Development of the nitride-based UV/DUV LEDs
The commercialization of the blue LED started in 1993 , and three primary- color LEDs are now available commercially. Since 1993, the applications of LEDs in various fields have expanded widely. Three years after the commercialization of the blue LED, violet blue LD was developed , which is the shortest-wavelength LD to be realized in the visible region. The use of the white LED, which is composed of a blue LED with phosphors, has led to general lighting applications of LEDs in addition to conventional display applications. All these newly developed devices, including blue LEDs and violet blue LDs, are composed of group III nitride semiconductors. Research activity and commercial activity have stimulated each other, forming a highly effective feedback loop. Breakthroughs made through research have rapidly resulted in the placement of new devices on the market. These new devices are applied in ways not always expected by researchers. The demand for new applications is also acting as a motivating force for new research and development. Therefore, the development of the group-III-nitride-related field is progressing very well.
The successful use of the group III nitrides in blue and the white LEDs and the violet blue LD has led to attempts to realize LEDs and LDs in much shorter wavelength regions such as the ultraviolet (UV) and deep-ultraviolet (DUV) regions. UV/DUV LEDs and UV LDs will be useful for sterilization, lithography, laser-induced fluorescence spectroscopy, biomedical applications, chemical sensing, chemical reaction control, prototyping, drilling, and optical storage, and for use as an excitation source for phosphors and catalysts. For example, the application of a UV/DUV LED to the treatment of skin diseases has been reported .
Attempts to fabricate UV/DUV LEDs and UV LDs have been made by many researchers, and rapid progress has been attained. The first UV LED was reported by Pankove in 1972 and used colloidal carbon and nominally undoped n-type GaN . Since the first pn-junction UV LED that emitted 375 nm light in 1989 , the emission wavelength of nitride-based LEDs has become shorter. In 2006 it reached 210 nm—the target shortest wavelength based on semiconductor light emitters . AlN was used as the active region. The emission wavelength of UV LDs has also become shorter. Figure 1.1 shows the emission wavelength of
Fig. 1.1. Emission wavelength of laser diodes as a function of calendar year.
short-wavelength UV LDs as a function of calendar year. In 2008 the wavelength reached 336 nm of the shortest-wavelength semiconductor LD, though it was operated upon the pulsed mode . This value has not been improved yet.
However, the performances of UV/DUV LEDs and UV LDs are still inferior to those of blue/white LEDs. Figure 1.2 summarizes the development of the EQE of UV/DUV LEDs as a function of calendar year. In 2012, the EQE reached 14.3% at 2 mA with an emission wavelength of 285 nm, which is still inferior to that of state-of-the-art blue LEDs by a factor of six.
Clarifying the factors that limit the EQE of UV/DUV LEDs and UV LDs is necessary to improve their performance. In general, there are three factors
FiG. 1.2. Improvement of the external quantum efficiency of UV (300-350 nm) and DUV (250-300 nm) LEDs as a function of calendar year (as of October 2012).
Fig. 1.3. Definition of each efficiency in this chapter, where Jtotai, Jrad., Jnon rad., and Jleakage denote total, radiative, non-radiative, and leakage current density, respectively.
that limit the EQE of LEDs: (1) internal quantum efficiency (niQE), (2) carrier injection efficiency (qIE), and (3) light extraction efficiency (nLEE). Figure 1.3 shows schematically the efficiency-limiting process of LEDs.
Among these factors, IQE is dependent on the radiative recombination rate of each material and the rate of non-radiative recombination, including Auger recombination. Theoretically, the radiative recombination coefficient of free ex- citons in AlN is larger than that for GaN or InN . Therefore, AlGaN-based QWs should be more radiative or should have a higher IQE than InGaN-based QWs. One of the most important factors for improving IQE is the growth of AlGaN with low defect density. In the following section, the development of the growth of AlGaN and the relationship between IQE and defect density are described.
Carrier injection efficiency is strongly dependent on the hole concentration in the p-type cladding layer as well as the alloy composition. With increasing Al content in p-type AlGaN, the hydrogen-like effective mass acceptor activation energy becomes large. Therefore, the room-temperature hole concentration decreases with increasing Al content. Of course, hole concentration can be increased by increasing the acceptor concentration. However, the heavy doping of Mg causes serious problems. Therefore, there is an optimum Mg concentration for obtaining a high hole concentration. In the following section, the effect of a low hole concentration in a p-type cladding layer on the performance of UV/DUV LEDs is discussed.
The light extraction efficiency is the most important parameter that limits the performance of UV/DUV LEDs. In general, face-up or face-down configurations are used for LEDs. The face-up configuration requires a transparent electrode. Unfortunately, suitable electrode materials that are transparent in the UV/DUV region are uncommon. In contrast, the face-down configuration requires a highly reflective electrode. There are some suitable materials that are reflective in the DUV region. High contact resistivity to p-type AlGaN is one of the most serious issues. In the case of conventional Ni/Au, Indium-Tin-Oxide, or Ag-Pd-Cu electrodes, p-GaN, which is opaque to UV and DUV light, is required to realize low contact resistivity.
In the following section, issues regarding IQE that are related to the crystal growth of AlGaN, IE and LEE are discussed in more detail.