Chapter Growth and Characterizations Chapter Growth and Characterizations In this work, the GaN-based quantum wells and quantum dots were all grown by an EMCORE D125 metal organic chemical vapor deposition (MOCVD) system Section 2.1 will present the essential theories of this growth technique and highlight the major components of our MOCVD system To fabricate a broadband GaN-based SESAM, the SiO2/Si3N4 dielectric DBR was deposited by the plasma enhanced chemical vapor deposition (PECVD) system The principle of the PECVD technique will be presented in Section 2.2 Section 2.3 will be devoted to the primary characterization techniques used in this work, including photoluminescence (PL) (Section 2.3.1), spectrophotometry (Section 2.3.2), atomic force microscopy (AFM) (Section 2.3.3), scanning electron microscopy (SEM) (Section 2.3.4), transmission electron microscopy (TEM) (Section 2.3.5), and X-ray diffraction (XRD) (Section 2.3.6) The PL technique was used to investigate the emission properties of the quantum well and quantum dot samples To study the linear transmission and reflection spectra of different samples, the spectrophotometry technique was applied The AFM technique was used to investigate the surface morphology of the samples, 38 Chapter Growth and Characterizations and the values of surface roughness can be obtained But the accuracy of the surface profiles obtained by this technique depends on the size of the tip compared to the sizes and aspect ratios of the surface features In comparison, the SEM technique gives more trustable images than the AFM technique, but the values of surface roughness cannot be obtained by using SEM In addition, the cross-sectional SEM technique was also used to examine the quality of the interfaces and estimate the layer thickness When the feature sizes scale down to a few nanometers or tens of nanometers, the TEM technique was applied to study the nanostructures Either the plan-view or the cross-sectional crystal structures can be investigated Also, to precisely study the crystal quality, structural property and the stress development of the samples, the XRD technique was applied 2.1 Metal organic chemical vapor deposition (MOCVD) The MOCVD growth, also called metal organic vapor phase epitaxy (MOVPE), is a non-equilibrium thin film epitaxial growth technique It relies on the vapor transport of precursors and the subsequent chemical reactions of the precursors in a heated zone The reaction chamber of an MOCVD system can either be a quartz tube or a stainless steel chamber, which contains a heated substrate susceptor The hot susceptor has a catalytic effect on the decomposition of the gaseous products and the following growth of the materials on this hot surface Similar to other crystal growth techniques, fundamental processes in MOCVD are commonly divided into thermodynamic and kinetic regimes 39 Chapter Growth and Characterizations Thermodynamics determines the driving force for the overall growth process, while kinetics defines the rates at which the various processes occur Hydrodynamics and mass transport, which are intimately linked, control the material transportation rate to the interface between the growing solid and the vapor In the meantime, the chemical reaction rates at the growing interface also play a role Actually, each of these factors dominates a certain stage of the overall growth process Since the last decade, MOCVD has become the most established growth technique for III-nitride material growth and device application [Dupuis1997; Keller2003] Figure 2.1 Simplified schematic diagram of GaN growth process Figure 2.1 shows the simplified schematic diagram of the GaN growth 40 Chapter Growth and Characterizations process Although the actual growth process of GaN is complicated, it basically consists of the following four steps: 1) Ga(CH3)3 diffusion through the boundary layer to the substrate, 2) surface reactions, 3) formation of GaN, and 4) removal of the reaction products All the GaN-based samples discussed in this thesis were grown with an EMCORE D125 vertical-geometry rotating-disk reactor (RDR) MOCVD system The schematic diagram of this system is shown in Fig 2.2 This system can be grouped into four major parts, i.e., (1) the gas handling system consisting of alkyl and hydride sources, necessary gases and all instruments used to control the gas flows and mixtures; (2) the reaction chamber, where the pyrolysis reaction and deposition occur, and the loadlock; (3) the heating and temperature controlling system; and (4) the exhaust, pumping and pressure controlling system [Ludowise1985] bubblers Figure 2.2 Schematic diagram of an EMCORE D125 RDR MOCVD system (Courtesy of EMCORE corporation) For the gas handling system, it mainly comprises the sources (precursors and 41 Chapter Growth and Characterizations other necessary gases), and the gas mixing system (or manifold) The purpose of the gas handling system is to deliver precisely metered amount of uncontaminated reactants into the growth chamber without the transients due to the changes of pressure or gas flows In this work, trimethylgallium (TMGa), trimethylindium (TMIn) (from Epichem) were used as the group-III precursors of gallium (Ga) and indium (In), respectively Highly purified ammonia (NH3) (ammonia blue from Solkatronic) was used as the group-V precursor of nitrogen (N) Highly purified H2 and highly purified N2 (from National Oxygen) were used as the carrier gases Extra purifiers are also used for further reduction of oxygen and other contaminants in the highly purified gases of NH3, N2 and H2 to the order of ppb (parts per billion) They are Matheson Nanochem NH3 purifier, Mykrolis Aeronex N2 purifier and Johnson Mathey H2 purifier In order to control the reaction precisely, a specially designed mass flow controllers (MFC) can accurately and reliably measure and control the molar flow rates of the gases For gas phase sources such as NH3, H2 and N2, which are stored in pressurized cylinders, the MFC can be solely used to control the molar mass flow rates precisely The maximum flow rates for NH3, H2 and N2 are 20 L, 10 L and 10 L, respectively For solid or liquid phase sources such as metalorganic precursors of TMGa and TMIn, which are stored in the stainless-steel cylinders called bubblers, the exact amount of sources is also controlled by the MFC, as well as the pressure controller and the temperature controller (thermal bath) More specifically, the partial pressure of the source vapor is actually regulated by precisely controlling the 42 Chapter Growth and Characterizations temperature of the metal organic source bubbler, and the temperature of metalorganic sources can be regulated by means of precisely controlling the temperature of the thermal bath where the metalorganic source bubbler is held In addition, the pressure of the bubbler is regulated by the pressure controller Thus, a controlled amount of metalorganic precursor can be transported by controlling the exact amount of abducting gas flow through the bubbler using the MFC: FMO = Fabduct × Pvapor Pbubbler − Pvapor , (2.1) where FMO is the molar flow rate per minute of the metalorganic precursor, Fabduct is the molar flow rate of abducting gas, Pvapor is the pressure of the source vapor and Pbubbler is the pressure of the bubbler In our system, there are two TMGa cylinders and two TMIn cylinders With the control by MFC, the maximum flow rates are 50 sccm from the TMGa#1 cylinder, 200 sccm from the TMGa#2 cylinder, 200 sccm from the TMIn#1 cylinder, and 1000sccm from the TMIn#2 cylinder It should be noticed that, at normal operating temperatures, the TMGa is liquid and TMIn is solid, and the bubbler temperatures for TMGa and TMIn are -10 oC and 30 oC, respectively The other important part of the gas handling system is the gas mixing system, or the manifold At the manifold, the individual reactants are either combined and switched to the growth chamber or vented to the exhaust line The manifold also keeps the delivery pressures of the growth chamber and exhaust line in balance This pressure balance is very important to stabilize the gas flow to obtain the atomically abrupt interfaces in the multi-layer heterostructures In our EMCORE D125 MOCVD system, there are three manifolds, one for the group-V precursor (NH3) and carrier 43 Chapter Growth and Characterizations gases (H2 or/and N2), the second for the trimethylaluminum (TMAl) precursor, and the third for the other precursors The purpose of separating TMAl from other alkyl precursors is to prevent parasitic reactions of Al because Al is very reactive, and TMAl is not used in this work A push line that always injects H2 is also added to each manifold to achieve the fast delivery of the reactants through the gas tube to the growth chamber and to minimize the switching time during the growth of an abrupt interface For the reaction chamber, our EMCORE D125 MOCVD system has a vertical-geometry RDR It is made of stainless steel and is water-cooled A loadlock maintained at high vacuum with a turbo molecular pump is connected to the reaction chamber and it can transfer wafers in/out of the reaction chamber In RDR, substrates sit on a