Open Access
Add to BibSonomy
Abstract
We have observed the transient behavior in the AlGaN photoluminescence. Under an excitation of 325 nm He-Cd laser beam, the blue luminescence (BL) bands and yellow luminescence (YL) bands of AlGaN vary with increasing illumination time. We propose that the chemical reactions between BL-related CNON-Hi (CN-Hi) and YL-related CN-Hi (CN) defect states are the cause of such a phenomenon. The BL transition temperature (Tt) is defined as the temperature at which the intensity of BL bands induced by CNON-Hi is equal to that originated from CN-Hi. Only at Tt, BL shows a peak energy variation due to the exposure. The Tt of AlGaN is higher than what is similarly detected in GaN because of the high reactivity of Al to O.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
GaN-based ultraviolet (UV) devices have multiple important applications in medical, industrial, and environmental fields [1–4]. AlGaN is generally used as active region of UV light-emitting diodes (LEDs) and laser diodes (LDs), especially in UV-B and UV-C region [5–9]. Hence, studying the photoelectric characteristics of AlGaN is essential for the realization of GaN-based UV device with high performance. Recently, a so-called transient behavior (or fatigue behavior) where the peak energy and intensity of blue luminescence (BL) bands and yellow luminescence (YL) bands are changed with time, the intensities may be weakened or enhanced during the laser excitation in the photoluminescence (PL) test has been reported for GaN material [10–12]. This behavior is thought to be associated with metastable defect centers [12]. We have also reported and discussed the different transient behaviors of GaN observed in a temperature range from 10 K to 300 K in a previous study [13]. It is very possible that AlGaN material, like the GaN material, will also show such a transient behavior. In this case, the transient behavior in AlGaN may influence the performance of GaN-based UV devices. However, the transient behavior of AlGaN is rarely reported or discussed till now.
In this study, we have observed the transient luminescence behavior in the AlGaN sample. The difference between the transient behavior of AlGaN and GaN has been discussed.
2. Experiment
2 AlGaN and GaN film samples are grown by metal-organic chemical vapor deposition (MOCVD) on sapphire substrates. The sample A contains a 1 µm AlGaN layer with 12.1% aluminum concentration on an AlN buffer layer while sample B only contains a 1 µm GaN layer. During the growth process, trimethylaluminum (TMAl), trimethylgallium (TMGa) and NH3 are used as aluminum, gallium and nitrogen precursors respectively. The AlGaN layer is grown at 1070 °C and 75 Torr while the GaN layer is grown at 1000 °C and 200 Torr. The C, H, O impurity concentrations of sample A are 1.0 × 1017 cm−3, 4.1 × 1017 cm−3 and 1.1 × 1017 cm−3 respectively while the C, H, O concentrations of sample B are 3.0 × 1017 cm−3, 5.0 × 1017 cm−3 and 6.0 × 1016 cm−3 respectively according to the secondary ion mass spectroscopy (SIMS) tests.
Two serials of PL testes, exposure-time dependent PL (EDPL) tests and temperature dependent PL (TDPL) tests, have been taken using a continuous wave (CW) 325 nm He-Cd laser as the excitation light source to analyze transient behaviors of the studied samples. It is noted that sample A during each run of PL measurement is exposed under the laser illumination nearly for 2 minutes. For EDPL tests, several runs of PL tests are taken with the increasing laser exposure time duration. The zero exposure time (marked as 0 min.) represents the case where the sample was not exposed under the laser illumination before the EDPL measurement. In other EDPL measurements, each run has an exposure time which lasts from 2 min. 4 min., 6 min. and 8 min., i.e. the sample has been experienced 1-4 runs of PL measurement and hence laser exposure for sample A. As for sample B, the duration of each run of PL measurement is nearly 1.5 min as the whole measurement time of each PL spectrum needed for GaN sample is a little shorter than for AlGaN. Hence, for sample B each EDPL run has an exposure time which lasts from 0 min., 1.5 min. 3 min., 4.5 min., 6 min. and 9 min. When the measurement temperature is changed, at each test temperature, a different spot site is chosen on the sample to be excited in order to avoid the influence of the history of previous illumination. In TDPL tests, the same spot site of sample is excited at all temperature. Besides, laser illumination is always blocked during the temperature lifting process.
