Citation
Solid State, Transparent, Cadmium Sulfide-Polymer Nanocomposites

Material Information

Title:
Solid State, Transparent, Cadmium Sulfide-Polymer Nanocomposites
Creator:
KOTHURKAR, NIKHIL K. ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Acetates ( jstor )
Aggregation ( jstor )
Cadmium ( jstor )
Nanocomposites ( jstor )
Nanoparticles ( jstor )
Particle size classes ( jstor )
Polymers ( jstor )
Solvents ( jstor )
Sulfur ( jstor )
Transmittance ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Nikhil K. Kothurkar. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
2/28/2005
Resource Identifier:
436098585 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

SOLID STATE, TRANSPARENT, CADMIUM SULFIDE-POLYMER NANOCOMPOSITES By NIKHIL K. KOTHURKAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Nikhil K. Kothurkar

PAGE 3

.`. nih }anen sz< pivimh iv*te, tTSvy< yaegs kalenaTmin ivNdit. Verily, there is nothing in the world more purifying than knowledge. He who prepares his mind through dedicated efforts is sure to find it within. —Bhagavad Gita IV 38 . mm AaTmSvpay gurve smipRtimdm!. This dissertation is dedicated to my Guru, who is the Indwelling Spirit of all.

PAGE 4

ACKNOWLEDGMENTS I would like to thank my doctoral adviser and guide, Dr. Anthony B. Brennan for being the wonderful adviser he has been. He inculcated the value for independent thinking and creativity in me from the beginning. I am glad that he gave me the freedom to choose and pursue my research objectives and guidance when needed. In my opinion he has been the perfect combination of freedom and guidance. I would also like to thank my supervisory committee for their time and efforts. Dr. Paul Holloway has been a constant source of inspiration and encouragement to me and I owe him a big thank you. I also appreciate Dr. Douglas for the very informative and useful discussions that I had with him. I thank Dr. Ronald Baney, who has been very kind and understanding in agreeing to substitute on my supervisory committee at a short notice. I thank Dr. Joseph Simmons whose advice, help and encouragement were invaluable during the initial stages of my research. I thank my friends and colleagues from the Brennan Research Group, Clayton Bohn, Dr. Brian Hatcher, Dr. Adam Feinberg, Leslie Wilson, Michelle Carman, Jim Schumacher, Kiran Karve, Thomas Estes, Dr. Charles Seegert, Licheng Zhao, Kenneth Williams and Wade Wilkerson, who have been very cooperative, understanding and helping. I thank, Drs. Jeanne Macdonald, and Jeremy Mehlem who trained me on several instruments and introduced me to graduate research. My thanks to Sherly Jules, who helped me carry out some experiments important for my research. Jennifer Wrighton deserves a special mention for being the best secretary I have met. It is no wonder that iv

PAGE 5

her understanding, resourcefulness and consideration for everyone, has made her the favorite of everyone who interacts with her. All my thanks and best wishes for her. I owe thanks to the Major Analytical Instrumentation Center for the training and use of their equipment and to Matt Olzta and Dr. Kerry Siebein for their patience and efforts in doing TEM on my samples. I thank Dr. Eric Van Stryland, Dr. David J. Hagan, Joel Hales, Miheala Balu and William Shensky of the Center for Research in Optics and Lasers, University of Central Florida for providing their advice, expertise and equipment for the nonlinear optical measurements. My special thanks to Dr. Jill Verlander Reed at the Electron Microscopy Core, Health Science Center, University of Florida, for providing training and use of the transmission electron microscope (TEM). I also thank Melissa Ann-Lewis, Wencui Zheng, Debbie Akin, for their painstaking efforts in microtoming my samples for TEM I would like to mention my family including my parents, Kishore and Jayshree Kothurkar, my sister, Shirin Navlani and my grandparents, Professor V. K. Kothurkar and Indu Kothurkar. Without their efforts and sacrifices, I would not have been able to pursue my graduate studies here. Murali Rangarajan and Sheshadri Thiruvenkadam have been more than brothers to me as the entire Maitri family has been more than a family to me. I thank them from the bottom of my heart. My wife, Devi has had to go through a lot of hardships and difficulties during the course of my graduate studies. I cannot fully make it up to her but I thank her for being ever so supportive, understanding and just being herself. My daughter Arsha has taught me a lot of patience and given me a sense of responsibility, which has helped me tremendously during my studies. v

PAGE 6

Lastly I thank and salute to Swami Adhyatma Chaitanya, who is my Guru and best friend. He has been a glorious source of wisdom and inspiration for me throughout my studies. What he has given me transcends all that health, wealth and human relations can give. vi

PAGE 7

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................xi LIST OF FIGURES.........................................................................................................xiii ABSTRACT.....................................................................................................................xvi CHAPTER 1 INTRODUCTION........................................................................................................1 Introduction...................................................................................................................1 Applications..................................................................................................................2 Optical Limiting....................................................................................................3 Semiconductor-Polymer Nanocomposite Solar Cells...........................................4 General Requirements of Nanocomposites for Device Applications...........................5 Specific Aims................................................................................................................6 2 BACKGROUND..........................................................................................................7 Review of Various Efforts to Synthesize Metal ChalcogenidePolymer Nanocomposites.......................................................................................................7 Quantum Size Effects in the Absorption Spectra of Semiconductor Nanoparticles..10 Summary.....................................................................................................................16 3 CADMIUM SULFIDE-POLYSULFONE NANOCOMPOSITES............................17 Introduction.................................................................................................................17 Criteria for Selection of Materials.......................................................................19 Materials and Methods...............................................................................................20 Sulfonation of Polysulfone..................................................................................20 Synthesis of CdS-PSF Nanocomposites..............................................................21 Phenolated CdS-PSF nanocomposites—Method I (PhCdS-PSF-I).............22 Phenolated CdS-PSF nanocomposites—Method II (PhCdS-PSF-II)...........23 Amine-capped CdS-SPSF nanocomposites—Method I (AmCdS-SPSF-I).24 Amine-capped CdS-SPSF nanocomposites—Method II (AmCdS-SPSF-II)25 Fluorothiophenol-capped CdS SPSF nanocomposites (FPhCdS-SPSF)......27 vii

PAGE 8

Uncapped CdS-PSF nanocomposites (UCdS-PSF)......................................28 Characterization...................................................................................................29 Results and Discussions..............................................................................................30 Optical Properties................................................................................................35 Scanning Electron Microscopy (SEM)........................................................36 Transmission Electron Microscopy (TEM)..................................................38 Conclusions.................................................................................................................39 4 CDS-EPOXY NANOCOMPOSITES........................................................................41 Introduction.................................................................................................................41 Materials and Methods...............................................................................................43 Synthesis..............................................................................................................43 Characterization...................................................................................................47 Results and Discussion...............................................................................................49 Optical Images.....................................................................................................49 UV-Visible Spectroscopy....................................................................................50 Fourier Transform Infrared Spectroscopy (FTIR)...............................................55 X-Ray Diffraction................................................................................................57 Transmission Electron Microscopy.....................................................................61 Two Photon Absorption Spectrum (TPA)...........................................................64 Thermogravimetric/Differential Thermal Analysis (TG/DTA)..........................65 Conclusions.................................................................................................................65 5 AGGREGATION CONTROL IN CDS-EPOXY NANOCOMPOSITES.................68 Introduction.................................................................................................................68 Materials and Methods...............................................................................................68 Synthesis..............................................................................................................68 Characterization...................................................................................................69 Results and Discussions..............................................................................................70 Stabilizer Molar Mass and Sol Stability..............................................................70 Stabilizer Concentration and Sol Stability..........................................................71 Order of Addition of Ingredients.........................................................................72 CdS Concentration and Sol Stability...................................................................73 Conclusions.................................................................................................................75 6 FACTORS AFFECTING TRANSMITTANCE AND BAND EDGE OF CDS-EPOXY NANOCOMPOSITES..................................................................................78 Introduction.................................................................................................................78 Identification and Measurement of Responses....................................................78 Identification of Factors......................................................................................79 Materials and Methods...............................................................................................80 Levels of Factors.................................................................................................80 Design of Experiments........................................................................................81 Synthesis..............................................................................................................82 viii

PAGE 9

Characterization...................................................................................................83 Results and Discussions..............................................................................................84 Response—Percent Transmittance......................................................................84 Response—Band Edge (onset)............................................................................87 Conclusions.................................................................................................................92 7 TEMPERATURE DEPENDENCE OF BAND EDGE IN CDS-EPOXY NANOCOMPOSITES................................................................................................94 Introduction.................................................................................................................94 Materials and Methods...............................................................................................94 Synthesis..............................................................................................................94 Characterization...................................................................................................95 Results and Discussions..............................................................................................95 Measurement of Band Edge................................................................................95 Particle Size Estimation from Band Edge Using Theoretical Models................97 Comparison of Experimental Data with Regression Model................................99 Causes of Temperature Dependence of Band Edge..........................................100 Transmission Electron Microscopy of Films Containing Sols Synthesized at Different Temperatures..................................................................................102 Conclusions...............................................................................................................105 8 HIGH CDS CONCENTRATION NANOCOMPOSITES.......................................108 Introduction...............................................................................................................108 Materials and Methods.............................................................................................109 Synthesis............................................................................................................109 Amine stabilized, in situ method (ASI)......................................................110 Prepolymer stabilized, two step method (PST)..........................................111 Prepolymer stabilized, in situ method (PSI)...............................................112 Characterization.................................................................................................113 Results and Discussions............................................................................................114 Optical Images...................................................................................................114 UV-Visible Spectroscopy..................................................................................115 X-Ray Diffraction..............................................................................................118 Conclusions...............................................................................................................119 9 CONCLUSIONS......................................................................................................121 APPENDIX A MEASUREMENT OF BAND EDGE......................................................................129 B SENSOR PROTECTION AND NONLINEAR OPTICS........................................133 Sensor Protection......................................................................................................133 Requirements of an Optical Limiter.........................................................................134 ix

PAGE 10

Current Research in Optical Limiting.......................................................................136 Nonlinear Optics.......................................................................................................137 Physical Processes Leading to the Nonlinear Absorption or Index of Refraction...138 Nonlinear Absorption........................................................................................138 Nonlinear Refraction.........................................................................................140 Electronic Polarization......................................................................................141 Band Renormalization:......................................................................................141 Raman Induced Kerr Effect...............................................................................142 Electrostriction..................................................................................................142 Molecular Reorientation....................................................................................142 Thermal Effects.................................................................................................142 Optical Kerr Effect............................................................................................143 Nonlinear Scattering:.........................................................................................143 Summary...................................................................................................................144 LIST OF REFERENCES.................................................................................................145 BIOGRAPHICAL SKETCH...........................................................................................153 x

PAGE 11

LIST OF TABLES Table page 3-1. Materials used for the synthesis of sulfonated polysulfone........................................20 3-2. Capping agents for CdS..............................................................................................21 3-3. Materials used for the synthesis of phenolated CdS-PSF—Method I........................22 3-4. Materials used for the synthesis of phenolated CdS-PSF—Method II.......................23 3-5. Materials used for the synthesis of amine capped CdS-SPSF—Method I.................25 3-6. Materials used for the synthesis of amine capped CdS-PSF—Method II..................26 3-7. Materials used for the synthesis of fluorothiophenol-capped CdS-SPSF...................27 3-8. Materials used for the synthesis of uncapped CdS-PSF nanocomposites..................29 4-1. Materials used for the synthesis of CdS-epoxy nanocomposites...............................44 4-2. Typical recipe for making CdS–epoxy nanocomposites with 1% w/w CdS..............46 4-3. FTIR peak identification and assignment...................................................................55 5-1. Stability of sols using different Jeffamine stabilizers.............................................70 5-2. Stability and morphology of particles with amine concentration...............................72 5-3. Synthesis of CdS sols in mixed stabilizers.................................................................73 6-1 Design Summary.........................................................................................................81 6-2 Design Matrix..............................................................................................................81 6-3. Factors affecting band edge and % transmission in CdS-epoxy nanocomposites......81 6-4. Measurement of Response variables viz. band edge and % transmittance................82 6-6. Model Fit....................................................................................................................85 6-7. Model Coefficients.....................................................................................................85 xi

PAGE 12

6-8. Analysis of variance table [Partial sum of squares] for Selected Factorial Model.....88 6-9.Model Fit.....................................................................................................................88 6.10. Model Coefficients....................................................................................................88 7-1. Measured band edge values for sols and films...........................................................97 7-2. Comparison of experimental data with model............................................................99 8-1. Materials used for the synthesis of CdS-epoxy nanocomposites.............................110 8-2. Amine stabilized, in situ method for films with 17 wt % CdS.................................111 8-3. Prepolymer stabilized, two step method for thick films with 21 %wt CdS............112 8-4. Prepolymer stabilized, two step method for thin films with 21 wt% CdS..............112 8-5. Prepolymer stabilized, in situ process for thick films with 21 wt% CdS................113 8-6. Prepolymer stabilized, in situ process for thin films with 21% wt CdS..................113 xii

PAGE 13

LIST OF FIGURES Figure page 1-1. Schematic diagram of an optical limiter used for optical sensor protection................3 1-2. Schematic diagram of CdSe nanorod-polymer solar cell. ...........................................4 2-1. Band edge vs. particle radius of CdS using tight binding calculation with effective mass approximation for comparison........................................................................15 3-1. UV-visible absorption spectrum of uncapped CdS-PSF nanocomposite film...........35 3-2. Optical photograph using Sony DCR-TRV20 camera on printed text.......................35 3-3. SEM SEI of CdS-PSF film at 15KV, 25000x. Bar = 2 m........................................37 3-4. ED X-ray spectra of CdS-PSF (resolution = 68 eV, Quantitative method: ZAF) showing A) Cadmium-rich and B) Polymer-rich regions........................................37 3-5. TEM micrographs of CdS-PSF at 80KV, on Zeiss EM 10A showing A) CdS small particles and B) aggregates of CdS made in PSF solution without capping agent and C) Small particles formed on heating partially dried PSF film containing Cd ac and thiourea.....................................................................................................................38 4-1. Optical photographs taken on a Sony DCR-TRV20 camera on printed text. A) 1.11 mm thick CdS-epoxy film. B) 0.36 to 0.16 mm thick CdS-PSF film......................49 4-2. UV-visible-near-infrared absorption spectrum of A) Pure (neat) epoxy and CdS epoxy film. B) UV-visible spectrum showing DMF + D2000 and CdS sol synthesized at 69 C. C) Pure (neat) epoxy and CdS-epoxy film with CdS synthesized at 69C..................................................................................................50 4-3. UV-visible spectra of reaction mixture during synthesis of CdS sols.......................53 4-4. CdS band edge as a function of progress of reaction. Red shift in band edge is indicative of particle growth....................................................................................54 4-5. FTIR spectra of DMF, cadmium acetate and thiourea solutions, and D2000 showing the important peaks and the complexation of cadmium by DMF............................55 xiii

PAGE 14

4-6. FTIR spectra of reaction mixture containing cadmium acetate, thiourea solutions in DMF and D2000 showing the complexation of cadmium by DMF and its recovery as the reaction progresses.........................................................................................56 4-7. XRD spectra of CdS-epoxy films. Sample 0 = pure epoxy. Sample 1 = CdS epoxy film (1 % CdS w/w, synthesized at 70 C)...............................................................59 4-8. XRD spectra of CdS sols. Sample 0 = DMF + D2000. Sample 1 = CdS in DMF+D2000 (2 % CdS w/w, synthesized at 70 C)...............................................59 4-9. Electron diffraction on a large cluster in CdS-epoxy films with 0.7 % w/w CdS on JEOL 2010F, 200KV................................................................................................61 4-10. TEM micrographs of 1.5% w/w CdS sols in DMF +D2000 showing small particles, medium and large sized aggregates..........................................................61 4-11. TEM micrographs of CdS-epoxy films with 0.7 % w/w CdS films showing: A) Small particles on Zeiss EM 10A, 80KV. B) Large aggregates and their vicinity on JEOL 2010F, 200KV................................................................................................62 4-12. The two-photon absorption spectrum (TPA coefficient () as a function of the equivalent 1-photon wavelength) of CdS-epoxy films with 0.75 and 1.9% CdS synthesised at different temperatures.......................................................................64 4-13. Thermogravimetric analysis of CdS-epoxy films containing 0.75% w/w CdS.......66 5-1. Band edge of CdS-epoxy nanocomposites as a function of CdS concentration.........74 5-2. Percent transmittance of CdS-epoxy films as a function of CdS concentration.........74 5-3. TEM micrographs of CdS-epoxy films in DMF+D2000 on Zeiss EM 10A, 80KV. Bar = 500 nm............................................................................................................75 6-1. One-factor plot of % transmittance as a function of temperature in K.....................86 6-2. One-factor plot of % transmittance as a function of stablizer weight in gm for a total sol weight of 6.0 gm.................................................................................................87 6-3. One-factor plot of % transmittance as a function of CdS in gm for a total sol weight of 6.0 gm..................................................................................................................87 6-4. Band edge as a function of temperature.....................................................................89 6-5. Band edge as a function of stabilizer content (gm in 6 gm of sol)............................90 6-5. Band Edge as a function of CdS content (gm in 6 gm of sol)....................................90 6-7. Band Edge as function of temperature at different stabilizer levels depicting the temperature-stabilizer interaction.............................................................................91 xiv

PAGE 15

6-8. Band Edge as function of temperature at different CdS levels depicting the temperature-CdS interaction....................................................................................91 7-1. UV-visible spectra for different sol synthesis temperatures. A) CdS sols in D2000 and DMF. B) CdS-epoxy films Note the red-shift in the band edge with increasing temperature...............................................................................................................96 7-2. Band edge of CdS sols and films vs. sol synthesis temperature. Also provided is an estimate of the particle size based on the tight binding calculation of CdS90..........97 7-3. Comparison between the temperature dependence of band edge as predicted by the regression model and experimental data................................................................100 7-4. TEM micrographs of CdS-epoxy films in DMF+D2000 on JEOL 3010F, 200KV. samples synthesized at A) 69 C; B, C and D)65 C E, F) 83 C.........................103 8-1. Optical photographs on a Sony DCR-TRV20 camera. Left column shows images taken by covering the lens of the camera with different films while the right column shows the photographs of the films on printed text. A and B) 0.63 mm thick, 21% CdS, prepolymer stabilized, two step (PST), CdS-epoxy film. C and D) 0.85 mm thick, 21% CdS, prepolymer stabilized, in situ (PSI), CdS-epoxy film. E and F) 1.11 mm thick, 0.75% CdS, D2000 stabilized CdS-epoxy film.............................115 6-2. UV-visible absorption spectra of A) 10 m thick, 17%CdS amine stabilized (ASI) CdS epoxy films. B) 10 m thick, 21% prepolymer stabilized (PSI and PST) CdS-epoxy films C) 0.65 and 0.85 mm thick, 21% prepolymer stabilized (PSI and PST) CdS-epoxy films.....................................................................................................117 8-3. X-ray diffraction plot of 0.85 mm thick prepolymer stabilized in situ (PSI) film with 21% CdS.................................................................................................................119 A-1 UV visible spectra of solid film samples from design of experiments in Chapter 5. A to H showing the measured onset of band edge for replicates...............................129 xv

PAGE 16

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SOLID STATE, TRANSPARENT, CADMIUM SULFIDE-POLYMER NANOCOMPOSITES By Nikhil K. Kothurkar August 2004 Chair: Anthony B. Brennan Major Department: Materials Science and Engineering This dissertation deals with the preparation and characterization of solid state, transparent CdS-polymer nanocomposites for potential applications in optical limiting, lenses and photovoltaics. The focus of this research was controlling CdS aggregation to obtain 1 mm thick, films with a transmittance >85%. CdS-polysulfone nanocomposites gave transparent sols using thiol-capping agents however solid films obtained, were translucent to opaque. CdS-epoxy nanocomposites gave 1 mm thick films with >85% transmittance. Oligomeric polyoxypropylene diamine stabilizers were used to shield the inter particle forces and increase the viscosity of the medium to give air stable sols. Films were obtained by curing the CdS sols with an epoxy resin system. The band edge of the nanocomposites showed a shoulder corresponding to the absorption from the CdS particles. A majority of small amorphous and non-stoichiometric (richer in Cd) particles <5 nm were observed in TEM micrographs. Relatively fewer CdS aggregates in the range xvi

PAGE 17

of tens to hundreds of nanometers were also detected. Aggregates were generally more numerous in films as compared to sols. The two-photon absorption coefficient of the CdS particles in the films showed a 30-fold reduction compared to bulk CdS. This was attributed to the lack of crystallinity of particles. The films were stable up to 300C. CdS aggregation in the CdS-epoxy system depends on a number of factors including temperature, stabilizer concentration, stabilizer molar mass and CdS concentration. Effective aggregation control yields transparent films. Factors affecting the band edge and transparency of the films were detected and regression models were fitted to the data. Temperature, stabilizer concentration and CdS concentration had significant effects on the band edge and transparency. Interactions between temperature-stabilizer concentration and temperature-CdS concentration had significant effects on the band edge. Band edge tunability with temperature was demonstrated. An attempt to estimate the particle size from the band edge was made but owing absence of any suitable theoretical models, an accurate estimate was not possible. However a rough estimate was provided. High concentration CdS-epoxy thin and thick films containing up to 21% CdS were synthesized using two different methods. High transparency of about 85% was obtained in some of them. xvii

PAGE 18

CHAPTER 1 INTRODUCTION Introduction Polymer-semiconductor nanocomposites offer the promise of a new generation of hybrid materials with numerous possibilities of applications such as in optical limiting1-3, optical data storage,4,5 optical computing,6 optical displays,7 catalysis,8 photovoltaics,9,10 and gas sensors.11 There are some distinct advantages of using organic-inorganic nanocomposites for device applications, over both the conventional inorganic as well as purely organic (polymeric) systems. These are discussed in detail by Godovsky.12 In general, organic-inorganic nanocomposites are low cost in terms of both materials and processing techniques compared to the clean room techniques and expensive materials of the conventional semiconductor processing industry. This enables easy production of large area devices like LEDs or solar cells from organic-inorganic nanocomposites13. There is considerable latitude with respect to morphologies and geometries that are often unattainable in conventional techniques. It is possible to achieve nanoscale devices with relatively low effort using self-assembly techniques14 and optical lenses with temperature stability of refractive index.15 There is enormous potential to produce a breadth of device configurations from the sheer number of different organic and inorganic materials available. Within each combination, variations in processing temperatures, concentrations or similar parameters, a significant change in properties can be achieved. One of the most important advantages of inorganic-organic nanoparticulate composites in particular, is a dramatic increase in 1

PAGE 19

2 long-term stability over purely polymer devices. P-N nanojunctions can be made with nanocomposites that are not possible with all polymer systems. The nanocomposites allow fractalization of the space charge layer and interpenetrating networks of nanoparticles and other peculiar morphologies that are valuable in device applications. Highly doped nanoparticles allow electric fields as high as 108 V/cm. Nanosize semiconductor particles immobilized in a polymer matrix show quantum size effects.16,17 Processing parameters such as temperature, concentration, stoichiometry, additives, etc, can be used to vary intermediate properties such as particle size,18,19 size distribution, particle morphology,20 interfaces21 and phase miscibility.22 The intermediate properties in turn govern the end properties such as band gap, transparency, nonlinear coefficients, and operating wavelength range. Band edge tailoring is one of the most significant. This provides an easy means to control the characteristics of devices such as photodiode emission wavelength12,23,24 or absorption range of a solar cell.12 The control achieved is much broader than achieved by stoichiometric control in conventional synthesis of chalcogenide diodes. The band edge tailoring and inorganic content in the composite also allows formation of non-scattering ‘single crystals’ of tailor made refractive and absorptive properties.12 The nanoparticles provide an extremely high interface area, which is of great use in sensor11,25,26 and catalytic27 applications. Applications Two specific applications namely, optical limiting and nanocomposite solar cells, will be discussed for which this research would be most relevant.

