超快時間解析反射技術研究氮砷化銦鎵單層量子井

全文

(1)國立交通大學光電工程研究所碩士論文. 超快時間解析反射技術研究氮砷化銦鎵單層量子井. The Ultrafast Time-Resolved Photoreflectance Measurement of InGaAsxN1-x Single Quantum Wells. 研. 究. 生：林. 指導教授：謝. 盈文. 秀峰教授. 指導教授：. 中華民國九十四年六月. i.

(2) 超快時間解析反射技術研究氮砷化銦鎵單層量子井 The Ultrafast Time-Resolved Photoreflectance Measurement of InGaAsxN1-x Single Quantum Wells. 研究生：林盈秀. Student: Ying-Hsiu Lin. 指導教授：謝文峰教授. Advisor: Dr. Wen-Feng Hsieh. 國立交通大學光電工程研究所碩士論文 A Thesis Submitted to Institute of Electro-Optical Engineering College of Electrical Engineering and Computer Science National Chiao Tung University in Partial Fulfillment of the Requirements for the Degree of Master in Electro-optical Engineering June 2005 Hsin-chu, Taiwan, Republic of China. 中華民國九十四年六月 ii.

(3) 超快時間解析反射技術研究氮砷化銦鎵單層量子井研究生：林盈秀. 指導教授：謝文峰教授. 國立交通大學光電工程研究所. 摘要. 在本論文中，我們比較由金屬有機化學氣相沈積法(MOCVD)成長的砷化銦鎵和氮砷化銦鎵單層量子井的超快載子動力學。從螢光光譜的量測，我們可以發現氮砷化銦鎵的能隙因為氮的加入比砷化銦鎵的能隙小很多，可以發出波長為 1.3μm 的光。我們利用飛秒時析反射量測研究載子的鬆弛過程。比較這兩個樣品的量測結果，我們發現在氮砷化銦鎵的載子生命期比在砷化銦鎵的短。這可能是因為載子鬆弛過程中，載子被加入氮所造成的缺陷捕捉所導致。. I.

(4) The Ultrafast Time Resolved Photoreflectance Measurement of InGaAsxN1-x Single Quantum Wells Student: Ying-Hsiu Lin. Advisor: Dr. Wen-Feng Hsieh. Institute of Electro-Optical Engineering National Chiao Tung University. Abstract The comparative analysis of the ultrafast carrier dynamics of metal-organic chemical vapor deposition (MOCVD) grown In0.4Ga0.6As and In0.4Ga0.6As0.995N0.005 single quantum wells (SQW) is reported.. From photoluminescence (PL). measurement we conclude that the incorporation of a low content N in InGaAsN reduces the band-gap energy significantly and allows emission wavelengths shift to 1.3 µm.. Carrier dynamics were investigated by femtosecond time-resolved. photoreflectance measurement at 800 nm with various pump intensities. Comparison with these two samples indicates that the carrier relaxation in the In0.4Ga0.6As0.995N0.005 SQW is dominated by defect traps associated with the N incorporation.. II.

(5) 致謝時光飛逝，兩年的碩士生活，隨著這本論文寫出來，已接近尾聲。兩年來不管是專業知識方面還是待人處事的態度都有進步，當然要感謝最多的就是我的指導教授謝文峰教授。如果沒有老師不厭其煩的教導，我想我不會得到這麼多。另外一方面老師帶領學生的風氣也很棒，讓實驗室不會死氣沈沈，而是在努力研究之餘還能充滿歡笑！隨著兩年的朝夕相處，和實驗室的大家，從不熟悉到熟悉，從有形象到不顧形象，現在，要離開還真有點捨不得呢！家弘學長、政哥、奎哥、小戴學長、智章學長、鄭信民學長、黃董、維仁、松哥、碩一唯一的學弟：展榕、實驗室另一朵花：潘晴如學姐，謝謝你們在實驗上、課業上和生活上的幫助。另外，同甘苦共患難的碩二的同伴：國峰、俊毅、家斌、小豪、文勛，謝謝你們在實驗和課業方面的互相切磋討論。我會一直一直記得國峰幫我殺了很多小強，和小毅非常直接的意見的！真的是謝謝大家，讓我的碩士生涯如此豐富。本論文得以完成，要感謝的人還很多：子安學長、亞銜學長、伯宗學弟在實驗技巧上的說明、贊伯出借儀器、親愛的爸媽的贊助讓我生活不虞匱乏，我只能說，沒有你們，就沒有我啊！再多的言語也沒有辦法表達清楚我心中滿溢的感謝！最後感謝國科會計畫 NSC-93-2112-M-009-035 對於此研究的支持以及贊助。. 盈秀 (阿笑). III. 風城交大筆 94.7.5.

(6) Content Abstrate(in Chinese)……………………………………………….…………………..I Abstrate (in English)…………………………………………………………………..II Acknowledgements…………………………………………………………………..III Content……………………………………………………………………………….IV List of figures………………………………………………………………………..VI List of Tables……………………………………………………………………..…VII. Chapter 1 1.1. Introduction.......................................................................1. Introduction..........................................................................................................1. 1.1.1. InGaAsN/GaAs Heterostructures.......................................................................1. 1.1.2. Semiconductor Carrier Dynamics......................................................................1. 1.1.3. Femtosecond Time-Resolved Spectroscopy ......................................................2. 1.2. Aim of This Thesis...............................................................................................3. Chapter 2. Femtosecond Time-Resolved Systems ............................7. 2.1. Pump-Probe Transmission/Reflection Spectroscopy...........................................7. 2.2. Time-Resolved Photoluminescence (PL) Spectroscopy ......................................8. 2.3. Time-Resolved Terahertz-Probe Spectroscopy....................................................9. Chapter 3 3.1. Ultrafast Spectroscopy of Semiconductor ....................12. Basic Concepts...................................................................................................12. 3.1.1. Scattering Processes in Semiconductors..........................................................12. 3.1.2. Carrier Relaxation in Photoexcited Semiconductors .......................................13. 3.2. Ultrafast Phenomena of Semiconductors...........................................................15. 3.2.1. Bulk GaAs........................................................................................................15. 3.2.2. InGaAsN/GaAs ................................................................................................16. 3.3. Carrier-Induced Change in Refractive Index .....................................................18 IV.

(7) 3.3.1. Band Filling .....................................................................................................18. 3.3.2. Bandgap Renormalization................................................................................19. 3.3.3. Free-Carrier Absorption ...................................................................................21. Chapter 4 4.1. Experiments.....................................................................23. InGaAsN Single Quantum Well Structures........................................................23. 4.1.1. Mode-Locked Ti:sapphire laser .......................................................................24. 4.1.2. Autocorrelator ..................................................................................................24. 4.1.3. Experimental Setup of Femtosecond Time-Resolved. 4.1.4. Zero Delay Point ..............................................................................................25. Chapter 5. Measurement ...........24. Results and Discussion ...................................................31. 5.1. Ultrafast Time Resolved Photoreflectance of Bulk GaAs .................................32. 5.2. Ultrafast Time Resolved Photoreflectance of InGaAs1-xNx SQW .....................34. Chapter 6. Conclusion and Perspectives..........................................40. Appendix I. Fitting results of bulk GaAs ...........................................41. Appendix II. Fitting results of SQW (I)..............................................42. Appendix III. Fitting results of SQW (II)...........................................43. References................................................................................................44. V.

(8) List of Figures Fig. 1-1 The electron-hole pair creation following excitation of a semiconductor with laser radiation of energy hω0 ..........................................................................4 Fig. 1-2 The block diagram illustrating the follow of energy in a photoexcited semiconductor. ................................................................................................4 Fig. 1-3 Illustrating different time-scale measurements require different time-resolution tools. ......................................................................................5 Fig. 1-4 Schematic diagram of the shortest pulses reported versus year .......................5 Fig. 2-1 Schematic for a generic setup for a two-beam nonlinear experiment ............10 Fig. 3-1 Four temporally-overlapping relaxation regime in photoexcited semiconductors. ............................................................................................22 Fig. 4-1 Cross-sectional schematic conduction band diagram of the In0.4Ga0.6As1-xNx SQW..............................................................................................................27 Fig. 4-2 HRTEM image of In0.4Ga0.6As0.995N0.005/GaAs SQW....................................27 Fig. 4-3 Result of photoluminescence (PL) spectra measurement. .............................28 Fig. 4-4 The optical beam path within Coherent Mira-900. ........................................28 Fig. 4-5 Sketch diagram of the intensity autocorrelator...............................................29 Fig. 4-6 The autocorrelation trace................................................................................29 Fig. 4-7 Femtosecond time-resolved photoreflectance measurement system. ............30 Fig. 5-1 Pump-intensity dependent time-resolved photoreflectance of bulk GaAs.....36 Fig. 5-2 The measured ∆R peak amplitude versus pump power in bulk GaAs. ..........36 Fig. 5-3 The normalized ∆R traces of bulk GaAs with various pump power. .............37 Fig. 5-4 The carrier lifetime versus pump power in bulk GaAs. .................................37 Fig. 5-5 The pump-intensity dependent time-resolved photoreflectance of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW at 800 nm. ........................................38 Fig. 5-6 The measured ∆R peak amplitude versus pump power of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW....................................................................39. VI.

