Emission internal quantum efficiency (IQE) is an important index evaluating the quality of active layers in optical devices. The IQE is usually estimated from temperature dependence of photoluminescence (PL) intensity by assuming that the IQE at cryogenic temperature is 100 %. III-nitride films, however, usually have high-density crystal defects, and it cannot be said that the assumption is valid. In 2016, we proposed a new method to estimate accurate IQE values by simultaneous PL and photo-acoustic (PA) measurements, and demonstratively evaluated the IQE values for various GaN samples. After that, we applied the method to InGaN quantum-well active layers and estimated the IQE values and their excitation carrier-density dependence in the layers. In this paper, we will introduce the principle of our simultaneous PA/PL measurement and will present some examples of absolute measurement of IQE values in III-nitride semiconductor samples.

Group III-V nitride semiconductors were evaluated using photothermal deflection spectroscopy (PDS), which detects the heat generated by non-radiative recombination through the defect levels with high sensitivity. This method has the characteristic of being able to evaluate even electrically and optically inactive samples such as ultrathin films and ion-implanted samples. Since the PDS can evaluate the defect levels in the band gap with deeper range, it gives a new picture of the level in the band gap and valence band structure. Then, PDS method has been applied to the development of III-V nitride semiconductors as a relatively simple defect evaluation method. In this article, the improvement of the carrier transport of two-dimensional electron gas at the InGaN channel was demonstrated by detecting the alloy disorder with PDS method. Also, the behavior of Mg ion-implanted GaN caused by thermal annealing was discussed for the requirement to realize p-type conduction in term of the valence band structure and defects in the band gap evaluated by PDS.

Omnidirectional photoluminescence (ODPL) spectroscopy, which can determine the internal quantum efficiency (IQE) of freestanding semiconductor crystals with direct bandgap without the need for model calculations, is reviewed. ODPL can determine IQE from the external quantum efficiency (EQE) based on the fact that light with self-absorption energy in the vicinity of the band-edge of a freestanding semiconductor crystal is emitted mainly in the normal direction of crystal surface. The procedure and applications of IQE determination are discussed using gallium nitride, zinc oxide, and metal halide perovskite as example cases. In particular, the ability to measure a large crystal sample outside an integrating sphere is one of the main features of ODPL spectroscopy, and it is expected to be combined with mapping measurements of the entire surface of semiconductor wafers and various types of nonlinear and microscopic spectroscopy.

Generation of heat energy or non-thermal-equilibrium phonons in the active regions of light-emitting diodes and transistors deteriorates their functional properties. We show phonon transport characteristics for defective In_xGa_1-xN (∼100nm) / GaN heterointerfaces studied by microscopic Raman imaging using a double-laser irradiation system. Simultaneous irradiation of two lasers enables the detection of site-selected crystal heating and transport of phonons in lateral and cross-interface directions from the position of the laser-induced phonon generation to the position of Raman-scattering detection. Two-dimensional images of the decrease in the Stokes-signal peak energy of the E₂(high) modes by the laser-heating give us the direct observation of local blocking of phonon-transport in the vicinity of misfit dislocations for the sample of x = 0.16 and threading dislocations in the sample of x = 0.05. Lateral transport over 20 μm, indicating diffusive transport, is observed for the latter sample. Non-diffusive or ballistic phonon transport under the heterointerface is expected from the images showing the defect arrangement in the InGaN layer in the image of the E₂(high)-mode energy of the GaN underlayer.

Phonon-related properties of nitride semiconductors, especially the effects of defects on them, are of significance in performance improvement of various devices based on nitride semiconductors. Since tackling this issue using first-principles calculations often demands huge computational costs, machine learning potentials have attracted much attention recently as a method to achieve high prediction accuracy and low computational cost simultaneously. In this article, we describe our recent attempts in developing high-dimensional neural network potential (HDNNP), one of machine learning potentials, for analyzing phonon-related properties of GaN. We show that the developed HDNNP reproduces the results of first-principles calculations of phonon band structures and thermal conductivities. In addition, preliminary results on the effects of nitrogen vacancy on thermal conductivity are shown. Future tasks and prospects of machine learning potentials are also discussed.