For the development of III-nitride-semiconductor-based monolithic micro-light-emitting diode (LED) displays, Eu,O-codoped GaN (GaN:Eu,O) is a promising material candidate for the red LEDs. The luminescence efficiency of Eu-related emission strongly depends on the local atomic structure of Eu ions. Our previous research has revealed that post-growth thermal annealing is an effective method for reconfiguring luminescent sites, leading to a significant increase in light output. We observed the preferential formation of a site with a peak at ∼2.004 eV by the annealing process. In this study, we demonstrate that it is a previously unidentified independent site (OMVPE-X) using combined excitation–emission spectroscopy and time-resolved photoluminescence measurements. In addition, we perform excitation power-dependent photoluminescence measurements and show that this OMVPE-X site dominates the emission at a low excitation power region despite its small relative abundance, suggesting a high excitation efficiency. Most importantly, applying our annealing technique to an LED exhibits a reasonably increased electroluminescence intensity associated with OMVPE-X, confirming that this site has a high excitation efficiency also under current injection. These results demonstrate the importance of OMVPE-X as a notable luminescent site for brighter and more efficient GaN:Eu,O-based LEDs.

Since the creation of high-efficient blue and green light-emitting diodes (LEDs) based on III-nitride semiconductor material systems, III-nitride-based high-efficient red light-emitting material has been in an increasing demand for the monolithic integration of three primary colors LEDs. The monolithic integration technique is attractive for micro-LED (μ-LED) displays, which are crucial for next-generation smartwatches, smartphones, and augmented and virtual reality (AR and VR) devices, including smart glasses.1 Although the luminous efficiency of InxGa1−xN/GaN-based red LEDs has been improving recently,2–5 it is still limited due to the low crystalline and strong internal electric fields in the well layers.6–8 In addition, InxGa1−xN/GaN-based red LEDs suffer from a strong quantum-confined Stark effect, which results in a significant wavelength shift with changing injection current levels.9 Eu,O-codoped GaN (GaN:Eu,O) is one of the leading candidates as an alternative approach to address these issues because it exhibits efficient and frequency stable red luminescence at λ ∼ 620 nm.10–14 The emission originates from the intra-4f electron levels shielded by filled outer shells, thus the Eu3+ emission is highly stable against current injection levels, as well as temperature. Moreover, GaN:Eu,O shows a high luminous efficiency even with a reduced device size, which is useful for μ-LED applications.15 This nature stems from the large carrier capture cross section of Eu3+ ions doped in GaN, which can suppress the effect of non-radiative surface recombination. We have fabricated GaN:Eu,O-based LEDs using an organometallic vapor phase epitaxy (OMVPE) method and demonstrated a light output of 1.25 mW at 20 mA operation with a maximum external quantum efficiency (EQE) of 9.2%.16,17 However, the quantum efficiency is still much lower as compared to InxGa1−xN/GaN-based blue and green LEDs;18,19 therefore, it is highly desirable to further improve the luminous efficiency for the future implementation of III-nitride-semiconductor-based monolithic μ-LED displays. In particular, μ-LED displays are expected to be used in the low current injection density regime (<∼1 A/cm2), and the quantum efficiency under such conditions is critical.

The radiative lifetime of Eu3+ ions does not significantly depend on the kind of luminescent site; therefore, the luminescence efficiency is greatly determined by carrier-mediated excitation efficiency, including free carrier capture cross section and energy transfer rate to Eu3+ ions.20,21 GaN:Eu,O grown using the OMVPE method has several luminescent sites with different local atomic structures around the Eu3+ ions. In particular, OMVPE7 (Eu2) and OMVPE8 (Eu2*), which are assumed to have an oxygen atom in the vicinity of the Eu3+ ion, play a key role in enhancing the light output of LEDs due to their much higher excitation efficiencies as compared to other sites.22 OMVPE7 and OMVPE8 are believed to be charged states of the same local atomic structure.23 OMVPE8, which is considered to have an additional electron compared to OMVPE7, shows low luminescence intensity at a cryogenic temperature; however, the intensity increases with elevating temperature.24 This behavior is considered to be caused by a change in the charge state of intrinsic defects in GaN (e.g., VN, VGa, and their complexes) with increasing temperature, and a part of OMVPE7 is converted into OMVPE8.23,25,26 Time-resolved photoluminescence (TR-PL) has revealed that OMVPE8 shows “afterglow” behavior, where the emission intensity from OMVPE8 increases after the excitation laser pulse has terminated.24 