PAGE 20

3 Optical Limiting An optical limiter is a device that is used to protect optical equipment like sensors from damaging high intensity radiation28. It is highly transparent at low intensities and when the intensity reaches above a certain threshold, they limit the light falling on the sensor28-30 by scattering, absorption or refraction or combination of these. Nonlinear optical behavior of materials is used to achieve optical limiting. Aperture Lens Sensor Infrared beam Optical Limter Figure 1-1. Schematic diagram of an optical limiter used for optical sensor protection Solid-state devices are preferable for this application although most efforts in optical limiting have focused on liquid systems2. Semiconductor-polymer nanocomposites have been used as optical limiters.31-36 The semiconductor nanoparticles on account of their good nonlinear optical properties are responsible for the working of the device. The nonlinear optical effects are typically very low in magnitude and about 1 mm path length37,38 and a high concentration of the semiconductor particles are required for sufficiently high optical effects. However a high loading of the nanoparticles often leads to poor transparency as a result of aggregation of the nanoparticles, within the polymeric matrix especially in thick films. So a nanocomposite used for optical limiting needs to be optimized for high transparency at a high CdS loading. Additionally a control over the working range by changing the band edge may also be required.30 The band edge also has an effect on the nonlinear optical properties.39

PAGE 21

4 Semiconductor-Polymer Nanocomposite Solar Cells Polymer-based solar cells have the capability of being used to make cheap large area panels, flexible panels or futuristic solar suits.9,12,40,41 The only downside is substantially low efficiency (about 1 to 2%) compared to commercial solar cells (10 to 15%). Cadmium chalcogenides such as CdS, CdSe, and CdTe nanoparticles in polymers have been used to make solar cells.9 Numerous efforts are being taken to improve the efficiency of the cells. In general a higher density of the nanoparticles is advantageous. However particles at higher concentrations tend to aggregate. Nanorods (7x80 nm), which enable better carrier transport and a lower percolation limit have been used instead of nanospheres. Sunlight Figure 1-2. Schematic diagram of CdSe nanorod-polymer solar cell. However there are difficulties in stabilization of the rods with increasing aspect ratios. The rods tend to get bunched up leading to a loss in efficiency. Thus control of aggregation of nanoparticles in the matrix is crucial to the production of efficient solar cells. Control of the band edge of the particles is also very important factor controlling the efficiency of the solar cell12. The size of the particles governs the absorption

PAGE 22

5 characteristics. This phenomenon is called the quantum size effect.42-45 Thus band edge control can be achieved by controlling particle size. General Requirements of Nanocomposites for Device Applications The general requirements of semiconductor-polymer nanocomposites for device applications12 can be summarized as the following: 1. Transparency. Generally a transmittance >85% is required for devices operating in the near infrared region. Effective dispersion of the nanoparticles in the matrix without large-scale aggregation or precipitation leads to high transparency in the films. Aggregation leads to scattering of light from large aggregates leading to opacity. 2. Tunability. The ability to tune the band edge and particle size is desirable for several applications. Control over particle size enables one to control the band edge of semiconductor nanoparticles.42,44 3. Monodispersity. Some applications such as nanoelectronics,5,46,47 optoelectronics,6 require a narrow size distribution—ideally monodisperse particles. 4. Particle morphological control. Applications such as nanoelctronics, photovoltaics,9 may require control of morphology of the particles; mainly aspect ratio—from nearly spherical to rods and wires. 5. Order. Ordering of particles into 2D or 3D arrays and aligning or orienting particles with higher aspect ratios are important requirements from the point of view of nanoelectronics,5,46,47 optoelectronics.6 Self assembly techniques are used to achieve order in nanoparticles.14,48-55 6. Processability. The material has to be processable into bulk films or thin films, coatings, and fibers to be of any practical use. 7. High temperature resistance. Applications such as optical limiting require that the materials have good thermal resistance.2,56,57 These factors have to be kept in mind while evaluating any nanocomposite. It is difficult to find a system that will meet all the above requirements. Some of these requirements are in fact the goals of rapidly growing fields. Practically one has to make compromises based on the application. For example, ordering of particles may not be of much use for a nanocomposite used for optical limiting, but excellent aggregation control

PAGE 23

6 is required in order to obtain high transparency at low intensities. High concentration of CdS is another requirement important for optical limiting as the nonlinear effects are proportional to the concentration of CdS in the nanocomposite. However this requirement has its limitations in that higher the concentration of CdS in the films the lower is stability of the particles and lower is the transparency of the films. The ability to produce thick films (~ 1 mm) is essential so as to get a high nonlinear signal.37,38 The matrix polymer needs to have a high temperature resistance to reduce laser damage. In general polymers have a lower damage threshold as compared to inorganic materials such as glasses. However polymers such as epoxies, polysulfones, and polyimides are transparent and have relatively good stability to temperatures as high as 300C. They will be preferred for this application. Specific Aims Following are the specific aims of this dissertation: 1. To synthesize and characterize CdS-polysulfone nanocomposite films with transparency > 85% 2. To synthesize and characterize CdS-epoxy nanocomposite films with thickness on the order of 1 mm and % transmittance >85% 3. To evaluate the effects of factors affecting the band edge and % transmittance of CdS epoxy nanocomposites using a design of experiments and demonstrate the tunability of optical band edge from 460 to 500 nm by varying the temperature of synthesis from 65 to 85C 4. To synthesize and characterize CdS epoxy films with CdS > 20% w/w and transparency > 85%

PAGE 24

CHAPTER 2 BACKGROUND Review of Various Efforts to Synthesize Metal ChalcogenidePolymer Nanocomposites A number of methods have been tried to incorporate inorganic particles into polymeric matrices. Liu et al.51 and Kickelbick58 independently give a detailed reviews of the various nanofabrication techniques adopted to incorporate nanoparticles into polymer matrices. A variety of systems using low MW surfactants, polymeric surfactants, block copolymer based micellar templates, cross-linkable copolymers, dendrimers,59,60 biopolymer templates, electric field induced synthesis, membrane based nanofabrication, etc have been discussed. Jelinski61 and Fendler62 have separately reviewed the various bio-inspired routes to make nanocomposites. These involve the synthesis and assembly of the nanoparticle using biological materials, such as amino acids, proteins, lipids, and nucleic acids. The biological materials, due to their diversity and specificity can be polymerized or otherwise organized into an enormous variety of structures. They exhibit hierarchical order and also have functional groups that are capable of coordinating or covalently linking with metal ions or particle surfaces. Soten and Ozin53 have reviewed the different means to self assemble quantum dots in polymer or organic matrices. Esteves and Trindade63 have classified the chemical procedures used to synthesize quantum dots or semiconductor nanoparticles into three 7

PAGE 25

8 broad categories: controlled precipitation, cluster buildup and molecular precursor methods. The controlled precipitation methods are those in which solutions containing metal cations and chalcogenide anions are mixed together and the precipitation of the nanoparticles is controlled by the presence of polymers,64 surfactants,65-74 Lewis bases18 or an electrical double layer. The cluster buildup approach uses inorganic clusters of a definite chemical composition and makes them assemble or phase transform to the final nanostructure, leading to the formation of single crystals or superlattices of Cd containing units.75-77 These units often contain thiol or similar groups and the metal chalcogenide particles thus produced22 often tend to be non-stoichiometric. The molecular precursor technique involves application of a thermal or chemical treatment to a molecular species containing the metal complex in a coordinating solvent, to yield the nanoparticles. This includes the solvothermal20,78 and the trioctylphosphine oxide-capped nanoparticles79-81 techniques. Most synthetic techniques will fall in at least one of the three above-mentioned categories. Nanoparticles of metal chalcogenides such as CdS have been used for a variety of applications like solar cells, optical limiting, catalysis and biological applications. Some of the reasons for its wide use are the ease and flexibility in synthesis and the relative simplicity of chemistry needed to modify the surface or otherwise stabilize the particles. This allows for numerous ways to immobilize the nanoparticles into a polymeric matrix in ordered arrays or random dispersions.

PAGE 26

9 In addition, the optical properties of CdS nanoparticles such as quantum size effects,42-44 photovoltaic effect,9 photoluminescence82 and the nonlinear optical effects such as two photon absorption,83 nonlinear refraction,84-86 and exciton absorption bleaching87 have been studied in great detail. This makes the system ideal for the development of nanocomposites for various optoelectronic devices. A number of systems are unable to stabilize the CdS particles in solution for a considerable period of time. Even if it is possible to stabilize them as sols, incorporating them into polymers and making transparent materials in the solid state is very difficult. Most efforts of obtaining stable sols of CdS, involve confining the particles in micelles or at interfaces or compatibilizing the surface of the particles by using capping agents such as thiols or similar molecules. However once synthesized, phase segregation is common in that the particles. Thiols are susceptible to oxidation forming sulfide linkages. This decreases their ability to stabilize the nanoparticles. Phase segregation leads to opacity that is detrimental to any optic or optoelectronic application. Hence these sols are not stable in air and tend to precipitate unless maintained under an inert atmosphere. This leads to inconvenience in processing and additional equipment and running costs. Solvent cast films from such sols tend to be poor in transparency due to phase segregation. Incorporating sols that are prone to precipitation, with polymers and making films thereof, compromises the optical properties. One of the ways of minimizing this problem is to make thin films of these nanocomposites. Spin casting semiconductorpolymer sols, can be very fast giving very little time for the nanoparticle clusters to

PAGE 27

10 aggregate or grow. Thus transparency can be improved by limiting the growth of the nanoparticles by making thin films. However the same systems fail if the film thickness is higher. Drying takes longer and that gives enough time for the aggregates to grow to dimensions sufficient to scatter light. Solid transparent semiconductor-polymer nanocomposites are possible only if the system is stable in the solution phase (preferably air stable) as well as during solidification (crosslinking or drying). Quantum Size Effects in the Absorption Spectra of Semiconductor Nanoparticles Semiconductor particles show a significant change in certain fundamental electronic properties such as the free electron and exciton absorption binding energy with reduction in particle size as compared to properties in the bulk. This phenomenon is called the quantum size effect. In general the quantum effect causes a blue shift in the band edge or an increase in the binding energy compared to the bulk. It arises when the dimensions of the particles approach the dimensions of the free electronic or excitonic Bohr radius. The Wannier exciton Bohr radius for CdS is about 2.8 nm.39 At these dimensions, the confinement effects of the electrons or the excitons become significant and must be taken into account. UV-visible spectra of unordered CdS nanoparticles show the excitonic transition as a bump to the right of the band edge at room temperature. The breadth of the transition in real crystals is a result of the inhomogeneous broadening due to structural defects and the polydispersity of the particles45. Only in case of monodisperse perfect crystals, the homogeneous line shape would be Lorentian39. In the case of relatively large nanoparticles, the line shape is similar to the absorption spectrum of bulk semiconductors.

PAGE 28

11 The band edge region may show discrete lines corresponding to different excited states. The CdS excitonic peak is visible in the absorption spectrum as a peak to the right (at lower energy) of the absorption band edge of CdS. The exciton peak has a relatively low energy and is seen only at low temperatures while the peak is not distinguishable from the absorption band edge at room temperature in bulk semiconductors such as CdS.45,39 Reducing the particle size changes the absorption spectrum of the semiconductor. The exciton peak increases in breadth and ultimately becomes indistinguishable from the absorption edge when the particle diameter equals its Bohr exciton diameter. The band edge region becomes diffuse and develops a broadband edge, which is progressively more blue-shifted (towards lower wavelengths) as the particle size is reduced. In general this characteristic is maintained irrespective of temperature. One can also get a qualitative idea of the particle size of the sols based on visual observation. Color in CdS sols is in fact subtractive color due to the absorbance edge lying in the visible region. Semiconductors act as low-pass filters. The fraction of white light with energy higher than the absorption edge will be absorbed while the rest transmitted.39 As seen above, the position of the exciton absorbance peak is a function of particle size. As particle size reduces there is a blue shift in this absorbance. If the particle size is very small (approximately 1 nm in case of CdS) such that the absorption is at a wavelength lower than the visible threshold ~ 380nm, none of the visible light is absorbed by it. Therefore, the sol would appear colorless and the precipitate would appear white. As the particle size increases the exciton absorbance moves to longer

PAGE 29

12 wavelengths into the violet and blue regions of the visible spectrum. When this happens, the violet and blue light is cut off from coming to our eyes due to absorption. Thus the sol takes on a light yellow hue. As the CdS particle size increases further, the exciton absorbance moves further into in the visible and the sol becomes progressively deeper yellow. A deep yellow indicates that the exciton absorption is in the green.82 When the color is deep yellowish orange, the particles have the same band edge as bulk CdS and may be regarded as bulk. In this case there is no quantum confinement. Sometimes a greenish tinge is seen to the colloids, especially those with small to intermediate particle sizes. This is due to the photoluminescence of the CdS particles in the green,88 which also depends upon particle size. Different models are proposed to relate the blue shift in the excitonic absorption with the particle sizes.42,44,89 There are two different approaches that may used to treat a nanoparticle as a part of a large crystal and the other is to consider it as a large molecule. The former proceeds by considering the quasiparticles (namely de Broglie wavelength and exciton Bohr radius) of an infinite crystal and then include the boundary conditions for finite boundaries. One of the simplest models of this type is called the effective mass approximation and is as given below. This model predicts the energy shift as a function of particle radius. The shift in the binding energy44 is given by the following equation: E = (22)/(2R2) (1.786 e2)/(R)-(0.248 e4)/(222) (2.1) Where, E = exciton binding energy shift (eV); = h/2; h = Planck’s constant 4.1357 x 10-15 eV seconds; = 3.14; R = particle radius (A); = reduced effective mass of exciton (gm); = dielectric constant of material.

PAGE 30

13 The above equation is obtained by solving the Schrodinger’s equation for spherical boundary conditions i.e. taking the confinement boundaries in all three dimensions. In case of nanorods or nanowires, where the length dimension is much larger than the confinement Bohr radius, similar equations may be obtained by solving the Schrodinger’s equation for two dimensions. The other approach is the quantum chemical approach in which the nanoclusters are essentially considered as large molecules. This approach involves studying the development of crystal-like properties from the atomic and molecular level. The atoms in the clusters may be looked at as having localized sets of bond orbitals. As the clusters grow in size, the highest occupied molecular orbital (HOMO) becomes top of the valence band while the lowest unoccupied molecular orbital (LUMO) becomes the bottom of the conduction band. In order to deal with the properties of such a cluster, one has to begin with the Schrodinger’s equation containing a Hamiltonian which includes the kinetic energies of all electrons and nuclei and potential energies of attractive interaction of electron-nucleus, and repulsive nucleus-nucleus interaction. It is a problem of N nuclei and n electrons. It can be solved analytically only for N = n = 1 in case of a hydrogen atom. The quantum chemical approaches are essentially numerical and based on several assumptions and approximations. For more complicated systems, in order to avoid the difficulties of ab initio calculations a number of semi-empirical techniques have been developed. One such technique is the Hueckel technique.45 The electronic structure of CdnSm, CdnSem has been computed using this technique. Another version of the Hueckel technique, called the tight

PAGE 31

14 binding approach has been successfully used by Lippens and Lannoo89 to determine the structure and properties of CdS and ZnS clusters. The quantum chemical approaches such as the tight binding calculation. This method has several advantages over the effective mass approximation and tends to fit experimental data more closely. The effective mass approximation, although the first and the simplest technique of its kind, tends to overestimate the particle size at sizes lower than a couple of nanometers nm. The tight binding model is preferable in this region. Sapra and Sarma90 have calculated the electronic structure of CdS using the tight binding calculation (TB) using sp3d5 model. Figure 2-1 shows the curve of the best fit for the data points calculated using the tight binding model. This fitted curve can be used as an empirical model to estimate CdS particle size from its band edge. The empirical equation mentioned is as follows: Eg = a1e( d/b1) + a2e (d/b2) (2.2) Where, a1 = 2.83 b1 = 8.22 a2 = 1.96 b2 = 18.07 d = diameter in nm Figure 2-1 also shows the plot of the band edge as a function of particle radius as per the tight binding calculation as well as the effective mass approximation for comparison.

PAGE 32

15 250300350400450500550024681012Particle Radius nmBand Edge nm22.533.544.55Band Energy eV TB sp3d5-Band Edge EMA-Band Edge TB sp3d5-Band Energy Figure 2-1. Band edge vs. particle radius of CdS using tight binding calculation with effective mass approximation for comparison.39 Thus the particle size may be calculated from the onset of the excitonic absorption. The particle size discussed here is the dimension in which the carriers are confined. The particles will seldom retain that dimension without aggregation. In practice, aggregates of the above-mentioned particles will be formed subsequently under the influence of secondary bonding forces. If the aggregation is not controlled there will be precipitation of the particles. The exciton absorption onset however will remain unaffected in case of CdS particles aggregating in complexing solvents. Thus the size of these aggregates cannot be deduced by measuring the excitonic band edge. TEM on the other hand will reveal the aggregate size but resolving the individual nanoparticles in the aggregates may be difficult.

PAGE 33

16 Summary The current status of inorganic-polymer nanocomposites was reviewed. Quantum size effects in semiconductor particles were discussed. This enables one to estimate the particle size based on the band edge of the material. With this background, we proceed to the actual synthesis and characterization of CdS-polymer nanocomposite films.

PAGE 34

CHAPTER 3 CADMIUM SULFIDE-POLYSULFONE NANOCOMPOSITES Introduction Semiconductor-polymer nanocomposites such as CdS-polymer nanocomposites in particular, have been synthesized mostly as thin films,8,65,91-93 in solution59,94 or as precipitates17,95 but not as transparent bulk films. This chapter will deal with the synthesis and properties of CdS polysulfone nanocomposites. The main obstacle to obtaining transparent bulk polymeric films was the tendency of small inorganic particles to aggregate into large agglomerates. When the size of the particles approaches the wavelength of light they tend to scatter light. This leads to a reduced transmittance in the films. Stabilization of the particles is essential to producing transparent nanocomposite films. Stabilization can be achieved by a number of techniques including functionalization of the matrix as well as the particles. The functionalization of the matrix involves introducing a reactive functional group in the polymer backbone or as side groups. The functionalization of the particle requires terminating the particle surface with groups,8,18,22,96-98 which preferentially react covalently with the corresponding groups on the polymer or interact with the matrix through secondary bonding forces. This prevents the aggregation of the particles in the polymer matrix and also improves the solubility in organic solvents. The particles and the polymer can be co-dissolved in an appropriate solvent by which films of the composite material may be cast. For example, functionalization CdS 17

PAGE 35

18 particles normally involves introduction of thiols in the synthesis of the particles. The Cd:S ratio is usually maintained close to 1. Part of the sulfur required may be supplied in the form of the thiol. This ensures a coating of the capping agent on the particles. The capping agent may further have functional groups that can react with the corresponding groups on the polymer. Several thiols have been used including thiophenol,8,71 aminoethanethiol18 and 4-flourothiophenol.22 These capping agents can give soluble powders of CdS. The capping agents improve the dispersibility of the particles in organic solvents in which pure CdS is normally not soluble. Organic solvents such as acetone, DMF, THF, have been used as dispersion media for capped CdS particles. A suitable solvent is chosen which also dissolves the polymer. A dispersion of the capped CdS particles may be taken and mixed with a solution of the polymer in the same solvent. The two can be mixed together and then films may be cast out in molds and dried to obtain films. Nanocomposites of semiconductors in several polymers such as polystyrene,99 polyurethane,93 Nafion,87 polyacrylamide,95 polythiophene10 have been made. However all these polymers have a low temperature resistance or are not transparent. For optical limiting in solid films high temperature resistance is essential. Failure due to thermal degradation in thick polymer films subjected to high intensity laser pulses is greater than in liquid state optical limiters such as CS2 or dispersions of carbon particles100,101 including nanotubes, fullerenes. This is because in liquids even though local heating may destroy or vaporize liquid or particles on the dimensions of the beam diameter, mixing and redistribution of the particles and liquid makes the limiter self-repairing.

PAGE 36

19 Criteria for Selection of Materials Inorganic phase. The most important criterion for the choice of the inorganic phase is the end-application, which is optical limiting. Optical limiting in the region of the transparency window is dictated by the nonlinear index of refraction (NL index). Materials with high NL index such as GaAs, cadmium chalcogenides, titanium dioxide may be considered.102 Other criteria are the compatibility and stability with polymers and organic solvents. Most semiconductor particles such as metal chalcogenides or oxides are incompatible with polymers. The particles used must therefore be amenable to functionalization to achieve good compatibility between the polymer matrix and the particles. Cadmium sulfide (CdS) nanoparticles are among the most studied nanoparticles in terms of their optical and nonlinear optical properties. They have a high NL index of refraction102 and their chemistry is well known. CdS nanoparticles have been successfully functionalized using a number of thiol based capping agents as mentioned before. These can give soluble powders of CdS. Thus CdS is an ideal candidate for making transparent thick nanocomposite films. Organic phase. The polymer used for the nanocomposite has to be soluble in common solvents and amenable to chemical functionalization. It should be easily processable into transparent films by solvent casting and should have a high temperature resistance. Polysulfones are amorphous polymers with a high glass transition temperature [heat deflection temperature of poly(aryl ether sulfone) =174C]. They are soluble in a variety of organic solvents such as toluene, THF, DMF and others.103 It can be readily processed into transparent films by solvent casting.

PAGE 37

20 Polysulfones have been functionalized successfully with polar groups such as sulfonic acid or its corresponding acid salts.103 Poly(aryl ether sulfone) was chosen as a suitable candidate since it satisfies all of the above requirements. This chapter details some of the important procedures that were used to make CdS-polysulfone hybrids. Some of the syntheses used poly(aryl ether sulfone) Udel, while the others used its sulfonated, free-acid form. Materials and Methods Sulfonation of Polysulfone Following the procedure of Noshay and Robeson,103 10 to 12% sulfonation of polysulfone (100% sulfonation = 1 -SO3H group per repeat unit) was carried out using a sulfonating complex of sulfur trioxide and triethyl phosphate. The polysulfone, sulfur trioxide and triethyl phosphate mole ratio of 1:1:1, was taken. This ratio yields a 10 to 12% sulfonation in the polymer.103 Table 3-1. Materials used for the synthesis of sulfonated polysulfone. Chemical name F unction C hemical formula Source A bbreviation P oly(aryl ether sulfone) P olyme r A moco P SF Sulfur trioxide Sulfonating agent SO3 F isher scientific SO3 T riethyl phosphate C omplexing agent (C2H5O)3PO F isher scientific T EP 1, 2 Dichloro ethane Solvent C 2H4Cl2 F isher scientific N /A M ethanol N on-solvent for precipitation C H3OH F isher scientific M eOH Procedure: The sulfonating complex was prepared by mixing 7 gm SO3 very slowly and with continuous stirring, to a solution containing 32 gm triethyl phosphate in 53.6 ml dichloroethane. The solution was maintained below 25C by dipping the flask in

PAGE 38

21 an ice bath. A polymer solution was prepared by dissolving 37.51 gm polysulfone in 250 ml dichloroethane. The polymer solution and the sulfonating complex were added at similar rates to a flask containing 275 ml of dichloroethane with vigorous stirring in a span of 10 min. The temperature of the mixture rose from 25C to 34C. Stirring was continued for another 30 minutes. The polymer was precipitated by adding excess of methanol. The slurry was filtered and the filtrate was dried under vacuum at 60C. This dried filtrate is sulfonated polysulfone.103 The above sulfonated polymer was stored in closed polyethylene containers and used whenever needed. Synthesis of CdS-PSF Nanocomposites The essential step involved in the synthesis of these composites is the capping of the CdS nanoparticles. A variety of capping agents were used to make functionalized (capped) CdS clusters. These have two functional groups; one of them being the thiol group and the other one different in each case. Table 3-2. Capping agents for CdS Capping agent Chemical formula Terminal functional group Stability in air Thiophenol C6H5SH Phenyl No 2-Amino thiophenol HSC6H4NH2 Aminophenyl No 4-Amino thiophenol HSC6H4NH2 Aminophenyl No Aminoethane thiol hydrochloride HSC2H4NH2.HCl Aminoethyl No 4-Fluro thiophenol HSC6H4F Fluorophenyl No In all cases it was assumed that the thiol group attaches covalently to the surface of a growing CdS particle and arrests further growth. The other functional group remains pendent from the surface of the particle and interacts with the immediate environment of the particle consisting of solvent, polymer, or other particles. Depending on the nature of

PAGE 39

22 the nature of the pendent functional group, it is possible to vary the solubility of the particles in various solvents and the compatibility with different polymer matrices. They may also be made to react with complementary functional groups placed on the matrix polymer. Depending upon the different capping agents and the PSF (sulfonated or unsulfonated) a number of different nanocomposites were synthesized. Following are procedures of making some of the more successful nanocomposites. Phenolated CdS-PSF nanocomposites—Method I (PhCdS-PSF-I) Table 3-3. Materials used for the synthesis of phenolated CdS-PSF—Method I. Chemical name Function Chemical formula Source Abbreviation Cadmium nitrate tetrahydrate S ource of cadmium Cd(NO3)2H2O Fisher scientific Cd-nitrate Sodium Sulfide Source of sulfur Na2S9H2O Fisher scientific N/A Water Solvent H2O Nanopure N/A Tetrahydrofuran Solvent C4H8O Fisher scientific THF Poly(aryl ether sulfone) Polymer Amoco PSF Thiophenol Capping agent C6H5SH Fisher scientific PhSH Procedure: A Cd++ solution was prepared by adding 0.0161 gm Cd(NO3)2H20 to 25ml THF. An S-solution was prepared by adding 0.0106 gm Na2S9H2O to 24 ml THF and 1 ml H2O. 20ml THF was taken in an Erlenmeyer flask. Thiophenol, 0.003 gm, was added to it and stirred well. Then 2.5 ml of each of the Cd++ and S-solutions were added to it drop-wise with constant stirring over a period of 10 min. A transparent yellow sol was obtained. Final volume of this sol was 25 ml and the concentrations of Cd++ and S-added, work out to 2x10-4 M each.

PAGE 40

23 Stirring was continued and the polymer pellets (5.037 gm) were added to this sol over a period of 20 min and stirring continued till a homogeneous solution was obtained. This solution was poured out into glass molds and left to dry slowly in the fume hood by placing in a cardboard box on it so as to maintain a high partial pressure of THF. The films took over 4 days to dry to the extent that they were not tacky and could be detached from the mold. Phenolated CdS-PSF nanocomposites—Method II (PhCdS-PSF-II) This method was different from the earlier in that the capped CdS particles were made separately, precipitated and then dried; after Herron et al.44,104 The particles were redispersed in an appropriate solvent and incorporated in the polymer. Table 3-4. Materials used for the synthesis of phenolated CdS-PSF—Method II. Chemical name Function Chemical formula Source Abbreviation Cadmium acetate, dihydrate Source of cadmium Cd(CH3COO)2H2O Fisher scientific Cd-acetate Sodium Sulfide Source of sulfur Na2S 9H2O Fisher scientific N/A Acetonitrile Solvent (CH3CO)2CN Fisher scientific AcN Methanol Solvent CH3OH Fisher scientific MeOH Water Solvent H2O Nanopure N/A Tetrahydrofuran Solvent for polymer C4H8O Fisher scientific THF Poly(aryl ether sulfone) Polymer Amoco PSF Thiophenol Capping agent C6H5SH Fisher scientific PhSH Procedure: A Cd solution was prepared by taking 0.1 M cadmium acetate in an 80:20 w/w mixture of methanol and acetonitrile. A Na2S solution was prepared by taking 0.1 M sodium sulfide in a 50:50 w/w mixture of water and methanol. A 0.2 M PhSH

PAGE 41

24 solution was prepared in acetonitrile. Fifty ml of Na2S solution was mixed with 50 ml of PhSH solution and stirred continuously. A 100 ml of Cd solution was added to this mixture slowly with constant stirring. Stirring was continued for 15 min. A precipitate was formed and was separated by filtration. The filtrate was vacuum dried, to give dry powder. Capped CdS clusters from this powder were extracted with acetonitrile leaving behind sodium acetate. The capped CdS clusters were obtained by evaporating off the acetonitrile and drying under vacuum. The capped CdS clusters were soluble in THF and a solution of the clusters was prepared in it. A solution containing 0.25 gm of CdS clusters in 25 ml THF was taken in an Erlenmeyer flask and stirring was started. Five gm of polymer were added to it in small parts over a period of 20 minutes with stirring till homogeneous solution is obtained. The solution was poured out in glass molds. The samples in the molds were left to dry gradually in chamber with high partial pressure of THF. When the films were free of tackiness, they were detached from the molds and the residual solvent was evaporated by heating in oven at 50C. Amine-capped CdS-SPSF nanocomposites—Method I (AmCdS-SPSF-I) The cadmium sulfide particles can be modified such that they have terminal reactive groups like amino. These groups may enable us to achieve chemical bonding between the polymer and the particles. Procedure: A Cd solution was prepared by taking 0.1 M cadmium acetate in an 80:20 w/w mixture of methanol and acetonitrile. A Na2S solution was prepared by taking 0.1 M sodium sulfide in a 50:50 w/w mixture of water and methanol. Aminoethane thiol solution (0.2 M) was prepared in acetonitrile. Fifty ml of Na2S was mixed with 50 ml of the aminoethane thiol solution and stirred continuously. A 100 ml of the Cd solution was added to this mixture slowly with constant stirring. Stirring was continued for 15 min. A

PAGE 42

25 precipitate was formed and was separated by filtration. The filtrate was vacuum dried, to give dry powder. This powder was insoluble in a variety of organic solvents and films could not be made by co-dissolving with polymer. Table 3-5. Materials used for the synthesis of amine capped CdS-SPSF—Method I. Chemical name Function Chemical formula Source Abbreviation Cadmium nitrate, tetrahydrate Source of cadmium Cd(NO3)2H2O Fisher scientific Cd-nitrate Sodium Sulfide Source of sulfur Na2S 9H2O Fisher scientific N/A Acetonitrile Solvent (CH3CO)2N Fisher scientific AcN Methanol Solvent CH3OH Fisher scientific MeOH Water Solvent H2O Nanopure N/A Tetrahydrofuran Solvent for polymer C4H8O Fisher scientific THF Sulfonated poly(aryl ether sulfone) Polymer N/A Amoco PSF 2-Aminoethane thiol hydrochloride Capping agent H2NC2H4SH Fisher scientific Aminoethane thiol Amine-capped CdS-SPSF nanocomposites—Method II (AmCdS-SPSF-II) This procedure used 4-aminothiophenol as the capping agent and was carried out while bubbling nitrogen through the reaction mixture so as to avoid contact with air. Procedure: A Cd++ solution was prepared by adding 1.31 gm Cd(NO3)2.4H20 to 50 ml DMF in a reaction vessel with a thermometer and a water-cooled condenser. The top end of the condenser was covered with parafilm in order minimize solvent losses. A needle carrying nitrogen was made to pass through it to bubble nitrogen in the reaction mixture. 4.83 gm of SPSF was added to it and stirred using a magnetic stirrer till it

