. 7
Bandgap.eV
particularly blue LEDs for keypads and white LEDs for liquid crystal display (LCD) backlight. Nitride-based LEDs also offer great promise for solid-state lighting applications such as architectural lighting, machine vision, illumination for signage, flashlights, decorative lights, and so on. The electronic device market includes high-power RF sources for base stations, transmitter-receiver modules, sensors and other high-power, high-temperature applications. The market for high-brightness LEDs (HB-LEDs) in lighting applications has grown rapidly with an overall market in the billions of dollars [3]. Based on continuing positive trends in this dynamic industry, the market for HB-LEDs is projected to grow to ~ $5.0 billion by 2007 [4]. Emerging applications of III-Nitrides in power electronics may significantly increase and expand the potential market. Electric vehicles, power station electronic switches and electric ships are a few of the areas that can be drastically changed by these group III-Nitride semiconductors. In the next sections, we will provide an overview of group III-Nitride semiconductor materials properties, growth techniques, main achievements in the fields of optoelectronics and microwave devices. We will also show how group III-Nitride materials can revolutionize the area of high-power electronics and review the challenges in this emerging technology.
6 5 4 3 2 1 0 Si
GaAs AlAs
SiC
GaN
AlN
Figure 3. The bandgaps of the semiconductor materials
II. MATERIAL PROPERTIES AND GROWTH The group of III-Nitride semiconductors includes three main materials, Gallium Nitride (GaN), Aluminum Nitride (AlN) and Indium Nitride (InN). All three materials crystallize mostly in so-called wurtzite structures that have a hexagonal unit cell. An example of GaN crystal structure is shown in Figure 2. An important property of this crystal cell is the lack of inversion symmetry, which leads to very strong polarization effects in group III-Nitride materials.
Figure 2. Schematic of the wurtzite GaN crystal structure [5]
Another fundamental property that makes group IIINitride materials extremely attractive for high-power, hightemperature applications is a very large bandgap, i.e. the energy required to ionize atoms and create free electrons.
There is one more property of group III-Nitride materials that is of crucial importance for optoelectronic and electronic applications. GaN, AlN and to some extent, InN can form alloys by mixing different fractions of Ga, Al and In in ternary and quaternary compounds, e.g. Ga xAl1-xN, GaxIn1-xN, Gax,InyAl1-x-yN. The bandgap energy of these compound materials is approximately equal to a weighted linear combination of the bandgap energies of the binary materials forming the alloy (so-called Vegard's law). For example, adding different fractions of Al into GaN compound, the bandgap energy can be varied from 3.4 eV up to 6.2 eV. This approached is called “band-gap engineering” [ 6 ] Normally, the ternary or quaternary materials are grown on top of GaN or AlN substrates. Practical implementations of bandgap engineering are limited by changes in lattice constants that create a strain between the layers of different compositions. Group III-Nitride materials belong to so-called directbandgap materials. This essentially means that the energy released by electron-hole recombination in these materials can be directly converted into light. This property makes group III-Nitrides extremely important for optoelectronic applications. Figure 4 shows the range of optical wavelength that can be covered by AlN - GaN - InN alloys. As seen, ternary and quaternary compounds of AlGaNInN can produce light not only in the entire visible part of the spectrum (0.7 – 0.4μm) but also in the deep UV range of it.
Figure 4. Optical spectrum covered by III-Nitride materials.
For high-power electronic applications, high breakdown voltages achievable in group III-Nitride materials are of primary importance. Figure 5 compares the critical electric fields for the avalanche breakdown mechanism in the most important semiconductor materials. As one can see Nitride based semiconductors possess characteristic breakdown fields exceeding nearly ten times those of Si or GaAs. The extremely high breakdown fields results in record-high breakdown voltages even when applied across micrometer scale spacing.
V/cm
4.00E+06 3.00E+06 2.00E+06 1.00E+06 0.00E+00 Si
GaAs
SiC
GaN
AlN
Figure 5. Breakdown fields of different semiconductor materials.
To-date, most of AlGaN/GaN and AlInGaN/GaN heterostructures are grown by Metal Organic Chemical Vapor Deposition (MOCVD). MOCVD systems are capable to operate in both conventional and atomic layer deposition regimes. Conventional deposition regime wherein precursors enter the growth chamber simultaneously is used to deposit GaN layers. Triethylgallium and ammonia are used as the precursor gases. AlxGa1-xN layers are deposited in atomic layer regime when precursors enter the chamber in a cyclic fashion. Triethylgallium, triethylaluminum and ammonia are used as precursors. The precursors are introduced into the chamber using hydrogen or nitrogen as a carrier gas. Epilayers are deposited on sapphire or SiC substrates placed on graphite susceptor, which is heated to the growth temperature by rf-induction.
AlGaN/GaN heterostructures are typically grown on basal (0001) sapphire, conducting 6H-SiC and semi-insulating 4H-SiC substrates. The substrates are slightly off-axis (12o) production grade 6H and 4H silicon carbide available from Cree Research. The doping levels of both n- and p-type 6H-SiC are on the order of 2x1018 cm-3. The micropipe density is approximately 30 cm-2. The deposition of approximately 50 nm thick of AlN buffer is followed by the growth of insulating/modulation doped GaN, which is finally capped with AlGaN barrier layer. The major difference between the AlGaN/GaN growth on sapphire and SiC substrates is the thickness of the insulating GaN layer. The cross-sectional TEM analysis of GaN grown on sapphire and SiC reveals strong dependence of growth defect distribution along the growth direction on the substrate material. The significant reduction in the number of threading dislocations in GaN on sapphire is observed for layer thicknesses above 1.5-2 m. The similar improvement in material quality for GaN grown on SiC was achieved at the thickness as low as 0.5 m or even lower.
III. OPTOELECTRONIC APPLICATIONS OF III-NITRIDES III.1. LIGHT EMITTERS The advent of blue light emitting diodes (LEDs) and short-wavelength blue-violet laser diodes using the group III-nitrides alloys (AlN, GaN, and InN) have opened up an era of new optical devices, which was not possible with the established III-V material systems based on GaAs, AlAs, GaP, InAs and related alloys. The direct bandgap and a wide spectrum of bandgap energies (0.7 eV for InN, 3.4 eV for GaN, and 6.2 eV for AlN) of group III-nitride semiconductors make them suitable for a wide range of optoelectronic applications [7]. While “visible” technology has been reaching its maturity, many groups have shifted the research focus towards shorter wavelength ultraviolet devices. Visible and solar blind nitride detectors have reached outstanding performance at very short cut-off wavelengths. The quantum efficiency of near ultraviolet (UV) LED received a huge boost with advanced growth, metallization and packaging techniques. The emission wavelengths of LEDs now cover the majority of the UV spectrum ranging from 400 nm to 250 nm. Finally, UV LED emission has shifted towards shorter wavelengths with viable operation demonstrated at 340 – 350 nm. Due to their enormous potential applications several research groups are actively developing deep UV LEDs. Nishida et al. have reported on milliwatt power UV LEDs with emission around 342 nm and 352 nm over SiC and hydride vapor phase epitaxial GaN substrates, respectively [8, 9]. The University of South Carolina (USC) has achieved the milliwatt powers in the wavelength range of 260 - 280 nm and emission at shortest wavelength to date of 254 nm [10,11]. A typical UV LED structure is shown in Figure 6. Both SiC and GaN substrates provided good heat removal
allowing the device to operate at high dc currents, but the optical absorption in the substrate (top p-layers and semitransparent contacts) significantly reduced the external quantum efficiency.
p-AlGaN AlGaN (5) QW
n+-AlGaN barrier AlN/AlGaN Nested SL MEMOCVD AlN Buffer
sapphire Figure 6. Typical design and layer structure of AlGaN/GaN based UV LED on sapphire substrate (USC)
The LED structure forms a p-n junction made out of wide bandgap material. The active (recombination) area is located in the space-charge region of the p-n junction and consists of a material with slightly narrower bandgap (socalled “quantum well”). The narrower bandgap effectively confines the electron-hole pairs thus greatly increasing the probability of recombination. Deep UV LEDs require AlGaN layers with Al content greater than 40 %. The internal quantum efficiency of such LEDs is mainly limited by the material quality and low doping efficiency, especially for Mg doped p-AlGaN. In device structures with high Alcontent AlGaN layers, these two factors determine the nonradiative recombination rate and injection efficiency. Novel USC devices utilized an innovative AlN/Al0.5Ga0.5N superlattice (SL) approach for deposition of 2 m thick AlxGa1-xN (x>0.35) buffer layers with significantly lower defect densities [ 12 , 13 ]. The number of non-radiative defects is itself a strong function of the buffer and the active layers material quality. In our recent devices the buffer layer and strain relief AlN/Al0.5Ga0.5N SL were grown using PALE approach allowing for the high structural quality [14]. This growth approach led to a significant increase of the LED quantum efficiency.
5.6 5.4 5.2
Normalized Intensity, a.u.
p+-GaN
Energy, eV 5
4.8
4.6
4.4
4.2
4
20 mA dc
220 230 240 250 260 270 280 290 300 310 320
Wavelength, nm Figure 7. Wavelength tuning in AlGaN/AlGaN Deep UV LEDs with different Al fraction.
By changing the Al fraction in the active and cladding regions of the LED, the emission wavelength can be tuned in the wide range as shown in Figure 7. The LED operates by applying forward bias to the p-n junction. The bias leads to an injection of electrons and holes into a quantum well. Ideally, all the electron-hole pairs should recombine in the quantum well resulting in complete conversion of the electrical current into a photon flux. However, as stated above, the conversion efficiency greatly depends on the defect concentration, the concentration of donor and acceptor impurities creating the electron-hole pairs required for efficient injection and on energy band profile in the cladding and active layers. Another fundamental problem in achieving high conversion efficiency is the extraction of the photon produced in the semiconductor. All semiconductor materials have refractive indices, n, greater than unity. For example, for GaAs and AlGaAs compunds, n 3.5, for GaN and AlGaN compounds, n 2.5. When the photons traveling inside the semiconductor reach the semiconductor - air interface under certain angle, they experience full internal reflection. Only the optical beams propagating perpendicularly to the surface or at the angles less than a certain critical angle can be extracted. As a result, the overall efficiency of LEDs is well below 100%. Material quality, doping, epitaxial layer design, dye shape and packaging play crucial roles in achieving high efficiency LEDs. The highest output powers obtained in the 255 nm 280 nm deep UV LEDs are shown in Figure 8 [15,16].
3.5
dc
3.0
280 nm
Power, mW
2.5 2.0 1.5
265 nm
1.0
255 nm
0.5 0.0
0
20
40
60
80 100 120 140 160 180 200
Current, mA Figure 8. Optical powers of AlGaN/GaN deep UV LEDs
III.2. PHOTODETECTORS Along with light emitters, light detectors are equally important elements of many optoelectronic systems. Traditionally, the detection of light in the ultraviolet (UV) part of electromagnetic spectrum has been realized using photo-multiplier tubes (PMT). These devices offer unprecedented sensitivity in UV range, low dark currents and high speeds. The high internal gain (> 10 6) achieved through the photomultiplication process makes PMTs very effective for detection of low intensity signals. However, very high operating voltages (typically > 1 kV) requiring special (and sometimes very bulky) power supplies make the PMT much less attractive for portable applications. In addition, PMTs are very fragile vacuum tube devices that are poorly suited for many space and field applications. An alternative approach uses UV-enhanced Si detectors that are typically p-i-n diodes with a special anti-reflection coating designed for the UV wavelength range. Si has a direct band gap at 3 eV, so photons in the UV ranges are absorbed efficiently near the front surface of the device. These semiconductor detectors can be made miniaturized, highly reliable, robust, and well integrated with other semiconductor components. Although the UV-enhanced Si detectors have many important advantages, their dark currents are high (typically in the nA range) and quantum efficiencies are poor. Because Si has a very high absorption constant beyond 3 eV, carriers are generated very near the front surface, and minority carriers may be lost through surface state recombination. For operation in the deep UV, the efficiency of these detectors may suffer from degradation of the SiO2/Si interface after prolonged UV exposures, due to the generation of surface states. Both PMT and Si detectors are sensitive throughout the UV and visible wavelengths thus requiring costly UV filters to achieve visible-blind or solar-blind operation. Certain outdoor detector applications would benefit from operation in total darkness to reduce or eliminate background currents, and since the solar spectrum terminates at 280 nm, they need to be non-responsive to wavelengths greater than 280 nm (hence “solar blind”). Other applications in the near UV
(below 300 nm) would also benefit from being nonresponsive to visible light, be it sunlight or indoor lighting. Thus an alternative to PMTs and Si is desirable. Other semiconductor detector materials have also been pursued by researchers including Ge, GaAs, and SiC, in order to combine the sensitivity of a PMT and robustness of Si with “blindness” to longer wavelengths. Ge and GaAs offer no obvious advantages over Si, but 4H-SiC is certainly visible blind. However, 4H-SiC has only one fixed band edge and thus cannot be tailored to the application. Wide band gap group III-nitride semiconductors, in particular, the AlxInyGa1-x-yN material system, have emerged as the most promising material for realization of solar-blind detectors. Initial development of group III-nitride photodetectors was started in the early 1990s [ 17 ] due to the fact that photodetector structures are relatively simple to grow and fabricate. Over the last decade significant progress had been made in the development of AlGaN-based photodetectors, which are able to detect light in part of the UV-C wavelength range (~ 200-280nm). Very high quantum efficiencies, low dark currents and high speeds have been achieved along with several orders of magnitude rejection of sensitivity beyond the solar-blind region. Numerous applications for such devices include flame detection, furnace control, engine monitoring, UV radiation dosimetry, pollution monitoring, and early missile threat warnings [ 18 , 19 ]. Recent progress in the development of LEDs emitting in the wavelength range from 240 nm to 365 nm and LEDs emitting around 340 nm has stimulated the prototyping of chemical/biological battlefield reagent detectors, space communications and non-line-of-sight covert communications, where AlGaN based photodetectors may also play a key role [20]. Photovoltaic detectors such as Schottky barrier detectors, metal-semiconductor-metal (MSM) photodetectors, p-n junction detectors and p-i-n photodetectors have also been demonstrated using the group III-nitride material system (see Figure 9) [21]. The first photovoltaic detector was realized using a Ti/Au Schottky contact to p-GaN doped with Mg up to about 5x1017 cm-3 [22]. The detector had absorbing electrodes on the top surface and was illuminated from the bottom (through the sapphire substrate). Later, Schottky barrier detectors with transparent contacts were also realized using both GaN [ 23] and AlGaN [ 24 -26 ]. Normalized responsivity of the AlGaN photoconductive detectors with different Al mole fractions showed a sharp peak at the cut-off wavelength (see Figure 10).
