Power dependent photoluminescence

The Fig. 4.4.12 shows the PL intensity as a function of excitation power at 15 K. The PL intensity is linear with excitation power at lower excitation power range, which indicates that the radiative recombination dominates the recombination process at 15 K for all samples in this excitation power range [56][57][58]. But as excitation power increases above about 20 mW, the PL intensity exhibits a tendency of saturation, this phenomenon will be discussed later. The Fig. 4.4.13 shows the emission energy and FWHM as a function of excitation power at 15 K for all samples. The emission energy gradually increases with increasing excitation power. On the other hand, the linewidth of spectra shrinks with increasing excitation power

from 0.05 mW to about 1 mW, and if injected power increase continually, then broadening of spectra is observed. In generally, the left side of red dash line in Fig. 4.4.13 is dominated by the coulomb screening of the QCSE and the right side of red dash line is dominated by the band filling effect of localized states [33][34][35]. (More detailed explanation will be stated in Chapter 5) The emission energy blue-shift resulted by bang filling effect is calculated to examine the degree of carrier localization, and the values are 55.5, 45.2, 47.2 and 133.3 meV for 0^o, 0.2^o, 0.35^o and 1^o sample, respectively. In general, the degree of emission energy blue-shift with increasing excitation power is proportion to the effective potential barrier of localized states, more deep of localized states will result in more blue shift of emission energy with increasing excitation power. The results indicate that the effective potential barrier of localized states decreases as misorientation angle increases from 0^o to 0.2^o, but it increases as misorientation angle increases above 0.2^o. And the results are good agree with the analysis of TRPL and thermal activation energy extracted from temperature dependent PL.

On the other hand, the PL efficiency (will be defined in Chapter 5) as a function of excitation power at 15 K was studied, which is shown in Fig. 4.4.14. The PL efficiency curve in Fig. 4.4.14 reflects the IQE, because the light extraction efficiency (LEE) does not depend on excitation power of laser light. From figure we can see the IQE increases with increasing excitation power under lower excitation power range, when the excitation power further increases, then the IQE decreases. From the discussion in Chapter 5, we know that the

increasing of PL efficiency at 15 K is due to the coulomb screening of the QCSE, and the decreasing of IQE is due to band filling effect of localized states. The former effect will increase the electron-hole wavefunction overlap, resulting in increasing of IQE. The latter effect will make the carriers more easily escape from localized states to extended states, which deteriorates the IQE. From the figure we can observe two phenomena: (i) The more pronounced decreasing of IQE for 0.2^o and 0.35^o. (ii) The IQE for 0.2^o and 0.35^o sample are more easy saturate than it for 0^o and 1^o sample

From the analysis of previous section we can know this phenomenon is due to that the 0.2^o and 0.35^o sample has smaller effective potential barrier of localized states, making the carriers more easily escape from localized states to extended states, resulting in more easy saturation and more pronounced decreasing of IQE. On the contrary, the 0^o and 1^o sample has larger effective potential barrier of localized states, making the carriers more difficultly escape from localized states to extended states, resulting in the peak of efficiency curve shifts to higher excitation power and PL efficiency decreases more weakly at higher excitation power.

4.5 Electroluminescence intensity as a function of injected current

Fig. 4.5.1 shows the output power as a function of injected current for the LED with different misorientation angle. The light output powers at 20 mA are 10.3, 13.3, 12.1 and 7.1

mW for the LED with misorientation angle of 0^o, 0.2^o, 0.35^o and 1^o, respectively, i.e., an improvement factor of approximate to 1.29 was achieved by increasing misorientation angle from 0^o to 0.2^o. But as the misorientation angle increases above 0.2^o, the degradation of output power is observed. The improvement of luminescence efficiency in 0.2^o sample could be due to the good surface morphology and the low dislocation density in the device, lowering the leakage current and contact resistant. On the other hand, p-type doping of GaN is a very important parameter in achieving high quality devices. The smoother surface morphology and the low dislocation density will improve the p-type doping level [56][57]. Although the InGaN/GaN MQWs with misorientation angle of 0^o, 0.35^o, and 1^o has larger effective potential barrier of localized states, but due to the increasing in dislocation density, current leakage and contact resistant, and the decreasing of p-type doping level, the performance of device is degraded.