PAGE 43

26 completely dissolved. The temperature of the vessel was maintained at 85 to 90C. by heating in an oil bath. A sulfide solution containing 0.5 ml aminothiophenol and 0.32 gm thiourea in 10 ml of de-aerated DMF was prepared in an argon filled glove box. The mole ratio of S:SH was kept at 0.5. The sulfide solution was added to drop wise with stirring to the reaction vessel containing the cadmium nitrate and polymer solution at 85-90C. The reaction vessel was stoppered without allowing contact with air and was moved to the glove box filled with argon. The reaction mixture was poured out in molds in the glove box. Thick films on the order of 1 mm, thin films on the order of 0.1 mm were cast. Some were moved outside the glove box while some were allowed to stay in the glove box. The films in molds that were left in the glove box took a few days to dry to the extent that they were not tacky and could be detached from the mold. Drying was accelerated by allowing flow of argon over the films and by periodically putting them under vacuum (20 mm Hg). If evacuated to a lower pressure there was tendency of bubble entrapment in the thick films. On drying till they were no more tacky, they were put in a vacuum oven at 90C to remove entrapped bubbles. Table 3-6. Materials used for the synthesis of amine capped CdS-PSF—Method II Chemical name Function Chemical formula Source Abbreviation Cadmium nitrate, tetrahydrate Source of cadmium Cd(NO3)2H2O Fisher scientific Cd-nitrate Thiourea Source of sulfur H2NCSNH2 Fisher scientific thiourea N, N, dimethyl formamide Solvent (CH3)2NCOH Fisher scientific DMF Sulfonated poly(aryl ether sulfone) Polymer N/A Synthesized SPSF 4-Amino thiophenol Capping agent H2NC6H5SH Fisher scientific Amino thiophenol

PAGE 44

27 Fluorothiophenol-capped CdS SPSF nanocomposites (FPhCdS-SPSF) This method followed the procedure of Gacoin et al.22,105,106 to synthesize the CdS particles, which are then incorporated into the polymer. Table 3-7. Materials used for the synthesis of fluorothiophenol-capped CdS-SPSF Chemical name Function Chemical formula Source Abbreviation Cadmium nitrate, tetrahydrate Source of cadmium Cd(NO3)2H2O Fisher scientific Cd-nitrate Sodium sulfide Source of sulfur Na2S9H2O Fisher scientific N/A Acetone Solvent CH3COCH3 Fisher scientific N/A Water Solvent H2O Fisher scientific N/A Triethyl amine Complexing agent (C2H5)3N Fisher scientific N/A N, N, dimethyl formamide Solvent (CH3)2NCOH Fisher scientific DMF Sulfonated poly(aryl ether sulfone) Polymer N/A Synthesized SPSF 4-Fluoro thiophenol Capping agent H2NC6H5SH Fisher scientific Fluoro thiophenol The CdS particles were first prepared in an inert atmosphere in a glove box in an acetone-water solvent system using fluorothiophenol as capping agent and triethyl amine as complexing agent. The formed particles were precipitated by adding excess of water and the precipitate recovered, dried and stored. Required quantity of the stored powder could be dissolved in an organic solvent such as acetone to give a clear sol. Procedure: A Cd solution containing 0.6062 gm cadmium nitrate Cd(NO3)2. 4H2O in 40 ml acetone and 10 ml distilled water was prepared. The solution was de-aerated by bubbling nitrogen and was taken to glove box with argon atmosphere. A Na2S solution containing 0.3049 gm sodium sulfide, in 20 ml acetone and 15 ml distilled water

PAGE 45

28 was prepared. Solution was de-aerated by bubbling nitrogen and was taken to the glove box with argon atmosphere. 0.64ml of stabilizer 4-fluorothiophenol and 0.538 ml triethylamine as complexing agent were added to the sodium sulfide solution. These two solutions were not soluble and the mixture appeared cloudy. Additional 15ml de-aerated acetone was added to it to dissolve it and a clear solution was obtained. This solution was added to the Cd solution slowly with constant stirring. Mole ratio of Cd++/S-was 2/1. Solution turned to opaque yellow-white color. Stirring was continued for 15 min. The CdS particles formed were precipitated completely by adding 20 ml water. The precipitate was centrifuged and separated and left to dry under vacuum after repeatedly purging chamber with argon. A dry yellow powder of the capped CdS was obtained, which had turned brown in some places. A sol of 0.37 gm of this powder in 10 ml DMF was prepared. A small fraction of the powder that had turned brown did not dissolve. It was separated by centrifugation. A solution of 5 gm of sulfonated polysulfone was made in 20 ml DMF. 10 ml of the CdS sol was added to the polymer solution and stirred. It formed a homogeneous transparent yellow solution. A few drops of the polymer-CdS solution were spread on a glass slide to make thin films. Some of it was poured into molds in order to make thick films. The films were dried under vacuum after purging with argon repeatedly. Uncapped CdS-PSF nanocomposites (UCdS-PSF) Another strategy to obtain transparent nanocomposite films was to mix solutions of cadmium acetate, thiourea and polymer in the same solvent and cast out films. Then when the films are partially dry, they could be heated to above their glass transition temperature. The heating would decompose the thiourea leading to the formation of CdS.

PAGE 46

29 However due to the high viscosity associated with the polymer above its glass transition temperature the aggregation of CdS would be severely restricted. The heating would also evaporate the remaining solvent. Table 3-8. Materials used for the synthesis of uncapped CdS-PSF nanocomposites Chemical name Function Chemical formula Source Abbreviation Cadmium acetate, dihydrate Source of cadmium Cd(CH3COO)2H2O Fisher scientific Cd ac Thiourea Source of sulfur H2NCSNH2 Fisher scientific Thiourea N, N, dimethyl formamide Solvent (CH3)2NCOH Fisher scientific DMF Poly(aryl ether sulfone) Polymer Amoco PSF Procedure: A Cd++ solution was prepared by adding 0.076 gm Cd acetate to 2 ml DMF. An S-solution was prepared by adding 0.021 gm thiourea to 2 ml DMF 20 ml DMF. A polymer solution was prepared by dissolving 2.5 gm PSF in 20 ml DMF. The Cd and S solutions were added to the PSF solution and stirred well. The solution was transparent and colorless. Films are cast out of this solution by pouring into molds and left to dry for about 12-24 hrs. They were still colorless but somewhat hazy. When partially dry, the films are moved to an air-circulating oven at 90C for 2 hrs. This heating was carried out to decompose the thiourea and release the sulfur, which would form CdS particles with the Cd acetate in the polymer. Characterization UV-Visible Absorption Spectroscopy. The UV-visible spectrum of 0.16 to 0.36 mm thick UCdS-PSF film, measured on a Shimadzu UV 2401 PC spectrometer with air

PAGE 47

30 as reference. A 0.2 mm thick film of pure PSF was also measured on the same equipment for comparison. Optical images. A digital image of UCdS-PSF was taken using a Sony DCR-TRV20 camera on printed text. This film was about 0.36 to 0.16 mm thick (from left to right) as shown in Figure 3-2. Scanning electron microscopy (SEM). Samples of polysulfone-CdS composite for SEM were prepared by casting a film containing Cd-acetate and thiourea in DMF followed by heating of the dried film as per procedure of UCdS-PSF films. A sample of the ~0.3 mm thick film was mounted on a mild steel SEM sample holder, carbon paint applied and coated with carbon. Secondary electron images (SEI) of the films were obtained on a JEOL 6400 microscope. Energy dispersive x-ray spectroscopy (EDS). EDS spectra on the same UCdS-PSF sample used for SEM were obtained on JEOL 6400 microscope at a resolution of 68 eV using the ZAF quantitative method. Transmission electron microscopy (TEM). The TEM micrographs were obtained on a Zeiss EM 10A microscope operating at 80 KV. TEM samples for were prepared by heating a solution of Cd ac, thiourea, and PSF in DMF at 90C. This solution was diluted, dropped on TEM copper grids and then dried under vacuum. Another set of samples used a method similar to UCdS-PSF in which the solution of Cd ac, thiourea, and PSF was diluted and dropped on a TEM grid. The gird was allowed to dry for 10 minutes under ambient conditions and then heated in an air-circulating oven at 100C for 30 minutes. Results and Discussions PhCdS-PSF-I and PhCdS-PSF-II methods did not yield good quality transparent thick films at a thickness of 0.5 to 1 mm. The films obtained were warped, and opaque.

PAGE 48

31 The opacity is a result of phase segregation. The CdS particles, being totally incompatible with the polymer phase separate forming large size aggregates, which scatter light. Aggregation takes place not only when the particles were formed in THF by mixing the Cd++ and S-solutions but also after the polymer is added and when the solvent is being evaporated. While it is necessary for the solvent evaporation rate to be high in order to prevent aggregation, it was necessary to have a low rate of solvent evaporation to obtain good film quality, free of bubbles and voids in thick films. One was often obtained at the cost of another. This is a very complex system containing several components whose solubility and phase behavior is very difficult to predict compared to the familiar binary systems. However one of the main issues is the insolubility of CdS in organic systems, which necessitates the presence of capping agent or an alternative means of controlling aggregation of the precipitated particles. In spite of the presence of the capping agent phase separation occurred. Another challenge was to find a common organic solvent for Cd(NO3)2H20, Na2S9H2O and the polymer. It is quite reasonable to assume that instead of looking for a single solvent that will dissolve the three, a mixed solvent should be tried. Since Na2S is soluble in water and not in any of the solvents that dissolve PSF, one of the components of the solvent system needs to be water. THF is a good solvent for PSF and can also be used in conjunction with water. The problem however is that even at low levels of water in THF (1 to 2 ml in 25 ml), the polymer does not dissolve in spite of its good solubility in pure THF.

PAGE 49

32 Even if the presence of Cd salt and sodium sulfide were neglected it becomes a ternary system consisting of PSF, organic solvent such as THF and water. PSF phase separates at very low concentrations of water. Taking this into account the amount of water was reduced in the system to 0.1 ml in 25 ml of polymer solution in THF. This made the amount of water in 25 ml of the polymer + CdS sol equal to. The polymer remained in solution without phase separating. With a view to preventing the phase separation of the CdS it was sought to covalently link CdS to the polymer. Amine terminated CdS particles could be reacted with sulfonated polysulfone to give a sulfonamide bridging between the particles and the matrix. This covalent coupling between these two phases of CdS and polymer would improve their compatibility and reduce phase separation. AmCdS-SPSF-I failed to yield any nanocomposites as the as the precipitates of capped particles were totally insoluble in a number of solvents. More importantly they were insoluble in all the solvents that are known to dissolve SPSF. No nanocomposites could therefore be made. The insolubility of the precipitated particles pointed to the susceptibility of the thiol based capping agents to oxidation. The thiols can get oxidized, in air, into corresponding disulfides during synthesis and drying. The CdS particles, no longer being coated by the thiols, aggregate and form bulk CdS that is insoluble in organic solvents. Therefore it is desirable to carry out the reactions under an inert atmosphere. AmCdS-SPSF-II films made and dried in the glove box containing argon were dried under argon were somewhat more successful in that transparent sols of CdS in SPSF were obtained. However films cast from these sols showed poor transparency. Additionally, there was some bubble entrapment. It was somewhat reduced by heating in

PAGE 50

33 vacuum. However heating lead to darkening of the films. The AmCdS-SPSF-II films dried in air were opaque and yellowish in color. The thinner films (~10 m) dried in air were also opaque and yellow in color. Thus it appears that that air is indeed detrimental to obtaining transparent films. FPhCdS-SPSF films showed a similar behavior. Clear sols of CdS particles in SPSF were obtained. The films however had a poor transparency. This tendency was aggravated by drying at ambient pressure and reduced when done under vacuum. However, below 20 mm Hg pressures, there was a strong tendency of formation of bubbles of the evaporating solvent, leading to poor surface quality. For thick films (1 mm), after finally evaporating all the solvent, the resultant film was cloudy, but thinner (10 m) films, were relatively more transparent. Since carrying out the reaction and drying in a glove box gave was difficult, time consuming and gave only partially transparent films an alternative strategy was needed. UCdS-PSF films were obtained by a relatively simple process. They appeared translucent yellow in color and showed superior transparency as compared to all the other films made. When the partially dry films containing polymer, solvent, CdS and thiourea are heated to 90C the thiourea decomposes releasing sulfur, which reacts with the cadmium salt and forms CdS. However the results were not very reproducible. It was difficult to determine to what stage the films should be dried before they were put in the oven and to what temperature they should be heated. The fraction of solvent present in the polymer determines the glass transition temperature of the film. In general a higher the solvent fraction lowers the glass transition temperature. However since the fraction of solvent was not measured or controlled, it is

PAGE 51

34 difficult to predict the glass transition temperature. Heating of the partially dry film probably raises the temperature of the system above the glass transition temperature. This allows an increased diffusion of the Cd++ and S-in the system, which is necessary for formation of particles. However the viscosity of the system heated above its glass transition temperature is far greater than a liquid polymer solution. Particles made by heating Cd salt and thiourea in a polymer solution (liquid state) show a very strong tendency to aggregate (Figure 3-5B). On the other hand particles made in a partially dry film represent an arrested precipitation of the particles owing to the high viscosity of the medium. A film with a larger fraction of solvent when heated, gave yellow opaque films. The yellow color indicates that CdS was formed in the films. But due to the larger fraction of solvent, the viscosity was low and aggregation was high. On the other hand, insufficient solvent fraction in the film tends to yield films that are very light yellow and hazy. In these films the CdS probably did not form owing to the oven temperature being below the glass transition temperature of the polymer. A similar situation exists in case of the heating temperature. The temperature to which they were heated had to be lower than the boiling point of the solvent to avoid boiling of solvent and void formation in the films. If the temperature is too high then formation of CdS is good but aggregation is high and there is the chance of the solvent boiling as well as degradation of the polymer. At lower temperatures the CdS is not likely formed. An optimum temperature and solvent fraction probably exists. However a more detailed investigation of this system was not pursued.

PAGE 52

35 All the films produced by different films given above, were either opaque or partially transparent. These films used for optical limiting need to be fairly transparent in the operating wavelength range at low intensities. The opaque films are not practically useful as optical limiters and were not characterized further. Characterization of only the best UCdS-PSF films obtained was done. Optical Properties 00.511.52300400500600700800Wavelength (nm)Absorbance (arb. units) PSF CdS-PSF Figure 3-1. UV-visible absorption spectrum of uncapped CdS-PSF nanocomposite film. Figure 3-2. Optical photograph using Sony DCR-TRV20 camera on printed text. The UCdS-PSF film shown (Figure 3-2) was about 0.36 to 0.16 mm thick (from left to right) and showed a transmittance of 75 % at 800 nm (Figure 3-1). It appears translucent to the naked eye as shown in Figure 3-2. Pure PSF shows a transparent region with low absorption (extinction) from beyond 800 nm right up to its band edge, which is 325 nm (Figure 3-1). The nanocomposite also shows a similar transparency region

PAGE 53

36 stretching from beyond 800 nm. The transmittance is somewhat lower than the pure polymer due to scattering from the particle aggregates. In the band edge region the CdS-PSF nanocomposite shows an excitonic absorbance peak, which appears as a broad shoulder before the polymer band edge. The onset of the excitonic feature is 436 nm (Figure 3-1). This feature belongs to the CdS particles present in it. A more detailed discussion on this point is given in Chapter 4. Scanning Electron Microscopy (SEM) The cause of the low transmittance of the films was mentioned earlier as phase separation between the polymer and CdS. Scanning electron microscopy in the topographical mode and energy dispersive characteristic x-ray spectroscopy and transmission electron microscopy were used to elucidate the microstructure and the verify phase segregation as the cause of the low transparency of the CdS-PSF films. Bright and dark regions could be seen in the SEI images of the films. In the bright regions contrast was good while the largely dark regions were quite featureless and contrast could be obtained only at high probe current. This difference in the topographical contrast images (SEI) indicates large variations in topography. The dark regions were relatively smooth. Given in Figure 3-3A is the SEM SEI image of one of the bright regions that show good contrast. Energy dispersive X-ray spectrum (EDS) analysis revealed a greater content of cadmium in the bright regions (Figure 3-4A). On the other hand, on traces of Cd were detected in the dark regions (Figure 3-4B). This evidence as well as the greater topographical variation in this region indicates that the bright regions probably contain clusters or aggregates of CdS. The relative absence of topographical variation in the dark

PAGE 54

37 regions indicates that this region is mainly composed of the polymer matrix with only traces of cadmium. Figure 3-3. SEM SEI of CdS-PSF film at 15KV, 25000x. Bar = 2 m A B Figure 3-4. ED X-ray spectra of CdS-PSF (resolution = 68 eV, Quantitative method: ZAF) showing A) Cadmium-rich and B) Polymer-rich regions.

PAGE 55

38 Thus the films show a phase-segregated morphology composed of the cadmium-rich and the polymer-rich phases respectively. These were found to be the polymer-rich regions with only traces of cadmium. Transmission Electron Microscopy (TEM) The TEM micrographs (Figure 3-5A) of CdS particles synthesized by heating Cd acetate and thiourea in PSF without a capping agent reveal some small particles from 10 to 15 nm in diameter. However the predominant structure is that of an aggregate (Figure 3-5B). This micrograph appears to be a snapshot of an Oswald ripening process in which larger particles are seen to grow at the expense of smaller ones. At the same time some bridging is observed in between the large particles. This may be the remnant of two large particles growing at the expense of the same small particle. A B C Figure 3-5. TEM micrographs of CdS-PSF at 80KV, on Zeiss EM 10A showing A) CdS small particles and B) aggregates of CdS made in PSF solution without capping agent and C) Small particles formed on heating partially dried PSF film containing Cd ac and thiourea. It is clear from the above that aggregation of CdS particles occurs in absence of any capping agent or aggregation control strategy. The samples prepared with the UCdS-PSF method (Figure 3-5C) show that the aggregation is controlled effectively. Particle size in these films with arrested precipitation is considerably smaller (<10 nm). The important thing to note is that the samples prepared for TEM were on a grid and were very thin

PAGE 56

39 films (on the order of microns). Solvent drying is bound to be very fast and aggregation is expected to be very low. Thicker films will naturally have greater aggregation than this. Conclusions The aim of this chapter was to synthesize thick transparent CdS-PSF films. A variety of techniques involving thiol based capping agents for particles were used. Functionalization of the polymer was carried out with the intention of covalently bonding to the functionalized (capped) particles. Phase separation in the films could not controlled in order to get transparent films. The main factor causing phase separation were insolubility of CdS in organic systems. Functionalization of the matrix as well as the particles was not completely successful in controlling phase separation. The presence of water in the system was another cause of phase separation since polysulfone is insoluble in water. Reducing the amount of water from the system gave somewhat better results. Air was found to be detrimental to obtaining transparent films. The thiol capping agents of the CdS particles are known to get oxidized in air. The oxidized products of the capping agents are ineffective in controlling the aggregation of CdS particles, which precipitate out. When the synthesis of the thiol-capped CdS particles was carried out in an inert atmosphere, transparent sols of CdS in solvents such as DMF and acetone were obtained. Sols of CdS were also obtained in polymer solutions of polysulfone or sulfonated polysulfone in dimethyl formamide or tetrahydrofuran. However transparent films could not be obtained. During the drying of the solvent cast films, phase segregation occurred and the dried films were translucent to opaque. This was attributed to the oxidation of the thiols with residual and leaking air in the dry box filled with argon. Films made by heating partially dried films of PSF solution in DMF, containing thiourea and cadmium acetate

PAGE 57

40 were translucent. This method is an arrested precipitation of CdS in a medium of high viscosity. This method gave the best results (transmittance of 75%). UV-visible spectra of the films show a shoulder at the foot of the band edge of the polymer. This shoulder is the exciton absorbance of the CdS particles in it. Scanning electron microscopy of these translucent films reveals two distinct phases, namely the polymer –rich phase and the cadmium-rich phase. The energy dispersive characteristic x-ray analysis confirms the presence of the two phases. These phases are large enough in size to cause a large amount of scattering of light and make the films translucent. Transmission electron microscopy shows that CdS particles present are predominantly in the form of aggregates in the absence of a capping agent or an aggregation control strategy. The CdS particles aggregate in polymer solution on formation. Films of CdS synthesized by heating thiourea and cadmium acetate in a partially dried film of PSF showed very small particles in the TEM micrographs. The aggregation was lower compared to the particles synthesized in polymer solutions. Thus although the objective of making transparent thick films was not achieved completely there are some very important conclusions that can be made from this attempt. Opacity in the nanocomposites is the result of particles aggregating to form large clusters and in extreme cases forming a precipitate. The thiol based capping agents are unsuitable due to their oxidative instability in addition to their bad odor and toxicity. The use of thiourea to replace sodium sulfide as source of sulfur and arresting aggregation by increasing the viscosity of the matrix, are worth further investigation.

PAGE 58

CHAPTER 4 CDS-EPOXY NANOCOMPOSITES Introduction The chief cause of the failure of the CdS-polysulfone nanocomposites was the phase separation and aggregation of the CdS particles to form large-sized precipitates. This led to opacity in the films. CdS nanoparticles tend to aggregate together in organic systems due to their incompatibility with organic systems. Even though it was possible to obtain clear sols of CdS in organic solvents (acetone, DMF, THF) and in solutions of polysulfone, using the thiol-based stabilizers such as fluorothiophenol, solid films with thickness on the order of 1 mm were not possible. The cause of this was thought to be the oxidation and consequent of the thiols making them ineffective as stabilizers. In order to minimize contact with air the particle synthesis, solution casting of films and drying were carried out in a dry box purged with argon. However due to the slow solvent evaporation rate from the thick films in a closed box, there was an increase in the time of exposure to traces of oxygen leaking into the system. This destabilized the particles by oxidizing the thiol based capping agents. It would be desirable if the thiol based capping agents, which are toxic, malodorous and susceptible to oxidation in air, could be replaced by alternative strategies to control the aggregation of CdS. Nevertheless, the partial success of being able to produce clear sols of CdS in presence of polysulfone is noteworthy. If it were possible to achieve a similar sol of the CdS particles in a polymer, oligomer or precursor,99 which could be rapidly cross-linked, it would be possible to physically limit the growth of aggregates. This method, as 41

PAGE 59

42 opposed to the slow drying in solvent cast thick films would limit the time of exposure to oxygen even if the stabilizers were not highly effective or were unstable in air. Hence crosslinkable systems are worth investigating further. Another factor causing opacity in the films could be the presence of water in the system. Polysulfone phase segregates in the presence of water and hence transparent films cannot be obtained. The use of Na2S as the source of sulfur necessitated water to be a solvent or co-solvent for the synthesis of the nanoparticles. However the presence of water in the synthesis rules out the direct incorporation of the sol in solution of most important high temperature polymers, which are soluble only in organic solvents. It was found that even small quantities of water could lead to a phase segregated and opaque film on drying. Therefore the particles had to be precipitated by addition of a non-solvent, washed several times and dried to form a powder. This powder could be redispersed to form a clear sol in organic solvents such as acetone and incorporated in the polymer solution. However at each stage the exposure to air could no be avoided entirely and some of the CdS unavoidably precipitated. An alternative source of sulfur that was soluble in organic solvents would eliminate the extra steps of precipitating, drying and re-dispersing the CdS particles. The CdS particles could be formed in an organic solvent and the polymer could be dissolved in the same solvent. Thiourea is soluble in several organic solvents and can release sulfur on heating. Cadmium acetate hydrate can also be dissolved in organic solvents such as DMF. A solution containing cadmium acetate and thiourea will give CdS nanoparticles on heating.20,21,78,107,108

PAGE 60

43 Although we have eliminated water from the system the particles will aggregate and form a precipitate. We are then left with the same problem of controlling aggregation of the particles. One of the ways could be to use thiols to obtain capped particles, which would form stable sols in organic systems such as DMF. However it was worthwhile to seek an alternative synthesis method, which would be easy, less hazardous and more effective. Such a method was discovered accidentally while working with different crosslinkable systems. Cadmium acetate and thiourea solutions in DMF in presence of poly(oxypropylene) diamines and triamines of molar mass 2000 to 5000, were found to yield air stable sols. This family of oligomer are curing agents for epoxies at lower molar mass and are available under the brand name Jeffamine form Huntsman Chemical Company. Higher molar mass (2000, and 5000) Jeffamines are used as flexiblizing curing agents in epoxies. The sols obtained were stable for several days in air. These sols could be mixed with epoxy resins and amine curing agents and cured to get transparent thick or bulk films of CdS-epoxy nanocomposites. This chapter deals with the synthesis and characterization of 1 mm thick CdS-epoxy nanocomposites Materials and Methods Synthesis Polyoxypropylene polyamines Jeffamine are curing agents for epoxies. They are clear liquids with low Brookfield viscosities at 25C = 9 to 819 cp depending on the molar mass. The nomenclature of the Jeffamines is as follows: The letter stands for the number of amine groups (D-diamines or T-triamines) while the number stands for the molar mass. D2000 for example is a diamine with

PAGE 61

44 approximate molar mass of 2000 while T5000 is a triamine with a molar mass of approximately 5000. In the synthesis of CdS-epoxy nanocomposites, the higher molar mass Jeffamines are used as stabilizers for CdS sols while the lower molar mass Jeffamines are used as curing agents with an epoxy resin. Table 4-1. Materials used for the synthesis of CdS-epoxy nanocomposites. Chemical name Function Chemical formula Source Abbreviation Cadmium acetate, dihydrate Source of cadmium Cd(CH3COO)2H2O Fisher scientific Cd-ac Thiourea Source of sulfur (NH2)2CS Fisher scientific Thiourea N, N dimethyl formamide Solvent (CH3)2NCHO Fisher scientific DMF Polyoxypropylene diamines (Jeffamine D2000) Stabilizer x = 33.1 Huntsman chemical company D2000 Polyoxypropylene triamine (Jeffamine T5000) Stabilizer x +y+z = 81 Huntsman chemical company T5000 Polyoxypropylene diamine (Jeffamine D230, D400) Curing agent x = 2.6-6.1 Huntsman chemical company D230, D400 Diglycidyl ether of bisphenol A (Epon 828) Resin Shell chemical company Epon 828 Procedure outline. Thiourea and cadmium acetate dihydrate are dissolved in DMF and the stabilizer (Jeffamine D2000 or T5000) is added to it. The solution is heated to a certain temperature in the range of 65-80C for 20 minutes in a water bath with constant stirring with a magnetic stirrer till solution turns yellow. This is a sol of the CdS

PAGE 62

45 particles. Actual concentrations of the amine, CdS and the heating temperature decide the particle size obtained. Epoxy resin and additional curing agent (D400) are added to the sol, the mixture is stirred well, degassed under vacuum, poured in a polyethylene mold and cured at 60C for 2 hours followed by a post cure at 80C for 3 hrs. Optically transparent, flexible films about 1 mm thick with about 0.05-1 % CdS by weight were obtained. Calculations. Cadmium acetate dihydrate and thiourea react with each other in solvents such as DMF to form CdS nanoparticles20,78,107. With a view to obtaining CdS particles with the Cd:S ratio = 1, the same ratio was maintained in the reactants by taking stoichiometric amounts of cadmium acetate dihydrate and thiourea. The reaction was carried out in a closed test tube. However escape of some amount of sulfur from the system in the form of H2S cannot be ruled out. The amount of D2000 stabilizer added, corresponded to a 3 equivalents of amine hydrogen atoms per Cd atom. The epoxy and curing agent added to the system were added in amounts such that the CdS content in the films was maintained in the vicinity of 1% w/w. The CdS content in % w/w was calculated based on weights of added ingredients and assuming that all the solvent evaporated during cure of resin and that there were no other volatiles that escaped the system or any other components that entered the system during cure. Some volatiles other than DMF may have escaped from the system during synthesis of sol but their weight is assumed to be negligible. Typical procedure. Dissolve 0.285 gm of thiourea in 1 gm DMF in a test tube and mix on a vortex mixer after closing it with a polyethylene closure. Take 0.1000 gm of Cd

PAGE 63

46 -ac in another test tube. Add remaining amount of DMF (1.4000 gm) to it. Close it with a polyethylene closure. Mix it on a vortex mixer till all the Cd-ac dissolves. Add 3.6000 gm of D2000 to the Cd-ac solution. Mix it till a homogenous solution is obtained. Heat it at 75C for 20 minutes on an oil bath with constant stirring using a magnetic stirrer. Table 4-2. Typical recipe for making CdS–epoxy nanocomposites with 1% w/w CdS Cd-ac gm Thiourea gm DMF gm D2000 gm Epon 828 gm D400 gm 0.1000 0.0285 2.4000 3.6000 2.2748 0.8218 Remove the test tube from the bath, close it and cool by holding under flowing water from outside for 2-3 minutes till it attains room temperature. This sol may be stored in a cool dark place for use up to a few days later. In order to make films, take 3.0 gm of sol in a test tube and add 2.2 gm epoxy to it. The epoxy is very viscous and needs to be mixed thoroughly mixed with the sol till a clear, viscous and homogeneous solution is obtained. Add 0.8218 gm of D400 curing agent to the above solution and mix on a vortex mixer. Pour out the mixture in polyethylene (HDPE) lids used as molds and hold under vacuum for about 15 min till all the froth and bubbles from the mixing disappear. Then transfer the molds into an air-circulating oven and cure at 60C for 2 hours followed by 80C for 3 hrs. Determination of the function of stabilizer. In order to determine the function of the stabilizer and the solvent in the formation of the nanoparticles an experiment was conducted in which a mixture of Cd-ac, thiourea solution in DMF and D2000 at room temperature was heated in a water bath at 75C. Samples were taken from this mixture before heating (mix before heat), after 5, 10, 15, 20 minutes at 75C. Also samples of pure solvent DMF, stabilizer D2000, Cd-ac solution in DMF, Cd-ac solution in DMF and D2000, thiourea solution in DMF. Cd-ac and thiourea solutions in DMF both freshly

PAGE 64

47 mixed and after 2 days at room temperature were taken for comparison. These samples were analyzed suing UV-visible spectroscopy and FTIR spectroscopy. Characterization Optical images. The digital image of a CdS-epoxy nanocomposite film with thickness of 1.11 mm containing 0.74% w/w CdS. was taken on printed text, by a Sony DCR-TRV20 camera. UV-visible spectroscopy (UV-vis). UV-Vis spectra were obtained on a Shimadzu UV 2401 PC spectrometer. The UV-visible-NIR spectra were obtained on a Perkin Elmer UV-vis-NIR spectrometer. Solid films were mounted on a transmission stage with air reference. Liquid samples were taken in a standard optical glass cell (Starna Cells Inc., prod. # 21-G-1, 1 cm path length) with an identical empty cell as reference. The band edge of the CdS absorption onset was measured as the x-coordinate of the point of intersection of the tangents drawn on the horizontal region between 500-600 nm and the inclined region between 400-500 nm of the curve. The % transmittance of the samples was measured at 800 nm. Fourier transform infrared spectroscopy (FTIR). FTIR spectra were obtained on a FTIR Nicolet 20 SXB spectrometer in the transmission mode by dropping solution on NaCl discs. The same samples from the experiment mentioned above were tested using this technique. All spectra were normalized by setting the relative intensity of the C-H peak at 2390 cm-1 in proportion to the molar ratio of C-H present in each solution with respect to DMF. The underlying assumption being that the C-H stretching mode at 2930 cm-1 (not shown in Figures 4-4 and 4-5) is not affected in any way throughout the reaction.