Figure 9. Schematic representation of photodetectors realized using the IIIN materials system (after ref. 21)
external quantum efficiencies have been recently reported [38-39]. Avalanche detectors potentially offer high internal gain via an avalanche multiplication process. As opposed to photoconductive and MSM detectors, which, in some cases, show slow response speed associated with the carrier trapping, avalanche detectors demonstrate a very high speed operation. The realization of avalanche detectors is very challenging due to the issues related to the non-uniform breakdown via material defects. Nevertheless, GaN avalanche detectors with multiplication ratios of 10-25 have been successfully demonstrated without the microplasma formation [40-42]. Tremendous progress in the arena of group III-nitride visible blind and solar blind photodetectors has led to the developments of phototransistors, imaging detectors, cameras and sensors. High speed p-i-n and Schottky detectors with 3-dB bandwidth of about 1 GHz as well as MSM detectors with about 5 GHz bandwidth have been shown recently [43]. An 8x8 GaN Schottky photodiode array was described as early as 1997 [25]. Recently, large area focal plane arrays [ 44 , 45 ] and cameras [ 46 ] were demonstrated owing to the developments in back illuminated structures. A flame sensor using a solar blind AlGaN p-i-n photodiode has been recently reported [47]. Despite of all the technological advances a wide-spread commercialization of group III-nitride based photodetectors is still delayed by two major research challenges: the requirement of native substrates for the growth of low defect density films and effective p-doping of AlxGa1-xN layers with high Al content.
III.2. SURFACE ACOUSTIC WAVE DEVICES
Figure 10.Spectral response of AlGaN/GaN photodetectrs with different Al fractions.
Various configurations of visible-blind GaN [27-33] and solar-blind AlGaN [ 34 - 37 ] p-n junction based photodetectors have been reported by many groups. These p-i-n detectors offer low bias voltage, low dark currents (due to large hetero-barriers), and high speed. Major disadvantages of p-i-n detectors are related to the difficulties of p-type doping of GaN, especially AlGaN layers with a high Al content [36] and to a high resistance of ohmic contacts to the p-type AlGaN layers. This resistance can be decreased by using p-GaN as the contact layer with iAlxGa1–xN (x > 0.4) as the active layer. However, the GaN contact layer absorbs a significant fraction of the optical beam reducing the device responsivity and, hence, UV/visible selectivity. Therefore, AlGaN/GaN heterostructure-based devices with transparent bottom window layers have attracted significant attention, and such devices with very low dark currents and very high zero-bias
Strong piezoelectric properties of nitride semiconductors enable their applications in Surface Acoustic Wave (SAW) and acousto-optic devices [48]. The SAW propagation in these materials is very sensitive to UV radiation, which enables the implementation of SAW-based UV and solarblind sensors. A remote sensor for wavelengths 365 nm using a GaN-based SAW oscillator was first demonstrated by Ciplys et al. [49]. The output signal of this sensor is in the radio-frequency range enabling the remote wireless signal pickup. Similar SAW oscillator-based sensors for wavelengths both above and below 300 nm were also reported recently [ 50 ]. A perturbation by the SAW of photogenerated charge carrier system in group III-nitride semiconductors leads to a variety of interesting effects, which have a potential for applications in UV sensors of new types.
IV. HIGH-FREQUENCY ELECTRONIC III-N DEVICES IV. 1 BASIC PRINCIPLES An AlGaN/GaN Heterostructure Field Effect Transistor (HFET) is one of the most promising electronic devices for the 21st Century, and has been a topic of intensive investigations since the first report in 1991. A typical
AlGaN/GaN HFET structure is shown in Figure 11; it consists of an AlGaN barrier layer grown on GaN channel layer. The device structures are mainly grown on sapphire or semi-insulating SiC substrates. Since neither of these substrates is lattice-matched to GaN, an AlN buffer layer is generally used to isolate the channel layer from the substrate. Defects and dislocations occur in the epitaxial layers due to lattice mismatch, with typical dislocation densities on the order of 108 to 109 /cm2. The lattice mismatches of GaN on sapphire and SiC substrates are 13% and 3.1%, respectively; therefore, the GaN crystal quality is better on SiC. With the thermal conductivity of SiC being about 10 times that of sapphire, it is evident that SiC is better suited for highpower and high-temperature applications.
Source
Gate
Drain
AlxGa1-xN 2DEG i-GaN
achievements, the potential of nitride based HFETs has not been fully realized as yet. The RF powers expected from the fundamental properties of nitride based materials significantly exceed the experimental data. One of the key problems limiting the HFETs RF characteristics is high gate leakage currents causing DC and RF parameter degradation. When the gate voltage goes positive the forward leakage current shunts the gate-channel capacitance and thus limits the maximum device current. When the gate voltage is negative, high voltage drop between the gate and the drain causes premature breakdown and thus limits maximum applied drain voltage. In addition, gate leakage currents increase the device sub-threshold currents, which decrease the achievable amplitude of the RF output. These limitations become even more severe at high ambient temperatures. The characteristics of group III-Nitride HFETs can be considerably improved by implementing a new approach, which results from the demonstration of good quality of SiO2/AlGaN and Si3N4/AlGaN interfaces. This approach opens up a way to fabricate Insulated Gate Heterostructure Field-Effect Transistors (IGHFETs), which have gate leakage currents several orders of magnitude below those of regular HFETs, and exhibit better linearity and higher channel saturation currents. Source
SiO2
Drain
SiC Figure 11. Typical AlGaN HFET structure.
The bandgap of AlGaN layer is larger than that of GaN layer. Due to this, the electrons from the AlGaN diffuse into GaN and form a very thin highly conducting layer (so-called 2 Dimensional Electron Gas, 2DEG layer). The maximum electron concentration in the 2DEG layer depends on the bandgap energy difference between the barrier and the underlying buffer layer. For AlGaN/GaN heterostructures, this bandgap offset is much higher than that of AlGaAs/GaAs structures. As a result, group III-Nitride HFETs have nearly ten times higher electron concentration in the 2DEG layer. Formation of the 2DEG in AlGaN/GaN HFETs is also affected by the strong polarization that occurs at the AlGaN - GaN interface. The total polarization is the sum of two components: spontaneous (Psp) and piezoelectric (Ppz). Spontaneous polarization arises due to the lack of inversion symmetry in the group III-Nitride crystal structures. In addition, piezoelectric polarization is the result of a strain in the AlGaN layer grown over GaN buffer due to the difference in the crystal lattices. The polarization effects further increase the 2DEG density. Moreover, the polarization effects allow achieving significant 2DEG densities without using any doping in the AlGaN barrier layer. This unique effect is referred to as “polarization doping” in group III-Nitride heterostructures [51]. Several groups demonstrated high power operation of AlGaN/GaN HFETs at microwave frequencies [52,53,54], including a 100 W output power single chip amplifier developed by Cree, Inc. and devices with 100 GHz cut-off frequency reported in [55]. However, in spite of impressive
AlGaN 2DEG
i-GaN
SiC Figure 12. Typical MOSHFET structure.
In the past, different insulating materials were proposed and studied for use as the gate insulator in the IGHFETs. Khan et al. and Binari et al. have reported on GaN-based Metal Insulator Semiconductor Field Effect Transistors (MISFETs) using i-GaN [56], i- AlGaN/GaN [57] and Si3N4 [58] as the gate insulator. These devices were operational but exhibited the current-voltage characteristic collapse at high drain biases related to a large density of interface states. Ren at el. have also reported on a GaN MOSFET using Ga2O3 and Gd2O3 oxides for the gate insulator [59]. All these initial devices suffered from low transconductances and/or high threshold voltages due to large density of the interface states and poor control of the dielectric thickness. In the year 2000, novel AlGaN/GaN Metal-OxideSemiconductor Heterostructure Field Effect Transistors (MOSHFETs) on sapphire [60] and on SiC [61] substrates with excellent gate control were demonstrated. Later on it was also shown that high performance insulated gate HFETs can be made using Si3N4. These devices are referred to as Metal Insulator Semiconductor Heterostructure Field Effect Transistors (MISHFETs). Operation of both device types is based on the high quality of material deposited at the
interface between SiO2 or Si3N4 layer and the AlGaN barrier layer. 0.1
HFET MOSHFET
1E-3
The operating frequency of a FET is determined by the gate length given by
IGS, A
1E-5
fT
1E-7 1E-9 1E-11 1E-13 1E-15 -12
-9
-6
-3
0
3
6
VGS, V Figure 13. Leakage currents comparison for identical geometry HFET and MOSHFET fabricated on the same wafer.
As shown in Figure 12, the MOSHFET design is very close to that of the HFET. The only difference is the presence of insulating layer under the gate that reduces the gate leakage currents by several orders of magnitude. Figure 13 compares the gate currents of identical geometry HFET and MOSHFET fabricated from the same wafer. As seen, both the reverse and forward gate currents of the MOSHFET are much lower than those of the HFET. The suppression of gate currents leads to several key advantages of MOSHFETs for high-power applications. First, it allows for application of positive gate voltages thus further increasing the maximum device currents. For typical HFETs, the peak currents are 1 - 1.2 A/mm. For the MOSHFETs, the currents up to 2 A/mm can be achieved [62]. The most important advantage of the MOSHFET is related to device reliability under high RF power stress. The difference in the device stability is clearly seen from the data of Figure 14. Detailed studies show that the degradation of HFET device types is due to forward gate currents caused by a large-amplitude input signal applied to the gate. 15
RF Power, W/mm
IV.2 HIGH-FREQUENCY PERFORMANCE
12
MOSHFET 9 6
HFET
3 0 0
20
40
60
80
100
Time, Hr Figure 14. RF power stability of identical geometry MOSHFET and HFET at 2 GHz room temperature stress.
Cg
s
2 g m Lg
where f T is the unity gain frequency, Cg is the gate capacitance, gm is the transconductance, Lg is the gate length. Gate length needs to be as short as possible to achieve high operating frequencies. High-frequency performance also depends on electron saturation velocity (s), which is the maximum speed an electron can reach at high electric fields. High saturation velocity is one more advantage of group IIINitride materials, as illustrated in Figure 15.
Figure 15. Velocity - field dependencies for different semiconductor materials.
Recently, AlGaN/GaN Hign Electron Mobility Transistors (HEMTs) with a unity current gain cutoff frequency (fT) of 110 GHz and a maximum frequency of oscillation (f max) of over 140 GHz have been demonstrated by reducing the gate length down to 50 nm [63]. However, this reduction of gate length leads to a decrease in breakdown voltage. To achieve higher performance, indium has been utilized as a surfactant during the growth of AlGaN/GaN HEMTs [61]. A demonstration of the concept used a MOCVD-grown epilayer which consists of a 100 nm AlN buffer, 2 m undoped GaN, a 5 nm undoped Al.25Ga.75N spacer, a 10 nm Si-doped (~5 1018 /cm3) Al.25Ga.75N charge supply layer, and a 10 nm undoped Al.25Ga.75N barrier layer. The layer was grown on 4H-SiC. Hall measurements showed a sheet carrier concentration of 1.1 × 1013 /cm2 and an electron 2 mobility of 1300 cm /V-s at room temperature on as-grown wafers. Device fabricated from this layer demonstrated a typical drain current density of 1.23 A/mm, peak transconductance of 314 mS/mm, and a gate-drain breakdown voltage of over 60 V. Figure 14 shows the shortcircuit current gain (|h21|) and maximum stable gain/maximum available gain (MSG/MAG) derived from on-wafer S-parameter measurements as a function of frequency for the 0.12 µm gate-length device.
(FP) AlGaN/GaN HFETs. Ando et. al. [65] demonstrated microwave powers of 10 W/mm at 2 GHz (gate-length 1 µm, source-drain opening 4.5 µm and Vds = 65 V). Similar fieldplated HFET devices were later reported by Thompson et. al [66] (16.5 W/mm at Vds = 60 V) and Wu et. al. [67] (12.4 W/mm at 48 V and 30 W/mm at Vds = 120 V). The first FPMOSHFET was recently demonstrated at USC [68] with the RF power - drain bias dependencies quite close to those expected from the I-V characteristics. At 55 V drain bias the FP MOSHFET delivered nearly 20 W/mm RF powers at 2 GHz. The device also demonstrated stable operation at 19 W/mm output power level during 100+ hours without any noticeable performance degradation.