(9) List of Tables Table 1-1 Fundamental processes in semiconductors....................................................6 Table 1-2 List of the application of time-resolved spectroscopy. ..................................6 Table 5-1 List of carrier lifetimes of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW with various pump power. ................................................................39. VII.

(10) Chapter 1. Introduction. 1.1 Introduction 1.1.1. InGaAsN/GaAs Heterostructures. InGaAsN/GaAs heterostructures are attracting much interest because of their highly nonlinear effects upon nitrogen incorporation [1] and their potential for long wavelength optoelectronic device application [2][3][4]. The large band offsets in InGaAsN quaternary alloy lead to the improvement of temperature performance, better than conventional InP-based materials [5][6]. The challenge to demonstrate devices of superior performance is at present dependent on improving the optical and structural quality of the dilute nitride quantum wells.. Structural imperfections are. most likely responsible for the significant increase of the monomolecular recombination rate in InGaAsN compared with InGaAs [7].. In addition, the carrier. dynamics is strongly affected by N incorporation, as shown in the few studies reported in the literature [8][9][10][11].. Consequently, understanding the influence. of nitrogen in InGaAsN is very valuable.. 1.1.2. Semiconductor Carrier Dynamics. On the past few decades, there have been tremendous advancements in the field of ultrafast carrier dynamics in semiconductors [12]. The driving force behind this movement other than the basic fundamental interest is the direct application of semiconductor devices and the endless need for the faster response and the faster processing of information.. To improve and to develop high-speed microelectronics. devices, there must be a basic understanding of the various dynamical processes in the 1.

(11) semiconductors.. Therefore, the excitation of semiconductors out of their equilibrium. and the subsequent relaxation processes with various rates become a key area of semiconductor research. Linear absorption takes place when the photon energy is larger than the fundamental energy gap of the semiconductor as shown in Fig. 1-1. Following optical excitation, electrons and holes undergo spatial and temporal evolution with characteristic times which depend on the various relaxation processes. Table 1-1 lists the fundamental processes and Fig. 1-2 shows the energy relaxation follows in a photoexcited semiconductor.. With the development of lasers that can. generate pulses as short as a few femtoseconds the excitation and the subsequent probing of semiconductors on an ultrafast time scale have become routine.. 1.1.3. Femtosecond Time-Resolved Spectroscopy. Fig. 1-3 illustrating that different time-scale measurements require different time-resolution tools.. For measuring the velocity of a running tortoise, it is enough. to use minute-resolution clock.. For measuring the velocity of a running human, a. second-resolution stopwatch is required.. However, only the femtosecond-resolution. pump-probe technique with ultrashort laser could achieve the measurement of electron lifetime between various energy bands.. In the past decade, the technology. of generating and detecting short optical pulses has advanced substantially.. As. shown schematically in Fig. 1-4, the width of optical pulses has fallen by more than 3 orders of magnitude since 1966. Systems such as the colliding pulse mode-lock (CPM) laser developed by Fork and Shank [13] have made possible ultrafast semiconductor measurements with femtosecond resolution.. The applications of. time-resolved spectroscopy in several fields are listed in Table 1-2.. 2.

(12) 1.2 Aim of This Thesis In this thesis, we report the ultrafast time-resolved photoreflectance of In0.4Ga0.6As1-xNx/GaAs single quantum wells.. We compare the ultrafast. time-resolved photoreflectance measurements in samples differing in the N incorporation (x = 0 % and x = 0.5 %). photoreflectance of bulk GaAs.. We also measure the ultrafast time-resolved. All the results of our measurements are shown and. discussed in chapter 5. In chapter 2 and chapter 3, we briefly describe the ultrafast time-resolved pump-probe system and ultrafast spectroscopy of semiconductor. In chapter 4, the investigated sample structure and the experimental setup of ultrafast time-resolved measurement are discussed in details.. Finally, the conclusion and perspectives are. both given in chapter 6.. 3.

(13) Fig. 1-1 The electron-hole pair creation following excitation of a semiconductor with laser radiation of energy hω0 . Energy relaxation follows via optical phonon emission (hω ph ) .. hω 0. hω0 − E g. Eg. Tc > Tl. Fig. 1-2 The block diagram illustrating the follow of energy in a photoexcited semiconductor.. 4.

(14) Fig. 1-3 Illustrating different time-scale measurements require different time-resolution tools.. Fig. 1-4 Schematic diagram of the shortest pulses reported versus year. (Ref. 12). 5.

(15) Table 1-1 Fundamental processes in semiconductors.. Table 1-2 List of the application of time-resolved spectroscopy.. Physics. Intraband &intersubband relaxation Phonon & Electron dynamics Energy gap dynamics Ionization dynamics Coherent polarization & population dynamics Laser-induced surface disordering Characterization of excited states dynamics Exciton dynamics. Chemistry. Solvent-solute interaction Polymer structure and dynamics Reaction kinetics in micelles Identification and detection of single molecule Electron and proton transfer Molecular vibration. 6.

(16) Chapter 2. Femtosecond Time-Resolved Systems. The advantages in the generation of ultrashort laser pulses in the past decade have allowed realizing the long standing dream of probing ultrafast carrier dynamic directly in the time domain.. The direct measurement of the temporal evolution of. the scattering/relaxation processes with fs time-resolution has become possible. Several femtosecond time-resolved techniques will be described later.. 2.1 Pump-Probe Transmission/Reflection Spectroscopy The most common adopted ultrafast spectroscopy technique is the pump-probe transmission/reflection technique. A schematic diagram of a generic pump-probe setup is shown in Fig. 2-1.. The probe beam is typically much weaker than the pump,. and the spot size of the probe at the sample is ideally considerably smaller than that of the pump so that it probes a region of uniform photoexcited density.. For. pump-probe transmission/reflection spectroscopy, one generally measures the change in the transmitted/reflected probe intensity induced by the pump as a function of the time delay between the pump and the probe pulses by chopping the pump beam and using a lock-in amplifier. The principle of monitoring an ultrafast event by the pump-probe technique is illustrated in Fig. 2-2, where the pump and probe pulses are synchronized.. For an. event, n(t), to be measured via pump-probe technique, it’s lifetime must not exceed the separating time. between two pulses (13 ns in our case).. The. transmitted/reflected probe intensity from sample will vary with n(t) while the probe pulses temporally overlapped with n(t).. However, this change due to n(t) is very. small, typically between 10-4~10-7, and is very difficult to detect directly under noisy 7.

(17) background (including laser noise, electronic noise, mechanic vibration etc.) by photodiode.. Hence, the phase-lock technique was usually used to provide a. background-free signal. The small signal was modulated by an optical chopper in order to eliminate the quite-large noise. As noted, the modulation was commonly applied to the pump pulse train to make sure that the signal merely came from n(t) induced by pump pulses. Then, the probe pulse train was also modulated by n(t) on the constant intensity of the probe pulse train, i.e. a AC signal (∆I(t)) is added to a DC signal (I0(t)). This signal was detected by the photodiode and fed to the lock-in amplifier, which was phase-locked to the optical chopper.. By varying the time delay t between the pump and probe pulses,. ∆I(t) would change as a function of time delay.. Therefore, the temporal evolution of. n(t) could be measured by assuming that the relation between n(t) and ∆I(t) is linear.. 2.2 Time-Resolved Photoluminescence (PL) Spectroscopy Time evolution of photoluminescence can also be obtained by scanning the relative delay between the two laser beams.. The time-resolved PL technique is used not only. to investigate carrier relaxation process, but also carrier transport process in semiconductors.. The simplest method is to measure photoluminescence dynamics. directly with a fast detector.. The schematic diagram of time-resolved PL is shown in. Fig. 2-1.. The best technique for direct measurement is the used of streak cameras. [14][15].. Two-dimensional streak cameras which provide spectral and temporal. information simultaneously have become very attractive and versatile tools for the time-resolved PL Spectroscopy.. Time-resolved photoluminescence spectra can also. be obtained by keeping the delay fixed and scanning the angle of the sample.. Some. time-resolved PL experiment results of InGaAs(N) are discussed in Ref. [8][11][16]. 8.

