One-dimensional metals: physics and materials science

Denis Fichou

Handbook of Oligo- and Polythiophenes

@ WILEYWCH

Related titles from WILEY-VCH

K. Miillen / G. Wegner Electronic Materials: The Oligomer Approach ISBN 3-527-29438-4,WILEY-VCH 1998.

S. Roth One-Dimensional Metals ISBN 3-527-26875-8,WILEY-VCH 1995.

Denis Fichou

Handbook of Oligo- and Polythiophenes

@ WILEY-VCH Weinheim New York . Chichester . Brisbane * Singapore Toronto

Dr. Denis Fichou Laboratoire des MatCriaux MolCculaires C.N.R.S. 2, rue Henry-Dunant F-94320 Thiais France

This book was carefully produced. Nevertheless, authors, editor and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Library of Congress Card No. applied for. A catalogue record for this book is available from the British Library.

Deutsche Bibliothek Cataloguing-in-Publication Data: Fichou, Denis: Handbook of oligo- and polythiophenes I Denis Fichou. - Weinheim ; New York ; Chichester ; Brisbane ; Singapore ; Toronto : Wiley-VCH, 1999 ISBN 3-527-29445-7

0 WILEY-VCH Verlag GmbH, D-69469 Weinheim (Federal Republic of Germany), 1999

Printed on acid-free and chlorine-free paper All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form - by photoprinting, microfilm, or any other means nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Composition: Alden Bookset, Oxford, England Printing: betzdruck, D-64291 Darmstadt Bookbinding: GroBbuchbinderei J. Schaffer, D-67269 Griinstadt Printed in the Federal Republic of Germany

Preface

At the eve of the 21st century and after twenty years of maturation, the world of conjugated polymers and oligomers is a flourishing branch of materials science with many opportunities for applications in electronics and photonics. Polyaniline, poly( p-phenylenevinylene) and polythiophene are among the most investigated conjugated polymers that combine the electronic and optical properties of semiconductors with the processing advantages and mechanical plasticity of conventional polymers. Depending upon their doping level, these versatile materials behave either as metallic conductors or semiconductors, can be chromophores or luminophores and may even develop large optical nonlinearities. When doped to metallic levels, conjugated polymers become highly conducting and may find applications in batteries, electrochromic or smart windows, electromagnetic shields, antistatic coatings and various types of sensors. On the other hand, when in the semiconducting form they exhibit similar electrical and optical properties as inorganic semiconductors. High performance optoelectronic devices fabricated from conjugated polymers such as light emitting diodes, field-effect transistors, photodetectors, photovoltaic cells, optocouplers and light modulators have been demonstrated. Although most of these polymer products still face technical problems and are not yet commercialized, they preconceive what could be in a near future the world of “plastic electronics”. Oligo- and polythiophenes (PT) present all aspects of a rich and homogeneous family of conjugated compounds, thanks to the extraordinary fecundity of thiophene chemistry. Since the discovery of conducting PT in 1982 at CNRS in Thiais, France, a tremendous number of substituted derivatives have been synthesized and their electronic properties investigated. If one of the early goals has been to improve the conductivity by controlling the growth and structure of the polymer, very rapidly new targets emerged. Grafting an adequate substituent on the main PT chain on a lateral carbon site provides an additional property such as solubility which is required to prepare freestanding films on any surface. Other substituents allow to introduce optical, magnetic or liquid crystalline properties. Beside, “functionalized” PTs combine electrical conductivity together with a second activity that can be triggered by electricity. Depending on this functionalization, PT derivatives can operate complex functions like for example selective recognition of biomolecules (DNA, oligonucleotides). Another important research field aims at controlling the molecular and structural ordering of semiconducting PT in view of improving its charge transport properties. A major advance in this direction has been realized in 1987 at CNRS, Thiais, with the synthesis of sexithiophene (ST), the linear hexamer of thiophene, and its use to fabricate an organic transistor whose performances are close to those of siliconbased devices. The spectacular increase of the carrier mobility in polycrystalline

VI

Preface

6T films as compare to disordered PT is the result of three criteria generally met by low-molecular weight oligomers: 1. high molecular order (defect-free molecules), 2. high chemical purity (up to electronic grade) and 3. high structural order in the solid state (up to single crystals). The concept of well-defined oligomers was born and rapidly extended to other compounds (arylenevinylenes, polyenes, acenes, etc.. . .) to turn into one of the most successful routes in the modern world of conjugated organics. This Handbook summarizes in ten chapters all aspects of oligo- and polythiophenes as they developed over the last twenty years, from chemistry to physics and applications. It has been written by the most reknown experts in the field worldwide, from both academics and industrial origins, with a constant care of clarity through concise texts and an extensive use of figures and tables. This first review on PTs and oligomers constitutes a comprehensive tool not only for researchers but also for advanced students and anyone willing to get informations on this novel class of materials. Denis Fichou October 1998

Contents

1

The Chemistry of Conducting Polythiophenes: from Synthesis to Self-Assembly to Intelligent Materials 1 Richard D. McCullough

1.1 1.2 1.3 1.3.1 1.3.1.1 1.3.1.1.1 1.3.1.1.2 1.3.1.2 1.3.1.3 1.3.1.3.1 1.3.1.4 1.3.1.4.1 1.3.1.4.2 1.3.1.4.3 1.3.1.4.4 1.3.1.4.5 1.3.1.4.6 1.3.1.4.7 1.3.1.4.8 1.3.1.4.9

Introduction 1 Chemical Synthesis of Unsubstituted Polythiophene (PT) 2 Chemcial Synthesis of Polyalkylthiophenes (PATS) 5 Straight Alkyl Side Chains 5 Chemical Synthesis of PATS 5 Metal Catalyed Cross-coupling Polymerizations 5 FeC13 Method for the Polymerization of PATs 6 Comparison of the Above Methods 8 Regioregular PATs 9 Regioregular HH-TT and TT-HH PATs 10 Regioregular, Head-to-Tail Coupled PATs 12 The McCullough Method 12 The Rieke Method 13 The Mechanism and Catalyst Choice 15 NMR Characterization of HT-PAT 15 IR and UV-Vis 16 Self-Assembly, X-ray, and Electrical Conductivity in HT-PATS 18 Other Methods 19 Random Copolymers of Alkyl Thiophenes 20 Head-to-Tail Coupled, Random Copolymers of Alkyl Thiophenes 20 Branched Alkyl PATs 21 PTs with Phenyl Sidechains 24 Chemical Synthesis of Heteroatomic Functionalized Substituents on PTs: Recognition Sites for Self-Assembly and Chemical Sensing 24 Chemical Synthesis of Alkoxy Polythiophenes 25 Chemical Prepared Alkoxy PTs as Conducting Polymer Sensors 27 Chiral Substituents on PT 32 Carboxylic Acid Derivatives: Self-Assembly and Sensors 33 Other Derivatives of PT 34 Fused Rings Systems 38 Conclusion 39 References 39

1.3.1.5 1.3.1.6 1.4 1.4.1 1.4.2 1.4.3 1.4.4 1.4.5 1.5 1.6

VIII

Contents

2

Electronic Properties of Polythiophenes 45 Shu Hotta and Kohzo It0

2.1 2.2 2.2.1 2.2.2 2.3 2.3.1 2.3.2 2.3.3

General Aspects of Conducting Polymers 45 Structure and Conformation of Polythiophenes 48 Morphology and Crystal Structure 48 Conformational Features 52 Electronic Processes of Polythiophenes 57 Charge Excitations in Polythiophenes 57 Charge Transport in Polythiophenes 60 Carrier Recombination: Photoluminescence and Electroluminescence 63 Spectroscopic Studies of the Charged States 65 Charge Storage Configurations in Solids and their Anisotropic Properties 65 Properties in Solutions 73 Concluding Remarks and Future Outlook 80 Acknowledgments 82 References 82

2.3.4 2.3.4.1 2.3.4.2 2.4

3

The Synthesis of Oligothiphenes 89 Peter Bauerle

3.1 3.2 3.2.1 3.2.1.1 3.2.1.2 3.2.1.3 3.2.1.4 3.2.2 3.2.2.1 3.2.2.2 3.2.2.3 3.2.2.4 3.2.2.5 3.2.2.6 3.3

Introduction 89 Synthesis of Oligothiophenes 93 Unsubstituted Oligothiphenes 93 Arene/arene-Coupling Methods by Oxidative Couplings 93 Transition Metal Catalyzed Coupling Methods 97 Ring Closure Reactions from Acyclic Precursors 104 Physical Properties of a-oligothiphenes and Isomers 111 Substituted Oligothiophenes 118 ,&@'-Substituted Oligothiophenes 119 a,a'-Substituted Oligothiphenes 139 a$-Substituted Oligothiphenes 145 Functionalized Oligothiphenes 155 Amphiphilic Oligothiphenes 170 Transition Metal Complexes of Oligothiophenes 171 Conclusion 172 Acknowledgement 173 References 173

4

Structure and Properties of Oligothiophenes in the Solid State: Single Crystals and Thin Films 183 Denis Fichou and Christiane Ziegler

4.1 4.2

Introduction 183 Single Crystals 184

Contents

4.2.1 4.2.2 4.2.2.1 4.2.2.2 4.2.2.3 4.2.2.4 4.2.2.5 4.2.2.6 4.2.2.7 4.2.3 4.2.3.1 4.2.3.2 4.2.3.3 4.2.3.4 4.3 4.3.1 4.3.1.1 4.3.1.2 4.3.2 4.3.2.1 4.3.2.2 4.3.2.3 4.3.2.4 4.3.2.5 4.3.2.6 4.3.2.7 4.3.3 4.3.3.1 4.3.3.2 4.3.4 4.3.4.1 4.3.4.2 4.3.4.3 4.3.4.4 4.3.4.5 4.3.4.6 4.3.4.7 4.3.4.8

General Description 184 X-ray Structures 187 Bithiophene (a-2T) and Derivatives 187 a-Terthiophene (a-3T) and Derivatives 191 a-Quaterthiophene (a-4T) and Derivatives 193 a-Quinquethiophene (a-5T) and Derivatives 202 a-Sexithiophene (a-6T) and Derivatives 203 a-Octithiophene (a-8T) 207 Polythiophene and 3-Alkylated Derivatives 2 13 Optical and Electrical Properties 214 General Remarks 214 Dimethylquarterthiophene 2 14 a-Sexithiophene (a-6T) 215 a-Octithiophene (a-8T) 216 Thin films 220 Deposition Techniques 220 Vacuum deposition Techniques 220 Preparation from Solution 220 Morphology 221 General Remarks 221 Monothiophene (IT) and Derivatives 222 Small Oligomers (a-2T-a-4T) and Derivatives 223 Quinquethiophene (a-5T) and Derivatives 226 Sexithiophene (a-6T) and Derivatives 234 Longer Obligothiophenes (a-7T, a-8T) and Derivatives 244 Polythiophene and Derivatives 245 Optical Characterization 247 Undoped Oligothiophenes 247 Charges in Oligothiophenes 257 Electrical Characterization 266 General Remarks 266 Contacts, I/V-Curves, Carrier Injection 267 Influence of the Structure on Conductivity Data 269 Influence of the Structure on Mobility Data 270 Temperature Dependence 27 1 Conjugation Length Influence 272 Influence of Dopants 272 Photoconductivity 274 References 274

5

Charge Transport in Semiconducting Oligothiophenes 283 Gilles Horowitz and Phillippe Delannoy

5.1 5.1.1

Basic Models 284 The Band Model 284

TX

X

Contents

5.1.2 5.1.2.1 5.1.2.2 5.1.2.3 5.1.3 5.1.3.1 5.1.3.2 5.1.3.3 5. I .4 5.1.5 5.2 5.2.1 5.2.2 5.2.3 5.2.3.1 5.2.3.2 5.2.3.3 5.2.4 5.3 5.3.1 5.3.1.1 5.3.1.2 5.3.1.3 5.3.2 5.4

Hopping 288 Localization 288 Temperature Dependence 289 Field Dependent Mobility 291 Polarons 291 Small Polaron 291 Molecular ‘Nearly Small’ Polaron 293 Polarons in n-Conjugated Polymers and Oligomers 295 Multiple Trapping 297 Summary 297 Measurement of the Mobility 298 Conductivity 298 Time of Flight 299 Space-Charge-Limited Current 300 Profile of Injected Charges 301 Estimation of the Space-Charge Limited Current 302 Effect of Traps 303 Field-Effect 305 Transport properties of Oligothiophenes 306 Conductivity, Mobility and Carrier Density 307 Variation with Chain Length 307 Carrier Density 308 Variation with Temperature 309 Traps 311 Concluding Remarks 3 12 References 3 13

6

Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes: Relation to Their Use in Electro-optic and Photonic Devices 317 J . Cornil, D. Beljonne, V. Parente, R. Lazzaroni, and J. L. Brtdas

6.1 6.2 6.3

Introduction 3 17 Theoretical Methodology 320 Electronic and Linear Optical properties of Neutral Oligothiophenes 322 Nature of the Lowest Excited States 322 Intersystem Crossing Processes 324 Lattice Relaxation Phenomena 326 Effects of Substitution 328 Electronic and Linear Optical Properties of Charged Oligothiphenes 333 Characterization of Metal/polymer Interfaces 339 Geometric Structures 340 Electronic Structures 343

6.3.1 6.3.2 6.3.3 6.3.4 6.4 6.5 6.5.1 6.5.2

Contents

6.5.3 6.6 6.6.1 6.6.2 6.7

Vibrational Signature 344 Nonlinear Optical Properties of Neutral Oligothiophenes 347 Chain Length Dependence of the third-order Polarizabilities in Thiophene Oligomers 349 Dynamic Third-order Response of Th7: Two-photon Absorption and Third-harmonic Generation 352 Synopsis 355 Acknowledgements 355 References 357

7

Electronic Excited States of Conjugated Oligothiophenes 361 Carlo Taliani and Wolfram Gebauer

7.1 7.2 7.2.1 7.2.2 7.3 7.3.1 7.3.2 7.3.2.1 7.3.2.2 7.3.2.3 7.3.2.4 7.3.3 7.4 7.4.1 7.4.2 7.4.2.1 7.4.2.2 7.4.3

Introduction 361 Electronic Structure of Conjugated Polymers 362 General Concept 362 Polythiophene 363 Oligothiophene Model Structure 364 Molecular Structure 364 Singlet States 367 Assignments 367 Chain Length Dependence 368 Nature of the Lowest Singlet Transition 369 Franck-Condon Coupling 370 Triplet States 372 Solid State Properties 373 Molecular Packing 373 Theoretical Approach 374 The Exciton Concept and the Lowest Excited State in 6T 374 Higher Transitions - Extended States 379 Experimental Evidence for the Nature of the Lowest Excited States 380 Structural and Morphological Aspects of Polycrystalline Thin Films 380 Optical Properties of Thin Polycrystalline Films 384 Highly Ordered Systems 387 Two-photon Excitation 392 Extended States 392 Triplet States 393 Excited States Ordering 394 Polarized Electroluminescence 395 Nonlinear Optical Properties of Polythiophene and Thiophene Oligomers 397 Acknowledgements 400 References 400

7.4.3.1 7.4.3.2 7.4.3.3 7.4.3.4 7.4.3.5 7.4.3.6 7.4.3.7 7.5 7.6

XI

XI1

Contents

8

Electro-optical Polythiophene Devices 405 Magnus Granstrom, Mark G. Harrison, and Richard H . Fiend

8.1 8.1.1 8.2 8.2.1 8.2.2 8.2.3 8.2.4 8.2.4.1 8.2.4.2 8.2.5 8.3 8.3.1 8.3.2 8.3.2.1 8.3.2.2 8.3.2.3 8.3.3 8.3.3.1 8.3.3.2 8.3.4 8.3.4.1 8.3.4.2 8.3.5 8.4 8.4.1 8.4.2 8.4.2.1 8.4.2.2 8.4.3 8.4.4 8.4.4.1

Overview 405 Relationship Betweenpolymers and Oligomers 405 Preparation of Thin Film Devices 408 Introduction 408 Polymers 408 Oligomers 409 Relative Merits of the Different Methods to Achieve Solubility 410 Substitution with Side-chains 410 Using a Soluble Partially-conjugated Precursor Polymer 412 Blends Between Polymer and Oligomers 413 Electronic Excitations in Oligothiophenes 413 Introduction 413 Intra-molecular Non-radiative Decay Channels 413 Internal Conversion 415 Intersystem Crossing 4 16 Singlet Fission 417 Inter-molecular Non-radiative Decay Channels in Thin Films 41 7 Aggregation and Davydov Splitting 417 Charge-transfer Excitons 418 Effects of Inter-ring Torsion and Coplanarity of Oligomers 419 Solution 420 Solid State 420 Concluding Remarks 42 1 Electroluminescent Devices 421 Introduction 421 Historical Survey of Organic LEDs 424 LEDs Based on Molecular Semiconductors 424 Polymeric LEDs 425 LEDs Based on Oligothiphenes 427 LEDs Based on Polythiophenes 429 Polythiophene LEDs Covering the Whole Visible Spectrum and a Bit More 430 Intrinsically-polarised Polymer LEDs 432 Polythiophenes in Microcavity Structures 434 Sub-wavelength Size Polymer LEDs 436 Voltage-controlled Colours 437 Photoconductive and Photovoltaic Devices 439 Introduction 439 Mechanism of Photoconductivity in Sexithiophene 440 Photovoltaic Applications (Solar Cells) 441 Photovoltaic Devices Based on Polythiophenes 443 Electro-Optical Modulator Devices 444 Optical Probing of Field-induced Charge in a-Sexithiphene 446

8.4.4.2 8.4.4.3 8.4.4.4 8.4.4.5 8.5 8.5.1 8.5.2 8.5.3 8.5.4 8.6 8.6.1

Contents

8.7

All-optical Modulator and Memory Devices 449 References 452

9

Oligo- and Polythiophene Field Effect Transistors 459 H. E. Katz, A . Dodabalapur and Z . Bao

9.1 9.2 9.3 9.3.1 9.3.2 9.3.3 9.3.4 9.3.5 9.3.6 9.4 9.4.1 9.4.2 9.4.3 9.4.4 9.5 9.5.1 9.5.2 9.5.3 9.6 9.7

Introduction 459 Operation of a Field-effect Transistor 460 Modeling of Oligothiphene TFTs 461 Analytical Modeling 46 1 Numerical Modeling 463 Interface Effects 464 Short-channel Effects 465 Sub-threshold Characteristics 466 Energy Levels 467 Oligothiophene FETs 468 Synthesis and Purification 468 Morphology 47 1 Substituted Oligothiophnenes 473 Fused Ring Materials 475 FETs Based on Polythiophenes 476 Regiorandom Polythiophene FETs 478 Regioregular Polythiophene FETs 478 All-printed Plastic FETs 481 Heterojunction FETs 483 Summary 485 Acknowledgments 486 References 486

10

Application of Electrically Conductive Polythiophenes 491 Gerhard Kossmehl and Gunnar Engelmann

10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9 10.9.1 10.9.2 10.9.3

Introduction 491 Conducting Materials 492 Antistatic Coatings 495 Electromagnetic Shielding Materials 496 Materials for Rechargeable Batteries, Capacitors 497 Junction Devices and Rectifying Bilayer Electrodes 501 Resists, Recording Materials and Fabrication of Patterns Electrochromic Devices 503 Sensors 506 Sensors for Gases 506 Sensors for Ions in Aqueous Solution 507 Sensors for Organic Materials 508

50 1

XI11

XIV

Contents

10.9.4 10.10 10.10.1 10.11

Sensors for Bio-organic Materials 512 Other Applications 512 General Consideration 513 Summary, Conclusions and Future Trends References 5 17

Index 525

516

List of Contributors

Peter Bauerle Institute of Organic Chemistry II University of Ulm Albert-Einstein-Allee 11 D-89081 Ulm Germany

Richard D. McCullough Department of Chemistry Carnegie Mellon University 4400 Fifth Avenue Pittsburgh, PA 15213-2683 USA

Z. Bao AT & T. Bell Laboratories Lucent Technologies 600 Mountain Avenue Murray Hill NJ 07974 USA

Philippe Delannoy Groupe de Physique des Solides Universites Paris 6 (Pierre et Marie Curie) et Paris 7 (Denis Diderot) 2 Place Jussieu 75251 Paris Cedex 05 France

D. Beljonne Service de Chimie des MatCriaux Nouveaux Universite de Mons-Hainaut Place du Parc 20 7000 Mons Belgium

A, Dodabalapur AT & T. Bell Laboratories Lucent Technologies 600 Mountain Avenue Murray Hill, NJ 07974 USA

Jean-Luc Bredas Service de Chimie des Mattriaux Nouveaux Universitt de Mons-Hainaut Place du Parc 20 7000 Mons Belgium

Gunnar Engelmann Institute of Organic Chemistry Freie Universitat Berlin Tokustrasse 3 D-14195 Berlin Germany

J. Cornil Service de Chimie des Mattriaux Nouveaux Universitk. de Mons-Hainaut Place du Parc 20 7000 Mons Belgium

Denis Fichou Laboratoire des Mattriaux Moliculaires CNRS 2 rue Henry-Dunant 94320 Thiais France

XVI

List of Contributors

Richard H. Friend University of Cambridge Department of Physics Cavendish Laboratory Madingly Road Cambridge CB3 OHE United Kingdom

K o b o Ito Department of Applied Physics Faculty of Engineering University of Tokyo 7-3-1 Hongo, Bunkyo-ku Tokyo 113 Japan

Wolfram Gebauer C.N.R. Instituto di Spettroscopia Molecolare Via Castaguloi 1 40126 Bologna Italy

H. E. Katz AT & T. Bell Laboratories Lucent Technologies 600 Mountain Avenue Murray Hill, NJ 07974 USA

Magnus Granstrom University of Cambridge Department of Physics Cavendish Laboratory Madingly Road Cambridge CB3 OHE United Kingdom

Gerhard Kossmehl Institute of Organic Chemistry Freie Universitat Berlin Takustrasse 3 D-14195 Berlin Germany

Mark G. Harrison University of Cambridge Department of Physics Cavendish Laboratory Madingly Road Cambridge CB3 OHE United Kingdom

L. Lazzaroni Service de Chimie des MatCriaux Nouveaux Universitk de Mons-Hainaut Place du Parc 20 7000 Mons Belgium

Gilles Horowitz Laboratoire des Matkriaux Molkculaires CNRS 2 rue Henry-Dunant 94320 Thiais France

V. Parente Service de Chimie des MatCriaux Nouveaux Universitt: de Mons-Hainaut Place du Parc 20 7000 Mons Belgium

Shu Hotta National Institute of Materials and Chemical Research Japan High Polymer Center 1-1 Higashi, Tsukuba, Ibaraki 305 Japan

Carlo Taliani C.N.R. Instituto di Spettroscopica Molecolare Via Castagnoli, 1 40126 Bologna Italy

List ojcontributors

Christiane Ziegler University of Tubingen Institute of Physical and Theoretical Chemistry Auf der Morgenstelle 8 D-72076 Tubingen Germany

XVII

Biography

Denis Fichou is a directeur de recherche at CNRS in Thiais, France. He received a Doctorat de 36me Cycle in organic chemistry at the University of Rennes, France, in 1981 and a Doctorat d’Etat in physical sciences at the University of Paris VI in 1986. He joined CNRS in 1982 at the Laboratory of Molecular Materials in Thiais. In 1986and again in 1992,he spent two years in T6ky8, Japan, as the CNRS Advisor of the Chemistry Department. In 1987, he initiated the successful “oligothophenes route” at CNRS, Thiais. His current research interests focus on material chemistry and the fabrication of electronic and photonic organic devices, particularly thin film transistors and laser crystals.

List of Symbols

a a

C

C D D d

D e E

4 Epa

For F f(E1

f*

G h H I j or J J k k

L m0

M MN MW

n n N N Nch

Nf Ns

P

lattice constant lattice vector capacitance electron-continuum coupling density dichroism thickness of sample diffusion coefficient electron charge transition enegry intrinsic semiconductivity oxidation potential applied electric field Fermi function relaxation frequency charge generation Planck constant Hamiltonian operator current current density overlap integral wave vector Boltzmann constant Kerr response function non-radiative decay rate channel length electron mass dipole moment number average molecular weight weight average molecular weight density of carriers refractive index number of repeat units in a chain number of molecules number of injected charges density of states at the Fermi level number of spins charge density

xx

List of Symbols

4 I

R R

s

T V V

V V

z z

A(r) 0 (cap theta) 0 a!

P X 6

E

4 4F

Y 77 n

A ~max

P U

0 P U

7

w

1c. a,

charge separation between molecular centres charge recombination distance between sites strain constant temperature or absolute temperature vibrational frequency velocity voltage intermolecular interactions number of molecules/unit cell channel width distortion angle between surface and molecules Debye temperature lattice constant Poole-Frenkel factor susceptibility tensor molar absorption coefficient dielectric constant potential fluorescence quantum yield cubic nonlinearity internal quantum efficiency absorption coefficient mean free path wavelength of absorption mobility band maximum fraction of charges free to move resistivity conductivity relaxation time optical phonon frequency polaron wave function wave function fluorescence

List of Abbreviations

acac AFM AM 1 BBN BCB BZ CASSCF CASPTZ CB CI CNDOjCI cod CTE

cv cv

DDQ DFT DFWM DMAC DMF dmso DOS dPPP EDOT EFISH EL ELS EPR ESR EVS FEBS FET FTIR GPC HB HCM HH HOMO HOPG

acetylacetonate atomic force microscopy Austin model 1 9 borabicyclo[3.3.llnonane benzocyclobutene Brillouin zone Complete active space self-consistent field Multiconfiguration second-order perturbation theory conduction band configuration interaction complete neglect of differential overlap/configuration interaction cyclooctadiene charge transfer electrons capacitance-voltage cyclic voltammetry dichlorodicyanoquinone density functional theory degenerate four wave mixing N,N’-dimethylacetamide N,N-dimethylformamide dimethyl sulfoxide density of states 1,3-diphenylphosphinopropane 3,4-ethylenedioxythiophene electronic field induced second harmonic generation electrolurninescenc electron energy 10s electron paramagnetic resonance electron spin resonance electrochemical voltage spectroscopy frequency domain electric birefringence spectroscopy field effect transistor Fourier transform infrared gel phoresis chromatography herringbone hydroquinonemethylether head-to-head (coupling) highest occupied molecular orbital highly oriented pyrolytic graphite

XXII

List of Abbreviations

HPLC HREELS HT HV INDO IR ISC IT0 L.R. LDA LED LEED L.R. LSDA LUMO M-I MIS MNDO MO MOS MP2 MRD-CI NBS NEXAFS NLO NMP NMR OASLM ODMR OFET OLED P3-BTSNa P3-ETSNa P3-TPSNa PAT PBD PBT PC PC PCHMT PCHT PDBBT PDDT PDDUT PDHBT PDOBT

high pressure liquid chromatography high resolution energy electron loss spectroscopy head-to-tail (coupling) high vacuum intermediate neglect of differential overlap infrared inter system crossing indium-doped tin oxide Lawesson’s reagent lithium diisopropylamide light emitting diodes low energy electron diffraction Lawesson’s Reagent local spin density approximation lowest unoccupied molecular orbital metal-insulator metal-insulator-semiconductor modified neglect of differential overlap molecular orbital metal-oxide-semiconductor Moller-Plesset perturbation theory multi reference double configuraton interaction N- bromosuccinimide near edge X-ray absorption fine structure nonlinear optics 1-methyl-2-pyrrolidone nuclear magnetic resonance optically-addressed SLM optically detected magnetic resonance organic FET organic LED sodium poly(3-thiophene-~-butanesulfonate) sodium poly(3-thiophene-/3-ethanesulfonate) sodium poly(3-(3-thienyl)propanesulfonate) poly(3-alkylthiophene) 2-(4-biphenylyl)-5-(4-tert-butylphenyl)-1,3,4-0xadiazole poly(3-butylthiophene) photoconductivity propylene carbonate poly(3-cyclohexyl-4-methylthiophene) poly(3-cyclohexylthiophene) poly(4,4’-dibutyl-2,2’-bithiophene) poly(3-dodecylthiophene) poly(3-(2-(N-dodecyl-carbamoyloxy)ethyl)thiophene) poly(3,3’-dihexyl-2,2‘-bithiophene) poly(4,4/-dioctyl-2,2’-bithiophene)

PDT PEDOT PHT PiBT PL PLED PMMA PMT POPT POT PPV PT PTOPT PVK RJ ROHF RPE SAM SCF SCLC SCRF SE SERS SFM SHG SLM SOMO SNOM SSH STM TCNQ TEB TFT THF THG THP TMS TOF TPE TT UHV UPS UV VB VEH Vis

poly(3-decylthiophene) poly(3,4-ethylenedioxythiophene) poly(3-hexylthiophene) poly(3-isobutylthiophene) photoluminescence polymer light emitting diodes pol ymethylmethacrylate poly(3-methylthiophene) poly(4-octylpheny1)thiophene poly(3-octylthiophene) poly(p-phenylenevinylene) polythiophene poly[3-(4-octylpheny1)2,2’-bithiophene] poly(9-vinyl carbazole) refractive index restricted open shell Hartree-Fock electron paramagnetic resonance self-assembled monolayer self consistent field space-charge limited current self consistent reaction field stimulated emission surface enhanced Raman spectroscopy scanning force microscopy second harmonic generation spatial light modulators singly occupied molecular orbital scanning nearfield optical microscopy Su, Schrieffer and Heeger scanning tunneling microscopy 7,7,8,8-tetracyanoquinodimethane transient electric birefringence thin film transistor tetrahydro furan third harmonic generation tetrahydrop yrany 1 trimethylsilyl time of flight two photon excitation tail-to-tail (coupling) ultra high vacuum UV photoelectron spectroscopy ultraviolet valence band valence effective Hamiltonian visible

XXIV Vis VRH XRD

List of Abbreviations

visible variable range hopping X ray diffraction

1 The Chemistry of Conducting Polythiophenes: from Synthesis to Self-Assembly to Intelligent Materials Richard D. McCullough

1.1 Introduction In the late 1970s, conjugated polymers were proclaimed as futuristic new materials that would lead to the next generation of electronic and optical devices. It now appears with the discoveries of, for example, polymer light emitting diodes (LEDs) [l] and organic transistors [2] that new technologies are eminent. Polythiophenes are an important representative class of conjugated polymers that form some of the most environmentally and thermally stable materials that can be used as electrical conductors, non-linear optical devices, polymer LEDs, transistors, electrochromic or smart windows, photoresists, antistatic coatings, sensors, batteries, electromagnetic shielding materials, artificial noses and muscles, solar cells, electrodes, microwave absorbing materials, new types of memory devices, batteries, nanoswitches, optical modulators and valves, imaging materials, polymer electronic interconnects, nanoelectronic and optical devices [3,4].Polythiophene and its derivatives work very well in some of the above applications and less impressively in other devices. Creative new design and development strategies of new polythiophenes has led to captivating new materials and enhanced performance in certain devices. The ability of molecular designers to begin to understand how to gain control over the structure, properties, and function in polythiophenes continues to make the synthesis of polythiophenes a critical subject in the development of new advanced materials. Here we attempt to review the synthesis of polythiophenes comprehensively. Due to the enormous literature on the synthesis of polythiophenes, we are sure that excellent work in this area will be inadvertently overlooked. However, we will highlight both the pioneering work and the frontier in the synthesis of pol ythiophenes. It is important to realize that, as it has become clear that structure plays a dominant role in determining the physical properties of conducting polymers, more research has focused on directing structure and function of these materials with synthesis. Synthesis can help to determine the magnitude of 7r overlap along the backbone and eliminate structural defects. Materials assembly (and/or processing) determines interchain overlap and dimensionality. Planarization of the backbone and assembly of the backbone in the form of 7r stacks lead to better materials and enhanced device performance in almost every category ranging from electrical conductivity to stability. Therefore, both remarkable enhancements in the electronic and photonic properties of the resultant materials and the creation of new functions,

2

1 The Chemistry of Conducting

such as new sensory materials, critically depends on the synthesis of the polythiophene. This of course leads to the exciting prospect that the properties of polythiophenes can be selectively engineered through synthesis and assembly. A large portion of both the pioneering and future work in conjugated polymers strongly depends on synthetic chemists creating new polymers that can be fabricated into new devices and whose physics and chemistry can be deeply understood.

1.2 Chemical synthesis of unsubstituted polythiophene (PT) One of the first chemical preparations of unsubstituted polythiophene (PT) was reported in 1980 by two groups [5, 61. Both synthesized polythiophene by a metal catalyzed polycondensation polymerization of 2,5dibromothiophene (Scheme 1). Yamamoto's synthesis treats 2,5-dibromothiophene (1) with Mg in THF in the presence of nickel(bipyridine) dichloride. The Mg reacts with either bromide to form either 2-bromo-5-magnesiobromothiopheneor 2-magnesiobromo-5-bromothiophene, which is self-coupled with the Ni(1I) catalyst to form a thiophene dimer carrying a MgBr at one end and a Br at the other. This condensation reaction is propagated and eventually low molecular weight PT is formed. The polymerization is the extension of Kumada coupling of Grignard reagents to aryl halides [7]. Since PT, even at low molecular weights, is insoluble in THF, the precipitation of the polymer under the above reaction conditions limits the formation of higher molecular weights. The PT synthesized by this method leads to 78% insoluble polymer that does not melt. The soluble fraction is lower molecular weight oligomers. Polythiophene polymer of molecular weight greater than 3000 are not soluble in hot chloroform [8]. The elemental analysis of this polymer indicated 1-3% Mg remains in the polymer sample. Similar results were found by Lin and Dudek. Polymerization of 2,5-dibromothiophene in the presence of Mg in THF using either Yamatnoto Route

L b and Dud& ROW MflH F

Scheme 1. The first chemical syntheses of polythiophene.

Chemical synthesis of unsubstituted polythiophene ( P T ) e.g. Mg or Zn

Niocatalyst

3 X = CI, Br, or I

3

wn 2

Scheme 2. Polycondensation dehalogenation route to polythiophene.

palladium(a~ac)~ (acac = acetylacetonate) or Ni(acach or C o ( a c a ~ )or~ Fe(acac)3 catalyst yields low molecular weight PT containing at 3% impurities as determined by elemental analysis. Polymerization of 2,5-dihalothiophene can be accomplished by reacting the generated bromo-Grignard of thiophene with Ni(I1) catalyst such as Ni(dppp)C12 (dppp = l73-dipheny1phosphinopropane)or the 2,5dihalothiophene can be polymerized by a polycondensation dehalogenation reaction with Ni(0) (Scheme 2). Systematic studies of the polymerization of 2,5-dihalothiophene (3) have subsequently been done by primarily Yamamoto [8-121 and others [13-151. Varying the amounts of Mg [13], the solvent [lo, 14, 151, the type of metal (i.e. Mg, Zn, etc.) [lo], concentration of monomer [13], the type halogen on the monomer [8, 12-15], the temperature [S, 9, 121, reaction time [8], and the type of catalyst used [8-131 has led to some good chemical methods for the synthesis of PT. The extension of these chemical methods to the synthesis of poly(3-alkylthiophene)~(PATS)and other polythiophenes will be later noted. It is seen in a paper by Wudl [14] that very good samples of PT can be prepared by the polymerization of highly purified 2,5diiodothiophene (Scheme 3). First 2,5diiodothiophene (4) is reacted with Mg in ether at reflux. The preformed

Wudl

,A, 1. Mg/ether/reflux '

\

-

S

/

*

2. anisole

*&sA \ 'S'

4

Ni(cod)$PPha

B r S a B r

1

DMF 60"-80"C, 16h

2

sugimoio and Yoshlm

FeC13 CHC13

5

2

Scheme 3. Specific examples of the synthesis of polythiophene.

4

I The Chemistry of Conducting

iodomagnesioiodothiophene is isolated as a residue and redissolved in hot anisole, whereupon Ni(dppp)C12is added and the mixture heated at 100°Cfor 5 h to induce polymerization. Extensive washing of the isolated PT with methanol, chloroform, THF, and chlorobenzene leads to the isolation of PT with elemental analysis within 0.3% of the calculated values for C188H971S46 (molecular weight z 4 K or 46 thiophene rings and 1 butadiene unit). This h g h purity PT sample contains barely 50ppm of Mg and Ni. However, it is proposed that the one butadiene unit arises from a desulfurization reaction promoted by Ni(0) intermediates [7]. Polymerization of the 2,5-dibromothiophene yielded PT that analyzed 2-3% low in sulfur, apparently due to said desulfurization. The Wudl sample of PT was characterized by IR, ESR, conductivity and thermopower measurements. The conductivity of the AsF5-doped material was about 10 S cm-' . Work on the polycondensation polymerization of 2,5-dihalothiophenes by Y amamot0 has shown that essentially a quantitative yield of PT can be made from 2,5-dibromothiophene, N i ( ~ o d )(cod ~ = cyclooctadiene), and PPh3 at 60-80°C in DMF (Scheme 3) [8]. It is also reported that the percentage of Br end groups decreases as reaction times are increased from 8 to 16 h, indicating that perhaps some seemingly insoluble PT continues to grow. Both less active catalysts such as Ni(PPh3)4 and less active monomers such as 2,s-dichlorothiophene lead to lower yields of PT. The PT synthesized is exclusively coupled at the 2,s-carbons as indicated by solid state 13C NMR which exhibits peaks at 136 and 125ppm only. Other synthetic methods can produce the conjugation disrupting 2,4-coupled polythiophene structure. Whle the elemental analyses for carbon and hydrogen are within 0.3%, the sulfur content of the PT is off by 3%. Vacuum deposition of PT (estimated molecular weight of 1.5-2K) onto carbon, gold, KBr, or aluminum at 250-300°C at Pa can be accomplished. Electron diffraction patterns of PT on carbon indicates the formation of crystalline PT with the PT chains arranged perpendicular to the carbon substrate - similar to oligothiophene films. Vacuum deposition of PT on rubbed polyimide films gave PT chains oriented parallel to the polyimide substrate with a dichroic ratio of 1.5. The PT films are further characterized by IR, X-ray, and conductivity measurements. Powder conductivity measurements on iodine doped samples gave a maximum conductivity of 50 S cm-' . Although the above methods have been generally used to prepare high quality PT (and PATS), other methods have been reported. An early report by Sugimoto reported the synthesis of PT by treating thiophene (5) with FeC13 (Scheme 3). The treatment of thiophene with butyl lithium provides 2,s-dilithiothiophene that can be polymerized with CuClz [16]. Thiophene can also be polymerized by trifluoroacetic acid in the presence of thallium(II1) trifluoroacetate [ 171. The acid-induced polymerization of thiophene was reported as early as 1883, yet produced tetrahydrothiophene units [18]. A novel polymerization of thiophene vapor can produce encapsulated PT in transition metal-containing zeolites [19]. Despite the lack of processability, the expected high temperature stability [14] and potential for very high electrical conductivity of PTJiZms (if made) still make it a highly desirable material. Perhaps precursor routes to PT will eventually lead to processable PT films.

1.3 Chemical synthesis of polyalkylthiophenes (PATs)

5

1.3 Chemical synthesis of polyalkylthiophenes (PATs) 1.3.1 Straight alkyl side chains In the quest for a soluble and processable conducting polythiophene, alkylthiophenes were polymerized. Poly(3-methylthiophene) (PMT) was chemically synthesized and was found to be insoluble [9, 20-221. The first chemical synthesis of environmentally stable and soluble poly(3-alkylthiophenes) (PATs) [23-251 was reported by Elsenbaumer in 1985 (Scheme 4). Very shortly after this report, other groups [26, 27, 281 also reported both the chemical and electrochemical preparation of PATs. Poly(3-alkylthiophene), with alkyl groups longer than butyl, can readily be melt- or solution-processed into films which after oxidation can exhibit reasonably high electrical conductivities of 1-5 S cmpl [23-281.

1.3.1.1 Chemical synthesis of PATs 1.3.1.1.1 Metal catalyzed cross-coupling polymerizations The first poly(3-alkylthiophenes) were prepared via Kumada cross-coupling [23-251, using a method similar to that used for the chemical preparation of polythiophene [5, 141. In this synthesis a 2,5-diiodo-3-alkylthiophene(6) (Scheme 4) is treated with one equivalent of Mg in THF, generating a mixture of Grignard species. A catalytic amount of Ni(dppp)C12 is then added and the polymer (7) is generated by a halo-Grignard coupling reaction. Large quantities of PATs can be prepared by this method and, though initially reported to have low molecular weights (e.g. M , = 3-8K, PDI = 2), later reports show that high molecular weights are possible [29]. 'H NMR of poly(3-butylthiophene) (PBT) showed that the polymer contains only 2,5-linkages, with random regiospecificity.No 2,4- (or ,B) couplings are observed presumbly due to steric blocking provided by the 3-alkyl group. Homopolymers of 3-alkylthiophenes with alkyl groups equal to or greater than butyl are soluble in common organic solvents such as chloroform, THF, xylene, toulene, methylene chloride, anisole, nitrobenzene, benzonitrile, nitropropane, etc. Casting from any of the aforementioned solvents affords thin films of PATs.

8

R

R

R

R

7

R

Scheme 4. The synthesis of poly(3-alkylthiophene).

6

I The Chemistry of Conducting R I

Ni(cod)gPPh3 X

DMF, A

X = Br or I in general 3

7

DenmmurrtlonPolymerlratin

R

dgCII CUI PdCI,

ClHg

S pyridine 9 A R = alkyl or esters

*

V

s

V

n

7

Scheme 5. Other chemical methods for the synthesis of PATs.

Poly(3-alkythiophene) can also be synthesized from 2,5-diiodo-3-alkylthiophene and zerovalent nickel catalysts (Scheme 5 ) [8]. Essentially the same conditions (monomer, Ni(cod)*, and PPh3 in DMF at 60°C for 16-48h) that are used to prepare PT give yields of 60-95% (Scheme 3). The reactions times are longer for PATs and diiodothiophenes are found to be more active monomers than dibromothiophenes. It is reported [8] that this type organometallic coupling polymerization proceeds with predominantly 5-to 5’- (head-to-head) type of couplings. This would give a PAT with mainly head-to-head and tail-to-tail (2- to 2’-) couplings. This is interpreted as selective oxidative addition of Ni to the less sterically hindered 5-position on the alkylthiophene farthest from the bulky alkyl group. Molecular weights by light scattering were reported to be: number average molecular weight (M,) equal equal to 30K (polydispersity to 7.4K and weight average molecular weight (Mw) index (PDI) = 4). The M , measured by GPC (CHCI3) was reported to be 52K.

1.3.1.1.2 FeCI, method for the polymerization of PATs Yoshino and Sugimoto [26] reported in 1986 a very simple method to prepare PATs (Scheme 4). The monomer, 3-alkylthiophene (8), is dissolved in chloroform and oxidatively polymerized with FeC13 [26], MoClS, or RuC13 [30]. Generally the ‘FeC13 method’ has been used to prepare PATs 131-361. Materials prepared by the FeC13 method produce a PATs with molecular weight ranging from Mn = 30-300K with polydispersities ranging from 1.3-5 [35, 361. The FeC13 method does not appear to generate 2,4-couplings in PATs. One very good paper on the synthesis of PATs via the FeC13 method has been reported by Leclerc and Wegner [35]. This paper provides a detailed synthesis, molecular weight data, ‘H and 13C NMR spectra, X-ray, electrochemistry, UVVis, and electrical conductivity data on PATs synthesized chemically with FeC13 and electrochemically. In this paper, alkylthiophenes in dry CHC13 (0.15 M) were treated dropwise with FeC13 in chloroform (0.4M). The mixture was stirred for 24h under a gentle argon stream (to help to remove generated HCl(g)). The

1.3 Chemical synthesis of polyalkylthiophenes ( P A T s )

7

polymer was then precipitated into methanol, filtered, redissolved in CHCl,, and the CHC1, slowly evaporated to give a free-standing film. The film was washed by Soxhlet extraction using methanol and acetone which yields a dedoped polymer containing < 0.10% Fe. The yields ranged from 75-80% with molecular weights of M , = 30-50K with PDIs around 5. The regioregularity, measured by the headto-tail content (discussed in detail later) ranges from 70-80% by this method. This paper reports that the PATs synthesized with FeC1, are more crystalline and regular than electrochemically prepared polymers. Very high molecular weights have been reported in the synthesis of PATs using the FeC13 method by bubbling dry air through the reaction mixture during the polymerization [36]. After isolation and dedoping of the polymer with concentrated ammonia solutions, and washing of the PATs, the molecular weights were determined by GPC using UV-Vis, refractive index (RI), and light scattering detectors. All three values were compared. As an example, a single sample of poly(3-hexylthiophene) had M , of 175K (RI), 124K (UV-Vis), and 398K (light scattering). The PATs had molecular weights ranging from 68K-175K (RI), 77K-146K (UVVis), and 204K-398K (light scattering). One of the major problems with the FeC1, method is that the method gives variable results. The reproducibility of the reaction has been examined, for example by Pomerantz and Reynolds [36]. The polymerization of 3-octylthiophene with FeCl, was repeated under identical reaction conditions five times. Investigations of the molecular weights of the five samples of poly(3-octylthiophene) revealed molecular weights that ranged from 54K to 122K (UV-Vis) with PDIs ranging from 1.6-2.7. Holdcroft [37] has reported that three identical preparations yielded three polymer samples containing three different levels of Fe impurities. The %Fe impurities found in the three samples were 9.6mol%, 4.15mol%, and 0.15molYo. The Fe impurity affects device performance of PT in field effect transistors [37] and in LEDs [38]. The Finnish company, Neste Oy has been working on cost effective methods to synthesize PATs, has reported on the mechanism of the FeC13 synthesis of PATs [39]. The FeC1, initiates an oxidation of the alkylthiophene to produces radical centers predominantly at the 2- and 5- position of thiophene that proprogate to form polymer. Systematic studies on the optimization of the reaction conditions leading to PATs [40] and improvements in the method [41] have been reported. A new synthesis of octylthiophene, followed by FeC13 polymerization led to a poly(3-octylthiophene) (POT) containing 84% head-to-tail couplings in a 70% yield. The molecular weight was reported as M , = 70K (PDI = 2.6). The iron content was only 0.008% and the chlorine content was 0.5%, compared with 1% observed in conventional POT [42]. The electrical conductivity of FeC1, doped POT was 47 S cm-'. The FeC13method is a well-established method to polymerize thiophenes [30-461 and even polydeuterated PATs [47] and continues to be the most widely used and straightforward method to prepare PT and its derivatives, despite the limitations and drawbacks to the method. A related synthetic method developed by Hotta appears to provide PATs that are of high molecular weight and very little metal impurity. The method is the dehydrohalogenation polymerization of 2-halothiophenes with anhydrous metal halogenides (AlCl,, FeC13, etc.) [48]. Molecular weights (M,) of 250K were reported,

8

I The Chemistry of Conducting

along with electrical conductivities of 200 S cm-’ in stretch oriented (5X) after doping with iodine. The elemental analysis of these films shows no detectable metal or chlorine. The NMR spectra, however still indicates an irregular structure. A new PAT synthesis has been recently reported by Curtis and co-workers (49) (Scheme 5). The method is reminiscent of the Yamamoto dehalogenation polycondensation polymerization of PT and PATs that was discussed above. The Curtis method polymerizes 2,5-bis(chloromercurio)-3-alkylthiophenes(9) using Cu powder and a catalytic amount of PdCl;?in pyridine. This method generates homopolymers as well as random copolymers with 3-alkyl and 3-estericsubstituents. Molecular weights are reasonably high for a coupling method (for PBT M, = 26K, PDI = 2.5). In copolymers the proportion of the alkyl groups in the copolymers matched the ratio of their respective monomers in the reaction mixture. Furthermore, this method is tolerant to the presence of the carbonyl functionality. Most of the methods to prepare PT derivatives tolerate very few functional groups. 1.3.1.2 Comparison of the above methods It is important to point out that the methods discussed above produce irregular PATs. That is to say that the self-coupling of 3-alkylthiophene occurs with no regiochemical control which produces a defective PAT. This point will be discussed in detail later. However, comparison of the above synthetic methods (FeC13 method, the Grignard coupling method using Ni(dppp)C12, and the electrochemical methods) has been done by several groups. LeClerc and Wegner [35] have compared the electrochemical method to the FeC13 method and report that the FeC13 method produces better samples of PAT than the electrochemical method. Roncali has indirectly compared chemical methods to electrochemical methods in a review on PTs [50] and says that the electrochemical synthesis of PT has led to the most conjugative and most high conductive materials as of 1992. Stein, Botta, Bolognesi, and Catellani [511 have recently compared the electrochemical method, the Grignard or ‘Ni’ method, and the FeC13 method. Poly(3-dodecylthiophene) (PDDT) is prepared by all three methods and the molecular weights, NMR, UV-Vis, and IR are all compared. The regioregularity and molecular weights for the methods were as follows (HT means head-to-tail): ‘Ni’ method gave 70% HT couplings with M , = 14.5K (PDI = 4), ‘Fe’ method gave 65% HT couplings with an M , = 191K (PDI = 9), and the electrochemical method gave 62% HT couplings with M , = 300K (PDI=6). It was found that the ‘Ni’ method gives the most conjugated PAT as judged by the sharpest UV-Vis peak at 510nm (film) and by the C=C stretch in the IR. The ‘Fe’ method gave a film, , ,A of 494nm. The electrochemically prepared sample gave a film, , ,A of 486nm. However this sample had the broadest UV-Vis peak, with red shifted tail absorbing at lower energies relative the other samples. This implies that the electrochemical method gives a broad range of conjugation lengths, some of which are quite conjugated. It is important to note that care must be taken in determining the molecular weights of PATs using GPC. It has been reported by Berry, Yue, and McCullough that extensive aggregation occurs in regioregular PATs and this can lead to errors in the GPC [52]. We have seen samples of PATS that have M , of 130K by GPC (THF).