4.6 Conclusion

From our material analysis results and some references, when the GaN grown on sapphire substrate with slight misorientation angle of 0.2^o, the condition for spiral annihilation can be satisfied, then the growth mode of GaN will be shifted from spiral-dominated to step-flow. Therefore, due to suppression of spiral growth, the dislocation density in underlying GaN layer can be effectively decreased, making upper layer with better crystal

quality and smoother surface, and better homogeneity of InGaN/GaN MQWs. On the other hand, the incorporation of group III atoms, the epitaxial growth of III-nitrides, is closely related to the density of dangling bonds at step edges. Therefore, when the misorientation angle is too large, the density of step edges is too high and it becomes difficult to grow with smooth surface, and the strain relaxation in the interface increases, resulting in increasing of dislocation density.

And due to the dislocation density and surface morphology of underlying GaN layer varies as a function of misorientation angle of sapphire substrate, the homogeneity of upper InGaN/GaN MQWs is influenced, because the indium incorporation rate in InGaN is strong depended on the dislocation density and surface morphology in underlying GaN layer. So from our optical properties analysis results, we observed that the degree of carrier localization in InGaN MQWs varies with the misorientation angle of sapphire substrate. And the InGaN MQWs with misorientation angle of 0.2^o has the smallest effective potential barrier of localized states, which have weaker carrier localization effect and lager dimension of nanostructure.

Moreover, due to decreasing of dislocation density and better surface morphology of sample with misorientation angle of 0.2^o, it may improve the p-type doping level, lower the contact resistant, and reduce the current leakage in the device, therefore enhance the

luminescence efficiency of LED. Although the InGaN/GaN MQWs with misorientation angle of 0^o, 0.35^o, and 1^o has larger effective potential barrier of localized states, but due to poor surface morphology and high dislocation density in the device, the performance of device is deteriorated.

Fig. 4.1.1 Diagram of sapphire substrate with misorientation angle toward [1120] direction.

Fig. 4.2.1 Sample structure.

Fig. 4.3.1 The AFM images of p-type GaN grown on sapphire substrate with different misorientation angle.

0.0 0.2 0.4 0.6 0.8 1.0 0.0

0.5 1.0 1.5 2.0 2.5 3.0

Rms (nm) 5 µ m x 5 µ m

Mis - orientation angle (degree)

Fig. 4.3.2 The surface roughness of p-type GaN as a function of misorientation angle.

0 ^o

p - GaN

MQWs

n - GaN

0.2 ^o

p - GaN

MQWs

n - GaN

1 ^o

p - GaN

MQWs

n - GaN

Fig. 4.3.3 HRTEM images for the LED grown on sapphire substrate with different misorientation angle.

0 ^o 0 ^o

0.2 ^o 0.2 ^o

1 ^o 1 ^o

Fig. 4.3.4 HRTEM images for the MQWs grown on sapphire substrate with different misorientation angle.

0.0 0.2 0.4 0.6 0.8 1.0 1

2 3 4 5 6 7

Dislocation density (cm-2 ) / 108

Mis - orientation angle (degree)

Fig. 4.3.5 The dislocation density as a function of misorientation angle.

0.0 0.2 0.4 0.6 0.8 1.0 240

250 260 270 280 290 300

(0 02) FWHM (arcsec)

Mis - orientation angle (degree)

Fig. 4.3.6 FWHM of XRC for GaN (002) reflections plotted as a function of misorientation angle.

10

-4000 -3000 -2000 -1000 0 1000 2000 3000 4000

ω

Fig. 4.3.7 The ω-2θ scan (0002) for the samples grown on sapphire substrate with different misorientation angle.

Fig. 4.3.8 The IRN of InGaN/GaN MQWs.

(a)

+2

GaN

-1

(b) +2

GaN

0

-1

-2

-3

-4

(c) +2 +1 GaN

-1

-2

(d)

+2

GaN

Fig. 4.3.9 Reciprocal space mapping measured around the (1015) reflection (a) 0^o (b) 0.2^o (c) 0.35^o (d) 1^o.

Fig. 4.3.10 The schematic diagram illustrating the effect of strain and composition gradients in the symmetric and asymmetric RLPs of InxGa1-xN.

500 550 600 650 700 750 800

A1(LO)

Raman intens ity (a.u.)

Raman shift (cm

^-1

)

0^o 0.2^o 0.35^o 1^o E2(High)

Fig. 4.3.11 The Raman spectra for the InGaN/GaN MQW LEDs grown on sapphire substrate with different misorientation angle.

