New Applications for GaN-Based Semiconductors
S.J. Peartona, C.R. Abernathya, M.E. Overberga, G.T. Thalera, A.H. Onstinea, B.P. Gilaa, F. Renb, B. Loub and J. Kimb
a Department of Materials Science and Engineering
University of Florida, Gainesville, FL 32611 USA
b Department of Chemical Engineering
University of Florida, Gainesville, FL 32611 USA
ABSTRACT
GaN-based visible light-emitting diodes and laser diodes are already commercialized for a variety of lighting and data storage applications. This materials system is also showing promise for microwave and high power electronics intended for radar, satellite, wireless base stations and utility grid applications, for biological detection systems and for a new class of spin-transport electronics (spintronics) in which the spin of charge carriers is exploited.
The explosive increase in interest in the AlGaInN family of materials is recent years has been fueled by the application of blue/green/UV light-emitting diodes (LEDs) in full-color displays, traffic lights, automotive lighting and general room lighting using so-called white LEDs.(1) In addition, blue/green laser diodes can be used in high storage-capacity digital versatile disks (DVDs) systems.(2) AlGaN-based photodetectors are also useful for solar-blind UV detection and have applications as flame sensors for control of gas turbines or for detection of missiles. There are currently major development programs in the US for three newer applications for GaN-based materials and devices, namely
(i) UV optical sources capable of operation down to 280 nm for use in airborne chemical and biological sensing systems, allowing direct multi-wavelength spectroscopic identification and monitoring of UV-induced reactions.
(ii) Power amplifiers and monolithic microwave integrated circuits (MMICs) for use in high performance radar units and wireless broadband communication links and ultra high power (>1 MW) switches for control of distribution on electricity grid networks.
(iii)
Room temperature, ferromagnetic semiconductors for use in
electrically-controlled magnetic sensors and actuators, high density, ultra-low
power memory and logic, spin-polarized light emitters for optical encoding,
advanced optical switches and modulators and devices with integrated magnetic,
electronic and optical functionality.
Properties of the GaN-Based Semiconductors
The three binary compounds in the AlGaInN system, namely GaN, InN and AlN, span the range of bandgaps from 1.9 to 6.2 eV. In addition, they have significantly smaller lattice constants than Si, the more established III-V compound semiconductors (such as GaAs, InP and GaP) and the II-VI semiconductors with wide bandgaps (Figure 1). The ternaries InGaN and AlGaN have been produced over the entire composition range between their component binaries, while InAlN is less well-explored. Note that GaN and AlN are fairly well lattice-matched to SiC substrates, which have the concomitant advantages of dopability and high thermal conductivity relative to the more commonly used Al2O3 substrates.
Table I compares the physical properties of GaN and AlN with other well-known semiconductors. The nitrides are well-suited to high temperature applications because of their wide bandgaps and low intrinsic carrier concentrations. The electron mobility in GaN is quite high considering the magnitude of the bandgap and is even higher in selectively doped AlGaN/GaN heterostructures where 2-dimensional electron gases (2DEG) may form, producing high sheet carrier densities.(3) These densities are enhanced by the strong piezoelectric and polarization effects present in the AlGaN/GaN structures. These properties make the nitrides well-suited for high-frequency applications. Finally, the GaN possesses a very high breakdown field, allowing the devices to support large voltages for high power operation.
Note that SiC is also well-suited for these applications and has been developed for electronics for a much longer period than GaN. However, as shown in Figure 2, GaN still has greater potential for these applications because of its larger bandgap and higher carrier velocity and mobility. The top of Figure 2 shows theoretical and experimental data for the critical breakdown field for various semiconductors as a function of their bandgaps. While diamond supports the largest field strength, the difficulty of synthesizing large electronic grade crystals and of subsequently doping and contacting to form devices has limited progress for this material.(4)
A combined figure-of-merit (CFOM)
for high frequency, high power/high temperature applications can be defined as(5)
where 
is the thermal conductivity, 
the dielectric
constant, 
the electron mobility
and 
the breakdown field.