3. Result
The EDPL results of sample A are shown in Fig. 1 and Fig. 2. The EDPL result includes the PL spectra measured at 4 different temperatures and after 5 different exposure time durations of 0-8 minutes (Fig. 1), and the variations of normalized integral intensity (a, c) and peak energy (b, d) of BL and YL with the exposure time, measured at 5 different temperatures of 30 K, 70 K, 180 K, 220 K, and 300 K (Fig. 2). It is found that the PL peak energy of AlGaN band-band transition is 3.73 eV at 30 K. The BL band centered at 3.0-3.2 eV and the YL band centered at 2.4-2.5 eV show various transient behaviors at different test temperatures. At 30 K, the BL intensity decreases with increasing exposure time but its peak energy hardly changes. Meanwhile, YL intensity rises with a decreasing peak energy during the exposure. At 70-160 K, BL variation is similar to at 30 K, but YL intensity starts to decline with increasing luminescence time. At 180 K, BL is weakened slightly but YL is enhanced during the exposure. At 220 K, both BL intensity and peak energy decrease after few-minute of luminescence, but YL intensity rises slightly. At 260-300 K, YL intensity declines while BL hardly changes during the exposure.
Fig. 1. The EDPL spectra of sample A at 30 K (a), 180 K (b), 220 K (c), 300 K (d) under different exposure (and luminescence) time durations of 0-8 min.
Fig. 2. The variations of normalized integral intensity (a) and peak energy of BL (b) with exposure time in sample A; the variation of normalized integral intensity (c) and the peak energy of YL (d) with exposure time in sample A. The variations of YL and BL at five representative temperature values, 30 K (squares, purple), 70 K (circles, blue), 180 K (upward triangles, green), 220 K (downward triangles, yellow), and 300 K (rhombs, red), are shown in the figure.
The TDPL test results of sample A in Fig. 3 show various temperature dependencies of the BL and YL peak energies for AlGaN sample A. As can be seen in Fig. 3(a), BL peak shows an abrupt peak energy drop, about 100 meV, from 160 to 200 K. At the temperature below 140 K, the peak energy of BL is higher than 3.10 eV (called as BL1). However, when the temperature reaches 220 K, the peak energy decreases to near 3.00 eV (called as BL2). This phenomenon is similar to the one we previously have observed in a GaN sample [13]. This indicates that BL emission at temperature over 220 K may be mainly generated by an electronic transition related to a level different from one of the BL observed at temperature lower than 140 K. In other words, the main sources of BL at lower temperature and at higher temperature may be different from each other.
Fig. 3. (a) The temperature-dependent BL (circles) and YL (triangles) peak energies dependence of sample A in TDPL tests (before the correction of eliminating exposure time-induced effect). (b) The temperature dependent BL and YL peak energies of sample A measured with 0-minute exposure time in EDPL test. (c) The temperature-dependent PL integral intensities of BL of sample A. The green and red dotted lines represent the fitting line of BL1 and BL2 respectively; the black solid line represents the fitting line of the sum of BL1 and BL2 intensity. (d) The temperature-dependent PL integral intensities of YL of sample A. The green and red dotted lines represent the fitting line of YL1 and YL2 respectively; the black solid line represents the fitting line of the sum of YL1 and YL2 intensity.
On the other hand, it seems that such peak energy drop does not happen for YL if only considering the TDPL results in Fig. 3(a). However, the influence of exposure time is unavoidable because the same spot site of sample is illuminated by the laser beam and the sample is excited during the whole process of TDPL test. In other words, the PL result at each temperature for TDPL test also has been influenced by different illumination time. In a more ideal condition, the accurate TDPL test result should be corrected and the measurement should be made with 0-minute exposure time for each test temperature considering that the PL spectra may be changed by the increasing exposure time. Hence, we add the temperature-dependent peak energy data with 0-minute exposure time in EDPL in order to figure out the real peak energy variation with the changing temperature as shown in Fig. 3(b). Actually, a similar peak energy drop for YL is observed in Fig. 3(b). Below 160 K, YL peak energy is near 2.44 eV (called as YL1), and it hardly changes with the rising temperature. However, when the temperature reaches 200 K, YL peak energy suddenly decreases to 2.41 eV (called as YL2) and then this value is kept unchanged up to 300 K. This indicates that the YL may also have two different sources at lower and higher temperature, respectively. Besides, the variation of BL peak energy also shows a little difference from what is observed in Fig. 3(a). The main difference between Fig. 3(a) and Fig. 3(b) is that in Fig. 3(b), the BL peak energy drop happens in the temperature range of 180-260 K instead of between 160 and 200 K as shown in Fig. 3(a).