Figure 16. Short-circuit current gain (|h21|) and maximum stable maximum available gain (MSG/MAG) of a 0.12 µm × 100 µm AlGaN/GaN HEMT on SiC substrates. After 64
IV.3 HIGH-POWER RF PERFORMANCE Due to high peak currents and high breakdown voltages, group III-Nitride HFETs promise record high output RF powers. For example, a typical AlGaN/GaN HFET with the maximum saturation current about IDSM 1A/mm and the knee voltage VKN 5 V at a moderate operating point of VD = 35 V, should deliver an output power POUT I0(VD - VKN)/2 7.5 W/mm, where I0 IDSM/2 is the operating DC current. However, for a typical device, even under pulsed drain bias and pulsed RF drive conditions eliminating the device self-heating, the measured RF output powers are typically between 2 to 4 W/mm (see Figure 17). The reduction in the RF power is referred to as RF current collapse and is due to intensive electron trapping by the defects in the material at high electric fields.
Figure 17. RF Power - Drain bias dependencies for regular and field-plated III-Nitride HFETs.
The RF current collapse remains the major challenge in the III-Nitride technology for high-power applications in the past ten years. It can be completely eliminated only by achieving extremely high quality materials and by perfect surface passivation. Recently a significant reduction of the current collapse was demonstrated in so-called field-plated
Figure 18. Field-Plated AlGaN/GaN MOSHFET.
V. III-NITRIDE ELECTRONICS
MATERIALS
FOR
HIGH-POWER
Given its unique characteristics, GaN has great potential also for power electronics applications. The most important material characteristics in which GaN has an advantage with respect to Si are wide bandgap (see Figure 3), high electron mobility, large saturation velocity, high sheet carrier concentration at heterojunction interfaces, and high breakdown field (see Figure 5). Besides its electrical characteristics, GaN presents significant advantages from a thermal point of view. First of all, GaN exhibits high thermal conductivity when grown on SiC substrates. Moreover, the coefficient of thermal expansion (CTE) of GaN is very close to the CTE of commonly used heatspreader materials such as AlN (see Figure 19) which should allow the realization of high-reliability thermal packages [69]. All these material advantages hold the promise of realizing high-speed, high-voltage, and high-temperature switching devices. The availability of such devices should allow power electronic specialists to realize power converters operating at higher frequencies, with reduction in the size of passive elements, such as filter inductors and capacitors, and with improvements in terms of achievable bandwidth. The high-voltage devices would simplify high-voltage converter structures, reducing the need for seriesing of devices with associated voltage-sharing problems. The ability of GaN devices to operate at higher temperatures without performance deterioration should make extreme temperature
important result is given by the fact that the switching power density well exceeds 20W/mm.
V.1 ECONOMIC IMPACT OF ONGOING GAN RESEARCH ON POWER ELECTRONICS APPLICATIONS
Si
SiC-4H
GaN
BeO
MMC
Alumina
AlSiC
8 7 6 5 4 3 2 1 0
AlN
CTE (ppm/K)
applications possible and allow heatsink size reduction in more conventional applications.
Figure 19. Thermal coefficient of expansion (CTE) for package substrates and semiconductors.
The use of GaN devices for power electronics is in its infancy. However, in recent times there have been a growing number of publications in this area. A number of publications report results for GaN devices specifically built for power electronics [70-71-72]. A few publications report experimental results related to the performance of GaN devices used in power converter applications. S. Yoshida et al. in [73] presented an application of GaN switches for an inverter application. The HFET operated at 20A with an on-state resistance of 8 mcm2 at 370 V. In terms of performance it is also significant to look at the switching times achieved: 10ns for the turn-on time and 11ns for the turn-off time. The operating frequency of the inverter was 0.37GHz for DC/AC conversion from 30 V to 100V AC. The converter topology consisted of two stages: an input converter with 4 switches operating in parallel and a DC/AC stage in an H-bridge configuration. Saito et al. [74] demonstrated the application of AlGaNGaN HEMT transistor on sapphire substrate. The transistor was designed with 470V breakdown and used in a DC/DC converter topology. The maximum switching frequency achieved in this application has been 200kHz. Matocha et al [75] demonstrated GaN Mosfet with very low on-resistance (1.9 mcm2) and high breakdown voltage (700 V). The USC group presented the first integrated solution based on GaN [76]. In this case an H-bridge topology was used for DC/AC conversion. A photo of the four devices is shown in Figure 20. The circuit has been successfully tested in half bridge operation with a resistive load operating at constant duty-cycle. The most significant limitation experienced in the laboratory testing was due to the driver circuits that limit the switching frequency to well below the GaN capability. For the sake of simplicity, in the first prototype, commercial IGBT drivers have been used in order to easily adjust the control voltage to test the switching behavior of the integrated components. The most
The on-going developmental work in the areas of optoelectronics, sensors, and communications will provide the stimulus needed to carry this technology into Power Electronics for the 21st Century. This economic leveraging should not be underestimated for providing impetus into GaN power device development. This alone may cause GaN to surpass any on-going research efforts in SiC, where the potential and actual economics may limit the investment into the technology [77]. Strategies Unlimited have projected the 2007 GaN market to exceed $5 billion dollars US. The market for GaN is vast and extremely diverse ranging from optoelectronics to microwave electronics to power electronics. Just as GaAs provided the technological tool for the advancement of cellular handsets in the 1990’s, a previously unforeseen application for the technology, GaN is positioned to created electronic markets and products yet to be envisioned.
Figure 20: AlGaN/GaN H-bridge converter. The overall dimensions of the four devices are 1.90 x 1.25mm.
VI. CONCLUSION The paper gives an overview of the state-of-the-art of GaN technology. The significant optoelectronics and RF applications of this material are discussed. The significant investment and material development spurred by these applications are leading to fast improvements in material characteristics. Power electronics applications can leverage on these developments. The first power converter applications of GaN devices are beginning to appear in the recent literature.
ACKNOWLEDGMENT This work was supported by the U.S. Office of Naval Research under Grant N00014-02-1-0623.
REFERENCES
23
1
M. A Khan, J. M. Van Hove J. N. Kuznia and D. T. Olsen, High electron mobility GaN-AlGaN heterostructures grown by LPMOCVD, Appl. Phys. Lett. 58, 2408 (1991) and also: M.A. Khan, J. N. Kuznia, J. M. Van Hove N. Pan and N. Carter, Observation of a two-dimensional electron gas in low pressure MOCVD deposited GaN/AlGaN heterojunctions, Appl. Phys. Lett. 60, 3027 (1992) 2 S. Rumyantsev, M. Shur and M. Levinshtein, Material Properties of Nitrides. Summary, Int. Journ. of High Speed Electronics and systems, V. 14, N1, 1-19 (2004) 3 “HB-LED market reaches $1.8 billion,” Compound Semicond., vol. 9, no. 2, pp. 17, 2003. 4 “High-brightness LEDs 2003 – Market review and forecast,” Strategies Unlimited, June 2003. 5 O. Ambacher, J. Smart, J. R. Shealy, N. G. Weimann, K. Chu, M. Murphy, W. J. Schaff, L. F. Eastman, R. Dimitrov, L. Wittmer, M. Stutzmann, W. Rieger and J. Hilsenbeck, Twodimensional electron gases induced by spontaneous and piezoelectric polarization charges in N- and Ga-face AlGaN/GaN heterostructures, Journ. Appl. Phys. 85, 32223233 (1999) 6 M. Asif Khan, J. W. Yang, G. Simin, R. Gaska, M. S. Shur, Hans-Conrad zur Loye, G. Tamulaitis, A. Zukauskas, David J. Smith, D. Chandrasekhar, and R. Bicknell-Tassius, Lattice and Energy Band Engineering in AlInGaN/GaN Heterostructures , Appl. Phys. Lett 76 (9) pp. 1161-1163 (2000) 7 J. W. Lee and I. Adesida, “Gallium Nitride for Electronics” in: Encyclopedia of RF and Microwave Engineering, John Wiley & Sons, Inc., 2005 ISBN: 0-471-27053-9 8 . T. Nishida, H. Saito, N. Kobayashi, Appl. Phys. Lett. 78, 39273928 (2001). 9 . T. Nishida, H. Saito, N. Kobayashi, Appl. Phys. Lett. 79, 711712 (2001). 10 W. Sun, V. Adivarahan, M. Shatalov, Y. Lee, S. Wu, J. Yang, J. Zhang, and M. A. Khan: Jpn. J. Appl. Phys. 43 (2004) L 1419. 11 J. P. Zhang, X. Hu, Y. Bilenko, J. Deng, A. Lunev, R. Gaska, M. Shatalov, J. Yang, and M. A. Khan, Appl. Phys. Lett. 85 5532 (2004). 12 . J. P. Zhang, H. M. Wang, M. E. Gaevski, C. Q. Chen, Q. Fareed, J. W. Yang, G. Simin, M. Asif Khan, Appl. Phys. Lett. 80, 3542-3544 (2002). 13 . H. M. Wang, J. P. Zhang, C. Q. Chen, Q. Fareed, J. W. Yang, M. Asif Khan, Appl. Phys. Lett. 81, 604-606 (2002). 14 . J. P. Zhang, M. Asif Khan, W. H. Sun, H. M. Wang, C. Q. Chen, Q. Fareed, E. Kuokstis, J. W. Yang, Appl. Phys. Lett. 81, 4392-4394 (2002). 15 W. Sun, V. Adivarahan, M. Shatalov, Y. Lee, S. Wu, J. Yang, J. Zhang, and M. A. Khan: Jpn. J. Appl. Phys. 43 (2004) L 1419. 16 J. P. Zhang, X. Hu, Y. Bilenko, J. Deng, A. Lunev, R. Gaska, M. Shatalov, J. Yang, and M. A. Khan, Appl. Phys. Lett. 85 5532 (2004). 17 M. A. Khan, J. N. Kuznia, D. T. Olson, J. M. Van Hove, M. Blasingame, and L. F. Reitz: Appl. Phys. Lett. 60 (1992) 2917. 18 M. P. Ulmer, M. Razeghi, and E. Bigan: Proc. SPIE 2397 (1995) 210. 19 P. Schreiber, T. Dang, G. Smith, T. Pickenpaugh, P. Gehred, and C.Litton, Proc. SPIE, 3629 (1999) 230. 20 G. A. Shaw and M. Nischan: Proc. SPIE 4396 (2001) 31. 21 M. S. Shur and M. A. Khan: MRS Bull. 22 (1997) 44. 22 M. A. Khan, J. Kuznia, D. T. Olson, M. Blasingame, and A. R. Bhattarai: Appl. Phys. Lett. 63 (1993) 2455.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Q. Chen, J. W. Yang, A. Osinsky, S. Gangopadhyay, B. W. Lim, M. Z. Anwar, M. A. Khan, D. V. Kuksenkov, and H. Temkin: Appl. Phys. Lett. 70 (1997) 2277. A. Osinsky, S. Gangopadhyay, B. W. Lim, M. Z. Anwar, M. A. Khan, D. V. Kuksenkov, and H. Temkin: Appl. Phys. Lett. 72 (1998) 742. B. W. Lim, S. Gangopadhyay, J. W. Yang, A. Osinsky, Q. Chen, M. Z. Anwar, and M. A. Khan: Electron. Lett. 33 (1997) 633. A. Teke, S. Dogan, F. Yun, M. A. Reshchikov, H. Le, X. Q. Liu, H. Morkoc, S. S. Zhang, W. B. Wang, and R. R. Alfano: Solid-State Electron. 47 (2003) 1401. M. A. Khan, Q. Chen, C. J. Sun, M. S. Shur, M. F. Macmillan, R. P. Devaty, and J. Choyke: Proc. SPIE 2397 (1995) 283. X. Zhang, P. Kung, D. Walker, J. Piotrowski, A. Rogalski, A. Saxler, and M. Razeghi: Appl. Phys. Lett. 67 (1995) 2028. Q. Chen, M. A. Khan, C. G. Sun, and J. W. Yang,: Electron. Lett. 31 (1995) 1781. A. Osinsky, S. Gangopadhyay, R. Gaska, B. Williams, M. A. Khan, D. Kuksenkov, and H. Temkin: Appl. Phys. Lett. 71 (1997) 2334. G. Y. Xu, A. Salvador, W. Kim, Z. Fan, C. Lu, H. Tang, H. Morkoc¸ G. Smith, M. Estes, B. Goldenberg, W. Yang, and S. Krishnankutty: Appl. Phys. Lett. 71 (1997) 13. D. Walker, A. Saxler, P. Kung, X. Zhang, M. Hamilton, J. Diaz, and M. Razeghi: Appl. Phys. Lett. 72 (1998) 3303. J. C. Carrano, T. Li, P. A. Grudowski, C. J. Eiting, D. Lambert, J. D. Schaub, R. D. Dupuis, J. C. Campbell: Electron. Lett. 34 (1998) 692. G. B. Parish, S. Keller, P. Kozodoy, J. P. Ibbetson, H. Marchand, P. T. Fini, S. B. Fleischer, S. P. DenBaars, U. K. Mishra, and E. J. Tarsa: Appl. Phys. Lett. 75, (1999) 247. D. Walker, V. Kumar, K. Mi, P. Sandvik, P. Kung, X. H. Zhang, and M. Razeghi: Appl. Phys. Lett. 76 (2000) 403. S. C. Jain, M. Willander, J. Narayan, and R. Van Overstraeten: J. Appl. Phys. 87, (2000) 965. E. J. Tarsa, P. Kozodoy, J. Ibbetson, B. P. Keller, G. Parish, and U. Mishra: Appl. Phys. Lett. 77 (2000) 316. P. Sandvik, K. Mi, F. Shahedipour, R. McClintock, A. Yasan, P. Kung, and M. Razeghi: J. Cryst. Growth 231 (2001) 366. J. C. Campbell, C. J. Collins, M. M. Wong, U. Chowdhury, A. L. Beck, and R. D. Dupuis: Phys. Stat. Solidi (a) 188 (2001) 283. K. A. McIntosh, R. J. Molnar, L. J. Mahoney, A. Lightfoot, M. W. Geis, K. M. Molvar, I. Melngailis, R. L. Aggarwal, W. D. Goodhue, S. S. Choi, D. L. Spears, and S. Verghese: Appl. Phys. Lett. 75 (1999) 3485. J. C. Carrano, D. J. H. Lambert, C. J. Eiting, C.J. Collins, T. Li, S. Wang, B. Yang, A. L. Beck, R. D. Dupuis, and J. C. Campbell: Appl. Phys. Lett. 76 (2000) 924. K. A. McIntosh, R. J. Molnar, L. J. Mahoney, K. M. Molvar, N. Efremow, Jr., and S. Verghese: Appl. Phys. Lett. 76 (2000) 3938. E. Ozbay, N. Biyikli, I. Kimukin, T. Kartaloglu, T. Tut, and O. Aytür: IEEE J. of Sel. Topics in Quant. Electron. 10 (2004) 742. B. Yang, B. Yang, K. Heng, T. Li, C. J. Collins, S. Wang, R. D. Dupuis, J. C. Campbell, M. J. Schurman, and I. T. Ferguson: IEEE J. Quant. Electron. 36 (2000) 1229. P. Lamarre, A. Hairston, S. Tobin, K. K. Wong, M. F. Taylor, A. K. Sood, M. B. Reine, M. J. Schurman, I. T. Fergusson, R. Singh, and C. R. Eddy Jr.: Proc. Mat. Soc. Symp. 639 (2001) G10.0.1. J. D. Brown, J. Boney, J. Matthews, P. Srinivasan, J. F. Schetzina, T. Nohava, W. Yang, S. Krishnankutty: MRS Internet J. Nitride Semicond. Res. 5 (2000) 6.