The bottom of Figure 2 shows the resulting values for GaN, 4H-SiC, GaAs
and Si and illustrates the potential of GaN for these applications.
GaN-Based UV Optical Sources
The current generation of visible GaN-based LEDs and laser diodes generally employs InGaN quantum well active regions, with AlGaN cladding layers. Replacing the InGaN with GaN quantum wells typically produces operating wavelengths in the 350-360 nm region.(6) The use of AlGaN quantum wells is capable of shifting the LED emission to below 340 nm.(7) However these devices typically show large decreases in optical output intensity with increasing operating temperature. In some cases the use of quaternary AlGaInN can produce improved lattice matching to the clad layer and better UV emission efficiency.(12)
There are many difficulties inherent in pushing to even shorter wavelengths. Hiragana et al.(13) reported photoluminescence measured at 77 K in the range 230-280 nm from AlxGa1-xN(AlyGa1-yN multiquantum wells grown on SiC substrates. However the intensity at room temperature from the AlGaN quantum wells was much lower than from InGaN-based structures. There is much work to be done on improving the material quality of AlGaN-based structures, such as the reduction of native point defects and impurities, control of strain and stress and increasing the conductivity of p-AlGaN which is essential for achieving acceptable p-ohmic contact resistance. In actual LED structures, there is a need to optimize current injection cavity design and light extraction through the contact layers. The large band-offsets in this system are also a challenge in terms of designing layer structures that minimize the turn-on voltage and enhance carrier transport.
The UV LEDs or laser diodes can be employed in systems for real-time measurement of fluorescence spectra from airborne biological and chemical particles. This data can then be employed as part of a compact biological or chemical agent warning sensor with fast response and high detection sensitivity. The UV wavelengths are necessary in order to cause fluorescence in many of the targeted chemicals and biological agents.
GaN-Based Power Electronics
There is increasing interest in the replacement of mechanical relays in power flow control circuits used in the electricity grid and in other applications such as electric automobiles and hybrid electric military vehicles. Presently, design and limitation of Si material limits use of power electronics in utility transmission and distribution systems.(14) Devices based on SiC and especially GaN offer a factor of 10 higher power handling capability.(15)
Most of the Si devices of the last decade are based on gate turn-off thyristors, with the emerging devices being insulated gate bipolar transistors (IGBTs) and mos-controlled thyristors (MCTs). There is also the beginnings of serious interest in the utility industry for static compensators (to provide voltage support on lines by generating or absorbing reactive power without need for large external reactors or capacitor banks), verified power controllers (comprehensively controls power flow) and dynamic voltage restorers (protects sensitive load from line disturbances). These would all be components of a flexible ac transmission system.
Inclusion of power electronics in such a system enables achievement of soft-switching to eliminate harmonics and dramatically improve power controllers and converters for ship propulsion and avionics.
Currently the realization of such systems is limited by the inadequacy of Si-based semiconductor-controlled rectifiers (SCRS) and gate turn-off thyristors (GTOs) in terms of
§ Maximum voltage ratings < 7 kV – multiple devices must be placed in series for high voltage systems.
§ Insufficient current-carrying capability – multiple devices must be placed in parallel for typical grid applications.
§ Conductivity in one direction only – identical pairs of devices and assemblies must be installed in antiparallel for switchable circuits.
§ Inadequate thermal management – heat damage is a primary cause of failure and expense.
Power rectifiers are key components of inverter modules, which are used in power flow control circuits. Recent progress in AlGaN-based Schottky and p-i-n rectifiers have shown reverse blocking voltages up to 9.7 kV on lateral structures.(16) The figure-of-merit (VB/2/RON, where VB is the reverse breakdown voltage and RON is the on-state resistance, was as high as 268 MW·cm-2 for these devices. While these results are impressive, the current density in lateral devices is limited and both the Schottky and p-i-n rectifiers displayed breakdown voltages that decreased as a function of temperature T, with the relation
where 
had magnitude
–(0.3-10) V·K-1.(17)
The negative temperature coefficient appears to be related to the high
defect density in the heteroepitaxial GaN on sapphire.
