2 2 11
( GaN.
Figure 2.3 shows the experimental flow chart of this chapter. By scanning electron
microscope (SEM) and cathodoluminescence (CL), the microstructures and nano-photonics
of samples are investigated. By atomic force microscopy (AFM) measurement, the surface
morphologies of samples are measured. By polarization- and temperature-dependent photoluminescence (PL), the optical properties of the N-polar GaN and semipolar (1122)
GaN samples are investigated. The frequency shift of phonon lines and strain inside the
samples are investigated by Raman scattering measurements. According to the results of the
experiments, we will discuss the anisotropic characteristics of the N-polar GaN and semipolar
) 2 2 11
( GaN grown on r-sapphire.
2.3 Sample Structures and Growth Conditions
Figure 2.4 shows sample structures of (a) N-polar GaN and (b) semipolar (1122) GaN.
Table 2.1 shows growth conditions of (a) N-polar GaN and (b) semipolar (1122) GaN.
N-polar GaN growths was carried out on 2°miscut sapphire substrate in a horizontal
metalorganic chemical vapor deposition reactor. The substrate was 2°miscut towards the
] 0 2 11
[ direction. For N-polar GaN growth, sapphire was heated up in a mixture of NH3 (3
15
SLM) and N2 (4 SLM) to 950 °C for 30 s. A LT-GaN buffer (30 nm thick) was grown on
nitridized sapphire at 600 °C in H2, followed by N-polar GaN growth at 1050 °C and 100
mbar.
Growth of semipolar (1122) GaN on these patterned substrates was performed using
an Aixtron 200/4RF-S horizontal metalorganic chemical vapor deposition (MOCVD) reactor.
Trimethylgallium (TMGa), Trimethylaluminum (TMAl), and ammonia were used as the
precursors for Ga, Al, and N, respectively. The etched r-sapphire was thermally cleaned in H2
ambient at1100 °C, followed by deposition of a 20 nm low-temperature (LT) AlN buffer at
500 °C and 100 mbar. Temperature was then ramped up to anneal the buffer layer, and growth
of GaN commenced at 1030 °C. The design of amultiple step growth procedure was
implemented in stages (1)-(3)：(1) to achieve selective growth on the sapphire sidewall (350
nm GaN, 1030 C, 200mbar, high V/III (low growth rate)), (2) to shape the stripe for optimum
coalescence (2.3um GaN, 1030 C, 200mbar, low V/III), (3) and to obtain a flat surface
morphology (2.3um GaN, 995 C, 300mbar, high V/III). The growth conditions necessary for
each stage are discussed in Reference [1].
2.4 Anisotropic Characteristics of N-polar GaN and Semipolar ( 11 2 2 ) GaN Grown on r-sapphire
2.4.1 Scanning Electron Microscope (SEM) and Cathodoluminescence (CL)
16
Studies of N-polar GaN and Semipolar ( 11 2 2 ) GaN Grown on r-sapphire
In other to investigate the relation between optical property and microstructures, scanning electron microscope (SEM) and cathodoluminescence (CL) measurements wereconducted at the same regions. As shown in Figure 2.5, SEM and CL images were acquired
with a Gatan monoCL3 spectrometer in a JEOL SEM system (model JSM 7000F) under
room temperature. The excitation electron voltage for CL measurement ranges from 5 to 11
kV.
Figure 2.6 (a) and (b) show the SEM images for the N-polar GaN and semipolar (1122)
GaN samples, respectively. In these samples, striation features were observed in the
morphology. The surface roughness of N-polar GaN is flatter than semipolar (1122) GaN
sample. Two region structures are easy to be observed in the semipolar (1122) GaN sample.
