SYNTHESIS AND CHARACTERIZATION OF COLLOIDAL INDIUM NITRIDE NANOCRYSTALS AND STUDY OF THEIR ELECTRONIC STRUCTUR E AND SIZE AND SHAPE DEPENDENT OPTICAL PROPERTIES By Basudeb Chakraborty A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemistry-Master of Science 2014 ABSTRACT SYNTHESIS AND CHARACTERIZATION OF COLLOIDAL INDIUM NITRIDE NANOCRYSTALS AND STUDY OF THEIR ELECTRONIC STRUCTUR E AND SIZE AND SHAPE DEPENDENT OPTICAL PROPERTIES By Basudeb Chakraborty In this work the synthesize colloidal Indium Nitrid e (InN) nanoparticles and characterization techniques to explain their morphology and differen t properties have been presented. The efficiency of previously reported synthetic methods have been tested and modifications have been attempted to enhance the quality of the nanocr ystals. A novel approach has been successfully developed to synthesize 4-12 nm diamet er, wurtzite phase InN nanoparticles with high crystallinity and quantitative photoluminescen ce quantum yield. This new synthetic method is based on the in-situ chemical reduction of indium(III) bidentate amine complexes in high- boiling point solvents. We demonstrate that the ele ctronic structure of these material is strongly dependant on the size and shape of the nanoparticle s. Due to quantum confinement playing a significant role the optical band gap of InN can be tuned to be up to 0.4 eV higher than the bulk material. The size of the nanoparticles can be cont rolled by the varying the precursor ratio, rate of addition of reactant whereas the shape can be mo dified by adjusting the nature and stoichiometry of the surfactants used during the sy nthesis. Dedicated to.... Ma, Baba and Gutul ACKNOWLEDGEMENTS Firstly, I would like to thank my supervisor Dr. Ré mi Beaulac for his guidance, patience, motivation, enthusiasm, and valuable insights. I am truly grateful to Dr. Richard J. Staples , Dr. Kathryn G. Severin and Dr. Askeland Per for their help with the X-ray diffraction, optical and X- ray photoelectron spectroscopy measurements. I woul d also like to thank my committee members Dr. Viktor Poltavets, Prof. Aaron L. Odom a nd Prof. James E. Jackson for their support, guidance and suggestions which motivated m e to work harder to achieve my research goals. This thesis would not have been possible without th e help and support of the Beaulac Group members and Department of chemistry, Michigan State University. Last but certainly not the least, I would like to t hank my family and friends for always being there for me, and supporting me in whatever pursuit I have chosen. TABLE OF CONTENTS LIST OF TABLES..................................... ................................................... ................................ vii LIST OF FIGURES.................................... ................................................... .............................. viii KEY TO ABBREVIATIONS............................... ................................................... ................... xii CHAPTER 1 : INTRODUCTION .......................... ................................................... .................... 1 GROUP III-V SEMICONDUCTOR NANOMATERIAL ............ ................................................ 1 GROUP III-NITRIDE SEMICONDUCTORS .................. ................................................... ......... 3 SIGNIFICANCE AND APPLICATION OF III-V MATERIALS AND INDIUM NITRIDE (InN) .......––––––––––––––––– ................................................... ................ 4 DIRECT AND INDIRECT BAND GAP ...................... ................................................... ..............6 PREVIOUSLY REPORTED SYNTHETIC METHODS FOR INDIUM N ITRIDE (InN) ......... 7 MOTIVATION FOR MOVING TO THE NANO REGIME .......... .............................................. 9 COMPARISON TO THE BULK MATERIAL ................... ................................................... ..... 11 THE BAND GAP CONTROVERSY OF INDIUM NITRIDE ................................................... 14 SCOPE OF THE THESIS ............................... ................................................... .......................... 18 CHAPTER 2 : EXPERIMENTAL SECTION .................. ................................................... ........ 20 MATERIALS ......................................... ................................................... ................................... 20 INSTRUMENTS ....................................... ................................................... ............................... 21 SYNTHESIS USING INDIUM NITRIDE NANOPARTICLES USING SODIUM AMIDE .....23 CRYSTAL SIZE ANALYSIS ............................. ................................................... ..................... 30 SYNTHESIS OF INDIUM NITRIDE BY THERMOLYSIS METHOD . .................................. 32 NEW APPROACH : SYNTHESIS OF INDIUM NITRIDE USING TM ED AS NITROGEN SOURCE ............................................ ................................................... ....................................... 34 MODIFICATIONS TOWARDS IMPROVED QUALITY COLLOIDAL NA NOCRYSTALS ................................................... ................................................... ................................................. 3 9 REVELATION OF SHAPE OF PARTICLES .................. ................................................... ....... 46 CHAPTER 3: CONTROLLING THE SHAPE OF NANOPARTICLES ––––––––...49 CHAPTER 4: INSIGHT INTO REACTION MECHANISM ........ ............................................. 61 SYNTHESIS OF INDIUM NITRIDE USING INDIUM METAL AS T HE STARTING MATERIAL .......................................... ................................................... .................................... 61 XPS (X-RAY PHOTOELECTRON SPECTROSCOPY) STUDIES TO I NVESTIGATE THE REACTION PATHWAY .................................. ................................................... ....................... 62 SYNTHESIS OF INDIUM NITRIDE WITH EXCESS TMED ...... ........................................... 65 CHAPTER 5: STUDY OF OPTICAL PROPERTIES ............ ................................................... . 69 PHOTOLUMINESCENCE (PL) FROM INDIUM NITRIDE NANOCRYS TALS ................... 69 QUANTUM YIELD (QY) .......––––––––––––––––––––––– –– 74 MANIFESTATION OF SIZE DEPENDANT OPTICAL PROPERTIES ................................. 76 CHAPTER 6 ......................................... ................................................... .................................... 79 ELEMENTAL ANALYSIS ................................ ................................................... ...................... 79 INDUCTIVELY COUPLED PLASMA ATOMIC EMISSION SPECTROS COPY ––.–... 79 CHN ANALYSIS ..................................... ................................................... ............................... 80 CONCLUSION......................................... ................................................... ................................. 83 FUTURE WORK........................................ ................................................... ............................... 84 BIBLIOGRAPHY....................................... ................................................... ............................... 85 LIST OF TABLES Table 1: Calculated QY (relative) for the InN sampl es using OLA:OA ratio 3:1.........................7 5 Table 2: Calculated QY (relative) for the InN sampl es using OLA:OA ratio 10:1.......................76 Table 3: Empirical data of bulk wurtzite InN....... ................................................... ......................77 Table 4: Results from the CHN analysis............. ................................................... ....................... 81 LIST OF FIGURES Figure 1: Plot of the solar spectrum. The range bet ween the dotted lines indicate the portion of the spectrum that can be achieved by InGaN......... ................................................... ...................... 4 Figure 2: Schematic image of a InN-based field-effe ct transistor. InN was epitaxially grown on YSZ (Yttria-stabilized zirconia) substrates. This w as performed using conventional photolithography and dry etching. The cubic InN cha nnel layers were covered with Au (source and drain electrodes) and amorphous HfO 2 (gate insulator). ––– .......................–––––. 5 Figure 3: Demonstration of direct (left) and indirect band ga p (right)............––– .....––..... 7 Figure 4: Increase in band gap with decrease in siz e of the semiconductor nanomaterial or quantum dot due to the quantum conf inement effect .................................... .... 12 Figure 5: Origin of the Burstein-Moss Shift ....... ................................................... ...................... 17 Figure 6: Molecular structure of some chemicals use d for different synthesis ......................... .. 21 Figure 7: PXRD pattern of InN nanoparticles; Crysta llite Size ~20 nm Structure: Wurtzite; System: Hexagonal; PDF: 01-074-0244; Space Group: P 6 3mc. ............................................... ... 26 Figure 8: TEM image of InN nanoparticles in hexane (A), (B). The Selected Area Electron Diffraction pattern corresponding to the wurtzite p hase of nanocrystals (C) .......................... .... 27 Figure 9: The PXRD pattern of the InN sample usi ng tributylamine, agrees to wurtzite phase of InN ............................................... ................................................... ............................................. 28 Figure 10: The EDS pattern of the sample prepared with TBA which shows presence of Indium............................................. ................................................... ........................................... 29 Figure 11: TEM images of InN nanoparticles synthesi zed using TBA ..................................... .. 29 Figure 12: Scherrer analysis to find crystallite si ze. Plugging in the values above the size of the nanocrystal is calculated to be 10.9 nm.(However th is kind of analysis can give us 20%-30% accuracy at best................................... ................................................... ....................................... 31 Figure 13: PXRD pattern of InN sample prepared by t hermolysis method ................................. 33 Figure 14: Absorption spectra of InN samples in TCE (Solvent)......................................... ........ 33 Figure 15: TEM images of InN sample prepared by the rmolysis method .................................. 33 Figure 16: The absorption spectra of aliquots drawn at 10, 20 and 30 min (Left). Aliquots drawn at 45, 60, 75 and 90 min (Right)............. ................................................... ........................ 37 Figure 17: The PXRD pattern of nanoparticles synthe sized in the method described in the section above. The black dotted line shows a Gaussian fit t o only 101 signal to find out crystallite size ................................................... ................................................... ................................................. 3 7 Figure 18: A)TEM image of InN TMED aliquot at 75 mi n (Avg. size = 4.4 nm) ; B) HRTEM image of the same ; C) Lattice fringes in the nanop articles are0.549 nm apart (consistent with the lattice parameter obtained from the wurtzite phase) , D) TEM image of InN TMED aliquot at 90 min (Avg. size = 5.62 nm) ; E) HRTEM image of the s ame –......–.................................–. 38 Figure 19: Scheme: Schematics of the reaction using TMED and n-BuLi ................................. 40 Figure 20: (Left) Aliquots taken at 25, 35, 50, 70 and 90 min . (Right) NIR absorption spectra of colloidal InN nanoparticles, showing the lowest exc itonic transition for different average particle diameters ......................................... ................................................... .......................................... 41 Figure 21: PXRD pattern of the sample agrees well t o wurtzite phase of InN ........................... . 42 Figure 22: (Left) Scherrer analysis to find out siz e of the nanocrystal, (Right) HRTEM image of a nanoparticle from the final sample............... ................................................... ........................... 42 Figure 23: Raman spectra of InN nanoparticles. Both A1(LO) and E2h peaks, which are related to a longitudinal optical phonon mode and a doubly deg enerated mode, respectively, were observed at 596cm -1 and 495cm -1 ................................................... ................................................... ........... 43 Figure 24: Size-dependence of the first absorption peak of nanometer-sized InN colloidal particles. The dashed line corresponds to the trend expected from the Brus equation ................ 44 Figure 25: Magnified TEM images of InN nanoparticle s. From these images it looks like the particles actually have a propensity to acquire pl atelet or disc like profiles ...................... ........ 46 Figure 26: (Left) Lattice fringes in a HRTEM image shows the 101 facet. (Right) AFM image of the particles synthesized by this process ...... ................................................... ........................ 47 Figure 27: (Left) Absorption spectra of aliquots ta ken out of the reaction mixture synthesized with OLA:OA ratio 5:1. (Right) Size-dependence of t he first absorption peak of these samples. The dashed line corresponds to the trend expected f rom the Brus equation ............................. .. 49 Figure 28: PXRD pattern of the TEM images synthesiz ed by OLA:OA = 5:1 and and HRTEM image showing (A) 101 plane and (B) 100 plane ..... ................................................... ................ 50 Figure 29: TEM images of the two other aliquots dra wn at 30 (left) and 45 min (right) from InN sample aliquots made with 5:1 ligand ratio ........ ................................................... ...................... 51 Figure 30: AFM image of nanoparticles prepared usin g OLA:OA = 5:1 ratio. Average height comes out to be as ~7.5 nm ......................... ................................................... ............................. 52 Figure 31: (Left) Absorption spectra of aliquots ta ken out of the reaction mixture synthesized with OLA:OA ratio 5:1. (Right) Size-dependence of t he first absorption peak of these samples ................................................... ................................................... ................................................. 5 3 Figure 32: PXRD pattern of the sample is consistent with wurtzite crystal structure ................ 54 Figure 33: Scherrer analysis (Left) and HRTEM image showing 101 plane (Right) .................. 54 Figure 34: TEM (Left) and AFM (Right) image of the final aliquot ..................................... ...... 55 Figure 35: (Left) Absorption spectra of aliquots ta ken out of the reaction mixture synthesized with OLA:OA ratio 10:1. (Right) Size-dependence of the first absorption peak of these samples............................................ ................................................... ........................................... 56 Figure 36: PXRD pattern of the sample consistent wi th wurtzite crystal structure ..................... 57 Figure 37: Scherrer analysis of 101 signal of the P XRD pattern ....................................... .......... 58 Figure 38: TEM images of nanocrystals from aliquot 3 (radius ~ 3.82 nm) (Left) and final aliquot (radius ~ 5.98 nm) (Right) ......................... ................................................... ............................... 59 Figure 39: AFM images of the nanoparticles from sam e aliquot (left- aliquot 3) and (right- final aliquot) .......................................... ................................................... ............................................ 59 Figure 40: (Left) 101 plane with a d value of 0.270 nm, (Right) a) 100 plane with a d value of 0.308 nm) 101 plane with a d value of 0.271 nm ................................ ....................................... 60 Figure 41: TEM image and selected area diffraction pattern of the the sample of In metal nanoparticles ..................................... ................................................... ....................................... 62 Figure 42: XPS data of the In 3d 5/2 of aliquots taken out at 30 min (A) and 60 min (B) from the reaction mixture using molar ratio of InBr 3, TMED, n-BuLi taken as 1:3:3 which shows relative concentration of In (III) and In (0) .............. ................................................... .............................. 63 Figure 43: : XPS data of the In 3d 3/2 and In 3d 5/2 of aliquots taken out at 90 min (A) from the reaction mixture using molar ratio of InBr 3, TMEDA, n-BuLi taken as 1:3:3 (B) shows relative concentration of In (III) to In (0) in the sample . ................................................... ...................... 64 Figure 44: XPS data of the In 3d 5/2 of aliquots taken out at different time from the s ynthesis using excess TMED ............................... ................................................... ................................. 66 Figure 45: Absorption spectra of aliquots from synt hesis using excess TMED .......................... 67 Figure 46: PXRD pattern of InN quantum dots synthes ized with excess TMED confirm the wurtzite nature of the crystal structure .......... ................................................... ............................ 68 Figure 47: Absorbance and PL Spectra of an aliquot of InN (A5) from the synthesis using OLA:OA ratio of 10:1 .............................. ................................................... ................................. 70 Figure 48: PL spectra of five aliquots of the sampl e from the synthesis using OLA:OA ratio 10:1 ................................................... ................................................... ................................................. 