PAGE 65

48 Transmission electron microscopy (TEM). TEM micrographs were obtained on a Zeiss EM 10A with an accelerating voltage of 80 KV. Some images, electron diffraction images and energy dispersive X-ray spectroscopy (EDS) data were obtained on a JEOL 2010F at 200 KV. The solid films were cryo-ultramicrotomed and the sections mounted on carbon coated, Formvar (SPI ChemTM) supported copper grids. The liquid samples containing CdS nanoparticles in D2000 and DMF were dropped (5L) on the similar grids, left for 1 minute, blotted out and dried at 60C for 15 minutes. X-ray diffraction (XRD). The x-ray diffraction data was obtained on a JEOL 3720 diffractometer. Solid XRD samples were prepared by cutting solid films of nanocomposites into approximately 1 cm2 pieces and pasting them onto glass slides using double-sided tape. Liquid samples were prepared by dropping a drop of the CdS sol on the glass slide and allowing it to spread over an area of approximately 1 cm2. The slides are then dried under vacuum to evaporate the DMF. The slides remain tacky. These slides are mounted on the diffractometer. The 2 range of 10 to 60 with a step size of 0.01. Sample 0 is the reference material in each case. In the solid films shown in Figure 4-7, neat epoxy film used as a reference while in the liquid samples a mixture of DMF and D2000 stabilizer is used as a reference. The references were made of the same composition as the samples except that they did not contain any cadmium acetate or thiourea. Two photon absorption spectroscopy (TPA). The two-photon spectrum was measured for three of the sample films containing 0.75% CdS, synthesized at 72, 76, 80C and 1.9% CdS synthesized at 75C in a pump-probe experiment. The pure polymer was also tested (not shown in graph). The pump wavelength was 1200 nm.

PAGE 66

49 Thermogravimetric/differential thermal analysis (TG/DTA). CdS-epoxy nanocomposites were tested on a Seiko TG/DTA 320 at 10C/min with 99.9% alumina as reference under nitrogen. Results and Discussion Optical Images The CdS-epoxy nanocomposite films appear very clear to the naked eye (Figure 4-1A). The slight haze is a result of surface roughness caused by curing of the films in open molds. Solvent (DMF) evaporation taking place simultaneously with the curing of the films probably leads to the surface roughness observed. The surface roughness was not quantified or controlled. The opacity due to light scattering from particle aggregates in the films appears to be minimal. A visual comparison between the CdS-PSF films and the CdS-epoxy films demonstrates the superior clarity of the latter (Figure 4-1B). A B Figure 4-1. Optical photographs taken on a Sony DCR-TRV20 camera on printed text. A) 1.11 mm thick CdS-epoxy film. B) 0.36 to 0.16 mm thick CdS-PSF film. The yellow color of the films is the result of the CdS absorption band being in the blue region of the visible spectrum. Absorption of the blue light from the spectrum leaves the transmitted light yellow in color. This phenomenon is called subtractive color. Neat epoxy on the other hand lacks the CdS absorption band and hence transmits all the visible light and is therefore colorless.

PAGE 67

50 UV-Visible Spectroscopy The UV-visible-NIR spectrum of pure epoxy and CdS-epoxy films shown in Figure 4-2A display the whole transparency window of the materials from the band edge on the left hand side (300 to 400 nm) to the multi phonon absorption peaks on the right hand side (2300 nm). The region in-between shows very little absorption, which indicates that the films are transparent throughout this region. The band edge region is of greatest interest here and will be discussed in greater detail. The band edge represents the excitation of valence band electrons across the forbidden gap (band gap) into the conduction band of insulators. It is observed as a steep or nearly vertical line as is seen in the vicinity of 350 nm for the epoxy polymer as well as DMF+D2000 in Figure 4-2A, B and C 00.511.523007001100150019002300wavelength nmabsorbance Pure epoxy CdS-epoxy A Figure 4-2. UV-visible-near-infrared absorption spectrum of A) Pure (neat) epoxy and CdS epoxy film. B) UV-visible spectrum showing DMF + D2000 and CdS sol synthesized at 69 C. C) Pure (neat) epoxy and CdS-epoxy film with CdS synthesized at 69C

PAGE 68

51 012345300400500600700800wavelength nmabsorbance DMF+D2000 CdS sol (69C) B 01234300400500600700800wavelength nmabsorbance Pure epoxy 69C C Figure 4-2. Continued The foot of this transition is called as the onset of the band edge and was measured by taking the intersection of the tangents drawn on the horizontal region between 500-600 nm and the inclined region between 400-500 nm of the curve. The band edge (onset) of the CdS sol (Figure 4-2B) is at 472 nm and is shifted to the right (red-shift) compared to the solution of D2000 and DMF, which has a band edge at about 350 nm. A similar behavior is seen in case of the solid films (Figure 4-2C) where the nanocomposite has its band edge red-shifted compared to the neat epoxy, which has a band edge at 325 nm. The band edges of the CdS sols and nanocomposites appear to have a shoulder on the right hand side of the polymer or solution band edge. This shoulder and the resulting red shift in the band edge are due to the absorbance from the CdS particles in the nanocomposite. A low intensity, broad CdS absorption band is superimposed on the band edge of the

PAGE 69

52 polymer. It appears as a broad shoulder seen between 400-500 nm (Figure 4-2A, B and C). In crystalline nanoparticles, this absorbance band is attributed to a bound exciton.39,44,76,90 The bound excitonic peak is sharp for monodisperse particles. It has been assumed in this dissertation that this absorption band is due to a bound exciton. A sharp peak for the exciton absorption could not be obtained for the CdS particles mainly due to two reasons: the broadening of the exciton peak and the proximity to the absorption edge of the solvent system. Broadening of absorption lines in semiconductors is classified as either homogeneous or inhomogeneous. Homogeneous broadening of the absorption lines is due to thermal lattice vibrations and hence is temperature dependent.39,45 Inhomogeneous line broadening is temperature independent and is the dominant broadening in polymer-inorganic nanocomposite systems. It is caused by a number of factors, the primary among them being particle size distribution.45 Other physical processes causing inhomogeneous broadening are stoichiometric variations, shape, surface structure, defect concentration, charge and local environment of the particles. Spectrally selective techniques such as spectral hole burning, selective emission spectroscopy etc. can be used to identify and measure inhomogeneous broadening. No attempt has been made in this dissertation to determine which of these physical processes have lead to the broadening of the exciton features seen in the spectra of the CdS-epoxy nanocomposites. Examination of the materials by (TEM) reveals a large variation in particle size. This distribution of particle sizes is assumed to contribute to the broadening.

PAGE 70

53 An estimate of the mean particle size may be obtained from the exciton absorption edge as measured by UV-visible absorption spectroscopy using the tight binding model and the effective mass model as discussed in Chapter 2. Particles with a band edge in the range of 400-500 nm generally are <6 nm, although there are some differences in the exact sizes calculated based on different models (Figure 2-1) the sizes are generally in the < 6 nm range which is called the strongly confined regime.45 The band edge-particle size correlation is dealt with in greater detail in Chapter 5. Here it will suffice to say that based on the quantum size effect,42-45,89,90,109,110 the closer the onset of the excitonic transition of the particles is to the blue region of the spectrum, the smaller are the particles while the closer the transition is the bulk band edge of CdS at 525 nm,39 the larger are the particles in size. 00.20.40.60.811.21.41.61.82800750700650600550500450400350300Wavelength (nm)Absorbance Cdacetate+thio fresh mixed Cd-acetate+thiourea 2days Cd-ac+thio+D2000 5 min 10 min 15 min 20 min Figure 4-3. UV-visible spectra of reaction mixture during synthesis of CdS sols The pure DMF, solutions of Cd-acetate and thiourea and their fresh mixture have no measurable absorbance before the band edge, which is near 300 nm (Figure 4-3). However, the UV-vis spectra of the mixture of cadmium acetate and thiourea solutions in DMF after two days at room temperature exhibits a sharp shoulder with an onset at 410

PAGE 71

54 nm. This is characteristic of CdS nanoparticles and shows that CdS particles are formed even without heating the mixture. The particle sizes are in the range of 1 to 5 nm based on the location of the band edge. The mixture of cadmium acetate, thiourea and D2000 in DMF has a band edge in the vicinity of 400 nm owing to the band edge of D2000 which s also at about 400nm. No shoulder is seen in the vicinity of the band edge indicating that the particles are not yet formed. However on heating the mixture for 5, 10, 15 and 20, minutes, a shoulder at the foot of the band edge of the mixture develops, which makes the band edge shift consistently to the red (to the right in the graph shown in Figure 3. The shift in the band edge to the right as the reaction progresses is an indication of an increase in particle size.42-44 45045546046547047548048549049505101520Time (min)Band Edge nm Band Edge Figure 4-4. CdS band edge as a function of progress of reaction. Red shift in band edge is indicative of particle growth. The red shift in the band edge as a function of the progress of the reaction (Figure 4-4). The shift in the band edge appears to level off at about 20 minutes. It can be inferred based on this that the particles are formed on heating of the mixture and they continue to grow till about 20 minutes when their growth slows down considerably.

PAGE 72

55 Fourier Transform Infrared Spectroscopy (FTIR) All spectra containing DMF showed a great degree of resemblance with the spectrum of DMF111,112 while those containing DMF and D2000 showed a resemblance with the spectra of both. The important peaks to be considered are the DMF peaks located at 1675, 1385 and 1090 cm-1 as shown in Table 4-3. 0.000.501.001.502.002.5099110881184128113771473157016661763wavenumber cm-1Absorbance DMF Cd-acetate Thiourea Cd+S fresh Cd+S 5days D2000 mixnoheat 5 min 10 min 15 min 20 min Figure 4-5. FTIR spectra of DMF, cadmium acetate and thiourea solutions, and D2000 showing the important peaks and the complexation of cadmium by DMF. Table 4-3. FTIR peak identification and assignment Sample Peak location cm-1 Intensity strong (s), medium (m), weak (w), broad (br), shoulder (sh) Assignment (% contribution of functional group) DMF 1675 s 30CO, 35NCH, 13CN asy DMF 1507 m, br 26HCN, 16C-N, 22ON asy DMF 1456 sh 97HCH DMF 1440 m 38HCN, 25HCH, 16CNs DMF 1385 m 39HCN, 25HCH, 25CN DMF 1259 w 42HCN, 22HCH, 14C0 DMF 1093 m 51NCH, 23CO D2000 1460 w, br 97HCH D2000 1380 w -D2000 1110 m, br CO ether

PAGE 73

56 The spectrum of cadmium acetate in DMF in Figure 4-5 shows a broadening and reduced relative intensity of the peak at 1675 cm-1 and to a smaller extent in the peaks at 1385 and 1090 cm-1. This indicates a complexing of the cadmium ions by the DMF either or both of the carbonyl and N-C groups. The thiourea solution in DMF showed a strong absorbance in the region 1600 to 1700. The absorbance was so strong that the spectrum was truncated above certain intensity as the detector signal saturated. This is probably the result of a superimposition of the peaks associated with the DMF peak at 1675 and the first overtone of NH stretching modes of thiourea.113,114 0.000.200.400.600.801.001.201.4099110881184128113771473157016661763wavenumber cm-1Absorbance BeforeHeating 5 min 10 min 15 min 20 min Figure 4-6. FTIR spectra of reaction mixture containing cadmium acetate, thiourea solutions in DMF and D2000 showing the complexation of cadmium by DMF and its recovery as the reaction progresses. The reaction mixture containing cadmium acetate, thiourea, DMF and D2000, before heating also shows similar broadening and reduction in peak intensity of the peaks at 1675, 1385 and 1090 cm-1 (Figure 4-6). This confirms the fact that DMF does complex Cd-ac. The samples at different periods of heating show a gradual narrowing and recovery in the intensity of the peaks at 1675, 1385 and 1090 cm-1. Again the effect was

PAGE 74

57 more pronounced in case of the peak at 1675 cm-1, as compared to the peaks at 1385 cm-1 and 1090 cm-1. The spectra appear progressively more similar to the spectrum of DMF as compared to the spectrum of the mixture before heating. Thus, the solvent DMF appears to complex cadmium in its ionic form while as the reaction progresses and the ionic cadmium is consumed to form insoluble CdS, the DMF is not able to complex it as much as ionic cadmium. This leads to the narrowing and increase in intensity of the peaks mentioned above. On the other hand there was no evidence that there was any complexation of cadmium acetate at the D2000 ether linkage. The peaks associated with the terminal amine groups could not be detected in this experiment, which is perhaps owing to their low molar concentration in the polymer. These terminal amine groups could play a role in complexing the cadmium acetate or CdS particles but it was not clear from this experiment. In the absence of any such evidence, the conclusion regarding the role of the stabilizer D2000 in preventing the aggregation of CdS particles formed from the reaction of cadmium acetate and thiourea is that it screens the secondary bonding forces that lead to aggregation and increases the viscosity of the medium, which slows down diffusion of particles. X-Ray Diffraction X-ray diffraction is one of the most widely used techniques for evaluating the crystallinity and structure of films and powders. It is essentially a method used for evaluating the lattice spacings, also known as the d-spacings, present in a crystalline sample. The technique is analogous to an optical diffraction experiment whereby the line spacings of the grating can be determined by the resulting diffraction of light. In x-ray diffraction, the incident energy is x-rays.

PAGE 75

58 Due to the small wavelength, relative to visible light, x-rays can interact with the atomic planes in the internal structure of materials, causing the x-rays to diffract at certain angles of incidence, and create a unique set of diffracted or reflected x-ray peaks. The technique is based on Bragg’s law, which states that these reflected x-rays from a set of lattice planes constructively interfere only at specific angles. These angles can be predicted using the mathematical expression of Bragg’s law, stated as: 2d sin = n (4.1) where n is an integer, is the angle of incidence for the x-rays, is the wavelength of the diffracted x-rays, and d is the spacing of the planes in the sample. The x-ray diffraction peaks of nanocrystals are usually broadened as compared to bulk samples of the same materials. With reference to the XRD technique, individual crystals smaller than 100nm are called particles.115 Based on the extent of broadening, it is possible to measure the nanocrystallite dimensions. The broadening of the peaks is given by the Scherrer equation.115 B = 0.9/(tcos ) (4.2) B2 = Bm2-Bs2 (4.3) and t = 4/3 d (4.4) where, t = coherence length of particle in nm, d = diameter of spherical crystal in nm, = wavelength of X-ray radiation in nm, Bm = 2 breadth (FWHM) of the sample in radians, Bs = 2 breadth (FWHM) of the standard in radians. B is the broadening that can be attributed to the particle size effect alone and not machine or temperature broadening. B is zero when particle size exceeds 1000 A. In general, relatively crude techniques suffice for estimating particle sizes in the range 0

PAGE 76

59 500 , On the other hand, for larger particles, which show very little line broadening, accuracy of the diffractometer and the sample preparation etc. can pose a challenge. Non-uniform strain in crystals can interfere with the size determination. Uniform strain causes no broadening but causes shift in peak location.115 0100200300400500600700800900010203040506072Intensity (arb. Units) 0 Sample 0 Sample 1 Figure 4-7. XRD spectra of CdS-epoxy films. Sample 0 = pure epoxy. Sample 1 = CdS epoxy film (1 % CdS w/w, synthesized at 70 C). 20040060080010001200140010203040502Intensity (arb. units) Sample 0 Sample 1 Figure 4-8. XRD spectra of CdS sols. Sample 0 = DMF + D2000. Sample 1 = CdS in DMF+D2000 (2 % CdS w/w, synthesized at 70 C).

PAGE 77

60 The solid films recorded two very broad peaks centered at about 20 and 45 (2) (Figures 4-7 and 4-8). There appeared to be no significant difference between the reference neat epoxy film and the CdS nanocomposites. This is indicative of a largely amorphous structure. Similarly, in case of the liquid samples a single amorphous peak centered at around 20 was obtained. Again, there was no noticeable difference between the sample and the reference. This confirms the amorphous nature of the materials. The estimate of CdS particle size based on the exciton onset as measured by UV visible spectroscopy is on the order of 1 to 5 nm. The breadth of the XRD diffraction peaks of the size range of particles would be several degrees as calculated from the Scherrer Equation (Equation 4-3). Additionally, the overall concentration of the particles is only about 1% w/w in the films and 2% w/w in the sols. The intensity of the peaks corresponding to the CdS would therefore be very low. Therefore it is unlikely that any of these rather broad and low intensity (due to low concentration) peaks could be differentiated from the amorphous peaks in the samples. Thus it was not possible to conclude based on merely XRD data whether the clusters in the films were completely amorphous or whether there was any crystallinity associated with the particles that was below the detection limit of the technique. However electron diffraction of the nanocomposites showed an amorphous material (Figure 4-9), which confirms the XRD data. Neither the small nanoparticles nor the larger clusters showed any crystallinity.

PAGE 78

61 Figure 4-9. Electron diffraction on a large cluster in CdS-epoxy films with 0.7 % w/w CdS on JEOL 2010F, 200KV. Transmission Electron Microscopy Figure 4-10. TEM micrographs of 1.5% w/w CdS sols in DMF +D2000 showing small particles, medium and large sized aggregates. A large majority of evenly distributed polydisperse particles in the size range of 2 to 10 nm is evident in the TEM micrographs of CdS sols (Figure 4-10). A few clusters ranging tens of nanometers across and occasional clusters of hundreds of nanometers are also visible in the micrographs. In the solid films, obtaining good samples by microtoming was very difficult. The films were semi-flexible and sectioning them at room temperature was not possible.

PAGE 79

62 Potting them in a hard epoxy resin followed by sectioning at room temperature was also not successful. Finally the potted samples were cryo-ultramicrotomed to get the samples required. Even these sections obtained were rather thick and uneven and hence getting good images was a challenge. Images could therefore be obtained only in regions where the section was thin enough to get an image. A B Figure 4-11. TEM micrographs of CdS-epoxy films with 0.7 % w/w CdS films showing: A) Small particles on Zeiss EM 10A, 80KV. B) Large aggregates and their vicinity on JEOL 2010F, 200KV A region was chosen from one of the best sections obtained that is representative of the whole section (Figure 4-11A). A mean aggregate size of 10.5 2.3 nm was estimated for this sample using a sample size of 80 particles. In general better images could be obtained at low accelerating voltages of 80 KV, which is customarily used for biological and polymeric samples rather than 200 KV commonly used for examining metals and inorganics. In fact the smaller particles 2 to 20 nm, which form the majority of the particles seen at 80 KV, could barely be seen at 200 KV. Larger aggregates were more prominent in the solid films (Figure 4-10B) compared to the sols. Yet the aggregation is not to the extent that it would cause opacity in the films. The fact that the films are transparent proves that the aggregation of the particles

PAGE 80

63 has been minimal and that the polymeric stabilizer is successful in preventing the aggregation of the CdS particles and thereby obtaining transparent films. Most of the particles are irregularly shaped with an aspect ratio close to unity. There appears to be a somewhat greater concentration of particles and clusters as observed visually, in the vicinity of large aggregates. The particles are concentrated in certain locations where Oswald ripening occurs in which the larger particles grow at the expense of smaller ones. Energy dispersive X-ray spectroscopy (EDS) of large clusters also taken on JEOL 2010F at 200 KV, was used to detect the presence of cadmium and sulfur. The CdS content of the sample films was 0.7 to 1% as calculated from the reactants. So quantitative results from the EDS are not very reliable and may be inaccurate by up to a factor of 2 for such low concentrations. In fact it was not possible to detect Cd or S unless the beam was focused on a fairly large cluster where the local concentration of Cd and S was high enough to be detected above the noise. Quantitative EDS analysis of the particles indicates that cadmium content of the clusters is higher than stoichiometric amount as compared to sulfur i.e. Cd:S ratio is greater than 1. Actual atomic ratios of Cd:S were 70:30. The reaction mixture contained stoichiometric amounts of Cd and S. Some losses of sulfur are anticipated during the synthesis of the nanoparticles when the test tube containing thiourea and cadmium acetate solutions was heated in a closed (not sealed) test tube. However it is unlikely that the loss of sulfur at this stage is sufficient to account for the stoichiometric mismatch observed in the EDS results. Another possibility is that heating the reaction mixture for 20 min is not sufficient for the whole amount of thiourea present in the mixture to decompose into sulfur. The

PAGE 81

64 decomposition of thiourea in the reaction mixture is probably a first order reaction. Therefore they may still be some unreacted thiourea in the matrix. However being at a low concentration below the detection limit of the EDS technique, it is not detected. Thus the loss of sulfur from the system may be lower than one would infer merely on the basis of the EDS quantitative results. Solid microtomed samples showed serious beam damage at the location where the beam was focused for EDS. Two Photon Absorption Spectrum (TPA) There was no significant nonlinear signal in the pure polymer, signifying that the nonlinearity of the pure polymer may be neglected and that all the nonlinearity of the nanocomposite may be safely attributed to the CdS particles alone. 00.30.60.91.21.5340360380400420440460480500Eq. 1-photon wavelength (nm) (cm/GW) 72C; 0.74% w/w 76C; 0.74% w/w 80C; 0.74% w/w Bulk CdS (Eg=2.42eV) 75 C; 1.9% w/wPump = 1200 nm Bulk CdS data 30x reduced Figure 4-12. The two-photon absorption spectrum (TPA coefficient () as a function of the equivalent 1-photon wavelength) of CdS-epoxy films with 0.75 and 1.9% CdS synthesized at different temperatures. The nonlinear signal strength was found to be very small (2% at the most). This was attributed to the low loading fraction of the CdS in the films. Based on the concentration of CdS in the films and the film thickness the equivalent path length

PAGE 82

65 through CdS was calculated and the two-photon absorption coefficient was measured. The curve of TPA coefficient () as a function of the equivalent 1-photon wavelength of the nanocomposites (Figure 4-12) shows a trend similar to bulk CdS except that the values were about 30 times less than that of the bulk CdS. It is not clear as to what might be the cause of this reduction in the value of the TPA coefficient. One of the probable reasons is that the CdS particles formed are not crystalline. This is evidenced from the XRD and electron diffraction data, which confirm the amorphous nature of the particles. The non-stoichiometry of the CdS could be factor contributing to the amorphous nature or could independently suppress the nonlinear properties of the particles. There are other probable sources from which this reduction could arise such as including presence of ligands, surface defects particles, temperature and pressure. Thermogravimetric/Differential Thermal Analysis (TG/DTA) The onset of a drastic weight loss is seen at about 300C (Figure 4-13). This is the onset of degradation of the material. There is no significant loss in weight before the onset of degradation, which signifies that there are no volatiles present in the film. The films have good thermal stability compared to most polymers. Conclusions In a novel synthesis, highly transparent, solid, thick nanocomposite films of CdS nanoparticles in an epoxy matrix were produced from air stable CdS sols. A high transparency from 500 nm to 2300nm was observed in the UV-visible NIR spectra of the material. A broad exciton feature can be seen in the band edge of the material. This is indicative of a broad particle size distribution. Transparent films were possible only by a

PAGE 83

66 successful prevention of aggregation of the CdS nanoparticles in the sol and during the curing of the epoxy. 0400080001200060206367507810899Temperature CSample Weight ug Degradation = 300C Figure 4-13. Thermogravimetric analysis of CdS-epoxy films containing 0.75% w/w CdS. Oligomeric stabilizers such as Jeffamine D2000 were used to make the air stable sols and thick films. FTIR (infrared spectroscopy) showed that that the cadmium in its ionic form, was complexed with the DMF while CdS particles were probably not complexed by DMF. No interaction with the stabilizer D2000 was detected with FTIR. It was hence inferred that the oligomeric stabilizer contained in the sols prevented aggregation of the CdS particles by screening the secondary bonding forces leading to aggregation and increasing viscosity, which slows the diffusion of the particles. Transmission electron microscopy (TEM) of the sols revealed that majority of the particles were in 2-20 nm range while there were medium (20 to 50 nm) and large sized clusters (>50 nm). TEM of the films similarly revealed small particles in the few nanometer range but relatively more medium and large sized aggregates as compared to the sols. EDS (energy dispersive characteristic x-ray) data revealed that the Cd:S ratio of

PAGE 84

67 the particles is about 70:30. Electron diffraction and x-ray diffraction showed that the particles were amorphous. Two photon absorption spectra of the films were obtained. The values of the nonlinear coefficients of the nanoparticles in the films were about 30 times lower than that of the bulk values. One of the major causes of this could be the lack of crystallinity of the particles. The thermal stability was evaluated using thermogravimetric analysis. The films were found to be stable upto 300C before which there was no detectable loss in weight from degradation or out gassing of volatiles.