47
A. Hirano, C. Pernot, M. Iwaya , T. Detchprohm, H. Amano, and I. Akasaki: Phys. Stat. Sol. (a) 188, (2001) 293. 48 D. Ciplys, R. Gaska, M. S. Shur, R. Rimeika, J. Yang, and M. Asif Khan: Proc. SPIE 4453 (2001) 37. 49 D. Ciplys, R. Rimeika, M. S. Shur, S. Rumyantsev, R. Gaska, A. Sereika A., J. Yang and M. A. Khan: Appl. Phys. Lett. 80 (2002) 2020. 50 D. Ciplys, M. S. Shur, N. Pala, A. Sereika, R. Rimeika, R. Gaska, and Q. Fareed, Proc. IEEE Sensor Conference, Vienna (2004) (to be published). 51 M. Asif Khan, J. W. Yang, G. Simin, R. Gaska, M. S. Shur, A. Bykhovsky, Piezoelectric Doping in AlInGaN/GaN Heterostructures , Appl. Phys. Lett., 75 (18) , pp. 2806-2808, 1999 52 T. Sheppard, W. L. Pribble, D.T.Emerson, Z. Ring, R. P. Smith, S. T. Allen, and J. W. Palmour, Device Research Conference Digest (2000) p. 37 53 N.X. Nguyen, M. Micovic, W.-S. Wong, P. Hashimoto, L.-M. McCray, P. Janke and C. Nguyen, Electron. Lett. 36 (2000) p. 468 54 Yi-Feng Wu; Kapolnek, D.; Ibbetson, J.P.; Parikh, P.; Keller, B.P.; Mishra, U.K., Very-high power density AlGaN/GaN HEMTs, IEEE Transactions on Electron Devices, V. 48, 586590 (2001) 55 Wu Lu; Jinwei Yang; Khan, M.A.; Adesida, I. AlGaN/GaN HEMTs on SiC with over 100 GHz fT and low microwave noise, IEEE Transactions on Electron Devices, V. 48, 581 -585 (2001) 56 M. Asif Khan, J. N. Kuznia, A. R. Bhattarai, and D. P. Olson, GaN/AlGaN Heterostructure Deposition by Low Pressure MOCVD for MISFET Devices, Material Research Society Proceedings, vol. 281, p.769 (1993) 57 M. Asif Khan, Michael S. Shur, Q. C. Chen, and J. N. Kuznia, Current-Voltage Characteristic Collapse in AlGaN/GaN Heterostructure Insulated Gate Field Effect Transistors at High Drain Bias, Electronics Letters, Vol. 30, No. 25, p. 2175, Dec. 8, 1994 58 . S. C. Binari, L. B. Rowland, G. Kelner, W. Kruppa, H. B. Dietrich, K. Doverspike, D. K. Gaskill, "DC, Microwave, and High-Temperature Characteristics of GaN FET Structures," in Int. Sym. Compound Semiconductors, ed. H. Goronkin (IOP Publishing, Bristol, 1995), p.459 59 F. Ren, M. Hong, S. N. G. Chu, and M .A. Marcus, M.J. Schurman, A. Baca, S. J. Pearton and C.R. Abernathy, Effect of Temperature on Ga2O3 (Gd2O3)/GaN Metal Oxide Semiconductor Field Effect Transistor, Appl. Phys. Lett. 73, 3893 (1998) 60 M. Asif Khan, X. Hu, G. Simin, A. Lunev, and J. Yang, R. Gaska and M. S. Shur, "AlGaN/GaN Metal-OxideSemiconductor Heterostructure Field Effect Transistor", IEEE Electron Device Letters, vol. 21, pp.63-65 (2000) 61 M. Asif Khan, X. Hu, G. Simin, and J. Yang, R. Gaska and M.S. Shur, AlGaN/GaN Metal-Oxide-Semiconductor Heterostructure Field Effect Transistors on SiC Substrates, Appl. Phys. Lett. v. 77 pp 1339-1341 (2000) 62 A. Koudymov, H. Fatima, G. Simin, J. Yang, M. Asif Khan, A. Tarakji, X. Hu, M. S. Shur, & R. Gaska Maximum Current in Nitride-Based Heterostructure Field Effect Transistors Appl. Phys. Lett, V. 80 pp. 3216-3218 (2002) 63 M. Micovic, N. X. Nguyen, P. Janke, W. S. Wong, P. Hashimoto, L. M. McCray, and C. Nguyen, “GaN/AlGaN high electron mobility transistors with fT of 110 GHz,” Electron. Lett., vol. 36, pp. 358–359, Feb. 2000. 64 V. Kumar, W. Lu, R. Schwindt, A. Kuliev, G. Simin, J. Yang, M. A. Khan, and I. Adesida, “AlGaN/GaN HEMTs on SiC
with fT of over 120 GHz,” IEEE Electron Device Lett., vol. 23, No. 8, pp. 455–457, Aug. 2002. 65 . Y. Ando, Y. Okamoto, H. Miyamoto, T. Nakayama, T. Inoue, and M. Kuzuhara, “10-W/mm AlGaN–GaN HFET With a Field Modulating Plate”, IEEE EDL, V. 24, 289-291 (2003) 66 R. Thompson, T. Prunty, V. Kaper, and J. R. Shealy, “Performance of the AlGaN HEMT Structure With a Gate Extension”, IEEE Trans. on ED, V. 51 292 - 295 (2004) 67 . Y.-F. Wu, A. Saxler, M. Moore, R. P. Smith, S. Sheppard, P. M. Chavarkar, T. Wisleder, U. K. Mishra, and P. Parikh, ”30W/mm GaN HEMTs by Field Plate Optimization”, IEEE EDL, V. 25, 117-119 (2004) 68 A. Koudymov et. al. Current collapse removal in field-plated AlGaN/GaN MOSHFETs. Submitted to IEEE EDL, April 2005 69 J.L. Hudgins, G. Simin, E. Santi and M.A. Khan, "An assessment of wide bandgap semiconductors for power devices," IEEE Trans. Power Electronics, Vol. 18, No. 3, pp. 907-914, May 2003 70 C. Park, P.L. Chapman, S.H. Rhee, S.J. X.HongZhang, P.T. Krein, K. Kim, GaN power switching device growth by plasma assisted molecular beam epitaxy, Conference Record of the 2002 Industry Applications Conference, 37th IAS Annual Meeting, Volume 1, 13-18 Oct. 2002, pp. 576 - 579 vol.1 71 S. Karmalkar, Jianyu Deng, M.S. Shur, IEEE Electron Device Letters, RESURF AlGaN/GaN HEMT for high voltage power switching, Volume 22, Issue 8, Aug. 2001 pp. 373 - 375 72 Naiqian Zhang, V. Mehrotra, S. Chandrasekaran, B. Moran, Likun Shen; U. Mishra, E. Etzkorn, D. Clarke, Large area GaN HEMT power devices for power electronic applications: switching and temperature characteristics, 2003 IEEE 34th Annual Power Electronics Specialist Conference, PESC '03 Volume 1, 15-19 June 2003, pp. 233 - 237 vol.1 73 S. Yoshida, J. Li, T. Wada and H. Takehara. “High-power AlGaN/GaN HFET with lower on-state resistance and higher switching time for an inverter circuit, on IEE Proc. On Circuits Devices and Systems, vol. 151, N. 3, June 2004, pp. 207-210 74 W. Saito, M. Kuraguchi, Y. Takada, K. Tsuda, Ichiro Omura,and T. Ogura, High Breakdown Voltage Undoped AlGaN–GaN Power HEMT on Sapphire Substrate and Its Demonstration for DC–DC Converter Application IEEE Transactions On Electron Devices, VOL. 51, NO. 11, Nov. 2004, pp. 1913-1917 75 K. Matocha, T. P. Chow and R. J. Gutmann, High-Voltage Normally Off GaN MOSFETs on Sapphire Substrates, IEEE Transactions On Electron Devices, VOL. 52, NO. 1, Jan. 2005, pp. 6-9 76 S.G. Pytel, S. Lentijo, A. Koudymov, S. Rai, H. Fatima, V. Adivarahan, A. Chitnis, J. Yang, J.L. Hudgins, E. Santi, A. Monti, G. Simin, and M. Asif Khan, “AlgaN/GaN MOSHFET Integrated Circuit Power Converter”, on Proc. of IEEE PESC 2004, Aachen (Germany), June 2004. 77 J.L. Hudgins, Gallium nitride – the other wide bandgap semiconductor, IEEE Power Electronics Society Newsletter, 2005
*Department of Electrical Engineering 209N WSEC, PO Box 880511 University of Nebraska Lincoln, NE 68588 [email protected]
I. INTRODUCTION The interest in III-Nitride semiconductors dates back to 1970s. It was initially stimulated by pioneering works by Pankove, Akasaki, Nakamura, Khan and many others showing tremendous potential of these materials for optoelectronics. Nearly at the same time, field-effect transistors have been recognized as the most promising electronic devices, which has been a topic of intensive investigations since the first report in 1991 [1]. The major
advantages of nitride-based devices that make them extremely promising for high-power, high-temperature applications are high electron mobility and saturation velocity, high sheet carrier concentration at heterojunction interfaces, high breakdown field, and low thermal impedance when grown over SiC substrates. The chemical inertness and excellent temperature stability of nitrides are the key properties to provide high reliability. Owing to these unique properties, group III-Nitride semiconductors provide great promise for overcoming the fundamental limitations associated with silicon and gallium arsenide semiconductors.
1000
600 400 200 0 1970
1975
1980
1985
2D electron gas
800
1990
Blue LED
1200
Breakthrough in growth
Number of publications
Abstract: Wide bandgap III-Nitride semiconductor materials possess superior material properties as compared to silicon, GaAs and other III-V compound materials. Recent achievements in Gallium Nitride (GaN) technology for optoelectronics have resulted in ultra-bright blue light emitting diodes and lasers, ultraviolet emitters, and solar-blind optical detectors. In the electronic area, drastic improvement of microwave device performance has been achieved, yielding record high power densities of 20 - 30 W/mm. Novel applications of these materials in high-power electronics for switching, energy conversion and control are just emerging. This paper provides an overview of the state-of-the-art IIINitride wide bandgap technologies and it explores power electronic applications while illustrating the enormous potential that GaN based devices have for overcoming the major challenges of power electronics in the 21st Century. The paper discusses the unique material and device properties of GaN-based semiconductors that make them promising for high-power, high-temperature applications. These include high electron mobility and saturation velocity, high sheet carrier concentration at heterojunction interfaces, high breakdown voltages, and low thermal impedance (when grown over SiC or bulk AlN substrates). The chemical inertness and radiation hardness of nitrides are other key properties. As applied to power electronics, the III-Nitride technology allows for high-power switching with sub-microsecond and nano-second switching times. The paper will present the innovations that further improve the performance of highpower DC-DC converters, switches and other building blocks. These include novel insulated gate HFET design that significantly expands the allowable input voltage amplitude, further increases the device peak currents, and most importantly, tremendously improves the large-signal stability and reliability. Insulated gate switching devices have been shown to operate at up to 300C with no noticeable parameter degradation. Novel monolithic integrated circuits of high-power switches and DC-DC converters and their performance parameters will be presented. The paper also discusses the major challenges associated with modern GaN technology and work in progress to overcome them.