1.3 Chemical synthesis of polyalkylthiophenes (PATs)

9

This experiment was reproduced at least five times. However, eventually we observed (in the same sample) a M , of 25K. Therefore, some of the reported molecular weights could be the molecular weights of aggregates.

1.3.1.3 Regioregular PATs While all of these methods reduce or eliminate 2,4-linkages, they do not solve the lack of regiochemical control over head-to-tail couplings between adjacent thiophene rings. Since 3-alkylthiophene is not a symmetrical molecule, there are three relative orientations available when two thiophene rings are coupled between the 2- and 5-positions. The first of these is 2,5’ or head-to-tail coupling (referred to herein as HT); the second is 2,2’ or head-to-head coupling (HH); the third is 5,5’ or tail-to-tail coupling (TT). All of the above methods afford products with three possible regiochemical couplings: HH, TT, and HT (Scheme 6). This leads to a mixture of four chemically distinct triad regioisomers when 3-substituted (asymmetric) thiophene monomers are employed [53, 541. These structurally irregular polymers will be denoted as irregular or non-HT. Irregular, substituted polythiophenes have structures where unfavorable HH couplings cause a sterically driven twist of thiophene rings resulting in a loss of conjugation. On the other hand, regioregular, head-to-tail (HT) poly(3-substituted)thiophene can easily access a low energy planar conformation leading to highly conjugated polymers. Increasing torsion angles between thiophene rings leads to greater bandgaps, with consequent destruction of high conductivity and other desirable properties.

A. FeC13 In-SnO:!

HH and other couplings polymerization

1. Mg polymerization

OR Ni’, PPh3

A Head-to-Tail-Head-to-Tail Coupling (HT-HT)

Head-to-TaiCHead-t+Head Coupling (HT-HH)

A Tail-teTail-Head-to-Tail Coupling (IT-HT)

A Tail-to-Tail-Head-toHead Coupling (lT-HH)

Scheme 6. Possible regiochemical couplings in PATs.

10

I The Chemistry of Conducting

The deleterious effects of non-HT coupling was first examined by Elsenbaumer et al. [55]. It was shown that the polymerization of a 3-butylthiophene-3’-methylthiophene dimer which contains a 63 : 37 mixture of a HT : HH couplings leads to a three-fold increase in electrical conductivity over the random copolymerization of butylthiophene and methylthiophene (50 : 50) reaction. Therefore increase in HT coupling leads to a more highly conductive PAT. 1.3.1.3.1 Regioregular HH-TT and TT-HH PATs Another approach to the preparation of a regiochemically defined PAT was to polymerize either the HH-dimer of alkylthiophene (3,3’-dialkyl-2,2‘-bithiophene) (10) [56] or the TT-dimer of alkylthiophene (4,4’-dialkyL2,2’-bithiophene) (11) [57, 58, 591 to yield essentially the same PAT, namely a HH-TT coupled PAT. SoutoMaior and Wudl compared the physical properties of e.g. poly(3-hexylthiophene) (PHT) and poly(3,3’-dihexyl-2,2‘-bithiophene) (PDHBT) (11) [56]that are synthesized by treating 3-hexylthiophene and 3,3’-dialkyl-2,2’-bithiophenewith FeC13, respectively (Scheme 7). The same polymers are also made electrochemically for comparison. The PHT prepared chemically had M , of 140K (PDI = 4.4) and PDHBT had an M , of 120K (PDI = 5.3). The PHT was determined to have 80% HT couplings and the PDHBT was a regochemically defined polymer containing alternating , , ,A of 398 nm for HH-TT couplings. Films cast from each polymer gave UV-Vis PDHBT and 508nm for PHT. This very large difference in conjugation length is attributed to the intrachain sulfur-alkyl interactions. However, it was also pointed out that it is not obvious why the 3,3’-substituted polymer should be much more sterically hindered than the polymer substituted mostly in a 3,4’-fashion, as is found in PHT. Evidently the steric effects do not seem to affect the electrical conductivity of the HH-TT polymer, PDHBT. The conductivity of the PDHBT oxidized with NOPF6 is 4 S cm-’, whereas PHT’s conductivity is 15 S cm-l. Poly(3-methylthiophene) and poly(3,3’-dimethyl-2,2’-bithiophene)were also compared. Krische and Hellberg were the first to prepare regiochemically defined TT-HH PATs. The TT-dimer (12) of alkylthiophene (4,4’-dialkyl-2,2’-bithiophene) was

10 HH dirner

hn 11

electrochem.or FeC13

R

0-Q 12

R

l7 dimer Scheme 7. Regioregular HH-TT and TT-HH PATs.

13

R

1.3 Chemical synthesis of polyalkylthiophenes (PATs)

11

first prepared electrochemically [59] and then prepared chemically by both Krische [58]and Pron and co-workers [57]. The polymers poly(4,4/-dibutyl-2,2/-bithiophene) (PDBBT) (13) and poly(4,4/-dioctyl-2,2/-bithiophene) (PDOBT) were prepared by polymerizing the appropriate dimer with FeC13 (Scheme 7). Again the TT-HH polymers PDBBT and PDOBT gave a solid state UV-Vis with,a, A of 392 nm and Amax of 388 nm, respectively, with M , of 15K for PDBBT (PDI = 1.5). The, , ,A of e.g. poly(3-butylthiophene) was 494 nm ( M , of 14K, PDI = 1.4). The cyclic voltammetry shows a single oxidation for both PDBBT and PBT at 0.96 V and 0.78 V vs. Ag/AgCl (CH3CN) - nother clear indication that the PBT is more conjugated in the solid with alkyl groups of octyl state. Pron [57] made poly(4,4/-dialkyl-2,2/-bithiophene)s and decyl. The results were virtually the same as Krische. Interestingly enough the FeC13 polymerization of 3-alkylthiophenes gave M , of 144K (PDI = 5.5) for PHT, M , of 142K (PDI = 3.1) for poly(3-octylthiophene) and M , of 250K (PDI = 4.2), which is in stark contrast to Krische [58].This again points to the variability of the FeC13 method unless specific conditions are implemented [35, 36,411. A partial list of other recent methods that are focused on regioregular non-HT PTs include the coupling of 5,5/-dilithiobithiophenes with CuC12[60], a similar coupling of a 5,5'-dilithiobithiophene with F e ( a ~ a c )in~ refluxing THF [61] and Stille coupling of 2,5/-dibromobithiophenes with 2, 5'-bis(trimethylstanny1)bithiophenes using a catalytic amount of PdC12(AsPh3)2[61]. The conformational energetic consequences of each of the possible regioisomers that can occur in PATs (the four oligomeric triads) have been modelled in the gas phase by molecular mechanics and Qb initio methods [62, 631. For the HT-HT example, both methods indicate that the thiophene rings prefer a trans co-planar orientation. Structures with the rings twisted up to 20" (molecular mechanics) or up to 50" (ab initio-STO-3G) from coplanarity all lie within less than 1kcal of each other on a very flat potential energy surface and accordingly are easily accessible [57]. The advantage of HT coupling is supported by crystallographic evidence from HT-HT oligomers of 3methylthiophene [64]. HT trimers of 3-methylthiophene are calculated to have a torsional angle of 7"-8" between conjoined rings [62]. This compares favorably with the 6"-9" observed in X-ray structure of unsubstituted a-terthienyl [64]. Introduction of a head-to-head coupling, as in the HT-HH example, dramatically alters the calculated conformation at the defective HH junction. The thiophene rings maintain a trans conformation, but they are now severely twisted approximately 40" from coplanarity [62] and even a 20" twist is not favored by over 5 kcal. Planarity is impossible as indicated by gas phase calculations. The calculations indicate that head-to-head couplings destroy conjugation inhibiting intrachain charge mobility [65] and can result in poor electrical conductors in polythiophenes that contain non-HT connectivity. It is important to point out that Bredas has reported that the 7r orbitals must be within 30" of coplanarity in order to achieve enough overlap to generate conducting polymer band structure [66]. Structurally homogeneous PTs, denoted as regioregular PTs, can be obtained by one of two general strategies. The obvious, currently most common approach, is to polymerize symmetric thiophene monomers or oligomers. The number of available publications is too large to be fairly considered here and will not be addressed

12

1 The Chemistry of Conducting

[62-651. This review is directed at the alternative strategy: the utilization of asymmetric coupling of asymmetric monomers in order to achieve regioregular HT coupled structures of polythiophene derivatives.

1.3.1.4 Regioregular, head-to-tail coupled PATs 1.3.1.4.1 The McCullough method The first synthesis of regioregular head-to-tail coupled poly(3-alkylthiophenes) (PATs) was reported by McCullougb [67] early in 1992 (Scheme 8). The PATs synthesized by this method contain =loo% HT-HT couplings. This new synthetic method regiospecifically generates 2-bromo-5-(bromomagnesio)-3-alkylthiophene (17, Scheme 9) (from monomer 15 [68-70]), which is polymerized with catalytic amounts of Ni(dppp)C12 using Kumada [7, 7 1-73] cross-coupling methods to give PATs with 98-1OO0h HT-HT couplings [62, 67, 74-79]. In this approach, HTPATs were prepared in yields of 44-69% in a one-pot, multistep procedure. Molecular weights for HT-poly(3-alkylthiophenes) are typically in the range of M , = 20-40K (PDI M 1.4). Some key features of the synthesis are the selective metallation of 15,with LDA [80-811 to generate 16. The organolithium intermediate 16 is stable at -78°C and does not undergo metal halogen exchange via any process including the halogen dance mechanism [82-841. In addition, thienyl lithiums are relatively poor organolithium reagents and therefore are unlikely to undergo metal-halogen exchange reactions with 2-bromo-3-alkylthiophenes. The intermediate 16 is then reacted with recrystallized (from Et20, in a dry box) MgBr2.Et20which results in the formation of 17, which does not rearrange at higher temperatures. Quenching studies

Br2I AcOH/15"C 75%

RMgBr

/

NBVACOW CHCl3 70-85% or NBWHF 7040%

Of

14

Et20, 35°C 75%

4 Br

15

8

1. LDA I THF I-40°C I 4 0 min 2. MgBr2-OEt21-60" to -4O"CI 40 min

3. -40°C + -5°C I 20 min

*

4. 0.5-1 mole % Ni(dppp)C12 -5"+25"C/18 h 45-70%

100% Head-to-Tall PATs RGroup % HT %yield 44% 16a: n-Dodecyl 99% 1 6 b ll-octyl 99% 65% 16c:fFHexyl 99% 5896 16d: n-Butyl 99% 69%

Scheme 8. The McCullough method for the regioregular synthesis of poly(3-alkylthiophenes) (PATs) with 100% head-to-tail couplings.

1.3 Chemical synthesis of polyalkylthiophenes (PATS)

15

TMSCl

99%

R=Hexyl

1%

99%

R = Dodecyl

1%

TMSCl

99%

13

r‘ R = Dodecyl 1%

Scheme 9. Organometallic intermediates and quenching reactions.

performed upon the intermediates 16 and 17 indicate 98-99% of the desired monomer and less than 1-2% of the 2,5 exchange product are observed [62] (Scheme 9). The subsequent coupling polymerization also occurs without any scrambling. The resulting HT-PAT is precipitated in MeOH, washed (fractionated) with sequential MeOH and hexanes Soxhlet extractions and then recovered by Soxhlet extraction with chloroform. 1.3.1.4.2 The Rieke method The second synthetic approach to HT-PAT was subsequently described by Chen and Rieke [85-891 (Scheme 10). This related coupling approach differs primarily in the synthesis of the asymmetric organometallic intermediate. In the Rieke method, a 2,5-dibromo-3-alkylthiophene (20) is added to a solution of highly reactive ‘Rieke zinc’ (Zn*). This metal reacts quantitatively to form a mixture of the isomers, 2-bromo-3-alkyl-5-(bromozincio)thiophene (21) and 2-(bromozincio)-3alkyl-5-bromothiophene (22). The ratio between these two isomers is dependent upon the reaction temperature, and to a much lesser extent, the steric influence of the alkyl substituent. Although there is no risk of metal-halogen exchange, cryogenic conditions must still be employed because the ratio of isomers 21 and 22 produced is affected by the temperature. The addition of a Ni cross-coupling catalyst, Ni(dppe)C12,leads to the formation of a regioregular HT-PAT, whereas addition of a Pd cross-coupling catalyst, Pd(PPh&, will result in the formation of a completely regiorandom PAT. As an alternative approach, a 2-bromo-3-alkyl-5-iodothiophene (23) will react with Rieke Zn to form only 2-bromo-3-alkyl-5-(iodozincio)thiophene (24). This species will then react in an identical fasluon to form either a regioregular HT-PAT or the regiorandom equivalent, depending upon the catalyst that was

14

I The Chemistry of Conducting

Ni(dppe)CI,

-78°C f i r

Br 20

ZnBr BrZn 21

Br

16

22

Pd(PPh3)4

Reglorandom PAT, 7

-

Ni(dppe)CI,

Zn'KHF

I4

B 23

r

-78°C)

IZn

16

24

Scheme 10. The Rieke method to prepare HT-PATS.

used for the polymerization [88]. After precipitation and Soxhlet extraction, the yield for these reactions are reported to be ~ 7 5 % .Molecular weights for polymers prepared by this method are M , = 24-34K (with a PDI = 1.4). One advantage of the Rieke method is that highly reactive Rieke zinc affords a functional group tolerant synthesis. The McCullough method and Rieke method of synthesis of regioregular HT-coupled polythiophenes produce comparable materials which are not spectroscopically distinct. Both methods appear to be generally applicable to thiophenes that are tolerant to organolithium, Grignard reagents, or zinc reagents. About the same time as regioregular, HT-PATS were prepared [67], Holdcroft investigated the reaction of 2,5-diiodo-3-alkylthiophene with Mg and the subsequent polymerization of this reaction mixture. It was found that the reaction conditions 2-iodo-5and time lead to a variation in the amount of 2,5-diiodo-3-alkylthiophene, magnesioiodo-3-alkylthiophene,and 2,5-dimagnesioiodo-3-alkylthiophene. This, of course greatly effects the regioregularity of the PAT synthesized. Quenching experiments on the reactive intermediates, as well as the endgroups were reported. Holdcroft was then able to use the above information in order to synthesize PATs containing a range of HT-HT couplings [90]. The relative configuration of the PAT can be examined from a triad (trimer repeat of alkylthiophenes) labelled %HT-HT couplings or from a dyad (dimer repeat of alkylthiophenes) labelled as %HT. The two are often confused and are only important in PATs with < 100% HT couplings. Holdcroft prepared PATs containing 35-58% HT-HT couplings (52-63 % HT couplings, respectively). It is shown systematically how the optical and electronic properties vary with the regioregularity of the PAT, i.e. the larger the % of HT-HT couplings the more conjugated and more conductive the PAT will be. These results are mirrored by the physical properties of regioregular HT-PAT that have been studied by McCullough and Rieke.

I .3 Chemical synthesis of polyalkylthiophenes ( P A T S )

15

1.3.1.4.3 The mechanism and catalyst choice The polymerizations described above utilize a metal catalyzed cross-coupling technique which has been extensively investigated [71, 91-95]. The reaction is believed to proceed by (i) oxidative addition of an organic halide with a metal-phosphine catalyst, (ii) transmetallation between the catalyst complex and a reactive organometallic reagent (or disproportionation) to generate a diorganometallic complex, and (iii) reductive elimination of the coupled product with regeneration of the metal-phosphine catalyst. Numerous organometallic species (including organomagnesium (Grignard), organozinc, organoboron, organoaluminum, and organotin) demonstrate sufficient efficiency to be utilized in cross-coupling reactions with organic halides. It should be noted that the choice of catalyst is a critical concern. It has been observed that the proportion of cross-coupling to homocoupling of the substrate, indicated by the degree of regioregularity in the product PT, can be dependent upon both the metal and the ligands used in the catalyst [85,88]. A comparison of Ni and Pd with monodentate (PPh,) or bidentate (Ph2PCH2CH2PPh2; dppe) ligands suggested that cross-coupling selectivity was a function of the steric environment of the catalyst. The catalyst with the greatest steric congestion, Ni(dppe)C12,produced almost exclusively cross-coupled product; the catalyst with the least congestion, Pd(PPh& produced a random mixture of cross- and homo-coupled product. However, in Rieke case the polymerization proceeds from a mixture of regioisomers and the success of obtaining 100% HT couplings depends on catalyst selectivity [85, 881. Where 2,5-dibromo-3-methylthiophene is reacted with Rieke Zn, the proportion of the isomers is 80 : 20. When 2,5-dibromo-3-cyclopentylthiopheneor 2,5-dibromo3-phenylthiophene is reacted with Rieke Zn, the proportions of isomers are 71 : 29 and 66 : 34 respectively, whereas in the McCullough method the polymerization proceeds from one regiospecificallygenerated monomer and therefore only HT couplings are found in all cases. 1.3.1.4.4 NMR characterization of HT-PAT Since poly(3-alkylthiophenes) and poly(3-substitutedthiophenes) are soluble in common organic solvents 'H and I3C NMR can be used to determine their structure and regiochemistry [24,35, 51,53-56,62,63,67,75,85]. In a regioregular, HT-PAT for example, there is only one aromatic proton signal in the ' H NMR due to the 4-proton on the aromatic thophene ring at 6 = 6.98 corresponding to only HT-HT triad sequence. Proton NMR investigations of regioirregulur, electrochemically synthesized PAT reveal that four singlets exist in the aromatic region that can be clearly be attributed to the protons on the 4-position of the central thiophene ring in each configurational triad: HT-HT, TT-HT, HT-HH, TT-HH. Synthesis of the four isomeric trimers by Barbarella and co-workers unambiguously assigned the relative chemical shift of each triad, with each trimer being shielded by about 0.05ppm relative to the polymer [63]. In this analysis the HT-HT (6 = 6.98), TT-HT (6 = 7.00), HT-HH (6 = 7.03), TT-HH (6 = 7.05) couplings are readily distinguished by a 0.02-0.03 ppm shift (Table 1). These assignments are the same as the assignment by Holdcroft [90]and different than those proposed by Sat0 [53-541. The relative ratio of HT-HT couplings to non-HT-HT couplings can also be determined

16

I The Chemistry of Conducting

Table 1. Relative chemical shift of triads. Configuration

Chemical shift 6 (ppm)

HT-HT TT-HT HT-HH TT-HH

6.98 7.00 7.02 7.05

by an analysis of the protons that are on the a-carbon of the 3-substituent on thiophene [24,55]. Relative integration of the HT-HT peak relative to the other non-HT resonances can give the % of HT-HT couplings. From this type of NMR analysis, HT-HT couplings [62, 67, 76, 85, 881. Samregioregular HT-PATS contain ~ 1 0 0 % ples from the FeC13 method contain 50-70% HT-HT couplings, in general [62,67]. End-groups have also been identified in both PATs and HT-PATS [88, 901 by both NMR and MALDI-MS [96]. The degree of structural regularity is likewise apparent in the 13CNMR in that only four resonances are apparent in the aromatic region, and they are attributable to the four carbons of a HT coupled thophene ring (e.g. PHT, S = 128.5, 130.5, 134.0, and 140.0ppm). Examination of the I3C NMR spectrum of regioirreguhr PHT shows an abundance of resonances from 125-144ppm. 1.3.1.4.5 IR and UV-Vis A measure of the conjugation length can be determined by both IR and W-Vis. The intensity ratio of the symmetric IR band at -1460 cm-' to the asymmetric band at M 1510cm-' C=C ring stretches decreases with increasing conjugation length. For HT-PATS this ratio is 6-9, less than half of the 15-20 value measured for regiorandom samples [88]. In the UV-Vis the red shift of the maximum absorption which is the T-T* transition for the conjugated polymer backbone is an indication of an increase in conjugation in the polymer. If a qualitative comparison of the solution UV-Vis of PATS is made, a red shift of the A,, was found in regioregular HT-PATS when compared to regiorandom, regioirregular PATs [62, 67, 75-76, 85, 881 (Table 2). As an example, the solution Table 2. Red shift of ,A,

for HT-PATS.

3-substituent

butyl hexyl octyl dodecyl

nA (m ,)

solution

Randomg4 50% HT

FeC1357 70% HT

Electrochem4' ?%HT*

428 428 428 428

436 436 436 436

434 434 440

440

Riekeg4

McCullough5'

98-99% HT

98-99% HT

449 456 45 1 453

450 442 446 460

* Holdcroft (ref. 90) has shown that 69% HT-HT poly(3-hexylthiophene) has a A,, HT-HT has a ,A, of 440.

of 434 and 80%

1.3 Chemical synthesis of polyalkylthiophenes (PATS)

17

, , ,A for HT-poly(3-dodecylthiophene) (PDDT) is about 450 nm, whereas (CHC13) PDDT from FeC13 has a, , ,A of 436nm and regiorandom PDDT has a, , ,A of 428nm [62, 67, 75-76, 881. Thus the regioregular HT-PATS have a lower energy 7r to 7r* transitions, indicating longer conjugation length. Spectra of regioirregular PAT films contain little structure and consist of a single broad absorption for the 7r to 7r* transition. Drop cast films of HT-PAT, in contrast, are red-shifted and vibronic spectra is apparent in the 7r to 7r* transition [62, 67, 75-76, 881. It is interesting to note that this fine structure may range from shoulders to well-defined peaks with definite absorption frequencies but varying intensities, mainly varying with film thickness or processing conditions. Films of irregular PAT polymers prepared from the FeC13 route have a, ,A = 480nm. Films of the regioregular, HT-PATS have solid state absorptions ranging from , , ,A = 562nm for ‘thin’ films of HT-PDDT to 530nm films for ‘thick’ films of HT-PDDT (Table 3). Analysis of thin films of HT-PATS by UV-Vis spectroscopy reveal the presence at least three distinct peaks. For example, HT-poly(3-dodecylthiophene) has a, , ,A = 562nm in the solid state. The lower-intensity peaks appear at X = 530 and 620 nm [57]. These non-A,,, peaks are quite substantial in , , ,A intensity (depending on film thickness) intensity ranging from 60-100% of the (Table 3). The band edge for regioregular PT range from 1.7-1.8eV [67, 76, 881, a 0.3-0.4eV improvement over the 2.1 eV reported for a regiorandom sample. Bjornholm and co-workers [97] have done NLO studies on HT-PATS and as part of that study they have made an estimate of the conjugation length based on linear optical data. In irregular PDDT the conjugation length was determined to be 7 thiophene rings, while regioregular, HT-PDDT has a conjugation length of 40 thiophene rings (80 7r electrons). The saturation was estimated to be roughly 100 7r electrons. Therefore, simply by eliminating coupling defects, thereby minimizing unfavorable steric interactions of the substituents, the solid state order and conjugation length is markedly improved resulting in a reduction in the band gap. Table 3. Solid state absorption for PAT polymers. A,

(nm) solid state

Polymer

McCullough ‘thick film’ >98% HT74

McCullough ‘thin film’ >98% H T ~ O

Rieke ‘thin film’ >98% HTg4

FeC13 70% HT57

PBT 4d

500* 580 610 504* 550 600 520* 553 603 526* 56 1 609

525 560* 608 525 555* 610 525 559* 610 530 562* 620

522 556* 605 526 556* 608 522

480

PHT 4c POT 4b PDDT 4a

480 480

556*

608 524 560* 610

480

18

I The Chemistry of Conducting

The differences in the degree of conjugation and macroscopic morphological order in HT-PATS is a function of film thickness (Table 3). A similar observation was reported by Roncali and Garnier [98, 991, who found that regioirregular poly(3methylthiophene) (PMT) prepared electrochemically exhibited a thickness dependent solid state UV-Vis spectra, which correlated with the conjugation lengths and electrical conductivity. They found that 0.2 pm thin films of poly(3-methylthiophene) (PMT) had a, , ,A of 510 nm, while the thinner films of 0.006 pm had , , ,A values as high 552 nm. The observation was explained by noting that as the polymer film thickens, morphological disorder can increase leading to a more disordered film relative to ultra-thin films. These thin films of non-HT PMT had a very high degree of structural order and extended 7r conjugation lengths with conductivities of up to 2000 S cm-'. In much thicker films (1-3 pm) of HT-PAT'S, the , , ,A of POT is 559 nm and PDDT is 562 nm, therefore the conjugation lengths of HT-PATS are comparable to the very highly ordered, very thin films (0.006 pm) of non-HT poly(3-methylthiophene). Therefore the morphological order found in the 'thin film regime' has been greatly extended from 0.006 pm to ~ 1 - pm 3 by the regular placement of the sidechains in HT-PATS. 1.3.1.4.6 Self-assembly, X-ray, and electrical conductivity in HT-PATS One of the most fascinating physical property differences between irregular PATs and regioregular HT-PATS is that supramolecular ordering occurs in regioregular HT-PATS and not in irregular PATs. Self-assembly in regioregular, HT-PATS were first discovered by McCullough [76-791. Solution light scattering studies by Berry [52] coupled with X-ray studies by McCullough [76] and Winokur [loo] on thin films shows that macromolecular self-assembly occurs in these conducting polymers [loll (Fig. 1). The self-assembly structure leads to a large increase in electrical condutivity in HT-PATS relative to irregular PATs. While the measured conductivity of HT-PAT

stacked polythiophenes

Figure 1. Self-assembled conducting polymer superstructures form from regioregular polythiophenes as confirmed by X-ray and light scattering studies.

1.3 Chemical synthesis of polyalkylthiophenes (PATs) 1000

-

500

-

200

-

100

-

400

300

200 Slcm regioreg. HT-POT as cast

i20 slcm stretch oriented POT from FeC13

100 Scm regioreg HT-PHT as cast

19

130 S/cm stretch oriented PHT from FeC13

Figure 2. Maximum electrical conductivities found in PATs. A 20 to 50 increase in the electrical e.g. the PAT samples below which were stretch conductivity is found by stretch orienting PATS oriented (Polymer 1992, 33, 2340) had initial conductivities of 10-20 Scm-'. ~

films cast from the same sample can differ markedly as a result ofvarying morphology from film to film, conductivities of HT-PDDT (doped with 12) the maximum values have been reported to be 1000 S cm-' ,[76].Values for other HT-PATSsynthesized by McCullough [62, 671 exhibited maximum electrical conductivities of 200 S cm-' for POT and 150 S cm-' for PHT [76]. McCullough et al. have routinely measured conductivities of 100-200S cm-' in these samples in HT-PDDT [78] (Fig. 2). In contrast, PDDTfrom FeCl, generally gave conductivities of 0.1-1 S cm-' (58-70% HT). Rieke and co-workers have also reported that the electrical conductivities for their HT-PATShave conductivities of 1000 S cm-' [88]. Rieke also reports that the average conductivit for HT-PBT is 1350 Scm-', with a maximum conductivity of 2000 S cm- [86].

Y

1.3.1.4.7 Other methods Following the reports by McCullough and Rieke, other groups have found that specific oxidative conditions with a limited number of thiophene monomers can

20

I The Chemistry of Conducting R

R

Scheme 11. Two monomers that give regioregular polymers by the FeCI, method.

lead to an increase in the number of HT couplings in polythiophene derivatives. Anderson reports [1021 that the combination of a sterically-hindered, activating 3-substituent and the slow addition of FeC13 leads to a regioselective synthesis of phenyl substituted polythiophenes with PDIs of 2.5 (Scheme 11). It has also been shown recently by LeClerc [ 1031that the preparation of poly(3-alkoxy-4-methylthiophenes) by FeC13coupling can lead to highly regioregular materials (Scheme 11). This may be due to an asymmetric reactivity of the oxidized monomers.

1.3.1.4.8 Random copolymers of alkyl thiophenes Random copolymers of 3-octylthiophene and 3-methylthiophene were prepared electrochemically and upon doping these copolymers were more thermally stable than the homopolymers [104]. 1.3.1.4.9 Head-to-tail coupled, random copolymers of alkyl thiophenes Regioregular HT-random copolymers of PAT have been prepared [ 1051. Head-totail coupled PAT random copolymers were synthesized by the route shown in Scheme 12. The Grignard compounds 29 and 30 were generated using the standard procedure and polymerized to give polymers 31-35, by simply mixing aliquots of 29 and 30 in direct proportion to the amount of incorporation desired. These copolymers were very soluble in typical organic solvents and possess excellent film forming abilities. The solution UV-Vis data for the copolymers 31-33 indicated that increasing the amount of dodecyl side chains increases the solution disorder leading to a nonplanar structure [lOS]. However, in the solid state, the polymers with a higher percentage of dodecyl side chains self-assembled in order to form planar structures with

1.3 Chemicul synthesis of polyalkylthiophenes (PATs)

X

dir

1. L D N THF/-GO"C -X 2. MgBr2*OEtz

21

BrMg

29 1. LDN THFI-60°C 2. MgBrz-OEt2

X

Y

Scheme 12. Synthesis of random HT-copolymers of PATs

long conjugation lengths. The conjugation in the solid state was greatest in polymer 31 (2: 1, C12H25: CH,;, , ,A = 565 nm), and decreased in polymer 32 (1 : 1, C12H25 : CH3; , , ,A = 550 nm), and polymer 33 (1 : 2, C12H25 : CH,;, , ,A = 545 nm). In addition, cyclic voltammetry of thin films of 31-33 indicated that there are longer conjugation lengths in the solid state for 31 and that the oxidation potential decreases as the proportion of dodecyl side chains decreases (Fig. 3). Conductivity results indicate that I2 doped thin films (0.5-4pm) of polymers 31 and 33 exhibited electrical conductivities in the range of 50-200 S cm-' [76]. In addition, the physical properties of these polymers were unchanged over molecular weights of number average molecular weights (M,) ranging from 9-28K (PDI = 1.6). Therefore, apparently the physical properties appeared not to be a function of the molecular weight. 1.3.1.5 Branched alkyl PATs Early reports that 3-isopropylthiophene could not be electrochemicallypolymerized and therefore few branched alkyl PATs have been chemically prepared. However in 1992, Wegner and co-workers [ 1061 prepared poly(3-cyclohexylthiophene)(PCHT)

22

I The Chemistry of Conducting

%

4

Conjugation Oxidatlon Conductivity Length Potential

DD

Figure 3. Tuning the properties of electronic and photonic conjugated polymers. As the percentage of dodecyl side-chains in regioregular HT-coupled 3-dodecylthiophene/3-methylthiphenecopolymers increases, the conjugation length increases, the oxidation potential decreases, and the conductivity increases.

(36) (Fig. 4)and compared the band gap, non-linear optical properties, and the electrical conductivity of the material with poly(3-hexylthiophene)prepared with FeC13. The M , (73K) of the PCHT was roughly the same as the PHT. However, the PCHT was much less conjugated, by 80 nm in the solid state relative to PHT. The two polymers were electrochemically doped in the presence of ClO, and the conductivity of PHT was 0.4 S cm-’ and PCHT was 0.0017 S cm-’ . These results indicated that sterically bulky side chains on PATs cause a steric twisting of the backbone and reduce the conjugation and electrical conductivity in PATs. A PT containing an annelated cyclohexyl ring, 37, has also been prepared by Wegner and found to have a lower electroactivity, conductivity, etc. than poly(3,4-dihexylthiophene) (38) and PHT as prepared from FeC13 [lo71 (Fig. 4).

n

36

37

Figure 4. Sterically encumbered PTs.

1.3 Chemical synthesis of polyalkylthiophenes (PATS)

23

Poly(3-cyclohexyl-4-methylthiophene)(PCHMT) (39) (Fig. 4) and poly(3-cyclohexylthiophene) (PCHT) (36) have been made by the FeC13 method in order to test these polymers as polymers LEDs [108]. The PCHMT gave an Mw of 72K (PDI 2.8) and a film, , ,A of 303 and PCHT had an M , of 56K with a very large of 426nm, with 77% H T couplings. It was found that PDI of 9 and a film, , ,A by varying the steric environment of the PT, that LEDs having from blue to near-infrared emission could be made. Recently Guillerez and co-workers [ 1091 have prepared regioregular chiral HT-PATS namely poly(3-(S-3’,7’-dimethyloctyl)thiophene)(40) (Fig. 5). They have shown that if the steric group is far enough removed from the backbone then the conjugation is relatively unaffected. The study of this PT reveals that circular dichroic spectroscopy can be used, in regioregular HT-PTs bearing a chiral group, to probe aggregation behavior as was first pointed out by Meijer [110]. In addition the above polymer exhibited large conformational changes induced by minute solvent variation. Fluorinated PTs have been prepared using the FeC13 method by LeClerc and the ability of these polymers to form monolayers was investigated 11 111 (41, 42) (Fig. 5). In addition, the polymerization of 2-(3-thienyl)ethyl perfluoroalkanotes (43-45) (Fig. 5 ) by the FeC13 method has been accomplished by Collard [112]. These polymers are being investigated for their ability to self-assemble via fluorophobic association.

1

*

Regioregular40

41

42

Figure 5. Chiral alkyl and flouroalkyl substituted PTs.

24

1 The Chemistry of Conducting

1.3.1.6 PTs with phenyl sidechains Phenyl substituted PTs were synthesized by Ueda [ 1 131using basically the polycondensation polymerization of 2,5-dihalothiophenes developed by Yamamoto [S- 131 using NiC12,PPh3, bipyridine, Zn, and N,N-dimethylacetamide (DMAC). The effect of solvent, amount of DMAC, NiC12,PPh3, and bipyridine, and temperature were investigated. The solution , , ,A of the poly(3-phenylthiophene) was 430 nm, which is essentially the same as sexithiophene. The polymer was soluble in typical organic solvents and films were found to be very thermally stable (10% weight loss at 550°C in air). Poly(3-phenylthiophene) was also prepared by the Rieke method from a mixture of regioisomers of bromo-zincbromothiophenes [87]. Nearly regioregular (94% HT reported) poly(3-(4-octylphenyl)thiophene) has been prepared [lo31 using FeC13 to oxidatively polymerize 3-(4-octylphenyl)thiophene. The film, , ,A of the poly(3-phenylthiophene) was 493 nm, which increased to 602 nm upon exposure to CHC13. Later papers reported on copolymers of thiophene and octylphenylthiophene [1071 and the parent poly(3-(4-0ctylphenyl)thiophene).

1.4 Chemical synthesis of heteroatomic functionalized substituents on PTs: recognition sites for self-assembly and chemical sensing The flexibility of PT to change its color and electrical conductivity in response to various analytes, solvents, and environments make the conjugated polymer an ideal candidate as an all polymer sensor. The linear alkyl-substituted polymers have been most widely studied because of their ease of synthesis. However, increasingly heteroatom substituted PTs are being designed, synthesized and explored in order to engineer intelligent properties into the conducting polymer. The possibility of merging host-guest chemistry, biological macromolecular assembly, organic self-assembly, and inorganic structural chemistry to create new conjugated polymer devices and smart materials is a rapidly expanding area. In the 1980s, Wrighton, Murray, Baughman, and Garnier were pioneers who published general discussionsof conducting polymer sensors and early demonstrations of the phenomena [114-1191. A perspective paper by Garnier [114] specifically suggested that, due to the ease of synthesis of a large variety of PTs, functionalized polythiophenes could be used as molecular sensors. Garnier and Roncali initially demonstrated that alkoxy substituted and chiral alkoxy substituted PTs that were synthesized electrochemically acted as sensors for cations. The alkoxy PT (47) (Fig. 6 ) showed a 200mV shift in the oxidation potential of the polymer when Bu4NC104was used an electrolyte instead of LiC104 in the solid state cyclic voltammetry (CV) experiment. T h s indicates that the electrochemical response of the PT could be altered by environmental effects. The chiral PT electrochemically synthesized in the study was found to recognized the complementary enantiomeric anion used as a dopant

Chemical synthesis of heteroatomicfunctionalized substituents on PTs

+ Can Recognize Li' and alkali metals and Bu~N*

P' 1

25

Can Recognize Lit

47

46 Disubstitution with Oxygen Directly Connected to Thiophene Ring Does Not lnhibil Conjugation of the PT

n 48

49

50

Figure 6. Alkoxy PTs.

during redox cycles in the CV experiment. These results demonstrated the use of alkoxy substituted PTs as conducting polymer sensors. A large body of work on the electrochemical synthesis of alkoxy substituted PTs has been reviewed by RonCali [50, 1201 and is beyond the scope of this review.

1.4.1 Chemical synthesis of alkoxy polythiophenes Alkoxy PTs and related derivatives were first reported to be electrochemically synthesized [120-1271. Alkoxy derivatives have several advantages over PATS. The first is that if the oxygen is directly attached to the ring, the band gap in the conducting polymer can be reduced by a substantial amount [55]and the conducting state of the polymer is stabilized [124,128].In addition, the side chains can act a molecular recognition units for chemical sensing or as arms for directed self-assembly of the polymer. Poly(3-methoxy-2,5-thiophenediyl)was chemically synthesized using the FeC13 method [ 1291. The polymer had limited solubility in DMSO, 1-methyl-2-pyrrolidinone (NMP), and propylene carbonate (PC). The polymer was characterized by UV-Vis-NIR, IR, CV, and electrical conductivity measurements. Short alkoxy chains on PTs that have been synthesized chemically or electrochemically generally led to low molecular weight and insoluble materials. Two solutions to the solubility problem were provided by Bryce (46) [122], RonCali and Garnier (47) [127], and Leclerc [128, 1301 (Fig. 6). Long chain alkoxy substituents lead to large increases in solubility [ 122, 1271, while 3,4-disubstituted PTs lead to the same effect [128, 1301. A comparison of poly(3-butoxythiophene) (48),

26

1 The Chemistry of Conducting

poly(3,4-dibutoxythiophene) (49), and poly(3-alkoxy-4-methylthiophene)(50) prepared by the FeC13method [128, 1301shows that poly(3-butoxy-4-methylthiophene) showed the highest electrical conductivity (2 S cm-’ , after doping with FeC13) among this group of polymers. It was found that despite 3,4-substitution on PT, poly(3-butoxy-4-methylthiophene)did not appear to exhibit a large steric twist of PT backbone. In fact, the, , ,A was 420 nm (CHC13) and 545 nm (film) indicating a highly conjugated PT. This is contrary to both poly(3,4-dimethylthiophene) and poly(3,4-dibutoxythiophene)polymers where the 3,4-disubstitution causes a sterically driven twist of the conjugated backbone and leads to very low conductivity in the doped polymer. This study indicated that certain 3,4-disubstituted PTs could be highly conductive polymers. The importance of this finding led to the development of very stable conducting polymers based on PT, namely poly(3,4-ethylenedioxythiophene) (PEDOT) (51) (Fig. 7) that will be discussed later. It is important to note that evidently oxygens directly attached to the ring, not only reduced the band gap of the polymer, but also do not cause a detrimental steric twist of the polymer out of conjugation. Leclerc [131] followed the above study with the chemical synthesis of poly(3,3’dibutoxy-2,2’-bithiophene)(3,3’-PDBBT) (52) (Fig. 7) by polymerizing 3,3’-dibutoxy-2,2’-bithiophene with FeC13 using the Sugimoto/Yoshino method. The PDBBT , , ,A of about 580 nm and an was a very highly conjugated PT which showed a film electrical conductivity of 2 S cm-’ . An interesting finding was that upon electrochemical or chemical oxidation of the polymer, the film became nearly transparent in the visible region. Further development [ 1321 of poly(bithiophene) derivatives led to the FeC13 synthesis of poly(3-butoxy-3’decyl-2,2’-bithiophene)(54), poly(4butoxy-4’-decyl-2,2’-bithiophene)(55), poly(3,3’-dibutoxy-2,2’-bithiophene)(3,3’PDBBT) (52), and poly(4,4’-dibutoxy-2,2’-bithiophene)(4,4’-PDBBT) (53). The 3,3’-PDBBT is made from the HH dimer of dibutoxybithiophene, whereas the 4,4‘PDBBT is made from the TT dimer of dibutoxybithiophene. The 4,4’-PDBBT was

‘ 0

\ /

\ I

n

BuO 52

51 PEDOT

54

Figure 7. More alkoxy PTs including PEDOT.

OBU

53

55

Chemical synthesis of’ heteroatomic functionulized substituents on PTs

27

found to have a molecular weight three times higher than that of the 3,3/-PDBBTand the 4,4’-PDBBT was more conjugated by 20 or so nm. However, there was very little difference in electrical conductivity or the redox behavior of the two polymers. The asymmetrically substituted polybithiophenes showed lower electrical conductivities (0.01-0.25 S cm-’ vs. 2-3 Scm-’) than both 3,3/-PDBBT and the 4,4/-PDBBT due to the presence of non-regioregular couplings. The presence of 25% HH couplings was reported. Other alkoxy PTs have been prepared with FeC13 [133] and solubilities, molecular weights, X-ray, TGA, etc. has been reported. Alkoxy PT have also been prepared by a Cu(C104)*oxidation of bithiophenes [ 1341. Poly(3-(alkoxyphenyl)thiophene) with meta- and para-alkoxy substituents have been prepared using the FeC13 method and the third order non-linear optical properties of these polymers were studied [135]. Langmuir-Blodgett films alkoxy PTs have been also reported to be prepared and studied [136, 1371. Grignard preparation via Kumada coupling has led to alkoxy PTs [138, 1391 that have been studied in devices such as polymer LEDs. In 1991, poly(3,4-ethylenedioxythiophene)(PEDOT) (Fig. 7) was prepared [ 1401. The very stable conducting polymer was initially chemically prepared by BASF as a thin film coating in antistatic plastics. A polycarbonate film was coated with a thin layer of polyvinylacetate containing an iron (111) salt. A second coating of EDOT on top led to a PEDOT in a polyvinylacetate matrix. The material showed excellent stability in the conductive state. The polymer PEDOT can also be prepared in bulk [ 1411 by polymerizing 3,4-ethylenedioxythophene (EDOT) with FeCl,. The PEDOT (doped with FeC13) prepared in refluxing CH3CN gave a conductivity of 15 S cm-’, while the polymer prepare in refluxing benzonitrile had a conductivity of 19-31 Scm-’, depending on reaction time. A later paper [142] reported that chemically prepared PEDOT (using the FeC13 method) led to a polymer was, for the most part, not soluble. On the other hand, when EDOT was electrochemically polymerized, a soluble PT resulted. A recent paper by Reynolds [143] found that in the synthesis of tetradecyl-substituted poly(ethy1enedioxythiophene)s (PEDOTC14H29)that the solubility of the resultant polymer was very sensitive to the FeC13/monomer ratio. When the ratio was 1, 100% solubility was found and the solubility decreased to 0 % solubility when the ratio was 5. The polymer PEDOTC14H29 formed deep purple films and upon electrochemical oxidation gave light green transparent films at a very low oxidation potential of 0.3V vs. Ag wire. Crown ether annelated PTs have also been polymerized using nickel catalysts [ 1441.

1.4.2 Chemical prepared alkoxy PTs as conducting polymer sensors In the preparation of conducting polymer sensors [114-119, 145-1471, it is critical that the control of the structural regularity be maintained. That is to say that a PT with a regular placement of the side chains or molecular recognition units will maximize the planarity of the PT and allow for the largest response window. In addition, the chemical synthesis of PTs allows for large amounts of material to be generate. Two reports of a tunable conducting polymer and sensors that can be

28

I The Chemistry of Conducting

prepared chemically were published in 1993 by McCullough [ 1481 and Swager [149] (Fig. 8). McCullough reported that the conductivity, electrochemical and optical response in regioregular alkoxy substituted PTs (e.g. 56) are highly sensitive to the nature and regiospecificity of the side chains [76, 1481. It was found that small conformational changes due to analyte or ion detection or minute solvent polarity changes produce large effects in regioregular PTs. Swager reported on PTs where adjacent thophene rings are linked by a crown ether-like unit (57). It was found that upon complexation of ions that the PT conjugation is greatly reduced, thus exhibiting a sensory response. Three related polymers 56, 64-66 (Scheme 13) were designed and prepared in order to investigate whether the polyethylene glycol like side-chains could help to increase electrical conductivity by increasing the ionic conductivity of the material. In addition, it was thought that since the regioregular placement of the side chains allows for binding cavities to be created, then these 'sensory arms' could be used to detect various analytes. The substituent on 65 is too short to allow solubility in the growing polymer therefore only low molecular weight materials are produced. Similarly 64 yielded marginally better results ( M , = 6000; PDI = 2). However, polymer 56, determined to be > 99% HT by NMR, was markedly different. The product was regularly of high molecular weight ( M , = 71 000; PDIPDI = 2 M 160 thiophene rings/chain). When doped with I2 this polymer possessed very high electrical conductivity of between 500- 1000S cm-' on average, and a maximum conductivity

57 X = 0 or -CH,Oz=1.2

w

Regioregular

M coupled

R

-@

11

+@

Figure 8. PTs as chemoselective sensors.

Chemical synthesis of heteroatomic functionalized substituents on PTs

2 s

Br21AcOH

R X Na'

~

51%

58

aBr

29

*

59

2. 1. LDA MgBr2.0Et2 J THF I-78% 4h 145 min "+ >

S

60-63

3. Ni(dppp)C121-78" + 25°C I 36 h

RX . .. .