The obvious striation feature and roughness were characterized by surface-pits and SFs. In
Figure 2.6 (a) and (b), the N-polar GaN sample has striation in lateral direction, yet the
semipolar (1122) GaN sample has two kinds of striation on the surface. One can observe
the shallow striation in lateral direction of the Figure 2.6 (b). It is suggested the N-polar GaN
sample has a higher defect density than that of the semipolar (1122) GaN sample. In
addition, Figure 2.6 (c) and (d) show the panchromatic CL images for the corresponding
SEM region with 11 kV excitation electron voltages under room temperature. Figure 2.6 (d)
shows the meshed structure on the surface of the semipolar (1122) GaN sample. The
17
semipolar (1122) GaN sample is brighter than the N-polar GaN sample. The result shows
that the density of SF for the N-polar GaN sample is larger than that of the semipolar (1122)
GaN sample.
Figure 2.7 shows the CL spectra taken from the same SEM regions under room
temperature. In the N-polar GaN sample, the weak CL intensities for the emission band reveal
a poor sample quality. Due to the lower SF density and a better sample quality, the semipolar
) 2 2 11
( GaN sample has higher CL intensities than that of the N-polar GaN sample. As a
result, the results are consistent with the brightness of CL images that mentioned above.
2.4.2 Atomic Force Microscopy (AFM) Studies of N-polar GaN and Semipolar ( 11 2 2 ) GaN Grown on r-sapphire
In other to study the surface morphology of the N-polar GaN and semipolar (1122) GaN samples, AFM measurement was conducted. Figure 2.8 shows experimental setup of
AFM measurement. AFM images were acquired with XE-70 and XE control electronics
under non-contact mode.
Figure 2.9 (a) and (b) show AFM images of the N-polar GaN and semipolar (1122)
GaN samples, while Figure 2.9 (c) and (d) show 3D AFM images taken from the same
regions, respectively. Surface roughness (Rq) of the N-polar GaN and semipolar (1122)
GaN samples are 0.5826 and 1.906 nm, respectively. For the semipolar (1122) GaN
18
samples, the higher roughness and higher CL intensity were observed. In general, sample
with high CL intensity usually has lower roughness. It is suggested that the roughness of the
semipolar (1122) GaN sample is contributed from two structures of the semipolar (1122)
GaN sample in Figure 2.6.
2.4.3 Polarization- and Temperature-dependent Photoluminescence (PL) Studies of N-polar GaN and Semipolar ( 11 2 2 ) GaN Grown on r-sapphire
PL is the most fundamental measurement for understanding the optical property of a
material. The experimental setup for PL measurement is shown in Figure 2.10. PL
measurements are carried out with the 325 nm line of a 55 mW He-Cd laser for excitation.
The samples are placed in a cryostat for temperature-dependent PL measurement.
2.4.3.1 Low-temperature Polarized PL and Degree of Polarization (DOP)
In other to investigate the effect of anisotropic strain on the optical properties, the
polarized PL measurement was conducted. Figure 2.11 (a) and (b) show the polarized PL
measurement for the N-polar GaN and semipolar (1122) GaN samples, respectively. The PL
spectra show a high polarization anisotropy. The degree of polarization (DOP), ρ, can be
expressed as
19
where I[1100] and I[1123] are the PL intensities for E//[1100] and E⊥[1123] ,
respectively [16]. Figure 2.12 (a) and (b) show the DOP for the two samples. The DOP of the
semipolar (1122) GaN sample is larger than that of the N-polar GaN sample.
2.4.3.2 Temperature-dependent PL Study
Figure 2.13 (a) and (b) show the PL spectra of the N-polar GaN and semipolar (1122) GaN samples as a function of temperature, respectively. In Figure 2.13 (a), the N-polar GaN
sample has three peaks ~359 nm (3.454 eV), ~382 nm (3.246 eV), and ~391 nm (3.171 eV)
under temperatures 10 ~ 100K. The dominate peak is around ~359 nm (3.454 eV). In Figure
2.13 (b), the dominate peak of semipolar (1122) GaN is 357.8 nm (3.466 eV). The emission
peak around ~358 nm is neutral acceptors (A0BE) recombination [18], while the emission
peaks around ~382 nm and ~391 nm is the relative intensities of the DAP recombination [19].