1995
2000
Year Figure 1. Progress in the number of publications per year in group IIINitride material and device research (after [2]).
Figure 1 shows the number of publications per year related to the III-Nitride semiconductors during past thirty years. As seen, after 1980s - 1990s when significant progress in MOCVD epitaxial growth was achieved, and the first electronic and optoelectronic devices were successfully demonstrated, the interest in the field increased tremendously. In the past ten years, the research in the IIINitride device applications has mainly focused on optoelectronics and microwave electronics. This resulted in several novel unique device types, which include blue and ultraviolet light emitters and lasers, visible- and solar-blind photodetectors and high-power microwave field-effect transistors. The device improvement goes alongside with material growth development, which led to tremendous improvement in the material quality and significant cost reduction. Following the tremendous progress in the material quality and device processing, a worldwide market for nitride-based devices has developed. One of the highest profile markets for nitride-based LEDs is in mobile phones,
The bandgap energy values for the most important semiconductor materials are shown in Figure 3. As one can see the bandgap energy of GaN, EG 3.4 eV, is more than three times higher than that of Si (1.1 eV) and more than two times higher than that of GaAs (1.41 eV). Since the probability of ionization and many other processes depend exponentially on the bandgap energy, a large bandgap is the key factor for high temperature operation, chemical inertness, and high breakdown voltage of GaN devices. The bandgap energy of AlN is even higher, ~ 6.2 eV. The bandgap value of the third important III-Nitride material, InN, is yet somewhat unclear, between 0.95 and 1.9 eV. 7
Bandgap.eV
particularly blue LEDs for keypads and white LEDs for liquid crystal display (LCD) backlight. Nitride-based LEDs also offer great promise for solid-state lighting applications such as architectural lighting, machine vision, illumination for signage, flashlights, decorative lights, and so on. The electronic device market includes high-power RF sources for base stations, transmitter-receiver modules, sensors and other high-power, high-temperature applications. The market for high-brightness LEDs (HB-LEDs) in lighting applications has grown rapidly with an overall market in the billions of dollars [3]. Based on continuing positive trends in this dynamic industry, the market for HB-LEDs is projected to grow to ~ $5.0 billion by 2007 [4]. Emerging applications of III-Nitrides in power electronics may significantly increase and expand the potential market. Electric vehicles, power station electronic switches and electric ships are a few of the areas that can be drastically changed by these group III-Nitride semiconductors. In the next sections, we will provide an overview of group III-Nitride semiconductor materials properties, growth techniques, main achievements in the fields of optoelectronics and microwave devices. We will also show how group III-Nitride materials can revolutionize the area of high-power electronics and review the challenges in this emerging technology.
6 5 4 3 2 1 0 Si
GaAs AlAs
SiC
GaN
AlN
Figure 3. The bandgaps of the semiconductor materials
II. MATERIAL PROPERTIES AND GROWTH The group of III-Nitride semiconductors includes three main materials, Gallium Nitride (GaN), Aluminum Nitride (AlN) and Indium Nitride (InN). All three materials crystallize mostly in so-called wurtzite structures that have a hexagonal unit cell. An example of GaN crystal structure is shown in Figure 2. An important property of this crystal cell is the lack of inversion symmetry, which leads to very strong polarization effects in group III-Nitride materials.
Figure 2. Schematic of the wurtzite GaN crystal structure [5]
Another fundamental property that makes group IIINitride materials extremely attractive for high-power, hightemperature applications is a very large bandgap, i.e. the energy required to ionize atoms and create free electrons.
There is one more property of group III-Nitride materials that is of crucial importance for optoelectronic and electronic applications. GaN, AlN and to some extent, InN can form alloys by mixing different fractions of Ga, Al and In in ternary and quaternary compounds, e.g. Ga xAl1-xN, GaxIn1-xN, Gax,InyAl1-x-yN. The bandgap energy of these compound materials is approximately equal to a weighted linear combination of the bandgap energies of the binary materials forming the alloy (so-called Vegard's law). For example, adding different fractions of Al into GaN compound, the bandgap energy can be varied from 3.4 eV up to 6.2 eV. This approached is called “band-gap engineering” [ 6 ] Normally, the ternary or quaternary materials are grown on top of GaN or AlN substrates. Practical implementations of bandgap engineering are limited by changes in lattice constants that create a strain between the layers of different compositions. Group III-Nitride materials belong to so-called directbandgap materials. This essentially means that the energy released by electron-hole recombination in these materials can be directly converted into light. This property makes group III-Nitrides extremely important for optoelectronic applications. Figure 4 shows the range of optical wavelength that can be covered by AlN - GaN - InN alloys. As seen, ternary and quaternary compounds of AlGaNInN can produce light not only in the entire visible part of the spectrum (0.7 – 0.4μm) but also in the deep UV range of it.
Figure 4. Optical spectrum covered by III-Nitride materials.
For high-power electronic applications, high breakdown voltages achievable in group III-Nitride materials are of primary importance. Figure 5 compares the critical electric fields for the avalanche breakdown mechanism in the most important semiconductor materials. As one can see Nitride based semiconductors possess characteristic breakdown fields exceeding nearly ten times those of Si or GaAs. The extremely high breakdown fields results in record-high breakdown voltages even when applied across micrometer scale spacing.
V/cm
4.00E+06 3.00E+06 2.00E+06 1.00E+06 0.00E+00 Si
GaAs
SiC
GaN
AlN
Figure 5. Breakdown fields of different semiconductor materials.
To-date, most of AlGaN/GaN and AlInGaN/GaN heterostructures are grown by Metal Organic Chemical Vapor Deposition (MOCVD). MOCVD systems are capable to operate in both conventional and atomic layer deposition regimes. Conventional deposition regime wherein precursors enter the growth chamber simultaneously is used to deposit GaN layers. Triethylgallium and ammonia are used as the precursor gases. AlxGa1-xN layers are deposited in atomic layer regime when precursors enter the chamber in a cyclic fashion. Triethylgallium, triethylaluminum and ammonia are used as precursors. The precursors are introduced into the chamber using hydrogen or nitrogen as a carrier gas. Epilayers are deposited on sapphire or SiC substrates placed on graphite susceptor, which is heated to the growth temperature by rf-induction.
AlGaN/GaN heterostructures are typically grown on basal (0001) sapphire, conducting 6H-SiC and semi-insulating 4H-SiC substrates. The substrates are slightly off-axis (12o) production grade 6H and 4H silicon carbide available from Cree Research. The doping levels of both n- and p-type 6H-SiC are on the order of 2x1018 cm-3. The micropipe density is approximately 30 cm-2. The deposition of approximately 50 nm thick of AlN buffer is followed by the growth of insulating/modulation doped GaN, which is finally capped with AlGaN barrier layer. The major difference between the AlGaN/GaN growth on sapphire and SiC substrates is the thickness of the insulating GaN layer. The cross-sectional TEM analysis of GaN grown on sapphire and SiC reveals strong dependence of growth defect distribution along the growth direction on the substrate material. The significant reduction in the number of threading dislocations in GaN on sapphire is observed for layer thicknesses above 1.5-2 m. The similar improvement in material quality for GaN grown on SiC was achieved at the thickness as low as 0.5 m or even lower.
III. OPTOELECTRONIC APPLICATIONS OF III-NITRIDES III.1. LIGHT EMITTERS The advent of blue light emitting diodes (LEDs) and short-wavelength blue-violet laser diodes using the group III-nitrides alloys (AlN, GaN, and InN) have opened up an era of new optical devices, which was not possible with the established III-V material systems based on GaAs, AlAs, GaP, InAs and related alloys. The direct bandgap and a wide spectrum of bandgap energies (0.7 eV for InN, 3.4 eV for GaN, and 6.2 eV for AlN) of group III-nitride semiconductors make them suitable for a wide range of optoelectronic applications [7]. While “visible” technology has been reaching its maturity, many groups have shifted the research focus towards shorter wavelength ultraviolet devices. Visible and solar blind nitride detectors have reached outstanding performance at very short cut-off wavelengths. The quantum efficiency of near ultraviolet (UV) LED received a huge boost with advanced growth, metallization and packaging techniques. The emission wavelengths of LEDs now cover the majority of the UV spectrum ranging from 400 nm to 250 nm. Finally, UV LED emission has shifted towards shorter wavelengths with viable operation demonstrated at 340 – 350 nm. Due to their enormous potential applications several research groups are actively developing deep UV LEDs. Nishida et al. have reported on milliwatt power UV LEDs with emission around 342 nm and 352 nm over SiC and hydride vapor phase epitaxial GaN substrates, respectively [8, 9]. The University of South Carolina (USC) has achieved the milliwatt powers in the wavelength range of 260 - 280 nm and emission at shortest wavelength to date of 254 nm [10,11]. A typical UV LED structure is shown in Figure 6. Both SiC and GaN substrates provided good heat removal
allowing the device to operate at high dc currents, but the optical absorption in the substrate (top p-layers and semitransparent contacts) significantly reduced the external quantum efficiency.
p-AlGaN AlGaN (5) QW
n+-AlGaN barrier AlN/AlGaN Nested SL MEMOCVD AlN Buffer
sapphire Figure 6. Typical design and layer structure of AlGaN/GaN based UV LED on sapphire substrate (USC)
The LED structure forms a p-n junction made out of wide bandgap material. The active (recombination) area is located in the space-charge region of the p-n junction and consists of a material with slightly narrower bandgap (socalled “quantum well”). The narrower bandgap effectively confines the electron-hole pairs thus greatly increasing the probability of recombination. Deep UV LEDs require AlGaN layers with Al content greater than 40 %. The internal quantum efficiency of such LEDs is mainly limited by the material quality and low doping efficiency, especially for Mg doped p-AlGaN. In device structures with high Alcontent AlGaN layers, these two factors determine the nonradiative recombination rate and injection efficiency. Novel USC devices utilized an innovative AlN/Al0.5Ga0.5N superlattice (SL) approach for deposition of 2 m thick AlxGa1-xN (x>0.35) buffer layers with significantly lower defect densities [ 12 , 13 ]. The number of non-radiative defects is itself a strong function of the buffer and the active layers material quality. In our recent devices the buffer layer and strain relief AlN/Al0.5Ga0.5N SL were grown using PALE approach allowing for the high structural quality [14]. This growth approach led to a significant increase of the LED quantum efficiency.
5.6 5.4 5.2
Normalized Intensity, a.u.
p+-GaN
Energy, eV 5
4.8
4.6
4.4
4.2
4
20 mA dc
220 230 240 250 260 270 280 290 300 310 320
Wavelength, nm Figure 7. Wavelength tuning in AlGaN/AlGaN Deep UV LEDs with different Al fraction.
By changing the Al fraction in the active and cladding regions of the LED, the emission wavelength can be tuned in the wide range as shown in Figure 7. The LED operates by applying forward bias to the p-n junction. The bias leads to an injection of electrons and holes into a quantum well. Ideally, all the electron-hole pairs should recombine in the quantum well resulting in complete conversion of the electrical current into a photon flux. However, as stated above, the conversion efficiency greatly depends on the defect concentration, the concentration of donor and acceptor impurities creating the electron-hole pairs required for efficient injection and on energy band profile in the cladding and active layers. Another fundamental problem in achieving high conversion efficiency is the extraction of the photon produced in the semiconductor. All semiconductor materials have refractive indices, n, greater than unity. For example, for GaAs and AlGaAs compunds, n 3.5, for GaN and AlGaN compounds, n 2.5. When the photons traveling inside the semiconductor reach the semiconductor - air interface under certain angle, they experience full internal reflection. Only the optical beams propagating perpendicularly to the surface or at the angles less than a certain critical angle can be extracted. As a result, the overall efficiency of LEDs is well below 100%. Material quality, doping, epitaxial layer design, dye shape and packaging play crucial roles in achieving high efficiency LEDs. The highest output powers obtained in the 255 nm 280 nm deep UV LEDs are shown in Figure 8 [15,16].