(18) 2.3 Time-Resolved Terahertz-Probe Spectroscopy Ultrafast lasers can generate coherent mm- and sub mm-wave radiation by a variety of different techniques.. This radiation can be collimated and transmitted over. reasonable distances, and can be detected using optically-gated photoconductive antennas. By adjusting the delay between the THz signal and the gating pulse, information about the amplitude and the phase of the THz signals can be obtained. The advantage of this technique by using an optical THz probe is that the THz probe photons have energy in the meV range and cannot substantially modify the electronic distribution in the sample.. Thus the THz radiation propagates practically without. absorption and dispersion in unexcited semiconductors and can detect even low concentrations (～1014 cm-3).. This method investigated in GaAS is discussed in Ref.. [17][18][19]. A novel experimental scheme employing an internal reflection of the THz pulses is shown in Fig. 2-3 [20]. First, a THz pulse incident on the input surface is led to propagate through the sample in equilibrium.. It is followed by the optical pump. pulse, which generates free carriers near below the surface.. The part of the THz. pulse which passes through directly is not affected by the pump pulse and it is used as a reference.. The part reflected on the output surface serves as a probe:. It. propagates back to the input surface, which has been excited meanwhile, and reflects from its inner side.. Then, it can be detected as the echo, carrying a signature of the. excited input surface (and equilibrium output surface).. 9.

(19) Fig. 2-1 Schematic for a generic setup for a two-beam nonlinear experiment. Fig. 2-3 [20]).. The time sequence of pulses and their propagation through the sample (Ref.. 10.

(20) n (t). Pump pulses. time 13 ns Probe pulses time t (delay) Optical chopper @ 3000 Hz at pump pulses. I0 (t). ∆ I (t). 0.33 ms. time. detecter & lock-in amplifier @ 3000 Hz ∆ I(t) / I (t) = ∆ R(t) / R (t) ∝ n (t) 0 0. Fig. 2-2 The illustration for the principle of pump-probe technique.. 11.

(21) Chapter 3. Ultrafast Spectroscopy of Semiconductor. 3.1 Basic Concepts 3.1.1. Scattering Processes in Semiconductors. The dynamics of electrons, holes, excitons and phonons is influenced by their interaction with one another, as well as with defects and interfaces of the system.. In. the following subsections, we will discuss the various scattering mechanisms [21]. (a) Carrier-Phonon Interaction Interaction of carriers with phonons plays a major role in the exchange of energy and momentum between carriers and the lattice, and hence determines the relaxation of photoexcited semiconductors.. Optical phonon scattering and intervalley. scattering are the most frequent of all the carrier-phonon scattering events.. These. processes always proceed on a picosecond time scale. On the other hand, acoustic phonon scattering of carriers is much less frequent than either optical phonon or intervalley scattering. The acoustic phonon processes take place on a nanosecond time scale. (b) Carrier-Carrier Scattering Carrier-carrier scattering determines the exchange of energy between carriers, and is primarily responsible for the thermalization of photoexcited non-thermal carriers. This process is mediated by Coulomb interaction and includes electron-electron, hole-hole and electron-hole scattering.. The difference in the masses between. electrons and holes reduces the energy change between these two species.. The rate. of carrier-carrier interactions can be found by using the Fermi golden rule of time-dependent perturbation theory.. 12.

(22) The conditions under which carrier-carrier scattering has a significant effect on the photoexcited carrier distribution function can now be discussed.. There are three. conditions that must be satisfied. First the carrier-carrier scattering rates must be greater than the generation-recombination rates.. Second, the exchanges of energy. involved in carrier-carrier scattering should be greater than the energy exchanges in carrier-phonon scattering.. Third, the carrier-carrier scattering rates should be greater. than the carrier-phonon scattering rates.. 3.1.2. Carrier Relaxation in Photoexcited Semiconductors. After a semiconductor is excited by an ultrashort pulse, it undergoes several stages of relaxation processes before it returns once again to the thermodynamic equilibrium [21].. The carrier relaxation can be classified into four temporally overlapping. regimes： (a) Coherent Regime The laser pulse can create either real or virtual excitation.. Excitations of a. semiconductor created by an ultrashort laser pulse have well-defined phase relation within the excitation. This coherent regime exhibits many interesting phenomena which are elegant manifestations of basic quantum mechanics in semiconductors. The scattering processes that destroy the coherence are extremely fast in semiconductors and pico- and femtosecond techniques are required for studying the coherent regime in semiconductors. (b) Non-thermal Regime In the case of real excitation, the distribution of the excitation after the destruction of coherence through dephasing is very likely to be non-thermal, i.e. the distribution. 13.

(23) function can not be characterized by a temperature. provides. information. about. various. processes,. Investigation of this regime such. as. carrier-carrier. or. exciton-exciton scattering that bring the non-thermal distribution to a hot, thermalized distribution. (c) Hot-Carrier Regime Carrier-carrier or exciton-exciton scattering is primarily responsible for redistributing the energy within the carrier (exciton) system, and leads to a thermalized distribution function of carrier (exciton), i.e., a distribution that can be characterized by a temperature. The temperature is usually higher than the lattice temperature and may be different for different sub-systems (electrons, holes, or excitons). The thermalization time depends strongly on many factors such as carrier density. Investigation of this hot carrier regime focuses on the rate of cooling of carriers to the lattice temperature and leads to information concerning various carrier-phonon, exciton-phonon and phonon-phonon scattering processes. (d) Isothermal Regime At the end of the hot-carrier regime, all the carriers are in equilibrium with each other, i.e., can be described by the same temperature of the lattice.. However, there is. still an excess of electrons and holes compared with the thermodynamic equilibrium. These excess electron-hole pairs recombine either radiatively or non-radiatively and return the semiconductor to the thermodynamic equilibrium. These temporally-overlapping regimes are schematically illustrated in Fig. 3-1 with some typical processes that occur in each relaxation regimes.. This time scale for. each event depends very strongly on the excitation (free carrier vs. exciton), the density of excitation, and the lattice temperature and so on. It should be emphasized that many of the physical processes leading to relaxation in the different regimes are 14.

(24) occurring simultaneously.. For example, the processes that destroy coherence may. also contribute to thermalization of carrier distribution function, and emission of phonons may occur while the electron and holes are thermalizing to a hot distribution. The non-thermal carrier distribution function created by a femtosecond pulse is influenced by the dephasing of coherent polarization during the pulse. Nonetheless, this description in terms of four relaxation regimes does provide a convenient framework for describing and discussing the dynamics of relaxation in semiconductors.. 3.2 Ultrafast Phenomena of Semiconductors 3.2.1. Bulk GaAs. Since the realization that ultrashort optical techniques allow the direct observation of carrier and phonon dynamics in semiconductors there have been an enormous amount of researches performed in group III-V materials.. The direct bandgap. characteristics of GaAs along with the availability of good quality samples had made GaAs the material of good choice in studying ultrafast carrier and phonon dynamics. Another reason for choosing group III-V as opposed to group IV materials is that the III-V materials were thought simpler to understand and to model due to their direct absorption and emission [12].. We will provide a brief description of the ultrafast. spectroscopy of GaAs. Time-resolved reflectance measurements with picosecond pulses were carried out by Auston et al. [22] in GaAs.. From the experiments, Auston et al. deduced the rate. of energy relaxation of hot electrons in GaAs to be approximately 0.4 eV/ps.. A short. time after Auston’s experiments, Shank et al. [23] carried out some interesting. 15.

(25) absorption measurement in GaAs.. They reported on time-resolved absorption. measurement in GaAs at 80 K, following an ultrashort laser pulse excitation.. In. 1988 Schoenlein et al. [24] investigated hot carrier dynamics in GaAs using the femtosecond pump and continuum probe technique.. They observed spectral. hole-burning due to a transient nonthermal carrier distribution and a rapid energy thermalization on a time scale of several tens of femtoseconds.. Zhou et al. [25] used. femtosecond luminescence to study ultrafast photoexcited holes in n-doped III-V compounds in 1992.. The dynamics of electrons and holes were studied separately. using n-doped samples and setting the excitation density much lower than the doping density.. A number of investigations on the relaxation of photoexcited carriers of. bulk GaAs have been reported in the past several years. We can easily discuss the carrier dynamics of bulk GaAs by comparing with these reports.. 3.2.2. InGaAsN/GaAs. It has been discovered that a partial replacement of the column V elements with nitrogen leads to large changes in the material properties of group III-V compounds. The InGaAsN semiconductor has been intensively studied for its both fundamental properties and potential applications in the telecommunication field.. Some. experimental results of time-resolved measurements by other groups will give later. In 1994, Sucha et al. [26] presented time-resolved measurements of carrier dynamics in bulk and quantum-well InGaAs using differential absorption spectroscopy.. They found that the carrier thermalization time is 200-300 fs. regardless of layer width for sample thickness ranging form 100 to 6000 Å.. Borri et. al. [27] studied the temperature and density dependences of exciton dephasing time in In0.18Ga0.82As/GaAs single quantum well by the time-integrated four-wave mixing. 16.