In the N-polar GaN sample, PL intensity of neutral acceptors (A0BE) recombination decays
quicker than that of DAP recombination with increasing temperature.
The PL peak positions as a function of temperature for the two samples are shown in
Figure 2.14. The peak positions of the N-polar GaN sample are distinguished into two parts.
One spectral range is from 335 nm to 370 nm, and the other is from 370 nm to 460 nm. The
N-polar GaN and semipolar (1122) GaN samples are 359 and 357.8 nm when the peak
position is at 10K. With increasing temperature from 10 to 300 K, both PL peak positions of
20
the first part (wavelength 335 nm ~ 370 nm) of the N-polar GaN and semipolar (1122) GaN
samples are red-shifted, while that of the second part (370 nm ~ 460 nm) of the N-polar GaN
sample does not change greatly.
2.4.3.3 Temperature-dependent Polarized PL Study
Figure 2.15 (a) and (b) show the PL spectra with polarization degree set at the largest
intensity φPLmax for the N-polar GaN and semipolar (1122) GaN samples as a function of
temperature, respectively. Figure 2.16 (a) and (b) show the PL spectra with polarization
degree set at the smallest intensity φPLmin for the N-polar GaN and semipolar (1122) GaN samples as a function of temperature, respectively. Figure 2.17 (a) and (b) show the
normalized PL integral intensities without polarization and with polarization degrees set at
PLmax
φ and φPLmin for the N-polar GaN and semipolar (1122) GaN samples as a function of
temperature, respectively. As the temperature increases, the integral PL intensity of normal,
MAX (φPLmax), and MIN (φPLmin) decrease. For the semipolar (1122) GaN sample, integral PL intensity of φ is larger than that of PLmin φPLmax and the integral PL intensity of φPLmax is
stronger than that of Normal.
2.4.4 Estimation of Strain by Raman Scattering Measurement
As shown in Figure 2.18, Raman spectra were recorded in the backscattering
21
configuration using a Jobin Yvon-Horiba Micro-Raman system (model T64000) under a 532
nm excitation laser. A polarizer was set between the laser and the measured sample to
polarize the laser light. Raman scattering can measure the frequency shift of phonons of the N-polar GaN and semipolar (1122) GaN sample. The strain distribution in (1122) GaN can
be estimated by Raman scattering.
In Figure 2.19 (a) and (b), the Cartesian axesx′, y′, and z′ correspond to the [1100],
] 23 1 1
[ , and [1122] directions of GaN, respectively [10]. The z′(x′x′)z' and z′(y′y′)z'
called Porto’s notation. It was used to describe the scattering geometry：the letters outside the
parentheses represent the direction of the incident and scattered light, while the letters inside
the parentheses represent the incident and scattered polarizations [20]
Figure 2.19 shows the Raman scattering spectra for the two samples. The spectra displays A1(TO) and E2(high) mode for GaN. The solid lines at 531.8 cm-1 and 568.0 cm-1
show the strain-free GaN of A1(TO) and E2(high) mode, respectively. Compressive stress in
the GaN thin film tends to shift Raman peak to a larger wave number, while tensile stress
moves the peak to an opposite direction. Based on the deformation potential approximation, the frequency shift, ∆ω =ω−ω0, can be defined as a function of the strain, where ω0
represents the phonon frequency of strain-free. The ∆ω and in-plane strain can be
influenced by the growth parameters. Table 2.2 shows ∆ω of A1(TO) and E2(high) mode
for the N-polar GaN and semipolar (1122) GaN samples.