These drawbacks point to the need for vertical geometry rectifiers, preferably fabricated on low defect density GaN substrates.(18) Figure 3 (top) shows a 200 mm thick, free-standing GaN substrate which was grown by vapor phase epitaxy on a c-plane Al2O3 substrate and then removed by differential heating from a laser beam.(19,20) The specific on-state resistance was in the range 1.7-3.0 MW·cm2 for Schottky rectifiers, with a turn-on voltage of ~1.8 V. The switching performance was analyzed using the reverse recovery current transient waveform, leading to an estimate of the high injection level hole lifetime of ~15 ns. Large-diameter (~7 mm) packaged Schottky rectifiers showed measurement system-limited forward currents as high as 1.65 A, as shown at the bottom of Figure 3.(21) Continued improvements in the production of GaN free-standing substrates will lead to even higher currents being possible.
For microwave power amplification in satellite links and wireless communication applications, the AlGaN/GaN high electron mobility transistor (HEMT) has emerged as the most promising device.(22-29) A typical Al0.3Ga0.7N/GaN HEMT shows a sheet carrier density up to ~1.5x1013 cm-2 (roughly an order of magnitude higher than for GaAs HEMTs), a unity current gain frequency (fT) of 65 GHz and a maximum oscillation frequency (fMAX) of 180 GHz. Output powers up to 9.8 W/mm at 8 GHz (power-added efficiency at 47%) have been reported for large periphery GaN HEMTs(29), while it has also been shown that these devices can achieve low microwave added noise figures (NF = 0.6 dB at 10 GHz), while maintaining a large breakdown voltage (>60 V) and hence a large dynamic range.(30)
Since one of the potential applications for HEMTs is in broad-band satellite transmission for communications, television and weather forecasting systems, it is necessary that they be radiation-resistant. Figure 4 shows the effect of 40 MeV proton irradiation at closer corresponding to 10 or 100 years in low-earth orbit on the dc characteristics of AlGaN/GaN HEMTs.(31) Even for the highest proton dose, the transconductance only decreases by 30%. Similarly the rf performance was only degraded by 30% for these closer. The main degradation mechanism was removal of carriers from the device channel by radiation-induced traps. The devices are also very resistant to g-ray damage.(32)
One frequently reported problem in these devices is that the rf power obtained is still much lower than expected from the dc characteristics.(33-35) This problem is manifested by a collapse in drain current or by frequency dispersions in the transconductance and output resistance, leading to severely reduced output power and power-added efficiency. Several mechanisms have been identified as the causes, including the presence of surface states between the gate and drain which deplete the channel in this region over a time constant long enough to disrupt modulation of the channel charge during large signal operation or of trap states in the buffer layer (Figure 5, top). Several studies have shown that the use of SiNx passivation layers can be effective in reducing the effects of surface states.(34-36) One drawback of typical plasma enhanced chemical vapor deposited SiNx is the high hydrogen content, which could migrate into the GaN or the gate metallization. Two alternative candidates for HEMT passivation are MgO and Sc2O3, which are under development as gate dielectrics for GaN.(37) These materials have larger band gaps (8 eV for MgO and 6.3 eV for Sc2O3) than the previously reported Gd2O3 (5.3 eV) and smaller lattice mismatches to GaN (6.5% for MgO and 9.2% for Sc2O3 vs 20% for Gd2O3). We find that layers of these materials deposited by molecular-beam epitaxy (MBE) on AlGaN/GaN HEMTs prevent much of the lag response found in unpassivated devices.(38)
As an example, Figure 5 (bottom) shows HEMT power measurement data before and after deposition of a thin (100 Å) Sc2O3 passivation film. The increase in power output was significantly larger than for SiNx passivation on the same devices.