PAGE 85

CHAPTER 5 AGGREGATION CONTROL IN CDS-EPOXY NANOCOMPOSITES Introduction The CdS-epoxy nanocomposites, that were synthesized and studied in the earlier chapter, are a complex system involving a polymeric diamine along with other cadmium complexing entities limiting the growth and aggregation of the particles. The ability to make 1mm thick, transparent CdS-epoxy nanocomposites is dependent on the ability to control aggregation of CdS in the sols and during the curing of the films. CdS synthesized in DMF by heating a solution of Cd-acetate and thiourea, in the absence of a stabilizer, forms a precipitate. CdS aggregates and precipitates in organic solvents due to its insolubility in organic solvents. This phase separation behavior was noted in Chapter 3. Although no effort is taken to prevent the phase separation in this chapter, the focus is to control aggregation, which is kinetic phenomenon. In presence of Jeffamine stabilizers, aggregation of the newly formed CdS nanoparticles is limited giving rise to clear, stable, yellow sols. Control of aggregation is possible in this system by adjusting different parameters in this system. This chapter deals with the controlling CdS aggregation by adjusting the concentration and the molar mass of the oligomeric Jeffamine stabilizer, the order of addition of reactants and concentration of CdS. Materials and Methods Synthesis Stabilizer molar mass and sol stability. CdS sols were prepared using the general procedure outlined in Chapter 4 with the exception that the stabilizers with different 68

PAGE 86

69 molar mass were used. D230, D400, D2000 and T5000 were used and the stability of the sols was observed visually. Stabilizer concentration and sol stability. CdS nanoparticle sols, samples 1-7 were prepared in DMF with Jeffamine D2000 as stabilizer, CdS concentration: 61.1 mM (molal – moles per 1000 gm of solvent) and CdS stoichiometry: Cd:S = 1.143 for each of the samples. Sample 8 was prepared with the same amount of CdS in pure DMF, without the D2000 stabilizer. The sols were prepared by heating the solutions of Cd-acetate and thiourea in DMF with Jeffamine D2000 as stabilizer from room temperature to 98C at about 1.5C/min. CdS concentration and sol stability. CdS-epoxy hybrids with different concentrations of CdS based on final film weight were made using D2000 stabilizer at 75C. Samples 0 to 5 were made with 0 (pure epoxy), 0.94, 1.35, 1.91, 2.39, 2.91% CdS w/w. The films were cured at 60C for 2 hr and 80C for 3 hr. All the films were approximately 1mm in thickness. The amount of D2000 corresponded to 3 times the equivalents of cadmium acetate. Characterization UV-visible spectroscopy (UV-vis). UV-Vis spectra were obtained on a Shimadzu UV 2401 PC spectrometer. Solid films were mounted on a transmission stage with air reference. Liquid samples were taken in a standard optical glass cell (Starna Cells Inc., prod. # 21-G-1, 1 cm path length) with an identical empty cell as reference. The band edge of the CdS absorption onset was measured as the x-coordinate of the point of intersection of the tangents drawn on the horizontal region between 500-600 nm and the inclined region between 400-500 nm of the curve. The % transmittance of the sample films was measured at 800 nm.

PAGE 87

70 Transmission electron microscopy (TEM). TEM micrographs were obtained on a Zeiss EM 10A with an accelerating voltage of 80 KV. The solid films were cryo-ultramicrotomed and the sections mounted on carbon coated, Formvar (SPI ChemTM) supported copper grids. The liquid samples containing CdS nanoparticles in D2000 and DMF were dropped (5L) on the similar grids, left for 1 minute, blotted out and dried at 60 C for 15 minutes. Results and Discussions Stabilizer Molar Mass and Sol Stability Stabilizer molar mass is the most critical factor in obtaining stable sols. CdS sols were made with different grades of Jeffamine polyamines (Table 1) indicate that lower molar mass polyamines such as D230 and D400, were unable to stabilize the CdS particles and precipitates were obtained. Stable sols were obtained with D2000 and T5000. Table 5-1. Stability of sols using different Jeffamine stabilizers. Stabilizer Sol/precipitate D5000 Sol D2000 Sol D400 Precipitate D230 Precipitate The higher molar mass stabilizers of CdS (D2000 and T5000) act as reactive flexibilizers in the cured epoxy. Generally, higher the molar mass of the flexibilizer, lower is the stiffness of the epoxy. Therefore D2000 was chosen over T5000 for all the syntheses with the view to obtaining stiffer films. A very high content of these flexibilizers makes the cured films very flexible or even tacky. It is necessary to limit the content of the stabilizers in the CdS sols. It is necessary to determine the concentration

PAGE 88

71 regime in which stable sols may be obtained and useful films of nanocomposites may be synthesized. Stabilizer Concentration and Sol Stability Clear yellow sols were formed in samples 1 to 7 (having decreasing concentration of stabilizer). Sample 8, which contained no stabilizer, showed a light yellow precipitate immediately on synthesis (Table 5-2). However on letting the test tubes stand for 1 month in a closed cabinet at 72, samples 5, 6, 7 also showed precipitates that had settled at the bottom of the test tubes while samples 1 to 4 had needle-like crystals growing on the glass surface inside the test tubes (Table 5-2). These needles were assumed to be CdS. The supernatant liquid in the test tube after 1 month was colorless and clear in samples 7 and 8, yellowish in 6, yellow in 4 and 5 and deep yellow in 1,2, and 3. This indicated that no CdS was suspended in them any more in 7, and 8, while some was still suspended in 4, 5, 6 and even more in 1, 2, and 3. As the concentration of the stabilizer increases (samples 7 through1), the stability of the particles in the sols improves. At relatively higher concentrations such as in Samples 1 to 4 the sols were relatively stable sols of CdS and that only a gradual solution crystallization was observed leading to the formation of needles. This indicates that the sols are not thermodynamically stable but there is a finite driving force towards nucleation and growth of the nanocrystals. However the time scale for this aggregation may be adjusted by adjusting the concentration of the sols. If the concentration of CdS in the sols is adjusted so that aggregation is negligible in a matter of a few hours (such as in samples 1 to 4) required for curing of the films, then transparent films may be obtained. There also appears to be a possible correlation between the particle size and the amine concentration as revealed by the difference in hue of the sols in the samples 1 to 8,

PAGE 89

72 observed visually. Subtractive color in CdS nanoparticles is explained in Chapter 2. A deep yellow-brown sol or precipitate means that particle size is relatively large. On the other hand, a lighter yellow means that particle size is relatively small. Sometimes a greenish tinge is seen to the colloids. This is due to the photoluminescence of the CdS particles, which luminesce82,116 in the green region of the visible spectrum. In general, sols with lower concentration of stabilizer lead to a smaller particle size while those with a higher concentration of the stabilizer lead to a higher particle size. Table 5-2. Stability and morphology of particles with amine concentration Sample D2000 mol % Stability after 1 day Sol color after 1 day Stability after 1 month Sol color after 1 month 1 19.19 Stable Deep yellow Numerous tiny needles Deep yellow 2 9.57 Stable Deep yellow Many tiny needles Deep yellow 3 5.32 Stable Deep yellow Several large, needles Deep yellow 4 3.43 Stable Yellow Few large, flat needles Yellow 5 2.10 Stable Yellow Yellow ppt Yellow 6 1.22 Ppt Yellow Yellow ppt Yellowish 7 0.52 Ppt Lemon yellow Yellow ppt Colorless 8 0.00 Ppt Light lemon yellow Yellow ppt Colorless Order of Addition of Ingredients The order of addition of ingredients was found to have a profound effect on the stability of the sols. In some of the earlier syntheses a mixture of the stabilizer (D2000) and curing agent (D400) was used while synthesizing the sols. After the synthesis of the sols, the required amount of epoxy was merely added and the sols. The mixture was then poured out in moulds and cured. In some cases, the sols obtained, were cloudy yellow, in some they were clear and in some cases precipitates were obtained. Mainly four such combinations were made and their results are summarized in Table 5-3

PAGE 90

73 Table 5-3. Synthesis of CdS sols in mixed stabilizers. Stabilizer Curing agent Sol Sol clarity D2000 D400 Clear Slightly hazy D2000 D230 Hazy Translucent D5000 D400 Clear Slightly hazy D5000 D230 Hazy Translucent Transparent films (>85% transmittance) were obtained only when the CdS particles were prepared using a single high molar mass stabilizer (either of D2000 or T5000). The epoxy could be added to the sol and mixed thoroughly before the addition of the curing agent. Equivalent results were obtained when a mixture of curing agent and epoxy was added to the sols and mixed thoroughly. However the sol would turn turbid if the curing agent (D230 or D400) were added to the sol before the addition of epoxy. Thus low molar mass Jeffamines not only are ineffective in controlling precipitation during synthesis but also reduce the stability of CdS sols and induce precipitation of CdS after their synthesis. Instability of the sols was observed to be more profound in case of D230 than D400. CdS Concentration and Sol Stability The absorption spectra of the films having 0.9, 1.3, 1.9, 2.3, 2.9% CdS w/w were measured (Figure 5-1) and an increasing blue shift of the band edge onset with increasing CdS content was observed. The % transmittance of the samples at 800 nm was also measured and samples with 0.9, 1.3, 1.9% CdS, show a nearly constant transmittance at about 82-85% (Figure 5-2). However for the samples with 2.3 and 2.9% w/w CdS, the fall in the transmittance is rather drastic. This is indicative of precipitation of CdS within the matrix and the scattering of light from the larger sized particle aggregates. The transition takes place around 1.9% w/w of CdS. Thus in the current system, the maximum CdS loading possible if about 1 mm thick transparent films are desired, is 1.9%.

PAGE 91

74 y = -6.2287x + 457.66R2 = 0.9744384404424444464484504524540123CdS % w/wBand Edge (nm) 4 Figure 5-1. Band edge of CdS-epoxy nanocomposites as a function of CdS concentration. y = 1.7012x + 81.377R2 = 0.8788y = -75.667x + 223R2 = 0.9928010203040506070809001234CdS % w/w% Transmittance Figure 5-2. Percent transmittance of CdS-epoxy films as a function of CdS concentration. The TEM pictures in (Figure 5-3A & B) shows a majority of 2-20 nm particles in sample with 0.94% CdS and aggregates of 20-50 nm in sample with 1.35% CdS. The presence of large aggregates (>100 nm ) was to a minimum in these samples. However in places large aggregates were detected as shown in C, D. Sample with 0.94% CdS showed fewer large aggregates as compared to sample with 1.3% CdS. In sample with 1.91% CdS, a relatively larger number of large aggregates were detected as shown in D.

PAGE 92

75 Increasing CdS concentration is related to an increasing tendency to form large sized aggregates (>100 nm). These large aggregates have the potential to reduce transmittance of the films. A B C D Figure 5-3. TEM micrographs of CdS-epoxy films in DMF+D2000 on Zeiss EM 10A, 80KV. Bar = 500 nm. Conclusions Control of aggregation of CdS in the sols and during the curing of the films is essential to obtaining 1 mm thick transparent CdS-epoxy nanocomposite films. Aggregation of CdS was controlled by adjusting the concentration and the molar mass of

PAGE 93

76 oligomeric Jeffamine stabilizer, the order of addition of reactants and concentration of CdS. Jeffamines of molar mass 2000 and 5000 were effective as stabilizers and yielded clear sols while lower molecular weights—D230 and D400—were found to be ineffective as stabilizers. D2000 was chosen over T5000 due to its lower molar mass and consequently more rigid films. CdS sols made using single stabilizer (D2000) followed by addition of epoxy resin and curing agent (D230 or D400) yielded transparent films. Preparation of CdS particles in a mixture of D2000 and curing agent (D230 or D400) followed by addition of epoxy resin yielded translucent films. Adding of D230 or D400 to a clear sol of CdS precipitates the sol. Hence the curing agent should be added only after mixing in the epoxy with the sol or the curing agent should be pre-mixed with the epoxy and then added to the sol. Sols of CdS require at least a 3.43 mol % D2000 to be stable. Sols with less than this amount precipitate immediately on synthesis or on standing for weeks to a month. At concentrations greater than 3.43 mol % a gradual crystallization of needles of what is assumed to be CdS proceeds. Indicating that the sol is not thermodynamically stable and that there is diving force towards the formation of crystals. A probable effect of stabilizer concentration on band edge of the sols was detected and requires further investigation. A decrease in band edge with increasing CdS content in films was observed. Percent transmittance of films remains nearly constant at 82-85% for CdS content below 1.9% while it drastically reduces at CdS concentrations above 1.9%. Low CdS content (0.9%) was associated with a majority of 2 to 20 nm particles as observed by TEM and relatively fewer aggregates larger than 100 nm. An increase in CdS content was found to

PAGE 94

77 increase the abundance of aggregates >100nm, which are responsible for reducing transparency of the films.

PAGE 95

CHAPTER 6 FACTORS AFFECTING TRANSMITTANCE AND BAND EDGE OF CDS-EPOXY NANOCOMPOSITES Introduction The CdS-epoxy nanocomposites, that were synthesized and studied in the earlier chapter, are a complex system involving a polymeric diamine along with other cadmium complexing entities limiting the growth and aggregation of the particles. Synthesis parameters can be varied to control the particle properties such as particle size and the aggregation, which in turn control the band edge and the transparency of the nanocomposites formed. Several factors including temperature, CdS concentration, amine concentration may have an effect on the band edge and % transmittance of the nanocomposites. Owing to the complexity of the system, interactions between these factors are likely. A design of experiments is necessary to determine the significant factors, detect interactions, evaluate the effect of each of the factors and fit a regression model to the results. Identification and Measurement of Responses The ability to control the band edge of a material is important in determining the working range and efficiency of devices such as optical limiters,1,28,30 nanocomposite solar cells9,10 and lenses.15 Additionally, measurement of the band edge allows us to estimate the particle size of CdS nanoparticles in the nanocomposite. The first exciton absorption onset can be measured from the spectra and the particle size can be calculated 78

PAGE 96

79 as before using the tight binding approximation45,89,90 or other models.39,43 The onset of the band edge was chosen as a measurable response of this design of experiments. Percent transmittance of the films is another response that extremely important from an application point of view. Transmittance >85% is essential for operation of optical limiters1,28,30 and lenses.15 Transmittance is adversely affected by aggregation of CdS particles. Therefore it is an indirect indicator of the stability of CdS in the matrix. Controlling aggregation of CdS particles is also required for improving efficiency of nanocomposite solar cells. It should be noted that surface roughness also contributes to the scattering of incident light and a loss in transparency. Therefore the function of transmittance as an indicator of particle aggregation may be limited. From the application point of view, the net transmittance is more important than what causes the scattering and consequent reduction in transparency. Percent transmittance is measured in the transparency window of the material. One can see from the absorption spectra of the nanocomposite films shown in Chapter 4 that the onset of the band edge is between 400 and 500 nm. By about 800 nm the absorption settles down to a low and almost constant value that is maintained throughout the transparency window of the material. Thus the % transmittance, measured at 800nm,was chosen as the other measured response of this design of experiments (DOE). Identification of Factors In Chapter 5, possible correlations between factors such as CdS concentration, and stabilizer concentration were shown to influence the band edge and transmittance of the nanocomposites were indicated. These factors will be included in the DOE. Other important factors such as temperature and molar mass of stabilizer may also be expected to influence band edge and transmittance. However stabilizer molar mass was not be

PAGE 97

80 considered for this DOE, due to a limited choice of stabilizers. The only suitable stabilizer among the ones available was Jeffamine D2000. Temperature on the other hand was considered as one of the factors in the DOE. Materials and Methods A full factorial design with two levels, three factors (23) and two replicates was planned. The important issue was to choose the low and high levels for each of the factors. Levels of Factors Sol synthesis temperature. The CdS sols were found to be stable even up to 100C. Normally at temperatures below 60C the decomposition of thiourea is very slow and the formation of colloids takes longer and particles formed are in the vicinity of 1 nm as the sols thereof are colorless and the precipitates are whitish. On the other hand, the upper limit is 110C, which is the boiling point of the solvent. Getting as high as the boiling point of the solvent increases solvent losses and flammability hazards. Hence the higher and lower values of 70C and 80C were chosen. CdS concentration. The CdS concentration range for transparent nanocomposites is 0 to 1.9% w/w CdS from Chapter 5. Here the lower and higher values chosen are approximately equidistant from the bounds of this range, viz. 0.8 and 1.2% w/w. Stabilizer concentration. Stable sols were obtained in the concentration range of Jeffamine D2000 stabilizer, 37 to 86%. w/w in Chapter 5. The lower and higher values are chosen such that they are approximately equidistant from the bounds of this range, viz. 45 and 75%. w/w.

PAGE 98

81 Design of Experiments The software used for the design of experiments was Stat-Ease, Design-Expert 6.0.5. A full factorial design model was fitted to the data. Table 6-1 Design Summary No. of experiments 16 Design Full factorial Levels 2 Factors 3 Replicates 2 Blocks None Table 6-2 Design Matrix Factors Responses Run # Temperature (C) Stabilizer (% w/w) CdS (% w/w) Band edge (nm) Transmittance (%) 1 80 45 0.8 485 75 2 80 75 1.2 479 74 3 80 45 1.2 479 66 4 70 75 1.2 468 77 5 80 75 0.8 488 90 6 70 45 0.8 477 68 7 70 75 0.8 462 77 8 80 45 0.8 483 91 9 70 45 0.8 474 86 10 70 75 0.8 465 86 11 70 45 1.2 480 51 12 80 45 1.2 475 89 13 70 75 1.2 465 83 14 70 45 1.2 480 47 15 80 75 0.8 488 93 16 80 75 1.2 481 83 Table 6-3. Factors affecting band edge and % transmission in CdS-epoxy nanocomposites Factor Factor Name Units Low level High level A Temperature C 70.00 80.00 B D2000 concentration Weight % 45 75 C CdS concentration Weight % 0.8 1.2

PAGE 99

82 Table 6-4. Measurement of Response variables viz. band edge and % transmittance Response Units Measurement technique Equipment Sample form Transmittance % UV-visible absorption spectrometry Shimadzu UV 2401 PC spectrometer solid films Band Edge nm UV-visible absorption spectrometry Shimadzu UV 2401 PC spectrometer solid films Synthesis Polyoxypropylene polyamines (Jeffamine D2000 and D230; Huntsman Chemical Company), N, N dimethyl formamide (DMF), thiourea and cadmium acetate dihydrate (Fisher Scientific), were used as received. In total, 16samples were made. Six ml of each of the samples were made. This amount was determined by the volume of the molds used to cast films. Two concentrations of CdS were taken. Two stock solutions of Cd-acetate and thiourea in DMF corresponding to the two concentrations of CdS were made such that 1 ml of each were added to each sample test tube. Appropriate amounts of stabilizer D2000 was added to the sample. Additional DMF was added to make the weight of each sample 6 gm. An oil bath with cooling cooper coils and manually controlled cold circulating water was placed on a magnetic stirrer-hotplate. Temperature was maintained to C of the desired temperature. In order to prepare the samples at 70C, the oil bath was heated to a temperature of 78C. A centrifuge tube containing the cadmium acetate solution in DMF and the D2000 stabilizer were added to it. The solution was allowed to remain in the bath for 5 minutes in which time the temperature of its contents rose and stabilized at 75C. Then 1 ml of

PAGE 100

83 thiourea solution containing the stoichiometric amount of thiourea was added. The temperature of the contents dropped to 70C. Cold water was circulated through the cooling coils of the bath to reduce the temperature of the bath to 70C where it was maintained for the rest of the reaction. The centrifuge tubes were removed at the end of 20 minutes after the addition of thiourea solution and cooled to room temperature by dipping in cold water. The reaction was assumed to have stopped at this time. A similar procedure was followed in case of samples that were prepared at 80C in which the bath was first heated to 80C in which the bath was first heated to 88C. The solution in the centrifuge tubes turned yellow on heating. This is indicative of the formation of CdS nanoparticles, which are yellow in color. 4 gm of the sol was mixed with a mixture of 2.4 gm Epon 828 and 0.4 gm D230, poured out in polyethylene lids used as molds and cured at 60C for 3 h followed by a post cure at 80C for 2 h. Characterization UV-visible spectroscopy (UV-vis). UV-Vis spectra were obtained on a Shimadzu UV 2401 PC spectrometer. The wavelength range chosen was 300 to 800 nm. The mode was absorbance and the sampling interval was set on ‘fast’. Special care was taken to ensure that each of the films was cleaned with ethanol and wipes to remove any dirt, or fingerprints that may affect the measurements. Solid films were mounted on a transmission stage with air reference. The onset of the absorption edge was measured by taking vertical projection of the point of intersection of two tangents on the absorption curve. The two tangents are taken on the flat region in the range of 450 to 550 nm and the inclined region in the range of 350 to 450 nm.

PAGE 101

84 The thickness of the films was measured and the average thickness was determined. The absorbance at 800 nm was normalized to the absorbance at 1 mm thickness of film. The % transmittance was calculated by using the relations: A = -log10(I/I0) (6.1) A = -log10(T) T = 10-A %T = 100*10-A (6.2) Where, A = absorbance of the films at 1 mm thickness; T = transmittance of the films; I = intensity of light transmitted through the film; I0 = intensity of light incident on the film Results and Discussions Response—Percent Transmittance A full factorial design model was fitted to the data. The model F-value of 3.64 implies the model is significant and that there is only a 4.53% chance that a model F-value this large could occur due to noise. Values less than 0.0500 in the “Probability > F” column (Table 6-5) indicate model terms are significant while values greater than 0.0500 indicate the model terms are not significant. In this case A, B and C are significant model terms. The predicted R2 of 0.0437 was not as close to the adjusted R2 of 0.5517 as one would have expected. This may indicate a large block effect or a possible problem with the model and/or data. Adequate precision measures the signal to noise ratio. A ratio greater than 4 is desirable. The ratio of 6.558 indicates an adequate signal. This model can be used to navigate the design space.

PAGE 102

85 Table 6-5. Analysis of variance [Partial sum of squares] for Selected Factorial Model Source Sum of squares Degrees of freedom Mean squares F value Probability > F Significance Model 2101.22 7 300.17 3.64 0.0453 Significant A 445.53 1 445.53 5.40 0.0486 Significant B 482.80 1 482.80 5.85 0.0419 Significant C 580.27 1 580.27 7.03 0.0292 Significant AB 180.15 1 180.15 2.18 0.1778 Not significant AC 35.50 1 35.50 0.43 0.5303 Not significant BC 85.20 1 85.20 1.03 0.3393 Not significant ABC 291.78 1 291.78 3.54 0.0969 not significant Pure Error 660.18 8 82.52 Cor Total 2761.39 15 Table 6-6. Model Fit Std. dev. 9.08 R2 0.7609 Mean 77.35 Adjusted R2 0.5517 C.V. 11.74 Predicted R2 0.0437 Precision 2640.71 Adequate precision 6.558 Table 6-7. Model Coefficients Factor Coefficient estimate Degrees of freedom Standard error 95% CI low 95% CI high VIF Intercept 77.35 1 2.27 72.11 82.58 A-Temperature 5.28 1 2.27 0.040 10.51 1.00 B-Stabilizer 5.49 1 2.27 0.26 10.73 1.00 C-CdS -6.02 1 2.27 -11.26 -0.79 1.00 AB -3.36 1 2.27 -8.59 1.88 1.00 AC 1.49 1 2.27 -3.75 6.73 1.00 BC 2.31 1 2.27 -2.93 7.54 1.00 ABC -4.27 1 2.27 -9.51 0.97 1.00 Final equation in terms of coded factors: % transmittance = +77.4 + 5.4(A) + 5.5(B) -6.0(C) (6.3) Final equation in terms of actual factors: % transmittance = + 5292.9 14.8(A) 1398.4(B) 1.08983(C) (6.4) Where,

PAGE 103

86 A = temperature in K; B = Stabilizer in gm; C = CdS content in gm The above results (Tables 6-5 and 6-7) show that temperature, stabilizer concentration and the CdS concentration have significant effects on the % transmittance while the interactions do not have significant effects. The % transmission shows an increase with increasing temperature (Figure 6-1) Figure 6-1. One-factor plot of % transmittance as a function of temperature in K. Transmittance also increases with stabilizer concentration. The probable cause is that at low concentrations of the stabilizer the particles tend to aggregate forming larger clusters, which scatter light. This leads to lower transmittance. At higher concentrations of the stabilizer the aggregation is probably reduced. This leads to a greater transmittance with higher concentration of the stabilizer (Figure 6-2). In case of CdS concentration, higher the concentration of CdS, greater is the tendency of the CdS particles formed to aggregate and form a precipitate. This leads to lower transmittance with increasing CdS content (Figure 6-3).

PAGE 104

87 Figure 6-2. One-factor plot of % transmittance as a function of stablizer weight in gm for a total sol weight of 6.0 gm. Figure 6-3. One-factor plot of % transmittance as a function of CdS in gm for a total sol weight of 6.0 gm. Response—Band Edge (onset) A full factorial design model was fitted to the data. The model F-value of 71.03 in (Table 6-8) implies the model is significant. There is only a 0.01% chance that a model F-value this large could occur due to noise. In this case A, B, C, AB and AC are significant

PAGE 105

88 model terms while BC and ABC are not significant. The lack of fit F-value of 0.25 implies the lack of fit is not significant relative to the pure error. There is a 78.11% chance that a lack of fit F-value, this large could occur due to noise. A non-significant lack of fit is desirable, as we want the model to fit. Table 6-8. Analysis of variance table [Partial sum of squares] for Selected Factorial Model Source Sum of squares DF Mean square F value Probability > F Significance Model 963.31 5 192.66 71.03 < 0.0001 Significant A 473.06 1 473.06 174.40 < 0.0001 Significant B 85.56 1 85.56 31.54 0.0002 Significant C 14.06 1 14.06 5.18 0.0460 Significant AB 264.06 1 264.06 97.35 < 0.0001 Significant AC 126.56 1 126.56 46.66 < 0.0001 Significant Residual 27.13 10 2.71 ---Lack of fit 1.63 2 0.81 0.25 0.7811 Not significant Pure error 25.50 8 3.19 Cor total 990.44 15 Table 6-9.Model Fit Standard deviation. 1.65 R2 0.9726 Mean 476.81 Adjusted R2 0.9589 C.V. 0.35 Predicted R2 0.9299 Precision 69.44 Adequate precision 24.416 Table 6.10. Model Coefficients Factor Coefficient estimate Degrees of freedom Standard error 95% CI low 95% CI high VIF Intercept 476.81 1 0.41 475.90 477.73 A-Temperature 5.44 1 0.41 4.52 6.35 1.00 B-Stabilizer -2.31 1 0.41 -3.23 -1.40 1.00 C-CdS -0.94 1 0.41 -1.85 -0.020 1.00 AB 4.06 1 0.41 3.15 4.98 1.00 AC -2.81 1 0.41 -3.73 -1.90 1.00 The predicted R2 of 0.9299 (Table 6-9) is in reasonable agreement with the adjusted R2 of 0.9589. Adequate precision measures the signal to noise ratio. The ratio of ratio

PAGE 106

89 greater than 4 is desirable. The ratio of 24.42 indicates an adequate signal. This model can be used to navigate the design space Final equation in terms of coded factors: Band onset = 476.81 + 5.44(A) 2.31(B) 0.94(C)+ 4.06(A)(B) 2.81(A)(C) (5.5) Final equation in terms of actual factors: Band onset = 264.55 + 0.65(A) 316.74(B) + 16234.38(C) + 0.90(A)(B) 46.88(A)(C) (5.6) Where, A = temperature in K; B = Stabilizer in gm; C = CdS content in gm The band edge (Figure 6-4, 6-5 and 6-6), increases with temperature, reduces with stabilizer concentration and reduces slightly with CdS concentration. Figure 6-4. Band edge as a function of temperature.