7 0 Figure 49: PL spectra of three aliquots of the InN sample synthesized using 3:1 ratio of OLA:OA ................................................... ................................................... ................................................. 7 2 Figure 50: PL Lifetime decay of InN sample from two different synthesis using OLA:OA ratio of 10:1 and 3:1 respectively ......................... ................................................... ................................. 72 Figure 51: Size vs. band gap plot for InN nanocryst als synthesized by using 3:1 OLA:OA ratio (Left), 10:1 ratio (Right) ........................ ................................................... ................................... 78 KEY TO ABBREVIATIONS AFM : Atomic force microscopy Al : Aliquot AlN : Alumunium Nitride Avg : Average CB : Conduction band CBM : Conduction band minimum CD : Compact disc CHN : Carbon hydrogen nitrogen CdSe : Cadmium Selenide DVD : Digital video disc EDS : Energy-dispersive X-ray spectroscopy FTIR : Fourier transform infrared FWHM : Full width at half maximum GaAs : Gallium Arsenide GaN : Gallium Nitride GaP : Gallium phosphide GaSb : Gallium Antimonide HD : Hexadecane HEMT : High electron mobility transistor HRTEM : High resolution Transmission electron spect roscopy ICP-AES : Inductively coupled plasma atomic emissio n spectroscopy ICP-OES : Inductively coupled plasma optical emissi on spectroscopy ICP : Inductively coupled plasma InGaN : Indium gallium nitride InN : Indium nitride InP : Indium phoshide InSb : Indium Antimonide IR : Infrared JCPDS : Joint Committee for Powder Diffraction Stan dards LED : Light emitting diode MBE : Molecular beam epitaxy MOCVD : Metal organic chemical vapor deposition NC : Nanocrystal NIR : Near infrared OA : Oleic acid ODE : Octadecene OLA : Oleylamine PbS : Lead Sulfide PbSe : Lead Selenide PDF : Powder diffraction file PL : Photoluminescence PXRD : Powder X-ray diffraction QD : Quantum dot QY : Quantum yield RF : Radiofrequency RPM : Revolutions per minute SAED : Selected area electron diffraction SI : Supporting information SPR : Surface plasmon resonance TBA : Tributylamine TCE : Tetrachloroethylene TCSPC : Time-correlated single photon counting TEM : Transmission electron microscopy TMED : N,N,N',N" tetramethylethylenediamine TOA : Trioctylamine TOP : Trioctylphosphine UV : Ultra violet VB : Valence band VEELS : Valence electron energy loss spectroscopy YSZ : Yttria-stabilized zirconia XRD : X-ray diffraction XPS : X-ray photoelectron spectroscopy ZnS : Zinc Sulfide CHAPTER 1 : INTRODUCTION GROUP III-V SEMICONDUCTOR NANOMATERIALS Quantum dots(QDs) are semiconductor nanocrystals wi th typical dimensions in the range of 1-100 nm that exhibit size-dependent optical and el ectronic properties. 1-4 With the absorption of a photon of energy equal or larger to the band gap ( Eg), the excitation of an electron leaves an orbital hole in the valence band. The negatively ch arged electron and positively charged hole constitute an electro-statically bound electron-hol e pair, known as the exciton. Recombination of the electron and hole annihilates the exciton and m ight be accompanied by the emission of a photon, a process known as radiative recombination. The exciton has a finite size within the crystal characterized by the excitonic Bohr radius (aexc ), which can vary from one to few hundreds of nanometers depending on the type of sem iconductor material. If the size of a semiconductor nanocrystal is comparable or smaller than the size of the exciton, the charge carriers become spatially confined giving rise to s ize-dependent optical properties (quantum confinement). QDs are constrained in all three dimensions. It is the exciton size that delineates the transition between the bulk and confined regime . The electronic properties of QDs are intermediate between those of bulk semiconductors a nd of discrete molecules and show absorption features corresponding to discrete elect ronic transitions. These discrete transitions are reminiscent of atomic spectra and have led to QDs a lso being known as artificial atoms 5 The electronic properties of QDs are dependent on the s ize and shape of the individual crystals. The main advantage of QDs is the ability to obtain diff erent electronic properties by tuning the size of the dots which make them so important for many appl ications. Owing to their nanoscale dimensions, they have very sharp density of states and excellent transport and optical properties. QDs have been applied in photodetectors, 6 lasers 7, sensors, displays and amplifiers. 8 Semiconductor materials composed of elements from g roup III and group V (or, group 13 & 15) of the periodic table are termed as III-V semicondu ctors and include materials aluminium nitride (AlN), indium nitride (InN), gallium nitride (GaN), gallium arsenide (GaAs) and indium phosphide (InP). These have advantages over other s emiconductor materials such as enhanced carrier mobility and ability to create ternary and quaternary materials that have fine-tunable electronic and optical properties based on the comp osition of material. III-V materials can also be utilized for specialized growth phenomena such a s sacrificial layers or strained layers in novel nanostructures such as semiconductor nanotubes, 9 nanowires, 10 and self assembled QDs. 11 As far as semiconductor physics is concerned, properties s uch as high electron mobility in some III-V semiconductors give them edge over elemental semico nductors for utilization in high speed devices. 12 Many III-V semiconductors have direct energy gaps ( AlN, GaN, GaAs, GaSb, InN, InP, InSb, InAs) making them useful for optoelectronics . The feature that distinguishes these III-V material s from other traditional semiconductors is their dire ct band gap. In semiconductors like Si or Ge, the conduction band minima and valence band maxima do not share the same point in k-space. Consequently, a conduction band electron can only r adiatively recombine with a valence band hole through the involvement of a phonon, where the phon on momentum equals the difference between the conduction band and valence band moment um. The involvement of the phonon makes this process much less probable, leading to s lower radiative recombination in indirect band gap materials than direct band gap ones. This results in radiative recombinations being only a small proportion of total recombinations, with mo st of them being non-radiative in nature. A great amount of light conversion efficiency is lost in indirect band gap materials due to these non-radiative recombination such as the Auger effec t or phonon emission. 13 This is not the case with direct band gap semiconductors since they do n ot require intermediate phonon involvement step and hence have much higher light conversion ef ficiencies, making them the choice materials for optical devices. GROUP III-NITRIDE SEMICONDUCTORS AlN, InN, GaN and their alloys are main components of the III-nitride semiconductor series. These prove to be very promising materials for mult iple electronic and optoelectronic applications. Using these materials, a wide range o f band gaps can be accessed (Fig. 1). For instance AlN has a band gap of 6.2 eV, whereas GaN has 3.4 eV (ultra violet (UV) region) and InN 0.7 eV (near infrared (NIR) band gap). III-V ma terials also possess very high decomposition temperatures; making them suitable for application in high-temperature electronics. 14 These are usually resistant to traditional fabrication techni que but on the other hand these materials offer appreciable chemical stability, one of the most des irable properties for fabrication in devices. GaN has a very large breakdown field (electric fiel d that a pure material can withstand) of ~ 3 x 10 6 V/cm 15 and InN is known to have quite high electron mobil ity (4400 cm 2/Vs). 16 These exceptional properties make these III-V materials d esirable in high-frequency and high-power applications. Alloys of these materials can be tune d to have possible combination of both the aforementioned properties. Figure 1: Plot of the solar spectrum. The range between the dotted lines indica te the portion of the spectrum that can be achieved by InGaN SIGNIFICANCE AND APPLICATION OF III (InN) III- V semiconductor materials have huge contribution in transforming our everyday life . These materials ground- breaking advancements such as the internet, wireles s communications, mobile phones and data stor age media like CD, DVD etc. These technological applications, such as solar cells, significant improvements in healthcare sector in bi o IR- laser detection of explosives and weapons). Some very distinct properties make these applications . These materials posses high carrier mobility, the y are mostly materials and some of th em have very narrow band gap mak spectrum. The range between the dotted lines indica te the portion of the spectrum that can be achieved by InGaN 17 SIGNIFICANCE AND APPLICATION OF III -V MATERIALS AND INDIUM NITRIDE V semiconductor materials have been utilized in many devices and instruments that have huge contribution in transforming our everyday life . These materials are largely responsible for breaking advancements such as the internet, wireles s communications, mobile phones age media like CD, DVD etc. These also have been utilized in technological applications, such as solar cells, light emi tting devices (LED) for displays, significant improvements in healthcare sector in bi o -imaging and also anti- terrorist activit laser detection of explosives and weapons). make these III- V semiconductors effective toward . These materials posses high carrier mobility, the y are mostly em have very narrow band gap mak ing them versatile for use over a spectrum. The range between the dotted lines indica te the portion of V MATERIALS AND INDIUM NITRIDE been utilized in many devices and instruments that have are largely responsible for breaking advancements such as the internet, wireles s communications, mobile phones also have been utilized in important tting devices (LED) for displays, terrorist activit ies (by V semiconductors effective toward s these particular . These materials posses high carrier mobility, the y are mostly direct band gap ing them versatile for use over a huge range of energies. With efficient synthetic m ethods they can be highly crystalline and are expected to exhibit striking optoelectronic propert ies. Figure 2: Schematic image of a InN-based field-effect transis tor. InN was epitaxially grown on YSZ (Yttria-stabilized zirconia) substrates. This w as performed using conventional photolithography and dry etching. The cubic InN cha nnel layers were covered with Au (source and drain electrodes) and amorphous HfO 2 (gate insulator). 18 InN is a direct band gap III-V semiconductor. The p rospect of InN having a band gap ~0.7 eV has caused a great deal of excitement because devic es that can utilize the entire visible spectrum and parts of the infrared spectrum have been envisi oned based on the In xGa 1-x N system. Such a material system would hold great promise for applic ations in solar cells and optoelectronics. 19 In Fig. 2 we see the demonstration of first field effect transistor (FET) device which uses InN . In this device ultrathin InN is epitaxially deposited on insulating oxide yttria-stabilized zirconia substrates. With various other applications, InN re mains a potential material for applications in high electron mobility transistors (HEMT), and as d ilute magnetic semiconductors. DIRECT AND INDIRECT BAND GAP The band gap depicts the energy difference between the valence band maxima and the conduction band minima. However, these do not alway s occur at the same value of electron momentum and this has massive impact on properties of crystal. In a direct band gap semiconductor, the top of the valence band and the bottom of the conduction band have the same momentum as shown in Fig. 3 (left). On the other ha nd, in an indirect band gap semiconductor, the valence band maxima and the conduction band min ima do not align (they have different values of electron momentum) as shown in Fig. 3 (ri ght). For a transition between the valence band and the c onduction band to occur the energy supplied must be greater than or equal to the band gap and t he momentum must be conserved. In a direct band gap semiconductor, a photon of energy Eg (where Eg is the band gap energy) can lead to the transition of an electron from the valence band to the conduction band resulting in the formation of a bound electron-hole pair because the the top o f the valence band and the bottom of the conduction band have the same momentum. However, in an indirect band gap semiconductor, this transition requires the addition or subtractio n of momentum in order to satisfy conservation of momentum condition. The transition in indirect b and gap semiconductor occurs via interaction of the electron with a photon as well as a phonon ( lattice vibration) in order to gain/lose momentum. The indirect transition generally has low er excitation probability than the direct one because the former one involves three-particle inte raction. Figure 3: Demonstration of direct (left) and indire ct band gap (right) PREVIOUSLY REPORTED SYNTHETIC METHODS FOR INDIUM NI TRIDE ( This thesis will mainly focus on report of III- nitrides synthesis was in 1938 by Juza crystallites. 20 V arious groups have attempted the synthesis of InN n anocrystals using so phase routes, including hydrothermal, molecular precursors. 25-29 The outcome of nanocrystalline InN, with non There are still no reports of high aqueous solvents. The precise control over the size and shape of the nanomaterials achievable by other methods. However, the chemistry behind is ra ther different than that of the more conventional reason behind this is the greater covalent nature of the III organometallic precursors often used for t Figure 3: Demonstration of direct (left) and indire ct band gap (right) PREVIOUSLY REPORTED SYNTHETIC METHODS FOR INDIUM NI TRIDE ( This thesis will mainly focus on synthetic strategies to prepare colloidal indium nitride. nitrides synthesis was in 1938 by Juza et al . who synthesized arious groups have attempted the synthesis of InN n anocrystals using so phase routes, including hydrothermal, 21 solvothermal, 22-24 and therma l decomposition of outcome s of these approaches are generally large agglomerati ons of nanocrystalline InN, with non -uniform size and morphology distribution of nanocry stals. There are still no reports of high -yield or colloidal solubility of these nanocrystals in organic or The precise control over the size and shape of the nanomaterials by other methods. However, the chemistry behind the formation of III ther different than that of the more conventional II- VI semiconductor nanocrysta the greater covalent nature of the III -V materials and hence the organometallic precursors often used for t heir synthesis are not stable and can form complexe s PREVIOUSLY REPORTED SYNTHETIC METHODS FOR INDIUM NI TRIDE ( InN) indium nitride. The first . who synthesized GaN and InN arious groups have attempted the synthesis of InN n anocrystals using so lution- l decomposition of of these approaches are generally large agglomerati ons uniform size and morphology distribution of nanocry stals. yield or colloidal solubility of these nanocrystals in organic or The precise control over the size and shape of the nanomaterials is often not the formation of III -V compounds VI semiconductor nanocrysta ls. The V materials and hence the heir synthesis are not stable and can form complexe s with the solvent. 30 The nucleation and growth events which need to be separated in order to achieve monodispersity become high-energy processes and tend to occur simultaneously in the synthesis of these materials. 31-33 Other techniques for synthesis of indium nitride n anocrystals include ammonolysis, 34 dc-arc plasma, 35 reactive laser ablation 36 and vapor-phase methods. 37 In these methods, nanocrystals often tend to agglomera te and exhibit poor size distribution and crystallinity. Recently, by molecular beam epitaxy (MBE) process, some groups have been able to grow high-quality crystalline InN which has wurt zite lattice type and a band gap of ~0.65-0.7 eV. 38-41 However a low cost, straightforward method with re adily available precursors was yet to be developed. Hsieh et al .17 and Xie et al .22 have reported synthetic methods to prepare InN nanocrystal by a simple blend of solution route and thermal decomposition method. However, no effect of quantum confinement have been demonstrate d in these nanocrystals and superior size distribution has not been achieved either. 