Chapter 2 Characteristics of III-Nitride quantum well structure
2.4 In-plane light polarization effect in nonpolar nanostructure
2.4 In-plane light polarization effect in nonpolar nanostructure
In addition to the advantage of no internal electric fields in growth direction, nonpolar orientation of the wurtzite crystal structures are also good for polarization-sensitive devices. In contrast to GaN films grown along the polar c-plane direction, do not exhibit any in-plane polarization anisotropy, m- and a-plane GaN films, where the c-axis lies in the film plane, exhibit significant polarization anisotropy in its light emission.
When GaN films are grown on c-plane substrates with similar hexagonal symmetry, there is no linear polarization emission happened under isotropic in-plane strain. However, a GaN film with an m-plane or a-plane orientation experience anisotropic in-plane strain resulted from the inherent lower symmetry of m and a
plane. An in-plane polarization anisotropy has been observed in the PL spectra of GaN/(Al,Ga)N MQWs. [12, 21]
The coordinate system used for the investigation of the polarization properties is
x//a[112 0]、y//m[1− 100]、and z//c[0001], the three planes are perpendicular to each − other. Take the m-plane GaN films for example, m-plane strain lifts the symmetry in the x-y plane of the wurtzite crystal and separates the original |X±iY>-like HH and LH states of unstrained WZ-GaN to |X>-like and |Y>-like states. A compressive strain along x induces a dilatation along y so that the energy of the |X>-like state is raised while the |Y>-like state is lowered. Valence bands are reconstituted to |X>-like,
|Z>-like, and |Y>-like states in order of decreasing electron energy. Therefore, the valence band structure is altered that the lowest transition energy is totally linearly polarized in the x-direction (E⊥C), while the second lowest transition energy is totally linearly polarized in the z-direction (E║C). The third lowest transition energy which is linearly polarized in the y-direction will not show any PL intensity difference when we put the polarizer in x-z plane to get the polarization information.
The optical polarization anisotropy in GaN films with a nonpolar orientation can be used for polarization-sensitive photodetectors [22] and static as well as dynamic polarization filtering [23]. It has also been applied to achieve polarized emission in (In,Ga)N/GaN light-emitting diodes grown along nonpolar orientations [24].
(a) (b)
Fig. 2.3.1 A schematic energy band diagram of (a) nonpolar quantum well and (b) polar quantum well under QCSE.
C.B
V.B
InGaN GaN
GaN C.B
V.B
InGaN GaN
GaN
Chapter 3 Sample Preparation and Experiment Setup 3.1 Sample growth and HRXRD quantification
All epitaxial films were grown on r-plane sapphire in MOCVD reactor.
Trimethylgallium, trimethylindium, and ammonia were the precursors used for sources of Ga, In and N in whole epitaxy process. At first, the r-plane sapphire substrate was treated by thermal annealing at 1090 . Subsequently, a 30 nm ℃ thickness AlN nucleation layer was deposited at 600 . The growth temperature was ℃ ramped up to 1120 to grow a℃ -plane bulk GaN of 2 µm thickness. The MQWs structure was grown at 700 which con℃ sisted of 10 pairs of GaN barriers and InxGa1-xN wells. We controlled different growth time ranging from 1~3 minutes on growing quantum well to get different active layer thickenss. Finally, the 50 nm capping layer of GaN was deposited. A schematic diagram of sample structure was shown in figure 3.1.1
The ω-2θ scan measurements of high resolution x-ray diffraction (HRXRD) made with a Philips MRD X’pert PRO diffractometer using CuKα1 radiation were applied to quantify the quantum well and barrier thicknesses. The In composition
could also be made sure. Figure 3.1.2 is a schema of the (112 0) a-plane orientation − which defines the specific crystallographic index used in the HRXRD analysis.
The dynamic diffraction simulation shown in figure 3.1.3(a)~(d) models the
peak locations of the HRXRD experimental results. The X-ray diffraction analysis confirmed that the In composition of the quantum well is around 23%, the GaN barrier is around 12 nm thick and the approximate well width for different growing time samples are 3, 6, 9, and 12 nm, respectively. We concluded the HRXRD measured result in table 3.1.1.
Fig. 3.1.1 A schematic drawing of sample structure
Fig. 3.1.2 The specific crystallographic index used in the HRXRD analysis.
10 periods GaN/InGaN MQWs
r-plane Sapphire AlN nucleation layer
Bulk GaN GaN capping layer
(a) 1min (b) 2min
(c)2.5min (d)3min
Fig. 3.1.3 Experiment and simulation results of HRXRD ω-2θ scans for a-plane InGaN/GaN MQWs.