GaN for Spintronic Applications
There is currently a lot of interest in the science and potential technological applications of spin-transport electronics (or spintronics), in which the spin of charge carriers (electrons or holes) is exploited to provide new functionality for microelectronic devices.(39-41) The phenomena of giant magnetoresistance and tunnelling magnetoresistance have been exploited in all-metal or metal-insulator-metal magnetic systems for read/write heads in computer hard drives, magnetic sensors and magnetic random access memories.(42) The development of magnetic semiconductors with practical ordering temperatures could lead to new classes of device and circuits, including spin transistors, ultra-dense non-volatile semiconductor memory and optical emitters with polarized output.
Most of the past attention on ferromagnetic semiconductors focussed on the (Ga,Mn)As (40) and (In,Mn)As (43) systems. In samples carefully grown single-phase by Molecular Beam Epitaxy (MBE), the highest Curie temperatures reported are ~110 K for (Ga,Mn)As and ~ 35 K for (In,Mn)As.(40) A tremendous amount of research on these materials systems has led to some surprising results, such as the very long spin lifetimes and coherence times in GaAs(44) and the ability to achieve spin transfer through a heterointerface(45), either of semiconductor/semiconductor or metal-semiconductor. One of the most effective methods for investigating spin-polarized transport is by monitoring the polarized electroluminescence output from a quantum-well light-emitting diode into which the spin current is injected. Quantum selection rules relating the initial carrier spin polarization and the subsequent polarized optical output can provide a quantitative measure of the injection efficiency.(46)
There are a number of essential requirements for achieving practical spintronic devices in addition to the efficient electrical injection of spin-polarized carriers. These include the ability to transport the carriers with high transmission efficiency within the host semiconductor or conducting oxide, the ability to detect or collect the spin-polarized carriers and to be able to control the transport through external means such as biasing of a gate contact on a transistor structure. Nitta et al.(47) demonstrated that a spin-orbit interaction in a semiconductor quantum well could be controlled by applying a gate voltage. These key aspects of spin injection, spin-dependent transport, manipulation and detection form the basis of current research and future technology. The use of read sensors based on metallic spin valves in disk drives for magnetic recording is already a $US100 B per year industry. It should also be pointed out that spintronic effects are inherently tied to nanotechnology, because of the short (~1 nm) characteristic length of some of the magnetic interactions. Combined with the expected low power capability of spintronic devices, this should lead to extremely high packing densities for memory elements. While the progress in synthesizing and controlling the magnetic properties of III-arsenide semiconductors has been astounding, the reported Curie temperatures are too low to have significant practical impact. A key development that focused attention on wide bandgap semiconductors as being the most promising for achieving high Curie temperatures was the work of Dietl et al. (48) They employed the original Zener model of ferromagnetism(49) to predict TC values exceeding room temperature for materials such as GaN and ZnO containing 5% of Mn and a high hole concentration (3.5x1020 cm-3). In the subsequent period after appearance of the Dietl et al (48) paper, remarkable progress has been made on the realization of materials with TC values at or above room temperature.
There are a number of existing models for the observed magnetism in semiconductors and conducting oxides. The Dietl et al.(48) near-field model considers the ferromagnetism to be mediated by delocalized or weakly localized holes in the p-type materials.(50) The magnetic Mn ion provides a localized spin and acts as an acceptor in most III-V semiconductors so that it can also provide holes. This treatment assumes that the Mn-doped III-V materials are charge transfer insulators and does not apply when d-shell electrons participate in charge transport. The spin-spin coupling is assumed to be a long-range interaction, allowing use of a mean-field approximation. The Curie temperature for a given material, Mn concentration and hole density is then determined by a competition between the ferromagnetic and antiferromagnetic interactions. The model takes into account the anisotropy of the carrier-mediated exchange interaction related with the spin-orbit coupling in the host material. The TC is proportional to the density of Mn ions and the hole density.