41,42 Very recently Chen et al . has successfully prepared monodispersed InN nanoparticles by combining soluti on- and vapor-phase methods under silica shell confinement. These nanocrystals are reported to have cubic lattice type and band gap around 0.58 eV. 43 In this particular report too, the matter of contr ol over size has not been addressed. We have attempted to resolve these signi ficant concerns by developing a new synthetic approach that produces monodispersed InN nanocrystals with uniform size distribution and better colloidal solubility in organic media. Recently Nathan R. Neale, et al .44 have modified a colloidal synthesis of 4-10 nm diameter indium ni tride (InN) nanocrystals that exhibit both a visible absorption onset ( 1.8 eV) and a strong localized surface plasmon reso nance absorption in the mid-infrared ( 3000 nm). This material was prepared keeping the go al in mind to make a nanostructure which will combine both metal and sem iconductor properties which is believed to be of importance in the field of telecommunication and optical tuning of plasmonic devices. They were able to change the absorption onset and t he Plasmon energy by oxidation-reduction technique which does not alter the chemical composi tion of the substance but pushes the Fermi level of the semiconductor material. Thus we can se e that though so far several attempts have been made to synthesize optically tuned colloidal I nN nanocrystal much success has not been achieved in controlling these properties from funda mental aspect that is by manipulation of the size and morphology of the material. In our work we would like to shed some light on this area. MOTIVATION FOR MOVING TO THE NANO REGIME The origin of quantum confinement in QDs arises fro m the spatial confinement of electrons within the crystallite boundary. This effect is obs erved when the size of the particle is comparable to the wavelength of the electron. With decrease in size of material to nano-dimensions, the decrease in confining dimensio n make the energy levels discrete; which in turn leads to an increase of the band gap energy. Qualitatively the quantum confinement effect is analogous to the problem of a particle in a box and efforts to quantify confinement effects have been topic of considerable research for more than t hree decades now. 45,46,47 In this regard, exploring the properties of InN and other III-V mat erials in nanoregime would be really interesting. Other than studying quantum physics in these small structures; another urge that boosted a lot of attention in this field was the ne cessity to miniaturize devices. The advantage of small devices being the higher density of integrati on, faster response, lower cost and low power consumption. The increasing demand for faster compu ter processors and high-capacity memory devices has been motivating the microelectronics in dustry to decrease the size of individual parts and components. COMPARISON TO THE BULK MATERIAL Bulk semiconductors are characterized by a composit ion-dependent band gap energy ( Eg). With the absorption of a photon, an exciton is created. The minimum energy required to excite an electron from the valence band into the conduction band is Eg-Eb (where Eb is the binding energy of the exciton). The exciton has a finite size with in the crystal characterized by the Bohr excitonic radius (distance between the electron-hol e pair aexc ), which can vary from less than 1 nm to more than 100 nm depending on the material. )1( 0Bsexc ama where aexc is the Bohr excitonic radius of the semiconductor material s = static dielectric constant of the material m0 = mass of the free electron aB = Bohr radius of the hydrogen atom 2hehemmmm µ = reduced mass of the exciton me (h) = effective mass of electron (hole) If the size of a semiconductor nanocrystal is small er than the size of the exciton, the charge carriers become spatially confined, which raises th eir energy due to the inclusion of the confinement energy term (energy required to confine the exciton to a radius smaller than the Bohr exciton radius). Dimension(s) of this confinem ent in the case of spherical nanoparticles refers to the diameter of the nanocrystal which nee ds to be comparable or less than twice the Bohr radius of excitons in the bulk material. 3 The energy of the exciton in a QD is greater than t he bulk semiconductor as shown in (Equation 3) where Econfinement is the confinement energy of the exciton and Ecoulombic is the coulombic energy term resulting due to the attracti on between the electron-hole pair. Thus, the smaller the size of the crystal, the greater the qu antum confinement and greater the difference in energy between the highest valence band (VB) and th e lowest conduction band (CB) energy levels, and thus greater energy is required to exci te the electron from VB to CB. )3( )(tconfinemen bandgap exciton Ebulk EE It is a well known fact that when charge carriers ( electrons and holes) are restricted by a potential barrier leading to quantization effect they manifes t exciting properties like size-tunable absorption and emission wavelengths, discrete elect ronic transitions (Fig. 4), high quantum yield. 1,2 The radiative recombination rate constants of semic onductor QDs are substantially larger than those of their bulk counterparts becaus e of the localization of the exciton in the QDs compared to the bulk semiconductors. 4,48 Narrow band gap semiconductors, including III-V compounds are emerging as an important class of mat erials and are attracting lot of attention because of their potential as emitters in the infra red region, and thus possible applications in telecommunication and biomedical fields. 48-50 The prospect of generating multiple excitons induced by spatial confinement in narrow band gap s emiconductor nanocrystals like InN results in considerable enhancement in the performance of o ptoelectronic devices, such as solar cells and low-threshold lasers. 51,52 It is very important to note that owing to their co valent nature, the III- V semiconductor nanocrytallites constant which res ults in a larger Bohr excitonic radius materials. This can potentially lead to properties. 1,48-50,53 This makes the III- nitrides and their respective alloy explore. However, t hese are among one of the least studied class of ma terials challen ges encountered in synthesis of nanoparticles. 41,54 InN is chemically band gap ~0.65- 0.7 eV. 41,55,56 Figure 4: Increase in band gap with dec quantum dot due to the quantum confinement effect V semiconductor nanocrytallites generally have smaller effective mass and ults in a larger Bohr excitonic radius (Equation 1) compared to II materials. This can potentially lead to some phenomenal, unprecedented quantum confinement nitrides and their respective alloy s a very inter esting class hese are among one of the least studied class of ma terials primarily ges encountered in synthesis of high quality, crystalline, monodispersed InN chemically quite inert and is a direct band gap semicon band gap with dec rease in size of the se miconductor nanomaterial or to the quantum confinement effect have smaller effective mass and larger dielectric compared to II -VI or IV-VI quantum confinement esting class of materials to primarily due to the high quality, crystalline, monodispersed InN inert and is a direct band gap semicon ductor with the miconductor nanomaterial or However, the band gap of InN was once thought to be 1.9 eV for quite some time; until it was found to be incorrect in 2001 reported by Davydov e t. al and following studies. 57-61 This controversy regarding the band gap of InN will be a ddressed in details in the next section. By alloying with other III-nitride material, InN can b e transformed into an excellent candidate for broad-spectrum solar cell 62 and equally promising as nanocrystalline phosphors spanning a wide range of the optical spectrum. Thus, exploring the synthetic possibilities of InN and understanding its intrinsic material properties has provided a great deal of motivation to develop a method to prepare size tunable indium nitride nan oparticles. THE BAND GAP CONTROVERSY OF INDIUM NITRIDE The optical band gap value of InN was first traced in 1986. For the first time a distinct optical absorption feature at ~1.9 eV was observed by Tansle y and Foley. 59 The InN samples showing that particular absorption edge were grown by radio frequency (RF) sputtering of a metallic indium target in nitrogen atmosphere. Optical analy sis was conducted by optical absorption and electrical measurements were performed on photo lit hographically defined clover leaf contacts. These polycrystalline InN films were found to be n-type in nature and these films were reported to exhibit a maximum mobility of 5000 cm 2/Vs at 150 K and room temperature electron mobility of 2700 cm 2/Vs. 63 The value 1.9 eV was accepted as an official b and gap for a while; as other reports supported this finding. 58,64,65 With increasing sophistication in MOCVD and MBE techniques it was possible to grow epitaxial InN fi lms. In a short communication published in 2001; Davydov et al . claimed to observe a prominent absorption peak as well as strong photoluminescence (PL) signal located below 1 eV (c lose to 0.9 eV) for InN single crystal grown by MBE. 66 This was subsequently supported by different publi cations from Wu et al .67 and Matsuoka et al .55 who also found evidence in InN grown by MBE and M OCVD methods with a band gap near 0.7 eV. These reports triggered re-ev aluation of InN properties and progressively lower optical band gap in the range of 0.65-0.80 eV was confirmed and other crucial parameters were naturally re-investigated. There are several issues that can be identified as reasons for the wide range in reported values for the InN band gap. GaN and AlN are two other nitride semiconductors known to show evidence of absorption onsets and PL feature. These properti es originate from deep traps within the band gap. 68 In some cases these effects can play a significant role in governing optical measurements. Thus a similar phenomenon seemed very possible for InN. Within some InN films traces of metallic indium were found and Shubina et al . argued that the luminescence from InN films increased to a great extent in regions around these metallic clusters. 69 This raises the possibility that the emission ~0.7 eV could be a result of Mie r esonances. Later, Specht et al . used valence electron energy loss spectroscopy (VEELS) technique to demonstrate that InN could apparently show an absorption onset near 1.7 eV. After this; a dditional doubts were raised that this process could induce electron damage on the sample its resu lting effect. 70 Even the ratio of In:N in the InN cluster can affect the band gap of the material which has been illustrated by Butcher et al . They have shown that the stoichiometry of indium an d nitrogen in InN films can be different and not 1:1 necessarily. Their work suggests that InN w ith excess In metal in the cluster or a different In:N ratio (other than unity) can be considered as In xNy alloys with varying composition and could be possible explanation behind wide variety o f results in determining the band gap. 68,71 Davydov et al . and Wu et al . have however claimed that the higher absorption f eatures observed can be described in terms of the Burstein-Moss effe ct. 72,73 The Burstein-Moss effect : When all states close to the conduction band of a s emiconductor material become populated (by electrons) and as a result the apparent band gap of a semiconductor is increased by driving the absorption edge towards higher energy, the phenomen on is known as Burstein-Moss effect. In some cases where degenerate semiconductors (a semic onductor with such a high level of doping that the material starts to act more like a metal t han a semiconductor) are involved (in this case InN with the presence of excess In metal) degenerat e electron distribution is found to take place and it is called Burstein-Moss shift (Fig. 5). This kind of shift takes place when the electron ca rrier concentration surpasses the conduction band edge density of states. This factor actually h appens to depend on the degenerate doping level in semiconductors. Semiconductors, that are d oped by a small amount, for those the Fermi level lies between the conduction and valence bands . As the doping concentration is increased, electrons populate states within the conduction ban d. This results in pushing the Fermi level higher in energy and in the case of degenerate leve l of doping, the Fermi level lies inside the conduction band. In these cases the "apparent" band gap of a semiconductor can be determined using various spectroscopic methods. If we consider a degenerate semiconductor, the Fermi level lies in conduction band and all the states below th e Fermi level are occupied. An electron from the top of the valence band can only be excited int o conduction band above the Fermi level. Since the Fermi level resides inside the conduction band Pauli's exclusion principle forbids excitation into these occupied states. For this rea son an increase in the apparent band gap is observed. So the new band gap that we see is the su m of actual band gap + Moss-Burstein shift (as shown in Fig. 5). Figure 5: Origin of the Burstein - Effect of oxygen impurity : It is known that indium oxide ( In In several cases InN films have been found to posse s due to drawbacks of the synthetic procedure after preparation or while performing InN films are prone to oxi dation at the surface electrical and optical prope rties is still an unresolved puzzle. It has that the higher oxygen concentration in these films reason behind the higher band gap This is simply because InN- In InN. 73,75,76 This theory is indeed supported by the fact that ma ny polycrystalline fi strong correlation between oxygen content and absor ption onset Yoshimoto et al .75 Later, Bhuiyan -Moss shift In 2O3) is an indirect gap material with a band gap near 3.75 eV In several cases InN films have been found to posse s s oxygen contamination. Some of them were synthetic procedure whereas some were due to oxidation of the sample performing measurements. Though it has been known for a while that dation at the surface ; yet how this influence s the outcome of rties is still an unresolved puzzle. It has been previously suggested higher oxygen concentration in these films results from sputtering and is the in gap that is observed du ring the characterization of this In 2O3 alloys will definitely have wider absorption edge than This theory is indeed supported by the fact that ma ny polycrystalline fi strong correlation between oxygen content and absor ption onset which has been shown Bhuiyan et al . used different growth techniques to produce fi gap near 3.75 eV .74 oxygen contamination. Some of them were some were due to oxidation of the sample measurements. Though it has been known for a while that s the outcome of the previously suggested and is the in herent ring the characterization of this material. absorption edge than This theory is indeed supported by the fact that ma ny polycrystalline fi lms exhibit which has been shown by to produce fi lms of variable oxygen content and it was found that absor ption edges again appeared to correlate with oxygen composition. 76 However, it has been suggested by Monemar et al ., 69 that the measured oxygen content in these films appears to be inadequ ate to account for a shift from 0.7 eV to 1.9 eV assuming common levels of band gap bowing. I t has also been noted that some reports show that In 2O3 in fact separate out within the InN lattice and do not form alloy. If that is the scenario then additional absorption feature should be observed near 3.75 eV which happens to be the band gap of In 2O3 and not 1.9 eV. Davydov et al . also entertained the possibility of oxygen inducing background electron concentration that re sults in occupied states well above the conduction band minimum (CBM). In such cases, absor ption transitions to empty conduction band states occur at energies larger than the band gap energy value which lead to Burstein-Moss effect. Theoretical simulation has predicted that the Moss-Burstein occupation can extend the absorption energy as much as 2 eV for InN. 72 SCOPE OF THE THESIS Despite the prospect of utilizing InN in many inter esting applications, this material is still in its infancy in terms of our understanding of the prepar ation and optoelectronic properties of this material. The highest quality epitaxial layers prep ared to this date still have high dislocation densities and heavy n-type conductivity. Understanding the origin of thi s n-type conductivity and eventually producing p-type InN will be a major step forward. On the othe r hand, resolving the band gap controversy would reveal the utility of th is material for infrared optics and high power HEMTs. This thesis considers all these aspects of I nN in details. Particular emphasis has been given on developing a very easy, scalable, inexpens ive method to synthesize high quality crystalline colloidal InN QDs. This has been attemp ted by different research groups all over the world without much success until now. Attempts have been made to address the band gap query by using the traditional techniques like absorption and photoluminescence (PL) spectroscopy. A number of techniques like transmission electron mic roscopy, electron diffraction, X-ray photoelectron spectroscopy have been used to charac terize the synthesized materials and to find out the size, shape, crystallinity and further rela te them to the electronic band structure. The enormous synthetic challenge to prepare a good quality material was the biggest inspiration to work in this area. The greater covalent nature o f the III-V material makes these more difficult to prepare. The list includes InN, InP and InAs. In order to achieve a narrow size distribution in colloidal particle synthesis, temporal separation b etween particle nucleation and particle growth is necessary. This can be easily obtained with II-V I materials by rapid introduction of the required precursors (usually discrete ions) into a solvent of choice, which results in immediate nucleation followed by slow particle growth. 77 The III-V materials being more covalent, it often becomes really difficult to find precursors that ca n achieve similar results like the II-VI nanocrystal synthesis. The organometallic precursor s often used in the preparation of III-V materials are not stable enough and can complex wit h the solvent instead. 78 The nucleation and growth steps in these cases have individual high en ergy barriers and separating them from each other become rather difficult. The materials prepar ed by such methods are generally of a lower quality unlike the II-VI analogues. Despite the dif ficulties in preparation, the III-V materials seem ideal candidates to examine size quantization effects, as the excitonic Bohr radius is much larger than for the analogous II-VI compounds. This means that particles should display phenomena such as the ‚blue shift™ in the band edge at relatively large crystal sizes, making investigations into strongly confined nanostructure s achievable. CHAPTER 2: EXPERIMENTAL SECTION MATERIALS The following chemicals were used for the different synthesis protocols. Indium (III) bromide, (InBr 3), anhydrous (99.999% In; Strem Chemicals) [MW: 354 .530 g/mol]; sodium amide (NaNH 2)(94%; Alfa Aesar) [MW : 39.01 g/mol], oleylamine ( OLA) (min. 40% assay; VWR), (min. 70% assay Sigma Aldrich) [MW: 267.49 g/mol, d ensity : 0.813 g/mL], octadecene (ODE) (90%; Sigma Aldrich) [MW: 252.48 g/mol, density: 0. 