22
2.5 Discussion and Summary
In summary, we have shown the experimental results of SEM, CL, AFM, PL, and
Raman measurements of the N-polar GaN and semipolar (1122) GaN samples. The surface
roughness of the semipolar (1122) GaN sample is larger than that of the N-polar GaN
sample. The higher surface roughness of the semipolar (1122) GaN sample was attributed
the meshed structure on the surface. The obvious striation feature and roughness were
characterized by of SF and TDs. In other words, the defect density of the N-polar GaN
sample is higher than that of the semipolar (1122) GaN sample. The result of CL intensity is
consistent of that of the AFM measurement. The semipolar (1122) GaN sample reveals the
higher CL intensity and the better crystal quality. For the N-polar GaN sample, the intensity
of the neutral acceptors (A0BE) recombination decays quickly than that of the DAP
recombination from 10 K to 300 K.
23
References
1. B. Leung, Q. Sun, C. Yerino, Y. Zhang, Jung Han, B. H. Kong, H. K. Cho, K. Y. Liao, Y.
L. Li, J. Cryst. Growth 341 (2012) 27-33.
2. H. Amano, N. Sawaki, I. Akasaki, Y. Toyoda, Appl. Phys. Lett. 48, 353 (1986).
3. I. Akasaki, H. Amano, Y. Koide, K. Hiramatsu, N. Sawaki, J. Cryst. Growth 98, 209
(1989).
4. J.N. Kuznia, M. Asif-Khan, D.T. Olson, R. Kaplan, J. Freitas, J. Appl. Phys. 73, 4700
(1993)
5. L. Sugiura, K. Itaya, J. Nishio, H. Fujimoto, Y. Kokubun, J. Appl. Phys. 82, 4877 (1997).
6. K. Hiramatsu, S. Itho, H. Amano, et al., J. Cryst. Growth 115, 628 (1991).
7. X.H. Wu, D. Kapolnek, E.J. Tarsa, et al., Appl. Phys. Lett. 68, 1371 (1996).
8. Nichia Chemical Industries, Ltd., 491 Oka, Kaminaka, Anan, Tokushima 774 Jpn. J. Appl.
Phys. 30, 1705 (1991)
9. Qian Sun, Benjamin Leung, Christopher D. Yerino, Yu Zhang, and Jung Han, Appl. Phys.
Lett. 95, 231904 (2009).
10. M. Funato, D. Inoue, M. Ueda, Y. Kawakami, Y. Narukawa, and T. Mukai, J. Appl. Phys.
107, 123501 (2010).
11. Benjamin Leung, Yu Zhang, Christopher D Yerino, Jung Han, Qian Sun, Zhen Chen,
Steve Lester, Kuan-Yung Liao and Yun-Li Li, Semicond. Sci. Technol. 27, 024016
24
(2012).
12. Sun Q, Yerino C D, Ko T-S, Cho Y S, Lee I-H, Han J and Coltrin M E, J. Appl. Phys.
104, 093523 (2008).
13. A. Kahouli, N. Kriouche, J. Brault, B. Damilano, P. Venne´gue`s, P. de Mierry, M.
Leroux, A. Courville, O. Tottereau, and J. Massies, J. Appl. Phys.110, 084318 (2011)
14. T. Wernicke, S.Ploch, V. Hoffmann, A. Knauer, M. Weyers, and M. Kneissl, Phys. Status
Solidi B. 248, 574 (2011).
15. Q. Sun, B. H. Kong, C. D. Yerino, T. S. Ko, B. Leung, H. K. Cho, and J. Han, J. Appl.
Phys. 106, 123519 (2009).
16. K. Hiramatsu, K. Nishiyama, A. Motogaito, H. Miyake, Y. Iyechika, and T. Maeda, Phys.
Status Solidi A 176, 535 (1999)
17. M. Ueda, K. Kojima, M. Funato, and Y. Kawakami, Appl. Phys. Lett. 89, 211907 (2007).
18. V.Kirikyuk, A.R.A. Zauner, P.C.M. Christianen, J. L. Weyher, P.R. Hageman, P.K.
Larsen, J. Cryst. Growth 230 (2001) 477-480.