Recently, we have shown that post-growth thermal annealing at a high temperature (>1000 °C) reconstructs luminescent sites and converts OMVPE1 and OMVPE2 with small excitation efficiency into OMVPE7 and OMVPE8 with higher excitation efficiency, leading to a significant increase in the photoluminescence (PL) intensity.27 In addition, it was found that a luminescent site with a peak at ∼2.004 eV is also preferentially formed by annealing (by a factor of ∼20 after annealing at 1100 °C for 10 min). This peak has been tentatively assigned as OMVPE8 so far.23,27 However, in this study, we determined that this emission peak is caused by an independent site, which we labeled as OMVPE-X. PL measurements have suggested that OMVPE-X is an efficiently excited luminescent site; thus, further investigation of the optical properties is important to realize highly efficient μ-LEDs. Several peaks have also been observed in Eu-implanted GaN in a similar emission energy region to OMVPE-X.28 The emission energies of GaN:Eu fabricated by different methods slightly vary; therefore, a more careful analysis, such as the combined excitation–emission spectroscopy (CEES) measurements,29 is needed to identify the sites; however, it is possible that this site also can be formed when other growth methods are used.

To elucidate the nature of the enhanced luminescence peak, we first perform CEES and time-resolved and TR-PL to clearly identify each luminescent site. Then, we investigate the detailed optical properties of OMVPE-X using temperature and excitation power-dependent PL measurements. Furthermore, we apply the annealing technique to LED fabrication and study the optical characteristics of OMVPE-X under current injection.

The sample investigated in this study for PL measurements was grown using an OMVPE method on a (0001) sapphire substrate. Trimethylgallium (TMG), trimethylindium (TMI), trimethylaluminum (TMA), and ammonia (NH3) were used as precursors of Ga, In, Al, and N, respectively. For the GaN:Eu,O layers, bis(normal-propyl-tetramethylcyclopentadienyl)europium (EuCppm2) and Ar-diluted oxygen were used as the Eu and O sources. The growth was initiated with a low-temperature GaN buffer layer, followed by a 2-μm-thick undoped GaN layer, a 600-nm-thick Al0.19In0.81N cladding layer, 220-nm-thick GaN:Eu,O layer, and a 10-nm-thick GaN capping layer. The growth temperature of the GaN:Eu,O layer was 960 °C. The sample was thermally annealed at 1100 °C in N2 + NH3 atmosphere at 100 kPa for 10 min with a 50-nm-thick SiO2 cap. The sample is the one studied in Ref. 27.

To begin with, we performed CEES measurement in the excitation region of phonon-assisted absorption at 10 and 150 K.29,30 The CEES technique allows us to selectively excite specific luminescence sites by resonantly exciting them with a wavelength-tunable laser. For the CEES measurements, a wavelength-tunable dye laser using Rhodamine 590 Chloride, with an output power of ∼400 mW, was utilized.

As shown in the CEES map measured at 10 K [Fig. 1(a)], each luminescent site can be clearly observed. The resonant excitation energies of OMVPE8 and OMVPE-X are almost identical, which is why OMVPE-X has been assigned initially to be a part of OMVPE8.23,27 For the CEES map measured at 150 K [Fig. 1(b)], emission from OMVPE8 is observed at the resonant excitation energies of OMVPE7 (∼2.1735, 2.1751 eV) as previously reported.23 However, at these excitation energies, OMVPE-X is not excited, suggesting that OMVPE-X is a different luminescent site from OMVPE8.