789 g/mL], oleic acid (OA) (90%; Alfa Aesar) [MW: 282.47 g/mol, density: 0.8946 g/mL], N,N,N,N-tetramethylethylenediamine (TMED) (99%; Sigma Aldrich) [MW : 116.20 g/mol, den sity : 0.775 g/mL], tributylamine (TBA) (98.5%; Sigma Aldrich) [MW: 185.35 g/mol] trioctylamine (TOA) (98%; Sigma Aldrich) [MW : 353.67 g/mol], n-Butyllithium soluti on; 2.5 M in hexanes [density of solution 0.693 g/mL at 25 °C, MW of butyllitium 64.06] and l ithium dimethylamide LiN(CH 3)2 (95% ; Sigma Aldrich). All the chemicals were used as rece ived. (TBA) (TOA) (TMED) Figure 6: Molecular structure of some chemicals use d for different synthesis INSTRUMENTS The Powder X-ray diffraction (PXRD) pattern was rec orded on the Bruker Davinci Diffractometer (Cu K = 0.154 nm). The QD solution was drop-cast on zero -background silica plates for PXRD measurements. The Transmission elec tron microscopy (TEM) images were recorded on a JEOL2200FS transmission electron micr oscope operating at 200 keV. Formvar- coated copper grids were used as nanocrystal suppor ts for TEM. Absorption spectra were measured on a Hitachi U-4001 spectrometer and a OLI S17 UV/VIS/NIR spectrometer. The Perkin Elmer Phi 5600 ESCA system was used for X-ra y photoelectron spectroscopy (XPS) analysis with a magnesium K X-ray source at a take-off angle of 45°. Raman spe ctra were measured with a Labram ARAMIS Horiba Jobin Yvon Ram an spectrometer, equipped with an Ar + ion laser as a light source. Argon ion lasers emit at 13 wavelengths through the visible, ultraviolet, and near-visible spectrum, including: 351.1 nm, 363.8 nm, 454.6 nm, 457.9 nm, 465.8 nm, 476.5 nm, 488.0 nm, 496.5 nm, 501.7 nm, 5 14.5 nm, 528.7 nm, 1092.3 nm. For the PL experiments, the emission of the QDs was monitor ed using a Horiba Jobin Yvon Fluorimeter with a InGaAs array (IGA-1024*1-25-1700-3LS). The g rating used is 300 grooves/nm, blazed at (ODE) (OA) (OLA) 2000 nm. The lifetime measurements were recorded us ing a Time-correlated single photon counting (TCSPC) set up. The acquisition electronic s is Pico-Harp 300 from Picoquant. The detector is a single photon detector from AUREA for infrared range (900-1700 nm). The laser used for the experiment is a 405 nm laser from Opto Engine LLC (MLL-III-405nm-100mW) and 532 nm laser from LASOS Lasertechnik GmbH (GLK-3250 T01). ICP-OES was done using 710-Axial ICP instrument. SYNTHESIS USING INDIUM NITRIDE NANOPARTICLES USING SODIUM AMIDE Based on previous reports it seemed that formation of indium metal as a by-product of InN nanoparticle synthesis is a common occurrence. 79-81 Some researchers have deliberately avoided the use of InBr 3 as a precursor stating the weaker covalent bond nat ure of InBr 3 (lower melting point than the chloride and fluoride) to be respons ible for formation of In metal. 82,83 Belcher et al . attempted to synthesize InN from thermolysis of I n(NH 2)3 complex and had some success, though it was not a stable colloidal solut ion and the samples did not exhibit any quantum confinement effects. 41 We have revisited this synthetic method as a start ing point towards our goal of synthesizing high quality cryst alline colloidal InN nanoparticles. The added advantage of this synthetic method was the commerci al availability and relatively inexpensive nature of the precursors used in this procedure. It has been observed that the indium metal impurities could be easily removed by oxidizing the mixture using dilute nitric acid (~4%). Following this process with slight modifications in the synthetic procedure, colloidal InN nanocrystals were successfully obtained. In this synthetic procedure 1 mmol (0.3543 g) of In Br 3 is taken in 7.59 mmol 2.5 mL OLA and 5.0 mL ODE (both degassed with nitrogen gas). This mixture is stirred and heated at 150 oC until a clear mixture is obtained; indicating that InBr 3 has been solvated in the nonpolar solvent. 3 mmol (0.127g) NaNH 2 is taken in 10 mL hexadecane (HD) and heated up to 220 oC. The sodium amide salt which is melted and dissolved in the hex adecane, results in a homogeneous dispersion of the reactive amide anion ( -NH 2). The InBr 3 solution is then injected into the amide solution using a glass syringe at 180 oC and the temperature is slowly ramped to 250 oC over a period of two hours. The solution remained at this temperatur e for 4 hours and then was slowly cooled down to room temperature over a time period of 2 ho urs. The transparent solution starts turning grey over time and after about two hours of heating turns com pletely black. The product is centrifuged at 3000 rpm and washed with bromide salt and then redispersed in metal impurities which are oxidized to soluble nitr ate form by addi acid. The sample is washed with not obtained then the sample is VWR) at ~40 oC to effectively ligate the surface of the nanocrys tals. The InN nanocrystals are finally dispersed in toluene. It is important to make sure that during the washin g procedure only the sodium bromide and indium nitrate is removed an d not the indium nitride to avoid removal of indium nitride particles contro lling the centrifugation rate is crucial. Indium nitride particles do not precipitate out if the cen trifugation is done at 3000 rpm or at a lower speed. The reaction pathwa y can be des We have adopted the above method reaction media and to have most of the precursor ma terials react at once. method involves the injection o molecules into hot solution of another precursor. T he temperature of injection (the temperature of the hot solution at which the other solution is injected) should be high enough for the nucleation to occur. Thus, the injection leads to the desired product. Due to a drop in temperature on addition o f the cold component, the formation time and after about two hours of heating turns com pletely black. The product is and washed with 5.0 mL hot ethanol (~50 oC) several times redispersed in 5.0 ml hexane. At this point the product contains some In metal impurities which are oxidized to soluble nitr ate form by addi ng 2 ml dilute sample is washed with 5 mL ethanol and then sonicat ed and if colloidal suspension is not obtained then the sample is heated and stirred with 5.0 mL OLA (min. 40% assay from C to effectively ligate the surface of the nanocrys tals. The InN nanocrystals are It is important to make sure that during the washin g procedure only the sodium bromide and indium nitrate is removed an d not the indium nitride particles. In order to avoid removal of indium nitride particles contro lling the centrifugation rate is crucial. Indium nitride particles do not precipitate out if the cen trifugation is done at 3000 rpm or at a lower y can be des cribed by the following (Equation 4): method in order to bring the effect of hot- injection method in the reaction media and to have most of the precursor ma terials react at once. The hot injection method involves the injection o f a ficoldfl (room temperature) solution of one precu rsor molecules into hot solution of another precursor. T he temperature of injection (the temperature of the hot solution at which the other solution is injected) should be high enough for the occur. Thus, the injection leads to the fiinstantaneousfl formation of nuclei of the product. Due to a drop in temperature on addition o f the cold component, the formation time and after about two hours of heating turns com pletely black. The product is several times to get rid of hexane. At this point the product contains some In dilute (4%) nitric ed and if colloidal suspension is (min. 40% assay from C to effectively ligate the surface of the nanocrys tals. The InN nanocrystals are It is important to make sure that during the washin g procedure only particles. In order to avoid removal of indium nitride particles contro lling the centrifugation rate is crucial. Indium nitride particles do not precipitate out if the cen trifugation is done at 3000 rpm or at a lower injection method in the The hot injection f a ficoldfl (room temperature) solution of one precu rsor molecules into hot solution of another precursor. T he temperature of injection (the temperature of the hot solution at which the other solution is injected) should be high enough for the formation of nuclei of the product. Due to a drop in temperature on addition o f the cold component, the formation of new nuclei is prevented. The result is a suspens ion of reasonably monodispersed nuclei together with considerable amounts of free precurso rs. Increasing the temperature to higher values, but maintaining it below the nucleation tem perature, results in slow growth of the existing nuclei. The greatest advantage of this met hod is that the nucleation and growth events are well-separated resulting in a very reasonable s ize dispersion of the nanocrystals. Though, by using this procedure we are able to synt hesize pure InN nanocrystals, the main drawback of this procedure is that despite carrying out several variations in the solvent used in synthesis, reaction temperature, post-synthesis tre atments and washing procedures we could not obtain good quality absorption data. We thought, th is can be result of two factors, either the nanocrystals were so large in size that it is not p ossible for them to stay dispersed in the solvent or the polar nature of the surface of the nanocryst als which prevents them from being efficiently solubilized in the nonpolar media. The structure of these materials obtained from Powder X-ray diffraction (PXRD) measurement correspond to wurtzi te phase InN with crystallite size over 20 nm (obtained through Scherrer analysis). This is consistent with that observed in the transmission electron microscopy (TEM) images as sh own respectively in Figs 7 and 8. Figure 7: PXRD pattern of InN na System: Hexagonal; PDF: 01- 074 The PXRD pattern of the synthesized nanocrystals ma tches very well with previously published wurtzite phase bulk InN semiconductor and 3.162, 2.954, 2.528 Å co rresponding to a = b = 3.558Å, c = 5.752Å . From the TEM images to be around 15~20 nm. The crystal fringes were not visible due to the large size of particles selected area electron diffraction (SAED) pattern c onfirmed the presence of crystalline InN. PXRD pattern of InN na noparticles; Crystallite Size ~20 nm Structure: Wurtzite; 074 -0244; Space Group: P6 3mc The PXRD pattern of the synthesized nanocrystals ma tches very well with previously published bulk InN semiconductor and InN nanostructures. 83,84 The first three rresponding to [100], [002] and [101] plane with lattice parameters . From the TEM images ; the size of the nanoparticles were confirmed to be around 15~20 nm. The crystal fringes were not visible due to the large size of particles selected area electron diffraction (SAED) pattern c onfirmed the presence of crystalline InN. Structure: Wurtzite; The PXRD pattern of the synthesized nanocrystals ma tches very well with previously published three d values are plane with lattice parameters the size of the nanoparticles were confirmed to be around 15~20 nm. The crystal fringes were not visible due to the large size of particles but selected area electron diffraction (SAED) pattern c onfirmed the presence of crystalline InN. Figure 8: TEM image of InN nanoparticles in hexane (A), (B). The S Diffraction pattern corresponding to the wurtzite phase of the nanocrystals (C) To obtain good quality colloidal nanoparti scavenger is introduced to strip of presumably should make them more soluble in non pol ar solvents. So, we included 5 mmol (1.25 mL) tributylamine (TBA) in the synthesis which has the possibility of acting also as a ligand. Post synthetic procedure for this experiment was ca rried out the same way as described for the previous experiment. The product of this method seems to be more stable in non pola and methylene chloride (CH 2Cl solubilize the nanoparticles. The absorption spectr a are recorded in tetrachloroethylene (TCE) and the excitonic feature is observed at 0.873 confinement in InN nanocrysta ls (Fig sample prepared by this method confirms the wurtzit e phase of the lattice. X-ray spectroscopy EDS (Fig. 1material. From the TEM images ( TEM image of InN nanoparticles in hexane (A), (B). The S elected pattern corresponding to the wurtzite phase of the nanocrystals (C) 55 To obtain good quality colloidal nanoparti cles an additional ligand which can also act as a p roton scavenger is introduced to strip of f the protons from the surface of the particles which presumably should make them more soluble in non pol ar solvents. So, we included 5 mmol (1.25 (TBA) in the synthesis which has the possibility of acting also as a ligand. Post synthetic procedure for this experiment was ca rried out the same way as described for the The product of this method seems to be more stable in non pola Cl 2) appears to work as the most efficient one to succes sfully solubilize the nanoparticles. The absorption spectr a are recorded in tetrachloroethylene (TCE) and the excitonic feature is observed at 0.873 eV which is the first direct evidence of quantum ls (Fig . 14). As we can see in Fig. 9, the PXRD pattern of the sample prepared by this method confirms the wurtzit e phase of the lattice. The e10) acts as an evidence for the presence of the eleme nt In in the From the TEM images ( Fig. 11 ) we can see that the average diameter of these par ticles elected Area Electron cles an additional ligand which can also act as a p roton the protons from the surface of the particles which presumably should make them more soluble in non pol ar solvents. So, we included 5 mmol (1.25 (TBA) in the synthesis which has the possibility of acting also as a ligand. Post synthetic procedure for this experiment was ca rried out the same way as described for the The product of this method seems to be more stable in non pola r solvents appears to work as the most efficient one to succes sfully solubilize the nanoparticles. The absorption spectr a are recorded in tetrachloroethylene (TCE) the first direct evidence of quantum the PXRD pattern of the energy-dispersive ) acts as an evidence for the presence of the eleme nt In in the ) we can see that the average diameter of these par ticles are around 8 nm which is smaller than the ~10 nm. 85,86 Figure 9: The PXRD pattern of InN which is smaller than the Bohr radius of In N, generally considered to be The PXRD pattern of the InN sample using tributylamine, agrees to wurtzite phase of N, generally considered to be agrees to wurtzite phase of Figure 10: The EDS pattern of the sample prepared w ith TBA which shows presence of Indium Figure 11: TEM images of InN nanoparticles synthesi zed using TBA CRYSTAL SIZE ANALYSIS The Scherrer Equation is a popular method to estima te crystallite size of nanoparticles. 87,88 This equation relates the size of sub-micrometre particl es, or crystallites, in a solid to the broadening of a peak in a diffraction pattern. Here, in Fig. 1 2 the peak corresponding to 110 plane (at 2 = 51.9 o) is chosen for Scherrer analysis. The following fo rmula is used to calculate crystal size. L = Size or thickness of crystallite K = constant dependent on crystallite shape (0.89) = x-ray wavelength (0.154 nm) B = FWHM (full width at half max) or integral breadt h (in radian); 0.79 o = 0.014 radian B = Bragg angle (51.874 o /2 = 25.937 o ) 5cos BLKBFigure 12: Scherrer analysis to find crystallite size. nanocrystal is calculated to be 10. accuracy at best) One of the limitations of scherrer formula is that it provides lower bound on the particles size. Various factors contribute to the broadening of the peak, out of which most important ones are inhomogeneous strain and crystal lattice impe Scherrer analysis to find crystallite size. Plugging in the values above the size of 10. 9 nm.