0.0 0.2 0.4 0.6 0.8 1.0

540 550 560 570 580 590 600

Raman intensity (a.u.)

Fig. 4.3.12 The Raman shift of E2(high) mode as a function of misorientation angle. The inset shows the the Raman spectra for the LED grown on sapphire substrate with different

misorientation angle.

Fig. 4.3.13 The calculated compressive stress as a function of misorientation angle.

2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6

PL In ten sity (a .u .)

Emission energy (eV)

R.T.

0^o 0.2^o 0.35^o

1^o 0.0 0.2 0.4 0.6 0.8 1.0

128 132 136 140

FWHM (meV)

Mis-orientation angle (degree)

Fig. 4.4.1 The PL spectra for the LED grown on sapphire substrate with different misorientation angle at room temperature, and the inset shows the FWHM of spectra as a

function of misorientation angle.

Fig. 4.4.2 Normalized emission energy mapping of µ-PL from InGaN/GaN MQWs grown on sapphire substrate with different misorientation angle.

0.0 0.2 0.4 0.6 0.8 1.0 0

10 20 30 40 50 60 70

Fluctuation of emission energy (meV)

Mis - orientation angle (degree)

Fig. 4.4.3 The fluctuation of emission energy as a function of misorientation angle.

2.4 2.6 2.8 3.0

Fig. 4.4.4 The temperature dependent PL spectra over a temperature range from 15 K to 300 K for the LED grown on sapphire substrate with different misorientation angle.

0 50 100 150 200 250 300 350

Fig. 4.4.5 The emission energies as a function of temperature for InGaN-related emission for the LED grown on sapphire substrate with different misorientation angle.

0.0 0.2 0.4 0.6 0.8 1.0

Fig. 4.4.6 The broadening parameter as a function of misorientation angle.

0 10 20 30 40 50 60 70 80

Fig. 4.4.7 The temperature dependent PL intensity for the LED grown on sapphire substrate with different misorientation angle.

Fig. 4.4.8 The thermal activation energy as a function of misorientation angle.

0 50 100 150 200 250 300 350

Fig. 4.4.9 The temperature dependent radiative recombination lifetime, nonradiative recombination and carrier lifetime for the LED grown on sapphire substrate with different misorientation angle.

Fig. 4.4.10 The temperature dependent radiative recombination lifetime for the LED grown on sapphire substrate with different misorientation angle.

0.0 0.2 0.4 0.6 0.8 1.0 12

14 16 18 20 22

Carrier lifetime (ns)

Mis - orientation angle (degree)

Fig. 4.4.11 The carrier lifetime detected at peak energy as a function of misorientation angle at 15 K.

0.01 0.1 1 10 100

PL inte ns ity (a.u.)

Excitation power (mW)

0^o 0.2^o 0.35^o 1^o

Fig. 4.4.12 The PL intensity as a function of excitation power at 15 K for the LED grown on sapphire substrate with different misorientation angle

1E-3 0.01 0.1 1 10 100

Fig. 4.4.13 The emission energy and FWHM as a function of excitation power for the sample grown on sapphire substrate with different misorientation angle.

1E-3 0.01 0.1 1 10 100

Fig. 4.4.14 The PL efficiency as a function of excitation power at 15 K.

0 20 40 60 80 100 0

10 20 30 40 50 60

Output power (mW)

Current (mA)

0^o 0.2^o 0.35^o 1^o

Fig. 4.5.1 Output power as a function of current for the LED grown on sapphire substrate with different misorientation angle.

Chapter 5 Physical mechanisms of excitation power dependent internal quantum efficiency in InGaN/GaN multiple quantum well light emitting diodes

5.1 Introduction

The InGaN/GaN material system has attracted much attention due to their tremendous potential for fabricating light emitting diodes operated from visable to ultraviolet energy range. In spite of this striking advanced technology, the emission process of this materials system is still under debate.

To improve the performance of InGaN/GaN LEDs, it is very important to know what physical mechanisms affect the IQE in this material system. In tradition, the IQE is estimate by measuring temperature dependent PL at a certain excitation condition, and assume IQE at low temperature is equal to 100%, then the relative IQE at room temperature can be obtained [67]. However, IQE is to be strongly dependent on injected carrier density, especially in InGaN-based QW system, because the existence of large internal electrical field and potential fluctuation. Therefore, it is important to measure and discuss IQE as a function of excitation power density. S. Watanabe et al. proposed a method to determine IQE by performing excitation power density and temperature dependent PL [20].