FIG. 1.

CEES map measured at (a) 10 K and (b) 150 K. Right: the enlarged CEES map.

FIG. 1.

CEES map measured at (a) 10 K and (b) 150 K. Right: the enlarged CEES map.

Close modal

Subsequently, TR-PL was performed at 180 K to investigate the detailed optical properties of OMVPE8 and OMVPE-X. We used a He–Cd laser (λ = 325 nm) pulsed by an acousto-optic modulator. Luminescence was detected by a streak camera equipped with a 30-cm spectrometer. Figure 2 shows the TR-PL signal for OMVPE8 and OMVPE-X measured at 180 K. OMVPE8 shows afterglow behavior.24 Contrary to OMVPE8, emission from OMVPE-X monotonically decreases after the laser pulse with a single-exponential function, confirming that OMVPE-X is not a part of OMVPE8.

FIG. 2.

TR-PL profile for OMVPE8 and OMVPE-X measured at 180 K.

FIG. 2.

TR-PL profile for OMVPE8 and OMVPE-X measured at 180 K.

Close modal

Now that OMVPE-X has been identified as a distinct luminescent site, we investigate the optical properties in detail. First, we conducted temperature-dependent PL using a He–Cd laser. Figure 3 shows PL intensity at the emission peak energy for each luminescent site. As it has been reported, PL intensity from OMVPE8 rises with increasing temperature and reaches a maximum intensity at ∼150 K.

FIG. 3.

Temperature-dependent integrated PL intensity ranging from 10 K to room temperature.

FIG. 3.

Temperature-dependent integrated PL intensity ranging from 10 K to room temperature.

Close modal

OMVPE-X also exhibited an increase in PL intensity with the rise in temperature (<∼150 K), which indicates that the local atomic structure of OMVPE-X stabilizes with increasing temperature. Due to its thermally activated characteristics, the thermal quenching ratio of the integrated PL intensity of OMVPE-X is smaller compared to other luminescent sites. Similar “anti-thermal-quenching” behavior has been observed for Eu-implanted Mg-codoped GaN, though a hysteric behavior was not observed.31–33 The authors have suggested the existence of two unique metastable states with different acceptor states. These results imply the presence of a metastable state related to OMVPE-X as well as OMVPE8. In the case of OMVPE8, the metastable state can be OMVPE7.

Then, we performed excitation power density-dependent PL at room temperature to evaluate the luminescence properties of OMVPE-X. As shown in Fig. 4, OMVPE-X shows a predominant emission at a low excitation power region.

FIG. 4.

Normalized excitation power density-dependent PL spectra.

FIG. 4.

Normalized excitation power density-dependent PL spectra.

Close modal
To discuss the relative luminescence efficiency of OMVPE-X, we roughly estimated the existing ratio of OMVPE-X from the CEES map measured at 10 K in the excitation region of zero-phonon-assisted region27 using the following equation:34,
NEuiσiΦτradi+1σiϕIi,
(1)
where NEui is the number of a luminescent site i, σi is the excitation cross section under resonant excitation, Φ is the photon flux, τradi is the radiative lifetime, and Ii is the PL intensity. For calculation, we used reported values of σi and τradi for OMVPE1-8.34 For OMVPE-X and other unidentified sites, we utilized the average values of OMVPE1-8. As a result, the existing ratio of OMVPE-X was roughly estimated to be ∼3%. The fact that OMVPE-X shows a predominant emission despite its small existing ratio suggests its quite high excitation efficiency.

It is known that the luminescence efficiency of GaN:Eu,O is typically low under low excitation power conditions because free carriers are captured by efficient non-radiative traps with large carrier capture cross section.22,27 Hence, the creation of Eu3+ luminescent site with high excitation efficiencies is highly important to compete with non-radiative traps in the carrier capture process and to enhance the luminescence efficiency at low current injection density region, where μ-LED displays are commonly utilized.