XR Group 60 & 56 -OCH&H20CH&H20CH3 61 & 64 -OCH,CH,OCH? - 628165 -0CHi 6 3 & 6 6 -SCHa

64-66,& 56

Scheme 13. The synthesis of head-to-tail coupled, heteroatorn-substituted polythiophenes.

of 5500 S cm-' for one sample of exceptional film quality [148]. These results indicate that HT-56 exhibits higher electrical conductivities relative to HT-PATS. This is in line with early reports on the high conductivity found in irregular 26 [I221 and contrary to reports on other studies on irregular 56 [120]. It is possible that solid state ion-dipole binding led to a highly ordered structure for irregular 56 polymerized in the presence of Bu4N+ and hence led to a highly conductive sample. Molecular models show that the Bu4N+fits well in a cavity formed by polyether arms. In addition, high conductivity in 56 may also be due to a predicted increase in the ionic conductivity, facilitated by the etheric side chains, thereby increasing the charge carrier mobility. As was the case with the PATS, film morphology appears to be the limiting factor in reproducing high conductivity. Polymer 56 does exhibit ion binding properties and an ionochromic response occurs upon exposure to Li+ in acetonitrile, and leads to a blue shift of up to 11 nm [148]. A dramatic chemoselective sensory response to Pb2+ and Hg2+ in CHC13 was discovered for 56 as indicated by a 200 nm blue shift in dilute solution [77, 101, 1501 upon addition of PbC12 or HgC12. There is no colormetric or optical response to Li', Zn2+,Cd2+,and a host of other ions. In concentrated solutions the effect is different. Concentrated solutions of HT-56 are deep magenta without added Pb-salts,,A,( = 575-600 nm; band edge at 700 nm). Again a striking transformation occurs upon the introduction of Pb(BPh4)2to the solution of HT-56. The conjugation length immediately decreases as indicated by a blue shift of 50-1OOnm upon introduction of Pb2+ ion accompanied by a 50nm blue shift in the band edge ,,A,( = 480-550 nm; band edge at 650 nm). A film cast from HT-56 and Pb(BPh4)2 = 440 nm; band edge at 550 nm). In contrast, films cast from is yellow in color,,A,( = 520nm; band edge at the salt-free solution are deep crimson in color,,A,( 720nm). There is a huge (170nm) difference in the band edge between films cast in the presence and the absence of Pb2+.Since films of pure Pb(BPh4)2are colorless, the Pb2+ ions induced a large disordering of the polymer. Comparison, by X-ray

30

I The Chemistry of Conducting

analysis, of a film of HT-56 to a similar film which had been cast in the presence of Pb(OAc)2indicates that the Pb2+ ions cause a significant amount of disorder in the film. X-ray diffraction shows that the very strong, wide-angle reflection, that represents interchain stacking of thiophene rings, has a half-width of 0.23 A for a film that contained Pb2+. In contrast, the corresponding half-width for the uncontaminated film is 0.11 In addition, iodine-doped films of HT-56 that had been cast from a solution (in CHC13) saturated with P ~ ( O A Cshowed )~ a z 10 000-fold decrease in electrical conductivity when compared to similar samples that contained no Pb salts (azO.001-0.01 vs. a = 100-1000 for samples without Pb) [101, 1501. Regioregular PT 56 solutions of M are able to colorirnetrically detect Pb and Hg at minute concentrations of M even in the presence of multiple ions making these PTs outstanding Pb and Hg detectors. Regular polymers (e.g. 57) (Fig. 8) were prepared and ionochromatic responses were measured in 0.1 M salt solutions in CH3CN. Polymer 57 (X=O, z = 1) had a ,,A of 497nm in solution and exhibited a chemoselective response to Na' (Ax,,, = 91nm) versus Li+ (Ax,,, = 46nm), and K' (Ax,, = 22nm). Biothiophene copolymers containing units of 57 (X = 0, z = 1) show selectivity for K+ (Ax,,, = 45 nm) versus Na' (Ax,,, = 30 nm), and Li+ (Ax,,, = 13 nm). Biothiophene copolymers containing units of 57 (X = CH20, z = 1 and z = 2) exhibited no response, indicating that the conformational restrictions are greater when a methylene bridges the thiophene ring and the polyether chain [149]. The design and development of conducting polymer sensory materials that are chemically prepared has led to conducting pseudopolyrotaxanes based on PT and PTs functionalized with calix[4]arene. These two co-polymers of PT (67, 69) (Fig. 9) are equipped to either recognize molecules or ions and therefore transduce molecular interactions into a measurable response ether ionochromatically, votammetrically, through fluorescent responses, and both iono- and chemoresistive responses [151-1531. The PT 69 exhibits a chemoresistiveresponse to paraquat derivatives and forms a pseudorotaxane upon binding. A flow cell experiment indicated that the PT is 'reset' after 2 min in a unoptimized demonstration of a real-time sensory device [152]. While the calix[4]arene PT had very large binding constants for alkali metals, they exhibited no changes in the UV-Vis spectrum and only minimal changes in their voltammetric responses, however large decreases in the polymer's conductivities were found upon exposure to Lif and Kf [ 1531. These PTs are exciting as potential new sensory materials and clearly demonstrate new creative approaches to all plastic sensors. Although the Swager and McCullough polymers are expensive to make (from an industrially point of view), minute amounts of each polymer is needed to prepare sensory solutions or films, essentially making them cost-effective. Regioregular, head-to-tail poly(3[w-( p-methoxyphenyoxy)-hexyllthiophenes[ 1541 has been recently prepared using the McCullough method. NMR analysis show that the polymer is 99% head-to-tail coupled. The regioregular polymer is much more soluble than the cross-linked irregular polymer, also prepared in this report. The polymer is similarly thermochromic to HT-56. Four probe conductivity measurements on pressed pellet samples show that the regioregular polymer shows a 1000-fold increase in the electrical conductivity versus the irregular sample. However, no report was made on the recognition features of the polymer.

A.

Chemical synthesis of heteroiitomic junctionalized substituents on PTs

I

1

31

} Calixarene As a Recognition Unit

e.g, 0.5 mM Na' response +100mVchange in Oxidation Potential and 99% decrease in drain current in Id-V, measurements

R = O(CH,CH,O),CH, e.g. 10 mM Li' response AX,,,

1

i-

Charge Transfer Binding As a Recognition Force to Form Pseudorotaxanes

e.g. 50 mM Paraquat response 4 0 m V change in Oxidation Potential and 26% decrease in drain current in Id-V, measurements

= 150 nrn

d 70

exposure to organic solvents and thermal response completely stereomutates polymer to give mirror image circular dichroism response

Figure 9. Other PTs as chemoselective sensors

The thermochromic, solvatochromic, and piezochromic responses of polythiophenes is well-known [16, 52, 155-1581 and is thought to be related to the reversible phase transition between a highly conjugated coplanar PT and a less conjugated, less planar conformation along the PT backbone. The chromic phenomena in both solid state and solutions of semicrystalline regioregular and amorphous nonregioregular polythiophene derivatives has been investigated by Leclerc [159]. A number of alkoxy PTs were tested and polymer 50, regioregular PMEEMT was found to sense alkali-metal ions. It turns out that K+ causes a solution self-assembly to occur and planarizes the backbone of the polymer. The response is sensitive to K+ concentrations ranging from 1 x 1OW6 M to 1 x l OP4 M. Above 1OP4 M then response is saturated. Changing the recognition unit from (2-methoxyethoxy)ethoxy group on PT to a poly(ethyleneglyco1) methyl ether (68) unit placed in a 95% regioregular M to [62,67,76] fashion changes the response concentration window from 1 x 1 x loP2M. The polymer exhibits an isobestic point indicative of biphasic conformation order/disorder and ion induced self-assembly [1601. The regioregular PT sensor was found to chemoselective, having a larger sensitivity to K+ ions over Na' over NH4+ over Li+. It was found the sensing response to the ions by the PT is 'molecularly amplified' by causing a 'domino effect' of PT conformational twisting to occur [161]. This type of discovery is yet another indication that regioregularity is an important factor in the development of ionoselective sensors. The transfer of conducting polymer sensors to surfaces is a next step. Attempts at the preparation of Langmuir-Blodgett films of regioregular alkoxy PTs [ 148, 1501

32

1 The Chemistry of Conducting

have met limited success [162] and variation in stability and formation of the LB films has been found. Surface bound sensory devices for biomedical applications have recently been prepared with polydiacetylene [1631.

1.4.3 Chiral Substituents on PT Most polymers are known to adopt helical conformations in solution, in the solid state, or both. It is possible to induce optical activity in the main chain of such a polymer by substituting with enantiomerically pure side chains. In such a system significant chirality is induced in the backbone only when the polymer forms a well-ordered aggregate, a state that is common in self-assembled, regioregular poly(3-alkylthiophene)~[52, 761. The role of chirality and optical activity as a function of regioregularity in polythiophenes can be examined by comparing the work of Roncali and Meijer and Boumann. Roncali [ 1641 has electrochemically prepared chiral polythiophenes by polymerizing (S)(+)- and (R)(-)-2-phenylbutyl ether of 3-pro lthiophene to yield chiral polythiophenes with reported specific rotations a ! ' = 3000 [110, 1651. Using the McCullough method, Meijer and Boumann of [ [ 1 10, 1651 have recently synthesized an optically active regioregular polythiophene 70 (Scheme 14) that exhibits a s ecific rotation of [a]22 = 140000 for A = 513nm and at the sodium D-line of [a] = -9000 for X = 589nm. This of course points to the variation in the optical rotation as a function of wavelength and regioregularity and assembly. Polymer 70 also undergoes stereomutation in the solid state. Solvatochromic studies of polymer 70 ( M , = 16 900; PDI = 1.4) show that varying the solvent composition dramatically affects the shape and A,, of the 7r to 7r* transition by altering the distribution of disordered and ordered, aggregated structures. The solid state thermochromism in 70 is typical for a polythiophene, except that at the

4

&la&

/ S\

t

KOHlTHF

71

4

NSSIDMF

~

72

dsr

1. LDA/THF 2. MgBrZ-OEtp

S

3. Ni(dppp)C12

73

R

R Group

72&73 =

70

Scheme 14. The synthesis of regioregular polythiophenes with chiral side chains.

Chemical synthesis of heteroatomic functionalized substituents on PTs

33

melting point of the polymer a complete loss of optical activity is observed. Even more interesting is the observation that when the polymer is cooled very fast from the disordered melt (by pouring the sample into a water bath at OOC). The absorption spectra is unchanged, but a mirror image CD spectrum (relative to the original sample before melting) is found. Therefore the regioregular, chiral polythiophene 70 undergoes stereomutation. This process is thought to be driven by an aggregation effect. The effect is reversible, affording the opportunity to tune the chirality of the spectrum simply by controlling the cooling rate. Irregular chiral polythiophenes do not show this effect. The same effects are also found in poly(3,4-di[(S)-2-methylbutoxylthiophene), which has been studied in some detail by both CD and circular polarized luminescence [ 166, 1671. Other chiral HT-PATS namely poly(3-(S-3’,7’-dimethyloctyl)thiophene) [lo91 confirms that CD spectra can be used, in chiral regioregular HT-PTs to probe aggregation states [110]. This particular polymer can also act as a sensor, by exhibiting large conformational changes induced by minute solvent variation.

1.4.4 Carboxylic acid derivatives: self-assembly and sensors It has already been demonstrated that the torsion between thiophene rings is extraordinarily sensitive to the steric interactions of the side chains. McCullough et al. have designed regioregular carboxylic acid derivatives in order to promote self-assembly through self-molecular recognition forming carboxylic acid dimer pairs between PT chains, ‘zipping up’ the ordered conducting polymer structure (Fig. 10). In addition, receptor sensing could lead to a huge signal amplification due to cooperative Analyte Driven Disassembly

PT disassembled by large metal detection, disordemd and not conjugated COLOR-YELLOW color tunableyellow, orange, red

Self-Assembled PT disassembling by large metal detection -very rapidly losing conjugation

Figure 10. PT zipper sensors.

Analyte Driven Self-Assembly

PT in metal-driven self-assembled state, after detection of small metal, very highly conjugated COLOR-PURPLE

PT in disordered but conjugated State COLOR-RED

34

I The Chemistry of Conducting

‘unzipping’ and twisting of the polymer causing both a colorimetric response and change in the electrical conduction in the molecular wire. Carboxylic acid salts were chosen as the substituent because it is trivial to dramatically change the steric demands of the function simply by changing the size of the counter cation. In addition, carboxylate substituted polythiophene should be water soluble. Regioregular carboxylic acid derivatives of PT have been prepared by employing the sturdy oxazoline protecting group. Regioregular, HT polymers 75, 76, 77, and 78 were synthesized as shown in Scheme 15 [168,169]. Deprotection of 75 in aqueous HCl yields the desired product 76 as a dark purple precipitate that is completely ‘zipped up’ into an ordered self-assembled conducting polymer. Upon deprotonation polymer 76 is converted into polymer 78 which is completely water soluble. Most interestingly, the predicted self-assembly and ionochromatic response is dramatically evident. Polymer 78 is a chemoselective ionochromatic sensor in water. The colorimetric response signal ranges from the self-assembled purple state to the disassembled, twisted yellow state and the A,, of the polymer changes over a 130 nm range simply by varying the counterion from NH; (purple), to Me4N+ (magenta), to Et4Nt (red), to Pr4N+ (orange), Bu4N+ (yellow) [168, 1691. The observed chemoselective chromism is not merely counterion size dependent, but is also related to the hydrophobicity of the counterion. What happens is that a protein-like hydrophobic assembly occurs in the PT with small counter-ions, whereas larger counter-ions break up the self-assembled state, effectively disassembling the PT and amplifying the signal response [168]. Irregular polythiophenes carrying carboxylic acids [ 1701have been prepared electrochemically and dramatic ionochromism or ionic self-assembly was not reported. However, other reports have looked at irregular PT carboxylic acid derivatives in competitive immunoassays for antigens and haptens [ 1711. The carboxylate function attached directly to the thiophene ring has been reported to be prepared [172] from the 2,5-dichloro-3-methylthiophenecarboxylate by the Yamamoto route [8].

1.4.5 Other derivatives of PT Regioregular esters of PT of any type can be prepared from the McCullough method [168, 1691 as shown in Scheme 15. The polymer molecular weights range from M,, = 12K to M , = 5K with PDIs = 2 with 100% HT couplings. Irregular polythiophenes [173] containing ester side chains have been prepared with limited success. The FeC13 method leads to partial deesterification of the isolated polymer and some ester functionalized thiophenes does not lead to polymer. Modifications of the reaction media lead to ester functionalized PTs that have been well-characterized [1741, having regioregularities of around 65%. The reported molecular weights are very large with DPs of 180-1250. Aggregation was not discussed. The conductivities of FeC13 doped materials were in the 0.1-0.0001 Scm-’ range. Ullmann coupling polymerization of 2,5-dibromo-3-alkylthiophenecarboxylates by Pomerantz is an excellent way to prepare ester derivatives of PTs [175] (Scheme 16). The M,’s of these materials were in the 4K region with PDIs around 2. Subsequent electroluminescent studies and bilayer devices were prepared and

Chemical synthesis of heteroatomic functionalized substituents on PTs

2

-

35

1. LDAITHFI-78°C 2. MgBr2.0Et2

s

3. 3. Ni(dppp)CI2 Ni(dppp)CI2

n

74

0

* 77

base I

-yo H,O soluble 78

Scheme 15. Synthesis of PT zipper sensor polymers water soluble, highly ionochromic, regioregular HT-polythiophenes.

studied by Pomerantz and Elsenbaumer [176]. Efficiencies of 0.018% were found in the ester PT polymers which was much better than irregular poly(3-octylthiophene) (5 x lop5%)and regioregular poly(3-hexylthiophene) ( 5 x lop4%)[38]. Poly(3-(2-(methacryloyloxy)ethyl)thiophene) (82) has been prepared by Holdcroft by the FeC13 method [26] (Fig. 11) and used to prepare an electronically conducting pattern by photolithography [ 1771. Related urethane-substituted PT has been prepared by Gregory [178] (Fig. 11) (83). Such a PT is thought to have very good solubility and show improved processability and may be able to be used in conducting elastomers. One example would be a blend of urethane-substituted PT and a polyurethane elastomer that could be used for numerous applications including, electromagnetic shielding and antistatic coatings. Another application 1. SOClp

79

2. ROH pyridine

80

Cu, DMF

81

150°C

7 days

82

Scheme 16. Synthesis of PT esters by Ullman coupling.

36

I The Chemistry of Conducting

of urethane substituted PT would be reversible thermal recording. Poly(3-(2-(Ndodecyl-carbamoy1oxy)ethyl)thiophene) (PDDUT) (Fig. 11) (84) has been prepared [179] by the FeC13 route [26]. The polymer has been characterized by NMR, IR, UVVis, and X-ray. All of the data is shown and reveals 80% HT couplings in the PDDUT. The thermochromic response of PDDUT was cast onto poly(ethy1ene terephthalate) and recorded letters were imprinted using a thermal recording head. Water soluble, self-doped [1801PTs were first prepared electrochemically by Wudl [181, 1821 in 1987. Sodium poly(3-thiophene-P-ethanesulfonate)(P3-ETSNa) and sodium poly(3-thiophene-P -butanesulfonate) (P3-BTSNa) are soluble in water in the doped and undoped states. The methyl esters were first polymerized and then the esters were converted to the acids with MeI. Similar sulfonate PTs have been prepared by Aldissi. It was reported that long alkyl chains attached to the sulfonate group can induced lyotropic liquid crystalline behaviour in the PT “31. Ordered PTs were reported in the patent possess high conductivities. Water soluble sulfonated PTs have also been prepared using the FeC13 [26] method by Ikenoue [184]. Later, Holdcroft [185] has chemically prepared poly(3(3-thieny1)propanesulfonate) (P3TPSNa) also using the FeC13 method (Fig. 11) (85). The focus of this study was to develop water-based photoresists that could used in photolithiography. Polythiophene films can be cast onto solid substrates and irradiated through a photomask. The non-radiated polymer is dissolved away, leaving a negative photoimage of photocrosslinked PT [186]. The HT:HH ratio was found to be 79 :21. The conductivities of the FeC1, doped materials was about 0.001-0.0001 S cm-’. Photoimaging experiments found that P3TPSNa was able to form both negative and positive images, depending on the media conditions of the irradiation and oxidation state of the polymer. One very interesting observation is that P3TPSNa films give a featureless X-ray diffraction spectra implying a completely amorphous polymer, whereas irregular poly(3-hexylthiophene) shows some crystallinity. Holdcroft postulates that ion-pairing promotes random disorder in the polyelectrolyte leading to amorphology. This is opposite of the self-assembly found in regioregular carboxylate PT [168].

Chemical synthesis of heteroatomic functionalized substituents on PTs

37

MeOH, HCI reflux 88 R1 = C&lfs, Ra = (CH&iOH Xll Y =I

Scheme 17. Synthesis of random HT-copolymers of polar PTS.

Hydroxydecyl-functionalized PT has been prepared using the FeC13method [ 1871 (86) (Fig. 11). The aim was to prepare a polymer that could self-assemble similar to the regioregular carboxylate PT [168]. The polymer is found to be =SO% HT coupled and is quite conjugated-with a maximum absorption around 510nm and a estimate gap of 1.8 eV that can be compared to 2.2 eV and 1.7 eV found in regioirregular and regioregular PATS, respectively. Two probe conductivity measurement of iodine doped samples gave values of 0.01-0.1 S cm-' . Similar copolymers bearing a hydroxy functionality have been prepared in a regioregular fashion by Holmes et al. A substituted HT-polythiophene copolymer containing alkyl and w-hydroxyalkyl side chains was prepared [188] (Scheme 17). The w-hydroxyalkyl side chain was first protected with a tetrahydropyranyl (THP) ether, polymerized, and deprotected to give the random copolymer 88 (Scheme 17). The copolymer 88 contains a free alcohol at the end of the side chain and can be functionalized by a number of reagents in order to tailor the properties of the conjugated polymer. Substitution of a sulfur atom directly on the thiophene ring is expected to lower the oxidation potential of the conjugated polymer. Examples with varying alkyl chain length (Scheme 18) (polymers HT-92, HT-93, HT-94, and HT-95) have been synthesized with greater than 90% HT-HT linkages [89]. The solubility is notably poor though, suggestinga stronger affinity between polymer chains and that the lack of solubility leads to low molecular weights ( M , = 4417) as determined by GPC. This lack of solubility is in contrast to the regioirregular polymer from 3-ethylmercaptothiophene prepared by Reynolds [1891 and co-workers. This polymer is soluble in common solvents and has M , = 2200 ( M , = 13000) and a broad polydispersity. The solution UV-Vis spectra in chloroform for HT-93 for example, exhibited three peaks at 263, 324, and 513, with a shoulder at 605 nm. The destabilization of the HOMO reduces the HOMO-LUMO gap and leads to a red shift in these polymers in solution. However, the solid state UV and conductivity do not vary markedly from the alkyl substituted model [76]. The conductivity of I2 doped thin films of these polymers were reported to range from 450-750 S cm-' . The conductivity in irregular poly(3-ethylmercaptothiophene)powder samples is about lop3S cm-'

38

1 The Chemistry of Conducting

1. Rieke Zn I THF 2. Ni(dppe)C12

Brd

91 SB

r

RT to reflux

HT-92 HT-93 HT-94 HT-95

R BwYl HexYl

OW1

Dodecyl

Scheme 18. The synthesis of regioregular PTs with thioether side chains.

[ 1891. A recently study has shown that polymerization of (2,5-dibromo-3-butylthio)thiophene using Mg, followed by Ni(dppp)C12in refluxing anisole leads to an apparently regioregular poly[3-(butylthio)thiophene] that has a molecular weight of 5K. This polymer in contrast to the above Rieke polymer [190] is soluble in CHC13, CCl,, toluene, benzene, THF, and CS2 [191] Kanatzidis has recently prepared the thio analogue to PEDOT, namely poly(3,4ethylenedithiathiophene) [1921 by the FeC13 method. Unfortunately the molecular weight of the polymer was found to be around 3-4K and the solubility was a bit low. However, this exceptionally interesting structure exhibited conductivities of around 0.4 S cm-' and very interesting thermopower behavior reminiscent of a metallic state. Other very interesting PTs have recently been prepared including those containing liquid crystalline side chains [ 1931, dicyanoPTs [1941, and copolymers bearing side chain non-linear optical chromophores [ 1951.

1.5 Fused rings systems A number of fused rings systems containing thophene rings systems have been prepared. The pioneering design and synthesis of poly(is0thianapthene) (PITN) by Wudl [196] propelled the synthesis and study of a large number of new materials. Since isothianapthene contains essentially a four electron cyclohexadiene system, when the benzenoid thiophene is oxidized to the conducting state and forms a quinoidal structure the cyclohexadiene will convert to a benzene structure. This means that the conducting state structure will be very stable energetically and force a large portion of quinoidal structure to be formed, leading a highly planarized polymer with a very low band gap of around l e v . Other methods for the synthesis of

References

39

PITN and its derivatives have been reported including those that are improved [197- 1991. Copolymers containing the isothianapthene unit and thiophene rings has been prepared by Cava and Loray [200]. A comprehensive presentation of PTs containing fused rings is beyond the scope of this review, however a few examples are discussed below. Pomerantz has prepared from FeC13 a dialkylated pyrazine fused PT [201]. Pomerantz has also prepared soluble fused thiophene PTs [202]. Even more elaborate fused bithiophene PTs have been reported by Collard [203]. Fused heterocycles on PTs and copolymers have been prepared by many including dithiole derivatives and others [204, 2051.

1.6 Conclusion Polythiophenes remain as one of the most versatile conjugated polymer systems. The ease of synthesis of a very large number of PT derivatives that can be engineered as new materials in limited only by the imagination. Polythiophenes will continue to lead the way to new unique sensory materials, to highly stable and efficient all-polymer transistors, to very highly conductive plastics, and to new nanoelectronic and nanooptical materials. New advances in the synthesis of regioregular PTs and the discovery of self-assembly in regioregular PTs provide well-defined building blocks that have increased the importance of PTs among conducting polymers. As well-defined materials become more increasingly available, new structure-property relationships through systematic studies of structure/physical property correlations will continue to unfold. This allows chemists, physicists, materials scientists, and engineers to have a better grasp on the development of new technologies. The ease and low cost of processing these polymers can then be exploited for future technologies and continued commercial applications.

References 1. J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H. Friend, P. L Burns, A. B. Holmes, Nature, 1990, 347, 539. 2. H. E. Katz, J . Muter. Chem. 1997, 7, 369. 3. Handbook of Conducting Polymers 2nd Edition (Eds.: T . Skotheim, J. Reynolds, R. Elsenbamer), Marcel Dekker, Inc., New York, NY, USA, 1998. 4. Handbook of Conducting Polymers (Ed.: T . Skotheim), Marcel Dekker, Inc., New York, N Y , USA, 1986. 5. T. Yamamoto, K. Sanechika, A. Yamamoto, J . Polym. Sci., Polym Lett. Ed. 1980, 18, 9. 6. J. W. P. Lin, L. P. Dudek, J. Polym. Sci., Polym. Chem. Ed. 1980, 18,2869. 7. K. Tamoa, K. Sumitani, M. Kumada, J. Am. Chem. Soc. 1972, 94,4376. 8. T. Yamamoto, A. Morita, Y. Miyazaki, T. Maruyama, H. Wakayama, Z. H. Zhou, Y. Nakamura, T. Kanbara, S. Sasaki, K. Kubota, Macromolecules 1992, 25, 1214. 9. T. Yamamoto, K. Sanechika, A. Yamamoto, Bull. Chem. SOC.Jpn. 1983,56, 1497. 10. T. Yamamoto, K. Osakada, T. Wakabayashi, A. Yamamoto, Makromol. Chem., Rapid Commun. 1985,6, 671.

40

1 The Chemistry of Conducting

11. T. Yamamoto, A. Morita, T. Maruyama, Z. H. Zhou, T. Kanbara, K. Saneckika, Polym. J. (Tokyo) 1990,22, 187. 12. T. Yamamoto, T. Maruyama, Z. H. Zhou, Y. Miyazaki, T. Kanbara, K. Saneckika, Synth. Met. 1991, 41, 345. 13. C. 2. Hotz, P. Kovacic, I. A. Khoury, J. Polym. Sci., Polyrn. Chem. Ed. 1983, 21, 2617. 14. M. Kobayashl, J. Chen,T. C. Chung, F. Moraes,A. J. Heeger, F. Wud1,Synth. Met. 1984,9,77. 15. I. Colon, G. T. Kwiatkowski, J. Polym. Sci., Polym. Chem. Ed. 1990, 28, 367. 16. A. Berlin, G. A. Pagani, F. Sannicolo, J. Chem. Soc., Chem. Commun. 1986, 1663. 17. L. Julia, A. G. Davies, D. R. Rveda, F. J. Balta Calleja, Chem. Znd. 1989, 78. 18. V. Meyer, Chem. Ber. 1883, 16, 1465. 19. P. Enzel, T. Bein, J. Chem. SOC.,Chem. Commun. 1989, 1326. 20. T. Yamamoto, K. Sanechika, Chem. Znd. (London) 1982, 301. 21. K. Kaneto, K. Yoshino, Y. Inuishi, Solid State Commun. 1983, 46, 389. 22. A. Amer, H. Zimmer, K. J. Mulligan, H. B. Mark, Jr., S. Pons, J. F. McAleer, J. Polym. Scz., Poly. Lett. Ed. 1984, 22, 77. 23. K. Y. Jen, R. Oboodi, R. L. Elsenbaumer, Polym. Mater. Sci. Eng.1985, 53, 79. 24, R. L. Elsenbaumer, K.-Y. Jen, R. Oboodi, Synth. Met. 1986, 15, 169. 25. G. G. Miller, R. L. Elsenbaumer, J. Chem. SOC.,Chem. Commun. 1986, 1346. 26. R. Sugimoto, S. Takeda, H. B. Gu, K. Yoshino, Chem. Express 1986,1, 635. 27. M. Sato, S. Tanaka, K. Kaeriyama, J. Chem. Soc., Chem. Commun. 1986,873. 28. S , Hotta, S. D. D. V. Rughooputh, A. J. Heeger, F. Wudl, Macromolecules 1987, 20, 212. 29. S. A. Chen, C. C. Tsai, Macromolecules 1993,26, 2234. 30. K. Yoshino, S. Hayashi, R. Sugimoto, Jpn. J. Appl. Phys. 1984,23, L899. 31. J.-E. Osterholm, J. Laakso, P. Nyholm, H. Isotalo, H. Stubb, 0. Inganas, W. R. Salaneck, Synth. Met. 1989, 28, C435. 32. S. Hotta, M. Soga, N. Sonoda, Synth. Met. 1988,26, 267. 33. K. Yoshino, S. Nakajima, R. Sugimoto, Jpn. J. Appl. Phys. 1987, 26, L1038. 34. I. Kulszewicz-Bajer,A. Pawlicka, J. Plenkiewicz, A. Pron, S . Lefrant, Synth. Met. 1989, 30, 335. 35. M. Leclerc, F. M. Diaz, G. Wegner, Makromol. Chem. 1989, 190, 3105. 36. M. Pomerantz, J. J. Tseng, H. Zhu, S. J. Sproull, J. R. Reynolds, R.Uitz, H. J. Arnott, H. I. Haider, Synth. Met. 1991,41-43, 825. 37. M. S. A. Abdou, X. Lu, Z. W. Xie, F. Orfino, M. J. Deen, S . Holdcroft, Chem. Muter. 1995, 7, 631. 38. F. Chen, P. G. Mehta, L. Takiff, R. D. McCullough, J. Muter. Chem. 1996,6, 1763. 39. V. M. Niemi, P. Knuuttila, J.-E. Osterholm, J. Korvola, Polymer 1992, 33, 1559. 40. J. Laakso, H. Jarvinen, B. Sagerberg, Synth. Met. 1993,55-57, 1204. 41. H. Jarvinen, L. Lahtinen, J. Nasman, 0. H o d , A.-L. Tammi, Synth. Met. 1995,69,299. 42. T. Taka, P. Nyholm, J. Laakso, M. T. Loponen, J. E. Osterholm, Synt. Met. 1991,41-43,899. 43. K. Yoshino, S. Nakajima, S. Fuji, R. Sugimoto, Polym. Commun. 1987, 28, 309. 44. 0. Inganas, W. R. Salaneck, J. Osterholm, J. Laasko, Synth. Met. 1988, 22, 395. 45. J. R. Reynolds, J. P. Ruiz, A. D. Child, K. Nayak, D. S. Marynick Macromolecules 1991,24, 678. 46. K. Yoshino, S. Nakajima, M. Onada, R. Sugimoto Synth. Met. 1989, 28, C349. 47. 0. R. Gatum, P. H. J. Carlsen, E. J. Samuelsen, J. Mardalen, Synth. Met. 1993,58, 115. 48. S. Hotta, M. Soga, N. Sonoda, Synth. Met. 1988, 26, 267. 49. M. D. McClain, D. A. Whittington, D. J. Mitchell, M. D. Curtis, J. Am. Chem. SOC.1995, I1 7, 3887. 50. J. Roncali, Chem. Rev. 1992,92, 711. 51. P. C. Stein, C. Botta, A. Bolognesi, M. Catellani, Synth. Met. 1995, 69, 305. 52. S. Yue, G. C. Berry, R. D. McCullough, Macromolecules 1996, 29, 933. 53. M. Sato, H. Morii, Polym. Commun. 1991,32, 42. 54. M. Sato, H. Morii, Macromolecules 1991, 24, 1196. 55. R. L. Elsenbaumer, K-Y. Jen, G. G . Miller, H. Eckhardt, L.W. Shacklette, R. Jow, in Electronic Properties of Conjugated Polymers, (Eds.: H. Kuzmany, M. Mehring, S. Roth), Springer Series in Solid State Sciences, Springer, New York 1987, vol. 76, p. 400.

References

41

56. R. M. Souto Maior, K. Hinkelmann, F. Wudl, Macromolecules 1990, 23, 1268. 57. M. Zagorska, 1. Kulszewicz-Bajer, A. Pron, L. Firlcj, P. Berier, M. Galtier, Synth. Met. 1991, 45, 385. 58. M. Zagorska, B. Krishe, Polymer 1990, 31, 1379. 59. B. Krische, J. Hellberg, C. Lilja, J . Chem. Soc., Chem. Commun. 1987, 19, 1476. 60. A. Berlin, G. A. Pagani, F. Sannicolcj, J . Chem. SOC.,Chem. Commun. 1986, 1663. 61. M. J. Marsella, T. M. Swager, J . Am. Chem. Soc. 1993, 115, 12214. 62. R. D. McCullough, R. D. Lowe, M. Jayaraman, D. L. Anderson, J. Org. Chem. 1993,58,904. 63. G. Barbarella, A. Bongini, M. Zambianchi, Macromolecules 1994, 27, 3039. 64. G. Barbarella, M. Zambianchi, A. Bongini, L. Antolini, Adv. Muter. 1994, 6, 561. 65. R. H. Baughman, R. R. Chance, J . Appl. Phys. 1976,47,4295. 66. J. L. Bredas, J. Chem. Phys. 1985,82, 3809. 67. R. D. McCullough, R. D. Lowe, J . Chem. Soc., Chem. Commun. 1992,70. 68. C. Van Pham, H. B. Mark, Jr.; H. Zimmer, Synth. Commun. 1986, 16, 689. 69. D. D. Cunningham, L. Laguren-Davidson, H. B. Mark, Jr.; C . Van Pham, H. Zimmer, J. Chem. SOC.,Chem. Commun. 1987, 1021. 70. G. Consiglio, S. Gronowitz, A-B. Hornfeldt, B. Maltesson, R. Noto, D. Spinelli, Chemica Scripta 1977, IZ, 175. 71. K. Tamao, K. Sumitani, M Kumada, J . Am. Chem. Soc. 1972,94,9268. 72. K. Tamao, K. Sumitani, Y. Kiso, M. Zembayashi, A. Fujioka, S. Kodama, I. Nakajima, A. Minato, M Kumada, Bull. Chem. Soc. Jpn. 1976, 49, 1958. 73. S. Kodama, I. Naajima, M Kumada, A. Minato, K. Suzuki, Tetrahedron 1982, 38, 3347. 74. R. D. McCullough, R. D. Lowe, Polym.. Prepr. 1992, 33, 195. 75. R. D. McCullough, R. D. Lowe, M. Jayaraman, P. C. Ewbank, D. L. Anderson, S. TristramNagle, Synth. Met. 1993, 55, 1198. 76. R. D. McCullough, S. Tristram-Nagle, S. P. Williams, R. D. Lowe, M. Jayaraman, J . Am. Chem. Soc. 1993, 115,4910. 77. R. D. McCullough, J. A. Belot, S. P. Williams, Molecular Engineering of Advanced Materials (Eds.: J. Becher, K. Schaumburg) NATO Adv. Res. Workshop Series, Series C: Math. and Phys. Sci., 1995, Vol. 456, p. 349. 78. R. D. McCullough, S. P. Williams, S. Tristram-Nagle, M. Jayaraman, P. C. Ewbank, L. Miller, Synth. Met. 1995, 69, 279. 79. R. D. McCullough, S. P. Williams, M. Jayaraman, J. Reddinger, L. Miller, S . Tristram-Nagle, Electrical, Optical, and Magnetic Properties of Organic Solid State Materials (Eds.: L. Dalton, C. Lee), Mater. Res. SOC.,Pittsburgh, PA 1994, Vol. 328, p. 215. 80. C. Soucy-Breau, A. MacEachern, L. C. Leitch, T. Arnason, P. Morand, J. Heterocycl. Chem. 1991, 28, 41 1. 81. A. MacEarchern, C. Soucy, L. C. Leitch, T. Arnason, P. Morand, Tetrahedron 1988,44,2403. 82. C. van Phan, R. S. Macomber, H. B. Mark, Jr., H. Zimmer, J. Urg. Chem. 1984,49, 5250. 83. G. M. Davies, P. S. Daview, Tetrahedron Lett. 1972, 8507. 84. S. Kano, Y. Yuasa, T. Yokomatsu, S . Shibuya, Heterocycles 1983, 20, 2035. 85. T-A. Chen, R. D. Rieke, J. Am. Chem. SOC.1992,114, 10087. 86. T-A. Chen, R. D. Rieke, Synth. Met. 1993, 60, 175. 87. T-A. Chen, R. A. O’Brien, R. D. Rieke, Macromolecules 1993,26, 3462. 88. T-A. Chen, X. Wu, R. D. Rieke, J . Am. Chem. SOC.1995, 117,233. 89. X. Wu, T-A. Chen, R. D. Rieke, Macromolecules 1995, 28,2101. 90. H. Mao, B. Xu, S . Holdcroft, Macromolecules 1993, 26, 1163. 91. M. Kumada, Pure Appl. Chem. 1980,52, 669. 92. E. Negishi, Pure Appl. Chem. 1981,53, 2333. 93. N. Bumagin, I. P. Beletskaya, Russ. Chem. Rev. 1990, 59, 1174. 94. J. K. Stille, Angew. Chem., Int. Ed. Engl. 1986, 25, 508. 95. V. N. Kalinin, Synthesis 1992, 413. 96. R. D. McCullough, R. D. Loewe, unpublished results. 97. T. Bjnrnholm, D. R. Greve, T. Geisler, J. C. Petersen, M. Jayaraman, R. D. McCullough, Adv. Mater. 1996, 8, 920. 98. J. Roncali, A. Yassar, F. Garnier, J . Chem. SOC.,Chem. Commun. 1988, 581.

42

1 The Chemistry of Conducting

99. A. Yassar, J. Roncali, F. Garnier, Macromolecules 1989, 22, 804. 100. T. J. Prosa, M. J. Winokur, R. D. McCullough, Macromolecules 1996,29, 3654. 101. R. D. McCullough, P. C. Ewbank in Handbook of Conducting Polymers (Eds.: T. Skotheim, R. L. Elsenbaumer,J. R. Reynolds), Marcel Dekker Inc., New York, USA 1997, Vol. 2,225. 102. M. R. Anderson, D. Selse, M. Berggren, H. Jarvinen, T. Hjertberg, 0. Inganas, 0.Wennerstrom, J. E. Osterholm, Macromolecules 1994, 27, 6503. 103. I. Levesque, M. Leclerc, J. Chem. SOC.,Chem. Commun. 1995,2293. 104. M. Pei, 0. Inganas, G. Gustafsson, M. Granstrom, M. Anderson, T. Hjertberg, 0. Wennerstrom, J. E. Osterholm, J. Laasko, H. Jarvinen Synth. Met. 1993, 55-57, 1221. 105. R. D. McCullough, M. Jayaraman, J. Chem. SOC.,Chem. Commun., 1995, 135. 106. W. A. Goedel, N. S. Somanathan, V. Enkelmann, G. Wegner, Makromol. Chem. 1992, 193, 1195. 107. N. Somanathan, G. Wegner, Synt. Met. 1995, 75, 123. 108. M. R. Anderson, M. Berggren, G. Gustafsson, T. Hjertberg, 0. Inganas, 0. Wennerstrom, Synth. Met. 1995, 71, 2183. 109. G. Bidan, S. Guillerez, V. Sorokin, Adv. Muter. 1996, 8, 157. 110. M. M. Bouman, E. W. Meijer, Adv. Muter. 1995, 7, 385. 111. L. Robitaille, M. Leclerc, Macromolecules 1994, 27, 1847. 112. J. S. Middlecoff, D. M. Collard, Synth. Met. 1997,84, 221. 113. M. Ueda, Y. Miyaji, T. Ito, Y. Oba, T. Sone, Macromolecules 1991, 24, 2694. 114. F. Garnier, Adv. Muter. 1989, 1, 513. 115. J. W. Thackeray, H. S. White, M. S. Wrighton, J. Phys. Chem. 1985,89, 5133. 116. E. D. Chidsey, R. W. Murray, Science 1986,231, 25. 117. R. H. Baughman, R. L. Elsenbaumer, Z. Iqbal, G. G. Miller, H. Eckardt in Electronic Properties of Conjugated Polymers (Eds. H. Kuzmany, M. Mehring, s. Roth), Springer-Verlag, Berlin, 1987,432. 118. R. M. Wightman, Science 1988,240,415. 119. R. H. Baughman, Makromol. Chem., Macromol. Symp. 1991,51, 1. 120. J. Roncali, Chem. Rev. 1997, 97, 369. 121. R. L. Blankespoor, L. L.; Miller, J. Chem. SOC.,Chem. Commun. 1985, 90. 122. M. R. Bryce, A. Chissel, P. Kathirgamanthan, D. Parker, N. R. M. Smith, J. Chem. SOC., Chem. Cornmun. 1987,466. 123. A. C. Chang, R. L. Blankespoor, L. L. Miller, J. Electroanal. Chem. 1987,236, 239. 124. S. Tanaka, M. A. Sato, K. Kaeriyama, Synth. Met. 1988, 25, 277. 125. M. Feldhues, G. Kampf, H. Litterer, T. Mechlenburg, P. Wegener, Synth. Met. 1989, 28, C487. 126. T. Yamamoto, A. Kashiwazaki, K. Kato, Makromol. Chem. 1989, 190, 1649. 127. J. Roncali, P. Marque, R. Garreau, F. Gamier, M. Lemaire, Macromolecules 1990,23, 1347. 128. G. Daoust, M. Leclerc. Macromolecules 1991, 24, 455. Jpn. 1989,62, 1908. 129. S. Tanaka, K. Kaeriyama, Bull. Chem. SOC. 130. M. Leclerc, G. Daoust, J. Chem. Soc., Chem. Commun. 1990,273. 131. R. Cloutier, M. LeclercJ. Chem. SOC.,Chem. Commun. 1991, 1194. 132. K. Faid, R. Cloutier, M. Leclerc, Macromolecules 1993, 26, 2501. 133. S. A. Chen, C. C. Tsai, Macromolecules 1993, 26, 2234. 134. M. C. Gallazzi, L. Castellani, R. A. Marin, G. Zerbi, J. Polym. Sci., Poly. Chem. Ed. 1993,31, 3339. 135. L. Robitaille, M. Leclerc, C. L. Callender, Chem. Muter. 1993,5, 1755. 136. A. Bolognesi, G. Bajo, A. Geng, W. Porzio, F. Speroni, Thin Solid Films 1994,243, 683. 137. C. L. Callender, C. A. Carere, G. Daoust, M. Leclerc, Thin Solid Films 1991,204, 451. 138. A. Bolognesi, C. Botta, Z. Geng, C. Flores, L. Denti, Synth. Met. 1995, 71, 2191. 139. L. Belobrzeckaja, G. Bajo, A. Bolognesi, M. Castellani Synth. Met. 1997, 84, 195. 140. F. Jonas, L. Schrader, Synth. Met. 1991, 41-43, 831. 141. G. Heywang, F. Jonas, Adv. Muter. 1992, 4 , 116. 142. Q. Pei, G. Zuccarello, M. Ahlskog, 0. Inganas, Polymer 1994,35, 1347. 143. A. Kumar, J. Reynolds, Macromolecules 1996,29, 7629. 144. Y. Miyazaki, T. Kanbara, K. Osakada, T. Yamamoto, K. Kubota, Polym. J. 1994, 26,509.

References

43

145. T. M. Swager, M. Marsella, Adv. Muter. 1994, 6, 595. 146. G. Zotti, Synth. Met. 1992, 51, 373. 147. P Bauerle, G. Gotz, M. Hiller, S. Scheib, T. Fischer, A. Segelbacher, M. Bennati, A. Grupp, M. Mehring, M. Stoldt, C. Seidle, F. Geiger, H. Schweizer, E. Umbach, M. Schmelzer, S. Roth, H. Egelhaaf, D. Oelkrug, P. Emele, H. Port, Synth. Met. 1993, 61, 71. 148. R. D. McCullough, S. P. Williams, J. Am. Chem. SOC.1993,115, 11608. 149. M. J. Marsella, T. M. Swager, J. Am. Chem. SOC.1993, 115, 12214. 150. R. D. McCullough, S. P. Williams, Chem. Muter. 1995, 7, 2001. 151. M. J. Marsella, P. J. Carroll, T. M. Swager, J. Am. Chem. SOC.1994, 116, 9347. 152. M. J. Marsella, P. J. Carroll, T. M. Swager, J. Am. Chem. SOC.1995, 117, 9832. 153. M. J. Marsella, R. J. Newland, P. J. Carroll, T. M. Swager, J. Am. Chem. SOC.1995,117,9842. 154. A. Iraqi, J. A. Crayston, J. C. Walton J. Muter. Chem. 1995, 5, 1831. 155. S. D. D. V.Rughooputh, S. Hotta, A. J. Heeger, F. Wudl, J. Polym. Sci., Polym. Phys. Ed. 1987,25, 1071. 156. P. 0. Ekeblad, 0. Inganas, Polym. Commun. 1991,32,436. 157. D. J. Sandman, Trends in Polym. Sci. 1994, 2, 44. 158. 0. Inganas, Trends in Polym. Sci. 1994, 2, 189. 159. K. Faid, M. Frechette, M. Ranger, L. Mazerolle, I. Levesque, M. Leclerc, T. A. Chen, R. D. Rieke, Chem. Mater. 1995, 7, 1390. 160. I. Levesque, M. Leclerc, J. Chem. Soc., Chem. Commun. 1995,2293. 161. I. Levesque, M. Leclerc, Chem. Muter. 1996, 8, 2843. 162. L. Belobrzeckaja, G. Bajo, A. Bolognesi, M. Catellani, Synth. Met. 1997, 84, 195. 163. D. H. Charych, J. 0. Nagy, W. Spevak, M. D. Bednarski, Science 1993,261, 585. 164. M. Lemaire, D. Delabouglise, R. Garreau, A. Guy, J. Roncali, J. Chem. SOC.,Chem. Commun. 1988,658. 165. M. M. Bouman, E. E. Havinga, R. A. Janssen, E. W. Meijer, Mol. Cryst. Liq. Cryst. 1994,256, 439. 166. B. M. W. Langeveld-Voss,R. A. J. Janssen, M. P. T. Christiaans, S. C. J. Meskers, H. P. J. M. Dekkers, E. W. Meijer, J. Am. Chem. SOC.1996, 118, 4908. 167. B. M. W. Langeveld-Voss, E. Peeters, R. A. J. Janssen, E. W. Meijer, Synth. Met. 1997, 84, 611. 168. R. D. McCullough, P. C. Ewbank, R. S. Loewe, J. Am. Chem. SOC.1997, 119, 633. 169. R. D. McCullough, P. C. Ewbank, Synth. Met. 1997, 84, 311. 170. P. Bauerle, K.-U. Gaudl, F. Wurthner, N. S. Sariciftci, H. Neugebauer, M. Mehring, C. Zhong, K. Doblhofer, Adv. Muter. 1990, 2, 490. 171. P. Englebienne, M. Weiland, J. Chem. SOC.,Chem. Commun. 1996, 1651. 172. H. Masuda, K. Kaeriyama, Makromol. Chem., Rapid Commun. 1992, 13,461. 173. F. Andreani, P. C. Bizzari, C. Della Casa, E. Salatelli, Polym. Bull. 1991, 27, 117. 174. P. C. Bizzari, F. Andreani, C. Della Casa, M. Lanzi, E. Salatelli, Synth. Met. 1995, 75, 141. 175. M. Pomerantz, H. Yang, Y . Cheng, Macromolecules 1995,28, 5706. 176. M. Pomerantz, Y. Cheng, R.K. Kasim, R. L. ELsenbaumer, Synth. Met. 1997,85, 1235. 177. J. Lowe, S. Holdcroft, Macromolecules 1995, 28, 4608. 178. M. Liu, R. V. Gregory, Synth. Met. 1995, 72, 45. 179. N. Hirota, N. Hisamatsu, S. Maeda, H. Tsukahara, K. Hyodo, Synth. Met. 1996, 80, 67. 180. Y. Ikenoue, J. Chiang, A. 0. Patil, F. Wudl, A. J. Heeger, J. Am. Chem. SOC.1988,110,2983. 181. A. 0. Patil, Y. Ikenoue, F. Wudl, A. J. Heeger, J. Am. Chem. SOC.1987, 109, 1858. 182. Y. lkenoue, N. Uotani, A. 0. Patil, F. Wudl, A. J. Heeger, Synth. Met. 1989, 30, 305. 183. M. Aldissi, U. S. Patent 155,450, 1988. 184. Y. lkenoue, Y. Saida, M. Kira, H. Tomozawa, H. Yashima, M. Kobayashi, J. Chem. SOC., Chem. Commun. 1990, 1694. 185. M. I. Arroyo-Villan, G. A. Diaz-Quijada, M. S. A. Abdou, S. Holdcroft, Macromolecules 1995,28,975. 186. M. S. A. Abdou, G. Diaz-Qijada, I. Arroyo, S. Holdcroft, Chem. Muter. 1991, 3, 1003. 187. C. D. Casa, F. Bertinelli, P. C. Bizzarri, E. Salatelli, Adv. Muter. 1995, 7, 1005. 188. K. A. Murray, S. C. Moratti, D. R. Baigent, N. C. Greenham, K. Pichler, A. B. Holmes, R. H. Friend, Synth. Met. 1995,6Y, 395.

44 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205.