1min 2min 2.5min 3min
In composition 0.23 0.23 0.23 0.23
well
thickness 3.3nm 5.9nm 8.8nm 12.1nm
Table 3.1.1 The HRXRD measured layer thickness and composition
3.2 Photoluminescence (PL)
Photoluminescence spectroscopy is an un-contact, nondestructive method of examining the electronic structure and optical characteristics of materials.
Photoluminescence is the emission of light from a material under optical excitation. It needs an excited light source to induce the emission and the energy of the excited light source should be higher than the band gap energy of the material. Light is directed onto a sample where it is absorbed and electrons in the valence band would get the energy to jump into the conduction band then relatively produce a hole in the valence band. When an excited electron in the conduction band returns back to the valence band, it releases the energy which includes a radiative process and a nonradiative process.
The energy of the emitted light (photoluminescence) relates to the difference in energy levels between two states involved in the transition which are the excited state and the equilibrium state. The intensity of the emitted light is related to the contribution of the radiative process. The intensity and spectral content of a photoluminescence spectrum is a direct inspection of material properties.
When it comes to radiative recombination process in semiconductors, there are many different path it could get through .We can recognized these various transitions by PL emission at low temperature, since low temperature rules out the influence of
thermal energy. Before recombination, an electron and a hole usually form a quasi-particle that is a bound state with Coulomb interaction between them. The quantum of this electronic polarization is called as exciton. The introduction of the exciton was made by Frenkel in 1931.
Generally in semiconductor, a weakly bound e-h pair whose wavefunction propagates more than the lattice spacing is formed and is called as Mott-Wannier exciton (free exciton, FE). [25] Conversely in ionic crystals or molecular crystals, e-h pair is strongly bound at the matrix atom or localized at its neighborhood. This type of exciton is called as Frenkel exciton. (bound exciton, BE).
Stable excitons will only be formed if the attractive potential is sufficient to protect the exciton against collisions with phonons. Mott-Wannier excitons have small binding energy due to their large radius, with typical values of around 0.01eV.
Since KBT ~0.025eV at room temperature, where KB is Boltzmann’s constant, these excitons are only observed clearly at low temperature in many materials. Frenkel excitons, on the other hand, have larger binding energies of the order 0.1-1eV, which makes them stable at room temperature.
The setup of our PL system is shown in Fig. 3.2.1. The pumping light source was a multi-mode and non-polarized Helium-Cadmium laser operated on 325nm with
20mW. After reflected by three mirrors, the laser light was focus by a lens which focal length was 5cm, to 0.1mm in diameter and the luminescence signal was collected by the same lens. The probed light was dispersed by 0.32 monochromator (Jobin-Yvon Triax-320) equipped with 1800, 1200, and 300 grooves/mm grating and which maximum width of the entrance slits was 1mm. The resolution was controlled in 1nm by selecting 300 grooves/mm grating and slit of 0.1mm. We use long pass filter to avoid the laser coupling with the PL spectrum.
3.3 Photoluminescence Excitation (PLE)
In PL measurement, which is peformed at a fixed excitation energy, the luminescence properties are generally investigated. In PL excitation (PLE) spectroscopy, which is carried out at fixed detection energy, provides mainly information about the absorption properties. Apart from PL experiments, the PLE measurement is a widely used spectroscopic tool for the characterization of optical transitions in semiconductors.
It is very important to note that the PLE experiment also depends strongly on the different carrier relaxation processes that connect the absorbing state to the luminescent state. Nevertheless, in many cases it is difficult to separate the influence of relaxation from that of absorption. The PLE spectrum is strongly influenced by the
relaxation depending on different samples.
The setup of our PLE system is shown in Fig 3.3.1. Except for the excited light source, the whole light collection setup and spectrometer (Triax320) are the same as PL system.
The pumping source of PLE was Xe lamp with 450W separated by a double-grating monochoramator (Jobin-Yvon Gemini180) and then coupled to samples at an angle about 450 by two focal lenses. We fixed the detection energy through the spectrometer Triax320, and changed the excitation energy range of Xe lamp to get PLE spectrums. At the exit of the spectrometer Triax 320, a high sensitive Hamamatsu photomultiplier tube (PMT) with GaAs photocathode was placed to detect the luminescence signals.