(However this kind of analysis can g ive us 20% One of the limitations of scherrer formula is that it provides lower bound on the particles size. Various factors contribute to the broadening of the peak, out of which most important ones are inhomogeneous strain and crystal lattice impe rfection. in the values above the size of the ive us 20% -30% One of the limitations of scherrer formula is that it provides lower bound on the particles size. Various factors contribute to the broadening of the peak, out of which most important ones are SYNTHESIS OF INDIUM NITRIDE BY THERMOLYSIS METHOD In this method 1 mmol (0.3543 g) of InBr 3 is taken in 7.59 mmol (2.5 mL) OLA and 5.0 mL ODE and 3 mmol (0.153 g) of lithium dimethylamide, LiN(CH 3)2 was taken in 15.0 mL hexane and stirred at room temperature (r.t.) for 2 days a nd 7 days in two different experimental setup. Then the supernatant solution from this mixture was carefully taken out and was heated at 200 oC under inert atmosphere (nitrogen gas) for 48 hours. The transparent colorless solution which contains the indium precursor turns black almost af ter a day of heating and the color turns deeper with time indicating formation of InN nanoparticles . In this process it is considered that In(N(CH 3)2)3 is formed and then it dissociates at high temperat ure to form crystalline InN. Aliquots were taken during the process to see any o bservable change. After the reaction was complete the reaction mixture was cooled down to ro om temperature. Then the aliquots were added to 5 ml hexane. The suspension of particles d id not seem to be stable in hexane or any other usual non-polar solvent such as toluene or me thylene chloride. PXRD was taken by drop casting the hexane solution containing the particle s on a silica plate. The PXRD (Fig. 13) corresponds to wurtzite phase and TEM images (Fig. 15) shows crystallite size to be around 15-20 nm. We thought the particles were too large t o be suspended in a solvent for long. The absorption spectra was taken by suspending these pa rticles in tetrachloroethylene (TCE). Though it does not result in a good quality colloidal susp ension, In the absorption spectrum (Fig. 14) we did see some feature but could not be convincingly assigned to any excitonic signature from the semiconductor. Since if a trace of the reactants us ed was left in the solution to be measured they could absorb in that region as well. If these were peaks from InN semiconductor then band gap of the material would determined to be ~0.77-0.79 e V; which is very close to the bulk InN in wurtzite phase. The reaction scheme c an be represented Figure 15: TEM images of InN sample pre Figure 13: PXRD pattern of InN sample prepared by thermolysis method an be represented by the following equation. (Equation 6) : TEM images of InN sample pre pared by thermolysis method 2.0 1.5 1.0 0.5 0.0 Absorbance 1.0 0.9 0.8 0.7 Energy (eV) Sample using Tributylamine Sample_Thermolysis 2 days Sample_Thermolysis 7 days Gaussian fit PXRD pattern of InN sample prepared by thermolysis method Figure 14: Absorption spectra of InN samples in TCE (Solvent) : 1.2 1.1 1.0 Sample using Tributylamine Sample_Thermolysis 2 days Sample_Thermolysis 7 days Gaussian fit Absorption spectra of InN samples in TCE (Solvent) NEW APPROACH : SYNTHESIS OF INDIUM NITRIDE USING TMED AS NITROGEN SOURCE As it is quite evident, all the previous methods at tempted have residual N-H bonds on the surface of the nanoparticles which leads to n-type doping of the material. This makes these part icles more polar in nature and makes them unstable in col loidal suspension using non-polar solvent. We also hypothesized that absence of these polar ex tensions could be the reason for not having enough control over the nucleation and growth proce ss. So, we wanted to find a precursor which will not contain these ŒN-H bonds which hints towar ds substituted amides. This led us to a precursor Tetramethylethylenediamine (TMED) which i s stable in non-polar media. Recently research group led by Horst Weller 89 published a method to prepare III-V semiconductor nanocrystals by transmetallation rout e where n-butyllithium ( n-BuLi) is used as a reducing agent to produce metal nanoparticle in-sit u which forms an organometallic precursor and later produces the desired compound by dissocia ting at high temperature. In this method trioctylphosphine (TOP) was used as the phosphorous source which is hypothesized to form a complex with indium (In) metal and then later form indium phosphide via thermolysis. We tried to replicate this method to prepare InN and attempt ed the synthesis using tributylamine and trioctylamine as the nitrogen source which resulted in mostly formation of In metal particles. Since, tetramethylethylenediamine (TMED) is known t o form complex with metals it seemed to be logical to use it as a starting material. With T MED we not only hoped to take care of the solubility issue but being a bidentate ligand it ac ts as an effective source for nitrogen. Besides, previously used amide precursors used to serve the purpose of both nitrogen source as well as the reducing agent. Here, having TMED as the source of nitrogen and n-BuLi as the reducing agent we expected the method to be simpler as well as mor e controlled. In this procedure 1 mmol (0.3543 g) of InBr 3 is taken in 2.5 mL OLA and 5 mL ODE and stirred at around 150 oC until a transparent solution (homogenous) was obt ained. 3 mmol (0.3 mL) TMED was taken in 3 mL ODE and mixed into the previ ous solution. This mixture is stirred for 1 hour and the temperature is ramped to 250 oC; in the meantime 6 mmol n-BuLi (2.4 ml 2.5(M) solution in hexane) was dissolved in 3.6 mL ODE and was then drop wise added to the mixture of InBr 3 and TMED over an hour at a rate of 6 ml/hour using a syringe pump. The mixture starts to turn brown after about 15-20 min and then black precipitate starts to form which is believed to be In metal. The concentration of this black precip itate increases with time and the solution gets darker. After the addition of n-BuLi is complete; the reaction mixture is heated a nd stirred for another half an hour and then cooled to room temper ature. The aliquots from reaction mixture is then precipit ated and washed with 5 ml ethanol and dispersed in hexane several times. The product show s maximum stability in methylene chloride but does not result into a good quality colloidal s uspension which might be due to the residual polar nature of the surface of the particles. So, t hese particles are stirred at a little elevated temperature of 50 oC with excess OLA (~10 mL) which increases the solu bility of these particles in non polar media. The excess OLA is removed by wa shing with ethanol and then the particles are dispersed in 5 mL TCE. The PXRD pattern of the final sample corresponds to wurtzite phase of InN (JCPDS no.:PDF- 01-070-2547) (Fig. 17) [Refer to SI Table 1]. In th e absorption measurement of the first two aliquots taken at 10 min and 20 min after starting addition of n-BuLi; no signature of InN band edge has been observed rather it manifests occurren ce of In metal in solution (Fig. 16(Left)). Absorption at ~400 nm is due to surface plasmon res onance absorption of In metal nanoparticles. 90,91 The aliquot taken after half an hour of beginning the addition first shows some near IR absorption which is probably indication of formation of InN. The SPR peak still subsists which notify co-existence of In and InN in the mixt ure. 89 The aliquot taken at 45 min shows clear absorption feature of InN and semiconductor nature of the material. The band edge absorption of the samples range from 0.86 to 1.0 eV(Fig. 16(Right )). The PXRD of the sample taken at 30 min was amorphous meaning predominant presence of In me tal nanoparticles (Melting point 159 oC, supposed to be even lower for nanoparticles)(SI-Fig . S1). Significant broadening is observed in the initial sample compared to the final one which is indicative of the growth of nanocrystals.(Fig. S2). The size of nanocrystals range from 5-9 nm (Fig. 18 ). The size distribution of the synthesized particles is quite narrow (10-12 %) (Figure S3). Fr om the SAED pattern and also by measuring the distance between lattice fringes (the spacing b etween two lines is 0.27 nm which 101 plane in the hexagonal lattice) we can confirm the presen ce of InN (Fig. S4). In the absorption spectra, a shift of max towards higher wavelength (lower energy) is observ ed in the aliquots taken at increasing time intervals (with time the nanopartic les grow in size) (Fig. 16(Right)). This is a manifestation of quantum confinement in the synthes ized nanoparticles. Figure 16: (Left) Aliquots d rawn at 10, 20 and 30 min. (Right 90 min. The absorption spectra were Figure 17 : PXRD pattern of nanoparticles synthesized in the method described above dotted line shows a Gaussian fit to only 101 sign rawn at 10, 20 and 30 min. (Right ) Aliquots drawn at 45, 60, 75 and spectra were measured with TCE as a solvent : PXRD pattern of nanoparticles synthesized in the method described above dotted line shows a Gaussian fit to only 101 sign al to find out crystallite size ) Aliquots drawn at 45, 60, 75 and : PXRD pattern of nanoparticles synthesized in the method described above . The black MODIFICATIONS TOWARDS IMPROVED QUALITY COLLOIDAL NA NOCRYSTALS The primary problem that we encountered with this s ynthetic method was obtaining stability of the nanoparticles in a solution as a colloidal susp ension. It seemed hard to have these InN particles stabilized in the non-polar media. So, we tried different surfactants and different ratios as the passivating media. The composition that seem ed to work a bit efficiently for the first time was a 3:1 ratio of OLA and OA while dissolving the InBr 3. This combination made dispersing the particles in the desired organic solvent; in ou r case TCE comparatively easier. The colloidal stability of the particles increase which is the fo remost requirement for study of optical properties. In a typical synthesis(Fig. 19), 1 mmol of InBr 3 (0.354 g) is solubilized in a mixture of oleic aci d (OA, 2 mL), oleylamine (OLA, 6 mL, min. 70% assay f rom Sigma-Aldrich) and octadecene (ODE, 5 mL) under air free conditions (N 2 Schlenck line) by heating to 150 oC. The nitrogen precursor, tetramethylethylenediamine (TMED, 3 mmol , 0.45 mL) in 3 mL ODE, is then added to the main solution. The mixture is stirred for ab out 1 hr, at which point the temperature is raised to 250 oC. A mixture of n-butyllithium ( n-BuLi, 1.2 mL of 2.5 M hexane solution) dissolved in 3 mL ODE is then added slowly at a ra te of 4.2 mL/hr to the InBr 3/TMED solution using a syringe pump. The mixture starts to darken about 20 min after the beginning of the n-BuLi addition, indicative of formation of InN ( vide infra ). Once the addition is over, the reaction is kept under heating/stirring for half an hour, and then cooled down to room temperature. The colloidal suspension is crashed ou t and washed several times with ethanol, and dispersed in tertrachloroethylene (TCE). During the washing procedure the centrifugation to crash the particles out needs to be done at a speed of 2500 rpm or lower for about 4-5 minute. At this speed only larger In metal impurities crash ou t along with the bromide salt. After removing the impurities to the supernatant TCE solution abou t 2 mL of etha centrifugation is done at 4500 rpm or higher. This usually crash out the larger InN nanoparticles. These cras hed out particles then undergo suspension is generally improved substantially 120-480 min. (2-8 hours). To crash out and then re to the rest of the solution and the speed for centr ifugation was increased up to 5700 rpm. The time for centrifugation was varied accordingly. NIR- Vis absorption spectroscopy measurem into 5 mL tetrachloroethylene. After wash aliquots d rawn with time is becoming where the final one is complete feature of the aliquots drawn with suggests that the aliquots that were drawn later contain lar ger particles. less confined their band gap energy is not much hig her than the bulk material whereas the smaller particles exhibit higher amount of confinem ent and their band gap energy can be 0.4 higher than the bulk material. Figure 19 : Schematics of the reaction The reaction p athway can be described as following the impurities to the supernatant TCE solution abou t 2 mL of etha nol is added and the centrifugation is done at 4500 rpm or higher. This usually crash out the larger InN nanoparticles. hed out particles then undergo surface treatment. The quality of the colloidal suspension is generally improved substantially by treating the sample in excess OLA at 40 To crash out and then re -suspend the smaller particles ethanol is added to the rest of the solution and the speed for centr ifugation was increased up to 5700 rpm. The time for centrifugation was varied accordingly. Vis absorption spectroscopy measurem ents were done by transferring the colloid particle s After wash ing, as we can see in Fig. 20 (left) rawn with time is becoming darker. The aliquot taken at 25 min is light brown in color final one is complete ly black. The absorption spectra also show that the excitonic feature of the aliquots drawn with increasing time; shifts to lower energy (Fig. that the aliquots that were drawn later contain lar ger particles. Larger particles being less confined their band gap energy is not much hig her than the bulk material whereas the smaller particles exhibit higher amount of confinem ent and their band gap energy can be 0.4 : Schematics of the reaction using TMED and n-BuLi athway can be described as following scheme: nol is added and the centrifugation is done at 4500 rpm or higher. This usually crash out the larger InN nanoparticles. The quality of the colloidal by treating the sample in excess OLA at 40 oC for suspend the smaller particles ethanol is added to the rest of the solution and the speed for centr ifugation was increased up to 5700 rpm. The ents were done by transferring the colloid particle s the color of the min is light brown in color also show that the excitonic (Fig. 20; right). This Larger particles being less confined their band gap energy is not much hig her than the bulk material whereas the smaller particles exhibit higher amount of confinem ent and their band gap energy can be 0.4 eV Figure 20: (Left) Aliquots taken at 25, 35, 50, 70 and 90 min. colloidal InN nanoparticles, showing the lowest diameters The gradual increase in the darkness of the color c an actually come from two factors. Firstly, due to the increase in size of the particles their inhe rent band gap energy becomes lower as the quantum confinement becomes less prominent. Secondly , larger particles can absorb more visible light than the smaller ones (they have high er extinction coefficient) which also contribute towards the darker color of the aliquots. Aliquots taken at 25, 35, 50, 70 and 90 min. (Right) NIR absorption spectra of colloidal InN nanoparticles, showing the lowest excitonic transition for different average particle The gradual increase in the darkness of the color c an actually come from two factors. Firstly, due to the increase in size of the particles their inhe rent band gap energy becomes lower as the uantum confinement becomes less prominent. Secondly , larger particles can absorb more visible light than the smaller ones (they have high er extinction coefficient) which also contribute towards the darker color of the aliquots. NIR absorption spectra of excitonic transition for different average particle The gradual increase in the darkness of the color c an actually come from two factors. Firstly, due to the increase in size of the particles their inhe rent band gap energy becomes lower as the uantum confinement becomes less prominent. Secondly , larger particles can absorb more visible light than the smaller ones (they have high er extinction coefficient) which also contribute Figure 21: PXRD patter n of the sample of InN Figure 22 : (Left) Scherrer analysis to find out size of the nanocrystal, (Right) HRTEM image of a nanoparticle from the sample taken from final aliquot of InN Intensity 30 29 28 27 Angle (2 )FWHM=1.015 deg 2theta=29.16 deg Size=7.994 nm Crystallite size ~ n of the sample drawn from final aliquot agrees well : (Left) Scherrer analysis to find out size of the nanocrystal, (Right) HRTEM image of taken from final aliquot of InN 32 31 to wurtzite phase : (Left) Scherrer analysis to find out size of the nanocrystal, (Right) HRTEM image of The PXRD pattern (Fig. 21 ) of the sample is in good Taking into account the width of the peak correspon ding to the 100 plane the angle 2 = 29.16 o, full width at half maxima (FWHM) is 1. the crystallite size of the material (Fig. 22). The HRTEM image ( Fplan es of InN. The distance between two planes is 0.308 nm which again is in agreement with (PCPDF - 01-074-0244). 83 Figure 23: Raman spectra of InN nanoparticles. Both a longitudinal optical phonon mode and a doubly deg enerated mode, respectively, were observed at 596cm -1 and 495cm -1 92,93 The Raman spectrum (Fig. 23) provide additional characterization information for InN. from Raman spectra proves the existence of In structure to be wurtzite. The intense peak at 300 ) of the sample is in good agreement with the wurtzite phase InN. Taking into account the width of the peak correspon ding to the 100 plane the we see that at the half maxima (FWHM) is 1. 015 o. From this dat a we can calculate size of the material using scherrer formula (Eqn. 5), which comes out to be ~ Fig. 22 ) shows the lattice fringes which corresponds to th e 101 es of InN. The distance between two planes is 0.308 nm which again is in agreement with Raman spectra of InN nanoparticles. Both A1(LO) and E2h peaks, which are related to a longitudinal optical phonon mode and a doubly deg enerated mode, respectively, were observed provide additional characterization information for InN. the existence of In -N bond and also verifies the nature of the crystal The intense peak at 300 cm -1 correspond to some internal deformation of agreement with the wurtzite phase InN. we see that at the a we can calculate comes out to be ~ 8 nm ) shows the lattice fringes which corresponds to th e 101 es of InN. The distance between two planes is 0.308 nm which again is in agreement with peaks, which are related to a longitudinal optical phonon mode and a doubly deg enerated mode, respectively, were observed provide additional characterization information for InN. The data N bond and also verifies the nature of the crystal correspond to some internal deformation of the lattice. 89 Raman spectra were taken with a Labram ARAMIS Hori ba Jobin Yvon Raman spectrometer, equipped with an Ar + ion laser as a light source operating at a wavelen gth of 514 cm -1 and focused on the sample through an optical micros cope. The infrared spectrum was recorded in the wavenumber range of 4000-400 cm -1 with a Fourier transform infrared (FTIR) spectrometer. Size of these InN nanocrystals has been determined by TEM and HRTEM images (see SI) taken from different aliquots of this synthesis and we wi ll relate the average size of these particles to the band gap of those nanocrystals. Figure 24: Size-dependence of the first absorption peak of nanometer-sized InN colloidal particles. The dashed line corresponds to the trend expected from the Brus equation The lowest-energy feature in the NIR spectrum is de finitely an excitonic transition here because this transition directly depends on the size of the nanocrystal (Fig. 24). As we have seen earlier, this is expected for quantum confined systems, as t he excitonic transition is supposed to shift to 1.4 1.2 1.0 0.8 0.6 Average Bandgap in eV 10 86420Average radius in nm 'Average bandgap of the particles from different aliquots' 'fit_Brus Equation' ˘ˇˆ˙ larger energies as the particle size is decreased f rom 9.2 nm to 4.5 nm diameters. The system becomes gradually more confined as we tend to decre ase the size of the particles which actually results into restricting the movement of the electr ons. The entire shift in this spectrum extends to 350 meV which happens to be 50% of the bulk band ga p of InN. This significantly manifests the magnitude of quantum confinement effects in these I nN nanoparticles. Brus predicted this behavior for nanocrystals as they experience confin ement by having a smaller size than their Bohr radii. 