I The Chemistry of Conducting

J. P. Ruiz, K. Nayak, D. S. Marynick, J. R. Reynolds, Macromolecules 1989, 22, 1231. X. Wu, T. A. Chen, R. D. Rieke, Macromolecules 1996, 29, 7671. F. Goldoni, D. Iarossi, A. Mucci, L. Schenetti, M. Zambianchi, J. Muter. Chem. 1997, 7,593. C. Wang, J. L. Schindler, C. R. Kannewurf, M. G. Kanatzidis, Chem. Muter. 1995, 7, 58. R. Toyosima, M. Narita, K. Akagi, H. Shirakawa, Synth. Met. 1995, 69, 289. F. Hide, Y. Greenwald, F. Wudl, A. J. Heeger, Synth. Met. 1997,85, 1255. K. G. Chittibabu, L. Li, M. Kamath, J. Kumar, S. K. Tripathy, Chem. Muter. 1994, 6, 475. F. Wudl, M. Kobayashi, A. J. Heeger, J . Org. Chem. 1984,49, 3382. T. Iyoda, M. Kitano, T. Shimidzu, J. Chem. SOC.,Chem. Commun. 1991, 1618. F. Wang, A. B. Kon, T. L. Rose, Synt. Met. 1997,84, 69. H. Paulussen, D. Vanderzande, J. Genaner, Synt. Met. 1997,84, 415. D. Loray, M. P. Cava, Adv. Muter. 1992, 4, 562. M. Pomerantz, B. Chalomer-Gill, L. 0. Harding, J. J. Tseng, W. J. Pomerantz, Synth. Met. 1993, 55-57, 960. M. Pomerantz, X. Gu, Synt. Met. 1997,84,243. S. Inaoka, D. M. Collard, Synt. Met. 1997,84, 193. M. Kozaki, S. Tanaka, Y. Yamashita, J . Chem. SOC.,Chem. Commun. 1992, 1137. M. Karikomi, C. Kitamura, S. Tanaka, Y. Yamashita, J. Am. Chem. SOC.1995, 117,6791.

2 Electronic Properties of Polythiophenes Shu Hotta and Kohzo Ito

2.1 General aspects of conducting polymers Conducting polymers are characterized by a (quasi)one-dimensional system along whch mobile electrons extend. Most of them possess a .rr-electronic backbone with a few exceptions sorted as a-electronic materials such as polysilanes. In the linearly extended ideal system comprising polymer chains with very long sequences, translational symmetry is satisfied along these chains and band structure (bandwidth, band gap, etc.) is determined in a k-space (Brillouin zone) [l]. Tetrahedral covalent solids whose electronic system consists of s- and p-electrons tend to have an energy gap when the atomic separation is small [2].This is probably the case with any linear conducting polymers, characterizing them essentially as semiconductors. In the case of polythiophene, the translational symmetry is satisfied by the fullystretched S-anti conformation [3]. Individual repeated units comprise a couple of thiophene rings in this planar conformation and are arranged with a period approx. 7.8 A along the chain direction (Fig. 1). In a rigid lattice of the polythiophene, the planar polymer chains having the S-anti conformation are packed in a well-known herringbone structure [4]where the polymer chains turn aside somewhat from the face-to-face arrangement. More recently polythiophenes appropriately substituted with various chemical groups (such as alkyl) [5,6] showed interesting structural characteristics on account of the presence of these substituents. We count among these characteristics a peculiar morphology based upon a rigorous face-to-face packing of the polymer backbones also featured by the S-anti conformation. This morphology is probably responsible for e.g. unusually high conductivity up to 5500 S cm-’ [7] achieved for some regioregular alkyl- or alkoxy-substituted polythiophenes. Unlike conventional inorganic semiconducting materials such as silicon and germanium, however, the polymer chains are weakly bound through the van der Waals force between them. This weak force is, therefore, easily destroyed by the intrachain rotation around the a-bonding interconnecting the thiophene rings. As a result, structural disorder or defects are introduced. Various interesting aspects arise in relation to this structural characteristic, because the strong interaction between electronic state and backbone conformation is a general feature of the conducting polymers. This can be visualized as, for instance, chromism. More interesting features of the conducting polymers emerge when elementary excitations (both neutral and charged) are introduced on the polymer chains. The ‘domain-wall’ concept (see Fig. 2 of Ref. 8 for the structure) proposed earlier in the history of the conducting polymers research not only plays a crucial role in establishing physical concepts of the conducting polymers but provides an appropriate chemical insight into the structure of those excitations.

46

2 Electronic Properties of Polythiophenes

....

.... I

r

....

I

7.8 A

7

....

Figure 1. (a) Fully-stretched S-nnti co9formation of polythiophene. A repeat unit consisting of two thiophene rings has a period 7.8A. (b) B (benzenoid) and Q (quinoid) phases in the polythiophene. The Q phase of a finite length is sandwiched between the two domain-walls with the B phase on either side. The B phase has a lower energy per unit length than the Q phase (nondegeneracy). The diagram schematically depicts two unpaired electrons (radicals) that are located on (and near) the two domain-walls. Either of these can be charged positive (hole) or negative (electron).

In this context the conducting polymers are categorized in the following two groups: the first group is characterized by the presence of the degenerate ground state [8], including trans-polyacetylene as the most typical substance; the other group includes cis-polyacetylene, poly(p-phenylene), polythiophene, polypyrrole, etc., which are characterized by the nondegenerate ground state [8, 91. Although a single isolated domain-wall is stable in the degenerate polymer chain with very long sequences, the energy requirement prevents such sequences from existing in the nondegenerate polymer systems. The situation exemplified by polythiophene is as follows: suppose in Fig. 1 there are two phases (domains) B and Q with B energetically more stable. In the diagram, B and Q phases represent the benzenoid and quinoid structures, respectively. The stability can be measured as an energy difference per repeated unit between the two phases. Consequently, the Q phase of very large length separated from the B phase by the domain-wall would have enormously larger energy than the corresponding B phase, and so such a Q phase cannot be stable. However, if the Q phase of a finite length is sandwiched between the two domain-walls with the B phases on either side, the situation is totally different. In this case the Q phase can be present, even though destabilized relative to the B phase. In particular, when charge is injected on the polymer chains, the electron-lattice coupling may well stabilize the elementary excitations located on (and near) the Q phase. This is analogously the case when a pair of a positive charge (hole) and a negative charge (electron) is introduced at once via e.g. photoexcitation. These elementary

2.1 General aspects of conducting polymers

47

excitations both neutral and charged may be termed ‘confined soliton pairs’ [lo]. Note here that the chemical structure similar to Fig. 1b representing the (charge) excitations in the nondegenerate polymers was originally proposed by BrCdas and co-workers for poly(p-phenylene) [9]. These excitations are characterized by strength of confinement which directly measures the extent of lifting of the degeneracy and determines the energy levels positioned within the energy gap. Furthermore, the theories predict that the charged elementary excitations are classified into species with or without a spin. The above-mentioned arguments were originally developed for a single isolated polymer chain that satisfies the translational symmetry. Nonetheless, the discussion can basically apply without loss of generality to polymer systems where either the interchain interaction is responsible or the translational symmetry has been broken, apart from the fact that the equation of motion can no longer be solved analytically [I 11. This is probably true of real polymer systems. Experimentally, the energy levels occurring within the energy gap can be determined relative to the band edge by (optical) transition energy below the gap (subgap transition). Furthermore, measurement of magnetic susceptibility provides a powerful clue to determining whether the charge excitation carries a spin. This stimulated many researchers to examine the applicability of the theories. Having these situations as a background, we describe in this chapter the electronic properties of the polythiophenes. Special attention is directed to (i) the structure and conformation (section 2.2) and (ii) various electronic processes (section 2.3) of these materials. The latter section mainly deals with the electronic properties relevant to the charged states. When we deal with the real polymer systems, disorder effects have to be taken into consideration. This is because in the (quasi) one-dimensional system the disorder tends to make the electronic states localized in cooperation with the electron-lattice coupling [12]. If the disorder is severe, the charges will be transported via hopping among the localized states accompanied by disorder potential as in the case of classical amorphous or non-crystalline media. This will be discussed in sections 2.3.2 and 2.3.3 in relation to the charge transport and recombination in the materials. Another interesting aspect in the studies of the conducting polymers emerges when their properties are investigated in solution. This is because the nature of the conducting polymers is essentially of one-dimensional character and this character shows up prominently in the solution. In particular, when an average separation among the polymer chains is larger than their hydrodynamic diameter [13], one is essentially dealing with the single-chain phenomena. In other words, the intrachain electronic processes can be surveyed in detail by separating them from interchain processes. Since (dis)ordering of the polymers can be tuned by suitable selection of the solvent, one is enabled to study the electronic properties in connection with such (dis)ordering. A further advantage of the solution studies of the conducting polymers lies in the fact that the relative stability between the elementary excitations can be altered by modulating physical parameters of the liquid media. Dielectric constants of solvents may be typically counted as such. These issues will be discussed in section 2.3.4.

48

2 Electronic Properties of Polythiophenes

2.2 Structure and conformation of polythiophenes 2.2.1 Morphology and crystal structure In general, the conducting polymer materials have no single-crystalline structures unlike inorganic semiconductors, except for polydiacetylene. In most of the conducting polymers, polymer chains are bundled to form a fibril or microfibril, and the fibrils are densely entangled with one another, which yields higher-order structures in the solid. This inhomogeneity strongly affects the electronic property of conducting polymers. In the case of polythiophene, electron microscopy shows the fibrillar structure composed of fibrils with diameters of about 25 nm [14]. The fiber diameter increases up to 80nm with doping [14]. The existence of such a fibrillar structure is also confirmed for other polythiophene derivatives using the atomic force microscopy (AFM) [15]. Similarly, the AFM images reveal the nanometer-scale surface morphology that consists of fibrils or bundles with an average size of 40 nm for undoped materials. The average size increases up to lOOnm for doped ones. This change of morphology is explained partly by aggregation of the polymer chains caused by the doping [15]. The structure and conformation of polymer chains in polythiophene and its derivatives have been studied intensively by X-ray diffractometry [3, 5, 6, 16-27], neutron scattering [28], and scanning tunneling microscopy (STM) [29-331. The crystals of the polythiophene (without substituents) represent the parent structure of the polythiophene family. Bruckner and Porsio [3] concluded from the X-ray diffraction profiles of polythiophene that rod-like polymer chains consisting of rigorously coplanar thiophene rings form a well-known heFringbone structure in an orthorhombic cell of a = 7.807, b = 5.526, and c = 7.753 A with the c axis paralleling the chain axis [3]. The size of c corresponds to the period of the planar zigzag conformation along the chain direction (Fig. 1). Mo et al. [18] and Yamamoto et al. [19] independently proposed related structures with a somewhat longer lattice constant c. The polythiophene used by the latter researchers was formed into a thin film by vacuum deposition on a suitable substrate (e.g. carbon, gold, aluminum, etc.) and had a molecular weight approx. 2000 corresponding to a polymerization degree of 24 [19]. Interesting structural variations have recently been found in poly(3-alkylthiophene)’s (PATs) with long alkyl side chains at the 3-position of the thiophene ring. These structures should be referred to the above-mentioned parent structure of the polythiophene. The analysis of the X-ray diffraction patterns has revealed a bundled structure where planar-zigzag rod-like polymer chains form a nearly faceto-face configuration, similar to comblike liquid crystalline polymers with mesogenic side groups [5, 6, 20-271. Note here that the planar-zigzag conformation in the polythiophene backbones of PATs is associated with a close packing of the alkyl side chains [5,21]. The lattice constant along the c-axis is reported very close to 7.8 and related to that for the polythiophene (without substituents), implying again the stretched

A

2.2 Structure and conformation of polythiophenes

49

n

cn

c,

c 3

0 0

m

0

0.0

I

I

I

I

10.0

20.0

30.0

40.0

1 50.0

2 8 Angle (degrees) Figure 2. X-ray diffraction profiles of various poly(3-alkylthiophenes) showing the influence of the side group on structure. Adapted from Synth. Met. 28 (Nos. 1 and 2), C419-26, 1989 (Ref. 20), M. J. Winokur, D. Spiegel, Y. Kim, S. Hotta, and A. J. Heeger, Structural and Absorption Studies of the Thermochromic Transition in Poly(3-hexylthiophenej, with kind permission from Elsevier Science S . A.

S-anti form in the chain. Deviation from this conformation would result in another repetition distance, exemplified by carefully conducted diffraction measurements and peaks assignment [23]. The a-axis spacing increases linearly with the contour length of the side chain, which indicates that the alkyl side chains with the extended trans conformation are oriented along the lateral a-axis direction [5]. This progression is evidenced in Fig. 2 [20] that displays increasing diffraction spacings (i.e. decreasing diffraction angles) with increased side-chain length up to the number of carbons six. The line shape getting sharper in this order is indicative of the structural ordering produced by long alkyl groups. Many researchers have also confirmed the a-axis progression [5] using PATS with longer alkyl-chain lengths. The lattice constant b representing a separation between the face-to-face backbone stacks, however, appears to differ from material to material without regular progression. In orthorhombic crystals [5] the separation may be measured by b/2 (b = 7.75 while this could be b sin Q! in monoclinic ones [23] (where a is a lattice constant weich is not a right angle). The separation can thus be estimated to be arouad 3.8 A. Bolognesi et al. [34], however, reported an unusually large separation (>7 A) for poly(3-decylthiophene) (PDT) and concluded accordingly that interchain electronic coupling is weak. Temperature-dependent X-ray profiles give further important information on the polymer morphology. Figure 3 [20] shows thermal cycling of poly(3-hexylthiophene), PHT. The major peak is reversibly sharpened at lower temperatures with concomitant increase in the diffraction angle, indicating that melting transition occurs, accompanied by the lattice expansion. Winokur et al. [20] inferred

A),

50

2 Electronic Properties of Polythiophenes

65C

1 -

2

4

6

8

1012

29 Angle (degrees)

4

6

8

10 12

14

28 Angle (degrees)

Figure 3. X-ray diffraction profiles of solution cast unoriented poly(3-hexylthiophene), PHT on (a) heating and (b) cooling cycles. Adapted with permission from Ref. 20. Copyright 1989 Elsevier Science S. A.

from the persistence of the diffraction peak around 5" that nematic alignment of the polymer exists even at high temperatures. On the basis of the peak profile analysis, they also pointed out that this diffraction is a superposition of two Gaussians, a narrow peak sitting atop another broad diffuse peak. The analysis of the width and intensity of these two peaks is likely to rationalize the assumption that two phases coexist as crystalline and nematic regions 1201. Further microscopic change in morphology can be probed by the infrared (IR) spectroscopy through examining specific modes of vibration. Tashiro et al. [5] showed that as temperature increases, the gauche band of the alkyl side chains, which represents disordering, is observed clearly in IR spectra above the transition temperature. In combination with the results of the X-ray diffraction experiments, they remarked that although at low temperatures the skeletal chains (i.e. polythiophene backbones) and the alkyl side-chains assume an almost fully extended trans conformation, the conformational disordering in both the skeletal and side chains becomes significant with temperature increasing and approaching the transition region. Note that this transition temperature ranged 0 to 80°C for poly(3-dodecylthiophene), PDDT, for example. Tashiro et al. [5] attributed the side-chain disordering to the trans-gauche transformation and sought the origin of the skeletal-chain disordering from torsional rotations around the a-bonding interconnecting the thiophene rings. Meanwhile, Winokur et al. [20] observed temperaturedependent intensity change of a vibrational band peaking at 1261 cm-' for PHT. They ascribed this band to the stretching mode of the C-C a-bonding mentioned

2.2 Structure and conformation ojpolythiophenes

51

above. This mode, which arose when the material was scanned from the room temperature, did not revert to its initial form through a reversed cooling scan, but retain its intensity the material exhibited at high temperatures. This permanent change may be referred to symmetry breaking within the skeletal chains of PHT [20]. As outlined above, most of the X-ray diffraction experiments show that the planar zigzag S-anti conformation prevails in the polythiophenes. Recent STM investigations, however, directly revealed the existence of other peculiar conformations such as (super)helical one [31] besides the 'normal' S-anti form. On the basis of the STM observation of PHT and poly(3-octylthiophene) (POT) adsorbed on cleaved graphite, Lacaze et al. [33] reported a helical structure with a 19A pitch (for POT) consisting of eight thiophene rings. They also observed a planar S-syn conformation (for PHT) as well as the S-anti form (for PHT) that comprises twelve rings aligned epitaxially on a graphite surface [33]. According to the theoretical calculations by the modified neglect of diatomic overlap (MNDO) method, there existed two minima at dihedral angles (SC,C,rS) of 180" and approx. 35" in the torsional potential curve of polythiophene [35]. The former conformation, which corresponds to the planar zigzag or rodlike conformation of S-anti forms (Fig. l), is more stable than the latter twisted one. The torsional potential curve of poly(3-methylthiophene), on the other hand, exhibited two minima at 160" and 34"with almost the same energy levels [35]. The change in the potential curve arises from the strong repulsion between sulfur and the adjacent methyl group. Since tke dihedral angle of 34" corresponds to a helical structure with a radius of approx. 8 A, this calculation may support the existence of a stable helical conformation in PATs. Doping produces specific effects upon structures of the polythiophenes according to how the dopants are accommodated in the polymer matrices. Winokur et al. [6] traced the structural changes of POT and PDDT with increasing iodine uptake. They found structural transformation upon doping to be accompanied by continuous change in the interlayer lateral separations (represented by the lattice constant a). They assumed from these observations that the polyiodide anions (mostly 13) are located alongside the backbone main chains. This is contrasted with a model proposed by Yamamoto et al. [19a] where the polyiodide anions (mostly I?) parallel the backbone chains of the polythiophene (without substituents). This contrast testifies a specific role of the alkyl side-chains of PATs. Kawai et al. [36] proposed another model where the dopant anions (toluene-p-sulfonate) are accommodated in a cuboidal vacant space surrounded by the alkyl chains and backbones. These peculiar structures are thus believed to result from the topological characteristics of the host polymers and dopants. In relation to the specific effects described above, electrochemically synthesized polythiophenes also exhibit interesting morphologies [ 17, 30, 3 1, 331. The STM images again indicate that the (super)helical structures are involved. The cooperation between the host polymers and dopant ions seems responsible for those structures as well. In this context, Hotta [37] pointed out that rapid deposition of the polymer chains that follows immediately after the electrochemical generation of those polymers may well lead to rather unusual crystal structures.

52

2 Electronic Properties of Polythiophenes

2.2.2 Conformational features The strong interaction between electronic state and backbone conformation is a unique and general feature of conjugated polymers. This feature shows itself most prominently in the electronic absorption spectra. This is because a well-extended polymer backbone causes a redshift in absorption spectra, whereas a disordered backbone conformation results in a blueshift (with concomitant loss of the sharp feature near the absorption edge). The redshift and blueshift are believed to result from extension and segmentation of the electronic wave functions, respectively [38]. Such spectral change reversibly takes place typically either by changing temperatures (thermochromism) or by addition/removal of solvent (solvatochromism). This reversible change can be understood in terms of the order-disorder phenomenon. These chromisms were first recognized for a series of polydiacetylene [39] and polysilane [40] compounds. In these systems, extent of the disorder can be measured by deviation from the fully-stretched all-trans backbone conformation [41]. In the case of the polythiophene and its derivatives, the conformational disorder is caused by the distortion around the a-bonding interconnecting the thiophene rings [42]. The 7r-delocalization along the polythiophene backbone will be maximized when the polymer chains assume the fully-stretched S-anti form. This delocalization will be hampered by the ring distortion of any amount. Figures 4 [42] and 5 [43] show typical illustrations of the thermochromism. A series of spectra in both these figures were taken using two kinds of PHTs with different molecular weight. The PHT for Fig. 4, which was synthesized electrocheof about two [38] mically, had Mw = 48000 with a polydispersity index

Figure 4. Absorption spectra of PHT (-1.4 x M) in 2,5-dimethyltetrahydrofuranat various temperatures. (a) -28"C, (b) -7S"C, (c) 5.5"C,(d) 7.OoC,and (e) 22°C. The polymer concentration is referred to the thiophene ring unit. Chromism of Soluble Polythienylenes, S . D. D. V. Rughooputh, S. Hotta, A. J. Heeger, and F. Wudl, Journal of Polymer Science: Part B: Polymer Physics, Copyright 1987 John Wiley & Sons, Inc. Reprinted from Ref. 42, pp. 1071-8, by permission of John Wiley & Sons, Inc.

2.2 Structure and conformation of polythiophenes

53

where A?w and MNdenote the weight-average molecular weight and number-average molecular weight, respectively. Meanwhile, PHT for Fig. 5 synthesized via a chemical route had A?w = 250000 with MW/MNof 5.5 [44]. The spectra of the former compound were recorded in 2,5-dimethyltetrahydrofuranand those of the latter were measured in dichloromethane. What can be seen in common for the two sets of spectra is that the absorption bands grow with decreasing temperatures at 520, 560, and 605nm. The two sets of spectra, however, differ in the following aspects: (i) Although the band at 520 nm is noted barely as a shoulder in Fig. 4, the corresponding band in Fig. 5 continues growing and is resolved as a real peak. At the lowest temperatures, as a result, the absorption maximum around 450 nm has been replaced with the band at 520 nm. (ii) An isosbestic point is clearly observed in Fig. 5 for temperatures ranging up to 285 K. However, it could not be noticed in Fig. 4, but the intersection of a series of spectra moves continuously toward the longer wavelength region with decreasing temperatures. Notice in this connection that in Fig. 5 a transition takes place quite suddenly within a narrow temperature range. The isosbestic point no longer subsists after the disordering transition has taken place, suggesting the occurrence of another phase that is more disordered. Similar spectral change can also be noticed when poor solvent is added to the system (solvatochromism). The color change caused by dissolving the polymer or removing the solvent from the polymer solution can also be referred to as the solvatochromism [38]. Recently Leclerc et al. [45] and Sandstedt et al. [46] reported essentially the identical features using the regioregular polythiophenes. They found that both in solids and solutions those polymers display dramatic thermochromism. The

'

01 200

1

1

400

600

-

-.__

1 ,

800

Wavelength (nm)

M) in dichloromethane at various temperatures. Figure 5. Absorption spectra of PHT (-4.0 x The polymer concentration is referred to the thiophene ring unit. Note that PHTs in Figs 4 and 5 have different molecular weights (see the text).

54

2 Electronic Properties of Polythiophenes

absorption bands grow with decreasing temperatures in the wavelength region covering approx. 520-600 nm, in good accordance with Fig. 5. Interestingly, moreover, Leclerc et al. [45] demonstrated a similar chromism upon addition of inorganic salts such as sodium chloride. They also studied the thermochromism using regiorandom polythiophenes [45]. They reported, however, that those regiorandom polymers did not show as dramatic chromism as the regioregular ones display, but exhibited a monotonous blueshift in the absorption maximum with increasing temperatures. The manifestation of the ordered phase in the solution is remarkable considering that dissolution of the materials in general brings about the disorder. This obviously explains how important role the energy stabilization due to the n-delocalization plays. At the same time, a sharp transition noted in Fig. 5 is contrasted with the spectra observed in the solid phase [5] where the transition is broadened. Efforts of researchers have made a variety of soluble polythiophenes available up to this date. Various researches relevant to the chromisms have been carried out accordingly. To study the effects of side groups upon the conformation of the polythiophene backbone can be counted among them [47]. It is easily understood by intuition that if, for example, the straight-chain alkyl groups are chosen, they produce essentially the same packing scheme and conformation in the backbones with lateral separations between them changed according to the alkyl-chain length. This very likely results in related spectral features of the materials. In fact, Fig. 6a [38] for polythiophene and PATS clearly demonstrates that their electronic spectra closely resemble one another, even though slight shifts in the absorption edge (around 650 nm) and in the weak shoulder just above the edge (around 600 nm) are observed along with the shifts in the absorption maxima (460-500nm). These spectra are compared with Fig. 6b [47] which shows the spectra of the soluble polythiophenes including a copolymer and a polymer having branched side groups. In Fig. 6b, the absorption edge is significantly broadened, even though the spectral profiles for these polymers are also related as a whole. Concomitantly, the location of the absorption maxima is significantly shifted and ranges from approx. 420 to 505 nm (yellow to red). These variations can be understood in terms of variation in the steric hindrance produced by the side groups. In particular, isobutyl groups introduce the coil conformation to the backbone, owing to the large steric hindrance resulting from a bulky and branched (disordered) structure. This conformation is responsible for a blueshift (approx. 30 nm) in the peak position of poly(3-isobutylthiophene), PiBT, relative to an analog, poly(3-butylthiophene). A concomitant decrease in the shoulder at 600 nm is also noticed for PiBT. Thus the spectroscopic feature of PiBT in the solid is virtually identical to that in the solution; compare its solid-state spectrum with Fig. 7 [42]that shows a typical spectrum of the soluble polythiophenes taken in good solvent (e.g. chloroform, toluene, etc.). In other words the segmentation and localization of the wave function have been essentially achieved in the solid state. In relation to the side-group associated chromism, Hotta [47] also showed that the copolymerization yields a specific effect relevant to a change in the backbone conformation. The aforementioned chromisms obviously indicate that no matter what processes and driving forces are responsible, the conformational change or transition of the n-conjugated backbone is the major origin of the dramatic color change. In t h s

2.2 Structure and conformation of polythiophenes

55

respect it is worth while comparing the thermochromism of the polythiophenes with that of polyfurans. Nishioka and Yoshino [48] reported that the thermochromism of the latter compounds is less clear than the former. Wang et al. [49]sought this origin from increments of two-center energy related to the a-carbons on the adjacent rings against the ring distortion. That is, on the basis of the quantum chemical

-k

I

I

I

I

I

(4

m

z -

-

WAVELENGTH (nm) Figure 6. (a) Absorption spectra of polythiophene and its derivatives substituted with normal-alkyl groups on the 3-position: polythiophene (------),poly(3-methylthiophene) (- --), poly(3-butylthiophene) (- - -), and PHT (). The materials were directly formed into thin films by the electrochemical synthesis (see section 2.3.1). Reprinted with permission from Ref. 38. Copyright 1987American Chemical Scoiety. (b) Absorption spectra of the solution-cast films of various soluble

polythiophenes: poly(3-isobutylthiophene) (-), poly(3-butylthiophene) (- - -), PHT -( 1, and a copolymer, poly(thiophene-co-3-hexylthiophene)[3-HT: thiophene = 3 : I] (- - -). Reprinted from Synth. M e t . 22(2), 103-13, 1987 (Ref. 47), S . Hotta, Electrochemical Synthesis and Spectroscopic Study of Poly (3-alkylthienylenes), with kind permission from Elsevier Science S . A.

56

2 Electronic Properties of Polythiophenes I

I

I

WAVELENGTH (nm)

Figure 7. Absorption spectrum for PHT in chloroform (approx. 2.3 x lop4M) at room temperature. Reprinted with permission from Ref. 42. Copyright 1987 John Wiley & Sons, Inc.

calculations the polyfurans were found to exhibit larger energy increments than the polythiophenes. Naturally, the polymer system with shallow potential against the ring distortion causes more dramatic chromism than that having deep potential. Using a regioregular polythiophene, Bouman and Meijer [50] observed a mirror image circular dichroic spectra for the samples that underwent different thermal treatments. They referred this observation to stereomutation associated with main-chain chirality. This might be related to the existence of the (super)helical structure [311 mentioned in the previous section. In a highly ordered polymer, a very sharp peak arises near the absorption edge and this peak is often referred to the excitonic absorption [39c]. This feature makes the conducting polymers stand out from the low molecular-weight compounds [51] and is most prominently represented by polydiacetylenes. This is partly because the polydiacetylenes can be obtained as single crystals [52]. Recent investigation, however, has demonstrated that this outstanding feature is not limited to the polydiacetylenes alone. Yoshimura et al. [53] showed that some polyazomethine compounds, made by evaporation polymerization, exhibit similar spectroscopic characteristics. Carefully recorded temperature-dependent spectra of polysilanes also show a similar sharp spike near the absorption edge [41]. Electroabsorption spectroscopy [54] provides a powerful tool for determining the origin of the absorption bands. The manifestation of the quadratic field dependence (Stark effect [55]) of modulated absorption intensity is thought to be a sign that the optical transition has the excitonic origin. Since the modulated intensity is expected to reflect the polarizability of the polymer backbone, spatial extent of the excitons can be evaluated. As an example, Bassler et al. [56] estimated the spatial extent of the excitons to be at most a few thiophene rings in a polythiophene derivative, PDDT. In a more ordered polymer system, furthermore, deviation from the quadratic field dependence can be observed. This deviation accompanied by the peak broadening in the modulated spectra is often related to the Franz-Keldysh effect [57] where the

2.3 Electronic processes of polythiophenes

51

coherence length of band states plays a role. Analyzing this behavior on polydiaFetylene, Horvath et al. [58] estimated the coherence length of 7r states at 200-400 A. At the same time, they estimated an associated effective mass of a free particle to be only 0.05nz0,where mois an electron mass. It should be emphasized that such large coherence length associated with the very small effective mass comes out of isolated polymer strands embedded in a single crystal monomer matrix [58]. Disorder effects, in turn, must be taken into account. Binh et al. [59] estimated from the line shape analysis the width of the 0-0 band associated with the excitonic transition. As a result, they found the band broadening with increasing temperatures, which they attributed to broadened excitonic density of states (DOS). Binh et al. [59] noted, furthermore, that the thermochromic spectral change is accompanied by this band broadening. They also related the width of the excitonic DOS to a characteristic energy parameter for the charged localized states [59] that represents the disorder of the system. Thus they hinted that the electronic processes both in the neutral and charged states result from the localized states of the same origin. Although it is of fundamental importance to understand how charges are accommodated when they are injected into the polymer chains, this question has not yet been addressed or investigated adequately. Recently, however, e Silva et al. [60] carried out theoretical investigations to approach this issue. They introduced a novel soliton conception based upon an equation of motion that explicitly takes account of the torsional degrees of freedom between adjacent thiophene rings. In that model the soliton is thought to consist of a ‘phase-wall’ between two phases in the chain; in other words before the soliton the even sites are twisted in one sense and the odd sites in the opposite sense, and after the soliton the torsion of even and odd rings is interchanged [60]. Using this conception e Silva et al. [60] stated that when the chain is twisted the excess charge concentrates on the soliton. In the planar configuration, on the other hand, the excess charge is supposed to spread out over the chain. This seems natural by intuition also, since the ring distortion (i.e. the defect creation) gives rise to energy instability [49, 601 and, hence, the excess charge is injected more easily into the defect than into the extended part of the chain. In this context, an ‘induced-rigidity’ concept is noteworthy. It is suggested that in the vicinity of a bipolaron on a doped polymer chain in dilute solution the straight-chain conformation is restored [6 11. Moreover, doping is expected to induce changes in the local chain conformation through creating the quinoidlike regions within the bipolaron that are more rigid against inter-ring rotations [28] (see Fig. 1). Further experimental and theoretical approaches toward clarifying this issue will be highly desired.

2.3 Electronic processes of polythiophenes 2.3.1 Charge excitations in polythiophenes The charge excitations in the conducting polymers are quite different from those of usual inorganic semiconductors such as crystalline silicon, even though the

58

2 Electronic Properties of Polythiophenes

conducting polymers exhibit the band structure at the ground state, similar to the semiconductors. The charges generated in the conducting polymers are strongly coupled with distortions of the polymer chain, which results in formation of solitons, polarons, or bipolarons. These peculiar charge excitations arise from the (quasi)one-dimensionality of the conducting polymers [62]. Historically, the theories were first developed by Su et al. [63] using polyacetylene (either trans or cis) on the basis of the tight-binding formalism with the 0-bond compressibility model. This was followed by Br6das et al. [64] who endeavored to make the theories applicable to polymers with the nondegenerate ground state typified by poly(p-phenylene), polypyrrole, and polythiophene. The success of the formalism developed by Su et al. [63] and Bredas et a/. [64] encouraged researchers to further develop the theory by explicitly including various physical quantities in the model Hamiltonian. As an example, effects of interchain coupling [65] and the influence of the electron-lattice coupling on the site energies [66] were taken into account. The phase-wall model by e Silva et al. [60] was developed along similar lines. The charge excitations can be introduced in the polymers through various methods of doping and the resulting polymers are rendered highly conducting. At a very low doping level, a charged polaron is energetically stable in polythiophene. That is, the charge stays at one domain-wall between the B and Q phases and a radical, i.e., an unpaired electron at the other domain-wall (see Fig. 1). Hence the charged polaron carries both the charge and the spin. In the band structure, the polaron yields two energy levels in the band gap, and in the case of a positive polaron the lower energy level is singly occupied with the higher one left empty (Fig. 8 [67]). At higher doping levels, the total amount of the Q phase on a polymer chain increases with the increasing number of polarons. What ensues from this situation is that two polarons tend to fuse together into a bipolaron to reduce the Q phase energetically unfavorable, overcoming the Coulombic repulsive force between two charges [64, 68-72]. This is particularly the case when the dopant ions screen the Coulomb repulsion between the charges on the polymer chain. As a result, the bipolaron in which the two charges with the same sign are located on each domain-wall

CONDUCTION

I

I

CONDUCTION BAND

fiwp

-nu,

Figure 8. Band diagram showing the gap states and allowed transitions for (a) a self-localized bipolaron and (b) a self-localized polaron. Both the species are charged positive (see the text). Reprinted with permission from Ref. 67. Copyright 1987 The American Physical Society.

2.3 Electronic processes of polythiophenes

59

(Fig. 1) is expected to be the most stable excitation for the nondegenerate conducting polymers [62, 7 1, 721. Similar to the polaron, the bipolaron produces two states in the band gap and, for example, a positive bipolaron has two empty levels (Fig. 8). Thus, the bipolaron holds zero spin and does not exhibit para-magnetism, unlike the polaron. At the highest doping levels of the nondegenerate conducting polymers? theory predicts that bipolaron bands are formed within the band gap as a consequence of the growth of the bipolaron levels [64, 731. When a pair of charges with unlike signs is introduced by e.g. photoexcitation, these charges are also expected to be located at the domain-walls (polaron-exciton) [74]. All these charge excitations can be regarded as confined soliton pairs [lo]. Strength of the confinement of the soliton pair depends partly upon strength of the Coulomb interaction between the charges on the polymer chain. This will be influenced by the presence or absence of dopant ions which screen that Coulomb interaction. Meanwhile? the charge excitations couple with the lattice vibrations and allow some symmetrical vibrational modes (Raman-active modes) to become infrared active by breaking the local symmetry [75]. This was first recognized in the doping and photoexcitation studies of trans-polyacetylene [62, 76, 771. Moreover? the amplitude mode formalism developed by Horovitz [78] has been successful in explaining the one-to-one correspondence between the photoinduced and the doping-induced IR-active vibrational modes and their relationship to the Raman modes of the pristine polymer. The theoretical models thus developed explain experimental results appreciably satisfactorily. The development of the theories, the other way around, stimulated researchers to examine their validity experimentally, particularly through various spectroscopic methods. Since the energy levels induced within the band gap by formation of the polaron or bipolaron can be probed and determined from electronic spectra of the charged species, this method has been used most versatilely. The optical transitions are schematically depicted in Fig. 8, where the positive polaron shows three optical transitions, hwl and hw2 from the valence band to midgap levels and hw3 between them. The positive bipolaron, on the other hand, exhibits only two transitions, hwl and hw2.These transitions satisfy the following sum rules assuming the electron-hole symmetry [77]: hwl

+ hw2 = Eg,

where Eg (= hwI in Fig. 8) is the band gap, and hwz - hW1 = hWj.

(1)

(2) Equation (2) applies only to the case of the polaron. Consequently, to examine whether actually observed results satisfy those relationships would be an intuitive criterion for guaranteeing the validity of the theory. The details on this issue will be presented in section 2.3.4.1 along with the results of other spectroscopic studies including IR and electron spin resonance (ESR). Prior to the recently developed theories on the conducting polymers, Holstein [79] described how a single excess charge (polaron) behaves in a one-dimensional molecular crystal. Note that approaches of the recent theories are somewhat related

60

2 Electronic Properties of Polythiophenes

to his in that extension of the polaron wave function varies depending on the electron-lattice coupling strength. Later, Emin [80] extended Holstein’s theory to a deformable continuum. Using a model of a coupled electron-continuum system bearing a charge, Emin [80] showed in a generalized manner that the following relationship holds (independent of spatial extent and dimensionality of the continuum): =Cl~f~)I2/S

(3) where A(r) is the distortion at a given position r, C the electron-continuum coupling constant, $(r) the polaron wave function, S a strain constant which reflects stiffness of the continuum. This relationship probably supplies us with a reliable basis for thinking of materials of a smaller spatial extent; in the case of a (quasi)one-dimensional system such materials may be represented by conducting oligomers like oligothiophenes. In fact, recent results of their X-ray experiments and MO calculations seem to be in good agreement with Eq. (3) [Sl-831. The inherently localized nature [66] of the charge excitations in the conducting polymers would permit us to assume similar excitations in the conducting oligomers as well [84]. The doping methods include both chemical and electrochemical techniques. For both the cases the polymers exhibit dramatic color change on account of occurrence of the subgap peaks and the shift of oscillator strength from the T-T* band to the subgap modes. In the polythiophene case, the color change between magenta and cyan can usually be observed. The chemical doping can be done by making the conducting polymers be in close contact with chemical species of dopants either in liquid phase or in vapor phase [85]. Various doping levels can be achieved by suitably varying the contact time or changing the dopant concentration in the liquid media [86]. If the polymers are soluble, the doping can be readily carried out in a homogeneous liquid phase [47,61,87]. The electrochemical methods, in turn, are based upon standard electrochemical processes [88]. The polymer materials (films, pressed pellets, etc.) adhering to or in close contact with the working electrode can be doped with various ionic species [89] of a supporting electrolyte at desired doping levels by changing the potentials of that electrode. Electrochromic devices were proposed on the basis of the color change due to reversible electrochemical doping and undoping [90]. The polymers such as the polythiophenes and polypyrroles can be electrochemically synthesized directly in the conducting form, this method being regarded as an option of a wide variety of the electrochemical doping techniques. Usually the p-type doping predominantly occurs in the case of polythiophenes. Occasionally, however, does the n-type doping take place for them [91]. A pair of charges both positive (hole) and negative (electron) can be introduced at once through photoexcitation [92]. Other special techniques of doping are presented in subsequent sections in connection with the properties of the charged states.

2.3.2 Charge transport in polythiophenes To understand how charges are transported is of fundamental importance in the materials science along with studying structures and morphologies of the materials.

2.3 Electronic processes of polythiophenes

61

Emin [12] pointed out that in the one-dimensional system the disorder encourages localization of the charge in cooperation with the electron-lattice interaction. This should be true of the conducting polymers with severe disorder. In such a case the charge transport may be described analogous to hopping transport in classical amorphous or non-crystalline media [93]. In this section we deal with several aspects of the charge transport in the polythiophenes. The charge injection from electrodes to the materials should also be properly dealt with in relation to the charge transport studies. The influence of the disorder is also mentioned in comparison with another class of organic semiconductors with weaker disorder, i.e. oligothiophenes. Recently, using the regioregular polythiophenes Yoon et al. [94] extensively studied the charge transport in these materials in the insulating regime where the carrier conduction is characterized as the hopping transport [93]. They classified the materials as follows according as the materials are 'close' to or 'far' from the metal-insulator (M-I) boundary [95]: If the materials are very close to the M-I boundary, their resistivity follows the temperature dependence characteristic of Mott's variable-range-hopping (VRH) conduction [93]. The materials relatively close to the M-I boundary exhibit a crossover from the Mott to Efros-Shklovskii hopping mechanism [96]. In the materials away from the M-I boundary, however, the extensive disorder and formation of inhomogeneous 'metallic islands' mask the above-mentioned hopping processes. Yoon et al. [94] observed similar charge transport features for polypyrrole and polyaniline as well. We add, however, that special care must be taken in determining from the experimental data which transport mechanism is likely. This is because resistivity ( p ) often exhibits exponential dependence on temperature T: i.e. p o( exp[(T/To)-"], on the basis of which the transport mechanism is inferred. The exponents x= l/2, 1/3, and 1/4 frequently arise in the non-crystalline or disordered media of semiconductors. When x = 1/3 happens, for instance, it is usually difficult to distinguish various conduction mechanisms including two-dimensional VRH, tunneling of carriers, interchain conduction, etc. [97]. Polythiophene and its derivatives are characterized as p-type semiconductors, and so they are expected to form a Schottky-like junction with a metal having low work function [98]. In this case the charge injection over this metal/polymer interface plays an important role. The charge injection mechanism can be studied by measuring temperature and voltage dependence of current. For instance, such measurements were carried out by Braun et al. [99] and Garten et al. [loo] independently using POT. In both the cases the said metal/polymer interface is responsible for the rectifying behavior [loll. Braun et al. [99] inferred from the results obtained on calcium/POT/ITO configurations (where I T 0 stands for indium/tin oxide) that the charge injection takes place via thermal fluctuation induced tunneling through a parabolic barrier. Garten et al. [IOO], on the other hand, concluded on the basis of the data obtained from the aluminum/POT/ITO devices that in the 'forward' mode of operation (where I T 0 is positive) the charges are injected through thermionic emission [ 1021 over the Schottky-like contact formed between POT and aluminum. They estimated a Schottky barrier height to be approx. 0.7eV, in good agreement with that evaluated by Turut and Koleli

62

2 Electronic Properties of Polythiophenes

[103]. Moreover, Garten et al. [loo] pointed out that virtually the same electroluminescent spectra can be observed from both the forward mode of operation and the ‘reverse’mode (where IT0 is negative). They attributed the latter mode of operation to direct tunneling of carriers into the transport bands. On the basis of those observations, they further concluded that the light emission results from the decay of the same kind of excited state in both the modes of operation [loo]. The interaction between the polythiophene and metals was recently investigated using quantum chemical calculations [104]. Such approaches will be important to improve the electrical contact between the polymer materials and electrodes. To study the charge transport in the conducting polymers, interchain mechanism and intrachain one must be distinguished. In this context, for instance, pressuredependent transport studies pursued by Ahlskog et al. [ 1051 are noteworthy. They observed increased mobility and conductivity with increasing pressures, which they referred to increased overlap between localized states. Such studies might give a clue to separating the two processes. This point will be dealt with in section 2.3.4.2 in further detail. From a point of view of how the disorder affects the transport phenomena, it is worth while comparing the charge transport characteristics in the polythiophenes with those in oligothiophenes and also comparing morphological features of these two classes of materials. Waragai et al. [83] studied the charge transport in thin films of the oligothiophenes with various molecular weight at varying temperatures on the field-effect transistor (FET) [ 1021 configurations. As a result, they admitted that the simple Arrhenius relationship holds for the oligothiophenes better than Mott’s VRH scheme. Relating the transport results to the optical data, they concluded that the (molecular) polarons play a role, even though the disorder effects obviously participate [83]. We note, on the other hand, that Binh et al. [59] concluded that the disorder effects overwhelm the polaron effects in the charge transport in PDDT. X-ray diffraction measurements, in turn, show that the thin films of the oligothiophene compounds consist of regular molecular layered structure [ 106,1071. Regarding an a,cu’-dimethylsubstituted quaterthiophene, for instance, Fig, 9 [ 1061indicates a very sharp and intense primary diffraction spacing (of 18.1 A) together with

0

28 Angle (degrees)

Figure 9. 8-20 profile for an a,cu’-dimethyl substituted quaterthiophene thin film in neutral form. Reprinted with permission from Ref. 106. Copyright 1991 The Royal Society of Chemistry.

2.3 Electronic processes of polythiophenes

63

higher-order reflections up to the thirteenth order. These features can be contrasted with the corresponding X-ray diffraction patterns of the PATS as shown in Fig. 2. Comparing these diagrams, we judge that the disorder in the oligothiophenes is weaker than that of the polythiophenes. This difference in the extent of disorder may well lead to the difference in the charge transport mechanism. A pioneering work by Koezuka and co-workers [lo81 revealed a mobility approximating 10-5 cm2/Vs for polythiophene, which was measured on metal-oxide-semiconductor (MOS) [lo21 devices. This mobility is comparable to that of organic molecules embedded in host polymer matrices [lo91 and smaller by two or three orders of magnitude than that of the oligothiophenes [83, 1101. Further thorough investigation will be strongly needed to clarify this difference in mobility level and to understand how the disorder influences the charge transport in the conducting polymers and related materials.

2.3.3 Carrier recombination: photoluminescence and electroluminescence Photoluminescence and electroluminescence result from carrier recombination. Whereas an electron-hole pair generated by photoexcitation causes the photoluminescence, the electroluminescencetakes place through the recombination of an electron and a hole injected from electrodes. Therefore, the origin of the two processes is expected to be essentially the same and this has been confirmed by related spectra observed from those processes [l 1I]. The notable discovery of electroluminescence in poly(p-phenylenevinylene) [ 1121 has accelerated the researches of both the photoluminescence and electroluminescence of the conducting polymers. Although transpolyacetylene exhibits very weak photoluminescence, some conducting polymers with the nondegenerate ground state show photoluminescence with high efficiency. This relates to the existence of their electroluminescence[112]. Examples of the conducting polymers that are potentially useful as the luminescent materials include poly(p-phenylenevinylenes) [l 11-1 191, poly(p-phenylenes) [ 120,1211, and polythiophenes [99, 100, 122-1331. Usually orange to red emission is observed for the polythiophenes [122]. The photoluminescence takes place via radiative recombination of photoexcited charges, which might otherwise decay through a nonradiative channel. When photoexcited electron-hole pairs are produced on a single chain of the nondegenerate conducting polymers, they form a neutral bipolaron with both the lower and higher levels within the gap singly occupied with a hole and an electron, respectively [62] (for the band structure see Fig. 8). Alternatively, this state may be envisaged as a polaronexciton with an electron and a hole located at each end of Q phase (see Fig. 1) [74]. These species can be (quasi)stabilized partly because a confinement potential based on the nondegeneracy of the ground state prevents the charge dissociation [ 1341. This kind of exciton can undergo a rapid radiative decay and cause fast luminescence because of the confinement potential [ 1351. This idea was confirmed by the fast luminescence (less than 9 ps) observed in the nondegenerate conducting polymers such as

64

2 Electronic Properties of Polythiophenes

cis-polyacetylene [1361 and polythiophene [ 1371. A series of peaks on the low energy side are observable in the photoluminescence spectra for polythiophene [ 138,1391and poly(pphenyleneviny1ene) [92]. Vardeny et al. [1381 suggested that these additional peaks are related to the slow recombination process resulting from intrachain excitons bound to spin defects. Meanwhile, McKenzie and Wilkins [140] explained the occurrence of those peaks by a model based on the multiphonon emission accompanying the electronic transition from the excited state to the ground one. The interchain coupling may stabilize photoexcited electron-hole pairs on different polymer chains. Distortion around the electrons and holes promotes formation of the negative and positive polarons on the polymer chains, respectively. These charged polarons are subsequently transformed into charged bipolarons. In this case, the charges decay nonradiatively without luminescence [92, 1411. Those polarons [141, 1421 and bipolarons [lo, 1421 were found to be generated in polythiophene and poly(3-methylthiophene), evidenced by their photoinduced absorption spectra. Another nonradiative charge recombination is caused by the interaction between excitons and injected charges. Such charges may be produced either by electrochemical doping [141,143] or by light irradiation [144];these charges can be introduced on an FET device configuration as well [145].The photoluminescence quenching is probably associated with the fact that charged polarons and/or bipolarons produced by the charge injection act as a trap for exciton, i.e. a quenching center [143, 1451. Since the discovery of electroluminescence [112], various kinds of polymer light emitting diodes (PLED) have been proposed and developed so that various colors from blue to (infra)red can be visualized. Some of these colors would be mixed to yield white light. By virtue of versatility of the polymer materials, it is quite possible that these devices will develop into large-area flat panel displays in the near future. In these devices, the conducting polymer films can simply be sandwiched between high and low work function metals which act as hole and electron injectors, respectively. The former can be chosen from among e.g. I T 0 and gold; the latter from among aluminum, calcium, etc. The applied electric field moves the injected hole and electron in the opposite direction to form a singlet or triplet exciton, of which the singlet one can decay radiatively with luminescence [I 121. A marked enhancement of electroluminescence efficiency was achieved for poly(pphenyleneviny1ene) by inserting a layer allocated both the electron-transporting and hole-blocking functions between the conducting polymer and the negative electrode [113]. Since the hole transport is blocked, holes are confined in close vicinity to the boundary between the said layer inserted and the conducting polymer film, and hence unfavorable effects such as exciton migration are inhibited. The resulting electroluminescence efficiency was reported to be ten times as high as that of devices without the blocking layer and to reach up to about 1% [113]. Fabrication of the heterostructures will be further worth pursuing along with approaches based on organic superlattices [146]. The progress in the PLEDs stimulated researchers into further efforts to apply a variety of potentially electroluminescent materials including the polythiophenes to this promising class of electronic devices [130, 131, 1331. As mentioned in section 2.2.2, the thermochromic behavior of PATS is a consequence of the main-chain conformational change. The coplanarity of the thiophene rings and deviation from

2.3 Electronic processes of polythiophenes

65

this conformation are both responsible. In a similar way, colors from PLED are controllable by using polythiophene derivatives with various side groups such as alkyl, cycloalkyl, and alkylphenyl [ 1291. These substituents change the main-chain conformation by the steric hindrance, leading to a large variation in the absorption maxima (305 to 594 nm) [129]. Similarly, electroluminescences ranging from blue, green, orange, and even down to near-infrared were observed with quantum efficiencies 0.01 to 0.6% [129]. The applications of the soluble polythiophene derivatives to the PLED are advantageous in that their polymer blends or composites can be easily formed [147]. Berggren et al. [ 1261have recently reported that these polymer blends (e.g. a mixture of different kinds of polythiophene derivatives) exhibit voltage controlled colors in electroluminescence. This effect was explained by the assumption that a number of nano-PLEDs of 50-200 nm in size yielded by micro phase separation are coupled parallel and operate in a specific voltage range depending upon the polymer species [126, 130, 1311. In addition, these nano-PLEDs dispersed in an insulator matrix such as poly(methy1 methacrylate) display white light emission with quantum efficiency of 0.4-0.6% [133].