In their study, the variation of IQE of InGaN/GaN LEDs with increasing excitation power at low and high temperature has been observed. But they did not say anything more to explain what physical mechanisms occur in it. For this study, we used their measured method

to determine IQE of InGaN/GaN MQW LEDs and also observed that the IQE changes with increasing excitation power density. By observing emission energy and FWHM of spectra as a function of excitation power density and carrier recombination dynamic by TRPL measurement, the physical mechanisms of excitation power dependent IQE for InGaN/GaN LEDs have been confirmed.

5.2

5.3

Sample preparation

The samples in this study are commercial InGaN/GaN MQW blue LEDs and grown by metal organic chemical vapor deposition (MOCVD). The sample in this study grown on c-plane (0001) sapphire substrates, consisting of 30-nm-thick AlN nucleation layer, a 2 µm Si-doped n-type GaN, and an unintentionally doped active layer with InxGa1-xN/GaN MQWs, and 0.2 µm Mg-doped p-type GaN. The doped concentration of n- and p-type GaN is nominally 5 x 10¹⁸ and 1x10¹⁹ cm^-3, respectively, the MQWs layer comprise 16 periods In0.2Ga0.85N well (~2 nm) and GaN barrier (~16 nm). The sample structure is shown in Fig.

5.2.1.

The measurement of the internal quantum efficiency of InGaN/GaN MQW LEDs

For this study, we used S. Watanabe et al. proposed method to determine the IQE of InGaN/GaN MQW LEDs. The PL quantum efficiency can be calculated by

EX

where IPL and IEX are PL intensity and excitation intensity, respectively. EPL and EEX are PL photon energy and excitation photon energy, respectively. C is a constant affected by mostly carrier injection efficiency by laser, light extraction and correction efficiency of PL, and does not depend on either excitation power density or measurement temperature. By performing temperature and excitation power dependent PL, and equation 5.3.1, the relative PL quantum efficiency curves can be obtained. And the constant C can be canceled out by normalizing the curves to the peak value at the lowest temperature, because it is independent on temperature or excitation power. From this normalization, the PL efficiency curves will not depend on carrier injection efficiency by laser, light extraction and correction efficiency of PL. Thus the PL efficiency obtained from temperature and excitation power dependent PL reflects IQE.

In tradition, the IQE is estimated by assuming that IQE at low temperature is 100%

regardless of excitation power density. However, IQE is strongly dependent on injected carrier density, therefore, it is more reasonable to assume peak of PL efficiency at lowest temperature is 100%, then, the IQE curves as a function of excitation power and temperature can be obtained.

Moreover, to avoid the absorption of GaN, the frequency doubled femtosecond pulsed Ti : sapphire laser of 390 nm was used for resonant excitation, the excitation power density was changed from 0.005 mW to 80 mW, and calculated injected carrier density is about 5 x

10¹³ to 8 x 10¹⁷ cm^-3.

Fig. 5.3.1 displays the IQE of the InGaN/GaN MQW LEDs as a function of injected carrier density at 15 K and 300 K. The IQE increases with increasing injected carrier density to reach its maximum. As injected carrier density further increases, then the IQE decreases.

The tendency of two efficiency curves at 15 K and 300 K is very similar. But under low injected carrier density range, the IQE at 300 K increases more pronounced than it at 15 K.

Moreover, the peak (arrow in Fig. 5.3.1) of efficiency curve in 300 K is at carrier density of about 1 x 10¹⁷ cm^-3, which is lager than it at 15 K, which is about 1 x 10¹⁶ cm^-3. The IQE at low temperature condition is more easy saturate than it at higher temperature. The detailed physical mechanisms will be discussed later.