3.5
dc
3.0
280 nm
Power, mW
2.5 2.0 1.5
265 nm
1.0
255 nm
0.5 0.0
0
20
40
60
80 100 120 140 160 180 200
Current, mA Figure 8. Optical powers of AlGaN/GaN deep UV LEDs
III.2. PHOTODETECTORS Along with light emitters, light detectors are equally important elements of many optoelectronic systems. Traditionally, the detection of light in the ultraviolet (UV) part of electromagnetic spectrum has been realized using photo-multiplier tubes (PMT). These devices offer unprecedented sensitivity in UV range, low dark currents and high speeds. The high internal gain (> 10 6) achieved through the photomultiplication process makes PMTs very effective for detection of low intensity signals. However, very high operating voltages (typically > 1 kV) requiring special (and sometimes very bulky) power supplies make the PMT much less attractive for portable applications. In addition, PMTs are very fragile vacuum tube devices that are poorly suited for many space and field applications. An alternative approach uses UV-enhanced Si detectors that are typically p-i-n diodes with a special anti-reflection coating designed for the UV wavelength range. Si has a direct band gap at 3 eV, so photons in the UV ranges are absorbed efficiently near the front surface of the device. These semiconductor detectors can be made miniaturized, highly reliable, robust, and well integrated with other semiconductor components. Although the UV-enhanced Si detectors have many important advantages, their dark currents are high (typically in the nA range) and quantum efficiencies are poor. Because Si has a very high absorption constant beyond 3 eV, carriers are generated very near the front surface, and minority carriers may be lost through surface state recombination. For operation in the deep UV, the efficiency of these detectors may suffer from degradation of the SiO2/Si interface after prolonged UV exposures, due to the generation of surface states. Both PMT and Si detectors are sensitive throughout the UV and visible wavelengths thus requiring costly UV filters to achieve visible-blind or solar-blind operation. Certain outdoor detector applications would benefit from operation in total darkness to reduce or eliminate background currents, and since the solar spectrum terminates at 280 nm, they need to be non-responsive to wavelengths greater than 280 nm (hence “solar blind”). Other applications in the near UV
(below 300 nm) would also benefit from being nonresponsive to visible light, be it sunlight or indoor lighting. Thus an alternative to PMTs and Si is desirable. Other semiconductor detector materials have also been pursued by researchers including Ge, GaAs, and SiC, in order to combine the sensitivity of a PMT and robustness of Si with “blindness” to longer wavelengths. Ge and GaAs offer no obvious advantages over Si, but 4H-SiC is certainly visible blind. However, 4H-SiC has only one fixed band edge and thus cannot be tailored to the application. Wide band gap group III-nitride semiconductors, in particular, the AlxInyGa1-x-yN material system, have emerged as the most promising material for realization of solar-blind detectors. Initial development of group III-nitride photodetectors was started in the early 1990s [ 17 ] due to the fact that photodetector structures are relatively simple to grow and fabricate. Over the last decade significant progress had been made in the development of AlGaN-based photodetectors, which are able to detect light in part of the UV-C wavelength range (~ 200-280nm). Very high quantum efficiencies, low dark currents and high speeds have been achieved along with several orders of magnitude rejection of sensitivity beyond the solar-blind region. Numerous applications for such devices include flame detection, furnace control, engine monitoring, UV radiation dosimetry, pollution monitoring, and early missile threat warnings [ 18 , 19 ]. Recent progress in the development of LEDs emitting in the wavelength range from 240 nm to 365 nm and LEDs emitting around 340 nm has stimulated the prototyping of chemical/biological battlefield reagent detectors, space communications and non-line-of-sight covert communications, where AlGaN based photodetectors may also play a key role [20]. Photovoltaic detectors such as Schottky barrier detectors, metal-semiconductor-metal (MSM) photodetectors, p-n junction detectors and p-i-n photodetectors have also been demonstrated using the group III-nitride material system (see Figure 9) [21]. The first photovoltaic detector was realized using a Ti/Au Schottky contact to p-GaN doped with Mg up to about 5x1017 cm-3 [22]. The detector had absorbing electrodes on the top surface and was illuminated from the bottom (through the sapphire substrate). Later, Schottky barrier detectors with transparent contacts were also realized using both GaN [ 23] and AlGaN [ 24 -26 ]. Normalized responsivity of the AlGaN photoconductive detectors with different Al mole fractions showed a sharp peak at the cut-off wavelength (see Figure 10).
Figure 9. Schematic representation of photodetectors realized using the IIIN materials system (after ref. 21)
external quantum efficiencies have been recently reported [38-39]. Avalanche detectors potentially offer high internal gain via an avalanche multiplication process. As opposed to photoconductive and MSM detectors, which, in some cases, show slow response speed associated with the carrier trapping, avalanche detectors demonstrate a very high speed operation. The realization of avalanche detectors is very challenging due to the issues related to the non-uniform breakdown via material defects. Nevertheless, GaN avalanche detectors with multiplication ratios of 10-25 have been successfully demonstrated without the microplasma formation [40-42]. Tremendous progress in the arena of group III-nitride visible blind and solar blind photodetectors has led to the developments of phototransistors, imaging detectors, cameras and sensors. High speed p-i-n and Schottky detectors with 3-dB bandwidth of about 1 GHz as well as MSM detectors with about 5 GHz bandwidth have been shown recently [43]. An 8x8 GaN Schottky photodiode array was described as early as 1997 [25]. Recently, large area focal plane arrays [ 44 , 45 ] and cameras [ 46 ] were demonstrated owing to the developments in back illuminated structures. A flame sensor using a solar blind AlGaN p-i-n photodiode has been recently reported [47]. Despite of all the technological advances a wide-spread commercialization of group III-nitride based photodetectors is still delayed by two major research challenges: the requirement of native substrates for the growth of low defect density films and effective p-doping of AlxGa1-xN layers with high Al content.
III.2. SURFACE ACOUSTIC WAVE DEVICES
Figure 10.Spectral response of AlGaN/GaN photodetectrs with different Al fractions.
Various configurations of visible-blind GaN [27-33] and solar-blind AlGaN [ 34 - 37 ] p-n junction based photodetectors have been reported by many groups. These p-i-n detectors offer low bias voltage, low dark currents (due to large hetero-barriers), and high speed. Major disadvantages of p-i-n detectors are related to the difficulties of p-type doping of GaN, especially AlGaN layers with a high Al content [36] and to a high resistance of ohmic contacts to the p-type AlGaN layers. This resistance can be decreased by using p-GaN as the contact layer with iAlxGa1–xN (x > 0.4) as the active layer. However, the GaN contact layer absorbs a significant fraction of the optical beam reducing the device responsivity and, hence, UV/visible selectivity. Therefore, AlGaN/GaN heterostructure-based devices with transparent bottom window layers have attracted significant attention, and such devices with very low dark currents and very high zero-bias
Strong piezoelectric properties of nitride semiconductors enable their applications in Surface Acoustic Wave (SAW) and acousto-optic devices [48]. The SAW propagation in these materials is very sensitive to UV radiation, which enables the implementation of SAW-based UV and solarblind sensors. A remote sensor for wavelengths 365 nm using a GaN-based SAW oscillator was first demonstrated by Ciplys et al. [49]. The output signal of this sensor is in the radio-frequency range enabling the remote wireless signal pickup. Similar SAW oscillator-based sensors for wavelengths both above and below 300 nm were also reported recently [ 50 ]. A perturbation by the SAW of photogenerated charge carrier system in group III-nitride semiconductors leads to a variety of interesting effects, which have a potential for applications in UV sensors of new types.
IV. HIGH-FREQUENCY ELECTRONIC III-N DEVICES IV. 1 BASIC PRINCIPLES An AlGaN/GaN Heterostructure Field Effect Transistor (HFET) is one of the most promising electronic devices for the 21st Century, and has been a topic of intensive investigations since the first report in 1991. A typical
AlGaN/GaN HFET structure is shown in Figure 11; it consists of an AlGaN barrier layer grown on GaN channel layer. The device structures are mainly grown on sapphire or semi-insulating SiC substrates. Since neither of these substrates is lattice-matched to GaN, an AlN buffer layer is generally used to isolate the channel layer from the substrate. Defects and dislocations occur in the epitaxial layers due to lattice mismatch, with typical dislocation densities on the order of 108 to 109 /cm2. The lattice mismatches of GaN on sapphire and SiC substrates are 13% and 3.1%, respectively; therefore, the GaN crystal quality is better on SiC. With the thermal conductivity of SiC being about 10 times that of sapphire, it is evident that SiC is better suited for highpower and high-temperature applications.
Source
Gate
Drain
AlxGa1-xN 2DEG i-GaN
achievements, the potential of nitride based HFETs has not been fully realized as yet. The RF powers expected from the fundamental properties of nitride based materials significantly exceed the experimental data. One of the key problems limiting the HFETs RF characteristics is high gate leakage currents causing DC and RF parameter degradation. When the gate voltage goes positive the forward leakage current shunts the gate-channel capacitance and thus limits the maximum device current. When the gate voltage is negative, high voltage drop between the gate and the drain causes premature breakdown and thus limits maximum applied drain voltage. In addition, gate leakage currents increase the device sub-threshold currents, which decrease the achievable amplitude of the RF output. These limitations become even more severe at high ambient temperatures. The characteristics of group III-Nitride HFETs can be considerably improved by implementing a new approach, which results from the demonstration of good quality of SiO2/AlGaN and Si3N4/AlGaN interfaces. This approach opens up a way to fabricate Insulated Gate Heterostructure Field-Effect Transistors (IGHFETs), which have gate leakage currents several orders of magnitude below those of regular HFETs, and exhibit better linearity and higher channel saturation currents. Source
SiO2
Drain
SiC Figure 11. Typical AlGaN HFET structure.
The bandgap of AlGaN layer is larger than that of GaN layer. Due to this, the electrons from the AlGaN diffuse into GaN and form a very thin highly conducting layer (so-called 2 Dimensional Electron Gas, 2DEG layer). The maximum electron concentration in the 2DEG layer depends on the bandgap energy difference between the barrier and the underlying buffer layer. For AlGaN/GaN heterostructures, this bandgap offset is much higher than that of AlGaAs/GaAs structures. As a result, group III-Nitride HFETs have nearly ten times higher electron concentration in the 2DEG layer. Formation of the 2DEG in AlGaN/GaN HFETs is also affected by the strong polarization that occurs at the AlGaN - GaN interface. The total polarization is the sum of two components: spontaneous (Psp) and piezoelectric (Ppz). Spontaneous polarization arises due to the lack of inversion symmetry in the group III-Nitride crystal structures. In addition, piezoelectric polarization is the result of a strain in the AlGaN layer grown over GaN buffer due to the difference in the crystal lattices. The polarization effects further increase the 2DEG density. Moreover, the polarization effects allow achieving significant 2DEG densities without using any doping in the AlGaN barrier layer. This unique effect is referred to as “polarization doping” in group III-Nitride heterostructures [51]. Several groups demonstrated high power operation of AlGaN/GaN HFETs at microwave frequencies [52,53,54], including a 100 W output power single chip amplifier developed by Cree, Inc. and devices with 100 GHz cut-off frequency reported in [55]. However, in spite of impressive
AlGaN 2DEG
i-GaN
SiC Figure 12. Typical MOSHFET structure.
In the past, different insulating materials were proposed and studied for use as the gate insulator in the IGHFETs. Khan et al. and Binari et al. have reported on GaN-based Metal Insulator Semiconductor Field Effect Transistors (MISFETs) using i-GaN [56], i- AlGaN/GaN [57] and Si3N4 [58] as the gate insulator. These devices were operational but exhibited the current-voltage characteristic collapse at high drain biases related to a large density of interface states. Ren at el. have also reported on a GaN MOSFET using Ga2O3 and Gd2O3 oxides for the gate insulator [59]. All these initial devices suffered from low transconductances and/or high threshold voltages due to large density of the interface states and poor control of the dielectric thickness. In the year 2000, novel AlGaN/GaN Metal-OxideSemiconductor Heterostructure Field Effect Transistors (MOSHFETs) on sapphire [60] and on SiC [61] substrates with excellent gate control were demonstrated. Later on it was also shown that high performance insulated gate HFETs can be made using Si3N4. These devices are referred to as Metal Insulator Semiconductor Heterostructure Field Effect Transistors (MISHFETs). Operation of both device types is based on the high quality of material deposited at the
interface between SiO2 or Si3N4 layer and the AlGaN barrier layer. 0.1
HFET MOSHFET
1E-3
The operating frequency of a FET is determined by the gate length given by
IGS, A
1E-5
fT
1E-7 1E-9 1E-11 1E-13 1E-15 -12
-9
-6
-3
0
3
6
VGS, V Figure 13. Leakage currents comparison for identical geometry HFET and MOSHFET fabricated on the same wafer.
As shown in Figure 12, the MOSHFET design is very close to that of the HFET. The only difference is the presence of insulating layer under the gate that reduces the gate leakage currents by several orders of magnitude. Figure 13 compares the gate currents of identical geometry HFET and MOSHFET fabricated from the same wafer. As seen, both the reverse and forward gate currents of the MOSHFET are much lower than those of the HFET. The suppression of gate currents leads to several key advantages of MOSHFETs for high-power applications. First, it allows for application of positive gate voltages thus further increasing the maximum device currents. For typical HFETs, the peak currents are 1 - 1.2 A/mm. For the MOSHFETs, the currents up to 2 A/mm can be achieved [62]. The most important advantage of the MOSHFET is related to device reliability under high RF power stress. The difference in the device stability is clearly seen from the data of Figure 14. Detailed studies show that the degradation of HFET device types is due to forward gate currents caused by a large-amplitude input signal applied to the gate. 15
RF Power, W/mm
IV.2 HIGH-FREQUENCY PERFORMANCE
12
MOSHFET 9 6
HFET
3 0 0
20
40
60
80
100
Time, Hr Figure 14. RF power stability of identical geometry MOSHFET and HFET at 2 GHz room temperature stress.