2.3.4 Spectroscopic studies of the charged states We have described in the previous sections how the charges are injected and transported in the conducting polymers. It is also as important as this to understand how the charges injected are relaxed and stored in these polymers. Meanwhile, physical measurements of the conducting polymers in solution also give important information about the charged states. In this section we study these subjects through various spectroscopic means and their combination. These involve unique techniques of electrochemical voltage spectroscopy [ 1481 and frequency-domain electric birefringence spectroscopy [149]. 2.3.4.1 Charge storage configurations in solids and their anisotropic properties

In this subsection, we deal with the charge storage configurations on the polythiophenes in the solids and their anisotropic features. These investigations also serve the purpose of examining whether the charged states share the nature of the charge excitations predicted by theories. A direct way to this end is to measure the magnetic susceptibility of polymers during the course of the doping and undoping [67,88, 150-1531. Most of the experimental results showed that, as dopant concentration increases, the magnetic susceptibility is also increased in a light doping region, exhibits a maximum, and then decreases in a higher doping region. This behavior would be explained by that polarons generated and increased first are subsequently transformed into spinless bipolarons with increasing dopant concentration f62, 64, 68-72, 150-1521. Colaneri et al. [67] conducted ESR measurements in situ in carefully constructed electrochemical cells. Figure 10 [67] shows the magnetic susceptibility of poly(3methylthiophene), PMT, as a function of cell voltage. The reversible change of

66

2 Electronic Properties of Polythiophenes 3.0 I

2.5

I

I I

h

2 2.0

8

\ $

$

1.5

Q

5 u

1.0

'

0.5

.

o

o

~

o

o

o

0 .o

2.35

2.65

2.95

3.25 Voltage ws.

3.55

3.85

4.15

Li

Figure 10. Magnetic susceptibility of poly(3-methylthiophene), PMT, as a function of the cell voltage. Reprinted with permission from Ref. 67. Copyright 1987 The American Physical Society.

the susceptibility with varying cell voltages is confirmed by repeatedly cycled measurements. Around 2.5V vs Li a weak ESR signal is observed corresponding to an integrated magnetic susceptibility of -2 x lop6emu/mol. Above 3.4V vs Li, the susceptibility begins to drop pretty suddenly to a value less than emu/mol. This voltage corresponds to a doping level of a few mol% per thiophene ring [67]. Important information on the charge injection can be obtained from measurements of the electrochemicalvoltage spectroscopy (EVS) whose results are displayed in Fig. 11 [67]. This technique of EVS [148] involves slowly stepwise incrementing the voltage of the electrochemical cell and recording the charge removed from (or injected into) the polymer after each voltage step. The cell is displaced from equilibrium by a small potential step and the current through the cell is monitored via an ammeter. When the current falls below a designated value (sufficiently small to assure quasi equilibrium), the current is integrated, yielding the charge AQ that flowed on the decreasing (or increasing) cell voltage from Voto Vo A V [148b]. Figure 11 demonstrates that the charge injection threshold is located around 3 V vs Li. A somewhat diffused threshold would reflect molecular weight distribution and disorder of the material. The EVS results indicate that most of the charge transfer to the polymer occurs in the range of electrochemical voltages in which the ESR data set an upper limit of the magnetic susceptibility emulmol. This value can be translated into 8 x lop5 spins per thiophene ring. Note, for example, that as shown in Fig. 1 1 , a cell voltage of 3.6V corresponds to an injected charge of about 200 mC; i.e. approximately 10 mol% doping. Combining this with the

+

2.3 Electronic processes of polythiophenes

67

3.7

3.4

3

2 a, Po

3.1

2

9

2.8

2.5

2.2

0

50

100

150 Charge (mC)

200

250

300

Figure 11. Relationship between the cell voltage and charge for a 2mg sample of PMT. Reprinted with permission from Ref. 67. Copyright 1987 The American Physical Society.

number of spins determined at the same cell voltage leads to a spin-to-charge ratio of Ns/NchN 8 x lop4,where N , is the number of spins in the polymer resulting from Nch injected charges. This small value directly and unambiguously demonstrates that the charge is stored predominantly in the spinless species (bipolarons). emu/mol observed at 2.7 V vs Li corThe maximum susceptibility -2 x responds to about 0.2mol% spin, i.e. about one spin per 500 thiophene rings, or about one spin per 2000 carbons that participate in 7r-conjugated system along the polythiophene backbone (a value comparable to that found in trans-polyacetylene [154]). At this voltage the charge injected is limited to about one charge per 500 thiophene rings, estimated from the EVS diagram (Fig. 11). Combining this with the above ESR data implies that an entity carrying a spin is the dominant charged species in the above voltage (2.7V vs Li). Since the mean molecular weight of the PMT synthesized electrochemically is expected to be equivalent to about 300 thiophene rings [38], the maximum susceptibility is likely to be translated into the presence of at most a spin per macromolecular chain. Correspondingly the EVS results imply the presence of at most a positive charge per chain as well. The susceptibility data combined with the EVS results thus seem to indicate that at most one polaronic defect is allowed to exist in each polymer chain at a very low doping level. Although the above-mentioned charged species carrying a spin may well be associated with a polaron, Colaneri et al. [67]referred the origin of these charged species to defect states in the energy gap localized by disorder. Furthermore, they remarked that the nature of these species is pretty different from that of the self-localized polarons characteristic of a charge added to a polymer chain without the disorder.

2 Electronic Properties of Polythiophenes

s.9v us

3.5v vs

3.0V vs L 2.5v vs Li

0.0

0.5

i

1.0

d

J

1.5

2.0

2.5

3.0

3.5

4.0

Energy (eV)

Figure 12. Absorption spectra taken as the PMT film (approx. 0 . 2 p n thickness) was electrochemically reduced from amaximum doping level to the neutral state (for calibration of the cell voltages, see Fig. 11). Reprinted with permission from Ref. 67. Copyright 1987 The American Physical Society.

Thus in actual polymer systems represented by PMT, the dominant charge storage configurations are believed to be bipolarons, ill-defined charged states somewhat resembling polarons being involved at very dilute doping regimes. The sudden change of the susceptibility near 3.4V vs Li (corresponding to a doping level of a few mol% per thiophene ring) is apparently related to a bend noted in the relationship between conductivity and dopant concentration for the same material [37]; this bend also occurs at a doping level of a few mol% per ring [89]. This is suggestive of the transition associated with the formation of bipolaron lattices in PMT [ 1551. Further information on the charged states is supplied by the UV-vis. spectroscopy, this method being widely exploited by many researchers [156,1571. This is based upon analyzing energy levels of the subgap states introduced by the charge injection. As an example, Fig. 12 [67] shows four absorption spectra taken as the polymer was electrochemically reduced from maximum doping to the neutral state. The spectrum of the neutral polymer (at 2.5 V vs Li) shows an interband transition peaking at 2.4 eV. The onset of absorption (i.e. the three-dimensional energy gap) is relatively sharp at about 1.9 eV. At the relatively higher doping levels (above about 3 V vs Li), Colaneri et al. [67] observed two well-defined subgap absorptions with maxima at Awl 0.65 eV and hwz 1.6 eV (see Fig. 12). By inclusion of the electron-electron repulsion associated with the double charge on the bipolaron, and the binding energy of the charged bipolaron to the counterions (2EB)[lo, 1521, Eq. (1) will be modified as follows: N

N

fiW1

+ AW2 = Eid - 2(uB - EB),

(4)

2.3 Electronic processes of polythiophenes

69

where Eid is the one-dimensional energy gap and 2UB is the difference in Coulomb energy between the initial state (double charge) and the final state (single charge). Ef could be estimated from the spectra to be about 2.2eV for PMT [lo, 671. Substituting this value and the above subgap transition energies into Eq. (4), one finds UB M EB.The fact that the hw, band is a real peak is demonstrated in Fig. 13 [158] that depicts a spectrum of a related polymer PHT doped with PF; ions. To investigate the nature of the charge excitations of the conducting polymers, choice of the charge injection techniques is of great importance. From the point of view of quantitative accuracy, the charge injection on metal-insulator-semiconductor (MIS) devices [74, 1021 is one of the most desirable means along with EVS. Since this method is based on charging of a capacitor, any accurate amount of charge can be injected by suitably regulating an applied voltage. Furthermore, Ziemelis et al. [74] emphasized additional advantages of this method as follows: The conditions are very different from those achieved with chemical doping, for which the role of the counterion is considered to be important; it is likely to stabilize the bipolaron by providing screening of the two like charges, and also through the strain energy associated with separation of the polymer chains to accommodate these ions. They also noted that photoinduced absorption experiments detect long-lived excitations, and that bipolarons, being less mobile and hence longer lived than polarons, are preferentially detected. Making the best use of this technique based on the MIS device, Ziemelis et al. [74] measured field-induced optical spectra of PHT. As a result, they detected three field-induced bands at <0.45, 2.16, and 1.8eV along with two other bands at -0.5 and 1.18 eV. They attributed the former three modes to the polarons and the latter two modes to the bipolarons.

Photon energy (eV) Figure 13. Change in the absorption coefficient of PHT after being doped with 1 mol% of PF; ions. The inset shows chemical structure of PHT. Adapted from Ref. 158. Copyright 1987 The American Physical Society.

70

2 Electronic Properties of Polythiophenes

It will be interesting to compare on the same material PHT the results of Ziemelis et al. [74] with those of Kim et al. [158],the latter being obtained by the conventional method of chemical doping. Since in the former case the counterions (dopants) are absent in the system, EBin Eq. (4) should be zero. Here we make a premise that in Eq. (4) EF of PHT is identical to that of PMT, viz. 2.2 eV [67]. This may be justified by close resemblance between their electronic spectra, in particular between positions of the absorption edges [38] (Fig. 6a). Substituting into Eq. (4)the experimental values borrowed from the literature, i.e. hwl = 0.5eV and hw2 = 1.18eV from Ref. 74 and EB = 0, UB could be estimated to be -0.25eV. This value is in good agreement with that given by Vardeny et al. for polythiophene [lo], probably representing the Coulomb correlation associated with the optical processes of the bipolaron in the polythiophene systems. Meanwhile, substituting hwl = 0.45 eV and Flu2= 1.65eV of Ref. 158 into Eq. (4), we find UB "N EB again within an error of -0.05 eV. Notice moreover that both UB and E B are small relative to the size of the energy gap [10,152]. The energy difference between hwl and fiw2 may be the measure of the degeneracy-lifting [62] and be regarded as identical to 2wo in the formalism of Fesser et al. [69]. Thus 2w0/Eldis expected to stand for extent of the confinement of the soliton pairs (bipolaron) [69, 741. By comparing the two different sets of spectral data, we judge that the

d

L

i

/ I I\

1

400 0

I

,

800.0

,

,

I

1

1200 0

1

induced

,

I

I

I

1600.0

I

,

2000 0

Frequency (em-l) Figure 14. Detailed photoinduced (top), 4% PF, doped (middle), and pristine (bottom) absorption spectra of PMT. The bottom spectrum has been multiplied by four times for easy comparison. Adapted with permission from Ref. 75. Copyright 1987 The American Physical Society.

2.3 Electronic processes of polythiophenes

71

extent of the confinement is stronger in the PHT involving counterions [158] than in the PHT without them [74]. This seems quite reasonable considering the screening of the two like charges of the bipolarons by the counterions [74]. The IR analysis of the charged polymers will provide another powerful tool of studying the nature of the charge excitations of the polymers (refer to section 2.3.1). Earlier IR studies demonstrated that a doped polythiophene exhibits four intense vibrational peaks at approx. 1030, 1120, 1200, and 1330cm-' together with additional weaker peaks ranging approx. 640 to 840 cm-' [75, 159, 1601. These IRactive vibrational (IRAV) modes are considered to reflect the presence of the quinoidal rings associated with formation of polarons or bipolarons [62, 157, 1611. As already indicated in Fig. 13, these IRAV modes are clearly resolved in the midIR region, almost joining the tail of the lower-energy mode of the two subgap transitions due to the bipolarons. More detailed structure of the IRAV modes is depicted in Fig. 14 [75] for PMT. The one-to-one correspondence between those modes induced either through doping or by photoexcitation is quite evident. Notice that because of the complete suppression of the background, much cleaner and sharper IRAV modes are found in the photoinduced absorption spectrum [75] (Fig. 14). Once the charge storage configurations are characterized as described above, it will then be of major importance to study their anisotropic properties. This is because enhancement in the conductivity caused by chain alignment is a general

)I(

4.0

3.0

RI nI

Es

2.0

1.6

1.2

0.8

0.4

WAVE NUMBER ( ~ 1 0 ~ c r n ' ~ )

Figure 15. Dichroic IR spectra of the drawn film of neutral PHT. Ell ( E l ) denotes the electric field vector of a probe ray whose vibration plane is polarized parallel (perpendicular) to the drawn direction. The drawn ratio of the film was five. The bands at 3058 and 1512cm-' are polarized along the drawn direction, while the band at 824cm-' is polarized perpendicular to that direction. Since the bands due to the hexyl group (at 2960-2860cmp') exhibit nearly the same intensity with weak dichroism in the two spectra, they are illustrated for short in the EL spectrum. Adapted with permission from Ref. 164. Copyright 1989 American Chemical Society.

72

2 Electronic Properties of Polythiophenes

feature of the conducting polymers [ 162, 1631. Hotta ei al. [ 1641carried out dichroic IR measurements using aligned films of polythiophenes to address these issues. Among these polymers, in particular, PHT exhibits good mechanical properties and drawability that result from mitigation of the strong interchain backbone interaction on account of introducing the hexyl side groups [164]. The dichroic spectra of the five-time drawn film of neutral PHT are compared in Fig. 15 [164]. Detailed assignments of the absorption bands can be seen in Ref. 164. The bands observed at 3058,1512, and 824cm-' show conspicuous dichroism. The first two bands are highly polarized along the drawn direction, while the last band is polarized perpendicular to that direction. The bands at 1512 and 3058 cm-' are due to v(ring) mode with bl-symmetry species and the aromatic v(CH) mode, respectively; that at 824 cm-' is attributed to the y(CH) mode. Implication of these observation is that the PHT chains are highly aligned along the drawn direction 11641. The dichroic spectra of polythiophene are well related to those of PHT, indicating that the presence of the aligned structure comprising the polythiophene chains is again evident. Figure 16 [164] shows dichroic IR spectra of the PHT films (drawn ratio of five) partially-oxidizedwith iodine. As can be seen in the spectra, five bands (marked with asterisk in Fig. 16) are raised or intensified by the doping around 1320, 1200, 1160, 1080, and 970 cm-' and strongly polarized along the drawn direction. Of these, the bands at 1320, 1160, and 970 cm-' would be related to the translational modes of the bipolarons as in the case of PMT [75]. The two other modes appear to be associated with the modes intrinsic to the pristine polymer [1641. Correspondingly, a

I-

f2, >.

a a a

Lia 4

W 0

z a m

a 0 cn m a

Figure 16. Dichroic IR spectra of the drawn film of partially-oxidizedPHT (the drawn ratio: five). M I*. The bands The PHT film was doped with iodine in an acetonitrile solution containing marked with asterisks are doping-induced or intensified modes. Reprinted with permission from Ref. 164. Copyright 1989 American Chemical Society.

2.3 Electronic processes of polythiophenes

73

broadband around 4000 cm-' exhibits strong dichroism, polarized along the drawn direction as well. This mode is assigned to the bipolarons, analogous to the case of PMT [67]. Thus the electronic processes associated with the charged states (mostly bipolarons) are characterized as fundamentally one-dimensional along the polymer chains. This conclusion is fully consistent with enhanced conductivity along the drawn direction in the partially-oxidized films [ 162, 1631. In particular, the five-time drawn film of PHT exhibited a conductivity as high as 200Scm-' [44, 164, 1651; this conductivity was roughly an order of magnitude higher than that of an undrawn film (-30 S cm-') [44, 164, 1651. 2.3.4.2 Properties in solutions PATs with relatively long side groups are soluble in common organic solvents with both their neutral and doped forms [38, 166-1681. Therefore, their conformations and various properties such as optical absorption and spin susceptibility can be investigated in solutions and compared with those in the solid state [38, 42, 61, 87, 88, 90, 167-1691. As a result, it turns out that the major electronic features of PATs in the solid state are also reproduced in the solution [61, 871. For instance, the ESR results for PHT in chloroform imply that the bipolarons are increasingly dominant with increasing doping levels, with polarons always present as secondary species [6I]. Furthermore, optical absorption spectra exhibit chromic behavior due to doping [61,87]. At the same time, the solution studies reveal interesting aspects of the conducting polymers different from those in the solid state. As mentioned in section 2.2.1, the conducting polymers have complex and inhomogeneous morphologies such as a fibrillar structure in the solid state (bulk or film), which strongly affects the transport property of electronic carriers. Specifically, at least three kinds of the carrier transport processes, i.e. the intrachain (intramolecular), interchain (inter-molecular), and inter-fibrillar modes, can be responsible for the electric conductivity of the conducting polymers (Fig. 17a). Nonetheless, it is difficult to distinguish individual contributions by analyzing macroscopic conductivities in the solid state. In particular, the intrachain mode which reflects onedimensionality of the polymer chain is necessarily smeared by the two other slow limiting processes in the electric conduction. If the conducting polymer is dilutely dissolved in a solvent and an average separation among the polymer chains is larger than their hydrodynamic diameters [13], the interchain mode is completely inhibited, but the intrachain one still contributes to the electric transport (Fig. 17b). As a result, we can exclusively obtain information on the single-chain phenomena, specifically on the intrachain transport process of the electronic carriers. It is to be noted that the electric polarizability should be measured instead of the conductivity to detect the intrachain mode based on carrier transport restricted within a contour length of a polymer chain. Other related advantages of the solution studies lie in the following aspects: (i) Depending on the solvent species and temperatures, the conformation of the polymer chains of PATs can be tuned widely, that is, from the rod (ordered) to the coil (disordered). Doping is also responsible for the conformational change.

74

2 Electronic Properties of Polythiophenes

Figure 17. Schematic diagram of the carrier transport processes in (a) the solid state and (b) the solution. The intra-molecular, inter-molecular, and inter-fibrillar processes can be all responsible for the electric conductivity. Amongst these processes, only the intra-molecular one contributes to the electric polarizability in a dilute solution of the conducting polymers where an average separation among the polymer chains is larger than their hydrodynamic diameters. The inter-molecular process, on the other hand, is completely inhibited in the dilute solution.

This effect can be seen e.g. for a nitrobenzene solution of poly(3-butylthiophene) [28]. In a chloroform solution of PHT where the PHT chains assume a coiled conformation [42,61], the polaron wave function is expected to be localized on account of the disorder and the electron-lattice interaction [12] both of which act cooperatively. In fact, Nowak et al. [61] inferred from analysis of the ESR line with hyperfine splitting that the polaron wave function is localized primarily on a single thiophene ring with only a small spin density on the neighboring rings on either side. They

2.3 Electronic processes of polythiophenes

75

mentioned that this wave function appears to be somewhat more tightly localized than expected from calculations based upon a straight-chain conformation for the neutral polymer [61, 1701. (ii) The relative stability between the polarons and bipolarons will be changed, depending on the solvent species also and the polymer concentration [61, 871. For instance, a polar solvent such as dichloromethane inhibits the transformation of polarons into bipolarons at a low doping level, in sharp contrast with the case of the chloroform solution [61, 871. Nowak et al. [87] associated this finding with the fact that the strength of the counterion screening is altered by varying the solvent. Having these circumstances as a background, we describe in this subsection how the intrachain carrier transport can be dealt with by the frequency-domain electric birefringence spectroscopy (FEBS) [ 1491. The FEBS technique utilizes the electric and optical anisotropies of molecules dissolved in the solution. When the external electric field is applied to a solution containing molecules with the anisotropic polarizability, the molecules tend to be oriented in the direction of the applied electric field. If moreover the molecules have the optical anisotropy, the solution exhibits the birefringence, which is called the Kerr effect [171] as a nonlinear optical effect. In the FEBS, where the sinusoidal electric field is applied, the Kerr constants are measured by varying the frequencyf of the sinusoidal wave. In this sense, the FEBS closely resembles the dielectric relaxation spectroscopy [172]. Two kinds of dielectric responses due to the permanent and induced dipole moments are expected in the dilute solution of the conducting polymers. If carries move along a polymer chain even more slowly than the rotation of the chain, the inhomogeneous distribution of the carriers yields the permanent (or quasi-permanent) dipole moment on the polymer chain. Thus, the electric polarizability arises from the orientation of the permanent dipole moment towards the direction of the external field. On the other hand, if the carriers move much faster than the rotation, the external electric field induces the electric polarizability and exerts a different type of torque on the polymer chain. These two different responses can be clearly distinguished by FEBS [149]. There are two kinds of electric birefringence techniques, the FEBS and the transient electric birefringence (TEB) method [ 171, 1731. The TEB method was applied to the solution of polydiacetylene to investigate the rod-coil conformational transition of the polymer chains [174]. The FEBS, on the other hand, has the advantage of giving us the mobility of the carriers along the polymer chain separately from the hydrodynamic radius of the polymer chain (usually referred to as the polymer conformation) [149]. Shimomura et al. [175] have recently applied the FEBS technique to the solutions of dilutely doped PHT to study the intrachain conduction in the conducting polymer and its relation to the main-chain conformation. Figure 18 [175a] shows a typical example of the FEB spectra, where Kdc is the dc component in the Kerr response function K(w), and K L and Klu are the real and imaginary parts of the complex 2w component KL(=K;, - X;,), respectively. The relationship among these quantities is expressed as K(w) = lydC

+ Re[K;uexp(i2wt)].

(5)

76

2 Electronic Properties of Polythiophenes

4

2

I-Erelaxation 0

loo

iol

f, 10’

I o4

lo3

lo5

lo6

Frequency (Hz) Figure 18. FEB spectra taken at room temperature with the solution of dilutely doped PHT in dichloromethane with the polymer concentration cp of 0.01 wt%. Kdc represents the dc component and K L and K L are the real and imaginary parts of the 2w component, respectively. Thef, denotes the rotational relaxation frequency. The low-frequency (LF), the middle-frequency (MF), and the high-frequency (HF) relaxation modes are observed. Adapted with permission from Ref. 175a. Copyright 1994. The American Physical Society.

Here, the angular frequency w = 2 r f is of the sinusoidal electric field applied to the solution. According to the theoretical analysis [149b], the relaxation time T’ of &c is given as follows by a harmonic mean of the rotational relaxation time 7; and the transport relaxation time rt of carriers along the polymer chain: =(

r p

+

(6) If the carrier transport is much faster than the rotation of the polymer chain, i.e., T~<< T ~r’ , approaches to T ~Then, . we can evaluate the intrachain mobility of the carrier directly from the K d c spectrum. The 2w component K . is given by (T’)-’

(Tt)-l

K. = +*(1 + i2wTr/3)-’

(7) where +* is a response function of which the real part is equal to Kdc, that is, KdC = Re(+*). Therefore, the rotational relaxation time is obtained from the K& spectra by the analysis combined with the K d c spectrum. Since rr is proportional to the cube of the hydrodynamic radius of the polymer chain [176], we can detect by the FEBS a conformational change of the chain with h g h accuracy and sensitivity. The K d c spectrum in Fig. 18 shows three relaxations, the low-frequency (LF) one with a relaxation frequency of approx. 40 Hz, the middle- frequency (MF) one with a frequency of 1 kHz, and the high-frequency (HF) one with no dispersion in the frequency range below 100kHz. Since the LF relaxation frequency is nearly equal to the rotational relaxation frequency fr in the K. spectra, the LF relaxation is due to the rotation of the quasi-permanent dipole moment. On the other hand, both

2.3 Electronic processes of polythiophenes

77

the MF and HF relaxations are attributed to the intrachain carrier transports, since these relaxation frequencies are much higher than f,. In fact, the frequency dispersion of the H F relaxation is observed in a higher frequency range by the dielectric relaxation spectroscopy as shown in Fig. 19 [175a]. Consequently, it turns out that there exist slow and fast modes in the intrachain conduction of the conducting polymer (vide infru). The FEB spectra of dilutely doped PHT have been measured in solutions with different kinds of solvents [175a] or by varying temperatures [175b] or by changing contour lengths of the polymer chain [175c]. Figure 20 [175a] indicates the dependence of the effective length LeK,i.e. the hydrodynamic radius, on the fraction of (a)

1.4-

1.2 -

L

7.0

-

I

I

1

o5

o6

1

I

10’

lo8

109

1O’O

lo9

10’O

Frequency (Hz)

1o5

1o6

10’

los

Frequency (Hz)

Figure 19. Dielectric relaxation spectra taken at room temperature with the solution of the dilutely doped PHT in dichloromethane with cp = 0.13 wt%. The ordinates (a) E’ and (b) E” are the real and imaginary parts of the complex dielectric constant E * , respectively. Reprinted with permission from Ref. 175a. Copyright 1994. The American Physical Society.

18

2 Electronic Properties of Polythiophenes 1.4 4

I

I

I

I

I

I

4

1.2 -

1 .o

3

It

I

0.8

0

ilo

0.6 -

-

0.4

0.2

70

65

80 85 90 Fraction of dichloromethane (%)

75

95

1

Figure 20. Effective (hydrodynamic) length Leffof PHT with cp 0.01 wt% in mixtures of dichloromethane and toluene at room temperature. Adapted with permission from Ref. 175a. Copyright 1994 The American Physical Society.

dichloromethane in solutions mixed with toluene. Figure 21 [43] depicts the temperature dependence of L,r. One can see crossover behaviors of the main-chain conformation from an extended form to a coiled one, that is, the rod-coil transition. Note that as previously shown in Figs 4 and 5 , the solution exhibits concomitant thermochromism associated with the conformational transition. Figure 22 [43] indicates the dependence of the M F relaxation time .rM on L,r. In this case where pure dichloromethane is used as a solvent, the polymer chain has 1.2 -

a a

1.0 -

0

a

3

0

v

4% 0.8

-

0.

0 0 0

-

0.6

230

240

250

260

270

Temperature (K) Figure 21. Temperature dependence of Leffof PHT with cp = 0.002 wt% in an equivolume mixture of dichloromethane and toluene.

2.3 Electronic processes of polythiophenes

0.4

0.6

0.8

1.o

79

1.2

L,,, (Pm)

Figure 22. Transport relaxation time TMofthe MF mode as a function of Leffof PHT in dichloromethane with cp = 0.004wtY0.

an extended form, and hence L,K can be regarded as the contour length L, of the polymer chain. Since the intrachain transport time in the M F relaxation increases with the chain length, the M F relaxation arises from an intrachain transport of carriers within the range up to L,. The diffusion constant of the M F relaxation was evaluated to be 1 x 10-9m2s-1[175a]. The analysis of the dielectric increment and the relaxation time in the dielectric relaxation spectra indicates that the H F relaxation is ascribable to a much more localized carrier transport along the polymer chain [175a]. The diffusion constant DH of the HF relaxation was about 2 x 10-7m2s-' with the transport range LH of 50nm. This LH corresponds to approx. 100 thiophene rings in length, which may represent the effective size of the 7r-conjugation. Note that the value of DH is of the same order as the diffusion constant of neutral solitons in polyacetylene determined by NMR [177] and ESR measurements [178] or that of polarons in polyaniline estimated by ESR [179]. Thus, the diffusion constant DMof the M F relaxation turned out to be one or two orders of magnitude smaller than DH.This strongly suggests that defects in the conducting polymer chain hinder the carrier transport encompassing the long range up to the contour length. It should be emphasized here that both DH and DM obtained above are most likely relevant to the polarons [87]. In the dilute solution of conducting polymers where the interchain separation among polymer chains is larger than their hydrodynamic radii, the conformational change should be driven by a single-chain mechanism [28,42, 1741. One of possible mechanisms is that the interaction between the conjugated electrons of the polymer backbone and the surrounding polarizable molecules of solvent stabilizes an extended conformation of the conducting polymer chain [180]. In this model, the chain conformation is determined by a balance between the interaction energy and the entropy of the polymer chain, similar to the helix-coil transition [181]. As temperature increases, the polymer conformation changes from an extended form to a coiled

80

2 Electronic Properties of Polythiophenes

one which is entropically favored. This results in the generation of a number of defects or disorder on the chain that breaks up the 7r-conjugation. The extended form at low temperature has a smaller number of defects than the coiled one. This permits us to envisage this extended form as a ‘broken-rod’ in which a relatively smaller number of defects are interspersed. These defects or disorder should be responsible for the existence of the two intrachain transport modes as mentioned above. This is one of the prominent disorder effects in the one-dimensional intrachain process, unlike the interchain one. At higher polymer concentration, an interchain interaction yields aggregation [42] or the lamellar structure similar to the comblike one in the solid state [182]. If the electronic processes are studied in the semidilute solution, the interchain carrier transport in the solution is expected to be detected and compared with that in the solid state (refer to section 2.3.2). Moreover, the intrachain carrier transport in heavily doped polymers in the dilute solution is also of great interest. The change in the electronic state or in the Coulombic repulsion between carriers may well affect the intrachain carrier transport as well as the conformation of the polymer chain [28].

2.4 Concluding remarks and future outlook We have studied the electronic properties of the polythiophenes in relation to their structural and conformational characteristics. Their electronic properties come out of the 7r-conjugated system extended along the polythiophene backbone. Since the electronic state is strongly coupled to the backbone conformation, a variety of electronic properties show themselves depending on the conformations of polymer chains. The structural variation can be measured as against a perfect crystal lattice where the polythiophene backbones satisfy the translational symmetry along the chain direction and form the herringbone structure spreading two-dimensionally perpendicular to that direction. In this respect polythiophene derivatives, especially PATS, provide a variety of peculiar structures. Examples include a bundled structure in which the rodlike polymer chains form nearly the face-to-face configuration. Helical and superhelical structures can be noted as well. The polymers having chiral side chains even exhibit stereomutation associated with the main-chain chirality. The interplay of the electronic structure and the backbone conformation is observed as chromisms both in the solid state and solution. The reversible appearance and disappearance of the fine structure linked to the peak shift in the electronic spectra obviously demonstrate that the order-disorder transformation in the backbone conformation plays an important role. Charges can be injected in the conducting polymers through various methods of doping. Of those, chemical and electrochemical methods are widely used. If the doping is achieved on the electronic devices such as MIS configurations, an accurate amount of charge can be injected. Upon the charge injection, energy levels are invoked within the gap on account of the electron-lattice coupling. In polymers

2.4 Concluding remarks andftcture outlook

81

with a nondegenerate ground state, two defects created on the domain-walls interact with each other to form a confined soliton pair. The energy levels are determined by strength of the confinement of the soliton pair, its net charge (polaron, bipolaron, polaron-exciton, etc.), physical parameters of surrounding media (e.g. dielectric constant), and so on. Spectroscopic studies including ESR, EVS, and UV-Vis. enable the detailed characterization of the charge storage configurations and show that in most cases the bipolarons are the most dominant excitations in the polythiophenes. At the same time, anisotropic features of the charge excitations are quite evident. This has been confirmed by the dichroic IR measurements of the aligned films of the polythiophenes. Further interesting features of the conducting polymers manifest themselves in the carrier recombination processes such as photoluminescence and electroluminescence. In the conducting polymers with the nondegenerate ground state, a photoexcited electron-hole pair forms the polaron-exciton, which is readily confined within the Q phase. The confinement potential prevents the dissociation and migration of the electron-hole pair and yields the fast photoluminescence. Meanwhile the electroluminescent studies have led to the development of the polymer light emitting diodes that emit a variety of colors, from blue to (infra)red. These colors can be varied by tuning the backbone conformation. When we deal with the real polymer systems, the materials are necessarily suffered from disorder to a greater or lesser extent. In the (quasi)one-dimensional systems, in particular, the disorder encourages localization of the charges in concert with the electron-lattice interaction. As a consequence, the charge transport is best characterized as the hopping transport in disordered media, this situation allowing us to treat the transport as that in amorphous or non-crystalline media following a classical scheme. Recent topics about this have been summarized and discussed. In this context we added a few remarks by comparing the polythiophenes and oligothiophenes from a point of view of how the disorder influences the charge transport. The studies of the conducting polymers in solution offer a good opportunity to inspect the charge transport limited within a single isolated polymer chain. Exclusively the intrachaira electronic processes can be studied when the conducting polymers are dispersed dilutely in the solution. The unique technique of FEBS has verified the presence of two levels of intrachain transport processes, in which defects or disorder plays a specific role. In the history of research on conducting polymers, high conductivity was pursued during the first decade. During the second decade, these researches were followed by extensive survey on a variety of physical properties as semiconductors. The discovery of the electroluminescence from the conducting polymers is a typical example. Efforts toward structuring ordered materials have been made throughout these decades. The contents we have described in this chapter can be summarized as outcomes of the past two-decade researches carried out along these lines. These results clearly indicate the importance of the promising class of conducting polymers, polythiophenes. The aforementioned research trend will be reinforced in the coming third-decade of research. Of these, highly aligned structures including single-crystallinepolymers are

82

2 Electronic Properties of Polythiophenes

worth pursuing above all. Much improved properties (e.g. high conductivity, high mobility, high luminescent efficiency, etc.) will be expected from these ordered structures. In this sense, the observation of steeply arising absorption edges in the regioregular polythiophenes is notable. This might be an initial sign of the excitonic transition, which has been established only for single-crystalline polydiacetylenes so far. Once the efforts to make the materials of high quality are connected with keen interest in discovering novel phenomena, we will arrive at a more advanced stage in the researches of conducting polymers.

Acknowledgments We thank Professor T. Yamabe, Professor A. J. Heeger, Professor F. Wudl, and Professor R. Hayakawa for their continuous enlightening discussions and suggestions. Thanks are also due to Professor S. D. D. V. Rughooputh, Professor Y. H. Kim, Professor M. J. Winokur, Dr. M. J. Nowak, Dr. N. Colaneri, Dr. D. Spiegel, and Dr. T. Shimomura for their helpful discussions and earnest experimental collaboration.

References 1. C. Kittel, Introduction to Solid State Physics, 6th ed., John Wiley & Sons, New York, 1986, Chap. 2. 2. W. A. Harrison, Electronic Structure and the Properties of Solids, Dover Publications, New York, 1989, pp. 38-40. 3. S. Briickner, W. Porzio, Makromol. Chem. 1988, 189,961. 4. (a) A. Gavezzotti, G. Filippini, Synth. Met. 1991,40,257. (b) J. Bernstein, J. A. R. P. Sarma, A. Gavezzotti, Chem. Phys. Lett. 1990,174,361. (c) G. R. Desiraju, A. Gavezzotti, Actu Crystallogr. 1989, B4.5, 473. (d) W. Porzio, S. Destri, M. Mascherpa, S. Briickner, Actu Polym. 1993, 44, 266. 5. K. Tashiro, K. Ono, Y. Minagawa, M. Kobayashi, T. Kawai, K. Yoshino, J. Polym. Sci., Polym. Phys. Ed. 1991, 29, 1223. 6. M. J. Winokur, P. Wamsley, J. Moulton, P. Smith, A. J. Heeger, Macromolecules 1991, 24, 3812. 7. (a) R. D. McCullough and R. D. Lowe, J. Chem. Soc., Chem. Commun. 1992, 70. (b) R. D. McCullough, S. P. Williams, S. Tristram-Nagle, M. Jayaraman, P. C. Ewbank, L. Miller, Synth. Met. 1995, 69, 279. 8. A. J. Heeger, in Handbook of Conducting Polymers (Ed.: T. A. Skotheim), Marcel Dekker, New York, 1986, Chap. 21. 9. R. R. Chance, D. S. Boudreaux, J.-L. Bredas, R. Silbey, in Handbook of Conducting Polymers (Ed.: T. A. Skotheim), Marcel Dekker, New York, 1986, Chap. 24. 10. Z. Vardeny, E. Ehrenfreund, 0. Brafman, M. Nowak, H. Schaffer, A. J. Heeger, F. Wudl, Phys. Rev. Lett. 1986, 56, 671. 11. A. R. Bishop, D. K. Campbell, in Nonlinear Problems: Present and Future (Eds.: A. R. Bishop, D. K. Campbell, B. Nicolaenko), North-Holland, Amsterdam, 1982, pp. 195-208.

References

83

12. D. Emin, in Handbook of Conducting Polymers (Ed.: T. A. Skotheim), Marcel Dekker, New York, 1986, Chap. 26. 13. J. des Cloizeaux, G. Jannink, Polymers in Solution: Their Modeling and Structure, Clarendon Press, Oxford, 1990, p. 57. 14. G. Tourillon, F. Garnier, J. Polym. Sci., Polym. Phys. Ed. 1984, 22, 33. 15. D. Y. Zhang, T. L. Porter, Synth. Met. 1995, 74, 55. 16. J. P. Montheard, T. Pascal, G. Seytre, G. Steffan-Boiteux, A. Douillard, Synth. Met. 1984, 9, 389. 17. F. Garnier, G. Tourillon, J. Y. Barraud, H. Dexpert, J . Muter. Sci. 1985, 20, 2687. 18. Z . Mo, K.-B. Lee, Y. B. Moon, M. Kobayashi, A. J. Heeger, F. Wudl, Macromolecules 1985, 18, 1972. 19. (a) T. Yamamoto, A. Morita, Y. Miyazaki et al., Macromolecules 1992,25, 1214. (b) T. Yamamoto, T. Kanhara, C. Mori, Synth. Met. 1990, 38, 399. 20. M. J. Winokur, D. Spiegel, Y. Kim, S. Hotta, A. J. Heeger, Synth. Met. 1989, 28, C419. 21. T. J. Prosa, M. J. Winokur, J. Moulton, P. Smith, A. J. Heeger, Macromolecules 1992, 25, 4364; Synth. Met. 1993,55-57, 370. 22. G. Gustafsson, 0. Inganas, H. Osterholm, J. Laakso, Polymer 1991,32, 1574. 23. J. Mirdalen, E. J. Samuelsen, 0. R. Gautun, P. H. Carlsen, Solid State Commun. 1991, 77, 337; Solid State Commun. 1991, 80, 687; Synth. Met. 1992, 48, 363. 24. H. J. Fell, E. J. Samuelsen, J. Mirdalen, E. Bakken, P. H. Carlsen, Synth. Met. 1995,69, 301. 25. (a) K. Tashiro, K. Ono, Y. Minagawa, K. Kobayashi, T. Kawai, K. Yoshino, Synth. Met. 1991, 41-43, 571. (b) K. Tashiro, M. Kobayashi, S. Morita, T. Kawai, K. Yoshino, Synth. Met. 1993, 55-57, 321. 26. K. Tashiro, K. Kobayashi, S. Morita, T. Kawai, K. Yoshino, Synth. Met. 1995, 69, 397. 27. S. Chen, J. Ni, Macromolecules 1992,25, 6081. 28. J. P. Aime, F. Bargain, M. Schott, H. Eckhardt, G. G. Miller, R. L. Elsenbaumer, Phys. Rev. Lett. 1989, 62, 55. 29. T. L. Porter, S. Jeffers, G. Caple, B. L. Wheeler, S. Swift, Surf. Sci. Lett. 1990, 238, L433. 30. G. Caple, B. L. Wheeler, R. Swift, T. L. Porter, S. Jeffers, J. Phys. Chem. 1990,94, 5639. 31. R. Yang, D. F. Evans, L. Cristensen, W. A. Hendrickson, J. Phys. Chem. 1990, 94, 6117. 32. S. Z . Dong, Q. Cai, P. Liu, A. R. Zhu, Appl. Surf. Sci. 1992, 60-61, 342. 33. E. Lacaze, K. Uvdal, R. Bodo, J. Garbarz, W. R. Salaneck, M. Schott, J. Polym. Sci., Polym. Phys. Ed. 1993, 31, 111. 34. A. Bolognesi, M. Catellani, S. Destri, W. Porzio, Makromol. Chem., Rapid Commun. 1991, 12, 9. 35. C. X. Cui, M. Kertesz, Phys. Rev. B 1989, 40, 9661. 36. T. Kawai, M. Nakazono, K. Yoshino, J . Muter. Chem. 1992,2,903. 37. S. Hotta, in Handbook of Organic ConductiveMolecules and Polymers (Ed.: H. S. Nalwa), John Wiley & Sons, Chichester, 1997, Vol. 11, Chap. 8. 38. S. Hotta, S. D. D. V. Rughooputh, A. J. Heeger, F. Wudl, Macromolecules 1987, 20, 212. 39. (a) G. N. Patel, R. R. Chance, J. D. Witt, J. Polym. Sci., Polym. Lett. Ed. 1978, 16, 607. (b) R. R. Chance, G. N. Patel, J. D. Witt, J . Chem. Phys. 1979, 71, 206. (c) Y. Tokura, T. Mitani, T. Koda, Chem. Phys. Lett. 1980, 75, 324. 40. (a) L. A. Harrah, J. M. Zeigler, J . Polym. Sci., Polym. Lett. Ed. 1985,23,209. (b) P. Trefonas, J. R. Damewood, R. West, R. D. Miller, Organometallics 1985, 4, 1318. 41. S. S. Bukalov, M. V. Teplitsky, L. A. Leites, C.-H. Yuan, R. West, Mendeleev Commun. 1996, 135. 42. S. D. D. V. Rughooputh, S. Hotta, A. J. Heeger, F. Wudl, J. Polym. Sci., Polym. Phys. Ed. 1987,25, 1071. 43. T. Shimomura, Master Thesis, University of Tokyo, 1995, p. 97. 44. S. Hotta, M. Soga, N. Sonoda, Synth. Met. 1988, 26, 267. 45. M. Leclerc, M. FrCchette, J.-Y. Bergeron, M. Ranger, I. LCvesque, K. FaYd, Macromol. Chem. Phys. 1996,197, 2077. 46. C. A. Sandstedt, R. D. Rieke, C. J. Eckhardt, Chem. Mater. 1995, 7, 1057. 47. S. Hotta, Synth. Met. 1987, 22, 103. 48. Y. Nishioka, K. Yoshino, Jpn. J. Appl. Phys. Part 2 1990, 29, L675.

84

2 Electronic Properties of Polythiophenes

49. S. Wang, K. Yoshino, K. Tanaka, T. Yamabe, J. Phys. SOC.Jpn. 1991,60,2002. 50. M. M. Bouman, E. W. Meijer, Adv. Muter. 1995, 7, 385. 51. Y. Yamaguchi, M. Kimura, K. Hanabusa, H. Shirai, Polym. Prepr. Jpn. 1996,45, 1730. Some phthalocyanine compounds are known to show sharp spectral feature near the absorption edge. 52. (a) G. Wegner, Makromol. Chem. 1971, 145, 85; Makromol. Chem. 1972, 154, 35. (b) R. H. Baughman, J. Polym. Sci., Polym. Phys. Ed. 1974, 12, 1511. 53. T. Yoshimura, S. Tatsuura, W. Sotoyama, A. Matsuura, T. Hayano, Appl. Phys. Lett. 1992, 60, 268. 54. (a) S. D. Phillips, R. Worland, G. Yu et al., Phys. Rev. B 1989, 40, 9751. (b) A. HorvAth, H. Bassler, G. Weiser, Phys. Status Solidi B 1992, 173, 755. 55. L. I. Shiff, Quantum Mechanics, 3rd ed., McGraw-Hill, New York, 1968, p. 252. 56. H. Bassler, M. Gailberger, R. F. Mahrt, J. M. Oberski, G. Weiser, Synth. Met. 1992, 49-50, 341. 57. L. Sebastian. G. Weiser. Phvs. Rev. Lett. 1981, 46, 1156. 58. A. Horvath,’G. Weiser,’ C. iapersonne-Meyer, M. Schott, S. Spagnoli, Phys. Rev. B 1996, 53. 13 507. 59. N.‘T. Binh, L. Q. Minh, H. Bassler, Synth. Met. 1993,58, 39. 60. G. M. e Silva, A. N. de Brito, N. Correia, Phys. Rev. B 1996, 53, 7222. 61. M. J. Nowak, S. D. D. V. Rughooputh, S. Hotta, A. J. Heeger, Macromolecules 1987,20,965. 62. A. J. Heeger, S. Kivelson, J. R. Schrieffer, W.-P. Su, Rev. Mod. Phys. 1988,60, 781. 63. W. P. Su,J. R. Schrieffer, A. J. Heeger, Phys. Rev. B 1980,22,2099. 64. J. L. Bredas, B. Themans, J. G. Fripiat, J. M. Andri, R. R. Chance, Phys. Rev. B 1984, 29, 6761. 65. S.-J. Xie. L.-M. Mei. D. L. Lin. Phvs. Rev. B 1994., 50., 13364. 66. D. Giri, K. Kundu, Phys. Rev. b 1996,53, 4340. 67. N. Colaneri, M. Nowak, D. Spiegel, S. Hotta, A. J. Heeger, Phys. Rev. B 1987, 36, 7964. 68. D. K. Campbell, A. R. Bishop, Nucl. Phys. B 1982,200,297. 69. K. Fesser, A. R. Bishop, D. K. Campbell, Phys. Rev. B 1983, 27, 4804. 70. J.-L. Brkdas, R. R. Chance, R. Silbey, Phys. Rev. B 1982,26, 5843. 71. D. Bertho, C. Jouanin, Phys. Rev. B 1987, 35, 626. 72. D. Bertho, A. Laghdir, C. Jouanin, Phys. Rev. B 1988, 38, 12 531. 73. J.-L. Bredas, B. Thkmans, J. M. Andrk, Phys. Rev. B 1982, 26, 6000; Phys. Rev. B 1983, 27, 7827. 74. K. E. Ziemelis, A. T. Hussain, D. D. C. Bradley, R. H. Friend, J. Riihe, G. Wegner, Phys. Rev. Lett. 1991, 66, 2231. 75. Y. H. Kim, S. Hotta, A. J. Heeger, Phys. Rev. B 1987,36, 7486. 76. C. R. Fincher, Jr., M. Ozaki, A. J. Heeger, A. G. MacDiarmid, Phys. Rev. B 1979, 19,4140. 77. D. Baeriswyl, D. K. Campbell, S. Mazumdar, in Conjugated Conducting Polymers (Ed.: H. Kiess), Springer-Verlag,Berlin, 1992, Chap. 2. 78. B. Horovitz, Solid State Commun. 1982,41, 729. 79. T. Holstein, Ann. Phys. ( N . Y . ) 1959, 8, 325. 80. D. Emin, in Electronic and Structural Properties of Amorphous Semiconductors (Eds.: P. G. Le Comber, J. Mort), Academic, New York, 1973, Chap. 7. 81. S. Hotta, H. Kobayashi, Synth. Met. 1994, 66, 117. 82. K. Tanaka, Y. Matsuura, Y. Oshima, T. Yamabe, S. Hotta, Synfh. Met. 1994, 66, 295. 83. K. Waragai, H. Akimichi, S. Hotta, H. Kano, H. Sakaki, Phys. Rev. B 1995, 52, 1786. 84. (a) D. Fichou, G. Horowitz, B. Xu, F. Gamier, Synth. Met. 1990, 39, 243. (b) K. Waragai, S. Hotta, Synth. Met. 1991, 41, 519. 85. J. C. W. Chien, Polyacetylene: Chemistry, Physics and Material Science, Academic Press, Orlando, 1984, p. 326. 86. S. Hotta, T. Hosaka, M. Soga, W. Shimotsuma, Synth. Met. 1984, 9, 381. 87. M. J. Nowak, D. Spiegel, S. Hotta, A. J. Heeger, P. A. Pincus, Macromolecules 1989,22,2917; Synth. Met. 1989, 28, C399. 88. T.-C. Chung, J. H. Kaufman, A. J. Heeger, F. Wudl, Phys. Rev. B 1984,30, 702. 89. S. Hotta, W. Shimotsuma, M. Taketani, S. Kohiki, Synth. Met. 1985, 11, 139.