circular horizontal disk, called a wafer carrier For our system, the wafer carrier is 5.25-inch in diameter, which allows up to three 2-inch wafers to be grown simultaneously Also, the wafer carrier is supported by a spindle which is connected to a motor Thus, during the growth, the wafer carrier can rotate at a high speed, i.e., 1000 revolution per minute (rpm) in our growth cases This fast rotation rate during the growth improves the gas flow hydrodynamics over the wafer carrier, leading to the improvement of the growth rate uniformity across the wafer surface In addition, it is noticed that, for the growth of GaN-based quantum wells and quantum dots, the uniform thickness and composition are essential In the reaction chamber of our system, the top flow flange consists of several inlets, through which the precursors and carrier gases are injected downward Upon arrival in the growth 44 Chapter Growth and Characterizations chamber, the separated alkyl and hydride source flows are distributed to different zones in the top flow flange The alkyl flow is distributed among the inner and outer zones, while the hydride and carrier gas flow is distributed among the inner, middle, and outer zones The distribution of alkyl flows to each zone is controlled with two MFCs, while the distribution of hydride and carrier gas flows is adjusted by three needle valves Hence, a customized flow distribution can be created and optimized with setting the MFCs and the individual needle valves The uniformity of growth rate and composition can therefore be controlled and optimized by setting the flow distribution Besides the careful control of the growth rate and the composition, accurate and uniform control of the temperature is also critical For the growth of quantum wells, the small oscillation in the temperature may result in the non-uniformity of the growth rate as well as the composition The careful temperature control is even more important for the growth of quantum dots, because the temperature window for the formation of quantum dots is rather narrow Hence, in our system, a two-zone resistive heater system is used, and it consists of an inner heater and an outer heater, providing uniform temperature across the wafer carrier [Walker1995] Also, each heater has its own power supply, proportional-integral-differential (PID) temperature controller, and a thermocouple The thermocouples are installed under the heaters, which measure the corresponding heater temperature The two heaters, therefore, can be controlled separately for better growth temperature uniformity [Gurary1994] In addition, since the thermocouples not directly measure the temperature of the 45 Chapter Growth and Characterizations wafer carrier, two pyrometers are also installed on the top of the reaction chamber to monitor the temperatures of the inner and outer parts of the wafer carrier separately Because the thermocouples only control the heater temperature and the pyrometers only monitor the temperature of the wafer carrier without the ability to directly control the heater temperature, a temperature calibration run is usually conducted before an actual run to ensure the better temperature control during the actual growth The temperature calibration run is exactly the same as the actual run except that no group-III precursors flow into the growth chamber Thus, with the usage of two-zone resistive heater system, the growth temperature can be controlled uniformly cross the wafer carrier up to about 1100oC Finally, the pumping and low-pressure control system (or vacuum system) is an important part for controlling the pressure of the reaction chamber The vacuum system basically consists of three parts, i.e., (1) a rotary pump, (2) two stage gas exhaust particle filters, and (3) a throttle valve The reaction chamber pressure is controlled by adjusting the position of the throttle valve flap with a MKS pressure controller unit The exhaust gases are pumped out of the reaction chamber by the pumps The exhaust particles are trapped by the two-stage particle filters The toxic parts in the exhaust gases are then absorbed in the toxic gas absorber unit before the exhaust gases are released into the environment Thus, accurate thickness and composition control in the growth of GaN-based quantum wells and quantum dots can be effectively achieved through the careful control of the flow rate, temperature and the pressure in our MOCVD system 46 Chapter Growth and Characterizations The uniform growth temperature and pressure during the MOCVD growth will also help improve the crystal quality of the