5.4

IP

L α

The analysis of physical mechanisms

In general, the PL intensity L can be written as powers of excitation power intensity I as:

[58][59][60]

(5.4.1)

where parameter P physically reflect the various recombination processes. If P = 1, it indicates the radiative recombination dominates. If P = 2, the Shockley-Read-Hall (SRH) recombination dominates, which is generally nonradiative recombination. And P = 2/3 indicates Auger recombination dominates. Fig. 5.4.1 shows the L-I characteristic of our

sample. At 15 K, L is linear with excitation power intensity (P ~ 1), which indicates that the radiative recombination dominates the recombination process at all injected carrier density range and the nonradiative centers are quenched at low temperature. The unchanged parameter P shows that the variation of IQE at 15 K does not associate to defect related physical mechanisms. However, under low excitation power density at 300 K, the superlinear dependence of L on I (P ~ 2) is observed, which shows that the defect relative nonradiative recombination dominates in this low injected power range. But as excitation power density continual increase, the linear dependence of L on I is exhibited. This phenomenon can be explained as following: The fact that the radiative recombination rate corresponds to the carrier density squared (Rrad α Bnp), whereas the nonradiative recombination corresponds to only the carrier density (Rnonrad α An) [5], where A and B are Shockley-Read-Hall recombination and radiative recombination coefficients, respectively. For GaN based LED, a large number of dislocation exist in the device, therefore, the nonradiative recombination rate is much higher than radiative recombination in the low carrier density. In order to achieve Rrad

>> Rnonrad, the enough injected carrier density is necessary. In our case, as injected carriers increase, the nonradiative recombination is gradually suppressed, therefore, the radiative recombination starts to dominate the recombination process, resulting in pronounced increasing of IQE, which is observed in Fig. 5.4.1.

To further understand what physical mechanisms occur as injected carrier density

increases, the injected carrier density dependent spectra were observed. Fig. 5.4.2 (a) shows the emission energy and the FWHM of spectra as a function of injected carrier density at 15 K.

The emission energy gradually increases with increasing carrier density. On the other hand, the linewidth of spectra shrinks with increasing carrier density from 5 x 10¹³ cm^-3 to 5 x 10¹⁶ cm^-3, and if injected carriers increase continually, then broadening of spectra is observed.

These phenomena have been observed in InGaN/GaN QW for several articles [33][34][35]. In generally, there are two possible mechanisms to explain the blue-shift of emission energy with increasing excitation power.

(1) Coulomb screening of the QCSE.

Several articles have reported that the internal electric fields existing in InGaN/GaN QW structure. This internal electric field through the QW tilts the potential band and leads to a spatial separate of electrons and holes in the QW, resulting a decreasing in degree of wave function overlap, decreasing of transition energy and broadening the linewidth of spectra, which is called the QCSE. The internal electric field in the QW can be screened by photogenerated carriers. The increasing of carrier density weakens the QCSE, resulting in increasing of transition energy, increasing of carrier recombination rate and shrinking of linewidth of spectra.

(2) Band filling effect of localized states.

Due to composition inhomogeneity and monolayer thickness fluctuation of InGaN QWs,

self-organized In-rich region is generated in InGaN active region, resulting in potential fluctuation of energy bandgap. As injected carrier density increases, an occupation of high energetic localized centers will be enhanced, inducing a blue-shift of emission energy.

Moreover, this effect will broaden the linewidth of spectra.

From above explanation, we can include that the coulomb screening of the QCSE dominates the region (i) in Fig. 5.4.2 (a), the effect will increase the overlap of wavefunction of electron and hole, which increases the IQE. And the band filling of localized states dominates the region (ii), the effect will make the carriers more easily escape from localized states to extended states, which deteriorates the IQE.

Otherwise, the recombination dynamics of the injected carrier density dependent TRPL have been studied to clarify our thought. Fig. 5.4.2 (b) shows the carrier lifetime as a function of injected carrier density at 15 K, the carrier lifetime decreases with increasing injected carrier density is observed. The decreasing of carrier lifetime with increasing injected carrier density has been investigated in other article [35], and it is ascribed to the coulomb screening of internal electric in InGaN QW. As carrier density in the QW increases, more excited carriers can screen the internal electric field in QW to increase the overlap of wavefunction for electron and hole, thus enhances the carrier recombination rate, so carrier lifetime is decreased. After the coulomb screening of QCSE (the right side of red dash line in Fig. 5.4.2), we can observe the carrier lifetime keep decreasing, but it exhibits a saturated tendency at

higher carrier density. This phenomenon is an evidence of band-filling effect of localized states. Some articles have investigated the emission energy dependent TRPL in InGaN

higher carrier density. This phenomenon is an evidence of band-filling effect of localized states. Some articles have investigated the emission energy dependent TRPL in InGaN

在文檔中 氮化銦鎵/氮化鎵多重量子井發光二極體之光學特性與內部量子效率研究 (頁 64-113)