References

85

90. K. Kaneto, K. Yoshino, Y. Inuishi, Jpn. J . Appl. Phys. Part 2 1983, 22, L412. 91. M. Aizawa, S. Watanabe, H. Shinohara, H. Shirakawa, J. Chem. Soc., Chem. Commun. 1985, 264. 92. R. H. Friend, D. D. C. Bradley, P. Townsend, J . Phys. D 1987, 20, 1367. 93. N. F. Mott, E. A. Davis, Electronic Processes in Non-crystalline Materials, Clarendon Press, Oxford, 1979, Chaps 2 and 3. 94. C. 0. Yoon, M. Reghu, D. Moses et al., Synth. Met. 1995, 75, 229. 95. N. F. Mott, Metal-Insulator Transitions, 2nd edn, Taylor & Francis, London, 1990, pp. 50-55. 96. A. L. Efros, B. L. Shklovskii, J . Phys. D 1975, 8, L49. 97. I. Youm, M. Cadene, D. Laplaze, J. Muter. Sci. Lett. 1995, 14, 1712. 98. (a) S. Glenis, G. Horowitz, G. Tourillon, F. Garnier, Thin Solid Films 1984, I l l , 93. (b) M. Kaneko, A. Yamada, J . Polym. Sci., Polym. Lett. Ed. 1985, 23, 629. (c) S. Glenis, G. Tourillon, F. Garnier, Thin Solid Films 1986, 139, 221. 99. D. Braun, G. Gustafsson, D. McBranch, A. J. Heeger, J . Appl. Phys. 1992, 72, 564. 100. F. Garten, J. Vrijmoeth, A. R. Schlatmann, R. E. Gill, T. M. Klapwijk, G. Hadziioannou, Synth. Met. 1996, 76, 85. 101. D. M. de Leeuw, E. J. Lous, Synth. Met. 1994, 65,45. 102. S. M. Sze, Physics of Semiconductor Devices, 2nd ed., John Wiley & Sons, New York, 1981, Chaps 5, 7, and 8. 103. A. Turut, F. Koleli, J. Appl. Phys. 1992, 72, 818. 104. F. Elfeninat, C. Fredriksson, E. Sacher, A. Selmani, J . Chem. Phys. 1995, 102, 6153. 105. M. Ahlskog, J. Paloheimo, H. Stubb, A. Assadi, Synth. Met. 1994, 65, 77. 106. S. Hotta, K. Waragai, J. Muter. Chem. 1991, I , 835. 107. S. Hotta, K. Waragai, Adv. Muter. 1993, 5, 896. 108. A. Tsumura, H. Koezuka, T, Ando, Appl. Phys. Lett. 1986,49, 1210. 109. (a) H. Bassler, Phys. Status Solidi B 1993,175, 15. (b) L. B. Schein, Philos. Mag. B 1992,65,795. 110. (a) F. Garnier, F. Deloffre, G. Horowitz, R. Hajlaoui, Synth. Met. 1993, 55-57, 4747. (b) F. Garnier, G. Horowitz, X. Peng, D. Fichou, Adv. Muter. 1990, 2, 592. (c) K. Waragai, H. Akimichi, T. Inoshita, S. Hotta, H. Sakaki, in Proc. 51st Ann. Techn. Con$ Soc. Plastic Engineers, The Society of Plastics Engineers, Brookfield, 1993, pp. 233 1-4. 111. D. Braun, A. J. Heeger, Appl. Phys. Lett. 1991, 58, 1982. 112. J. H. Burroughes, D. D. C . Bradley, A. R. Brown et al., Nature 1990,347, 539. 113. A. R. Brown, D. D. C. Bradley, J. H. Burroughes et al., Appl. Phys. Lett. 1992, 61, 2793. 114. P. L. Burn, A. B. Holmes, A. Kraft, D. D. C. Bradley, A. R. Brown, R. H. Friend, J. Chem. Soc., Chem. Commun. 1992, 32. 115. N. C. Greenham, S. C. Moratti, D. D. C. Bradley, R. H. Friend, A. B. Holmes, Nature 1993, 365, 628. 116. A. Kraft, P. L. Burn, A. B. Holmes, D. D. C. Bradley, A. R. Brown, R. H. Friend, Synth. Met. 1993,55-57, 936. 117. C. Zhang, H. von Seggern, K. Pakbaz, B. Kraabel, H.-W. Schmidt, A. J. Heeger, Synth. Met. 1994, 62, 35. 118. C. Zhang, S. Hoger, K. Pakbaz, F. Wudl, A. J. Heeger, J. Electron. Muter. 1993, 22, 413. 119. C. Zhang, S. Hoger, K. Pakbaz, F. Wudl, A. J. Heeger, J . Electron. Muter. 1994, 23, 453. 120. G. Grem, G. Leditzky, B. Ullrich, G. Leising, Adv. Muter. 1992, 4, 36. 121. G. Grem, G. Leising, Synth. Met. 1993, 55-57, 4105. 122. (a) Y. Ohmori, M. Uchida, K. Muro, K. Yoshino, Solid State Commun. 1991, 80, 605. (b) T. Yamamoto, H. Wakayama, T. Fukuda, T. Kanbara, J. Phys. Chem. 1992, 96, 8677. 123. N. C. Greenham, A. R. Brown, D. D. C. Bradley, R. H. Friend, Synth. Met. 1993,55-57,4134. 124. G. G. Malliaras, J. K. Herrema, J. Wildeman, R. Wieringa, R. E. Gill, S. S. Lampoura, G. Hadziioannou, Adv. Muter. 1993, 5, 721. 125. R. E. Gill, G. G. Malliaras, J. Wildeman, G. Hadziioannou, A h . Muter. 1994, 6, 132. 126. M. Berggren, 0. Inganas, G. Gustafsson et al., Nature 1994, 372, 444. 127. M. Berggren, G. Gustafsson, 0. Inganas, M. R. Anderson, 0. Wennerstrom, T. Hjertberg, Appl. Phys. Lett. 1994, 65, 1489. 128. M. Berggren, G. Gustafsson, 0. Inganas, M. R. Anderson, T. Hjertberg, 0. Wennerstrom, J . Appl. Phys. 1994, 76, 7530.

86

2 Electronic Properties of Polythiophenes

129. M. R. Andersson, M. Berggren, 0. Inganas et al., Macromolecules 1995,28,7525. 130. 0 .Inganas, M. Berggren, M. R. Andersson et al., Synth. Met., 1995, 71, 2121. 131. M. Berggren, 0. Inganas, G. Gustafsson, M. R. Andersson, T. Hjertberg, 0. Wennerstrom, Synth. Met. 1995, 71,2185. 132. M. Granstrom, M. Berggren, 0. Inganas, Synth. Met. 1996, 76, 141. 133. M. Granstrom, 0. Inganas, Appl. Phys. Lett. 1996,68, 147. 134. S. A. Brazovskii, N. N. Kirova, JETP Lett. 1981, 33, 4. 135. L. Lauchlan, S. Etemad, T.-C. Chung, A. J. Heeger, A. G. MacDiarmid, Phys. Rev. B 1981, 24, 3701. 136. W. Hayes, C. N. Ironside, J. F. Ryan, R. P. Steele, R. A. Taylor, J. Phys. C 1983, 16, L729. 137. K. S. Wong, W. Hayes, T. Hattori et al., J. Phys. C 1985, 18, L843. 138. Z. Vardeny, E. Ehrenfreund, J. Shinar, F. Wudl, Phys. Rev. B 1987,35, 2498. 139. J. Poplawski, E. Ehrenfreund, S. Glenis, A. J. Frank, Synth. Met. 1989, 28, C335. 140. R. H. McKenzie, J. W. Wilkins, Synth. Met. 1991, 41-43, 3615. 141. K. Kaneto, F. Uesugi, K. Yoshmo, J. Phys. SOC.Jpn. 1987,56,3703. 142. Z. V. Vardeny, X. Wei, Synth. Met. 1993, 54, 99. 143. S. Hayashi, K. Kaneto, K. Yoshino, Solid State Commun. 1987,61,249. 144. D. D. C . Bradley, R. H. Friend, J . Phys.: Condens. Matter 1989, 1, 3671. 145. P. Dyreklev, 0. Inganas, J. Poloheimo, H. Stubb, Synth. Met. 1993, 54, 99. 146. (a) J. Kido, K. Hongawa, K. Okuyama, K. Nagai, Appl. Phys. Lett. 1993, 63, 2627. (b) K. Uchiyama, H. Akimichi, S. Hotta, H. Noge, H. Sasaki, in Materials Research Society Symposium Proceedings (Eds.: A. F. Garito, A. K.-Y. Jen, C. Y.-C. Lee, L. R. Dalton), Materials Research Society, Pittsburgh, 1994, 328, pp. 389-93. (c) H. Akimichi, T. Inoshita, S. Hotta, H. Noge, H. Sakaki, Appl. Phys. Lett. 1993, 63, 3158. (d) F. F. So, S. R. Forrest, Phys. Rev. Lett. 1991, 66, 2649. (e) A. Saxena, J. D. Gunton, Synth. Met. 1987, 20, 185. 147. S. Hotta, S. D. D. V. Rughooputh, A. J. Heeger, Synth. Met. 1987, 22, 79. 148. (a) A. H. Thompson, Physica (Utrecht) 1980, 99B, 100; Phys. Rev. Lett. 1978, 40, 1511. (b) J. H. Kaufman, T.-C. Chung, A. J. Heeger, J. Electrochem. SOC.1984, 131, 2847. (c) L. W. Shacklette, J. E. Toth, Phys. Rev. B, 1985, 32, 5892. 149. (a) N. Ookubo, Y. Mori, R. Hayakawa, Y. Wada, Jpn. J. Appl. Phys. 1981, 19, 2271. (b) N. Ookubo, Y. Hirai, K. Ito, R. Hayakawa, Macromolecules 1989, 22, 1359. 150. K. Kaneto, S. Hayashi, S. Ura, K. Yoshino, J. Phys. SOC.Jpn. 1985,54, 1146. 151. S. Hayashi, K. Kaneto, K. Yoshino, R. Matsushita, T. Matsuyama, J. Phys. SOC.Jpn. 1986, 55, 1971. 152. J. Chen, A. J. Heeger, Solid State Commun. 1986,58, 251. 153. M. Scharli,H. Kiess, G. Harbeke, W. Berhnger, K. W. Blazey, K. A. Miiller, Synth. Met. 1988, 22, 317. 154. B. R. Weinberg, E. Ehrenfreund, A. J. Heeger, A. G. MacDiarmid, J. Chem. Phys. 1980, 72, 4749. 155. (a) S. Stafstrom, J.-L. BrBdas, Phys. Rev. B 1988, 38, 4180. (b) M. Logdlund, R. Lazzaroni, S. Stafstrom, W. R. Salaneck, J.-L. Bredas, Phys. Rev. Lett. 1989, 63, 1841. 156. D. L. Greenaway, G. Harbeke, Optical Properties and Bandstructure of Semiconductors, Pergamon, Oxford, 1968. 157. H. Kiess, G. Harbeke, in Conjugated Conducting Polymers (Ed.: H. Kiess), Springer-Verlag, Berlin, 1992, Chap. 4. 158. Y. H. Kim, D. Spiegel, S. Hotta, A. J. Heeger, Phys. Rev. B 1988,38, 5490. 159. S. Hotta, W. Shimotsuma, M. Taketani, Synth. Met. 1984/1985, 10, 85. 160. W. Hayes, F. L. Pratt, K. S. Wong, K. Kaneto, K. Yoshino, J. Phys. C 1985, 18, L555. 161. Y. S. Lee, M. Kertesz, J. Chem. Phys. 1988, 88, 2609. 162. H. Kahlert, G. Leising, Mol. Cryst. Liq. Cryst. 1985, 117, 1. 163. A. G. MacDiamid, A. J. Heeger, Synth. Met. 1979/80, I , 101. 164. S. Hotta, M. Soga, N. Sonoda, J. Phys. Chem. 1989,93,4994. 165. S. Hotta, M. Soga, in Proc. MRS Intern. Mtg. Adv. Mater. (Eds.: M. Doyama, S. Somiya, R. P. H. Chang), Materials Research Society, Pittsburgh, 1989, 1, pp. 225-30. 166. R. L. Elsenbaumer, K. Y. Jen, R. Oboodi, Synth. Met. 1986,15, 169. 167. K. Kaeriyama, M. Sato, S. Tanaka, Synth. Met. 1987, 18, 233.

References

87

168. M. Sato, S. Tanaka, K. Kaeriyama, J. Chem. SOC.,Chem. Cornmun. 1986, 873. 169. (a) S. D. D. V. Rughooputh, M. J. Nowak, S. Hotta, A. J. Heeger, F. Wudl, Synth. Met. 1987, 21,41. (b) K. Yoshino, S. Nakajima, R. Sugimoto, Jpn. J . Appl. Phys. Part 2 1987,26, L1371. 170. J.-L. Brtdas, G. B. Street, Acc. Chem. Res. 1985, 18, 309. 171. C. T. O'Konski, S. Krause, in Molecular Electro-Optics (Ed.: C. T . OKonski), Marcel Dekker, New York, 1976, Part 1. 172. J. R. MacDonald, Impedance Spectroscopy, John Wiley & Sons, New York, 1987. 173. I. Teraoka, R. Hayakawa, J. Chem. Phys. 1989, 91,4920. 174. K. C. Lim, A. Kapitulnik, R. A. Zacher, A. J. Heeger, J. Chem. Phys. 1985, 82, 516. 175. (a) T. Shimomura, H. Sato, H. Furusawa, Y. Kimura, H. Okumoto, K. Ito, R. Hayakawa, S. Hotta, Phys. Rev. Lett. 1994,72,2073.(b) T. Shimomura, Y. Kimura, K. Ito, R. Hayakawa, S. Hotta, Synth. Met. 1995, 69, 689. (c) T. Shimomura, Y. Kimura, H. Okumoto, K. Ito, R. Hayakawa, Rep. Prog. Polym. Phys. Jpn. 1994, 37, 447. 176. M. Doi, S. F. Edwards, The Theory of Polymer Dynamics, Oxford University Press, Oxford, 1986. 177. N. Nechtshein, F. Devreux, F. Genoud, M. Guglielmi, K. Holczer, Phys. Rev. B 1983,27, 61. 178. K. Mizoguchi, K. Kume, H. Shirakawa, Solid State Commun. 1984, 50, 213. 179. K. Mizoguchi, N. Nechtshein, J. P. Travers, C. Menardo, Phys. Rev. Lett. 1989,63, 66. 180. K. S. Schweizer, J. Chem. Phys. 1986,85, 1156. 181. P. J. Flory, Statistical Mechanics of Chain Molecules, John Wiley & Sons, New York, 1969. 182. S. Yue, G . C. Berry, R. D. McCullough, Macromolecules 1996, 29, 933.

This Page Intentionally Left Blank

3 The Synthesis of Oligothiophenes Peter Bauerle

3.1 Introduction The discovery of highly conducting polyacetylene in 1977 [l] prompted the synthesis of other polymers with conjugated .ir-systemssuch as polypyrrole [2], poly-p-phenylene [3], poly-p-phenylenevinylene[4], polyphenylenesulfide [5, 61, polyaniline [7], as well as of other polyheterocycles such as polythiophene [8], polyfuran [S], polyselenophene [9], or more extended polyaromatics such as polyazulene [lo]. Because they can have a variety of structural variations, polythiophenes have become the type of conducting polymer most frequently investigated because conductivity is mostly unaffected by substituents [l 11. In addition, both conducting and semiconducting polythiophenes are very stable and readily characterized. Although the polymer properties can be influenced and tailored by the variation of monomeric building blocks to a certain extent, precise predictions about unequivocal structure/property relationships are not possible. The physical properties of such conducting polymers cannot be correlated directly to the structural parameters, which frequently depend on the preparation conditions. Due to statistical chain length distribution and interruption of the conjugated chain by mislinkages and other defects, these materials, like all polymers, lack a rigidly defined structural principle. Therefore, the conjugation and conducting pathways are interrupted and severely disturbed. Figure 1 depicts the ‘real’ structure of a polythiophene schematically. The synthesis and investigation of well-defined model oligomers has therefore recently become useful to gain insight into the structural and electronic pecularities of the corresponding polymers. In the meantime, for nearly all basic conducting polymers homologous series of defined oligomers have been synthesized. Especially oligothiophenes have reached more and more prominence in recent years [ 111. By assembling defined mono- and oligomers step by step, via well understood organic reaction sequences, materials are obtained where both chain and conjugation length are well controlled and rigorously defined [12]. Since oligomers may serve as model compounds for the respective polymers, the painstaking synthetic procedure appears well justified. Depending on their size and substitution pattern they are usually more soluble than polymers and are stable in various redox states. The precise characterization of the electronic and geometric structure succeeds both in solution and in the solid state. The physical properties are now well correlated to the (conjugated) chain length and thus ‘real’ structure/property relationships become available. The data may be compared to those of the corresponding polymers in order to estimate their mean conjugation length or to extrapolate to a (hypothetical) infinite chain length. This information is not accessible from investigations on the polymeric systems [13].

90

3 Oligothiophenes

Figure 1. Typical ‘real’ structure of polythiophene including mislinkages and defects in comparison with a defined and all-a-linked oligothiophene. Conjugation pathways are drawn with bold lines.

The study on defined oligomers also provides information which may be used to improve strategies for the development of novel materials. Some of their physical properties even surpass those of the polymers. For this reason, oligothiophenes have been recently advanced as components for molecular electronics [141 and optical devices [ 151. In fact, an all-a-linked sexithiophene has been successfully employed as an active component in an organic field-effect transistor [14] and in a light-modulating device [ 161. The mobility of the charge carriers and the transistor characteristics were found to be superior to those of an analogous poly(bithiophene) transistor and even approach those of transistors based on amorphous silicon. Prior to 1960, bi- and polythienyls and their derivatives were not easily attainable, and consequently only few investigations on these compounds had been made. It is intriguing to see that even in the 1930s a whole series of a-oligothiophenes or apolythienyls (as they were also called) up to a-septithiophene were laboriously isolated and well characterized, by Steinkopf and coworkers. They were interested to discover the difference between thiophenes and corresponding benzenes on the one hand, and with aliphatic conjugated chains, polyenes, on the other hand. Since both series were already known up to septi-p-phenylene [17] and dimethyldodecahexaene [18], respectively, Steinkopf et al. synthesized a whole series of aoligothiophenes 2-7 up to the heptamer, and expected thiophenes to be the missing link between olefins and benzene [19]. Ullmann’s biaryl synthesis was applied, and 2iodothiophene 8 was treated with copper bronze. The main product, a-bithiophene 2, could be isolated by steam distillation (<20% yield), and also after lengthy purification through extraction and vacuum sublimation, the whole series of aterthiophene 3 up to a-septithiophene 7 was obtained in very small amounts [Eq. (111 [19c, 201.

3.1 Introduction

91

Cu-bronze,

S

n 2-7 (n= 0-5)

8

By comparing melting points, absorption maxima, and the qualitative assessment of fluorescence and ‘halochromie’ (reaction with concentrated sulfuric acid) the various oligomers were characterized and, for the first time, their structures were qualitatively correlated with their properties. Remarkably, 60 years after the first report on a-oligothiophenes, these properties are of high current interest with respect to the application of conjugated materials in organic light emitting diodes [21]. After several a-oligothiophenes had been isolated in small amounts from yellow marigold [22], Gronowitz and coworkers started a new era of oligothiophenes in 1960. 2,2/-Bithiophene 2 and the isomeric 3,3/-derivative 11 thus became easily available by symmetric coupling of the corresponding lithiated thiophenes 9 and 10 with cupric chloride [Eq. (2)] [23]. This breakthrough gave rise to a vast amount of synthetic work [24]. In the following years, the main interest has been devoted to bi- and terthiophenes also because they were found to occur frequently as acetylenic derivatives in certain plants (family of Compositae and Asteraceae) [25]. Many of the a-conjugated oligothiophenes which have been isolated and synthesized exhibit biological activity: some of them generate skin pigmentation, act as herbicides or inhibit feed germination, while others are phototoxic against nematodes, algae, human erythrocytes, insect larvae and eggs [26]. This phototoxic activity rests on the ability of the thiophene oligomers to develop singlet oxygen when illuminated. This was proven by many mechanistic and photophysical studies [27].

2

9

Li

10

11

Finally, the third and current era of oligothiophenes was initiated in 1982 with the first synthesis of polythiophene P1 as an electrically conductive polymer [8]. Somewhat later, a-bithiophene 2 and a-terthiophene 3 were used as monomers, instead of thiophene 1 itself, as substrates for electropolymerization in order to control the structure of the polymer [Eq. (3)] [28]. Owing to their lower oxidation potential,

92

3 Oligothiophenes

electropolymerization can be performed under milder conditions [29]. Furthermore, since the thiophene rings are exclusively a-linked in the starting molecule, a polymer with less a,,L?-defectsis to be expected [12, 301. Nevertheless it was found in most studies, that electropolymerization of thiophene on the one hand, and of thiophene oligomers on the other hand, does not lead to the same polymer and that, contrary to what could be expected, the use of conjugated precursors yields less conjugated and less conducting material [30b, 311.

1-3 (n 1-3)

PI

Conversely, due to the defined structure of a-oligothiophenes it was quickly realized that they constitute interesting models of the electronic properties of polythophene [32]. The first spectral characterization of oligothiophene radical cations and dications in 1989 [32c, 331 and the use of a-sexithiophene 6 as active component in an all-organic electronic device one year later [34] have triggered a renaissance of interest and intensive research work in the synthesis and characterization of these materials that had been neglected for so many years [35]. Several types of nomenclature have been established for the oligothiophenes. Short notations like the original proposed by Bredt [36] (e.g. a,a,a-quaterthiophene) and later on by Steinkopf [19a] (e.g. a-quaterthiophene) were later followed by the general and more convenient abbreviations nT or a-nT (e.g. 4T or a-4T). However, only the much more complicated systematic nomenclature used by Chemical Abstracts allows the precise denotation of linkages between rings and positions of substituents. The numbering scheme of thiophene, of the three isomers of bithiophene, and of a substituted terthiophene is given in Fig. 2. Since for conjugated materials only oligothiophenes connected in a-positions are relevant, and a consistency with other oligomeric series should be given, in this chapter we will use a more precise but also convenient short notation which considers the repeating units according to polymer nomenclature and the endgroups: X-T,-Y (e.g. H-T,-H). The early synthesis and characterization of oligothiophenes entailed an immense number and variety of derivatives and related compounds. Due to their structural variability, stability, and well-balanced properties, it seems likely that oligothiophenes have developed to be ideal model compounds for the corresponding conjugated polymers. This chapter intends to give an overview of the synthesis of various classes of oligothiophenes. They will be arranged by the different methods

I

2

11

12

13

Figure 2. IUPAC-nomenclature of thiophene 1, the isomeric bithiophenes 2, 11, 12, and 3’-substituted terthiophene 13.

3.2 Synthesis of Oligothiophenes

93

of preparation, in contrast to the three extensive reviews which have been published up to now and cover the literature up to 1988 [37].

3.2 Synthesis of Oligothiophenes 3.2.1 Unsubstituted Oligothiophenes In the short time in which oligothiophenes have been investigated with respect to conjugated materials, several general categories have been established. Unsubstituted all-a-linked oligothiophenes (this section) and substituted derivatives (section 3.2.2) represent the most interesting derivatives. Regioisomeric oligothiophenes including P-linkages have a strongly decreased conjugation and are only discussed to a minor extent. The main interest of unsubstituted oligothiophenes lies in their solid-state properties and in their application as organic semiconductors in electronic devices. Oligothiophenes may be synthesized via two different routes, either by C-C-linkage reactions between thiophenes and/or oligothiophenes or by ring closure reactions from acyclic precursor molecules. One of the most useful procedures in the formation of C-C-a-bonds which is also effective in the synthesis of oligothiophenes is the metal-promoted coupling reaction of organic halides. The reactions of active organometallic reagents, such as organolithium or Grignard reagents with salts of transition metals, are described below. These reactions, however, have the disadvantage of requiring at least stoichiometric amounts of transition metals. More recently, not only very effective, but also purely catalytic transition metal-catalyzed methods have been developed which provide oligothiophenes of different chain lengths and of different connectivities. Mainly, low-valent transition metals and their complexes are important coupling agents for such processes. Representative examples are the already mentioned Ullmann reaction, and the Ni(0)- and Pd(0)-catalyzed cross coupling of organometallics with organic halides (see below). Ring closure reactions of acyclic precursor molecules like e.g. diacetylenes, 1,4diketones, or 1,Cdithiins to form thiophene rings are also widely used to synthesize various types of oligothiophenes. These reactions are described in section 3.2.1.3. 3.2.1.1 Arene/arene-Coupling Methods by Oxidative Couplings

Copper(Il)-promoted oxidative coupling The oxidative coupling of organolithum compounds with cupric chloride [38] was the first reaction that was found to be considerably more effective than the Ullmann analogous reactions utilized by Steinkopf [19] and Sease et al. [22] [Eq. (4)].In 1930 Steinkopf reported that the reaction of (2-thieny1)magnesium bromide with CuC12 resulted in 2,2’-bithiophene H-T2-H 2, in a 44% yield [39]. Lipkin could increase the yield up to 50% [40]. Gronowitz [41] and later

94

3 Oligothiophenes

Kauffmann [42] and Kagan [43], further improved the yields of oligothiophenes by using the corresponding organolithium compounds rather than the Grignard reagents. Lithiation of thiophenes can be effected either by simple deprotonation or by halogen-metal interconversion. Thus, the reaction of monothiophene H-TI -H 1 with n-butyl lithium (n-BuLi) followed by the addition of one equivalent of CuC12 afforded H-T2-H 2 in 54% yield [41, 421. Application of two equivalents of CuC12 even raised the yield to 85%, indicating that the effect of the salt is more than catalytic [42]. Copper salts also surpass the effect of other transition metals: the use of catalytic amounts of cobalt chloride either with Li-T1-H 9 [44] or the corresponding Grignard derivative [45] led only to 30% and 26% formation of H-T2-H 2, respectively. The oxidative coupling of organolithium compounds is not restricted to bithiophenes, but can also be applied to higher, even-numbered oligothiophenes up to a-sexithiophene H-T6-H 6. While the coupling of H-T2-H 2 to a-quaterthiophene H-T4-H 4 succeeds in 64% yield, the attempt of the repetitive coupling of the latter to result in a-octithiophene H-T8-H 14 failed [42]. Probably the reduced solubility of the polyaromatic precursors in etheral solvents made the dimerization of H-T8-H 14 to an a-sedecithiophene impossible [46]. Meanwhile, the coupling of much better soluble a-octi-(N-methylpyrrole) to the corresponding hexadecamer became possible [47].

9,15,16 (n=1,2,3)

1,2,3(n=1,2,3)

2,4,6 (n=1,2,3)

When lithiumdiisopropylamide (LDA) was used as base the yield of the oxidative dimerization of H-T2-H 2 to H-T4-H 4 was raised to 86%. In the same way, H-T6-H 6 was obtained from a-terthiophene H-T3-H 3 in 65% [48] and 73% yield [43] after purification. Nevertheless, lithiation of oligothiophenes always includes the problem, that with the use of equimolar amounts of base a mixture of the desired product and its dimer is always found. Thus, the reaction of Li-T1-H 9 with CuC12 gave 41% of H-T2-H 2 and 30% of H-T4-H 4. The equilibrium obtained in a mixture of H-T2-H 2 and Li-T1-H 9 evidently favors the lithium salt of the Li-T2-H 15 [Eq. (5)]. This clearly indicates that the a-protons of a-oligothiophenes H-T,-H exhibit greater acidity compared to those of H-T1 -H 1. The use of half equimolar amount of base, however, led to the nearly exclusive formation of the desired oligomer. The excess of unreacted H-T1-H 1 could mostly be recovered [43].

9

2

15

1

This method has furthermore been used to synthesize a quaterthiophene regioisomer. Thus, the reaction of 3,3'-bithiophene 11 with one equivalent of n-BuLi affords the 2-lithio compound 17. Coupling of the latter with CuC12 led in 6%

95

3.2 Synthesis of Oligothiophenes

yield to 3,3‘: 2’,2’’: 3”,3”’-quaterthiophene 18 which includes one a-a-linkage and two @-@-linkages[Eq. (6)]. A crystallographic study showed that the thiophene rings were not coplanar. The angle between the adjacent rings in the 3,3’-bithienyl moieties was found to be about 20” [49].

n-Buli

2

2

2x

CUCI,

17

11

Cyclopolyarenes have also been synthesized in moderate yields by oxidative coupling of Grignard or lithium derivatives with halides of transition metals (CuC12, NiC12) [50]. In this manner, Kauffmann et al. described the only examples of cyclo(o1igothiophenes) reported so far [42, 46, 5 11. The required dilithiated thiophene species were obtained by metal/halogen exchange of the corresponding dibromobithiophenes. Interestingly, both the coupling of 3,3‘-dilithio-2,2‘-bithiophene 19 and 2,2’-dilithio-3,3’-bithiophene 20 with CuC12 or FeC13 afforded the cyclo(tetrathophene) 21 in 23% and 24% yield, respectively [Eq. (7)]. The macrocyclic oligothiophene (cycloocta [1,2-b :4,3-b’ :5,6-b”:8,7-bN’]tetrathiophene)21 comprises two a-a-linkages and two @-,Blinkages between the thiophene rings involved. As side product a cyclohexathiophene was isolated in 4% yield in one experiment.

fi s

s

-w

CUCI,

CUCI,

_.__, s

u t i

s 21

19

(7)

S

20

GS

S

a 22

s

s

23

n-Buli

CUCI,

-

1

&Ql I

(9)

21

s

s

96

3 Oligothiophenes

4,4’-Dilithio-3,3’-bithiophene 22 was converted to the isomeric all-&linked cyclo(tetrathiophene) 23 (cycloocta [1,2-c:4,342’: 5,6-c”:8,7-~”’]tetrathiophene)[Eq. (S)]. Its structure was determined by X-ray crystallography [42, 511. The planes of the corresponding macrocyclic adjacent thiophene rings form an angle of 53.7” which is smaller than that in cyclotetrabenzene. Even the oxidative coupling of two macrocycles 21 with the system n-BuLilCuC1, can be achieved in 28% yield forming the corresponding bis-macrocycle 25 [Eq. (9)] [42, 511. The synthesis of the first all-a-linked cyclo(o1igothiophenes)are presently under way [52]. Finally, the oxidative coupling with CuC12 has also been used for the chemical synthesis of various polythiophenes. Thus the reaction of the bifunctional Li-T2-Li 26, obtained in 92% yield by deprotonation of H-T2-H 2, with two equivalents of n-BuLi, with cuprous chloride in anisole lead to poly(bithiophene) P2, an insoluble brown precipitate [Eq. (lo)]. After extraction, the polymer was obtained in yields ranging from 25 to 50%. Its ‘doping’ with AsF, afforded a polymer with a conductivity of 5 S cm-’ which is somewhat lower than that determined for films grown electrooxidatively [53].

Chemical and electrochemical oxidative coupling Another synthesis of H-T6-H 6 is described in a patent where the chemical oxidation of H-T3-H 3 with iron(II1) chloride via the dimerization of the radical cation results in H-T6-H 6 in 84% yield [54]. In a very recent paper the synthesis of H-T8-H 14 from H-T4-H 4 is described by the same oxidation procedure, but no yield is given [55]. This method is also used for the chemical polymerization of thiophenes to polythiophenes [54]. The related electrooxidative dimerization of H-T4-H 4 and a-quinquethophene H-T5-H 5 extended the series of a-conjugated oligothiophene up to H-T8-H 14 and a-decithiophene H-Tlo-H 27,respectively [Eq. (ll)]. Since in this homologous row the increase in chain length goes along with a dramatic decrease in solubility, it is doubtful and not reported whether or not these derivatives were easy to purify and characterize [56].

4,5 (054.5)

14,27 (n=4,5)

Oxidative coupling of organoboranes Kagan reported in 1983 the synthesis of the whole series of H-T,-H via organoboranes [26d]. The oligomers can be prepared separately in acceptable yields with organoborane reagents. To a solution of Li-T,-H 9, 15, 16, 28, the 9-methoxy derivative of 9-borabicyclo[3.3.llnonane is added and a corresponding boronated thiophene 29-32 is formed [Eq.(12)]. After neutralization with boron trifluoride etherate, a second Li-T,-H, which may differ from the previous one, is added to boranes 33-36 and the resulting complex 37-41 is oxidized with iodine in order to

3.2 Synthesis of Oligothiophenes

97

couple the two thiophene units attached to the boron moiety. Thus, 8 1% H-T2-H 2, 37% H-T3-H 3, 50% H-T4-H 4, 55% H-T,-H 5, and 59% H-T6-H 6, could be synthesized in a particularly simple procedure which allows the choice of reagent in each of the steps [Eq. (13)].

9,15,16,28 &=12.3,4)

29,30,31,32 @=1.2.3.4)

33,34,35,36 @51,2.3.4)

Li+

r-i

R

H 33,34,35,36

elA3.4)

R

H

37,38,39,40,41 @[email protected],4.5.6)

2,3,4,5,6 (n=2,3,4,5.6)

Davies et al. synthesized H-T2-H 2 in 50% yield reacting the ethanolamine ester of di(2-thieny1)borinic acid with N-bromosuccinimide (NBS) as brominating agent. The investigation aimed at the use of the easily isolable borinic acids as intermediates in synthetic organic chemistry [57].

3.2.1.2 Transition Metal Catalyzed Coupling Methods The 'Ullmann reaction' Although several improvements to the original Ullmann procedure have been made, for instance: using DMF as solvent to avoid the formation of higher oligomers [58], using copper acetate instead of copper [591, or using directly organocopper derivatives [43, 601; the Ullmann biaryl synthesis is still best suited for the coupling of acceptor substituted and electron deficient arenes [611. Also, the Ullmann reaction has recently been employed with great success in the synthesis of well-defined oligopyrroles from N-protected a,a-dibromo(oligopyrro1es) and elemental copper in DMF [62]. Mechanistic studies supported the evidence that arylcopper compounds are intermediates in the synthesis of biaryls [60, 611. This observation smoothed the way for the efficient synthesis of unsymmetric biaryls in a two-step procedure consisting, first, in the preparation of an, e.g. thienylcopper derivative followed by the treatment with different aryl halides. 2-Thienylcopper 42, for example, is prepared from the corresponding Li-T1-H 9 or Grignard derivative and a copper(1)halide and is then reacted with iodo- or bromoarenes in pyridine or quinoline. By this procedure e.g. H-T2-H 2 is formed in 42% yield by the reaction of Cu-TI-H 42 and 2-iodothiophene 8 [61] [Eq. (14)]. Analogously, 2-(p-nitropheny1)thiophene is obtained in 70% when Cu-TI-H 42 is treated with p-iodonitrobenzene. Note that the formation of undesired symmetric biaryl products is avoided [60, 61, 631. Qu

+ 9

cut

-

+

QC" 42

,/o Pyridine_ Q y ) 75 "C

8

u

2

(14)

98

3 Oligothiophenes

Nickel- and palladium-catalyzed cross coupling reactions - the ‘Kumada reaction’ The real breakthrough in oligothiophene synthesis was achieved when very effective transition metal catalyzed cross coupling reactions were applied to the C-C-bond formation between heterocyclic systems. Cross-coupling means the formation of C-C-single bonds on the basis of the reaction of an organometallic compound with an organic halide. The classical ‘Wurtz reaction’ includes a possible metalhalogen exchange and hence, the formation of homocoupled products. Unsaturated organohalides with CSpz-Hal bonds do not enter non-catalyzed C-C-bond formation. Although it has been known for a long time that catalytic amounts of transition metals induce the coupling of Grignard reagents with organic halides, the so-called ‘Kharash reaction’ was seldom employed in synthesis because complex mixtures of cross-coupling, homo-coupling, and disproportionation products were formed [64]. More attention was paid in 1971 to a publication of Kochi et al. who reported that soluble catalysts of silver, iron, or copper were very effective for the selective coupling of Grignard reagents with organic halides [65]. One year later Corriu et al. [66] and Kumada et al. [67] discovered independently that phosphine complexes of nickel catalyze the selective cross-coupling of Grignard reagents with aryl and alkenyl CSpz-halides. Murahashi found later, that also phosphine complexes of palladium exhibit catalytical activity in the reaction of alkyllithium compounds with haloolefins [68]. This opened up the possibility for the preparation of a wide variety of unsaturated organic compounds from two different organic halides [69] [Eq. (15)]. These reactions, which are described in more detail in the following paragraphs, allow the synthesis of even- and odd-numbered all-a-linked oligothiophenes and regioisomers with various chain lengths and substitution patterns [70].

In 1982, at a time when much attention was already paid to conducting polymers, and polythiophene was first synthesized [8], an extensive report of Kumada et al. was published describing Grignard cross-coupling reactions including heterocyclic compounds, mainly thiophene and pyridine derivatives [71]. It was shown that some oligothiophenes are most effectively and conveniently prepared by the reaction of the Grignard reagent of 2-bromothiophene 42 with various bromo- or dibromothiophenes. By adding the Grignard solution to a mixture of the organic halide and catalytic amounts of nickel-complexes (0.1- 1mol.%) in ether solution, the reaction proceeds under very mild conditions and gives the coupling product in h g h yields. The ‘Kumada reaction’ has become the most frequently used method in the synthesis of various types of thiophenes. Since its mechanism is basically valid also for other reactions catalyzed by transition metals, it will be elucidated more thoroughly [69, 701. The catalytic cycle is depicted in Fig. 3. (1) In the introductory step, the dihalophosphinenickel L2NiX2 reacts with two equivalents of the Grignard reagent to form the intermediate bis-organo complex L2NiR2. Its formation is easily observed by the dissolution of the insoluble catalyst to a reddish-brown solution.

3.2 Synthesis of Oligothiophenes

RX

99

’ i (2)

i A

R’X’ I R bNi<

LNi’

‘R

R

RX‘

Figure 3. Catalytic cycle of the organonickel complex promoted cross-coupling reaction of Grignard reagents with haloarenes by Kumada [69,70].

(2) On the action of an organic halide R’X’ the bis-organo complex L2NiR2releases the homo-coupling product R- R to form L2NiR’X’,the actual catalytic species. (3) The halogen X’ is readily replaced by the organic group R of another equivalent of Grignard reagent. The resulting complex L2NiRR’ comprises the two coupling components. (4) A second haloarene is taken up to form the fivefold coordinated intermediate L2Ni(. . .R’X’)RR’. (5) The catalytic cycle closes when the cross-coupling product R-R’ is set free and the catalytic species L2NiR’X’recovered. Due to the starting reaction in the catalytic cycle the homo-coupling product may be found to a certain extent which is typically around 0.5-1.5% [72]. Nevertheless, the reaction is usually very selective and gives the cross-coupling products in high yields. The reaction is applicable to various types of Grignard reagents (e.g. aryl, alkyl) and organic halides bearing a CSp2-carbon(e.g. aryl, vinyl). The reactivity order of the halide component was found to be Ar-I > Ar-Br > Ar-Cl > Ar-F. Grignard reagents are equally prepared either in Et20 or in THF, however, the

100

3 Oligothiophenes

v\M,,P% 4

‘5

PhP-(CH,),-PPh, PA,:

PPh,

C I C I 43 (M = Ni, Pd)

n=2, dpp. n=3. dppp

dppt

Figure 4. Structures of organonickel and -palladium complexes which are active catalysts in crosscoupling reactions of Grignard reagents with haloarenes.

reaction proceeds considerably faster in E t 2 0 than in THF and the dehalogenation product is formed in larger amount in THF. For poorly soluble compounds, either the Grignard or the halogenated reagent of oligothiophenes, benzene works very successfully as cosolvent [73]. From the large variety of catalysts 43 examined for thiophene synthesis, Ni(dppp)Cl2 [74], Ni(dppf)C12 [75], and Ni(dppe)C12 [74] were the most effective (Fig. 4). However, in some cases where these Ni-catalysts are not very reactive or side reactions occur, the analogous Pd-complexes L2PdX2 gave much better results [71]. In general, the Pd-complexes are less reactive, but more selective. Rossi et al. discovered that depending on the ratio of the reacting compounds, the Pd(dppf)C12-promotedreaction of the Grignard reagents of 2- or 3-bromothiophene with dibromothiophenes results in either mono-coupling to synthetically very valuable bromobithiophenes or in a twofold reaction to tertluophene isomers [76]. Thus, H-T2-H 2 (go%), 3,3’-bithiophene 11 (100% GC), H-T3-H 3 (80% GC), and H-T4-H 4 (64%) were synthesized by the reaction of the Grignard reagent of 2-bromothiophene BrMg-T1 -H 42 with 2-bromothiophene Br-TI -H 44, 3-bromothiophene, 2,5-dibromothiophene Br-T, -Br 45, or 5,5’dibromobithiophene Br-T2-Br 46, respectively [Eq. (16, 17)] [71]. In a systematic study by Zimmer et al., some yields were increased (2: 81%; 3: 86%; 4: 89%) and the series was extended to H-T5-H 5 (60-70% [49], 91% [77]). The synthesis of H-T6-H 6 failed due to the inherent insolubility of a,a-dibromoquaterthiophene Br-T4-Br 48 in ether or tetrahydrofuran (THF). Nevertheless, H-T6-H 6 could then be obtained in 56% yield by the ‘Kumada reaction’ of the Grignard reagent BrMg-T2-H 49 with Br-T2-Br 46 [Eq. (18)]. However, the synthesis of monobrominated oligothiophenes, the precursors of the Grignard reagents, is still rather difficult (see section 2.3) and Br-T2-H 50 could finally be obtained by selective metal/halogen exchange of Br-T2-Br 46 with one mole of n-BuLi and successive aqueous work-up [77].

42

45,46,41,40(n=l,2,3,4)

3,4,5 (n=1.2,3)

3.2 Synthesis of Oligothiophenes

49

46

101

6

Naarmann et al. used exactly the same reactions to synthesize the whole series of a-oligothiophenes up to H-T6-H 6. No yields are given, but the authors paid very much attention to the purification procedure and the physical constants of the various thiophene oligomers [12]. The use of iodo compounds proved to be even more effective, since H-T2-H 2 was obtained by the nickel-catalyzed coupling [Ni(dppe)C12]of I-TI-H 8 and its Grignard derivative in 90% yield [78]. The fourteen regioisomers of terthiophene, which exhibit an interesting photoenhanced toxic activity, have been satisfactorily synthesized mainly by nickel- or palladium-catalyzed Grignard cross-coupling reactions. The synthesis of these isomers has been reviewed very thoroughly [37b,c] and their significance with respect to conjugated materials is rather incidental. Table 3 (section 3.2.1.4) gives only a short overview including yields and physical properties. Generally, the yields are somewhat higher for the phosphine complexes of nickel as catalyst than for the palladium derivatives. The Ni-catalyzed homo-coupling of 2-bromo-3,3’-bithiophene to form the P,a,Plinked quaterthiophene isomer 18 also proved to be far more effective (84% yield) than the oxidative coupling of the corresponding lithiated bithiophene 17 with CuC12 (6% yield) [49]. A method originally developed by Colon and Kelsey for the symmetric coupling of arenes was used for the synthesis of even numbered a-oligothiophenes [79]. The catalyst is prepared in situ by the reduction of NiC12 in the presence of PPh, or Ni(PPh3)2C12with zinc in DMF as solvent. Thus, Br-T1-H 44, Br-T2-H 50, and the corresponding terthiophene derivative Br-T3-H 51 have been coupled to H-T2-H 2, in 41% [80], to H-T4-H 4 in 66% [Sl] and 87% [82], and to H-T6-H 6 in 48% yield [12,81], respectively [Eq. (19)]. Compared to the oxidative coupling of the corresponding lithiated oligothiophenes to the a-oligothiophenes 2 85% [42], 4 86% [43], 6 73% [43] the yields in this type of reaction turn out to be somewhat lower. With this procedure, however in only 18% yield, even H-T8-H 14, could be obtained from 5-bromo-a-quaterthiophene Br-T4-H 52 which was synthesized from H-T4-H 4 in 51% yield [81]. Besides H-Tlo-H 27 whose existence could not be proven unequivocally [57], H-T8-H 14 is the highest member of a-oligothiophenes synthesized so far. The octamer is practically insoluble in all solvents which prevents closer examination, e.g. of its UV spectrum in solution. However, a UV spectrum of an evaporated thin film of H-T8-H 14 could be investigated [56].

44,50,51,52 (n=1,2,3,4)

2,4,6,14 (n-1.2.3.4)

Finally, both methods of Ni(0)-catalyzed reactions were also used for the chemical synthesis of various polythiophenes. For instance, reaction of the bifunctional

102

3 Oligothiophenes

Br-TI -Br 45 either with Ni(dppe)C12 and Zn in HMPT at 150°C or with Mg and Ni in THF led to polythiophene P1 as insoluble powders [Eq. (2011 [83].

PI

45

A limitation of the applicability of the nickel-catalyzed reactions may occur either when the heterocyclic ring does not effectively add the Grignard reagent, or the halide does not give efficient halogen-magnesium exchange which then would lead to symmetric by-products [84]. The most serious limitation of the reactions described is, however, that substituents on both the organic halide and the Grignard reagent, are restricted to those which do not react with Grignard reagents. Thus e.g. the reaction of carbonyl groups with Li- or Mg-reagents leads to side products [71, 851. In order to circumvent this problem, either protected components [86] or, more versatile, organo derivatives of less electropositive metalloids may be used. The cross-coupling reactions especially for aryl/arylcouplings became applicable to compounds with a wide variety of functional groups by using less reactive, since less nucleophilic, tin-, boron-, or zincorganyls, in combination with appropriate palladium catalysts. These metal-organic arylating agents are moreover advantageous since they are relatively stable against oxygen and water and may be synthesized from the corresponding lithiated derivative and purified seperately. Copper-, zirconium-, or aluminum-organic compounds are less favorable since their synthetic availability is restricted or low catalytic activity is found. The ‘Stille reaction’ The palladium-promoted coupling of organotin compounds turned out to be especially fruitful because of the possible use of various organic electrophiles, the mild reaction conditions, the regioselectivity, and the tolerance of many functional groups (e.g. C02R, CN, OH, CHO, NOz). General aspects of this so-called ‘Stille reaction’ and of its applicabilty have been extensively reviewed [87]. With respect to heterocyclic compounds the literature has been compiled up to 1991 [88]. In the ‘Stille-type reaction’ the organometallic component R’SnRg may be synthesized from the corresponding lithiated derivative R’Li by the reaction with trialkylstannylchloride R”SnX or by the reaction of organic halides R’X with hexaalkyldistannane R;Sn-Sn3RN [89]. R’ may be alkinyl, alkenyl, aryl, benzyl, or ally1 and R” methyl or butyl which are typically not transferred [Eq. (21)]. In general, aryliodides, -bromides, and -triAates add even at moderate temperatures to the Pd(0)-complex, whereas arylchlorides must be activated by electron withdrawing substituents. R-X

+ R‘-SnR‘,

P45

R-R + e,SnX

3.2 Synthesis of Oligothiophenes

103

Pd2(dba)3. As a rule, catalytically active complexes of Pd(0) are derived from Pd(I1) in situ. The catalytic cycle is basically similar to that of the Ni-catalyzed ‘Kumada reaction’. It involves the conversion of Pd(0) to Pd(II), begins with the oxidative addition of the electrophilic reagent R-X to Pd(O), continues with transmetalation of R’ from the tin to the palladium compound, and closes with the reductive elimination of R-R’ from the palladium complex [Eq. (22)] [MI. R-X

Pd4,

R ‘ S n R r L2PdRX

LZRPdR’

% LzPdRR’

RR‘

An acceleration of the reaction rates has been achieved by the use of AsPh3 or P(f~ry1)~ as ligands [90] or cocatalytic Cu(1)- [91] and Ag(1)-species [92]. With respect to oligothiophenes the ‘Stille reaction’ has mostly been applied to the synthesis of substituted oligothiophenes (section 3.2.2). However, Crisp describes the synthesis of H-T2-H 2 (80%) and H-T3-H 3 (61%) by the Pd(PPh3)2C12catalyzed coupling of tributyl(2-thieny1)tin 53 with I-T1 -H 8 and 2,Sdiiodothiophene I-TI-I 54, respectively [Eq. (23)], [93]. The yields are somewhat lower than for the corresponding nickel-catalyzed coupling reactions. Reaction of the organotin compound 53 with Br-T1-Br 45 and Pd(PPh3)4 as catalysts yields H-T3-H 3 in 59% yield [82].

