Luminescence properties of defects in GaN

Abstract

Gallium nitride ͑GaN͒ and its allied binaries InN and AIN as well as their ternary compounds have gained an unprecedented attention due to their wide-ranging applications encompassing green, blue, violet, and ultraviolet ͑UV͒ emitters and detectors ͑in photon ranges inaccessible by other semiconductors͒ and high-power amplifiers. However, even the best of the three binaries, GaN, contains many structural and point defects caused to a large extent by lattice and stacking mismatch with substrates. These defects notably affect the electrical and optical properties of the host material and can seriously degrade the performance and reliability of devices made based on these nitride semiconductors. Even though GaN broke the long-standing paradigm that high density of dislocations precludes acceptable device performance, point defects have taken the center stage as they exacerbate efforts to increase the efficiency of emitters, increase laser operation lifetime, and lead to anomalies in electronic devices. The point defects include native isolated defects ͑vacancies, interstitial, and antisites͒, intentional or unintentional impurities, as well as complexes involving different combinations of the isolated defects. Further improvements in device performance and longevity hinge on an in-depth understanding of point defects and their reduction. In this review a comprehensive and critical analysis of point defects in GaN, particularly their manifestation in luminescence, is presented. In addition to a comprehensive analysis of native point defects, the signatures of intentionally and unintentionally introduced impurities are addressed. The review discusses in detail the characteristics and the origin of the major luminescence bands including the ultraviolet, blue, green, yellow, and red bands in undoped GaN. The effects of important group-II impurities, such as Zn and Mg on the photoluminescence of GaN, are treated in detail. Similarly, but to a lesser extent, the effects of other impurities, such as C, Si, H, O, Be, Mn, Cd, etc., on the luminescence properties of GaN are also reviewed. Further, atypical luminescence lines which are tentatively attributed to the surface and structural defects are discussed. The effect of surfaces and surface preparation, particularly wet and dry etching, exposure to UV light in vacuum or controlled gas ambient, annealing, and ion implantation on the characteristics of the defect-related emissions is described.

Realising epitaxial growth of GaN on (001) diamond

“…The second peak cannot be related to acceptor bound recombination (ABE) because the energy distance (78 meV) from the DBE peak is too large. Instead, it is attributed to the less commonly observed Y 2 line, 49 normally detected at 3.40 eV, but, again, the emission can be shifted toward lower energy due to tensile stress. The Y 2 line is reported to be influenced by acceptor doping 49 (like carbon) and is strongly related to stacking faults parallel to the substrate/film interface, 50,51 which originate from the overgrowth in the lateral direction of GaN islands with different altitudes, as described above in Sec.…”

High intensity solid‐state UV source for time‐gated luminescence microscopy

“…Luminous output of the UV LED at a distance 2 cm from the LED face was 30 mW cm 22 and this was reduced to 27.2 mW cm 22 after passage through the collimating lens. Power exiting from the filter cube (following reflection from the dichroic) was measured at 14.1 mW cm 22 , thus only about half the output from the LED was actually directed into the microscope objective.…”

“…Power exiting from the filter cube (following reflection from the dichroic) was measured at 14.1 mW cm 22 , thus only about half the output from the LED was actually directed into the microscope objective. Filter-cubes from two different manufacturers were equipped with UV LEDs and both assemblies produced similar optical outputs (highest values shown here).…”

“…Relative power output was reduced 50% at 60°off-axis for the engineering samples compared to 45°for the production devices with better directivity. Although rated output of the devices at 500 mA was 100 mW cm 22 , only a portion of the output was captured by the optical sensor that was situated about 2 cm from the emitting face of the LED. In Figure 5a, the fluorescence of the central desmid (Chlorophyta staurastrum; the trigonal organism) is strong by comparison with the immunofluorescently labelled Giardia cysts when viewed in conventional epifluorescence mode.…”

“…Typically LEDs can be switched within a couple of nanoseconds, however, as a consequence of trapped states within the gallium nitride lattice (22,23), some residual long-lived emission is observed in the visible for the UV LEDs. The broad visible luminescence represents unwanted optical output from the excitation source and must be removed to ensure that the key advantage of timeresolved techniques, low background, is maintained.…”