GaN-based structures 2.2 Plasma enhanced chemical vapor deposition (PECVD) The PECVD technique － i.e., film growth using gas phase precursor activated in a glow discharge environment － has been widely employed in semiconductor device fabrication for several decades It is heavily used for deposition of nitrides and oxides The typical diagram of a standard PECVD system is shown in Fig 2.3 (a) Figure 2.3 Diagrams of (a) stand PECVD and (b) PECVD with a plasma box (Courtesy of Unaxis Semiconductor) In a PECVD system, the plasma is generated by applying a radiative-frequency (RF) field to a low-pressure gas The electrons in the reactor, which gain sufficient energy from the electric field, collide with gas molecules Then 47 Chapter Growth and Characterizations Figure 2.9 (b) Figure 2.9 (c) 63 Chapter Growth and Characterizations Figure 2.9 (d) Figure 2.9 AFM images from an InGaN/GaN quantum well sample surface (a) 3-dimensional image, (b) flatten analysis, (c) roughness analysis, and (d) section analysis Finally, it should be noticed that although the AFM can measure the height variation with sub-angstrom resolution, the accuracy of the surface profile still depends on the scanning rate, as well as the size of the tip compared to the sizes and aspect ratios of the features Especially when both the feature sizes and the aspect ratios are small, fine scanning has to be conducted at a low speed to achieve the better resolution In this work a typical scanning rate of 1Hz (one sample per second) is used, and there are 256 or 512 samples in each scanning Here, one sample represents one line of scanning, which then consists of 256 or 512 sampling points 2.3.4 Scanning electron microscopy (SEM) The SEM technique is a powerful and versatile instrument It has many advantages over traditional microscopes The SEM has a large depth of field, which 64 Chapter Growth and Characterizations allows the simultaneous observation of the cross-sectional and surface images It also has much higher resolution, enabling the closely spaced (separated by several tens of nanometers) small features to be resolved Also, because the SEM uses electromagnets rather than lenses, the researcher has much more control on the degree of magnification All of these advantages, as well as the strikingly clear images, make the SEM one of the most useful instruments in materials research today The schematic setup of an SEM system is shown in Fig 2.10 In the SEM measurement, the largely magnified images are formed by using electrons instead of light A beam of electrons is produced at the top of the microscope by an electron gun The electron beam follows a vertical path through the microscope, which is evacuated down to the pressure of ~ 10-4 Pa in this work So the electron beam can travel through electromagnetic fields and lenses, which focus the beam down toward the sample Once the beam hits the sample, various signals are generated, which will be discussed in the next paragraph, and the detection of the specific signals can produce the surface images or information of the elemental composition in the sample 65 Chapter Growth and Characterizations Figure 2.10 Schematic setup of an SEM system [Courtesy of Iowa State University SEM homepage] As illustrated in Fig 2.11, the three main signals generated during the SEM measurement are secondary electrons, backscattered electrons, and X-rays All these signals can provide a great amount of information about the sample properties For secondary electrons, they are emitted from the atoms on the top surface and therefore can produce a readily interpretable image of the surface The contrast of the image is determined by the sample morphology A high resolution image can therefore be obtained because of the small diameter of the secondary electron beam The backscattered electrons are primary beam electrons which are ‘reflected’ from atoms in the solid The contrast of the image is then determined by the atomic number of the 66 Chapter Growth and Characterizations elements in the sample The image will therefore show the distribution of different chemical phases in the sample However, because these electrons are emitted from a certain depth in the sample, the resolution of such images is not as good as that from secondary electrons In addition, the interaction of the primary beam with atoms in the sample also causes the electron orbital transitions, resulting in the emission of the X-ray The emitted X-ray has an energy characteristic of the parent element Detection and measurement of this energy therefore allow for the elemental analysis (Energy