The approximate concentration of OMVPE-X for the as-grown state was estimated to be ∼1 × 10−1%, indicating that the concentration was greatly increased by annealing at 1100 °C. Thus, it is expected that the relative abundance of OMVPE-X can be further increased through future optimizations of growth and annealing condition, leading to higher-efficient GaN:Eu,O luminescence even under high excitation power condition. Despite being a minority site, OMVPE-X showed a predominant PL emission. It would be challenging to make OMVPE-X a majority site, however, a small increase in the concentration of OMVPE-X is likely to lead a large increase in PL intensity due to the high excitation efficiency.

Finally, we fabricated GaN:Eu,O-based LED with annealing treatment to investigate the optical properties of OMVPE-X under current injection. To suppress damage introduction to p-type GaN, annealing was performed after the growth of the GaN:Eu,O active layer. A low-temperature GaN buffer layer was grown on a (0001) sapphire substrate followed by an undoped GaN layer and an n-type GaN layer. A 280-nm-thick GaN:Eu,O layer was grown at 960 °C, then annealed at 1100 °C for 20 min in the reactor. The in situ reflectivity monitoring system revealed that ∼90 nm of the active layer was thermally decomposed during the annealing. Afterward, a p-GaN and p+-GaN layer was grown. As the reference, we prepared a conventional LED without annealing treatment with a 190-nm-thick GaN:Eu,O active layer. To control the n-type and p-type conductivity, we used monomethyl silane (CH3SiH3) and bis-cyclopentadienyl magnesium (Cp2Mg) to dope Si and Mg, respectively.

Normalized EL spectra are displayed in Fig. 5(a). Similar to the PL spectra, OMVPE-X exhibits a dominant emission, especially at low injection current density region for the annealed LED. This indicates that OMVPE-X is efficiently excited even under current injection, which is an essential characteristic for the implementation of μ-LED displays. The conventional LED (w/o anneal) shows a much smaller EL intensity related to OMVPE-X even under low current density region, implying the relatively low abundance of OMVPE-X. External quantum efficiency (EQE) of LEDs is displayed in Fig. 5(b). The annealed LED exhibits higher EQE due to the formation of luminescent sites with high excitation efficiencies, including OMVPE-X.

FIG. 5.

(a) Normalized injection current density-dependent EL spectra for the annealed LED. The black line shows the EL spectrum of the conventional LED as the reference. (b) EQE as a function of injection current density for the annealed LED and the conventional LED.

FIG. 5.

(a) Normalized injection current density-dependent EL spectra for the annealed LED. The black line shows the EL spectrum of the conventional LED as the reference. (b) EQE as a function of injection current density for the annealed LED and the conventional LED.

Close modal

The EQE demonstrated in this research is smaller than previously reported values.17 Optimization of LED structure (e.g., introducing electron blocking layer, increasing the thickness of the active layer) is important to improve the output power. Especially, multilayer structure is crucial to be employed because it greatly improves the luminescence efficiency of GaN:Eu,O-based LEDs.17,35 In terms of the annealing process, annealing time, temperature, atmosphere, and pressure should be optimized to suppress thermal decomposition and Eu desorption during the process.