94 In this case for InN we see that it follows the ex pected trend quite well even in strong confinement regime. REVELATION OF SHAPE OF PARTICLES Though we can see the material exhibiting excitonic feature and confinement behavior very evidently, there were some facts during the charact erization process which raised some questions in our mind about the shape of the nanoparticles. T he relation between the size and the excitonic feature does not seem to follow the nature predicte d by Brus properly as the material close to hit the bulk band gap value. It seems that band gap of the particles do not tend to change after obtaining a certain size of around 8 nm. Ideally it should reach progressively for the bulk band gap as the size come close to the Bohr radii of tha t substance. But from Fig. 24 it's very evident that after reaching radii of about 4 nm the band ga p of the particles still stays pretty much at the same energy level. Some PXRD pattern and TEM images also indicated in the same way. Figure 25: Magnified TEM images of InN nanoparticle s. From these images it looks like the particles actually have a propensity to acquire pl atelet or disc like profiles The 002 peak in the PXRD stays a bit hidden under t he strong 101 feature and in some of the PXRD data actually cannot distinguish the 002 signa l from the 101 (Fig. 20). Looking carefully at the TEM images we see that the boundaries of the particle can be seen inside another particle. This means that the particles are probably stacked on one another and this is easily visible in some cases. Had all these crystal have the equal di mension it'd very unlikely as a sphere can not be piled on the other. It would be a very unstable structure and boundaries of them also should not be this evident. The other reason to think abou t the irregular shape is the HRTEM images (Fig.s 20 and 25) which show preference towards one particular plane i.e. the 101 plane. So, we performed AFM analysis on these particles and it tu rned out they were really very flat as predicted. Figure 26 : (Left) Lattice fringes in a HRTEM image shows the 101 facet. (Right) AFM the partic les synthesized by this process From this AFM image (Fig. 26 ) of a sample from the final aliquot of the product InN, we can get the average height of the nanoparticle images show that the average size of these nanocrystals is 9.2 height of the nanomaterials is smaller strongly depends on direction and it plays a dominant role in this c ase. The electron will experience confinement in the smallest dimension av ailable and that will play the most fundamental role in deciding the extent of confinem ent in a certain particle. This phenomenon supports the saturation of band gap after a certain size Oleic acid as a ligand or surfactant is known to fa cilitate growth of particles in manner. 95-98 Here , introduction of OA might have administered a simi lar effect. In that case growth in a par ticular facet of crystal is more favored than other s or in other way on plane the : (Left) Lattice fringes in a HRTEM image shows the 101 facet. (Right) AFM les synthesized by this process ) of a sample from the final aliquot of the product InN, we can get the average height of the nanoparticle s which in this case is about ~6 nm. However, the TEM the average size of these nanocrystals is 9.2 nm. That clearly shows the nanomaterials is smaller than their average diameter. Quantum confinement thus on direction and it plays a dominant role in this c ase. The electron will experience confinement in the smallest dimension av ailable and that will play the most fundamental role in deciding the extent of confinem ent in a certain particle. This phenomenon saturation of band gap after a certain size in the size vs. Eg (band gap e Oleic acid as a ligand or surfactant is known to fa cilitate growth of particles in , introduction of OA might have administered a simi lar effect. In that case ticular facet of crystal is more favored than other s or in other way on plane the : (Left) Lattice fringes in a HRTEM image shows the 101 facet. (Right) AFM image of ) of a sample from the final aliquot of the product InN, we can get nm. However, the TEM nm. That clearly shows that the diameter. Quantum confinement thus on direction and it plays a dominant role in this c ase. The electron will experience confinement in the smallest dimension av ailable and that will play the most fundamental role in deciding the extent of confinem ent in a certain particle. This phenomenon (band gap e nergy) plot. Oleic acid as a ligand or surfactant is known to fa cilitate growth of particles in a uni-directional , introduction of OA might have administered a simi lar effect. In that case ticular facet of crystal is more favored than other s or in other way on plane the growth is more restricted. In most cases the crysta l plane with highest amount of surface energy undergoes preferential growth and which helps it to stabilize. 99 There has not been enough study to investigate the growth process of different crys tal planes of wurtzite phase nanomaterials. We wanted to explore this which actually provided us w ith an opportunity to alter the shape of particles. CHAPTER 3: CONTROL Keeping in mind the ability of the ligands to influ ence the shape of nanomaterials we wanted to see if by changing the ratio of them we can actuall y maneuver the shape of these particles. If our assumption were true putting different amount of li gands in t effect on the morphology of the product. Also, the other objective was eventually to reach the bulk band gap by increasing the effective ratio of OLA and OA and see what effects it bring to the morphology. Since OLA was proven to be an efficient surfactant system we wanted to proceed with that first. For the previous method we had a ratio of OLA to OA as 3:1. It was varied to 5:1, 7.5:1 and 10:1. Since, OA is the component which favors growth on a certain f acet it's ratio to OLA was decreased gradually to observe the resulting effects. Figure 27: (Left) Absorption spectra of aliquots taken out of the reaction mixture synthesized with OLA:OA ratio 5:1. (Right) The dashed line corresponds to the trend expected f rom the Brus equation CONTROL LING THE SHAPE OF NANOPARTICLES Keeping in mind the ability of the ligands to influ ence the shape of nanomaterials we wanted to see if by changing the ratio of them we can actuall y maneuver the shape of these particles. If our assumption were true putting different amount of li gands in t he synthesis should have a variable effect on the morphology of the product. Also, the other objective was eventually to reach the bulk band gap by increasing the effective Bohr radius of the material. So, we decided to vary the what effects it bring to the morphology. Since OLA was proven to be an efficient surfactant system we wanted to proceed with that first. For the previous method we had a ratio of OLA to OA as 3:1. It was varied to 5:1, 7.5:1 and 10:1. Since, is the component which favors growth on a certain f acet it's ratio to OLA was decreased gradually to observe the resulting effects. (Left) Absorption spectra of aliquots taken out of the reaction mixture synthesized with OLA:OA ratio 5:1. (Right) Size-dependence of the first absorption peak of these samples. The dashed line corresponds to the trend expected f rom the Brus equation (Equ ation 10 SHAPE OF NANOPARTICLES Keeping in mind the ability of the ligands to influ ence the shape of nanomaterials we wanted to see if by changing the ratio of them we can actuall y maneuver the shape of these particles. If our he synthesis should have a variable effect on the morphology of the product. Also, the other objective was eventually to reach the ohr radius of the material. So, we decided to vary the what effects it bring to the morphology. Since OLA -OA mixture was proven to be an efficient surfactant system we wanted to proceed with that first. For the previous method we had a ratio of OLA to OA as 3:1. It was varied to 5:1, 7.5:1 and 10:1. Since, is the component which favors growth on a certain f acet it's ratio to OLA was decreased (Left) Absorption spectra of aliquots taken out of the reaction mixture synthesized of these samples. ation 10 ) Figure 29: TEM images of the two other aliquots dra wn at 30 (left) and 45 min (right) from InN sample aliquots made with 5:1 ligand ratio. As we see in the absorption spectra the excitonic f eature shows up at different energies due to formation of different size particles in the aliquo ts drawn in time interval. It seems to follow the trend predicted by brus more closely and we can als o reach the bulk band gap value using these particles. From the PXRD data we can see that with the OLA:OA ratio 5:1 we can get very similar quality nanocrystals with same crystalline phase (Fig. 28). The particles posses a diameter of ~12.2 nm and they also have good size d istribution (Fig. 29) which is clear from the TEM images. An interesting fact is in the HRTEM ima ge (Fig. 28) we can see two different types of planes though which was very rare in case of the particles synthesized using 3:1 OLA:OA ratio. However the preferred orientation eff ect still remains and it can be observed in the PXRD pattern of the material. The 002 feature i s not very distinct and associated with the peak from 101 plane (Fig. 28). From the AFM images (Fig. 30) also we can see that the average height of the particles is ~7.5 nm which is still s maller than their diameter. The AFM images also reveal the disc like structure of the particles. So , we kept on investigating effects of changing ratio of the ligands in the synthesis. Figure 30: AFM image of nanoparticles prepared using OLA:OA = 5:1 ratio comes out to be as ~7.5 nm Since increasing the ratio of OLA:OA led us to the right direction the next step was exploring further raised proportion of OLA in the mixture. So, we went on with OLA:OA 7 .5:1 and 10:1. In these methods we kept everything same apart from the amount of OLA and OA introduced. For 7.5:1 rati o the amount of OLA used was 7 notice in the absorption spectra very similar trend is observed. Th min, 45 min, 60 min , 75 min. We can find the average size of the parti cles from the TEM images (see SI) and band gap of each particle size from th eir optical absorption. The reveal the disc like structure of the particles. So , we kept on investigating effects of changing in the synthesis. AFM image of nanoparticles prepared using OLA:OA = 5:1 ratio . Average height Since increasing the ratio of OLA:OA led us to the right direction the next step was exploring of OLA in the mixture. So, we went on with OLA:OA 7 .5:1 and 10:1. In these methods we kept everything same apart from the amount of OLA and OA introduced. o the amount of OLA used was 7 mL and 1 mL of OA was included. As we can absorption spectra very similar trend is observed. Th e aliquots were drawn at 30 , 75 min. We can find the average size of the parti cles from the TEM images (see SI) and band gap of each particle size from th eir optical absorption. The reveal the disc like structure of the particles. So , we kept on investigating effects of changing . Average height Since increasing the ratio of OLA:OA led us to the right direction the next step was exploring of OLA in the mixture. So, we went on with OLA:OA 7 .5:1 and 10:1. In these methods we kept everything same apart from the amount of OLA and OA introduced. of OA was included. As we can e aliquots were drawn at 30 , 75 min. We can find the average size of the parti cles from the TEM images (see SI) and band gap of each particle size from th eir optical absorption. The plot of average band gap vs. radius tag along the nature suggested by the saturation at a certain size. PXRD of the sample confirms it to be InN ( Figure 31: (Left) Absorption spectra of aliquots taken out of the rea ction mixture synthesized with OLA:OA ratio 5:1. (Right) Size In order to compare the morphology of these InN nan ocrystals synthesized we want to average diameter and average height of the particle s. In the PXRD pattern ( material we don't see 002 plane as a shoulder of 101 features are present for both signals. As mentioned earlier, preferred orientation could be the reason for the XRD signals to pile up. radius tag along the nature suggested by the Brus model and we do not see any PXRD of the sample confirms it to be InN ( Fig. 32). Absorption spectra of aliquots taken out of the rea ction mixture synthesized Size- dependence of the first absorption peak of these sa mples. In order to compare the morphology of these InN nan ocrystals synthesized we want to average diameter and average height of the particle s. In the PXRD pattern ( plane as a shoulder of 101 plane in this case. Rather, very distinct features are present for both signals. As mentioned earlier, preferred orientation could be the reason for the XRD signals to pile up. rus model and we do not see any Absorption spectra of aliquots taken out of the rea ction mixture synthesized dependence of the first absorption peak of these sa mples. In order to compare the morphology of these InN nan ocrystals synthesized we want to know the average diameter and average height of the particle s. In the PXRD pattern ( Fig. 32) of the plane in this case. Rather, very distinct features are present for both signals. As mentioned earlier, preferred orientation could be the ˘ˇˆ˙ Figure 32: PXR D pattern of the sample is consistent with wurtzite crystal structure Figure 33: Scherrer analysis (left) and HRTEM image showing 10 1 Intensity 29 28 27 Angle (2 )2theta=29.24 deg FWHM=0.6 deg Size=13.53 nm Crystallite size ~ D pattern of the sample is consistent with wurtzite crystal structure Scherrer analysis (left) and HRTEM image showing 10 1 plane (right) 31 30 D pattern of the sample is consistent with wurtzite crystal structure plane (right) So, we see that the average size of the particles i n the fin (Fig. 3 4). The Scherrer analysis (Fig. we get the average size to be 13.53 the solution of nanopart icles is drop casted on the sample holder. It is ve ry likely that when the solvent is evaporated the nanoparticles form an agg lomerate which would be much larger in size than usual nanoparticles. Since in the scherrer ana lysis a collection of quantitative calculated the average size that comes up as the re sult is larger. Figure 34 : TEM (left) and AFM (right) image of the final ali quot In the AFM image taken using the same sample aliquo t we see that the a particle is ~8.5 nm. This still points toward the fact that all the dimensions are equal for the nanomaterial. Though in the AFM image the particles don't look like platelets and in the PXRD pattern also, there's no indication of preferred or ientation it seems that we have quite spherical particles this time as well. to be residues from organic solvents. fringes which correspond to 101 So, we see that the average size of the particles i n the fin al aliquot for this method is ~ 4). The Scherrer analysis (Fig. 33) also supports this number because from the peak width average size to be 13.53 nm. This is can be explained as during the sample preparation icles is drop casted on the sample holder. It is ve ry likely that when the solvent is evaporated the nanoparticles form an agg lomerate which would be much larger in size than usual nanoparticles. Since in the scherrer ana lysis a collection of quantitative calculated the average size that comes up as the re sult is larger. : TEM (left) and AFM (right) image of the final ali quot In the AFM image taken using the same sample aliquo t we see that the a verage height for this nm. This still points toward the fact that all the dimensions are equal for the nanomaterial. Though in the AFM image the particles don't look like platelets and in the PXRD pattern also, there's no indication of preferred or ientation it seems that we have quite spherical particles this time as well. The very small spots on the AFM image to be residues from organic solvents. The HRTEM image (Fig. 3 3) exhibit presence of lattice fringes which correspond to 101 plane. We did not have enough HRTEM data to draw a aliquot for this method is ~ 12.25 nm from the peak width can be explained as during the sample preparation icles is drop casted on the sample holder. It is ve ry likely that when the solvent is evaporated the nanoparticles form an agg lomerate which would be much larger in size than usual nanoparticles. Since in the scherrer ana lysis a collection of quantitative data is verage height for this nm. This still points toward the fact that all the dimensions are equal for the nanomaterial. Though in the AFM image the particles don't look like platelets and in the PXRD pattern also, there's no indication of preferred or ientation it seems that we have not obtained The very small spots on the AFM image are believed 3) exhibit presence of lattice enough HRTEM data to draw a conclusion on whether this can be attributed to a p referred orientation effect. So, we decided to look further with a different ligand ratio. However, as we can see, in case of 5:1 and 7.5:1 bo th the final sample shows excitonic feature very close to the band gap of the bulk material. So , the first objective of reaching the size featuring bulk band gap is completed. This would be very useful to pin-point the Bohr radius of InN which has not been yet experimentally determine d. In the next procedure everything was kept the same except for the amounts of ligands introduced. 