Dispersive X-ray Spectroscopy or EDX), providing rapid qualitative and quantitative analysis of the elemental composition with a sampling depth of 1-2 μm X-rays may also be used to form mapping or line profiles, and show the elemental distribution in the sample surface Figure 2.11 Summary of the ranges of backscattered electrons, secondary electrons, X-ray, and Auger electrons for electrons incident on a solid [Goldstein1984] The SEM images presented in this thesis were taken from a JEOL JSM-6700F field emission SEM The extremely fine electron source generated from the field emission gun makes it possible to obtain images of much higher resolution than those from a conventional SEM High magnifications in excess of 200,000 times 67 Chapter Growth and Characterizations are obtainable, which translate to the resolution of ~ 1.5 nm at an accelerating voltage of 30 kV Also, the high brightness of this source allows for the high resolution imaging (in excess of 100,000 times) and characterization of beam sensitive materials (e.g fragile plastics and integrated circuits) at very low accelerating voltages (ie 0.2 to kV) In addition, because the SEM utilizes vacuum conditions and uses electrons to form an image, special sample preparations may be required if the sample is not conductive To improve the surface conductivity and avoid the charging effect, a thin layer of gold is usually sputtered on the sample surface But the gold coating will change the material contrast in the images In our work, because the ultra-high magnifications were not required, and the GaN-based semiconductor material has sufficient conductivity, the sample was only grounded by attaching the surface to the ground terminal, i.e., the sample surface was connected to the specimen holder through several copper tapes In this case, no severe charging effect was observed, and the material contrast in the images remains unchanged Compared to the AFM technique, the SEM technique gives more faithful reproduction of the surface, because its resolution is mainly limited by the conductivity of the sample Also, SEM can be used to obtain the cross-sectional images of the sample, instead of the height variation profile taken by the AFM technique Hence, by conducting the SEM measurements, the quality of the interfaces can be analyzed and the film thickness can be estimated But it should be noted that, when using the SEM for surface profile study, no roughness rms values can be 68 Chapter Growth and Characterizations obtained 2.3.5 Transmission electron microscopy (TEM) The TEM technique serves as one of the most direct means of observing submicron structures It is employed in this work for the observation of GaN-based quantum structures The schematic setup of a TEM system is shown in Fig 2.12 In the TEM measurements, electrons are accelerated by the same method as that in the SEM, but they travel through the specimen The specimen must be very flat and thin (approximately 0.1 μm or less) Similar to a slide projector, more electrons can pass through the less dense regions, and fewer electrons pass through the regions that are denser The diffracted electron beams are brought to focus in the back focal plane of the objective lens, which is the plane of the diffraction pattern When the microscope is operated in the diffraction mode, the diffraction lens is focused on the back focal plane and the subsequent lenses project a magnified diffraction pattern on the fluorescent screen, allowing the user to observe the diffraction pattern The objective lens also produces an inverted image of the specimen in the first image plane If the diffraction lens is focused on this plane, the microscope is in the imaging mode and produces magnified image on the florescent screen for observation 69 Chapter Growth and Characterizations Figure 2.12 Schematic setup of a TEM system (Courtesy of Rensselaer Polytechnic Institute TEM homepage) The TEM technique has great advantages It can image individual atoms and their relative positions, and also give compositional information over an area of interest There are, however, drawbacks to this technique as well Prior to the imaging, tedious sample preparation steps from polishing, dimpling and ion milling have to be conducted, because the sample has to be very flat and thin for observation, as mentioned above It usually requires a skilled operator to handle the actual imaging process too Furthermore, TEM is a destructive characterization technique as the sample would be destroyed after sample preparation and imaging The TEM images, in this work, were obtained