It is certainly fair to say that the origin of ferromagnetism in wide bandgap semiconductors is still not totally understood. Many aspects of the experimental data can be explained by the mean-field model (which is based on the Ruderman-Kittel-Kasuya-Yosida (RKKY) interaction). However, ferromagnetism has been observed in samples that have very low hole concentrations, in insulting material and more recently in n-type material. Recent models have taken account of these observations of ferromagnetism in non-degenerate samples, and included consideration of indirect exchange interactions caused by virtual electron excitations from magnetic impurity acceptors to the valence band(51) or the effects of positional disorder(52) which lead to unusual spin and charge transport properties and the shape of the magnetization curve. In some cases, positional disorder of the magnetic impurities is found to enhance the ferromagnetic transition temperature.(53)
An alternative approach using local density functional calculations indicated that the magnetic impurities may form small nano-size clusters (just a few atoms in dimension) that produce the observed ferromagnetism.(154) This would be difficult to detect by most characterization techniques.
Additional studies have predicted which magnetic dopants should be most effective in GaN (eg. V, Cr, Mn, Fe) without additional doping to produce carriers(55), and chemical trends were identified. Numerous recent reports have found room temperature ferromagnetism in (Ga,Mn)N.(56-58)
Note that not all of the materials are degenerately-doped p-type. In the mean-field theories it is difficult to achieve ferromagnetism in n-type semiconductors due to their generally smaller s-d interaction. In these types of theory, the sp-d interactions are regarded as the effective magnetic field acting on the carriers, so that when spontaneous magnetization and holes are present, the resultant spin-splitting in the valence band lowers the system energy.
The initial work on this material involved either microcrystals synthesized by nitridization of pure metallic Ga in supercritical ammonia or bulk crystals grown in reactions of Ga/Mn alloys on GaN/Mn mixtures with ammonia at ~1200°C.(59,60) These samples exhibit ferromagnetic properties over a broad range of Mn concentrations, as did some of its early MBE-grown films.(61)
More recent reports on epi growth of (Ga,Mn)N have detailed a range of growth conditions producing single-phase material and the resulting magnetic properties.(62) In general, no second phases are found for Mn levels below ~10% for growth temperatures of ~750°C. The (Ga,Mn)N retains n-type conductivity under these conditions and is single-phase as measured by e-ray diffraction. When the Mn concentration is increased significantly, peaks from tetragonal Mn0.6Ga0.4 become visible.
In accordance with most of the theoretical predictions, magnetotransport data showed the anomalous Hall effect, negative magnetoresistance and magnetic resistance at temperatures that were dependent on the Mn concentration. For example, in films with very low (<1%) or very high (~9%) Mn concentrations, the Curie temperatures were between 10-25 K. The sheet resistance showed negative magnetoresistance below 150 K, with the anomalous Hall coefficient disappearing below 25 K. When the Mn concentration was decreased to 3 at.%, the (Ga,Mn)N showed the highest degree of ordering per Mn atom.
The local structure and effective chemical valency of Mn in MBE-grown (Ga,Mn)N samples has been investigated by Extended X-Ray Absorption Fine Structure.(63) It was concluded that most of the Mn was incorporated substitutionally on the Ga sub-lattice with effective valency close to +2 for samples with ~2 at.% Mn.(63) There was also evidence that a fraction (from 1-36%,depending on growth condition) of the total Mn concentration could be present as small Mn clusters.
(Ga,Fe)N films grown by MBE have also been reported.(64) EXAFS data showed that the Fe was substitutional in Ga sites, with ferromagnetic ordering present below ~100 K for samples grown at ~380°C.(64) Direct evidence for implanted Fe on substitutional Ga sites in GaN has come from emission channelling measurements.(165) It is possible that with further optimization, GaN films doped with Fe, Ni and other impurities might also show room temperature ferromagnetism.