PAGE 107

90 Figure 6-5. Band edge as a function of stabilizer content (gm in 6 gm of sol). Figure 6-5. Band Edge as a function of CdS content (gm in 6 gm of sol). The interaction AB (Figure 6-7) shows that band edge increases steeply at high stabilizer concentration while it increases less steeply at a low stabilizer concentration. Thus increase in stabilizer content in the sols increases the sensitivity of the temperature dependence of the band edge.

PAGE 108

91 Figure 6-7. Band Edge as function of temperature at different stabilizer levels depicting the temperature-stabilizer interaction. Figure 6-8. Band Edge as function of temperature at different CdS levels depicting the temperature-CdS interaction. The interaction plot of temperature and CdS concentration (Figure 6-8) shows that temperature dependence of the band edge is stronger at lower CdS concentrations while it is less strong at higher CdS concentrations. This result is also important as it implies that

PAGE 109

92 while it may be desirable from the device point of view to increase CdS content of the films, it will necessarily be at the expense of the band edge tunability with temperature. Conclusions Band edge and % transmittance were found to be properties of relevance in devices such as optical limiters, nanocomposite solar cells and nanocomposite lenses. These properties were identified as measurable responses of a design of experiments. Temperature, stabilizer concentration and CdS concentration were identified as factors in the design. A 23 factorial design of experiments was conducted. The results showed that % transmittance has significant effects from temperature, stabilizer concentration and the CdS concentration while the interactions did not have a significant a effects on it. The % transmittance showed an increase with increasing temperature and stabilizer concentration while it undergoes a decrease with increasing CdS concentration. The greater transmittance with increased stabilizer content is probably the result of better aggregation control of the particles in the presence of the stabilizer. The decrease in transmittance with in increase in CdS content is probably due to the greater tendency of CdS to form aggregates in the available amount of stabilizer. The band edge increased with temperature, reduced with stabilizer concentration and reduced slightly with CdS concentration. The interaction of temperature and stabilizer concentration showed that band edge increased steeply at high stabilizer concentration while it increased less steeply at a low stabilizer concentration. The interaction between temperature and CdS concentration showed that temperature dependence of the band edge was stronger at lower CdS concentrations while it was less strong at higher CdS concentrations. This result also important as it implies that while it

PAGE 110

93 may be desirable from the device point of view to increase CdS content of the films, it will necessarily be at the expense of the band edge tunability with temperature.

PAGE 111

CHAPTER 7 TEMPERATURE DEPENDENCE OF BAND EDGE IN CDS-EPOXY NANOCOMPOSITES Introduction The design of experiments in Chapter 6 clearly shows that band edge of the material depends on all three factors considered above, namely, temperature, CdS concentration and D2000 concentration, and that there are significant interactions among the factors as well. However from a practical point of view, the processing temperature is the factor that is easiest to adjust and the band edge is also most sensitive to changes in temperature. Adjusting the concentration of stabilizer and CdS to control band edge may result is compromising mechanical properties and transparency. Therefore it would be useful to demonstrate the tuning of the band edge with temperature and also compare the results of this experiment with the empirical regression model obtained from the designed experiment in Chapter 6. An attempt to estimate the particle sizes from the band edge of the CdS nanoparticles will also be made and the limitations of the estimate will be discussed. The microstructure of the nanocomposites will be studied for different temperatures using transmission electron microscopy. Materials and Methods Synthesis Sols of CdS-epoxy containing were prepared and part of the sols were mixed with appropriate quantities of epoxy and D230 curing agent to obtain films containing about 1 94

PAGE 112

95 wt. % CdS after curing. The procedure used was identical with the typical procedure detailed in Chapter 4. Characterization UV-visible spectroscopy (UV-vis). UV-vis spectra were obtained on a Shimadzu UV 2401 PC spectrometer. The UV-visible-NIR spectra were obtained on a Perkin Elmer UV-vis-NIR spectrometer. Solid films were mounted on a transmission stage with air reference. Liquid samples were taken in a standard optical glass cell (Starna Cells Inc., prod. # 21-G-1, 1 cm path length) with an identical empty cell as reference. The band edge of the CdS absorption onset was measured as the x-coordinate of the point of intersection of the tangents drawn on the horizontal region between 500-600 nm and the inclined region between 400-500 nm of the curve. The % transmittance of the samples was measured at 800 nm. Transmission electron microscopy (TEM). TEM micrographs were obtained on a JEOL 2010F at 200 KV. The solid films were potted in a hard epoxy resin and cryo-ultramicrotomed and the sections mounted on carbon coated, Formvar (SPI ChemTM) supported copper grids. The sections obtained were rather thick (>200 nm) and uneven and hence getting good images was a challenge. Images could therefore be obtained only in regions where the section was thin enough to get an image. Results and Discussions Measurement of Band Edge The band edge of sols and films was measured (Table 7-1). The red-shift in the band edge of the sols and the films as a function of temperature is depicted in Figure.7-1A and B. As temperature increases, the band edge approaches the CdS bulk band edge, which is about 525 nm (not shown in Figure 7-1).

PAGE 113

96 00.511.522.533.544.55300400500600700800Wavelength nmAbsorbance DMF+D2000 65C 69C 74C 79C 83C A 01234300400500600700800Wavelength nmAbsorbance Pure epoxy 65C 69C 74C 79C 83C B Figure 7-1. UV-visible spectra for different sol synthesis temperatures. A) CdS sols in D2000 and DMF. B) CdS-epoxy films Note the red-shift in the band edge with increasing temperature.

PAGE 114

97 Table 7-1. Measured band edge values for sols and films Temperature C Band edge—sols nm Band edge—films nm Particle radius estimate nm 65 454 456 1.8 69 472 473 2.1 74 481 480 2.3 79 499 498 3 81 504 502 3.5 430440450460470480490500510606570758085Temperature CBand Edge nm1.522.533.54Particle radius nm Sol Band Edge Film Band Edge TB radius estimate Figure 7-2. Band edge of CdS sols and films vs. sol synthesis temperature. Also provided is an estimate of the particle size based on the tight binding calculation of CdS.90 A similar trend in the band edge as a function of sol synthesis temperature is observed in the sols as well the films (Figure 7-2). This indicates that there is no shift in the onset of the band edge of the materials when the sols are cured in an epoxy matrix. In other words the band edge is determined by the sol synthesis temperature and is independent of the curing temperature. Particle Size Estimation from Band Edge Using Theoretical Models According to the quantum size effect, it is possible to estimate the particle size if the band edge is known.

PAGE 115

98 However particle size estimation from the band edge of the CdS particles in the sols and films is possible only if the crystal structure of the particles is known. In fact the models discussed in Chapter 2, namely the effective mass approximation and the tight binding model, both assume that the CdS is present in a regular crystal lattice while calculating the electronic structure of the particles. The XRD and electron diffraction data from Chapter 4 failed to reveal any crystallinity in the particles. The particles are amorphous in nature. This brings us to a profound question as to whether or not quantum size effects are to be expected in amorphous semiconductor particles. There is undeniable evidence that the shoulder in the range 400-500 nm seen in the UV visible spectrum belongs to the particles. It is also true that the band edge of these clusters is blue shifted from that of pure bulk crystalline CdS that is around 525 nm. If we assume that this blue shift in and edge is not quantum size effect then there appears to be no explanation to the origins of this effect. Secondly, the stoichiometry of Cd:S in the CdS-epocy nanocomposites is uncertain. EDS data in Chapter 4 indicates a 70:30 Cd:S stoichiometry. However EDS data at CdS loadings as low as 1 to 2% is subject to an error as high as a factor of 2. In this case the CdS particles may as well be stoichiometric. However some caution is needed in applying theoretical models designed for stoichiometric particles. The, the direct application of any of the above models to determine the particle size is therefore not possible. Using the tight binding model, it is possible to calculate the electronic structure (and hence correlate band edge and particle size) of non-stoichiometric clusters.45

PAGE 116

99 However whether it is possible to correlate band edge and particle size for amorphous clusters remains a question to be answered. Leaving this question unanswered, we could proceed to obtaining a rough estimate the particle sizes. The estimated particle sizes of are in the range of 1.7 to 3.5 nm (Table 7-1) for band edges between 454 to 504 nm (Figure 7-2) as shown in Figure.5-7. It should also be noted that the distribution of the particle sizes is quite broad as observed in TEM images in Chapter 4. The theoretical models mentioned above agree with experimental values when the particle size distribution is narrow. Therefore the particle size estimates obtained (Table 7-1) pertain to the mean size and are not expected to be very accurate. Comparison of Experimental Data with Regression Model The values of band edge vs. temperature as predicted by the model and the experimental data (Table 7-2) were obtained. The important point to be noted is that while the model was based on a design of experiment done with temperature levels 70 and 80C respectively, the temperature dependence study was carried out over a wider range of 65 to 83C. The model was extrapolated to this wider range and the predicted profile was obtained. Table 7-2. Comparison of experimental data with model Temperature C Band edge (model) nm Band edge (experiment) nm 65 464 456 69 469 473 74 476 480 79 483 498 83 488 502 A comparison of this extrapolated prediction with the experimental data shows a good qualitative similarity between the model and the data (Figure 7-3). The monotonic

PAGE 117

100 increase of the band edge with increasing temperature was correctly followed by the model. In general, the slope of the experimental data was somewhat greater than the model. 440450460470480490500510606570758085Temperature CBand Edge Model Experiment Figure 7-3. Comparison between the temperature dependence of band edge as predicted by the regression model and experimental data The model was reasonably close to the data in the range of 75 to 80C. On the other hand, the model deviates from the data at both ends of the range viz. 65 and 83C. This is probably the result of a nonlinear relationship between the band edge and the temperature. Better models with 3 levels instead of 2 may be able to model the behavior of the system more accurately. However such models would need more number of experiments to be performed. An alternative strategy would be to fit different 2 level models in different temperature ranges. Causes of Temperature Dependence of Band Edge The solvent DMF was shown to complex the cadmium acetate in Chapter 4. The stabilizer showed no such complexation or interaction with the cadmium acetate. CdS

PAGE 118

101 particles formed on heating a mixture of thiourea and cadmium acetate in DMF are probably still complexed by DMF and thus are prevented from growing further. As the sulfur radical is evolved from thiourea its local concentration increases at certain locations where, it will attack the cadmium ions, which are complexed by DMF. It has to displace the DMF in order to do so. In doing so however, the local concentration of DMF increases and it becomes more and more difficult to displace any more DMF and finally the growth of the particle stops. In pure DMF the viscosity of the medium is relatively low and so the diffusion of thiourea and DMF to and from the vicinity of the growing particle is easier. So there is relatively lower buildup of concentrations of either of them and their concentrations are more or less averaged throughout the solution. Also the concentration of DMF is high (nearly 100%) so the crowding of DMF around growing particles is high. This leads to small partilces which cannot grow to large sizes. D2000 has a much higher viscosity as compared to DMF. So mixtures of D2000 and DMF are generally more viscous than DMF alone. In presence of D2000 stabilizer, the diffusion of thiourea and DMF to and from the vicinity of the growing particle is slower due to a higher viscosity. There is a greater likelihood of a buildup of either of these moieties. This makes the particle size more sensitive to the rate of evolution of sulfur or in short the temperature. At lower temperatures, the rate of evolution of sulfur is low and even a small amount of crowding of DMF molecules around the particle can stop the growth of the particles. Thus small particles are formed at lower temperatures. At high temperatures the rate of evolution of sulfur is high. This leads to a greater buildup of sulfur concentration

PAGE 119

102 locally, which in turn requires more crowding of the DMF molecules around it to halt the growth. This leads to larger particle sizes (and more red-shifted band edges) at higher temperatures. This mechanism also explains the interaction between the temperature and stabilizer concentration in which the temperature dependence is stronger at a higher stabilizer concentration. Transmission Electron Microscopy of Films Containing Sols Synthesized at Different Temperatures Transmission electron microscopy of the samples synthesized at different temperatures could not measure different mean particle sizes or their distribution due to difficulties in preparing the samples. The microtomed sections were too uneven and in general too thick to obtain good images despite all efforts to obtain good sections. However particles of dimensions on the order of a few nanometers were detected in all samples tested so far not restricted to this experiment. As far as this particular experiment is concerned, all the samples tested too showed small particles in the range of few nanometers (Figure 7-4C, D and F). Larger clusters of these particles were seen in all samples too. Almost all the clusters observed, from tens to hundreds of nanometers were not monolithic particles but were in fact clusters of much smaller particles (Figure 7-4E). The fact that the large clusters are not monolithic particles but aggregations of smaller particles is significant. It follows from this that the electronic properties, band edge in particular, of even the aggregates are going to be same as the small particles. The size and the number of the large clusters would however affect the transmittance of the films. Larger and more numerous the aggregates, more would be the scattering from them and lower would be the transmittance.

PAGE 120

103 A B C D Figure 7-4. TEM micrographs of CdS-epoxy films in DMF+D2000 on JEOL 3010F, 200 KV samples synthesized at A) 69C; B, C and D) 65C E, F) 83C. Although the mean particle size could not be measured and a trend in the particle size or its distribution as a function of the sol synthesis temperature could not be detected, an interesting difference in the vicinity of large clusters was detected. Samples synthesized at 65 and 69C show regions with a low density of cadmium, which appear like a mist. Note that cadmium in the system is the only source of a detectable contrast in this largely organic system.

PAGE 121

104 E F Figure 7-4. Continued Although the mean particle size could not be measured and a trend in the particle size or its distribution as a function of the sol synthesis temperature could not be detected, an interesting difference in the vicinity of large clusters was detected. Samples synthesized at 65 and 69C show regions with a low density of cadmium, which appear like a mist. Note that cadmium in the system is the only source of a detectable contrast in this largely organic system. These regions, owing to their dark color are cadmium-rich regions. These regions span several microns in the vicinity of large aggregates (Figure 7-4A). Even at higher magnifications and no graininess or particles are visible in them (Figure 7-4B). In the samples synthesized at 83C however, these low cadmium density regions were totally absent (Figure 7-4E and F). Numerous small clusters (sizes < 20 nm) were detected in the vicinity of the larger aggregates. If we recall, from the discussion of the design of experiments in Chapter 6, higher temperatures were associated with a greater transmittance while lower temperature are associated with lower transmittance. At low temperature the rate of degradation of

PAGE 122

105 thiourea is low. So lower amount of sulfur is released and only a small fraction of the cadmium acetate forms particles with it. The rest of the cadmium acetate remains unreacted for want of more sulfur. When the sols are mixed with epoxy and curing agent, and cured at an elevated temperature, some of this cadmium acetate finds some sulfur that is released from thiourea due to the heating during curing. However at this stage, the viscosity of the resin is probably very high, as the resin has probably reached its gel point by then. Therefore it cannot form any clusters. This cadmium sulfide that is formed is visible as the amorphous ‘misty’ regions in between the larger clusters. These regions can be very large in size, on the order of several microns and contribute to scattering. Thus sols synthesized at lower temperatures yield films with a lower transmittance. On the other hand for films containing sols synthesized at higher temperatures, the decomposition of thiourea is more rapid. A greater fraction of thiourea has already given up the sulfur contained in it and most of the particles that could have been formed are already formed. On curing this sol with an epoxy there is very little sulfur released at this stage and hence the ‘misty’ regions are not seen in samples which has sols synthesized at a higher temperature. Conclusions The temperature dependence of band edge was studied and it revealed that the band edge (wavelength) or the band gap (energy) were determined by the sol synthesis conditions and was independent of the curing conditions. The temperature dependence band edge of the nanoparticulate sols was thought to be the result of nanoparticle size. The temperature determined the particle size and the particle size determined the band edge. The particle size was hypothesized to be a kinetic

PAGE 123

106 phenomenon. It was controlled by two competing factors namely, the rate of evolution of sulfur from thiourea which is a function of temperature and the diffusion of sulfur and DMF molecules which is a function of the viscosity of the medium. At lower temperatures the latter dominates and small particles are formed while at higher temperatures, the former dominates and larger particles are formed. The experimental data obtained was compared to the regression model obtained in the design of experiments. The model qualitatively follows the data well in that it shows a monotonic increase in the band edge with increase in temperature. The model deviates from experimental data towards the extremities of the temperature range studied. The causes of this deviation are the possible nonlinearity of the temperature dependence of band edge curve and the extrapolation of the model beyond its designed range. A more sophisticated model would perform better and is highly recommended. Accurate particle size estimation of CdS nanoparticles could not be made owing to the absence of theoretical models for amorphous and non-stoichiometric particles. A rough estimate was obtained by using a model calculated for stoichiometric CdS nanocrystals. Variations of the tight binding calculation allow for calculation of band edge in non-stoichiometric nanocrytallites and merit further investigation. The causes of temperature dependence of the band edge were investigated and an explanation was proposed. The increased viscosity due to the presence of the stabilizer was responsible for the phenomenon. Mean particle size and size distribution could not be calculated using TEM due to poor samples. However particles with sizes less than 5 nm were detected in all samples. Also larger clusters from tens to hundreds of nanometers were detected. Regions of low

PAGE 124

107 cadmium density spanning several microns were detected in films made from sols synthesized at 65 and 69C. These regions are responsible for the reduced transparency in sols synthesized at low temperatures.

PAGE 125

CHAPTER 8 HIGH CDS CONCENTRATION NANOCOMPOSITES Introduction Increasing the concentration of CdS in CdS-epoxy nanocomposite films is of great importance as the linear and relevant nonlinear optical properties of the films are dependent on the concentration of CdS. The concentration of CdS in thick films of CdS-epoxy nanocomposites with thickness on the order of 1 mm could go only as high as 1.9% after which precipitation of the nanoparticles caused opacity. The polyoxypropylene diamines (Jeffamine) act as stabilizers and prevent the aggregation of CdS nanoparticles when precipitated from solution of Cd-acetate and thiourea in DMF. The cadmium ions and the newly formed CdS particles are complexed by DMF. However the particles tend to aggregate probably under the influence of secondary bonding forces. The Jeffamine stabilizers such as D2000 and D5000, due to their oligomeric size keep the particles apart and screen these forces and prevent aggregation. The stabilizers, due to their high viscosity also slow down all diffusion of ions and particles during synthesis and thus provide particle size control with temperature. However the presence of the stabilizer automatically places a limit on the CdS content of the system. Higher the content of stabilizer, lower is the content of CdS. Now in order to improve the concentration of the CdS in the system one can reduce the size of the stabilizer molecule allowing CdS particles to come closer to each other and increase the CdS content of the films. This means that the molar mass of the molecule has to be reduced. In that case however, there is a likelihood of the CdS 108

PAGE 126

109 particles coming too close together increasing the chances of aggregation and precipitation. The simultaneous reduction the viscosity of the reaction mixture in case of a lower molar mass stabilizer facilitates the diffusion of the particles and hence makes the system more susceptible to aggregation and precipitation. If it were possible to reduce the molar mass of the stabilizer and yet, somehow prevent aggregation of the particles, then it would be possible to increase the CdS loading without compromising on the transparency. Materials and Methods Synthesis One of the strategies for making transparent films with greater concentration of CdS is to make thin films. A dilute solution of a lower molar mass stabilizer such as D230 or D400 (molar mass = 230, 400 respectively), solvent, DMF and the epoxy resin, Epon 828 is poured out on a glass slide or spin coat on a substrate and allowed to dry partially. The stabilizer also acts as curing agent for epoxy. Then it is put in an oven at about 80C to nucleate the particles and simultaneously cure the resin. The remaining solvent also evaporates off at this stage in a thin film. The curing of the resin and nucleation of the particles takes place simultaneously. Due to the reduced thickness, the evaporation of the solvent while curing is faster and the gel point is reached sooner. This should limit the growth of the aggregates as the matrix gels before the particles can grow substantially. If the aggregates are limited to size in the range of tens of nanometers, the films will be transparent. Another strategy is to eliminate the amine stabilizer entirely. An amine terminated, linear prepolymer of the epoxy and the low molar mass curing agent such as D400 (taken

PAGE 127

110 in molar ratio 1:2 respectively) is first prepared with DMF as solvent. The Cd-acetate and thiourea are dissolved in it. This is mixed with additional amount of epoxy to fully cure the system in a thin or thick film. The particle formation and the curing of the resin take place simultaneously. A slight variation of this process is to form the particles in the prepolymer solution first and then mix it with epoxy and cure. The highlight of this method is that both thick and thin films with a high concentration of CdS can be made. Table 8-1. Materials used for the synthesis of CdS-epoxy nanocomposites. Chemical name Function Chemical formula Source Abbreviation Cadmium acetate, dihydrate Source of cadmium Cd(CH3COO)2 .2H2O Fisher scientific Cd-acetate Thiourea Source of sulfur (NH2)2CS Fisher scientific Thiourea N, N dimethyl formamide Solvent (CH3)2NCHO Fisher scientific DMF Polyoxyethylene diamine (Jeffamine D400, D230) Stabilizer and curing agent Huntsman chemical company D230, D400 Diglycidylether of bisphenol A (epon 828) Resin Shell chemical company Epon 828 Amine stabilized, in situ method (ASI) Cd-acetate and thiourea were mixed with DMF in a test tube and a clear solution was obtained on mixing them on a vortex mixer. The epoxy resin and D400 were added to it one after another and mixed thoroughly till a clear solution was obtained. A few drops of the solution were poured on glass slide and the solution was allowed to spread. The slides were dried partially under vacuum for 2 h. The slides were moved to an oven at 80C for 1 h to cure the films. This is a very simple method and yields highly transparent films with particles having a reasonably narrow distribution.

PAGE 128

111 Table 8-2. Amine stabilized, in situ method for films with 17 wt % CdS Cd-acetate gm Thiourea gm DMF gm Epon 828 gm D400 gm 0.100 0.028 5.400 0.097 0.090 Microwave processing. In a variation of the same method, some samples were cured using microwave heating instead of the conventional air-circulating oven. One of the slides was put in a Panasonic (1200 W) microwave oven at power level 5 (600 W) for four 1-minute runs. A block of SiC, which is a highly microwave active material was placed about 10-15 cm away from the slide. At the end of the first minute itself, there was discoloration and crazing observed in some regions of the film, indicating degradation. The same increased after every 1 minute run. The samples definitely show good potential for a very rapid curing using microwave radiation. One of the problems with this process was the uneven heating in the microwave oven, which led to degradation in certain regions while allowed a good cure in others. The microwave-cured samples show a very sharp transition of the CdS exciton absorption indicating a narrow particle size distribution. Prepolymer stabilized, two step method (PST) Half the required amount of epoxy was added to the entire required amount of curing agent (D230) and mixed thoroughly with 1 ml of DMF on a vortex. The mixture was left in an oven at ~80C for 1/2 h for the epoxy to form a prepolymer. The solution was cooled by holding test tube under flowing water at room temperature. We assume that an amine-terminated prepolymer was formed and there was increase in molar mass compared to the reactants. Thiourea and cadmium acetate were mixed with the remaining amount of DMF on a vortex mixer. Then remaining epoxy was added to it and mixed on a vortex mixer till a clear solution was obtained. This solution was mixed with the

PAGE 129

112 prepolymer solution. A clear, colorless solution was obtained. It was left in a test tube and kept in the oven at 80C for an hour. A clear yellow solution was obtained. This indicated that it was a stable sol of CdS nanoparticles. In case of thick films this solution was poured into polyethylene lids used as molds to form thick films. The molds were put in an oven at 80 C for 3 h. In case of thin films, a few drops were poured on glass slides and allowed to spread out to form thin films. The solution for the thin films is considerably more dilute as compared to the solution for thick films. The slides were put in an oven at 80 C for 1h. Table 8-3. Prepolymer stabilized, two step method for thick films with 21 %wt CdS Cd-acetate gm Thiourea gm DMF gm Epon 828 gm D230 gm 0.500 0.143 2.077 0.472 0.172 Table 8-4. Prepolymer stabilized, two step method for thin films with 21 wt% CdS Cd-acetate gm Thiourea gm DMF gm Epon 828 gm D230 gm 0.100 0.028 5.400 0.095 0.042 A great advantage of this procedure is the possibility of controlling the band edge by varying the synthesis temperature of the sols as demonstrated in Chapter 7. Prepolymer stabilized, in situ method (PSI) Half the amount of epoxy required, was added to the entire required amount of curing agent (D230) and mixed thoroughly with 1 ml of DMF on a vortex. The mixture was left in an oven at ~80 C for 1/2 h for the epoxy to form a prepolymer. The solution was cooled by holding test tube under flowing water at room temperature. We assume that an amine-terminated prepolymer was formed and there was increase in molar mass compared to the reactants. Thiourea and cadmium acetate were mixed with the remaining amount of DMF and added to the remaining half of epoxy resin. The two were

PAGE 130

113 mixed on a vortex mixer till a clear solution was obtained. This solution was mixed with the prepolymer solution. A clear, colorless solution was obtained. In case of thick films this solution was poured into polyethylene lids used as molds to form thick films. The molds were put in an oven at 80 C for 3 h. In case of thin films, a few drops were poured on glass slides and allowed to spread out to form thin films. The solution for the thin films is considerably more dilute as compared to the solution for thick films. The slides were put in an oven at 80 C for 1h. Table 8-5. Prepolymer stabilized, in situ process for thick films with 21 wt% CdS Cd-acetate gm Thiourea gm DMF gm Epon 828 gm D230 gm 0.500 0.143 2.077 0.472 0.172 Table 8-6. Prepolymer stabilized, in situ process for thin films with 21% wt CdS Cd-acetate gm Thiourea gm DMF gm Epon 828 gm D230 gm 0.100 0.028 5.400 0.095 0.042 Characterization UV-visible spectroscopy (UV-vis). UV-Vis spectra were obtained on a Shimadzu UV 2401 PC spectrometer. Solid films were mounted on a transmission stage with air reference. The band edge of the CdS absorption onset was measured as the x-coordinate of the point of intersection of the tangents drawn on the horizontal region between 500-600 nm and the inclined region between 400-500 nm of the curve. The % transmittance of the samples was measured at 800 nm. X-ray diffraction (XRD). The x-ray diffraction data was obtained on a JEOL 3720 diffractometer. Solid XRD samples were prepared by cutting solid films of nanocomposites into approximately 1cm2 pieces and pasting them onto glass slides using double-sided tape. The 2 range of 10 to 80 with a step size of 0.01.

PAGE 131

114 Results and Discussions Optical Images The optical clarity of the prepolymer stabilized (PSI and PST) thick films with CdS concentrations as high as 21% (as calculated from the amounts of Cd-acetate and thiourea) are represented pictorially in two ways (Figure 8-1A, B, C and D). The left column (Figure 8-1A, C) shows images taken with the camera with the films covering the camera lens. This is intended to roughly simulate their configuration in an optical limiting application. The right column (Figure 8-1B and D) is an images of the films placed on black printed white paper. A D2000 stabilized film, (85% transmittance and 0.74% CdS w/w), made with the procedure given in Chapter 4, is also shown in a similar manner for comparison (Figure 8-1E and F). The PST films with 21% CdS depicted in Figure 8-1A show a transparency of 52% at 800 nm. The greenish yellow hue of the films is a result of the shift in the band edge of the polymer owing the presence of the CdS absorption band in the vicinity of the band edge of the polymer. This is discussed in greater detail in Chapter 4. The distortion of the image in Figure 8-1A is due to warping of the films during processing. The haziness in the films indicates that there is some amount of aggregation in the films. The PSI films with 21% CdS are colorless and show a high transmittance of 85% transparent comparable to the D2000 stabilized films of a much lower CdS concentration (0.74% CdS). There is no noticeable haziness due to aggregation in these films. The superiority in optical clarity of the D2000 stabilized films over the high concentration films is obvious however that is understandable as it contains over 20 times less amount of CdS.