19. A. Y. Polyakov, N. B. Smirnov, A. V. Govorkov, Q. Sun, Y. Zhang, Y. S. Cho, I.-H. Lee,
J. Han, Materials Science and Engineering B 166 (2010) 83–88
20. Bottcher T, Einfeldt S, Figge S, Chierchia R, Heinke H, Hommel D and Speck J S 2001
Appl. Phys. Lett. 78 (1976).
25
Figure 2.1 LT-AlN growth on sapphire (a) inclined SEM and (b) AFM imaging of top surface of r-plane sapphire (1 μm×1 μm). Selectivity of GaN on the AlN buffer is shown by (c) inclined and (d) top view SEM images [1].
Figure 2.2 Controlled growth evolution of semipolar GaN viewed by growth interruptions and cross-sectional SEM imaging. Under different growth conditions (a), (b), and (c) as labeled, Column (i) first step condition, Column (ii) second step growth conditions are changed to control relative facets growth rates. Column (iii) growth is continued under the same conditions until before coalescence [1].
26
Figure 2.3 Experimental flow chart of this chapter.
N-polar GaN and Semipolar (1122)
GaN grown on r-sapphire by MOCVD
Scanning Electron Microscope (SEM) and Cathodoluminescence
(CL) studies
Atomic Force Microscopy (AFM) Measurements
To investigate microstructures and nanophotonics of samples
To measure the surface morphology of samples
To investigate optical properties of samples
To investigate the frequency shifts of phonon lines and in-plane
strains
To discuss the anisotropic characteristics of N-polar GaN and Semipolar (1122) GaN grown
on r-sapphire
27
(a) (b)
Figure 2.4 Sample structures of (a) N-polar GaN and (b) semipolar (1122) GaN.
N-polar Semipolar
GaN, 995oC, 300 mbar, 2.3 μm, high V/III GaN, 1030oC, 200 mbar, 2.3 μm, low V/III
GaN GaN, 1050oC, 100 mbar, low V/III, 1 μm
GaN, 1030oC, 200 mbar, 350 nm, high V/III
buffer
layer LT-GaN, 600oC, 30 nm LT-AlN, 500 oC, 100 mbar, 20 nm substrate 2°miscut sapphire r-sapphire
Table 2.1 Growth conditions of (a) N-polar GaN and (b) semipolar (1122) GaN.
28
(a)
(b)
Figure 2.5 (a) SEM and (b) CL experimental setups in a JEOL SEM system (model JSM 7000F).
29
Figure 2.6 SEM images of the (a) N-polar GaN and (b) semipolar (1122) GaN samples and the panchromatic CL images (c) and (d) taken over the same regions with 11kV excitation electron voltage under room temperature, respectively.
0 5000 10000 15000 20000
300 350 400 450 500 550 600
0 the excitations of 5, 7, 9, and 11kV electron voltages under room temperature.
30
Figure 2.8 Experimental setup of AFM measurement.
Figure 2.9 AFM images of the (a) N-polar GaN (Rq：0.5826nm) and (b) semipolar (1122) GaN (Rq：1.906nm) samples, and 3D AFM images (c) and (d) taken from the same regions, respectively. Surface roughness of each sample, Rq, is shown in the parentheses.
31
Figure 2.10 Experimental setup of PL measurement.
0.0 0.1 0.2 0.3 0.4
350 400 450 500 550
0.0
Figure 2.11 Polarization-dependent PL at 10K for the (a) N-polar GaN and (b) semipolar )
2 2 11
( GaN samples.
32
350 400 450 500 550
0.0 degree of polarization for the two samples is also shown.
33
350 400 450 500 550
0.0
Figure 2.13 PL spectra as a function of temperature for the (a) N-polar GaN and (b) semipolar )
Figure 2.14 PL peak position as a function of temperature for the N-polar GaN (black and red) and semipolar (1122) GaN (green) samples.
