VARIABLE BAND GAP POLY(3,4-AL KYLENEDIOXYTHIOPHENE)-BASED POLYMERS FOR PHOTOVOLTAIC AND ELECTROCHROMIC APPLICATIONS By BARRY C. THOMPSON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005
Copyright 2005 by Barry C. Thompson
To my parents
ACKNOWLEDGMENTS Graduate school is one of the most difficult times in a personâ€™s life. Without the support of others I could not have achieved this seemingly insurmountable goal. First and foremost I would like to thank my parents, Barry and Judithe Thompson. Never more than a phone call away, they were the one constant and they supported me through these challenging years, encouraging me in my work and helping me through the difficult times. I do not hesitate to say that I could not have completed graduate school without them. With no less sincerity, I would also like to thank my advisor Prof. John R. Reynolds. In our long history together, dating back to my days as an undergraduate student, Dr. Reynolds has encouraged me, supported me, and taught me to achieve at the highest level. From Dr. Reynolds I have learned to accept nothing less than my very best. Dr. Reynolds embodies the model of what an advisor and a scientist should be. I would also like to thank the members of my committee: Prof. Kenneth B. Wagener, Prof. Kirk S. Schanze, Prof. Daniel R. Talham, and Prof. Paul H. Holloway. Many members of the Reynolds group have been influential in my development as a graduate student. Philippe Schottland was my mentor in the lab during my first summer in Gainesville and he imparted a strong foundation in conjugated polymer research and became a close friend. C.J. DuBois, Carl Gaupp, Avni Argun, Shane Waybright, Jeremiah Mwaura, Ben Reeves, Christophe Grenier, and Kyukwan Zong, have all contributed to my work and I am grateful for their help and advice over the years. Special iv
thanks go to Nisha Ananthakrishnan for her help with AFM and optical microscopy, Genay Jones for his help with GPC, and Young-Gi Kim for his help in optimizing the multicomponent devices in chapter 3 and for sharing his expertise in the fabrication of photovoltaic devices. I would also like to thank Garry Cunningham, Mauricio Pinto, and Richard Farley from the Schanze group for their help in the photophysical characterization of several polymer samples. My experience on the polymer floor has been an extremely positive one. For this I would like to thank Sara Klossner, Tasha Simmons, and Lorraine Williams. I would also like to thank George and Jospehine Butler. The Butlerâ€™s generosity has enhanced the quality of my education in many ways. I am glad that I had a chance to meet Prof. Butler and catch a glimpse of a pioneer in the field of polymer chemistry. I would also like to thank the many collaborators that I have interacted with over the years. Thanks go to David Rauh at EIC Laboratories for his helpful discussions and incites. Prof. Charles Spangler at Montana State University not only provided the compounds investigated in chapter 4, but he also provided interesting discussions and ideas. Thanks are also due to Tracy McCarley at LSU, who provided the MALDI data in chapter 3. Tracyâ€™s superb abilities as a mass spectrometer and her tireless work ethic were a real asset to my work. I also acknowledge Khalil Abboud for solving X-ray structures for several compounds and for his willingness to discuss all aspects of molecular structure. I would also like to thank Prof. Pierre Audebert and Gilles Clavier from the ENS Cachan in France. I would especially like to thank Pierre, with whom I have worked closely on several projects over the past several years. I would like to thank v
Pierre for his advice, for being my personal tour guide in Paris, and for a continuing friendship. I cannot omit my undergradute advisor and mentor, Ronald Tucceri. Without the support and guidance of Ron Tucceri, I would not be a chemist. He turned a positive undergraduate experience at the University of Rio Grande into an outstanding education. For 5 years, my best friend has been Garrett Oakley. Together, we survived courses, cumes, orals, and everything else that life threw our way. I was proud to be the best man at his wedding. In the end, it is our close friends that get us through the difficult times. For this I would like to thank Garrett Oakley, Jennifer Oakley, and Emine Boz. I will miss the Friday lunches, hanging out, and all the good memories. vi
TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES .............................................................................................................ix LIST OF FIGURES ...........................................................................................................xi ABSTRACT ....................................................................................................................xvii CHAPTER 1 VARIABLE BAND GAP CONJUGATED POLYMERS: PRINCIPLES AND APPLICATIONS.....................................................................................................1 1.1 Fundamental Properties of Conjugated Polymers........................................1 1.2 Methods of Band Gap Control in Conjugated Polymers.............................5 1.3 Selected Applications of Conjugated Polymers...........................................7 1.3.1 Chromism in Conjugated Polymers.................................................8 1.3.2 LEDs..............................................................................................12 1.3.3 Photovoltaics..................................................................................14 1.4 Poly(3,4-alkylenedioxythiophene)-Based Polymers (PXDOTs)...............53 2 EXPERIMENTAL METHODS.............................................................................57 2.1 Introduction................................................................................................57 2.2 General Synthetic Methods........................................................................57 2.3 Electrochemical Methods...........................................................................59 2.4 Optical and Spectroscopic Methods...........................................................63 2.4.1 Spectroelectrochemistry and Colorimetry.....................................63 2.4.2 Photophysics..................................................................................64 2.5 Surface Morphology Characterization.......................................................65 2.6 Photovoltaic Devices.................................................................................65 2.7 Other Characterization Techniques............................................................68 3 CYANOVINYLENE-BASED CONJUGATED POLYMERS FOR PHOTOVOLTAIC DEVICES...............................................................................70 3.1 Introduction and Literature Overview.......................................................70 vii
3.2 Synthesis and Physical Characterization of Cyanovinylene Polymers......79 3.2.1 Soluble CNV Polymers via the Knoevenagel Polycondensation..80 3.2.2 Soluble CNV Polymers via Transition Metal Mediated Polymerization.............................................................................102 3.3 Electronic Band Structure Determination and Electronic Characterization.......................................................................................114 3.4 Photovoltaic Devices...............................................................................140 3.5 Conclusions and Outlook.........................................................................167 3.6 Synthetic Details......................................................................................169 4 DONOR FUNCTIONALIZED FULLERENES FOR BULK HETEROJUNCTION SOLAR CELLS...............................................................195 4.1 Soluble Donor-Functionalized Fullerenes...............................................195 4.2 Triphenylamine Functionalized Fullerenes..............................................204 4.3 Thienylene-vinylene Functionalized Fullerenes......................................214 4.4 Conclusions..............................................................................................225 4.5 Photovoltaics: Outlook and Perspective..................................................225 5 ELECTROPOLYMERIZABLE VARIABLE BAND GAP CONJUGATED POLYMERS........................................................................................................231 5.1 Developing New Routes toward High Electron Affinity Narrow Band Gap Polymers...........................................................................................231 5.1.1 Nitrogen Containing Heterocycles and Tetrazine........................232 5.1.2 Synthesis of Electropolymerizable Tetrazine Monomers for DA Polymers......................................................................................235 5.2 Synthesis of Bis-Heterocylcle Monomers for Electropolymerization.....242 5.3 Electronic and Electrochemical Properties of Tetrazine Compounds.....252 5.4 Conclusion and Outlook..........................................................................256 5.5 Closing Statement....................................................................................258 5.6 Synthetic Procedures................................................................................259 APPENDIX CRYSTALLOGRAPHIC INFORMATION FOR COMPOUNDS...........266 LIST OF REFERENCES.................................................................................................277 BIOGRAPHICAL SKETCH...........................................................................................295 viii
LIST OF TABLES Table page 1-1 Solar cell characteristics for the best polymer-PCBM devices reported in the literature...................................................................................................................50 3-1 Molecular weight (GPC) data for Knoevenagel polymers.......................................95 3-2 Photophysical data for CN-PPV analogues in toluene solution.............................117 3-3 Summary of the band structure data for the investigated polymers.......................130 3-4 Colorimetric data for CN-PPV analogues..............................................................140 3-5 Summary of polymer-PCBM (1/4) solar cell results.............................................158 3-6 Results for devices consisting of PProDOT-Hx 2 :CN-PPV, MEH-PPV, and PCBM measured under AM1.5 conditions............................................................160 3-7 Results for devices consisting of PBProDOT-Hx 2 -CNV, MEH-PPV, and PCBM measured under AM1.5 conditions............................................................166 4-1 Performance of TPA-C 60 / MEH-PPV solar cells relative to analogous PCBM / MEH-PPV devices.................................................................................................209 4-2 Photovoltaic device results for devices based on MEH-PPV, PCBM, and TPA-C 60 (0.8 / 3.9 / 0.3)..................................................................................................212 4-3 Device results for TV-C 60 containing devices measured under AM1.5 conditions...............................................................................................................222 5-1 Peak oxidation potentials of tetrazine compounds.................................................256 A-1 Atomic coordinates ( x 104) and equivalent isotropic displacement parameters (2x 103)for BEDOT-TZ-H 2 ...............................................................................268 A-2 Bond lengths  for BEDOT-Tz-H 2 ....................................................................269 A-3 Bond angles  for BEDOT-Tz-H 2 .......................................................................271 A-4 Anisotropic displacement parameters (2x 103) for BEDOT-Tz-H 2 ..................274 ix
A-5 Hydrogen coordinates ( x 104) and isotropic displacement parameters (2x 10 3) for BEDOT-Tz-H2.............................................................................275 A-6 Hydrogen bonds for BEDOT-Tz-H2 [ and ].....................................................276 x
LIST OF FIGURES Figure page 1-1 Representative repeat unit structures of conjugated polymers...................................2 1-2 Degenerate and nondegenerate ground states in conjugated polymers......................3 1-3 Conceptual model of the buildup of electronic energy bands in conjugated polymers and a comparison of the major types of band structures............................5 1-4 Schematic representation of the donor-acceptor approach to narrow band gap conjugated polymers..................................................................................................8 1-5 The effects of planarity on the band gap of conjugated polymers...........................10 1-6 Electrochromism in conjugated polymers................................................................11 1-7 Processes of photoluminescence (a) and electroluminescence (b) in conjugated polymers...................................................................................................................13 1-8 Schematic representation of the operation of a PLED.............................................13 1-9 Solar cell concepts....................................................................................................16 1-10 Standard Air Mass (AM) conditions for solar cell evaluation.................................18 1-11 Schematic representation of a DSSC under short circuit conditons.........................21 1-12 Illustration of the process of exciton generation and dissociation in an excitonic solar cell...................................................................................................................23 1-13 Schematic representation of a two layer organic solar cell......................................27 1-14 Chemical structures of MEH-CN-PPV and POPT...................................................32 1-15 The bulk heterojunction concept for polymer solar cells.........................................35 1-16 Schematic illustration of the four fundamental processes occurring in polymer based solar cells........................................................................................................37 1-17 Structures of narrow band gap polymers used in polymer / PCBM solar cells.......40 xi
1-18 Schematic representation of the processes in an optimized bulk heterojunction solar cell...................................................................................................................48 1-19 Different morphologies of heterojunction solar cells...............................................52 1-20 Representative PXDOT polymers............................................................................55 2-1 Relationship between electrode and absolute potentials..........................................62 2-2 Schematic representation of photovoltaic device composition................................67 3-1 Band diagram for an ideal donor polymer for PCBM..............................................74 3-2 Examples of narrow band gap CN-PPV analogues reported in the literature..........76 3-3 Representative structures from the three main types of proposed CN-PPV analogues for investigation in the development of an ideal donor polymer for photovoltaic devices.................................................................................................79 3-4 Family of compounds targeted for the synthesis of Knoevenagel polymers and model compounds....................................................................................................80 3-5 Synthesis of aldehyde monomers and model compound precursors........................81 3-6 Synthetic routes to compound 7...............................................................................84 3-7 Final step in the intended synthesis of compound 7 via route B in Figure 3-6........84 3-8 Model reaction with TosMIC...................................................................................85 3-9 Mitsunobu model reaction........................................................................................85 3-10 Synthesis of thiophene diacetonitrile.......................................................................86 3-11 Synthetic route for the synthesis of ProDOT-Hx 2 -(CH 2 CN) 2 ..................................87 3-12 Family of model compounds and polymers synthesized via the Knoevenagel methodology.............................................................................................................88 3-13 Synthesis of PProDOT:CN-PPV............................................................................90 3-14 Synthesis of Knoevenagel model compounds..........................................................91 3-15 Synthesis of PProDOT-R 2 :CN-PPV......................................................................92 3-16 Synthesis of Th-CN-PPV, CN-TV, and CN-PPV..................................................94 3-17 Absorption spectra for molecular weight fractions of Knoevenagel polymers........96 xii
3-18 MALDI MS of PProDOT-Hx 2 :CN-PPV...............................................................99 3-19 MALDI MS of CN-PPV........................................................................................100 3-20 MALDI MS of Th-CN-PPV..................................................................................101 3-21 MALDI MS for CN-TV.........................................................................................102 3-21 Mechanism of the Yamamoto coupling polymerization........................................105 3-22 Representative electropolymerizable CNV polymers............................................106 3-23 Family of precursors synthesized for Yamamoto coupling monomers..................107 3-24 Synthesis of compound 22.....................................................................................108 3-25 Synthesis of compounds 24 and 25........................................................................108 3-26 Synthesis of Yamamoto monomer precursors by Knoevenagel condensation......110 3-27 Synthesis of monomers for Yamamoto polymerization.........................................110 3-28 Yamamoto polymerization of PBProDOT-Hx 2 -CNV..........................................112 3-29 MALDI MS for PBProDOT-Hx 2 -CNV................................................................114 3-30 Solution absorbance of CN-PPV analogues in toluene..........................................115 3-31 Absorbance and photoluminescence spectra of CN-PPV analogues in toluene solution...................................................................................................................116 3-32 Thin-film optical absorbance of CN-PPV analogues.............................................118 3-33 Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for electropolymerized model compounds...................................................................122 3-34 Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for PBProDOT-Hx 2 -CNV..........................................................................................126 3-35 Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for PProDOT-Hx 2 :CN-PPV and Th-CN-PPV..........................................................127 3-36 Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for CN-TV and CN-PPV..........................................................................................................129 3-37 Summary of polymer band structures incorporating optical and electrochemical data.........................................................................................................................132 xiii
3-38 Spectroelectrochemistry of PProDOT-Hx 2 :CN-PPV (a) and PBProDOT-Hx 2 -CNV (b)..................................................................................................................134 3-39 Spectroelectrochemistry of PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNV.......................................................................................................................136 3-40 Spectroelectrochemistry for CN-TV, Th-CN-PPV, and CN-PPV.......................138 3-41 Spectroelectrochemistry for PBEDOT-CNPV......................................................139 3-42 Schematic representation of PL quenching in polymer PCBM blends..................142 3-43 Photoluminescence quenching of PProDOT-Hx 2 :CN-PPV and Th-CN-PPV with PCBM.............................................................................................................143 3-44 Photoluminescence quenching of CN-PPV and CN-TV with PCBM..................145 3-45 Excitation and absorption spectra for CN-TV/PCBM blend.................................146 3-46 Photovoltaic properties of PProDOT-Hx 2 :CN-PPV / PCBM solar cells based on a 1/4 blend (w/w) of polymer relative to PCBM...............................................148 3-47 AFM height image of PProDOT-Hx 2 :CN-PPV / PCBM blend (1/4)..................150 3-48 Photovoltaic properties of PBProDOT-Hx 2 :CNV / PCBM solar cells based on a 1/4 blend (w/w) of polymer and PCBM.................................................................151 3-49 AFM height image of a PBProDOT-Hx 2 :CN-PPV / PCBM blend (1/4)............152 3-50 Photovoltaic properties of Th-CN-PPV / PCBM solar cells based on a 1/4 blend (w/w) of polymer relative to PCBM......................................................................153 3-51 AFM height image of Th-CN-PPV / PCBM blend (1/4)......................................154 3-52 Photovoltaic properties of CN-TV / PCBM solar cells based on a 1/4 blend (w/w) of polymer relative to PCBM......................................................................155 3-53 AFM height image and optical microscopy of a CN-TV / PCBM blend (1/4).....156 3-54 Thin film absorption spectra for MEH-PPV (orange line), PProDOT-Hx 2 :CN-PPV (blue line), and PBProDOT-Hx 2 -CNV (solid black line) relative to the AMO (black dashed line) and AM1.5 (red line) spectrum.....................................159 3-55 Current voltage characteristics for PProDOT-Hx 2 :CN-PPV / MEH-PPV / PCBM devices........................................................................................................160 3-56 IPCE data for blends consisting of PProDOT-Hx 2 :CN-PPV, MEH-PPV, and PCBM.....................................................................................................................162 xiv
3-57 AFM height images for three component blends of MEH-PPV/PProDOT-Hx 2 :CN-PPV/PCBM.............................................................................................163 3-58 IPCE data for an all-polymer device based on a 1:1 blend of PProDOT-Hx 2 :CN-PPV and MEH-PPV................................................................................164 3-59 Absorbance and photoluminescence spectra for a 1:1 blend of PProDOT-Hx 2 :CN-PPV and MEH-PPV................................................................................165 3-60 IPCE data for devices with blends consisting of PBProDOT-Hx 2 -CNV, MEH-PPV, and PCBM.....................................................................................................167 4-1 Structure of fullerne-C 60 (a) and isomeric products upon cycloadditon (b)...........196 4-2 Overview of general synthetic methods employed in the synthesis of substituted fullerene derivatives...............................................................................................198 4-3 Structures of DA-dyads that yield efficient single-component PVDs...................201 4-4 Structure of Zn-Pc-C 60 DA-dyad used in three-component devices with MDMO-PPV and PCBM......................................................................................................204 4-5 Synthesis of TPA-C 60 ............................................................................................205 4-6 Absorbance and photoluminescence of TPA-C 60 in dichloromethane..................207 4-7 Cyclic voltammetry of TPA-C 60 in 0.1 M TBAP / dichloromethane....................208 4-8 Photovoltaic device results for devices based on MEH-PPV, PCBM, and TPA-C 60 (0.8 / 3.9 / 0.3)..................................................................................................213 4-9 Structures of 5TV and TV-C 60 ...............................................................................215 4-10 Normalized absoprtion and emission spectra of TV-C 60 (0.1 mg/mL) and 5TV (0.1 mg/mL) measured in dichloromethane solution.............................................216 4-11 Normalized absorption spectra of TV-C 60 in a thin film.......................................218 4-12 Solution electrochemistry of TV-C 60 and 5TV......................................................219 4-13 Thin-film electrochemistry of TV-C 60 ...................................................................221 4-14 Photocurrent action spectra for devices containing TV-C 60 with PCBM and PVK........................................................................................................................223 4-15 IPCE data for a series of PVDs containing TV-C 60 ...............................................224 5-1 Reduction potentials of common nitrogen containing heterocylces......................233 xv
5-2 Stucture of BBiThTz and PBBiThTz...................................................................234 5-3 Synthesis of BEDOT-PyrPyr as an illustrative examples of a DA monomer synthesis via metal catalyzed coupling..................................................................235 5-4 Synthesis of 3,6-diphenyl-terazine.........................................................................237 5-5 Synthesis of bis-N-pyrrolyl-tetrazine by nucleophilic aromatic substitution.........239 5-6 Structures of donor-acceptor tetrazine monomers with terminal electropolymerizable groups reported in the literature..........................................240 5-7 Structures of the investigated tetrazine family.......................................................243 5-8 Synthesis of BThTz and BPhTz............................................................................243 5-9 Synthesis of BEDOT-Tz.......................................................................................244 5-10 X-ray crystal structure of BEDOT-Tz-H 2 . Dashed lines represent hydrogen bonds......................................................................................................................245 5-11 Synthesis of 5-EDOT-2-thiophenecarbonitrile (EThCN).....................................246 5-12 Synthetic route for B(EDOT-Th)-Tz....................................................................249 5-13 Synthesis of B(BiEDOT)-Tz.................................................................................251 5-14 Solution absorbance spectra of tetrazine compounds in dichloromethane.............253 5-15 Solution reduction of tetrazine compounds by DPV..............................................254 A-1 Crystal structure for BEDOT-Tz-H 2 .....................................................................266 A-2 Representation of intramolecular and intermolecular hydrogen bonding in BEDOT-Tz-H 2 ......................................................................................................266 xvi
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 VARIABLE BAND GAP POLY(3,4-ALKYLENEDIOXYTHIOPHENE)-BASED POLYMERS FOR PHOTOVOLTAIC AND ELECTROCHROMIC APPLICATIONS By Barry C. Thompson May 2005 Chair: John R. Reynolds Major Department: Chemistry The major focus of this work was the synthesis of new narrow band gap (< 1.8 eV) conjugated polymers that can be used to effectively absorb sunlight for application in solar cells. The investigated polymers were designed to absorb solar photons, and also to have the proper frontier orbital energies to ensure air stability and effective electron transfer to a fullerene-based acceptor on photoexcitation. Based on these criteria, a family of soluble poly(arylene-cyanovinylenes) were synthesized by either Knoevenagel polycondensation or Yamamoto coupling polymerization. Here the aromatic rings in the polymer backbone were varied using benzene, thiophene, 3,4-ethylenedioxythiophene, and dialkyl-substituted 3,4-propylenedioxythiophene, and combinations thereof. Optically determined electronic band gaps of the resulting polymers varied from 1.4 to 2.1 eV. It was found that the narrowest band gaps resulted from the strongest donor acceptor interactions along the conjugated backbone. xvii
The polymers were characterized by electrochemical and spectroscopic means. Solar cells were constructed and characterized using a solar simulator for power conversion efficiencies and monochromatic illumination to measure the incident photon to current efficiency (IPCE). Several types of devices were investigated: primarily those based on two-component blends comprising the poly(arylene-cyanovinylenes) and soluble fullerene derivatives, and three-component devices based on the addition of a high-molecular-weight commercial polymer to the aforementioned blends. Commercial conjugated polymers such as MEH-PPV (poly(2-methoxy-5-(2â€™-ethylhexyloxy)para-phenylenevinylene)) were found to be effective as blending agents, as the high-molecular-weight polymers improved film quality, and the high band gap of MEH-PPV provided a broadening of the absorption spectrum of the blend. Device efficiencies of ca. 1% were achieved in these cases. The approach taken here of blending several polymers with varying band gaps is a little-used approach and suggests a means of maximizing solar spectrum overlap. The electrochromic properties of this family of polymers were also investigated, indicating that these are multifunctional materials. The utility of so-called donor-acceptor dyads based on compounds that contain conjugated oligomers covalently attached to fullerenes, was also investigated. Additionally, synthetic routes toward a new class of donor-acceptor polymer based on tetrazine and 3,4-ethylenedioxythiophene were explored. xviii
CHAPTER 1 VARIABLE BAND GAP CONJUGATED POLYMERS: PRINCIPLES AND APPLICATIONS 1.1 Fundamental Properties of Conjugated Polymers Although conjugated polymers have long been the focus of intense research, this field has become especially well known in recent years, perhaps owing to the awarding of the 2000 Nobel Prize in Chemistry to Hideki Shirakawa, Alan MacDiarmid, and Alan Heeger for the discovery of the field. Conjugated polymers (often referred to as conducting polymers because they are able to achieve a high level of electrical conductivity in their doped form) are attractive for such applications as light emitting diodes (LEDs), thin film transistors (TFTs), photovoltaic devices (PVDs), and electrochromic devices (ECDs), among many others, primarily because of the semiconducting nature of this class of organic polymer. While the electronic properties are the focus of most applications, it is the underlying synthetic organic flexibility that is the most attractive feature of conjugated polymers. Through the manipulation of monomeric and polymeric structure, physical, optical, electronic, and electrochemical properties can be tuned for a specific application. Exploration of this structural diversity over the years has led to the development and investigation of numerous classes of conjugated polymers. Figure 1-1 shows the parent structures of a few representative classes of conjugated polymers (CPs) along with a few specific examples of interest. For a thorough overview of the historical development of the field, numerous reviews are available. 1 1
2 nPAN H nPPyS nPTh nPPV PPP n S n PITNS nOO PEDOT n n nS nP3HTO O O O OR RO NC OR RO CN CN-PPVMDMO-PPVMEH-PPV HN nPANI Figure 1-1. Representative repeat unit structures of conjugated polymers. PA = polyacetylene, PPy = polypyrrole, PTh = polythiopehene, PPP = poly(p-phenylene), PPV = poly(p-phenylene vinylene), PANI = polyaniline, PITN = polyisothianapthene, PEDOT = poly(3,4-ethylenedioxythiophene), CN-PPV = cyano-PPV, MEH-PPV = poly(2-methoxy-5-(2-ethylhexoxy)-1,4-phenylenevinylene), MDMO-PPV = poly(2-methoxy-5-(3,7-dimethyloctyloxy)-1,4-phenylenevinylene), and P3HT = poly(3-hexylthiophene). While the chemical structures of CPs vary widely, the electronic structures of CPs fit into two main classes as defined by the delocalization of -electrons along the conjugated backbone: those with degenerate ground states and those with non-degenerate ground states. With the exception of polyacetylene (degenerate ground state), the other conjugated polymers in Figure 1-1 have a non-degenerate ground state (note that PANI is a special case that does not fit neatly into this model and will not be considered here 2 ). Figure 1-2 shows a schematic representation of the ground state electronic structures of conjugated polymers of both types. While PA is predicted to have a fully delocalized
3 structure due to the two equivalent ground-state resonance forms, it is in fact subject to a Peierls distortion that removes the perfect bond length alternation. S S S S S S S S Degenerate Ground StatesNondegenerate Ground StatesAromaticQuinoidab Figure 1-2. Degenerate and nondegenerate ground states in conjugated polymers. (a) Illustration of the degenrate ground states of polyacetylene. (b) Illustration of the nondegenerate ground states of polythiophene as a representation for all aromatic conjugated polymers. Here there is an energetic difference between the localized aromatic form and the delocalized quinoid form. To fully understand the electronic properties of CPs, one must understand the band model that is used to explain the electronic structure of conjugated polymers. Figure 1-3a shows the fundamental aspects of this band model for the case of polythiophene. In conjugated polymers, the delocalization of -electrons along the backbone results in the formation of continuous energy bands when the effective degree of conjugation becomes sufficient. No definitive guidelines specify when the effective conjugation length is enough to consider the molecule a polymer. Often, the saturation of a specific property (e.g. electrochemical or optical) in a series of well-defined oligomers is used to estimate the convergence to polymeric behavior. For the case of PPV, it has been estimated that 7 to 10 repeat units is sufficient, 3 while other studies suggest 10-15 repeat units. 4 Such measurements are however dependant on the polymer structure and environment. Additionally, different properties can saturate at different effective conjugation lengths.
4 As such, there is no simple answer to the question of when an oligomer becomes a polymer. Nonetheless, at some point, the result of increasing the conjugation length is the conversion from a molecular system with well-defined HOMO and LUMO energies to a system with a band gap (E g ) that separates fully populated electronic states in a valence band (VB) from empty electronic states in the conduction band (CB). Numerous factors (discussed in the following section) are known to control the magnitude of the band gap in conjugated polymers. As with other materials that can be described according to the band model, the magnitude of the band gap is critical in determining the ultimate properties of the material. As seen in Figure 1-3b, a material with zero band gap is a metal, as it contains a partially filled band of electrons allowing for a high conductivity. A material with a small band gap (arbitrarily defined as < 3 eV, although some definitions place the upper limit at 4 eV) 5 is referred to as a semiconductor. Such materials often absorb visible light and thus are potentially colored. In this case, thermal population of CB states results in increasing conductivity with increasing temperature. When the band gap reaches a prohibitively large value, thermal energy is not sufficient to induce electronic conductivity. These large band gap materials (E g > 3 eV) are insulators, and are expected to be colorless due to the fact that they do not absorb light at wavelengths greater than 400 nm. Even though conjugated polymers can achieve near metallic conductivity in their doped forms (section 1.3.1 describes doping and the effects on band structure), in their neutral state, conjugated polymers are either insulators or weak semiconductors. The precise magnitude of the band gap affects numerous properties such as color and conductivity.
5 CBVB EgHOMO LUMO Th2Th4Th6ThPTh CBVB CBVB Eg CBVBEg ab Eg = 0metalEg < 3 eVsemiconductorEg > 3 eVinsulator Figure 1-3. Conceptual model of the buildup of electronic energy bands in conjugated polymers and a comparison of the major types of band structures. (a) Here thiophene (Th) is used as the model and only the frontier molecular orbital bands are considered. The spacing between the energy levels are intended to correspond with precise energies. For a more complete and accurate illustration of the molecular orbital structure of polythiophene see Salzner, U.; Lagaowski, J. B.; Pickup, P. G.; Poirier, R. A. Synth. Met. 1998, 96, 177-189. (b) Band diagrams showing the difference between metals, semiconductors, and insulators. 1.2 Methods of Band Gap Control in Conjugated Polymers As discussed in Section 1.1, the organic nature of conjugated polymers allows for the manipulation and variation of their electronic properties. One important tunable electronic property is the band gap. For many applications such as LEDs and PVDs, the value of the band gap is pivotal, as it plays a role in the color of the emitted light in LEDs and the effectiveness with which solar radiation is absorbed in photovoltaic devices. The structural basis for manipulating the magnitude of the band gap has been investigated and reviewed. 6 Several major factors are known to directly correlate the chemical structure of the polymer backbone with the magnitude of the band gap. The most fundamental cause of a band gap in conjugated polymers is the bond length alternation. In polyheterocylces, this bond length alternation is best envisioned as a competition between the nondegenerate quinoid and aromatic ground states represented in Figure 1-2. If the two
6 structures were exactly equal in energy, they would be degenerate, and the complete delocalization of electrons along the backbone would, in principle, lead to a zero-band-gap polymer. However, even for polyacetylene, the band gap is greater than zero (1.4 eV) based on the Peierls distortion (which is the instability of one-dimensional systems that causes a structural distortion resulting in the creation of an energy gap at the Fermi level). As a consequence, even polyacetylene exhibits definite bond-length alternation. For aromatic polymers, synthetic efforts have been made to enhance the stability of the quinoid form relative to the aromatic form, in hopes of reducing the magnitude of the band gap, using resonance effects as in the case of poly(isothianapthene) (PITN) (Figure 1-1). In this case the larger energy of aromatization of benzene relative to thiophene results in the increased stability of the quinoidal form of the polymer backbone and a band gap of 1.1 eV (compared with 2.0-2.5 eV for polythiophene 7 ). The control of backbone planarity through both intrachain (i.e., in ladder-type or fused-ring polymers and interchain effects (i.e., induced order, as in regioregular poly(alkylthiophenes)) are also routes that have been explored for controlling the magnitude of the band gap. One important synthetic target has been the development of low band gap (or narrow band gap) conjugated polymers (E g < 1.5 eV). Narrow band gap polymers are promising candidates for solar cells 8 as well as n-type conductors and multicolor electrochromic polymers. 9 Among the successful approaches to a narrow band gap system, the donor-acceptor approach has received significant attention. 10 In this approach, the combination of an electron-rich donor and an electron-deficient acceptor in close conjugation results in a polymer with a compressed band gap. By carefully selecting the appropriate structure of the donor and the acceptor, one can tune the
7 magnitude of the band gap, the absolute energies of the frontier orbitals, and the solubility of the resulting polymer. For the realization of narrow band gap polymers, the donor-acceptor approach has by far the most synthetic utility while avoiding the necessity of controlling interchain effects or generating insoluble, rigidly planar polymer backbones. Figure 1-4 shows the principle of the donor-acceptor concept. Here it can be seen that a hybridization of HOMO and LUMO energy levels of the donor and the acceptor results in the formation of the compressed band gap. The specific example shown is for the case of poly(bis-EDOT)-cyanovinylene (PBEDOT-CNV). Here it can be seen that not only is the band gap compressed in the DA polymer relative to the pristine donor and acceptor polymers, but that the HOMO and LUMO energy of the DA polymer strongly resembles that of the donor polymer and the acceptor polymer respectively. 11 1.3 Selected Applications of Conjugated Polymers In the early days of conjugated polymer research, conductivity was the primary focus for application-based research. It was envisioned that a combination of high conductivity and lightweight could lead to a revolution in electronics technology. While polyacetylene was eventually able to achieve a conductivity approaching that of copper, the environmental instability of polyacetylene along with the significantly decreased conductivity of more robust conjugated polymers, led to a shift in the focus of application driven research. At the present time, a plethora of possible applications exist for conjugated polymers, with little or no focus bestowed on the conducting nature that made these polymers so famous. While applications based on the conducting nature of conjugated polymers rely on the use of highly doped polymers, applications sought after today focus more on the neutral, semiconducting form of the polymers.
8 Among the possible applications at the forefront of research today, those based on the optical properties of conjugated polymers or the interaction of the polymers with light are among the most investigated. Here three examples are discussed, including the concept of chromism, the principles of light emitting diodes (LEDs), and an extensive review of the polymer-based photovoltaic devices. EnergyLUMOHOMOLUMOHOMO DonorAcceptorDonor-AcceptorS OO n nS OO S OO CN CN n HOMO = 4.6 eV LUMO = 3.0 eVEg = 1.6 eVHOMO = 4.7 eV LUMO = 3.6 eVEg = 1.1 eVHOMO = 5.8 eV LUMO = 3.7 eVEg = 2.1 eVPEDOTPCAPBEDOT-CNVE= 0 eV Eg Figure 1-4. Schematic representation of the donor-acceptor approach to narrow band gap conjugated polymers. The energy diagram indicates the effective hybridization of molecular orbitals that occurs in donor-acceptor polymers that results in the narrowing of the band gap. The energy values of PBEDOT-CNV are given as a representative example. All energies are reported relative to vacuum. 1.3.1 Chromism in Conjugated Polymers Chromism can be defined as a reversible change in the transmitted or reflected color of a material, as a result of an applied stimulus. This stimulus can be any external force such as light, heat, or electrical current. Conjugated polymers are well known to
9 display many types of chromic phenomena both in solution and in the solid state. 12 Solvatochromism, thermochromism, photochromism, and electrochromism are among the best-known chromic phenomena in conjugated polymers. With the exception of electrochromism, the other major types of chromism in conjugated polymers can be defined as conformation-induced chromism. Figure 1-5 shows the central concept of conformation-induced chromism for the example of polythiophene. Here, any external stimulus that will change the effective conjugation length of a polymer, by altering the planarity of the backbone and the degree of -orbital overlap, will alter the effective gap and ultimately the color of the material, as the band gap determines the minimum energy light the polymer will absorb. In solvatochromism, changes in the solvent quality affect the aggregation behavior of the polymer in solution, while in thermochromism (both in solution and in films), heating the polymer has a similar effect on the overall chain conformation. Photochromism operates based on a distinct mechanism, but the ultimate result is again an overall change in the polymer conformation. In this case a photochromic moiety must be incorporated into the polymer and a reversible chemical change (e.g., an isomerization) in this photochromic group serves to perturb the main chain conformation of the polymer backbone. Azobenzene 13 and diarlyethene 14 are among the most commonly studied photochromic moieties, which are either incorporated as pendant or main chain functionalities along the polymer backbone. Potential applications of conformation-induced chromism are primarily in the areas of sensors, 15 memories, and switches. 14
10 Energy Eg S S S S S S S S S PlanarTwisted Long effective conjugation lengthShort effective conjugation length Figure 1-5. The effects of planarity on the band gap of conjugated polymers. Adapted from McCullough, R. D.; Ewbank, P. C. In Handbook of Conducting Polymers, 2 nd ed.; Skotheim, T. A.; Elsenbaumer, R. L.; Reynolds, J. R., Eds.; Marcel Dekker: New York, 1998, chapter 9. In recent years, electrochromic phenomena have gained special interest because of possible applications in smart windows, displays, optical shutters, and mirror devices. 16 Electrochromism is defined as a reversible change in the absorption/transmission characteristics of a material (throughout the electromagnetic spectrum) that is induced by an external voltage. All conjugated polymers are potentially electrochromic in thin-film form. Figure 1-6 shows the principle of electrochromism in conjugated polymers. 17 In Figure 1-6a it can be seen that the neutral polymer displays a band gap and the minimum energy transition is the HOMO-LUMO transition or * transition. On oxidation (or analogously, on reduction) (Figure 1-6b), new charge-carrier associated states are induced in the band gap of the polymer, giving rise to lower energy transitions. At high levels of oxidation, the * transition of the neutral polymer is depleted at the expense
11 of the lower energy charge-carrier associated transitions as the polymer is converted to its doped (conducting) state. These charge carriers are generally considered to be polarons (delocalized radical cations) or bipolarons (delocalized dications). The nature of these charge carriers has been discussed extensively in the literature. 18 Figure 1-6c shows a typical case of the spectral changes that occur in a conjugated polymer on electrochemical oxidation. Electrochromism in conjugated polymers has recently been reviewed. 19 Figure 1-6. Electrochromism in conjugated polymers. (a) The consequences of oxidation (doping) one the band structure of conjugated polymers for increasing levels of doping. Here allowed optical transitions are shown for the various states of the polymer. (b) Structural representation of the basic charge carrier types for the example of polythiophene. (c) Spectral consequences of oxidative doping for the representative case of PProDOT-Bu 2 :CN-PPV (see chapter 3). In this case the * transition is centered at ~600 nm and it is bleached as the potential is increased from 0.60 to 0.95 V vs. Fc/Fc + .
12 1.3.2 LEDs Perhaps the most highly publicized and well-known application of conjugated polymers is as the active material in Polymer Light Emitting Diodes (PLEDs). The discovery of electroluminescence in conjugated polymers in 1990 can be considered at least somewhat responsible for the rapid growth of research in the field that was witnessed throughout the 1990s, transforming conjugated polymers from an academic research field to one with real applications. 1 c,20 Figure 1-7 summarizes the basic concept of electroluminescence in conjugated polymers in comparison to photoluminescence. In photoluminescence, a conjugated polymer absorbs light of sufficient energy to promote an electron from the HOMO to the LUMO of the polymer. After relaxation to the lowest energy singlet excited state, the polymer can then emit light through fluorescence as it relaxes to the ground state. In a simple single-layer PLED, electrons are injected at the cathode to yield radical anions, and holes are injected at the anode to yield radical cations. The generated charges migrate under the influence of the applied electric field, and recombine to yield a singlet excited state that can emit light upon relaxation to the ground state. Figure 1-8 shows a schematic of a typical two-layer PLED incorporating both electron-transport and hole-transport layers. In this device, electrons are injected from a low-work-function metal (Ca, Al) into the LUMO of an acceptor polymer with a high electron affinity. At the same time, holes are injected from the high-work-function electrode (ITO) into the HOMO of a donor polymer with a low ionization potential. Holes are then transported through the donor, and electrons through the acceptor. Recombination and emission then occur at the interface between the donor and acceptor. A great deal of effort has been expended to optimize the device structure, balance the
13 charge transport properties, and synthesize stable polymers that emit light across the electromagnetic spectrum. LUMOHOMO h h' absinglet excited state e-Cathode radical anion h' singlet excited state radical cation e-Anode Figure 1-7. Processes of photoluminescence (a) and electroluminescence (b) in conjugated polymers. CBCBVBVB Anode (ITO)Cathode (Al)EFEFHole Transport LayerElectron Transport Layer h+eh Figure 1-8. Schematic representation of the operation of a PLED.
14 1.3.3 Photovoltaics After the discovery of electroluminescence in conjugated polymers, it became apparent that conjugated polymers could be used as the active material in other optoelectronic devices. While much of the early research focus was bestowed on LEDs, some work on conjugated polymer-based photovoltaic devices (PVDs) began in the early 1990s. The concept certainly follows, as a PVD operates in the inverse manner of an LED. Early work on polymer PVDs was also spurred on by contemporary work on small molecule organic PVDs. Over the last decade, the interest in polymer-based PVDs has certainly grown as efficiencies have improved dramatically. Additionally, the promise of achieving a lightweight, flexible solar cell is attractive from a scientific, economic, and ecological point of view. Basic concepts of photovoltaic devices. A photovoltaic device provides a source of electrical current under the influence of light or similar radiation. 21 In essence, a photovoltaic device (PVD or solar cell) converts sunlight into electrical power. The photovoltaic effect was first observed by Bequerel in 1839 22 using electrode-bound silver halides in acidic solution. However, over the course of the next 100 years, little progress was made in the development of solar energy technology. With the advent of the semiconductor revolution and silicon technology, solar cells were finally able to achieve the status of viable devices with practical applications. The first report of a silicon-based solar cell came from Bell Labs in 1941, 23 although it was not until 1954 that solar cells became the focus of intense research based on the work of Chapin concerning single crystal silicon cells. 24 Although numerous other types of solar cells exist today, devices based on silicon comprise ~99% of the world market for photovoltaic materials. 25
15 Silicon-based solar cells. The prototypical silicon-based solar cell is the pn heterojunction device, as shown schematically in Figure 1-9. In this type of device, a pn heterojunction is created by taking a wafer of p-type silicon (silicon doped with a group III element such as Gallium or Indium) and diffusing a group V element (Phosphorous or Arsenic) into the surface. 26 The p-type region has an excess of holes because of the electron vacancies generated by the incorporation of the group III element. Conversely, the n-type region has an excess of electrons. Both materials however are electrically neutral when they are isolated from each other. As a consequence of the difference in the Fermi energy in the nand p-type regions, a depletion layer is established at the interface of the nand p-type materials. In this depletion region, excess electrons from the n-type material have filled the holes in the p-type region near the interface. As a result, there are no free charge carriers in the depletion region and additionally there is a built in potential resulting from generation of a net positive charge on the n-type side of the depletion region and a net negative charge on the p-type side of the depletion region. When the pn junction is irradiated with light of sufficient energy (greater than or equal to the magnitude of the band gap) an electron is excited from the valence band to the conduction band, resulting in a free electron-hole pair. As a consequence of the built-in-potential (E bi ) in the depletion region, electrons are swept toward the n-type region, and holes are swept toward the p-type region. If the heterojunction is connected to an external circuit, a current will be measured flowing in the reverse bias sense, from p to n, and the current is thus defined as negative (notice the difference in the direction of current flow and electron flow in Figure 1-9a). If the bias is swept from negative to positive, the I-V characteristic of the diode will be as seen in Figure 1-9b. Here, it can be
16 seen that the dark current (current that flows under no external irradiation) is illustrative of a prototypical diode in which no current flows until a sufficient bias (turn-on voltage) is applied to overcome the built-in potential of the depletion region, and drive electrons in the direction of positive current flow. The light current shows that when such a diode is irradiated, current will flow in the negative direction until a bias equal to V oc (open circuit voltage) is applied. The short-circuit current (I sc ) is the current that flows under irradiation when no bias is applied. Figure 1-9. Solar cell concepts. (a) Representation of a silicon-based pn heterojunction PVD under short circuit conditions, illustrating the built in electric field (E bi ) that exists in the depletion region near the pn junction and the direction of current flow upon irradiation with light. (b) Current-voltage characteristic of a silicon solar cell for dark and light conditions. (c) Illustration of the fill factor (FF), showing the open circuit voltage (V oc ), the short circuit current (I sc ), and the voltage and current at the maximum power point (V m and I m ). The important characteristics of such a device are the V oc , I sc (or more appropriately J sc ), FF (fill factor), and (power conversion efficiency). The FF, defined in Equation 1-1, reflects the diode properties of the solar cell. While an ideal solar cell would have a FF
17 of 1, real devices show smaller values based on the effects of resistances that exist within the device, rendering the diode non-ideal. The power conversion efficiency, as defined in Equation 1-2, is the most important parameter in a solar cell when considering the relationship between production cost and power output. The efficiency is simply a ratio of the power output (electrical power) relative to the power input (solar radiation). As a set of informal guidelines, it has been proposed that it is desirable for a practical device to have an efficiency of 5 25% along with a FF of 0.6 to 0.8, a V oc of 0.5 to 1.0 V, and a J sc of 10 to 40 mA/cm 2 . 27 However, the requirements for device performance are ultimately based on the application and a general set of standards cannot be proposed. For example, the requirements for a solar cell on a satellite and a solar cell in a handheld calculator are certainly different. FF = (Im) (Vm)(Isc) (Voc) PoutPinFF(Isc) (Voc) Pin=(1-1)(1-2) To compare the efficiency of devices, a standard set of illumination conditions has been devised, based on the modification of the solar spectrum upon its interaction with differing amounts of air mass (AM) (Figure 1-10). The condition defined as AM0 is the condition that is encountered in outer space with sunlight impinging from directly overhead with an intensity of 1.35 kW/m 2 . With sunlight directly overhead at sea level, AM1 conditions are encountered, where AM1 means that the sunlight has passed through exactly one air mass. Here the total power density has been reduced to 1 kW/m 2 . 27 The most common simulator condition used is the AM1.5 (or AM1.5 global or AM1.5G). Under these conditions, sunlight has passed through 1.5 air masses because of the
18 incident angle of the sunlight relative to the normal. Standard reporting conditions also require an incident intensity of 1 kW/m 2 (100 mW/cm 2 ). The AM1.5G spectrum incorporates both the direct and diffuse (scattered) portions of the solar spectrum encountered on the earthâ€™s surface. 28 Figure 1-10 shows the relationship of AM0 and AM1.5G radiation. Here it can be seen that the atmosphere reduces the intensity of solar radiation in general, especially in the UV region. Additionally, atmospheric components, primarily water, carbon dioxide, oxygen, ozone, and methane, introduce several characteristic absorption bands into the spectrum. In the visible and near-infrared portion of the solar spectrum shown in Figure 1-10a, the atmospheric absorption spectrum is dominated by the interaction of sunlight with water vapor. 29 The small peak at ~825 nm as well as the large peaks at 940 and 1140 nm are directly attributed to absorption by water vapor. The sharp peak at 750 nm is due to ozone and oxygen. 30 Figure 1-10. Standard Air Mass (AM) conditions for solar cell evaluation. (a) Comparison of the solar spectrum under AM0 and AM1.5 conditions. (b) Schematic representation of the irradiation geometry represented by standard irradiation conditions.
19 Another important measure of the efficiency in solar cells is the external quantum efficiency (EQE) or Incident Photon to Current Efficiency (IPCE) defined in Equation 1-3. The importance of this measure is that it is a single-wavelength measurement that indicates the number of photogenerated charge carriers (electrons) produced by a photon of a given energy. When the IPCE is plotted vs. wavelength, a photocurrent action spectrum is obtained that indicates the relative contribution of various spectral regions (and thus the materials absorbing in those regions) to the overall photocurrent generated by the device. =IPCE (%) = 0.124 x Isc (A/cm2) (nm) x Pin (W/cm2) # electrons# photonsx 100(1-3) Over the last 50 years, the efficiency of silicon-based PVDs has risen dramatically. Coupled with the concomitant decrease in manufacturing costs, silicon photovoltaics have come to dominate the world market. However, despite efforts minimize the cost of silicon devices, this class of solar cell accounts for less than 0.1% of the worldâ€™s energy market, mostly because of the expensive methods of device fabrication. 31 As mentioned previously, silicon based solar cells account for 99% of the world photovoltaic market. Nonetheless, there are a variety of cutting edge technologies based on other inorganic materials that truly define the state of the art in solar cells. At the present time, the highest reported efficiency for a single crystal silicon solar cell is ~24%. 32 However, the highest reported efficiency for a solar cell is ~37.3% for the case of a triple junction device reported by Spectrolab. 32 In a multijunction cell, individual heterojunction layers are stacked and connected in series or in parallel, in order to increase the V oc or I sc of the device, and the ultimate power output. Additionally such
20 devices are designed to contain three semiconductor layers with distinct band gaps. The use of several semiconductors with variable band gaps allows a thorough overlap with the solar spectrum. For the case of the 37% devices from spectrolab, GaInP 2 , GaAs, and Ge are used as the active materials, which ensures absorption of light from 300-1600 nm. Additionally, such devices are equipped with concentrators, which are optical elements that concentrate direct sunlight by tracking the sun. Multijunction cells currently give the highest efficiency for devices in both space and terrestrial application. 33 Another class of inorganic solar cells that are receiving increasing attention are the thin film devices, primarily those based on Cu(In, Ga)Se 2 (â€œCIGSâ€) and CdTe. The CIGS devices are of great interest because they have been to shown to exhibit efficiencies as high as 19%. These semiconductor alloys are attractive because they can be sputtered on metal films to yield thin film solar cells. The heterojunction in these devices is often formed by depositing a thin layer of CdSe on top of the CIGS layer. In this case, CdSe has a band gap of 2.4 eV and CIGS have a band gap of ~1-1.1 eV. Such devices are however still in the developmental stage. Dye-sensitized solar cells. Over the last 10 years, another class of solar cell has emerged as a possible competitor for silicon-based solar cells in some niche applications. The dye-sensitized solar cell (DSSC) was first reported by Grtzel in 1991 34 with a power conversion efficiency of 7.1%. Currently, DSSCs have achieved power conversion efficiencies greater than 10% under AM1.5 conditions. 35 The DSSCs offer an interesting alternative to silicon devices, becasue the principle of operation and methods of device fabrication are completely different. A DSSC is a photoelectrochemical cell in which the processes of light absorption and charge transport are decoupled. Figure 1-11 shows the
21 structure and mode of operation of a typical DSSC. In these devices, the critical component is a mesoporous layer of TiO 2 , which is deposited on a conductive glass substrate, using various techniques (such as screen-printing, from a colloidal suspension of TiO 2 nanoparticles). 36 The resulting TiO 2 layer is about 50 to 65% porous and the rough, nanostructured surface has a large surface area to which a monolayer of charge-transfer dye is attached via chemisorption. The most commonly used dyes are polypyridyl complexes of ruthenium or osmium. 37 Between the dye-functionalized TiO 2 layer and the cathode, an electrolyte (liquid, solid, or gel) based on the iodide/triodide redox couple is incorporated into the device. Figure 1-11. Schematic representation of a DSSC under short circuit conditons. The mode of operation is illustrated by the numbered processes. (1) Light is absorbed by the dye molecules bound to the TiO 2 , generating a singlet excited state. (2) Electron transfer from excited dye molecule to TiO 2 layer. (3) The electron is collected at the anode (conducting glass) and enters the external circuit. (4) The electron is injected into the electrolyte. (5) The injected electron reduces an oxidized species (ox to red). (6) The reduced species (red) is then converted back to ox as it donates an electron to the now oxidized dye molecule to complete the circuit.
22 The operation of the device is very simple as outlined in Figure 1-11. Light is absorbed by the sensitizing dye and electron transfer from the excited dye molecule to the TiO 2 follows. Electrons are then collected at the conductive glass substrate and pass into the external circuit. The now-oxidized dye is then reduced to its neutral form via electron transfer from the electrolyte, which is used to complete the circuit in the photochemical device. DSSCs are still an active area of research. It can be seen that the fundamental mode of operation is distinct from silicon devices in which electron-hole pairs are created simultaneously upon excitation with light. This basic difference lies in the fact that DSSCs are excitonic solar cells, a concept that will be explained further in the next section on organic solar cells. Current efforts in DSSC research are focused on the development and optimization of solid-state solar cells. 38 If highly efficient, all solid state devices could be realized, DSSCâ€™s could provide a viable technology. Small-molecule organic solar cells. Another class of solar cells that have received recent attention as a potential alternative to silicon-based cells are photovoltaic devices based on organic molecules. While organic photovoltaics have not achieved the efficiencies of silicon devices or even dye-sensitized solar cells, this class is of interest based on the promise of low processing and production costs. While anthracene was the first organic compound observed to exhibit photoconductivity in 1906, 31 it was not until 1986 that the first viable organic solar cell was described by Tang. 39 In this case a power conversion efficiency of 1% was observed. More important than the magnitude of the efficiency is the conceptual design of the device presented. In this work, Tang employed a two-layer device design based on copper phthalocyanine (CuPc) and a perylene
23 derivative and observed that the interface between the two organic materials was the key parameter in determining the photovoltaic performance. The importance of the interface in an organic device is at the very heart of the photovoltaic process in organic materials. Unlike inorganic semiconductors where light absorption leads to the direct formation of free electrons and holes, absorption of light by organic molecules generates mobile, electron-hole pairs, or excitons. 40 Hence, organic solar cells belong to the class of excitonic solar cells (along with DSSCâ€™s). 41 For such devices, the electron-hole pair (exciton) that results from light absorption is relatively strongly bound and thus requires a driving force in order to affect dissociation. The interface between two materials of different electron affinity and/or ionization potential is an ideal site for exciton dissociation. Figure 1-12 illustrates the process of exciton dissociation. Additionally, Figure 1-12 illustrates two types of electron transfer; forward electron transfer from excited donor to the ground state acceptor (Figure 1-12a) and forward electron transfer to an excited acceptor from the ground state donor (Figure 1-12b). Figure 1-12. Illustration of the process of exciton generation and dissociation in an excitonic solar cell.
24 An important parameter in the design of an excitonic solar cell is the selection of complementary donor and acceptor materials that will ensure effective photoinduced charge transfer. Thus an appropriate energetic relationship must exist between the donor and acceptor. In a simplified view of the process of exciton dissociation, for the case illustrated in Figure 1-12a, one can look at the offset of the conduction bands (LUMOs) between the donor and acceptor in order to determine if sufficient driving force exists for charge transfer to occur. To a first approximation, the band offset must be greater than the exciton binding energy (generally estimated as the difference between the optical and electrochemical band gap). Although this is the common method used to approximate if charge transfer will occur between a given donor-acceptor pair, this is an oversimplification of the process as it ignores the simple fact that the excited electron in the donor exciton does not have the exact same energy as if it were in the LUMO of the ground state donor. Stated more precisely, electron transfer from donor to acceptor can occur when the conduction band (or LUMO level) of the acceptor is sufficiently lower in energy (as determined by the exciton binding energy) than the excited electron in the donor exciton. 1c However, in order to formally answer the question of whether or not electron transfer will occur from a donor (D) to an acceptor (A), one must first focus on the thermodynamic considerations of the reaction shown below. D+A DA+Go = -nF( EoA/A EoD /D) = -RTlnKK + + Here an equilibrium exists between the forward process of electron transfer and the reverse process of back electron transfer which is governed by the equilibrium constant
25 K. The simple thermodynamic answer to the question is thus to look at the difference in the standard electrode potentials (E o ) for the reduction of A and the oxidation of D, or equivalently to calculate the G o for the reaction, as shown above. It is generally accepted that electron transfer from donor to acceptor will occur (to the extent allowed by the equilibrium constant K) if the difference in standard electrode potentials is more positive than .4 V (K > 10 -7 ) and that the reaction will proceed to completion if the E o is more positive than 0.4 V. 42 This thermodynamic model ignores the kinetics of the reaction, and it is simply assumed that the reaction will be rapid as long as the thermodynamic conditions for charge transfer are met. This model also ignores the difference in the excited state energies of the donor and acceptor (D * and A * ) relative to the ground states (D and A). Nonetheless this model serves as a more sophisticated first approximation for the determination of which donor-acceptor pairs might operate efficiently in photovoltaic devices. Measured ionization potentials, electron affinities, experimental E 1/2 values, and calculated HOMO or LUMO energies are often used in place of the standard electrochemical potentials of the donor and acceptor materials. Thus it is often difficult to calculate the precise driving force for electron transfer. Back electron transfer must also be considered, although this process is less important under the influence of an electric field in a photovoltaic device. Much more sophisticated and complete descriptions of electron transfer have been developed for charge transfer in organic systems. However, such descriptions exceed the scope of the present discussion. As such, electron transfer will be assumed possible if the band offset is greater than the exciton binding energy.
26 The basic device structure for an organic solar cell is as shown in Figure 1-13. While both dye sensitized solar cells and organic devices are excitonic solar cells, the major difference between the two is that in organic solar cells, excitons must first diffuse to an interface before dissociation can occur. Estimates of the exciton diffusion lengths in polycrystalline films of perylene bis(phenylethylimide) place the value as high as 1.8 m. 43 As such, the main processes that must occur within an organic solar cell are: light absorption, exciton diffusion, and charge transfer. Once charge transfer has occurred, free charge carriers (holes and electrons) must then be transported through the respective layers of the device for collection at opposite electrodes. In this case ITO serves as the anode (as defined under forward bias) and a low work function metal (e.g. aluminum) serves as the cathode. Organic solar cells based on small molecules have improved considerably since the seminal work of Tang in 1986. The best reported devices today have efficiencies approaching 6%. 44 The most efficient single heterojunction devices (see Figure 1-13) are based on sequentially thermally evaporated layers of donor and acceptor molecules on an ITO substrate with a thermally evaporated aluminum counter electrode. In one of the best reported devices, copper phthalocyanine is used as the donor and fullerene (C 60 ) is used as the acceptor in order to achieve 3.6% power conversion efficiency under AM1.5 (100 mW/cm 2 ) irradiation. 45 Current work in this area is focused on the development of multiple heterojunction (tandem) devices in order to improve the efficiency of the devices. 46 While the fabrication of organic solar cells based on small molecules offers the potential for cheaper device production relative to silicon devices, the fact that thermal
27 deposition of small molecules under high vacuum is required is a drawback. Ultimately, device fabrication techniques such as spin coating or inkjet printing offer a more attractive route to low cost solar cells. For this reason much interest in organic solar cells has been focused on polymer-based devices. Figure 1-13. Schematic representation of a two layer organic solar cell. Polymer-based solar cells. Among all the organic photovoltaic devices, those based on conjugated polymers are perhaps the most attractive from an application standpoint based on the materials properties of polymers. â€œPlasticâ€ photovoltaics present the possibility of flexible, large area devices that can be solution processed and printed in a roll-to-roll fashion, which could result in significant cost-savings relative to silicon technologies. 47 Based on these potential advantages, a growing research effort has been directed toward realizing such devices over the past decade. While the best reported devices are only able to achieve ~4% power conversion efficiency, the growth of this field and the development of a fundamental understanding of the parameters associated with polymer PVDs has been remarkable. As with small molecule organic solar cells, conjugated polymer-based devices are excitonic solar cells. The origins of the polymer solar cell field are based on single layer
28 devices in which the photoactive material (a film of a conjugated polymer in this case) is sandwiched between two electrodes of different work function. The first reported observations of the photovoltaic effect in organic polymers appeared in the 1960s. 48 These early reports focused on the photoconductivity of conjugated polymers such as poly(phenylenediacetylene), although no definitive explanation could be proposed at the time for the observed photoconductivities in these systems. 49 It was not until well after the birth of the modern conducting polymer field 50 that significant efforts were made to develop and understand the photovoltaic effect in conjugated polymers. In the early 1980s limited efforts were focused on developing single layer devices based on polyacetylene 51 or electrochemically grown polythiophene and poly(3-methylthiophene). 52 , 53 However, by the end of the 1980â€™s little progress had been made in developing conjugated polymer photovolatics . 54 After the first reports of electroluminescence in conjugated polymers, 55 efforts began to examine the photovoltaic characteristics of the polymer diodes. Indeed, much work had already been done concerning the process of photogeneration of charges in conjugated polymers, especially for PPV. 56 ,, 57 58 From this early work it was deduced that photogenerated excitons in a conjugated polymer can be dissociated to yield free charge carriers in photodiodes with two electrodes of different work function. In the earliest thorough study of a PPV single-layer PVD, 59 100 nm thick films were sandwiched between ITO and a low work-function electrode (Ca, Mg, or Al). While V oc values of more than 1V were observed, photocurrents of only a few microamperes limited the devices to maximum external quantum efficiencies (IPCE) of less than 1%. This low level of performance was problematic for all of the early single-layer photovoltaic
29 devices. 57,, 60 61 With such devices, the overall power conversion efficiency is generally limited to the order of 10 -3 â€“ 10 -2 %. 62 A significant improvement in device performance occurred as a result of a parallel research track concerning the process of photoinduced electron transfer from conjugated polymers to fullerene (C 60 ). The first reported observation of photoinduced electron transfer from a conjugated polymer to C 60 appeared in 1992 63 for the case of an MEH-PPV-C 60 composite (see Figure 1-1 for the structure of MEH-PPV). This observation quickly led to the incorporation of C 60 into polymer photovoltaic devices. In the earliest such report, 64 heterojunction diodes were fabricated by first spin coating MEH-PPV onto ITO and then vacuum evaporating C 60 followed by gold as the back contact. In this case, a power conversion efficiency of only 0.04% was reported (at 514.5 nm), however the spectral dependence of the photocurrent (IPCE) strongly paralleled the absorption spectrum of MEH-PPV, indicating that photoexcitations of MEH-PPV contributed to the photocurrent. While the devices presented in this work gave only limited performance, the concept of using an acceptor in conjunction with the conjugated polymer brought the polymer PVD field up to the point found in the work of Tang 39 concerning small molecule heterojunction devices. Upon incorporation of the powerful electron acceptor C 60 into a polymer device, a driving force for exciton dissociation is introduced. Importantly, the sub-picosecond time scale required for photoinduced charge transfer from polymer to C 60 , is fast enough to suppress many loss mechanisms (e.g., photoluminescence) that impede photovoltaic performance. One important observation, even in this seminal work, is that the length scale for exciton diffusion is extremely small in conjugated polymers (estimated to be on the order of only a few angstroms in this early
30 work), indicating that charge transfer occurs only at the donor-acceptor interface. As such, in this two-layer device, most of the excitons generated in the MEH-PPV layer do not give rise to free electrons and holes. A positive point is that the use of two layers exhibiting differing electron mobilities creates a means for efficiently removing the few photogenerated charges. After the initial report of C 60 -based polymer PVDs, a number of groups began to explore the use of heterojunction devices in the mid-1990s. Using two-layer devices based on PPV and sublimation-deposited C 60 , single wavelength efficiencies as high as 9% were observed. 65 Another notable two-layer approach was the use of donor-type polymers and acceptor-type polymers sequentially spin coated onto a substrate. Here selective solubility of the two polymers was critical in order to avoid re-dissolution of the initially spin-coated layer. In the work of Zakidov and Yoshino, et al. 66 poly(p-pyridyl vinylene) was spin coated from formic acid, followed by spin coating of poly(3-hexylthiophene) from chloroform. Here device performance was quite poor, with FF = 0.23 and Jsc = 6.0 A/cm2, although significant improvements were observed when compared to single layer devices based on poly(3-hexylthiophene) alone. During this period, the group of Richard Friend reported the most efficient two-layer device in 1998. 67 In this case the electron deficient MEH-CN-PPV (see Figure 1-14) was used as the acceptor material and POPT (poly[3-(4-octylphenyl)thiophene]) was used as the donor. This work extended the concept of the two-layer device by using a donor layer that consisted of a blend of 95% POPT and 5% MEH-CN-PPV, while the acceptor layer was formed from a blend of 95% MEH-CN-PPV and 5% POPT. The donor and acceptor layers were spin coated onto ITO and calcium or aluminum respectively. The
31 donor half of the device was then heated to 200C and the acceptor half was applied with light pressure. Heating of the donor half in the pre-lamination step not only made lamination possible, but the thermal annealing of POPT served to induce longer wavelength absorption based on the formation of a more ordered structure. 68 The overall power conversion efficiencies of such devices reached 1.9% under AM1.5 conditions with external quantum efficiencies of 29% at 480 nm. These devices far outperformed analogous two layer devices consisting of pure donor and acceptor layers as well as single layer blends of donor and acceptor. The benefit of incorporating a small amount of the acceptor into the donor layer (and vice-versa) is that excitons can be dissociated within the layers. However, due to the incorporation of only 5% of the minority phase, limited pathways exist for removing both electrons and holes from a single layer. Perhaps the key aspect of these devices is the diffuse interface that is undoubtedly produced during the lamination process. In this case acceptor and donor materials are expected to form an intermixed network at the heterojunction. Here, the enhanced performances relative to a single layer blend device of the same donor and acceptor materials are based on the specific properties of POPT, which forms a highly ordered structure upon thermal annealing that allows for long wavelength absorption and enhanced charge-carrier mobility. With a 50/50 blend of the two components in a single layer, macrophase separation (loosely defined as phase separation on a scale large enough that it impedes exciton diffusion and dissociation as well as charge transport) occurs upon annealing and the device operates poorly.