3.4 Confocal Optical image
"Confocal" is defined as having the same focus. What this means in the microscope is that the final image has the same focus as or the focus corresponds to the point of focus in the object. The object and its image are "confocal". The microscope is able to filter out the out-of-focus light from above and below the point of focus in the object. Normally when an object is imaged in the fluorescence microscope, the signal produced is from the full thickness of the specimen which does
not allow most of it to be in focus to the observer. The confocal microscope eliminates this out-of-focus information by means of a confocal pinhole situated in front of the image plane which acts as a spatial filter and allows only the in-focus portion of the light to be imaged. Light from above and below the plane of focus of the object is eliminated from the final image. A schematic diagram of the confocal principle is shown in Fig 3.4.1. [26]
Confocal microscopy offers several advantages over conventional widefield optical microscopy, including the ability to control depth of field, elimination or reduction of background information away from the focal plane (that leads to image degradation), and the capability to collect serial optical sections from thick specimens.
The basic key to the confocal approach is the use of spatial filtering techniques to eliminate out-of-focus light or glare in specimens whose thickness exceeds the immediate plane of focus. In fact, confocal technology is proving to be one of the most important advances ever achieved in optical microscopy.
The confocal microscope experiment setup is shown in Fig.3.4.2.We chose a 40x objsctive and a 100um fiber to collect the optical signal from the samples that gave a space resolution about 3um. The excitation light source was a multi-mode Helium-Cadmium laser operated on 325nm with 40mW.After reflecting by four
mirrors, the laser light went through a beam expander and then again reflected by three mirrors to arrive the objective .After the laser light excitation, photoluminescence from the sample was collected by the same objective and finally met the fiber which play the role as the confocal pinhole. All the optical signals was transmitted to a high sensitive Hamamatsu photomultiplier tube (PMT) and dealt with by a computer to form a optical image. The optical images we saw reflect the spacital distribution of the luminescence intensity from the sample we probed on.
Fig. 3.2.1 The setup of PL system
Fig. 3.3.1 The setup of PLE
monochromotor Mirror1
sample
Lens4 Filter Lens3
Triax320 Xe lamp Lens1
Lens2
Mirror1 Mirror 2
Lens2
Filter 360nm Lens1
Triax320
25mW He-Cd laser (325nm) InGaN MQWs
Fig.3.4.1 A schematic diagram of the confocal principle [26]
Fig. 3.4.2 The confocal microscope experiment setup
Chapter 4 Optical properties of a-plane InGaN/GaN MQWs (I) 4.1 Introduction
The optical characteristics of c-plane InGaN/GaN multiple quantum wells (MQWs) have been studied extensively. [8~9] However, the optical properties of a-plane InGaN/GaN MQWs are still worth investigating for assistance of fabrication
due to the lack of the internal field and possible different growth parameters. Since the interface roughness and treading dislocations in a-plane heterostructures are more complicated than those in c-plane heterostructures, [20] the luminescence mechanism requires further clarification of the dependence of the optical characteristics on the different InGaN/GaN quantum well widths. Craven et al. had investigated optical characteristics of GaN/AlGaN MQWs with different well widths [7]; however, the issues related to well width dependence of a-plane InGaN/GaN MQWs have not been conferred yet.
In section 4.2, CW photoluminescence (PL) was performed to investigate the emission peak position and the material quality. Excitation power dependent PL measurement shows that the relation between laser excitation power and sample luminescence intensity is linear direct proportion which proves the absence of built-in electric field in these a-plane MQWs.
As mentioned in section 2.2, localization effect plays an important role in polar
c-plane InGaN-based structure luminescence. It is widely accepted that the high luminescence efficiency of polar InGaN-based structures is due to the exciton localization effect. Thus, in section 4.3, we performed temperature dependent PL and photoluminescence excitation (PLE) experiments to examine if localization effect also be an important role in luminescence efficiency of nonpolar a-plane InGaN/GaN MQWs. We see an increasing PL intensity decay in thicker well width sample when the temperature is increasing from 20K to 300k. S-curve shift of emission peak energy with increasing temperature which passes for a result of localization effect is also observed in 9nm and 12nm well width samples. In PLE experiment, a larger Stokes shift and a broadening PLE spectrum are observed when the well width gets wider.
We attribute the Stokes shifts to inhomogeneous broadening localization effect which is expected due to the samples’ structure and composition non-uniformity. This result also implies that localization effect possesses a crucial role for luminescence efficiency.
Final, in section 4.4, the in-plane light polarization effect was observed. An average degree ρ of linear polarization about 60% was got in these four a-plane InGaN/GaN MQWs.