The PL intensity variation with test temperature can be fit by Arrhenius model generally [14,15]:
where I0 is the PL intensity at 0 K, C is a constant, EA is the activation energy, and k is the Boltzmann constant. Because BL may have two different sources and two different peak energies, the total BL intensity IBL will be the sum of these two BL bands, BL1 and BL2 (IBL1 and IBL2). Assuming IBL1 and IBL2 are independent variants, they can be fitted by Arrhenius model respectively. Therefore, the total IBL will be:The same model can be applied to YL intensity analysis:
Noteworthily, the intensity of BL1 and BL2 will be equal at a certain temperature as shown in Fig. 3(c). We call such a temperature as the temperature of BL transition. Near the temperature of BL transition, the BL of AlGaN shows many interesting phenomena which will be discussed in detail later.
To further understand the difference between AlGaN’s transient behaviors and GaN’s ones, EDPL spectra of GaN sample B are measured and the data are analyzed. As shown in Fig. 4(a) and (b), at the temperature below 160 K, BL intensity of sample B becomes weaker with increasing exposure time. Meanwhile, BL peak energy decreases with increasing exposure time at 120-140 K. When the temperature rises and reaches 160 K, BL intensity and peak energy stops changing during the illumination. On the other hand, as shown in Fig. 4(c) and (d), at the temperature below 120 K, YL intensity changes in an un-monotonic way while it rises with increasing exposure time at both 140 and at 160 K. Then YL intensity hardly changes with increasing exposure time until the temperature rises to reach 260 K. Then YL intensity becomes weaker with increasing exposure time. In addition, from 30 to 140 K, YL peak energy shows a red-shift with increasing exposure time, and no obvious variation of YL peak energy is observed with increasing exposure time when the temperature is above 160 K.
Fig. 4. The variation of normalized integral intensity (a) and peak energy of BL (b) with increasing exposure time in sample B; the variation of normalized integral intensity (c) and the peak energy of YL (d) with increasing exposure time in sample B. The variations of YL and BL are measured at six representative temperature values, 30 K (squares, purple), 120 K (circles, blue), 140 K (upward triangles, green), 160 K (downward triangles, yellow), 220 K (rhombs, orange), and 300 K (leftward triangles, red).
4. Discussions
It is found that the transient behaviors of AlGaN sample A are similar to the ones of GaN sample B which were reported previously in [13]. Hence, a similar mechanism may be also applied on the explanation about the transient behaviors of the AlGaN sample: 1) Both BL and YL are generated from two different kinds of carbon-related point defects, respectively; 2) The chemical transformations between two defect states related to BL and the other two states related to YL cause the transient behaviors of AlGaN.
The BL bands and YL bands in GaN PL are generally considered to originate from electronic transitions related to defect states, such as impurities [17,18], vacancies [19,20], dislocations [21,22]. Particularly, carbon-related defects are often found in MOCVD grown GaN due to the application of organic sources. The relative high C impurity concentrations in the samples A (1.0 × 1017 cm−3) and B (1.0 × 1017 cm−3) also suggest that the carbon-related defects are very likely the sources of BL and YL. The peak energy of BL observed in the AlGaN film sample is peaked at 2.99 eV (BL2) to 3.16 eV (BL1). It is known that CNON-Hi complex may yield a BL band maximum at 3.03 eV in GaN [11]. Considering that AlGaN band gap is near 0.2 eV broader than the one of GaN (GaN 3.50 eV at 30 K [23], the AlGaN sample 3.73 eV at 30 K), CNON-Hi may be assigned as the source of AlGaN BL1. In addition, CN-Hi will generate a BL band peak at 2.95 eV in GaN so it may also induce a BL band maximum at 3.15 eV in AlGaN [11]. This value seems to be a little larger than the one of BL2 in AlGaN. However, because the peak energy of BL2 is measured at 300 K, the band gap of AlGaN will be narrower at higher temperature (a temperature-induced shrinkage of about 0.1 eV). Hence, CN-Hi may cause a BL band with a peak energy close to BL2 as what is observed in AlGaN. At the temperature below 180 K, the intensity of BL band induced by CNON-Hi is far higher than the one caused by CN-Hi. In other words, CNON-Hi is the main source of BL in this low temperature range. However, when the temperature rises and reaches 180 K, the intensity of CNON-Hi-related BL decreases abruptly while the intensity of BL induced by CN-Hi changes only slightly. Hence, the peak energy of BL will be mainly determined by BL2, declining suddenly when temperature increases from 180 K to 260 K. Then CN-Hi becomes the main source of BL and this BL peak energy hardly changes with the further temperature lifting at the temperature above 260 K. The fitting results in Fig. 3. (c) also support above analysis.