32 NC CN O O O O nMEH-CN-PPVS nPOPT Figure 1-14. Chemical structures of MEH-CN-PPV and POPT. While the work of Friend described above illustrated a highly effective two-layer approach to polymer solar cells, it also indicated that an intimate mixing of donor and acceptor materials was a critical factor in maximizing the device efficiencies. This is directly related to the limited exciton diffusion length in conjugated polymers and the necessity of a donor-acceptor interface for effective exciton dissociation. As a consequence, those devices with the maximum donor-acceptor interfacial area and the thinnest polymer layer (within the limit of not compromising light absorption) or smallest polymer domains are expected to be the most efficient. As described below, this realization gave rise to the â€œbulk-heterojunctionâ€ technique of device fabrication in which the donor and acceptor are blended together to give a bicontinuous (percolated) network of donor and acceptor. For the past ten years this technique has been the most successful technique employed in the fabrication of plastic solar cells. 69 Before the name bulk-heterojunction was developed to describe devices based on donor-acceptor blends, the technique of blending a conjugated polymer with and acceptor had been utilized to study photoinduced charge transfer as well as for the fabrication of photovoltaics. The earliest studies concerning photoinduced charge transfer from conjugated polymer to C 60 were performed on composite films cast from blends of
33 conjugated polymer and C 60 . 63,,, 70 71 72 The first report of a polymer photovoltaic device based on a polymer-C60 blend appeared in 1994 73 and can be considered the precursor of the bulk-heterojunction technique. In this work, photodiodes were fabricated based on single layer MEH-PPV-C 60 blends sandwiched between ITO and calcium electrodes. The active layer was cast from solutions that were 90/10 (w/w) MEH-PPV/C 60 in xylenes. While the devices were not thoroughly evaluated for their photovoltaic performance, it was observed that the photoresponse under short-circuit conditions was enhanced by more than an order of magnitude relative to devices made from pure MEH-PPV .61 In developing the bulk heterojunction concept, two parallel approaches set the standard. The first approach is based on blends of donor and acceptor conjugated polymers and the second approach is based on blends of donor polymer and solubilized fullerene as the acceptor. The groups of Richard Friend 74 and Alan Heeger 75 reported the first approach almost simultaneously, but independently in 1995. In both cases, blends of MEH-PPV and CN-PPV (see Figure 1-1) were used as the active layer in the devices. The introduction of the cyanovinylene moiety in CN-PPV was known to increase the electron affinity of PPV derivatives and thus improve its electron transport properties. 76 In both papers, photoluminescence quenching in blends of MEH-PPV and CN-PPV was deemed indicative of photoinduced charge transfer. While neither group reported an AM1.5 power conversion efficiency for the devices, under short-circuit conditions, Heeger reported a power conversion efficiency of 0.9% at 430 nm under microwatt illumination and Friend reported an external quantum efficiency of 6% at 550 nm under milliwatt illumination levels. The lack of a uniform standard for device evaluation detracts from the impact of these results. Perhaps more important than the quantitative
34 results presented in these two papers were the concepts and qualitative observations brought forth. Heeger clearly put forth the bulk-heterojunction concept in this work and actually coined the term â€œbulk-heterojunction materialâ€ to describe the blend of donor and acceptor polymers as applied to photovoltaic devices. Both groups studied a variety of blend compositions and while neither group presented evidence of a bicontinuous network, Friend showed that phase separation did occur over a broad range of weight ratios from 7:1 to 1:10 of MEH-PPV and CN-PPV respectively, using TEM with selective staining of the MEH-PPV phase. One interesting observation by Heeger was that the magnitude of the V oc in such devices was not consistent with the work function difference between the two electrodes in the device as had been observed for single-layer polymer devices. While Heeger proposed a preliminary explanation of this phenomenon, future work would more fully elucidate the origin of the V oc in bulk heterojunction devices. The next step in the development of polymer solar cells was the introduction of solubilized fullerene derivatives as electron acceptors in bulk heterojunction devices. The development of such devices was first reported by Heeger and Wudl in 1995 77 and established the platform for device fabrication that is still the most reproducibly effective method for high performance devices even ten years later. Here the solubilized fullerene derivative PCBM 78 ([6,6]-phenyl C 61 -butyric acid methyl ester) was used in blends with MEH-PPV as seen in Figure 1-15. This soluble fullerene derivative was developed in the group of Wudl, and it is not altogether clear whether PCBM was synthesized specifically for use in bulk heterojunction devices or if this application was a serendipitous use of the new molecule. Nonetheless, the use of soluble fullerenes presented a real advantage over
35 any other polymer based solar cell known at the time. Fullerene had long been known to be an extremely effective electron acceptor in the presence of photoexcited conjugated polymers, but the low solubility of C 60 in organic solvents had limited its use in devices to either two-layer structures in which the C 60 had been vacuum deposited or to blend devices in which the concentration was limited by insolubility or the tendancy of C 60 to aggregate and crystallize. With PCBM however, blends with a conjugated polymer could incorporate as much as 80% of the fullerene by weight. In this seminal report devices consisting of 1:4 MEH-PPV:PCBM were found to give the best performance with a 2.9% power conversion efficiency at 430 nm (20 mW/cm 2 ). Figure 1-15 illustrates the concept of a bicontinuous network of donor and acceptor in a bulk heterojunction device. Figure 1-15. The bulk heterojunction concept for polymer solar cells. In this case the donor is MEH-PPV and the acceptor is PCBM. In the device diagram, note that current flow is in the direction opposite of electron flow.
36 Using this same platform of blends consisting of 80% PCBM and 20% conjugated polymer as the active layer, continuous improvement over the last 10 years has led to devices with 2.5% (AM1.5) 79 efficiency based on MDMO-PPV (see Figure 1-1) and PCBM as well as devices with efficiencies of 2.9% (AM1.5) 80 based on MEH-PPV and PCBM. Also, devices based on poly(3-hexylthiophene) (P3HT) have reached efficiencies of 3.5% (AM1.5) 81 in blends with PCBM and devices based on MDMO-PPV and PCBM-C70 have reached efficiencies of 3.0% (AM1.5). 82 The highest reported efficiency for a device is 3.85% for a P3HT/PCBM device. A tremendous research effort has been devoted to the optimization of these devices and improvements in the understanding of the device physics, device engineering, and materials chemistry have led to these advances. In the following, an overview will be presented of the key factors in device optimization that led to the 2-4% percent efficiencies that have become commonplace. Additonally, those factors still in need of optimization will be discussed. In the attempts to optimize polymer-based solar cells, researchers have focused on the four fundamental processes that occur in such devices: light absorption, charge transfer, charge transport, and collection, as illustrated in Figure 1-16. In order to optimize conjugated polymer-PCBM bulk heterojunction devices it is imperative to gain a detailed understanding of all the processes that are occurring. The first step in the photovoltaic process is the absorption of light. The conjugated polymer is the primary source of light absorption across the visible portion of the solar spectrum, while PCBM absorbs most strongly in the UV. A good overlap with the solar spectrum is desired (see Figure 1-10) and the peak of photon flux form the sun occurs at 1.8 eV (~700nm). Thus a polymer with a band gap of < 1.8 eV is desired for use in these devices. However MEH
37 PPV and P3HT have band gaps of 2.2 and 2.0 eV respectively, suggesting the need for other polymers or approaches to increase the number of absorbed photons. LUMOHOMO h aDonor Exciton b Exciton Diffusion Donor(Conjugated Polymer)Donor ExcitonDonor ExcitonAcceptor(PCBM)LightAbsorptionc e Charge Transferd AcceptorRadical AnionCharge Carrier DonorRadical CationCharge Carrier Charge TransportHole TransportLayerCharge TransportElectron TransportLayereCharge collection: electrons at the cathode, holes at the anode Figure 1-16. Schematic illustration of the four fundamental processes occurring in polymer based solar cells. Once light has been absorbed, charge transfer can then take place at the interface between a donor and acceptor. While the process of charge transfer from conjugated polymers (with sufficient donor character) to C 60 is essentially 100% efficient, the primary limitation for charge transfer is the exciton diffusion length of the polymer, which is on only the order of 5-10 nm. 83 The major concern here is thus the length scale
38 of the phase-separation between polymer and PCBM, as polymer domains larger than ~100 nm will result in the generation of large fractions of excitons that never effectively dissociate. Closely tied to charge generation is charge transport and collection. Again, this is largely a morphological issue to a first approximation. Bicontinuous networks of donor and acceptor are required to effectively transport charges to the electrodes (holes in the donor polymer and electrons in the PCBM acceptor phases). On a deeper level, the charge carrier mobilities in the polymer and PCBM phases are critical for enhancing the charge transport and avoiding charge recombination. Finally, the process of charge collection is an interfacial process in which charges must be effectively transported from one medium to another. Optimization of electrode properties has greatly enhanced the efficiency of devices by improving this process. The importance of the interfacial processes in these devices is the key to their operation. Looking at these processes in greater detail, there are several means by which the number of photons absorbed by a device can be increased. The most simplistic approach is to simply increase the thickness of the active layer. Conjugated polymers have very high absorption coefficients (~10 4 -10 5 cm -1 ) and thus a film of P3HT, for example, will absorb 95% of the photons in the wavelength range of the polymer absorption with a film thickness of only 240 nm, as compared to silicon solar cells which are ~100 m thick. 84 However, not only will the absorption of photons saturate when a critical thickness is reached, but charge carrier mobility and resistances in the films become important when the polymer device active layers exceed an optimum thickness of ~100 nm. 85 The most useful approach is to pursue narrow band gap polymers (alternately referred to as low
39 band gap polymers), which will more effectively blanket the solar spectrum and thus absorb photons over a much broader range of wavelengths. 86 Several reports have appeared concerning the use of narrow band gap polymers (defined here as polymers with an Eg < 1.8 eV) in photovoltaic devices. 8 ,87 The most efficient devices reported to date are based on 1:4 blends of the alternating polyfluorene copolymer PFDTBT (poly(2,7-(9-(2ethylhexyl)-9-hexyl-fluorene)-alt-5,5-(4,7-di-2-thienyl-2,1,3-benzothiadiazole)) 87d (Figure 1-17) and PCBM, which give power conversion efficiencies of 2.2% (AM1.5). While this donor-acceptor copolymer and its analogues are finding increasing success in photovoltaic devices, with a band gap of ~1.9 eV, PFDTBT is not technically a narrow band gap polymer even though it is touted as such. A true narrow band gap polymer that has shown reasonable success in solar cells is PTPTB (poly(2,5-bis(2-thienyl)-N-dodecylpyrrole-co-2,1,3-bezothiadiazole) (shown in Figure 1-17) which possesses a band gap of 1.6 eV and gives ~1% efficiency (AM1.5) in 1:3 blends with PCBM. 87e However, no devices based on narrow band gap polymers have been reported to exceed the value of 2.5% which is taken as a benchmark for MDMO-PPV:PCBM solar cells. Several other approaches have been taken to broaden the wavelength range over which photons are absorbed. These approaches include the addition of long-wavelength absorbing dye molecules to the polymer-PCBM blend, 88 as well as the use of several polymers of varying band gaps blended with PCBM. 88b Ultimately though none of these approaches to broaden the spectrum of absorbed photons has yet to rival the PPV:PCBM platform.
40 S NSN S nPFDTBTS N C12H25 S NSN R NSN R nPTPTBR = H or Br Figure 1-17. Structures of narrow band gap polymers used in polymer / PCBM solar cells. After light absorption, the main processes in polymer-PCBM photovoltaic devices are charge transfer, followed by charge transport and collection. The complications for charge transfer arising from the minimal exciton diffusion length in conjugated polymers have already been discussed. However, from exciton dissociation to charge collection, the device performance is governed by the nature of the interfaces within the device as well as the charge transport properties of the various materials. The interfaces include the donor-acceptor interfaces within the polymer-PCBM blend, as determined by the blend morphology, as well as the interfaces between the photoactive layer and the electrodes. Ultimately, the morphology of the donor-acceptor blend is the critical parameter on which the success of the bulk heterojunction methodology hinges. If the active layer blend is engineered to give a bicontinuous network in which domain sizes facilitate the ~10 nm polymer exciton diffusion length and allow for effective charge transport to the electrodes, then a device has a chance to operate efficiently. If these requirements are not fulfilled, the device efficiency will certainly suffer. While numerous parameters affect the morphology of a polymer-PCBM blend, those that are the most simple to control are the blend composition (weight ratio of components), the solvent used for depositing the film, and the means and conditions employed for depositing the film, with spin coating being the most widely employed method. Most efforts toward the optimization of these
41 parameters have focused on blends of MDMO-PPV and PCBM. In the first report of a 2.5% efficient device, 79 striking evidence of the dramatic effect of solvent choice was presented. In this case devices were fabricated by spin coating a 1:4 blend of MDMO-PPV:PCBM from either toluene or chlorobenzene. For the toluene device, efficiency peaked at ~0.9% (AM1.5) with short circuit current densities on the order of 2.3 mA/cm 2 . However when the same composition was spin coated from chlorobenzene, devices gave efficiencies of 2.5% with short circuit current densities of 5.3 mA/cm 2 . The dramatic increase in efficiency was directly related to the morphology of the blends as studied by AFM. In this case it was found that the toluene-cast film had surface features with horizontal dimensions on the order of 0.5 m while the largest features in the chlorobenzene cast films were on the order of 0.1 m. Additionally, the toluene film had a surface roughness of about 10 nm while the chlorobenzene film had a surface roughness of only 1nm. All this points to the fact that the chlorobenzene film displays a more intimately mixed morphology with smaller domain sizes than in the toluene cast film. The enhancements are attributed to the better solubility of PCBM in chlorobenzene, which suppresses aggregation. It is postulated by the authors that large clusters of PCBM reduce the electron mobility in the devices, as continuous pathways for electron transport are sparse. Additionally, hole mobility in the polymer phase is affected by the choice of solvent as films of MDMO-PPV show twice the mobility of films cast from toluene. The enhanced FF in the chlorobenzene device (0.61 vs. 0.50 for toluene) is further evidence of enhanced charge carrier mobilites and reduced resistances in the active layer. Overall
42 this study was important for describing the importance of morphology and charge carrier mobility on device performance. Since this original work by Sariciftci, a great deal of work has been done recently to study the morphology in MDMO-PPV:PCBM devices in detail. 89 Using TEM, AFM, and secondary ion mass spectrometry Janssen et al. 89c studied the morphology evolution of MDMO-PPV:PCBM blends (cast from chlorobenzene) containing PCBM concentrations between 33 and 90% by weight. In this case it was found that with up to 50% PCBM, no phase separation was observed. With more than 67% PCBM, a nanoscale phase separation is observed in which domains of PCBM are found within a relatively homogeneous matrix of 50:50 MDMO-PPV:PCBM. Devices based on these blends showed that device performance (efficiency, FF, IPCE peak currents) increased for blends with 67-80% PCBM, but reached a maximum at 80% PCBM followed by a drop-off at higher PCBM concentrations. The enhanced device performance between 67 and 80% PCBM is attributed to optimization of the blend morphology with PCBM domain sizes increasing from 40-65 nm (AFM) at 67% PCBM to 60-80 nm at 80% PCBM. Beyond 80%, PCBM the domain sizes increase to ~110-200 nm and the optical density of the films becomes so high that light does not effectively penetrate the thin films. This work has been supported by recent work form the Sariciftci group 89a using AFM and SEM, to show that PCBM forms nanoclusters (~ 20 nm diameter) in a continuous matrix when films are cast from chlorobenze with 67-80% PCBM in MDMO-PPV. Such studies also indicate that annealing of these films results in macrophase separation as the polymer chains become highly mobile and the PCBM clusters tend to aggregate. This
43 suggests a possible degradation mechanism for the performance of polymer-PCBM devices under continuous illumination where thermal stability may be a problem. 89d As identified in the preceding discussion of morphology, the charge transport characteristics are strongly affected by the film morphology. Once charge transfer has occurred, charges must then be transported to the respective electrodes under the influence the internal electric field of the device. Thus, charge carrier mobility in the donor and acceptor phases is the critical parameter. An important goal towards the improvement of polymer PVDs is to increase and balance the charge transport characteristics of the two phases. An electron mobility for PCBM 90 has been measured to be 2 x 10 -7 m 2 V -1 s -1 , while measurements of the hole mobility in MDMO-PPV 91 have revealed a value of ~5 x 10 -11 m 2 V -1 s -1 . As such, it is necessary to find methods of increasing the hole mobility in MDMO-PPV or finding new higher mobility polymers, as the unbalanced charge transport limits the ultimate current that can be produced by the device. Recent measurements of the hole mobility of 1:4 MDMO-PPV:PCBM indicate a value of 2 x 10 -8 m 2 V -1 s -1 , which suggests that the transport properties in devices may not be as unbalanced as previously thought. 92 Nonetheless, higher mobility polymers are thought to be one way to ultimately improve the efficiency of polymer PVDs. The highest reported hole mobility in a conjugated polymer is 1 x 10 -5 m 2 V -1 s -1 for regioregular P3HT. 93 Recent efforts on the production of regiospecific MDMO-PPV have yielded polymers with hole mobilities 3.5 times greater than regiorandom MDMO-PPV. 94 Devices based on these regiospecific polymers gave 2.65% efficiency as opposed to the 2.5% efficiency with regiorandom polymer. More impressive than this modest improvement in efficiency is the improvement in FF from 0.61 to 0.71, indicative of a
44 decrease in the series resistance of the active layer which can be directly attributed to the enhanced hole mobility in the regiospecific MDMO-PPV. One additional practical concern about mobility is the process of charge recombination. Even in a device with ideal blend morphology, charge carrier recombination will occur to a certain extent as determined by the mean carrier drift length, which is based on the charge carrier lifetime. Measurements have shown that the charge carrier lifetime in MDMO-PPV:PCBM blends is on the order of microseconds to milliseconds. 95 In principle, recombination is only a serious problem if the mean carrier drift length is less than the thickness of the film. However devices do not operate like theoretical models; recombination does occur and this parameter suggests another constraint on the optimal thickness of the photoactive layer and emphasizes the need for polymers with high hole mobilities. In the broader sense of charge transport, the device as a whole must be considered. Thus the role of the electrodes and the interfaces between the active layer and the electrodes must be considered in addition to the transport properties within the active layer. Modification of the electrodes in polymer-PVDs has led to considerable enhancements in device performance over the last 10 years. In the first report of MEH-PPV:PCBM bulk heterojunction devices by Heeger in 1995, the active blend was cast directly onto ITO coated glass and then a layer of aluminum or calcium was vacuum deposited on top of the blend. Direct contact between the polymer blend and the electrode has been shown to be problematic for several reasons. For ITO, the work function is ~4.7 eV (see Figure 1-18). In order to more effectively align the electrode work function with the HOMO energy of the polymer and thus facilitate charge transfer by reducing contact
45 resistance, ITO is most often coated with PEDOT-PSS [poly(3,4-ethylenedioxy-thiophene)-polystryenesufonate] which is found to have a work function of ~5.1 eV. 96 The other role of PEDOT-PSS is to smooth the ITO layer and prevent shorts often caused by the rough ITO surface. 88b PEDOT-PSS has been shown in many cases with small molecule and polymer organic PVDs to improve the efficiency of the device. 62,80 Another critical parameter of the ITO (anode) is the conductivity. If devices are to be able to deliver large currents, highly conductive electrodes are required. It has been shown that decreasing the surface resistance of ITO from 50 / to 12 / can increase the efficiency of MEH-PPV:PCBM devices from 1.4% to 2.9% (AM1.5), and the Jsc from 5.8 mA/cm 2 to 8.4 mA/cm 2 in devices in which both types of ITO were coated with PEDOT-PSS. 80 Currently ITO is the most commonly used transparent electrode in organic and polymer PVDs. The cathode (or back electrode) also exerts an influence on the performance of polymer PVDs. The most commonly used cathode material is aluminum. It has been found in the preparation of polymer LEDs that applying a thin layer of LiF between the polymer and the aluminum improves the properties of devices 97 and the same observation has been made with polymer PVDs. While the LiF layer has been shown to dramatically improve electron transport across the interface, the exact mechanism of operation is still not fully understood, although several explanations have been postulated. 98 An interesting explanation is proposed by Spanggard based on the formation of a ~30 insulating layer that is observed to form at the interface between PPV polymers and aluminum electrodes when no LiF is used. This layer is thought to be formed by aluminum atoms that diffuse through the polymer matrix and react with vinylene linkages in the polymer, disrupting
46 the conjugation and ultimately rendering the polymer insulating. It has been proposed that with a thin layer of LiF (5 ), an aluminum fluoride compound is formed at the interface with a small amount of concomittant n-type doping of the polymer by the Li + , which enhances the conductivity at the interface. 99 Regardless of the mechanism, the addition of LiF to devices has been reported to result in as much as 20% relative increase the power conversion efficiency. 100 In addition to the effects of the protective LiF layer, the nature of the metal electrode is also important. Several groups have done studies to evaluate the consequences on device performance of using cathodes of different work functions. 101 In this case, conflicting views exist concerning the effect of the cathode material on the device parameters, specifically the open circuit voltage. Such experiments have also raised questions about the fundamental origin of the open circuit voltage in bulk heterojunction devices. In the simplest model of a single-layer polymer PVD, the V oc is ascribed to the difference in the work functions of the two electrodes. The situation however, becomes more complicated with heterojunction devices (two-layer or bulk heterojunction) as the HOMO and LUMO energies of the donor and acceptor begin to play a role along with the work function of the electrodes. 101 ,102 One school of thought purported by the Sariciftci group is that the nature of the cathode has little influence on the magnitude of the open circuit voltage in bulk heterojunction solar cells. 101d This conclusion is based on a comparative study of MDMO-PPV:PCBM devices in which the work function of the cathode was varied from 2.87 eV with calcium to 5.1 eV with gold in a device with a PEDOT-PSS/ITO anode. The result was approximately a 150 mV decrease in V oc for the gold device relative to the calcium device despite a 2.2 eV shift in
47 the work function of the cathode. As such, it was concluded that the work function of the cathode was relatively less important in bulk heterojunction devices. In the same work the primary source of the V oc was attributed to the difference in energy of the donor HOMO and the acceptor LUMO. This conclusion was based on the observed dependence of the magnitude of the V oc on the acceptor strength of several soluble fullerene derivatives. It was proposed that the Fermi level of the cathode is pinned by the LUMO of the acceptor. Similar work has shown that the V oc of bulk heterojunction devices is likewise controlled by the HOMO energy (oxidation potential) of the donor polymer. 102a The conclusion of the Sariciftci group is in contrast to the work of Mihailetchi 101a which indicates a strong correlation between the work function of the back electrode and the V oc of the device, evidenced by 0.5 V decrease in the V oc of MDMO-PPV:PCBM devices with a Pd electrode relative to the same device with a LiF/Al electrode. Regardless, it is clear that the identity of the electrode as well as the band structures of the active materials play a role in determining the V oc of a device. This is important as the V oc of a device is directly related to its maximum power output (P = IV). Figure 1-18 gives an overview of the optimized device structure based on the advances in device engineering that have occurred over the past 10 years. Based on the previous discussion of all the parameters affecting the characteristics of polymer-based solar cells, several things are clear. First, it is evident that a complete understanding of the operation of these devices has not been reached. However it is clear that a complex set of interrelated parameters govern the device operation. Beyond the absorption of light and the dissociation of electrons, optimizing charge transport is found time and again to be the critical factor and one that is tied intimately to I sc , FF, and ultimately the power
48 conversion efficiency. Thus reducing resistances in the active layer as well as at the interfaces and increasing and balancing hole transport with electron transport will lead to devices with higher fill factors and efficiencies. PCBM CBPolymer CBPolymer VBPCBM VB Anode (ITO)Cathode (Al)ITOF = 4.7 eV h+eAlF = 4.3 eVLiF (~5A) PEDOT-PSS F = 5.1 eVPEDOT-PSS h+ e-h+ h 123344 Figure 1-18. Schematic representation of the processes in an optimized bulk heterojunction solar cell. (1) Light absorption. (2) Exciton diffusion and charge transfer. (3) Charge transport. (4) Charge collection as facilitated by the buffer layers LiF and PEDOT-PSS on the cathode at the cathode and anode respectively. As an example of a highly efficient device, devices based on P3HT and PCBM have gained interest over the past few years as a result of the report of a 3.5% efficient device by Sariciftci. This device utilizes all the advances of the last 10 years and is made up of a PEDOT-PSS coated ITO anode on which the active layer (P3HT and PCBM in an unreported ratio) is cast, before a LiF/Al electrode is vacuum deposited. When the device is measured as fabricated, an efficiency of 0.4% (AM1.5) is found along with a V oc of 0.3V, a J sc of 2.5 mA/cm 2 , and a FF of 0.4. In this case, the IPCE shows a maximum of ~25% at 420 nm. If the device is instead annealed at 75C for several minutes after deposition of the LiF/Al electrode, the V oc increases to 0.5 V and the J sc increases to 7.5 mA/cm 2 . The result is a device with an efficiency of 2.5% and a FF of 0.57. It is
49 suspected that annealing the blend results in an enhanced ordering of the P3HT and a concomitant increase in the hole mobility of the polymer. An increased crystallinity of P3HT has been observed previously upon annealing, 103 which supports this assertion. Additionally, enhanced field-effect mobilities have been observed in thermally annealed samples of P3HT relative to samples that were not annealed. 104 Evidence for the ordering of the polymer phase in these P3HT:PCBM devices is found in a definite red-shift in the IPCE, which shows a maximum of 60% at 500 nm, and an IPCE of nearly 30% at 625 nm. This red-shift in the IPCE is attributed to the appearance of the suspected highly ordered polymer phase, which gives rise to new long wavelength absorptions in the polymer. Alternatively, it was observed that if the P3HT:PCBM device is thermally annealed while simultaneously being subjected to an external voltage greater than the V oc (2.7 V), the device gives an efficiency of 3.5%, a V oc of 0.56 V, a J sc of 8.5 mA/cm 2 , and a FF of 0.6. Here the IPCE shows a maximum of 70% at 500 nm. The authors assume that the application of a bias results in an enhanced ordering of the polymer chains in the direction of the applied electric field, although no additional evidence is reported to support this claim. Previous work on LEDs has shown that such electrical annealing can be effective for enhancing the efficiency of the devices. 105 In the case of LEDs, it is proposed that ionic impurities in the polymer film play a role in this electrical annealing, as they are thought to diffuse to the electrode interfaces upon application of a bias and enhance charge injection at the electrodes. The ultimate cause of the enhanced efficiency in P3HT:PCBM solar cells upon electrical annealing remains unknown. An important point to stress is that for these P3HT:PCBM devices (as with most devices reported in the
50 literature) fabrication and testing is carried out entirely in an argon filled glovebox. The stability of these devices under ambient conditions is quite poor, and is one of the major limitations as these device move toward use in applications. More recent reports of analogous P3HT:PCBM devices indicate efficiencies of ~4% (AM1.5) with IPCE values peaking at 75%, which is the highest IPCE value reported for a polymer-PCBM device. Little detail is provided concerning the specifics of device construction for the 3.5% or the ~4% P3HT:PCBM devices. For example, in the case of the 3.5% device discussed above, the weight ratio of P3HT and PCBM is not reported and neither is the source or the percent regioregularity of the P3HT. A more detailed study 106 has shown that a 1:1 blend of P3HT and PCBM gives 3% (AM1.5) efficient devices after thermal annealing using P3HT with 96% regioregularity. Such devices represent the forefront of polymer-PCBM device technology. Table 1-1 provides a summary of some of the most efficient polymer-PCBM devices reported in the literature to date. Table 1-1. Solar cell characteristics for the best polymer-PCBM devices reported in the literature. Blend Composition V oc (V) J sc (mA/cm 2 ) FF (%) MDMO-PPV / PCBM (1/4) 0.82 5.25 0.61 2.5 MEH-PPV / PCBM (1/5) 0.87 8.4 0.40 2.9 MDMO-PPV / PCBM (C 70 ) (1/4.6) 0.77 7.6 0.51 3.0 P3HT / PCBM 0.55 8.5 0.6 3.5 P3HT / PCBM ---15 ---3.85 P3HT / PCBM (1:1) 0.61 9.4 0.53 3 All reported values were measured under AM1.5 conditions.
51 While the polymer-PCBM platform has totally dominated the field for the past 10 years, with most efforts focused on MEH-PPV and MDMO-PPV, several other approaches are gaining interest. As mentioned earlier, a push toward lower band gap polymers for more effective photon harvesting is underway. Additionally several other types of polymers are under investigation for use in polymer solar cells. An interesting approach has been the development of block copolymers that consist of a donor block and an acceptor block. Most notable is an approach based on the formation of PPV-polystyrene copolymers bearing pendant C 60 molecules on the polystyrene block (Figure 1-19). 107 Such polymers are designed to ultimately allow precise control over the donor-acceptor ratio while relying on the nanoscale phase separation that is well known in block copolymers to give well defined morphologies that could improve charge transfer as well as charge transport and collection. This concept of linking the donor and acceptor has been extended to several other approaches (see chapter 4). One approach based on modifying the polymer architecture is the so-called double-cable approach. 108 In this case acceptor moieties are covalently attached as pendant functionalities to the donor polymer backbone (see Figure 1-19). The concept is that the covalent attachment of the acceptor to the donor will enforce a large interfacial area for effective charge transfer while allowing effective bi-directional charge transport along the â€œcables.â€ In another â€œall-polymer approachâ€ several groups are working on the development of electron accepting polymers that could be blended with donor conjugated polymers and used in the place of fullerenes or other types of small molecule acceptors. 109 Likewise the use of alternative small molecule acceptors to be used in place of C 60 is an active area of research. 110 Carbon nanotubes also show promise as acceptors and electron transport materials for use
52 in bulk heterojunction photovoltaic devices. 111 A similar approach that is gaining great interest is the development of hybrid organic-inorganic solar cells based on blends of conjugated polymer donors and inorganic acceptor/electron transport materials. The work of Alivisatos has shown great promise with the development of devices based on CdSe nanocrystals and P3HT, which give efficiencies of 1.7% (AM1.5). 112 Likewise, the work of McGehee on the development of ordered bulk hetrojunction devices based on the use of patterned TiO 2 as an acceptor has gained interest as a viable approach to polymer solar cells. As a final note, polymer solar cells may ultimately see great increases in efficiency by employing engineering based approaches such as the utilization of tandem or multijunction solar cells, which have had a great impact on the field of small molecule organic solar cells. 113 Figure 1-19. Different morphologies of heterojunction solar cells. Top left: two-layer heterojunction. Top right: bulk heterojunction. Here the acceptor could be a small molecule, a carbon nanotube, an inorganic nanocrystal, or another polymer. Botton left: Diblock copolymer consisting of a donor block and an acceptor block. In this case, the acceptor block contains pendant small molecule acceptors. Bottom right: double-cable morphology. In this case the backbone is a conjugated donor polymer and the pendant groups are small molecule acceptors.
53 These alternative approaches to the polymer-PCBM solar cell offer exciting new possibilities for the realization of viable polymer photovoltaic devices. However, at the present time, none of these approaches are competitive with the polymer-PCBM platform. In order to advance this very successful approach, new polymers are needed to overcome the limitations of MDMO-PPV and P3HT. Designing new polymers with narrow band gaps and high hole mobilities will require the development of a new family of polymers. One class of polymers that have gained interest in recent years are the poly(3,4-alkylenedioxythiophenes) (PXDOTs). Polymers based on XDOTs offer an avenue to the realization of a new family of photovoltaic materials. 1.4 Poly(3,4-alkylenedioxythiophene)-Based Polymers (PXDOTs) Poly(3,4-ethylenedioxythiophene) (PEDOT) was first reported in the late 1980â€™s as a product of Bayer AG research laboratories. 114 This polymer was originally designed to yield a soluble conjugated polymer that could be synthesized via oxidative polymerization without the incorporation of and couplings that had been known to plague polythiophenes. While PEDOT is not a soluble polymer, it does exhibit a highly conductive and stable conducting state. The favorable properties of PEDOT have led to the commercialization of PEDOT-PSS [PEDOT-poly(styrenesulfonate)] as an aqueous dispersion that has been used as an antistatic coating by AGFA and has become a key component of PLEDs and PVDs. Additonally, the success of PEDOT has led to the development of literally hundreds of polymers based on EDOT or other alkylenedioxythiophene (XDOT). 115 There are several major structural features of EDOT that lead to the interesting and useful properties of the polymer. First, the monomer is 3,4-disubstituted, which
54 eliminates defects during oxidative polymerization caused by coupling of thiophene through the 3 and 4 positions. Additionally, the two substituents are electron-donating oxygens, which causes PEDOT to be very electron rich and allow oxidation at low potentials to yield highly stable and conductive conjugated polymers. This electron donating ability of the oxygens also raises the HOMO of the polymer without much affect on the LUMO energy, yielding a polymer with a reduced band gap of 1.6 eV relative to the 2.2-2.3 eV band gap of polythiophene. Also of importance is that the ethylenedioxy ring effectively pins the substituents in place and reduces any steric hindrance that would affect coupling during polymerization. Due to its low oxidation potential, EDOT has been incorporated into numerous electropolymerizable multi-ring monomers (see Figure 1-20). This approach allows a combination of the properties exhibited by the different units present in the monomers. Several donor-acceptor monomers have been developed to yield low band gap polymers based on EDOT. In this case PBEDOT-PyrPyr and PBEDOT-CNV 116 show band gaps of 1.2 and 1.1 eV respectively. Several different approaches have also been used to synthesize XDOT monomers with varying ring size and substitution. 117 Poly(3,4-propylenedioxythiophene) (PProDOT) and symmetrically substituted analogues such as PProDOT-Hx 2 have been shown to yield soluble, processable polymers upon Grignard metathesis polymerization (GriM) of the dibromo monomer. 118 The GriM method overcomes the harsh oxidative polymerization conditions encountered when using ferric chloride. 119 In these cases molecular weights approach 50,000 g/mol (M n ). The GriM approach has also proven
55 successful for the polymer LPEB, synthesized form an unsymmetrical multi-ring monomer. 120 N S S OO OO nNN S OO nS OO nS OO n NC S OO S OO n S OO n OC12H25 O H25C12PEDOTPProDOTPBEDOT-PyrPyrPBEDOT-CNVPProDOT-Hx2LPEBRepresentative Electropolymerizable PXDOT PolymersRepresentative Soluble PXDOT Polymers Figure 1-20. Representative PXDOT polymers. Alkylenedioxythiophene polymers have proven to be highly successful as electrochromic materials based on the high contrast ratios, fast switching times, low switching potentials, and electrochemical stability exhibited by these polymers. Electrochromic devices have been constructed using a variety of devices architectures such as absorptive/transmissive as well as reflective, based on electropolymerized polymers as well as soluble, spray-coated polymers. 19, 121 The ability to synthesize soluble, processable PXDOT polymers and the ability to tune the band gap in these materials via the donor-acceptor approach, makes these polymers excellent materials for electrochromics. However these same characteristics make PXDOTs a promising class of polymers to investigate for photovoltaic applications.
56 In this work the applicability of PXDOT-based polymers for use in solar cells will be explored. Towards this end, the development of a family of soluble, narrow band gap PXDOT polymers will be discussed in chapter 3. Here a series of cyanovinylene donor-acceptor polymers are discussed not only in the context of their use in solar cells, but in the broader sense of the optoelectronic properties that make soluble, narrow gap polymers attractive for numerous applications. In chapter 4 the theme of polymeric solar cells is continued. Here two novel fullerene-based donor-accpetor dyads are investigated for their optical, electrochemical, and photovoltaic properties. Such novel materials present a possible improvement over the polymer-PCBM platform that has dominated the polymer solar cell field. In chapter 5, the concept of PXDOT based donor acceptor polymers is revisited. In this chapter routes toward the development of narrow band gap polymers are explored using EDOT as the donor and tetrazine as the acceptor. Tetrazine is the most electron poor of the nitrogen containing heterocycles and has received little attention as a building block for donor-acceptor polymers. In the following chapter (chapter 2), the experimental techniques used in this work will be briefly surveyed.
CHAPTER 2 EXPERIMENTAL METHODS 2.1 Introduction In chapter 1, the fundamental properties of conjugated polymers were described and related to the numerous optical applications that have garnered so much interest in the chemistry, physics, and materials science communities. The scope of conjugated polymers research has expanded well beyond the boundaries of these traditionally separate disciplines. The breadth of the research effort spans the range of monomer and polymer synthesis, to electronic and electrochemical materials characterization, up to device fabrication, testing, and optimization. In order to complete a full study of a new conjugated polymer, a variety of specialized techniques are required. Herein is presented an overview of the experimental techniques used in the preparation of this dissertation. Previous dissertations from the Reynolds group have given extensive overviews of the techniques used for the synthesis and fundamental characterization of conjugated polymers. 122 The focus here will be on the techniques that have been developed or expanded in the preparation of this dissertation. 2.2 General Synthetic Methods Chemicals were purchased from Aldrich or Acros and used without further purification unless otherwise noted. Reactions were performed under argon or nitrogen using standard Schlenk techniques. Dry THF and ether were obtained from a Fisher keg system and were dried by passing through a column of Al 2 O 3 . The detailed procedures followed in the synthesis of the various compounds described herein, are found in the 57
58 respective chapters. All new monomers were characterized by 1 H NMR (300 MHz, Mercury 300), 13 C NMR (75 MHz, Mercury 300), high resolution mass spectroscopy (HRMS) on a Finnigan MAT 95Q mass spectrometer, elemental analysis, and melting point. Infrared spectroscopy was also used for functional group identification. Single crystals for X-ray analysis were obtained via slow solvent diffusion techniques in which the compound was dissolved in a good solvent (dichloromethane) and exposed to the vapors of a poor solvent (ethanol) in a closed vessel. Data were collected (see Appendix A) at 173 K on a Siemens SMART PLATFORM equipped with A CCD area detector and a graphite monochromator utilizing MoK radiation ( = 0.71073 ). Cell parameters were refined using up to 8192 reflections. A full sphere of data (1850 frames) was collected using the -scan method (0.3 frame width). The first 50 frames were re-measured at the end of data collection to monitor instrument and crystal stability (maximum correction on I was < 1 %). Absorption corrections by integration were applied based on measured indexed crystal faces. The structure was solved by the Direct Methods in SHELXTL6, and refined using full-matrix least squares. Polymer synthesis was also performed using standard Schlenk techniques. Polymers were purified by either reprecipitation or Soxhlet extraction. Characterization was performed by 1 H NMR (500 MHz Inova 500), 13 C NMR (75 MHz Mercury 300), along with IR spectroscopy. Polymer thermal stability was assessed by thermogravimetric analysis (TGA) using a Perkin-Elmer TGA 7. Polymers were also characterized by elemental analysis. Polymer molecular weights were estimated by GPC on two 300 x 7.5 mm Polymer Laboratories PLGel 5m mixed-C columns with a Waters Associates liquid chromatography 757 UV absorbance detector. Polymer solutions were
59 prepared in either chloroform or THF and a constant flow rate of 1mL / minute was used. Reported molecular weights are relative to polystyrene. Molecular weights were also investigated by matrix assisted laser desorption quadrupole-time-of-flight mass spectrometry (MALDI QqTOF) with an Applied Biosystems QSTAR XS hybrid quadrupole-time-of-flight (QqTOF) mass spectrometer equipped with a vacuum MALDI source (Louisiana State University). HABA (2-(4-hydroxyphenylazo)benzoic acid) or terthiophene matrices were used. 2.3 Electrochemical Methods Electrochemistry has been widely used for the characterization of conjugated polymers 123 and a detailed overview of the methods used in the Reynolds group has been prepared. For the work reported herein, a standard three-electrode electrochemical cell was used with a platinum button working electrode, a platinum wire counter electrode, and a silver wire pseudo-reference electrode calibrated vs. Fc/Fc + . All potentials are reported vs. Fc/Fc + in accord with the IUPAC standard for electrochemistry in organic solvents. 124 Electrochemistry was performed using an EG&G Princeton Applied Research model 273A potentiostat / galvanostat operated with Corrware II software from Scribner and Associates. The primary techniques used were cyclic voltammetry (CV) and differential pulse voltammetry (DPV). Polymers were adsorbed to the working electrode by electropolymerization or drop casting of a soluble polymer from a 1-2% (w /w) solution of the polymer in dichloromethane, chloroform, or toluene. For electrochemical experiments, acetonitrile was dried over CaH 2 and distilled prior to use. Electrolytes tetrabutylammonium perchlorate (TBAP), tetrabutylammonium hexafluorophosphate (TBAPF 6 ), and lithium perchlorate (LiClO 4 ) were purchased, or in the case of TBAP, synthesized from TBABr and HClO 4 .
60 For the comparison of electrochemical data reported in the literature, it is often necessary to convert potentials reported relative to one reference electrode to another reference electrode scale or to vacuum scale. The interrelationship between various reference electrodes has been tabulated in a variety of sources, 125 , 126 although there is some disagreement in the literature based on the different methods used for evaluating these relationships. Especially important is the relationship between the electrochemical value of a potential (V vs. reference electrode) and the absolute energy of this value (eV vs. vacuum). It is clear that a consistent set of standards is needed for comparison of data across the literature. For the work reported herein, the benchmark will be the normal hydrogen electrode (NHE) as 4.5 eV relative to vacuum, as this is a generally accepted value. It should be noted that two schools of thought exist on this topic. 127 One group uses a variety of conceptual and experimental data to support the value of ~4.5 eV, 125 ,128 while another group supports a value of ~4.8 eV. 129 However, the value of ~4.5 eV (4.44 eV) is supported as the IUPAC recommended value. 130 Nonetheless, electrochemical potentials are frequently reported vs. saturated calomel electrode (SCE) rather than NHE, and this electrode is generally accepted to be +0.24 V vs. NHE. As ferrocene (Fc/Fc + ) is the reference of choice for organic electrochemistry, the potential of Fc/Fc + vs. SCE is a critical value. Recent measurements indicate that Fc/Fc + is +0.38 V vs. SCE (i.e. +0.62 V vs. NHE). 131 This value is in accord with previously reported values that place Fc/Fc + at +0.6-0.7 V vs. NHE in organic electrolytes. 126 As a consequence, the Fc/Fc + redox couple can be considered ~5.1 eV below the vacuum (here the convention of +5.1 eV will be used).
61 This value of 5.1 eV for Fc/Fc + differs from some previously published values, 132 , 133 which place Fc/Fc + at 4.8 eV relative to vacuum. This discrepancy is easily solved by tracing the literature origins of the value of 4.8 eV for Fc/Fc + . In a recent work, Heeger reported the frontier orbital energies for a series of polymers relative to vacuum by converting electrochemical data based on the assumption that Fc/Fc + is 4.8 eV relative to vacuum. This value was obtained by an electrochemical calibration that placed Fc/Fc + at 0.38 V vs. SCE and the assumption that SCE is 4.4 eV relative to vacuum. In doing so, the value of 4.4 eV for SCE vs. vacuum was attributed to de Leeuw. 134 In de Leeuwâ€™s work, a value of 0.24 V for SCE vs. NHE was also used, as is generally accepted. Therefore de Leeuw indirectly suggests that NHE is only 4.16 eV relative to vacuum. Two references are used by de Leeuw to support this claim that SCE is 4.4 eV relative to vacuum. The first is the work of Bredas 135 and the second is the work of Lohman. 136 However, in the work of Bredas, a value of 4.7 eV is given for SCE vs. vacuum and the same work of Lohman is cited as the source! Further, in the work of Lohman a value of 4.48 eV is reported for NHE relative to vacuum. Thus, this data is actually consistent with the value of 5.1 eV for Fc/Fc + relative to vacuum. The inaccurate estimate of Fc/Fc + to be 4.8 eV relative to vacuum has also been propagated by the Cambridge group (Holmes and Friend), but by a different set of events. The Cambridge group has reported that electrode potentials can be converted to vacuum by assuming that Fc/Fc + is 4.8 eV below the vacuum. In doing so, the authors cited the work of Bssler. 137 In this case, the value of 4.8 eV had been derived by the self proclaimed â€œrather crude approximationâ€ of assuming that Fc/Fc + was 0.2 V vs. NHE and that NHE was 4.6 eV relative to vacuum. Based on the more recent measurements cited
62 above, a value of 0.2 V for Fc/Fc + relative to NHE is clearly in error. It is thus clear that even though a value of 4.8 eV for Fc/Fc + has become accepted in the literature, this value is inaccurate. For the work reported here, Fc/Fc + will be taken as 5.1 eV relative to vacuum. Figure 2-1 shows the proper relationship between the various electrodes and the vacuum scale. In order to convert from a measured potential vs. Fc/Fc + to the absolute potential, the measured value is simply added to 5.1 and the result is the absolute energy in electron volts. For example, if the E 1/2 of a given polymer is measured to be +0.55 V vs. SCE, this corresponds to an electrochemical potential of +0.17 V vs. Fc/Fc + and an absolute energy of ~5.3 eV below the vacuum. In this work, HOMO and LUMO energies are estimated from the onset of oxidation and reduction (CV and/or DPV) respectively. vacuumNHESCEAg/Ag+Fc/Fc+4.5 eV4.74 eV5.00 eV5.12 eVESCE = ENHE 0.24 VEAg/Ag+ = ESCE 0.26 VEFc/Fc+ = ENHE 0.62 VEFc/Fc+ = ESCE 0.38 VEFc/Fc+ = EAg/Ag+ 0.12 VReferenceElectrodeEnergyElectrochemical ConversionsElectrochemical Potential-4.5 V+0.24 V+0.50 V+0.62 V0 V Figure 2-1. Relationship between electrode and absolute potentials. Values are taken from the literature 125, 131 . Electrochemical conversions allow one to directly convert a measured potential relative to one reference electrode, to the value that would be measured according to another reference electrode. The energetic relationship gives a comparison of how far below the vacuum level each standard redox couple lies.
63 2.4 Optical and Spectroscopic Methods Critical for the application of conjugated polymers in electro-optical devices is a thorough understanding of the spectroscopic features of the polymers and the interaction of the polymers with light. The experiments conducted here were focused on elucidating fundamental structure-property relationships as well as the direct evaluation of properties for electrochromic and photovoltaic applications. 2.4.1 Spectroelectrochemistry and Colorimetry Spectroelectrochemical experiments were performed using either a Cary 500 UV-Vis-NiR specrophotometer for bench-top experiments or a Stellarnet diode-array Vis-NiR spectrophotometer equipped with a InGaAs diode array detector with fiber-optic capabilities for dry-box studies. In all cases, a three-electrode cell was utilized, as described above, with indium tin oxide (ITO) coated glass used as the working electrode (Delta Technologies 8-12 ). Polymer films were deposited by electropolymerization or by spray coating of the soluble polymer from a 1-2% (w/w) solution of the polymer in dichloromethane using an Iwata HP-BC airbrush. In this experiment, the evolution of the polymer absorption spectra is monitored as a function of applied potential. Polymer films were also investigated by in-situ colorimetric analysis. 138 In this case polymer films were deposited on ITO as for spectroelectrochemistry. The precise color coordinates were then recorded as a function of potential using a Minolta CS-100 Chroma-Meter and the CIE (Commission Internationale de lEclairage) recommended (0/0) illuminating/viewing geometry for transmission measurements. The sample was illuminated from behind with D50 (5000K) light source in a light booth specially designed to exclude external light. A background measurement was taken using blank ITO in the appropriate electrolyte solution in a quartz cuvette. The Yxy values of the
64 standard illuminant (Y 0 , x, y) were thus recorded. CIE 1931 Yxy values are reported along with the relative luminance (%Y = Y/Y 0 x 100) as a function of potential. Relative luminance offers an accurate measure of the perceived transmissivity of a material across the entire visible region according to the natural sensitivity of a standard human observer. 2.4.2 Photophysics Fluorescence and excitation spectra were obtained using a Jobin-Yvon Fluorolog-3 PL spectrophotometer. For polymer solutions, emission quantum yields were measured relative to zinc phthalocyanine (Zn-PC) in pyridine 139 or zinc tetraphenylporphryn (Zn-TPP) in toluene. 140 The optical density of all solutions was kept below A = 0.1. Thin film fluorescence data was collected with the same instrument. Polymer thin films were deposited on PEDOT-PSS coated glass by spin-coating from dichlorobenzne solution. Molar absorptivities and absorption coefficients were also calculated for all polymers in solution and in thin film form respectively. For thin film measurements, the absorbance of polymer thin films was recorded on PEDOT-PSS coated glass (in order to mimic device conditions) using a Cary 500 UV-Vis-NiR spectrophotometer. In this dual beam instrument, the baseline was measured with two PEDOT-PSS coated glass slides and the reference beam was passed through a PEDOT-PSS coated glass slide during the measurement. The film thickness was then measured using a Dektak 3030 profilometer and the absorption coefficient was calculated using the relationship A = x, where A is the measured absorbance, is the absorption coefficient, and x is the thickness of the film.
65 2.5 Surface Morphology Characterization In addition to measuring the thickness of polymer thin films, the surface morphology of the films was investigated using AFM and optical microscopy. For both measurements, films were spin-coated on PEDOT-PSS coated glass substrates. Substrate preparation involved cleaning the glass by sonicating with SDS (sodium dodecyl sulfate) in 18 M deionized water, followed by sonicating with successively 18 Mdeionized water, acetone, and methanol. Substrates were then spin-coated with PEDOT-PSS (Baytron-P) (400 L at 4000 rpm for 30s) and annealed in a vacuum oven for 2 hours at 150C. Polymer films were then coated onto the substrates at a concentration of 30 mg/mL from o-dichlorobenzene and spin-coated at 500 rpm for 18 s, followed by 1000 rpm for 60 s. AFM measurements were recorded on a Nanoscope IIIa Dimension 3100 AFM and optical microscope images were recorded on an Olympus IX70. 2.6 Photovoltaic Devices Photovoltaic devices were constructed in accord with the state-of-the-art in device fabrication as described in chapter 1. Indium tin oxide coated glass (1 x 1) was purchased from Delta Technologies (Rs = 8-12 /). A mask was then constructed by affixing two 0.5 cm wide strips of tape across the center of the ITO slide. The two strips of tape were placed in parallel and separated by approximately 0.3 cm. In this way, the ITO electrode pattern, shown in Figure 2-2b, was protected. The exposed ITO (ITO not covered with tape) was then removed by etching. Here, a vapor technique was used for etching, as the sample was placed ITO-face down above a 3:1 mixture of concentrated HCl and HNO 3 . Complete etching of the exposed ITO generally took 10-15 minutes. The tape was then removed from the etched substrate and the substrate was rinsed with 18
66 M water. The substrate was then cleaned by successively sonicating with SDS in 18 M water (10 minutes), 18 M water (10 minutes), acetone (10 minutes), and isopropanol (10 minutes). The substrates were then plasma treated with oxygen plasma for 15 minutes using a plasma cleaner (Harrick PDC-32G). This method has been observed to be important for removing impurities that are adsorbed to the surface of the ITO and ensuring a uniform surface composition. 141 After this step, 400 L of PEDOT-PSS (Bayer Baytron P VP Al 4083) was spin coated onto the ITO surface at 4000 rpm for 30 s, which resulted in a film that was ~40 nm thick as evaluated by profilometry. The PEDOT-PSS coated substrates were then annealed in a vacuum oven for 2 hours at 150C, as has been shown to be effective for increasing the conductivity of PEDOT-PSS. 142 After the annealing step, the devices were then assembled in the layered architecture as seen in the schematic in Figure 2-2a. The photoactive layer was generally a conjugated polymer-PCBM blend, spin-coated from a solution with a total concentration of 30 mg/mL. For a device, 300 L of this solution was coated onto the PEDOT-PSS coated substrate in the glovebox and the film was formed by spin coating at 500 rpm for 18 s followed by 1000 rpm for 60 s. Film thicknesses varied from sample to sample and specific data will be discussed in chapter 3. Lithium Fluoride (LiF, 0.5 nm) and Aluminium (Al, 200 nm) were then sequentially deposited by thermal evaporation on the photoactive layer (the role of LiF was discussed in chapter 1). A stainless steel mask with 1 cm x 1inch openings was utilized in the deposition of these two top electrodes that run perpendicular to the ITO electrodes on the substrate (see Figure 2-2b). As a consequence, four pixels are generated with an active area of 0.25 cm 2 each. The center
67 of the device (the area containing the four active pixels) is subsequently encapsulated in epoxy and allowed to cure in air. Figure 2-2. Schematic representation of photovoltaic device composition. (a) Layer-by-layer view. (b) Top view indicating the pixel patterns. Figure indicates that the ITO had been etched to leave two strips of ITO that were 0.5 cm wide. LiF and Al are deposited through a mask to give two strips that are 0.5 cm wide. The points at which the strips of ITO and Al coincide define the active area of each of the four pixels (0.25 cm 2 ). The devices were then evaluated by two techniques. The first was to measure the AM1.5 power conversion efficiency (see chapter 1). In this case, the currentvoltage (IV) characteristics were measured with a Keithley 2400 source measurement unit (SMU) under AM1.5G illumination with an incident power density of 100 mW/cm 2 using a 150 W Xe arc lamp and appropriate air mass filters (Oriel instruments). The power output of the light source was calibrated with a power energy meter (Ophir 2A-SH photodiode). In these measurements, the positive lead of the SMU was attached to the ITO electrode and the negative lead was attached to the aluminum electrode.
68 Efficiencies were reported as measured, although it is noted that several groups apply a spectral mismatch factor to AM1.5 solar simulators in order to make the match of the simulator more precise relative to the true AM1.5 spectrum. 143 In that case, measured efficiencies are multiplied by a factor of 0.76 for simulators using a metal-halogenide lamp and 0.90 for simulators using a Xenon lamp (for measurements made at 80 mW/cm 2 ). As no mismatch factor has been calculated for AM1.5G Xe solar simulators that use the standard 100 mW/cm 2 , no mismatch factor was applied. The external quantum efficiency of the photovoltaic devices was also evaluated by measuring the incident photon to current efficieny (IPCE% see chapter 1). In this case device pixels were irradiated with monochromatic light from a 75 W tungsten-halogen lamp. Here, the light from the lamp was passed through a monochromator (Instruments SA Inc., 1200VIS), and the wavelength of light was selected. The power output of the lamp was recorded at 10 nm wavelength intervals between 350 and 750 nm using a UDT instruments S350 Power-Energy Meter equipped with a UDT 221 Silicon Sensor Head. The current response under short circuit conditions was then recorded for each pixel at 10 nm intervals using a Keithley 2400 SMU (positive lead to ITO and negative lead to aluminum). 2.7 Other Characterization Techniques Several other techniques are also often used for the characterization of polymers for use in photovoltaic devices. Perhaps one of the most important measurements is to establish if charge transfer is occurring between the donor and the acceptor in photovoltaic devices. While IPCE measurements can indicate which components in the device are contributing to photogenerated charge carriers, several more fundamental methods can also be utilized. Among these methods, photoluminescence quenching (PL
69 quenching), light induced electron spin resonance (LESR), photoinduced absorption (PIA), and time resolved transient absorption (TA) are the most commonly used. In this work PL quenching was used to study the interaction between a conjugated polymer donor and a PCBM acceptor. In this case, the PL of a pristine conjugated polymer thin film is measured relative to a blend of the polymer and PCBM. The decrease in the emission of the blend relative to the pristine polymer when using the same excitation wavelength is indicative of charge or energy transfer (see chapter 3, Figure 3-42 and associated text for a full discussion). While this technique is the least demanding in terms of required instrumentation, the result cannot definitively prove charge transfer or decouple charge transfer from energy transfer.
CHAPTER 3 CYANOVINYLENE-BASED CONJUGATED POLYMERS FOR PHOTOVOLTAIC DEVICES 3.1 Introduction and Literature Overview At the present time conjugated polymer-based photovoltaic devices (PVDs) are able to give efficiencies of only 2.5-4% under the most optimal conditions. As seen in chapter 1, the best results are achieved with devices based on dialkoxy-PPVs (MEH-PPV or MDMO-PPV 79,82 ) or poly(3-alkylthiophenes) (P3AT) as electron donors (or simply donors) in combination with soluble fullerene acceptors, most notably PCBM. By optimizing film morphology and composition as well as device construction (choice of electrodes, buffer layers, etc.), efficiency appears to peak at ~4%. While further, physical device optimization may lead to small improvements in device efficiency; dramatic increases in performance can be sought by altering the fundamental makeup of these devices. The most glaring deficiency in the operation of PPV or P3AT devices is the limitation induced by the relatively high polymer band gaps of 2.2 and 2.0 eV respectively. While the peak of photon flux from the sun occurs at ~1.8 eV (700 nm), such polymers do not have an appreciable absorbance in this wavelength range. In order to increase the number of photons absorbed at peak solar wavelengths, it is necessary to shift the absorbance spectrum of the donor polymer toward the red end of the visible spectrum. However, no amount of processing can ultimately overcome this limitation of PPV and P3AT polymers and thus the synthesis of new, narrow band gap polymers is required. 70
71 It should be noted that such polymers are not only potentially useful for polymeric solar cells, but are interesting in a more general sense as potentially multifunctional materials. Based on their narrow band gap, these polymers may have accessible p and n type redox states, which presents the possibility for use in electrochromic devices or as electron or hole transporting materials in LEDs. Such polymers could also find use in field effect transistors (FETs). From a fundamental point of view, a limited number of conjugated polymers are known to show p and n-type conductivity and these narrow band gap polymers present interesting opportunities to gain further understanding of the process of n-type doping and conductivity. Nonetheless, several groups have developed narrow band gap polymers specifically for use in PVDs. 8,87e, 144 However, to the best of our knowledge, no systematic approach for the development of an â€œidealâ€ narrow band gap donor polymer has been published. An ideal donor polymer can be defined as a polymer with the minimal band gap allowed by frontier orbital requirements for air stability and charge transfer to a given acceptor. The polymer must also exhibit a suitably high hole mobility in order to work effectively in concert with the acceptor for charge transport in a device. Our approach to an ideal donor, as outlined below, focuses on developing a polymer with the ideal electronic band structure and relies on optimizing the structures of proven polymers. Here measurement and optimization of hole mobility is considered a second level of polymer optimization in structures that exhibit an ideal electronic structure. As such, charge carrier mobility is not measured in this work, as the development of a polymer with the ideal electronic structure is the focus. In order to establish a point of reference, by far the most successful approach to polymer-PVDs has been the bulk heterojunction approach. Further, in this
72 approach, PCBM has been established as the most effective electron acceptor. As such, we will focus on optimizing PCBM-based bulk heterojunction devices through the development and incorporation of new (ideal), narrow band gap polymers. As mentioned previously, MEH-PPV, MDMO-PPV, and P3AT are the most successful polymers in PCBM devices and are thus the primary platforms for potential structural modification towards the development of an ideal system. In P3AT derivatives, enhanced photovoltaic performance is the result of â€œpost-productionâ€ treatment (described in chapter 1), which primarily involves annealing the active layer after the device has been fabricated. Such treatment results in a modification of the blend morphology based on an increased crystallinity in the polymer, and a red shift in the polymer absorption spectrum. This treatment not only allows for the absorption of additional photons, but the induced changes in morphology increase the hole conductivity of P3AT significantly. While the use of annealed P3AT-PCBM blends has proven highly efficient on several occasions, 145 there are reasons why this platform is less than ultimately desirable for modification to a narrow band gap platform. First, with the P3AT family, there is little structural modification that will maintain the polythiophene backbone and decrease the band gap appreciably without significantly changing the HOMO and LUMO energies or significantly affecting the ordering properties that are so advantageous in these systems. Second, any attempt to develop multi-component blends containing P3AT, PCBM, and a narrow band gap polymer, are greatly hindered by the necessity of annealing the blend in order to achieve optimal performance. Annealing of such multi-component blends can lead to macrophase separation, which could drastically
73 reduce the PV performance. Finally, and perhaps most importantly, the performance of P3AT devices is often difficult to reproduce. Clearly then, it is much more attractive to focus on the use of PPV derivatives and attempt to structurally transform them into ideal donors. Here variation in the nature of the aromatic rings and the ring substituents as well as vinylene substitution and copolymerization are readily available tools in the development of a narrow band gap polymer. Figure 3-1 illustrates the band structure of several polymers including MEH-PPV and an ideal donor relative to PCBM. The proposed ideal donor has the minimal band gap as allowed by frontier orbital requirements for air stability and photoinduced charge transfer to a fullerene acceptor. These requirements arise for several reasons. First, it is desired that a polymer be air stable (i.e. resistant to oxidation) in order to allow ease of handling and processing by having a fairly low-lying HOMO (~5.2 eV or lower, assuming that the energy level of SCE is 4.7 eV below the vacuum level 146 ). Additionally, a donor polymer intended for use with a soluble fullerene acceptor (i.e. PCBM) should be capable of effective electron transfer to the fullerene upon photoexcitation. As stated in chapter 1, electron transfer is assumed possible if the band offset between the donor LUMO and the acceptor LUMO is greater than the exciton binding energy. Exciton binding energies in conjugated polymers are estimated to be 0.4-0.5 eV, 147 , 148 although LUMO offsets of as little as 0.3 eV have been shown sufficient for photoinduced electron transfer. 87e,88 b Thus, the donor polymer should have a LUMO energy of 3.8 eV or above, for the case of a PCBM acceptor (LUMO = 4.2 eV). In order to transform MEH-PPV into an ideal donor for PCBM, the LUMO must be lowered from 3.2 eV to ~3.8 eV, while minimizing any effect on the HOMO energy as the band gap is
74 compressed. The ideal polymer, with a LUMO of 3.8 eV and a HOMO energy of 5.2 eV will allow a minimal band gap (~1.4 eV), without appreciably compromising the V oc of the PVD (recall from chapter 1 that the V oc is related to the energetic difference between the donor HOMO and the acceptor LUMO). As one avenue towards this end, the replacement of vinylene linkages in dialkoxy-PPVs with cyanovinylene linkages is known to lower both the LUMO and the HOMO by ~0.5 eV while having little effect on the magnitude of the band gap as seen for CN-PPV in Figure 3-1. 74,76,132 If a method of HOMO compensation could be found that would retain the HOMO energy of the parent dialkoxy-PPV while showing the LUMO energy of the CN-PPV, a reduced band gap polymer could be realized by structurally adapting CN-PPV. Figure 3-1. Band diagram for an ideal donor polymer for PCBM. Also shown are donors MEH-PPV, CN-PPV, and PProDOT-Hx 2. Dashed lines indicate the thresholds for air stability (5.2 eV) and effective charge transfer to PCBM (3.8 eV). Note that these values are all approximate as some values come from the literature and the method of determining HOMO-LUMO energies varies from sample to sample. In all cases orbital energies are given based on the assumption that the energy of SCE is 4.7 eV vs. vacuum, 146 and Fc/Fc + is +0.380 V vs. SCE (i.e. 5.1 eV relative to vacuum).