(26) Bhattacharya et al. [28] performed pump-probe differential transmission spectroscopy on In0.4Ga0.6As/GaAs/AlGaAs heterostructures. Jiang et al. [8] investigated carrier dynamics in an InxGa1-xAs1-yNy (x~0.03, y~0.01) epilayer grown on GaAs by MOCVD.. Time-resolved photoluminescence (PL). measurement, performed for various excitation intensities and sample temperatures, indicates that the broad PL emission at low temperature is dominated by localized exciton recombination. Grenouiliet et al. [29] reported an anomalous temperature dependence of the PL spectrum of a 7 nm InGaAsN/GaAs single quantum well. Kaschner et al. [10] investigated the recombination process in InGaAsN/GaAs multiple quantum wells by time-resolved measurements. Tansu et al. succeeded in developing several dilute-nitride quantum well (QW) on GaAs substrate by metal-organic chemical vapor deposition (MOCVD) in the past few years.. These high-performance InGaAs(N) QW lasers have been realized with. lasing performance up to emission wavelength of 1233 – 1300 nm [30][31][32]. Their temperature dependent PL can be found in Ref. [33].. Carrier recombination. was investigated from the behavior of the differential carrier lifetime with carrier density on two InGaAs1-xNx single quantum well laser diodes, that differ only in the nitrogen context x in the well [7].. It is shown that the recombination lifetime is. significantly reduced when nitrogen is added into the QW. In CLEO 2005, these samples were reported that a quantum-dot like behavior and a strong localization affect the carrier dynamics for temperature from 15 to 150 K by time-resolved PL [34]. The interesting behaviors in III-V-N compounds are full of vitality studied at present.. 17.

(27) 3.3 Carrier-Induced Change in Refractive Index 3.3.1. Band Filling. The electrons and holes as Fermions can occupy each quantum state only once according to the Pauli exclusive principle.. Hence each k state in a semiconductor. band can be occupied twice, once with a spin-up and once with a spin-down carrier. An occupied state is no longer available as a final state in an optical absorption process.. Due to the principle of energy minimization, the carriers in. quasi-equilibrium occupy the available states from the bottom of the band, so that the energetically lowest states are occupied first. This results in filling the states near the bottom of the conduction band by electrons and the top of the valance band by holes.. Hence, there is a decrease in the absorption coefficient at energies above the. bandgap. If parabolic bands are assumed or under the effective-mass approximation, the optical absorption near the bandgap in a direct-gap semiconductor is given by the square-root. law:. α 0 (E) =. C E − Eg , E ≥ Eg E. and. α 0 ( E ) = 0, E ≤ E g , where. E = hω is the photon energy, Eg is the bandgap energy, and C is a constant involving material’s parameters, matrix elements between periodic parts of the Bloch states at the band edges, and the fundamental constants. In the case of band filling, there is a finite probability that a state in the conduction band will be occupied by an electron and/or a state in the valance band will be empty (of electrons).. If we denote an energy in the valance band by Ea and. an energy in the conduction band by Eb, then the absorption coefficient of an injected semiconductor is α ( N , P, E ) = α 0 ( E )[ f v ( Ea ) − f c ( Eb )] .. Here N and P are the. concentrations of free electrons and holes, respectively, α0 represents the absorption of 18.

(28) pure materials in the absence of injection, fc(Eb) is the probability of a conduction band state of energy Eb being occupied by an electron, and fv(Ea) is the probability of a valance band state of energy Ea being occupied by an electron.. The change in. absorption induced by the band filling is ∆α ( N , P, E ) = α 0 ( E )[ f v ( Ea ) − f c ( Eb ) − 1] . For the case of injection, we will have fc > 0 and fv < 1 close to the respective band edges.. This formula predicts ∆α is negative, i.e., the band filling decreases the. absorption coefficient at a fixed incident photon energy. The real and imaginary parts of the refractive index n + iκ (or of the dielectric constant ε 1 + iε 2 ) are not independent. integrals n( E ) = 1 +. They are related by the Kramers-Kronig. 4πκ 2ch ∞ α ( E ' ) , c is the speed of light, e is P∫ dE ' , where α = 2 2 2 0 E' −E λ e. the electron charge, E is the photon energy, and P indicates the principal value of the integral.. In analogy with the change of absorption coefficient we defined. ∆n( N , P, E ) = n( N , P, E ) − n0 ( E ) , where n0 is the refractive index of uninjected material.. The. ∆n ( N , P , E ) =. change. in. refractive. 2ch ∞ ∆α ( N , P, E ' ) P dE ' . e 2 ∫0 E '2 − E 2. index. is. then. given. by. The refractive index decreases for energies. near and below Eg, and ∆n becomes positive for energies well above Eg.. 3.3.2. Bandgap Renormalization A consequence of the Coulomb interaction among the laser-excited electronic. excitations is the effect of the bandgap skrinkage (bandgap renormalization); i.e., the energies of the electrons and holes in their respective bands are reduced.. This. energy reduction is a consequence of the exchange effect for particles with equal spin, and Coulomb correlation effect for all particles. The exchange effect is caused by the Pauli exclusive principle.. The probability that two Fermions with identical 19.

(29) quantum numbers are at the same point in real space is zero.. For increasing. separation between the particles, the probability slowly approaches unity. Hence, the Pauli exclusion leads to a reduction of the probability that equally charged particles come close to each other, and this in turn reduces the repulsive (i.e., positive Coulomb energy) contribution. An expression for the bandgap renormalization, originally derived by Wolff, is. ∆E g = −(. e 2πε 0ε s. )(. 3neh. π. )1 / 3 , where neh is the concentration of free electrons or holes, ε0. is the permittivity of free space, and εs is the relative static dielectric constant of the semiconductor.. The estimated shrinkage is proportional to the cube-root of the. carrier concentration or the average interparticle spacing. We modeled the bandgap renormalization to cause a rigid translation (red shift) of the absorption curve. The change in absorption due to bandgap renormalization is predicted to be ∆α (neh , E ) =. C C E − E g − ∆E g (neh ) − E − Eg . E E. It predicts ∆α which is always. positive, the largest near the bandgap, and rapidly decreasing for higher energies. As with band filling, the change in refractive index was calculated by applying the Kramers-Kronig integral to the ∆α from bandgap renormalization, the largest change in refractive index is near the bandgap. Unlike band filling, however, the bandgap renormalization causes negative change in refractive index (∆n < 0) for energies above the bandgap as a result of decreasing in absorption coefficient for a fixed incident photon energy.. 20.

(30) 3.3.3. Free-Carrier Absorption Thus far, we have considered changes of the interband absorption due to band. filling and bandgap renormalization effects. In addition, a free carrier can absorb a photon and move to a higher energy state within a band. In the Drude model, this intraband free carrier absorption, also known as the plasma effect, is modeled as being directly proportional to the concentration of electrons and holes and the square of the wavelength. ∆n = −(. The corresponding change in refractive index is given by. e 2 λ2 N P )( + ) , where λ is the photon wavelength, me is the effective 2 2 8π c ε 0 n me mh. mass of electrons, and. m1 / 2 + mlh1 / 2 1 = 3hh/ 2 with mhh and mlh being the effective mh mhh + mlh3 / 2. masses of heavy holes and light holes, respectively. The sign of ∆n from the plasma effect is always negative.. Hence, the change of reflectance ∆R in pump probe. measurement is negative.. 21.

(31) (≤ 200 fs ). (≤ 2 ps). (∆E > hω LO ) (~ 1 − 100 ps ). (∆E < hω LO ) (≥ 100 ps ). Fig. 3-1 Four semiconductors.. temporally. overlapping. 22. relaxation. regime. in. photoexcited.

(32) Chapter 4. Experiments. 4.1 InGaAsN Single Quantum Well Structures The investigated samples were grown by low-pressure metalorganic chemical vapor. deposition. (LP-MOCVD). by. Tansu. et. al.. Trimethylgallium,. trimethylaluminium, and trimethylindium are used as the group-III sources. The group-V precursor used here are AsH3 and PH3, and the N precursor was U-dimethylhydrazine. p dopants, respectively.. The dopant sources are SiH4 and dielthylzinc for the n and The active regions are all based on the 60Å. In0.4Ga0.6As1-xNx QW, sandwiched with various barrier materials. The In and N contents of the InGaAsN material were determined from high-resolution x-ray diffraction and secondary mass ion spectroscopy.. The N composition x in the. structures is 0% and 0.5% for samples A and B, respectively.. The schematic. diagrams of the structure investigated here are shown in Fig. 4-1 [30]. The details of growth optimization can be found in Ref. [30][30][31]. High resolution transmission electron microscopy (HRTEM) image shown in Fig. 4-2 is performed for the interface quality of In0.4Ga0.6As0.995N0.005/GaAs SQW [33]. The crystal quality and interface of quantum are directly investigated. The rough interface shows a deteriorated material quality, compared with the conventional III-V materials. The room-temperature PL spectra of the samples are presented in Fig. 4-3. The band gap of these samples are 1.04eV (1.2µm) and 0.97 eV (1.3µm) for In0.4Ga0.6As and In0.4Ga0.6As0.995N0.005 SQW, respectively.. 23.