The peak energies of YL band, on the other hand, are located from 2.41 eV (YL2) to 2.44 eV (YL1) in the AlGaN sample. YL bands may be generated from carbon-related point defects CNON and CN which can cause the YL bands of 2.25 eV and 2.59 eV in GaN [24,25]. At the low temperature below 160 K, the intensity of YL band induced by CNON is close to the one caused by CN. Hence, the total YL band is a mixture originated from these two sources. In this particular case, it causes a YL band maximum at nearly 2.44 eV (YL1). When temperature rises from 160 to 200 K, the YL generated from CN decays quickly, which induces a decrease of YL peak energy with the rising temperature. When the temperature is over 200 K, the intensity of YL induced by CN is no longer comparable with CNON-related one’s. Therefore, the peak energy of YL stops varying and is fixed at 2.41 eV (YL2). This analysis also conforms with the fitting results show in Fig. 3(d).
Based on the analysis mentioned above, the variation of YL bands and BL bands with the increasing exposure time may be induced by the chemical transformations between defect states. It is known that under the excitation of 325 nm laser, BL-related CNON-Hi (or CN-Hi) transitions may gradually turn into YL-related CNON (or CN) ones due to the recombination enhanced defect reaction mechanism [11,26]. The transition between defect states of luminescence origin will cause the variation of PL intensity with the exposure time if the involved defects are the main source of BL (or YL). Moreover, if the intensities of BL (or YL) generated from CNON-Hi and CN-Hi (or CNON and CN) are comparable, the transition behavior will cause the peak energy variation of the whole BL (or YL) band. As shown in Table 1, at the low temperature below 180 K, the main source of BL is CNON-Hi. Under the excitation of the 325 nm laser, the source turns from CNON-Hi to CNON, which causes the decline of BL intensity during the exposure. On the other hand, at the same time CN changes into CN-Hi. The reducing CN defect states and increasing CNON states contribute to the decrease of YL peak energy since the CN state has a higher energy level than CNON. Moreover, the quantity variation of CN and CNON centers will also play a role on the YL intensity. The increase of CNON will enhance YL while the decrease of CN will weaken it. These two effects are competitive so that the intensity of the whole YL band rises with increasing temperature at 30 K but declines at 70-160 K during the exposure. Since the reaction between CN and CN-Hi may reach a chemical equilibrium at 180 K, only the increase of CNON centers contributes to the YL intensity. Hence, the YL is enhanced with a slight decline of peak energy as observed at 180 K.