34
0.0 0.1 0.2 0.3
350 400 450 500 550
0.0
Figure 2.15 Temperature-dependent PL with polarization degree set at φPLmax for the (a) N-polar GaN and (b) semipolar (1122) GaN samples.
0.0 0.1 0.2
350 400 450 500 550
0.0
Figure 2.16 Temperature-dependent PL with polarization degree set at φPLmin for the (a) N-polar GaN and (b) semipolar (1122) GaN samples.
35
Figure 2.17 The normalized PL integral intensities without polarization and with polarization degrees set at φPLmax and φPLmin for the (a)N-polar GaN and (b) semipolar
) 2 2 11
( GaN sample as a function of temperature.
36
Figure 2.18 Experimental setup of Jobin Yvon-Horiba Micro-Raman system (model T64000).
37
500 525 550 575 600 625 650 0
38
Chapter 3 Carrier transport study of Ga-polar, N-polar, and semipolar InGaN/GaN multiple quantum well LEDs by using quantum efficiency and time-resolved electroluminescence measurements
3.1 Introductions
III-Nitride compounds, such as InGaN and AlGaN, catch much attention. They cover
the spectral range from ultraviolet (UV) to infrared (IR). In most light emitting diodes (LEDs)
and laser diodes in the visible range, InGaN/GaN quantum wells (QWs) structures are used
as the active layers. Since 1993, when the InGaN high-brightness blue LEDs were fabricated,
III nitrides became a subject of a great interest that was regarded as “the blue-UV revolution”
in optoelectronics. GaN and AlGaN compounds became the key materials in
short-wavelength optoelectronics as well as in high-power electronics. According to
Optoelectronics Industry Development Association of USA (see Figure 3.1) [1], high
brightness LEDs exhibit a 30-fold increase in efficiency per decade and start superseding
incandescence bulbs and discharge tubes in general lighting applications around 2008.
In order to achieve a high efficiency, dislocations should be eliminated. Radiative
recombination is quenched in a capture zone around dislocations (illustrated as purple zones
in the cartoon at the bottom of Figure 3.2) [1]. As dislocation density ρdisl increases, internal
39
radiative efficiency ηint decreases. The capture zone may be viewed as having a radius on the
order of the minority carrier diffusion length Lo. However, the minority carrier diffusion
length itself depends on dislocation density. Hence, the dependence of internal radiative
efficiency on dislocation density must in general be solved self-consistently.
Figure 3.3(a) shows the energy band diagrams of Ga-polar and N-polar InGaN MQW
LEDs. The energy band diagrams were calculated using a self-consistent
Schrodinger-Poisson solver that incorporates spontaneous and piezoelectric polarization. In
the Ga-polar LED, the polarization fields are reversed with respect to the p-n junction
depletion field, resulting a wider depletion region. However, in the N-polar LED, the
polarization field is in the same direction as the depletion region, leading to the reduced
depletion region. This is expected to result in a lower turn-on voltage for the N-polar LED
p-n junction. In Figure 3.3(b), due to the reverse direction of polarization field in a N-polar
LED, a larger forward bias voltage assists the field in QW regions to approach flat-band
conditions, leading to less stark effect across the QW. However, in a Ga-polar LED,
increasing forward bias voltage decrease the electron and hole wave-function overlap in the
quantum wells [2].
Figure 3.4 (a) and (b) shows a schematic representation of the Ga-face and N-face
surfaces, respectively, including the dangling bonds available at <1010> and <1120>
step edges. For a single-carbon atom bonding in a tetrahedral configuration, there are two
40
sites (site AGa and site BGa) on the Ga-face surface, where a carbon atom on a nitrogen site is
likely due to the higher number of available dangling bonds. However, on the N-face, there is
only a single site (site AN), which is likely to allow a carbon to bond in the tetrahedral
configuration. Alternatively, there are two sites on the Ga-face (site CGa and site DGa), where
carbon incorporation as part of a monomethyl-gallium group is likely, whereas, the N-face
only has a single site (site BN). In either case, since there are more available sites for
incorporation, it is expected that carbon incorporation would be higher on the Ga-face
surface [3].