(33) 4.1.1. Mode-Locked Ti:sapphire laser. A commercial Kerr-lens-mode-locked Ti:sapphire laser (Coherent Mira900) is used as the light source. Typically, the pulse width is 175 fs and the pulse repetition rate is 75 MHz and the wavelength is tunable from 700 nm to 900 nm. Fig. 4-4 shows the optical path within the laser cavity.. The Ti:sapphire laser is pumped by a. diode-pumped frequency-doubled Nd:YVO4 laser (Coherent Verdi-8). The pump power is 8 W (?) and the output power is smaller than 1 W.. 4.1.2. Autocorrelator. Autocorrelation measurement using second harmonic generation (SHG) is the most common method for measuring the pulse-width of ultrashort laser pulse.. A. schematic of intensity autocorrelator is shown in Fig. 4-5. The laser beam is split into two with identical power by 50/50 beamsplitter.. One beam is temporally. delayed by a translation stage. Two beam focus non-collinearly onto a beta barium borate (BBO) nonlinear optical crystal. The autocorrelation signal comes from the second harmonic generation (SHG) of BBO is detected by a silicon detector. The signal is than recorded on a personal computer. The result of the autocorrelation trace is shown in Fig. 4-6. The red curve is the Gaussian fit to the data. The measured pulse-width of our laser is 230 fs.. 4.1.3. Experimental Setup of Femtosecond Time-Resolved Measurement. Figure 4-7 shows the schematic diagram of the femtosecond time-resolved setup.. 24.

(34) The ultrafast time-resolved photoreflectance experiments are performed at room temperature using the pulsed Ti:sapphire laser described in Sec.4.2.1. The laser beam is split into pump and probe beams by 50/50 beamsplitter (BS). The variable neutral density filter (ND1) is used to attenuate the power of probe beam to keep the pump probe ratio 20:1. Another variable neutral density filter (ND2) is used to adjust the power of the pump beam. The 1.25 µm-stepping resolution translation stage, controlled by a computer, is used to change the optical path length of the pump beam, resulting in a (positive or negative) time delay between the pump and the probe pulses. The minimal time delay is 8.33fs. In order to minimize the optical coherent artifacts, the polarization of the pump beam is rotated perpendicularly to that of probe beam by a half-wave (λ/2) plate. The pump and probe beams then are focused separately by two identical convex lenses with focal length 10 cm and 5 cm.. The. focused pump beam spot size ~ 250 µm. The spatial overlap of pump and probe beams on the sample was monitored by a CCD camera. The reflected pump beam is rejected using an iris. A silicon photodetector (Newport 818-SL series) detects the reflected intensity of the probe beam.. With an optical chopper in the pump beam. path, the detected signal is processed with a lock-in amplifier (Stanford Research System, Model SR830) to perform a phase locked measurement. The probe is then recorded on a personal computer as a function of temporal delay between pump and probe beams.. 4.1.4. Zero Delay point. In this experiment, it is important to find out the zero-delay point (∆t=0). In order to find out the zero point, we focused the laser pulses into a beta barium borate (BBO) crystal to produce frequency-doubled visible light. The type-I phase matching 25.

(35) condition of second harmonic generation (SHG) is given by ∆k = k (ω 2 ) − 2k (ω1 ) = 0 , where. k (ω1 ) =. 2πn1 (ω1 ). n2 (ω 2 ) = ne (ω 2 ,θ ) = (. λ1. ,. k (ω 2 ) =. 2πn2 (ω 2 ). λ2. ,. cos 2 θ sin 2 θ −1 / 2 , respectively. + ) no2 (ω 2 ) ne2 (ω 2 ). λ1 = 2λ2. ,. and. Here, θ is the phase. matching angle, ω1 is the fundamental angular frequency, ω2=2ω1, no is the ordinary refractive index, and ne is the extra-ordinary refractive index.. 26.

(36) Fig. 4-1 Cross-sectional schematic conduction band diagram of the In0.4Ga0.6As1-xNx SQW (Ref. [30]).. Fig. 4-2 HRTEM image of In0.4Ga0.6As0.995N0.005/GaAs SQW (Ref. [33]).. 27.

(37) 100 (a) In0.4Ga0.6As PL Intensity (a.u.). 80 60 40 20 0 0.95. 1.00. 1.05. 1.10. 1.15. 1.20. 1.25. Energy (eV) 18. (b) In0.4Ga0.6As0.995N0.005. PL Intemsity (a.u.). 16 14 12 10 8 6 4 2 0 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20. Energy (eV). Fig. 4-3 Result of photoluminescence (PL) spectra measurement.. Fig. 4-4 The optical beam path within Coherent Mira-900.. 28.

(38) Fig. 4-5 Sketch diagram of the intensity autocorrelator.. Autocorrelation signal (a.u.). 22 20 18 230 fs 16 14 12 -800 -600 -400 -200. 0. 200 400 600 800. Delay time (fs). Fig. 4-6 The autocorrelation trace.. 29.

(39) λ. Fig. 4-7 Femtosecond time-resolved photoreflectance measurement system.. 30.

(40) Chapter 5. Results and Discussion. The time-resolved photoreflectance response of the semiconductor to optical pulses can be broadly attributed to the effects of the carriers generated by the pump pulses via interband transition.. The change in the complex dielectric function. induced by these carriers is governed by the various carrier relaxation and decay mechanisms inside the semiconductor. This essentially determines the reflectivity of the probe pulse. This change can be modeled by considering the absorption change due to band filling (BF), the band gap shift due to the presence of carriers (better known as band-gap renormalization or BGR), plasma screening effect, enhancement of absorption due to the Coulomb enhancement factor (CEF), and intraband absorption of photons by the free carriers (FCA). The change in the reflectivity is due to the change in the complex refractive index (n + iκ ) . In general, the major contributions to the change in n can be written as ∆n = ∆nBF + BGR + ∆nFCA .. The change in the reflectivity (∆R) which is the. difference between the reflectivity R with the pump ON and the pump OFF, can be calculated as ∆R(hν ) = R (n' , κ ' , hν ) − R(n, κ , hν ) . Here the reflectivity R(n, κ , hν ) is given by R (n, κ , hν ) = relation κ ( hν ) =. [ n( hν ) − 1]2 + κ 2 ( hν ) , and κ is related to α by the [ n( hν ) + 1]2 + κ 2 ( hν ). λ λ α (hν ) and ∆κ (hν ) = ∆α ( hν ). 4π 4π. The values of n’ and κ’ are. the modified real and imaginary parts of refractive index, respectively, in the presence of the carrier density N (or in the presence of the pump beam, i.e. pump-ON condition). We will show our results of bulk GaAs and InGaAsN/InGaAs SQWs in the followings.. 31.

(41) 5.1 Ultrafast Time Resolved Photoreflectance of Bulk GaAs Figure 5-1 shows the experimental result for bulk GaAs pumped at 800 nm (or photon energy 1.55eV) with various pump power. Initially, a fast positive peak was observed and then it follows a long tail. The physical reason for the fast positive peak is absorption bleaching.. The rising is due to the generation of carriers. occupying the optical-coupled transition states. As a result, the real part of the complex refractive index increases at frequencies which are high compared with the probing frequency and hence the reflectance increases. The decay is due to carrier being scattered out of their initial states and subsequent relaxation toward the band edges through carrier-carrier scattering and carrier-phonon scattering. In GaAs, the carrier-carrier scattering time of 30 fs were observed with carrier density of 1017cm-3 [35]. In our experimental condition, 230 fs pulses used in Fig. 5-1, the carrier-carrier scattering process may not be time-resolved [36]. Therefore, the decaying of the transient is attributed to cooling of hot carriers by emitting LO phonons. Since the carrier-phonon scattering rate in semiconductor is independent of carrier density, the decay rate of the transient is expected to be constant at different excitation densities [21]. The pump intensities of Fig. 5-1 vary from 0.29–2.9 MW/cm2. At the pump wavelength (800 nm or 1.55 eV), the absorption coefficient of GaAs is about 1.5 µm-1 [37]. The corresponding peak excitation densities are estimated to be 0.4×1015 – 4.04×1016 cm-3.. The excess energy of initial carriers is 0.13 eV for each. electron-hole (e-h) pair with bandgap of 1.42 eV for GaAs.. For a LO phonon energy. of 37 meV in GaAs, each hot electron emits 3 LO phonons in cascade before reaching the band edge [38]. In Fig. 5-2 we found that the peak amplitude of ∆R does increase with increasing 32.