PAGE 132

115 A B C D E F Figure 8-1. Optical photographs on a Sony DCR-TRV20 camera. Left column shows images taken by covering the lens of the camera with different films while the right column shows the photographs of the films on printed text. A and B) 0.63 mm thick, 21% CdS, prepolymer stabilized, two step (PST), CdS-epoxy film. C and D) 0.85 mm thick, 21% CdS, prepolymer stabilized, in situ (PSI), CdS-epoxy film. E and F) 1.11 mm thick, 0.75% CdS, D2000 stabilized CdS-epoxy film UV-Visible Spectroscopy Thin films. The 10 m ASI film containing 17%CdS, transmits 86% of 800 nm energy (Figure 8-2A). The spectrum shows a sharp transition corresponding to the CdS exciton absorbance at about 365 nm, which is indicative of particle sizes on the order of 1 nm. This is concurrent with the observation of the design of experiments in Chapter 6, which shows a negative dependence of the band edge with CdS concentration. The concentration of this ASI film is substantially higher (17%) than the films that were made for the design of experiments (0.8 to 1.2%). Therefore one would expect a

PAGE 133

116 considerable blue shift in the band edge of the particles in the high concentration films. Although the absorption of the particles is very close to the band edge of the polymer it can still be distinguished from the polymer band edge owing to the relatively narrow particle size distribution leading to a sharp transition. However in general and the particle size reduces with an increasing CdS concentration, and the band edge shifts closer to the band edge of the polymer (~325 nm). At a point when the band edge of the CdS particles is very close to the band edge of the polymer, it becomes more and more difficult to distinguish it from the polymer band edge. The PSI film with 21% CdS particles shows a high transmittance of 85% with no shoulder corresponding to the CdS exciton transition (Figure 8-2B). It is likely that the particles formed are very small in size. As a result heir band edge is so blue shifted that it is indistinguishable from the band edge of the polymer. However there is also a possibility that particles may not be formed. It is likely that owing to the curing the formation being forced to take place at the same time, the matrix may have reached its gel point before the particles could form. With the matrix past its gel point, diffusion of the Cd and sulfur moieties would be so slow that the formation of particles may not be possible. However the PST film made by the two-step process does show a red shift in the band edge of the polymer compared to the PSI film. This indicates that CdS particles are formed in the films made by the two-step method. The onset of the transition is 415 nm, which implies that the particles formed are on the order of 1.3 nm. This film shows a transmittance of 82% at 800 nm.

PAGE 134

117 00.511.522.5800700600500400300Wavelength (nm)Absorbance A -0.50.51.52.53.54.5800750700650600550500450400350300Wavelength (nm)Absorbance Prepolymer stabilizedIn situ Prepolymer stabilizedTwo step B Figure 6-2. UV-visible absorption spectra of A) 10 m thick, 17%CdS amine stabilized (ASI) CdS epoxy films. B) 10 m thick, 21% prepolymer stabilized (PSI and PST) CdS-epoxy films C) 0.65 and 0.85 mm thick, 21% prepolymer stabilized (PSI and PST) CdS-epoxy films. Thick films. The thick films made by the prepolymer method (PSI and PST) show a similar trend (Figure 6-2C) as in case of their thin film counter parts(Figure 6-2B). The PSI film does not show any absorption feature belonging to CdS. Either the particles formed are very small in size (<0.8 nm) that their band edge is indistinguishable from the

PAGE 135

118 polymer band edge or the particles may not be have formed. The PSI film is highly transparent and shows 85% transmittance at 800 nm. The PST film shows a red shift in the band edge, which indicates the formation of the particles. However the breadth of the particle size distribution is very high in this film since a sharp transition is not obtained but a considerably broad shoulder is obtained instead. Also the transmittance of this film is only 52% at 800 nm, which is considerably lower than the PSI film. 00.511.522.53800750700650600550500450400350300Wavelength (nm)Absorbance Prepolymer stabilizedIn situ Prepolymer stabilizedTwo step C Figure 6-2. Continued. X-Ray Diffraction The x-ray diffraction of plot of PSI with 21% CdS films made by the prepolymer process shown in Figure 8-3 was unable reveal any crystalline structure. The material is largely amorphous as noted in case of the D2000 stabilized films in Chapter 4. Although the presence of a very small fraction of crystalline nanoparticles cannot be ruled out, it is undetectable above the noise. The XRD peaks of nanoparticles (nanocrystallites) are typically broadened as compared to their bulk counterparts. Even if the peaks were

PAGE 136

119 broadened they would still be detectable above the noise if they formed a substantial fraction of the material. Clearly, such is not he case here and therefore crystalline nanoparticles, even if they existed, would form an undetectably small fraction in the material. 501001502002501018263442505866742counts Figure 8-3. X-ray diffraction plot of 0.85 mm thick prepolymer stabilized in situ (PSI) film with 21% CdS Conclusions The main goal of this chapter was to make CdS-epoxy films with high concentrations in the range of 20% w/w and high transparency of about 85%. Films with 85% transmittance—both thin and thick films—were made. Two strategies were employed to stabilize this high concentration of CdS in the films. One of them involved the use of a lower molar mass polymeric amine stabilizer in a dilute solution of the epoxy resin and cadmium acetate and thiourea to make thin films. This process was also modified in that it was cured using microwave radiation. The cured films degraded in some placed due to uneven heat distribution within the microwave oven. The amine stabilized films were highly transparent, contained 17% CdS and had a relatively sharp

PAGE 137

120 transition. The estimated particle size was about 1 nm and there was no control over particle size or the band edge. In another novel method the amine stabilizer was replaced by an amine terminated prepolymer of the epoxy resin. This method had two variations. One of was an in situ particle formation while another was a two step method in which the particles were formed and the resin was cured in separate steps. The former gave thin and thick films with high transmittance of 86 and 85% respectively but the band edge did not show the familiar absorption of the CdS particles. The particles, if formed, would be of a particle size estimated to be less than 0.8 nm. The latter method gave thin and thick films with lower transmittance of 82 and 52% respectively. The particle band edge could be seen clearly in the thin films while it could be seen as a very broad shoulder in the thick films. X-ray diffraction of the prepolymer stabilized films showed an amorphous structure. Although the goals of this chapter are fulfilled on the whole, more work needs to be done primarily to verify the presence of CdS particles in the films.

PAGE 138

CHAPTER 9 CONCLUSIONS Solid state, transparent CdS-polymer nanocomposites are an interesting material with several applications including optical limiting, solar cells, and lenses. This work was primarily inspired by optical limiting applications of the nanocomposites. The important requirements for this application are high transparency, tunability of band edge and high CdS concentration in thick films. The purpose of this dissertation was to develop a nanocomposite system that satisfies the above requirements. The potential of a CdS-polysulfone system to yield transparent nanocomposite films was investigated. A variety of techniques involving thiol based capping agents for particles were used. Functionalization of the polymer was carried out with a view to achieving a covalent bonding with the functionalized (capped) particles. However CdS particles precipitated on synthesis. The oxidation of the thiol based capping agents in the presence of air was thought to be the result of destabilization and precipitation of the CdS particles. When the synthesis was carried out in an inert atmosphere, transparent sols of CdS in solvents such as DMF and acetone were obtained. Sols of CdS were also obtained in polymer solutions of polysulfone or sulfonated polysulfone in dimethyl formamide or tetrahydrofuran. However transparent films could not be obtained. During the drying of the solvent cast films, phase segregation occurred and the dried films were opaque. This was attributed to the oxidation of the thiols with residual and leaking air in the dry box filled with argon. Another probable cause of the phase separation could be the presence of small amounts of water in the system, which is required to dissolve sodium sulfide, 121

PAGE 139

122 which is the source of sulfur in most cases. However even though the water from the system was eliminated by replacing sodium sulfide with thiourea, which is soluble in organic solvents, the films obtained were still opaque. In an arrested precipitation of CdS particles films were made using thiourea and cadmium acetate in a polysulfone matrix in situ. These films were translucent yellow. This method is an arrested precipitation of CdS in a medium of high viscosity. This method gave the best results. However even these films were translucent and not highly transparent. UV-visible spectra of the films show a shoulder at the foot of the band edge of the polymer. This shoulder is the exciton absorbance of the CdS particles in it. Scanning electron microscopy of the translucent films reveals two distinct phases, namely the polymer-rich phase and the cadmium-rich phase. The energy dispersive characteristic x-ray analysis confirms the presence of the two phases. These phases are large enough in size to cause a large amount of scattering of light and make the films translucent. Transmission electron microscopy of CdS-PSF sols shows small particles on the order of 10 nm. However it also shows that the particles present are predominantly in the form of aggregates. The particles aggregate in polymer solution. Films of CdS synthesized in situ in the polysulfone (UCdS-PSF) showed very small particles in the TEM micrographs. The aggregation was very low in these films. Thus although the objective of making transparent thick films was not achieved completely there are some very important conclusions that can be made from this attempt. Opacity in the nanocomposites is the result of particles aggregating to form large clusters and in extreme cases forming a precipitate. The thiol based capping agents are unsuitable due to their oxidative instability in addition to their bad odor and toxicity. The

PAGE 140

123 use of thiourea to replace sodium sulfide as source of sulfur and arresting aggregation by making the viscosity of the matrix high, are worth further investigation. In a novel synthesis, highly transparent (>85%) and 1 mm thick solid nanocomposite films of CdS nanoparticles in an epoxy matrix was produced from air stable sols. The UV-visible NIR spectrum of the materials shows the high transparency from 500 nm – 2300nm. The band edge of the material shows a broad exciton feature, which indicates a broad particle size distribution. The oligomeric stabilizers such as Jeffamine D2000 were key in making the air stable sols and thick films. FTIR (infrared spectroscopy) shows that that the cadmium in its ionic form, complexes with the DMF while CdS particles are probably not complexed by DMF. No interaction with the stabilizer D2000 could be detected with FTIR. It was hence inferred that the oligomeric stabilizer contained in the sols prevented aggregation of the CdS particles by screening the secondary bonding forces leading to aggregation and increasing viscosity, which slows the diffusion of the particles. Transmission electron microscopy (TEM) of the sols reveals that majority of the particles are in the few nanometer range while there are medium and large sized clusters. TEM of the films also reveals small particles in the few nanometer range but more medium and large sized aggregates were found. EDS (energy dispersive characteristic x-ray) data reveals that the Cd:S ratio of the particles is about 70:30. Electron diffraction and x-ray diffraction show that the particles are amorphous. Two photon absorption spectra of the films were obtained. The nonlinear absorption coefficient of the CdS nanoparticles in the films was 30 times lower in magnitude than that of bulk CdS. One of the major causes of this could be the lack of crystallinity of the particles. The thermal stability was evaluated using

PAGE 141

124 thermogravimetric analysis. The films were found to be stable upto 300C before which there was no detectable loss in weight from degradation or out gassing of volatiles. It was observed that the band edge the stability and transparency of the sols and films were controlled by a number of factors such as temperature, CdS concentration, stabilizer type and concentration, and order of addition of ingredients. Some of the factors were screened based on practical considerations and with the end-use in mind. A 23 factorial design of experiments was conducted with temperature, stabilizer concentration and CdS concentration as the factors. The response variables measured were %transmittance and band edge. The results showed that % transmission has significant effects from temperature, stabilizer concentration and the CdS concentration while the interactions do not have a significant a effect on it. The % transmission shows an increase with increasing temperature and stabilizer concentration while it undergoes a decrease with increasing CdS concentration. The increased transparency at higher temperatures is caused by lower amounts of unreacted thiourea and Cd acetate in the sols, which are capable of forming clouds or mists of CdS during the curing of films. The greater transmittance with increased stabilizer content is the result of better aggregation control of the particles in the presence of the stabilizer. The decrease in transmittance with in increase in CdS content is due to the greater tendency of CdS to form aggregates in the available amount of stabilizer. . The band edge increases with temperature, reduces with stabilizer concentration and reduces slightly with CdS concentration. The interaction of temperature and stabilizer concentration shows that band edge increases steeply at high stabilizer concentration

PAGE 142

125 while it increases less steeply at a low stabilizer concentration. The interaction between temperature and CdS concentration shows that temperature dependence of the band edge is stronger at lower CdS concentrations while it is less strong at higher CdS concentrations. This result is also important as it implies that while it may be desirable from the device point of view to increase CdS content of the films, it will necessarily be at the expense of the band edge tunability with temperature. The temperature dependence of band edge was studied in greater detail and it reveals that the band edge is determined by the sol synthesis conditions and is independent of the curing conditions. The causes of temperature dependence of the band edge were investigated and an explanation proposed. The size of the particles formed is determined by two competing factors namely, the rate of evolution of sulfur from thiourea which is a function of temperature and the diffusion of sulfur and DMF molecules which is a function of the viscosity of the medium. At lower temperatures the latter dominates and small particles are formed while at higher temperatures, the former dominates and larger particles are formed. The experimental data obtained was compared to the regression model obtained in the design of experiments. The model qualitatively follows the data well in that it shows a monotonic increase in the band edge with increase in temperature. The slope pf the curve estimated by the model is somewhat less than the experiment and the model deviates from experimental data towards the extremities of the temperature range studied. The causes of this deviation are the possible nonlinearity of the temperature dependence of band edge curve and the extrapolation of the model beyond its designed range. A more sophisticated model would perform better and is highly recommended. An accurate particle size estimation of CdS nanoparticles could not

PAGE 143

126 be made owing to the absence of theoretical models for amorphous and non-stoichiometric particles. Rough estimates were obtained by using a model calculated for stoichiometric CdS nanocrystals. Variations of the tight binding calculation allow for calculation of band edge in non-stoichiometric nanocrytallites and merit further investigation. Mean particle size and size distribution could not be calculated using TEM due to poor samples. However particles with sizes less than 5 nm were detected in all samples. Also larger clusters from tens to hundreds of nanometers were detected. Misty regions spanning several microns were detected in films made from sols synthesized at 65 and 69C. These regions are responsible for the reduced transparency in sols synthesized at low temperatures. In an effort to increase the concentration of CdS in the films, CdS-epoxy films with high concentrations in the range of 20% w/w and high transparency of about 85% were made. Films with 85% transmittance—both thin and thick films—were made. Two strategies were employed to stabilize this high concentration of CdS in the films. One of them involved the use of a lower molecular weight polymeric amine stabilizer in a dilute solution of the epoxy resin and cadmium acetate and thiourea to make thin films. This process was also modified in that it was cured using microwave radiation. The films cured well but degraded in some placed due to uneven heat distribution within the microwave oven. These films were highly transparent, contained 17% CdS and had a relatively sharp transition. The estimated particle size was about 1 nm and there was no control over particle size or the band edge. In another novel method the amine stabilizer was replaced by an amine terminated prepolymer of the epoxy resin. This method had two variations. One of was an in situ

PAGE 144

127 particle formation while another was a two-step method in which the particles were formed and the resin was cured in separate steps. The former gave thin and thick films with high transmittance of 86 and 85% respectively but the band edge did not show the familiar absorption of the CdS particles. The particles, if formed, would be of a particle size estimated to be less than 0.8 nm. The latter method gave thin and thick films with lower transmittance of 82 and 52% respectively. The particle band edge could be seen clearly in the thin films while it could be seen as a very broad shoulder in the thick films. The most important contribution of this research is the demonstration of the ability to make solid state, transparent CdS nanocomposites. Although thin films or precipitates of such composites have been made, this is the first time that transparent, thick films of CdS-epoxy nanocomposites have been synthesized. The band edge tunability with temperature is a very important feature with regard to the application of these films. The same system, with some modifications is capable of making thick, high concentration films. Thus the CdS-epoxy nanocomposite system represents a unique system with potential applications not only in optical limiting but a variety of optical applications including coatings and lenses. As part of future work it will be worthwhile to investigate into making similar nanocomposites with crystalline CdS particles. Since non-stoichiometry of the CdS particles could be a major contributor to the non-crystallinity, particles with a more than stoichiometric amount of thiourea could be made, so as to tilt the stoichiometry in favor of sulfur. A systematic effort needs to be made to investigate quantum size effects in amorphous metal chalcogenide clusters such as CdS. Although there is a preponderance

PAGE 145

128 of such studies in case of nanocrystals, there is none for amorphous systems. This will involve accurate measurements of cluster sizes and distributions of CdS clusters using HRTEM and correlating the band edge-particle size relations with appropriate quantum chemical models. More work needs to be done, to verify the presence of CdS particles in the high concentration films. Similar nanocomposites using conducting polymer matrices should be investigated for use in solar cells.

PAGE 146

APPENDIX A MEASUREMENT OF BAND EDGE 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 6 Run 9 Onset = 477 nmOnset = 474 nm A 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 7 Run 10Onset = 462 nmOnset = 465 nm B Figure A-1 UV visible spectra of solid film samples from design of experiments in Chapter 5. A to H show the measured onset of band edge for replicates 129

PAGE 147

130 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 11 Run 14 Onset = 480 nmOnset = 480 nm C 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 14 Run 13 Onset = 468 nmOnset = 465 nm D Figure A-1 Continued

PAGE 148

131 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 1 Run 8 Onset = 485 nmOnset = 483 nm E 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 5 Run 15 Onset = 488 nmOnset = 488 nm F Figure A-1 Continued

PAGE 149

132 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 3 Run 12 Onset = 479 nmOnset = 475 nm G 00.511.522.533.54800750700650600550500450400350300WavelengthAbsorbance Run 2 Run 16 Onset = 479 nmOnset = 481 nm H Figure A-1 Continued

PAGE 150

APPENDIX B SENSOR PROTECTION AND NONLINEAR OPTICS Sensor Protection Lasers usage in day-to-day life as well as in defense applications has increased the possibility of lasers being used as weapons to destroy optical systems or eyes. The off-the-shelf availability of powerful commercial lasers makes it possible for even a layman to use them for destructive purposes with minimal knowledge of the physics of lasers. In any optical system, the most sensitive components, which are usually, the sensors or the detector elements, need to be protected from laser damage. "Sensor protection is defined as anything that can be used to protect optical systems and human eyes from debilitating laser effects."28 Sensor protection is required for three major effects that lasers have on optical systems, officially referred to as Level I, Level II and Level III.28 Level I. Optical Augmentation. Lasers can be used to retro-reflect the beam off an optical systems so as to reveal its location. Level II. Jamming or dazzle: It is the use of laser beam to temporarily disable an optical system by excessive illumination. It does not permanently disable the sensor or other components of the optical system. Level III. Laser Damage: It is the use of laser light to permanently damage or disable an optical system. In this chapter, we shall be concerned mainly with Level II and Level III protection. Following is a summary of the current fielded sensor protection devices.28 Sacrificial Systems. These are devices that fail below the damage threshold of the sensor and turn opaque on failure thus preventing any damage to the sensor. They offer only a one-time protection and need to be replaced each time damage 133

PAGE 151

134 occurs. They also render the optical system disabled till the device is replaced owing to the irreversible opacity on failure. They only prevent costlier equipment from being damaged. Fixed-line Filters. These are optical filters that are used to selectively filter out certain frequencies from the incident light. They generally have a low transmission 10-20%, similar to sunglasses and they lead to color distortion, which may interfere in the working of the sensor. This type of device has specific relevance to protective eyewear. However these are ineffective against 'frequency-agile' lasers that can rapidly tune off of the predicted frequencies. Shutter systems. Electro-optical shutters can be used to respond within 10s. This makes them effective only to protect against CW lasers and not short-pulsed lasers. The short pulses of destructive intensities will pass through the device and destroy the sensor even before the shutter activates and cuts out the radiation. Neutral Density Filters. These are simple optical filters that absorb over a broad range of wavelengths. They provide moderate protection. They will lead to jamming instead of damage by a potentially destructive laser. However, they drastically reduce contrast and overall transmission. They are acceptable only in cases where the low transmission is acceptable and where jamming is acceptable instead of damage. Requirements of an Optical Limiter In general, a protective device such as an optical limiter must have the least impact on the performance of the sensor or the optical system. Following are some of the specific requirements. Broadband operation. It should be effective over the entire wavelength band of the sensor. There are mainly four bands: Visible VIS (400-700 nm), near infrared NIR (0.75-3 m), middle infrared MIR (3-30 m), far infrared FIR (30-1000 m) High transmittance. It must have a high transmission in the ‘on’ position, in all bands and low transmission in the ‘off’ position. An acceptable high transmission in the ‘on’ state of any device in the VIS is 40 % while it is about 85-90 % for NIR-FIR. These values are useful only as general and reasonable goals and are by no means rigid. Maximum permissible exposure (MPE). The maximum permissible exposure for the human eye is the maximum safe level of total inter-ocular energy (TIE) that can be allowed into the eye. This value is 0.2 J (fluence of 0.5J/cm2) for pulses less than 17 ms for a dark-adapted eye. In general the MPE is determined by the damage threshold of the most sensitive components of the optical system. The device should start working at an energy level of equal to or less than the

PAGE 152

135 MPE (by an appropriate factor of safety) and should remain so (in the low transmission mode) till the failure of the device.28 Dynamic range. The dynamic range of any device is the ratio of the input energy at failure to the input energy at 'switch on', which is described as: Rd = Log10(Energymax. incid)/(Eon) (B.1) Optical density (OD). The optical density is the ratio of the energy transmitted to the maximum energy incident in terms of orders of magnitude of protection. The OD is represented by the following: ODsystem = -1*log10(ET/Emax. incid.)28 (B.2) Where: ET = is the transmitted energy Emax. incid = is the maximum energy incident before failure Dynamic range. The dynamic range of a shutter involving the human eye is generally required to be 4 Thus the OD of such a device also needs to be at least 4. Temporal bandwidth of protection. Laser sources come in a wide variety of pulse widths; ultra short pulse (10-100 fs) short pulses 0.1-40 ns, long pulse 40 ns-1ms, continuous wave (CW) >1ms. The protecting device should take into consideration attack from any of these laser sources. Environmental stability The stability of the device in the working conditions should be considered. Also, the toxicity or other hazards associated with the materials used in the device should be addressed. F-number. The device should be able to work under low F number conditions so as to give a wide field of view. The F number is defined as39: F number = f/D = 2/NA = 1/2tan (B-3) Where, f = focal length; D = diameter of lens; NA = numerical aperture Size. The size, weight and complexity of the device is also important as it determines the easy by which it can be incorporated into existing optical system which is not necessarily designed to accommodate any extra space to house the protecting device.

PAGE 153

136 Tendency for jamming. Optical protection systems using NL scatter and NL refraction rely on system apertures that end up blocking the incident light when it is scattered or refracted from the protecting device. This light however stays in the system and leads to a sharp reduction in contrast. In extreme cases, it may also lead to the jamming of the system. Current Research in Optical Limiting None of the currently available materials provide a satisfactory amount of protection.28 Active limiting devices are not feasible as the pulse width of the destructive lasers moves to the ultra fast regime (30 to 100 femto-seconds) Any device requiring external activation is bound to be slow and can at the most achieve a sub-microsecond response time. Hence it is quite clear any such devices with a sub nano second response time are bound to be passive devices or devices that do not require any external triggering but use the incident beam itself as the triggering signal.1,36,117 Such passive devices are intensity dependent. They utilize self-induced nonlinear optical effects to switch on at above certain intensities.1,100,117 This means that the incident beam itself provides the signal to the limiter to switch from the on to the off position. No external triggering or activation is needed and hence these devices are fast acting. Merely dealing with the material properties may not be sufficient to produce an effective optical limiter but it is also an engineering issue.36 Devices use one or more of the effects like nonlinear absorption,118-121 nonlinear refractive index change122-127 and nonlinear scattering29,100,128-131 to achieve optical limiting. Following is a brief review of nonlinear optics, and the various nonlinear optical processes used for passive optical limiting.102

PAGE 154

137 Nonlinear Optics The polarization P is the material response of a material to an electric field E. At low intensities, the relation between the two is given by a Taylor series expansion102 of the electric field E: P = (1) E + (2) E2 + (3) E3 + (B.4) (1) = first order susceptibility. Susceptibility is a tensor and a complex quantity. The real part is related to the refractive index while the imaginary part is related to the absorption coefficient. The first order susceptibility is responsible for linear optics. The magnitude of the susceptibility reduces as the order of the susceptibility increases. (2) = second order susceptibility that is responsible for second order nonlinear optics. Second order nonlinear effects are present in materials that are non-centrosymmetric. These are materials, which have molecules having one or more axes symmetry passing through the centroid. The even orders of the susceptibility vanish when the material is centrosymmetric. When present, this effects is far higher in magnitude than the third order nonlinear effects arising out of the (3). Photorefractive effect is a second order nonlinear effect. (3) = This susceptibility is much smaller in magnitude (|(3)| ~10-12-10-15 esu) than the second order susceptibility (|(2)| ~10-7-10-9 esu) (3) leads to the appearance of a nonlinear refractive index n2’ n2’ Re{ (3)1111(-; , -, )} There is also a corresponding nonlinear absorption coefficient 2 2 Im{ (3)1111(-; , -, )} Thus the net refractive index (n) and the absorption coefficient () are given by the following equation: n = n0 + n2’ I (B.5)

PAGE 155

138 = 0 + 2 I (B.6) Where, n0 and 0 are the linear refractive index and the linear absorption respectively Physical Processes Leading to the Nonlinear Absorption or Index of Refraction Nonlinear Absorption A number of nonlinear absorption processes that exist, viz. two-photon absorption, excited state absorption, reverse saturable absorption.29,102 Excited state absorption requires unoccupied states above the first excited state. The excited photon absorbs another photon and gets excited to the higher excited state. Since this appears as an increased absorption above the linear absorption and hence is called nonlinear. No electronic transitions occur when the incident energy of a laser is lower than the band gap energy. This is true at low intensities of incident radiation, which is the domain of linear optics. However at very high intensities, multi-photon processes are prevalent which involve the electrons in the valence band to simultaneously absorb two or more photons of a lower-than-band-gap-energy and jump to the higher state. This phenomenon is probable only at high intensities and can be safely neglected at lower intensities that are commonly used in all linear optical equipment. The presence of these multi-photon processes leads to a multi-photon (in particular, a two photon) absorption coefficient. According to Kramers-Kronig relations in linear optics,39,132-134 whenever there is a change in the absorption coefficient of a material undergoes a change, it is always accompanied with a corresponding change in the refractive index. Mathematically this is given by the following: =b]d-)]/[([[ )(c/ )n(220 (B.7)

PAGE 156

139 =b]d-)]/[(n[[ )(c/ )(220 (B.8) Where: = angular frequency of electric field = angular frequency of polarization n() b = linear refractive index as a function of angular frequency of polarization () = absorption coefficient = 2 ()/c Where: () is the imaginary part of the complex refractive index n*(). Thus, )( ))/(2( i )n( )( i )n( )(*n2D t= + = n + = (B.9) Where, D() = dielectric constant These relations are also applicable to the nonlinear absorption coefficient and the nonlinear refractive index. Whenever there is a change in the absorption coefficient due to any of the above nonlinear effects, there is a corresponding change in the nonlinear refractive index according to the Kramers-Kronig relations. The nonlinear refractive index arising out of the two-photon absorption coefficient is normally negative at energies below the band-gap energy but more than half the band-gap energy (two-photon resonance) and it decreases in magnitude as the energy reduces. Then there is a crossover point at energy somewhat higher than the two-photon energy, where the refractive index is zero.39 As energy reduces further, the refractive index rises steeply to a positive value and peaks just before the two-photon energy. Then it falls at energies below the two photon

PAGE 157

140 energy approaching the n = 0 line assymtodally.39 The overall absorption coefficient is the sum of the linear and nonlinear parts. Thus there is an increase in the absorption coefficient in the nonlinear regime when intensities are high enough to trigger the nonlinear response provided that the energy of the incident beam is equal to or lower than the two-photon resonance energy. Nonlinear Refraction The nonlinear refractive index arises out of the third order polarizability or second hyperpolarizability of a material. In general there is a change in the refractive index as a function of optical electric field or the intensity. Refractive index n = n0 + n (B.10) where n0 = linear refractive index n = change in refractive index In Kerr like materials, the change in refractive index change: n = 2n2fEf2 or n = n2` (I) since I = 2t0n0cfEf2 (B.11) where fEf= Electric field intensity n2 = nonlinear index coefficient n2` = nonlinear index intensity coefficient I = Intensity In non-Kerr like materials the index change is not a linear relation of intensity but some function of it. Hence in general, n = f(I) (B.12) A number of processes may be responsible for causing the phenomenon of a nonlinear refractive index. The chief among them are discussed below.102

PAGE 158

141 Electronic Polarization Electromagnetic radiation incident on polarizable atoms, leads to a polarization of the electron cloud. If the energy of the radiation is less than the absorption edge, then band-filling effects tend to dominate. The carriers are excited into the conduction band from where they rapidly thermalize and occupy the states at the bottom of the conduction band.102 In case the pulse width of the radiation is very short compared to the electron hole recombination time the states at the bottom of the conduction band are filled. This results in the bleaching of the absorption at lower energies. Thus the carriers have to be excited to a energy higher than the lowest states in the conduction band. This excitation creates a blue shift in the band-gap energy. The change in absorption coefficient leads to a corresponding change in the nonlinear refractive index through the Kramers Kronig relations. The NL index coefficient n2 is negative at less than band gap energy while it is positive at greater than band gap energy. Band Renormalization: If the intensity is much higher with energy in the vicinity of the band gap then electron and hole plasmas are formed. The plasmas shield the Coulomb attraction between the electrons and holes, which creates a red shift in the band gap. For less than band gap energies, the absorption increases and the nonlinear index coefficient is positive. In case of excitons in direct band semiconductors, there is a screening of the exciton Coulomb interaction, which reduces the binding energy of the exciton and bleaches the exciton.102 For less than band gap energy, n2 is negative while at higher than band gap energy, n2 is positive.