The expected advantages of spin devices include non-volatility, higher integration densities, lower power operation and higher switching speeds. The ability to control the ferromagnetism through manipulation of the carrier density in gated semiconductor structures has great promise for integrated logic/memory/sensor chips.
Additional work is needed on the synthesis and control of carrier density in wide bandgap semiconductor and on how the magnetic properties are related to the carrier type and density. The theoretical understanding of the origin of the ferromagnetism is progressing rapidly, but much work is needed to characterize the local environment around the Mn ions and the electrical properties of the Mn. This will require use of element-specific techniques such as transmission electron microscopy with Z-contrast, EXAFS and various tunnelling microcopies to correlate lattice position and the presence of any nanoclusters with the resultant magnetic properties. Finally, it is imperative to fabricate a wide variety of spin-related device structures in order to identify the factors that limit their performance, which in turn will inevitably lead to improved contacts, cleaner interfaces and a clearer picture of spin-injection, control and transport issues.
Figure 6 (top) shows a spin-LED structure, in which the top n-contact layer is ferromagnetic GaMnN. When bias is applied to the device (bottom), it operates as an efficient blue LED. We are currently determining the temperature dependence of the spin injection efficiency in these structures.
Acknowledgments
The authors gratefully acknowledge the HEMT power measurements from R.C. Fitch (WPAFB), the collaborations with A.F. Hebard, N. Theodoropoulou and Y.D. Park on spintronics, the supply of GaN substrates from S.S. Park and Y.J. Park (Samsung) and support from ONR and NSF.
REFERENCES
1. see for example, S. Nakamura and S.F. Chichibu (eds.), Introduction to Nitride Semiconductor Blue Lasers and Light-Emitting Diodes (Taylor and Francis, London, 2000).
2. see for example, S. Nakamura, S.J. Pearton and G. Fasal, The Blue Laser Diodes (Springer, Berlin, 2000).