4.2 Photoluminescence spectra of a-plane InGaN/GaN multiple quantum wells
4.2.1 Room temperature photoluminescence
Room temperature PL measurements were performed using the cw 325 nm He–Cd laser operating at an excitation level of 25 mW.
The CW PL spectra of these four samples measured at room temperature are shown in Fig. 4.2.1(a). The detailed MQW PL peak emission energy shown in Fig.
4.2.1(b) increases from 2.47eV to 2.79 eV with the decreasing well width could be fully attributed to the quantum size confinement effect. The similar peak energy of the samples of 9 and 12 nm is due to the weak quantum confinement effect in a larger well width. Figure 4.2.1(c) shows the PL peak emission intensity that gradually decreased with the increase of InGaN well width. When the well width is thicker than 6 nm, the PL intensity drops more quickly.The well width of optimal integrated PL intensity for a-plane would be thicker than relative thin polar quantum well [27]. In additional, the material quality, interface roughness and the excitonic Bohr radius would be considered in terms of the determination of optimal well width [7].
4.2.2 Power dependent photoluminescence
We then analyze different power dependences I ~Pα for the samples of different well widths over a wide range of excitation power where I is the PL intensity, P is the pumping power intensity, and α is the power index. Power dependent PL
measurement was carried out by using the CW 325 nm He–Cd laser which power density was controlled from 2 to 200 mW/cm2.
In Fig.4.2.2, we show that the PL spectrum as a function of the excitation power for the samples with different well width. We obtained unshifted PL peaks with the increasing pumping power density, which is well known for a-plane hexagonal MQWs with the nonpolar characteristic and the flat band structure. [3] The stable spectral peak position under a wide range of excitation powers clearly indicates that the a-plane InGaN/GaN MQWs on the r-plane sapphire are nonpolar and thus free from the built-in electrostatic fields.
Fig.4.2.3 shows the PL integrated intensity among different excitation power density. Generally, I ~ P relation would satisfy with the condition when nonradiative channels saturate and radiative recombination predominates at elevated excitation.
The power indices around 1 for our all samples indicate that the radiative recombination dominate in the optical transition [28] and is absolutely independent of InGaN well width, which also give the other evidence that no built-in electric field was observed within our a-plane InGaN/GaN MQWs with different well widths.
However, such the a-plane MQWs without the built-in electric field within should not exhibit strong PL intensity dependence on the well width. [29] Other determining factors should account for the PL intensity drop with a thicker quantum well.
4.3 The localization effect of a-plane InGaN/GaN multiple quantum wells
4.3.1 Temperature dependent photoluminescence
In order to further understand whether the localization effect plays an important role in nonpolar InGaN MQWs as well as in polar c-plane InGaN nanostructure or not, PL spectra were measured under different temperature in the range of 20K to 300K using the CW 325 nm He–Cd laser.
Figure 4.3.1 shows the evolution of PL spectra as a function of temperature for the a-plane InGaN/GaN MQWs with well width from 3nm to 12 nm. In these four samples, the decrease of PL intensity with increasing temperature is observed independent of quantum well width. Thermal quenching PL intensity with increasing temperature is a general phenomenon in III-V semiconductor nanostructure which is caused by carriers thermalization from the radiative recombination centers or/and localized states to the nonradiative recombination centers or/and delocalized states.
[30] This thermal quenching behavior will be discussed later.
Moreover, at low temperature, there reveals three separated peaks in the PL spectrum, the most high energy peak located around 3.35eV in all samples is suggested to the signal of bulk GaN. The middle energy peak which is obvious in the samples of 3nm and 6nm well width but is merged with the lowest-energy side signal in the samples of 9nm and 12nm well width is supposed to the signal coming
from shallow localized states. [31] The lowest energy signal comes from the deep localized states. As can be seen, only the PL emission from excitons in deep localized states dominates the luminescence from 20K to room temperature, the other two higher energy emissions suffer an apparent quickly thermal quenching when the temperature increases. Along with the increase in temperature, the nonradiative energy relaxation of excitons occurs at shallow localized states, and then the efficient radiative recombination of excitons occurs mainly at deep localized states. The exciton dynamics at shallow delocalized states is very sensitive to the lattice temperature. When the temperature goes up, it is hard for excitons to stay stable in shallow localized states, the thermal dissociation of excitons occurs at shallow localized states, and electrons and holes are thermally excited into the delocalized states, thus, more transfer and relaxation processes happen which quench the emission from high energy states. [31]
Figure 4.3.2 shows the Arrhenius plot of the normalized integrated PL intensity
Figure 4.3.2 shows the Arrhenius plot of the normalized integrated PL intensity