PAGE 159

142 Raman Induced Kerr Effect A Raman induced Kerr Effect is a phenomenon observed in a two-beam experiment. In linear Raman scattering a highly intense pulse from a laser scatters into a weak Stokes beam, which is downshifted in frequency from the laser frequency by a value corresponding to a Raman active mode. In case of stimulated Raman scattering, the Stokes beam is used to trigger the scattering of the intense laser or pump beam and enhance scattering 10-100 fold. In other words the weak beam draws gain from the pump beam. This phenomenon is also related to the third order susceptibility.102 Electrostriction An electrostrictive force is produced on the atoms or molecules when a spatially varying electric field or a field with inhomogeneous intensity is incident on a medium. The force is proportional to the gradient of the modulus of the square of the field. This produces a tendency for translation of dipoles to the regions of high intensity, leading to a modulation of the local refractive index. This is a relatively slow process of the order of nanoseconds. Molecular Reorientation Molecules having an anisotropic linear polarizability tensor appear optically isotropic when they are disordered in the bulk. However when an electric field is applied they experience a torque that tends to align them along the direction of the field. When an optical wave, polarized along the direction of the field is incident on the material it perceives a higher index of refraction in presence of the aligning field than without it. Thermal Effects A nonlinear refractive index may also be an effect of change in physical properties of a material due to the presence of an optical field. At longer wavelengths, due to

PAGE 160

143 coupling with the multi phonon modes (higher overtones of phonon modes), there may be absorption of the incident radiation leading to heating. This heating is localized in the region where the beam is incident. It causes a shift in the phonon mode leading to a shift in the absorption band and consequently a change in the refractive index. Physically, thermal vibrations cause an expansion of the unit cell leading to a reduction in the band gap. The change in refractive index due to this effect39 is given by the following equation: n2 = (T/ C)(dn/dT) (B.13) Where = characteristic diffusion time; T = thermal expansion coefficient; = density of material; C = specific heat; dn/dT = temperature dependence of the refractive index. Normally this coefficient is negative for gasses, liquids and most solids hence thermal nonlinearity in them will be negative. However in some materials this can be positive hence their thermal nonlinearity will be positive. Optical Kerr Effect It is a change in the refractive index due to change in phase of the incident light. Nonlinear Scattering: Nonlinear scattering is employed in two systems basically. The first is a colloidal dispersion of fullerenes,121,135,136 carbon black100 or carbon nanotubes29,101,137. These materials, being colloidal in size have fairly good transmission in a very broad range of the electromagnetic spectrum. However, they also have a very high absorption coefficient. At very high intensities, the materials reach extremely high temperatures locally causing vaporization of the solvent to form micro bubbles57 or send out plumes of plasma.100 This causes scattering of the incident light.

PAGE 161

144 The particles are essentially destroyed and in order that the damage to the limiter is 'self-healing' this type of limiters are liquid state. The colloid has to be replaced periodically to avoid any perceptible deterioration in the performance with time. The second system based on nonlinear scattering uses a two-phase system such as dye-doped micro-particles in a sol-gel matrix129 or a polymer blend128. The matrix and the particles have different nonlinear refractive indices. However they are index matched at the operational wavelength. Thus the composite appears transparent. At high intensities, the nonlinear effects become significant and owing to the different nonlinear indices of the two materials, a refractive index change results. This causes scattering of the light passing through the medium thereby limiting the light reaching the detector. Practically it is difficult to isolate the contribution of each of the above processes as a number of processes invariably go on simultaneously. However there are some processes that are more likely or have a greater effect, at certain specific wavelengths or a certain range of wavelengths. Summary The chapter discussed the effects of lasers, the need for sensor protection and the current devices used for sensor protection. For reliable protection from ultra-fast lasers passive optical limiters are required. These materials use nonlinear optical processes to achieve optical limiting. Their response time is on the order of pico-second to femtoseconds. The fundamentals of nonlinear optics were discussed and brief descriptions of different nonlinear processes were provided.

PAGE 162

LIST OF REFERENCES (1) Soileau, M. J.; Guha, S.; Stryland, E. V. Proceedings of the SPIEThe International Society for Optical Engineering 1985, 540, 2-8. (2) Xu, S.; Wang, H. L.; Malko, A.; Kohlman, R.; Smilowitz, L.; Klimov, V.; McBranch, D. E. 1999, 3798, 76-84. (3) Song, Y.; Zhang, C.; Wang, Y.; Fang, G.; Duan, C.; Liu, S.; Xin, X.; Ye, H. Optics Communications 1999, 168, 131-134. (4) Rochon, P.; Bissonnette, D.; Nattansohn, A.; Xie, S. Applied Optics 1993 December 10, 32, 7277-7279. (5) Bandyopadhyay, S., The University of California, Riverside, http://ndl.ee.ucr.edu/c1ab.pdf 2002, Last accessed July 27, 2004. (6) Wherrett, B. S. Synth. Met. (Switzerland), Synthetic Metals, Properties of Organic Materials of the 1995 E-MRS Spring Conference 1996, 76, 3-9. (7) Lee, M. J.; Judge, C. P.; Wright, S. W. Solid-State Electronics 2000, 44, 1431-1434. (8) Hirai, T.; Miyamoto, M.; Watanabe, T.; Shiojiri, S.; Komasawa, I. Journal of Chemical Engineering of Japan 1998, 31, 1003-1006. (9) Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002, 295, 2425-2427. (10) Huynh, W. U.; Peng, X.; Alivisatos, A. P. Advanced Materials 1999, 11, 923. (11) Flora, K.; Brennan, J. D. The Analyst 1999, 124, 1455-1462. (12) Godovsky, D. Y. Advances in Polymer Science 2000, 153, 163-205. (13) Bach, U.; Lupo, D.; Comte, P.; Moser, J. E.; Weiss, F.; Salbeck, J.; Spreitzer, H.; Graetzel, M. Nature 1998, 395, 583-585. (14) Fendler, J. H. Chemisty of Materials 2001, 13, 3196-3210. (15) Border, J.; McGovern, M. R., Patent, USA, 6,552,111, 2003, Eastman Kodak Company (Rochester, NY) (16) Gurin, V. S.; Artemyev, M. V. Journal of Crystal Growth 1994, 138, 993-997. 145

PAGE 163

146 (17) Berman, A.; Charych, D. Advanced Materials 1999, 11, 296-300. (18) Kamat, P. V.; Wijngaarden, M. D. L. v.; Hotchandani, S. Israel Journal of Chemistry 1993, 33, 47-51. (19) Foglia, S.; Suber, L.; Righini, M. Colloids and Surfaces A: Physiochemical and Engineering Aspects 2001, 177, 3-12. (20) Yu, S. H.; Yang, J.; Han, Z.-H.; Zhou, Y.; Yang, R. Y.; Qian, Y.-T.; Zhang, Y. H. Journal of Materials Chemistry 1999, 9, 1283-1287. (21) Pal, U.; Loaiza-Gonzalez, G.; Bautista-Hernandez, A.; Vazquez-Cuchillo, O. Superficies y Vacio 2000, 11, 40-43. (22) Gacoin, T.; Malier, L.; Boilot, J.-P. Journal of Materials Chemistry 1997, 7, 859-860. (23) Schlamp, M. C.; Peng, X.; Alivisatos, A. P. Journal of Applied Physics 1997, 82, 5837-5842. (24) Ammermann, A.; Bhler, A.; Dirr, S.; Johannes, H.-H.; Kowalsky, W. AE International Journal of Electronics and Communications 1996, 50, 327-333. (25) Krihak, M. K.; Shahriari, M. R. Electronics Letters 1996, 32, 240-242. (26) Dagani, R. Chemical & Engineering News 2000, 8, 9. (27) Hirai, T.; Suzuki, K.; Komasawa, I. Journal of Colloid and Interface Science 2001, 244, 262-265. (28) Miller, M. J.; Mott, A. G.; Ketchel, B. P. Proceedings of the SPIEThe International Society for Optical Engineering. 1998, 3472, 24-29. (29) Lee, W. T.; Bogess, T. F. Progress in Quantum Electronics 1993, 17, 299-338. (30) Powell, R. C.; Taheri, B.; Munoz, A. F.; Reeves, R. J. Proceedings of the SPIEThe International Society for Optical Engineering 1992, 1692, 21-36. (31) Miller, T. M.; Zhao, L.; Brennan, A. B. Journal of Applied Polymer Science 1998, 68. (32) Dong, F.; Koudoumas, E.; Couris, S.; Shen, Y.; Qiu, L.; Xingfa Fu. Journal of Applied Physics 1997 May 15, 81, 7073-7075. (33) Said, A. A.; Wei, T. H.; Salvo, R. D.; Wang, Z.; Sheik-Bahae, M.; Hagan, D. J.; Stryland, E. W. V. Proceedings of the SPIEThe International Society for Optical Engineering 1992, 1692, 37-43.

PAGE 164

147 (34) Fischer, G. L.; Boyd, R. W.; Moore, T. R.; Sipe, J. E. Optics Letters 1996 October 15, 21, 1643-1645. (35) Soileau, M. J.; Williams, W. E.; Stryland, E. W. V. IEEE Journal of Quantum Electronics 1983 April 4, QE-19, 731-735. (36) Wood, G. L.; III, W. W. C.; Miller, M. J.; Salamo, G. J.; Sharp, E. J. Proceedings of the SPIEThe International Society for Optical Engineering. 1989, 1105, 154-180. (37) Sheik-Bahae, M.; Said, A. A.; Wei, T. H.; Wu, Y. Y.; Hagan, D. J.; Soileau, M. J.; Van Stryland, E. W. Proceedings of the SPIE The International Society for Optical Engineering, 1148, 41-51. (38) Cha, M., Doctoral Dissertation "Third Order Nonlinear Spectroscopy of Organic Materials", 1994, Physics, University of Central Florida, M. PhD Thesis 1994. (39) Simmons, J. H.; Potter, K. S. Optical Materials; 2000 ed.; Academic Press: San Diego, CA, 2000. (40) Shaheen, S. E.; Brabec, C. J.; Sariciftci, N. S. Applied Physics Letters 2001, 78, 841-843. (41) Sandnes, B.; Rekstad, J. Solar Energy 2002, 72, 63-73. (42) Nozik, A. J.; Williams, F.; Nenadovic, M. T.; Rajh, T.; Micic, O. I. Journal of Physical Chemistry 1985, 89, 397-399. (43) Potter, B. G. J.; Simmons, J. H. Physical Review B 1988, 37, 10838-10845. (44) Wang, Y.; Herron, N. Physical Review B 1990, 42, 7253-7255. (45) Gaponenko, S. V. Optical Properties of Semiconductor Nanocrystals; 1998 ed.; Cambridge University Press:, 1998. (46) Li, Y.; Wan, J.; Gu, Z. Materials Science and Engineering A 2000, 286, 106-109. (47) Xu, J. M.; Dmitri, R.; Terry, B.; Martin, M. The Journal of Physical Chemistry 1996, 100, 14037-14047. (48) Ouellette, J. The Industrial Physicist 2000, December 2000, 26-29. (49) Shenton, W.; Pum, D.; Sleytr, U. B.; Mann, S. Nature 1997, 389, 585-587. (50) Ferrry, I.; Mikrajuddin; Kikuo, O. Nano Letters 2001, 1, 231-234. (51) Liu, T.; Burger, C.; Chu, B. Progress in Polymer Science 2003, 28, 5-26. (52) Liu, J.; Alvarez, J.; Kaifer, A. E. Advanced Materials 2000, 12, 1381-1383.

PAGE 165

148 (53) Soten, I.; Ozin, G. A. Current Opinion in Colloid and Interface Science 1999, 4, 325-337. (54) Schafer, A. H.; Siedel, C.; Chi, L.; Fuchs, H. Advanced Materials 1998, 10, 839-842. (55) Pum, D.; Sleytr, U. B. Trends in Biotechnology (TIBTECH) 1999, 17, 8-12. (56) Rosa, M. E. D.; Weijie Su, D. K.; Brant, M. C.; McLean, D. G. Proceedings of the SPIEThe International Society for Optical Engineering. 1997, 3146, 134-141. (57) Said, A. A.; Xia, T.; Hagan, D. J.; Wajsgrus, A.; Yang, S.; Kovsh, D.; Decker, M. A.; Khodja, S.; Stryalnd, E. W. V. Proceedings of the SPIEThe International Society for Optical Engineering 1996, 2853, 159-169. (58) Kickelbick, G. Progress in Polymer Science 2003, 28, 83-114. (59) Sooklal, K.; Hanus, L. H.; Ploehn, H. J.; Murphy, C. J. Advanced Materials 1998, 10, 1083-1087. (60) Sooklal, K.; Hanus, L. H.; Ploehn, H. J.; Murphy, C. J. Langmuir 2000, 16, 2621-2626. (61) Jelinski, L., WTEC Panel Report on Nanostructure Science and Technology R&D Status and Trends in Nanoparticles, Nanostructured Materials, and Nanodevices Chapter 7 1999, http://itri.loyola.edu/nano/07_01.htm, Last accessed 2004, July 29. (62) Fendler, J. H. Current Opinion in Solid State and Materials Science 1997, 2, 356-359. (63) Esteves, A. C. C.; Trindade, T. Current Opinion in Solid State and Materials Science 2002, 6, 347-353. (64) Qi, L.; Colfen, H.; Antonietti, M. Nano Letters 2001, 1, 61-65. (65) Hirai, T.; Watanabe, T.; Komasawa, I. The Journal of Physical Chemistry B 2000, 104, 8962-8966. (66) Santra, S.; Tapec, R.; Theodoropoulou, N.; Dobson, J.; Hebard, A.; Tan, W. Langmuir 2001, 17, 2900-2906. (67) Shiojiri, S.; Hirai, T.; Komasawa, I. Chemical Communications 1998, 1998, 1439-1440. (68) Hirai, T.; Okubo, H.; Komasawa, I. Journal of Colloid and Interface Science 2001, 235, 358-364. (69) Hirai, T.; Namba, M.; Komasawa, I. Journal of Colloid and Interface Science 2002, 252, 89-92.

PAGE 166

149 (70) Hirai, T.; Okubo, H.; Komasawa, I. Journal of Materials Chemistry 2000, 10, 2592-2596. (71) Shiojiri, S.; Hirai, T.; Komasawa, I. Journal of Chemical Engineering of Japan 1998, 31, 142-146. (72) Shiojiri, S.; Hirai, T.; Komasawa, I. Journal of Chemical Engineering of Japan 1997, 30, 137-145. (73) Hirai, T.; Shiojiri, S.; Komasawa, I. Journal of Chemical Engineering of Japan 1994, 27, 590-597. (74) Hirai, T.; Bando, Y.; Komasawa, I. Journal of Physical Chemistry B 2002, 106, 8967-8970. (75) Vossmeyer, T.; Reck, G.; Katsikas, L.; Haupt, E. T. K.; Schulz, B.; Weller, H. Science 1995, 267, 1476-1479. (76) Vossmeyer, T.; Giersig, L. K. M.; Popovic, I. G.; Diesner, K.; Chemsieddine, A.; Eychmuller, A.; Weller, H. Journal of Physical Chemistry 1994, 98, 7665-7673. (77) Rockenberger, J.; Troger, L.; Kornowski, A.; Vossmeyer, T.; Eychmuller, A.; Feldhaus, J.; Weller, H. Journal of Physical Chemistry B 1997, 101, 2691-2701. (78) Yu, S.-H.; Wu, Y.-S.; Yang, J.; Han, Z.-H.; Xie, Y.; Qian, Y.-T.; Liu, X.-M. Chemisty of Materials 1998, 10, 2309-2312. (79) Lazell, M.; O'Brien, P. Journal of Materials Chemistry 1999, 9, 1381-1382. (80) Barnett, S. A.; Blake, A. J.; Champness, N. R.; Nicolson, J. E. B.; Wilson, C. Journal of the Chemical Society, Dalton Transactions 2001. (81) Rajamathi, M.; Seshadri, R. Current Opinion in Solid State and Materials Science 2002, 6, 337-345. (82) Farmer, S.; Patten, T. E. Chemisty of Materials 2001, 13, 3920-3926. (83) Trindade, T.; O'Brien, P.; Pickett, N. L. Chemistry of Materials 2001, 13, 3843-3858. (84) Yu, Z.; Degang, F.; Xin, W.; Juzeng, L.; Zuhong, L. Colloids and Surfaces A: Physiochemical and Engineering Aspects 2001, 181, 145-149. (85) Schwerzel, R. E.; Spahr, K. B.; Kurmer, J.; Wood, V. E.; Jenkins, J. A. Journal of Physical Chemistry A 1998, 102, 5622-5626. (86) Omi, S.; Hata, C.; Uchida, K.; Kaneko, S.; Tanji, H.; Asahara, Y. Nonlinear Optics 1995, 13, 127-134.

PAGE 167

150 (87) Inoue, H.; Urquhart, R. S.; Nagamura, T.; Grieser, F.; Sagaguchi, H.; Furlong, D. N. Colloids and Surfaces A: Physiochemical and Engineering Aspects 1997, 126, 197-208. (88) Farmer, S. C.; Patten, T. E. Chemisty of Materials 2001, 13, 3920-3926. (89) Lippens, P. E.; Lanoo, M. Physical Review B 1989, 39, 10935-10939. (90) Sapra, S.; Sarma, D. D. Physical Review B 2004, 69, 125304. (91) Hirai, T.; Miyamoto, M.; Komasawa, I. Journal of Materials Chemistry 1999, 9, 1217-1219. (92) Hirai, T.; Okubo, H.; Komasawa, I. Journal of Physical Chemistry B 1999, 103, 4228-4230. (93) Hirai, T.; Komasawa, I. Journal of Materials Chemistry 2000, 10, 2234-2235. (94) Damle, C.; Gole, A.; Sastry, M. Journal of Materials Chemistry 2000, 10, 1389-1393. (95) Qiao, Z.; Xi, Y.; Li, G.; Zhu, Y.; Qian, Y. Journal of Materials Science 2000, 35, 285. (96) Hirai, T.; Saito, T.; Komasawa, I. Journal of Physical Chemistry 2001, 105, 9711-9714. (97) Shiojiri, S.; Miyamoto, M.; Hirai, T.; Komasawa, I. Journal of Chemical Engineering of Japan 1998, 31, 425-433. (98) Trollsas, M.; Hawker, C. J.; Hedrick, J. L.; Carrot, G.; Hillborn, J. Macromolecules 1998, 31, 5960-5963. (99) Yu, S.-H.; Yoshimura, M. Langmuir 2001, 17, 1700-1707. (100) Mansour, K.; Stryland, E. W. V.; Soileau, M. J. Proceedings of the SPIEThe International Society for Optical Engineering 1989, 1105, 91-102. (101) Vivien, L.; Anglarret, E.; Riehl, D.; Hache, F.; Bacaou, F.; Andrieux, M.; Lafonta, F.; Journet, C.; Goze, C.; Brunet, M.; Bernier, P. Optics Communications 2000, 174, 271-275. (102) Sutherland, R. Handbook of Nonlinear Optics, 1996; Vol. 52. (103) Noshay, A.; Robeson, L. M. Journal of Applied Polymer Science 1976, 20, 1885-1903. (104) Herron, N.; Wang, Y.; Eckert, H. Journal of the American Chemical Society 1990, 112, 1322-1326.

PAGE 168

151 (105) Thierry, G.; Laurent, M.; Guillaume, C.; Jean-Pierre, B. Proceedings of the SPIEThe International Society for Optical Engineering. 1997, 3136, 358-365. (106) Gacoin, T.; Malier, L.; Boilot, J.-P. Chemical Materials 1997, 9, 1502-1504. (107) Yang, J.; Zeng, J.-H.; Yu, S.-H.; Yang, L.; Zhou, G.-e. Chemisty of Materials 2000, 12, 3259-3263. (108) Nardelli, M.; Fava, G.; Branchi, G. Acta Crystallogrphica 1960, 13, 898. (109) Inoue, M. Japanese Journal of Applied Physics, Part 1 (Regular Papers, Short Notes & Review Papers) 2000, 39, 3971-3976. (110) Borgohain, K.; Mahamuni, S. Journal of Materials Research 2001, submitted not yet published. (111) Jao, T. C.; Scott, I. Journal of Molecular Spectroscopy 1982, 92, 1-17. (112) Jones, R. L. Journal of Molecular Spectroscopy 1963, 11, 411-421. (113) Badger, R.; Waldron, R. D. Journal of Chemical Physics 1957, 26, 255-256. (114) James E, S. Journal of Chemical Physics 1956, 26, 248-254. (115) Cullity, B. D. Elements of X-Ray Difraction; Second Edition ed., 1978. (116) Firth, A. V.; Cole-Hamilton, D. J.; Allen, J. W. Applied Physics Letters 1999, 75, 3120-3122. (117) Soileau, M. J.; Guha, S.; Stryland, E. V. Proceedings of the SPIEThe International Society for Optical Engineering 1986, 633, 144-147. (118) Khodja, S.; Hagan, D. J.; Lin, J. H.; Przhonska, O. V.; S. Yang, J. B.; Stryland, E. W. V. Proceedings of the SPIEThe International Society for Optical Engineering 1997, 3146, 12-22. (119) Xia, T.; Hagan, D. J.; Dogariu, A.; Said, A. A.; Stryland, E. V. Applied Optics 1997 June 20, 36, 4110-4122. (120) Wang Song, Q.; Chungping, Z.; Gross, R.; Birge, R. N. U. Optics Letters 1993, 18, 775-777. (121) Dou, K.; Du, J. Y.; Knobbe, E. T. Journal of Luminescence 1999, 83-84, 241-246. (122) Kowalski, G. J. Optical Materials 2000, 14, 155-160. (123) Ramanujam, P. S.; Gluckstad, J.; Lindvold, L. R.; Rasmussen, J. J. R. D. Optical Memory & Neural Networks 1994, 3, 321-327.

PAGE 169

152 (124) Ohke, S.; Umeda, T.; Cho, Y. Japanese Journal of Applied Physics, Part 2 1998, 37, L1312-1314. (125) Takei, H.; Shimizu, N. Applied Optics 1996, 35, 1348-1354. (126) Clays, K.; van Elshocht, S.; Persoons, A. Optics Letters 2000, 25, 1391-1393. (127) Chase, L. L.; Stryland, E. W. V. CRC Handbook of Laser Science and Technology 1995, Supplement 2Optical Materials, 269-288. (128) Bernini, U.; Stefano, L. D.; Feo, M.; Mormile, P.; Russo, P. Optics Communications 1994, 112, 169-174. (129) Hagan, D. J.; Yang, S.; Stryland, E. W. V.; Basanez, C.; Moreshead, W.; Nogues, J. L. Proceedings of the SPIEThe International Society for Optical Engineering. 1999, 3798, 17-21. (130) Joudrier, V.; Bourdon, P.; Hache, F.; Flytzanis, C. Applied Physics B 1999, August 19, 70, 105-109. (131) Knopf, K. K. Zeitschrift fur Physik A (Atomic nuclei) 1991, 338, 233-237. (132) Sheik-Bahae, M.; Hagan, D. J.; Said, A. A.; Young, J.; Wei, T. H.; Van Stryland, E. W. Proceedings of the SPIEThe International Society for Optical Engineering 1990, 1307, 395-400. (133) Sheik-Bahae, M.; Hagan, D. J.; Said, A. A.; Young, J.; Wei, T. H.; Van Stryland, E. W. Optical Engineering, 30, 1228-1235. (134) Hagan, D. J., University of Central Florida, School of Optics, CREOL, http://nlo.creol.ucf.edu/Hagan/OSE%205312%202002/Class%209%20sellmeier%20and%20KK.doc 2002, Last accessed August 23, 2002. (135) Sun, X.; Lu, Z.; Goh, S. H.; Ji, W. Fullerene Science and Technology 2000, 8, 55-64. (136) Xia, H.; Zhu, C.; Gan, F. Proceedings SPIEThe International Society of Instrumentation Engineers 1997, 3136, 57-61. (137) Mishra, S. R.; Rawat, H. S.; Mehendale, S. C.; Rustagi, K. C.; Sood, A. K.; Bandopadhyay, R.; Govidraj, A.; Rao, C. N. R. Chemical Physics Letters 2000, February 4, 317, 510-514.

PAGE 170

BIOGRAPHICAL SKETCH Nikhil K. Kothurkar was born on January 25, 1976 in Pune, India. He completed his schooling from Loyola High School, Pune, India in 1991 from where he graduated first class with distinction. He completed his higher secondary schooling in 1993 from the prestigious Ferguson College, Pune, India. He graduated first class with distinction and continued toward his undergraduate studies in Maharashtra Institute of Technology, University of Pune, India. He pursued a degree in Polymer Engineering. He completed his final year project entitled “Hollow Fiber-Reinforced Polyester, Wind Turbine Blades” under the guidance of Dr. Zurale and graduated first class with distinction in 1997. In 1999, he moved to Gainesville, Florida to pursue his graduate studies in the University of Florida. He joined the Materials Science and Engineering program and worked as a graduate research assistant under the guidance of Dr. Anthony B. Brennan. His research project was developing a matrix for inorganic-organic particulate nanocomposites for optical applications. He maintained a GPA of 3.89/4.00. He defended his doctoral dissertation entitled “Solid State Cadmium Sulfide-Polymer Nanocomposites” on July 15, 2004 and graduated with a doctoral degree in August 2004. 153