Cg
s
2 g m Lg
where f T is the unity gain frequency, Cg is the gate capacitance, gm is the transconductance, Lg is the gate length. Gate length needs to be as short as possible to achieve high operating frequencies. High-frequency performance also depends on electron saturation velocity (s), which is the maximum speed an electron can reach at high electric fields. High saturation velocity is one more advantage of group IIINitride materials, as illustrated in Figure 15.
Figure 15. Velocity - field dependencies for different semiconductor materials.
Recently, AlGaN/GaN Hign Electron Mobility Transistors (HEMTs) with a unity current gain cutoff frequency (fT) of 110 GHz and a maximum frequency of oscillation (f max) of over 140 GHz have been demonstrated by reducing the gate length down to 50 nm [63]. However, this reduction of gate length leads to a decrease in breakdown voltage. To achieve higher performance, indium has been utilized as a surfactant during the growth of AlGaN/GaN HEMTs [61]. A demonstration of the concept used a MOCVD-grown epilayer which consists of a 100 nm AlN buffer, 2 m undoped GaN, a 5 nm undoped Al.25Ga.75N spacer, a 10 nm Si-doped (~5 1018 /cm3) Al.25Ga.75N charge supply layer, and a 10 nm undoped Al.25Ga.75N barrier layer. The layer was grown on 4H-SiC. Hall measurements showed a sheet carrier concentration of 1.1 × 1013 /cm2 and an electron 2 mobility of 1300 cm /V-s at room temperature on as-grown wafers. Device fabricated from this layer demonstrated a typical drain current density of 1.23 A/mm, peak transconductance of 314 mS/mm, and a gate-drain breakdown voltage of over 60 V. Figure 14 shows the shortcircuit current gain (|h21|) and maximum stable gain/maximum available gain (MSG/MAG) derived from on-wafer S-parameter measurements as a function of frequency for the 0.12 µm gate-length device.
(FP) AlGaN/GaN HFETs. Ando et. al. [65] demonstrated microwave powers of 10 W/mm at 2 GHz (gate-length 1 µm, source-drain opening 4.5 µm and Vds = 65 V). Similar fieldplated HFET devices were later reported by Thompson et. al [66] (16.5 W/mm at Vds = 60 V) and Wu et. al. [67] (12.4 W/mm at 48 V and 30 W/mm at Vds = 120 V). The first FPMOSHFET was recently demonstrated at USC [68] with the RF power - drain bias dependencies quite close to those expected from the I-V characteristics. At 55 V drain bias the FP MOSHFET delivered nearly 20 W/mm RF powers at 2 GHz. The device also demonstrated stable operation at 19 W/mm output power level during 100+ hours without any noticeable performance degradation.
Figure 16. Short-circuit current gain (|h21|) and maximum stable maximum available gain (MSG/MAG) of a 0.12 µm × 100 µm AlGaN/GaN HEMT on SiC substrates. After 64
IV.3 HIGH-POWER RF PERFORMANCE Due to high peak currents and high breakdown voltages, group III-Nitride HFETs promise record high output RF powers. For example, a typical AlGaN/GaN HFET with the maximum saturation current about IDSM 1A/mm and the knee voltage VKN 5 V at a moderate operating point of VD = 35 V, should deliver an output power POUT I0(VD - VKN)/2 7.5 W/mm, where I0 IDSM/2 is the operating DC current. However, for a typical device, even under pulsed drain bias and pulsed RF drive conditions eliminating the device self-heating, the measured RF output powers are typically between 2 to 4 W/mm (see Figure 17). The reduction in the RF power is referred to as RF current collapse and is due to intensive electron trapping by the defects in the material at high electric fields.
Figure 17. RF Power - Drain bias dependencies for regular and field-plated III-Nitride HFETs.
The RF current collapse remains the major challenge in the III-Nitride technology for high-power applications in the past ten years. It can be completely eliminated only by achieving extremely high quality materials and by perfect surface passivation. Recently a significant reduction of the current collapse was demonstrated in so-called field-plated
Figure 18. Field-Plated AlGaN/GaN MOSHFET.
V. III-NITRIDE ELECTRONICS
MATERIALS
FOR
HIGH-POWER
Given its unique characteristics, GaN has great potential also for power electronics applications. The most important material characteristics in which GaN has an advantage with respect to Si are wide bandgap (see Figure 3), high electron mobility, large saturation velocity, high sheet carrier concentration at heterojunction interfaces, and high breakdown field (see Figure 5). Besides its electrical characteristics, GaN presents significant advantages from a thermal point of view. First of all, GaN exhibits high thermal conductivity when grown on SiC substrates. Moreover, the coefficient of thermal expansion (CTE) of GaN is very close to the CTE of commonly used heatspreader materials such as AlN (see Figure 19) which should allow the realization of high-reliability thermal packages [69]. All these material advantages hold the promise of realizing high-speed, high-voltage, and high-temperature switching devices. The availability of such devices should allow power electronic specialists to realize power converters operating at higher frequencies, with reduction in the size of passive elements, such as filter inductors and capacitors, and with improvements in terms of achievable bandwidth. The high-voltage devices would simplify high-voltage converter structures, reducing the need for seriesing of devices with associated voltage-sharing problems. The ability of GaN devices to operate at higher temperatures without performance deterioration should make extreme temperature
important result is given by the fact that the switching power density well exceeds 20W/mm.
V.1 ECONOMIC IMPACT OF ONGOING GAN RESEARCH ON POWER ELECTRONICS APPLICATIONS
Si
SiC-4H
GaN
BeO
MMC
Alumina
AlSiC
8 7 6 5 4 3 2 1 0
AlN
CTE (ppm/K)
applications possible and allow heatsink size reduction in more conventional applications.
Figure 19. Thermal coefficient of expansion (CTE) for package substrates and semiconductors.
The use of GaN devices for power electronics is in its infancy. However, in recent times there have been a growing number of publications in this area. A number of publications report results for GaN devices specifically built for power electronics [70-71-72]. A few publications report experimental results related to the performance of GaN devices used in power converter applications. S. Yoshida et al. in [73] presented an application of GaN switches for an inverter application. The HFET operated at 20A with an on-state resistance of 8 mcm2 at 370 V. In terms of performance it is also significant to look at the switching times achieved: 10ns for the turn-on time and 11ns for the turn-off time. The operating frequency of the inverter was 0.37GHz for DC/AC conversion from 30 V to 100V AC. The converter topology consisted of two stages: an input converter with 4 switches operating in parallel and a DC/AC stage in an H-bridge configuration. Saito et al. [74] demonstrated the application of AlGaNGaN HEMT transistor on sapphire substrate. The transistor was designed with 470V breakdown and used in a DC/DC converter topology. The maximum switching frequency achieved in this application has been 200kHz. Matocha et al [75] demonstrated GaN Mosfet with very low on-resistance (1.9 mcm2) and high breakdown voltage (700 V). The USC group presented the first integrated solution based on GaN [76]. In this case an H-bridge topology was used for DC/AC conversion. A photo of the four devices is shown in Figure 20. The circuit has been successfully tested in half bridge operation with a resistive load operating at constant duty-cycle. The most significant limitation experienced in the laboratory testing was due to the driver circuits that limit the switching frequency to well below the GaN capability. For the sake of simplicity, in the first prototype, commercial IGBT drivers have been used in order to easily adjust the control voltage to test the switching behavior of the integrated components. The most
The on-going developmental work in the areas of optoelectronics, sensors, and communications will provide the stimulus needed to carry this technology into Power Electronics for the 21st Century. This economic leveraging should not be underestimated for providing impetus into GaN power device development. This alone may cause GaN to surpass any on-going research efforts in SiC, where the potential and actual economics may limit the investment into the technology [77]. Strategies Unlimited have projected the 2007 GaN market to exceed $5 billion dollars US. The market for GaN is vast and extremely diverse ranging from optoelectronics to microwave electronics to power electronics. Just as GaAs provided the technological tool for the advancement of cellular handsets in the 1990’s, a previously unforeseen application for the technology, GaN is positioned to created electronic markets and products yet to be envisioned.
Figure 20: AlGaN/GaN H-bridge converter. The overall dimensions of the four devices are 1.90 x 1.25mm.
VI. CONCLUSION The paper gives an overview of the state-of-the-art of GaN technology. The significant optoelectronics and RF applications of this material are discussed. The significant investment and material development spurred by these applications are leading to fast improvements in material characteristics. Power electronics applications can leverage on these developments. The first power converter applications of GaN devices are beginning to appear in the recent literature.
ACKNOWLEDGMENT This work was supported by the U.S. Office of Naval Research under Grant N00014-02-1-0623.
REFERENCES
23
1