75 The conceptual solution to this problem and the basis of the systematic approach to an ideal donor presented here is to modify the chemical and electronic structure of CN-PPV. The incorporation of an electron-rich moiety in place of a dialkoxybenzene is intended to achieve the desired result and cause the HOMO of the polymer to be raised relative to that of CN-PPV while relying on the stronger donor-acceptor interaction to concomitantly compress the band gap. An additional advantage of this approach is that little effect on the LUMO energy of CN-PPV is expected upon increasing the electron-richness of the backbone, as detailed work by our group has shown that in a homologous series of donor-acceptor cyanovinylene (CNV) polymers, the donor strength controls the HOMO energy without significantly affecting the acceptor determined LUMO energy. As CN-PPV has a LUMO of ~3.8 eV, this route appears promising. While cyanovinylene is typically a structural motif used in electron accepting polymers, cyanovinylene polymers have been used as donors in PCBM based photovoltaic devices, albeit with limited success. 149 However, CN-PPV itself, to the best of our knowledge, has not been used as a donor in PVDs. While the concept of developing a systematic family of narrow band gap CN-PPV analogues for use as electron donors in photovoltaic devices is an original one, other narrow band gap CN-PPV analogues have been synthesized for various applications. Figure 3-2 illustrates a few representative examples of this class from the literature. Polymer (i) and (ii) were synthesized for use as high electron affinity polymers for LEDs and are both direct analogues of CN-PPV in which either one or both of the dialkoxybenzenes has been replaced with a 3-alkythiophene. 150 Here the Knoevenagel polycondensation of a dialdehyde monomer and a diacetonitrile monomer was employed,
76 as is the case for CN-PPV. The increased electron-richness of 3-alkylthiophene relative to dialkoxybenzene results in a stronger donor acceptor interaction in the polymer. As a result, the greater the thiophene content, the lower the band gap. Polymer (iii) and (iv), on the other hand, are not direct CN-PPV analogues as they are synthesized via the oxidative (FeCl 3 ) polymerization of bis-heterocycle CNV monomers where thiophenes or EDOTs were used as the polymerizable end groups. 144d However, polymers (iii) and (iv) are effectively CN-PPV analogues in which one dialkoxybenzene per repeat unit has been replaced by a bithiophene or biEDOT moiety. In effect, the CN-PPV backbone has been transformed from a donor-acceptor-donor-acceptor (DADA) structure to a DDADA structure. Despite the greater donor content (and stronger donor content for the case of the EDOT derivatives) in the polymer backbone, the band gap of (iii) and (iv) varies from 1.6-1.9 eV, similar to polymer (i) and (ii). It should be noted that polymers (iii) and (iv) have been utilized in photovoltaic devices, although only limited data has been published. CN S O C6H13 nS S C12H25 n(i) Eg = 1.8 eV(ii) Eg = 1.6 eVS O O C6H13 S n(iii)a R =H, Eg = 1.8 eV(iii)b R = C8H17, Eg = 1.9 eVS O O C6H13 S nOO OO R R (iv)a R = H, Eg = 1.6 eV(iv)b R = C14H29, Eg = 1.7 eVR R NC NC CN O CN H13C6H25C12 H25C12 NC CN H13C6NCH13C6 Figure 3-2. Examples of narrow band gap CN-PPV analogues reported in the literature. Polymers (i) and (ii) are described in the literature as are polymers (iii) and (iv) 144d .
77 The polymers presented in Figure 3-2 serve as a starting point in the development of a family of narrow band gap CN-PPV analogues. In the design of the ideal donor using CN-PPV as the parent structure, our primary choice for the electron-rich donor moiety used for replacing dialkoxybenzene is ProDOT-R 2 , as described in chapter 1. The band structure of PProDOT-Hx 2 is shown in Figure 3-1 and the high lying HOMO is indicative of the electron rich nature of polymers incorporating XDOT moieties. Many different avenues exist as to how this XDOT group can be incorporated into the polymer, as variation in sequence distribution and polymerization method lead to a variety of possible structures. Additionally, as this is a systematic study of approaches to the ideal donor polymer for PVDs, variation of the donor strength as well as the sequence distribution is required in order to establish the utility of the CN-PPV motif as the platform for the ideal donor. Figure 3-3 illustrates three main types of polymers proposed as CN-PPV analogues of variable electron richness based on the thiophene and/or ProDOT content of the polymer backbone. The first of these three types are the direct CN-PPV analogues of the DADA type, which are accessible via the Knoevenagel condensation of a dialdehyde and a diacetonitrile monomer. In each case, only one isomeric structure based on the monomers OHC-X-CHO and NCCH2-Y-CH 2 CN is shown, as it is known that reversing the position of the cyano group on the vinylene linkage has little effect on the resulting electronic structure. In this case, the choice of the actual route to employ is based solely on the ease of monomer synthesis. The second class of polymers shown in Figure 3-3 are the DA polymers that could be achieved via a homo-Knoevenagel polycondensation. To the best of our knowledge, no such polymers have been reported in the literature, with the
78 exception of a phenylene analogue that was not synthesized via the Knoevenagel route. 151 The third class of polymers consists of those of the DAD structure. Here the CNV moiety is set in place via a Knoevenagel condensation and then a dibromo-monomer is polymerized via a metal-mediated polymerization that is tolerant of the CNV functionality. Of the possible options, Suzuki, Yamamoto, or Rieke polymerizations are attractive choices, as will be discussed in more detail in section 3-2. Oxidative polymerization (e.g. FeCl 3 ) from the nonbrominated monomer is also a potential option for the polymerization in this class. The benefit of this class of polymer is that the harsh basic conditions of the Knoevenagel polycondensation are avoided, which should lead to polymers containing fewer defects. Another advantage is that the monomer precursors for this class (prior to bromination) are electropolymerizable, and thus allow an additional means of studying the electrochemical and spectroscopic properties of the polymer in thin film form. Such monomers could also allow X-ray analysis, which can provide information about the conformation of the polymer backbone. Additionally, using this route model electropolymerizable monomers can be synthesized with greater ease (due to the lack of necessity to include solubilizing groups) and can allow a quick determination of the polymer band structure. In section 3.2 the development of the synthetic routes toward the specific targets selected from Figure 3-3 will be presented. In section 3.3 the electronic band structures of the resulting polymers will be derived from electrochemical and spectroscopic characterization. Here, the focus is to establish the structure-property relationship between donor strength and sequence distribution on the electronic band structure in
79 these CN-PPV analogues. The results of photovoltaic devices containing these polymers will then be presented in section 3.4. CN NC OR RO OR RO nS R RO OR nS R S R nS RO OR nOO R R S S R nOO R R S S nOO R R OO R R OR RO S R CN S CN OO R R nnnS R CN S R S CN S OO R R OO R R nS RO OR OO R R S OO R R nnType IType IIType IIINC CN NC CN CN NC NC CN NC CN CN NC CN Figure 3-3. Representative structures from the three main types of proposed CN-PPV analogues for investigation in the development of an ideal donor polymer for photovoltaic devices. 3.2 Synthesis and Physical Characterization of Cyanovinylene Polymers The central structural attribute of CN-PPV analogues is the cyanovinylene (CNV) moiety. As described in the previous section, the Knoevenagel condensation is the key bond forming reaction in the synthesis of this functionality. When used in conjugated polymers, this functional group can either be pre-formed in the monomer (Figure 3-3 type III), or achieved through the copolymerization of dialdehyde and diacetonitrile monomers in the Knoevenagel polycondensation (Figure 3-3 type I and II). Here both methodologies
80 have been selected for use in the development of a family of electron donor-CN-PPV polymers for photovoltaics. 3.2.1 Soluble CNV Polymers via the Knoevenagel Polycondensation For the synthesis of soluble CNV polymers that are direct structural analogues of CN-PPV, the Knoevenagel polycondensation is the polymerization method of choice. For the development of a family of narrow band gap CN-PPV analogues, the monomers and model compound precursors shown in Figure 3-4 were selected for synthesis. These monomers were designed for the synthesis of Type I Knoevenagel polymers (see Figure 3-3), as the synthesis of unsymmetrical monomers for Type II Knoevenagel polymers is certainly a more challenging route. S O H OO S O H OO O H 12S O H O H OO R R 3a R = butyl3b R = hexyl3c R = ethylhexylS O H C12H25 O H 4 CN NC CN NC OC12H25 O H25C1267S CN C12H25 CN 8S CN OO CN 10S CN CN OO R R 11a R = butyl11b R = hexylS CN CN 9 O O OC12H25 O H25C125H H Figure 3-4. Family of compounds targeted for the synthesis of Knoevenagel polymers and model compounds. The synthesis of mono or dialdehyde derivatives of thiophenes or dioxythiophenes presented here (compounds 1-4 in Figure 3-4) is based on the initial formation of the mono or dianion of the heterocycle by reaction with n-butyllithium followed by
81 quenching with excess DMF. Subsequent acidic workup yields the desired product. Figure 3-5 shows the complete synthesis of the targeted carboxaldehyde monomers. The synthesis of compounds 1 and 2 has been previously reported via this route, 152 while compound 1 has also been synthesized using Vilsmeier chemistry. 11 Although carboxaldehye derivatives of ProDOT have not previously been reported, the methodology employed is the same as with EDOT and the reaction yields for 3b are comparable with those for 2. Note that compounds 3a and 3c were not synthesized via the lithiation route discussed above, but were instead synthesized using Mitsunobu chemistry that was reported for the synthesis of the parent dioxythiophenes. 153 S O H OO S O H OO O H 12S OO S O H O H OO R R S OO R R S O H C12H25 O H 4S C12H25 3b R = hexyl S C12H25 O H S C12H25 OO S C12H25 OO O H 1. nBuLi (1 equiv) 2. excess DMF1. nBuLi (2 equiv)2. excess DMF1. nBuLi (2 equiv)2. excess DMF1. nBuLi (2 equiv)2. excess DMFHOCH2CH2OHpTSA1. nBuLi (1 equiv)2. excess DMFH2SO461%58%79%121327%overall yieldS OO S C12H25 OO Figure 3-5. Synthesis of aldehyde monomers and model compound precursors.
82 For compound 4, the synthetic path required an extra step as the initial reaction of 3-dodecylthiophene with 2 equivalents of n-butyllithium, followed by reaction with excess DMF resulted in the isolation of primarily 4-dodecylthiophene-2-carbaldehyde (as detected by 1 H NMR), which is the expected major product of mono-lithiation in the case of 3-alkythiophenes. 154 The synthesis of compound 4 has been reported via this synthetic methodology in 80% yield, 155 however the reaction conditions were not exactly the same as those employed here. In the literature procedure, lithiation is in hexane, followed by refluxing subsequent to the formation of a precipitate. Additionally TMEDA (tetramethylethylenediamine) was added as a common chelating agent to enhance the reactivity of organolithium species. These two differences suggest the possibility that the dianion of 3-dodecylthiophene is more difficult to form than with XDOT monomers and that steric factors may cause the 5-position of the thiophene ring to be less reactive than the 2-position. Regardless, use of the same conditions employed for the diformylation of EDOT and ProDOT-Hx 2 resulted in the monoformylation of 3-dodecylthiophene. As such, subsequent acetal protection 156 was required before lithiation and formylation with DMF could be repeated. The product was ultimately obtained by deprotection using dilute sulfuric acid, whereas deprotection using acetic acid proved ineffective, despite the reported successful use of acetic acid to deprotect other 3-alkylthiophenes. 157 For the acetonitrile compounds shown in Figure 3-4, synthesis is more complicated, as no general method can be established. Here, numerous routes have been explored with varying degrees of success. Figure 3-6 details the numerous routes attempted for the synthesis of compound 7. While route A is the most straightforward, it is not readily generalized to the thiophene or dioxythiophene targets, due to the instability of the 2,5
83 bishalomomethyl derivatives of thiophene. 158 , 159 In the case of route A, alkylation, 160 bromomethylation, 161 and cyanide substitution 162 were performed based on literature procedures. As with route A, route B also passes through a bishalomethyl derivative and is designed to capitalize on the ability to form dialdehyde derivatives of thiophene, ProDOT and dialkoxybenzene. In this route hydroquinone is converted to 1,4-dibromo-2,5-dialkoxybenzene (16) in a two step process similar to that reported in the literature. 160 This intermediate was then converted into the 2,5-dialkoxyterephthalaldehyde (5) using a similar lithiation-formylation technique as described above. 163 Subsequent reduction with LAH yielded the bis-benzyl alcohol (17). Literature precedent suggested that it would be possible to convert this diol directly to the biscyanomethyl derivative using a procedure developed by Davis and Untch illustrated in Figure 3-7. 164 According to the literature, when benzyl alcohol is treated with two equivalents of TMSCl and two equivalents of sodium cyanide in the presence of catalytic sodium iodide, a nearly quantitative yield of benzyl cyanide is produced. However in this case, only the bischloromethyl product (18) was isolated. The mechanism for this conversion is not known and further study of this route was abandoned in favor of more successful routes. Nonetheless, substitution of the bis-choloromethyl derivative (17) with NaCN in DMF gave the desired product 7. Route C (Figure 3-6) follows a similar pathway as ultimately followed in A and B, using a mesylate rather than halogen as a leaving group for substitution with NaCN. Nonetheless, the low yield of this reaction and the increased number of synthetic steps does not give any advantage over path A.
84 OH OH OC12H25 OC12H25 OH OH OC12H25 OC12H25 Br Br OC12H25 OC12H25 CN NC 7 OC12H25 OC12H25 OC12H25 O Br Br Br Br O H H25C12O H OC12H25 OC12H25 OH HO OC12H25 OC12H25 OMs MsO OC12H25 OC12H25 Cl Cl ADBC 1. KOH2. C12H25Br67%(CH2O)n33% HBr / HOAc61%NaCNDMF51%B, C, DBr2HOAc / CH2Cl21. KOH2. C12H25Br41%1. nBuLi2. DMF29%TosMIC, tBuOKDME, CH2Cl225%LiAlH416586%17141518TMSCl, NaCN, NaICH3CN / DMFEt3N, MsCl42%NaCNDMF33%NaCNDMF21% from diol Figure 3-6. Synthetic routes to compound 7. OC12H25 OC12H25 OH HO TMSCl, NaCN, NaIDMF / Acetonitrile OC12H25 OC12H25 CN NC 717 Figure 3-7. Final step in the intended synthesis of compound 7 via route B in Figure 3-6. Route D appears to be a viable choice as a general method in the synthesis of aryl-acetonitriles. Here using tosylmethyl isocyanide (TosMIC), an aldehyde can be converted directly to the corresponding acetonitrile in a single step. 165 This approach had been previously used for the conversion of ketones into nitriles, 166 and is one of only a handful of methods for the direct conversion of aldehydes to nitriles. 167 This approach capitalizes
85 on the ability to synthesize the dialdehyde derivatives of a large number of thiophene, XDOT, and phenyl derivatives. While this reaction has been reported in the literature, it has not been reported for the conversion of a dialdehyde into a diacetonitrile. The model reaction shown in Figure 3-8, as well as the final step in pathway D, represent the first two examples of this synthetic conversion. In the investigated reactions, isolated yields of the bisacetonitrile product were only in the range of 25-33%. This route was also attempted for the conversion of 2,5-thiophene-dicarboxaldehyde into 2,5-thiopehendiacetonitrile, but no product or starting material was isolated. This result suggests further optimization of the reaction conditions is required for extension of this synthetic methodology to thiophene derivatives. If this reaction can be generalized to thiophene analogues, it presents a powerful tool for the synthesis of monomers for Knoevenagel polymerization. H H O O NC CN 19TosMIC, tBuOK DME33% Figure 3-8. Model reaction with TosMIC. A Mitsunobu route utilizing acetone cyanohydrin was also investigated using benzyl alcohol as a model compound as seen below. 168 Here low yields and the use of the diethyl or dimethylazodicarboxylate (DEAD or DMAD) preclude this reaction from being a viable alternative. HO NC CN OH PPh3, DEAD (DMAD)30-37%20 Figure 3-9. Mitsunobu model reaction.
86 For the synthesis of thiophene-diacetonitriles (compounds 8 and 9 in Figure 3-4), several routes were attempted, in addition to the TosMIC route described above. Limited literature precedent exists for the synthesis of either compound and in both cases the only route that has been successfully employed is the bischloromethylation of thiophene 169 or 3-alkylthiophene. The synthesis of compound 9 is shown below in Figure 3-10. Here the bischloromethyl derivative is highly unstable. Any attempts to purify the compound by distillation resulted in the formation of a suspected polymeric product. When this bischloromethyl intermediate is used without purification, compound 9 can be synthesized in low yield (9% from thiophene) by reaction with sodium cyanide followed by distillation and subsequent recrystallization. S S Cl Cl S CN CN HCl37% formalinNaCNDMF9% from thiophene9 Figure 3-10. Synthesis of thiophene diacetonitrile. For the case of alkylthiophene derivatives, the instability of the bischloromethyl intermediate is enhanced due to the electron donating nature of the alky substituent. Reaction of 3-methylthiophene under the same conditions employed for thiophene, resulted in the formation of a small amount of unstable product. With 3-dodecylthiophene the reaction yielded no product, despite literature precedent that the â€œhighly unstableâ€ compound can be isolated and used without further purification. The synthesis of compounds 10 and 11 was not attempted via the bischloromethylation route due to the extreme instability that is expected for such electron rich compounds. The first route attempted was conversion of the dialdehyde (2 and 3a) to the diol, followed by conversion to the mesylate and substitution with sodium cyanide. However the intermediate diol was not sufficiently stable to allow completion of
87 the reaction sequence. Figure 3-11 details the final route to compound 11 that was attempted. Here ProDOT-Hx 2 (12) was converted into the 2,5-dibromo derivative by bromination with NBS. 118a Then a palladium catalyzed reaction with cyanomethyl-tributlytin 170 was attempted, as has been reported in the literature for the conversion of numerous benzyl bromides to the corresponding cyanomethylbenzenes. 171 No product or starting material was isolated in this case. S CN CN OO R R S OO R R S OO R R Br Br NBS / CHCl3Bu3SnCH2CNPdCl2[P(o-tolyl)3]2xylenes 1211b R = hexyl Figure 3-11. Synthetic route for the synthesis of ProDOT-Hx 2 -(CH 2 CN) 2 . As such, compounds 1-7 and 9 from Figure 3-4 were synthesized, but compounds 8, 10, and 11 were found to be inaccessible. In general it is found that dialdehyde functionalized thiophenes are far more accessible than thiophene-diacetonitriles. With these building blocks in place, the family of polymers and model compounds shown in Figure 3-12 was synthesized using Knoevenagel methodology. The reported method for the original Knoevenagel synthesis of CN-PPV (the hexyloxy derivative) by the Cambridge group employed tetrabutylammonium hydroxide (Bu 4 NOH) (5 mol%) as the base in a 1:1 mixture of THF and t-butanol at 50 0 C to give a polymer with an M n of 4,000 g/mol. A revised method published by the same group for the polymerization of thiophene and phenylene monomers indicated that polymerization proceeded most smoothly in 3:1 mixture of t-butanol/THF at 50 0 C with 10 mol% Bu 4 NOH and 10 mol% potassium t-butoxide (tBuOK) as co-bases, although molecular weight data was not reported in this case. Further work by the Cambridge group on the
88 synthesis of thiophene-based Knoevenagel polymers reported molecular weights (M n ) of 8,000-13,000 g/mol using 25 mol% t-butoxide as the base in 1:1 THF/t-butanol. 172 S S NC CN BEDOT-CNPV (21)OO OO S OO R R NC CN nPProDOT:CN-PPVR = hexyl or ethylhexyl S OO R R O O C12H25NC CN nH25C12R = butyl or hexylPProDOT-R2:CN-PPVS S NC CN n C12H25CN-TV S C12H25 O O C12H25NC CN nH25C12Th-CN-PPV O O C12H25NC CN nH25C12 C12H25O OC12H25 CN-PPV Figure 3-12. Family of model compounds and polymers synthesized via the Knoevenagel methodology. Several other groups have reported the synthesis of CN-PPV analogues of various structures using Knoevenagel polycondensation reactions employing similar conditions. 173 Chen reported the synthesis of a series of CN-PPV analogues using 10 mol% Bu 4 NOH and 10 mol% tBuOK in 50/50 THF/t-butanol to achieve molecular weights on the order of 14,000 to 18,000 g/mol (M n ). 173c Using only Bu 4 NOH as the base, Suh reported molecular weights (M n ) on the order of 8,000 g/mol in the case of polyfluorenes. 173b Although Wang has reported molecular weights (M w ) of 300,000 g/mol using 10% Bu 4 NOH in THF/t-butanol, 173a cyanovinylene polymers synthesized via the Knoevenagel polycondensation reaction rarely result in molecular weights greater than a few tens of thousands. The low molecular weights and poorly defined polymers that are
89 often realized using this polymerization are often attributed to the severe nature of the polymerization conditions. While one report exists of a mild version of the Knoevenagel Polycondensation reaction employing a Ruthenium catalyst, 174 the vast majority of reported polymerizations utilize a strong base such as potassium t-butoxide or Bu 4 NOH in a mixed solvent consisting typically of an alcohol and THF at elevated temperature. As the reaction conditions are very harsh at high temperatures in the presence of a strong base, crosslinking and side reactions such as Michael or Thorpe additions are common, 175 and the reaction must be carefully controlled. Boucard et al. detail that crosslinking is maintained at a minimum when the concentration of base is equal to the concentration of cyanomethyl groups and the reaction is run at reflux in methanol/THF solvent mixtures. 175 Despite this, no definitive set of conditions can be certified as ideal for this polymerization without examining the results of polymerizations run under various conditions. Figure 3-13 shows the first polymerizations attempted using dioxythiophene monomers. Initially, compound 3a was copolymerized with the commercially available 1,4-phenylenediacetonitrile using 10 mol% Bu 4 NOH and 10 mol% tBuOK in 3:1 t-butanol/THF as has been proven effective for thiophene monomers. In this case the product was completely insoluble in all investigated solvents and no low molecular weight fractions could be isolated even with Soxhlet extraction. Anaylsis by IR revealed a strong peak at 2207 cm -1 , which is indicative of the carbon-nitrogen stretch of an unsaturated nitrile (for reference the carbon-nitrogen stretch of 1,4-phenylene-diacetonitrile is found at 2251 cm -1 and carbon nitrogen stretch of bis-EDOT-CNV is
90 found at 2207 cm -1 ). No residual peak at 2251 cm -1 is observed, indicating that within the detection limits of the FTIR, no saturated nitrile end groups are present. A peak at 1656 cm -1 can be attributed to an aldehyde in the polymer, while the aldehyde associated peak in the monomer was observed at 1649 cm -1 with a shoulder at 1666 cm -1 . It was thus suspected that polymerization had occurred, but no further characterization was possible due to the insolubility of the polymer. It was felt that the insolubility was due to lack of sufficient solubilizing groups on the polymer backbone, as the homopolymer of PProDOT-Bu 2 itself has limited solubility at high molecular weight. 118b As a means of overcoming this insolubility, the bisethylhexyl-substituted derivative of ProDOT (3c) was utilized in a second attempt using the same reaction conditions. Again however the product was totally insoluble, while showing a strong IR signal at 2208 cm -1 to indicate that the reaction had proceeded to an undetermined extent. In the case of the ethylhexyl polymer, no aldehyde associated peaks were observed in the IR nor were any saturated nitrile peaks. The absence of end groups suggests the possibility that higher molecular weights were achieved than for the dibutyl polymer. S OO R R NC CN nPProDOT:CN-PPVR = butyl or ethylhexylS OO R R O O H H CN NC +3a R = butyl3c R = ethylhexyltBuOK / Bu4NOHtBuOH:THF (3:1)92 % R = butyl76 % R = ethylhexyl Figure 3-13. Synthesis of PProDOT:CN-PPV. Based on the above results, several possibilities existed. First, the polymerâ€™s insolubility could be due to lack of solubilizing groups, with the unsubstituted phenylene moiety leading to strong pi-stacking. Second, the polymerization method could be too harsh and could result in Michael additions and crosslinking, although the lack of
91 saturated nitrile peaks (2250 cm -1 ) in the IR fails to confirm this assertion. As a means of testing the polymerization conditions, a model compound (BEDOT-CNPV, 21) was synthesized using two different sets of conditions as seen in Figure 3-14. In path A, Bu 4 NOH was omitted and 1.1 equivalents of potassium t-butoxide was used as the base. This change was inspired by reaction conditions that had been used in Knoevenagel condensations to make EDOT-CNV compounds in the past. 11 Additionally it had been observed during the synthesis of the PProDOT:CN-PPV polymers that addition of tBuOK casued a gradual color change in the solution, but upon addition of Bu 4 NOH the solution rapidly turned black and a precipitate was observed to form. From this observation it was surmised that Bu 4 NOH was more reactive than tBuOK and could thus be leading to unfavorable side reactions. Note that 1:1 THF:t-butanol was also used in method A. The result of path A was an 81% yield of a brick red solid. Path B, using the same conditions as employed in the synthesis of PProDOT:CN-PPV, gave only a 65% yield of the product. Most important was the observation that path B led to the formation of a large quantity of insoluble tar-like material, whereas path A led was a much cleaner reaction. As such, it was determined that the reaction conditions used in method A would be used for the polymerization of the CN-PPV analogues. S S NC CN BEDOT-CNPV21OO OO S OO O H 1tBuOKtBuOH / THF (1:1) 81%tBuOK/Bu4NOHtBuOH / THF (3:1)65%AB Figure 3-14. Synthesis of Knoevenagel model compounds. Figure 3-15 shows the polymerization of PProDOT-R 2 :CN-PPV derivatives using the dialdehydes of ProDOT-Bu 2 (3a) and ProDOT-Hx 2 (3b)in combination with 1,4
92 bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene (7) using the conditions established in the model reaction above. The addition of dodecyloxy-substituents to the phenylene ring was intended to enhance the solubility of the polymer relative to PProDOT:CN-PPV. As can be seen, polymer yields ranged from 36% for the dibultyl derivative to 67-89% (batch 1 gave 89% and batch 2 gave 67%) with the dihexyl derivative. The low yield for the dibutyl polymer can be attributed, at least in part, to the much smaller reaction scale that was used when compared to the dihexyl polymerization. S OO R R O O C12H25NC CN nH25C12R = butyl or hexylPProDOT-R2:CN-PPVR = butyl 36%R = hexyl 67-89%S O H O H OO R R 3a R = butyl3b R = hexyl CN NC OC12H25 O H25C127+ tBuOKtBuOH / THF (1:1) Figure 3-15. Synthesis of PProDOT-R 2 :CN-PPV. In both cases, a soluble polymer was obtained that was purified by reprecipitation from chloroform into methanol. Polymer molecular weights were estimated using GPC vs. polystyrene standards in THF. For PProDOT-Bu 2 :CN-PPV, M n and M w were found to be 20,000 g/mol and 31,100 g/mol respectively, for a PDI of 1.6. For PProDOT-Hx 2 :CN-PPV, M n and M w were found to be 13,500-17,900 g/mol and 20,700-28,500 g/mol for two separate batches of the polymer. Batch 1 gave the lower molecular weights and both batches showed a PDI of 1.5. Table 3-1 (presented in the ensuing discussion) summarizes the results along with molecular weight data estimated using chloroform as the GPC solvent. The polymer structures were also supported by data from IR spectroscopy, 1 H NMR, and elemental analysis (see the experimental data in section 3.6). Strong signals in the IR at 2204 cm -1 and 2206 cm -1 for the dibutyl and the dihexyl
93 derivatives respectively, were indicative of the cyanovinylene linkages in the polymer backbone and no end groups were observed by IR. Similarly, 1 H NMR data supports the proposed structure of the polymer, although a number of small, spurious peaks indicate that side reactions certainly did occur during the polymerization, although the degree and type of such reactions could not be determined. End groups were not visible by 1 H NMR and the polymer was found to give no 13 C NMR signals (other than alkyl peaks) even under extended measuring conditions due to the limited solubility of the polymer even at elevated temperatures. Elemental analysis was found to be consistent with the proposed structures. Figure 3-16 shows the synthesis of the three remaining CN-PPV-based polymers prepared by the Knoevenagel polycondensation. Here the same reaction conditions were employed as for PProDOT-R 2 :CN-PPV and yields ranged from 92% with CN-PPV to as low as 9% with CN-TV. All three polymers were initially purified by reprecipitation into methanol from chloroform, but a lack of solubility for Th-CN-PPV and CN-TV necessitated the use of Soxhlet extraction with CHCl 3 to isolate the soluble portions. With CN-TV the crude yield of 99% for the initially insoluble product was reduced to only 7% after Soxhlet extraction. Lack of alkyl substitution on both thiophene rings in the repeat unit drastically reduced to solubility of this polymer relative to other polymers in the family. For Th-CN-PPV, after Soxhlet extraction, the polymer was isolated in 43% yield. For Th-CN-PPV and CN-PPV, characterization by IR, 1 H NMR, and elemental analysis supported the proposed structures. For CN-PPV, the 1 H NMR was found to be consistent with values reported in the literature. 173c Again however, spurious peaks in the 1 H NMR of both polymers indicated that side reactions occurred in the polymerizations as was
94 seen with PProDOT-R 2 :CN-PPV. No end groups were observed by IR or 1 H NMR for either Th-CN-PPV or CN-PPV. For the case of CN-TV, the solubility after Soxhlet extraction was still extememely limited and no NMR data was obtained for the polymer. Analysis by IR indicated the presence of cyanovinylene linkages by the peak at 2209 cm -1 . No peak was observed at 2259 cm -1 , which corresponds to the saturated nitrile in monomer 9. Also for CN-TV, a peak at 1655 cm -1 suggested the presence of unreacted aldehydes. However the shift from 1671 cm -1 , observed in monomer 4, to 1655 cm -1 observed in CN-TV, is indicative of the transformation to a more fully delocalized system, as one would expect to see in a conjugated oligomer or polymer. All the polymers synthesized by Knoevenagel polycondensation were characterized by GPC analysis vs. polystyrene standards. The molecular weight data is summarized in Table 3-1, for samples run in THF and chloroform. S S NC CN n S O O C12H25NC CN nH25C12 O O C12H25NC CN nH25C12 O OC12H25 CN-TVTh-CN-PPVCN-PPV CN NC OC12H25 C12H25O + tBuOKtBuOH, THF43%S C12H25 O O H H + tBuOKtBuOH, THF99% crudeS C12H25 O O H H S CN CN O O H H CN NC OC12H25 O OC12H25 O + tBuOKtBuOH, THF92%494757H25C12H25C12H25C12H25C12H25C12 Figure 3-16. Synthesis of Th-CN-PPV, CN-TV, and CN-PPV.
95 Table 3-1. Molecular weight (GPC) data for Knoevenagel polymers. Polymer M n (g/mol) M w (g/mol) PDI PProDOT-Bu 2 :CN-PPV 20,000 (THF) 31,100 (THF) 1.56 (THF) PProDOT-Hx 2 :CN-PPV (batch 1) 13,500 (THF) 21,100 (CHCl 3 ) 20,700 (THF) 42,600 (CHCl 3 ) 1.53 (THF) 2.02 (CHCl 3 ) PProDOT-Hx 2 :CN-PPV (batch 2) 17,400 (THF) 14,700 (CHCl 3 ) 26,300 (THF) 28,700 (CHCl 3 ) 1.51 (THF) 1.95 (CHCl 3 ) Th-CN-PPV 25,500 (THF) 12,500 (CHCl 3 ) 102,500 (THF) 53,000 (CHCl 3 ) 4.02 (THF) 4.25 (CHCl 3 ) CN-TV 3,100 (THF) 4,500 (THF) 1.46 (THF) CN-PPV 13,700 (THF) 10,900 (CHCl 3 ) 29,900 (THF) 24,400 (CHCl 3 ) 2.17 (THF) 2.24 (CHCl 3 ) Molecular weights are estimated for the various samples by monitoring the absorbance of the polymer with an in-line UV-Visible photodiode array at the absorbance maximum of the polymer. To further evaluate the polymers, the elution of the samples on the GPC column was monitored with and in-line photodiode array detector in order to record the UV-visible absorption of the selected fractions of the polymers. Spectra were extracted at various times and are shown in Figure 3-17. The absorbance maximum is listed for each spectrum and the spectra are numbered to indicate the order in which the fractions came off the column. A molecular weight can be estimated relative to polystyrene based on the elution time for each specific fraction measured. For Figure 3-17a-c, spectra were recorded in chloroform, while Figure 3-17d shows the spectra recorded in THF for CN-TV, based on difficulties with aggregation for CN-TV in chloroform. In Figure 3-17a, PProDOT-Hx 2 :CN-PPV shows an absorbance maximum at ~630 nm for fractions with molecular weights > 16,000 g/mol (curves (i)-(iii)). With molecular weights lower than this, a definite blue shift in the UV-visible spectrum is observed. With a molecular weight of 6,000 g/mol (curve (iv)), max is found at 573 nm, which indicates that the electronic properties of the polymer are no longer saturated for this polymer which has a degree of polymerization of ~7, corresponding to 14 aromatic
96 rings. At this low degree of polymerization, the vibronic structure of the polymer has also changed, showing a single peak, rather than the two peaks observed for the higher molecular weight fractions. Figure 3-17. Absorption spectra for molecular weight fractions of Knoevenagel polymers. Molecular weights are reported in g/mol vs. polystyrene. (a) PProDOT-Hx 2 :CN-PPV, batch 2: (i) 37,500, (ii) 25,000, (iii) 16,500, (iv) 6,000, (v) 3,800. (b) Th-CN-PPV: (i) 93,000, (ii) 39,500, (iii) 14,900, (iv) 6,400, (v) 2,300. (c) CN-PPV: (i) 41,300, (ii) 15,500, (iii) 6,200. (d) CN-TV: (i) 8,600, (ii) 4,800, (iii) 2,500, (iv) 1,400. The peak at ~650 nm is an artifact. Figure 3-17b shows the photodiode array data for Th-CN-PPV. Here the spectrum extracted at the elution peak (curve (iii)) shows a maximum at 504 nm. Higher molecular weight fraction (curves (i) and (ii)) show maxima at 513 and 509 nm, while the fraction with a molecular weight of 6,400 (curve (iv)) shows a maximum at 496 nm. Even the spectrum extracted for a molecular weight of 2,300 g/mol shows a maximum at 504 nm. As such, there seems to be less dependence on chain length for Th-CN-PPV than for PProDOT-Hx 2 :CN-PPV, as the optical properties appear to be saturated even at
97 low molecular weights for Th-CN-PPV. Similar behavior is observed for CN-PPV, as the absorption spectra do not show any significant deviation in absorbance maximum or curve shape for molecular weights between 41,000 g/mol and 6,200 g/mol. There is a slight blue-shift in the absorption spectra from 471 nm to 458 nm, but nothing as dramatic as that observed for PProDOT-Hx 2 :CN-PPV. The main difference in these three cases is that the absorption spectra of PProDOT-Hx 2 :CN-PPV shows vibronic structure (Figure 3-17a). Notice that it is the lower energy peak that is depleted most rapidly in low molecular weight fractions of the polymer, whereas the high-energy peak is only slightly blue shifted at lower molecular weights. For CN-PPV and Th-CN-PPV, with no defined vibronic structure, the absorbance maximum shows less dependence on molecular weight. The significance of this finding is unclear at this time. Figure 3-17d shows the spectra extracted for CN-TV. Here again, no vibronic structure is observed in the polymer and for molecular weights between 2,500 and 8,600 g/mol little change in the max is observed, with a value observed between 565 and 567 nm in all cases. The lowest molecular weight fraction observed (curve (iv)) showed a maximum at 532 nm, for a molecular weight of 1,400 g/mol. This corresponds to a degree of polymerization of only ~3, and indicates that the electronic properties are not saturated at such a small oligomer length. As a final note, one must keep in mind that all the molecular weight data in Figure 3-17 is estimated relative to polystyrene, so the molecular weight data reported is only approximate and should be viewed in terms of the trends rather than the magnitude of the reported molecular weights.
98 As a further means of molecular weight analysis and structure proof, the polymers were characterized by matrix assisted laser desorption ionization quadrupole-time of flight mass spectrometry (MALDI QqTOF MS or MALDI MS). In recent years, MALDI has received increasing attention for the characterization of synthetic polymers. 176 Unlike GPC, MALDI MS gives structural information in addition to molecular weight. In the past, MALDI has been successfully used for the analysis of alkylenedioxythiophene polymers. 118a Figure 3-18 shows the mass spectrum measured for PProDOT-Hx 2 :CN-PPV using terthiophene as the matrix. The mass spectrum confirms the presence of polymer chains with masses up to nearly 30,000 amu. The spacing between the peaks corresponds to ~1738 amu, which corresponds to the calculated molecular weight of two repeat units of the polymer. It is unclear why this is the dominant series in the spectrum, but it does confirm the expected chemical makeup of the polymer backbone. Closer analysis of the major peaks in the spectrum can provide information about end groups of the polymer. If a single molecule of dialdehyde 3b and dinitrile 7 react to form a cyanovinylene linkage, with the loss of water and if both end groups remain unreacted, the molecular weight of the resulting oligomer (n = 1) is calculated to be 887.3 g/mol. For the case of n = 2, where an additional molecule of 3b and 7 add to the oligomer, the molecular weight is calculated to be 1756.7 g/mol, if the end groups do not react. If this pattern continues, the general formula for the polymer molecular weight is [887.3n â€“ 18.0(n â€“ 1)]. As such, for n = 10, the molecular weight is calculated to be 8711.3 g/mol and for n =12 a value of 10450.0 g/mol is found. Notice in Figure 3-18 that the peaks at m/z 8693 and 10432 correspond to n = 10 and n = 12 respectively, with the loss of an additional molecule of water in each case. This extra loss of water is observed throughout
99 the dominant series in the spectrum. The loss of water could be explained by a cyclization mechanism that results in cyclic polymer species as the dominant product, although this seems highly unlikely. Importantly, small molecules are sometimes lost under the relatively harsh conditions of MALDI-Qq-TOF and this could explain the loss of water, although the mechanism is not exactly clear. Keep in mind that the accuracy of the instrument is +/2 at masses greater than 4000, so the values shown in Figure 3-18 should not be taken as precise values. Nonetheless, the MALDI data supports the proposed structure for PProDOT-Hx 2 :CN-PPV. Figure 3-18. MALDI MS of PProDOT-Hx 2 :CN-PPV. Terthiophene was used as the matrix. The inset indicates the dominant spacing pattern in the spectrum, which corresponds to two repeat units. The MALDI-TOF MS of CN-PPV was run in HABA (2-(4-hydroxyphenylazo)-benzoic acid) as the matrix, and likewise gave insight into the repeat unit structure of the polymer as seen in Figure 3-19. Here a decaying signal at masses above 10,000 amu, is not necessarily a limiting value, as larger molecular weight chains may not be as easily volatilized. Additionally, the PDI for this polymer is ~2.2, and such a high value can
100 account for decaying signals at higher masses. The spacing between alternating peaks (e.g. m/z 4969 and 3978) corresponds to a mass difference of ~992 amu which is the mass difference expected for the addition of one repeat unit. The spacing between successive peaks corresponds to the addition of one monomer to the polymer chain. For example the spacing between the peak at 5961 and 5427 is 534 amu, which approximately corresponds to the addition of one diacetonitrile monomer. Although a great deal of fine structure is seen near each peak, no information about the precise nature of the end groups could be ascertained. Figure 3-19. MALDI MS of CN-PPV. HABA was used as the matrix. Figure 3-20 shows the MALDI-TOF MS for Th-CN-PPV obtained with HABA as the matrix. In this case, a well-defined peak pattern is not observed. For masses greater than 6000 amu, peak intensity becomes very weak. Nontheless, as seen in the inset in Figure 3-20, that alternating peaks are separated by ~798 amu (e.g. peaks at m/z 5638 and 4838), which corresponds to the molecular weight of the repeat unit.
101 Figure 3-20. MALDI MS of Th-CN-PPV. HABA was used as the matrix. In the same way, the MALDI MS was measured for CN-TV in a terthiophene matrix as seen in Figure 3-21. In this case, masses over 2500 amu were not observed. This is in accord with the molecular weights estimated by GPC, which indicated that CN-TV was the lowest molecular weight polymer in the family. Peak spacings in the mass spectrum corrspond to ~435 amu, which correlates well with the repeat unit molecular weight of 435 g/mol (e.g. m/z 2047 and 1612). The peak at m/z 1612 corresponds to n = 3, with the addition of one molecule of dialdehyde 4 and the peak at m/z 2047 corresponds to n = 4 with the addition of one molecule of 4. The intense peak at 1182 should be ignored, as it is a baseline artifact. However, the peaks at m/z 1319 and 1754 correspond to n = 3 and n = 4 for the case of each having one aldehyde end group and one acetronitrile end group. The exact masses for these species were calculated to be 1320.6 and 1754.8 amu respectively, showing excellent agreement within the error of the measurement (+/2).
102 Figure 3-21. MALDI MS for CN-TV. Terthiophene was used as the matrix. The thermal stability of the Knoevenagel polymers was also investigated by thermogravimetric analysis (TGA) in nitrogen atmosphere and in air. In all cases, the polymers are stable up to 300C under nitrogen, with onsets of decomposition between 300 and 325C. This first decompostion process results in a greater than 50% weight loss for all of the investigated polymers. This first decomposition process occurs in almost identical fashion in air, indicating the stability of these polymers in the presence of oxygen, even at elevated temperatures. A second decomposition process is observed to start between 400 and 450C for all polymers, with nearly complete mass loss observed by 600-700C. 3.2.2 Soluble CNV Polymers via Transition Metal Mediated Polymerization Based on the results of section 3.2.1, it is clear that while the Knoevenagel polymerization is capable of producing relatively high molecular weight (10,000-20,000 g/mol) CN-PPV analogues, the polymers contain a certain level of defects as detected by 1 H NMR. As a means of producing polymers with a more defect-free structure, the
103 polymerization of cyanovinylene monomers by transition metal mediated polymerization has also been explored. This route is attractive not only because it promises to yield well-defined polymers, but also because a large amount of literature exists concerning elctropolymerizable CNV monomers. As such, a monomer that is found to have suitable frontier orbital energies and a suitable band gap can be selected and functionalized to induce solubility upon chemical polymerization. The main caveat of this route is that a polymerization method must be found that is compatible with cyanovinylenes and alkylenedioxythiophenes. For the polymerization of single-ring ProDOT monomers, the most successful route has been to use Grignard metathesis polymerization (GriM) of the 2,5-dibromo monomer, 118a,118b as described in chapter 1. However it is known that Grignard reagents are not compatible with cyanovinylenes due to the high reactivity of nitriles in the presence of strong nucleophiles. 177 Several other metal-mediated polymerizations occur under milder conditions and present themselves as possibilities for the polymerization of ProDOT-CNV monomers. A great deal of literature exists on the use of the Suzuki coupling as a method for the synthesis of conjugated polymers derived from the copolymerization of a dibromo monomer and a bis-boronic acid (or ester) monomer. 178 Indeed this polymerization proceeds under mild conditions and does not involve the presence of strong nucleophiles that may cause side reactions with nitrile groups. Further, literature exists concerning the synthesis of alternating copolymers containing cyanovinylene moieties in their backbone. 179 A major drawback of this method for the synthesis of homopolymers is that one must be able to either synthesize both the dibromo derivative and the bis-boronate of
104 the same pre-monomer or synthesize an A-B type monomer that contains both halogen and boronate functionality. In either case, akylenedioxythiophenes are not known to efficiently form boronic acids or esters and thus the Suzuki method is only attractive if copolymers are desired. Further the synthesis of unsymmetrical A-B type monomers (i.e., those that contain a boronate and bromine) is often hampered by the difficulty of selectively introducing complementary reactive groups on the same monomer. Other possible methods include the Stille coupling, the Rieke polymerization, ferric chloride polymerization, and the Yamamoto coupling. For the Stille reaction, while compatible with cyanovinylenes, this method is only attractive if copolymers are desired. Additionally, it is often difficult to purify bis-trialkyltin derivatives, 180 making the polymerization difficult. The Rieke method 181 is a polymerization based on the Negishi coupling which utilizes aryl zinc halides (ArZnX) as the active species. It is known that nitriles are tolerant of these aryl-zinc species. Additionally, the Rieke method is a very good method for homopolymerizing dihalo monomers and is thus attractive for the case. However, the complex preparation of the active zinc species (Rieke Zinc) makes this method somewhat less desirable than the use of ferric chloride or the Yamamoto coupling which are also effective means of homopolymerization. Oxidative polymerization using ferric chloride is one of the oldest methods for synthesizing conjugated polymers and is still often used in the synthesis of soluble conjugated polymers. While thiophene, dioxythiophene, and even bis-dioxythiophene-CNV monomers 144d are known to polymerize efficiently by this method, there are several drawbacks. The most notable downside to using a ferric chloride polymerization is that it is an oxidative polymerization and thus the polymer is obtained in its oxidized form and
105 must subsequently be chemically reduced. It is often difficult to obtain the fully neutral product and often the harsh oxidizing conditions of the reaction lead to a material with numerous defects. An extremely versatile and mild method for the polymerization of conjugated polymers is the Yamamoto coupling polymerization. 182 This polycondensation reaction promoted by Ni(0) is used for the polymerization of dihalomonomers. The basic mechanism of the reaction is shown in Figure 3-21a. Notice that one equivalent of Ni(0) is converted irreversibly to Ni(II)X 2 for each coupling reaction. As such the Ni(0) species is a reagent and must be used in stoichiometric quantities. The degradation step shown in Figure 3-21b will be discussed in the context of the polymerization presented in the latter portion of this section. Ni(0)Lm+ArX X i Ni(II)Lm X ArX iComplex 1Complex 1+jComplex 1' Oxidative AdditionDisproportionationLmNi(II)X2Ni(II)Ar Lm ArX iComplex 2X j+Complex 2 Reductive Elimination+Ar Ar X iX j Ni(II)Lm X ArX i Ni(II)Lm X ArX Ni(II)Ar Lm ArX iX j Ni(0)LmabNi(0)Lm+ArX X Ni(II)Lm X ArComplex 3Oxidative Addition (II)NiLm X ArH X ArH H Decompositionor Figure 3-21. Mechanism of the Yamamoto coupling polymerization. (a) The general mechanism. (b) Illustration of one possible degradation/termination process.
106 One of the most common Ni(0) species used is Ni(COD) 2 , although other reagents have also found success in may cases. 183 This coupling polymerization has been used for the synthesis of electron poor and electron rich monomers alike, giving molecular weights of up to several hundred thousand in many cases. Importantly the Yamamoto coupling has been successful in the polymerization of alkyenedioxythiophene polymers 184 and the mechanism does not involve strong nucleophiles. Based on the strengths of this polymerization technique and the appropriateness to the task at hand, the Yamamoto polymerization is selected as the polymerization method of choice. Figure 3-22 shows selected electropolymerized polymers from the literature, along with band structure information. Considering that a band gap of less than 1.8 eV is desired for use in solar cells, polymers a, b, d, e, and f are of interest. However, as symmetrical monomers simplify synthesis (brominating unsymmetrical monomers could be potentially challenging) and the backbone structure of the polymer, polymers a, d, and f are selected for transformation into soluble polymers via Yamamoto polymerization. S S NC nS S NC nOO S S NC nOO S S NC nOO OO S O O S CH3NC CN H21C10 nS O O S CH3NC CN H21C10 nOO OO aEg = 1.5 eV (opt.)HOMO = 5.2 eVLUMO = 3.4 eVbEg = 1.4 eV (opt.)HOMO = 4.8 eVLUMO = 3.4 eVcEg = 1.8 eV (opt.)HOMO = 5.6 eVLUMO = 3.7 eVdEg = 1.1 eV (opt.)HOMO = 4.7 eVLUMO = 3.4 eVeEg = 1.2 eV (opt.)HOMO = 4.8 eVLUMO = 3.4 eVfEg = 1.6 eV (opt.)HOMO = 4.6 eVLUMO = 3.2 eV Figure 3-22. Representative electropolymerizable CNV polymers. Note that all HOMO and LUMO energies were measured electrochemically and that all values were corrected for E SCE = 4.7 V vs. vacuum. Compounds a, b, d, and e are reported by Reynolds et al. and compounds c and f are reported by Vanderzande et al. 185
107 Figure 3-23 shows the family of compounds synthesized as precursors for the monomers to be used in the Yamamoto polymerizations. Compound 7 was synthesized as described previously and 22 and 23 were synthesized in a manner analogous to compounds 1-4 discussed earlier. S O H OO C6H13 23S O H C8H1722 CN NC OC12H25 O H25C127H13C6 S CN OO C6H13 25H13C6 S CN 24 C8H17 Figure 3-23. Family of precursors synthesized for Yamamoto coupling monomers. The synthesis of 22 (Figure 3-24) was complicated based on the formation of two products. Here the 2-formyl-3-octyl isomer was desired as this would leave the 4-position open and allow unhindered polymerization through the 5-position in the eventual monomer. Accordingly 3-octylthiophene was monobrominated in acetic acid with NBS to give 26. 186 This reaction is known to be highly selective and 1 H NMR indicated that only the desired 2,3-isomer was obtained with no detectable 2,4-isomer. Formylation of 2-bromo-3-octlythiophene with n-butlylithium and DMF led to a mixture of the desired 2,3-product and the 2,4-product in a 91:9 ratio as determined by 1 H NMR. This mixture of isomers was found to be inseparable by chromatography, and while the mixture of isomers is undesirable, the product ratio of 91:9 exceeds literature precedent for the synthesis of 2-formyl-3-alkylthiophenes. 154 The scrambling of the product is due to intraor intermolecular reactions that occur subsequent to metal-halogen exchange on the 2
108 bromo-3-ocylthiophene and lead to the formation of a small amount of the more stable 2-lithio-4-octyl anion. S C8H17 NBSacetic acidS C8H17 Br 91-93% 1. nBuLi / THF2. DMFS C8H17 S C8H17 O H O H +91 : 9 ratio of 2,3 : 2,4 by 1H NMR82%2226 Figure 3-24. Synthesis of compound 22. For the synthesis of monoacetonitrile compounds 24 and 25, a different route was employed than for the bis-acetonitrile derivatives discussed earlier. Here literature precedent exists for the direct synthesis of arylmonoacetonitriles via the nickel catalyzed coupling of an aryl â€“zinc species with bromoacetonitrile. 187 Figure 3-25 illustrates the synthesis of compounds 24 and 25. The yields in these reactions of 15-30% are in accord with literature precedent for the synthesis of acetonitrile derivatives of thiophenes and dioxythiophenes. S C8H17 Br S C8H17 S C8H17 CN NC +1. nBuLi / THF2. ZnCl23. BrCH2CN,Ni(acac)2,CHDPP15% -30%85 : 15 ratio of 2,3 : 2,4 by 1H NMR S OO C6H13 S OO C6H13 CN 1. nBuLi / THF -78 oC2. ZnCl2 0 oC3. BrCH2CN, Ni(acac)227%24252612H13C6 H13C6 Figure 3-25. Synthesis of compounds 24 and 25. It should be noted that a Vilsmeier reaction was attempted using compound 25 in the hopes of generating an unsymmetrical monomer for a Type II Knoevenagel polymer
109 (see Figure 3-3). However, no product or starting material was isolated as it is suspected that the acidic conditions of the Vilsmeier reaction are too harsh for the unstable acetonitrile compound. With the monomer precursors in hand, the synthesis of monomers proceeded via Knoevenagel condensation followed by bromination. Figure 3-26 shows the premonomers synthesized by the Knoevenagel reaction. The synthesis of BProDOT-Hx 2 -CNPV (27) proceeded in only 9% yield following the reaction conditions used in the Knoevenagel polymerizations discussed earlier. The reaction was then rerun with ethanol as the solvent as had proven successful in the synthesis of other CNV monomers. The yield however decreased to the point that only trace product could be isolated. Similarly low yields have been reported in the synthesis of related compounds. Insufficient amounts of this monomer were obtained for subsequent bromination and polymerization. For premonomers 28 and 29 (BProDOT-Hx 2 -CNV), the Knoevenagel reaction proceeded in good to excellent yields in ethanol. A mixture of four isomers was observed for compound 28, as was expected. This mixture could not be separated and indeed the complex mixture prevented the ultimate purification of the compound, which was carried on to the next step in crude form. It is interesting to note that 1 H NMR revealed a mixture of isomers for compound 29 in a ration of 5:1, which are suspected of corresponding to the trans and the cis isomers. Bromination of compounds 28 and 29 was then carried out using NBS in DMF as is seen in Figure 3-27. The use of NBS in DMF has been reported as an especially mild and selective means of brominating reactive aromatic species. 188 This is especially important in the bromination of electron rich alkylenedioxythiophene monomers, which are
110 susceptible to oxidation in the presence of NBS. 189 Crude compound 30 contained such a complex mixture of isomers as observed by TLC and 1 H NMR that purification proved elusive and the monomer was never obtained in sufficient purity (as measured by elemental analysis) to allow polymerization. S OO C6H13 CN 25 CN NC OC12H25 O H25C127S O H OO C6H13 23S O H C8H1722H13C6 S CN 24 C8H17S O H OO C6H13 23H13C6 S CN S OO C6H13 OO C6H13 +S OO C6H13 O O S OO C6H13 C12H25NC CN H25C12BProDOT-Hx2-CNPVH13C6 tBuOKtBuOH, THF9%tBuOKEtOHS CN S C8H17 C8H17 67%++ 87%2829H13C6 H13C6 H13C6H13C6 27BProDOT-Hx2-CNVtBuOKEtOH Figure 3-26. Synthesis of Yamamoto monomer precursors by Knoevenagel condensation. S CN S C8H17 C8H17 2829NBSDMF 75%NBSDMFS CN S C8H17 C8H17 30Br Br BProDOT-Hx2-CNVS CN S OO C6H13 OO C6H13 Br Br H13C6H13C6 S CN S OO C6H13 OO C6H13 H13C6H13C6 31BProDOT-Hx2-CNV-Br2 Figure 3-27. Synthesis of monomers for Yamamoto polymerization.
111 The monomer BProDOT-Hx 2 -CNV-Br 2 (31) was then polymerized using a Ni(COD) 2 mediated Yamamoto coupling polymerization as seen in Figure 3-28. In this case, a modified Yamamoto procedure was employed as a means of maximizing the molecular weight of the polymer. In the established method of Yamamoto polymerization, Ni(COD) 2 , 2,2-bipyridine, and 1,5-cylclooctadiene (COD) are dissolved in DMF and heated before the monomer is added. However, it has recently been reported that reversing the order of addition by slowing adding a solution of Ni(COD) 2 , 2,2-bipyridine, and COD in DMF to a solution of the monomer results in a nearly 30-fold increase in the polymer molecular weight. 190 The reasoning behind this method can be explained according to the mechanism shown in Figure 3-21. The first step in the Yamamoto polymerization (Figure 3-21a) is the oxidative addition of Ni(0) into the Ar-X bond to yield the intermediate referred to as complex 1. According to the mechanism, the mono-oxidative addition product is formed and then subsequently undergoes disproportionation followed by reductive elimination. However, it can be contested that when monomer is added to a solution with a high concentration of Ni(0), a considerable number of monomer molecules will undergo two oxidative additions to yield the bis-Ni(II) complex, complex 3, shown in Figure 3-21b. Complex 3 is suspected to be less stable than complex 1, based on difficulties in isolating this complex in greater than 10% yield as opposed to the complex 1 which can be isolated in greater than 50% yield, using controlled methods for the synthesis of these organometallic species. 191 Although no definitive mechanism has been proposed, it is suspected that the bis-Ni(II) complex 3 is much more likely to undergo hydrolysis or decomposition, which serves to terminate the
112 polymerization. Thus the concept of reverse order addition favors the formation of complex 1 by maintaining a high concentration of monomer relative to Ni(0). Ni(COD)2, bpy, CODDMF60 oCS OO C6H13 CN S OO C6H13 nPBProDOT-Hx2-CNVS CN S OO C6H13 OO C6H13 BProDOT-Hx2-CNV-Br2Br Br 29%H13C6H13C6H13C6H13C6 Figure 3-28. Yamamoto polymerization of PBProDOT-Hx 2 -CNV. The polymerization proceeded for 24 h, before it was quenched by pouring into cold methanol. The precipitate was then placed in a Soxhlet thimble and washed with methanol followed by hexanes and then finally taken up in dichloromethane. After this, excess Nickel was removed using EDTA as a complexing agent. 192 The deep blue polymer (dichloromethane fraction) was isolated in 29% yield. It should be noted that the hexane fraction was also blue and undoubtedly contained low molecular weight polymer. This could help to explain why the yield was so low in this reaction. Polymerizations of other soluble XDOT polymers have reported yields of 90%, although without fractionation. 184 Characterization of PBProDOT-Hx 2 -CNV was then performed in a manner analogous to that for the Knoevenagel polymers reported in the previous section. The polymer structure is supported by 1 H NMR and IR (see experimental section 3.6). In this case elemental analysis revealed only trace Ni in the polymer, but CHN values did not correlate with the calculated values. It is suspected that problems with combustion in this polymeric sample led to spurious results by elemental analysis. The thermal stability of PBProDOT-Hx 2 -CNV was investigated by TGA and it was found that the onset of
113 decompostiton occurred at 335C in air and under nitrogen. In air, this first decomposition process resulted in a 90% mass loss between 335C and ~600C. The residual 10% mass was retained even up to 900C. Under nitrogen, the first degradation process at 335C resulted in a ~20% mass loss between 335C and 400C. A second decomposition process set in at ~400C which resulted in a further mass loss, leaving the same residual 10% at 900C which was observed in air. This residual 10% mass may explain the inaccurate elemental analysis results, as it is suspected that complete combustion of the polymer was not observed. Analysis by GPC revealed an M n of 14,300 g/mol and an M w of 40,000 g/mol for a PDI of 2.80 measured in THF vs. PS. Analysis of the polymer using an in-line photodiode array during elution gave only a very weak absorption that could not be separated from the noise of the baseline and thus useful data could not be obtained. No explanation is available to explain why PBProDOT-Hx 2 -CNV was more difficult to detect with the photodiode array than the Knoevenagel polymers described in the previous section. The polymer was further characterized by MALDI MS in a terthiophene matrix, as seen in Figure 3-29. Here the polymer gives a spectrum with well-defined peak spacings up to masses of greater than 10,000 amu. Peak spacing follows a regular pattern with peaks appearing at regular intervals of 696 amu, which corresponds to the repeat unit molecular weight of the polymer. The inset of Figure 3-29 also shows that MALDI reveals information about the end groups in the polymer. For a polymer chain with n = 5, the molecular weights are calculated to be 3482.3, 3561.2, and 3640.1 g/mol for the polymer chain with H/H, H/Br, and Br/Br end groups respectively. In this case, values of
114 3481, 3560, and 3640 amu are observed, in excellent agreement with the calculated values, considering the mass accuracy of the instrument is +/-2 amu. Figure 3-29. MALDI MS for PBProDOT-Hx 2 -CNV. Terthiophene was used as the matrix. 3.3 Electronic Band Structure Determination and Electronic Characterization The model compounds and soluble polymers presented in section 3.2 provide a systematic family, which can be used for establishing the utility of CN-PPV analogues for use in PVDs. Additionally, as mentioned in section 3.1, these donor-acceptor polymers offer the potential to be useful in LEDs, FETs, and electrochromic devices based on the broad set of electronic, conductive, and redox properties that they promise to possess. From a fundamental standpoint, the complete electronic band structure including HOMO and LUMO energies as well as the magnitude of the band gap must be evaluated if the role of these polymers in any application is to be understood. The electronic properties of the soluble polymers were first evaluated in solution and the UV-Vis absorption spectra of the polymers in toluene are shown in Figure 3-30. It can clearly be seen that modification of the parent CN-PPV structure leads to a red shift
115 in the absorption maxima of the polymers in toluene solution. All five polymers absorb strongly at their max with molar absorptivities of 10 4 L mol -1 cm -1 as listed in Table 3-2. For comparison, MEH-PPV has a molar absorptivity of 1.3 x 10 4 L mol -1 cm -1 (496 nm). Figure 3-30. Solution absorbance of CN-PPV analogues in toluene. (a) CN-PPV, (b) Th-CN-PPV, (c) CN-TV, (d) PProDOT-Hx 2 :CN-PPV, (e) PBProDOT-Hx 2 -CNV.
116 To further evaluate the photophysical properties of the polymers in solution, the fluorescence spectra of the polymers were recorded in toluene as seen in Figure 3-31. All five polymers are observed to show PL in solution and the fluorescence quantum yields are listed in Table 3-2. It is observed that PProDOT-Hx 2 :CN-PPV is the only polymer with a quantum yield of greater then 0.1. Figure 3-31. Absorbance and photoluminescence spectra of CN-PPV analogues in toluene solution. (a) CN-PPV, (b) Th-CN-PPV, (c) CN-TV, (d) PProDOT-Hx 2 :CN-PPV, (e) PBProDOT-Hx 2 -CNV.