These results clearly indicate the importance of increasing the amount of OMVPE-X to achieve even brighter and efficient GaN:Eu,O-based LEDs in the future. For this purpose, it is crucial to elucidate the atomic structure of OMVPE-X. OMVPE7 and OMVPE8, which have high excitation efficiencies, are considered to have an oxygen atom around Eu3+. Oxygen works as a donor-like defect in GaN:Eu,O and is believed to pair with VGa acceptor-like defect, thereby enhancing the efficiency of carrier capture and energy transfer to Eu3+ ions.36 Therefore, certain impurities such as oxygen may be associated with OMVPE-X. In order to provide hints for the identification of the atomic structure of OMVPE-X, we prepared two samples with different oxygen concentrations (see the supplementary material for details). The oxygen concentration was controlled by the flow rate of Ar-diluted oxygen. After annealing at 1200 °C, the sample grown with an increased flow rate (136 µmol/min) exhibited larger PL intensity associated with OMVPE-X as compared to the sample grown with a conventional flow rate (68 µmol/min). This result suggests that oxygen plays a key role in the formation of OMVPE-X. This is important knowledge to selectively form OMVPE-X in the future for the implementation of GaN:Eu,O-based μ-LED displays.

In order to determine the local atomic structure of OMVPE-X, first-principles calculations should be employed37,38 with sufficient supporting experimental data. Recently, Yamaga et al. have demonstrated crystal-field analysis of Eu3+ for Mg-codoped GaN:Eu grown by molecular beam epitaxy method.39 They have determined the local symmetry of Eu3+ for some specific luminescent sites. However, GaN:Eu,O grown by the OMVPE method has more kinds of luminescent sites and the spectrum is highly complicated, making accurate evaluation difficult using this method. We have reported polarization-dependent CEES measurements to study the local symmetry of sites and found that OMVPE-X has a similar symmetry as OMVPE7 and OMVPE8, which have the highest degree of C3v-like symmetry.40 However, OMVPE-X is likely to have several point defects in the vicinity including an oxygen atom; thus, further experimental and theoretical investigations are required to determine the local atomic structure of OMVPE-X.

In conclusion, we have revealed that the luminescent site with a peak at ∼2.004 eV (referred to as OMVPE-X), which can be preferentially formed by annealing, is an independent site using CEES and TR-PL measurements. OMVPE-X exhibits a thermally activated characteristic, which is likely due to the stabilization of the local atomic structure of OMVPE-X with increasing temperature. OMVPE-X shows a predominant emission under both photoexcitation and current injection when the free carrier density is small (weak excitation region). These results demonstrate that OMVPE-X has a high excitation efficiency and is a noteworthy luminescent site for even brighter and efficient GaN:Eu,O-based μ-LEDs.

See the supplementary material for the role of oxygen in the formation of OMVPE-X.

This work was partially supported by JSPS KAKENHI (Grant Nos. 18H05212, 22K14614, and 23H05449). This was also partially supported by the NSF (Award No. 2129183).

The authors have no conflicts to disclose.