53

54

3

The ‘Stille reaction’, however, is limited by a side reaction which has sometimes been observed. The detection of coupling products including a phenyl group derived from the palladium catalyst indicates a rearrangment in which a phenyl group of the phosphine ligand migrates to the metal. A systematic investigation showed that the use of triarylphosphines with electron withdrawing substituents prevents this side reaction [94]. The ‘Suzukireaction’ Suzuki [85a] and Miller [95] have described the palladium-catalyzed coupling of various bromobenzene derivatives with benzene boronic acid or its cyclic esters. These cross-couplings which use the less electropositive boron instead of magnesium or tin proceed with good to excellent yields even in sterically demanding positions. Gronowitz et al. modified the palladium-catalyzed ‘Suzuki coupling’ so that it became useful to the heterocyclic series [96]. This carbon-carbon coupling method, as does the ‘Stille-coupling’, tolerates a variety of functional groups both in the organometallic reagent and in the heterocyclic halide. In the series of unsubstituted oligothiophenes, 2-thiopheneboronic acid 55, easily prepared from Li-Tl -H 9 and boronic acid trimethylester, reacts with Br-TI -Br

104

3 Oligothiophenes

45 to H-T3-H 3 in 40% yield [Eq. (24)]. Without any necessary precautions against oxygen, the reaction is performed in a mixture of DME and aqueous sodium bicarbonate, using [Pd(PPh3)4]as catalyst [96]. Similarly, 3,3‘-bithiophene 11 and 2,3‘-bithiophene 12 were synthesized by Gronowitz et al. using 2- and 3-thiophene boronic acid as organometallic component [97, 981. 3,2’: 5’,”‘terthiophene 56 has been obtained in 48% yield from 3-thiopheneboronic acid and Br-T1-Br 45 [96].

55

45

3

Typically, an excess of 20% boronic acid is used in the coupling reaction. Less excess results in the formation of mono-coupled by-products which are difficult to separate. A typical side reaction that may occur is deboronation. Gronowitz observed the greatest tendency for deboronation in the case of electron rich heteroaromatics [97]. Only few examples are known in which organozinc or organomercuric derivatives instead of the more reactive Grignard reagents are used in palladium catalyzed heteroaryl/heteroaryl-couplingreactions. In most cases, they are used for couplings of substituted thiophenes [99] or mixed thiophene/(hetero)arene compounds [1001. The various examples of mixed heteroarene compounds, including oligomers of thiophenes combined with pyrroles [loll, furans [102], pyridines [98b, 1021, or pyrimidins [99, 103, 1041, are not considered in detail in this chapter.

3.2.1.3 Ring Closure Reactions from Acyclic Precursors Aside from the arene-coupling methods, oligothiophenes can also be synthesized by ring closure reactions from acyclic precursor molecules. Since there have been many reports using ring closure methods for the synthesis of mainly bi- and terthiophene derivatives [37], in the following only those examples which are related to conjugated materials are described. Arornatization of tetrahydrothiophenes (thiophanes) One of the first reports using the aromatization of saturated precursor molecules was the preparation of 3,3’-bithiophene 11 from the tetrasodium salt of 1,2,3,4butanetetracarboxylic acid by twofold cyclization with phosphorous hexa- or heptasulfide in only 5% yield [36]. Gronowitz et al. improved the yield to 17% by reaction of 3-ketotetrahydrothiophene with 3-thienyllithium followed by dehydration and aromatization of the tertiary alcohol with chloranil [105]. In the following, dehydration procedures were used for the synthesis of several terthiophene isomers, however, always in low yields. A similar reaction of 3-ketotetrahydrothiophene with the Grignard reagent of 5-iodo-2,2’-bithiophene affords a tertiary alcohol in 81 % yield. Nevertheless, the aromatization to 2,3’:4’,2”-terthiophene proceeds less effectively (1 8% yield) [106]. This is also the case for the reaction of 5-(2-thienyl)-3-ketotetrahydrothiophene57 with BrMg-T1 -H 42 which results

3.2 Synthesis of Oligothiophenes

105

in 2,2’: 4’,2”-terthiophene 59 in 19% after aromatization of the tertiary alcohol 58 [Eq. (2511 ~ 7 1 .

m

42

3

0

Chloranil ____,

\ I

s

1

59

59

80

The same terthiophene isomer 59 has been obtained by the treatment of thiophene with orthophosphoric acid under mild conditions. The trimer 60 comprising a saturated central tetrahydrothiophene ring was successively oxidized with chloranil to form the terthiophene 59 [Eq. (26)] [log]. Cyclization of I ,I-diketones Another more convenient pathway for the synthesis of thiophenes proceeds via 1,4diaryl-substituted 1,4-diketones which are cyclized to thiophene rings by treatment with hydrogen sulfide and hydrochloric acid. Some methods for the synthesis of thophene-substituted 1,Cdiketones begin with the transformation of acetylthiophene. 2-Acetylthiophene 61 is reacted in a ‘Mannich reaction’ with formaldehyde and dimethylamine to yield the corresponding Mannich base 62 in 70% yield. The Mannich base 62 is then subjected to a ‘Stetter reaction’ [lo91 which results in 1,4-di-(2’-thienyl)-l,Cbutanedione 63 in 70% yield via the cyanhydrine of 2-thiophenecarbaldehyde [110]. Reaction of Mannich base 62 with the isomeric 3-thiophenecarbaldehyde under the same conditions results in 1-(2’thienyl)-4-(3’-thenyl)-1,4-butanedione64 in lower (35%) yield [Eq. (27)] [ 1101.

A

NaCN I DMF 63

0 61

0 62

NaCN I DMF

64

Another approach to 1,4-diketones is the oxidative coupling of the lithium enolate of 2-acetylthiophene 61 with CuClz in DMF. 1,4-Di-(2’-thienyl)-1,4-butanedione 63 is formed in 85% yield [l 1 I].

106

3 Oligothiophenes

Similarly, the silyl ether of 2-acetylthiophene 65 obtained by the reaction of 61 with trimethylsilychloride in DMF is oxidized with Ag20 to yield the dithienyl1,Cdiketone 63 in 71% yield [Eq. (28)] [112].

61

65

63

The isomeric 1&di-(3‘-thienyl)- 1,4-butanedione 68 is synthesized in 60% yield by oxidation of the silyl ether of 3-acetylthiophene 66 with phenyl periodate and BF3-etherate [Eq. (29)] [113].

JcHa

S

IDA

___, PhlO

M&cH2

BF3. Et,O

h4e3SiCI

ws

(29)

S

67

86

68

The cyclization of the 1,4-dithienyl-substituted1,Cdiketones to oligothiophenes proceeds by treatment with H2S and HC1 (classical Paal-Knorr synthesis [114]). Phosporous(V) sulfide or Lawesson’s reagent (L.R.) are also used as sulfurization reagents. H-T3-H 3 was obtained from 1,4-bis-(2/-thienyl)-1,Cbutanedione 63 and H2S/H+ in 70% yield [115], with P2S5 in 66% yield [110, 1121, and with L.R. in 85% yield [110], respectively [Eq. (30)]. The isomeric 3,2’: 5’,3/’-terthiophene 56 is similarly synthesized by the ring closure of 1,4-di-(3/-thenyl)-1,4-butanedione 68 with P2S5 in 75% yield [Eq. (31)] [113]. In the same way the unsymmetric 1(2/-thienyl)-4-(3‘-thienyl)-l ,Cbutanedione 64 is closed to 2’2’: 5/,3”-terthiophene 69 in 84% yield [Eq. (32) [110].

63

3

68

56

69

64

I

Laweson‘sReagent

I

3.2 Synthesis of Oligothiophenes

107

More recently Ellinger and Merz described the synthesis of H-T3-H 3 and H-T5-H 5 via acylation by the FriedelLCrafts method. Thiophene H-TI-H 1 and H-T2-H 2 were converted into 1,4-di(2/-thienyl)-1,4-butanedione 63 and the corresponding bis(dithieny1)-substituted diketone 70 by acylation with succinyl chloride and aluminum chloride in 55% and 25% yield, respectively [116]. Reaction of the diketones 63 and 70 with L.R. gave H-T3-H 3 and H-T5-H 5 in high yields (87’3’0, 92%) [Eq. (33)]. However, the Friedel-Crafts route was not successful for higher homologs since analogous acylation of H-T3-H 3 with succinyl chloride resulted in the formation of 5,5’-bis-(2,2/:5/,2/’-terthien-5-yl)tetrahydrofuran-2-one 71. The formation of this cyclic lactone may be explained via rearrangement of the intermediate 4-0x0 acid. n

1.2 (n4.2)

3,5 (n=l,2)

63,70 (n4.2)

(33)

71

Cyclization of diacetylenes The cyclization of thienyl-substituted diacetylenes with hydrogen sulfide or sodium sulfide is also a very successful method for the synthesis of oligothiophenes. The key point here is the synthesis of the precursor diynes. Symmetrical diacetylenes can be obtained by the oxidative coupling of acetylenes with copper(1) or copper(I1) salts (‘Glaser, Hay, or Eglington coupling’). The synthesis of unsymmetric diynes proceeds best via the ‘Cadiot-Chodkiewicz procedure’ in which an acetylene is reacted with an acetylic halide in the presence of copper(1) salts [117]. Therefore, thienylsubstituted acetylenes and bromoacetylenes are important starting materials. Ethinyl-substituted (o1igo)thiophenes are e.g. obtained by the Wittig reaction of the corresponding aldehyde 72-74 with tetrabromomethane and triphenylphosphine which results in 1,l-dibromoethylene substituted thiophenes 75-77. Successive elimination of HBr and halogen/metal exchange with n-BuLi leads, after aqueous work-up, to the desired acetylenes 78-80 in good overall yields [Eq. (3411 [26dI.

72,73,74 (n=1.2.3)

75,76,77 (n=1,2.3)

78,79,80 (nd.2.3)

Another procedure starts from I-T1-H 8 which is reacted with the Grignard reagent of bromo(trimethylsily1)acetylene and [Pd(PPh,),] as catalyst to yield a TMS-protected thienylacetylene 81. Deprotection is effectively feasible by treatment

108

3 Oligothiophenes

with base and results in the desired thienylacetylene 78 in good overall yield (84%) [Eq. (3511 1761.

8

81

78

2-Ethinylthiophene 78, 5-ethinyl-2,2’-bithiophene 79, and 5-ethinyl-2,2’:5’,”’terthiophene 80, have been dimerized with cuprous or cupric chloride under ‘Glaser conditions’ to the corresponding diacetylenes 82 in 87% [80, 1181, 83 in 73% [26d], and 84 in 99% yield [26d], respectively [Eq. (36)]. Similarly, 3-ethinylthophene 85 gives the corresponding 1,4-di-(3’-thienyl)-1,3-butadiyne 86 in 73% yield [Eq. (37)] [1061. The unsymmetric 1-(2’-thienyl)-4-(3’-thienyl)-1,3-butadiyne 87 is obtained by coupling 3-ethinylthiophene 85 and bromo(2-thieny1)acetylene under ‘Cadiot-Chodkiewitz conditions’ in 91% yield [Eq. (38)] [106]. Similarly, 1-(2’thieny1)-4-(2,2’-bithien-5-~1)1,3-butadiyne 88 is synthesized by the coupling of 5-ethinyl-2,2’-bithiophene79 and bromo(2-thieny1)acetylene in 96% yield [Eq. (3% [26dI. cu+of cu2+

78,79,80 (nd.2.3)

85

85

79

- -

-

82,83,84 (*1.2.3)

86

87

88

Organoboranates of the type (T,-BR2 -TJ Li’ 37-41 have already been mentioned with respect to their significance in the preparation of oligothiophenes (see above). Similarly, the coupling of two acetylenes to a butadiyne via an organoborane intermediate which is treated with iodine can be utilized to synthesize thiophene-substituted diacetylenes. Kagan et al. reacted 5-(lithioethyny1)2,2’-bithiophene 89 successively with 9-borabicyclo[3.3.llnonane and BF3-etherate

3.2 Synthesis of Oligothiophenes

109

to form borane 90 and 2-(1ithioethynyl)thiophene to yield the above mentioned unsymmetric 1-(2’-thienyl)-4-(2,2’-bithien-5-yl)-l,3-butadiyne88 in 69% yield in a one pot procedure [Eq. (40)] [26d]. R

I

89

90

The final step in the synthesis of oligothiophenes is now the cyclization of the thienyl-substituted diynes with hydrogen sulfide or sodium sulfide. 1,4-Di-(2’thienyl)-1,3-butadiyne 82 was cyclized to H-T3-H 3 with H2S in ethanol by Schulte et al. in 51 YOyield [119], and with Na2S in methanol by Kagan et al. in 84% yield [118]. H-T5-H 5 and H-T7-H 7 were synthesized by cyclization of the corresponding 1,3-butadiynes 83 and 84 in 74% and 98% yield, respectively [Eq. (41)]. H-T4-H 4 could be obtained in quantitative yield from the unsymmetric diyne, 1-(2‘-thienyl)-4-(2,2’-bithien-5-yl)-l,3-butadiyne 88 [Eq. (42)] [26d]. Terthiophene isomers 69 and 56 are formed by the reaction of 1-(2’-thienyl)-4-(3’-thienyl)1,3-butadiyne 87 [ 1 191 and 1,4-bis-(3‘-thienyl)-1,3-butadiyne 86 with sodium sulfide in 87% and in quantitative yield, respectively [Eqs. (43, 44)] [80, 1181.

02,03,04 (n=l.2.3)

3,5,7 (n-1.2.3)

00

87

86

4

69

66

110

3 Oligothiophenes

Oligothiophenesfrom 1,I-dithiins Nakayama et d.used thienyl-substituted 1,4-dithiinswhich are obtained from easily accessible diketosulfides for the preparation of a-oligothiophenes and isomers up to the heptamer [37b, 1201. The dithienyldiketosulfides 91-93 are prepared by the reaction of chloroacetyl-substituted (o1igo)thiophenes and sodium sulfide in almost quantitative yield and are further cyclized to the corresponding 1,Cdithiins 94-96 with L.R. in 60% yield. The extrusion of sulfur from the dithiin moiety via ylide intermediates is achieved by refluxing the dithiin in o-dichlorobenzene and results in a mixture of two possible isomers [Eq. (45)]. In the case of 2,6-di(2’thienyl)-l,Cdithiin 94, a ratio of 13: 1 of H-T3-H 3 and the 2,3’”’,2’’-isomer 97 in 85% yield is obtained. The separation of the compounds by recrystallization turns out well since the a@-connected terthiophene is better soluble in hexane. Oxidation of the dithiin with rn-chloroperoxybenzoic acid and extrusion of SO from the resulting sulfoxide in the presence of DMSO afford a mixture of H-T3-H 3 and the isomer 97 in a ratio of 22: 1 and in a total yield of 90% [120].

91,92,93 (n=1.2,3)

94,95,96 (m1,2,3)

(45)

heat

3,5,7 (n=l-3)

100

97,98,99 (n=1.2,3)

59

The synthesis was extended to the corresponding penta- and heptathiophenes. Thus starting from the chloroacylated bi- and terthiophene, the corresponding diketosulfides 92 and 93 are both obtained in 98% yield. Cyclization with L.R. and extrusion of sulfur results in H-T5-H 5 and H-T,-H 7 in an overall yield of 33% and 38%, respectively. In these cases, the amount of isomeric by-products 98 and 99 are small (<2%). Interestingly, the 2,5-di-(2’-thienyl)-1,Cdithiin 100 which forms only one ylide intermediate gives the 2,2’: 4’,”’-terthiophene 59 in 12% yield upon heating to 200°C [Eq. (46)] [107]. Various diketosulfides91-93 were also transformed by Nakayama and coworkers in a ‘McMurry reaction’ to the corresponding cis-hydroxydithiolane 101 or dihydrothiophenes 102 and 103. The reaction proceeds via intramolecular reductive coupling released by a low-valent titanium reagent (83-89% yield). Acid-catalyzed

3.2 Synthesis of Oligothiophenes

1 11

dehydration of hydrothiophenes 101-103 or oxidation with dichlorodicyanoquinone (DDQ) give rise to the isomeric oligothiophenes 97-99 including a central thiophene ring which is connected in the P-positions to other thiophenes (83-95% yield) [Eqs. (47, 48)] [121].

of '9 TICI,/Zn THF

S

HO

TsOH

OH

91

92,93 (nd.3)

t\

- 2 HZO

S

S

101

97

(47)

102,103 (n-2.3)

When the complete reaction sequence is applied to 2,3': 4',2''-terthiophene 97, the @linked septithiophene 106 which includes alternating a,@-and @,a-bonds is obtained in excellent overall yield (62%) via the diketosulfide 104 and the dihydrothiophene 105 [Eq. (49)] [121]. In contrast to the isomeric H-T7-H 7 which is red and only poorly soluble in organic solvents, the heptamer 106 is a white crystalline solid, easily soluble, and exhibits a melting point 200°C lower than its linear counterpart. The attempted synthesis of a corresponding pentadecamer from the isomeric septithiophene 106 led to a glassy solid which could not be characterized.

o,,$)

____,

1) CICH,COCI 2 ) Nais

I \ S 97

w . " j f 'o H o TCI.IZn THF. r.t.'

I \

I \ S

S 104

(49)

105

106

It can be concluded that a-oligothiophenes are available by many powerful methods. For bi- and terthiophenes not only the a-linked, but also the remaining regioisomers have been prepared, whereas for the higher homologs only few oligothiophenes comprising a,@-or @,@-bondshave been synthesized so far. The next section summarizes the physical constants of oligothiophenes. The last paragraph gives experimental details of some selected examples. 3.2.1.4 Physical Properties of a-Oligothiophenes and Isomers

Modern chemistry can describe the electronic and structural features of conjugated materials in general and of oligothiophenes in particular more specifically. The

1 12

3 Oligothiophenes

Figure 5. Mesomeric resonance forms of polyenes, oligothiophenes, and oligo-p-phenylenes.

general structural principle of these systems is represented by a one-dimensional chain comprising conjugated double bonds which may also include heteroatoms. Therefore, oligo- and polythiophenes can be regarded as conjugated chains consisting of sp2p,-carbon atoms which have an analogous structure to cis-polyacetylenes and are stabilized by sulfur atoms. The capability of a conjugated system to transduce electronic effects depends on the delocalization of the charge carriers which are created along the molecular chain due to electronic or optical excitation. Delocalization of charge carriers is represented by resonance structures. In the case of conjugated systems an aromatic and a quinoid structure can be drawn (Fig.5). Among conjugated chains, resonance is most pronounced for polyenes. Due to the energetic adequacy of the resonance forms, polyenes have a degenerate ground state. In contrast, 7r-systems consisting of linked aromatic or heteroaromatic units exhibit a reduced conjugation between these moieties. This is on one hand caused by aromaticity and on the other hand by the reduced planarity due to the steric hindrance of o-hydrogen atoms. Therefore, in the case of aromatic systems like oligo- and polythiophenes, the quinoidal resonance form is of only minor significance and leads to non-degenerate ground state [1221. A diagnostic criterion for the efficiency of delocalization is the bond length alternation [1231. Furthermore, physical properties directly correlated to conjugation phenomena are optical transitions, redox potentials and nonlinear optical effects. The physical properties (melting point, absorption, fluorescence, and oxidation potentials) of the various oligothiophenes are summarized in Table 1. The highest melting points found are noted in the table. With increasing chain length of the oligothiophene the melting point increases as expected. With respect to the electronic properties, the longest wavelength absorptions [131, emissions [1241 and also the oxidation potentials [ 12,771gradually shift to lower energies with increasing number of thiophene rings. This reflects, as expected, an increasing conjugation which is also observed by the colors of the homologous row. The color deepens from colorless for bithiophene 2, to pale and chrome yellow, to orange, to red, and finally to wine red for septithiophene 7 and octithiophene 14 [81]. However,

3.2 Synthesis of Oligothiophenes

1 13

only a rough correlation of the transition energies with the (inverse) chain length can be determined. This type of correlation is predicted theoretically [ 1251 and confirmed in several oligomer series [ 131. The optical data and oxidation potentials of the higher homologs are rather unreliable due to their very low solubility. Nor are the redox data of the smaller members in the series precise, since the cyclic voltammograms are irreversible due to rapid follow-up reactions of the radical cations created. Therefore, precise determination of the redox potential is difficult. Nevertheless, a reasonably linear relationship of the oxidation potential versus the inverse chain length was obtained by Diaz el al. for oligothiophenes H-TI-H 1 to H-T5-H 5, oligophenylenes and oligopyrroles (Fig. 6) [126]. The values for polythiophene, however, are clearly blue-shifted with respect to the extrapolated potential of an (hypothetic) infinite chain length. Despite these drawbacks, Fichou et al. were able to determine absorption spectra and magnetic properties of the radical cations and dications, respectively, of H-T5-H 5 and H-T6-H 6 in highly diluted solutions [127]. The charged oligothiophenes are considered to be ideal model compounds for the investigation of the charge carriers responsible for the charge transport in conducting polymers. ESR investigations of radical cations and anions of H-T2-H 2 to H-T5-H 5 reveal that the oligomers exist not exclusively in an all-trans conformation of the thiophene rings [ 1281. Absorption spectra of oligothiophene radical cations were also obtained by oxidation of the corresponding oligothiophenes in the channels of zeolites where the subsequent reaction of the reactive cationic species is suppressed by the local environment. However, only mixtures of cations were created in the presence of large excess of unreacted neutral species, so that assignment of the absorption bands is rather difficult [32b, 1291. The development of rectifying Schottky diodes and field effect transistors (FET) which contain thin films of oligothiophenes as the active semiconducting component, has stimulated many investigations of the solid state properties of oligothiophenes. Investigations on silicon dioxide/oligothiophene FETs clearly reveal that the charge transport properties of the oligothiophenes correlate with the inverse chain length [ 1301. Thus, the conductivity and mobility increase gradually with increasing chain length of the oligomer and considerably exceed those of polythiophene [13I]. The photophysics of ter- and quaterthiophenes were thoroughly investigated against the background of their phototoxic activities [1321. Transient absorption spectra which are identical to those of the corresponding radical cations and fluorescence kinetics in the row H-T2-H 2 to H-T6-H 6 exhibit an increasing fluorescence life time with increasing chain length [ 1331. Fluorescence spectra and quantum yields of thin films of oligothiophenes reveal clearly that in the solid state the fluorescence efficiency is dramatically diminished by more than three orders of magnitude [134]. Due to their extended conjugated system, oligothiophenes are also promising materials for non-linear optical (NLO) applications [16]. In this respect, it was shown that the third order susceptibilities correlate with the chain length. For example, H-T6-H 6 exhibits a x3-value which is higher than this of H-T3-H 3 by a factor of ten [135]. The electronic structure of oligothiophene films has also been investigated by electron loss spectroscopy [136], UPS and XPS [137].

Method of preparation (paragraph)

3.2.1. le 3.2.1.1" 3.2.1.1e 3.2.1.2' 3.2.1.2h 3.2.1.1e 3.2.1.1e 3.2.1.2g 3.2.1. le 3.2.1.2J 3.2.1.2g 3.2.1.2' 3.2.1.le 3.2.1.2'

3.2.1.2g 3.2.1 .2k 3.2.1.3" 3.2.1.3"' 3.2.1.2' 3.2.1.2J 3.2.1.3" 3.2.1.3O 3.2.1.3"' 3.2.1.2' 3.2.1.3" 3.2.1.3"'

3.2.1.1g 3.2.1.1" 3.2.1.2' 3.2.1.2'

Oligothiophene (nT)

H-Tz-H 2

H-T3-H 3

H-Td-H 4 66

64

55 64

37 40 51 55 59 61 66 70 70 80 84 85

26 30 41 41 42 44 50 50 54 80 81 81 85 90

Yield ["/.I

[26dI [421 t711 [12, 811

126dI [971 [1191 Dl61 [821 [931 [110, 1121 [12Oa] 11441 ~711 176, 801 t1101

[41, 421 [931 [26dI [771 1421 ~711

[451 1441 [431 [801 1621 [391 1401

Ref.

215-216

95-96

33

rn.p.= ["CI

Table 1. Preparation and physical properties of a-oligothiophenes H-To-H.

407,426

437,478

390 (4.66)

1811

362

Fluorescence A,, [nrnlc

355 (4.40)

302 (4.10)

Absorption A,,b(lg E ) [nml

[12,120a, 1431

[19c, 39,731

Ref.

0.97(0.12)

1.05 (0.21)

1.28(0.43)

Md

Oxidation potentials

6

$

-8

5

x

50

ru

P

r

c

7

q

'

g

a

-

18

-

-

86 87 89 99 25 55 56 60-70 74 91 48 56 59 65 73 84 38 98

'

WI

1541 [120bl [26dI [541 [561 [811 1561

1431 [821 [771 [26dI 11161 WdI [120bl [481 [26dI [771 [12, 811 [771 WdI [481

-

364

328

307-309

258-259

PI1

[120b]

[I21

[121

-

-

-

-

522, 560

510p

482, 514

4404

432 (-4.78)

416 (4.74)

-

-

-

0.46 (0.04)

0.70 (0.08)

Highest melting points are given. Maximal absorptions and extinction coefficients in CHC13 (lit. 77). Fluorescence maxima in dioxane/acetonitrile at 298 K (lit. 124). Irreversible oxidation potentials, differential pulse voltammetry in CH,CN/TBABF vs. Ag/AgCI (lit. 77); in parentheses cyclic voltammetry in propylencarbonat/LiC104 (0.5 M) vs. Ag/Ag+ (lit. f12). Copper(I1)-promoted oxidative coupling. Chemical and electrochemical oxidative coupling. Oxidative coupling of organoboranes. U l m a n n coupling'. j 'StiIIe coupling'. 'Kumada coupling'. 'Suzuki coupling'. Aromatization of thiophanes. Cyclation of diacetylenes. Cyclation of 1,4-diketones. Predicted value based on a 1/n vs. Emaxplot. From 1,Cdithiins. Not reliable due to low solubility.

H-TI0-H 27

H-TB-H 14

H-T,-H

H-Ts-H 6

H-TS-H 5

3.2.l.le 3.2.1.21 3.2.1.2' 3.2.1.3" 3.2.1.3m 3.2.1.l P 3.2.1.30 3.2.1.2' 3.2.1.3n 3.2.1.21 3.2.1.21 3.2.1.2' 3.2. I.lg 3.2.1.1" 3.2.1.le 3.2.1.1f 3.2.1.3' 3.2.1 .lg 3.2.1.1 3.2.1. I f 3.2.1.2' 3.2.1.1f

ul

-

8

2

$-

52

=:

0

i;.

8

*$x

Figure 6. Plot of oxidation potentials (Epa)versus inverse chain length (1 /n) for aromatic oligomers [126].

Since for electronic and optical applications the solid-state structure of the conjugated materials and the interface of the semiconductor to the (metal) substrate are crucial, the bulk and surface properties of oligothiophenes were investigated by several methods. Angle dependent and polarized absorption and IR-measurements on sexithiophene single crystals elucidate the spatial arrangement of the molecules [1381. Optical measurements on oriented monolayers of various oligothiophenes reveal a different orientation of the molecules depending on the substrate surface [134]. XRD measurements show that H-T6-H 6 forms a liquid crystalline mesophase at higher temperatures [1391. Crystalline charge-transfer complexes are formed by the combination of oligothiophenes with TCNQ [140] or heteropolyanions [141]. In the former case the variation of the oligothiophene components do not affect the relatively low conductivities (lo-'' to Scm-'). The conductivity of a H-T6-H/PMol20;i charge-transfer complex, however, is notably high (2.7 S cm-I). Interestingly, oligothiophenes are oxidized as well in the solid state, as in the gas phase. The simultaneous evaporation of H-T6-H 6 and FeC13 results in thin films whose electronic structure corresponds to the one of 'doped' a-sexithiophene [142]. Table 2 also gives the data for the regioisomeric and cyclic oligothophenes described above. For the isomeric oligothiophenes which include a,@-or P,P-bonds

296

60-6 1 59-60 53-54 38-39 101-102 158-160 39-40 64 68-69 103-104 49-50 157.5-158.5 156- 158 193 81-82 83.5-84.5

85 83 95 93 72 87 95 76 83 94 77 61 48 100 85 50

3.2.1.2’ 3.2.1.3m 3.2.1.21 3.2.1.2’ 3.2.1.3n 3.2.1.2‘ 3.2.1.30 3.2.1.21 3.2.1.2; 3.2.1.21 3.2.1.2’ 3.2.1.3n 3.2.1.2‘

2,2/: 3’,2”

2,2’: 4’,2” 2,2’: 3’,3/’ 2,2‘: 4’,3” 2,2’: 5’,3” 2,3’: 2‘,3” 2,3’: 4’,2/’ 2,3’:4’,3” 2,3‘: 5‘,3” 3,2‘: 3’,3” 3,2’:4’,3“

3,2’: 5’,3” 3,3’: 4’,3”

3,3/:2’,2/’:3”,3/”

i-T, 107

i-T3 i-T3 i-T3 i-T3 i-T3 i-T3 i-T3 i-T3

59 108 109 69 110 97 111 112 i-TJ 113 i-T3 114

i-T3 56 i-T3 115

i-T4 18

88-89 203 126- 127

10 26 28 86 80

3.2.1.1e 3.2.1.1e 3.2.1.1” 3.2.1.3” 3.2.1.3’ 3.2.1.3”

C-T4 21 C-T4 23 (c-T4)Z 25

i-TS 98

i-T, 99 i-T, 106

98

255-266 300-301 355 (dec.)

6 84

3.2.1. le 3.2.1.2’

169.5- 171.5

283 260

68-68.4 133-134

55 88

370 288

262

324 250

282 295 310 33 1 292 272 242 272 275 262

[nml

-

(1g €1

3.2.1.I” 3.2.1.1”

& I , ,

2,3‘ 3,3’

Ref.

i-T2 12 i-T2 11

m.p. [“CI

[YO]

Yield

Method of preparation (paragraph)

Connectivity

Oligothiophene (i-Tn, c-T,)

Table 2. Preparation and physical properties of regioisomeric and cyclic oligothiophenes.

Ref.

118

3 Oligothiophenes

general structure/property-relationships are not as evident. Thus e.g. comparing the three septithiophene isomers 7, 99, 106, the melting points decrease the more pbonds are involved in the molecule which then becomes less linear (7:328°C; 99: 203°C; 106: 126- 127°C). Simultaneously, due to the distortion of the corresponding thiophene rings and the weaker overlap of the orbitals involved, the absorption maximum is shifted to shorter wavelength and higher transition energies (Arnm = 7: 440 nm; 99: 370 nm; 106: 288 nm). This seems to be a general trend and is also = 2: 302nm; 12: 283nm; 11: observed in the series of bithiophenes (2, 12, 11: A,, 260nm) and terthiophenes (e.g. 3, 109, 114: A,, = 3: 355nm; 109: 310nm; 114: 250 nm). However, in the bithiophene sequence 2, 12, 11 the melting points rise (2: 33°C; 12: 68-68.4"C; 11: 133-134°C).

3.2.2 Substituted Oligothiophenes While polythiophene itself is intractable and therefore not processable, in the neutral state poly(3-alkylthiophenes) are soluble in regular organic solvents and still exhibit electrical and optical properties comparable to the unsubstituted derivatives [ 1511. Since the description of the first poly(3-alkylthiophenes) in 1986, structure examination in solution has become possible and information about the type of connectivity between the thiophene rings and the mean chain length is available. Nevertheless, due to head-to-head couplings, adjacent alkyl substituents cause steric hindrance and diminution of the conjugation via non planar conformations. The degree of regularity and the structure of the polymers also depend strongly on the polymerization conditions. Besides the typical chain length distributions and conjugation defects for poly(3-alkylthiophenes), stereoregular factors play an additional role. Recently, the very effective synthesis of highly regioregular poly(3-alkylthiophenes) via selective metal-catalyzed coupling reactions has been reported independently by McCollough [1521 and Rieke [153]. The polymers consist of alkylated thiophene rings which are mostly linked head-to-tail (>95-99%). In comparison to the corresponding random polymers, these polythiophenes exhibit highly crystalline parts and strongly increased conjugation despite moderate molecular masses. Thus, the absorption maxima and the band aps are shifted to lower energies and the conductivities reach values (1350 S cm-') which are considerably higher [ 1541. In analogy to the polymers, the solubility of oligothiophenes decreases dramatically with increasing chain length, which is due to the stiffness of the conjugated 7rsystem and the strong interactions between the chains, they are difficult to purify. Especially the long members are virtually impossible to characterize. This problem of low solubility can be overcome by the synthesis of corresponding oligothiophenes which bear alkyl groups in /3-positions. As an example, the solubility of a dialkylated sexithiophene which is described in more detail later on, is higher than 400 g 1-' whereas that of H-T6-H 6 is lower than 0.05 gl-' [155]. This solubility provides an opportunity to synthesize much longer oligomers which would serve as desired models for the better understanding of polymeric systems. They therefore represent ideal model compounds for the investigation of charge carriers which are responsible for the charge transport along the conjugated chains. These properties

3.2 Synthesis of Oligothiophenes

119

may be correlated with the (defined) chain length of the oligomers and be compared to that of the ‘real’ polymer. However, the characterization of radical cations and dications as models for polarons and bipolarons demands the synthesis of longer oligomers because the shorter ones are inherently reactive and undergo follow-up reactions. Thus, the concept of conjugation length plays an important role in the theory of conductivity and the non-linear optical properties of conjugated polymers. The modification of conducting polymers with flexible side chains in order to improve their tractability has in the meantime become widespread, but the influence of such a modification on the conjugation length remains rather subtle. In order to study the interrelation between chemical structure and various physical properties in such polymers, starting in 1991, several series of partially alkyl-substituted oligothiophenes have been prepared as model systems which are described below. Although a variety of differently substituted bi- and terthiophenes have already been synthesized and charaterized before, e.g. with respect to their biological properties, only those which show electronic properties interesting for new organic materials or which served as starting material for polymerizations will be described. 3.2.2.1

p, fl-Substituted Oligothiophenes

The concept of conjugation length was a major motivation for the Dutch group of Wynberg to synthesize in 1991 the first series of a,a-coupled oligothiophenes which bear solubilizing alkyl side chains in the free P-positions [ 1561. A series starting from a trimer up to an undecamer including two oligomers with t-butyl endgroups in the terminal a-positions has been developed. The synthetic strategy chosen starts from 1,4-diketones whch are prepared via the ‘Stetter reaction’, followed by ring closure with L.R. (see above). Thus, a series of oligothiophenes 116-122 including 3 to 7 , 9 and 11 thiophene rings were built up. This synthetic strategy, however, leads in general to a mixture of isomers differing in the position of the n-alkyl groups in certain thiophene rings. Thus e.g. alkylsubstituted 1,4-diketone 116 is cyclized with L.R. to terthiophene 117. On one hand this is bis-acylated to terthiophene 118 which is further converted into the bis-Mannich base 120. On the other hand, terthiophene 117 is formylated under ‘Vilsmeyer-Haack’ conditions to the terthiophene carbaldehyde 119. Both compounds are subjected to a ‘Stetter reaction’ yielding the bis- 1,4-diketone 121 which is cyclized by L.R. to the undecamer 122 [Eq. (50)]. The properties depend critically on the effective conjugation length, structure and conformation, since adjacent alkyl substituents give rise to steric hindrance and, hence, to a nonplanar conformation. In this series, certainly the solubility is reduced with increasing chain length, but the different undecamers even exhibit a solubility of 2-10 g 1-’ in chloroform, whereas the conductivity (20 S cm-’) is in the same order of magnitude as for doped poly(3-alkylthiophenes) (10-100 Scm-’) [157]. This is a clear indication that the effective conjugation length in the corresponding polymers is not much longer than 11 units. The UV-absorption, however, is a more direct measure of the conjugation length. In this series, the absorption maxima are steadily = 345 nm for terthiophene 117 to red-shifted with increasing chain length (A, 462 nm for undecamer 122) and seems to approach saturation, without actually

120

3 Oligothiophenes

2

ww ""W +

\ I

0

0

R

R

119 (65%)

120 (70% (=%)I

R

121

1 k

LR

I

R

122 16% (44W (last tm, Steps)

R A = C.4 (C,Ji,J

reaching it. The conjugation lengths of the nonamer and undecamer 122 exceed that of poly(3-alkylthiophenes),,A,( = 430-440 nm) due to less steric interaction of the side groups which are more widely spaced in the case of the oligomers. Absorption spectra of solid films for both oligomers are much like poly(3-alkylthiophenes) = 520nm) than in solution. This is an and exhibit maxima at lower energies,,A,( indication of a larger conjugation length in the solid state. In 1992, the group of Garnier published the synthesis of oligothiophenes substituted with solubilizing decyl side chains in P-positions [1551. Their synthetic strategy was different from that of the Dutch group and started from the 3'-decylterthiophene 124 which is obtained by Kumada-coupling of two equivalents of the Grignard reagent of 2-bromothiophene 42 and 2,5-dibromo-3-decylthiophene 123 in 92% yield. The corresponding hexamer 125 bearing two alkyl side chains was isolated by the oxidative coupling of the lithiated trimer 124 with CuC12 in 50% yield. This in situ homo-coupling is not selective and resulted also in regioirregularly substituted products. Sexithiophene 125 could be purified and separated from longer oligothiophenes by chromatography. The optical spectra in solution and in the solid state, respectively, match with those of the corresponding sexithiophene synthesized by the Wynberg group. An unusually high fluorescence quantum

3.2 Synthesis of Oligothiophenes

121

yield of = 50% is reported for sexithiophene 125 which is double that for unsubstituted H-T6-H 6. A high solubility of 400gl-' (1 moll-') in chloroform is given for the dialkylated sexithiophene 125 and the electrochemical coupling of this hexamer is reported. The cyclovoltammograms show two quasi-reversible oxidation waves which are due to the isomeric mixture rather less structured (E? M 0.47 V, E; = 0.65 V vs. Ag/Agf) and correspond to the successive formation of radical cations and dications. By repetitive cycling, thin adsorbed films are obtained which exhibit redox waves with a shape typical for poly(3-alkylthiophenes) (Epa= 0.37 V, Epc = 0.20 V vs. Ag/Ag+). The peak potentials correspond to those of the best polythiophene films. Since the narrowness of the peaks (half-height width = 140mV) suggests homogenous conjugated chains and the optical characterizations revealed a red shift of the maximum absorption in comparison to the undecamer 122 prepared by Wynberg et al., it is concluded that by electrocoupling a corresponding dodecamer 126 is formed. These experiments which demonstrate the strength of the 'oligomeric approach', showed that stable and soluble radical cationic and dicationic model compounds can even be obtained from intermediate long oligomers and provide insight into the doping and conducting mechanism of polythiophene. Indeed, despite the relatively short conjugation length, they behave much like conducting polymers, and the synthesis of longer members would certainly result in an even better performance.

123

42

c'oyl

125 (SOSC)

Consequently, due to the good solubility, the Paris group was able to couple the dialkylated sexithiophene 125 via the same method to a tetraalkylated regioirregular dodecithiophene 126 in 15% yield [Eq. (51)] [158]. Since lithiation of oligothiophenes is not selective, duodecithiophene 126 had to be separated from higher homologs and hexamer 125 by liquid chromatography. Oligothiophene 126 is moll-') allowing optical and electrochemical measurereasonably soluble (3 x ments in solution. The absorption maximum is slightly red-shifted A(,, = 465 nm) with respect to undecamer 122. A correlation of the absorption data of the different

122

3 Oligothiophenes

oligothiophenes available up to this moment with their inverse chain length revealed an almost linear behaviour. In electrochemical experiments, dodecamer 126 showed less pronounced and rather ill-defined quasi-reversible redox waves. The (estimated) peak potentials in comparison to the corresponding sexithiophene 125 are shifted to negative potentials (EP M 0.33 V, E; M 0.62 V vs. Ag/Ag+). The cyclovoltammograms are stable upon continuous cycling, indicating that no further coupling of the radical cations and dications occurs. The CV of a solid film of duodecithiophene 126 is comparable to that of poly(3-alkylthiophene) films. The redox potentials lie very close to those measured for the material which was produced electrochemically by coupling hexamer 125 (Epa= 0.35 V, Epc = 0.20 V vs. Ag/Ag+). Dodecamer 126 was doped in solution with FeC13. In contrast to sexithiophene 125 which can be oxidized successively to the monocation and the dication, the absorption spectra of 126 showed a one-step oxidation leading to two near-IR peaks located at E M 0.8 and 1.7eV. These values are close to those of electrochemically doped polythiophene. Assisting in situ EPR measurement gave only very weak signals on oxidized dodecamer 126. The authors concluded that in the case of very long oligomers immediately a spinless dication as most stable state is formed, which would correspond to bipolarons in conducting polymers. A conductivity of CT = 5 S cm-' for iodone-doped solid dodecamer 126 is reported. The problem of regio-irregularity in such P-substituted oligothiophenes was first solved with the synthesis of isomerically pure alkyloligothiophenes by Bauerle et al. [159]. The decisive path was the regioselective monobromination of alkyl bi- and terthiophenes as key building blocks for the synthesis of higher oligomers. However, direct bromination of oligothiophenes, even under selective conditions, always gives some dibromo derivatives which cannot be separated on preparative scale [1601. The preparation of monobrominated oligothiophenes was done in a painstaking way either by employing protecting-group techniques [161, 1621 or by the use of the palladium-catalyzed cross-coupling of dibromothiophenes with one equivalent of the Grignard reagent of 2- or 3-bromothiophene catalyzed by Pd(dppf)Cl,. Rossi et al. could obtain monobromobithiophenes in moderate yields after tedious purification [76]. In contrast, bromination of oligothiophenes with the high selective system NBS in DMF resulted in a marked increase in selectivity. Under carefully chosen conditions, a-monobromo bithiophene Br-T2-H 50 and terthiophene Br-T3-H 51 could be isolated in good yields [73]. This technique was now applied to the bromination of 3'-dodecylterthiophene 13 which was obtained analytically pure in 31% yield by the 'Kumada-coupling' of two equivalents of BrMg-T1-H 42 and 2,5-dibromo-3-dodecylthiophene. Reaction of terthiophene 13 with one equivalent NBS at ambient temperature gave the monobromination products 127-128 in 56% and the dibromo compound in only 9% yield. The ratio of the two possible isomers 127 and 128 was determined independently to 90: 10 by 'HNMR and HPLC analysis indicating an unexpectedly high selectivity. Repetitive crystallization afforded the isomer 127 in 97% purity and 26% yield. The definitive structural assignment of bromoterthiophene 127 relied mainly on 2D correlation spectra. The isomerically pure bromo(dodecy1)terthiophene 127 is the key intermediate for the synthesis of the rigidly defined higher oligomers and thus was coupled to the corresponding regioregular didodecylsexithiophene 129 in 33 %

3.2 Synthesis of Oligothiophenes

123

yield [Eq. (52)]. The nickel catalyst was prepared in situ from nickel dichloride, triphenylphosphine and zinc (see above). As expected, the melting point of hexamer 129 (1 10- 1 11 "C) was distinctly higher than that reported for the isomeric mixture (80-84°C) prepared by Wynberg et al.

\ r

-!!Ez?%

&p,& ..+..p

C12H25

C12HZ5

13

127 (26%)

c12H25

128

129 (%)

Characterization of the electronic properties of the regioregular and easily soluble sexithiophene 129 revealed that it is stable even in four different redox states due to its well-defined structure [ 1631. In the cyclovoltammogram reversible waves due to the formation of the radical cation and dication (EP = 0.34V, E i = 0.54V vs. Fc/Fc+), additionally quasireversible waves for the radical anion and dianion (E: = -2.27V, ET = -2.40V vs. Fc/Fc+) are found. The stability of the redox states allowed the examination of their optical properties. The maximal absorption in the neutral state,,A,( = 416nm) is blue-shifted in comparison to the unsubstituted sexithiophene,,A,( = 432 nm) and reflects the steric interaction of the alkyl chains with the conjugated 7r-system. Accordingly, a crystal structure of sexithiophene 129 reveals a twisting of the relevant thiophene rings by 10.8". The transition energies of all the species in different redox states were determined and it was found by variable-temperature measurements that the absorption spectrum of the radical cation is reversibly blue-shifted with decreasing temperature. This phenomenon was also discovered by Miller et al. for a terthiophene radical cation and can be explained by a reversible equilibrium between the monomeric and a dimeric radical cation [Eq. (53)] [164]. The 7r-dimerization of radical cations (and anions) has been described before for other large aromatic radical ions like porphyrins or viologens. It was recognized by both groups that the formation of corresponding dimeric species is complementary to polarons and bipolarons and may be a new alternative to the description of the doping behavior of polythiophenes. 2 (H-T,,-H)+

a

(H-T,,-H)?

(53)

In situ EPR spectra were taken from the paramagnetic sexithiophene radical cation and anion and allowed the determination of the spin density distributions. In accordance with the postulated 7r-dimerization,for the radical cation the intensity of the EPR signal decreases substantially with decreasing temperature and indicates a spin pairing to the 7r-dimer. This fact easily explains the puzzling observation that

124

3 Oligothiophenes

a

Figure 7. STM images (area ca. 8.5 x 5.8 nm) of (a) non-regioregular dialkylsubstituted sexithiophene and (b) the regioregular dialkylsubstituted sexithiophene 129 [ 1651.

the EPR activity of polarons in conducting polymers is only detected at unexpectedly low levels of doping. As mentioned before, for the radical cation of tetraalkylated duodecithiophene 126 no ESR activity could be detected [159] and is now easily explained due to this dimerization. The disproportionation of the radical cations to a dication and a neutral species as well as the formation of a CTbond in the dimer can be excluded from the optical properties. STM-investigations by Rabe and Stabel on adsorbates of a non-regioregular sexithiophene synthesized by Wynberg et al. [ 1561 and the directly comparable regioregular sexithiophene 129 of Bauerle et al. [160] on highly oriented pyrolytic graphite (HOPG) showed for the first time the influence of the defined structure on ordered monolayers [ 1651. The STM images of both materials exhibit similarly ordered structures in which both the sexithiophene backbones and the alkyl chains are oriented in a lamella structure parallel to the graphite surface in such a way that the alkyl side chains can achieve maximal van-der-Waals interactions. The conjugated backbone is given by the white areas and the alkyl side chain by the dark areas (Fig. 7). The difference between the two hexamer monolayers, however, is the spatial demand of the repeating units. For both compounds the area per unit cell contains one molecule. This area is 3.25 nm2 in the case of the non-regioregular sexithiophene and nonuniform spacing is found. In contrast, for the regioselectively substituted hexamer 129 the structure is much more uniform and the demand of area is much smaller (2.38 nm2). More insight into the structural pecularities of P-substituted oligothiophenes was given by Hadziioannou et al. who synthesized two analogous stereoregular dialkysubstituted sexithiophenes 134 and 135. Their X-ray structures were determined more precisely than before [ 1661. The more straightforward synthesis started with the symmetric dialkyl quaterthiophenes 131 which were obtained in 47-66% yield by the nickel-catalyzed cross-coupling of the Grignard reagents of 3-alkyl-2-iodothiophene 130 and Br-T2-Br 46. Bromination of quaterthiophenes 131 with NBS in chloroform/acetic acid provided the dibromo quaterthiophenes 132 in 56-79% yield and were further coupled under ‘Kumada conditions’ with the Grignard reagent of 2-bromo- or 5-trimethylsilyl-2-bromothiophene133,

3.2 Synthesis of Oligothiophenes

125

respectively, to the regioregular sexithiophenes 134 and 135 which differ by the length of the alkyl side chain and the terminal end group (25-59% yield) [Eq. (54)]. ,R

R

R

R‘

134 [MiR=C,H, . (59%)] 135 IkSiMe,. !4=C8H,, (25%)]

X-ray data of oligothiophenes are normally scarce due to the difficulties in crystallization. In the case of bis(trimethylsily1)substituted hexamer 135, obviously the end-substitution helps in the crystallization process. Thus, good refinements could be obtained due to the lower rotational disorder in the molecules. In accordance with sexithiophene 129, butylated sexithiophene 134 shows torsional angles of 9 and 11’, respectively, for the thiophene rings bearing the alkyl chain and the connected inner and outer rings. In the case of the octyl-substituted hexamer 135 the deviation from planarity is smaller (5 and 7”). The alkyl chains have a planar zig-zag conformation and are lying almost in the plane of the thiophene backbone. In 1994, Sat0 et al. could increase the chain length of a-oligothiophenes up to 15 thiophene rings in the correct connectivity [167]. A homologous series of oligothiophenes bearing now two alkyl side chains per three thlophene units was synthesized through nickel-catalyzed reductive coupling of 5,5”-dibromo-3,3”-dihexylterthiophene 137 which was obtained by bromination of the corresponding dihexylterthiophene 136 in 95% yield. Due to the bifunctionality of the reagent, coupling with the system NiC12/Zn/PPh3/DMF/2,2’-bipyridine yields a mixture of higher oligomers [Eq. (55)]. The yields of the isolated hexyl-substituted oligothiophenes 138-141 are therefore low. Separation from higher molecular weight products could be achieved through liquid chromatography and afforded hexamer 138 in 1.8%, nonamer 139 in 2.8%, dodecamer 140 in 1.4%, and pentadecamer 141 in

126

3 Oligothiophenes

1.8%. GPC suggested that the oligomers are nearly monodisperse (M,/M, = 1.05-1.13). Due to their solubility, their structure could be proven by NMR spectroscopy.