(42) pump power. It is proportional to the carrier density at low pumping due to the band filling effect, but it becomes flattened even slightly decreases with increasing pump power at high pumping. The magnitude of ∆R peak is proportional to the density of carriers accumulated in the allowed optical transition states after photoexcitation. At high pumping, the rate of carrier-carrier scattering becomes faster due to more carriers being generated [35]. As a result, the injection carriers cause a bandgap shrinkage by the bandgap renormalization, which is given by ∆E g = −(. e 2πε 0ε s. )(. 3neh. π. )1 / 3 [39],. where neh is the e-h pair density in cm-3, ε0 is the permittivity of free space, and εs is the static dielectric constant of the semiconductor (12.9 in bulk GaAs [40]). The renormalized bandgaps are estimated to be 1.415 eV and 1.412 eV, for 10 mW and 100 mW pumping, respectively. For a fixed photon energy, as the bandgap becomes narrower, the range of allowed optical transition states would be enlarged, which in turn reduces the state-filling (band-filling) effect. Therefore, at high pumping, the nonlinear relation between the peak amplitude of ∆R and the injected carrier density is due to the enhance carrier-carrier scattering and the induced bandgap shrinkage to decrease ∆R. The normalized ∆R traces with various pump powers are shown in Fig. 5-3. To extract the carrier lifetime, we fitted the experimental ∆R traces (starting from the delay time; in most cases, 233 fs from the peak of the trace) by a double exponential decay function, A exp( −t / τ 1 ) + B exp( −t / τ 2 ) . By varying parameters A, B, τ1, and τ2, we can obtain the best fit to the data (Appendix I). The carrier lifetime versus pump power is shown in Fig. 5-4.. The fast component of carrier lifetime. approximates 0.2±0.04 ps, and it is not change with increasing pump intensity. This result confirms that the carrier relaxation is due to electron-phonon scattering [21]. The slow component of carrier lifetime approximates 2.15 ps, resulting from the 33.

(43) cooling of the hot carrier distribution to the lattice temperature [36][41].. 5.2 Ultrafast Time Resolved Photoreflectance of InGaAs1-xNx SQW Figure 5-5 shows the ultrafast time-resolved photoreflectance of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW with pump power of 10–50 mW at 800 nm. Although the thickness of the SQWs is only 6 Å within a 3000 Å wide confining layer (GaAs), we still can find quite different photoreflectance results of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW.. The negative change of transient reflectance. corresponds to the negative index change on the order of 10-5 – 10-4 for In0.4Ga0.6As SQW.. This is evaluated by using the simplified Kramers-Kronig relation (Esser et al.. 1990) to be. ∆R 4 ≈ 2 ∆n . R (n − 1). Here n is 3.43 of In0.4Ga0.6As SQW [42].. The. mechanism of negative change of transient reflectance is very complicated and still does not fully understand at present.. Bandgap renormalization, free-carrier. absorption (FCA) or re-excitation of trapped carriers into the conduction band may account for the negative change of transient reflectance [43]. The measured peak amplitude of ∆R versus pump power is shown in Fig. 5-6. The peak amplitude of ∆R increases linearly with the pump intensity.. The. magnitude of ∆R peak, which in In0.4Ga0.6As SQW is larger than in In0.4Ga0.6As0.995N0.005 SQW, is determined by the population of carriers accumulated in the allowed optical transition states after photoexcitation. In comparison, the carrier lifetimes of In0.4Ga0.6As0.995N0.005 SQW are shorter than those of In0.4Ga0.6As SQW. This is also found in the time-resolved PL measurements by Lifang et al. [34] and the below threshold modulation frequency response measurements by Anton et al. [7].. 34.

(44) Using a single exponential decay function, A exp( −t / τ 1 ) , the fitting results are shown in Appendix II. measurement results.. We can see that they do not quite match with our. So we fit the normalized ∆R traces of In0.4Ga0.6As and. In0.4Ga0.6As0.995N0.005 SQWs with various pump power by a double exponential decay function again (see Appendix III). List of carrier lifetimes of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW with various pump power given in Table 5-1. In order to explain both the smaller nonlinearity and the carrier lifetimes we suggest that defects are created with N incorporation. These defects form trap states within the bandgap.. These trap states very efficiently capture excited carriers and act as. nonradiative recombination centers. This capture takes place in a shorter time scale than the intraband relaxation, thus explaining the comparably shorter carrier lifetimes. Because of the localized nature of these traps, even carriers in excited states above the conduction band minimum can be captured.. Unpredictably, such thin SQW of only. 6 Å within a 300 Å wide confining layer, this effect is sensitive enough to contribute to the small maximum reflectivity changes of In0.4Ga0.6As0.995N0.005 SQW as compared to In0.4Ga0.6As SQW.. In addition, the band gap renormalization effect. decreases the band filling effect.. Similar results of negative ∆R and lifetime. shortening due to defect trapping had been reported for low-temperature grown GaAs (LT-GaAs), in which defects (As antisites) form efficient nonradiative recombination centers as compared with bulk GaAs.. 35.

(45) ∆R/R. 0.011 0.010 0.009 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000. 100mW 90 mW 80 mW 70 mW 60 mW 50 mW 40 mW 30 mW 20 mW 10 mW. 0. 2000 4000 6000 Delay time (fs). 8000. Fig. 5-1 Pump-intensity dependent time-resolved photoreflectance of bulk GaAs.. 0.009. Peak amplitude. 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.0. 0.5. 1.0. 1.5. 2.0. 2.5. 3.0. 2. power density (MW/cm ). Fig. 5-2 The measured ∆R peak amplitude versus pump power in bulk GaAs.. 36.

(46) Normalized. 10mW 40mW 70mW 100mW. 0. 2000. 4000. 6000. Delay time (fs). Fig. 5-3 The normalized ∆R traces of bulk GaAs with various pump power.. 0.40 2.4 0.35 2.2. 0.25. 2.0. 0.20. 1.8. 0.15 1.6 0.10 0.5. 1.0. 1.5. 2.0. 2.5. 3.0. 2. power density (MW/cm ). Fig. 5-4 The carrier lifetime versus pump power in bulk GaAs. 37. τ2 (ps). τ1 (ps). 0.30.

(47) (a) In0.4Ga0.6As SQW 0.000. ∆R/R. -0.001. 10 mW 20 mW 30 mW 40 mW 50 mW. -0.002 -0.003 -0.004 -0.005 -0.006 0. 2000. 4000. 6000. 8000. 10000. Delay time (fs) (b) In0.4Ga0.6As0.995N0.005 SQW. ∆R/R. 0.0000. -0.0002. +10 mW. -0.0004. {20 mW U30 mW V40 mW 50 mW. -0.0006 -1000. 0. 1000. 2000 3000. 4000. 5000. 6000. Delay Time (fs). Fig. 5-5 The pump-intensity dependent time-resolved photoreflectance of (a) In0.4Ga0.6As SQW. (b) In0.4Ga0.6As0.995N0.005 SQW at 800 nm.. 38.

(48) 6.0 5.5 -3. In0.4Ga0.6As0.995N0.005 SQW 0.8. 5.0 4.5. 0.6 4.0 3.5. 0.4. -3. Peak amplitude (×10 ). peak amplitude (×10 ). 1.0. In0.4Ga0.6As SQW. 3.0 2.5 0.2. 0.4. 0.6. 0.8. 1.0. 1.2. 1.4. 0.2 1.6. 2. power density (MW/cm ). Fig. 5-6 The measured ∆R peak amplitude versus pump power of In0.4Ga0.6As SQW (solid triangle) and In0.4Ga0.6As0.995N0.005 SQW (solid circle). Table 5-1 List of carrier lifetimes of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW with various pump power. Power density. In0.4Ga0.6As. In0.4Ga0.6As0.995N0.005. (MW/cm2). τ1(ps). τ2(ps). τ1(ps). τ2(ps). 0.5827. 2.92. --. 0.311. 10.697. 0.8741. 2.568. --. 0.31. 10.588. 1.1654. 2.371. 21.586. 0.348. 6.188. 1.4568. 1.872. 10.562. 0.254. 1.458. 39.