117 Table 3-2. Photophysical data for CN-PPV analogues in toluene solution. Polymer abs (nm) ( max , L mol -1 cm -1 ) Fl (nm) ( F ) PProDOT-Hx 2 :CN-PPV 628 1.9 x 10 4 660 0.11 a PBProDOT-Hx 2 -CN-PPV 612 1.1 x 10 4 723 0.033 a Th-CN-PPV 522 3.1 x 10 4 680 0.036 b CN-TV 572 nm 1.0 x 10 4 636 0.028 a CN-PPV 463 nm 2.0 x 10 4 557 0.050 b Quantum yield standards: (a) zinc phthalocyanine in pyridine, ex = 610 nm, F = 0.30. (b) zinc tetraphenylporphryn in toluene, ex = 420 nm, F = 0.033. In order to determine the electronic band structure of the polymers, solid-state measurements are required. Here two methods are used in order to determine the magnitude of the band gap. Optical measurements by thin-film UV-Vis absorption spectroscopy are used to approximate the magnitude of the band gap using the low-energy onset of the * transition. The magnitude of the band gap is also determined electrochemically by cyclic voltammetry (CV) and/or differential pulse voltammetry (DPV). Figure 3-32 shows the thin film optical absorbance of the polymers. Soluble polymers were spin coated onto PEDOT-PSS coated glass and electropolymerizable polymers were electrodeposited on ITO coated glass for the measurement. From the absorption onsets of the polymers, the optical band gaps are determined as listed in the figure. Band gaps vary from 2.1 eV as measured for the parent, CN-PPV, down to 1.5 eV for PBProDOT-Hx 2 -CNV and PBEDOT-CNPV. In all cases, increasing the content of electron rich donor moieties resulted in a concomitant decrease in the polymer band gap relative to CN-PPV as a consequence of the strengthened donor-acceptor interaction.
118 Figure 3-32. Thin-film optical absorbance of CN-PPV analogues with their respective band gaps as estimated from the onset of the -* transition. (a) Solution cast CN-PPV, (b) Solution cast Th-CN-PPV, (c) Solution cast CN-TV, (d) Solution cast PProDOT-Hx 2 :CN-PPV, (e) Solution cast PBProDOT-Hx 2 -CNV, (f) Electrodeposited PBEDOT-CNPV, and (g) Electrodeposited PBProDOT-Hx 2 -CNPV.
119 The band gaps measured for Th-CN-PPV (1.8 eV) and CN-TV (1.6 eV) are in accordance with the band gaps measured for structurally similar polymers reported in the literature (see Figure 3-2). 150 Note that the band gap of PProDOT-Hx 2 :CN-PPV (1.7 eV) is smaller than that of Th-CN-PPV (1.8 eV) as is expected from the incorporation of the more electron rich ProDOT-Hx 2 as opposed to the alkylthiophene. It is surprising that the difference is only 0.1 eV, as a 3,4-dialkoxythiophene is suspected to be a much stronger donor than an alkylthiophene. However, as will be discussed below, ProDOT is not as strong of a donor as EDOT, so the effect of substituting a ProDOT in place of a thiophene is expected to be less dramatic than for the substitution of an EDOT in place of a thiophene. Evidence of this is seen in the band gap of PBProDOT-Hx 2 -CNV, which was measured to be 1.5 eV and is considerably larger than the band gap of 1.1 eV determined optically for the bis-EDOT cyanovinylene analogue (Figure 3-22d). This difference can largely be attributed to the more electron rich and stronger donor nature of EDOT as compared to ProDOT, which is evidenced by a polymer oxidation potential for PEDOT that is ~0.2 V lower than for PProDOT. 117 a,193 This disparity in electron richness can potentially be attributed to the nonplanar twist and chair-like conformations of the propylenedioxy seven-membered ring that have been observed with X-ray crystallography. 118b Such a distorted ring could serve to reduce the electronic donation of the oxygen atoms into the -system of the thiophene ring in the case of ProDOT relative to the more planar EDOT. 117a Additionally, evidence from X-ray analysis of single crystals indicates that the dihedral angle between thiophene rings in a molecule of BiEDOT is 0 , 194 whereas this angle is measured to be ~10 in BiProDOT-Et 2 . 118b This evidence suggests that in a polymer containing ProDOT-ProDOT linkages, the electronic
120 delocalization may be reduced relative to a polymer containing EDOT-EDOT linkages. Concerning the other Bis-XDOT polymers, PBProDOT-Hx 2 -CNPV and PBEDOT-CNPV, the measured band gaps of 1.6 and 1.5 eV respectively, correlate well with structurally similar polymers reported in the literature. While BEDOT-CNPV has been previously reported, 195 the electropolymerization was not reported, and thus no band structure data. The dialkoxybenzene analogue of PBEDOT-CNPV has been reported to have a band gap (optical) of 1.58 eV, which is also very similar to that of PBPProDOT-Hx 2 -CNPV, indicating that the incorporation of a ProDOT has little effect on the magnitude of the band gap in this case. It is interesting to note that only PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNPV show vibronic structure, with the most well defined structure visible in PProDOT-Hx 2 :CN-PPV. The precise origin of this vibronic structure is not completely understood. The thin film absorption coefficients of the soluble polymers were estimated to be on the order of 2 x 10 4 cm -1 in all cases, calculated using thicknesses measured by profilometry for films deposited by spin coating on PEDOT-PSS coated glass. It is also interesting to note that PBProDOT-Hx 2 -CNV, PBProDOT-Hx 2 -CNPV, and PBEDOT-Hx 2 -CNPV show residual absorption at lower energies, while the Knoevenagel polymers show no residual absorption at lower energies. This residual absorption has also been observed in a family of electropolymerizable CNV polymers previously reported by our group. While this phenomenon may have a structural basis, with the exception of PBProDOT-Hx 2 -CNV, it could also be due to a substrate effect, with scattering caused by the rough ITO substrate in the case of electropolymerizable polymers.
121 As a more complete means of mapping the band structures of the polymers, electrochemical measurements were employed. In this case, not only can the magnitude of the band gap be determined, but the absolute energies of the HOMO and LUMO orbitals can be estimated from the onsets of polymer oxidation and reduction respectively. In this work two methods of electrochemical characterization were used. While cyclic voltammetry (CV) is the most commonly used method for electrochemical characterization of conjugated polymers, differential pulse voltammetry (DPV) is another method that can provide additional insights into the redox properties of a conjugated polymer as discussed in chapter 2. The polymers discussed here can be divided into three categories for the sake of discussing electrochemical properties. The first class consists of the electropolymerizable model compounds PBEDOT-CNPV and PBProDOT-Hx 2 -CNPV, which can only yield films upon electrochemical deposition. The second class consists of PBProDOT-Hx 2 -CNV in both its electrochemically deposited form and its solution cast chemically polymerized form. This polymer offers the unique ability to compare variation in electrochemical properties as a consequence of polymerization method. The final class consists of the four soluble polymers synthesized via the Knoevenagel polycondensation. Here a direct comparison of the electrochemical properties can be correlated with the repeat unit structure of the polymer. All electrochemical measurements were performed in the oxygen and water free environment of an argon-filled glovebox in order to avoid complications arising from the instability of the reduced form of the polymers. Figure 3-33 shows the CV and DPV for electrodeposited films of PBEDOT-CNPV and PBProDOT-Hx 2 -CNPV. Both polymers were deposited by repeated oxidative
122 cycling in solutions that were 0.005 M in monomer and 0.1 M in supporting electrolyte (TBAPF 6 ) in either dichloromethane (BEDOT-CNPV) or a 50/50 mixture of dichloromethane and acetonitrile (BProDOT-Hx 2 -CNPV). The peak monomer oxidation potentials were found to be +0.83 V and +0.69 V respectively for PBEDOT-CNPV and PBProDOT-Hx 2 -CNPV measured vs. Fc/Fc + . All polymer films were then broken in by cycling multiple times in monomer free electrolyte (0.1 M TBAPF 6 / acetonitrile) until the electrochemical response became constant. Figure 3-33. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for electropolymerized model compounds. (a) CV of PBEDOT-CNPV, (b) DPV of PBEDOT-CNPV, (c) CV of PBProDOT-Hx 2 -CNPV, (d) DPV of PBProDOT-Hx 2 -CNPV. Measurements were performed on a 0.02 cm 2 Pt button working electrode in 0.1 M TBAPF 6 / acetonitrile with a Pt wire counter electrode and a silver wire pseudo reference electrode calibrated vs. Fc/Fc + .
123 For recording the CV, the potential was swept in the anodic direction starting from a potential midway between the onset of reduction and the onset of oxidation where the polymer was not redox active. After the anodic sweep, the potential was swept in a cathodic direction. For the case of PBEDOT-CNPV, this method worked well (Figure 3-33a), however for PBProDOT-Hx 2 -CNPV (Figure 3-33c), the oxidation and reduction of the polymer were addressed separately, as cycling over the full potential range resulted in the appearance of prepeaks and the rapid degradation of the polymer film. For recording the DPV, the polymer film was first broken in (as described above) and then the reduction and the oxidation were examined separately. For PBEDOT-CNPV it can be seen that the magnitude of the band gap as estimated from the onsets of oxidation and reduction, is narrower as determined by CV (1.2 eV) and DPV (1.0 eV) as compared to the optically determined value (1.5 eV). The lack of precise agreement between band gap values determined by theses three methods has been observed previously with CNV polymers. For PBProDOT-Hx 2 -CNPV, the results indicate the opposite trend with CV and DPV determined band gaps values of 2.0 and 1.9 eV respectively, which are both larger than the optically determined value of 1.6 eV. Examination of the data presented in Figure 3-33 reveals one of the major strengths of DPV, which is the lack of prepeaks that serve to obscure the onset potentials of oxidation and reduction. With CV, the onset of oxidation for PBEDOT-CNPV is very broad and ill defined, while the onset of reduction is obscured by a prepeak at .5 V. However, cycling this polymer through only the cathodic redox process (not shown) allows a clearer determination of the onset of reduction. The prepeak is found to only
124 occur after the polymer has been cycled in the anodic direction on the previous scan. Such prepeaks have previously been attributed to trapped charges in the polymer film. 196 One other point of interest concerning the CV and DPV data in Figure 3-33 is the current response of the polymer films at potentials greater than the E 1/2 of the oxidation. For the case of PBEDOT-CNPV, the onset of oxidation is observed at .4 V vs. Fc/Fc + . As the potential is swept in the anodic direction, the current steadily increases until the E 1/2 of the polymer is reached at approximately 0 V. For the CV (Figure 3-33a), a small peak is observed corresponding to the peak potential of polymer oxidation. Beyond this peak, the current is gradually observed to increase giving a broad, flat shape that is commonly associated with capacitive charging of the polymer film. For the case of the DPV for PBEDOT-CNPV (Figure 3-33b), upon oxidation, no peak is observed and the current appears to level out at potentials greater than the E 1/2 of the polymer. Again this behavior has been previously observed by DPV and has been attributed to the capacitive charging of the polymer film. This is in contrast to the behavior of PBProDOT-Hx 2 -CNPV (Figure 3-33c and 3-33d), in which the current drops off at potentials greater than the E 1/2 of oxidation in both the CV and the DPV. The implication here is that the oxidized film of PBEDOT-CNPV is strongly capacitive and that PBProDOT-Hx 2 -CNPV is not. This perhaps suggests that thin films of PBEDOT-CNPV are more electrically conductive than thin films of PBProDOT-Hx 2 -CNPV. With a more highly electrically conductive polymer, injection of charges into the film at potentials greater than the E 1/2 is expected as further oxidation of the delocalized electronic structure is likely until very high doping levels are achieved. 197 This could account for the steady current response observed in the DPV of PBEDOT-CNPV after the E 1/2 is reached. In
125 the case of a polymer with a lower electrical conductivity, where the conductivity is dominated by site-limited redox mechanism, it is expected that at potentials anodic of the E 1/2 a decrease in current response will be observed as no further sites are available for oxidation. Ultimately though, such speculation about conductivity cannot be justified without measuring the in situ conductivity of the polymer films. 11,198,197 Figure 3-34 shows the CV and DPV for a solution cast film of PBProDOT-Hx 2 -CNV (Figure 3-34 a and b) and an electrodeposited film PBProDOT-Hx 2 -CNV (Figure 3-34c and d). The solution cast film was deposited by drop casting from a 1% (w / w) solution in dichloromethane, while the electrodeposited film was formed by repeated oxidative cycling from a 0.005 M solution in monomer in 0.1 M TBAPF 6 / acetonitrile. The peak of monomer oxidation was found to be +0.72 V vs. Fc/Fc + . Comparing the CV data for this polymer (Figure 3-34a vs. 3-34c), the magnitude of the polymer band gap is estimated to be 1.8 eV as opposed to the value of 1.5 eV determined optically. Overall, both CV and DPV estimate a larger band gap relative to that determined optically. Analysis by DPV also indicates distinct behavior for the soluble polymer relative to the electrodeposited polymer. While both polymer samples show a strong current response for oxidative electrochemistry, the electrodeposited film (Figure 3-34d) shows a much stronger current response upon reduction than the solution cast film (Figure 3-34b). Using both CV and DPV, the reverse scans of the cathodic electrochemical process show a much weaker current response than the forward scan for the case of the solution cast film compared to the electrodeposited film. This is potentially indicative of the more open morphology that is expected for the electrodeposited films. Such an open morphology is expected to facilitate the influx and efflux of dopant ions (TBA + in the
126 case of reduction) into and out of the polymer film. The large TBA + cations may face greater hindrance when trying to penetrate the dense, solution cast film. It is interesting to note that no current response is observed upon reduction when TBAP is used as the supporting electrolyte for films of either type. In previous studies in our group, it has been observed that the strongest current responses and the only examples of true n-type doping are observed upon reduction occur with TBA + as the cation, 11, 198 although in these cases, all films were electrodeposited. Figure 3-34. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for PBProDOT-Hx 2 -CNV. (a) CV of solution cast film, (b) DPV of solution cast film, (c) CV of electrodeposited film, (d) DPV of electrodeposited film. Measurements were performed on a 0.02 cm 2 Pt button working electrode in 0.1 M TBAPF 6 / acetonitrile with a Pt wire counter electrode and a silver wire pseudo reference electrode calibrated vs. Fc/Fc + . Figure 3-35 shows the cyclic voltammetry and DPV for solution cast films of PProDOT-Hx 2 :CN-PPV and Th-CN-PPV. In both cases the polymers were drop cast
127 on a Pt button working electrode from 1% (w / w) solutions in dichloromethane. Cyclic voltammetry of both polymers showed oxidation and reduction when scanning in the anodic and cathodic directions respectively. In both cases, the polymers were unstable to scanning over the full electroactive range and thus the electrochemical processes had to be addressed independently. The band gaps measured by CV are considerably larger than those measured optically at 2.3 eV vs. 1.7 eV for PProDOT-Hx 2 :CN-PPV and 2.5 eV vs. 1.8 eV for Th-CN-PPV. Figure 3-35. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for PProDOT-Hx 2 :CN-PPV and Th-CN-PPV. Films were drop cast from dichloromethane solution (1 wt %). (a) CV of PProDOT-Hx 2 :CN-PPV, (b) DPV of PProDOT-Hx 2 :CN-PPV, (c) CV of Th-CN-PPV, (d) DPV of Th-CN-PPV. Measurements were performed on a 0.02 cm 2 Pt button working electrode in 0.1 M TBAPF 6 / acetonitrile with a Pt wire counter electrode and a silver wire pseudo reference electrode calibrated vs. Fc/Fc + . Considering the DPV results for these two polymers, while Th-CN-PPV gives results that are in good agreement with the data obtained by CV, PProDOT-Hx 2 :CN
128 PPV shows no cathodic redox response by DPV. In a limited study to examine the effect of electrolyte on the reduction of PProDOT-Hx 2 :CN-PPV, LiClO 4 , TBAP, and TBAPF 6 were investigated. Films were also cast on Pt from dichloromethane and from dichloromethane containing the electrolyte salt, as has been shown to improve the redox response of solution cast films. 132 In all cases, no appreciable current response could be observed by DPV and only in the case of TBAPF 6 could a reduction be observed by CV. It should be noted that the current response of PProDOT-Hx 2 :CN-PPV is highly unstable for reduction and is only reproducible over a few scans. Figure 3-36 shows the CV and DPV for solution cast films of CN-TV and CN-PPV. Again, films of both polymers were deposited on a Pt button working electrode by drop casting from 1% (w / w) solutions in dichloromethane. For CN-TV, the electrochemical band gap found by CV (2.1 eV) and DPV (1.9 eV) is larger than the optically determined value of 1.6 eV. In Figure 3-36a and 3-36b it can be observed that the oxidation of CN-TV is irreversible and once the polymer is oxidized, no electroactivity is retained by the film. As such, for recording the CV of the polymer, the potential is first swept in the cathodic direction and then in the anodic direction. For DPV, after a forward scan in the anodic direction, no current response is observed on the reverse scan. Electrochemical reduction of the polymer however is reproducible over multiple cylces. The instability to oxidation is not observed for the previously discussed polymers in this chapter and the origin of the behavior is not clear. For CN-PPV, the electrochemical behavior can be directly compared to literature precedent as the CV data has been reported for the MEH-CN-PPV, 132 which differs from the presently studied samples by only the nature of the side chains. From the literature,
129 the onset of oxidation has been measured to be +0.8 V vs. Fc/Fc + and the onset of reduction has been measured to be .3 V vs. Fc/Fc + , for an electrochemical band gap of 2.1 eV. In the presently studied sample of CN-PPV, the onset of oxidation is observed to be ~1V vs. Fc/Fc + as determined by CV and DPV. The onset of reduction is also consistent between DPV and CV measurements, with a value of .7 V. As such, the electrochemically determined band gap is 2.7 eV, which is significantly higher than the value of 2.1 eV that was determined optically. Figure 3-36. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) for CN-TV and CN-PPV. Films were drop cast from dichloromethane solution (1 wt %). (a) CV of CN-TV, (b) DPV of CN-TV, (c) CV of CN-PPV, (d) DPV of CN-PPV. Measurements were performed on a 0.02 cm 2 Pt button working electrode in 0.1 M TBAPF 6 / acetonitrile with a Pt wire counter electrode and a silver wire pseudo reference electrode calibrated vs. Fc/Fc + . Based on the above results, the band structures of the investigated polymers can be summarized as in Table 3-3 and as seen in Figure 3-37. In Figure 3-37 the data from
130 optical and electrochemical measurements are used in tandem to establish a complete band picture. In all cases the electrochemically determined HOMO and LUMO energies are converted to vacuum energies based on the assumption that Fc/Fc + is 5.1 eV below the vacuum as discussed in chapter 2. The shaded areas in the band structures in Figure 3-37 represent the approximate positions of the band edges based on the combination of data from CV, DPV and optical measurements. In general, the maximum width of the shaded areas corresponds to the band edge positions determined by CV, which typically gave the widest measure of the band gap (except for the case of PBEDOT-CNPV in which the optically determined value of the band gap was the largest). The center of the optical band gap is represented as the overall center of the presented band structure. Table 3-3. Summary of the band structure data for the investigated polymers. Polymer E ox onset (CV) V HOMO (CV) eV E red onset (CV) V LUMO (CV) eV E ox onset (DPV) V HOMO (DPV) eV E red onset (DPV) V LUMO (DPV) eV E g (optical) eV CN-PPV 1.0 6.1 -1.7 3.4 1.0 6.1 -1.7 3.4 2.1 Th-CN-PPV 0.9 6.0 -1.6 3.5 0.9 6.0 -1.5 3.6 1.8 CN-TV 0.8 5.9 -1.3 3.8 0.7 5.8 -1.2 3.9 1.6 PProDOT-Hx 2 :CN-PPV 0.7 5.8 -1.6 3.5 0.6 5.7 --------------1.7 PBProDOT-Hx 2 -CNV (echem) 0.3 5.4 -1.5 3.6 0.2 5.3 -1.6 3.5 1.5 PBProDOT-Hx 2 -CNV (drop cast) 0.3 5.4 -1.5 3.6 0.3 5.4 -1.5 3.6 1.5 PBProDOT-Hx 2 -CNPV 0.3 5.4 -1.7 3.4 0.3 5.4 -1.6 3.5 1.6 PBEDOT-CNPV -0.3 4.8 -1.5 3.6 -0.4 4.7 -1.4 3.7 1.5 All potentials are reported vs. Fc/Fc + and all HOMO and LUMO energies are derived from the electrochemical data based on the assumption that the Fc/Fc + redox couple is 5.1 eV relative to vacuum. Looking at the overall band picture of this class of polymers presented in Figure 3-37, several key points can be understood. First, all of the polymers are donor-acceptor
131 polymers in which the acceptor is cyanovinylene. As such it is expected that the LUMO energy will be relatively constant across the series, which it indeed is. The average measured LUMO energy for all the polymers falls between 3.5 and 3.9 eV. As the proportion of electron rich heterocycle incorporated into the polymer is increased, the HOMO is observed to undergo a concomitant increase in energy level. Looking at the average HOMO energies, the value stays essentially constant for Th-CN-PPV relative to CN-PPV at ~6 eV despite the addition of the more electron rich 3-alkylthiophene. However for CN-TV, the HOMO energy of ~5.8 eV is a testament to the more electron rich nature of the all thiophene-heterocylce content of the backbone. Comparing PProDOT-Hx 2 :CN-PPV with its direct analogue Th-CN-PPV, the HOMO energy increases from 6 eV to 5.7 eV upon replacing alkylthiophene with the more electron rich ProDOT-Hx 2 . There is even a more marked change when progressing from PProDOT-Hx 2 :CN-PPV to PBProDOT-Hx 2 -CNPV in which the HOMO energy increases from 5.7 to 5.3 eV based on the equivalent of substituting a BiProDOT-Hx 2 moiety into the repeat unit. The HOMO energy of PBProDOT-Hx 2 -CNV remains at essentially the same level as PBProDOT-Hx 2 -CNPV, as the HOMO appears to be determined by the electron rich ProDOT in either case. Interestingly, PBEDOT-CNPV is the only polymer of the family to have a HOMO energy above 5.2 eV, at ~4.9 eV. This is further evidence that EDOT is a stronger donor than ProDOT. As discussed earlier, PBProDOT-Hx 2 -CNPV and PBEDOT-CNPV have similar optical band gaps (1.6 and 1.5 eV repectively), However, the effect of the powerful donor nature of EDOT is clearly observable when examining the HOMO energy measured electrochemically. It is evident that PBEDOT-CNPV is 0.6 to 0.7 V easier to oxidize than PBProDOT-Hx 2 -CNPV. As such, one can
132 conclude that ProDOT is more effective than EDOT for the synthesis of soluble narrow band gap polymers that are air stable. The accessibility of symmetrical ProDOT derivatives with solubilizing alkyl chains is another advantage of ProDOT over EDOT. Relative to alkylthiophene, ProDOT is a more effective donor, as it gives a stronger donor acceptor interaction and reduced band gaps relative to analogous polymers containing thiophene. Looking more generally at the derived band structures (Figure 3-37) it can be seen that PBProDOT-Hx 2 -CNPV and PBProDOT-Hx 2 -CNV come the closest to approaching the â€œidealâ€ electron donor (Figure 3-1) for PCBM. Especially for the case of PBProDOT-Hx 2 -CNV, the band edges almost perfectly overlap the defined energies for the HOMO and LUMO of the ideal polymer. Figure 3-37. Summary of polymer band structures incorporating optical and electrochemical data.
133 To further evaluate the electrochemical behavior of the polymers, a spectroelectrochemical series was measured for each polymer. In this case the spectral changes upon oxidation and reduction were recorded. These results not only give deeper insight into the processes of electrochemical oxidation and reduction in these polymers, they establish the electrochromic nature of the polymers in question. Figures 3-38 â€“ 3-41 illustrate these results and the potential of these polymers to find application in electrochromic devices as well as photovoltaic devices. Figure 3-38 shows the spectroelectrochemistry upon oxidation for PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNV, measured from films drop cast in ITO from dichloromethane (1 wt %). These measurements are bench-top measurements that rely on the stability of the neutral and oxidized forms of the polymer in air. In Figure 3-38a it can be seen that PProDOT-Hx 2 :CN-PPV shows behavior typical of a conjugated polymer upon oxidation. In the neutral form of the polymer (+0.15 V vs. Fc/Fc + ) a strong absorption is observed, centered at 600 nm that corresponds to the -* transition of the neutral polymer and imparts a deep blue color to the polymer film. The neutral polymer does not absorb in the near-infrared. Upon increasing the potential in a stepwise fashion (50 mV steps), the -* transition is observed to bleach at the expense lower energy charge-carrier associated peaks (see chapter 1) in the near infrared. The first decrease in the -* transition is observed at +0.5 V (as indicated in Figure 3-38a) and the film is fully oxidized by +0.85 V. During this process the polymer is undergoing p-type doping, in which the polymer is converted to a more conductive or doped form. Throughout this process, the polymer thin film is observed to become gradually more transparent and, at +0.85 V the polymer is a transmissive sky blue. The color change observed here is
134 similar to that observed in PEDOT, but the applied potentials required to affect bleaching are considerably higher, as the spectrum of PEDOT is observed to evolve at potentials as low as .4 V vs. Fc/Fc + . 117a From an electrochromic point of view, PProDOT-Hx 2 :CN-PPV is a form of PEDOT that has an air stable neutral form. Figure 3-38. Spectroelectrochemistry of PProDOT-Hx 2 :CN-PPV (a) and PBProDOT-Hx 2 -CNV (b). Polymer films were cast from dichloromethane solution onto ITO coated glass. All potentials are reported vs. Fc/Fc + . The supporting electrolyte consisted of 0.1 M TBAP/acetonitrile. In both cases, the potential was increased in 50 mV steps. Figure 3-38b shows the spectroelectrochemical series for PBProDOT-Hx 2 -CNV. Here, the neutral polymer is also deep blue and upon oxidation the visible absorption is gradually bleached, resulting in a polymer that is transmissive sky blue in the oxidized form. The two main differences observed for PBProDOT-Hx 2 -CNV relative to PProDOT-Hx 2 :CN-PPV are the lower switching potential and the much more intense near-infrared absorptions that are induced upon doping. The lower switching potential is expected based on the results presented in Table 3-3, which indicate a higher lying HOMO for PBProDOT-Hx 2 -CNV. Note that the fully neutral spectrum for PBProDOT-Hx 2 -CNV is only attainable by reducing the film with hydrazine. Such behavior indicates that once the film has been cycled, it is difficult to fully neutralize the polymer
135 electrochemically. The much more intense absorption band in the near-infrared suggests that a higher doping level is achieved in PBProDOT-Hx 2 -CNV relative to PProDOT-Hx 2 :CN-PPV. This phenomenon could be due to some inherent property of the two films or it could simply be attributed to the fact that PBProDOT-Hx 2 -CNV is stable over a wider range of potentials anodic of the onset of oxidation, possibly allowing higher doping levels. For PProDOT-Hx 2 :CN-PPV, the full electroactive range of the polymer is only ~0.4 V. Beyond applied potentials of 0.85 V for PProDOT-Hx 2 :CN-PPV, the polymer film is observed to irreversibly oxidize. To gain further insight into the electrochromic properties and fundamental spectroscopic features of this family of polymers, the spectral changes in the polymers were also observed upon reduction in the oxygen and water free environment of an argon-filled glovebox. For these measurements, polymer films were spray-coated onto ITO glass. Figure 3-39 illustrates the reductive spectroelectrochemistry for PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNV measured in the glovebox using a fiber-optic Vis-Near-infrared spectrophotometer. The quality of the results obtained from this instrument are not as high as those shown in Figure 3-38 for the bench-top spectrophotometer, but the ability to probe spectra changes upon reduction makes this fiber-optic setup more powerful. In Figure 3-39a, the spectral changes upon reduction for PProDOT-Hx 2 :CN-PPV are observed. At potentials between .9 V and .6 V, the spectrum of the polymer is constant and shows the strong -* transition of the neutral polymer. At .7 V a slight decrease in the -* transition is observed with a concomitant increase in the near infrared absorption. Over the narrow range of applied potentials from .7 to .9 V, the -* transition of the polymer is observed to bleach
136 and the absorption in the near infrared is observed to increase. During this process the polymer is changed from deep blue in the neutral state, to a transmissive grey in the reduced form. The gap in the data between 850 nm and 950 nm is due to a gap at the point of detector changeover from visible to near infrared. Notice also that the data is removed from 1125-1200 nm and from 1325-1425 nm (as illustrated with dashed lines) due to absorptions that appear and can be attributed to an incomplete baseline correction of the instrument. The positioning of these absorption bands suggests that they are due to water, although the reason for their appearance in these experiments is not clear. Figure 3-39. Spectroelectrochemistry of PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNV. (a) Reduction of PProDOT-Hx 2 :CN-PPV, (b) reduction of PBProDOT-Hx 2 -CNV. Note that the data between 1125-1200 nm as well as 1325-1425 nm (dashed lines) has been removed as described in the text. No data is shown between 860 nm and 950 nm as the detectors do not cover this wavelength range. Potentials were increased in 100 mV steps and all potentials are reported vs. Fc/Fc + . The supporting electrolyte consisted of 0.1 M TBAPF 6 /acetonitrile. The question of interest is whether or not the above spectral changes upon reduction can be attributed to a true n-type doping or not. Previous work in our group has indicated that simple electrochemical reduction in conjugated polymers cannot be directly attributed to an n-type doping processes. In order for a polymer to be considered n-doped, not only must the polymer show a reduction, but the polymer must show evidence for charge carrier formation on reduction, as can be assessed by spectroelectrochemistry
137 or in-situ conductivity measurements. The strong increase in the near-infrared absorption of PProDOT-Hx 2 :CN-PPV upon reduction suggests that a true n-type doping process is occurring. Previous work in our group on electropolymerizable cyanovinylene polymers has relied on in-situ conductivity measurements to establish n-doping as instrumentation was not available to probe the spectral response upon reduction. In that case, in-situ conductivity measurements showed that the conductivity in the polymers could better be attributed to redox conductivity upon reduction rather than electrical conductivity that is observed in a true doping process. A complete comparative study between the two sets of polymers based on spectroelectrochemistry and in-situ conductivity will be required in order to fully understand the suspected n-type doping process. Figure 3-39b shows the spectroelectrochemistry for PBProDOT-Hx 2 -CNV upon reduction. Interestingly, the polymer does show a well-defined spectral response upon reduction, which leads to a decrease in the * transition and a concomitant increase in the Near-infrared-absorption. The first spectral changes are observed at .6 V, similar to the case for PProDOT-Hx 2 :CN-PPV, indicating the similarity in LUMO energies for the two CNV polymers. Again, this behavior appears indicative of n-type doping. Figure 3-40 shows the spectroelectrochemical series for CN-TV, Th-CN-PPV, and CN-PPV. In the case of CN-TV (Figure 3-40a and b), a strong increase in the Near-infrared-absorption is observed upon reduction (Figure 3-40b), but the response is much weaker upon oxidation (Figure 3-40a). This appears consistent with the CV and DPV results shown in Figure 3-36a and 3-36b which indicate that CN-TV gives a much stronger electrochemical response upon reduction and only an irreversible response upon oxidation. As seen in Figure 3-40c and 3-40d, Th-CN-PPV shows spectral changes
138 characteristic of doping upon oxidation and reduction. For CN-PPV, Figure 3-40f appears indicative of n-type doping, as the * transition is bleached upon reduction and strong charge carrier associated transitions are observed in the near infrared. Upon oxidation, while the * transition is observed to bleach, only a minimal increase in the near infrared absorbance is observed. Figure 3-40. Spectroelectrochemistry for CN-TV, Th-CN-PPV, and CN-PPV. (a) Oxidation of CN-TV, (b) reduction of CN-TV, (c) oxidation of Th-CN-PPV, (d) reduction of Th-CN-PPV, (e) oxidation of CN-PPV, and (f) reduction of CN-PPV. Potentials were increased in 100 mV steps and all potentials are reported vs. Fc/Fc + . Other details are as described in the previous figure.
139 For the electropolymerized polymers, PBEDOT-CNPV and PBProDOT-Hx 2 -CNPV, films were deposited electrochemically on ITO. It was found to be very difficult to deposit a thick enough film of PBProDOT-Hx 2 -CNPV for spectroelectrochemistry. Using CV and potentiostatic growth methods no suitable film could be formed. On the other hand thick, highly absorptive films of PBEDOT-CNPV could easily be deposited using repeated potential cycling. This difference in behavior can be related to a suspected increase in the solubility of oligomers formed from PBProDOT-Hx 2 -CNPV relative to PBEDOT-CNPV. Nonotheless, spectroelectrochemical results are shown for PBEDOT-CNPV in Figure 3-41. Spectral changes associated with p and n-type doping are observed in Figure 3-41a and 3-41b respectively. Here the first spectral changes upon reduction are observed at .65 V, which is very similar to all of the polymers studied here, providing additional evidence that the LUMO energies of all the cyanovinylene polymers are essentially the same. Figure 3-41. Spectroelectrochemistry for PBEDOT-CNPV. (a) Oxidation and (b) reduction. Films were deposited on ITO-caoted glass via repeated oxidative potential scanning. For the spectroelectrochemistry, 0.1 M TBAPF 6 / acetonitrile was used as the electrolyte. Potentials were increased in 100 mV steps and all potentials are reported vs. Fc/Fc + . Other details are described in Figure 3-39.
140 Colorimetric analysis was also utilized to characterize the electrochromic properties of the polymer films. Table 3-4 shows the %Y and xy data for the polymers upon oxidation. Colors of the oxidized , neutral, and reduced states are also listed. Table 3-4. Colorimetric data for CN-PPV analogues. Polymer (State) E (V) %Y x y Observed color PProDOT-Hx 2 :CN-PPV (N) -0.6 50 0.317 0.365 blue PProDOT-Hx 2 :CN-PPV (Ox) 1.5 66 0.347 0.382 grey-blue PProDOT-Hx 2 :CN-PPV (Red) -1.9 ---------grey-blue PBProDOT-Hx 2 -CNPV (N) -0.6 59 0.342 0.371 blue PBProDOT-Hx 2 -CNPV (Ox) 0.7 76 0.356 0.390 colorless PBProDOT-Hx 2 -CNPV (Red) -2.1 ---------grey-blue Th-CN-PPV (N) -0.6 29 0.355 0.346 purple Th-CN-PPV (Ox) 1.3 56 0.380 0.397 grey Th-CN-PPV (Red) -2.0 ---------grey-blue CN-TV (N) -0.6 49 0.357 0.380 brown-purple CN-TV (Ox) 1.0 59 0.382 0.404 slightly yellow CN-TV (Red) -2.0 ---------gray CN-PPV (N) -0.6 60 0.443 0.405 orange CN-PPV (Ox) 1.5 80 0.383 0.404 light yellow CN-PPV (Red) -2.1 ---------grey-blue PBEDOT-CNPV (N) -1.0 22 0.293 0.321 blue PBEDOT-CNPV (Ox) 0.6 37 0.327 0.362 blue-grey PBEDOT-CNPV (Red) -2.0 ---------grey-blue State = neutral (N), oxidized (Ox), reduced (Red). Potentials are reported vs. Fc/Fc + . For reference the white light (D50) has color coordinates (x,y) of 0.357, 0.358. 3.4 Photovoltaic Devices Construction of polymer based PVDs is perhaps more of an art than a science. Numerous variables can potentially dramatically affect the device performance. As discussed in chapter 1, such factors as electrode choice, thickness of the photoactive
141 layer, blend composition, and solvent choice for spin coating, are key parameters that often prove difficult to optimize. Additionally, the role of all these factors and their precise interplay is, in many cases, not fully understood. As such, it is difficult to develop a concise, logical study that will definitively assess the performance of a given polymer in a PVD. The polymers described in the previous sections were designed and synthesized to function as electron donors for a PCBM acceptor in a photovoltaic device. As such, the primary goal here is to establish if these CN-PPV analogues will transfer electrons to PCBM upon photoexcitation and if devices can be engineered to exhibit a photovoltaic effect under AM1.5 conditions. The PCBM used here was synthesized according to the literature (see section 3.6 for synthetic details). Before incorporating a polymer into a device with PCBM, it is interesting to examine the energetic relationship between the two components in thin films. Proof for charge transfer can be established in a variety of ways using such techniques as ESR and PIA, as discussed in chapter 2. Another simpler technique is PL quenching. In this case the PL of the pristine polymer film is measured and then the PL of a polymer-PCBM blend film is measured under the same conditions. Figure 3-42 shows a schematic of the processes involved. After initial generation of an excited state (Figure 3-42a), two competing processes are seen in Figure 3-42b and 3-42c. In the pristine polymer, the excited state can return to the ground state by the emission of a photon. With a polymer-PCBM blend, if the energetic relationship between the donor and acceptor is appropriate, the process of charge transfer is much faster than PL, and thus a quenching of the PL should be observed. This quenching is generally so strong because electron transfer in
142 polymer-PCBM blends occurs on a sub-picosecond time-scale (estimated to be < 100 fs), 199 as opposed to PL which occurs on the time scale of ~200-300 ps. 200 This process could however be complicated if very large domain sizes are present in the film that could prevent exciton diffusion to the donor-acceptor interface. A quenching of the polymer PL is indicative of a likely charge transfer to PCBM, although the possibility of energy transfer cannot be ruled out. While this simple experiment is not definitive, the experiment can establish if charge transfer is likely, or in the case of unquenched emission, that charge transfer is highly unlikely. LUMOHOMO h aDonor Exciton b Donor(Conjugated Polymer)Donor ExcitonDonor ExcitonAcceptor(PCBM)LightAbsorptionc e Charge Transfer h' Figure 3-42. Schematic representation of PL quenching in polymer PCBM blends. (a) Light absorption to generate an exciton. (b) PL of the pristine polymer. (c) Charge transfer from excited donor to PCBM acceptor. The PL quenching in polymer-PCBM blends was measured for all the soluble polymers investigated. With the exception of PBProDOT-Hx 2 -CNV, all polymers showed photoluminescence in thin films. For PBProDOT-Hx 2 -CNV, the lack of thin film PL is not surprising based on the extremely weak PL observed for this polymer even in solution. As discussed previously, residual Nickel had been removed from the polymer by washing with EDTA, so nickel is not suspected to be responsible for the lack of
143 polymer PL in thin films. There is no apparent structural basis for the lack of PL. Figure 3-43 shows the PL quenching for PProDOT-Hx 2 :CN-PPV (a) and Th-CN-PPV (b) in 1/4 (w/w) blends with PCBM. These two polymers showed typical quenching behavior, as addition of PCBM reduced the intensity of emission by greater than 95% in both cases. Figure 3-43. Photoluminescence quenching of PProDOT-Hx 2 :CN-PPV and Th-CN-PPV with PCBM. (a) PProDOT-Hx 2 :CN-PPV with an excitation wavelength of 570 nm for pristine polymer and blend. (b) Th-CN-PPV with an excitation wavelength of 540 nm for pristine polymer and blend. Red lines show the normalized absorption of the pristine polymer film and black lines show the normalized absorption of the polymer/PCBM (1/4) blend film on PEDOT-PSS coated glass. Green lines show the normalized emission of the pristine polymers and blue lines show the photoluminescence of the films. In Figure 3-44, the PL quenching of CN-PPV and CN-TV is shown. With CN-PPV (Figure 3-44a), the measured PL of the pristine film with excitation at 465 nm was very broad with two peaks at 565 nm and at 707 nm of almost equal intensity. Similar broad, low energy emission at has been observed previously in thin film samples of dihexyloxy-CN-PPV. 201 In solution, dihexyloxy-CN-PPV was observed to show emission at 555 nm (toluene), which matches well with the data in Figure 3-31, which shows an emission maximum at 557 nm for CN-PPV in toluene. However, thin film emission with the dihexyloxy derivative was observed to show a broad emission with a single maximum at 690 nm. In the previously studied case, the emission in solution was attributed to
144 intrachain singlet excitions, or excited states that form and decay on a single chain. In a film, the low energy emission was attributed to interchain excitons (physical dimers or excimers). Such interchain species are of lower energy and thus give a red shifted emission. It has been noted that CN-PPVs have an especially strong tendancy to show excimer associated emission in thin films, although other PPVs exhibit this behavior as well. 202 Interestingly, the emission shown for the CN-PPV pristine film in Figure 3-44a, shows both intrachain emission (565 nm) and excimer emission (707 nm). A similar behavior is observed in the solution PL of CN-PPV (not shown) if the solution is not stirred for at least 24 hours, suggesting that excimer emission is suppressed when aggregates are broken apart upon extended stirring. Referring again to Figure 3-44a, upon excitation of the CN-PPV / PCBM blend film with 465 nm light, the maximum intensity of the emission was reduced by a factor of six and the emission maximum is located at 611nm. Additionally, the excimer-attributed emission band is suppressed. This behavior suggests that charge or energy transfer from polymer to PCBM is occurring. For CN-TV (Figure 3-44b) the result is quite different than that observed for the other polymers. Here upon excitation of the pristine polymer with 580 nm light, a broad, weak emission was observed with a maximum of 688 nm. Excitation of the PCBM-blend film with 580 nm light resulted in a sharp emission with a maximum of 719 nm that was observed to be 3.4 times as intense as the emission of the pristine polymer film. To further investigate the unusual PL behavior of CN-TV and its blends with PCBM, the excitation spectrum of the polymer-PCBM blend was measured for the emission at 719 nm. Figure 3-45 illustrates the excitation spectrum along with the full absorption spectrum of the polymer-PCBM blend. Here it can be observed that the dominant
145 contribution to the PL at 719 nm is the absorbance at ~275 nm which is attributed to PCBM. Notice in the excitation spectrum that there is essentially no contribution to the PL at 719 nm from CN-TV, which shows an absorption maximum of 581 nm. Figure 3-44. Photoluminescence quenching of CN-PPV and CN-TV with PCBM. (a) CN-PPV with excitation wavelength of 465 nm for pristine polymer and blend. (b) CN-TV with excitation wavelength of 580 nm for pristine polymer and blend. Red lines show the normalized absorption of the pristine polymer film and black lines show the normalized absorption of the polymer/PCBM (1/4) blend film on PEDOT-PSS coated glass. Green lines show the normalized emission of the pristine polymers and blue lines show the photoluminescence of the films. There are several possible explanations for this unusual behavior of CN-TV / PCBM blends. First, direct excitation of PCBM at wavelengths as long as 664 nm has been observed to result in emission at ~730 nm. 89a While it is possible that PCBM is being directly excited in this case, the fact that the polymer is strongly absorbing at 581 nm, makes this scenario seem unlikely. However, if the polymer-PCBM blend is macrophase-separated with very large domains of PCBM, it is possible that PCBM can be directly excited independently of the polymer, and the emission observed in Figure 3-44b is a combination of the emission of the two substances. In fact, as will be seen in the following section, very large clusters are observed by AFM in thin films of blended CN
146 TV and PCBM. Another less likely explanation for this strange result, is that energy transfer from CN-TV to PCBM is occurring upon excitation of the blend at 581 nm. Subsequent to energy transfer, PCBM is then emitting light at 719 nm. Regardless of the mechanism, it appears that if charge transfer is occurring from CN-TV to PCBM, it is not the dominant process in the blend upon photoexcitation. The photocurrent action spectrum for this blend is also discussed in the upcoming section, and it provides additional information about the generation of charge carriers in these blend films. 3004005006000.00.20.40.60.81.01.2 Intensity (a. u.)Wavelength (nm) Figure 3-45. Excitation and absorption spectra for CN-TV/PCBM blend. Excitation spectrum (red line) taken at 720 nm and absorption spectrum (black line) for a 1/4 blend (w/w) of CN-TV and PCBM on PEDOT-PSS coated glass. While no definitive results could be gleaned from the PL quenching measurements, it was apparent that charge transfer from polymer to PCBM is a distinct possibility for PProDOT-Hx 2 :CN-PPV, Th-CN-PPV, and CN-PPV, while efficient charge transfer appears unlikely with CN-TV. Attempts were made to study the process of charge transfer in more detail by measuring the photoinduced absorption spectrum (PIA) of the polymers and the polymer-PCBM blends. Several obstacles have thus far precluded these measurements. The investigated polymers absorb at long wavelengths, and accessing instrumentation with the necessary detector range as well as a suitable
147 time-scale for measurement, has thus far prevented the acquisitions of this data. Currently investigations are underway in the laboratory of Professor Emil List at the Technical University of Graz, Austria to further evaluate the processes of charge and energy transfer in these polymer-PCBM blends. To gain deeper insight into the energetic relationship of these polymers with PCBM, bulk heterojunction photovoltaic devices were constructed based on 1/4 (w/w) blends of each polymer with PCBM (as described in chapter 2). In all cases blends were spin coated from dichlorobenzene solutions. The chosen weight ratio of 1:4 is based on the preponderance of evidence discussed in chapter 1 that suggests that this weight ratio is the most effective for producing efficient devices. The measurement of the IPCE (external quantum efficiency) in preliminary devices is crucial, as it can offer strong proof of electron transfer from the polymer to PCBM through a photocurrent action spectrum that strongly parallels the polymer absorption spectrum. Additionally, evaluation of the solar cell performance under AM1.5 conditions can establish the utility of the polymers for photovoltaics. However, as described in chapter 1, a great deal of optimization is often required to maximize the performance of a polymer in a bulk heterojunction device. Figure 3-46 illustrates the photovoltaic performance of PProDOT-Hx 2 :CN-PPV / PCBM (1/4) solar cells. In Figure 3-46a, it can be seen that the diode possesses the characteristics of a photovoltaic device. While the efficiency of the device is only 0.15%, the J sc of ~1.3 mA/cm 2 is a reasonable current for this type of device. The most problematic feature of this device is the low FF of only 0.25. Such a low fill factor is indicative of a device with a high resistance. In fact, a fill factor of 0.25 suggests that the
148 current is essentially varying linearly with applied bias, as is the case for a resistor (V=IR). However, based on the curve shape of the dark current (black line in Figure 3-46a) and the overall curve shape of the light current (red line), it is clear that the device shows diode characteristics. Encouragingly, the IPCE data shown in Figure 3-46b indicates that PProDOT-Hx 2 :CN-PPV is acting as a photexcited electron donor and is thus the major contributor to the photocurrent in the device. The slight blue shift in the IPCE relative to the polymer absorption spectra is a consequence of blending the polymer with PCBM. As can be seen in Figure 3-43a, the absorption spectrum of the polymer in a blend with PCBM is blue shifted relative to the pristine film. The IPCE data shown in Figure 3-46b is for a representative device. For the construction of a series of devices based on PProDOT-Hx 2 :CN-PPV and PCBM (1/4), the magnitude of the IPCE is seen to vary somewhat, but the curve shape is found to be essentially constant. This level of reproducibility is general for all the devices discussed in this work. Figure 3-46. Photovoltaic properties of PProDOT-Hx 2 :CN-PPV / PCBM solar cells based on a 1/4 blend (w/w) of polymer relative to PCBM. (a) Current voltage characteristic of a representative device as evaluated under AM1.5 conditions (100 mW/cm 2 ). Black curve shows the dark current and the red curve shows the current upon illumination. (b) Photocurrent action spectra for the device. Black squares represent the IPCE value at the give wavelength. The solid curve is the absorption spectra of a pristine polymer film shown as reference.
149 For the devices based on PProDOT-Hx 2 :CN-PPV and PCBM discussed above, solutions in dichlorobenzene consisting of PProDOT-Hx 2 :CN-PPV and PCBM were prepared with a total concentration of 30 mg/mL with four parts PCBM and one part PProDOT-Hx 2 :CN-PPV (optimal for MEH-PPV / PCBM devices). The active layer was spin coated onto PEDOT-PSS coated ITO under the conditions of 500 rpm for 18 s, followed by 1000 rpm for 60 s. These conditions are defined as standard conditions for spin coating. The resulting films were found to have a thickness of ~50 nm. Changing the spin coating speed was found to have little effect on the thickness of the active layer, while increasing the concentration of the blend to 40 mg/mL was found to essentially double the film thickness, when spin-coating was performed under the same conditions. However, in the devices based on thicker films, short-circuit current densities were found to be 25% lower and AM1.5 efficiencies were similarly diminished. This is not surprising, as thicker films are expected to suffer from increased series resistances. The surface morphology of PProDOT-Hx 2 :CN-PPV / PCBM (1/4) blends spin coated under standard conditions from 30 mg/mL solutions (dichlorobenzene) were then studied on PEDOT-PSS coated glass slides by atomic force microscopy (AFM) as a means of understanding the phase separation. Figure 3-47 shows the height image of a sample based on PProDOT-Hx 2 :CN-PPV and PCBM. Here it does not appear that there is macrophase separation in the blend film, although it appears that the surface has a grainy nature. The surface roughness is calculated to be only 1 nm, indicating a very smooth surface (the bright spot is a piece of dust). For comparison, in blends used to model the most efficient MDMO-PPV / PCBM (1/4) devices, the maximum peak to valley heights measured by AFM are < 10 nm (on PEDOT-PSS coated ITO) and the
150 domain sizes are on the order of 100 nm (by TEM). 89c The overall appearance of such films on the micron scale (by AFM) is that of a smooth, grainy surface, similar to that seen in Figure 3-47. Analysis by SEM of such MDMO-PPV / PCBM films also shows that film surfaces are uniform, smooth, and continuous. 89a For the case of the PProDOT-Hx 2 :CN-PPV blends, a decrease in resolution for AFM images taken at smaller length scales prevented more detailed surface information from being collected. For a complete understanding of the blend morphology, SEM and TEM analysis is required. Nonetheless, the AFM results presented in Figure 3-47 indicate that there are no obvious attributes of the blend morphology that could adversely affect the device performance. Figure 3-47. AFM height image of PProDOT-Hx 2 :CN-PPV / PCBM blend (1/4). Figure 3-48 shows the photovoltaic performance of PBProDOT-Hx 2 -CNV / PCBM (1/4) solar cells. Here the current-voltage characteristic shown in Figure 3-48a is indicative of a diode exhibiting the photovoltaic effect. The FF of 0.32 indicates that the device is operating as a more ideal diode than seen with the devices based on PProDOT-Hx 2 :CN-PPV. A J sc of ~1.2 mA/cm 2 and a power conversion efficiency of 0.21% are reasonable values for an unoptimized device. The interesting result is seen in Figure 3
151 48b. Here the IPCE data (black squares) is shown relative to the polymer absorption spectra (solid line). Also shown are the IPCE values multiplied by a factor of five (red triangles). In this expanded view of the IPCE data, it can be seen that the IPCE mirrors the polymer absorption near the maximum absorption of the polymer. However, the IPCE at that point is only ~2%. At the same time, IPCE values of greater than 15% are observed at wavelengths in the UV region of the spectrum. Such behavior indicates the polymer is not acting primarily as a photoexcited electron donor for PCBM. Instead the IPCE data appears to suggest that PCBM is the primary contributor to the photocurrent. A mechanism based on electron transfer from the HOMO of the polymer to the HOMO of photoexcited PCBM may be predominating in this case, as was illustrated schematically in Figure 1-12. Figure 3-48. Photovoltaic properties of PBProDOT-Hx 2 :CNV / PCBM solar cells based on a 1/4 blend (w/w) of polymer and PCBM. (a) Current voltage characteristic of a representative device as evaluated under AM1.5 conditions (100 mW/cm 2 ). (b) Photocurrent action spectra for the device. Black squares represent the IPCE value at the give wavelength. The solid curve is the absorption spectra of a pristine polymer film shown as reference. Red triangles represent the IPCE of the device multiplied by a factor of 5 to show the relationship to the polymer absorption spectrum. As with PProDOT-Hx 2 :CN-PPV devices, some effort was given to the optimization of PBProDOT-Hx 2 :CNV devices. For films spin coated under standard
152 conditions from 30 mg/mL solutions (dichlorobenzene) that were 1/4 in polymer and PCBM, film thicknesses were found to be ~30 nm. Increasing the blend concentration to 40 mg/mL resulted in little change in the film thickness or in the device performance. Figure 3-49 shows the AFM results for a film of PBProDOT-Hx 2 :CNV and PCBM that was spin coated under standard device conditions on PEDOT-PSS coated glass from dichlorobenzene. Here also, a very smooth surface is observed with a surface roughness calculated to be less than 1nm. As with the PProDOT-Hx 2 :CN-PPV films discussed above, the AFM image in Figure 3-49 gives little information about the precise morphology of the PBProDOT-Hx 2 :CNV / PCBM films. However, the relatively smooth and uniform nature of the film surface suggests that there is no macrophase separation in the blend. Figure 3-49. AFM height image of a PBProDOT-Hx 2 :CN-PPV / PCBM blend (1/4). Figure 3-50 shows the photovoltaic performance of Th-CN-PPV / PCBM (1/4) solar cells. Here, the diode characteristics are similar to the devices based on PProDOT-Hx 2 :CN-PPV as seen in Figure 3-46a, with a FF of 0.25. However in this case the efficiency and the short circuit current are an order of magnitude lower. The IPCE data
153 for Th-CN-PPV is shown in Figure 3-50b. Here the photocurrent matches the polymer absorption spectrum very closely. The peak efficiency of only ~1% indicates that the overall process of photogeneration and collection of charge carriers is not efficient in the devices constructed under the employed conditions. The precise origin for such low photocurrent efficiency is unclear at this time as it could be related to any number of the processes occurring in the solar cell. Importantly though, it does appear that the photoexcited polymer is a contributor to the photocurrent in the device. Figure 3-50. Photovoltaic properties of Th-CN-PPV / PCBM solar cells based on a 1/4 blend (w/w) of polymer relative to PCBM. (a) Current voltage characteristic of a representative device as evaluated under AM1.5 conditions (100 mW/cm 2 ). (b) Photocurrent action spectra for the device. Black squares represent the IPCE value at the give wavelength. The solid curve is the absorption spectra of a pristine polymer film shown as reference. Figure 3-51 shows an AFM image of a blend film of Th-CN-PPV and PCBM spin coated under standard conditions from dichlorobenzene (1/4 polymer to PCBM) on PEDOT-PSS coated glass. Such films were approximately 50 nm thick. It can be seen in the height image that this film has a significantly different surface morphology than analogous films with PProDOT-Hx 2 :CN-PPV or PBProDOT-Hx 2 :CN-PPV. In this case the film has a roughness of only 3 nm, but it can be seen that there appear to be fissures or at least valleys in the surface. Speculation about the valleys in the AFM image
154 could lead one to suggest that an incompatibility between the two components is leading to cracks or fissures at the interfaces in the film. While such speculation may be reasonable, a distinction between phases in this blend is not possible, as the overall surface appears to be of relatively uniform composition. It is evident from Figure 3-51 that the film quality in the blends of Th-CN-PPV and PCBM is less than optimal. It is certainly possible that the relatively poor performance of the Th-CN-PPV devices is due to the irregular surface morphology of the film. Optimization based on solvent choice, solution concentration, etc., could possibly lead to an improvement in film quality and a means of testing if the device performance can be improved or if it is limited based on the fundamental interaction of Th-CN-PPV with PCBM. Figure 3-51. AFM height image of Th-CN-PPV / PCBM blend (1/4). Figure 3-52 shows the photovoltaic performance of CN-TV / PCBM (1/4) solar cells. Here the diode behavior of the device as seen in Figure 3-52a is similar to that observed with the Th-CN-PPV device. A low FF of 0.25 and an efficiency of only ~0.01%, are less than desirable characteristics for a solar cell. The cause of such poor performance can be more easily understood in this case when the IPCE measurements for
155 this device are considered as seen in Figure 3-52b. Here it can be seen that the photocurrent does not mirror the optical absorption of the polymer to any extent. This is not surprising considering the results of the PL quenching experiment with this polymer, shown in Figure 3-44. Indeed the LUMO offset of the polymer relative to PCBM is calculated to be only ~0.2-0.4 eV, as CN-TV has the lowest lying LUMO of any of the studied polymers, using the electrochemically determined frontier orbital picture (Figure 3-37). Based on this simple model for predicting the driving force for electron transfer, it is assumed that the driving force is insufficient for effective electron transfer and thus the IPCE data is as expected. Figure 3-52. Photovoltaic properties of CN-TV / PCBM solar cells based on a 1/4 blend (w/w) of polymer relative to PCBM. (a) Current voltage characteristic of a representative device as evaluated under AM1.5 conditions (100 mW/cm 2 ). (b) Photocurrent action spectra for the device. Black squares represent the IPCE value at the given wavelength. The solid curve is the absorption spectra of a pristine polymer film shown as reference. Figure 3-53a shows an AFM image of the surface of a CN-TV / PCBM blend spin coated under standard conditions as used for the previously described AFM samples. These films were found to be ~75 nm thick. Here again, as with Th-CN-PPV, some cracks or holes are observed in the surface of the blend by AFM. The surface roughness was calculated to be on the order of 1 nm. However, it should be noted that by optical
156 microscopy large clusters were observed on the surface of the blend film (Figure 3-53b) and the AFM data was collected in the areas between these clusters. The clusters were closely spaced so that an AFM image could not be collected over a 5 m area due to the interaction of the AFM tip with the clusters. Figure 3-53. AFM height image and optical microscopy of a CN-TV / PCBM blend (1/4). (a.) AFM height image of blend. (b.) Microscope image of blend. (c.) Microscope image of pristine polymer film. These large clusters (Figure 3-53b) may be responsible for the poor performance of the devices, as a very low quality, macrophase separated film is not expected to give high efficiency devices. Additionally, these large clusters may be responsible for the unusual PL response of polymer-PCBM blend observed in Figure 3-44b. If these large clusters correspond to PCBM, then direct excitation of PCBM may result in the observation of PL from PCBM without any interaction with the polymer. In any case, the optical micrograph suggests that the two components are likely macrophase separated, which could also possibly account for the blend fluorescence observed in Figure 3-44b, which
157 can be attributed to the overlapped emission of the two, non-interacting components. Attempts to verify this hypothesis by fluorescence microscopy proved unsuccessful based on the difficulty to probe emission wavelengths beyond 700 nm and the extremely weak emission observed in the pristine and blend films. As a reference, Figure 3-53c shows the optical microscope image for a pristine film of CN-TV, which indicates that large surface features are present even in the pristine polymer. The clusters observed on the surface of the pristine polymer film are generally smaller than for the case of the polymer-PCBM blend, but they suggest that aggregates observed in the blend film could be due to polymer and/or PCBM. It is apparent from Figure 3-53c that the root of the device problems with CN-TV may be based on the polymer itself, as it appears that CN-TV has a strong tendency to aggregate (also observed with GPC as discussed earlier). For the case of CN-PPV, no operational devices could be constructed with PCBM as the film forming properties of CN-PPV were observed to be extremely poor. Uniform films of pristine CN-PPV were also not realized. A variety of blend concentrations and spin coating conditions were attempted without success. Table 3-5 summarizes the results of the devices based on the polymer-PCBM blends. For reference (Table 1-1), the most efficient (AM1.5) polymer-PCBM devices show short circuit currents of more 5 mA/cm 2 with fill factors above 0.5 and efficiencies over 2%. Additionally, IPCE values as high as 75% have observed in the best devices. Nonetheless, such devices are the result of years of optimization. Among the polymers studied here, PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNV give adequate photovoltaic performance, with only PProDOT-Hx 2 :CN-PPV showing reasonable photocurrent (IPCE) that can be directly attributed to the role of the polymer as a
158 photoexcited donor. Values for MEH-PPV / PCBM devices constructed under the same conditions are also given in Table 3-5. Table 3-5. Summary of polymer-PCBM (1/4) solar cell results. Polymer Voc Jsc (mA/cm2) FF (%) (AM1.5) IPCE(%) (max) PProDOT-Hx2 : CN-PPV 0.45 1.0 0.25 0.11 11 600 PBProDOT-Hx2CNV 0.51 1.6 0.28 0.22 2.5 620 CN-TV 0.71 0.10 0.25 0.015 0.10 580 Th-CN-PPV 0.44 0.10 0.26 0.011 1.0 560 CN-PPV ----------------------------------MEH-PPV 0.81 2.6 0.39 0.81 26 490 Results are an average from multiple devices to illustrate reproducible values. Based on the above results, it is apparent that none of the polymers equaled or rivaled the performance of MEH-PPV when fabricated under identical conditions. However, PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNV did show efficiencies on the order of 0.1-0.2%. As an additional approach to producing an efficient solar cell with a narrow band gap donor, a multicomponent approach was explored. Here, it was envisioned that by blending a narrow gap polymer with MEH-PPV and PCBM, the combination of the two polymers could effectively absorb solar photons, while transferring electrons to PCBM as a common acceptor. The two polymers selected for this â€œspectral broadeningâ€ approach were PProDOT-Hx 2 :CN-PPV and PBProDOT-Hx 2 -CNV as these polymers showed the most promising photovoltaic properties in combination with PCBM. Figure 3-54 illustrates the absorption spectra of these two polymers along with MEH-PPV as seen relative to the solar spectrum. It is clear that either low band gap polymer working in concert with MEH-PPV will allow the absorption of low energy photons that MEH-PPV could not absorb alone.
159 0.00.51.01.52.0 Wavelength (nm)Absorbance (a. u.)30045060075090010501200 0100200300400500600 Photon Flux (1025 photons m2 s-1) Figure 3-54. Thin film absorption spectra for MEH-PPV (orange line), PProDOT-Hx 2 :CN-PPV (blue line), and PBProDOT-Hx 2 -CNV (solid black line) relative to the AMO (black dashed line) and AM1.5 (red line) spectrum. While PBProDOT-Hx 2 -CNV exhibits a lower band gap and broader absorption in the visible region than PProDOT-Hx 2 :CN-PPV, it was observed that the IPCE value at the absorption maxima of PBProDOT-Hx 2 -CNV in blends with PCBM was only on the order of 2%, whereas PProDOT-Hx 2 :CN-PPV gave a value of 11%. As such, it could be conlcluded that PProDOT-Hx 2 :CN-PPV is effectively able to transfer an electron to PCBM upon photoexcitation, while the PBProDOT-Hx 2 -CNV devices operated under a mechanism in which photoexcitation of the polymer was not the primary means of charge carrier generation. Nonetheless, blends consisting of PProDOT-Hx 2 :CN-PPV, MEH-PPV, and PCBM were investigated along with devices consisting of PBProDOT-Hx 2 -CNV, MEH-PPV, and PCBM. Table 3-6 lists the AM1.5 results for a typical series of devices consisting of PProDOT-Hx 2 :CN-PPV, MEH-PPV, and PCBM. Here device number one (I) serves as a reference device and is fabricated using a 1:4 blend of MEH-PPV and PCBM. The
160 weight ratio of PCBM in the devices was then held constant (at 80%) while the composition of the polymer fraction varied from 100% MEH-PPV (device I) to 75% MEH-PPV and 25% PProDOT-Hx 2 :CN-PPV (device II) to 50% each of the two polymers (device III) to 75% PProDOT-Hx 2 :CN-PPV and 25% MEH-PPV (device IV). With devices I-III, the performance is relatively constant, despite the variation in active layer composition. In device IV however, with PProDOT-Hx 2 :CN-PPV serving as the majority fraction of the polymer portion, the device performance drops off sharply, in terms of current, FF, and efficiency. Figure 3-55 shows a comparison of the current-voltage characteristics for devices I and III. It can be seen that both devices show similar behavior despite the difference in device composition. Table 3-6. Results for devices consisting of PProDOT-Hx 2 :CN-PPV, MEH-PPV, and PCBM measured under AM1.5 conditions (100 mW/cm 2 ). Device Composition PProDOT-HX 2 :CN-PPV : MEH-PPV : PCBM V oc J sc (mA/cm 2 ) FF I 0 : 1 : 4 0.84 3.8 0.30 0.94 II 0.25 : 0.75 : 4 0.83 3.8 0.32 1.01 III 0.5 : 0.5 : 4 0.82 3.6 0.31 0.92 IV 0.75 : 0.25 : 4 0.77 2.0 0.22 0.33 -0.20.00.20.184.108.40.206-6-5-4-3-2-101 Device III Device ICurrent (mA/cm2)Bias (V) Figure 3-55. Current voltage characteristics for PProDOT-Hx 2 :CN-PPV / MEH-PPV / PCBM devices. Data is shown for devices I and III from Table 3-6.