Takenori Iwaya: Conceptualization (lead); Data curation (lead); Investigation (lead); Methodology (equal); Validation (equal); Visualization (lead); Writing – original draft (lead); Writing – review & editing (lead). Shuhei Ichikawa: Investigation (equal); Resources (supporting); Supervision (equal); Validation (equal); Visualization (equal); Writing – review & editing (equal). Volkmar Dierolf: Methodology (equal); Resources (supporting); Supervision (equal); Writing – review & editing (equal). Brandon Mitchell: Investigation (equal); Resources (supporting); Writing – review & editing (equal). Hayley Austin: Investigation (supporting); Methodology (equal). Dolf Timmerman: Investigation (equal). Jun Tatebayashi: Supervision (supporting); Writing – review & editing (supporting). Yasufumi Fujiwara: Methodology (equal); Resources (lead); Supervision (lead); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
Z. Y.
Fan
,
J. Y.
Lin
, and
H. X.
Jiang
,
J. Phys. D: Appl. Phys.
41
,
094001
(
2008
).
2.
D.
Iida
,
K.
Niwa
,
S.
Kamiyama
, and
K.
Ohkawa
,
Appl. Phys. Express
9
,
111003
(
2016
).
3.
Z.
Zhuang
,
D.
Iida
, and
K.
Ohkawa
,
Appl. Phys. Lett.
116
,
173501
(
2020
).
4.
D.
Iida
,
Z.
Zhuang
,
P.
Kirilenko
,
M.
Velazquez-Rizo
,
M. A.
Najmi
, and
K.
Ohkawa
,
Appl. Phys. Lett.
116
,
162101
(
2020
).
5.
D.
Iida
,
Z.
Zhuang
,
P.
Kirilenko
,
M.
Velazquez-Rizo
, and
K.
Ohkawa
,
Appl. Phys. Express
13
,
031001
(
2020
).
6.
M. B.
Nardelli
,
K.
Rapcewicz
, and
J.
Bernholc
,
Appl. Phys. Lett.
71
,
3135
(
1997
).
7.
R.
Singh
,
D.
Doppalapudi
,
T. D.
Moustakas
, and
L. T.
Romano
,
Appl. Phys. Lett.
70
,
1089
(
1997
).
8.
M.
Auf der Maur
,
A.
Pecchia
,
G.
Penazzi
,
W.
Rodrigues
, and
A.
Di Carlo
,
Phys. Rev. Lett.
116
,
027401
(
2016
).
9.
T.
Takeuchi
,
S.
Sota
,
M.
Katsuragawa
,
M.
Komori
,
H.
Takeuchi
,
H.
Amano
, and
I.
Akasaki
,
Jpn. J. Appl. Phys.
36
,
L382
(
1997
).
10.
J.
Heikenfeld
,
M.
Garter
,
D. S.
Lee
,
R.
Birkhahn
, and
A. J.
Steckl
,
Appl. Phys. Lett.
75
,
1189
(
1999
).
11.
S.
Morishima
,
T.
Maruyama
,
M.
Tanaka
, and
K.
Akimoto
,
Phys. Status Solidi A
176
,
113
(
1999
).
12.
H. J.
Lozykowski
,
W. M.
Jadwisienczak
,
J.
Han
, and
I. G.
Brown
,
Appl. Phys. Lett.
77
,
767
(
2000
).
13.
J. H.
Park
and
A. J.
Steckl
,
Appl. Phys. Lett.
85
,
4588
(
2004
).
14.
J. H.
Park
and
A. J.
Steckl
,
J. Appl. Phys.
98
,
056108
(
2005
).
15.
D.
Denier van der Gon
,
D.
Timmerman
,
Y.
Matsude
,
S.
Ichikawa
,
M.
Ashida
,
P.
Schall
, and
Y.
Fujiwara
,
Opt. Lett.
45
,
3973
(
2020
).
16.
A.
Nishikawa
,
T.
Kawasaki
,
N.
Furukawa
,
Y.
Terai
, and
Y.
Fujiwara
,
Appl. Phys. Express
2
,
071004
(
2009
).
17.
B.
Mitchell
,
V.
Dierolf
,
T.
Gregorkiewicz