M. A Khan, J. M. Van Hove J. N. Kuznia and D. T. Olsen, High electron mobility GaN-AlGaN heterostructures grown by LPMOCVD, Appl. Phys. Lett. 58, 2408 (1991) and also: M.A. Khan, J. N. Kuznia, J. M. Van Hove N. Pan and N. Carter, Observation of a two-dimensional electron gas in low pressure MOCVD deposited GaN/AlGaN heterojunctions, Appl. Phys. Lett. 60, 3027 (1992) 2 S. Rumyantsev, M. Shur and M. Levinshtein, Material Properties of Nitrides. Summary, Int. Journ. of High Speed Electronics and systems, V. 14, N1, 1-19 (2004) 3 “HB-LED market reaches $1.8 billion,” Compound Semicond., vol. 9, no. 2, pp. 17, 2003. 4 “High-brightness LEDs 2003 – Market review and forecast,” Strategies Unlimited, June 2003. 5 O. Ambacher, J. Smart, J. R. Shealy, N. G. Weimann, K. Chu, M. Murphy, W. J. Schaff, L. F. Eastman, R. Dimitrov, L. Wittmer, M. Stutzmann, W. Rieger and J. Hilsenbeck, Twodimensional electron gases induced by spontaneous and piezoelectric polarization charges in N- and Ga-face AlGaN/GaN heterostructures, Journ. Appl. Phys. 85, 32223233 (1999) 6 M. Asif Khan, J. W. Yang, G. Simin, R. Gaska, M. S. Shur, Hans-Conrad zur Loye, G. Tamulaitis, A. Zukauskas, David J. Smith, D. Chandrasekhar, and R. Bicknell-Tassius, Lattice and Energy Band Engineering in AlInGaN/GaN Heterostructures , Appl. Phys. Lett 76 (9) pp. 1161-1163 (2000) 7 J. W. Lee and I. Adesida, “Gallium Nitride for Electronics” in: Encyclopedia of RF and Microwave Engineering, John Wiley & Sons, Inc., 2005 ISBN: 0-471-27053-9 8 . T. Nishida, H. Saito, N. Kobayashi, Appl. Phys. Lett. 78, 39273928 (2001). 9 . T. Nishida, H. Saito, N. Kobayashi, Appl. Phys. Lett. 79, 711712 (2001). 10 W. Sun, V. Adivarahan, M. Shatalov, Y. Lee, S. Wu, J. Yang, J. Zhang, and M. A. Khan: Jpn. J. Appl. Phys. 43 (2004) L 1419. 11 J. P. Zhang, X. Hu, Y. Bilenko, J. Deng, A. Lunev, R. Gaska, M. Shatalov, J. Yang, and M. A. Khan, Appl. Phys. Lett. 85 5532 (2004). 12 . J. P. Zhang, H. M. Wang, M. E. Gaevski, C. Q. Chen, Q. Fareed, J. W. Yang, G. Simin, M. Asif Khan, Appl. Phys. Lett. 80, 3542-3544 (2002). 13 . H. M. Wang, J. P. Zhang, C. Q. Chen, Q. Fareed, J. W. Yang, M. Asif Khan, Appl. Phys. Lett. 81, 604-606 (2002). 14 . J. P. Zhang, M. Asif Khan, W. H. Sun, H. M. Wang, C. Q. Chen, Q. Fareed, E. Kuokstis, J. W. Yang, Appl. Phys. Lett. 81, 4392-4394 (2002). 15 W. Sun, V. Adivarahan, M. Shatalov, Y. Lee, S. Wu, J. Yang, J. Zhang, and M. A. Khan: Jpn. J. Appl. Phys. 43 (2004) L 1419. 16 J. P. Zhang, X. Hu, Y. Bilenko, J. Deng, A. Lunev, R. Gaska, M. Shatalov, J. Yang, and M. A. Khan, Appl. Phys. Lett. 85 5532 (2004). 17 M. A. Khan, J. N. Kuznia, D. T. Olson, J. M. Van Hove, M. Blasingame, and L. F. Reitz: Appl. Phys. Lett. 60 (1992) 2917. 18 M. P. Ulmer, M. Razeghi, and E. Bigan: Proc. SPIE 2397 (1995) 210. 19 P. Schreiber, T. Dang, G. Smith, T. Pickenpaugh, P. Gehred, and C.Litton, Proc. SPIE, 3629 (1999) 230. 20 G. A. Shaw and M. Nischan: Proc. SPIE 4396 (2001) 31. 21 M. S. Shur and M. A. Khan: MRS Bull. 22 (1997) 44. 22 M. A. Khan, J. Kuznia, D. T. Olson, M. Blasingame, and A. R. Bhattarai: Appl. Phys. Lett. 63 (1993) 2455.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Q. Chen, J. W. Yang, A. Osinsky, S. Gangopadhyay, B. W. Lim, M. Z. Anwar, M. A. Khan, D. V. Kuksenkov, and H. Temkin: Appl. Phys. Lett. 70 (1997) 2277. A. Osinsky, S. Gangopadhyay, B. W. Lim, M. Z. Anwar, M. A. Khan, D. V. Kuksenkov, and H. Temkin: Appl. Phys. Lett. 72 (1998) 742. B. W. Lim, S. Gangopadhyay, J. W. Yang, A. Osinsky, Q. Chen, M. Z. Anwar, and M. A. Khan: Electron. Lett. 33 (1997) 633. A. Teke, S. Dogan, F. Yun, M. A. Reshchikov, H. Le, X. Q. Liu, H. Morkoc, S. S. Zhang, W. B. Wang, and R. R. Alfano: Solid-State Electron. 47 (2003) 1401. M. A. Khan, Q. Chen, C. J. Sun, M. S. Shur, M. F. Macmillan, R. P. Devaty, and J. Choyke: Proc. SPIE 2397 (1995) 283. X. Zhang, P. Kung, D. Walker, J. Piotrowski, A. Rogalski, A. Saxler, and M. Razeghi: Appl. Phys. Lett. 67 (1995) 2028. Q. Chen, M. A. Khan, C. G. Sun, and J. W. Yang,: Electron. Lett. 31 (1995) 1781. A. Osinsky, S. Gangopadhyay, R. Gaska, B. Williams, M. A. Khan, D. Kuksenkov, and H. Temkin: Appl. Phys. Lett. 71 (1997) 2334. G. Y. Xu, A. Salvador, W. Kim, Z. Fan, C. Lu, H. Tang, H. Morkoc¸ G. Smith, M. Estes, B. Goldenberg, W. Yang, and S. Krishnankutty: Appl. Phys. Lett. 71 (1997) 13. D. Walker, A. Saxler, P. Kung, X. Zhang, M. Hamilton, J. Diaz, and M. Razeghi: Appl. Phys. Lett. 72 (1998) 3303. J. C. Carrano, T. Li, P. A. Grudowski, C. J. Eiting, D. Lambert, J. D. Schaub, R. D. Dupuis, J. C. Campbell: Electron. Lett. 34 (1998) 692. G. B. Parish, S. Keller, P. Kozodoy, J. P. Ibbetson, H. Marchand, P. T. Fini, S. B. Fleischer, S. P. DenBaars, U. K. Mishra, and E. J. Tarsa: Appl. Phys. Lett. 75, (1999) 247. D. Walker, V. Kumar, K. Mi, P. Sandvik, P. Kung, X. H. Zhang, and M. Razeghi: Appl. Phys. Lett. 76 (2000) 403. S. C. Jain, M. Willander, J. Narayan, and R. Van Overstraeten: J. Appl. Phys. 87, (2000) 965. E. J. Tarsa, P. Kozodoy, J. Ibbetson, B. P. Keller, G. Parish, and U. Mishra: Appl. Phys. Lett. 77 (2000) 316. P. Sandvik, K. Mi, F. Shahedipour, R. McClintock, A. Yasan, P. Kung, and M. Razeghi: J. Cryst. Growth 231 (2001) 366. J. C. Campbell, C. J. Collins, M. M. Wong, U. Chowdhury, A. L. Beck, and R. D. Dupuis: Phys. Stat. Solidi (a) 188 (2001) 283. K. A. McIntosh, R. J. Molnar, L. J. Mahoney, A. Lightfoot, M. W. Geis, K. M. Molvar, I. Melngailis, R. L. Aggarwal, W. D. Goodhue, S. S. Choi, D. L. Spears, and S. Verghese: Appl. Phys. Lett. 75 (1999) 3485. J. C. Carrano, D. J. H. Lambert, C. J. Eiting, C.J. Collins, T. Li, S. Wang, B. Yang, A. L. Beck, R. D. Dupuis, and J. C. Campbell: Appl. Phys. Lett. 76 (2000) 924. K. A. McIntosh, R. J. Molnar, L. J. Mahoney, K. M. Molvar, N. Efremow, Jr., and S. Verghese: Appl. Phys. Lett. 76 (2000) 3938. E. Ozbay, N. Biyikli, I. Kimukin, T. Kartaloglu, T. Tut, and O. Aytür: IEEE J. of Sel. Topics in Quant. Electron. 10 (2004) 742. B. Yang, B. Yang, K. Heng, T. Li, C. J. Collins, S. Wang, R. D. Dupuis, J. C. Campbell, M. J. Schurman, and I. T. Ferguson: IEEE J. Quant. Electron. 36 (2000) 1229. P. Lamarre, A. Hairston, S. Tobin, K. K. Wong, M. F. Taylor, A. K. Sood, M. B. Reine, M. J. Schurman, I. T. Fergusson, R. Singh, and C. R. Eddy Jr.: Proc. Mat. Soc. Symp. 639 (2001) G10.0.1. J. D. Brown, J. Boney, J. Matthews, P. Srinivasan, J. F. Schetzina, T. Nohava, W. Yang, S. Krishnankutty: MRS Internet J. Nitride Semicond. Res. 5 (2000) 6.
47
A. Hirano, C. Pernot, M. Iwaya , T. Detchprohm, H. Amano, and I. Akasaki: Phys. Stat. Sol. (a) 188, (2001) 293. 48 D. Ciplys, R. Gaska, M. S. Shur, R. Rimeika, J. Yang, and M. Asif Khan: Proc. SPIE 4453 (2001) 37. 49 D. Ciplys, R. Rimeika, M. S. Shur, S. Rumyantsev, R. Gaska, A. Sereika A., J. Yang and M. A. Khan: Appl. Phys. Lett. 80 (2002) 2020. 50 D. Ciplys, M. S. Shur, N. Pala, A. Sereika, R. Rimeika, R. Gaska, and Q. Fareed, Proc. IEEE Sensor Conference, Vienna (2004) (to be published). 51 M. Asif Khan, J. W. Yang, G. Simin, R. Gaska, M. S. Shur, A. Bykhovsky, Piezoelectric Doping in AlInGaN/GaN Heterostructures , Appl. Phys. Lett., 75 (18) , pp. 2806-2808, 1999 52 T. Sheppard, W. L. Pribble, D.T.Emerson, Z. Ring, R. P. Smith, S. T. Allen, and J. W. Palmour, Device Research Conference Digest (2000) p. 37 53 N.X. Nguyen, M. Micovic, W.-S. Wong, P. Hashimoto, L.-M. McCray, P. Janke and C. Nguyen, Electron. Lett. 36 (2000) p. 468 54 Yi-Feng Wu; Kapolnek, D.; Ibbetson, J.P.; Parikh, P.; Keller, B.P.; Mishra, U.K., Very-high power density AlGaN/GaN HEMTs, IEEE Transactions on Electron Devices, V. 48, 586590 (2001) 55 Wu Lu; Jinwei Yang; Khan, M.A.; Adesida, I. AlGaN/GaN HEMTs on SiC with over 100 GHz fT and low microwave noise, IEEE Transactions on Electron Devices, V. 48, 581 -585 (2001) 56 M. Asif Khan, J. N. Kuznia, A. R. Bhattarai, and D. P. Olson, GaN/AlGaN Heterostructure Deposition by Low Pressure MOCVD for MISFET Devices, Material Research Society Proceedings, vol. 281, p.769 (1993) 57 M. Asif Khan, Michael S. Shur, Q. C. Chen, and J. N. Kuznia, Current-Voltage Characteristic Collapse in AlGaN/GaN Heterostructure Insulated Gate Field Effect Transistors at High Drain Bias, Electronics Letters, Vol. 30, No. 25, p. 2175, Dec. 8, 1994 58 . S. C. Binari, L. B. Rowland, G. Kelner, W. Kruppa, H. B. Dietrich, K. Doverspike, D. K. Gaskill, "DC, Microwave, and High-Temperature Characteristics of GaN FET Structures," in Int. Sym. Compound Semiconductors, ed. H. Goronkin (IOP Publishing, Bristol, 1995), p.459 59 F. Ren, M. Hong, S. N. G. Chu, and M .A. Marcus, M.J. Schurman, A. Baca, S. J. Pearton and C.R. Abernathy, Effect of Temperature on Ga2O3 (Gd2O3)/GaN Metal Oxide Semiconductor Field Effect Transistor, Appl. Phys. Lett. 73, 3893 (1998) 60 M. Asif Khan, X. Hu, G. Simin, A. Lunev, and J. Yang, R. Gaska and M. S. Shur, "AlGaN/GaN Metal-OxideSemiconductor Heterostructure Field Effect Transistor", IEEE Electron Device Letters, vol. 21, pp.63-65 (2000) 61 M. Asif Khan, X. Hu, G. Simin, and J. Yang, R. Gaska and M.S. Shur, AlGaN/GaN Metal-Oxide-Semiconductor Heterostructure Field Effect Transistors on SiC Substrates, Appl. Phys. Lett. v. 77 pp 1339-1341 (2000) 62 A. Koudymov, H. Fatima, G. Simin, J. Yang, M. Asif Khan, A. Tarakji, X. Hu, M. S. Shur, & R. Gaska Maximum Current in Nitride-Based Heterostructure Field Effect Transistors Appl. Phys. Lett, V. 80 pp. 3216-3218 (2002) 63 M. Micovic, N. X. Nguyen, P. Janke, W. S. Wong, P. Hashimoto, L. M. McCray, and C. Nguyen, “GaN/AlGaN high electron mobility transistors with fT of 110 GHz,” Electron. Lett., vol. 36, pp. 358–359, Feb. 2000. 64 V. Kumar, W. Lu, R. Schwindt, A. Kuliev, G. Simin, J. Yang, M. A. Khan, and I. Adesida, “AlGaN/GaN HEMTs on SiC
with fT of over 120 GHz,” IEEE Electron Device Lett., vol. 23, No. 8, pp. 455–457, Aug. 2002. 65 . Y. Ando, Y. Okamoto, H. Miyamoto, T. Nakayama, T. Inoue, and M. Kuzuhara, “10-W/mm AlGaN–GaN HFET With a Field Modulating Plate”, IEEE EDL, V. 24, 289-291 (2003) 66 R. Thompson, T. Prunty, V. Kaper, and J. R. Shealy, “Performance of the AlGaN HEMT Structure With a Gate Extension”, IEEE Trans. on ED, V. 51 292 - 295 (2004) 67 . Y.-F. Wu, A. Saxler, M. Moore, R. P. Smith, S. Sheppard, P. M. Chavarkar, T. Wisleder, U. K. Mishra, and P. Parikh, ”30W/mm GaN HEMTs by Field Plate Optimization”, IEEE EDL, V. 25, 117-119 (2004) 68 A. Koudymov et. al. Current collapse removal in field-plated AlGaN/GaN MOSHFETs. Submitted to IEEE EDL, April 2005 69 J.L. Hudgins, G. Simin, E. Santi and M.A. Khan, "An assessment of wide bandgap semiconductors for power devices," IEEE Trans. Power Electronics, Vol. 18, No. 3, pp. 907-914, May 2003 70 C. Park, P.L. Chapman, S.H. Rhee, S.J. X.HongZhang, P.T. Krein, K. Kim, GaN power switching device growth by plasma assisted molecular beam epitaxy, Conference Record of the 2002 Industry Applications Conference, 37th IAS Annual Meeting, Volume 1, 13-18 Oct. 2002, pp. 576 - 579 vol.1 71 S. Karmalkar, Jianyu Deng, M.S. Shur, IEEE Electron Device Letters, RESURF AlGaN/GaN HEMT for high voltage power switching, Volume 22, Issue 8, Aug. 2001 pp. 373 - 375 72 Naiqian Zhang, V. Mehrotra, S. Chandrasekaran, B. Moran, Likun Shen; U. Mishra, E. Etzkorn, D. Clarke, Large area GaN HEMT power devices for power electronic applications: switching and temperature characteristics, 2003 IEEE 34th Annual Power Electronics Specialist Conference, PESC '03 Volume 1, 15-19 June 2003, pp. 233 - 237 vol.1 73 S. Yoshida, J. Li, T. Wada and H. Takehara. “High-power AlGaN/GaN HFET with lower on-state resistance and higher switching time for an inverter circuit, on IEE Proc. On Circuits Devices and Systems, vol. 151, N. 3, June 2004, pp. 207-210 74 W. Saito, M. Kuraguchi, Y. Takada, K. Tsuda, Ichiro Omura,and T. Ogura, High Breakdown Voltage Undoped AlGaN–GaN Power HEMT on Sapphire Substrate and Its Demonstration for DC–DC Converter Application IEEE Transactions On Electron Devices, VOL. 51, NO. 11, Nov. 2004, pp. 1913-1917 75 K. Matocha, T. P. Chow and R. J. Gutmann, High-Voltage Normally Off GaN MOSFETs on Sapphire Substrates, IEEE Transactions On Electron Devices, VOL. 52, NO. 1, Jan. 2005, pp. 6-9 76 S.G. Pytel, S. Lentijo, A. Koudymov, S. Rai, H. Fatima, V. Adivarahan, A. Chitnis, J. Yang, J.L. Hudgins, E. Santi, A. Monti, G. Simin, and M. Asif Khan, “AlgaN/GaN MOSHFET Integrated Circuit Power Converter”, on Proc. of IEEE PESC 2004, Aachen (Germany), June 2004. 77 J.L. Hudgins, Gallium nitride – the other wide bandgap semiconductor, IEEE Power Electronics Society Newsletter, 2005