Table 1. Main sources of YL and BL and reactions between them in sample A
When the temperature rises and reaches 220 K, the intensity of BL induced by CNON-Hi is comparable to that generated from CN-Hi. In fact, a reduction of CNON-Hi will not only cause a decline of BL intensity, but also a decrease of its peak energy. Besides, YL intensity rises because of the increase of CNON defect states. Noteworthily, here 220 K is a special temperature value. As shown in Fig. 2(b) the transient variation of BL peak energy can be clearly observed only at 220 K. On the other hand, as shown in Fig. 3(b), the BL peak energy drops with rising temperature also happen in the temperature range of 180-260 K. Therefore, at 220 K, the intensities of BL induced by CNON-Hi and CN-Hi are equal. That is why we define this temperature 220 K as the so-called BL transition temperature. In this article, only the temperature of BL transition is emphasized to discuss in detail while the one of YL transition has not been mentioned so much as there may be no such a specific temperature at which the intensities of YL bands induced by CN and CNON are equal. As can be seen from Fig. 3(d), the intensities of YL originated from these two defects are comparable in a larger temperature range. The peak energy variation of YL from 30 to 180 K detected by the EDPL test also supports this assumption, i.e. there may be no such a temperature of YL transition. Since the chemical equilibrium conditions are related to the temperature, the equilibrium concentrations will change with temperature. At the temperature below 220 K, the initial concentration of CNON defect states is below its equilibrium concentration in this temperature range and thus CNON-Hi changes into CNON under the excitation. However, when the temperature reaches 260 K, the equilibrium concentrations change and the initial concentration of CNON defect states exceeds its equilibrium concentration. Therefore, CNON begins to change into CNON-Hi during the light exposure. The YL intensity is reduced due to the decrease of CNON. However, BL hardly changes because the main source of BL is CN-Hi in this temperature range.
The largest difference of transient behaviors between the AlGaN and GaN samples is the difference of their temperatures of BL transition. According to above discussions about the temperature of BL transition, the BL peak energy will only vary at the temperature of BL transition during the exposure. Hence, the temperature of BL transition for sample B is at 120-140 K, which is quite lower than that of sample A. At the temperature of BL transition:
5. Conclusion
AlGaN shows various transient behaviors at 30-300 K which are similar to what exist in GaN. The transient behavior in AlGaN is caused by the chemical transitions between BL-related CNON-Hi (CN-Hi) and YL-related CN-Hi (CN) defects. At the temperature below 160 K, BL mainly originates from CNON-Hi while YL is induced by CNON and CN. CNON-Hi changes into CNON and CN turns into CN-Hi during the illumination, which causes the intensity variations of BL and YL. On the other hand, YL peak energy reduces because CNON defect states have a lower level than CN. At 180 K, CN stops changing into CN-Hi and then YL intensity rises due to the increase of CNON defect states. At 220 K, the intensity of BL induced by CNON-Hi is comparable to that generated from CN-Hi and thus BL shows decreasing peak energy and intensity due to the transition from CNON-Hi to CNON. When the temperature rises and reaches 260 K, CNON begins to change into CNON-Hi, which cause a decline of YL intensity during the excitation. We have defined the temperature where the intensities of BL bands originated from CNON-Hi and CN-Hi are equal as Tt. Only at the temperature near Tt, BL shows a peak energy variation during the luminescence. Compared with GaN, the AlGaN sample shows a higher Tt and it is assigned to the high reactivity of Al to O in AlGaN films.
Funding
Beijing Municipal Science & Technology Commission, Administrative Commission of Zhongguancun Science Park (Z211100007921022, Z211100004821001); National Natural Science Foundation of China (62034008, 62074142, 62074140, 61974162, 61904172, 61874175, 62127807, U21B2061); Key Research and Development Program of Jiangsu Province (BE2021008-1); Beijing Nova Program (202093); Strategic Priority Research Program of Chinese Academy of Sciences (XDB43030101); Youth Innovation Promotion Association of the Chinese Academy of Sciences (2019115).
Acknowledgments
This work was supported by Beijing Municipal Science & Technology Commission, Administrative Commission of Zhongguancun Science Park (Z211100007921022, Z211100004821001), National Natural Science Foundation of China (Grant Nos. 62034008, 62074142, 62074140, 61974162, 61904172, 61874175, 62127807, U21B2061), Key Research and Development Program of Jiangsu Province (BE2021008-1), Beijing Nova Program (Grant No. 202093), Strategic Priority Research Program of Chinese Academy of Sciences (Grant No. XDB43030101), and Youth Innovation Promotion Association of Chinese Academy of Sciences (Grant No. 2019115).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. J. Walter, T. Cremer, K. Miyagawa, and S. Tashiro, “A new system for laser-UVA-microirradiation of living cells,” J. Microsc. 209(2), 71–75 (2003). [CrossRef]
2. M. Kneissl, T. Kolbe, J. Schlegel, J. Stellmach, C. Chua, Z. Yang, A. Knauer, V. Küller, M. Weyers, and N. M. Johnson, “AlGaN-based Ultraviolet Lasers - Applications and Materials Challenges,” in CLEO:2011 - Laser Applications to Photonic Applications, OSA Technical Digest (CD) (Optical Society of America, 2011), JTuB1.