using a Philips FEG 300CM 70 Chapter Growth and Characterizations system operating at 300 kV During the measurements, there are basically two different imaging modes: bright field and dark field In the bright field mode, as shown in Fig 2.13 (a), the electron beam would be incident along the optical axis The objective aperture would be located to allow only the center beam to pass The objective aperture would screen off the diffracted beam, and the image is formed by the center beam As the resulting image has a background which is generally bright, it is referred to as the bright field image In the dark field mode, as shown in Fig 2.13 (b), the objective aperture would be placed to block off the center beam Thus, only the diffracted beam would form the image In this mode, crystal lattice images may be obtained, utilizing the multiple beam interference imaging mode [Horiuchi1994] Also, with the active beam vector of b , which can be obtained using the diffraction mode, if a dislocation has the Burgers vector of g , it holds that this dislocation will be shown out of contrast (invisible) when g ⋅ b = [Shiojima2000] Other dislocations will be visible and shown as bright lines on the dark background The dislocations in our GaN-based quantum structures were therefore analyzed in this dark field mode by applying different active beam vectors For GaN-based materials, the Burgers vector for pure edge dislocations is b = [1120] , and the Burgers vector for screw dislocations is b = [ 0001] Hence, in the dark field mode, only the screw and mixed dislocations will be visible when the active beam vector is g = 0002 ; and only the pure edge dislocations will be visible when the active beam vector is g = 1120 71 Chapter Growth and Characterizations (a) (b) Figure 2.13 Two different modes of TEM: (a) bright field and (b) dark field (Courtesy of Electron Microscopy website, Swiss Federal Institute of Technology Zurich) In the analysis of TEM images, two types of contrast must be considered, which are illustrated in Fig 2.14 The first type is termed as the scattering contrast This is the result of mass difference or thickness difference in the sample With different materials, the absorption may be different due to the difference in densities This leads to the formation of a mass contrast image Then in thicker regions, fewer electrons would be transmitted, therefore resulting in a thickness contrast image The second type of contrast is termed as the phase contrast This is formed 72 Chapter Growth and Characterizations by interference of two or more electron beams In this sense, fringe patterns would be produced if crystalline materials were imaged (a) (b) (c) Figure 2.14 Contrasts in TEM images: (a) mass contrast, (b) thickness contrast, and (c) lattice fringe image (Courtesy of Electron Microscopy website, Swiss Federal Institute of Technology Zurich) 2.3.6 X-ray diffraction (XRD) The XRD is a well-developed, versatile and powerful technique for non-destructive ex-situ characterization of semiconductor structures Structure information, such as crystalline quality, strain relaxation, and lattice constant can be accurately obtained with high resolution XRD (HR-XRD) [Fewster1993] Figure 2.15 shows the typical components and angles of the goniometer for a θ-2θ X-ray diffractometer The principle of the XRD is according to the Bragg’s law: d hk l sin θ B = nλ (2.5) where dhkl is the spacing of the lattice planes with Miller indices {h k l}, θB is the corresponding Bragg angle, and λ is the wavelength of X-ray radiation The spacing of the lattice planes with Miller indices {h k l} can be calculated with following equation for wurtzite III-nitrides: dhk l ⎛ h + k + hk 3l ⎞ = ⎜ + 2⎟ ⎜ a2 4c ⎟ ⎠ ⎝ − (2.6) where a and c are lattice parameters of the wurtzite III-nitrides 73 Chapter Growth and Characterizations The XRD measurements in this work were carried out using a Philips X’Pert MRD HR-XRD system It includes a copper (Cu) anode X-ray tube as the X-ray source, a Ge (220) symmetric 4-crystal monochromator as the incident beam collimator to select the Cu Kα1 radiation (λ = 1.54Å) and to confine the angular divergence to be about 12 arcsec, a sample stage (the goniometer), and a proportional detector with an analyzer (receiving slit) As shown in Fig 2.15, the sample stage can rock (ω scan), tilt (χ scan), or rotate (φ scan) around the respective axes It can also move along the x-y-z directions to position the sample The χ axis allows for 180o tilting; and the φ axis normal to the sample surface can rotate through 360o Moreover, in the XRD system, the sample