(49) Chapter 6. Conclusion and Perspectives. In this thesis, the ultrafast time-resolved photoreflectance of GaAs and InGaAsxN1-x (x = 0 % and x = 0.5 %) single quantum wells are shown. In bulk GaAs, the initially short carrier lifetime from the time-resolved photoreflectance measurements was found not change with increasing the pump power. confirms that the decaying is due to carrier-phonon scattering.. This. The nonlinear. relation between the peak amplitude of ∆R and the pump power is due to carrier-carrier scattering and the induced bandgap shrinkage. In comparison with the ultrafast time-resolved photoreflectance of In0.4Ga0.6As SQW and In0.4Ga0.6As0.995N0.005 SQW, striking differences are found in the maximum nonlinearity and the carrier relaxation even with such thin single quantum wells of only 6nm within 300 nm wide defining layers. For fixed photon energy, the carrier relaxation for In0.4Ga0.6As0.995N0.005 SQW is faster than for In0.4Ga0.6As SQW. The carrier dynamics is strongly affected by N incorporation. The origin of negative dip is complicated and does not fully understand at the moment, it needs further investigation. Although we have succeeded in demonstrating the ultrafast time-resolved photoreflectance measurements, there are still many experimental knacks waiting for improvement.. First, the pump beam modulation by optical chopper could be. replaced by AOM modulator to increase the signal to noise ratio. Second, the pump lens with the shorter focal length may be used to increase the pump density to lift the photoexcitation densities. Third, to tune laser wavelength so that one can make the laser wavelength resonate with the bandgap of the samples.. 40.

(50) Appendix I. Fitting results of bulk GaAs. 1.2. y0 x0 A1 t1 A2 t2. 1.0 0.8. -0.166 ? 233 ? 0.10207 206.76148 1.09838 2450.20416. y0 x0 A1 t1 A2 t2. 1.0 ? ? ? ?. .19788 52.35342 .1732 31.13685. 0.8. -0.153 ? 233 ? 0.09589 148.99514 1.07957 2414.90249. ? ? ? ?. .1871 28.95594 .12883 66.92246. 0.6. 0.6 0.4 0.4 0.2. 0.2. 0.0. 0.0 -0.2 -2000. 0. 2000. 4000 y0 x0 A1 t1 A2 t2. 1.0 0.8. 6000. -0.2 -2000 1.2. 8000. -0.155 ? 233 ? 0.06878 159.79789 1.07653 2239.88761. ? ? ? ?. 2000. 4000 y0 x0 A1 t1 A2 t2. 1.0. .19052 66.56866 .14531 65.15816. 0.8. 0.6. 6000. 8000. -0.135 ? 233 ? 0.08255 ? .25361 296.10401 ? 592.52296 1.0791 ? .26169 2200.83196 ? 88.12882. 0.6. 0.4. 0.4. 0.2. 0.2. 0.0. 0.0. -0.2 -2000. 0. 0. 2000. 1.2. 4000 y0 x0 A1 t1 A2 t2. 1.0 0.8. 6000. -0.135 ? 233 ? 0.08427 275.82625 1.05403 2097.08528. -0.2 -2000. 8000. 0. 2000. 1.0 ? ? ? ?. .24454 435.71209 .2517 61.66101. 0.8 0.6. 0.6. 4000 y0 x0 A1 t1 A2 t2. 6000. -0.16 ? 233 ? 0.07961 183.3031 1.05869 2134.97202. 8000. ? ? ? ?. .19144 52.40385 .16604 74.13032. 0.4. 0.4. 0.2. 0.2. 0.0. 0.0 -0.2 -0.2 -2000. 0. 2000. 4000 y0 x0 A1 t1 A2 t2. 1.0 0.8. 6000. -2000. 8000. -0.118 ? 233 ? 0.10176 ? .192 205.8275 ? 13.20467 0.998 ? .17888 2116.17755 ? 00.67345. 0.4. 0.4. 0.2. 0.2. 0.0. 0.0. 2000. 4000 y0 x0 A1 t1 A2 t2. 1.0 0.8. 6000. -0.2 -2000 1.2. 8000. -0.09 ? 233 ? 0.32015 ? .17986 205.03614 ? 40.42862 0.7688 ? .1691 2061.37062 ? 75.06742. 2000. 0.6 0.4. 0.2. 0.2. 6000. -0.12 ? 233 ? 0.20285 188.91468 0.91601 2199.99569. 4000 y0 x0 A1 t1 A2 t2. 0.8. 0.4. 8000. ? ? ? ?. 6000. -0.25 ? 233 ? 0.69842 237.04956 0.56421 5997.88308. .18126 63.40372 .15799 22.55517. 8000. ? ? ? ?. .17684 23.15835 .1173 199.46753. 0.0. 0.0. -0.2. -0.2 -2000. 0. 1.0. 0.6. 4000 y0 x0 A1 t1 A2 t2. 0.8 0.6. 0. 2000. 1.0. 0.6. -0.2 -2000. 0. 0. 2000. 4000. 6000. -0.4 -2000. 8000. Delay time (fs). 0. 2000. 4000. Delay time (fs). 41. 6000. 8000.

(51) Appendix II. Fitting results of SQW (I). y0 x0 A1 t1. 1.2 1.0. 0.45 ? 474.98551 0.53117 3524.73152. y0 x0 A1 t1 A2 t2. 1.0. ??? 01.16434. 0.8. 0.8. 0.6. 0.6. 0.4. 0.4. 0.025 ? 150 ? 0.66191 254.79301 0.38468 1458.69436. ? ? ? ?. .21423 38.91828 .23196 98.05025. 0.2. 0.2. 0.0. 0.0. -0.2 -2000. 0. 2000. 1.2 1.0. 4000. 6000. y0 x0 A1 t1. 0.45 ? 550 ? 0.54384 3869.74375. 8000. -1000. 10000. 0. 1.0 ? .05192 ? 30.41082. 1000 2000 3000 4000 5000 6000 y0 x0 A1 t1 A2 t2. 0.8. 0.8. 0.6. -0.12951 130 ? 0.77838 348.30133 0.30401 6188.52741. ? ? ? ? ?. .14392 33.39927 .12058 715.38786. 0.6. 0.4. 0.4. 0.2. 0.2 0.0. 0.0. -0.2. -0.2. -2000. 0. 2000. 4000 y0 x0 A1 t1. 1.2 1.0. 6000. 8000. 0.45 ? 400 ? 0.53651 3800.98591. 10000. -1000. 0. ? .04813 ? 94.3196. 1.0. 0.8. 0.8. 0.6. 0.6. 0.4. 1000 2000 3000 4000 5000 6000 y0 x0 A1 t1 A2 t2. 1.2. -0.03969 130 ? 0.80127 310.79657 0.29945 10588.84005. ? ? ? ? ?. .10064 0.98489 .06991 823.22197. 0.4. 0.2. 0.2. 0.0. 0.0. -0.2. -0.2 -2000. 0. 2000. 4000 y0 x0 A1 t1. 1.0 0.8. 6000 0.31123 550.00122 0.61723 3965.50671. 8000 ? ? ? ?. -1000. 10000. .02879 90.98497. 0. 0.8 0.6. 0.6. 1000 2000 3000 4000 5000 6000 y0 x0 A1 t1 A2 t2. 1.0. -0.1488 ? 130 ? 0.6252 ? .12668 311.94998 ? 31.15416 0.23123 ? .08808 10697.8594 ? 3008.63184. 0.4. 0.4. 0.2. 0.2. 0.0 0.0. -0.2 -0.2 -2000 1.4. 0. 2000. 4000 y0 x0 A1 t1. 1.2 1.0. 6000 0.0001 ? 500 ? 0.91235 6523.94918. 8000. -1000. 10000. ? .0553 ? 21.12615. 0.8. 0.6. 0.6. 0.4. 0.4. 0.2. 0.2. 0.0. 1000 2000 3000 4000 5000 6000 y0 x0 A1 t1 A2 t2. 1.0. 0.8. -0.30907 130 ? 0.92237 174.92167 0.39519 6.0054E118. ? ? .21896 ? 7.69965 ? .03659 ?. 0.0. -0.2. -0.2. -0.4 -0.6 -2000. 0. 1.2. -0.4 0. 2000. 4000. 6000. 8000. 10000. -1000. 42. 0. 1000 2000 3000 4000 5000 6000.