, and
Y.
Fujiwara
,
J. Appl. Phys.
123
,
160901
(
2018
).
18.
Y.
Narukawa
,
M.
Ichikawa
,
D.
Sanga
,
M.
Sano
, and
T.
Mukai
,
J. Phys. D: Appl. Phys.
43
,
354002
(
2010
).
19.
S.
Saito
,
R.
Hashimoto
,
J.
Hwang
, and
S.
Nunoue
,
Appl. Phys. Express
6
,
111004
(
2013
).
20.
H.
Sekiguchi
,
Y.
Takagi
,
T.
Otani
,
H.
Okada
, and
A.
Wakahara
,
J. Appl. Phys.
113
,
013105
(
2013
).
21.
J. K.
Mishra
,
T.
Langer
,
U.
Rossow
,
S.
Shvarkov
,
A.
Wieck
, and
A.
Hangleiter
,
Appl. Phys. Lett.
102
,
061115
(
2013
).
22.
D.
Timmerman
,
B.
Mitchell
,
S.
Ichikawa
,
J.
Tatebayashi
,
M.
Ashida
, and
Y.
Fujiwara
,
Phys. Rev. Appl.
13
,
014044
(
2020
).
23.
B.
Mitchell
,
N.
Hernandez
,
D.
Lee
,
A.
Koizumi
,
Y.
Fujiwara
, and
V.
Dierolf
,
Phys. Rev. B
96
,
064308
(
2017
).
24.
R.
Wakamatsu
,
D.
Timmerman
,
D.
Lee
,
A.
Koizumi
, and
Y.
Fujiwara
,
J. Appl. Phys.
116
,
043515
(
2014
).
25.
C. G.
van de Walle
and
J.
Neugebauer
,
J. Appl. Phys.
95
,
3851
(
2004
).
26.
I. C.
Diallo
and
D. O.
Demchenko
,
Phys. Rev. Appl.
6
,
064002
(
2016
).
27.
T.
Iwaya
,
S.
Ichikawa
,
D.
Timmerman
,
J.
Tatebayashi
, and
Y.
Fujiwara
,
Appl. Phys. Lett.
122
,
032102
(
2023
).
28.
V.
Katchkanov
,
K. P.
O’Donnell
,
S.
Dalmasso
,
R. W.
Martin
,
A.
Braud
,
Y.
Nakanishi
,
A.
Wakahara
, and
A.
Yoshida
,
Phys. Status Solidi B
242
,
1491
(
2005
).
29.
V.
Dierolf
,
C.
Sandmann
,
J.
Zavada
,
P.
Chow
, and
B.
Hertog
,
J. Appl. Phys.
95
,
5464
(
2004
).
30.
N.
Woodward
,
A.
Nishikawa
,
Y.
Fujiwara
, and
V.
Dierolf
,
Opt. Mater.
33
,
1050
(
2011
).
31.
K. P.
O’Donnell
,
P. R.
Edwards
,
M. J.
Kappers
,
K.
Lorenz
,
E.
Alves
, and
M.
Boćkowski
,
Phys. Status Solidi C
11
,
662
(
2014
).
32.
A. K.
Singh
,
K. P.
O’Donnell
,
P. R.
Edwards
,
K.
Lorenz
,
M. J.
Kappers
, and
M.
Boćkowski
,
Sci. Rep.
7
,
41982
(
2017
).
33.
D.
Cameron
,
K. P.
O’Donnell
,
P. R.
Edwards
,
M.
Peres
,
K.
Lorenz
,
M. J.
Kappers
, and
M.
Boćkowski
,
Appl. Phys. Lett.
116
,
102105
(
2020
).
34.
R.
Wakamatsu
,
D. G.
Lee
,
A.
Koizumi
,
V.
Dierolf
, and
Y.
Fujiwara
,
J. Appl. Phys.
114
,
043501
(
2013
).
35.
W.
Zhu
,
B.
Mitchell
,
D.
Timmerman
,
A.
Koizumi
,
T.
Gregorkiewicz
, and
Y.
Fujiwara
,
MRS Adv.
2
,
159
(
2017
).
36.
B.
Mitchell
,
J.
Poplawsky
,
D.
Lee
,
A.
Koizumi
,
Y.
Fujiwara
, and
V.
Dierolf
,
J. Appl. Phys.
115
,
204501
(
2014
).
37.
K.
Hoang
,
Phys. Status Solidi RRL
10
,
915
(
2016
).
39.
M.
Yamaga
,
K. P.
O’Donnell
,
H.
Sekiguchi
, and
A.
Wakahara
,
J. Lumin.
262
,
119953
(
2023
).
40.
S.
Copelman
,
H.
Austin
,
D.
Timmerman
,
J. D.
Poplawsky
,
M.
Waite
,
J.
Tatebayashi
,
S.
Ichikawa
,
Y.
Fujiwara
,
V.
Dierolf
, and
B. J.
Mitchell
,
Proc. SPIE
11302
,
113021Z
(
2020
).

Supplementary Material