3. H. Amano, R. Collazo, C. D. Santi, S. Einfeldt, M. Funato, J. Glaab, S. Hagedorn, A. Hirano, H. Hirayama, R. Ishii, Y. Kashima, Y. Kawakami, R. Kirste, M. Kneissl, R. W. Martin, F. Mehnke, M. Meneghini, A. Ougazzaden, P. J. Parbrook, S. Rajan, P. Reddy, F. Römer, J. Ruschel, B. Sarkar, F. Scholz, L. Schowalter, P. Shields, Z. Sitar, L. Sulmoni, T. Wang, T. Wernicke, M. Weyers, B. Witzigmann, Y.-R. Wu, T. Wunderer, and Y. Zhang, “The 2020 UV Emitter Roadmap,” J. Phys. D: Appl. Phys. (2020).
4. D. B. Li, K. Jiang, X. J. Sun, and C. L. Guo, “AlGaN photonics: recent advances in materials and ultraviolet devices,” Adv. Opt. Photonics 10(1), 43–110 (2018). [CrossRef]
5. N. Susilo, J. Enslin, L. Sulmoni, M. Guttmann, U. Zeimer, T. Wernicke, M. Weyers, and M. Kneissl, “Effect of the GaN:Mg Contact Layer on the Light-Output and Current-Voltage Characteristic of UVB LEDs,” Phys. Status Solidi A 215, 1700643 (2018). [CrossRef]
6. N. Maeda, M. Jo, and H. Hirayama, “Improving the Light-Extraction Efficiency of AlGaN DUV-LEDs by Using a Superlattice Hole Spreading Layer and an Al Reflector,” Phys. Status Solidi A 215(8), 1700436 (2018). [CrossRef]
7. S. Tanaka, Y. Ogino, K. Yamada, T. Omori, R. Ogura, S. Teramura, M. Shimokawa, S. Ishizuka, A. Yabutani, S. Iwayama, K. Sato, H. Miyake, M. Iwaya, T. Takeuchi, S. Kamiyama, and I. Akasaki, “AlGaN-based UV-B laser diode with a high optical confinement factor,” Appl. Phys. Lett. 118(16), 163504 (2021). [CrossRef]
8. K. Sato, S. Yasue, K. Yamada, S. Tanaka, T. Omori, S. Ishizuka, S. Teramura, Y. Ogino, S. Iwayama, H. Miyake, M. Iwaya, T. Takeuchi, S. Kamiyama, and I. Akasaki, “Room-temperature operation of AlGaN ultraviolet-B laser diode at 298 nm on lattice-relaxed Al0.6Ga0.4N/AlN/sapphire,” Appl. Phys. Express 13(3), 031004 (2020). [CrossRef]
9. T. Omori, S. Ishizuka, S. Tanaka, S. Yasue, K. Sato, Y. Ogino, S. Teramura, K. Yamada, S. Iwayama, H. Miyake, M. Iwaya, T. Takeuchi, S. Kamiyama, and I. Akasaki, “Internal loss of AlGaN-based ultraviolet-B band laser diodes with p-type AlGaN cladding layer using polarization doping,” Appl. Phys. Express 13(7), 071008 (2020). [CrossRef]
10. R. Armitage, Q. Yang, and E. R. Weber, “Analysis of the carbon-related “blue” luminescence in GaN,” J. Appl. Phys. 97(2005).