stage is also a precision goniometer with optical encoders on the axes which can eliminate angular uncertainties on the two main axes of ω and 2θ Ф Incident Beam Slit χ ω Specime n X-ray Tube ω 2θ =0 o 2θ Ge(002) 4-crystal Monochromator Recie ving Slit Detecor 2θ =90o Figure 2.15 Schematic diagram of the goniometer for a θ-2θ x-ray diffractormeter (Courtesy of PANalytical) 74 Chapter Growth and Characterizations The most common XRD scan modes include the angular-scan (ω scan) and the radial scan (ω-2θ or 2θ-ω scan) In the angular-scan (ω scan), the detector (2θ) is fixed in position and the sample (ω) is rocked around the position which satisfies the Bragg condition Hence, the angular-scan is often referred as rocking curve scan In practical cases, as the crystal quality of semiconductor structures cannot be completely perfect, the diffraction peaks will then be broadened In general, the broadening of the diffraction peaks from the ω scan is related with the following possible factors: (1) gradient of the lattice parameter, (2) crystal mosaicity, (3) disruption of the X-ray coherence by extended defects, and (4) bending of the sample [Leszczynski1998] Therefore, the rocking curve is very useful for evaluating the crystalline quality of GaN-based semiconductor materials [Heinke2000], which normally have several threading dislocations (TDs) contents inside the films, namely, pure edge TDs, screw TDs, and mixed TDs In the radical scan (ω-2θ or 2θ-ω scan), the detector rocks twice as fast in the same direction as the sample rocks Precise lattice constant measurement can be performed by this technique Radical scan is commonly used to determine the strain and stress levels, layer composition of III-nitrides, and periods of multiple quantum well or superlattice structures [Fewster1998] 2.3.7 Pump-probe spectroscopy Pump-probe spectroscopy is a technique to study the dynamic processes in materials or chemical compounds by means of spectroscopic techniques With the help of ultrafast pulsed lasers, it is possible to study those short-lived events occurring 75 Chapter Growth and Characterizations on time scales as short as 10-14 seconds As sketched in Fig 2.16, in a pump-probe experiment, the output pulse train from an ultrafast laser is divided into two beams: the sample is excited by one pulse train (pump) and the changes it induces in the sample are probed by the second pulse train (probe), which is suitably delayed with respect to the pump beam The pump beam is sent at regular intervals that are longer than the response of the sample to ensure that there is no overlap of excitation events from the current pump beam and the previous pump beam Due to the excitation of the pump beam, the sample is disturbed from its equilibrium position and returns after some time to its initial position The probe beams are sent in with different time delays to “probe” the real-time response of the sample after the disturbance For example, to study the absorption property of a saturable absorber, the absorption of the sample will first be saturated by the strong pump beam; then the absorption recovery process is studied in real-time by the probe beams If the absorption saturation has not been recovered in the saturable absorber, the probe beam will pass through the sample with little absorption If the absorption saturation has been fully recovered, the probe beam will experience strong absorption while passing through the sample Therefore, by analyzing the probe beams arriving at the sample with different time delays, the absorption recovery process of a saturable absorber can be studied 76 Chapter Growth and Characterizations Figure 2.16 Schematic diagram of a pump-probe setup (Courtesy of Università degli Studi della Tuscia) In this work, we used the pump-probe technique to study the recovery times of GaN-based saturable absorbers as well as their nonlinear transmittance property The detailed experimental setup will be presented later in Chapter 77 ... second) is used, and there are 25 6 or 5 12 samples in each scanning Here, one sample represents one line of scanning, which then consists of 25 6 or 5 12 sampling points 2. 3.4 Scanning electron microscopy... excition transition and bound exciton transition Eb is the free exciton binding energy and Ebx is the bound exciton binding energy In addition to the radiative transitions, there are also non-radiative... certain defect level, the binding energy is determined by the bandgap, free exciton binding energy, and the bound exciton binding energy, as illustrated in Fig 2. 6 Figure 2. 6 Illustration for