7.2 mL OLA and 0.8 mL of OA were used in this attem pt. The aliquots shown in the absorption spectra were drawn at 25, 35, 45, 60, 75 and 90 min (Fig. 35). Figure 35: (Left) Absorption spectra of aliquots ta ken out of the reaction mixture synthesized with OLA:OA ratio 10:1. (Right) Size-dependence of the first absorption peak of these samples. Abs. (Arb. U) 1.4 1.2 1.0 0.8 0.6 Energy (eV) 1.2 1.0 0.8 0.6 Average Bandgap 10 86420 Avergae radius Average Bandgap fit_Brus ˝˛˚˛˘˜ The colloidal quality of the nanocrystals from their absorption spectra and the TEM images (s ee SI) gave us the information about their size. As we look into the average band gap vs avera ge size plot it seems to follow the trend predicted by Br us. In the PXRD pattern (Fig. preferred orientation which means the crystals must be regularly shaped and the incident x gets equivalent treatment in all directions. It als o falls in very good agreement crystal structure of InN. Figure 36: PXRD pattern of the sample is consistent quality of the nanocrystals was substantially good. The band gap energy was found from their absorption spectra and the TEM images (s ee SI) gave us the information about their size. As we look into the average band gap vs avera ge size plot it seems to follow the trend us. In the PXRD pattern (Fig. 36) we do no t really see any interference due to preferred orientation which means the crystals must be regularly shaped and the incident x gets equivalent treatment in all directions. It als o falls in very good agreement PXRD pattern of the sample is consistent with wurtzite crystal structure substantially good. The band gap energy was found from their absorption spectra and the TEM images (s ee SI) gave us the information about their size. As we look into the average band gap vs avera ge size plot it seems to follow the trend t really see any interference due to preferred orientation which means the crystals must be regularly shaped and the incident x -ray with the wurtzite with wurtzite crystal structure Figure 37: Scherrer analysis of 101 signal of the P XRD pattern shown above So, from both Scherrer analysis and TEM images, the size of nanoparticles were found to be ~11 nm. Here we see a set of TEM and AFM showing sa mple particles from the same aliquot so we can compare their morphology. Intensity 35 34 33 Angle Scherrer analysis of 101 peak Size of crystallite ~ 10.2nm Figure 38: TEM images of nanocrystals from al (radius ~5.98 nm) Figure 39: AFM images of the nanoparti aliquot) TEM images of nanocrystals from al iquot 3 (radius ~3.82 nm) [le ft] the nanoparti cles from same aliquot (left- aliquot 3) and (right ft] and final aliquot 3) and (right - final CHAPTER 4: INSIGHT INTO REACTION MECHANISM Since, the reaction scheme is not exactly a hot inj ection method we tried to investigate the mechanism of the reaction. Initially it was assumed that In metal is formed as an intermediate and TMED forms a complex with In metal which is lat er reduced by n-BuLi to produce InN. There are a set of experiments we did which support s this hypothesis. SYNTHESIS OF INDIUM NITRIDE USING INDIUM METAL AS T HE STARTING MATERIAL If In metal were the constituent complexing with TM ED to form the precursor complex In metal should be an efficient source of In in the reaction . So, In metal was synthesized in situ using n-BuLi as reducing agent and then TMED was added to it. Despite raising the temperature to the desired point and continuous heating the reaction f or over an hour any sign of evolution of InN was not observed. The In metal nanoparticles are bl ack in color after addition of TMED and raising the temperature beyond the melting point of In metal nanoparticles the solution starts to turn little translucent and it does not exhibit any further precipitation indicating development of InN nanocrystals. However when small amount of n-BuLi was added to the solution black color appears within 2 minutes which might be due to form ation of small quantity of InN in the solution or due to further reduction of an oxidized state. The PXRD pattern does not match with any traditiona l InN phase (See SI) and amorphous SAED (Fig. 41) pattern suggests those particles were not crystalline in nature meaning most of them we re left as unreacted In metal nanoparticle particles are formed as product. Figure 41: TEM image AND nanoparticles. XPS (X- RAY PHOTOELECTRON SPECTROSCOPY) STUDIES TO INVESTIG ATE THE REACTION PATHWAY To confirm our proposition about the reaction mechanism XPS st udies and ICP measurement was done on the aliquots obtained from different the XPS studies that in the initial aliquots concen tration of In(0) predominates and in the consecutive aliquots concentration of In (III) whic h is a component of InN To inspect the effect of TMED, In case 1, 3 and 3 mmol of InBr 3, TMED, re left as unreacted In metal nanoparticle . The TEM images (Fig. 41) show AND selected area diffraction pattern of the sample RAY PHOTOELECTRON SPECTROSCOPY) STUDIES TO INVESTIG ATE THE our proposition about the reaction mechanism XPS st udies and ICP measurement was done on the aliquots obtained from different reaction mixture. It turned out in the XPS studies that in the initial aliquots concen tration of In(0) predominates and in the consecutive aliquots concentration of In (III) whic h is a component of InN , grows continuously. To inspect the effect of TMED, In N was synthesized with varying amount of TMED. In t he first , TMED, n-BuLi and in the other one 12 mmol of TMED was show that 10-15 nm sample of In metal RAY PHOTOELECTRON SPECTROSCOPY) STUDIES TO INVESTIG ATE THE our proposition about the reaction mechanism XPS st udies and ICP -OES reaction mixture. It turned out in the XPS studies that in the initial aliquots concen tration of In(0) predominates and in the grows continuously. N was synthesized with varying amount of TMED. In t he first BuLi and in the other one 12 mmol of TMED was used. In Fig. 42 we see the XPS data of two aliquots taken at 30 and 60 synthesis mentioned in the previous paragraph. In the aliquot taken at 30 In(0) is predominant where as in the 60 Figure 42: XPS data of the In 3d reaction mixture using molar ratio of concentration of In (III) and In (0) 42 we see the XPS data of two aliquots taken at 30 and 60 min. for the first in the previous paragraph. In the aliquot taken at 30 min concentration of In(0) is predominant where as in the 60 min aliquot the opposite is observed. In 3d 5/2 of aliquots taken out at 30 min (A) and 60 min (B) from the molar ratio of InBr 3, TMED, n-BuLi taken as 1:3:3 which shows relative In (0) 100 min. for the first min concentration of of aliquots taken out at 30 min (A) and 60 min (B) from the which shows relative In the aliquot drawn at 90 min compared to that of In metal or In(0) Fig. 42 suggests that, most likely an In addition of the reducin g agent transforms concentration of In(0) at the beginning. Figure 43: XPS data of the In 3d reaction mixture using molar ratio of concentration of In (III) to In (0) in the sample min (Fig. 43) the relative ratio of In(III) increases even more compared to that of In metal or In(0) . The gradual rise in the re lative concentration of In(III) in suggests that, most likely an In -intermediate was formed during the process. Initial ly g agent transforms InBr 3 into In metal which contributes to the high concentration of In(0) at the beginning. In 3d 3/2 and In 3d 5/2 of aliquots taken out at 90 min (A) from the molar ratio of InBr 3, TMEDA, n-BuLi taken as 1:3:3 (B) shows relative concentration of In (III) to In (0) in the sample ratio of In(III) increases even more lative concentration of In(III) in intermediate was formed during the process. Initial ly into In metal which contributes to the high of aliquots taken out at 90 min (A) from the (B) shows relative Later, when InBr 3 and TMED forms enough amount of the intermediate an d its concentration reaches a threshold value. At this point this inter mediate dissociates to form InN which boosts the relative concentration of In(III) in the soluti on. This intermediate now becomes the source of formation of InN hence In(III) and the dissociation becomes the dominant process. With time by more conversion into In(III) the relative ratio of the In(III) to In metal continues to grow higher. SYNTHESIS OF INDIUM NITRIDE WITH EXCESS TMED In this case the ratio of InBr 3, TMED, n-BuLi was taken as 1:12:3. The Aliquots were drawn at 30, 45 and 75 min intervals. It is found that at sa me rate of addition of the reducing agent and at same time interval, the particles formed were large r in size. The average size of the nanoparticles from TEM was ~6.14 nm with reasonably good size dis tribution. The XPS data shows lower percentage of In metal component in these aliquots. A portion of the final sample was treated with dilute 0.2 mM nitric acid and this helps to ge t rid of most of the In metal part. PXRD of the washed nanoparticles agrees to the wurtzite phase o f InN (Fig. 46). Figure 44: XPS data of the In 3d using excess TMED So, from these XPS data (Fig. 4concentration of In(0) is not found. We suppose the reason for this is favorable condition for formation of the intermediate producing InN. metals. 101-104 This leads to the fact with more conversion rate of InBr later to InN, there is not much left to be reduced to In(0). This supports the very low concentration of In(0) at the start and almost nil at the end. This favors the assumption that InBr probably forms some kind of complex with TMED whic In 3d 5/2 of aliquots taken out at different time from the synthesis 44) it has been observed that in this case even at 30 min exce ss concentration of In(0) is not found. We suppose the reason for this is favorable condition for formation of the intermediate producing InN. TMED is known to form stable complexes with This leads to the fact with more conversion rate of InBr 3 into this later to InN, there is not much left to be reduced to In(0). This supports the very low concentration of In(0) at the start and almost nil at the end. This favors the assumption that InBr probably forms some kind of complex with TMED whic h is more favored in presence of excess different time from the synthesis been observed that in this case even at 30 min exce ss concentration of In(0) is not found. We suppose the reason for this is favorable condition for TMED is known to form stable complexes with into this intermediate and later to InN, there is not much left to be reduced to In(0). This supports the very low concentration of In(0) at the start and almost nil at the end. This favors the assumption that InBr 3 h is more favored in presence of excess reactant whereas when there's more unreacted InBr 3, n-BuLi reduces it directly to form more In(0). Also from the absorption spectra of the sample prep ared with excess TMED we see that aliquots taken out at the same time as other reaction mixtur es comprised of lower InBr 3 to TMED ratio incorporates smaller nanocrystals. Keeping everythi ng same by increasing relative quantity of TMED produced larger size particles (Figure S10). T his means the rate of formation of InN nanocrystals is faster in presence of excess TMED. This also supports the idea that most likely TMED complexes well with the In(III) component and thus forms InN faster upon addition of n-BuLi. Also the PXRD pattern (Fig. 46) matches very well with wurtzite InN. Figure 45 : Absorption spectra of the aliquots take n during synthesis with excess TMED Absorbance (Arb. Unit) 1.4 1.2 1.0 0.8 Wavelength 'Al 1' 'Al 2' 'Final' 'Gaussian fit' Figure 46: PXRD pattern of InN wurtzite nature of the crystal structure So, it occurred to us that if In metal itself yield of InN, it's probably not an intermediate tow ards the end product. The other option would be then if InBr 3 forms some kind of complex with TMED which is then reduced to InN by the reducing agent as TMED is known to form stable comp lexes with metals instances of TMED being used as a chelating or comp lexing complex intermediate involved the rate of formation of that complex should be higher in presence of excess TMED. So, by monitoring the rate and yield role of TMED was investigated. InN quantum dots synthesized with excess TMED wurtzite nature of the crystal structure us that if In metal itself cannot act as an efficient precursor leading to high yield of InN, it's probably not an intermediate tow ards the end product. The other option would forms some kind of complex with TMED which is then reduced to InN by the reducing agent as TMED is known to form stable comp lexes with metals . There are several instances of TMED being used as a chelating or comp lexing agent in literature. If there is omplex intermediate involved the rate of formation of that complex should be higher in presence of excess TMED. So, by monitoring the rate and yield role of TMED was investigated. synthesized with excess TMED confirm the act as an efficient precursor leading to high yield of InN, it's probably not an intermediate tow ards the end product. The other option would forms some kind of complex with TMED which is then reduced to InN by the There are several agent in literature. If there is a omplex intermediate involved the rate of formation of that complex should be higher in presence of excess TMED. So, by monitoring the rate and yield role of TMED was investigated. CHAPTER 5: STUDY OF OPTICAL PROPERTIES PHOTOLUMINESCENCE (PL) FROM INDIUM NITRIDE NANOCRYS TALS As we were able to determine the band gap of InN us ing the techniques of optical absorption, photoluminescence was next to be explored. In a rec ent work Neale et al .44 claimed that InN is very unlikely to demonstrate any photoluminescence due to its electronic structure. In all the previous reports, it has been shown that InN exhibi ts luminescence in visible region. This was explained considering the previously reported band gap of 1.9 eV. With more research and development in this field; the actual band gap of t his material has been determined to be close to 0.7 eV; 55 so now these explanations do not seem justified an y more. We have been able to demonstrate luminescence of th ese materials in NIR region and it can be varied by tuning the size of InN. This seems much m ore reasonable because we can show that the optical band gap as well as the wavelength of e mission is related to the nanocrystal size due to quantum confinement effect. The photoluminescenc e quantum yield (QY) of the material has been found to be reasonably high considering the fa ct that we have not yet explored all the possibilities to increase the emissive properties o f these materials. The fact that these materials (without additional surface treatments to enhance t he emissive property) tend to show quite high QY of emission (detailed calculations shown later) implies that the surface of these particles do not accommodate very large number of trap states. T hus the colloidal nature and superior optical quality of these nanocrystals makes them an incredi bly appealing subject to study. In the following discussion, the PL efficiency of these ma terials will be investigated. The PL efficiency of materials from different synthetic methods will also be compared. Figure 47: Absorbance and PL Spectra of an aliquot of InN (A5) from the synthesis using OLA:OA ratio of 10:1 Figure 48: PL spectra of five aliquots of the sampl e from the synthesis using OLA:OA ratio 10:1 70x10 -3 60 50 40 30 20 10 0Absorbance 1600 1500 1400 1300 1200 1100 1000 Wavelength (nm) 30x10 325 20 15 10 50 Luminescence Intensity Absorbance Luminescence Luminescence Intensity 1600 1500 1400 1300 1200 1100 Wavelength (nm) A1 A2 A3 A4 A5 Fig. 47 compares the absorbance and PL spectra of a n aliquot of InN from the synthesis using OLA: OA in 10:1 ratio as described in chapter 3. Th e dotted lines in the figure are the Gaussian fits of the absorbance and the emission peaks. The abrupt drop ~1600 nm is due to the limited detection capacity of the instrument above this wav elength range. Figs 48 and 49 depict the dependence of the PL position on the average QD siz e. The width of the luminescence peak is correlated to the size distribution of the nanocrys tals. The spectral bandwidth of the PL spectra in Fig. 47 is comparable to the absorption peak, sugge sting that the emission is from a single excitonic state rather than an ensemble of trap sta tes. Trapped state emission from a nanocrystal solution typically has a full width at half maximum (FWHM) signicantly larger than the FHWM of the Gaussian t of the band edge peak. 105 The shift between the lowest energy peak in the abs orbance spectrum of the QD and the corresponding emission is termed as the fiStokes shi ftfl. From Fig. 47 the Stokes shift of the InN (A5) is calculated to be ~106 nm. A larger Stokes s hift is desirable for applications of QDs in LEDs because larger Stokes shift means a smaller ov erlap area between absorption and emission spectra because reabsorption reduces the total e ciency of the QDs. 106 The Stokes shift has also been reported to be strongly dependent on the QD si ze. 107 In our case too, a similar phenomenon has been observed. 10 2468100 2468 Counts 5000 4000 3000 2000 1000 0 Time (ns) Figure 49: PL spectra of three aliquots of the InN sample synthesized using 3:1 ratio of OLA:OA Figure 50: PL Lifetime decay of InN sample (final a liquot) from two different synthesis using OLA:OA ratio of 10:1 and 3:1 respectively Intensity (Arb. U) 1600 1400 1200 1000 Wavelength (nm) A1 (X2.5) A3 A4 468100 24681000 2 Counts 5000 4000 3000 2000 1000 0 Time (ns) Fig. 50 shows the excited state decay of InN nanopa rticles. The lifetime decay was found to be mono-exponential in nature. To extract the lifetime of the exited state, the decay was fit to single exponential function (black solid lines show the fi t). The lifetimes of the nanoparticles have been calculated to be 1.4 µs and 1.6 µs respectively. Pb S nanocrystals have been reported to show a single-exponential time constant measured to be ~1 µs about two orders of magnitude longer than that reported for CdSe/ZnS nanocrystals. 105,108 Colloidal PbSe semiconductor nanocrystals have also shown long uorescence lifetimes of up to 0.88 µs. 109 The decay time of the InN reported here are similar to the IV-VI NCs and is l onger than expected for a dipole transition (Fig. 50), but shorter than that expected for emiss ion related to trap states. 