161 Based on only the AM1.5 data, the relative contribution of the two polymers to the photocurrent is not apparent. In order to investigate the role of the two polymers in these three-component devices, the photocurrent action spectrum was evaluated for devices with varying compositions of the two polymers and a constant amount of PCBM. Figure 3-56 shows the results of IPCE measurements for such films. Data is shown for an MEH-PPV / PCBM (1/4) device and a PProDOT-Hx 2 :CN-PPV /PCBM (1/4) device for reference. It can be seen that the MEH-PPV / PCBM device shows an IPCE with a maximum of ~26% at 490 nm. This data closely resembles the absorption spectrum of MEH-PPV, indicating that the polymer is the primary contributor to the photocurrent. An IPCE of ~3% is observed for this device in the wavelength range of 600-700 nm. This is suspected to be primarily due to PCBM, as it is observed (Figure 3-56 inset) that MEH-PPV / PCBM blends show residual absorbance at long wavelengths that is not observed in the pristine polymer. As described in Figure 3-46b, the IPCE for the PProDOT-Hx 2 :CN-PPV /PCBM (1/4) device closely matches the absorption spectrum of PProDOT-Hx 2 :CN-PPV. Interestingly, in a device consisting of 0.5 parts PProDOT-Hx 2 :CN-PPV, 0.5 parts MEH-PPV, and 4 parts PCBM, shows an IPCE of ~28% at 480 nm and a value of ~12% at 600 nm, indicating that the photocurrent in devices containing the two donor polymers, in combination with PCBM, consists of contributions from both polymers. For the case of a device that contains 0.9 parts PProDOT-Hx 2 :CN-PPV, 0.1 parts MEH-PPV, and 4 parts PCBM, the photocurrent action spectra indicates that PProDOT-Hx 2 :CN-PPV is contributing to the long wavelength photocurrent, with an IPCE of ~12% at 600 nm. At 480-500 nm, no definite MEH-PPV associated peak is observed, owing to the small amount of MEH-PPV in the device. For this device, the
162 photocurrent peaks at 440 nm with an efficiency of 22%. This short wavelength contribution is undoubtedly due to MEH-PPV, PCBM, and PProDOT-Hx 2 :CN-PPV, which all show appreciable absorbance in this wavelength range. It is unclear why the IPCE is higher at 440 nm with this 0.9/0.1/4 device than it is with the MEH-PPV / PCBM device. Nonetheless, the results presented in Figure 3-56 suggest that both polymers are able to act as photoexcited electron donors in the three component blends. 400500600700051015202530 4005006007000.00.20.40.60.81.0 MEH-PPV / PCBM (1/4)MEH-PPVAbsorbance (a. u.)Wavelength (nm)IPCE (%)Wavelength (nm) Figure 3-56. IPCE data for blends consisting of PProDOT-Hx 2 :CN-PPV, MEH-PPV, and PCBM. PProDOT-Hx 2 :CN-PPV / PCBM (1/4) (), MEH-PPV / PCBM (1/4) (), MEH-PPV / PProDOT-Hx 2 :CN-PPV / PCBM (0.5 / 0.5 / 4) (), and MEH-PPV / PProDOT-Hx 2 :CN-PPV / PCBM (0.1 / 0.9 / 4) (). Inset shows the absorption spectrum of MEH-PPV and a 1:4 blend of MEH-PPV and PCBM on PEDOT-PPSS coated glass. An interesting point in this series of tricomponent devices is the sharp disparity in the performance of devices III and IV shown in Table 3-6. With a 1/1 MEH-PPV / PProDOT-Hx 2 :CN-PPV polymer phase, the AM1.5 efficiency is essentially unchanged relative to the MEH-PPV / PCBM device. However, increasing the amount of
163 PProDOT-Hx 2 :CN-PPV to 75% of the polymer component results in a decrease in the device performance. It can be postulated that a change in the morphology of the blend upon changing the majority polymer component is responsible for this behavior. In Figure 3-57a, it can be seen that with a 1/1 polymer phase, the surface morphology of the film is very uniform and the surface roughness is calculated to be 0.5 nm. With 75% PProDOT-Hx 2 :CN-PPV in the polymer component, the surface morphology of the film is observed to have changed considerably (Figure 3-57b). In this case, depressions are apparent in the surface of the film with diameters of ~100-300 nm (the white spot is dust) and the surface roughness is calculated to be ~2nm. While no definite results can be gleaned from this data about the composition of the surface, it is clear that a change in morphology is observed on going from a composition with equal amounts of the two polymers to one in which PProDOT-Hx 2 :CN-PPV is the major polymer component. Figure 3-57. AFM height images for three component blends of MEH-PPV/PProDOT-Hx 2 :CN-PPV/PCBM. (a) PProDOT-Hx 2 :CN-PPV/MEH-PPV/ PCBM (0.5/0.5/4) (b) PProDOT-Hx 2 :CN-PPV/MEH-PPV/ PCBM (0.75/0.25/4).
164 As a control experiment for the above set of spectral broadening experiments, the properties of an all-polymer device consisting of a 1:1 blend of PProDOT-Hx 2 :CN-PPV and MEH-PPV were evaluated. These devices gave very poor photovoltaic performance with a V oc of 0.79 V, a J sc of 0.030 mA/cm 2 , a FF of 0.28, and a power conversion efficiency of 0.0066%, which is similar to values observed in devices based on a single, pristine polymer as the photoactive layer (see chapter 1). The IPCE data for this type of device is shown in Figure 3-58 along with the absorption spectra for pristine MEH-PPV and a 1:1 blend of the two polymers. It is evident that MEH-PPV is the dominant contributor to the limited photocurrent generated in the device. The results for this device strongly suggest that PCBM is the primary electron acceptor for both MEH-PPV and PProDOT-Hx 2 :CN-PPV in the three-component devices discussed above. This device also indicates that if PProDOT-Hx 2 :CN-PPV serves as an electron acceptor for MEH-PPV, the process is not an efficient one. 4005006007000.000.050.100.220.127.116.110.350.40 IPCE (%)Wavelength (nm) Figure 3-58. IPCE data for an all-polymer device based on a 1:1 blend of PProDOT-Hx 2 :CN-PPV and MEH-PPV. Curve with black squares represents the IPCE. Red line is the absorption spectra of MEH-PPV and blue line is the absorption spectra of a 1:1 blend of PProDOT-Hx 2 :CN-PPV and MEH-PPV.
165 To further evaluate the energetic interaction between PProDOT-Hx 2 :CN-PPV and MEH-PPV, the photoluminescence was investigated for thin film blends of the two polymers (Figure 3-59). However, no excitation wavelength exists at which MEH-PPV can be selectively excited in the presence of PProDOT-Hx 2 :CN-PPV, and thus it is difficult to draw definite conclusions concerning charge or energy transfer between the two polymers. However, based on the low IPCE values in the above all-polymer device, it is unlikely that a charge transfer interaction exists between the two polymers. Further, upon excitation of the blend at 475 nm, strong emission is observed from both components, suggesting that charge transfer from MEH-PPV to PProDOT-Hx 2 :CN-PPV is unlikely, as is energy transfer from MEH-PPV to PProDOT-Hx 2 :CN-PPV. Figure 3-59. Absorbance and photoluminescence spectra for a 1:1 blend of PProDOT-Hx 2 :CN-PPV and MEH-PPV. (a) Absorbance spectrum of MEH-PPV (dash), PProDOT-Hx 2 :CN-PPV (dot), and 1:1 blend of the two polymers (solid line) spin coated onto PEDOT-PSS coated glass. (b) Photoluminescence spectra of MEH-PPV excited at 475 nm (dash), PProDOT-Hx 2 :CN-PPV excited at 475 nm (dot), and a 1:1 blend of the two polymers excited at 475 nm (solid line). One interesting result is seen in the absorption of the blend in Figure 3-59a (solid curve). Here it can be seen that even though the two polymers were blended in equal proportions by weight, the absorbance near the max of MEH-PPV is ~4 times as great as that at the max of PProDOT-Hx 2 :CN-PPV. Certainly both polymers are absorbing in the
166 range of the MEH-PPV peak to give an additive effect, but the disparity in the peak intensities indicates one reason why MEH-PPV makes a larger contribution to the IPCE in blend devices. This is a surprising result, as the two polymers show similar molar absorptivities. Following the same spectral broadening approach as that described for PProDOT-Hx 2 :CN-PPV, devices were constructed based on PBProDOT-Hx 2 -CNV, MEH-PPV, and PCBM. The device results under AM1.5 conditions are shown in Table 3-7. Here it can be seen that even a small amount of added PBProDOT-Hx 2 -CNV is enough to drop the efficiency of the MEH-PPV / PCBM device to an almost constant value, and the result is a device with the approximate characteristics of the PBProDOT-Hx 2 -CNV / PCBM (1/4) device discussed previously (see Table 3-5). Table 3-7. Results for devices consisting of PBProDOT-Hx 2 -CNV, MEH-PPV, and PCBM measured under AM1.5 conditions (100 mW/cm 2 ). Device Composition PBProDOT-HX 2 -CNV : MEH-PPV : PCBM V oc J sc (mA/cm 2 ) FF I 0 : 1 : 4 0.82 4.1 0.37 1.23 II 0.25 : 0.75 : 4 0.64 1.2 0.28 0.20 III 0.5 : 0.5 : 4 0.61 1.4 0.25 0.22 IV 0.75 : 0.25 : 4 0.60 1.5 0.26 0.24 Figure 3-60 shows the photocurrent action spectra for devices consisting of PBProDOT-Hx 2 -CNV, MEH-PPV, and PCBM. Here it can be seen that the photocurrent contribution from PBProDOT-Hx 2 -CNV is relatively minimal. This is not surprising based on the proposed mechanism of charge transfer from PBProDOT-Hx 2 -CNV to PCBM, in which it is expected that electron transfer from the HOMO of ground state PBProDOT-Hx 2 -CNV to the LUMO of and excited PCBM is the primary means of photoinduced charge transfer.
167 4005006007000510152025 IPCE (%)Wavelength (nm) Figure 3-60. IPCE data for devices with blends consisting of PBProDOT-Hx 2 -CNV, MEH-PPV, and PCBM. PBProDOT-Hx 2 -CNV / PCBM (1/4) (), MEH-PPV / PCBM (1/4) (), PBProDOT-Hx 2 -CNV / MEH-PPV / PCBM (0.5 / 0.5 / 4) (). 3.5 Conclusions and Outlook Herein, the synthesis of a family of narrow band gap cyanovinylene polymers has been described using Knoevenagel or Yamamoto polymerizations to give soluble polymers, or electropolymerization to give insoluble films. For these polymers, band gaps from 1.5-1.8 eV are realized, as determined optically. As such, polymers with absorption spectra that strongly overlap the solar spectrum and meet the target of a band gap less than 1.8 eV have been synthesized. Band structure elucidation using optical as well as electrochemical methods, indicates that all of the polymers (except PBEDOT-CNPV) have HOMO and LUMO energies that fall within the specifications defined for the ideal electron donor for a PCBM acceptor. Specifically in the case of PBProDOT-Hx 2 -CNV, the band structure is almost in perfect agreement with that of the ideal donor. Importantly, the band structure determination illustrates the value of ProDOT as a donor heterocycle for the synthesis of donor-acceptor polymers. Not only does ProDOT give a
168 stronger donor-acceptor interaction than alkythiophene, ProDOT containing polymers are more stable under ambient conditions than EDOT containing polymers. The performance of the polymers was further evaluated through the construction of a series of photovoltaic devices based on blends of each respective polymer with PCBM. In those devices based on PProDOT-Hx 2 :CN-PPV, PBProDOT-Hx 2 -CNV, and Th-CN-PPV, the polymer acts as a photoexcited electron donor for PCBM, and thus as a direct contributor to the photocurrent in the devices as evidenced by IPCE measurements. Of all the polymers synthesized, only PProDOT-Hx 2 :CN-PPV exhibits an IPCE value of >10% at the absorption maximum of the polymer. Such behavior makes this polymer especially attractive for a spectral broadening approach to device construction in which two donor polymers with distinct absorption spectra are used in concert with a single, common acceptor in a blend device. In this case, devices based on MEH-PPV, PProDOT-Hx 2 :CN-PPV, and PCBM were found to give efficiencies of ~1% under AM1.5 conditions. Furthermore, IPCE data indicates that both MEH-PPV and PProDOT-Hx 2 :CN-PPV contribute to the photocurrent in such devices. While polymers with targeted band structures were realized, the device efficiencies did not reflect the near-ideal optical properties of the donor polymers. In order to improve the devices, several levels of optimization are required. At the synthetic level, higher molecular weight polymers with a minimum level of defects are desired. Specifically, further evaluation and optimization of Knoevenagel and Yamamoto conditions is required for these systems. Additionally, the question of hole mobility in the polymers must be addressed at the next level of structural optimization. Modification of the polymer structure and/or processing conditions in the device could enhance the hole
169 mobility in the polymer. Limited charge carrier mobility could certainly play a role in limiting the efficiency of the devices. At the device level, solvent choice, blend composition, and blend concentration need to be systematically optimized along with the selection of the most suitable spin-coating conditions. Based on the encouraging IPCE results for several of the investigated polymers, enhancement of device performance could certainly be realized through such a systematic process of optimization. The family of polymers presented here also opens the door to an extensive electrochromic study as the polymers are observed to exhibit changes in their optical properties upon oxidation and reduction. From a fundamental point of view, it is of great interest to evaluate the in situ conductivity of the polymers upon oxidation and reduction in order to determine if these polymers exhibit both p and n type doping. It is thus clear that the work presented here is merely the beginning, as the possibilities presented by soluble, air stable, narrow band gap polymers are numerous. 3.6 Synthetic Details 2,3-dihydrothieno[3,4-b][1,4]dioxine-5-carbaldehyde (EDOT-CHO) (1). In 100 mL of dry THF, 6.65 g (46.8 mmol, 1.0 equiv) of EDOT was dissolved and the solution was cooled to C. Then 21 mL of 2.5 M n-butyllithium (53 mmol, 1.1 equiv) was added dropwise. The solution was then warmed to 0C and was stirred at this temperature for 20 minutes. The solution was again cooled to C and 6.8 mL (88 mmol, 1.9 equiv) of anhydrous DMF was added rapidly via syringe. The solution was then warmed to room temperature and stirred for one hour, after which it was poured into ice-water acidified with HCl. The resulting precipitate was isolated by filtration and washed with water. After recrystallization from methanol, the product was obtained (4.82 g, 61%) as
170 yellow plates. mp 127-129 0 C (Lit. 142 0 C). 1 H NMR (300 MHz, CDCl 3 ) 9.91 (s, 1H), 6.80 (s, 1H), 4.37 (m, 2H), 4.28 (m, 2H). 2,3-dihydrothieno[3,4-b][1,4]dioxine-2,5-dicarbaldehyde (EDOT-(CHO) 2 ) (2). EDOT (3.50 g, 24.6 mmol) was dissolved in 50 mL of dry THF and cooled to C. Then 25 mL of 2.3 M n-butyllithium (58 mmol, 2.4 equiv) was added to the solution. The reaction was then warmed to 0C for 30 minutes and then 7 mL of dry DMF was added after cooling the solution back to C. The reaction mixture was then warmed to room temperature and poured into 3 M HCl and extracted three times with dichloromethane. The organic layer was then washed with saturated NaHCO 3 and dried with MgSO 4 . The compound was then purified by column chromatography on silica with dichloromethane and petroleum ether (3:2) as the eluent to yield 2.81 g (58%) of an off-white solid. mp 143-145 0 C (Lit. 141 0 C). 1 H NMR (300 MHz, CDCl 3 ) 10.05 (s, 2H), 4.46, (s, 4H). 3,3-dihexyl-3,4-dihydro-2H-thieno[3,4-b][1,4]dioxepine-6,8-dicarbaldehyde (ProDOT-Hx 2 -(CHO) 2 ) (3b). To 120 mL of dry THF containing 9.69g (29.9 mmol) of ProDOT-Hx 2 at C, was added 33 mL of 2.1 M n-butlylithium (69 mmol, 2.3 equiv). The solution was warmed to 0C for 30 minutes and then cooled back to C. Then 8.5 mL of dry DMF (110 mmol, 3.7 equiv) was added rapidly via syringe and the solution was then warmed to room temperature and stirred for 1 h. After this, the solution was pored into 3 M HCl and extracted with dichloromethane. The organic layer was washed with saturated NaHCO 3 and then dried with MgSO 4 . After removing the solvent by rotary evaporation, the crude product was purified by column chromatography with 2:1 dichloromethane and hexanes to yield 8.92 g (79%) of product as a white solid. mp 75-77C. 1 H NMR (300 MHz, CDCl 3 ) 10.04 (s, 2H), 4.11 (s, 4H), 1.4-1.2 (m, 20H), 0.90
171 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 182.25, 155.19, 127.96, 78.57, 44.03, 32.41, 31.89, 30.21, 23.03, 22.83, 14.27. IR (KBr cm -1 ) 2954, 2930, 2860, 1670, 1652, 1492, 1463, 1385, 1248, 1198, 1109, 1091, 1072, 727, 690. HRMS calcd for C 21 H 32 O 4 S (M + ), 380.2021; found 380.2020. Anal. calcd for C 21 H 32 O 4 S: C, 66.28; H, 8.48; O, 16.82; S, 8.43. Found: C, 66.32; H, 8.59; S, 8.38. 3-dodecylthiophene-2,5-dicarbaldehyde (4). Crude 4-dodecylthiophene-2-carbaldehyde (10.8 g, 39 mmol) was dissolved in 200 mL benzene with 0.145 g p-toluenesulfonic acid (0.8 mmol) and 4 mL of ethylene glycol (70 mmol) and refluxed for 24 h with a Dean-Stark trap to remove water. The solution was cooled, poured into 20 mL of 10% (w/w) NaHCO 3 /water and washed three times with 10 mL of the same solution. The organic layer was then dried with MgSO 4 and the solvent was removed by rotary evaporation and the product (brown solid) was dried under vacuum. The dried product was then dissolved in 250 mL of dry THF and cooled to C, at which point 18 mL of 2.4 M n-butyllithium (43 mmol) was added dropwise and the solution was stirred for 30 minutes at the same temperature. Then 20 mL of dry DMF (260 mmol) was added rapidly via cannula and the reaction was warmed to room temperature and stirred for 1 h before pouring into 300 mL 2 M HCl. The mixture was then extracted with ether and the organic layer was washed with saturated NaHCO 3 , followed by water before drying with MgSO 4 . The crude product was then stirred with 100 mL of 3M H 2 SO 4 at room temperature and then worked up again as described above. Column chromatography with 9:1 hexanes and ethyl acetate to give 3.27 g (27% from 3-dodecylthiophene) of the product as a white solid. mp 28-29C. 1 H NMR (300 MHz, CDCl 3 ) 10.14 (s, 1H), 9.97 (s, 1H), 7.65 (s, 1H), 2.44 (t, 2H), 1.70 (m, 2H), 1.4-1.2 (br,
172 18H), 0.88 (t, 3H). 13 C NMR (75 MHz, CDCl 3 ) 183.57, 183.20, 152.21, 148.09, 143.52, 137.34, 32.12, 31.43, 29.84, 29.82, 29.80, 29.69, 29.55, 29.54, 29.43, 28.73, 22.90, 14.33. IR (KBr cm -1 ) 2923, 2850, 1671, 1529, 1463, 1219, 1163, 863, 780, 685, 582, 481. HRMS calcd for C 18 H 28 O 2 S (M + ), 308.1810; found 308.1810. Anal. calcd for C 18 H 28 O 2 S: C, 70.08; H, 9.15; O, 10.37; S, 10.39. Found: C, 70.54; H, 9.43. 2,5-bis(dodecyloxy)terephthaldehyde (5). In 600 mL of dry ether, 88.9 g (147 mmol, 1 equiv) of 1,4-dibromo-2,5-bis(dodecyloxy)benzene was dissolved and cooled to 0C. Then 142 mL (369 mmol, 2.5 equiv) of 2.6 M n-butyllithium was added dropwise. After addition was complete, 90 mL of DMF (1160 mmol, 7.9 equiv) was added rapidly and the solution was stirred and warmed to room temperature overnight. The solution was then poured into water, which had been acidified with a few drops of HCl. The mixture was then extracted with dichloromethane and the organic layer was dried with MgSO 4 . Column chromatography on silica (1:1 hexanes and toluene) followed by recrystallization from hexanes gave the product as long yellow needles (21.16 g, 29%). mp 81-83C. 1 H NMR (300 MHz, CDCl 3 ) 10.52 (s, 2H), 7.43 (s, 2H), 4.08 (t, 4H), 1.83 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H). 13 C NMR (300 MHz, CDCl 3 ) 189.68, 155.48, 129.54, 111.87, 69.50, 32.13, 29.86, 29.84, 29.79, 29.77, 29.56, 29.54, 29.28, 26.24, 22.91, 14.33. IR (KBr cm -1 ) 2870, 2850, 1680, 1490, 1471, 1429, 1390, 1284, 1215, 1126, 996, 884, 717, 696. HRMS calcd for C 32 H 54 O 4 (M + ), 502.4022; found 502.4028. Anal. calcd for C 32 H 54 O 4 : C, 76.45; H, 10.83; O, 12.73. Found: C, 76.58; H, 11.19. 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene (Route A) (7). Sodium cyanide (2.41 g, 49 mmol, 2.5 equiv) and 1,4-bis(bromomethyl)-2,5-bis(dodecyloxy)benzene (12.0 g, 19 mmol, 1 equiv) were dissolved in 250 mL of dry
173 DMF and the solution was heated to 110C. The reaction was stirred for two days, during which time the reaction turned dark orange and a precipitate was observed to form. The reaction was then cooled to room temperature and poured into 750 mL of cold water that was 0.5 M in sodium hydroxide. The crude solid was isolated by filtration and then taken up in chloroform and washed with water. The organic layer was then dried with MgSO 4 . The crude solid was then purified by column chromatography on silica (1:1 hexanes and dichloromethane) followed by recrystallization from ethanol and chloroform (2:3) to give 5.12 g (51%) of the product as a white solid. mp 98-100C. 1 H NMR (300 MHz, CDCl 3 ) 6.91 (s, 2H), 3.97 (t, 4H), 3.70 (s, 4H), 1.79 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 150.27, 119.35, 118.05, 112.89, 69.27, 32.13, 29.88, 29.85, 29.82, 29.78, 29.59, 29.48, 26.82, 25.05, 22.90, 18.86, 14.33. IR (KBr cm -1 ) 2961, 2870, 2849, 2917, 2246, 1514, 1466, 1427, 1392, 1341, 1222, 1209, 1070, 999, 846, 720, 686. Anal. calcd for C 34 H 56 N 2 O 2 : C, 77.81; H, 10.76; N, 5.34; O, 6.10. Found: C, 77.85; H, 11.07; N, 5.18. 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene (Route B) (7). In 100 mL of dry DMSO, 1.0 g of sodium cyanide and 3.327 g of 1,4-bis(chloromethyl)-2,5-bis(dodecyloxy)benzene (6.13 mmol) were dissolved and the solution was heated to 50C for 2 h and then heated to 80C for 30 minutes. The reaction mixture was then poured into water and the precipitate was isolated by filtration. Chromatography on silica (1:1 hexanes and dichloromethane) gave the product as a white solid (0.618 g, 20%). 1 H NMR (300 MHz, CDCl 3 ) 6.91 (s, 2H), 3.97 (t, 4H), 3.70 (s, 4H), 1.79 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 150.27, 119.35, 118.05, 112.89, 69.27, 32.13, 29.88, 29.85, 29.82, 29.78, 29.59, 29.48, 26.82, 25.05, 22.90, 18.86, 14.33. HRMS
174 calcd for C 34 H 56 N 2 O 2 (M + ), 524.4342; found 524.4341. Anal. Calcd. for C 34 H 56 N 2 O 2 : C, 77.81; H, 10.76; N, 5.34; O, 6.10. Found: C, 77.55; H, 11.07; N, 5.34. 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene (Route C) (7). In 300 mL of dry THF, 3.32 g (6.60 mmol) of 1,4-bis(hydroxymethyl)-2,5-bis(dodecyloxy)benzene was dissolved. Then 4.8 mL of methanesulfonyl chloride was added at room temperature and the mixture was stirred for 1 h. Then 11.7 mL of triethylammine was added and the reaction was allowed to stir overnight. The reaction mixture was then poured into sat. K 2 CO 3 and extracted with ether. The organic layer was dried over MgSO 4 and the solvent was removed to give a viscous orange oil. This oil was then dissolved in 250 mL of anhydrous DMF and 1.2 g of potassium cyanide was added. The reaction was stirred at 50C for 24 hours. The reaction was then cooled to room temperature and 100 mL of 1 M NaOH was added, followed by extraction with ether. The organic layer was then dried on MgSO 4 . Column chromatography on silica (1:1 hexanes and dichloromethane) gave 0.709 g (21%) of a white solid. 1 H NMR (300 MHz, CDCl 3 ) 6.91 (s, 2H), 3.97 (t, 4H), 3.70 (s, 4H), 1.79 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 150.27, 119.35, 118.05, 112.89, 69.27, 32.13, 29.88, 29.85, 29.82, 29.78, 29.59, 29.48, 26.82, 25.05, 22.90, 18.86, 14.33. 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene (Route D) (7). In a Schlenk flask under Argon, 3.66 g (32.7 mmol, 4.1 equiv) of potassium t-butoxide was dissolved in 15 mL of dry DME. The solution was then cooled to C. At this time, a solution of 3.26 g (16.7 mmol, 2.1 equiv) tosylmethyl isocyanide (TosMIC) in 15 mL DME was added dropwise. The resulting brown-orange solution was then cooled to C. Then a solution of 2,5-bis(dodecyloxy)terephthalaldehyde (4.00 g, 8.0 mmol, 1 eqiuv) in 100 mL
175 of DME and 500 mL dichloromethane was added dropwise. The resulting orange solution was heated to 40C and then 50 mL of dry methanol was added dropwise. The solution was then heated at 40C for 1 h and then cooled to room temperature and poured into water which had been acidified with a few drops of acetic acid. The mixture was extracted with dichloromethane, after which the organic layer was washed with saturated sodium bicarbonate, followed by water. The organic layer was then dried with MgSO 4 . The crude product was purified by column chromatography on silica (1:1 hexanes and dichloromethane) to give 1.03 g (25%) of product. 1 H NMR (300 MHz, CDCl 3 ) 6.91 (s, 2H), 3.97 (t, 4H), 3.70 (s, 4H), 1.79 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 150.27, 119.35, 118.05, 112.89, 69.27, 32.13, 29.88, 29.85, 29.82, 29.78, 29.59, 29.48, 26.82, 25.05, 22.90, 18.86, 14.33. 2,5-thiophene-diacetonitrile (9). In a 100 mL flask, 40 mL of 37% formaldehyde solution and 10 mL of concentrated HCl were bubbled with HCl gas for 60 minutes. Then 9.93 g (118 mmol) of thiophene was added dropwise via an addition funnel. The initially clear solution became cloudy white and then a black lower layer formed and the solution became hot. The solution was stirred with continuous inlet of HCl gas for 10 minutes after the addition of thiophene was complete. The reaction mixture was then poured into water and extracted with chloroform. The organic layer was then washed with saturated NaHCO 3 (x2) followed by water and then dried with MgSO 4 . The dried solution was filtered through a plug of NaHCO 3 and the solvent was then removed under vacuum. The crude bischloromethyl product was then dissolved in 50 mL dry DMF and added dropwise to a solution of 200 mL dry DMF containing 14 g of NaCN (290 mmol). The reaction mixture was then stirred at room temperature overnight. The reaction was then
176 heated to 60C for 4 h, after which it was cooled to room temperature and poured into brine. After extraction with ether (x3), the organic layer was washed with water and dried with MgSO 4 . The crude product was then passed through silica gel with dichloromethane and hexanes (3:1) as the eluent, followed by distillation (105-115C, 0.006 Torr) and recrystallization (CHCl 3 and CCl 4 or pentane and dichlromethane) to give the product (1.72 g, 9% from thiophene) as long clear needles. mp 35-36C (Lit. 39-40C). 1 H NMR (300 MHz, CDCl 3 ) 6.95 (s, 2H), 3.88 (s, 4H). 13 C NMR (75 MHz, CDCl 3 ) 131.96, 127.65, 116.59, 18.93. IR (KBr cm -1 ) 3090, 2947, 2931, 2910, 2259, 1742, 1603, 1561, 1498, 1415, 1304, 1230, 1132, 1111, 1032, 921, 799, 727, 472. HRMS calcd for C 8 H 6 N 2 S (M + ), 162.0252; found 162.0256. Anal. calcd for C 8 H 6 N 2 S: C, 59.23; H, 3.73; N, 17.27; S, 19.77. Found: C, 59.19; H, 3.61; N, 17.10. 3,3-dihexyl-3,4-dihydro-2H-thieno[3,4-b][1,4]dioxepine 118a (ProDOT-Hx 2 ) (12). In 200 mL of dry toluene, 4.649 g (32.3 mmol) of 3,4-dimethoxythiophene, 18.5 g (67 mmol) of 2,2-dihexylpropane-1,3-diol, and 1.75 g (9.2 mmol) of p-toluenesulfonic acid were dissolved. The solution was then heated to refluxed over 4 molecular sieves for 12 h. The reaction mixture was then cooled and poured into water and extracted with toluene. The toluene was then removed by rotary evaporation and the crude oil was purified by column chromatography on silica gel eluting with 3:1 hexane:ethyl dichloromethane to yield 8.73 g (84%) of product as a clear oil. 1 H NMR (300 MHz, CDCl 3 ) 6.41 (s, 2H), 3.84 (s, 4H), 1.40-1.20 (m, 20H), 0.89 (t, 6H). 13 C NMR (75 MHz CDCl 3 ) 149.9, 104.9, 77.8, 43.9, 32.0, 31.8, 30.4, 23.0, 22.9, 14.3. 4-dodecylthiophene-2-carbaldehyde (13). At C, 38 mL of 2.3 M n-butyllithium (87 mmol, 2.2 equiv) was added dropwise to a solution of 9.838 g (39.0
177 mmol, 1 equiv) of 3-dodecylthiophene (Aldrich) in 200 mL dry THF. The solution was then stirred at this temperature for 45 minutes, after which 24 mL (310 mmol) of dry DMF was added rapidly via syringe. The resulting solution was warmed to room temperature and stirred for 1 h. The reaction was then poured into 2M HCl (300 mL) and extracted with dichloromethane. The organic layer was then washed with saturated NaHCO 3 and dried with MgSO 4 . Upon removal of solvent by rotary evaporation, 10.8 g of crude product were obtained. The crude product consisted of a 9:1 mixture of 4-dodecylthiophene-2-carbaldehyde and 3-dodecylthiophene-2-carbaldehyde and was used in the next step without further purification. 1 H NMR (300 MHz, CDCl 3 ) 10.04 (s, 0.13 H), 9.87 (s, 1.0 H), 7.65-7.63 (d, 0.16 H), 7.61 (s, 1.06 H), 7.37 (s, 1.0 H), 7.02-7.00 (d, 0.14 H), 2.64 (t, 2.24 H), 1.63 (m, 2.56 H), 1.4-1.2 (m, 23.85 H), 0.88 (t, 3.92 H). 1,4-bis(dodecyloxy)benzene (14). Hydroquinone (5.13 g, 47 mmol) was dissolved in 100 mL of ethanol and 5.73 g KOH (100 mmol, 2.2 equiv) in 50 mL was added to give an orange solution that was stirred for 1 h at room temperature. Then 27.11 g (110 mmol, 2.3 equiv) of 1-bromodecane in 25 mL of ethanol was added dropwise. The reaction was then heated to 50C for 24 h, during which time a white precipitate formed. The reaction was then poured into water and filtered. The crude solid was washed with water, dried, and recrystallized from ethanol to give 13.82 g (67%) of product. mp 70-71C (Lit. 70-72 0 C). 1 H NMR (300 MHz, CDCl 3 ) 6.81 (s, 4H), 3.89 (t, 4H), 1.78 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H). 1,4-bis(bromomethyl)-2,5-bis(dodecyloxy)benzene (15). To a solution consisting of 30 mL of 33% HBr in acetic acid and 110 mL of glacial acetic acid, 13.8 g (31 mmol, 1 equiv) of 1,4-bis(dodecyloxy)benzene and 2.80 g of paraformaldehyde (93
178 mmol, 3 equiv) were added. The white slurry was heated to 70-75C, resulting in the dissolution of all solids. After a few minutes at this temperature, a white precipitate began to form and stirring was continued for 2 h. The reaction mixture was then cooled to 0C, poured into cold water, and filtered. The crude off-white product was then recrystallized with hexane followed by recrystallization from dichloromethane-methanol to yield 12.02 g (61%) of the product as a white solid. mp 92-93C (Lit. 203 96-97C). 1 H NMR (300 MHz, CDCl 3 ) 6.85 (s, 2H), 4.52 (s, 4H), 3.98 (t, 4H), 1.81 (m, 4H), 1.5-1.2 (br 36H), 0.88 (t, 6H). 1,4-dibromo-2,5-bis(dodecyloxy)benzene (16). Hydroquinone (110 g, 1 mol) was added to a mixture of 250 mL of glacial acetic acid and 250 mL of dichloromethane. A solution of 110 mL of bromine (2.15 mol) in 75 mL dichloromethane was added dropwise with stirring. After addition was complete, the reaction was stirred overnight. The reaction mixture was filtered, washed with water, and recrystallized from isopropanol and water to give dibromohydroquinone as a white solid. A portion of the product (103.3 g, 386 mmol) was dissolved in 300 mL of ethanol and a solution of 56.5 g potassium hydroxide (1.4 mol, 2.6 equiv) in 400 mL of ethanol was added dropwise. The solution was stirred at room temperature for 1 h. Then, 295 mL (306 g, 1.23 mol, 3.2 equiv) of 1-bromodecane was added dropwise to the solution. The solution was then heated to 50C for 12 h. The reaction mixture was then cooled to room temperature and poured into water. The crude product was then isolated by filtration and recrystallized from 3:1 ethanol:benzene to give 95.8 g (41% from dibromohydroquinone) of the product as a white solid. mp 76-77C (Lit. 160 77-79C). 1 H NMR (300 MHz, CDCl 3 ) 7.08 (s, 2H), 3.94 (t, 4H), 1.79 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H).
179 1,4-bis(hydroxymethyl)-2,5-bis(dodecyloxy)benzene (17). Lithium aluminum hydride (4.0 g, 105 mmol, 4.6 equiv) was dissolved in 500 mL of dry THF and cooled to 0C. Then, 11.328 g (22.6 mmol) of 2,5-bis(dodecyloxy)terephthalaldehyde was added to the solution and the solution quickly changed from yellow to white. The solution was then stirred overnight at room temperature. Then 250 mL of 3 M HCl was added dropwise to the solution. The solution was then extracted with ethyl acetate. Rotary evaporation yielded a crude white solid that was recrystallized from 3:1 ethanol and benzene to give 9.759 g (86%) of the product as a white solid. mp 115-117C. 1 H NMR (300 MHz, CDCl 3 ) 6.84 (s, 2H), 4.67 (d, 4H), 3.98 (t, 4H), 2.38 (t, 2H), 1.78 (m, 4H), 1.5-1.2 (br, 36H), 0.88 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 150.87, 129.29, 112.57, 68.99, 62.40, 32.14, 29.88, 29.85, 29.82, 29.80, 29.65, 29.60, 29.57, 26.37, 22.91, 14.33. HRMS calcd for C 32 H 58 O 4 (M + ), 506.4335; found 506.4331. Anal. calcd for C 32 H 58 O 4 : C, 75.84; H, 11.54; O, 12.63. Found: C, 75.87; H, 11.88. 1,4-bis(chloromethyl)-2,5-bis(dodecyloxy)benzene (18). In 100 mL of dry DMF and 100 mL of dry acetonitrile, 9.256 g (18.3 mmol) of 1,4-bis(hydroxymethyl)-2,5-bis(dodecyloxy)benezene was dissolved along with 500 mg of sodium iodide and 3.689 g of sodium cyanide (75.3 mmol) and 10 mL of TMSCl (79 mmol). The solution was then heated at 65C for two days. At that time, water was added and the solution was extracted with hexane. The hexane was washed with brine. After removal of the solvent, the crude solid was recystallized from chlorform and ethanol (1:1) to give the product as a white solid (4.38 g, 42%). mp 91-93C. 1 H NMR (300 MHz, CDCl 3 ) 6.91 (s, 2H), 4.63 (s, 4H), 3.98 (t, 4H), 1.8 (m, 4H), 1.5-1.2 (b, 36H), 0.88 (t, 6H). HRMS calcd for
180 C 32 H 56 Cl 2 O 2 (M + ), 542.3657; found 542.4. Anal. calcd for C 32 H 56 Cl 2 O 2 : C, 70.69; H, 10.38; O, 5.89; Cl, 13.04. Found: C, 70.39; H, 10.10; Cl, 12.13. 1,4-phenylene-diacetonitrile (19). In a flask under argon was placed 3.06 g (15.7 mmol) of TosMIC. In a separate flask, 3.45 g (30.8 mmol) of tBuOK was added and both flasks were charged with 15 mL of dry DME. In a third flask, 1.008 g of terephthalaldehyde (7.5 mmol) was dissolved in 25 mL of DME. The tBuOK solution was then cooled to C and the TosMIC solution was slowly dripped into the flask. After addition was complete, the combined solution was cooled to C and then the solution of terephthalaldehyde was added dropwise. The reaction was then stirred for a full hour at C before 40 mL of anhydrous methanol were added and the reaction was subsequently heated to reflux for 15 minutes. The reaction was then cooled to room temperature and the solvent was removed. Water acidified with a few drops of acetic acid was then added and the mixture was extracted with dichloromethane. The organic layer was dried with MgSO 4 and after removal of the solvent, the crude product was purified by chromatography on silica (dichloromethane) to give 0.380 g (33%) of the product as a white solid. 1 H NMR (300 MHz, CDCl 3 ) 7.37 (s, 4H), 3.77 (s, 4H) (matched authentic sample, Aldrich (6)). HRMS calcd for C 10 H 8 N 2 (M + ), 156.0687; found 156.0685. Phenylacetonitrile (20). In 15 mL of ether, 0.543 g (5.0 mmol) of benzyl alcohol and 2.01 g (7.7 mmol) of triphenylphosphine were dissolved. Then 1.105 g (7.6 mmol) of dimethylazodicarboxylate (DMAD) was added dropwise at 0C. After stirring for 10 minutes, 0.639 g of acetone cyanohydrin (7.5 mmol) was added dropwise at 0C. The solution was then refluxed overnight. The solution was then poured in to water and extracted with dichloromethane. The organic layer was dried over MgSO 4 and a crude oil
181 was isolated upon removal of the solvent. Chromatography gave the product as a clear oil (0.216 g, 37%). 1 H NMR (300 MHz, CDCl 3 ) 7.35 (m, 5H), 3.73 (s, 2H) (matched authentic sample, Aldrich). Bis-EDOT-CN-PPV (Route 1) (21). In a solution consisting of 25 mL of 1:1 t-butanol and THF, 0.506 g (3.0 mmol) of EDOT-CHO and 0.239 g (1.5 mmol) of 1,4-phenylenediacetonitrile were dissolved at room temperature. Then 0.375 g (3.3 mmol, 1.1 equiv) of potassium t-butoxide was added, resulting in the immediate darkening of the solution. The solution was then heated to 70C for 2 h. At this time the reaction was cooled to room temperature and poured into water yielding a brick red precipitate. The precipitate was redissolved in chloroform and washed with water. Upon removal of the solvent, 0.562 g (81%) of a brick red solid was isolated. mp dec. > 220 0 C. 1 H NMR (300 MHz, DMSO-d 6 ) 7.86 (s, 2H), 7.71 (s, 4H), 7.08 (s, 2H), 4.39 (m, 4H), 4.28 (m, 4H). HRMS calcd for C 24 H 16 N 2 O 4 S 2 (M + ), 460.0551; found 460.0542. Anal. calcd for C 24 H 16 N 2 O 4 S 2 : C, 62.59; H, 3.50; N, 6.08; O, 13.90; S, 13.93. Found: C, 60.45; H, 3.50; N, 5.57. Bis-EDOT-CN-PPV (Route 2) (21). In a solution consisting of 30 mL of 3:1 t-butanol and THF, 0.534 g (3.1 mmol) of EDOT-CHO and 0.255 g (1.6 mmol) of 1,4-phenylenediacetonitrile were dissolved by heating to 50C. Then 0.035 g (0.31 mmol, 0.1 equiv) of potassium t-butoxide and 0.3 mL of 0.1 M tetrabutylammonium hydroxide (0.3 mmol, 0.1 equiv) was added, resulting in the immediate darkening of the solution. The solution was then heated to 50C for 2 h. At this time the reaction was cooled to room temperature and poured into water yielding a sticky black precipitate. The precipitate was redissolved in chloroform and washed with water. Upon removal of the solvent, 0.468 g
182 (65%) of a dark red solid was isolated. mp dec. > 220 0 C. 1 H NMR (300 MHz, DMSO-d 6 ) 7.88 (s, 2H), 7.72 (s, 4H), 7.10 (s, 2H), 4.40 (m, 4H), 4.29 (m, 4H). HRMS calcd for C 24 H 16 N 2 O 4 S 2 (M + ), 460.0551; found 460.0550. Anal. calcd for C 24 H 16 N 2 O 4 S 2 : C, 62.59; H, 3.50; N, 6.08; O, 13.90; S, 13.93. Found: C, 62.03; H, 3.59; N, 5.99. 3-octylthiophene-2-carbaldehyde (22). Freshly dried 2-bromo-3-octylthiophene (5.96 g, 21.7 mmol) was dissolved in 100 mL of dry THF and cooled to C. Then 9.5 mL of 2.5 M n-butyllithium (23.8 mmol, 1.1 equiv) was added dropwise via syringe. After 40 minutes, 8.5 mL of dry DMF (5 equiv) was added. The solution was then allowed to warm to room temperature and stirred overnight. The reaction mixture was then poured into 200 mL of 3 M HCl and extracted with dichloromethane. The organic layer was then washed with saturated NaHCO 3 followed by water, before drying with MgSO 4 . After the solvent was removed, the crude product was purified by chromatography on silica gel eluting with 9:1 (hexanes and ethyl acetate). The product was isolated in 82% yield (3.99 g) as a slightly yellow oil. 1 H NMR (300 MHz, CDCl 3 ) 10.04 (s, 1H), 9.87 (s, 0.1H), 7.64 (d, 1H), 7.61 (s, 0.1H), 7.37 (s, 0.1H), 7.01 (d, 1H), 2.96 (t, 2H), 2.64 (t, 0.2H), 1.67 (m, 2.2H), 1.4-1.2 (b, 11H), 0.88 (t, 3.3H). 13 C NMR (75 MHz, CDCl 3 ) 183.10, 182.34, 153.04, 137.83, 134.58, 130.88, 31.94, 31.56, 29.48, 29.46, 29.45, 29.41, 29.31, 29.30, 29.28, 28.60, 22.76, 14.20. HRMS calcd for C 13 H 20 OS (M + ), 224.1235; found, 224.1240. 3,3-dihexyl-3,4-dihydro-2H-thieno[3,4-b][1,4]dioxepine-6-carbaldehyde (ProDOT-Hx 2 -CHO) (23). In 40 mL of dry THF, ProDOT-Hx 2 (3.394 g, 10.5 mmol) was dissolved and cooled to C. Then 4.6 mL of 2.5 M n-butyllithium (11.5 mmol, 1.1 equiv) was added dropwise. The solution was then warmed to 0C and was stirred at
183 this temperature for 20 minutes. The solution was again cooled to C and 5 mL (65 mmol, 6.2 equiv) of dry DMF was added rapidly via syringe. The solution was then warmed to room temperature and stirred for 1 h, after which it was poured into ice-water acidified with HCl. The solution was then extracted with THF and the organic layer was dried with MgSO 4 . After rotary evaporation, the crude oil was purified by column chromatography on silica gel eluting with 10:1 hexanes and ethyl acetate (followed by filtering through activated carbon to remove persistent colored impurities) to give 2.15 g (58%) of the product as a faint-yellow oil. 1 H NMR (300 MHz, CDCl 3 ) 9.91 (s, 1H), 6.85 (s, 1H), 4.06 (s, 2H), 3.90 (s, 2H), 1.4-1.2 (m, 20 H), 0.89 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 181.13, 156.30, 148.39, 122.01, 115.09, 77.91, 77.73, 43.74, 32.15, 31.68, 30.04, 22.80, 22.62, 14.05. HRMS calcd for C 20 H 32 O 3 S (M + ), 352.2072; found 352.2055. Anal. calcd for C 20 H 32 O 3 S: C, 68.14; H, 9.15; O, 13.62; S, 9.10. Found: C, 68.06; H, 9.10; S, 9.02. 3-octylthiophene-2-carbonitrile (24). Freshly dried 2-bromo-3-octylthiophene (7.00 g, 25.4 mmol) was dissolved in 100 mL of dry THF and cooled to C. Then 10.5 mL of 2.6 M n-butyllithium (27.3 mmol) was added dropwise and the reaction was stirred for 1 h at C. The solution was then transferred via cannula to a solution of 27 mmol of ZnCl 2 in THF at 0C. The reaction was stirred for an additional hour at 0C and then transferred via cannula to a flask containing 0.663 g of Ni(acac) 2 , 0.684 g of cyclohexyl-diphenylphosphine, and 3.05 g (25.4 mmol) of bromoacetonitrile in 50 mL of THF. The reaction was then heated at 65-70C for two hours. The reaction was then cooled to room temperature and poured into 150 mL of 1 M HCl and extracted with ether. The organic layer was washed with water and dried over MgSO 4 . Chromatography
184 on silica gel (10:1 hexanes and ethyl acetate) gave 1.79 g (30%) of the product as a slightly yellow oil. 1 H NMR (75 MHz, CDCl 3 ) 7.17 (d, 1H), 6.89 (s, 0.1 H), 6.87 (d, 1H), 6.82 (s, 0.1H), 3.86 (s, 0.2H), 3.79 (s, 2H), 2.55 (t, 2.2H), 1.59 (m, 2.2H), 1.4-1.2 (b, 11H), 0.88 (t, 3.3H). 13 C NMR (75 MHz, CDCl 3 ) 143.93, 141.60, 129.45, 128.78, 124.26, 123.92, 120.47, 117.25, 32.13, 30.54, 29.70, 29.60, 29.50, 28.55, 22.90, 19.03, 16.63, 14.37. HRMS calcd for C 14 H 21 NS (M + ), 235.1395; found, 235.1395. 2-(3,3-dihexyl-3,4-dihydro-2H-thieno[3,4-b][1,4]dioxepin-6-yl)acetonitrile (ProDOT-Hx 2 -ACN) (25). In 150 mL of dry THF, ProDOT-Hx 2 (13.71 g, 42.3 mmol, 1 equiv) was dissolved and cooled to C. Then 18 mL of 2.5 M n-butyllithium (45 mmol, 1.1 equiv) was added and the reaction was stirred at C for 1 h. Then the solution was transferred via cannula to an addition funnel and subsequently added to a solution consisting of 50 mL of THF and 45 mL 1.0 M ZnCl 2 (45 mmol) at 0C. The resulting solution was stirred at 0C for 2 h. This solution was then transferred via cannula to a solution consisting of 5.18 g bromoacetonitrile (43.2 mmol), 1.14 g cyclohexyldiphenylphosphine (4.3 mmol), and 1.17 g Ni(acac) 2 (4.6 mmol) in THF. The resulting brown solution was heated to 70C overnight. After cooling to room temperature, the reaction was then poured into 500 mL of 1 M HCl. After extraction, the organic layer was washed with 0.2 M HCl and then dried over MgSO 4 . Column chromatography on silica gel eluting with hexanes and ethyl acetate (10:1) gave the product (4.35 g, 28%) as a viscous yellow oil. 1 H NMR (300 MHz, CDCl 3 ) 6.38 (s, 1H), 3.88 (s, 2H), 3.83 (s,2H), 3.70 (s, 2H), 1.4-1.2 (br, 20H), 0.89 (t, 6H). 13 C NMR (75 MHz, CDCl 3 ) 149.56, 147.87, 117.09, 108.97, 103.28, 77.99, 77.90, 44.19, 31.97,
185 31.91, 30.31, 22.93, 22.83, 15.44, 14.27. HRMS calcd for C 21 H 33 O 2 NS (M + ), 363.2232; found 363.2248. 2-bromo-3-octylthiophene (26). In 30 mL of glacial acetic acid, 5.063 g (25.8 mmol) of 3-octylthiophene was dissolved. Then 4.61 g (25.9 mmol) of freshly recrystallized NBS was added. As the internal temperature of the solution increased from 17 to 33C, all the NBS dissolved. The reaction mixture then quickly cooled back to room temperature. The mixture was then poured into 150 mL of water extracted with ether. The organic layer was then washed with 2 M NaOH (5x100 mL), followed by washing with water. The organic layer was then dried with MgSO 4 and the solvent ws subsequently removed under vacuum to give 7.00 g (99%) of the product as a clear oil. 1 H NMR (300 MHz, CDCl 3 ) 7.17 (d, 1H), 6.79 (d, 1H), 2.55 (t, 2H), 1.54 (m, 2H), 1.4-1.2 (b, 10H), 0.88 (t, 3H). Bis-ProDOT-Hx 2 -CNV-1,4-(2,5-didodecyloxybenzene) (BProDOT-Hx 2 -CNPV) (27). In 60 mL of THF and 60 mL of t-butanol, ProDOT-Hx 2 -CHO (0.964 g, 2.73 mmol) and 0.709 g ( 1.35 mmol) of 2,5-didodecyloxy-1,4-phenylene-diacetonitrile were dissolved and the solution was heated until all solids were dissolved (~50C). Then 0.65 g (5.8 mmol) of t-butoxide was added and the reaction was heated to 70C for 3 h. The reaction was then cooled and poured into water and extracted with ether. The organic layer was then dried with MgSO 4 . After chromatography on silica gel (1:1 heaxanes and dichloromethane) the product was dissolved in dichloromethane and precipitated by dripping into cold methanol. Filtration gave the product as a yellow-brown solid (0.138 g, 9%). mp 106-107C. 1 H NMR (300 MHz, CDCl 3 ) 8.25 (s, 2H), 7.09 (s, 2H), 6.69 (s, 2H), 4.04 (t, 4H), 3.96 (s, 4H), 3.89 (s, 4H), 1.85 (m, 4H), 1.4-1.2 (br, 76H), 0.89 (m,
186 18H). 13 C NMR (75 MHz, CDCl3) 1. HRMS calcd for C 74 H 116 N 2 O 6 S 2 (M + ), 1192.8275; found 1192.8247. Anal. calcd for C 74 H 116 N 2 O 6 S 2 : C, 74.45; H, 9.79; N, 2.35; O, 8.04; S, 5.37. Found: C, 72.95; H, 10.41; N, 2.27. 2,3-bis(3-octylthiophen-2-yl)acrylonitrile (28). In 100 mL of ethanol, 2.34 g (10.0 mmol) of 3-octylthiophene-2-carbonitrile and 2.28 g (10.2 mmol) of 3-octylthiophene-2-carbaldehyde were dissolved. Then 1.45 g (13 mmol) of tBuOK was added to the reaction and it was heated to reflux for 10 h. The reaction mixture was then poured into 200 mL of water and extracted with dichloromethane. The organic layer was washed with MgSO 4 and then the solvent was removed under vacuum. Chromatography on silica (1:1 hexanes and dichloromethane) gave 2.95 g (67%) of the product as a viscous yellow oil. 1 H NMR and 13 C NMR reveal a complex, overlapping mixture of isomers and the absence of starting material. HRMS calcd for C 27 H 39 NS 2 (M + ), 441.2524; found 441.2542. Anal. calcd for C 27 H 39 NS 2 : C, 73.41; H, 8.90; N, 3.17; S, 14.52. Found: C, 73.91; H, 9.50; N, 3.45. 2,3-bis(3,3-dihexyl-3,4-dihydro-2H-thieno[3,4-b][1,4]dioxepin-6-yl)acrylonitrile (BProDOT-Hx 2 -CNV) (29). In 100 mL of absolute ethanol, ProDOT-Hx 2 -ACN (3.45 g, 9.50 mmol) and ProDOT-Hx 2 -CHO (3.37 g, 9.57 mmol) were dissolved and 1.41 g (12.6 mmol) of potassium t-butoxide ws added. The solution immediately darkened on addition of the base and was then heated to reflux for 3 h. The reaction mixture was then cooled to room temperature and poured into water to give a red
187 precipitate, which was isolated by filtration. The precipitate was redissolved in dichloromethane, washed with water and dried with MgSO 4 . Chromatography on silica gel with hexanes and dichloromethane (1:1) gave the product (5.75 g, 87 %) as an orange solid. mp 68-74C. 1 H NMR (300 MHz, CDCl 3 ) 7.65 (s, 1H), 7.64 (s, 0.2 H), 6.63 (s, 1H), 6.59 (s, 0.2 H), 6.53 (s, 0.2 H), 6.41 (s, 1H), 3.96 (s, 4H), 3.87 (s, 4H), 1.4-1.2 (br, 40H), 0.88 (t, 12H). 13 C NMR (75 MHz, CDCl 3 ) 151.52, 150.41, 149.37, 147.31, 130.94, 117.81, 117.45, 117.23, 109.01, 103.73, 98.05, 78.28, 78.09, 78.01, 44.06, 43.95, 32.31, 32.13, 31.95, 31.93, 30.36, 30.32, 23.02, 22.85, 14.28.HRMS calcd for C 41 H 63 NO 4 S 2 (M + ), 697.4198; found 697.4196. Anal. calcd for C 41 H 63 NO 4 S 2 : C, 70.54; H, 9.10; N, 2.01; O, 9.17; S, 9.19. Found: C, 70.57; H, 9.43; N, 1.85; S, 9.16. 2,3-bis(5-bromo-3-octylthiophen-2-yl)acrylonitrile (30). In 125 mL of dry DMF, 1.533 g (3.48 mmol) of 2,3-bis(3-octylthiophen-2-yl)acrylonitrile was dissolved and the solution was cooled to 0C. Then 1.255 g (7.05 mmol, 2.03 equiv) of NBS was added and the solution was allowed to warm slowly to room temperature and then stirred overnight. The reaction mixture was then poured into 300 mL of brine and extracted with ether. The ether layer was then washed with water and dried with MgSO 4 . Chromatography on silica (25:1 petroleum ether and ether and 50:1 hexanes and ether) was inadequate to separate the components of the reaction mixture. 1 H NMR and 13 C NMR reveal a complex, overlapping mixture of isomers. HRMS calcd for C 27 H 37 Br 2 NS 2 (M + 79 Br 81 Br) 599.0714; found 599.0710. HRMS calcd for C 27 H 37 Br 3 NS 2 (M + 79 Br 81 Br 81 Br) 678.9798; found 678.9872. HRMS calcd for C 27 H 37 BrNS 2 (M + 79 Br) 519.1629; found 519.1804. Results of HRMS reveal that the complex mixture of isomers
188 in the starting material led to a complex mixture of isomeric mono, di, and tribromo products based on the differing reactivities of the isomeric starting compounds. 2,3-bis(8-bromo-3,3-dihexyl-3,4-dihydro-2H-thieno[3,4-b][1,4]dioxepin-6-yl)acrylonitrile (BProDOT-Hx 2 -CNV-Br 2 ) (31). In 125 mL DMF BProDOT-Hx 2 -CNV (2.294 g, 3.29 mmol) was dissolved and the solution was cooled to 0C. Then 1.23 g of freshly recrystallized NBS was added in one portion and the solution was then allowed to warm gradually to room temperature. After two hours the solution had changed from bright red to a dark black. At this point the reaction mixture was poured into 300 mL of brine and extracted with ether. The ether layer was washed with brine and then dried with MgSO 4 . Chromatography on silica eluting with 18:1 petroleum ether and ether gave the product as a dark red, highly viscous oil (2.11 g, 75 %). 1 H NMR (300 MHz, CDCl 3 ) 7.47 (s, 1H), 3.96-3.94 (m, 8H), 1.4-1.2 (br, 40H), 0.88 (t, 12H). 13 C NMR (75 MHz, CDCl 3 ) 150.34, 148.21, 147.20, 146.50, 146.22, 132.92, 129.60, 129.09, 117.04, 116.84, 116.43, 99.75, 92.86, 85.61, 78.46, 78.39, 78.36, 78.23, 43.95, 43.81, 31.70, 30.09, 30.04, 22.75, 22.64, 14.07. HRMS calcd for C 41 H 61 Br 2 NO 4 S 2 (M + ), 853.2406; found 853.2440. Anal. calcd for C 41 H 61 Br 2 NO 4 S 2 : C, 57.54; H, 7.18; Br, 18.67; N, 1.64; O, 7.48; S, 7.49. Found: C, 57.69; H, 7.22; N, 1.55. PProDOT:CN-PPV (R = butyl). In a solution consisting of 7 mL THF and 21 mL t-butanol, 0.886 g (2.73 mmol) of ProDOT-Bu 2 -(CHO) 2 and 0.425 g (2.72 mmol) 1,4-phenylenediacetonitrile were dissolved at 50C. Then 0.06 g (0.5 mmol) potassium t-butoxide was added along with 0.4 mL of 1.0 M tetrabutylammonium hydroxide (0.4 mmol) and the solution was heated to 50C for 20 minutes. The reaction mixture was then cooled to room temperature and poured into 500 mL of ice-cold methanol acidified
189 with 1 mL of acetic acid. The black precipitate was then isolated by filtration and dried under vacuum. The solid (1.114 g, 92%) was found to be insoluble in all investigated solvents. IR (KBr cm -1 ): 3034, 2956, 2931, 2862, 2207, 1656, 1582, 1510, 1480, 1450, 1376, 1254, 1208, 1080, 1048, 1021, 990, 902, 832, 728, 672, 624, 582, 496. PProDOT:CN-PPV (R = ethylhexyl). In a solution consisting of 25 mL THF and 75 mL t-butanol, 1.000 g (2.30 mmol) of ProDOT-EtHx 2 -(CHO) 2 and 0.354 g (2.30 mmol) 1,4-phenylenediacetonitrile were dissolved and the solution was cooled to 0C. Then 0.172 g (1.54 mmol) potassium t-butoxide was added along with 0.9 mL of 1.0 M tetrabutylammonium hydroxide (0.9 mmol) and the solution was heated to 50C for 20 minutes. The reaction mixture was then cooled to room temperature and poured into 500 mL of ice-cold methanol acidified with 1 mL of acetic acid. The black precipitate was then isolated by filtration and dried under vacuum. The solid (0.970 g, 76%) was found to be insoluble in all investigated solvents. IR (KBr cm -1 ): 3034, 2958, 2927, 2858, 2208, 1583, 1510, 1480, 1455, 1379, 1298, 1252, 1052, 1022, 990, 904, 832, 769, 728, 655, 624, 582, 497. PProDOT-Bu 2 :CN-PPV. In a solution consisting of 9 mL THF and 9 mL t-butanol, 0.068 g (0.21 mmol) ProDOT-Bu 2 -(CHO) 2 and 0.110 g (0.21 mmol) 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene were dissolved at 40C. Then 0.063 g (0.56 mmol) potassium t-butoxide was added and the solution was heated to 70C for 2 h. The reaction mixture was then cooled to room temperature and poured into 500 mL of ice-cold methanol acidified with 1 mL of acetic acid. The precipitate was then isolated by filtration and dried under vacuum. The solid was then reprecipitated from chloroform into methanol to give 0.058 g (36%) of the polymer as a black solid. 1 H NMR (300 MHz
190 CDCl 3 ) 8.75 (b, 2H), 6.90 (b, 2H), 4.20 (b, 4H), 3.97 (b, 4H), 2.02 (b, 4H), 1.5-1.0 (bm, 48H), 0.88 (bm, 12H). IR (KBr cm -1 ): 2926, 2854, 2204, 1552, 1510, 1478, 1381, 1262, 1081, 802. GPC analysis (THF vs. PS): M n = 20,000 g/mol; M w = 31,100 g/mol; PDI = 1.56. PProDOT-Hx 2 :CN-PPV (Batch 1). In a solution consisting of 36 mL THF and 36 mL t-butanol, 0.326 g (0.86 mmol) ProDOT-Hx 2 -(CHO) 2 and 0.450 g (0.86 mmol) 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene were dissolved at 40C. Then 0.26 g (2.3 mmol) potassium t-butoxide was added and the solution was heated to 70C for 2 h. The reaction mixture was then cooled to room temperature and poured into 600 mL of ice-cold methanol acidified with 1 mL of acetic acid. The precipitate was then isolated by filtration and dried under vacuum. The solid was then reprecipitated from chloroform into methanol to give 0.660 g (89%) of the polymer as a black solid. 1 H NMR (300 MHz, CDCl 3 ) 8.75 (b, 2H), 6.93(b, 2H), 4.21 (b, 4H), 3.95 (bm, 4H), 1.98 (b, 4H), 1.5-1.2 (bm, 56H), 0.88 (bm, 12H). IR (neat cm -1 ): 2924, 2854, 2206, 1571, 1513, 1479, 1450, 1384, 1285, 1261, 1225, 1058, 858, 802, 722. Anal. calcd: C, 75.81, H, 9.95, N, 3.21, S, 3.68. Found: C, 74.72, H, 10.00, N, 2.97, S, 3.84. GPC Analysis (THF vs. PS): M n = 13,500 g/mol; M w = 20,700 g/mol; PDI = 1.53. PProDOT-Hx 2 :CN-PPV (Batch 2). In a solution consisting of 37 mL THF and 37 mL t-butanol, 0.339 g (0.89 mmol) ProDOT-Hx 2 -(CHO) 2 and 0.467 g (0.89 mmol) 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene were dissolved at room temperature. Then 0.20 g (1.8 mmol) potassium t-butoxide was added and the solution was heated to 70C for 2 h. The reaction mixture was then cooled to room temperature and poured into 600 mL of ice-cold methanol acidified with 1 mL of acetic acid. The precipitate was then
191 isolated by filtration and dried under vacuum. The solid was then reprecipitated from chloroform into methanol to give 0.472 g (67%) of the polymer as a black solid. 1 H NMR (300 MHz, CDCl 3 ) 8.75 (b, 2H), 6.92 (b, 2H), 4.22 (b, 2H), 3.96 (bm, 4H), 1.99 (b, 4H), 1.5-1.2 (bm, 56H), 0.88 (bm, 12H).IR (neat cm -1 ): 2925, 2854, 2205, 1572, 1512, 1479, 1450, 1384, 1338, 1285, 1218, 1058, 917, 858, 814. Anal. calcd: C, 75.81, H, 9.95, N, 3.21, S, 3.68. Found: C, 75.49, H, 10.51, N, 2.96, S, 3.17. GPC Analysis (THF vs. PS): M n = 17,400 g/mol; M w = 26,300 g/mol; PDI = 1.51. Th-CN-PPV. In a solution consisting of 40 mL THF and 40 mL t-butanol, 0.291 g (0.945 mmol) 3-dodecylthiophene-2,5-dicarbaldehyde and 0.495 g (0.945 mmol) 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene were dissolved at room temperature. Then 0.28 g (2.5 mmol) potassium t-butoxide was added and the solution was heated to 70C for 2 h. The reaction mixture was then cooled to room temperature and poured into 500 mL of ice-cold methanol acidified with 1 mL of acetic acid. The precipitate was then isolated by filtration and dried under vacuum. The solid was then Soxhlet extracted with methanol for 24 h followed by hexanes for 24 h. The polymer was then isolated by Soxhlet extraction with chloroform to give 0.342 g (43%) of a deep red-purple solid. 1 H NMR (300 MHz, CDCl 3 ) 8.29 (b, 2H), 7.89 (b, 1H), 7.13 (b, 2H), 4.07 (bm, 4H), 2.77 (b, 2H), 1.9-1.2 (bm, 60H), 0.87 (bm, 9H). IR (neat cm -1 ): 2953, 2921, 2851, 2209, 1725, 1502, 1455, 1376, 1261, 1218, 1095, 1025, 870, 801. Anal. calcd: C, 78.14; H, 10.34; N, 3.50; S, 4.01. Found: C, 77.43; H, 9.07; N, 3.13; S, 4.31.GPC Analysis (THF vs. PS): M n = 25,500 g/mol; M w = 102,500 g/mol; PDI = 4.02. CN-TV. In a solution consisting of 70 mL THF and 70 mL t-butanol, 0.955 g (3.10 mmol) 3-dodecylthiophene-2,5-dicarbaldehyde and 0.502 g (3.10 mmol) 2,5
192 thiophenediacetonitrile were dissolved at room temperature. Then 0.92 g (8.2 mmol) potassium t-butoxide was added and the solution was heated to 70C for 2 h. The reaction mixture was then cooled to room temperature and poured into 1 L of ice-cold methanol acidified with 10 mL of acetic acid. The precipitate was then isolated by filtration and dried under vacuum. The solid was then taken up in 1,1,2,2-tetrachloroethane by Soxhlet extraction. After removal of the solvent, the polymer was again Soxhlet extracted with methanol for 24 h and then taken up in chloroform to give 0.098 g (7%) of the polymer as a dark purple solid. IR (neat cm -1 ): 2923, 2852, 2212, 1655, 1575, 1458, 1261, 1093, 1021, 876, 802. Anal. calcd: C, 71.51, H, 7.39, N, 6.42, S, 14.69. Found: C, 66.16, H, 7.09, N, 4.32, S, 11.63. GPC Analysis (THF vs. PS): M n = 3,100 g/mol; M w = 4,500 g/mol; PDI = 1.46. CN-PPV. In a solution consisting of 17 mL THF and 17 mL t-butanol, 0.200 g (0.40 mmol) 2,5-bis(dodecyloxy)terephthalaldehyde and 0.208 g (0.40 mmol) 1,4-bis(cyanomethyl)-2,5-bis(dodecyloxy)benzene were dissolved at room temperature. Then 0.090 g (0.80 mmol) potassium t-butoxide was added and the solution was heated to 70C for 2 h. The reaction mixture was then cooled to room temperature and poured into 600 mL of ice-cold methanol acidified with 1 mL of acetic acid. The precipitate was then isolated by filtration and dried under vacuum. The solid was then reprecipitated from chloroform into methanol to give 0.351 g (92%) of the polymer as a dark red solid. 1 H NMR (300 MHz, CDCl 3 ) b, 2H), 7.95 (b, 2H), 7.12 (b, 2H), 4.11 (b, 8H), 1.86 (b, 8H), 1.5-1.2 (bm, 72 H), 0.87 (b, 12H) IR (neat cm -1 ): 2955, 2924, 2853, 2211, 1725, 1604, 1505, 1466, 1422, 1377, 1261, 1217, 1095, 1022, 919, 866, 800, 710, 668. Anal.