(52) Appendix III. Fitting results of SQW (II). In0.4Ga0.6As SQW. In0.4Ga0.6As0.995N0.005 SQW y0 x0 A1 t1 A2 t2. 1.2 1.0 0.8. y0 x0 A1 t1 A2 t2. 1.0. 0.3929 600 0.35332 1827.04922 0.25841 10562.97904. 0.8 0.6. 0.6. 0.025 150 0.66191 254.79301 0.38468 1458.69436. 0.4. 0.4. 0.2. 0.2 0.0. 0.0 -2000. 0. 2000. 4000. 6000. 8000. -0.2. 10000. -1000. 0. 1000 2000 3000 4000 5000 6000. Delay time (fs) y0 x0 A1 t1 A2 t2. 1.2 1.0 0.8. 0.35974 600 0.41655 2371.56618 0.2369 21586.03705. y0 x0 A1 t1 A2 t2. 1.0 0.8 0.6. -0.12951 130 0.77838 348.30133 0.30401 6188.52741. 0.6 0.4. Normalized -∆R. 0.4 0.2. 0.2. 0.0. 0.0. -0.2 -2000. -0.2. 0. 2000. 4000. 6000. 8000. 10000. -1000. 0. 1000 2000 3000 4000 5000 6000. Delay time (fs) y0 x0 A1 t1 A2 t2. 1.2 1.0 0.8. 0.37205 600.00568 0.48789 2568.21706 0.14182 1.7755E97. y0 x0 A1 t1 A2 t2. 1.2 1.0 0.8. 0.6 0.6. 0.4. 0.4. 0.2 0.0. 0.2. -0.2. 0.0. -2000. 0. 2000. 4000. 6000. 8000. 10000. -0.2 -1000. Delay time (fs) y0 x0 A1 t1 A2 t2. 1.0 0.8. 0.33444 600 0.56158 2920.33623 0.04667 9.1652E136. 0. 0.8 0.6. 0.4. 0.4. 0.2. 0.2. 0.0. 0.0. 1000 2000 3000 4000 5000 6000 y0 x0 A1 t1 A2 t2. 1.0. 0.6. -0.2 -2000. -0.03969 130 0.80127 310.79657 0.29945 10588.84005. -0.1488 130 0.6252 311.94998 0.23123 10697.8594. -0.2. 0. 2000. 4000. 6000. 8000. 10000 -1000. Delay time (fs). 43. 0. 1000 2000 3000 4000 5000 6000.

(53) References [1] A. Polimeni, G. Baldassari Hoger von Hogersthal, M. Bissiri, M. Capizzi, A. Frova, M. Reinhardt, and A. Forchel, Semicond. Sci. Technol., 17, 797 (2002). [2] C. Ellmers, F. Hohnsdorf, J. Koch, C. Agert, S. Leu, D. Karaiskaj, M. Hofmann, W. Stolz, and W. W. Ruhle, Appl. Phys. Lett., 74, 2271 (1999). [3] N. Tansu, J. Y. Yeh, and L. Mawst, IEEE J. Sel. Topics Quantum Electron., 9, 1220 (2003). [4] J. Y. Yeh, N. Tansu, and L. J. Mawst, Electron Lett., 40, 739 (2004). [5] H. P. Xin, and C. W. Tu, Appl. Phys. Lett., 72, 2442 (1998). [6] G. L. Belenky, C. L. Reynolds, D. V. Donetsky, G. E. Shtengel, M. S. Hybertsen, M. A. Alam, G. A. Baraff, R. K. Smith, R. F. Kazarinov, J. Winn, and L. E. Smith, IEEE J. Quantum Electron., 35, 1515 (1999). [7] O. Anton, C. S. Menoni, J. Y. Yeh, L. J. Mawst, J. M. Pikal, and N. Tansu, IEEE Photon. Technol. Lett., 17, 953 (2005). [8] R. A. Mair, J. Y. Lin, H. X. Jiang, E. D. Jones, A. A. Allerman, and S. R. Kurtz, Appl. Phys. Lett. 76, 188 (2000). [9] I. A. Buyanova, G. Pozina, P. N. Hai, N. Q. Thinh, J. P. Bergman, W. M. Chen, H. P. Xin and C. W. Tu, Appl. Phys. Lett., 77, 2325 (2000). [10] A. Kaschner, T. Lu ttgert, H. Born, A. Hoffmann, A. Yu. Egorov and H. Riechert, Appl. Phys. Lett., 78, 1391 (2001). [11] A. Vinattieri, D. Alderighi, M. Zamfirescu, M. Colocci, A. Polimeni, M. Capizzi, [12] [13] [14] [15] [16] [17]. D. Gollub, M. Fischer, and A. Forchel, Appl. Phys. Lett., 82, 2805 (2003). Andreas Othonos, J. Appl. Phys., 83, 1789 (1998). R. L. Fork, B. I. Greene, and C. V. Shank, Appl. Phys. Lett., 38, 671 (1981). D. J. Bradely et al., Appl. Phys. Lett., 16, 53 (1970). M. Schelev et al., Appl. Phys. Lett., 18, 354 (1971). B. L. Liu, B. Liu, and Z. Y. Xu, J. Appl. Phys., 90, 5111 (2001). A. Leitenstorfer, C. Furst, A. Laubereau, W. Kaiser. Phys. Rev. Lett., 26, 1545 (1996).. [18] M. Schall, P. U. Jepsen, Opt. Lett., 25, 13 (2000). [19] M. C. Beard, G. M. Turner, C. A. Schmuttenmaer, Phys. Rev. B., 15, 15764 (2000). [20] F. Kadlec, H. Nemec, and P. Kuzel, Phys. Rev. B., 70, 125205 (2004). [21] J. Shah, Ultrafast Spectroscopy of Semiconducters and Semiconductor Nanostructures, Springer-Verlag, Berlin, (1996). [22] D. H. Auston, S. McAfee, C. V. Shank, E. P. Ippen, and O. Teschke, Solid-State 44.

(54) Electon., 21, 147 (1978). [23] C. V. Shank, R. L. Fork, R. F. Leheny, and J. Shah, Phys. Rev. Lett., 42, 112 (1979). [24] R. W. Schoenlein, W. Z. Lin, S. D. Brorson, E. P. Ippen, and J. G. Fujimoto, Solid-State Electron., 31, 443 (1988). [25] X. Q. zhou, K. Leo, and H. Kurz, Phys. Rev. B., 45, 3886 (1992). [26] G. Sucha, S. R. Bolton, D. S. Chemla, D. L. Sivco, and A. Y. Cho, Appl. Phys. Lett., 65, 1486 (1994). [27] P. Borri, W. Langbein, J. M. Hvam, and F. Martelli, Phys. Rev. B., 59, 2215 (1999). [28] P. Bhattacharya, S. Ghosh, S. Pradhan, J. Singh, W. zong-Kwei, J. Urayma, K. Kim, and T. B. Norris, IEEE J. Quantum Electronics., 39, 952 (2003). [29] L. Grenouillet, C. Bru-Chevallier, G. Guillot, P. Gilet, P. Duvaut, C. Vannuffel, A. [30] [31] [32] [33] [34] [35]. Million, and A. Chenevas-Paule, Appl. Phys. Lett., 76, 2241 (2000). N. Tansu, N. J. Kirsch, and L. J. Mawst, Appl. Phys. Lett., 81, 2523 (2002). N. Tansu, J. Y. Yeh, and L. J. Maest, Appl. Phys. Lett., 82, 3008 (2003). N. Tansu, J. Y. Yeh, and L. J. Maest, Appl. Phys. Lett., 82, 4038 (2003). H. C. Kuo, F. I. Lai, Y. H. Chang, M. Y. Tsai, S. Y. Kuo, S. C. Wang, N. Tansu, J. Y. Yeh, and L. J. Maest, J. J. Appl. Phys., in press. Lifang Xu, D. Patel, G. Vaschenko, O. Anton, C. S. Menoni, J. Y. Yeh, T. T. V. Roy, L. J. Mawst, N. Tensu, CLEO/QELS ’05, Baltimore, Maryland, USA. W. Z. Lin, R. W. Schoenlein, J. G. Fujimoto, and E. P. Ippen, IEEE J. Quantum. Electron., 24, 267 (1988). [36] W. Z. Lin, R. W. Schoenlein, E. P. Ippen and R. A. Logan, Appl. Phys. Lett., 50, 124 (1987). [37] S. U. Dankowski, D. Streb, M. Ruff, P. Kiesel, M. Kneissl, B. Knuepfer, G. H. Doehler, U. D. Keil, C. B. Sorenson, and A. K. Verma, Appl. Phys. Lett., 68, 37 (1996). [38] J. A. Kash, J. C. Tsang, and J. M. Hvam, Phys. Rev. Lett., 54, 2151 (1985). [39] P. A. Wolff, Phys. Rev., 126, 405 (1962). [40] M. Levinshtein, S. Rumyantsev, M. Shur, Handbook series on semiconductor parameters, World Scientific, (1996) [41] W. Z. Lin, J. G. Fujimoto, E. P. Ippen, R. A. Logan, Appl. Phys. Lett., 51, 161 (1987). [42] J. Micallef et al, Appl. Phys. Lett., 62, 3164 (1993). [43] B. R. Bennett, R. A. Soref, and J. A. D. Alamo, IEEE J. Quantum Electronics., 26, 113 (1990). [44] S. sinning, T. Dekorsy, and M. Helm, IEE Proc.-Optoelectron., 151, 361 (2004).. 45.

(55)