11. D. O. Demchenko, I. C. Diallo, and M. A. Reshchikov, “Hydrogen-carbon complexes and the blue luminescence band in GaN,” J. Appl. Phys. 119(3), 035702 (2016). [CrossRef]
12. S. J. Xu, G. Li, S. J. Chua, X. C. Wang, and W. Wang, “Observation of optically-active metastable defects in undoped GaN epilayers,” Appl. Phys. Lett. 72(19), 2451–2453 (1998). [CrossRef]
13. B. Wang, F. Liang, D. Zhao, Y. Ben, J. Yang, P. Chen, and Z. Liu, “Transient behaviours of yellow and blue luminescence bands in unintentionally doped GaN,” Opt. Express 29(3), 3685–3693 (2021). [CrossRef]
14. M. Leroux, N. Grandjean, B. Beaumont, G. Nataf, F. Semond, J. Massies, and P. Gibart, “Temperature quenching of photoluminescence intensities in undoped and doped GaN,” J. Appl. Phys. 86(7), 3721–3728 (1999). [CrossRef]
15. M. A. Reshchikov, D. O. Demchenko, J. D. McNamara, S. Fernández-Garrido, and R. Calarco, “Green luminescence in Mg-doped GaN,” Phys. Rev. B 90(3), 035207 (2014). [CrossRef]
16. W. Götz, N. M. Johnson, C. Chen, H. Liu, C. Kuo, and W. Imler, “Activation energies of Si donors in GaN,” Appl. Phys. Lett. 68(22), 3144–3146 (1996). [CrossRef]
17. H. S. Zhang, L. Shi, X. B. Yang, Y. J. Zhao, K. Xu, and L. W. Wang, “First-Principles Calculations of Quantum Efficiency for Point Defects in Semiconductors: The Example of Yellow Luminance by GaN: C-N + O-N and GaN:C-N,” Adv. Opt. Mater. 5, 1700404 (2017). [CrossRef]
18. M. A. Reshchikov, P. Ghimire, and D. O. Demchenko, “Magnesium acceptor in gallium nitride. I. Photoluminescence from Mg-doped GaN,” Phys. Rev. B 97(20), 205204 (2018). [CrossRef]
19. R. Armitage, W. Hong, Q. Yang, H. Feick, J. Gebauer, E. R. Weber, S. Hautakangas, and K. Saarinen, “Contributions from gallium vacancies and carbon-related defects to the “yellow luminescence” in GaN,” Appl. Phys. Lett. 82(20), 3457–3459 (2003). [CrossRef]
20. Z. Xie, Y. Sui, J. Buckeridge, A. A. Sokol, T. W. Keal, and A. Walsh, “Prediction of multiband luminescence due to the gallium vacancy-oxygen defect complex in GaN,” Appl. Phys. Lett. 112(26), 262104 (2018). [CrossRef]
21. Y. Linkai, Q. Haoran, H. Jialin, Z. Mei, Z. Degang, J. Desheng, Y. Jing, L. Wei, and L. Feng, “Influence of dislocation density and carbon impurities in i-GaN layer on the performance of Schottky barrier ultraviolet photodetectors,” Mater. Res. Express 5(4), 046207 (2018). [CrossRef]
22. D. G. Zhao, D. S. Jiang, J. J. Zhu, Z. S. Liu, H. Wang, S. M. Zhang, Y. T. Wang, and H. Yang, “Role of edge dislocation and Si impurity in linking the blue luminescence and yellow luminescence in n-type GaN films,” Appl. Phys. Lett. 95(4), 041901 (2009). [CrossRef]
23. K. B. Nam, J. Li, J. Y. Lin, and H. X. Jiang, “Optical properties of AlN and GaN in elevated temperatures,” Appl. Phys. Lett. 85(16), 3489–3491 (2004). [CrossRef]
24. D. O. Demchenko, I. C. Diallo, and M. A. Reshchikov, “Yellow Luminescence of Gallium Nitride Generated by Carbon Defect Complexes,” Phys. Rev. Lett. 110(8), 087404 (2013). [CrossRef]
25. M. A. Reshchikov, D. O. Demchenko, A. Usikov, H. Helava, and Y. Makarov, “Carbon defects as sources of the green and yellow luminescence bands in undoped GaN,” Phys. Rev. B 90(23), 235203 (2014). [CrossRef]
26. L. C. Kimerling, “Recombination enhanced defect reactions,” Solid-State Electron. 21(11-12), 1391–1401 (1978). [CrossRef]
27. H. W. Jang, M.-K. Lee, H.-J. Shin, and J.-L. Lee, “Investigation of oxygen incorporation in AlGaN/GaN heterostructures,” physica status solidi (c) n/a, 2456-2459 (2003).