193 calcd: C, 79.78; H, 10.96; N, 2.82. Found: 79.29; H, 11.75; N, 2.55. GPC Analysis (THF vs. PS): M n = 13,700 g/mol; M w = 29,900 g/mol; PDI = 2.17. PBProDOT-Hx 2 -CNV. In 100 mL of anhydrous DMF, 2.07 g of BProDOT-Hx 2 -CNV-Br 2 was dissolved and the solution was bubbled with argon for 30 minutes and heated to 60C. Then a solution of 0.803 g (2.9 mmol) of Ni(COD) 2 and 0.463 g (3.0 mmol) of 2,2'-bipyridine and 0.30 mL (2.4 mmol) of freshly degassed cyclooctadiene in 25 mL of DMF was stirred at 60C for 30 minutes and added to the monomer solution dropwise via cannula. The reaction mixture was then stirred at 60C for 24 h, cooled to room temperature and poured into 1 L of ice-cold methanol. The resulting precipitate was isolated by filtration into a Soxhlet thimble and then extracted with methanol for 24 h followed by hexane for 24 h. The polymer was then taken up in dichloromethane and repricipitated into methanol and washed successively with 200 mL of hot 0.01 M EDTA solution (pH = 3-4), 200 mL of hot 0.01 M EDTA solution (pH = 8-9), and 200 mL of water. The polymer was then dried under vacuum to give 0.487 g (29%) of a dark blue solid. 1 H NMR (300 MHz, CDCl 3 ) 7.77-7.71 (m, 1H), 3.99 (m, 8H), 1.5-1.2 (m, 40H), 0.89 (bs, 12H). IR (neat cm -1 ): 2961, 2926, 2868, 2206, 1725, 1547, 1454, 1375, 1261, 1094, 1020, 866, 800, 709, 662. Anal. calcd (assuming that all bromines have reacted): C, 70.54; H, 9.10; N, 2.01; S, 9.19. Found: C, 42.41; H, 7.75; N, 0.40; S, 2.81; Br, 0.37; Ni, 0.02. GPC Analysis: M n = 14,300 g/mol; M w = 40,000 g/mol; PDI = 2.80. (1-(3-methoxycarbonyl)-propyl-1-phenyl-[6,6]-C 61 (PCBM). 78 In 60 mL of freshly distilled pyridine, 3.00 g of methyl 4-benzoylbutyrate p-tosylhydrazone (prepared according to the literature) and 470 mg of sodium methoxide were combined and stirred for 15 minutes at room temperature. Then 2.88 g of C 60 was added in 200 mL of 1,2
194 dichlorobenzene and the solution was heated to 65-70 C. The dark purple solution was stirred for 24 hours under argon. At this time the reaction mixture was cooled to room temperature and solvent was subsequently distilled off under argon to reduce the reaction mixture to a volume of ~50 mL. The crude compound (the [5,6]-fulleroid) was then purified by column chromatography on silica using 100 mL of chlorobenzne as a pre-eluent, followed by dichlorobenzene to collect the unreacted C 60 , and finally with toluene. The solvents were then removed to give 1.74 g of the crude product, which was subsequently dissolved in 330 mL of dichlorobenzene and heated to reflux overnight. The dichlorobenzene was then removed by distillation to concentrate the crude product (PCBM) to ~20 mL. The crude slurry was then added to a centrifuge tube along with an equal volume of methanol. Three centrifugation-decantation steps with methanol added each time were perfomed until the methanol was completely clear. Chromatography on silica with toluene gave the product (1.627 g, 45%) as a black-brown solid. 1 H NMR (300 MHz, CDCl 3 ) 7.93 (d, 2H), 7.55 (m, 3H), 3.68 (s, 3H), 2.92 (m, 2H), 2.53 (m, 2H), 2.19 (m, 2H). HRMS (FAB, NBA) calcd for C 72 H 14 O 2 (M + ), 911.1071; found 911.1080. UV-Vis (hexane) max (nm): 208, 257, 327, 430, 494, 695.
CHAPTER 4 DONOR FUNCTIONALIZED FULLERENES FOR BULK HETEROJUNCTION SOLAR CELLS Bulk heterojunction solar cells based on a donor polymer and the soluble fullerene acceptor PCBM, represent the forefront of the research effort in the development of conjugated polymer solar cells, as described in chapter 1. The excellent properties of PCBM can be attributed to the high electron affinity of C 60 in combination with a solubilizing group, which allows thorough mixing and efficient blending with soluble conjugated polymers. While PCBM has been the most successful soluble fullerene derivative, numerous other derivatives exist and the library of compounds is growing on a daily basis. One interesting class of functionalized fullerenes contains those derivatives that can be described as donor-acceptor dyads. In this case an electron rich donor moiety is tethered to the C 60 core, not only enhancing the solubility of the fullerene, but also inducing interesting and relevant electronic consequences for photovoltaic applications. In this chapter, two novel donor-functionalized fullerenes will be examined in the context of application to solar cells following a brief discussion of the concepts and literature surrounding such donor-acceptor dyads. 4.1 Soluble Donor-Functionalized Fullerenes Since the discovery and isolation of C 60 in 1985, 204 the fullerenes have been the subject of intense research. One area that is of special interest for materials applications is the chemical functionalization of fullerenes for the purpose of inducing solubility as well as controlling the electronic properties. A picture of the chemical reactivity of C 60 was 195
196 developed at an early stage, 205 and numerous derivatives have since been synthesized 206 , 207 using a variety of synthetic approaches based on addition and cycloaddition chemistry. 208 In fact, addition and cycloaddition reactions are the only methods available for the functionalization of C 60 . The reactivity of fullerene is best described as that of an electrophile (analogous to an electron deficient olefin) and as such, it is susceptible to reaction with various nucleophiles. 209 Figure 4-1a illustrates the resonance structure of C 60 that most accurately explains its reactivity. In total, C 60 is made up of 20 six-membered rings and 12 five-membered rings. The [6,6] bond is identified as the junction between two six-membered rings and the [5,6] bond is the junction between a five and a six-membered ring. The electronic structure is such that the [6,6] bond has much more double bond character than the [5,6] bond, which is essentially a single bond in the minimum energy structure. 210 [6,6]-bond [5,6]-bond [5,6]-open[6,6]-open [6,6]-closed[5,6]-closed R R R R R R R R [5,6]-doublebonds [5,6]-doublebondsaC60 b Figure 4-1. Structure of fullerne-C 60 (a) and isomeric products upon cycloadditon (b).
197 Several different types of reactions are commonly used in the synthesis of substituted fullerene derivatives and Figure 4-1b illustrates the structures of the four possible isomers that are formed upon cycloaddition for the case of cyclopropanation. The products are named according to which bond has undergone addition, either [5,6] or [6,6] and whether the result of the addition has been to remove a bond within the C 60 framework (open) or leave a single bond at the point of addition (closed). Here the [6,6]-open product is not a reasonable product as it shows a highly strained structure with three [5,6] double bonds as well as bridgehead double bonds that place it in violation of Bredtâ€™s rule. In a similar manner, the [5,6]-closed isomer is a highly strained structure that requires the formation of two [5,6] double bonds, which are unfavorable to the overall stability of fullerene. The two fullerene isomers that are commonly isolated in the functionalization of C 60 are the [5,6]-open and the [6,6]-closed, which is the most stable of all the isomers. Notice that even though the [5,6]-open isomer is in violation of Bredtâ€™s rule, all the [5,6] bonds are single bonds and thus the compound is only 6 kcal/mol less stable than the [6,6]-closed isomer. The [6,6]-closed isomer is often referred to as [6,6]-methanofullerene. The three most commonly used addition reactions for fullerenes are dipolar cycloadditions, addition-elimination reactions, and the addition of carbenes. 208,210 Figure 4-2 shows the overall transformation of all three reaction types. Notice that with carbene (Figure 4-2a) and addition-elimination reactions (Figure 4-2b), the [6,6]-closed isomer is the expected product after only a single step. In Figure 4-2c, for the case of a dipolar cycloadditon reaction (the synthesis of PCBM is an example), after the formation of the initial five-membered pyrazoline addition adduct, a molecule of nitrogen is then
198 eliminated and the primary product is the [5,6]-open product also known as the [5,6]-fulleroid. The [5,6]-fulleroid is then isomerized to the lower energy [6,6]-closed isomer, better known as the [6,6]-methanofullerene. R R' R R' [6,6]-closed R' R X R R' X AdditionElimination R R' [6,6]-closed R R' N2 [6,6]-closed R N R' N -N2 [5,6]-open R R R R' [6,6]-closed abc Figure 4-2. Overview of general synthetic methods employed in the synthesis of substituted fullerene derivatives. (a) Addition of a carbene to the [6,6]-bond directly yields the [6,6]-closed isomer. (b) Addition followed by intramolecular elimination directly yields the [6,6]-closed isomer. (c) The [1,3]-dipolar cylcoadddition of a diazo compound yields the unstable pyrazoline intermediate which undergoes thermal elimination of nitrogen with a concomitant isomerization to the [5,6]-open isomer. Thermal isomerization yields the more stable [6,6]-closed isomer.
199 While [5,6]-fulleroids are isoelectronic with C 60 (possessing 60 electrons), [6,6]-methanofullerene derivatives are 58 electron systems, as a double bond has been replaced with a single bond in the C 60 framework. Nonetheless, for both the methanofullerene, PCBM and the corresponding [5,6]-fulleroid, the reduction potentials are shifted cathodic (making them more difficult to reduce) of C 60 itself, indicating that changing one bond in the parent C 60 is sufficient to perturb the electronic structure. 78 For the 58electron system of methanofullerenes, reduction potentials are generally shifted 100-150 mV cathodic of C 60 . 211 Specifically, the first reduction of PCBM is observed at .169 V vs. Fc/Fc + and the first reduction of the [5,6]-fulleroid analogue is observed at .135 V vs. Fc/Fc + , indicating little difference in the electron affinity of the two. For comparison, C 60 shows a first reduction at .056 V vs. Fc/Fc + , when measured under the same conditions as for PCBM. Numerous types of singly and multiply functionalized C 60 derivatives have been synthesized through the above-mentioned routes. While attachment of the side chain to the fullerene has an effect on the solubility of the molecule, 212 the nature of the pendant group can also have an affect on the electronic properties of the molecule. 213 For example, strong electron withdrawing groups (e.g. nitriles) on the cyclopropane ring in methanofullerene are observed to shift reduction potentials as much as 300 mV anodic of the diphenyl substituted analogue. An especially interesting class of functionalized C 60 derivatives are those which can be referred to as donor-acceptor dyads (DA-dyads). In this case, an electron donor is tethered to the strongly electron accepting fullerene. From a fundamental point of view, this type of molecule is interesting for the study of intramolecular charge and energy
200 transfer. 207 ,214 Such systems are also promising for numerous applications such as optical limiting and photovoltaics. 215 Potentially, the most attractive role for a fullerene donor-acceptor dyad in a photovoltaic device is as the sole photoactive material. The report of the application of a trimeric oligophenylene-vinylene (OPV) functionalized C 60 for use in solar cells in 1999, was the first report of a bulk heterojunction PVD in which the photoactive layer was a single component. 216 The strength of this approach is that it offers a means of overcoming the problems of macrophase separation encountered in classical bulk heterojunction devices, by guaranteeing a large donor-acceptor interfacial area through the chemical linkage of donor and acceptor. In bulk heterojunction devices, macrophase separation can lead to large domains of donor and acceptor, which not only inhibit charge generation (due to the limited exciton diffusion length), but also preclude effective charge transport due to the trapping of charges within the large domains. In this first report of a DA-dyad device, the monochromatic power conversion efficiency (400 nm, 5 mW/cm 2 ) was found to be only ~1%. Further work revealed that energy transfer was far more favorable in this system than charge transfer, which was the ultimate cause of the low power conversion efficiencies. Tetrameric OPV-C 60 dyads were subsequently found to give long-lived charge-separated states (on the order of 1 ms) in thin films, although photovoltaic device performance was modest (no efficiency was reported), with very low observed fill factors (0.25). 217 Charge transfer is favored with a longer oligomeric donor, due to the enhanced donating ability of the more extended conjugated system. The improved performance can also be credited to the stronger absorption in the visible range by the longer oligomer. Based on this precedent, a variety of other OPV, oligophenyleneethynylene (OPE),
201 oligonapthylenevinylene, and oligothiophene C 60 -DA-dyads have been tested for use in photovoltaic devices. 218 The three fulleropyrrolidines in Figure 4-3, have shown among the highest power conversion efficiencies reported in the literature for DA-dyad devices. Compound (i) 219 showed an AM1.5 (80 mW/cm 2 ) power conversion efficiency of 0.2% and a peak IPCE value of ~10% at 440 nm, which correlates well with the absorption maxima of the donor oligomer. Compound (ii) 220 was not evaluated under AM1.5 conditions, but is worth mentioning as it was found to show an IPCE value of ~10% at 456 nm ( max of the donor oligomer). Compound (iii) 221 showed a white light (80 mW/cm 2 , unidentified illumination source) power conversion efficiency of 0.37%, which is the highest reported efficiency for a DA-dyad device. All three of these compounds demonstrated efficient intramolecular photoinduced charge separation. N S OC6H13 O H13C6 S Br N S S S S H C6H13 H13C6 n N O 3 S Cl CN N N HN N O O (i)(ii)(iii) Figure 4-3. Structures of DA-dyads that yield efficient single-component PVDs. For compound (ii) n=4. Improving the efficiency of this type of device will certainly require an increase in the absorption of light across the visible spectrum as well as enhancements in the
202 charge carrier transport properties. By using longer oligomeric side chains in these compounds the hole mobility is observed to increase, however with more oligomer, the C 60 / oligomer mass ratio (or more generally the C 60 / donor mass ratio) is reduced, causing a decrease in electron mobility. 222 It has been observed that as the C 60 / oligomer mass ratio deviates from the 4:1 value that has been found to be optimal in most polymer-PCBM bulk heterojunctions, the electron mobility of the device decreases. In most DA-dyad materials the mass ratio is less than 1:1, indicating that the amount of acceptor in these materials is possibly inadequate for optimal operation of the devices. Optimization of this class of device will require new systems that allow maximization of hole and electron transport through improving the C 60 / oligomer mass ratio while absorbing the maximum amount of visible light. While such C 60 -DA-dyads offer the possibility of single component photoactive layers in PVDs, there are several other roles that these compounds can fulfill in a bulk heterojunction device. In particular, the use of blends consisting of a conjugated polymer, a DA-dyad, and PCBM offer an attractive means of increasing the spectral range over which light is absorbed by either donor component alone, as well as improving the C 60 / donor mass ration. Figure 4-4 illustrates one compound that has been used in the fabrication of multicomponent photovoltaic devices. 88a Here the donor portion of the DA-dyad is a zinc-phthalocyanine (Zn-Pc). Phthalocyanines are promising donors for fullerene acceptors, and long-lived charge-separated states have been observed in these dyads. 223 In this case, the Zn-Pc moiety can be viewed as an antenna, which not only absorbs light, but also collects light energy via energy transfer from MDMO-PPV, followed by electron transfer to C 60 . As such, the spectral range of the photocurrent in the
203 MDMO-PPV / PCBM device is enhanced via the addition of the Zn-Pc, which absorbs strongly at wavelengths near 700 nm. Covalent attachment of the Zn-Pc moiety to the C 60 ensures close contact between the donor and the acceptor, which guarantees that upon exciation of Zn-Pc (either by light absorption or energy transfer from MDMO-PPV) charge transfer can take place. In the three-component blend device, MDMO-PPV, PCBM, and Zn-Pc-C 60 are blended together in a 25%: 52.5%: 22.5% ratio by weight. This blend composition gives 50% donor and 50% acceptor by weight, based on the fact that Zn-Pc-C 60 is 50% donor and 50% acceptor by weight. Here, MDMO-PPV strongly absorbs photons between 350 and 600 nm, while Zn-Pc absorbs at 350 nm and 700 nm. Thus, in combination, the two effectively blanket the visible portion of the solar spectrum. Ultimately though, the blends consisting of MDMO-PPV and Zn-Pc-C 60 as well as the MDMO-PPV / PCBM / Zn-Pc-C 60 three component blends are convincingly outperformed by the MDMO-PPV:PCBM device. While AM1.5 results were not reported, the MDMO-PPV / PCBM (1/4) device showed a peak IPCE value of ~55% at 475 nm, while the MDMO-PPV / Zn-Pc-C 60 (1/1) device showed a peak IPCE value of only 2.5% at 350 nm. The tricomponent device showed two IPCE maxima at 490 and 690 nm of ~10% each. The authors attribute the decrease in device performance in the tricomponent blend relative to the MDMO-PPV / PCBM device to an unfavorable and unoptimized morphology in the films. However, the cause of the low efficiency may be more directly related to the C 60 / donor ratio, which is 1/1. Perhaps increasing the amount of PCBM in these blends will lead to enhanced efficiencies. Nonetheless, IPCE measurements indicate that both MDMO-PPV and Zn-Pc are operating as photoexcited donors in the devices.
204 N N N N N N N N N Zn-Pc-C60 Zn Figure 4-4. Structure of Zn-Pc-C 60 DA-dyad used in three-component devices with MDMO-PPV and PCBM. Based on this brief discussion it is evident that C 60 -DA-dyads are interesting for application to PVDs. Such systems can be used to fill a variety of roles in PVDs and can serve to control phase separation or act as an antenna in a polymer-based multicomponent blend. Here we investigate two such novel systems and their operation in photovoltaic devices. 4.2 Triphenylamine Functionalized Fullerenes As a first type of C 60 -DA-dyad, we examined a triphenylamine (TPA) functionalized compound, which will be referred to as TPA-C 60 (Figure 4-5). This compound was synthesized in the laboratory of Professor Charles W. Spangler (Montana State University) using methodology developed for the synthesis of similar compounds. 224 Here, the attached donor chromophore is diphenylamino-stilbene (Ph 2 AS), which is a compound that has been previously reported. 225 Compound 1 was obtained by reaction of the benzyl alcohol derivative of Ph 2 AS with malonyl dichloride. The DA-dyad was then synthesized via the addition elimination route described in Figure 4-2b. In this case, the combination of CBr 4 and the base DBU (1,8-diazabicyclo-[5.4.0]
205 undecane) is known to generate the -bromo anion of the malonate ester, which is effective for a one-step addition elimination reaction with C 60 . 226 O O O O N N C60DBU, CBr4toluene O O O O N N 1TPA-C60 Figure 4-5. Synthesis of TPA-C 60 . A relatively limited number of TPA-functionalized fullerenes have been reported in the literature and while photoinduced charge transfer has been shown to give rise to charge-separated states, the lifetime is strongly dependent on the structure of the dyad. 227 In this case, covalent functionalization of fulleropyrrolidines with a single TPA moiety gave fast charge recombination, while the use of rotaxane functionalized fullerenes with a noncovalently attached TPA gave long lived-charge separated states. No measurements were reported in the solid state, but this limited precedent shows that charge separation is possible in fullerene-TPA dyads and that the nature of the charge transfer process is strongly dependent on the structure of the dyad. For applications in PVDs, fullerene donor-acceptor dyads with TPA donors are promising candidates based on the well-known properties of TPA compounds. Compounds based on TPA are perhaps best known as hole transport materials for use in
206 LEDs. 228 However, derivatives of TPA are also known to function as electron donors in photovoltaic devices. 229 Importantly evidence exists for charge transfer from TPA to C 60 . 230 Concerning the compound TPA-C 60 , Figure 4-6 shows the absorption and photoluminescence measured in dichloromethane solution. Two peaks are observed in the absorption spectra at 257 nm and 371 nm. The peak at 257 nm corresponds to the methanofullerene, while the peak at 371 corresponds to the absorbance of the Ph 2 AS, which has been reported to absorb with a maximum of 362 nm in acetonitrile solution. 225a The DA-dyad was found to emit in dichloromethane solution with an emission maximum of 450 nm. This matches well with the observed emission of Ph 2 AS, which shows a maximum at 463 nm in acetonitrile solution. 225a More importantly, the wavelength of maximum emission was not found to vary with excitation wavelength, and excitation at 257, 300, and 370 nm showed the same emission, although the greatest intensity was observed upon excitation at 370 nm. Methanofullerenes are often observed to emit very weakly at ~700 nm in solution, 231 but in this case no emission was observed at 700 nm upon excitation at any of the investigated wavelengths. The photophysical properties of TPA-C 60 were not measured in thin film form due to the extremely poor film forming properties of the compound. Even at spin coating speeds as low as 500 rpm, all the material is ejected, with no coating of the substrate. Based on this limited solution study it can be observed that the photoluminescence of Ph 2 AS is not quenched in by the covalently attached fullerene. This indicates that charge transfer is not likely to occur in solution as a result of the photoexcitation of Ph 2 AS. Even if charge transfer did occur and recombination was fast, no emission of the Ph 2 AS should be observed. While thin film
207 measurements could not be performed, it is unlikely that effective charge transfer would occur in the films, based on the results observed in solution. 3004005006000.00.51.01.52.02.53.03.5 Absorbance (a. u.)Wavelength (nm)0.00.40.81.21.62.02.4 PL Intensity (a. u.) Figure 4-6. Absorbance and photoluminescence of TPA-C 60 in dichloromethane. The concentration was 0.1 mg/mL. The sample was excited at 370 nm. The electrochemical properties of TPA-C 60 were also investigated in dichloromethane solution with TBAP as the supporting electrolyte. Figure 4-7 shows the cyclic voltammetry (CV) obtained in an argon-filled glovebox and compared with the results for PCBM measured under the same conditions. Here the first two reduction waves of the methanofullerene TPA-C 60 are observed with E 1/2 values of .9 V and .3 V vs. Fc/Fc + respectively. For comparison, the first reduction of PCBM is observed to have an E 1/2 of .0 V vs. Fc/Fc + . In a different electrolyte (TBAPF 6 / dichlorobenzene), a value of .17 V vs. Fc/Fc + has been reported for the first reduction of PCBM. Nonetheless, the interesting point is the difference in the first reduction potential of TPA-C 60 relative to PCBM. It can be seen that TPA-C 60 is slightly easier to reduce than PCBM (by ~100 mV). Although small, this shift in reduction potential can be attributed to the presence of two electron withdrawing ester groups in TPA-C 60 relative to the
208 electron releasing phenyl and alkyl substituents of PCBM. It has been reported previously that increasing the electron-withdrawing ability of the substituents on the cyclopropane ring of methanofullerenes induces an anodic shift (making it easier to reduce) in the first reduction potential. The electrochemical results presented above cannot be directly correlated with the band structure of the solid material as the data was obtained in solution. Nonetheless, this data provides a relative estimate (relative to PCBM) of the ease of reduction of TPA-C 60 , and the first reduction is the value that is important for determining how easily a photoexcited donor will reduce this compound. -2.0-1.5-1.0-0.50.00.51.01.5-0.3-0.2-0.10.00.10.20.30.4 O O O O N N TPA-C60 O O PCBMCurrent (mA/cm2)E (V) vs. Fc/Fc+ Figure 4-7. Cyclic voltammetry of TPA-C 60 in 0.1 M TBAP / dichloromethane. The first reduction of PCBM is shown for reference. The CV of TPA-C 60 is represented by the solid line and PCBM is represented by the dashed line. Figure 4-7 also shows two oxidation processes for TPA-C 60 with E 1/2 values of +0.5 V and +0.7 V. These two redox processes can be independently addressed and correspond to the oxidation of Ph 2 AS. While calculating HOMO and LUMO energies based on solution measurements is not the most preferred method, one can gain a sense of the relative energies and relative ease of oxidation and reduction. From the data in Figure
209 4-7 it can be observed that the onset of the first reduction of the methanofullerene in TPA-C 60 occurs at .85 V vs. Fc/Fc + (LUMO = 4.25 eV), while the onset of the first oxidation of Ph 2 AS occurs at +0.5 V (HOMO = 5.6 eV). Although an effective intramolecular charge transfer process in TPA-C 60 is in doubt (based on the solution PL measurements), the compound still provides an interesting candidate for incorporation into solar cells. To this end several approaches were taken. In the first approach it was desired to use TPA-C 60 as the sole photoactive material. However, TPA-C 60 does not form a film upon spin coating, as described previously. In the second approach, a comparison was made between the performance of TPA-C 60 and PCBM in devices based on 4/1 blends (w/w) of fullerene compound relative to MEH-PPV (Aldrich). Table 4-1 lists the results measured under AM1.5 conditions for this set of devices. Notice that while the V oc is similar for both devices, but the short circuit current is decreased by a factor of greater than 400 when switching from PCBM to TPA-C 60 . The efficiency is likewise drastically reduced. Perhaps most telling is the FF of 0.18 in the TPA-C 60 device, which indicates that the device is not operating as a standard diode. Table 4-1. Performance of TPA-C 60 / MEH-PPV solar cells relative to analogous PCBM / MEH-PPV devices. Device composition V oc (V) J sc (mA/cm 2 ) FF (%) PCBM : MEH-PPV (4:1) 0.70 3.4 0.32 0.75 TPA-C 60 : MEH-PPV (4:1) 0.73 0.007 0.18 0.0009 Ultimately there are two possible explanations for the poor performance of the devices containing TPA-C 60 . First, the film quality in the devices containing TPA-C 60 is
210 inferior to the analogous PCBM devices. This film quality can be related to the fact that TPA-C 60 is observed to be significantly less soluble in dichlorobenzene than PCBM and based on the cloudy nature of the TPA-C 60 solutions at device concentrations, it is suspected that the compound never fully dissolves. For an accurate comparison of photovoltaic devices, it is a necessary prerequisite that all components be fully soluble in the same solvents and under the same conditions. If this solubility condition is not met, the results have little meaning. However, no solvent was found that could dissolve TPA-C 60 at device concentrations. Solubility aside, the second possible explanation for the poor device performance is related to the previously discussed point concerning the C 60 / donor mass ratio. In devices with TPA-C 60 and MEH-PPV, the TPA oligomers and MEH-PPV serve, in principle, as electron donors and hole transport materials, while the methanofullerene portion of TPA-C 60 serves as the lone electron acceptor. In this case the mass ratio of acceptor to donor is only 2:3, based on the fact that TPA-C 60 is ~50% by weight donor. In this case, the ratio of methanofullerene to donor is probably insufficient to yield a device with a bicontinuous morphology. To look more closely at the potential of TPA-C 60 for use in PVDs, two more types of devices were constructed. In the first, TPA-C 60 was blended with PCBM and poly(vinylcarbazole) (PVK, M w = 1,100,000 g/mol, Aldrich) in a ratio of 1.6 / 3.1 / 0.3 by weight in order to engineer a device with a 20% donor and 80% methanofullerene by weight. This weight ratio is based on the fact that TPA-C 60 is 50% donor and 50% acceptor by weight. In this case, 1.6 parts of TPA-C 60 gives 0.8 parts donor and 0.8 parts acceptor. Coupled with 3.1 parts of PCBM and 0.3 parts PVK results in 1.1 parts donor and 3.9 parts acceptor in total. Here, PVK was intended to function as a hole transport
211 material and as a high molecular weight polymer to aid in the film forming properties. Additionally, PVK does not absorb at wavelengths overlapping Ph 2 AS, and thus the photocurrent contribution of Ph 2 AS could in principle be measured directly by IPCE. A further advantage of using PVK is that this polymer is known to act as a donor for PCBM in photovoltaic devices. 232 In the reported case, evidence suggests that electron transfer from ground-state PVK to photoexcited PCBM occurs as indicated by PIA measurements. Thus, while PVK does not absorb visible light, the polymer can be considered as an active donor and hole transport material. However, devices based on TPA-C 60 , PCBM, and PVK were not found to operate as photovoltaic devices. Poor film quality is considered as a major reason for this result. As an alternative approach, devices consisting of MEH-PPV, PCBM, and TPA-C 60 in a weight ratio of 0.8 / 3.9 / 0.3 were constructed. These devices were designed to test if TPA-C 60 could be used in small concentrations (based on its low solubility) to effect an enhancement in the performance of a device based on the 1/4 weight ratio of donor to acceptor that has proven so effective. In this case, it was envisioned that a small amount of the DA-dyad might act as an interface modifier to enhance the interaction between PCBM clusters and conjugated polymer in the blend. Here MEH-PPV was selected as the conjugated polymer component over MDMO-PPV based on the superior solubility of our sample of MEH-PPV (Aldrich) relative to MDMO-PPV (Aldrich). It was desired to solubilize the components of the device as much as possible. With 0.8 parts MEH-PPV, 3.9 parts PCBM, and only 0.3 parts TPA-C 60 , the device gave a 4/1 blend of acceptor to donor, while containing enough MEH-PPV to ensure quality film formation and minimal TPA-C 60 to ensure a fully soluble blend. Table 4-2 summarizes
212 the results for these devices, relative to MEH-PPV / PCBM (1/4) devices as a reference. It can be seen that incorporation of a small amount of TPA-C 60 in these devices had essentially no effect on the operation of the device. Table 4-2. Photovoltaic device results for devices based on MEH-PPV, PCBM, and TPA-C 60 (0.8 / 3.9 / 0.3). Blend composition V oc (V) J sc (mA/cm 2 ) FF (%) MEH-PPV / PCBM / TPA-C 60 (0.8 / 3.9 / 0.3) 0.76 3.49 0.34 0.91 MEH-PPV / PCBM (1 / 4) 0.78 3.42 0.34 0.90 Diode characteristics measured under AM1.5 (100 mW/cm 2 ) conditions. Figure 4-8 shows more detailed results for these devices. In Figure 4-8a, the IPCE for the MEH-PPV / PCBM / TPA-C 60 device is shown relative to the absorption spectra of a thin film blend of MEH-PPV and PCBM (1/4). Here it can be seen that the IPCE values typical of an MEH-PPV / PCBM blend are observed, and no clear contribution from the Ph 2 AS side chains of TPA-C 60 can be seen. The AM1.5 data for a typical three-component blend device is shown in Figure 4-8b. These results suggest that the small amount of added TPA-C 60 that serves to maintain the 4/1 acceptor/donor ratio, has little effect on the performance of the device. At higher concentrations of TPA-C 60 , cloudy solutions result, so comparisons with other three-component blend devices is not possible. Based on the very narrow composition range where TPA-C 60 can be used, and the lack of definitive evidence that the compound is contributing to the photovoltaic effect, TPA-C 60 does not appear useful for enhancing the efficiency of polymer-PCBM devices.
213 Figure 4-8. Photovoltaic device results for devices based on MEH-PPV, PCBM, and TPA-C 60 (0.8 / 3.9 / 0.3). (a) IPCE results are shown as black squares and the absorption of an MEH-PPV / PCBM (1/4) film is shown as a solid line for reference. (b) Diode characteristics of a typical pixel measured under AM1.5 (100 mW/cm 2 ) conditions. The red line represents the dark current and the black line represents the current under illumination. Ultimately there are several reasons why TPA-C 60 is not the optimum DA-dyad for use in photovoltaic devices. First, and most importantly, it has a very low solubility in common organic solvents, less so that even PCBM and films of TPA-C 60 cannot be deposited by spin-coating. These facts proved to be an insurmountable obstacle. Additionally there is no obvious evidence of a strong intramolecular charge transfer in TPA-C 60 , perhaps owing to the limited size of the oligomer. Finally, the Ph 2 AS oligomers absorb light at wavelengths less than 400 nm. As such Ph 2 AS cannot be effectively used as an antenna to capture long wavelength light. Finally, TPA-C 60 does not seem to impart any beneficial properties through incorporation of small amounts into MEH-PPV / PCBM devices. For these reasons we have turned our attention toward DA-dyads that bear longer oligomers that will impart greater solubility and that will more strongly absorb visible light.
214 4.3 Thienylene-vinylene Functionalized Fullerenes Another interesting class of oligomers for use in DA-dyads are thienylene-vinylene compounds. Soluble, alkyl substituted poly(thienylene-vinylenes) (PTVs) are known to have band gaps on the order of 1.6-1.8 eV depending on the alkyl substitution and the regioregularity of the polymer. 233 This gives the polymers a very favorable overlap with the solar spectrum and thus it is surprising that they have received so little attention for use in photovoltaic devices. 144c,144e Initial efforts 144c have shown AM1.5 efficiencies of ~0.2% in devices based on PTV and PCBM, and charge transfer from the photoexcited polymer to PCBM appears to be the primary mode of operation in these devices. Another interesting class of related materials are the oligomeric thienylene-vinylenes (OTVs). It has been shown that well-defined oligomers consisting of up to 16 TV repeat units can be synthesized and fully characterized. 234 In this case the relationship between oligomer length and electrochemical and optical properties is clearly established, and it is estimated that these properties will saturate to the value of the polymer when ~20 repeat units are achieved. Such OPVs have also been used in DA-dyads and triads with methanofullerenes. 235 In this case, a comparative study was performed in solution with blends of OTVâ€™s and N-methylfulleropyrrolidine relative to OTV functionalized fulleropyrrolidines. It was observed by photoinduced absorption spectroscopy (PIA) in polar solvents that charge transfer occurred in both the blends of OTV and methanofullerene as well as the covalently linked DA-dyads. There was a chain length dependence of the charge transfer that indicated that an oligomer of at least three repeat units was necessary to affect charge transfer in the blends, while DA-dyads showed charge transfer in polar solvent with two or more repeat units. These facts make OTV
215 functionalized methanofullerenes attractive targets for use as DA-dyads in photovoltaic devices. Figure 4-9 shows the OTV functionalized DA-dyad that will be referred to as TV-C 60 along with the oligomer, 5TV, which was used for comparative experiments. These compounds were synthesized in the laboratory of Professor Charles W. Spangler (Montana State University). Here, the attached donor chromophore is 5TV, which was synthesized by a series of Wittig reactions based on the ability to functionalize thiophene in the 2 and/or 5 positions with an aldehyde or phosphonium salt. Compound 2 was obtained from 5TV via end functionalization with a thioether moiety bearing a THP protected alcohol. After deprotection, reaction of the alcohol with malonyl dichloride in the presence of pyridine gave compound 2. The DA-dyad was then synthesized via the addition elimination route described for TPA-C 60 . O O O O C60DBU, CBr4toluene 2 S S S S S S S Br S S S S S Br TV-C60 O O O O S S S S S S S Br S S S S S Br S S S S S Br Br 5TV Figure 4-9. Structures of 5TV and TV-C 60 .
216 The electronic spectroscopy of TV-C 60 and 5TV was investigated in dichloromethane solution and the absorption and emission spectra of the two compounds are shown in Figure 4-10. Here it can be seen that 5TV has an absorption maximum at 527 nm, as well as two smaller peaks at 309 and 362 nm, while TV-C 60 has a maximum at 256 nm associated with the methanofullerene as well as two shoulders at 309 and 362 nm along with a second maximum at 527 nm corresponding to the 5TV oligomers. Thus the absorbance of the 5TV portion of TV-C 60 is essentially identical with that of 5TV itself, indicating little ground state electronic interaction in solution between 5TV and methanofullerene or between the two 5TV groups on TV-C 60 . The max of 527 nm for 5TV matches well with values reported in the literature for the dihexylthiophene-based 4TV and 6TV oligomers that show absorption maxima at 489 nm and 548 nm respectively in dichloromethane. 3004005006007008000.000.250.500.751.001.25 5TV PL5TV AbsTV-C60 PLTV-C60 AbsIntensity (a. u.)Wavelength (nm) Figure 4-10. Normalized absoprtion and emission spectra of TV-C 60 (0.1 mg/mL) and 5TV (0.1 mg/mL) measured in dichloromethane solution. The absorption and emission of TV-C 60 are represented with solid lines and the absorption and emission of 5TV are represented with dashed lines.
217 The emission spectra of the two compounds (also shown in Figure 4-10) show a sharp disparity between the oligomer and the DA-dyad. Here excitation of TV-C 60 at 520 nm results in emission with a maximum at 635 nm, which is in contrast to the emission observed for 5TV, which shows a maximum at 572 nm when excited at the same wavelength. The larger Stokes shift observed with the DA-dyad when compared to the free oligomer shows the emission of a lower energy photon, which suggests that the lowest energy singlet excited state in 5TV is somehow stabilized in the presence of methanofullerene or possibly due to intramolecular excimer formation with the other 5TV oligomer on the same molecule. In fact, the observation that no electronic interaction is occurring in the ground state, but a noticeable interaction is occurring in the excited state, is consistent with excimer formation. 236 Without the monosubstituted version of TV-C 60 for comparison, the interaction cannot be precisely identified. The electronic spectroscopy of TV-C 60 was also investigated in thin films and the results are shown in Figure 4-11. No such data could be obtained for 5TV, as it did not show suitable film-forming properties. In thin film form, TV-C 60 shows a max of 537 nm and the onset of the methanofullerene peak (< 400 nm). The slight red-shift in the absorption maximum of the oligo-TV, is to be expected when comparing film measurements to solution measurements, due to extended conjugation and increased intermolecular interactions in the film. From this data one can estimate a band gap of 1.9 eV for the oligomer 5TV, which indicates a band gap slightly larger than PTV, as expected for this short, discrete oligomer. Interestingly, no measurable emission is observed for TV-C 60 in thin film form. This could be indicative of charge transfer or simple PL quenching due to aggregation in the thin film.
218 3004506007500.00.20.40.60.81.0 Absorbance (a. u.)Wavelength (nm) Figure 4-11. Normalized absorption spectra of TV-C 60 in a thin film. The film was spin-coated onto PEDOT-PSS coated glass from dichlorobenzene (6 mg/ml). To further investigate the electronic structure of TV-C 60 , the electrochemistry of TV-C 60 and 5TV were investigated in dichloromethane solution with TBAP as the supporting electrolyte using both CV and DPV as seen in Figure 4-12. In Figure 4-12a, the CV for TV-C 60 is shown. Here the E 1/2 for the first reduction is observed to be .95 V vs. Fc/Fc + which is very close to the value of .90 V observed for TPA-C 60 . The second reduction is also observed at .3 V and a third, poorly defined reduction is observed at approximately .8 V. Also present in this CV is the E 1/2 of oxidation, which is observed at +0.2 V. Notice in this CV that the peaks are very poorly defined and the onsets are even more poorly defined. In order to gain a clearer picture of the electrochemical processes in TV-C 60 , DPV was also performed on the compound in solution as seen in Figure 4-12b. In this case the E 1/2 values of the three observed reductions were found to be .95, -1.35, and -1.85 V vs. Fc/Fc + . For the oxidation, an E 1/2 of +0.20 V was also observed.
219 Figure 4-12. Solution electrochemistry of TV-C 60 and 5TV. (a) CV of TV-C 60 , (b) DPV of TV-C 60 , (c) CV of 5TV, (d) DPV of 5TV. All measurements were performed in 0.1 M TBAP / dichloromethane with a platinum button working electrode and a silver wire reference electrode calibrated relative to Fc/Fc+. All potentials are reported relative to Fc/Fc+. For comparison, the electrochemistry of 5TV was also evaluated under the same conditions as seen in Figure 4-12c and 4-12d. In the CV of 5TV, two redox processes can be observed at 0 V and 0.1 V. These two processes are more clearly defined in the DPV of 5TV seen in Figure 4-12d. For comparison, in the work of Roncali 234 on discrete TV oligomers, it was found that for the nTV oligomers with n =4, two reversible oxidation processes were observed, which were attributed to the formation of the radical cation and dication. These two redox processes were not observed upon oxidation of TV-C 60 measured under the same conditions. For TV-C 60 the potential was scanned up to +0.75 V and no second oxidation was observed. Furthermore the E 1/2 of the oxidation in TV
220 C 60 is observed to be shifted 100 mV anodic of the second oxidation in 5TV. This behavior indicates a subtle difference between 5TV in its molecular form and when it is attached covalently to another 5TV and a methanofullerene The fact that the oxidation potential of 5TV shifts in the anodic direction in TV-C 60 suggests that some electronic interaction may somehow destabilize or inhibit radical cation formation, making the 5TV portion of TV-C 60 more difficult to oxidize. The effect of the end groups on 5TV cannot be ignored, as one of the bromine end groups in 5TV has been replaced with sulfur in TV-C 60 . However if this anodic shift in the oxidation of TV-C 60 was merely an endgroup effect, the opposite behavior would be expected, with the sulfur-capped oligomer in TV-C 60 expected to oxidize more easily than the dibromo-5TV. The electrochemistry of TV-C 60 was also investigated in thin film form as shown in Figure 4-13. In Figure 4-13a, one can see that TV-C 60 gives a single oxidation with and E 1/2 of +0.2 V, which matches with the data obtained in solution (Figure 4-12 a). The reduction was however poorly defined in the thin film and an accurate E 1/2 value could not be estimated from the first reduction. Note that the oxidation and reduction were addressed independently in the thin film in order to avoid prepeaks, which appear when scanning the full range. Figure 4-13b shows the DPV for the film of TV-C 60 . In this case, the E 1/2 of the oxidation was also observed to be +0.2 V as was seen in Figure 4-12b from solution measurements. Here the reduction is very poorly defined and no values could be accurately obtained from the data. Based on thin film electrochemistry, the onset of oxidation for TV-C 60 is observed to be +0.1 V vs. Fc/Fc + (the same value is also measured in solution), which corresponds to a HOMO of 5.2 eV for the donor portion of
221 the molecule. The LUMO of the acceptor portion can be estimated to at 4.25 eV based on the onset of the first reduction (-0.85 V vs. Fc/Fc + ) measured in solution. Figure 4-13. Thin-film electrochemistry of TV-C 60 . (a) CV. (b) DPV. Measurements were made on a platinum button working electrode in 0.1 M TBAP / acetonitrile as the supporting electrolyte. Potentials were measured vs. a silver wire pseudo-reference electrode calibrated vs. Fc/Fc + . With a preliminary understanding of the electronic structure of TV-C 60 , the compound was then used to construct photovoltaic devices. As with TPA-C 60 , several approaches were used for incorporating TV-C 60 into devices. The main advantage of TV-C 60 relative to TPA-C 60 is that the longer wavelength absorptions allow it to be detected more easily using IPCE measurements. Table 4-3 summarizes the device results for a series of devices incorporating TV-C 60 measured under AM1.5 conditions. In the first approach, it was attempted to make devices using pristine films of TV-C 60 (Table 4-3, entry 1). As with TPA-C 60 , this approach failed based of the poor film quality of the pristine films. Likewise, blends with PCBM (Table 4-3 entry 2) failed to yield high enough quality films to yield operational devices. In order to generate a film of device quality based on TV-C 60 and PCBM, PVK was used as an additive. Here the device active layer was made up of TV-C 60 , PCBM, and PVK in a ratio of 1 / 3.7 / 0.3 in order
222 to ensure a 1/4 ratio of donor to acceptor. This is based on the fact that TV-C 60 is 70% donor and 30% acceptor by weight. The AM1.5 results for this type of device are shown in Table 4-3, entry 3. It can be seen that an efficiency of 0.16% was achieved, with a short-circuit current of 1 mA/cm 2 . More importantly, Figure 4-14 shows the photocurrent action spectra for this type of device relative to the absorption spectra of pristine TV-C 60 . It can be seen that a peak IPCE efficiency of ~7.5% is observed at 550 nm and that the shape of the IPCE curve matches the absorbance of the covalently attached 5TV oligomers, indicating them as the source of photocurrent in the device. While this is not definitive proof of intramolecular charge transfer, it is strong evidence that charge transfer from TV oligomer to methanofullerene is occurring. Table 4-3. Device results for TV-C 6 0 containing devices measured under AM1.5 conditions. Device # Device Compositon V oc (V) J sc (mA/cm 2 ) FF (%) 1 TV-C 60 ----------------2 TV-C 60 / PCBM (1.4 / 3.6) ----------------3 TV-C 60 / PCBM / PVK (1 / 3.7 / 0.3) 0.58 1.01 0.27 0.16 4 TV-C 60 / PCBM / MDMO-PPV (1 / 3.7 / 0.3) 0.49 1.08 0.29 0.17 5 MDMO-PPV / PCBM / TV-C 60 (0.8 / 3.9 / 0.3) 0.65 1.94 0.31 0.38 6 MDMO-PPV / PCBM (1 / 4) 0.69 3.83 0.42 1.10 PVK = poly(vinylcarbazole) M w = 1,100,000 g/mol, Aldrich.
223 400500600700012345678 IPCE (%)Wavelength (nm) Figure 4-14. Photocurrent action spectra for devices containing TV-C 60 with PCBM and PVK. Solid line represents the absorption spectra of TV-C 60 on PEDOT-PSS coated glass and the curve denoted by black squares represents the IPCE for a device consisting of TV-C 60 / PCBM / PVK (1 / 3.7 / 0.3). To further examine the utility of TV-C 60 for use in PVDs, the DA-dyad was blended with MDMO-PPV and PCBM in several combinations. MDMO-PPV was selected in favor of MEH-PPV based on the superior performance of MDMO-PPV devices in relation to MEH-PPV devices. Additionally, as TV-C 60 is much more soluble than TPA-C 60 , blends with MDMO-PPV were readily realized. Two different blend compositions were chosen to evaluate the effect of various TV-C 60 concentrations on the performance of MDMO-PPV/PCBM solar cells. Here a blend consisting of weight ratios of 1 / 3.7 / 0.3 of TV-C 60 / PCBM / MDMO-PPV (Table 4-3 entry 4) and 0.8 / 3.9 / 0.3 of MDMO-PPV / PCBM / TV-C 60 (Table 4-3 entry 5) were selected to give devices that were 4/1 acceptor to donor. In the case of device 4, TV-C 60 is the majority donor material and in device 5 MDMO-PPV is the majority donor material. As can be seen, device 4 gives almost identical performance to device 3, which only differed in the use of PVK rather than MDMO-PPV. The IPCE data for these devices is shown in Figure 4-15. It can
224 be seen that TV-C 60 is a strong contributor to the photocurrent in device 3 and device 4, with device 4 showing a maximum at 530 nm of nearly 12%. In device 5 however, MDMO-PPV is the majority donor and the AM1.5 efficiency essentially doubles along with a concomitant doubling of the short circuit current relative to devices 3 and 4. In this case, as can be seen in the IPCE data in Figure 4-15, the primary contributor to the photocurrent is now MDMO-PPV, as the maximum in the IPCE is at 480 nm with a value of ~20%. As a reference, a device based on a 1/4 blend of MDMO-PPV and PCBM was constructed. Table 4-3 entry 6 shows that this device gave an efficiency of 1.1% with a short circuit current of nearly 4 mA/cm 2 . The improvement in efficiency for device 6 relative to device 5 is nearly threefold, indicating that even a small amount of TV-C 60 is enough to significantly alter (decrease) the performance of the MDMO-PPV device. Figure 4-15 illustrates this contrast even more clearly based on the IPCE results. 40050060070005101520253035 IPCE (%)Wavelength (nm) Figure 4-15. IPCE data for a series of PVDs containing TV-C 60 . (): TV-C 60 / PCBM / PVK (1/3.7/0.3), (): TV-C 60 / PCBM / MDMO-PPV (1/3.7/0.3), (): MDMO-PPV / PCBM / TV-C 60 (0.8/3.9/0.3), (): PCBM / MDMO-PPV (4/1).
225 Based on the above results it is clear that TV-C 60 can operate as the photoactive material in organic solar cells. Optimizing the blending conditions with the appropriate polymer and PCBM could lead to enhanced photovoltaic performance. The polymer, PTV presents itself as a possible polymer for use in the development of three component blends based on the inherent compatibility it will offer for TV-C 60 . 4.4 Conclusions The results discussed in this chapter raise several important points concerning the use of C 60 -DA-dyads in photovoltaic devices. First, dyads represent an interesting alternative to the simple polymer-PCBM devices that were the focus of chapter 3. For the evaluation and application of such dyads, long wavelength absorbance and sufficient solubility are necessary. Additionally, based on the results presented here, dyads are best used in combination with PCBM and another donor polymer (such as PVK) in order to maintain the C 60 / donor mass ratio necessary to ensure a bicontinuous network, as well as to provide adequate film quality. Finally, the precise interplay between the various components in multicomponent blends must be thoroughly evaluated, as it is difficult to predict. In the case of MEH-PPV / PCBM / TPA-C 60 devices (0.8/3.9/0.3), the device characteristics were essentially the same as for devices based on 1/4 blends of MEH-PPV and PCBM. On the other hand devices of the same composition based on TV-C 60 and MDMO-PPV gave efficiencies that were nearly a factor of three lower than MDMO-PPV/PCBM (1/4) devices. In order to explain such phenomena, a precise understanding of the photophysics and morphology in these devices is required. 4.5 Photovoltaics: Outlook and Perspective This chapter brings to a close the discussion of photovoltaic devices, which has accounted for the majority of this dissertation. As such, this is an interesting point to
226 reflect on the general concepts and key points that have been brought out in this work. It should be clear to the reader that the development of polymer-based photovoltaics is an endeavor that is still in its infancy. Despite the hundreds (if not thousands) of papers that have been published, the field has grown ever so gradually over the past 10 years. In 1995, the first bulk heterojunction device based on MEH-PPV and PCBM was reported, but despite the tremendous research effort that has been focused on improving this basic device, little has changed (see chapter 1). The best devices are still based on blends of conjugated polymers and PCBM and the best, reproducible efficiencies are still less than 4%. However, there are a number of reasons why this field is so attractive for so many chemists, physicists, and materials scientists. The prospect of harnessing solar energy using lightweight flexible devices that can be fabricated with solution processing techniques, promises to be a very high impact development. From a scientific point of view, the subject is so intriguing because of its interdisciplinary nature that transcends the traditional boundaries of science. If conjugated polymer based photovoltaics ever find commercial applications, the success will draw from diverse areas of expertise. From a technological point of view, there are several major challenges that must be overcome for the realization of commercial devices. While device efficiency is important, efficiency is not the major obstacle. A lightweight, flexible device that operates at even 1% efficiency could be useful if it could be produced cheaply and operate at a constant level for a reasonable amount of time (weeks or years depending on the application). As such, the first and foremost problem is the device stability. The currently used platform for fabrication of conjugated polymer devices does not yield devices with any appreciable
227 long-term stability. Even for devices fabricated in the glovebox and measured in the glovebox, measured efficiencies drop off rapidly in a matter of only hours. This degradation can be credited to the action of intense UV light on the organic polymer. Certainly, UV filters could be used, however this would not overcome the problems of sample heating. If a material is going to be used as the active component of a solar cell, it must be able to withstand heat, without degrading. Polymer-PCBM blends are known to be unstable at elevated temperatures. When a polymer is heated above its T g , the polymer-PCBM blend is no longer static. Specifically, PCBM has a tendency to diffuse through the amorphous polymer and aggregate into large domains, 89d which irreversibly reduces the efficiency of the device. A discussion about blend morphology also brings up an interesting point about bulk heterojunction photovoltaic devices. As chemists and materials scientists, we seek to understand relationships between structure and property, whether it be chemical structure, blend structure, or device structure. For all the effort we put into synthesizing polymers with desired band gaps and frontier orbital energies, as well as engineering electrodes to have the appropriate work function and conductivity, we give away all this engineered control when the bulk heterojunction is employed. Every set of materials gives different blend properties. Not just every different type of polymer, but every single batch of every polymer. Additionally, solvent choice, concentration of the active material in solution, active layer composition, spin coating speed, annealing temperature and time, and so many other variables affect the morphology of the active layer and thus the characteristics of the device. In order to gain back the control, we need reproducible ways to develop controlled morphologies. In this area, donor-acceptor block copolymers, or
228 patterned donor or acceptor layers promise to yield interesting results. The morphology of the active layer is so critical as it plays a central role in the effectiveness of the processes of charge transfer and charge transport. At the current time, the bulk heterojunction is still the state of the art in polymer photovoltaics. There are several reasons why it is a worthwhile endeavor to seek to produce the highest efficiency bulk heterojunction device. First, by testing the limits of this platform we can learn all the advantages and disadvantages it offers. When we really, deeply understand these factors, we will be more prepared to move on to the next generation of polymer photovoltaics. Additionally, bulk heterojunction devices are not completely understood. It is definitely a worthwhile scientific task to understand how these devices operate. By doing so we learn much about organic electronics in general, as well as a great deal about the fundamental physics and chemistry of conjugated polymers and fullerenes. Would we know so much today about charge transfer in conjugated polymers or about the chemistry of fullerenes if it werenâ€™t for the research effort in photovoltaics? While the photovoltaic work presented in this dissertation did not improve on the highest reported efficiencies already found in the literature, several interesting points were brought out in this work that can serve as a guide to a novice in the field. Primarily, a polymer based bulk heterojunction device begins with the polymer. For improvements in polymer-PCBM devices, robust narrow band gap polymers are required. A systematic approach is the most useful in developing the structure-property relationships that can teach us which structural attributes of polymers give the best devices. In order to function as a donor in a PVD a polymer must absorb sunlight, transfer electrons, and transport
229 holes. Additionally, in order for a polymer to be useful in a blend device with PCBM, the polymer must fulfill several requirements beyond the electronic requirements discussed in chapter 3. First, the polymer must be of extremely high quality. A real effort needs to be undertaken to develop conjugated polymers that exhibit the molecular weights and mechanical properties of conventional polymers. With robust, durable materials, new opportunities in all areas of organic electronics will be possible. The polymer must be free of impurities and defects and the polymer must be of a suitably high molecular weight that it can form thick (~100 nm or more) films of high quality. The polymer must be completely soluble in a solvent that will solubilize PCBM (or the selected acceptor). Chlorobenzene or dichlorobenzene are excellent choices due to the favorable morphologies that are often observed when using these solvents to spin coat polymer-PCBM blends. Not only must the polymer be of high quality, but all materials used in PVDs must be of the highest quality. If PCBM is used, it must be pure, dry, and free of other trace solvent from the synthesis. Likewise, any solvents used for casting the active layer must be rigorously purified. A carefully distilled and degassed solvent is essential, whether spin coating, spray coating, or ink jet printing is used. The basic premise of using pure, clean materials applies equally to the materials used to construct the devices. As such, ITO, PEDOT-PSS, LiF, and aluminum must also be scrutinized. Using the currently known techniques, the production of high efficiency solar cells requires the fabrication, handling, and measuring of polymer PVDs under inert conditions. Another interesting avenue to improving bulk heterojunction devices is to replace PCBM. As observed in chapter 4, PCBM holds a special place among C 60 derivatives.
230 This may simply be due to the fact that it has received immeasurably more attention than any other fullerene derivative. An important lesson learned in chapter 4 is that with C 60 devices, the donor / C 60 weight ratio is critical. Sufficient acceptor must be included to generate the bicontinuous network that drives the bulk heterojunction. Additionally, C 60 is so hard to replace as the acceptor of choice because it has a very high electron affinity and it shows high electron mobilities. These two qualities have proven difficult (impossible) to match with other organic small molecules. Perhaps carbon nanotubes will displace fullerenes as the material of choice. Even inorganic nanoparticles may provide the answer, although the purist is resistant to abandoning the dream of an all-organic device. Ultimately this field is wide open. The successful scientist will not be trapped by the traditional boundaries between physics, chemistry, and engineering, but instead will seek to understand the device concept form the carbon-carbon bonds in the polymer backbone to the nature of the polymer exciton, to the effects of temperature during the thermal evaporation of the aluminum electrode. Only in this way can the field advance. The success of the field relies on individuals who are unwilling to dismiss poor device results as the consequence of an â€œunoptimized (undetermined) morphology.â€ In photovoltaic devices, the details make all the difference.