3. H. Morkoc et al., Solid-State Electron. 46, 157 (2002).
4. H.J. Looi et al., Mat. Res. Soc. Symp. Proc. 483, 63 (1998).
5. J.C. Zolper, Solid-State Electron. 42, 2153 (1998).
6. J. Han et al., Appl. Phys. Lett. 73, 1688 (1998).
7. T.J. Zhang et al., Jap. J. Appl. Phys. 40, L921 (2001).
8. S. Nagahama et al., Jap. J. Appl. Phys. 40, L788 (2001).
9. S. Nagahama et al., Phys. Stal. Solidi A 188, 1 (2001).
10. H. Hirayama et al., Phys. Stal. Solidi A 188, 83 (2001).
11. T. Nishida and N. Kobayashi, Phys. Stal. Solid A 188, 113 (2001).
12. J. Han et al., Jap. J. Appl. Phys. 39, 2372 (2000).
13. H. Hirayama et al., Appl. Phys. Lett. 80, 37 (2002).
14. R. Zehringer, A. Stuck and T. Lang, Solid-State Electron. 42, 2139 (1998).
15. J.B. Casady et al., Solid-State Electron. 42, 2165 (1998).
16. A.P. Zhang et al., Appl. Phys. Lett. 78, 823 (2001).
17. A.P. Zhang et al., IEEE Electron. Dev. ED 48, 407 (2001).
18. J.W. Johnson et al., IEEE Electron. Dev. ED 49, 32 (2002).
19. S.S. Park et al., Jap. J. Appl. Phys. 39, L1141 (2000).
20. E. Oh et al., Appl. Phys. Lett. 78, 273 (2001).
21. J.W. Johnson et al., Solid-State Electron. 46, 911 (2002).
22. M.S. Shur, Solid-State Electron. 42, 2131 (1998).
23. S.J. Pearton et al., Mat. Sci. Eng. R 30, 55 (2000).
24. N. Q. Zhang et al., IEEE Electron. Dev. Lett. EDL 21, 421 (2000).
25. G. Simin et al., IEEE Electron. Dev. Lett. EDL 22, 53 (2001).
26. M.S. Shur and M.A. Khan, Physica Scripta, T 69, 103 (1999).
27. S. Karmalkar and U.K. Mishra, IEEE Trans. Electron. Dev. 48, 1515 (2001).
28. O. Breitschadel et al., Mat. Sci. Eng. B 82, 238 (2001).
29. Y.F. Wu et al., IEEE Trans. Electron. Dev. 48, 586 (2001).
30. J.W. Johnson et al., Solid-State Electron. 45, 1979 (2001).
31. B. Lou et al., Appl. Phys. Lett. 79, 2196 (2001).
32. B. Luo et al., Appl. Phys. Lett. 80, 604 (2002).
33. L.F. Eastman et al., IEEE Electron. Dev. ED 48, 479 (2001).
34. B.M. Green et al., IEEE Electron. Dev. Lett. EDL 21, 268 (2000).
35. S.C. Binari et al., IEEE Electron. Dev. ED 48, 465 (2001).
36. R. Vetury et al., IEEE Electron. Dev. ED 48, 560 (2001).
37. B.P. Gila et al., Phys. Stal. Solidi B 228, 171 (2001).
38. B. Luo et al., Appl. Phys. Lett. 80, 1661 (2002).
39. J.M. Kikkawa and D.D. Awschalom, Science 287, 473 (2000).
40. H. Ohno, J. Vac. Sci. Technol. B 18, 2039 (2000).
41. S.A. Wolf et al., Science 294, 1488 (2001).
42. C.H. Tsang et al., IBM J. Res. Develop. 42, 103 (1998).
43. A. Oiwa et al., Appl. Phys. Lett. 78, 518 (2001).
44. D.D. Awschalom and J.M. Kikkawa, Phys. Today 52, 33 (1999).
45. Y.D. Park et al., Appl. Phys. Lett. 77, 3989 (2000).
46. B.T. Jonker et al., Phys. Rev. B 62, 8180 (2000).
47. J. Nitta et al., Phys. Rev. Lett. 78, 1335 (1997).
48. T. Dietl et al., Science 287, 1019 (2000).
49. C. Zener, Phys. Rev. 81, 440 (1950).
50. T. Dietl, H. Ohno and F. Matsukura, Phys. Rev. B 63, 195205 (2001).
51. V.I. Litvinov and V.K. Dugaev, Phys. Rev. Lett. 86, 5593 (2001).
52. M. Berciu and R.N. Bhatt, Phys. Rev. Lett. 87, 108203 (2001).
53. J. Schliemann et al., Appl. Phys. Lett. 78, 1550 (2001).
54. M. von Schilfgaarde and O.N. Myrasov, Phys. Rev. B 63, 233205 (2001).
55. H. Katayama-Yoshida et al., J. Cryst. Growth, 231, 438 (2001).
56. M.L. Reed et al., Appl. Phys. Lett. 79, 7473 (2001).
57. G.T. Thaler et al., presented at AVS Symp., Seattle, Washington, Nov. 2001.
58. S. Sonoda et al., J. Cryst. Growth (in press).
59. W. Gebicki et al., Appl. Phys. Lett. 76, 3870 (2000).
60. M. Zajac et al., Appl. Phys. Lett. 79, 2432 (2001).
61. S. Kuwabara et al., Jap. J. Appl. Phys. 40, L 724 (2001).
62. M.E. Overberg et al., Appl. Phys. Lett. 79, 1312 (2001).
63. Y.L. Soo et al.. Appl. Phys. Lett. 79, 3926 (2001).
64. H. Akinaga et al., Appl. Phys. Lett. 77, 4377 (2000).
Table I. Physical
properties of selected semiconductors.