109 The long radiative lifetime in the PbSe nanocrystals was attributed to the effects of dielectric screening. The screening of the radiating eld inside the nanocrys tal has the effect of weakening the internal eld and consequently increases the radiative lifet ime. Thus the long lifetimes of the InN QDs observed are consistent with the effect of dielectr ic screening for semiconductor nanocrystals with somewhat high dielectric constants. The lifeti mes observed for these III-V materials and IV-VI materials studied before are found to be much longer than that of the more conventional II-VI nanocrystals which show lifetime in the range of ~10-25 ns owing to the higher dielectric constant of these materials. The highly mono-expone ntial nature of the excited state decay suggests that the surface of these nanomaterials we re well-passivated and probably lack appreciable amount of trap states. Presence of surf ace traps generally leads to highly multi- exponential behavior of the excited state decay. Th erefore, absence of such behavior clearly reflects the good surface property of these particl es. II-VI QDs with really high quantum yields seldom show such mono-exponential behavior. QUANTUM YIELD (QY) The quantum yield ( ) of photoluminescence emission is the number of ph oton emission occurrence for each photon absorbed by the system. QY is a measure of the emissive property of the particles and hence an important quantity to be determined in our case. 7..absorbed photons of No emitted photons of No Here, we calculate the relative quantum yield of so me InN samples using a dye IR-26 (See SI for PL signal) (A commercially available dye emitting i n the NIR region; QY 0.05%) as a standard. 110 The QY of the nanoparticles are calculated using th e following equation 111 : )8( 05.0 2 sxxssxnnFFAA Here, the subscripts x and s refer to the sample (I nN) and the standard (IR26), respectively. ‚A™ refers to the integrated area under the emission sp ectrum, ‚F™ refers to the fraction of exciting light absorbed at the excitation wavelength (405 nm here) and n to the refractive index of the solvent (dichloromethane for IR26 and tetrachloroet hylene for InN) A sample QY calculation for Aliquot 2 (InN sample p repared using OLA:OA=10:1) by plugging in the absorbance and integrated PL intensity value s in Equation 8 is shown below: %1.5 12.1 277 .0737 .0 10 6.42 10 2.18 1.0 36 (QY calculations for other sample in SI) For InN sample prepared using OLA:OA=3:1 the QY tab le is given below: Aliquots max (nm) Calculated QY (%) 1 1350 4.77 2 1400 4.38 3 1425 4.28 Table 1: Calculated QY (relative) for the InN sampl es prepared using OLA:OA=3:1 The calculated QYs for the different InN aliquots f rom the synthesis using OLA:OA ratio 10:1 are listed here in the table below: Aliquots max (nm) Calculated QY (%) 1 1350 2.2 2 1400 5.1 3 1425 2.6 4 1475 4.0 Table 2: Calculated QY (relative) for the InN sampl es using OLA:OA ratio 10:1 MANIFESTATION OF SIZE DEPENDANT OPTICAL PROPERTIES THE BRUS EQUATION An intuitive approach to understanding quantum dots utilizing the particle-in-a-box approach was performed by L.E. Brus. 94,112 In the following section; the relationship between the size of the nanoparticles and the band gap will be demonstr ated with the help of Brus equation. 814786 .1 1800222rerhErEgg where Eg(r) is the band gap of a quantum dot of radius r, Eg(0) is the bulk band gap, h is Planck constant (6.626 10 -34 J s), e is the electrical unit charge (1.602 10 -34 C), 0 is the permittivity of vacuum (8.854 10 -12 C 2s2kg -1 m-3 ), is the optical dielectric constant of the material , and is the reduced mass of the electron-hole pair: 9hehemmmm where me(h) is the effective mass of the electron/hole. The Co ulomb term shifts E g to lower energy as r, while the quantum localization (confin ement) term shifts E g to higher energy as r 2. Thus the apparent band gap will always increase for small enough r. The relevant empirical parameters of the bulk InN are listed in Table 3 below. These parameters have been used to relate the size dependence of the observed band gap of the synthesized InN nanocrystals to that predicted from the Brus equati on. Parameters Value Reference Eg(0) 0.69 eV 113 6.0 114 me 0.1 m 0 115 mh 1.63 m 0 116 * m 0 is the rest mass of the electron, 9.109 10 -31 kg. Table 3: Empirical data of bulk wurtzite InN. Converting the units from joules to electron-volts and meters to nanometers, and substituting the above parameters in Equation 8, we get the followin g relationship for the band gap of the quantum confined InN (Equation 10): )10( .07.0 .99.3 )0( )(22rnm eV rnm eV ErEgg Figure 51: Size vs. band gap plot for InN nanocryst als synthesized by using 3:1 OLA:OA ratio (left), 10:1 ratio (right) The curved line (dotted) shows the size predicted b y the Brus equation (Equation 10). The agreement is really good, considering the limits of Brus equation. As explained in the previous discussion there is su btle difference between the ways the two synthesized materials follow the trend predicted by Brus. Due to their platelet like shape the QDs synthesized by using 3:1 molar ratio reaches a mini mum in the band gap energy whereas the QDs synthesized via 10:1 ratio of ligands can reach the band gap energy of the bulk material (Fig. 51). 1.4 1.2 1.0 0.8 0.6 Average Bandgap in eV 10 86420Average radius in nm 'Average bandgap of the particles from different aliquots' 'fit_Brus Equation' 1.2 1.0 0.8 0.6 Average Bandgap 10 86420 Avergae radius Average Bandgap fit_Brus ELEMENTAL ANALYSIS We carried out some elemental analysis to find out the composition of the nanocrystals or the relative molar ratio of In:N. According to the work carried out by Nathan R. Neale et al ., 44 they claim that this ratio had a big impact on how the s emiconductors behave. We tried to figure out if the materials synthesized with different methods ha d any difference in their relative composition and how significant this ratio is as far as their e lectronic properties are concerned. For this purpose the amount of In was found out by a techniq ue called inductively coupled plasma atomic emission spectroscopy (ICP-AES) and the quantity of nitrogen was calculated from CHN analysis. INDUCTIVELY COUPLED PLASMA ATOMIC EMISSION SPECTROS COPY Inductively coupled plasma atomic emission spectros copy (ICP-AES), also referred to as inductively coupled plasma optical emission spectro metry (ICP-OES) is one of the most powerful and popular analytical tools for the deter mination of trace elements in a myriad of sample types. 117 This technique is based upon the spontaneous emiss ion of photons from atoms and ions that have been excited in a radiofrequency (RF) discharge. Liquid and gas samples can be injected directly into the instrument, while sol id samples require extraction or acid digestion. The sample solution is converted to an aerosol and directed into the central channel of the plasma. At its core the inductively coupled plasma (ICP) sustains a temperature of approximately 10 000 K, which vaporizes the aerosol quickly and a nalytes are liberated as free atoms in the gaseous state. Further collisional excitation withi n the plasma imparts additional energy to the atoms, promoting them to excited states. Sufficient energy is often available to convert the atoms into ions and subsequently promote the ions to exci ted states. Both the atomic and ionic excited state species may then relax to the ground state vi a the emission of a photon. These photons have characteristic energies that are determined by the quantized energy level structure for the atoms or ions. Thus the wavelength of the photons can be used to identify the elements from which they originated and the total number of photons is direc tly proportional to the concentration of the originating element in the sample. CHN ANALYSIS CHN analysis is a form of elemental analysis concer ned with determination of only Carbon (C), Hydrogen (H) and Nitrogen (N) in a sample. The most popular technique behind the CHN analysis is combustion analysis where the sample is first fully combusted and then the products of its combustion are analyzed. 118 The complete combustion is usually achieved by pro viding abundant oxygen supply during the combustion proces s. In this method, the carbon, hydrogen and nitrogen atoms of the product under analysis ox idize and form carbon dioxide, water and nitrogen oxide, respectively. Then these combustion products are carefully collected and weighed. The weights are used to determine the elem ental composition, or empirical formula, of the analyzed sample. Findings from the ICP-AES Studies: Yield of reaction for 3:1 system = 31%; Conc. of In from ICP-AES = 15.68g/L Yield of reaction for 10:1 system = 38%; Conc. of I n from ICP-AES = 19.20g/L From CHN analysis, the percent quantities of these elements were established to be as following: Sample C H N In In:N Expected value 0.1-10% 0-0.25% 10-11% 80-90% OLA Only 3.12 -- 12.72 84.16 0.733:0.9 = 0.81:1 3:1 (A) 1.57 -- 10.55 87.88 0.765:0.75 = 1.015:1 3:1 (B) 2.31 0.10 11.07 86.62 0.75:0.79 = 0.948:1 5:1 (A) 1.47 -- 10.97 87.56 0.76:0.78 = 0.97:1 7.5:1 (A) 2.05 -- 12.87 85.08 0.74:0.92 = 0.8:1 10:1 (A) 2.43 0.16 11.44 86.13 0.75:0.817 = 0.917:1 10:1 (B) 2.35 -- 12.38 85.27 0.74:0.88 = 0.84:1 Table 4: CHN analysis of InN samples prepared using different OLA:OA ratio As we can see the data from ICP-OES analysis and fr om CHN analysis support each other in the sense that amount of In found by both technique is almost equivalent. Also, this set of data shows that the In:N ratio is >1. The material synth esized by Neale et al., 44 is reported to have excess metal [In(0)] content; and this is suspected to be the reason behind the electron rich surface and the Burstein-Moss effect which is respo nsible for the absorption onset at 1.8 eV. Due to this phenomenon, the excitonic feature of th e semiconductor is not revealed and the band gap transition cannot be noticed in the InN nanocry stals synthesized by the Neale group. However, the InN nanocrystals synthesized by us do not have substantial In(0) content and is thus unlikely to show this effect. CONCLUSION In this report the synthetic methods and mechanisti c details to obtain high quality, monodispersed, colloidal, wurtzite InN have been de scribed and the size dependant optical properties of indium nitride nanocrystals have been illustrated for the first time. It has also been depicted that not only the size but the shape of th ese nanocrystals can have significant effects on the electronic structure and the optical properties of the semiconductor. We have been able to synthesize different sizes and the shapes of the In N nanocrystals by varying the reaction conditions and changing the ratio of surfactants re spectively. These nanocrystals turned out to be emissive; which is an impressive breakthrough for t echnological applications. Lifetime measurements provide us vital insight towards the n ature of the exciton. We have obtained a very precise idea about the Bohr radius of our material of interest and have also been able to study the correlation of the size dependence of the band gap with that predicted by Brus. An outline for the reaction mechanism has also been provided by identi fying the undesired by-product which is In metal. Though we have not been able to isolate the complex formed as the intermediate during the synthesis but we have enough evidence to stand by the fact that a bi-dentate amine complex of In forms in the course of the reaction with TMED . This technique produces nanoparticles with a reason ably high yield. This method can be applied to synthesize various nitride nanocrystals and also utilized to dope these nanocrystals. This report illustrates the technique to obtain high quality, s ize and shape controlled and emissive colloidal InN nanocrystals for the first time along with the detailed study of its electronic structure and optical properties. FUTURE WORK A number of new techniques have been developed for doping of nanocrystals and it'd be an interesting step to explore their effectiveness tow ard InN. Also, since we don™t have very clear idea about the nature of the intermediate and the r eaction pathway, it's our aim to find a more proficient way to decipher it. The material shows t o be very promising as far as its emissive nature is concerned. It would be worth investing so me time to look into ways of increasing the quantum yield of InN since that is highly desirable for several industrial applications. Synthesis of colloidal, monodispersed III-V nanocrystals, suc h as; gallium nitride (GaN), indium and gallium phosphide (InP, GaP) and study of their siz e dependant optical properties would be our next priority. These materials having very differen t band gap values could prove to be very useful in synthesis of alloy materials covering a w ide range of spectrum. We also plan to investigate the effect of doping (transition metals ) on the magnetic and electronic properties of these materials for photonics and spintronics appli cation. BIBLIOGRAPHY BIBLIOGRAPHY !"˙#˝$%!&˙ #˝'%(˜ )ˇ)"˙#˝˝ # # *˙ ˆˆ#*%+˜ #' # # !"˙#˝$%(˜ )ˇ)"˙#˝˝ # ˘ # ˝, ˛ˆ˙ #˝- # ˇˆ˘ # .ˇ)˛,,#*/%0,˙#1$ ˙˝ # ˛ˇ # 2˛ˆ3˜"#-%*˙˚˛, "#˝ ˚˜ !"˚ # #ˇ # '˛˚˛ˆ "3#˝˝%'4˜#56%+˙7#58˝ %.44ˆˆ#9%:˙,˙ˆ˙ "˛3˛#.˝%*˛&˛,˙# !; ˚˜ !"˚ # #$ # <˛4˙,˘#=%-, ˇ)#2 % !!& #5˙,,; ˆ3# .˜#5$%>˜,˚#' '"""%˚(˜ # ˆ# ?˙ˆ˜˛#.˝%9#=%5)˜#$.%'#> (˝˚ # # -ˆ˙&&#-2%/+˛˛ #.- !)˚*˚˚!!˚ # ˘ # $=˙#(%0#<%:)˛˚#2%'#8%:)˜#?% @˜#'%0˙ )"˙#.%8 ˛#9%8˙"˛˙# 1%@˛"˙#2%("ˆ3˛4 "3#.%'#=5 '˚"˚&*˜ # -˛"˙#=$ +)˚!˚ %/˙AB7@˙"#B@# B˜˛3#/˝%!"˘ˆ#=6 *˚ # # /ˆ#1˝%$#06%5˛ˆ#5< %0˜Cˆ ˙#B$%0˛˛#!1 ,))˝ # # 2˙˛#+ !˚-*˚.*˚"˚ # ˘/ # < )#=5#2$8# ( "#2%("˜4˙#0%0˙4˛3˛ )#˝%()ˆ˛#= %?˜D˙"˛#< (!˚ # $# 5)ˇ)4˜#.%˝C˜)˛ˆ˛#8%.˙ˆ˛#8%B˛"˛˜˛ #. ,))˝˛0$˘1˘˛˛2 # ˛ # =˜C˛#*%<˛)#<%˝˙˚# ,˜ # ˇ#$ # >˙˚#@%>#@%'#%'#>%:)˛˚#* (3 # ˇ # >˛˙#=%>#@%8˛˚#*%'˜˙#9 '˚˜ # $ˇ # .˛˘˛#0%/˛"#?'%6˙˘˛˛D#˝%. ")#22%*˛˙#5B* # ˘# +˛#@%'˜#%>˜#>%5˜#/%<˛˙#>%? ˚#>%9˛˚#E ˚4˚3 # ˇ$˘ # ?˛"#˝5%.ˆ˙7˛ #?%.˜ "#<%-ˆ C"˙7#<%˝4˛ˇ)#5*%6 ˚#2%? ˇ)# *˝ , # ˘ˇ/ # .˛˘˛#0%/˛#2%.ˇ)7C#+%*˛˙#5B *˚ # ˘ # /˚˛#./%*˛ˆ)#B-%2˛"˙7ˆC#-/ %644˙ #-5%+˜)˙#9! ,˜30 '" # #˛ # '˜#*.%.)#5<%5)˛#8.%6˜˘˛"˛ ˛#*%<˜#.?%'#=6%<˜˛˚#5@ %5˜-"˜˚˜ # # +)˛˜˘#.%E#=%<˜#!%+,ˇ)#˝ 2 (˝ # ˆ# +˛"#*8%6˙˙˚,-˛ˆˆ A <˛ˆ)#=*%.)˛˚#== ˙ # ˇˆ # 0,˙#1$ (˚ !!& %˘˘%5*5- A+˙ˇ˛*˛ˆ˙#?'# ˝,,#-2%9˛,"˛#+=%+˛7˘#26 ,˜30'" # $˛ # .ˇ)7C#+%2#5%2˛ ˛,˛#(%. )˛˘ #*%/+˛˛ #.-%2 )˛#;0 *˚ # ˘ # '#<@#<%@˜#.%:˙˜#6%'#@%'˜# .%@˛˚#. ,))˝ # ˛ # 6˙˙˘7#8='#1=%* 4˜˘#.<%0 ˘3#$2%'#<9< ,))˝ # ˆ/ # +)˛˜˘#.%E#=%<˜#!%+,ˇ)#˝ 2 (˝ # ˆ# 5)˛˚#@'%'#?%?˛ˆ)#˝%2# (˜ # ˇ/ # 5)˛˚#@'%2#'#? ,6!*˚ # ˇ/ # 5 ˛#2%0#@E%=#9%6˜˙#<%2# 8 ,))˝ # ˛ # < )#=5%@˜#/.%<˜#!%+,ˇ)#˝ 2 *˚ # ˇ/ # '#'%<˜#.%'˜#@%>#@ (3 # ˇ˛ # 5)#%'%@%>˜%>%5˛˙#5 , # ˘#$ # -˛,˙˛"#-0+%2,,#!2%B˛,#B * , # ˘# # @˙&&#˝/ ,$ˇ1ˇ2˘˛˛#0˘ˆ# # $ˇ # 0˛ ˆ#2˝$ % %1˙,# ˝ )˙˙#*5 (!˚ # #ˆ˛ # *˙˚˛ˇ)#˝%0 )˛7#.%+˜ˆ#2%<˛ ˙# 2%0˙˙7 "#˝%!3ˇ)˜,,#˝%9,,#< ,*˚ # ˘˘ # .˛˚ˆ#!< ,*˚ # ˘ˆ # *˙˚˛ˇ)#˝'%!3ˇ)˜,,#˝%<ˇ"3#.6% 0 )˛7#.1 # ## B˙C"#˝ ˝ # $ˆ # B˙C"#˝=%+˛˘#25%'˜ˆ)#=2%'˛ 7#2%!,,˚ ˙#*=%=˙) ˙#=5 ˙ # ˘˘/ # '˜#9%8˛,,˛#/ , # ˘#$ # 8˛3,˙#-B%<&&˛#=%-˜4,ˇ˛ˆ˙#; .-˝#!˘%1˙,;.F˝ 2˛ˆ ˜˙"˛#8%("˛˙ˆ˙#<%B˛"˛˙#2%<˛˛# <%0˜˙ˆ˙#! ,))˝0 # ˘ # /˛3˘˙#1@%0,˙ˇ)")#˝˝ '!(˚˙, %''5#+˛ˆ˙*˙˜˚A?'# 83˛˚#1˝%!˜ ˆ˚#1˝%0˛ ,˙#˝ 2%˝˘˛#˝?%2˛,˛ˆ˘˙˜#1@ # ˘˘ # 9 ˆ˛#0'%'˛ ˙#*-9%+ˆˆ#2= 7%, # # 8˛ ,3#8'%?˙,3#5- ,)) # ˛ # 2˙"˙ˇ#< (˚!˚& %.˚A<˘,4˚# 2˛˘,˜˚#( !˚8&9:5 %˘˘%.˚# @˛˛˙ˆ˙#˝%8 ˜D˙#2%()"˜4˙#2%<˛ ) ˙ˆ˙#˝$ ˆ'˚ "˜˚˜-˚17".ˆ2 %!, .ˇˇAB˛˚˙3˛#=˛˛### 8˛ ,3#8'%?˙,3#5- "˝ # ˇ/ # =" #/9%/˙7#=/ ˙; # #˛ # $"˜ˆ˛#0%$˙˜#@%8˛"˛#( %6 # ##$ # /˛3˘˙#1@%0,˙ˇ)")#˝˝%. 3˛#* -%!ˆ #11%$˛˙#.1%+ˇ) ˆ˘ˆ#?% ?˜ˆ)˜,,#=%<˛˛#<%2˜˘3#˝1%˝˘)˙, ˘#=%.ˇ)˙˛#(%6˛˜,#= ; # ˇˇ˛ #* 9˜#=%9˛,˜"7ˇC#9%@˜#02%˝˚$$$# =9%<˛,,#!!%'˜#<%.ˇ)˛& = ,)) ˝ # / # +˜ˆˇ)#0.˝%8˛ ,3#8' !)˚*˚˚! # # # .)˜4˛#81%$˛˙#.1%2D"#1B% .˙,3 )"˙#//%1" )#1˝%0˙#-.% 1˛ ˙#˝%'3˛#=%0˛˙"#˝%˝˛˙#<%. )˙˙#0%0˛ ˇ#˝%2˙˛#+ ˙˝ # ˛ˇ # .ˇ)ˆ#-%<˙#=5%>˜#>%˝ˆ˛˚#*% 94#!*%!#*%0 ,˙7 "#5 # ˘# # .ˇ˙ˆˆ#0%+˜ˆˇ)#˝%9ˆ4ˆG?˙˜H˜ˆ#2 %5)#--8%-ˇ#0!%8 #<% .) ˆ)˛#.0%.)˜4˛#81%$˛˙#.1%9˜) #*%-),, #2*%2˙˛#+ # ˇ# 9˜#=%9˛,˜"7ˇC#9%.)˛#9%@˜#02% ˝˚$$$#=9%<˛,,#!!%'˜#<%.ˇ)˛&&# 9= ˙; # # /˛3˘˙#1@%0,˙ˇ)")#˝˝%!ˆ #1 1%0˜˘3˜"˙#/˝%$˛˙#.1%1" )# 1˝%+ˇ) ˆ˘ˆ#?%?˜ˆ)I,,#=%˝˘)˙,˘#= %6˛˜,#=%2˜˘3#˝1%<˛˛#<%<˛ )˙ˆ˙# ˝% @˛˛˙ˆ˙#˝%<˛,,#!! ; # ˇ#$ # <˛4˚#$%6˛H ˆ#56 ,)) # / #* @˙ )˙ˆ˙#2%@˛˛˙ˆ˙#<%<˜˛˚#9%<˛ ˛#<%.˛˛#=%5)˛3˛)˛˛#˝%<˙˙# @ ,))˝ # # # +)˜3˛#˝6%.˜˚ˆ˛#0%0˛ ˛ )˛#0%<˛ )˙ˆ˙#˝%@˛˛˙ˆ˙#˝%/˛3˘˙#1@ ,))˝ # # # 2,,˙/˙˚J#5˘%',D˙ˆ)#-%1˛˛", 4˚)#/˝2 # ˘# 6#2 !˚˚+)*˚ # # 5˜4,˛˘#*9%+,˛#*6%9˛,,˛ˇ#5 <%*3˙,˘ #80%0˛#*+ ; # ˘/ # +˛#@=%'˜#6%>˜#>6%5˜#/'% <˛˙#>-%?˚#>%9˛˚#E' ˚4˚3 # ˇ$˘ # 9,, #*'%=˛"#=? "!˚'˚˜ # ## # 9˜#5%'#89%'#'@%<˜#'@%' ˜#@%>#@ (3 # ˇ˛ # >˛˙#=-%>#@%8˛˚#*%'˜˙#9 '˚˜ # $ˇ # >˜#@B%5)˚#9@ ˙;8*˚*˚ # $ # $ )˘˛#2%2˙˙"˛#8%<˛˛˙"˛#/%8˛3˛# 2%?˜Dˆ˛#.%0˛7˛"˛#@%<˛˛˘˛#2% .˛ ˛"#8%2˙#@%6˙˙˚,-˛ˆˆ A ˝3ˇ)#>%/˙ˆ#!1#!˘%B˙˛.ˇˇ AB7@˙"# .ˇ)#- ;!˜˚4˚%.ˇ)&&#-!%.ˇ)˛˛"# *! , # ˘#/ # $4JMC#=%(,˛#*%2˛DN#?=%.˚˜˛# ˝%@˛˛˚˜ˇ)#8%B˛ )#@%5˜ ˇN#*%˝ˆO # ' ˙; # # /˛˚˙ˆ#?%?˛˘˙#=%-H˜#5%5˛˜ ˙ˆ#2 ˙; # # +˜ #'! # / # @#@%˝, ˛ˆ˙ #˝- (!˚ # $#ˆ # 07˙#.6%<3˙#8 ,˙ # $˘ # 9˛˚#>%-˚#E%'#@ ,˙ # $/ # /)#58%B˚˜3#8/%0,ˆC#?%/˙#8 ( ,( # ## =˜#@%5˛ ˜,˛#2?%.#=%0#.@% 5)˙#=%˝, ˛ˆ˙ #˝- , # ˘ˇ # 5)˛˚#@<%'˜#@.%<˙˚#@'%0˜˙#5 8%67˙#.%@)#=˝ ,)) # ˘/ˆ # 5#2*%*˙ 4˜˚)#'!%8˙ ˙#B 5%93)I,,#8%+˚˛#*6% 8˙ ˆ#?/%9˚)˛˘ˆ#0 +˚˜ # ˛ # 5˙˛˜#˝B%0)˛˘"˛#5+%0˙˛#' , # # 0)˛˘"˛#5+%!ˇ))K&#˝%9˚˘#?%5 ˙˚˛#=? '˚˜ # ˘ # 0)˛˘"˛%2˛ˇ/˙˛,˘#/6%'˛#@?#%%-˙ 7,,#˝0%5˙˚˛#=? '˚˜ # $˛ # 9˛#=<%8)˙ #!%9˛ˆˆ#˝*%<ˇ "4˚#B*%*˜4 CˆG/˜,˙#< (˜ # ˘ # <˜#=%'#'%@˛˚#9%2˛˛#'%9˛˚#' %˝, ˛ˆ˙ #˝- # ˇ˛ˇ # 0˜˙#2%'#=0%/˛44˙˜ #+(%2"˜, ˇ#?1%+˛7˘#26 # ˘/ # 2˙, #$%'˛4ˆ#0%.ˆ #/%/2˜3 ˇ"#/%B˙,,ˆ#8%2˛ˆ #=5%1˛)˛ˇ"# ?%1˛ˆ˙#˝%/,˜#5%˝,,˛#6%< # ,( # ## 9)4˚#+'%9˛˚#5%6˜3˙ˆG.˙ ˆ# - ; # ˘/ # .˙#(!%=˙) ˙#=5%'˜ˆ)#=2 %2˘˚ˆˆ#˝6%B˙C"#˝=%+˛˘#25 ˝ # ˘ B˚) 7˛˘G*,#.- 6!˚ # ˘ # +˜ #'! # ˛/ # /˛3˘˙#1@%0,˙ˇ)")#˝˝%. 3˛# *-%!ˆ #11%$˛˙#.1%+ˇ) ˆ˘ˆ#?% ?˜ˆ)I,,#=%<˛˛#<%2˜˘3#˝1%˝˘)˙, ˘#=%.ˇ)˙˛#(%6˛˜,#= 1:2 # ˇˇ˛ #* :˜4,˙#˝ ˚)˚-,!˚*˚4(0 ,(0'(0;(004 =˙)9,3P.˙ AB7@˙"# '˛4ˇ)ˆ#9*%.˚˛,,#+ ˙; # $ˆ # -˜˚)#.0%/˜˚˘˛,#/=%+˛˘#.%˝4˛ #*˝ % # ˘$ # <˙˜#>%=˙ #+8 '!!)>+)")˚ ˚ %=˙) 9,3P.˙ 'ˆ˘A5)ˇ) ˆ# ?˛˘˛#1-%8")˙˛#1/%B"˜,ˇ)˛# (B ?!˚,5=0 # # #