CHAPTER 5 ELECTROPOLYMERIZABLE VARIABLE BAND GAP CONJUGATED POLYMERS As discussed in chapter 1, there are numerous possible applications of conjugated polymers based on the optical properties exhibited by this class of materials. While the bulk of this thesis has been focused on the development of conjugated polymers for photovoltaic devices and the evaluation of such devices, the physical and materials constraints imposed by this application (i.e. solubility, processability, etc.) severely limit the scope of the target polymers that are evaluated. From an electronic standpoint, ideal electron donor polymers for PVDs possess several properties that are potentially attractive for numerous other applications. Specifically, these polymers have a narrow band gap and potentially accessible p and n-type redox states. These properties are especially attractive for electrochromic applications as well as numerous other applications that exploit the optoelectronic, redox, or conductive properties of conjugated polymers. This chapter deals with the development of new routes toward novel narrow band gap donor-acceptor polymers based on tetrazine acceptors. The application driven focus here is to develop electropolymerizable multi-color electrochromic polymers. The conceptual interest lies in expanding the scope of knowledge concerning DA polymers. 5.1 Developing New Routes toward High Electron Affinity Narrow Band Gap Polymers The conceptual basis of using the donor acceptor approach to achieve narrow band gap polymers was discussed in chapter 1. In this approach, the combination of an 231
232 electron-rich donor and an electron-deficient acceptor results in a conjugated polymer with a compressed band gap. By carefully selecting the appropriate structure of the donor and the acceptor, one can tune the magnitude of the band gap and the absolute energies of the frontier orbitals. One specific donor-acceptor approach to low band gap polymers has centered on the use of nitrogen containing heterocycles as effective acceptors. Heterocycles such as pyridine or quinoxaline (see Figure 5-1) which contain an imine nitrogen (C=N) are electron deficient and thus serve as efficient electron acceptors. 237 In specific examples, band gaps as low as 0.3 eV and 0.5 eV has been reported for the case of electropolymerized donor-acceptor-donor polymers based on thiophene as the donor and either thiadiazolothienopyrazine or benzothiadiazole as the acceptor. 238 , 239 The other advantage of incorporating electron deficient N-heterocycles into a conjugated polymer is that the LUMO energy of a conjugated polymer will be lowered due to the presence of the more electronegative nitrogen atom. 240 As such, the resulting polymer will be easier to reduce, more likely to exhibit reversible n-type doping, and show higher levels of electron mobility. Such a polymer could serve as an efficient electron accepting material in photovoltaic devices or as an electron transport material in LEDâ€™s, if a soluble processable analogue could be produced. 5.1.1 Nitrogen Containing Heterocycles and Tetrazine Figure 5-1 shows several common electron deficient N-heterocycles along with the reported reduction potentials. Naphthalene and biphenyl are shown as a reference for the reduction potentials of hydrocarbon small molecules. Notice that both hydrocarbons are slightly more easily reduced than pyridine based on the extended delocalization of the -system in these two molecules relative to pyridine. For the single-ring N-heterocycles, a
233 general trend toward an anodic shift in the reduction potential can be observed as the number of nitrogen atoms in the ring is increased from one in pyridine up to four in s-tetrazine (s = symmetrical) or 1,2,4,5-tetrazine (which will simply be referred to as tetrazine). In this case the reduction potential shifts nearly 2 V in the anodic direction indicating that tetrazine is significantly easier to reduce than pyridine as well as being nearly 1 V easier to reduce than either of the triazines. In fact tetrazine is the most electron deficient of all the simple N-heterocycles. 241 Even the delocalized fused-ring compound pteridine, which also contans four nitrogen atoms, is more difficult to reduce than tetrazine. While pyridine, 2,2-bipyridyl, 242 quinoxaline, 243 and pyridopyrazine 244 have been utilized in donor-acceptor conjugated polymers, relatively little attention has been given to tetrazine. N pyridine-3.10 V biphenyl-2.93 Vnaphthalene-2.88 VNN NN NN pyrimidine-2.72 Vpyridazine-2.56 Vpyrazine-2.47 VNNN NNN s-triazine-2.43 Vas-triazine-1.96 V NNNN s-tetrazine-1.19 V N N 2,2'-bipyridyl-2.50 V NN quinoxaline-2.00 VNN N pyrido[3,4-b]pyrazine-1.76 VNN NN pteridine-1.43 V Figure 5-1. Reduction potentials of common nitrogen containing heterocylces. The data in this table is adapted from the literature and the potentials listed are vs. Fc/Fc + by conversion from reported potentials vs. Hg pool, which is .52 V vs. SCE and assuming that Fc/Fc + is 0.38 V vs. SCE (see chapter 2). Indeed only one example of a tetrazine-containing conjugated polymer has been reported to date. 245 In this case, Audebert et al. reported a 3,6-bis-bithiophene-tetrazine
234 monomer, bis[5-(2,2-bithienyl)]-tetrazine or BBiThTz (Figure 5-2), and found that it could oxidaively electropolymerize to yield an electroactive conjugated polymer thin film which showed a reversible oxidation and reduction at +0.65 V and .25 V vs. Fc/Fc + respectively. The band gap of this polymer was optically determined to be ~1.9 eV, which is reduced relative to polythiophene (2.2-2.3 eV) based on the donor-acceptor interaction along the backbone. The reduction potential of the monomer (in solution) was observed to be .2 V vs. Fc/Fc + , almost the same as that observed for the polymer and that observed for tetrazine itself (Figure 5-1). This relationship between acceptor reduction potential, monomer reduction potential, and polymer reduction potential has been previously observed in N-heterocycle based donor-acceptor polymers and thus, knowledge of the acceptor reduction potential is a good approximation for the reduction potential of the polymer. 122a This relationship between reduction potentials is based on the direct correlation between reduction E 1/2 and LUMO energy. 246 NNNN S S S S BBiThTz +ENNNN S S S S PBBiThTz n Figure 5-2. Stucture of BBiThTz and PBBiThTz. Based on this precedent, it is of interest to develop additional polymerizable donor-acceptor tetrazine monomers. With the incorporation of stronger electron donors, polymers with band gaps less than 1.9 eV can be targeted. In addition to promising a high electron affinity and a narrow band gap, donor-acceptor polymers incorporating tetrazine are attractive targets for electrochromic applications based on the strongly colored nature of tetrazine. Tetrazine is known to show a * transition at 252 nm in cylcohexane similar to the 254 nm absorption of benezene. However, tetrazine also possesses n-*
235 transitions at 320 nm and an absorption at 542 nm, which gives rise to the deep red color of tetrazine. This visible absorption band suggests the possibility of interesting electrochromic phenomena in conjugated polymers. 5.1.2 Synthesis of Electropolymerizable Tetrazine Monomers for DA Polymers As a starting point in the development of DA polymers based on tetrazine, the synthesis of 3,6-bis-donor-heterocycle tetrazine monomers is targeted. Here electropolymerizable monomers are desired, as the synthesis is simplified relative to soluble polymers and simple polymeric systems can provide a first approximation of the utility of this class of polymers. Electropolymerizable monomers provide polymers as electrode bound films that are readily characterized. Limited literature precedent exists concerning the synthesis of 3,6-bis-substituted tetrazines with heteroaromatic substituents, but several routes have been used for the realization of such compounds. A general route for the synthesis of bis-donor substituted N-heterocycle monomers has been via the metal catalyzed coupling of the bis-halogenated acceptor with two equivalents of appropriately functionalized donor. 244, 247 Figure 5-3 shows the synthesis of BEDOT-PyrPyr as a representaive example of this well-known method of synthesis. Here the use of the Stille coupling is shown, but Negishi, Kumada, and Suzuki couplings are analogous reactions that find utility in similar syntheses. S OO Sn(CH3)3 2N Br Br + PdCl2(PPh3)2N S S OO OO NN NN BEDOT-PyrPyr Figure 5-3. Synthesis of BEDOT-PyrPyr as an illustrative example of a DA monomer synthesis via metal catalyzed coupling.
236 The situation becomes more complicated with tetrazines. Organometallic derivatives of tetrazines are known to be unstable 248 and thus the dihalo-tetrazine is the only viable starting material for cross-coupling reactions. Only 3-alkyl-6-bromo-tetrazines are known 249 , while 3,6-diiodotetrazine and 3,6-dibromotetrazine are unknown. However, 3,6-dichlorotetrazine is a known compound; 250 and Hiskey 251 has reported an effective route for its synthesis. The choice for coupling partner with 3,6-dichlorotetrazine is also rather limited, as Grignard reagents, arylzinc reagents, and boronic acids are known react with the tetrazine ring to give unwanted side products or degredation products. To date, only Sonogashira coupling conditions have proven effective for the substitution of tetrazines via cross coupling chemistry. 248 In this case 3,6-dichlorotetrazine was not found to give stable bis-acetylene compounds, but a series of 3-substituted-6-chlorotetrazines did give rise to stable products under Sonogashira conditions. Clearly this approach is limited in scope, and to date, no 3,6-diaryl-tetrazines have been reported via a cross coupling method. As such, metal catalyzed cross-coupling chemistry does not appear to be a viable route for the synthesis of DA tetrazine monomers. The simplest 3,6-diaryl-tetrazine is the 3,6-diphenyl derivative. Pinner first reported the synthesis of this compound in 1893 and the synthetic sequence is shown in Figure 5-4a. 252 In this case the aryl nitrile was reacted with HCl in the presence of ethanol under anhydrous conditions to give the hydrochloride salt of the imido ester (imidate). Further reaction of the imidio ester with hydrazine in the presence of aqueous base (KOH or NH 4 OH) gave rise to the amidrazone (not isolated), which underwent a bimolecular
237 condensation to give the dihydrotetrazine (shown here as the 1,2-dihydro derivative). 253 The dihydrotetrazine was found to oxidize readily to the tetrazine in air. CN HCl (g)ethanol NH OEt HCl NH4OH (aq)NH2NH2 NNH2 NH2 imido esteramidrazone NNH2 NH2 amidrazone2 N HN NH N [O] N N N N dihydrotetrazine a CN HCl (g)methanol NH OMe HCl imido ester methanol, NEt3NH2NH2 N HN NH N dihydrotetrazine N N N N tetrazinetetrazine bc CN NH2NH2 NNH2 NH2 amidrazone N HN NH N dihydrotetrazine [O] N N N N tetrazine [O] NNH2 NH2 amidrazone NNH2 NH2 amidrazone2 Figure 5-4. Synthesis of 3,6-diphenyl-terazine. Via the Pinner route (a), using the Wiley modification (b), and using the generalized one-step reaction (c). In all cases the dihydrotetrazine shown is the 1,2-dihydrotetrazine. In reality both 1,2 and 1,4-dihydro products may be formed, although neither is stable enough to fully characterize. The original Pinner conditions gave rise to low yields and difficult purification steps. However, an improvement in this method was reported by Wiley in 1957 which gave rise to the 3,6-diphenyl, 3,6-di-m-tolyl, and 3,6-di-p-biphenyl derivatives in greater than 50% yield, as shown in Figure 5-4b. 254 The only modification employed by Wiley was to react the imido ester with hydrazine in the presence of triethylamine under
238 anhydrous conditions in methanol, which allowed isolation of the dihydrotetrazine by precipitation from methanol in greater than 50% yield. As a more general reaction, it has been found that nitriles will react directly with hydrazine at elevated temperatures to form intermediate amidrazones, which undergo the subsequent bimolecular condensation to the dihydrotetrazine as seen in Figure 5-4c. 255 In most cases the dihydrotetrazine spontaneously oxidizes to the tetrazine. The route shown in Figure 5-4c is in fact is often employed in the synthesis of symmetrically substituted 3,6-dihyrotetrazines, and furthermore, aryl nitriles usually give satisfactory yields by this method. 256 The benefit of all of the routes shown in Figure 5-4 is that the tetrazine ring is formed during the reaction, and the use of air sensitive organometallic reagents is avoided. A variety of additional, methods have been developed for the synthesis of symmetrically substituted 3,6-tetrazines, 256 however most of these methods rely on a condensation similar to the original Pinner method to yield the dihydrotetrazine. Both 1,2 and the 1,4-dihydrotetrazines have been reported, but in most cases the dihydro-compound oxidizes so easily that it is either not isolated or characterized. Often the oxidation is aided by the use of an oxidizing agent such as nitrous acid, ferric chloride, hydrogen peroxide, or isoamyl nitrite. An interesting variation of the Pinner method has been the use of elemental sulfur in the presence of hydrazine, which has been shown effective as a method of converting aromatic nitriles into tetrazines. 257 In this case, reaction times on the order of 1-3 hours and yields in the range of 76-94% were reported for a variety of substituted benzonitriles. This represents an advantage over simply reacting an aromatic nitrile with hydrazine or hydrazine hydrate, as reaction times are often on the order of days in this case. Despite
239 some doubts about the validity of this approach, 258 it has been successfully employed in the synthesis of bis-(2-thienyl)-tetrazine (BThTz) and bis-(2-pyrrolyl)-tetrazine, 259 as well as bis[5-(2,2-bithienyl)]-s-tetrazine (BBiThTz) (Figure 5-2), 245 and a variety of other 3,6-substituted aryl tetrazines. 260 Although this reaction has been known for more than 30 years, a mechanism was only proposed recently by Audebert, 259 although the validity of this mechanism has yet to be rigorously tested. One other interesting approach for the synthesis of 3,6-diaryl-substituted tetrazines was also explored by Audebert and relied on the enhanced reactivity of tetrazines toward nucleophilic aromatic substitution. 261 In this case, bis(N-pyrrolyl)tetrazine was synthesized by reaction of two equivalents of the N-pyrrolyl anion on bis(3,5-dimethylpyrazol-1-yl)tetrazine as seen in Figure 5-5. The leaving group 3,5-dimethylpyrazol-1-yl has been shown to be highly effective for nucleophilic substitutions on tetrazines. 251,261, 262 While the reaction was successful with pyrrole using the nitrogen as the nucleophile, attempts to form the bis--heterocycles via a carbon nucleophile were unsuccessful. With the much more reactive carbon nucleophiles only tars could be obtained. NNNN N N N N N Li NNNN N N Figure 5-5. Synthesis of bis-N-pyrrolyl-tetrazine by nucleophilic aromatic substitution. Figure 5-6 shows the structures of all the compounds that have been reported in the literature that posses a suitable structure to allow for the realization of electro-polymerization to yield linear polymers with tetrazine in the main chain. Here compound 1 has been reported by Audebert and Soluducho. 263 In this case Soloducho synthesized 1
240 from the 2-carbonitrile of pyrrole in the presence of hydrazine, whereas Audebert utilized the sulfur-hydrazine route discussed previously. Yields for this compound were essentially the same by both methods, although reaction times were reduced from 18 hours to 1.5 hours using the sulfur modification. Compounds 2 and 3 were also reported by Soloducho and were synthesized by Stille coupling of the trialkyl-tin thiophene and trialkyl-tin EDOT with 3,6-bis-(4-halophenyl)-tetrazine. For these compounds, the 3,6-bis-(4-iodophenyl)-tetrazine or 3,6-bis-(4-bromophenyl)-tetrazine were synthesized via Sandmeyer chemistry from 3,6-bis-(4-amino)-tetrazine after Pinner chemistry was performed on 4-aminobenzonitrile. Compounds 4 and 5 were reported by Audebert as discussed above. NNNN HN NH NNNN S S NNNN S S S S NNNN S S NNNN S S OO OO 1234 BThTz5 BBiThTz Figure 5-6. Structures of donor-acceptor tetrazine monomers with terminal electropolymerizable groups reported in the literature. Concerning the compounds in Figure 5-6, no electrochemistry has been reported for compounds 2 and 3. The anodic electropolymerization of compounds 1 and 4 was attempted, but no polymer films were formed. The oxidation potential of compound 4 (BThTz) was found to be > 1.8 V vs. Ag/Ag + (> 1.7 V vs. Fc/Fc + ) and the oxidation was
241 not observed to give a well-defined peak. No explanation has been proposed to explain the inability of BThTz to electropolymerize, but it is possible that lack of film formation is a manifestation of the polythiophene paradox. 264 In the case of polythiophene, the monomer oxidation potential for thiophene is more than 1 V anodic of the polymer oxidation potential. As a consequence, while polythiophene is deposited on the electrode, the polymer concurrently degrades due to the high applied potentials. It is suspected that at such high potentials, the increased density of positive charges along the polythiophene backbone makes it highly susceptible to nucleophilic attacks, which introduce conjugation breaks into the backbone. The overoxidized polymer serves to passivate the electrode and prevents polymer film growth. The same phenomenon could be responsible for the lack of polymerization with BThTz. This is supported by the observed decrease in current response on scans subsequent to the initial scan during the attempted electropolymerization of BThTz by cyclic voltammetry. For comparison, compound 1 was found to give a well-defined oxidation at +0.6 V vs. Ag/Ag + (0.5 V vs. Fc/Fc + ), but again no polymerization occurred. It was proposed by the authors that a self-deprotonoation occurred in the cation-radical that led to an equilibrium between structures in which the pyrrole-hydrogen was either bonded to a nitrogen on the tetrazine ring or on the pyrrole-nitrogen. This delocalization of charge is thought to stabilize the radical-cation to the point that it will not couple with another radical cation. In the case of compound 5 (BBiThTz), electropolymerization does occur with a peak monomer oxidation potential of +0.85 V vs. Fc/Fc + , which is nearly 1 V lower than for compound 4. As this is the only conjugated polymer issued from an electropolymerizable tetrazine monomer reported to date, it is interesting to expand the
242 scope of the monomer family in order to discover what properties of this class of polymer can be exploited. Additionally it is interesting to investigate the fundamental chemistry of this class of monomer and determine what structural factors are the most effective for generating conjugated polymers. 5.2 Synthesis of Bis-Heterocylcle Monomers for Electropolymerization Based on the above discussion we targeted a family of bis-heterocycle-tetrazine monomers based on EDOT as the electropolymerizable moiety. As discussed in chapter 1, EDOT presents numerous advantages relative to thiophene for electropolymerization. Importantly, EDOT is substituted in the 3 and 4 positions, preventing the possibility of unfavorable -coupling. Additionally, EDOT is more electron-rich than thiophene and thus not only has a lower oxidation potential, but the stronger donor character gives a stronger donor-acceptor interaction in DA polymers. Figure 5-7 shows the structures of the compounds that have been targeted, along with the structures of the compounds that were studied as reference materials. The general synthetic procedure that was followed was the reaction of the aromatic nitrile with hydrazine and sulfur in ethanol at reflux in a sealed pressure tube. Figure 5-8 shows that the synthesis of BThTz and 3,6-bisphenyl tetrazine (BPhTz), starting from the commercially available nitriles. Note that a mixture of ethanol and THF were used in the synthesis of BPhTz. This cosolvent was used to examine if the reaction would still occur in the presence of solvents more suitable for solubilizing certain aromatic nitriles that exhibit limited solubility in ethanol. Both compounds were isolated by filatration after the reaction and subsequently oxidized with isoamyl nitrite, followed by recrystallization to yield the product as a bright red solid for BThTz and a bright violet
243 solid for BPhTz. In either case the dihydro intermediate was too unstable to isolate. The compounds BBiThTz and dichlorotetrazine (DClTz) were received (Pierre Audebert, ENS Cachan, France) and were synthesized according to the literature. 245,251 NNNN S S NNNN S S S S NNNN S S S S NNNN S S OO OO OO OO OO OO OO OO NNNN NNNN Cl Cl NNNN S S S S BEDOT-TzB(EDOT-Th)-TzB(BiEDOT)-TzBThTzBPhTzDClTzBBiThTz Figure 5-7. Structures of the investigated tetrazine family. NNNN S S NNNN BThTzBPhTz CN S CN 22 NH2NH2 H2OS / EtOH, THF52%25%1.2. isoamyl nitritedichloromethane NH2NH2 H2OS / EtOH1.2. isoamyl nitritedichloromethane Figure 5-8. Synthesis of BThTz and BPhTz.
244 Figure 5-9 illustrates the synthetic route for BEDOT-Tz. Here the first step was the synthesis of 2-EDOT-carbonitrile. For this step a modified version of the method developed by Heldrich et al 265 for the synthesis of aryl nitriles was employed. This method is based on the generation of the aryl-lithium species followed by nucleophilic substitution on p-toluenesulfonyl cyanide (TsCN) 266 to yield the aryl nitrile. The resulting EDOT-nitrile was used directly in the next step without purification. Using standard reaction conditions, BEDOT-Tz-H 2 was isolated in 27% yield from EDOT as a bright orange solid. This compound was purified by column chromatography on silica gel and was found to be remarkably stable, showing no signs of oxidation to the tetrazine. This stability of the dihydro-tetrazine is unprecedented and points to a novel property of BEDOT-Tz relative to BThTz. S OO 1. nBuLi / THF2. TsCNS OO CN NH2NH2 H2OS / ethanol27%(over two steps)HNNNHN S S OO OO BETz-H2BETz-H2 isoamyl nitritedichloromethaneNNNN S S OO OO BETz16% Figure 5-9. Synthesis of BEDOT-Tz. To further investigate the unusual stability of BEDOT-Tz-H 2, crystals of the compound (grown by vapor diffusion with dichloromethane / ethanol) were analyzed by X-ray analysis (see appendix A and chapter 2). Figure 5-10 shows the crystal structure of the compound, which offers an explanation of the unusual stability. In the crystal structure it can be seen that the tetrazine ring is the 1,4-dihydro isomer. Analysis revealed that intramolecular hydrogen bonding occurs between the oxygens of the ethylenedioxy
245 bridge and the hydrogens on the 1 and 4 positions of the dihydro-tetrazine ring. In this case the bond lengths for the hydrogen bonds defined by H1-O3 and H2-O1 are 2.340 and 2.130 respectively. This intramolecular hydrogen bonding offers an explanation for the enhanced stability of BEDOT-Tz-H 2 relative to the dihydro intermediates in the bis-thienyl and phenyl analogues. Interestingly, in the crystal structure, intermolecular hydrogen bonding is also observed, as H1 and H2 are involved in bifurcated hydrogen bonds with N2 and S2 respectively (see Appendix A), in addition to the intramolecular hydrogen bonds described above. It can also be seen that the molecule is not planar, as the dihydro-tetrazine ring has adopted a twist-boat conformation. The torsional angle between the thiophene ring (plane defined by C10, C9, S1, C12, and C11) and the plane defined by N1, C2, and N4 is found to be 24.2. The torsional angle between the other thiophene ring (plane defined by C4, C3, S2, C6, and C5) and the plane defined by N2, C1, and N3 is 9.5. Notice that the ethylenedioxy bridge of the second EDOT shows two conformations in the crystal structure. Figure 5-10. X-ray crystal structure of BEDOT-Tz-H 2 . Dashed lines represent hydrogen bonds. Despite its enhanced stability, BEDOT-Tz-H 2 was converted to the deeply red BEDOT-Tz upon refluxing with isoamyl nitrite in dichloromethane. Suitable single crystals of BEDOT-Tz were not obtained for X-ray analysis. The conversion from
246 dihydro-tetrazine to tetrazine was confirmed by UV-visible spectroscopy, NMR, as well as high-resolution mass spectrometry (see experimental details in section 5.6). The oxidation step proceeded in relatively low yield (16%) as determined after column chromatography. This low yield could either be due to incomplete conversion, degradation during the oxidation, or degradation on the column. Several other routes were also attempted as a means of synthesizing BEDOT-Tz based on the nucleophilic aromatic substitution of 2-lithio-EDOT or the Grignard of EDOT with dichlorotetrazine or bis(3,5-dimethylpyrazol-1-yl)tetrazine. As for the case of thiophene, these routes proved unsuccessful and resulted only in the formation of tars. For the synthesis of B(EDOT-Th)-Tz via the sulfur-hydrazine route, the necessary starting compound was EThCN shown in Figure 5-11, which has not been previously reported. This compound was realized by a multi-step procedure (Figure 5-11) that began with the conversion of commercially available 5-bromo-2-thiophenecarboxaldehyde to the aldoxime (6), 267 followed by conversion to 5-bromo-2-thiophenecarbonitrile (7) by reaction with acetic anhydride. 268 This conversion proceeded in 40% yield over two steps and gave only a single isomer. S Br O H NH2OH HClpyridineethanol58%S Br N H OH 66 O O O S CN Br 769%7+S OO Sn(CH3)3 Pd(PPh3)4DMFS OO S CN 94%EThCN Figure 5-11. Synthesis of 5-EDOT-2-thiophenecarbonitrile (EThCN).
247 It should be noted that attempts to synthesize 5-bromo-2-thiophenecarbonitrile 269 or 5-iodo-2-thiophenecarbonitrile 270 by halogenation of the 5-positon of 2-thiophene-carbonitrile resulted in a mixture of inseparable isomers in all cases, as has been observed previously for the bromination of 2-thiophencarboxaldehyde. 271 Likewise several methods for the direct conversion of aromatic-carboxaldehydes to carbonitriles also proved unsuccessful. 272 After isolation of the 5-bromo-2-thiophenecarbonitrile (7), Stille coupling with trimethyltin-EDOT yielded the desired monomer precursor in 94% yield. The conversion of EThCN to B(EDOT-Th)-Tz-H 2 was then attempted using the same conditions that had proven successful for the synthesis of BEDOT-Tz-H 2 , as seen in Figure 5-12. In this case, an orange solid was isolated by filtration in 33% yield. This compound, suspected to be B(EDOT-Th)-Tz-H 2 , was found to be unstable, as would be expected based on the bis-thienyl-tetrazine structure in which intramolecular hydrogen bonding with the EDOT oxygens cannot play a role in stabilizing the compound. Upon standing in air, or even with storage under inert conditions, the compound was found to change from a bright orange color to a brownish color. Nonetheless, 1 H NMR in CDCl 3 immediately after the reaction revealed a spectrum that was consistent with the proposed structure (see experimental details in section 5.6) and indicated the absence of starting material. The absence of starting material was also confirmed by TLC. The suspected product was not sufficiently soluble in CDCl 3 to give a 13 C NMR on a time scale over which the compound was stable, as a brown precipitate began to form in the NMR tube after only a few minutes. Likewise C 6 D 6 was unsuitable and DMSO-d 6 was found to rapidly degrade the compound. The compound was then analyzed by HRMS. In this case, the mass of the suspected dihydro monomer (B(EDOT-Th)-Tz-H 2 ) was calculated to be
248 528 g/mol. Not surprisingly, no peak was observed corresponding to this product. Interestingly, a small peak was found at 526 amu, suggesting that a small amount of the final product, B(EDOT-Th)-Tz had been formed by oxidation in air. However, this brownish compound was totally insoluble in all investigated solvents and could not be characterized by NMR. Furthermore, the brownish compound became black with time. It appears that if a small amount of B(EDOT-Th)-Tz had been formed, it was either too insoluble to characterize further, or too unstable to isolate. The darkening of the sample with time could be indicative of the conversion from dihydro-tetrazine (or tetrazine) monomer to an oxidized polymeric product. Electron rich monomers are sometimes observed to spontaneously polymerize in air to yield a doped conjugated polymer. This is well known for pyrrole, which forms pyrrole-black on oxidation in air 273 and it has also been observed with dibromo-pyrrole 274 and dibromo-EDOT. 275 However, further characterization was not attempted. The controlled oxidation of the orange product that was suspected to be B(EDOT-Th)-Tz-H 2, was also attempted with isoamyl nitrite in dichloromethane. However in this case, the compound was also observed to change from bright orange to brown and precipitate from the reaction mixture. Again the compound could not be characterized or purified due to its insolubility in organic solvents. In this case, HRMS did not reveal any identifiable peaks. The controlled oxidation was then attempted by passing air over the orange compound, in order to oxidize it in the most mild of conditions. The results were the same as above and product could not be isolated. Based on these results it was deemed that B(EDOT-Th)-Tz was inherently unstable. The cause of this instability is
249 still unknown and surprising considering that compounds 3 and 5 from Figure 5-6 were both reported to be stable. S OO S CN EThCN2 NH2NH2 H2OS / ethanol33% crudeNNHNHN S S S S OO OO B(EDOT-Th)-Tz-H2[O]DegradationNNNN S S S S OO OO B(EDOT-Th)-Tz Figure 5-12. Synthetic route for B(EDOT-Th)-Tz. As an alternate route, the Stille coupling route employed by Soloducho for the synthesis of compounds 2 and 3 in Figure 5-6 was employed. Here, the converion of 3,6-(5,5-dibromo-2,2-thiophene)-tetrazine (received from Prof. Pierre Audebert, ENS Cachan, France) into B(EDOT-Th)-Tz was attempted via Stille coupling with trimethyltin-EDOT. In this case only an insoluble brown solid was obtained and no product or tetrazine starting material was recovered. It is thus concluded that B(EDOT-Th)-Tz-H 2 is the suspected product of the reaction of EthCN with hydrazine and sulfur, although this compound is found to be inherently unstable. Conversion to B(EDOT-Th)-Tz has proven elusive, as oxidation in air or with isomyl nitrite results in the degradation of the compound to an insoluble product that has not been identified. Based on the fact that this product could not be accessed via Stille coupling, it is suspected that B(EDOT-Th)-Tz is inherently unstable, although the precise cause of this instability is not known.
250 As a final compound for investigation, B(BiEDOT)-Tz was targeted as seen in Figure 5-13. Here BiEDOT-CN was prepared from BiEDOT via reaction of lithio-BiEDOT with tosyl-cyanide in 65% yield. The attempted conversion to the dihydrotetrazine was performed using the sufur-hydrazine route, however in this case a cosolvent of ethanol and THF (1/1) had to be used to fully solubilize BiEDOT-CN. The use of cosolvents had proven effective in the synthesis of BPhTz as previously discussed. Upon reaction, a bright orange solid was isolated by filtration with a crude yield of 55%. This compound was found to be only sparingly soluble in CDCl 3 and while the product could be dissolved in DMSO-d 6 with heating, the solution rapidly darkened with the formation of a black precipitate. Analysis by 1 H NMR in CDCl 3 revealed a mixture that appeared to contain the suspected product (see experimental details in section 5.6). However, for this compound, HRMS revealed only a preponderance of starting material. Nonetheless, this does not prove that product was not formed, as aryl-nitrile is known to be a primary fragmentation product of tetrazines. 276 No hard evidence could be gathered that this compound was definitely the dihydrotetrazine, as IR, elemental analysis, HRMS, and 13 C NMR failed to support the structure. Only 1 H NMR suggests that the product was formed, but it also shows that if it was formed, it was not pure. Attempts to purify this product by column chromatography or recrystallization proved fruitless. It is possible that the use of THF and ethanol as cosolvents in the synthesis of B(BiEDOT)-Tz-H 2 resulted in the formation of an unidentified side product, but this does not seem likely as BPhTz was successfully synthesized in this manner. This suspected dihydro product was only found to be stable when maintained under inert conditions. Oxidation with air or isoamyl nitrite resulted in the decomposition of the compound and no tetrazine is isolated.
251 NNNN S S S S OO OO OO OO B(BiEDOT)-TzNNHNHN S S S S OO OO OO OO B(BiEDOT)-Tz-H2S S OO OO S S OO OO 1. nBuLi / THF2. TsCN NH2NH2 H2OS / EtOH, THF55% crude[O]Degradation65%BiEDOT-CNCN Figure 5-13. Synthesis of B(BiEDOT)-Tz. As a control experiment, BiEDOT was dissolved in dichloromethane under argon and a small amount of isoamyl nitrite was added. The solution immediately darkened and precipitate formed. It was clear that isoamyl nitrite, although a mild oxidizing agent, is too strong an oxidizing agent to be used in the presence of electron rich compounds like BiEDOT. This may serve to explain why isoamyl nitrite could not be used for the controlled oxidation of B(EDOT-Th)-Tz-H 2 or B(BiEDOT)-Tz-H 2 . However these compounds were also found to be unstable to air oxidation. It appears as though there is some inherent instability upon oxidation of the dihydrotetrazine. It may be due to the oxidation of the electron rich donors or it might be due to an instability of the resulting tetrazine compound. Such instability could be due to the very strong donor-acceptor interaction, which may lead to a ground-state charge transfer. The reason for the instability remains unclear. As such, of the targeted EDOT-tetrazine compounds shown in Figure 5-7, only BEDOT-Tz was isolated.
252 5.3 Electronic and Electrochemical Properties of Tetrazine Compounds The electronic properties of the obtained compounds were investigated by UV-visible spectroscopy in dichloromethane solution as seen in Figure 5-14. All of the compounds displayed a very intense absorption band that corresponds to the * transition. In the case of DClTz and BPhTz, this * maximum is located in the UV at 298 nm. With an increasing donor strength, the * transition is observed to red shift, with BThTz, BEDOT-Tz, and BBiThTz showing maxima of 347, 367, and 425 nm respectively. The shift in absorbance from 298 nm with BPhTz to 347 nm with BThTz is indicative of the enhanced donor nature of thiophene relative to benzene. The red shift in the * absorption is thus a strong indication of the strength of the donor-acceptor interaction between the donor heterocycle and the tetrazine ring. The other interesting comparison to note is the shift in the * transition when comparing BEDOT-Tz-H 2 (294 nm) and BEDOT-Tz (367 nm). This red shift of 73 nm is indicative of the conversion of the non-planar, non-aromatic dihydro compound into the aromatic BEDOT-Tz. The other absorption band that is observed in the tetrazine compounds is centered around 530 nm and corresponds to the n-* transition (see Figure 5-14 inset), which is responsible for the color in these tetrazine compounds. Notice that while the * transition for all of the compounds is characterized by a molar absorptivity of 10 4 L mol -1 cm -1 , the n-* transition is far weaker, with molar absorptivities on the order of 10 2 L mol -1 cm -1 . This behavior is known to be typical for tetrazine compounds, as tetrazines generally absorb in the range of 520-570 nm with a molar absorptivity of a few hundred. In this case, BPhTz shows an absorption at 551nm, whereas BThTz, BEDOT-Tz, and
253 DClTz show absorbances at 532, 535, and 533 nm respectively. The n-* absorption of BBiThTz is not well defined and is probably obscured by the strong * transition at 425 nm. It is interesting to note that BEDOT-Tz-H 2 shows no absorbance at wavelengths longer than 450 nm, indicative of the fact that the dihydro-tetrazine has not been converted into the tetrazine. 3004506007500.00.20.40.60.81.01.21.4 4505005506006500.000.010.020.030.04 Absorbance (a. u.)Wavelength (nm) DClTz BBiThTz BEDOT-Tz BPhTz BThTz BEDOT-Tz-H2 Absorbance (a. u.)Wavelength (nm) Figure 5-14. Solution absorbance spectra of tetrazine compounds in dichloromethane. The electrochemical properties of the tetrazine compounds were also investigated in solution. The solution reduction of each of the compounds was investigated in 0.1 M TBAP / acetonitrile under oxygen and water free conditions in order to assess the acceptor strength of the various compounds. As discussed previously, the reduction value of a monomer is often found to correlate with the reduction value of the corresponding
254 polymer. Figure 5-15 summarizes the DPV results for the reduction of the investigated compounds. Figure 5-15. Solution reduction of tetrazine compounds by DPV. Measurements were performed in 0.1 M TBAP / acetonitrile, on a Pt button working electrode with a Ag wire pseudo-refernence electrode calibrated vs. Fc/Fc + . (a) DClTz, (b) BPhTz, (c) BThTz, (d) BBiThTz, (e) BEDOT-Tz. It can be seen in Figure 5-15a that DClTz shows an E 1/2 of .90 V vs. Fc/Fc + . Recall from Figure 5.1 that the reduction potential for Tetrazine itself is .19 V vs. Fc/Fc + . This indicates that DClTz is easier to reduce than tetrazine, based on the electron withdrawing effect of the chlorine atoms. All the other compounds in Figure 5-15 show a
255 reduction between .1 and .2 V. Essentially the same reduction potential is observed for BPhTz (Figure 5-15b) and BThTz (Figure 5-15c) at .15 V, while BBiThTz (Figure 5-15d) is slightly easier to reduce (-1.10 V) and BEDOT-Tz (-1.20 V) (Figure 5-15e) is slightly more difficult to reduce. This behavior is not surprising as the extended conjugated system of BBiThTz relative to BThTz is expected to make the compound more likely to delocalize a negative charge, while the strong electron rich character of EDOT is expected to make the BEDOT-Tz slightly more difficult to reduce that BThTz. The value for the reduction of BThTz has been reported to be .2 V so the value of .15 V measured here is in reasonable agreement. Note that BEDOT-Tz-H 2 was not observed to show a reduction. Electropolymerization of the tetrazine compounds was also attempted. Table 5-1 reports the peak oxidation potentials of the compounds as measured by CV in 0.1 M TBAP / acetonitrile. The peak oxidation potential of DClTz is reported as a reference. With the exception of BBiThTz, which has been reported to electropolymerize, none of the compounds were observed to polymerize. This result was not surprising for BPhTz, as phenylene monomers generally are not observed to electropolymerize efficiently, or for BThTz, which has been previously reported not to polymerize. 259 For BEDOT-Tz, repeated potential cycling, potentiostatic, and galvanostatic deposition were attempted in a variety of solvent/electrolyte mixtures. In all cases, no electropolymerization was observed. For BEDOT-Tz-H 2 , it is interesting to note that the peak monomer oxidation potential is exactly the same as for EDOT. However, no polymerization of BEDOT-Tz-H 2 was observed.
256 Table 5-1. Peak oxidation potentials of tetrazine compounds. Compound DClTz BPhTz BThTz BBiThTz BEDOT-Tz BEDOT-Tz-H 2 E (V) vs. Fc/Fc + 1.7 1.7 1.4 1.3 1.3 1.1 Measurements were performed in 0.1 M TBAP / acetonitrile. As an alternative route, the polymerization of BEDOT-Tz and BThTz were also attempted utilizing BF 3 (Et 2 O), according to the method that was recently published. 277 It has been found that electropolymerization of high oxidation potential monomers occurs more efficiently in the presence of the strong Lewis acid. However, both tetrazine compounds were found to be unstable in BF 3 (Et 2 O) and black solutions resulted. Ultimately the question of interest is why BThTz and especially BEDOT-Tz are not observed to electropolymerize. Monomers based on EDOT and pyridine as well as EDOT and pyridopyrazine (see Figure 5-3) are observed to efficiently electropolymerize. Additionaly, BBiThTz is known to polymerize. As such it appears that there is no inherent reason why tetrazine monomers should not electropolymerize (e.g. instability). Perhaps a situation analogous to the polythiophene paradox is to blame. However, at the present time the question remains open. In order to answer this question theoretical calculations could offer insight, but the successful synthesis of a homologous family of tetrazine monomers could ultimately give a more satisfying answer. 5.4 Conclusion and Outlook In this chapter, routes toward the synthesis of donor-acceptor monomers based on EDOT as the donor and tetrazine as the acceptor were explored. The ultimate goal of achieving electropolymerizable monomers was not achieved. Synthetic difficulties and the inability to polymerize BEDOT-Tz, prevented the realization of this goal. One aspect
257 brought out in this work is that tetrazine is so chemically distinct from other N-heterocycles used in conjugated polymer synthesis that it is a mistake to assume that it can be utilized in an analogous way. It should be stressed here that tetrazine does possess characteristic reactivity and structural properties that set it apart from other N-heterocycles. First, crystal structures of tetrazine reveal a planar structure with C-N bonds that are 1.334 and N-N bonds that are 1.321 . Additionally, the N-C-N bond angles are found to be 127.2 0 and the C-N-N bond angles are found to be 115.6 0 . From this structural information it can be seen that tetrazine does not possess the perfectly symmetrical and fully delocalized electronic structure of benzene. Calculations suggest that tetrazine may have a resonance stabilization energy as high as 40 kcal/mol, which supports the aromatic structure assigned to the compound. Nonetheless, tetrazine shows reactivity that is not typical for aromatic compounds. Tetrazines are well known to undergo Diels-Alder reactions as well as nucleophilic substitution on the carbons or azaphilic addition to the nitrogens. 278 Further tetrazines are unstable in the presence of acids and bases, as well as carbon nucleophiles and organometallic reagents, as discussed earlier. In fact, tetrazine itself is so unstable that it must be stored under inert conditions, whereas 3,6-disubstituted tetrazines are often stable. As a final note on the reactivity of tetrazines, these compounds are often highly unstable and are known to violently decompose. This has led to the use of these compounds for the development of high-energy materials. These factors make the pursuit of tetrazine containing conjugated polymers a challenging, yet interesting, goal.
258 The work presented here suggests several areas for future work with DA tetrazine monomers. While EDOT is a strong donor and a commercially available compound, it is less attractive than ProDOT for use in DA compounds for several reasons. From a practical standpoint, dialkyl-ProDOTs impart greater solubility. For BEDOT-Tz, the low solubility limited the concentration of the monomer that could be used in electrochemical experiments and ultimately played a role in the inability to obtain single crystals of the compound for X-ray analysis. Perhaps more importantly, ProDOT imparts greater air stability to donor acceptor monomers and polymers, as discussed in chapter 3. As a potential target monomer, bis-dialkyl-ProDOT-tetrazine (BProDOT-R 2 -Tz) provides an interesting candidate. Even if the monomer could not be electrochemically polymerized, the potential for a chemical polymerization exists due to the presence of solubilizing alkyl groups on the monomer. Bromination of BProDOT-R 2 -Tz or bromination of the aryl nitrile prior to tetrazine ring formation should provide a suitable monomer for chemical polymerization. The Yamamoto coupling polymerization (chapter 3) provides the mild conditions that should be most applicable in the presence of the reactive tetrazine ring. The use of ProDOT could also lead to stable, isolable analogues of B(EDOT-Th)-Tz and B(BiEDOT)-Tz using the chemistry developed in this chapter. While alkylthiophenes do not possess the donor strength of ProDOT or EDOT, based on the precedent with BBiThTz, alkylthiophenes could provide an initial route to stable, soluble tetrazine polymers. 5.5 Closing Statement As conjugated (conducting) polymers approach their 30 th year, the field that has grown around them continues to expand at an ever-increasing rate. This interdisciplinary research effort has reached the point where conjugated polymers are being investigated in
259 industry and academia for use in real world applications. In this work, the optoelectronic properties of conjugated polymers were the primary application-based focus. It should be clear from chapter 1 that a tremendous research effort has been aimed at the development of polymeric solar cells. Understanding the physics behind these devices and mastering their construction is an endeavor that could possibly bring great rewards. However, it should also be clear form reading chapter 1, 3, and 4 that it is chemistry that provides the foundation upon which polymer solar cells are built. The organic basis of conjugated polymers is the real strength of this class of materials. The ability to tailor the organic structure of the polymer in a controlled way, offers the most powerful tool for optimizing the performance of any device or process based on a conjugated polymer. Beyond the development of solar cells, or electrochromic devices, it the task of the chemist to develop a firm, fundamental understanding of the structure-property relationships in conjugated polymers. Only with this understanding can polymers be optimized for use in an application. This fundamental chemical approach has led to the development of the donor-acceptor polymers presented in chapter 3 and 5. Although success has not yet been achieved in the development of dioxythiophene-tetrazine polymers, the work presented in chapter 5 illustrates the critical role of the polymer chemist: the synthesis of new polymers and the development of structure-property relationships. 5.6 Synthetic Procedures 3,6-di(thiophen-2-yl)-1,2,4,5-tetrazine (BThTz or 4). In 3mL of ethanol, 1.05 g (9.6 mmol) of thiophene-2-carbonitrile and 0.2 g of powdered sulfur (6 mmol in S) were dissolved. Then 1.5 mL of hydrazine hydrate was added (31 mmol). The mixture was then heated to 90C in a pressure tube (Aldrich 38 mL pressure tube with threaded plug)
260 for 2 h. During this time the reaction mixture turned black, followed by the formation of an orange precipitate. The reaction was then cooled to room temperature and carefully vented to release the H 2 S that was generated during the course of the reaction. The reaction mixture was then diluted with 35 mL of dichloromethane and 1 mL (~7mmol) of isoamyl nitrite was added. The reaction was then stirred for 10 minutes at room temperature (in air). During this time the solution was observed to change from brown orange to bright red. The mixture was then filtered and the filtrate was isolated. The dichloromethane was removed under vacuum and the crude solid was recrystallized from ethanol to give 0.609 g (52%) of the product as long red needles. mp dec. 172C. 1 H NMR (300 MHz CDCl 3 ) 8.29 (dd, 2H), 7.69 (dd, 2H), 7.28 (dd, 2H). IR (KBr cm -1 ) 3099, 1834, 1535, 1443, 1385, 1279, 1083, 1073, 1003, 913, 851, 717, 607, 549. UV-Vis (CH 2 Cl 2 ) max 347 nm (1.1 x 10 4 L mol -1 cm -1 ), 532 nm (150 L mol -1 cm -1 ). 3,6-diphenyl-1,2,4,5-tetrazine (BPhTz). 279 In a mixture of 2.5 mL of THF and 2.5 mL of ethanol was added 1.40 g (13.6 mmol) of bezonitrile and 0.21 g of sulfur powder (6.8 mmoles in S). Then 2 mL of hydrazine hydrate (~40 mmol) was added and the reaction mixture was heated to 100C for 1 h. The reaction mixture was cooled to room temperature and vented to release H 2 S. The reaction mixture was taken up in 75 mL of dichloromethane and 1 mL of isoamyl nitrite was added and the mixture was stirred for 15 minutes. Dichloromethane was removed under vacuum and the crude product was recrystallized from ethanol to give 0.40 g (25 %) of violet plates. mp 176-178C. 1 H NMR (300 MHz CDCl 3 ) 8.67 (m, 4H), 7.63 (m, 6H). HRMS calcd for C 14 H 10 N 4 (M + ), 234.0905; found 234.0902. IR (KBr cm -1 ) 3071, 1600, 1457, 1392, 1309, 1188, 1104,
261 1052, 1023, 919, 774, 689, 589. UV-Vis (CH 2 Cl 2 ) max 298 nm (1.3 x 10 4 L mol -1 cm -1 ), 551 nm (230 L mol -1 cm -1 ). 3,6-bis(2,3-dihydrothieno[3,4-b][1,4]dioxin-5-yl)-1,4-dihydro-1,2,4,5-tetrazine (BEDOT-Tz-H 2 ). In 50 mL of dry THF, 1.90 g (13.4 mmol) of EDOT was dissolved and the solution was cooled to C. Then 6 mL of 2.6 M n-butyllithium (16 mmol, 1.2 equiv) was added dropwsie and the reaction was allowed to stir for 1 h at C. Then the lithio-EDOT solution was transferred via cannula to a solution containing 4.80 g (26.5 mmol, 2 equiv) of p-toluenesulfonyl cyanide in 50 mL of THF at C. After stirring for 15 minutes at C, the reaction was allowed to warm to room temperature and 15 mL of concentrated ammonium hydroxide was added. The reaction was stirred for 15 additional minutes and then poured into 1 M NaOH. Extraction with ether followed by drying over MgSO 4 and removal of ether yielded the crude 2-EDOT-carbonitrile (EDOT-CN), which was used directly with no further purification. The crude EDOT-CN was then dissolved in 5 mL of ethanol and 0.21 g (6.7 mmol of S) of sulfur powder was added, along with 2 mL of hydrazine hydrate (41 mmol) and the reaction was heated to reflux for 1.5 hours. The reaction mixture was then filtered and the orange solid was washed with ethanol. Column chromatography on silica (dichloromethane) gave the product as an orange solid (0.658 g, 27%). mp = dec. 225C. 1 H NMR (300 MHz CDCl 3 ) 7.87 (s, 2H), 6.38 (s, 2H), 4.27 (m, 8H). 13 C NMR (75 MHz, CDCl 3 ) . HRMS calcd for C 14 H 12 N 4 O 4 S 2 (M + ), 364.0300; found 364.0303. UV-Vis (CH 2 Cl 2 ) max 294 nm (1.8 x 10 4 L mol -1 cm -1 ). IR (KBr, cm -1 ) 2924, 2868, 1587, 1534, 1497, 1438, 1363, 1167, 1085, 1071, 907, 882, 799, 721. Anal. calcd for
262 C 14 H 12 N 4 O 4 S 2 : C, 46.14; H, 3.32; N, 15.38; O, 17.56; S, 17.60. Found: C, 44.84; H, 3.37; N, 13.52. 3,6-bis(2,3-dihydrothieno[3,4-b][1,4]dioxin-5-yl)-1,2,4,5-tetrazine (BEDOT-Tz). In 100 mL of dichloromethane, 0.518 g of BEDOT-Tz-H 2 was dissolved along with 5 mL of isoamyl nitrite. The reaction was then heated to reflux in the presence of air for 12 h. During this time the reaction mixture changed from orange to deep red. The reaction was then cooled and the dichloromethane was removed. Chromatography on silica gel (dichloromethane) gave the product as a bright red solid (0.082 g, 16%). mp dec. >220C. 1 H NMR (300 MHz CDCl 3 ) 6.70 (s, 2H), 4.46 (m, 4H), 4.31 (m, 4H). 13 C NMR (75 MHz, CDCl 3 ) 165.80, 159.15, 142.60, 106.75, 104.56, 65.38, 64.28. HRMS calcd for C 14 H 10 N 4 O 4 S 2 (M + ), 362.0143; found 362.0140. UV-Vis (CH 2 Cl 2 ) max 367 nm (1.9 x 10 4 L mol -1 cm -1 ), 535 (220 L mol -1 cm -1 ). IR (KBr, cm -1 ) 3101, 2921, 1580, 1497, 1466, 1352, 1367, 1280, 1184, 1157, 1061, 1030, 940, 909, 760, 702, 584. Anal. calcd for C 14 H 10 N 4 O 4 S 2 : C, 46.40; H, 2.78; N, 15.46; O, 17.66; S, 17.70. Found: C, 46.29; H, 2.45; N, 14.14; S, 16.24. 5-bromothiophene-2-carbaldehyde oxime (6). To 25 mL of absolute ethanol, 2 mL of pyridine (25 mmol), 1.22 g (17.6 mmol) of hydroxylamine hydrochloride and 3.023 g (15.8 mmol) of 5-bromothiophene-2-carbaldehyde were added. The reaction was then heated to reflux under argon for 1 h. The reaction was then cooled to room temperature and poured into 30 mL of water, and the resulting precipitate was isolated by filtration. The solid was then recrystallized from water to 1.901 g (58%) of product as a white solid. 1 H NMR (300 MHz CDCl 3 ) 8.17 (b, 1H), 7.61 (s, 1H), 7.11 (d, 1H), 7.07 (d, 1H).
263 5-bromothiophene-2-carbonitrile (7). In a three neck flask, 3 mL of acetic anhydride (32 mmol) and 1.900 g (9.22 mmol) of 5-bromothiophene-2-carbaldehyde oxime were combined and heated to ~145C for 1 h. The solution was then cooled and 10 mL of water was added. The reaction was then heated to reflux for three minutes and then cooled to room temperature. The reaction mixture was neutralized with concentrated NaOH and subsequently extracted with ether. The ether layer was dried with MgSO 4 and the ether was removed under vacuum. Chromatography on silica gel with 1:1 hexanes and dichloromethane gave 1.200 g (69%) of the product as a pale yellow oil. 1 H NMR (300 MHz CDCl 3 ) 7.39 (d, 1H), 7.10 (d, 1H). [Lit. 7.38 (d, 1H), 7.08 (d, 1H)]. 5-(2,3-dihydrothieno[3,4-b][1,4]dioxin-5-yl)thiophene-2-carbonitrile (EThCN). In 100 mL of dry DMF, 1.20 g (6.4 mmol) of 5-bromothiophene-2-carbonitrile was dissolved along with 2.90 g (9.1 mmol, 1.4 equiv) of trimethyltin-EDOT, and 0.35 g (0.30 mmol, ~5 mol%) of Pd(PPh 3 ) 4 . The reaction was heated to 110-115C overnight. The reaction was then cooled to room temperature and poured into brine and subsequently extracted with ether. The ether layer was then washed with water and dried with MgSO 4 . Removal of ether followed by chromatography on silica (2:3 dichloromethane and hexanes) gave the product as a yellow crystalline solid (1.49 g, 94%). mp 133-134C. 1 H NMR (300 MHz CDCl 3 ) 7.50 (d, 1H), 7.12 (d, 1H), 6.36 (s, 1H), 4.39 (m, 2H), 4.27 (m, 2H). 13 C NMR (75 MHz, CDCl 3 ) 142.43, 142.07, 137.63, 121.88, 114.96, 110.60, 106.56, 99.91, 65.49, 64.68. HRMS calcd for C 11 H 7 NO 2 S 2 (M + ), 248.9918; found 248.9914. Anal. calcd for C 11 H 7 NO 2 S 2 : C, 52.99; H, 2.83; N, 5.62; O, 12.84; S, 25.72. Found: C, 52.75; H, 2.73; N, 5.33.
264 3,6-bis(5-(2,3-dihydrothieno[3,4-b][1,4]dioxin-5-yl)thiophen-2-yl)-1,4-dihydro-1,2,4,5-tetrazine (B(EDOT-Th)-Tz-H 2 ). In 4 mL of ethanol, 0.594 g (2.4 mmol) of EThCN, 0.10 g (3 mmol) of powdered sulfur, and 1.5 mL (31 mmol) of hydrazine hydrate were dissolved and heated to 90-95C in a sealed pressure tube for 1.5 h. During this time, the solution turned black and an orange precipitate formed. The reaction was then cooled and the mixture was carefully vented to remove H 2 S that had formed during the reaction. The reaction mixture was then filtered and an orange solid was isolated (0.210 g, 33%). 1 H NMR (300 MHz CDCl 3 ) 7.17 (d, 2H), 7.13 (d, 2H), 6.99 (s, 2H), 6.29 (s, 2H), 4.35 (d, 4H), 4.26 (d, 4H). IR (KBr cm -1 ) 3107, 2926, 1540, 1492, 1460, 1392, 1366, 1171, 1069, 1017, 968, 908, 804, 681, 604, 579. HRMS calcd for C 22 H 16 N 4 O 4 S 4 (M + ), 528.0054; Found 526.0030. 7-(2,3-dihydrothieno[3,4-b][1,4]dioxin-5-yl)-2,3-dihydrothieno[3,4-b][1,4]dioxine-5-carbonitrile (BiEDOT-CN). In 120 mL of dry THF, 3.98 g (14.1 mmol) of BiEDOT was dissolved and cooled to C. Then 6.2 mL of n-butyllithium was added via syringe and the solution was stirred at this temperature for 45 minutes. This solution was then transferred via cannula to a solution of 4.98 g (27.5 mmol, 1.95 equiv) of p-toluenesulfonic acid in 70 mL of THF at C. After 1.5 hours at C, 18 mL of NH 4 OH was added and the reaction was stirred for 15 minutes and then poured into 250 mL of 1 M NaOH. Extraction with dichlormethane and drying with MgSO 4 , gave the crude product. Chromatography on silica (2:1 dichloromethane and hexanes) gave the product as a light yellow solid (2.82 g, 65%). mp 252-254C. 1 H NMR (300 MHz CDCl 3 ) 6.41 (s, 1H), 4.37 (m, 6H), 4.26 (m, 2H). 13 C NMR (75 MHz, CDCl 3 ) 152.74,149.00, 141.42, 139.52, 135.11, 113.55, 109.99, 100.86, 82.00, 65.55, 65.52,
265 64.88, 64.74. HRMS calcd for C 13 H 9 NO 4 S 2 (M + ), 306.9973; found 306.9989. Anal. calcd for C 13 H 9 NO 4 S 2 : C, 50.80; H, 2.95; N, 4.56; O, 20.82; S, 20.87. Found: C, 50.24; H, 2.76; N, 4.41. 3,6-bis(7-(2,3-dihydrothieno[3,4-b][1,4]dioxin-5-yl)-2,3-dihydrothieno[3,4-b][1,4]dioxin-5-yl)-1,4-dihydro-1,2,4,5-tetrazine (B(BiEDOT)-Tz-H 2 ). In 7.5 mL of THF and 7.5 mL of ethanol, 0.460 g of BiEDOT-CN, 0.040 g (1.3 mmol) of sulfur were dissolved. Then 1.5 mL (31 mmol) of hydrazine hydrate was added and the reaction mixture was heated to 95C for 4 h. The reaction mixture was then cooled to room temperature and carefully vented to remove H 2 S that had formed during the reaction, and a bright orange solid was isolated by filtration (0.266 g, 55%). 1 H NMR (300 MHz CDCl 3 ) 9.5 (b, 2H), 6.39 (s, 2H), 6.31 (s, 0.5H), 4.88 (b, 4H), 4.46-4.3 (bm, 14H), 4.25 (m, 4H).
APPENDIX A CRYSTALLOGRAPHIC INFORMATION FOR COMPOUNDS Figure A-1. Crystal structure for BEDOT-Tz-H 2 . Figure A-2. Representation of intramolecular and intermolecular hydrogen bonding in BEDOT-Tz-H 2 . 266
267 R1 = (||Fo| |Fc||) / |Fo| wR2 = [w(Fo2 Fc2)2] / wFo22]]1/2 S = [w(Fo2 Fc2)2] / (n-p)]1/2 w= 1/[2(Fo2)+(m*p) 2 +n*p], p = [max(Fo2,0)+ 2* Fc2]/3, m & n are constants. Crystal data and structure refinement for BEDOT-Tz-H 2 . Empirical formula C14 H12 N4 O4 S2 Formula weight 364.40 Temperature 173(2) K Wavelength 0.71073 Crystal system Monoclinic Space group P2(1)/n Unit cell dimensions a = 9.5275(5) = 90. b = 11.4383(6) = 99.231(1). c = 13.7106(7) = 90. Volume 1474.81(13) 3 Z 4 Density (calculated) 1.641 Mg/m3 Absorption coefficient 0.391 mm-1 F(000) 752 Crystal size 0.21 x 0.11 x 0.09 mm3 Theta range for data collection 2.33 to 27.49.