The oxygen concentrations in different growth conditions measured by SIMS are shown
in Figure 3.5. It is found that the oxygen concentration in the N-face GaN samples was at
least 10 times higher than that in the Ga-face GaN samples for all growth conditions. For the
Ga-face GaN, atoms impinging on a group V site forms only a single bond to the Ga surface
atoms, while for the N-face GaN, an atom impinging on a group V site forms 3 bonds to the
Ga surface atoms, leading to a stronger bonding of oxygen atoms on the N-face GaN surface.
In general, the observed trends can also be explained by the substitution of nitrogen surface
atoms by an oxygen atom on the N-face GaN surface. As the V/III ratio increases, more
active nitrogen is supplied to the surface. As expected, the incorporation of oxygen decreases
[3]. Table 3.1 show structures and properties of Ga-poalr and N-polar GaN [4].
41
3.2 Motivation and Investigation Flow Chart
Solid-state lighting through LEDs is considered as the next generation lighting. The
external quantum efficiencies (EQEs) of InGaN LEDs at wavelengths between 365 and 450
nm have been greatly improved；however, the internal quantum efficiency of InGaN green
LEDs is relatively poor [6-7]. InxGa1-xN alloys at a high indium mole fraction often leads to
low crystalline quality because of indium aggregation or phase separation. In high indium
mole fraction, the V-shape defects are the most common defects that occur at InGaN/GaN
QWs. V-shape defects are easily formed in high indium MQW and triggered by treading
dislocations in the buffer layer. But sometimes, these defects are formed because of strain
relaxation associated with stacking faults or indium segregation [8-11]. The output power of
green LEDs can be greatly improved by reducing the V-shape defects through growth
parameter optimization.
Figure 3.8 shows the experimental flow chart of this chapter. These experiments not
only showed the electrical characteristics, but also physical phenomena in LEDs. In our
experiment, one Ga-polar LED (Ga-polar@sapphire LED), one N-polar LED
(N-polar@sapphire LED), and two Semipolar LEDs (semipolar@m-sapphire and
semipolar@r-sapphire LEDs) are prepared. After the samples are ready, we measure the
electroluminescence spectra. Electroluminescence spectra can show the basic optical
characteristics of the LEDs. Integrating sphere is used to characterize electroluminescence
42
(EL) spectra for LED devices. To more clearly understand the samples in different situations
under different applied voltage, we take the EL images of LEDs by CCD camera. With a
basic understanding of the LEDs, we feel interested in the luminescence behaviors of the
LEDs. For example, is the current distribution of the LED uniform ？ What emitting
mechanism of the LEDs？What the luminescence efficiency of LED is？etc. Therefore, we
conduct current-voltage, luminescence efficiency, and time-resolved electroluminescence
measurements. Time-resolved electroluminescence (TREL) measurement can characterize
response time and decay time for LED devices. Finally, the results are used to analyze the
polarization effect on device characteristics and carrier transport properties of LEDs.
3.3 Sample Structures and Growth Conditions of LEDs
Figure 3.9 shows the sample structures of (a) Ga-polar grown on sapphire
(Ga-polar@sapphire LED), (b) N-polar grown on sapphire (N-polar@sapphire LED), (c)
semipolar grown on r-sapphire (semipolar@r-sapphire LED), and (d) semipolar grown on
m-sapphire (semipolar@m-sapphire LED) InGaN LEDs. Table 3.2 shows the growth
conditions of Ga-polar, N-polar, and semipolar InGaN/GaN MQWs LEDs.
Both Ga-polar and N-polar undoped GaN growths were carried out on nominally
Both Ga-polar and N-polar undoped GaN growths were carried out on nominally