268 Index ranges -10h12, -14k14, -13l17 Reflections collected 9397 Independent reflections 3323 [R(int) = 0.0300] Completeness to theta = 27.49 98.2 % Absorption correction Integration Max. and min. transmission 0.9745 and 0.9336 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 3323 / 0 / 234 Goodness-of-fit on F2 1.254 Final R indices [I>2sigma(I)] R1 = 0.0298, wR2 = 0.0859  R indices (all data) R1 = 0.0381, wR2 = 0.0886 Largest diff. peak and hole 0.292 and -0.212 e.-3 Table A-1. Atomic coordinates ( x 104) and equivalent isotropic displacement parameters (2x 103)for BEDOT-TZ-H 2 . U(eq) is defined as one third of the trace of the orthogonalized Uij tensor. _______________________________________________________________________________ Atom x y z U(eq) __________________________________________________________________ S1 10056(1) 6860(1) -982(1) 33(1) S2 5561(1) 4524(1) 2879(1) 37(1) N1 6562(1) 5762(1) 7(1) 32(1) N2 6244(1) 5337(1) 931(1) 33(1) N3 7625(1) 6961(1) 1490(1) 32(1) N4 8560(1) 6762(1) 784(1) 30(1) O1 8216(1) 7071(1) 3601(1) 34(1) O2 7632(1) 5803(1) 5316(1) 48(1) __________________________________________________________________
269 Table A-1. Continued _______________________________________________________________________________ Atom x y z U(eq) __________________________________________________________________ O3 7539(1) 4185(1) -1502(1) 35(1) O4 9498(1) 4368(1) -2913(1) 40(1) C1 6817(1) 5979(1) 1655(1) 28(1) C2 7949(1) 6158(1) 45(1) 27(1) C3 6652(1) 5675(1) 2657(1) 29(1) C4 7333(2) 6127(1) 3530(1) 30(1) C5 7019(2) 5515(1) 4376(1) 35(1) C6 6069(2) 4634(1) 4132(1) 41(1) C7 8711(3) 7483(2) 4617(2) 36(1) C8 8973(3) 6470(2) 5303(2) 46(1) C7' 9176(8) 6932(8) 4510(5) 42(2) C8' 8323(10) 6813(7) 5348(6) 44(2) C9 8691(1) 5938(1) -785(1) 27(1) C10 8507(1) 5070(1) -1478(1) 28(1) C11 9464(2) 5159(1) -2177(1) 31(1) C12 10363(2) 6083(1) -1998(1) 34(1) C13 7875(2) 3215(1) -2107(1) 44(1) C14 8241(2) 3646(2) -3068(1) 46(1) __________________________________________________________________ Table A-2. Bond lengths  for BEDOT-Tz-H 2 . __________________________________________________________________ Bond Length __________________________________________________________________ S1-C12 1.7166(16) S1-C9 1.7285(14) S2-C6 1.7133(17) S2-C3 1.7343(14) N1-C2 1.3900(17) N1-N2 1.4334(17) __________________________________________________________________
270 Table A-2. Continued __________________________________________________________________ Bond Length __________________________________________________________________ N1-H1 0.882(16) N2-C1 1.2840(18) N3-C1 1.4008(18) N3-N4 1.4354(17) N3-H2 0.884(19) N4-C2 1.2853(18) O1-C4 1.3623(17) O1-C7' 1.432(7) O1-C7 1.475(2) O2-C8' 1.326(7) O2-C5 1.3673(19) O2-C8 1.490(3) O3-C10 1.3661(16) O3-C13 1.4517(19) O4-C11 1.3597(17) O4-C14 1.4420(19) C1-C3 1.450(2) C2-C9 1.456(2) C3-C4 1.367(2) C4-C5 1.428(2) C5-C6 1.360(2) C6-H6A 0.9500 C7-C8 1.488(4) C7-H7A 0.9900 C7-H7B 0.9900 C8-H8A 0.9900 C8-H8B 0.9900 C7'-C8' 1.515(11) C7'-H7'A 0.9900 C7'-H7'B 0.9900 C8'-H8'A 0.9900 ___________________________________________________________________
271 Table A-2. Continued __________________________________________________________________ Bond Length __________________________________________________________________ C8'-H8'B 0.9900 C9-C10 1.366(2) C10-C11 1.428(2) C11-C12 1.358(2) C12-H12A 0.9500 C13-C14 1.499(2) C13-H13A 0.9900 C13-H13B 0.9900 C14-H14A 0.9900 C14-H14B 0.9900 __________________________________________________________________ Table A-3. Bond angles  for BEDOT-Tz-H 2 . ________________________________________________________ Bond Angle  ________________________________________________________ C12-S1-C9 92.36(7) C6-S2-C3 92.22(8) C2-N1-N2 114.25(11) C2-N1-H1 114.8(10) N2-N1-H1 109.3(10) C1-N2-N1 111.52(12) C1-N3-N4 113.46(11) C1-N3-H2 112.7(12) N4-N3-H2 111.4(12) C2-N4-N3 111.41(12) C4-O1-C7' 106.0(3) C4-O1-C7 114.90(13) C7'-O1-C7 31.8(3) C8'-O2-C5 112.0(3) ________________________________________________________
272 Table A-3. Continued ______________________________________________________ Bond Angle  ______________________________________________________ C8'-O2-C8 29.9(4) C5-O2-C8 110.44(14) C10-O3-C13 111.75(12) C11-O4-C14 111.84(11) N2-C1-N3 120.62(14) N2-C1-C3 119.97(13) N3-C1-C3 119.40(13) N4-C2-N1 120.87(13) N4-C2-C9 119.60(13) N1-C2-C9 119.45(12) C4-C3-C1 129.07(13) C4-C3-S2 110.34(11) C1-C3-S2 120.43(11) O1-C4-C3 124.27(13) O1-C4-C5 122.52(14) C3-C4-C5 113.21(13) C6-C5-O2 125.28(14) C6-C5-C4 112.38(14) O2-C5-C4 122.34(14) C5-C6-S2 111.79(12) C5-C6-H6A 124.1 S2-C6-H6A 124.1 O1-C7-C8 110.1(2) O1-C7-H7A 109.6 C8-C7-H7A 109.6 O1-C7-H7B 109.6 C8-C7-H7B 109.6 H7A-C7-H7B 108.2 C7-C8-O2 110.4(2) C7-C8-H8A 109.6 ___________________________________________________
273 Table A-3. Continued ________________________________________________________ Bond Angle  ________________________________________________________ O2-C8-H8A 109.6 C7-C8-H8B 109.6 O2-C8-H8B 109.6 H8A-C8-H8B 108.1 O1-C7'-C8' 108.9(6) O1-C7'-H7'A 109.9 C8'-C7'-H7'A 109.9 O1-C7'-H7'B 109.9 C8'-C7'-H7'B 109.9 H7'A-C7'-H7'B 108.3 O2-C8'-C7' 112.2(7) O2-C8'-H8'A 109.2 C7'-C8'-H8'A 109.2 O2-C8'-H8'B 109.2 C7'-C8'-H8'B 109.2 H8'A-C8'-H8'B 107.9 C10-C9-C2 130.23(13) C10-C9-S1 110.54(10) C2-C9-S1 119.23(10) C9-C10-O3 124.47(13) C9-C10-C11 112.99(12) O3-C10-C11 122.52(13) C12-C11-O4 124.22(13) C12-C11-C10 112.73(13) O4-C11-C10 123.00(13) C11-C12-S1 111.38(11) C11-C12-H12A 124.3 S1-C12-H12A 124.3 O3-C13-C14 110.75(13) O3-C13-H13A 109.5 ______________________________________________________
274 Table A-3. Continued ________________________________________________________ Bond Angle  ________________________________________________________ C14-C13-H13A 109.5 O3-C13-H13B 109.5 C14-C13-H13B 109.5 H13A-C13-H13B 108.1 O4-C14-C13 111.34(14) O4-C14-H14A 109.4 C13-C14-H14A 109.4 O4-C14-H14B 109.4 C13-C14-H14B 109.4 H14A-C14-H14B 108.0 ____________________________________________________________ Table A-4. Anisotropic displacement parameters (2x 103) for BEDOT-Tz-H 2 . The anisotropicdisplacement factor exponent takes the form: -22[ h2 a*2U11 + ... + 2 h k a* b* U12 ]. ________________________________________________________________________ U11 U22 U33 U23 U13 U12 ________________________________________________________________________ S1 30(1) 33(1) 37(1) 0(1) 4(1) -9(1) S2 38(1) 29(1) 46(1) 2(1) 12(1) -5(1) N1 25(1) 39(1) 30(1) 0(1) -1(1) -7(1) N2 27(1) 37(1) 33(1) 0(1) 4(1) -7(1) N3 36(1) 29(1) 32(1) -2(1) 6(1) -7(1) N4 29(1) 30(1) 31(1) 1(1) 3(1) -6(1) O1 37(1) 34(1) 32(1) -3(1) 4(1) -6(1) O2 66(1) 46(1) 32(1) 4(1) 9(1) 1(1) O3 35(1) 30(1) 39(1) -2(1) 3(1) -9(1) O4 39(1) 39(1) 42(1) -8(1) 8(1) -1(1) C1 22(1) 27(1) 34(1) 2(1) 2(1) 1(1) _______________________________________________________________________
275 Table A-4. Continued ________________________________________________________________________ U11 U22 U33 U23 U13 U12 ________________________________________________________________________ C2 23(1) 26(1) 31(1) 5(1) -2(1) -1(1) C3 25(1) 25(1) 37(1) 3(1) 7(1) 2(1) C4 29(1) 27(1) 37(1) 1(1) 9(1) 5(1) C5 45(1) 29(1) 34(1) 2(1) 11(1) 8(1) C6 55(1) 31(1) 42(1) 6(1) 22(1) 3(1) C7 33(1) 39(1) 35(1) -7(1) -1(1) -1(1) C8 46(2) 51(2) 37(1) -3(1) -3(1) 3(1) C9 22(1) 27(1) 31(1) 6(1) -1(1) -2(1) C10 23(1) 26(1) 33(1) 5(1) -3(1) -1(1) C11 28(1) 31(1) 33(1) 2(1) 0(1) 4(1) C12 27(1) 37(1) 38(1) 4(1) 6(1) -1(1) C13 49(1) 33(1) 48(1) -6(1) 1(1) -6(1) C14 49(1) 42(1) 47(1) -11(1) 5(1) -10(1) _______________________________________________________________________ Table A-5. Hydrogen coordinates ( x 104) and isotropic displacement parameters (2x 10 3) for BEDOT-Tz-H2. ________________________________________________________________________ x y z U(eq) ________________________________________________________________________ H6A 5728 4140 4599 49 H7A 7984 8002 4829 43 H7B 9599 7938 4637 43 H8A 9345 6752 5978 55 H8B 9696 5949 5089 55 H7'A 9812 7620 4625 50 H7'B 9769 6228 4478 50 H8'A 8972 6876 5986 53 H8'B 7630 7462 5311 53 ________________________________________________________________________
276 Table A-5. Continued ________________________________________________________________________ x y z U(eq) ________________________________________________________________________ H12A 11073 6275 -2385 41 H13A 7048 2681 -2238 53 H13B 8688 2770 -1747 53 H14A 8400 2969 -3486 55 H14B 7433 4103 -3421 55 H1 6286(16) 5241(14) -459(11) 26(4) H2 8085(19) 7264(16) 2044(14) 47(5) ________________________________________________________________________ Table A-6. Hydrogen bonds for BEDOT-Tz-H2 [ and ]. D-H...A d(D-H) d(H...A) d(D...A) <(DHA) ________________________________________________________________________ N1-H1...O3 0.882(16) 2.340(15) 3.0026(17) 132.0(12) N3-H2...O1 0.884(19) 2.130(19) 2.8606(17) 139.5(16) N1-H1...N2#1 0.882(16) 2.483(15) 3.0458(17) 122.2(12) N3-H2...S2#2 0.884(19) 2.883(19) 3.4431(13) 122.7(14) ________________________________________________________________________ Symmetry transformations used to generate equivalent atoms: #1 -x+1,-y+1,-z #2 -x+3/2,y+1/2,-z+1/2
LIST OF REFERENCES 1. (a) Handbook of Conducting Polymers, 2 nd ed.; Skotheim, T. A.; Elsenbaumer, R. L.; Reynolds, J. R., Eds.; Marcel Dekker: New York, 1998. (b) Handbook of Organic Conductive Molecules and Polymers; Nalwa, H. S., Ed.; Wiley: New York, 1997. (c) Moliton, A.; Hiorns, R. C. Polym. Int. 2004, 53, 1397-1412. (d) Heeger, A. J. Synth. Met. 2002, 125, 23-42. (e) MacDiarmid, A. G.; Angew. Chem. Int. Ed. 2001, 40, 2581-2590. 2. (a) Stafstrom, S.; Bredas, J. L. J. Mol. Struct. (Theochem) 1989, 188, 393-427. (b) Trivedi, D. C. In Handbook of Organic Conductive Molecules and Polymers Vol. 2; Nalwa, H. S., Ed.; Wiley: New York, 1997, chapter 12. 3. Hsu, J.-H.; hayashi, M.-t.; Lin, S.-H.; Fann, W.; Rothberg, L. J.; Perng, G.-Y.; Chen, S.-A. J. Phys. Chem. B 2002, 106, 8582-8586. 4. Gettinger, C. L.; Heeger, A. J.; Drake, J. M.; Pine, D. J. J. Chem. Phys. 1994, 101, 1673-1678. 5. Ng, K. K. Complete Guide to Semiconductor Devices; McGraw-Hill: New York, 1995. 6. (a) Roncali, J. Chem. Rev. 1997, 97, 173-205. (b) Ajayaghosh, A. Chem. Soc. Rev. 2003, 32, 181-191. (c) Pomerantz, M. In Handbook of Conducting Polymers, 2 nd ed.; Skotheim, T. A.; Elsenbaumer, R. L.; Reynolds, J. R., Eds.; Marcel Dekker: New York, 1998, chapter 11. 7. Del Zoppo, M.; Castiglioni, C.; Zuliani, P.; Zerbi, G. In Handbook of Conducting Polymers, 2 nd ed.; Skotheim, T. A.; Elsenbaumer, R. L.; Reynolds, J. R., Eds.; Marcel Dekker: New York, 1998, chapter 28. 8. Dhanabalan, A.; van Duren, J. K. J.; van Hal, P. A.; van Dongen, J. L. J.; Janssen, R. A. J. Adv. Funct. Mater. 2001, 11, 255-262. 9. DuBois, C. J.; Reynolds, J. R. Adv. Mater. 2002, 14, 1844-1846. 10. van Mullekom, H. A. M.; Vekemans, J. A. J. M.; Havinga, E. E.; Meijer, E. W. Mater. Sci. Eng. 2001, 32, 1-40. 11. Thomas, C. A.; Zong, K.; Abboud, K. A.; Steel, P. J.; Reynolds, J. R. J. Am. Chem. Soc. 2004, 126, 16440-16450. 277
278 12. Leclerc, M. Adv. Mater. 1999, 11, 1491-1498. 13. (a) Izumi, A.; Teraguchi, M.; nomura, R.; Masuda, T. Macromolecules 2000, 33, 5347-5352. (b) Levesque, I.; Leclerc, M. Macromolecules 1997, 30, 4347-4352. 14. (a) Irie, M. Chem. Rev. 2000, 100, 1685-1716. (b) Stellachi, F.; Bertarelli, C.; Toscano, F.; Gallazzi, M. C.; Zotti, G.; Zerbi, G. Adv. Mater. 1999, 11, 292-295. 15. McQuade, D. T.; Pullen, A. E.; Swager, T. M. Chem. Rev. 2000, 100, 2537-2574. 16. Monk, P. M. S.; Mortimer, R. J.; Rossiensky, D. R. Electrochromism: Fundamentals and Applications; VCH: Weinheim, Germany, 1995. 17. Thomas, C. T. Donor-Acceptor Methods for Band Gap Reduction in Conjugated Polymers: The Role of Elctron Rich Heterocycles Ph.D. Dissertation, University of Florida, Gainesville, FL, 2002. 18. (a) Bredas, J. L.; Street, G. B. Acc. Chem. Res. 1985, 18, 309-315. (b) Sakamoto, A.; Furukawa, Y.; Tasumi, M. J. Phys. Chem. B 1997, 101, 1726-1732. 19. Argun, A. A.; Aubert, P.-H.; Thompson, B. C.; Schwendeman, I.; Gaupp, C. L.; Hwang, J.; Pinto, N. J.; Tanner, D. B.; MacDiarmid, A. G.; Reynolds, J. R. Chem. Mater. 2004, 16, 4401-4412. 20. (a) Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Marks, R. N.; Taliani, C.; Bradley, D. D. C.; Dos Santos, D. A.; Bredas, J. L.; Logdlund, M.; Salaneck, W. R. Nature 1999, 397, 121-128. (b) Kraft, A.; Grimsdale, A. C.; Holmes, A. B. Angew. Chem. Int. Ed. 1998, 37, 403-428. 21. Green, M. A. Physica E 2002, 14, 11-17. 22. Becquerel, A. E. C. R. Acad Sci. 1839, 9, 145,561. 23. Ohl, R. S. Light Sensitive Device Including Silicon, US patent No. 2,443,542, 27 th May, 1941. 24. Chapin, D. M.; Fuller, C. S.; Pearson, G. L. J. Appl. Phys. 1954, 25, 676. 25. Goetzberger, A.; Hebling, C.; Schock, H.-W.Materials Science and Engineering R 2003, 40, 1-46. 26. Komp, R. J. Practical Photocvoltaics: Electricity from Solar Cells; AATEC: Ann Arbor, 2002; p. 13. 27. Ng, K. K. Complete Guide to Semiconductor Devices; McGraw-Hill: New York, 1995; pp. 474-481. 28. Rostalski, J.; Meissner, D. Sol. Energy Mat. Sol. Cells 2000, 61, 87-95.
279 29. Albert, P.; Smith, K. M.; Bennartz, R.; Newnham, D. A.; Fischer, J. J. Quant. Spec. Rad. Transfer.2004, 84, 181-193. 30. Anthes, R. A.; Panofsky, H. A.; Cahir, J. J.; Rango, A. The Atmosphere, 2 nd ed.; Charles E. Merrill: Columbus, 1978, p. 30. 31. Spanggaard, H.; Krebs, F. C. Sol. Energy Mater. Sol. Cells 2004, 83, 125-146. 32. Shaheen, S. E.; Ginley, D. S.; Jabbour, G. E. Mater. Res. Bull. 2005, 30, 10-15. 33. Bailey, S. G.; Raffaelle, R.; Emery, K. Prog. Photovolt: Res. Appl. 2002, 10, 399-406. 34. Oâ€™Reagan, B.; Grtzel, M. Nature, 1991, 353, 737-739. 35. Grtzel, M. J. Photochem. Photobiol. B 2003, 4, 145-153. 36. Grtzel, M. J. Sol-Gel Sci. Technol. 2001, 22, 7-13. 37. Hagfeldt, A.; Grtzel, M. Acc. Chem. Res. 2000, 33, 269-277. 38. Grtzel, M. Mater. Res. Bull. 2005, 30, 23-27. 39. Tang, C. W. Appl. Phys. Lett. 1986, 48, 183-184. 40. Gregg, B. A.; Hanna, M. C. J. Appl. Phys. 2003, 93, 3605. 41. Gregg, B. A. J. Phys. Chem. B 2003, 107, 4688. 42. Chanon, M.; Hawley, M. D.; Fox, M. A. in Photoinduced Electron Transfer, Part A, Fox, M. A.; Chanon, M., Eds.; Elsevier: Amsterdam, 1988; pp. 1-59. 43. Gregg, B. A.; Sprague, J.; Peterson, M. J. Phys. Chem. B 1997, 101, 5362-5369. 44. Forrest, S. R. Mater. Res. Bull. 2005, 30, 28-32. 45. Peumans, P.; Forrest, S. R. Appl. Phys. Lett. 2001, 79, 126-128. 46. Peumans, P.; Yakimov, A.; Forrest, S. R. J. Appl. Phys. 2003, 93, 3693-3723. 47. Brabec, C. J. Sol. Energy Mater. Sol. Cells 2004, 83, 273-292. 48. Kanicki, J. in Handbook of Conducting Polymers, Vol. 1, Skotheim, T. A., Ed.; Marcel Dekker: New York, 1985; pp. 543-660. 49. (a) Kanda, S.; Pohl, H. A. In Organic Semiconducting Polymers; Marcel Dekker: New York, 1968, chapter 3. (b) Boguslavskii, L. I.; Vannikov, A. V. Organic Semiconductors and Biopolymers; Plenum: New York, 1970; pp. 21-35.
280 50. Shirakawa, H.; Louis, E. J.; MacDiarmid, A. G.; Chiang, C. K.; Heeger, A. J. J. Chem. Soc. Chem. Comm. 1977, 578-580. 51. Weinberger, B. R.; Gau, S. C.; Kiss, Z. Appl. Phys. Lett. 1981, 38, 555-557. 52. Glenis, S.; Horowitz, G.; Tourillon, G.; Garnier, F. Thin Solid Films 1984, 111, 93-103. 53. Glenis, S.; Tourillon, G.; Garnier, F. Thin Solid Films 1986, 139, 221-231. 54. Horowitz, G. Adv. Mat. 1990, 2, 287-292. 55. Burroughes, J. H.; Bradley, D. D. C.; Brown, A. R.; Marks, R. N.; Friend, R. H.; Burn, P. L.; Holmes, A. B. Nature 1990, 347, 539. 56. Pichler, K.; Halliday, D. A.; Bradley, D. D. C.; Burn, P. L.; Friend, R. H.; Holmes, A. B. J. Phys.: Condens. Matter 1993, 5, 7155-7172. 57. Karg, S.; Riess, W.; Dyakonov, V.; Schwoerer, M. Synth. Met. 1993, 54, 427-433. 58. Halliday, D. A.; Burn, P. L.; Bradley, D. D. C.; Friend, R. H.; Gelsen, O. M.; Holmes, A. B.; Kraft, A.; Martens, J. H. F.; Pichler, K. Adv. Mat. 1993, 5, 40-43. 59. Marks, R. N.; Halls, J. J. M.; Bradley, D. D. C.; Friend, R. H.; Holmes, A. B. J. Phys. Condens. Matter 1993, 6, 1379-1394. 60. Antoniadis, H.; hsieh, B. R.; Abkowitz, M. A.; Jenekhe, S. A. Synth. Met. 1994, 62, 265. 61. Yu, G.; Zhang, C.; Heeger, A. J. Appl. Phys. Lett. 1994, 64, 1540-1542. 62. Hoppe, H.; Sariciftci, N. S. J. Mater. Res. 2004, 19, 1924-1945. 63. Sariciftci, N. S.; Smilowitz, L.; Heeger, A. J.; Wudl, F. Science 1992, 258, 1474-1476. 64. Sariciftci, N. S.; Braun, D.; Zhang, C.; Srdanov, V. I.; Heeger, A. J.; Stucky, G.; Wudl, F. Appl. Phys. Lett. 1993, 62, 585-587. 65. Halls, J. J. M.; Pichler, K.; Friend, R. H.; Moratti, S. C.; Holmes, A. B. Appl. Phys. Lett. 1996, 68, 3120-3122. 66. Tada, K.; Onoda, M.; Zakhidov, A. A.; Yoshino, K. Jpn. J. Appl. Phys. 1997, 36, L306-L309. 67. Granstrom, M.; Petritsch, K.; Arias, A. C.; Lux, A.; Andersson, M. R.; Friend, R. H. Nature, 1998, 395, 257-260.
281 68. Berggren, M.; Gustafsson, G.; Inganas, O.; Andersson, M. R.; Wennerstrom, O.; Hjertberg, T. Appl. Phys. Lett. 1994, 65, 1489-1491. 69. Janssen, R. A. J.; Hummelen, J. C.; Sariciftci, N. S. Mater. Res. Bull 2005, 30, 33-36. 70. Smilowitz, L.; Sariciftci, N. S.; Wu, R.; Gettinger, C.; Heeger, A. J.; Wudl, F. Phys. Rev. B 1993, 47, 13835-13842. 71. Lee, C. H.; Yu, G.; Moses, D.; Pakbaz, K.; Zhang, C.; Sariciftci, N. S.; Heeger, A. J.; Wudl, F. Phys. Rev. B 1993, 48, 15425-15433. 72. Kraabel, B.; Lee, C. H.; McBranch, D.; Moses, D.; Sariciftci, N. S.; Heeger, A. J. Chem. Phys. Lett. 1993, 213, 389-394. 73. Yu, G.; Pakbaz, K.; Heeger, A. J. Appl. Phys. Lett. 1994, 64, 3422-3424. 74. Halls, J. J. M.; Walsh, C. A.; Greenham, N. C.; Marseglia, E. A.; Friend, R. H.; Moratti, S. C.; Holmes, A. B. Nature 1995, 376, 498-500. 75. Yu, G.; Heeger, A. J. J. Appl. Phys. 1995, 78, 4510-4515. 76. Greenham, N. C.; Moratti, S. C.; Bradley, D. D. C.; Friend, R. H.; Holmes, A. B. Nature 1993, 365, 628-630. 77. Yu, G.; Gao, J.; Hummelen, J. C.; Wudl, F.; Heeger, A. J. Science 1995, 270, 1789-1791. 78. Hummelen, J. C.; Knight, B. W.; Lepec, F.; Wudl, F. J. Org. Chem. 1995, 60, 532-538. 79. Shaheen, S. E.; Brabec, C. J.; Sariciftci, N. S.; Padinger, F.; Fromherz, T.; Hummelen, J. C. Appl. Phys. Lett. 2001, 78, 841-843. 80. Alem, S.; de Bettignies, R.; Nunzi, J.-M.; Cariou, M. Appl. Phys. Lett. 2004, 84, 2178-2180. 81. Padinger, F.; Rittberger, R. S.; Sariciftci, N. S. Adv. Funct. Mater. 2003, 13, 85-88. 82. Wienk, M. M.; Kroon, J. M.; Verhees, W. J. H.; Knol, J.; Hummelen, J. C.; van Hal, P. A.; Janssen, R. A. J. Angew. Chem. Int. Ed. 2003, 42, 3371-3375. 83. (a) Halls, J. J. M.; Pichler, K.; Friend, R. H.; Moratti, S. C.; Holmes, A. B. Synth. Met. 1996, 77, 277-280. (b) Halls, J. J. M.; Pichler, K.; Friend, R. H.; Moratti, S. C.; Holmes, A. B. Appl. Phys. Lett. 1996, 68, 3120-3122. 84. Coakley, K. M.; McGehee, M. D. Chem. Mater. 2004, 16, 4533-4542. 85. Sariciftci, N. S. Materials Today 2004, 7, 36-40.
282 86. Winder, C.; Sariciftci, N. S. J. Mater. Chem. 2004, 14, 1077-1086. 87. (a) Inganas, O.; Svensson, M.; Zhang, F.; Gadisa, A.; Persson, N. K.; Wang, X.; Andersson, M. R. Appl. Phys. A. 2004, 79, 31-35. (b) Yohannes, T.; Zhang, F.; Svensson, M.; Hummelen, J. C.; Andersson, M. R.; Inganas, O. Thin Solid Films 2004, 449, 152-157. (c) Zhou, Q.; Hou, Q.; Zheng, L.; Deng, X.; Yu, G.; Cao, Y. Appl. Phys. Lett. 2004, 84, 1653-1655.(d) Svensson, M.; Zhang, F.; Veenstra, S. C.; Verhees, W. J. H.; Hummelen, J. C.; Kroon, J. M.; Inganas, O.; Andersson, M. R. Adv. Mater. 2003, 15, 988-991. (e) Brabec, C. J.; Winder, C.; Sariciftci, N. S.; Hummelen, J. C.; Dhanabalan, A.; van Hal, P. A.; Janssen, R. A. J. Adv. Funct. Mater. 2002, 12, 709-712. (f) Vangeneugden, D. L.; Vanderzande, D. J. M.; Salbeck, J.; van Hal, P. A.; Janssen, R. A. J.; Hummelen, J. C.; Brabec, C. J.; Shaheen, S. E.; Sariciftci, N. S. J. Phys. Chem. B 2001, 105, 11106-11113. (g) Shaheen, S. E.; Vangeneugden, D.; Kiebooms, R.; Vanderzande, D.; Fromherz, T.; Padinger, F.; Brabec, C. J.; Sariciftci, N. S. Synth. Met. 2001, 121, 1583-1584. 88. (a) Neugebauer, H.; Loi, M. A.; Winder, C.; Sariciftci, N. S.; Cerullo, G.; Gouloumis, A.; Vazquez, P.; Torres, T. Sol. Energy Mat. Sol. Cells 2004, 83, 201-209. (b) Winder, C.; Matt, G.; Hummelen, J. C.; Janssen, R. A. J.; Sariciftci, N. S.; Brabec, C. J. Thin Solid Films 2002, 403-404, 373-379. 89. (a) Hoppe, H.; Niggemann, M.; Winder, C.; Kraut, J.; Hiesgen, R.; Hinsch, A.; Meissner, D.; Sariciftci, N. S. Adv. Funct. Mater. 2004, 14, 1005-1011. (b) Yang, X.; van Duren, J. K. J.; Rispens, M. T.; Hummelen, J. C.; Janssen, R. A. J.; Michels, M. A. J.; Loos, J. Adv. Mater. 2004, 16, 802-806. (c) van Duren, J. K. J.; Yang, X.; Loos, J.; Bulle-Lieuwma, C. W. T.; Sieval, A. B.; Hummelen, J. C.; Janssen, R. A. J. Adv. Funct. Mater. 2004, 14, 425-434. (d) Yang, X.; van Duren, J. K. J.; Janssen, R. A. J.; Michels, M. A. J.; Loos, J. Macromolecules 2004, 37, 2151-2158. 90. Mihailetchi, V. D.; van Duren, J. K. J.; Blom, P. W. M.; Hummelen, J. C.; Janssen, R. A. J.; Kroon, J. M.; Rispens, M. T.; Verhees, W. J. H.; Wienk, M. M. Adv. Funct. Mater. 2003, 13, 43-46. 91. Martens, H. C. F.; Brom, H. B.; Blom, P. W. M. Phys. Rev. B 1999, 60, R8489-R8492. 92. Melzer, C.; Koop, E. J.; Mihailetchi, V. D.; Blom, P. W. M. Adv. Funct. Mater. 2004, 14, 865-870. 93. Sirringhaus, H.; Brown, P. J.; Friend, R. H.; Nielsen, M. M.; Bechgaard, K.; Langeveld-Voss, B. M. W.; Priering, A. J. H.; Janssen, R. A. J.; Meijer, E. W.; Herwig, P.; de Leeuw, D. M. Nature, 1999, 401, 685-688. 94. Mozer, A. J.; Denk, P.; Scharber, M. C.; Neugebauer, H.; Sariciftci, N. S.; Wagner, P.; Lutsen, L.; Vanderzande, D. J. Phys. Chem. B 2004, 108, 5235-5242.
283 95. (a) Nelson, J. Phys. Rev. B 2003, 67, 155209-1-10. (b) Arndt, C.; Zhokhavets, U.; Gobsch, G.; Winder, C.; Lungenschmied, C.; Sariciftci, N. S. Thin Solid Films 2004, 451-452, 60-63. 96. Aernouts, T.; Geens, W.; Poortmans, J.; Heremans, P.; Borghs, S.; Mertens, R. Thin Solid Films 2002, 403-404, 297-301. 97. Jabbour, G. E.; Kawabe, Y.; Shaheen, S. E.; Wang, J. F.; Morrell, M. M.; Kippelen, B.; Peyghambarian, N. Appl. Phys. Lett. 1997, 71, 1762-1764. 98. Brown, T. M.; Millard, I. S.; Lacey, D. J.; Burroughes, J. H.; Friend, R. H.; Cacialli, F. Synth. Met. 2001, 124, 15-17. 99. Jonsson, S. K. M.; Carlegrim, E.; Zhang, F.; Inganas, O.; Salaneck, W. R.; Fahlman, M. â€œPhotoelectron Spectroscopy of Surfaces and Interfaces in Plastic Solar Cells: The Benefit of LiF/Al as Opposed to Al Electrodesâ€ Materials Research Society Proceedings, Fall 2004. 100. Brabec, C. J.; Shaheen, S. E.; Winder, C.; Sariciftci, N. S.; Denk, P. Appl. Phys. Lett. 2002, 80, 1288-1290. 101. (a) Mihailetchi, V. D.; Koster, L. J. A.; Blom, P. W. M. Appl. Phys. Lett. 2004, 85, 970-972. (b) Mihailetchi, V. D.; Blom, P. W. M.; Hummelen, J. C.; Rispens, M. T. J. Appl. Phys. 2003, 94, 6849-6854. (c) Ramsdale, C. M.; Barker, J. A.; Arias, A. C.; MacKenzie, J. D.; Friend, R. H.; Greenham, N. C. J. Appl. Phys. 2002, 92, 4266-4270. (d) Brabec, C. J.; Cravino, A.; Meissner, D.; Sariciftci, N. S.; Fromherz, T.; Rispens, M. T.; Sanchez, L.; Hummelen, J. C. Adv. Funct. Mater. 2001, 11, 374-380. 102. (a) Gadisa, A.; Svensson, M.; Andersson, M. R.; Inganas, O. Appl. Phys. Lett. 2004, 84, 1609-1611. (b) Snaith, H. J.; Greenham, N. C.; Friend, R. H. Adv. Mater. 2004, 16, 1640-1645. 103. Zhao, Y.; Yuan, G.; Roche, P.; Leclerc, M. Polymer 1995, 36, 2211-2214. 104. Zen, A.; Pflaum, J.; Hirschmann, S.; Zhuang, W.; Jaiser, F.; Asawapirom, U.; Rabe, J. P.; Scherf, U.; Neher, D. Adv. Funct. Mater. 2004, 14, 757-764. 105. Lee, T.-W.; Park, O. O. Appl. Phys. Lett. 2000, 77, 3334-3336. 106. Kim, Y.; Choulis, S. A.; Nelson, J.; Bradley, D. D. C.; Cook, S.; Durrant, J. R. Appl. Phys. Lett. 2005, 86, 063502. 107. Van der Veen, M. H.; de Boer, B.; Stalmach, U.; van de Wetering, K. I.; Hadziioannou, G. Macromolecules 2004, 37, 3673-3684. 108. Cravino, A.; Sariciftci, N. S. J. Mater. Chem. 2002, 12, 1931-1943.
284 109. Breeze, A. J.; Schlesinger, Z.; Carter, S. A.; Tillman, H.; Horhold, H.-H. Sol. Energy Mater. Sol. Cells 2004, 83, 263-271. 110. Deng, X.; Zheng, L.; Yang, C.; Li, Y.; Yu, G.; Cao, Y. J. Phys. Chem. B 2004, 108, 3451-3456. 111. (a) Bhattacharyya, S.; Kymakis, E.; Amaratunga, G. A. J. Chem. Mater. 2004, 16, 4819-4823. (b) Wu, W.; Zhang, S.; Li, Y.; Li, J.; Liu, L.; Qin, Y.; Guo, Z.-X.; Dai, L.; Ye, C.; Zhu, D. Macromolecules 2003, 36, 6286-6288. (c) Kymakis, E.; Alexandrou, I.; Amaratunga, G. A. J. J. Appl. Phys. 2003, 93, 1764-1768. (d) Kymakis, E.; Amaratunga, G. A. J. Appl. Phys. Lett. 2002, 80, 112-114. (e) Coleman, J. N.; Curran, S.; Dalton. A. B.; Davey, A. P.; McCarthey, B.; Blau, W.; Barklie, R. C. Phys. Rev. B 1998, 58, R7492-R7495. 112. Huynh, W. U.; Dittmer, J. J.; Alivisatos, A. P. Science 2002, 295, 2425-2427. 113. (a) Xue, J.; Uchida, S.; Rand, B. P.; Forrest, S. R. Appl. Phys. Lett. 2004, 84, 3013-3015. (b) Peumans, P.; Yakimov, A.; Forrest, S. R. J. Appl. Phys. 2003, 93, 3693-3723. 114. (a) Bayer AG, Eur. Patent 339 340, 1988. (b) Jonas, F.; Schrader, L. Synth Met. 1991, 41, 831-836. (c) Heywang, G.; Jonas, F. Adv. Mater. 1992, 4, 116-118. 115. (a) Groenendaal, L.; Jonas, F.; Freitag, D.; Pielartzik, H.; Reynolds, J. R. Adv. Mater. 2000, 12, 481-494. (b) Groenendaal, L.; Zotti, G.; Aubert, P.-H.; Waybright, S. M.; Reynolds, J. R. 2003, 15, 855-879. 116. Sotzing, G. A.; Thomas, C. A.; Reynolds, J. R.; Steel, P. J. Macromolecules 1998, 31, 3750-3752. 117. (a) Kumar, A.; Welsh, D. M.; Morvant, M. C.; Piroux, F.; Abboud, K. A.; Reynolds, J. R. Chem. Mater. 1998, 10, 896-902. (b) Welsh, D. M.; Kumar, A.; Meijer, E. W.; Reynolds, J. R. Adv. Mater. 1999, 11, 1379-1382. (c) Zong, K.; Madrigal, L.; Groenendaal, L.; Reynolds, J. R. Chem. Comm. 2002, 2498-2499. 118. (a) Reeves, B. D.; Grenier, C. R. G.; Argun, A. A.; Cirpan, A.; McCarley, T. D.; Reynolds, J. R. Macromolecules 2004, 37, 7559-7569. (b) Welsh, D. M.; Kleoppner, L. J.; Madrigal, L.; Pinto, M. R.; Thompson, B. C.; Schanze, K. S.; Abboud, K. A.; Powell, D.; Reynolds, J. R. Macromolecules 2002, 35, 6517-6525. 119. Mishra, S. P.; Krishnamoorthy, K.; Sahoo, R.; Kumar, A. J. Polym. Sci. Part A: Polym.Chem. 2005, 43, 419-428. 120. Wang, F.; Wilson, M. S.; Rauh, R. D.; Schottland, P.; Thompson, B. C.; Reynolds, J. R. Macromolecules 2000, 33, 2083-2091.
285 121. (a) Argun, A. A.; Cirpan, A.; Reynolds, J. R. Adv. Mater. 2003, 15, 1338-1341. (b) Cirpan, A.; Argun, A. A.; Grenier, C. R. G.; Reeves, B. D.; Reynolds, J. R. J. Mater. Chem. 2003, 13, 2422-2428. 122. (a) DuBois, C. J. Donor-Acceptor Methods for Band Gap Control in Conjugated Polymers. Ph.D. Dissertation, University of Florida, Gainesville, FL, 2003. (b) Gaupp, C. L. Structure Property Relationships of Electrochromic 3,4-alkylenedioxyheterocycle-based Polymers and Copolymers. Ph.D. Dissertation, University of Florida, Gainesville, FL, 2002. 123. Hanbook of Conducting Polymers; 2 nd ed.; Skotheim, T. A.; Elsenbaumer, R. L.; Reynolds, J. R. Eds.; Marcel Dekker: New York, 1998, chapter 20. 124. Gritzner, G.; Kuta, G. J. Pure Appl. Chem. 1984, 56, 461-466. 125. Bard, A. J.; Faulkner, L. R. Electrochemical Methods:Fundamentals and Applications, 2 nd Ed.; Wiley: New York, 2001. 126. Sawyer, D. T.; Sobkowiak, A.; Roberts, J. L. Electrochemistry for Chemists, 2 nd ed.; Wiley: New York, 1995. 127. (a) Villegas, I.; Gomez, R.; Weaver, M. J. J. Phys. Chem. 1995, 99, 14832-14839. (b) Wagner, F. T. In Structure of Electrified Interfaces, Lipowski, J.; Ross, P. N., Eds.; VCH: Weinheim, 1993, chapter 9. 128. (a) Farrell, J. R.; McTigue, P.; J. Electroanal. Chem. 1982, 139, 37-56. (b) Hansen, W. N.; Hansen, G. J. Phys. Rev. A 1987, 36, 1396-1402. 129. (a) Gomer, R.; Tryson, G. J. Chem. Phys. 1977, 66, 4413-4424. (b) Kotz, E. R.; Neff, H.; Muller, K. J. Electroanal. Chem. 1986, 215, 331-344. 130. Trassati, S. Pure Appl. Chem. 1986, 58, 955-966. 131. Pavlishchuk, V. V.; Addison, A. W. Inorganica Chimica Acta 2000, 298, 97-102. 132. Li, Y.; Cao, Y.; Gao, J.; Wang, D.; Yu, G.; Heeger, A. J. Synth. Met. 1999, 99, 243-248. 133. Morgado, J.; Cacialli, F.; Friend, R. H.; Chuah, B. S.; Rost, H.; Holmes, A. B. Macromolecules 2001, 34, 3094-3099. 134. de Leeuw, D. M.; Simenon, M. M. J.; Brown, A. R.; Einerhard, R. E. F. Synth. Met. 1997, 87, 53-59. 135. Bredas, J. L.; Silbey, R.; Boudreaux, D. S.; Chance, R. R. J. Am. Chem. Soc.1983, 105, 6555-6559. 136. Lohmann, F. Z. Naturforsch., Teil A 1967, 22, 843-844.
286 137. Pommerehne, J.; Vestweber, H.; Guss, W.; Mahrt, R. F.; Bssler, H.; Porsch, M.; Daub, J. Adv. Mat. 1995, 7, 551-554. 138. Thompson, B. C.; Schottland, P.; Zong, K.; Reynolds, J. R. Chem. Mater. 2000, 12, 1563-1571. 139. Seybold, P. G.; Gouterman, M. J. Mol. Spectroscopy 1969, 31, 1-13. 140. Strachan, J. P.; Gentemann, S.; Seth, J.; Kalsbeck, W. A.; Lindsey, J. S.; Holten, D.; Bacian, D. F. J. Am. Chem. Soc. 1997, 119, 11191-11201. 141. Armstrong, N. R.; Carter, C.; Donley, C.; Simmonds, A.; Lee, P.; Brumbach, M.; Kippelen, B.; Domercq, B.; Yoo, S. Thin Solid Films 2003, 445, 342-352. 142. Zhang, F. L.; Gadisa, A.; Inganas, O.; Svensson, M.; Andersson, M. R. Appl. Phys. Lett. 2004, 84, 3906-3908. 143. Sommeling, P. M.; Rieffe, H. C.; van Roosmalen, J. A. M.; Schonecker, A.; Kroon, J. M.; Wienke, J. A.; Hinsch, A. Sol. Energy Mater. Sol. Cells 2000, 62, 399-410. 144. (a) Zhou, Q.; Hou, Q.; Zheng, L.; Deng, X.; Yu, G.; Cao, Y. Appl. Phys. Lett. 2004, 84, 1653-1655. (b) Inganas, O.; Svensson, M.; Zhang, F.; Gadisa, A.; Persson, N. K.; Wang, X.; Andersson, M. R. Appl. Phy. A 2004, 79, 31-35. (c) Smith, A. P.; Smith, R. R.; Taylor, B. E.; Durstock, M. F. Chem. Mater. 2004, 16, 4687-4692. (d) Colladet, K.; Nicolas, M.; Goris, L.; Lutsen, L.; Vanderzande, D. Thin Solid Films 2004, 451-452, 7-11. (e) Henckens, A.; Knipper, M.; Polec, I.; Manca, J.; Lutsen, L.; Vanderzande, D. Thin Solid Films 2004, 451-452, 572-579. 145. (a) Al-Ibrahim, M.; Roth, H.-K.; Zhokhavets, U.; Gobsch, G.; Sensfuss, S. Sol. Energy Mat. Sol. Cells 2005, 85, 13-20. (b) Riedel, I.; Dyakonov, V. Phys. Stat. Sol. 2004, 201, 1332-1341. 146. Bard, A. J.; Faulkner, L. R. Electrochemical Methods:Fundamentals and Applications, 2 nd Ed.; Wiley: New York, 2001, p.54. 147. Meskers, S. C. J.; Hubner, J.; Oestreich, M.; Bassler, H. J. Phys. Chem. B 2001, 105, 9139-9149. 148. Arkhipov, V. I.; Bassler, H. Phys. Stat. Sol. 2004, 201, 1152-1187. 149. Zerza, G.; Rothler, B.; Sariciftci, N. S.; Gomez, R.; Segura, J. L.; Martin, N. J. Phys. Chem. B 2001, 105, 4099-4104. 150. Moratti, S. C.; Cervini, R. Holmes, A. B.; Baigent, D. R.; Friend, R. H.; Greenham, N. C.; Gruner, J.; Hamer, P. J. Synth. Met. 1995, 71, 2117-2120. 151. Yu, Y.; Lee, H.; VanLaeken, A.; Hsieh, B. R. Macromolecules 1998, 31, 5553-5555.
287 152. Mohanakrishan, A. K.; Hucke, A.; Lyon, M. A.; Lakshmikantham, M. V.; Cava, M. P. Tetrahedron 1999, 55, 11745-11754. 153. Zong, K.; Madrigal, L.; Groenendaal, L.; Reynolds, J. R. Chem. Comm. 2002, 2498-2499. 154. Detty, M. R.; Hays, D. S. Heterocycles 1995 , 40, 925-937. 155. Amari, C.; Pelizzi, C.; Predieri, G.; Destri, S.; Porzio, W.; Einsiedel, H.; Menges, B.; Mittler-Neher, S. J. Mater. Chem. 1996, 6, 1319-1324. 156. Carpenter, A. J.; Chadwick, D. J. Tetrahedron 1985, 41, 3803-3812. 157. Johnson, A. L. J. Org. Chem. 1976, 41, 1320-1324. 158. Holmes, A. B.; Friend, R. H.; Moratti, S. C.; Baigent, D. R.; Bradley, D. D. C.; Cervini, R.; Greenham, N. C.; Hamer, P. J. U.S. Patent 5,514,878, 1994. 159. Badger, G. M.; Elix, J. A.; Lewis, G. E. Aust. J. Chem. 1965, 18, 70-89. 160. Ruiz, J. P.; Dharia, J. R.; Reynolds, J. R.; Buckley, L. J. Macromolecules 1992, 25, 849-860. 161. Wang, B.; Wasielewski, M. R. J. Am. Chem. Soc. 1997, 119, 12-21. 162. Gill, R. E.; van Hutten, P. F.; Meetsma, A.; Hadziioannou, G. 1996, 8, 1341-1346. 163. Stalmach, U.; Detert, H. J. Prakt. Chem. 2000, 342, 10-16. 164. Davis, R.; Untch, K. G. J. Org. Chem. 1981, 46, 2985-2987. 165. van Leusen, A. M.; Oomkes, P. G. Synth. Comm. 1980, 10, 399-403. 166. Oldenziel, O. H.; van Leusen, D.; van Leusen, A. M. J. Org. Chem. 1977, 42, 3114-3118. 167. (a) Perosa, A.; Selva, M.; Tundo, P. J. Chem. Soc. Perkin Trans. 2 2002, 1033-1037. (b) Yoneda, R.; Harusawa, S.; Kurihara, T. Tetrahedron Lett. 1989, 30, 3681-3684. (c) Harusawa, S.; Nakamura, S.; Yagi, S.; Kurihara, T.; Hamada, Y.; Shioiri, T. Synth. Comm. 1984, 14, 1365-1371. (d) Yokoyama, M.; Ohteki, H.; Kurauchi, M.; Hoshi, K.; Yanagisawa, E.; Suzuki, A.; Imamoto, T. J. Chem. Soc. Perkin 1 1984, 2635-2640. 168. (a) Aesa, M.; Baan, G.; Novak, L.; Szantay, C. Synth. Comm. 1995, 25, 1545-1550. (b) Wilk, B. K. Synth. Comm. 1993, 23, 2481-2484. 169. Griffing, J. M.; Salisbury, L. F. J. Am. Chem. Soc. 1948, 70, 3416-3419.
288 170. Pereyre, M.; Bellegarde, B.; Valade, J. C. R. Seances Acad. Sci. 1967, 265, 939-941. 171. Kosugi, M.; Ishiguro, M.; Negishi, Y.; Sano, H.; Migita, T. Chem. Lett. 1984, 1511-1512. 172. Cervini, R.; Holmes, A. B.; Moratti, S. C.; Kohler, A.; Friend, R. H. Synth. Met. 1996, 76, 169-171. 173. (a) Li, H.; hu, Y.; Zhang, Y.; Ma, D.; Wang, L.; Jing, X.; Wang, F. J. Polym. Sci. Part A: Polym. Chem. 2004, 42, 3947-3953. (b) Jin, Y.; ju, J.; Kim, J.; Lee, S.; Kim, J. Y.; Park, S. H.; Son, S.-M.; Jin, S.-H.; Lee, K.; Suh, H. Macromolecules 2003, 36, 6970-6975. (c) Chen, S.-A.; Chang, E.-C. Macromolecules 1998, 31, 4899-4907. (d) Boucard, V.; Benazzi, T.; Ades, D.; Sauvet, G.; Siove, A. Polymer 1997, 38, 3697-3703. 174. Liao, J.; Wang, Q. Macromolecules 2004, 37, 7061-7063. 175. Boucard, V.; Ades, D. Siove, A. Romero, D.; Schaer, M.; Zuppiroli, L. Macromolecules 1999, 32, 4729-4731. 176. (a) Murgasova, R.; Hercules, D. M. Int. J. Mass. Spectrom. 2003, 226, 151-162. (b) Nielen, M. W. F. Mass Spectrom. Rev. 1999, 18, 309-344. 177. Yassar, A.; Demanze, F.; Jaafari, A.; El Idrissi, M.; Coupry, C. Adv. Funct. Mater. 2002, 12, 699-708. 178. Schluter, A. D. J. Polym. Sci. Part A: Poly. Chem. 2001, 39, 1533-1556. 179. (a) Cho, N. S.; Hwang, D.-H.; Jung, B.-J.; Lim, E.; Lee, J.; Shim, H.-K. Macromolecules 2004, 37, 5265-5273. (b) Beaupre, S.; Leclerc, M. Macromolecules 2003, 36, 8986-8991. (c) Beaupre, S.; Leclerc, M. Adv. Funct. Mater. 2002, 12, 192-196. 180. Apperloo, J. J.; Groenendaal, L. B.; Verheyen, H.; Jayakannan, M.; Janssen, R. A. J.; Dkhissi, A.; Beljonne, D.; Lazzaroni, R.; Bredas, J. L. Chem. Eur. J. 2002, 8, 2384-2396. 181. (a) Wu, X.; Chen, T.-A.; Rieke, R. D. Macromolecules 1996, 29, 7671-7677. (b) Chen, T.-A.; Oâ€™Brien, R. A.; Rieke, R. D. Macromolecules 1993, 26, 3462-3463. 182. (a) Yamamoto, T. Synlett 2003 425-450. (b) Yamamoto, T. Macromol. Rapid Comm. 2002, 23, 583-606. 183. Yamamoto, T.; Maruyama, T.; Zhou, Z.; Ito, T.; Fukuda, T.; Yoneda, Y.; Begum, F.; Ikeda, T.; Sasaki, S.; Takezoe, H.; Fukuda, A.; Kubota, K. J. Am. Chem. Soc. 1994, 116, 4832-4845.
289 184. (a) Yamamoto, T.; Shiraishi, K.; Abla, M.; Yamaguchi, I.; Groenendaal, L. Polymer 2002, 43, 711-719. (b) Yamamoto, T.; Abla, M. Synth. Met. 1999, 100, 237-239. 185. Wagner, P.; Aubert, P.-H.; Lutsen, L.; Vanderzande, D. Electrochem. Comm. 2002, 4, 912-916. 186. Hoffmann, K. J.; Carlsen, P. H. J. Synth. Comm. 1999, 29, 1607-1610. 187. Frejd, T.; Klingstedt, T. K. Synthesis 1987, 1, 40-42. 188. Mitchell, R. H.; Lai, Y.-H.; Williams, R. V. J. Org. Chem. 1979, 44, 4733-4735. 189. Donat-Bouillud, A.; Levesque, I.; Tao, Y.; Dâ€™Iorio, M.; Beaupre, S.; Blondin, P.; Ranger, M.; Bouchard, J.; Leclerc, M. Chem. Mater. 2000, 12, 1931-1936. 190. Zhang, Z.-B.; Fujiki, M.; Tang, H.-Z.; Motonaga, M.; Torimitsu, K. Macromolecules 2002, 35, 1988-1990. 191. Kim, Y.-S.; Sato, R.; Maruyama, T.; Osakada, K.; Yamamoto, T. J. Chem. Soc. Dalton Trans. 1994, 943-948. 192. Yamamoto, T.; Morita, A.; Miyazaki, Y.; Maruyama, T.; Wakayama, H.; Zhou, Z. H.; Nakamura, Y.; Kanabra, T.; Sasaki, S.; Kubota, K. Macromolecules 1992, 25, 1214-1223. 193. Groenendaal, L.; Zotti, G.; Aubert, P.-H.; Waybright, S. M.; Reynolds, J. R. Adv. Mat. 2003, 855-879. 194. Sotzing, G. A.; Reynolds, J. R.; Steel, P. J. Adv. Mat. 1997, 9, 795-798. 195. Pepitone, M. F.; Hardaker, S. S.; Gregory, R. V. Chem. Mater. 2003, 15, 557-563. 196. (a) Zotti, G.; Schavion, G.; Zecchin, S. Synth. Met. 1996, 72, 275-281. (b) Borjas, R.; Buttry, D. A. Chem. Mater. 1991, 3, 872-878. (c) Crooks, R. M.; Chyan, O. M.; Wrighton, M. S. Chem. Mater. 1989, 1, 2-4. 197. Ofer, D.; Crooks, R. M.; Wrighton, M. S. J. Am. Chem. Soc. 1990, 112, 7869-7879. 198. DuBois, C. J.; Abboud, K. A.; Reynolds, J. R. J. Phys. Chem. B 2004, 108, 8550-8557. 199. Brabec, C. J.; Sariciftci, N. S.; Hummelen, J. C. Adv. Funct. Mater. 2001, 11, 15-26. 200. Alam, M. M.; Jenekhe, S. A. Chem. Mater. 2004, 16, 4647-4656. 201. Samuel, I. D. W.; Rumbles, G.; Collison, C. J. Phys. Rev. B 1995, 52, 11573-11576.
290 202. Nguyen, T.-Q.; Martini, I. B.; Liu, J.; Schwatrz, B. J. J. Phys. Chem. B 2000, 104, 237-255. 203. Chen, Z.-K.; Meng, H.; Lai, Y.-H.; Huang, W. Macromolecules 1999, 32, 4351-4358. 203. Kroto, H. W.; Heath, J. R.; Oâ€™Brien, S. C.; Curl, R. F.; Smalley, R. E. Nature 1985, 318, 162-163. 205. Wudl, F. Acc. Chem. Res. 1992, 25, 157-161. 206. Nakamura, E.; Isobe, H. Acc. Chem. Res. 2003, 36, 807-815. 207. Bendikov, M.; Wudl, F.; Perepichka, D. F. Chem. Rev. 2004, 104, 4891-4945. 208. Wudl, F. J. Mater. Chem. 2002, 12, 1959-1963. 209. Prato, M. J. Mater. Chem. 1997, 7, 1097-1109. 210. Diederich, F.; Isaacs, L.; Philp, D. Chem. Soc. Rev. 1994, 23, 243-255. 211. Keshavarz-K., M.; Knight, B.; Haddon, R. C.; Wudl, F. Tetrahedron 1996, 5149-5159. 212. Zheng, L.; Zhou, Q.; Deng, X.; Yuan, M.; Yu, G.; Cao, Y. J. Phys. Chem. B 2004, 108, 11921-11926. 213. (a) Carano, M.; Da Ros, T.; Fanti, M.; Kordatos, K.; Maraccio, M.; Paolucci, F.; Prato, M.; Roffia, S.; Zerbetto, F. J. Am. Chem. Soc. 2003, 125, 7139-7144. (b) Echegoyen, L.; Echegoyen, L. E. Acc. Chem. Res. 1998, 31, 593-601. 214. (a) Guldi, D. M.; Luo, C.; Swartz, A.; Gomez, R.; Segura, J. L.; Martin, N. J. Phys. Chem. A 2004, 108, 455-467. (b) Martin, N.; Giacalone, F.; Segura, J. L.; Guldi, D. M. Synth. Met. 2004, 147, 57-61. (c) Martini, I. B.; Ma, B.; Da Ros, T.; helgeson, R.; Wudl, F.; Schwatrz, B. J. Chem. Phys. Lett. 2000, 327, 253-262. (d) van Hal, P. A.; Knol, J.; Langeveld-Voss, B. M. W.; Meskers, S. C. J.; Hummelen, J. C.; Janssen, R. A. J. J. Phys. Chem. A 2000, 104, 5974-5988. (e) Imahori, H.; Sakata, Y. Eur. J. Org. Chem. 1999, 2445-2457. (f) Knorr, S.; Grupp, mehring, M.; Grube, G.; Effenberger, F. J. Chem. Phys. 1999, 110, 3502-3508. (g) Martin, N.; Sanchez, L.; Illescas, B.; Perez, I. Chem. Rev. 1998, 98, 2527-2547. 215. Nierengarten, J.-F. New J. Chem. 2004, 28, 1177-1191. 216. Nierengarten, J.-F.; Eckert, J.-F.; Nicoud, J.-F.; Ouali, L.; Krasnikov, V.; Hadziioannou, G. Chem. Comm. 1999, 617-618. 217. Peeters, E.; van Hal, P. A.; Knoll, J.; Brabec, C. J.; Sariciftci, N. S.; Hummelen, J. C.; Janssen, R. A. J. J. Phys. Chem. B 2000, 104, 10174-10190.
291 218. Nierengarten, J.-F. Sol. Energy Mater. Sol. Cells.2004, 83, 187-199. 219. Guldi, D. M.; Luo, C.; Swartz, A.; Gomez, R.; Segura, J. L.; Martin, N.; Brabec, C.; Sariciftci, N. S. J. Org. Chem. 2002, 67, 1141-1152. 220. Negishi, N.; Yamada, K.; Takimiya, K.; Aso, Y.; Otsubo, T.; Harima, Y. Chem. Lett. 2003, 32, 404-405. 221. Maggini, M.; Possamai, G.; Menna, E.; Scorrano, G.; Camaioni, N.; Ridolfi, G.; Cassalbore-Miceli, G.; Franco, L.; Ruzzi, M.; Corvaja, C. Chem. Comm. 2002, 2028-2029. 222. Nierengarten, J.-F.; Gu, T.; Aernouts, T.; Geens, W.; Prrotmans, J.; Hadziioannou, G.; Tsamouras, D. Appl. Phys. A 2004, 79, 47-49. 223. Loi, M. A.; Denk, P.; Hoppe, H.; Neugebauer, H.; Winder, C.; Meissner, D.; Brabec, C.; Sariciftci, N. S.; Gouloumis, A.; Vazquez, P.; Torres, T. J. Mater. Chem. 2003, 13, 700-704. 224. Honciuc, A.; Jaiswal, A.; Gong, A.; Ashworth, K.; Spangler, C. W.; Peterson, I. R.; Dalton, L. R.; Metzger, R. M. J. Phys. Chem. B 2005, 109, 857-871. 225. (a) Yang, J.-S.; Lin, Y.-D.; Lin, Y.-H.; Liao, F.-L. J. Org. Chem. 2004, 69, 3517-3525. (b) Lin, T,-C.; He, G. S.; Prasad, P. N.; Tan, L.-S. J. Mater. Chem. 2004, 14, 982-991. 226. Camps, X.; Hirsch, A. J. Chem. Soc. Perkin Trans. 1 1997, 1595-1596. 227. (a) Sandanayaka, A. S. D.; Sasabe, H.; Araki, Y.; Furusho, Y.; Ito, O.; Takata, T. J. Phys. Chem. A 2004, 108, 5145-5155. (b) Sandanayaka, A. S. D.; Matsukawa, K.; Ishi-I, T.; Mataka, S.; Araki, Y.; Ito, O. J. Phys. Chem. B 2004, 108, 19995-20004. 228. (a) Van Slyke, S. A.; Chen, C. H.; Tang, C. W. Appl. Phys. Lett. 1996, 69, 2160-2162. (b) Yakushchenko, I. K.; Kaplunov, M. G.; Efimov, O. N.; Belov, M. Y.; Shamev, S. N. Phys. Chem. Chem. Phys. 1999, 1, 1783-1785. 229. Hong, Z. R.; Lee, C. S.; Lee, S. T.; Li, W. L.; Shiroata, Y. Appl. Phys. Lett. 2002, 81, 2878-2880. 230. Ghosh, H. N.; pal, H.; Sapre, V.; Mittal, J. P. J. Am. Chem. Soc. 1993, 115, 11722-11727. 231. Li, J.; Sun, N.; Guo, Z.-X.; Li, C.; Li, Y.; Dai, L.; Zhu, D.; Sun, D.; Cao, Y.; Fan, L. J. Phys. Chem. B 2002, 106, 11509-11514. 232. Brabec, C. J.; Johannson, H.; Padinger, F.; Neugebauer, H.; Hummelen, J. C.; Sariciftci, N. S. Sol. Energy Mater. Sol. Cells 2000, 61, 19-33.
292 233. Loewe, R. S.; McCullough, R. D. Chem. Mater. 2000, 12, 3214-3221. 234. Jestin, I.; Frere, P.; Mercier, N.; Levillain, E.; Stievenard, D.; Roncali, J. J. Am. Chem. Soc. 1998, 120, 8150-8158. 235. Apperloo, J. J.; Martineau, C.; van Hal, P. A.; Roncali, J.; Janssen, R. A. J. J. Phys. Chem. B 2002, 106, 21-31. 236. Schwatrz, B. J. Annu. Rev. Phys. Chem. 2003, 54, 141-172. 237. Yamamoto, T.; Zhou, Z.; Kanbara, T.; Shimura, M.; Kizu, K.; Maruyama, T.; Nakamura, Y.; Fukuda, T.; Lee, B.-L.; Ooba, N.; Tomaru, S.; Kurihara, T.; Kaino, T.; Kubota, K.; Sasaki, S. J. Am. Chem. Soc. 1996, 118, 10389-10399. 238. Tanaka, S.; Yamashita, Y. Synth. Met. 1995, 69, 599-600. 239. Kitamura, C.; Tanaka, S.; Yamashita, Y. Chem. Mater. 1996, 8, 570-578. 240. Wiberg, K. B.; Lewis, T. P. J. Am. Chem. Soc. 1970, 92, 7154-7160. 241. Katriztky, A. R. Handbook of Heterocyclic Chemistry; Pergamon Press: Amsterdam, 2000. 242. Wang, L.-H.; Chen, Z.-K.; Xiao, Y.; Kang, E.-T.; Huang, W. Macromol. Rapid. Comm. 2000, 21, 897-900. 243. Yamamoto, T.; Lee, B.-L.; Kokubo, H.; Kishida, H.; Hirota, K.; Wakabayashi, T.; Okamato, H. Macromol. Rapid Comm. 2003, 24, 440-443. 244. Lee, B.-L.; Yamamoto, T. Macromolecules 1999, 32, 1375-1382. 245. Audebert, P.; Sadki, S.; Miomandre, F.; Clavier, G. Electrochem. Comm. 2004, 6, 144-147. 246. Streitwieser, A.; Schwager, I. J. Phys. Chem. 1962, 66, 2316-2320. 247. (a) Sonmez, G.; Meng, H.; Wudl, F. Chem. Mater. 2003, 15, 4923-4929. (b) Kitamura, C.; Tanaka, S.; Yamashita, Y. Chem. Mater. 1996, 8, 570-578. 248. Novak, Z.; Kotschy, A. Org. Lett. 2003, 5, 3495-3497. 249. Counotte-Potman, A.; van der Plas, H. C.; van Veldhuizen, B.; Landheer, C. A. J. Org. Chem. 1981, 46, 5102-5109. 250. Gleiter, R.; Schehlmann, V.; Spanget-Larsen, J.; Fischer, H.; Neugebauer, F. A. J. Org. Chem. 1988, 53, 5756-5762. 251. Chavez, D. E.; Hiskey, M. A. J. Energetic Mater. 1999, 17, 357-377.
293 252. Pinner, A. Ber. 1893, 26, 2128-2135. 253. Roger, R.; Neilson, D. G. Chem. Rev. 1961, 61, 179-211. 254. Wiley, R. H.; Jarboe, C. H.; Hayes, F. N. J. Org. Chem. 1957, 22, 835-836. 255. Neilson, D. G.; Roger, R.; Heatlie, J. W. M.; Newlands, L. R. Chem. Rev. 1970, 70, 151-170. 256. Wiley, P. F. In Chemistry 1,2,3-Triazines and 1,2,4-Triazines, Tetrazines, and Pentazines, Neunhoffer, H.; Wiley, P. F., Eds.; Wiley: New York, 1978, pp. 1073-1283. 257. Adbel-Rahman, M. O.; Kira, M. A.; Tolba, M. N. Tetrahedron Lett. 1968, 3871-3872. 258. Bowie, R. A.; Gardner, M. D.; Neilson, D. G.; Watson, K. M.; Mahmood, S.; Ridd, V. J. Chem. Soc. Perkin I 1972, 2395-2399. 259. Audebert, P.; Sadki, S.; Miomandre, F.; Clavier, G.; Vernieres, M. C.; Saoud, M.; Hapiot, P. New J. Chem. 2004, 387-392. 260. Hu, W.-X.; Rao, G.-W.; Sun, Y.-Q. Biorganic Med. Chem. Lett. 2004, 14, 1177-1181. 261. Novak, Z.; Bostai, B.; Csekei, M.; Lorincz, K.; Kotschy, A. Heterocycles 2003, 60, 2653-2668. 262. Huynh, H. V.; Hiskey, M. A.; Archuleta, J. G.; Roemer, E. L.; Gilardi, R. Angew. Chem. Int. Ed. 2004, 43, 5658-5661. 263. Soloducho, J.; Doskocz, J.; Cabaj, J.; Roszak, S. Tetrahedron 2003, 59, 4761-4766. 264. Krische, B.; Zagorska, M. Synth. Met. 1989, 28, C263-C268. 265. Rutan, K. J.; Heldrich, F. J.; Borges, L. F. J. Org. Chem. 1995, 60, 2948-2950. 266. van Leusen, A. M.; Jagt, J. C. Tetrahedron Lett. 1970, 967-970. 267. Bercot-Vatteroni, M. Ann. Chim. 1962, 7, 303-337. 268. Meltzer, R. I.; Lewis, A. D.; King, J. A. J. Am. Chem. Soc. 1955, 77, 4062-4066. 269. Ikemeto, N.; Estevez, I.; Nakanishi, K.; Berova, N. Heterocycles 1997, 46, 489-501. 270. Dâ€™Auria, M.; Mauriello, G. Tetrahedron Lett. 1995, 36, 4883-4884. 271. Buu-Hoi, N. P.; Lavit, D. Journal of the Chemical Society 1958, 1721-1723.
294 272. (a) Olah, G. A.; Keumi, T. Synthesis 1979, 112-113. (b) Seed, A. J.; Toyne, K. J.; Goodby, J. W. J. Mater. Chem. 1995, 5, 653-661. 273. Street, G. B. In Handbook of Conducting Polymers, Vol 1; Skotheim, T. A., Ed.; Marcel Dekker: New York, 1986, chapter 8. 274. Audebert, P.; Bidan, G. Synth. Met. 1986, 15, 9-22. 275. Meng, H.; Perepichka, D. F.; Bendikov, M.; Wudl, F.; Pan, G. Z.; Yu, W.; Dong, W.; Brown, S.; J. Am. Chem. Soc. 2003, 125, 15151-15162. 276. Eicher, T.; Hauptman, S. The Chemistry of Heterocylces, Wiley-VCH: Weinheim, 2003. 277. Alkan, S.; Cutler, C. A.; Reynolds, J. R. Adv. Funct. Mater. 2003, 13, 331-336. 278. Farago, J.; Novak, Z.; Schlosser, G.; Csampai, A.; Kotschy, A. Tetrahedron 2004, 60, 1991-1996. 279. Kaszynski, P.; Young, V. G. J. Am. Chem. Soc. 2000, 122, 2087-2095
BIOGRAPHICAL SKETCH Barry C. Thompson was born in Milwaukee, Wisconsin in 1977. After a brief stay in Minnesota, he grew up in Gallipolis, Ohio. Barry received his B.S. in chemistry and physics from the University of Rio Grande in Rio Grande, Ohio in 2000. Also in 2000, Barry received a National Science Foundation (NSF) Graduate Research Fellowship and moved to the University of Florida to study electroactive polymers with Prof. John Reynolds. 295