137 (95%)

136

R = CSH,3 --

M,, m.DMF H

H n

(55)

138 ( ~ 2 1.8%) ; 139 (nt3 28%) 140 (-4: 1.4%) 141 (-5; 1.8%)

138-141

In the absorption spectra the T-T* transitions are gradually shfted to lower energies as the chain length increases [138 (A = 410nm), 139 (A = 440nm), 140 (A = 448 nm), 141 (A = 456 nm)]. The conjugation length of the pentadecamer 141 must thus be similar to poly(3,3/’-dihexylterthiophene)which exhibits the same absorption maximum. The absorption maxima of solid films also show a decrease of the transition energy with increasing chain length, but the maximum of pentadecamer 141 is conversely located at higher energies [138 (A = 408 nm), 139 (A = 456 nm), 140 (A = 494 nm), 141 (A = 448 nm)]. The authors conclude from this finding that the oligomers up to duodecithiophene 140 have coplanar structures whereas the structure of the longest oligothiophene 141 should be coiled, which was once proposed for polythiophene [ 1681. The W/VIS/NIR and ESR spectra of the corresponding cations were investigated by controlled oxidation with FeC13 [167a,c]. By adding two equivalents of oxidant the optical and ESR spectra showed that the oxidized species were radical cations and the charge seems to be spread over 6 to 8 thiophene units. The absorption spectra were similar to those of doped poly(3-alkylthiophenes) bearing long alkyl chains. Also in these cases, relatively low spin concentrations are detected in ESR experiments which decrease with increasing chain length. The low number of spins per molecule and the simultaneous change in the absorption spectra point also to the formation of diamagnetic -/r-dimers.The further oxidation of hexamer 138 led to absorption features which were interpreted as being due to the formation of a radical trication instead of the usual dication. The longest oligothiophene so far synthesized as a part of a homologous row of isomerically pure a-linked oligothiophenes 143-146 was recently reported by Bauerle et al. [169]. According to calculations, sedecithiophene 146 should be 64 long when extended. 3,3/”-Didodecylquaterthiophene143 is the main building block and was obtained in 75% yield by the ‘Kumada-coupling’ of two equivalents of the Grignard reagent of 2-bromo-3-dodecylthiophene 142 with Br-T2-Br 46. By oxidative coupling of the lithiated tetramer 143 with CuC12, the corresponding octithiophene 144 and duodecithiophene 145 were obtained in one step due to the formation of dilithiated products. Owing to their good solubility, 144 and 145

A

3.2 Synthesis of Oligothiophenes

127

could be obtained in analytically pure form in isolated yields of 30% and 8%, respectively, by repeated chromatography. Octamer 144 was then coupled again with the system n-BuLi/CuC12 to give the corresponding hexadecamer 146 in 19% yield of pure material after chromatographic work-up [Eq. (56)]. Nl(@iW2

H25c&

B

+

y

+

46

2 6%

&

,

S

C,,H2,

142

i

143 (75%)

n-Buuicfi,

i

146 (19%)

The melting points of the very stable oligothiophenes 143-146 increase with increasing chain length but lie markedly below those of the unsubstituted oligothiophenes. The alkyl side chains lead to increased steric interaction between the conjugated backbones and thus reduce K-K interactions between the molecules in the solid state. Additionally, the solubility is raised by the presence of the side chains in a way that all the homologs are soluble in hexane, aromatic and chlorinated hydrocarbons and could be fully characterized. Absorption spectroscopy of 143-146 showed that, as expected, the energy of the longest wavelength T-T* transition is red-shifted and the extinction coefficient linearly increased with increasing chain length. While a good correlation is normally found between the absorption energy and the reciprocal of the chain length for other oligothiophene series up to a chain length of seven [ 13, 1701, the relationship in this series is not linear. The longer members of the series, oligomers 145 and 146, deviate from linearity such that the absorption maxima are shifted to higher energies than expected. It is not clear at the moment, if this is due to a winding of the long thiophene backbone which would cause a reduction of the effective conjugation or if the effect observed is a general phenomenon and a saturation is approached. Theory predicts a linear behavior up to infinite chain lengths [ 1711. Temperature-dependent

128

3 Oligothiophenes

spectroelectrochemical investigations on this series showed the tendency that the radical cations dimerize more strongly with increasing chain length. Whereas the monomeric and the dimeric radical cations are in equilibrium at approximately -40°C for quaterthiophene 143, equivalent portions are formed at ~ 3 0 ° C for octithiophene 144, at e65"C for duodecithiophene 145, and even >80"C for sedecithiophene 146. Since oligothophenes are ideal model compounds for investigating the charge carriers responsible for the charge transport in conjugated polymers, the redox properties are of special interest. As expected, the first oxidation potential for oligomers 143-146 is shifted cathodically with increasing chain length. The oxidation potentials of dodecamer 145 (Epa= 0.19 V vs Fc/Fc+) and hexadecamer 146 (Epa= 0.12 V vs Fc/Fc+) are more negative than that of the structurally related poly(3-dodecyl-2,2'-bithiophene) (Epa= 0.3-0.35 V vs Fc/Fc+) [169]. This clearly indicates that longer oligomers exceed the average conjugated chain length of the parent polythiophenes which typically lie at 8- 10 properly linked monomer units. In multisweep experiments of all homologs, thin films are deposited on the working electrode. The films exhibit sharp and symmetric redox waves similar to the electrochemical response of redox polymers. Their first oxidation potential is once again at considerably lower potentials than those of the parent oligothiophenes (Epa= -0.1 1 to -0.13 V vs Fc/Fc+) and indicate the presence of fairly well-defined higher oligomerization products. Evidently, a preorganization of the molecules takes place on the metal surface and well-ordered products are formed by oxidation. An investigation of physisorbed monomolecular layers of these oligothiophenes on graphite by STM high resolution images revealed that each compound forms ordered supramolecular structures in which the molecules are oriented in lamellae (Fig. 8). The bright areas are due to the conjugated .Ir-systemand the darker regions to the alkyl chains. In all four cases the lamellae are separated by roughly 2.2 nm and the degree of overlap between two neighboring thiophene chains decreases from 35" for quaterthiophene 143 to 5" for sedecithiophene 146. Finally, for 146 the molecules are lined up almost linear and are oriented ideally for further oxidative coupling of the adjacent a-positions. This fact may explain the formation of extremely long 'molecular wires' in electrochemically prepared thin films.

Oligothiophenes synthesized for subsequent polymerization Since there are many examples of @-substitutedbi- and terthophenes only those which served for subsequent polymerization to the corresponding poly(alky1thiophenes) and functionalized poly(alky1thiophenes)will be listed. Oligomers including different substitution patterns and varying alkyl substituents up to sexithophenes have been synthesized in this context and are noted in Tables 3-7. In most cases, the oligomers were obtained by transition metal catalyzed cross-coupling reactions between organometallic reagents and thienyl halides (see section 3.2). Optical, electrochemical and X-ray diffraction characterizations of various substituted oligomers provide insight into the influence of the substituents on structural features and consequently on the electronic properties. Despite the alkyl groups impose a weak electron-donating effect on the conjugated 7r-system, e.g. with respect to H-T2-H 2, the absorption maxima of 3- and 3,3'-substituted bithiophenes are blue-shifted and the redox potentials are increased. This is due

3.2 Synthesis of Oligothiophenes

b

a

129

H 2 nm

H 1 nm

d C

H

H 2 nm

2 nm

Figure 8. STM images of 2D crystals of the whole series of oligothiophenes on HOPG. (a) Tetramer 143; (b) octamer 144; (c) dodecamer 145; (d) hexadecamer 146 [169].

to a deviation of the thiophene rings from coplanarity which causes a diminution of the 7r-orbital overlap (see e.g. 2,147,155).X-ray structure data indeed show that the conformational properties in the solid state are different from those in solution, but they are mainly determined by the system to relieve conformational strain which is dispersed over the whole molecule by succesive and cooperative bond and angle deformation [192]. Following the development in polythiophene chemistry and the development of regioregular poly(3-alkylthiophenes) [ 152- 1541, the first examples of regioregular oligo(3-alkylthiophenes) were very recently reported. The inherent problem in the synthesis of this model compounds is that isomer-free building blocks with a 2,4substitution pattern are necessary. However, 3,4”4”-trialkyl-2,2’:5’,”’-terthiophenes 193, 196, 200 are the longest regioregular oligo(3-alkylthiophenes) reported so far. Barbarella et al. succeeded in the synthesis and characterization of the corresponding methyl [I891 and hexyl derivatives [177, 1781. In order to get insight into the structural pecularities of poly(3-hexylthiophene) all the other possible regioisomeric terthiophenes 197-199 were also synthesized and the properties compared. The methyl-substituted regioregular bithiophene 162 and terthiophene 193 were obtained by successive ‘Kumada’ coupling reactions. However, the coupling of the corresponding brominated bithiophene and the Grignard reagent of 2-bromo4-methylthiophene affords the regioregular terthiophene 193 in only 20% yield [Eq. (57)]. The authors give for terthiophene 193 an (isomeric) purity of >98%. This compound was synthesized before by Zimmer et al. [173], but its melting

66 86 88 76 30 69 95 53 85 23 75

2 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170

3,3’-Diethyl-2,2’-bithiophene 3,3’-DihexyL2,2’-bithiophene 3-Ethyl-3’-methyl-2,2’-bithiophene 3,3‘-Di-[(2-tetrahydropyranyloxy)ethyl]-2,2’-bit~ophene

3,4’-DihexyL2,2’-bithiophene 3,4‘-Didodecyl-2,2’-bithiophene

3,4’-Di-[(2-tetrahydropyranyloxy)ethyl]-2,2’-bithiophene 3,4’-Di-[(2-hydroxyethyl]-2,2’-bithiophene 3,4’-Di-[6-( p-methoxyphenoxy)hexyl]-2,2’-bithiophene 4,4’-DimethyL2,2’-bithiophene

4,4’-DihexyL2,2’-bithiophene 3,3’,4,4’-Tetramethyl-2,2’-bithiophene

3,3’-Di-(2-sulfonatoethyl)-2,2’-bithiophene-sodium salt 3,4’-Dimethyl-2,2’-bithiophene

3,3’-Di-(2-hydroxyethyl)-2,2’-bithiophene

3-Ethyl-2,2’-bi thiophene 3-Hexyl-2,2‘-bithiophene 3-Dodecyl-2,2’-bithiophene 4-Methyl-2,2’-bithiophene 6-(2,2’-bithien-3-yl)hexanoicacid N-[6-(2,2’-bithien-3-yl)hexanoyloxy]pyrro~din-2,5-dion 2-[2-0xa-7-(2,2’-bithien-3-yl)heptyl]1,4,7,1O-tetraoxacyclododecane 3,3’-Dimethyl-2,2’-bithiophene

3-Methyl-2,2‘-bithiophene

6

75

50

70

50

80

81 72 89 89 85 11 70

[%]

2,2’-Bithiophene

Yield

Compound number

a-bithiophenes.

Bithiophenes

Table 3. Preparation and physical properties of ,&substituted

310 311 310

292

302 299 298

250 242

295 270 268 279 270

309

302 299 295

Absorption A,, [nm] in CHC13

Ref.

77,170,172,173 170,172,173 170 174 73 1.20 77 1.67 175 0.88(FC/FC+) 175 1.23 176 1.51,1.17 77,170,172,173 1.33 (SCE) 177 1.87 77 1.53(SCE) 177,178,179 172 180 180 180 1.13 77,172,173 1.22(SCE) 177 177,179 1.35(SCE) 181 180 180 182 77,170,172,173 1.10 171 1.21(SCE) 177 1.25(SCE) 148

1.24 1.20 1.75

Epa[VI vs. Ag/AgCl

Oxidation potential

3

2

$-

g2

?.

0

kl

0

w

c

3.2 Synthesis of’ Oligothiophenes

131

point (39°C) is considerably lower than this obtained by Barbarella et al. (61”C), indicating a far lower isomeric purity in the former case. X-ray structure determination of regioregular terthiophene 193 revealed two crystallographically independent, but identical molecules, characterized by torsional angles of 6-9” around the inter-ring bond. Interestingly, the molecule crystallizes in a chiral space group containing one single enantiomer. Due to the defined head-totail linkages of the thiophene rings in 193 and the resulting non-planarity, a pair of rod-like equienergetic enantioconformers or atropisomers with +w and -w interring twist angles exists. Single atropisomers cannot be observed in solution, since the rotational barriers around the inter-ring bond are too small. 3,4’,4’’-Trihexy1-2,2’:5’,”’-terthiophene 196 was similarly prepared, but by ‘Stilletype’ cross-coupling reactions [177]. In this reaction sequence, the less effective step is the Pd(0)-catalyzed coupling of the organotin thiophene and 2-bromo-3-hex163 in only 23%. Bromination ylthiophene to yield 3,4’-dihexyl-2,2‘-bithiophene of the latter and reaction with the stannylated thiophene under ‘Stille-conditions’ results in 69% yield of the desired regioregular terthiophene 196 [Eq. (57)].

R = Me.C,H13. C H ,.l,

(CH,),OTHP,

(CH2)20H, (CH,),0C,H,0CH3:

M = BrMg, Me3% B(OH),

(57)

n

R Me. CsH13. (CH,),OC,H,OCH,; I

R

M = BrMg, Me&

193,196,200

While there is only one report on regioregular poly(3-alkylthiophenes) functionalized with polar groups [198], the synthesis of corresponding oligomers was attempted recently [1801. Analogous ‘Stille-type’ reaction of the related tetrahydropyranyl(THP)-protected derivatives of 3-(2-hydroxyethyl)thiophene revealed another problem not mentioned before. A mixture of the desired cross-coupled 3,4’-disubstituted bithiophene 165 (50%) and the 4,4’-substituted regioisomer (lo%), which is formed due to homo-coupling reactions, is obtained. The authors were not able to separate the regioisomers completely, but finally the deprotected hydroxyethyl-substituted derivative 166 could be purified by chromatography. Regioregular bi- and terthiophenes which bear hydroquinonemethylether(HCM)-protected alkyl side chains were synthesized by Crayston et al. using successive ‘Kumada’ coupling reactions [182]. Ether cleavage of the HCM-group directly results in the corresponding bromo derivatives whch can be replaced by a great variety of functional groups [86]. Although the synthesis of the 3,4’-disubstituted dimer 167 succeeds in moderate yield (50%), the successive coupling of the

336 345

80 53 77 60 92 58 (26) 69 53 68 23 42 33 60 172 173 174 175 176 13 177 178 179 180 181 182 183

3’-Methyl-2,2’: 5’,2”-terthiophene 3’-Ethyl-2,2’: 5’,2”-terthiophene 3’-Butyl-2,2’: 5’,2”-terthiophene 3’-0ctyl-2,2‘:5’,2”-terthiophene 3’-Decyl-2,2’:5‘,2”-terthiophene 3’-Dodecyl-2,2’:5’,2”-terthiophene 3’-Phenyl-2,2’: 5’,2”-terthiophene 3’-(pMethoxyphenyl)-2,2‘: 5‘,2”-terthiophene 3’-( p-Cyanophenyl)-2,2’:5’,2‘’-terthiophene 3’-(Thien-2-~1)-2,2’: 5’,2’’-terthiophene 3’-(5-Methylthien-2-~1)-2,2‘: 5’,2”-terthiophene 3’-(Pyrid-4-~1)-2,2’: 5’,2”-terthiophene 3’-(3,6-Dioxaheptyl)-2,2’: 5’,2”-terthiophene

341 345 346 (MeCN) 350 (MeCN) 340 (MeCN) 345 (MeCN) 345 (MeCN) 344 (MeCN)

352

70

171

2,2’: 5’,2”-Terthiophene

3-Methyl-2,2’: 5’,2’’-terthiophene

Absorption A, [nm] in CHC13

355

Yield [%]

86

Compound number

3

~~

Terthiophenes

Table 4. Preparation and physical properties of P-substituted a-terthiophenes.

1.05 (SCE)

1.06 (SCE)

1.05 (SCE) 1.05 (SCE)

1.05 0.98 (SCE) 1.13 1.03 (SCE) 1.11

Oxidation potential EpaM vs. Ag/AgCl

77, 172 183 77, 172 183, 184 77, 172, 184 71 156, 183 184 158 156, 73 184, 185 185 185 185 185 185 184

Ref.

2

$-

e

3

2 s?

cu

76

35 69 51 24 27 26

194 195 196 197 198 199 200 201 202

3,4’,3’’-Trirnethyl-2,2’: 5’,”‘-terthiophene 3,3’,3’’-Trimethyl-2,2’: 5’,2’’-terthiophene 4’-Ethyl-3,”’-dimethyl-2,2’: 5’,2’’-terthiophene

3,4’,4’’-Trihexyl-2,2’: 5’,2”-terthiophene 4,4’,3’’-Trihexyl-2,2’: 5’,2’’-terthiophene 3,4’,3’’-Trihexyl-2,2’: 5‘,2”-terthiophene 4,4’,4’’-Trihexyl-2,2’: 5’,2’’-terthiophene 3,4‘,4‘’-Tris-[6-( p-methoxyphenoxy)hexyl]-2,2’:5’,”’-terthiophene 3,4,3”,4”-Tetramethyl-2,2‘: 5‘,”’-terthiophene 3,4,3’,4’,3’’,4”-Hexamethyl-2,2’: 5’,2”-terthiophene

20

193

3,4’,4’’-Trimethyl-2,2’: 5’,”’-terthiophene

87 67 48 95 78

184 185 186 187 188 189 190 191 192

3’-(3-Sulfonatopropy1)-2,2’: 5‘,2’’-terthiophene-potassiumsalt 6-(2,2’: 5’,”’-Terthien-3’-yI)hexanoic acid N-[6-(2,2‘5’,2’’-terthienthien-3’-yl)hexanoyloxy]pyrrolidin-2,~-dion 2-[2-0xa-7-(2,2’:5‘,2”-terthien-3’-yl)heptyl]-1,4,7,10-tetraoxacyclododecane 3,3’-Dimethyl-2,2’:5‘,2”-terthiophene 3,3”-Dimethyl-2,2’:5’,2’’-terthiophene 3’,4’-Dimethyl-2,2’:j’,”’-terthiophene 3’,4’-Dibutyl-2,2’:5’,2’‘-terthiophene 2,5-Di(thien-2-yl)-cyclopenta[c]thiophene 0.93 0.85 (SCE) 0.98

0.88 (SCE) 0.88 (SCE) 0.88 (SCE) 0.88 (SCE)

321 336 326 316 348 354

0.97 (SCE)

1.09 1.03 (SCE)

0.72 (Fco/+) 0.90

324 346 324

350

344

346

0.90 (SCE)

173,178,189 177 77 148 77 177, 178 177, 178 171, 178 177, 178 182 148 148

183 175 175 176 149 77, 186 148, 183 187, 188 183

Compound number

4 203 204 205 206 207 208

Quaterthiophenes

2,2’: 5‘,2’‘:5”,2‘”-Quaterthiophene 4’,3’’-Dimethyl-2,2‘: 5’,2‘’:5”,2”’-quaterthiophene 3,3”-Dimethyl-2,2’: 5’,2’‘:5”,2”’-quaterthiophene 4’,3”-Dimethyl-2,2’: 5’,2”:5”,2”’-quaterthiophene 3,4’,3’’,3’”-Tetramethyl-2,2’: 5’,2‘‘:5”,2”’-quaterthiophene 3,3’,4’’,3’’-Tetrarnethyl-2,2‘: 5’,2”:5”,2”’q~aterthiophene 4,4’,3‘’”’’’-Tetramethyl-2,2’: 5’,2”:5”,2’”-quaterthiophene

Table 5. Preparation and physical properties of P-substituteda-quaterthiophenes.

9

70

89 19 80

[%I

Yield

390 348 380 346 342 346 348

Absorption A,, [nm] in CHC13

0.90

1.05

EpaM vs. Ag/AgCl

Oxidation potential

77,190 190,192 77,173,190,192 190,192 191

77

77

Ref.

c)

P

W

212

3”,4”-Dibutyl-2,2’:5’,2”:5”,2”’:5”’,2””-quinquethiophene

5

209 210 211

Compound number

2,2‘: 5’,2“:5“,2“‘:5”’,2””-Quinquethiophene 3,3””-Dimethyl-2,2’:5’,2”:5”,2”’:5”’,2””-quinquethiophene 3’,4”’-Dirnethyl-2,2’:5‘,2“:5“,2“‘:5“‘,2““-quinquethophene 3,3’,4’”,3’”-Tetramethyl-2,2’:5’,2”: 5”,2”’:5”’,2’”-quinquethiophene

Quinquethiophenes

Table 6. Preparation and physical properties of P-substituted a-quinquethiophenes.

91

[%]

Yield

416

Absorption A,, [nm] in CHC13

0.97 0.80 (SCE) 0.80 (SCE) 0.86 (SCE)

Oxidation potential Epa[Vl vs. Ag/AgCl

194

77 193 193 193

Ref.

VI

w

c

h

C’

6

95

2

3’(4’”3’’’’(4’’‘’)-DioctyI-2,2’: 5’,2’’: 5”””’: 5”’,2’r”:5””,2””’-sexithiophene(irreg.) 3’(4’),3””“’’’’)-Di-(3,6-oxaheptyl)-2,2‘:5’,2’’: 5”,2”’:5’”,2”’: 5””,2””’-sexithiophene (irreg.) 3’,4’,3’’‘’,4’’’’-Tetrabuty1-2,2’: 5’,2‘’: 5”,2”’:5’”,2”’: 5‘”’,2”’-sexithiophene 3’,4’,3’‘‘’,4’’rr-Tetrahexyl-2,2‘: 5’,2’’: 5”,2”’:5”’,2””:5””,2”’”-sexithiophene 3,3’,4~‘~3’”,4rr’’,3’’’r-He~amethy1-2,2’: 5’,2”: 5”,2”’:5”’,2’’”:5n”,2‘’”r-sexithiophene 4,4~,3”,4”’~3”,4’’’’-Hexamethyl-2,2’: 5‘,2’’: 5”,2”’:5”’,2”’’:5”“,2””-sexithiophene

3’(4’)”’’’’(4””)-Dimethyl-2,2’: 5’,”’: 5”,2”’:5”’,2”’’:5””,2””’-sexithiophene(irreg.)

2,2’: 5’,2’’: 5”,2”’:5’”,2””:5””,2’‘”’-Sexithiophene

Sexithiophenes

Table 7. Preparation and physical properties of P-substituted a-sexithiophenes.

20 8

56

6 213 214 215 216 217 218 219

Yield [%]

Compound number

368 368

432

Absorption ,A, [nm] in CHCl3

(0.46) 1.06 (SCE) 0.98 (SCE) 0.99 (SCE)

Oxidation potential Epa[VI vs. Ag/AgCI

194 196 197 197

77 195 195 195

Ref.

2

3

9

2E.?

3

0

LJ

3.2 Synthesis of Oligothiophenes

137

brominated dimer leads in only 26% yield to the regioregular terthiophene 200. Also in this case, a minor amount of homo-coupling product (4,4’-disubstituted bithiophene) was detected. The synthesis of longer regioregular oligothiophenes would be extremely interesting with respect to their model character and their electronic properties in their own right. Evidently, this demands more rigidly defined regioselective and effective synthetic steps with a minimum formation of regioisomeric homo-coupling products. In this respect, the ‘Suzuki-type’coupling reaction of 3-dodecyl-5-thienylboronic acid 220 and 2-bromo-3-dodecylthiophene 142 proved to be more successful [181]. Since the bororganic component can be obtained pure and free of isomers by recrystallization, optimized coupling with [Pd(PPh3),] as catalyst in DME/NaHC03 affords 75% dimeric products in which the desired 3,4’-didodecyl bithiophene 164 is preferably formed in a ratio of 28 :1 (detected by ‘H NMR spectroscopy) in comparison to the corresponding 4,4’-isomer 221 [Eq. (58)]. Pure 3,4’-didodecyl-2,2’-bithiophene 164 was effectively separated from the 4,d-disubstituted homo-coupling product 221 by preparative HPLC.

Bridged oligothiophenes A strategy for controlling regioregularity, planarity and rigidity simultaneously is best realized in ,6,/3’-bridged oligothiophenes including a fixed conformation. Thus, cyclopentadithiophenes originally synthesized by Wynberg et al. as thiophene analogs of fluorene [199] have recently received much attention. In order to prove that the electronic properties of alkylated ologithiophenes are dependent on steric hindrance, Zimmer et al. synthesized and reinvestigated the rigidified 3,3‘-bridged bithiophenes 222 and 223 in which the thiophene rings are kept in a syn-cisconformation [ 1701. However, the synthesis of this type of compounds is tedious and includes many steps. In comparison to 3,3’-dimethyl-2,2’-bithiophene 155 (A, = 270 nm; Epa= 1.51 [ 1.171 V vs Ag/AgCl), the maximum absorption of the fully planar 4H-~yclopenta[2,1 -b;3,4-b’]dithiophene222 is considerably red-shifted , ,A,( = 31 1 nm) and exhibits the by far lowest oxidation potential in the alkylbithiophene series (Epa= 0.97 V). The less rigid analog 4H,SH-~yclohexa[2,1b;3,4-b1]dithiophene 223 exhibits still a rather long-wavelength absorption ,,A,( = 305 nm) and a somewhat higher oxidation potential (Epa= 1.20V). Zerbi et al. investigated this series including the homologous cycloheptadithiophene derivative 224 and their corresponding polymers by means of vibrational spectra [200]. The interpretation of the spectra revealed largest .rr-electron delocalization for the fully planar dithienocyclopentadiene system. A series of 4-alkyl- 225 and 4,4’-dialkyl-substituted cyclopentadithiophenes 226 synthesized and polymerized by Berlin and Zotti et al. resulted in soluble and highly conductive polymers. The partial rigidity of the polythiophene backbone causes anomalously red-shifted absorption spectra in the neutral state which

138

3 Oligothiophenes

indicates a high conjugation (A, = 545-680 nm). The monomers are obtained by one- or two-fold deprotonation of the parent bridged bithiophene with n-BuLi and subsequent alkylation with alkyl halides in 3 1-47% yield [201].X-ray structure analyses of 4H-cyclopenta[2, l -b;3,4-b’]dithiophene222 and the spiro-analog spiro[4Hcyclopenta[2,1-b;3,4-b1]dithiophene-4,1’-cyclopentane] 227 [ 1861 confirm the fully planar arrangement of the bridged bithiophenes [202].

fi S

S

& S

S

222-224 (n = 1-3)

As a further model compound for rigidified polythiophene, Roncali et al. have developed a new synthesis of the first fully planar terthiophene by bridging the internal P-positions [203].The synthesis is based on the cyclic ketone 228 which is oxidatively dimerized after deprotonation by CuClz to the 1,Cdiketone 229 (30% yield). Intramolecular cyclization of 229 with L.R. affords the rigid 4H,SH-dicyclopenta[2,1-b;3,4-b‘;2’1’-a’;3’,4’-bN]terthiophene 230 in 45% yield. The rigidification induces considerable changes in the optical and chemical properties and leads to a much smaller HOMO/LUMO-gap. Thus, the absorption spectrum of 230 differs from that of H-T3-H 3 by a fine structure typical of rigid conjugated systems and by a bathochromic shift of the absorption maximum of Ax,,, = 22 nm. A shoulder at X = 387 nm and a weak absorption tail proceeding up to about 850 nm confirms the considerable extension of the effective conjugation in 230.Simultaneously, the oxidation potential of rigid terthiophene 230 is shifted negatively to Epa= 0.60 V (H-T3-H 3: Epa=. 1.07V). Surprisingly, rigid terthiophene 230 also can be reduced relatively easily (Epc= -0.75V). The estimated HOMO/LUMO-gap is diminished from A E % 3.20eV for H-T3-H 3 to A E M 1.65eV for the rigid analog 230.

228

(30%)

230 (45%)

A further diminution of the HOMO/LUMO-gap, respectively the band-gap, in rigid bithiophenes and the corresponding polymers is found when electronwithdrawing groups are linked to the bridging carbon. This is verified in cyclopenta[2,l-b;3,4-b’]dithiophene-4-one] 231 where the carbonyl group should increase the quinoid character of the oligomer [204]. In this compound the longest wavelength absorption is reasonably red-shifted (AX = 161 nm) compared to the parent compound 222.The oxidation potential of ketone 231 is somewhat higher (AE,, = 0.28 V) than that of cyclopentabithiophene 222, indicating that the withdrawing carbonyl group has a moderate effect on the energy level of the HOMO.

3.2 Synthesis of Oligothiophenes

139

Several other substituted bridged bithiophenes 232-240 are in fact derivatives of ketone 231. Thus, the dioxolane 232 and the dithioacetale 233 have been synthesized by Roncali et al. [205]. Furthermore by condensation reaction of thienocyclopentanone 231 and malonic acid derivatives, the dicyanomethylene and cyano(nonafluorbuty1)sulfonyl-substituted derivatives 234 [206] and 235 [207], respectively, were recently described. In both compounds again the LUMO energy is lowered and the absorption red-shifted by about (AA = 100 nm) compared to ketone 231. Zotti, Berlin et al. synthesized similarly p-nitrobenzyl-, p-nitrobenzylidene-, 4-pyridyl-, and 4-(N-methylpyridinium)-substituted cyclopentadithiophenes 236-239 [208]. In these cases, the substituents cause a decrease in oxidation potential by AE = 0.25-0.3OV compared to the parent compound 222 [209]. Also starting from thienocyclopentanone 231, Roncali et al. synthesized via Wittig-Horner and Wittig olefination with phosphonate esters or phosphonium salts, respectively, a series of bridged bithiophenes 240 including a 1.3-dithiole moiety [210]. Here also the oxidation potential is decreased by AE = 0.1-0.3V and, , ,A red-shifted by 90 nm.

S 231

232 (R.R = -O(CH&O-) 233 (R,R = -S(CH&3-) 236 (R = kk R' = CHzCshNOz) 238 (R = H; R = CH,C&,N)

S

234 (R,R' = CN) 235 (R = CN: R' = SO,C+d 237 (R H; R' = C,H,NO,) 239 (R = H R' = C&i,NMe* CFSOj) 240 (R,R' = -S(R'C=CR')S)

The above mentioned examples prove that besides other strategies the rigidification of conjugated systems leads to a decrease of the HOMO/LUMO-gap and the extension of r-conjugation. However, new synthetic strategies seem to be necessary in order to develop longer rigid oligothiophenes or a totally planar (super)polythiophene.

3.2.2.2 a,a'-Substituted Oligothiophenes The different series of /?-substituted oligothiophenes described above showed clearly that alkyl side chains in /?-position lead to a strong increase of the solubility properties, particularly for the longer oligomers. The investigation of the electronic structure of different redox states as models for polarons and bipolarons in solution is nevertheless only possible for longer members (n 2 5), since radical ions of shorter oligomers tend inherently to dimerize or to oligomerize. Therefore, the introduction of solvating substituents at the reactive terminal a,a'-positions of the oligothiophenes should increase the stability of shorter members in the oxidized (and reduced) state and facilitate their investigation. Series of a-alkyl and a,a'-dialkyl-substituted oligothiophenes were synthesized and characterized by different research groups. Especially, monosubstituted derivatives are attractive candidates since they offer the possibility of dimerizing them to

140

3 oligothiophenes

the corresponding a,a’-disubstituted oligothiophenes with doubled conjugated chain length. However, due to the electron richness of the oligothiophene system they are difficult to obtain selectively. Zotti et al. synthesized a-methyl derivatives up to the pentamer by using Pd-catalyzed Grignard cross-coupling reactions 242 obtained from 5-methyl-2,2’[211]. Thus, 5-bromo-5’-methyl-2,2’-bithiophene bithiophene 241 was reacted with BrMg-Tl -H 42 to afford 5-methylterthiophene 243 in 93% yield. Bromination of Me-T3-H 243 with NBS led effectively to the 5bromo-5”-methylterthiophene244 (92%). Successive coupling with the same Grignard reagent led to 5-methylquaterthiophene 245 in 87% yield on one hand and with the reaction with BrMg-T2-H 49 to the corresponding pentamer Me-T5-H 246 in moderate 20% yield, on the other hand [Eq. (59)].

242

42

243 (93%)

244 (92%)

245 (87%)

246 (20%)

Zotti et al. then coupled oxidatively the singly blocked a-methyl oligothiophenes Me-T,-H 241, 243, 245, 246 to the corresponding a,a‘-disubstituted oligothiophenes Me-T2,-Me 247-250 [Eq. (60)]. The materials with doubled length up to the decamer were found as solids on the working electrode. From in situ EPR and conductivity measurements as a function of the potential it was concluded that the highest conductivity is obtained in a ‘mixed-valence’ state in which the oligothiophenes are partly in a radical cationic and a dicationic state [211].

241,243,245,246 (ne-5)

247-250 (n=2-5)

In an analogous manner, Bauerle et al. synthesized corresponding a-dodecyl- and a,a’-didodecyl oligothiophenes [73]. In order to obtain the monoalkylated derivatives, first the selective synthesis of monobrominated oligothiophenes in high yield and purity was performed. Due to the always present formation of disubstituted products which are difficult to separate, the mono-functionalization of oligothiophenes is an inherent problem. This was solved in this case by the use of the mild and selective brominating system NBS/DMF and by the careful choice

3.2 Synthesis of Oligothiophenes

141

of the reaction conditions in order to suppress the formation of the dibrominating products. Thus, Br-T2-H 50 and Br-T3-H 51 were isolated in pure form in 70% and 86% yield, respectively, starting from the unsubstituted H-T2-H 2 and H-T3-H 3 [Eq. (61)]. Normally, this type of compound has been synthesized by indirect methods and in moderate yields [162]. The monobromination of H-T4-H 4 was however problematic, since the dibrominated product Br-T4-Br 48 is instantaneously formed. In the same way, the dibromination of these oligothiophenes was performed with two equivalents of NBS in DMF and results in the a,a’dibromooligothiophenes Br-T,-Br 46-48 in 76-88% yield [Eq. (61)]. Br

m T

46,47,48

H

(n= 2-4)

M

2,3,4

nH

(n= 2-4)

NBS DMF

.#j$

(611

50,51 ( n = 2 . 3 )

The a-monoalkylated oligothiophenes 252, 254, 255 were obtained in 59-87% yield by ‘Kumada-coupling’ of BrMg-T1 -H 42 and 2-bromo-5-dodecylthiophene 251 [Eq. (62)] or of the Grignard reagent of the latter with Br-T2-H 50 and Br-T3-H 51, respectively [Eq. (62)].

50,51

(n=2.3)

253

254,255

(n= 2.3)

The same reaction of two equivalents of 5-dodecyl-2-thienylmagnesium bromide 253 with Br-TI-Br 45 and Br-T2-Br 46 led to the corresponding a,a’-dialkylated terthiophene 256 and quaterthiophene 257 in 81% and 70% yield, respectively [Eq. (64)]. a,@’-Didodecylsexithiophene 258 was prepared by oxidative coupling of lithiated terthiophene 16 with CuCI2 [Eq. (65)] [72].

253

51

45# (n= 1,2)

16

256,257 ( n = I.2)

258

( n = 4)

Hotta et a1 [212] realized a series of corresponding methyl-substituted oligothiophenes up to the hexamer. This homologous row was also synthesized using

142

3 Oligothiophenes

Kumada's aryl/aryl-coupling procedure. Thus, the Grignard reagent of 2-halo-5methylthiophene 260 was reacted with Br-T,-Br 45-48 under Ni(0)-catalysis to yield the a,a'-dimethyl-oligothiophenesMe-T,-Me 261, 262, 247, 263, 248 which were purified through recrystallization from alcohols or ketones (dimer to tetramer) and chlorobenzene (penta- and hexamer) [Eqs. (66), (67)]. Even single crystals could be obtained using the purified materials by further slow recrystallization process. An X-ray structure determination of Me-T4-Me 247 could be performed.

259

260

260

45-48 (n = 1-4)

261

262,247,263,248 (n = 1-4)

Furthermore, the doping of thin films and single crystals, respectively, of these or oligothiophenes with oxidating agents like iodine, nitrosyl salts NO'X-, acceptors like TCNQ was investigated and resulted in conductivities in the range of o = lop2 to lo-' Scm-' [213]. Additionally, the neutral and the doped oligothiophene films showed spectroscopic characteristics in the solid state which were different from those in solution. In the neutral state, as for nonsubstituted oligothiophenes the absorption bands show a fine structure due to vibronic couplings. The lowest and the second lowest energy modes are assigned to the 0-0 and 0-1 transitions, respectively. Their energy separation is about 0.2 eV or 1600cm-' and is attributed to the ring-stretchmg mode in the thiophene rings [214]. In the doped state, the spectra were interpreted by the association of two molecules. For the first time, the same features were observed in the solid state (secondary peaks or shoulders on the high energy side), as was found for the dimerization of radical cations in solution (see above) [215]. With respect to the solid-state properties in organic transistors and light emitting diodes, Garnier et al. synthesized a,a'-dihexylsexithiophene 267 by oxidative coupling of lithiated 5-hexylterthiophene 266 with CuC12 (55% yield). The monoalkylated terthiophene 266 was obtained by palladium-catalyzed coupling of 5-hexyl-2-thienylzinc-chloride264 and Br-T2-H 50 in 70% yield [216]. However, the solubility given for hexamer 267 is surprisingly low due to the large intermolecular interactions in the solid state. Structural characterization of vacuum-evaporated thin films of hexamer 267 by X-ray diffraction revealed molecular organization and layered structures with molecules standing with a tilting angle of 16" on the Si/Si02-substrate surface. Electrical characterization indicates a higher conductivity (factor of 3-6) and higher field-effect mobility (factor of 25) for the a,a'-disubstituted derivative than found for the parent H-T6-H 6. Furthermore, the conductivity in the oriented films is largely anisotropic with a ratio of 120 in favor of the conductivity parallel to the substrate plane.

3.2 Synthesis of Oligothiophenes

143

By a similar strategy, the corresponding (iPr)3Si-T6-Si(iPr)3 269 was obtained [217]. Palladium-catalyzed coupling of the silylated organozinc thiophene 265 and Br-T2-H 50 gave (iPr)3Si-T3-H 268 in 48% yield. This is further dimerized to the a,a'-disubstituted (iPr)3Si-T6-Si(iPr)3 269 with n-BuLi/CuC12in 43% yield. Single crystals of this compound could be obtained and X-ray analysis showed that, in contrast to other oligothiophenes and obviously due to the bulky triisopropyl groups, intermolecular interactions play a more important role than intramolecular ones. Thus, in this case, a non-planar anti conformation of the conjugated .rr-system is favored in which a gradual twist of the thiophene rings is observed. The terminal thiophene rings are twisted with even 37.4" in relation to their neighbors which themselves form a dihedral angle of 21.4" with the two inner coplanar thiophene rings. The efficiency of oligothiophene-based light emitting diodes could be enhanced by using a two-layer system of unsubstituted H-T6-H 6 and (iPr)3Si-T6-Si(iPr)3 269 [218]. The longest wavelength absorptions of both sexithiophenes are red-shifted in comparison to H-T6-H 6 (Ax,,, = 11-12nm) due the comparable electron-donating character of the alkyl and silyl substitutents.

264,265

266,268 (70%. 48%)

50

267,269 (55%. 43%)

R = CSY3. (iPr)$i

Parakka and Cava have reported long chain a,a'-disubstituted sexithiophenes 277-279, which were either obtained by oxidative dimerization of appropriate monoalkylated terthiophenes 270-272 or by reductive nickel-catalyzed coupling of the corresponding brominated terthiophenes 274-276 [2 191. Reaction of monoaldehyde OHC-T3-H 74 with hexadecylmagnesiumbromide yields the corresponding carbinol in 92% yield, which is very effectively further reduced with sodium cyanoborohydride to 5-heptadecylterthiophene 270 in 97% yield. This monoalkylated terthiophene was successively coupled with FeC13 in benzene to the apdiheptadecyl-sexithiophene 277. The blue oxidation product is finally reduced by hydrazine to yield 55% of the hexamer. 5-Hexadecyloxymethyl-terthiophene 271 was made by nucleophilic substitution of 5-hydroxymethyl-terthiophene. This was obtained from OHC-T3-H 74 by reaction with hexadecyl bromide in only 7 % yield. The olefinic 5-(heptadec-1-eny1)-terthiophene 272 was synthesized by dehydration of carbinol 5-(l-hydroxyheptadecyl)-terthiophene with p-toluenesulfonic acid. At ambient temperature, 30% of the pure trans derivative is obtained, at 85°C 58% of a cisltrans mixture (85: 15). The bromo compounds 274-276 were prepared by the same procedures as described above starting from 5-bromo-5"formyl-terthiophene Br-T3-CH0 273 which is obtained by bromination of OHC-T3-H 74 with NBS in 20% yield. In one case, direct bromination of

144

3 Oligothiophenes

alkylterthiophene 270 to bromoalkylterthiophene 274 was achieved with 1,3-

dibromo-5,5-dimethylhydantoinin 93% yield. The bromoterthiophenes 274-276 were coupled to the corresponding a,a’-dialkylated sexithiophenes 277-279 with the catalytic system [Pd(PPh3)4],zinc, potassium iodide, and triphenylphosphane in DMF in 56% 277, 81% 278, and 82% yield 279 [Eq. (6911.

74

273-

1

I -

L

AIEIN. lHF

274-276

270-272

R = C,&

,lE@R&+

(69)

(270,274,277)

R = qOC,,&

(271,275,278) R = (E)CH=Cl+C,,y,(272,276,279)

277-279

Absorption spectra and redox potentials of the monoalkylated terthiophenes and the dialkylated sexithiophenes were determined. Due to the conjugation of the adjacent double bonds in a-position, in comparison to the n-alkyl substituted terthiophenes 270 and 271, the olefinic counterpart 272 exhibited a bathochromic shift of the longest wavelength absorption (Ax = 22-25nm) and is oxidized at lower potentials (AE,, = 0.12-0.17 V). The same trend is found less pronounced for olefinic sexithiophene 279 in comparison to the alkylated hexamers 277 and 278 (Ax = 10-1 5 nm; AE,, = 0.04-0.08 V). Despite their solubilizing alkyl side chains, surprisingly the hexamers did not show a much higher solubility in organic solvents than the parent unsubstituted H-T6-H 6. Evidently, besides the T-T interaction of the conjugated system, additional van der Waals attractions of the long alkyl chains cause the low solubility of the rigid-rod type molecules. Some more alkylated and silylated oligothiophenes were synthesized with respect to their electrochemical and EPR properties of the corresponding radical ions, or their biological properties: 5,5’-dimethyl-, 5,5’-diisopropyl-, 5,5’-di-tbutyl- [220], 5-trimethylsilyl-, 5,5’-bis(trimethylsilyl)-2,2’-bithiophene [221], 5-methyl-, 53’dimethyl-, 5-tbutyl-, 5,5’-dLtbutyl-, 5-[(H3C)2=CHCH2-]-[24b, 2221, 5,5”-bis(trimethylsily1)-terthiophene, and 5,5“-bis(trimethylsily1)-quaterthiophene [221]. In their series of regioirregular P,$-alkyl substituted oligothiophenes, Wynberg et al. also included the septithiophene tBu-T,-tBu 280 which was synthesized similarly

3.2 Synthesis of Oligothiophenes

145

by ‘Stetter reaction’ of 5-formyl-5”-tbutyl-terthiophene and the corresponding Mannich base (47% yield) and subsequent cyclization of the resulting diketone with L.R. (26% yield) (see above) [156]. Also in this case, the solubility is drastically reduced in comparison to the P-alkyl substituted analogs and the longest wavelength absorption is red-shifted (Ax = 11-16 nm). Evidently, a,a’-disubstitution of oligothiophenes results in a nearly undisturbed conjugated m-system and therefore simultaneously in strong intermolecular interactions which on the other hand cause a low solubility.

280

3.2.2.3 @-Substituted Oligothiophenes Several series of oligothiophenes bearing substituents at both the a- and the ppositions have been developed recently. This class of compounds now comprises two factors which affect the properties of oligothiophenes. The substituents at the (terminal) a-positions block the reactive part when the oligothiophenes are transformed into cationic species as models for polarons and bipolarons, which are considered as the charge carriers in conducting polymers. Thus, also the investigation of shorter and normally reactive oligomers becomes available. As was seen in the previous paragraph, alkyl side chains in the a-positions do not cause an increase of the solubility and the longer members are scarcly soluble in common organic solvents due to their rigid-rod character, the additional introduction of solubilizing alkyl side chains in P-positions is straightforward. On the other hand, certainly, P-substituents at inner thiophene rings cause steric interactions with the adjacent rings and thus a certain reduction of the conjugation length should be taken into account. The synthesis and characterization of a complete series of ‘end-capped‘oligothiophenes (ECnT) up to a heptamer by Bauerle clearly revealed the usefulness of this approach [13, 721. Due to the blockmg of the reactive a- and ,&positions with a cyclohexene ‘cap’, on the one hand a more precise characterization of the oligomers in various oxidation states was possible, and on the other hand, due to the enhanced solubility, excellent correlations of the spectroscopic and electrochemical data with the (inverse) chain length were obtained. 4,5,6,7-Tetrahydrobenzo[b]thiophene281 is the key building block for this series and was synthesized in 75% yield by ether cleavage of 3-(p-methoxyphenoxybutyl)thiophenewith boron tribromide under dilution conditions which favors the intramolecular ring closure reaction [86]. The smallest member in this series, EClT 282, could be obtained with the same ether cleavage reactions from 3,4-di( p-methoxyphenoxybuty1)thiophene in 70% yield. Selective bromination of 281 with NBS yields the 2-bromo derivative 283 in 89% yield which can easily be transformed into the corresponding Grignard reagent. ‘Kumada-coupling’ of the latter with 283 itself, Br-T1-Br 45, Br-T2-Br 46, and Br-T3-Br 47, results in the bithiophene EC2T 284 (47% yield) [Eq. (70)], terthiophene EC3T 285 (64% yield) [Eq. (71)], quaterthiophene EC4T 286 (78% yield) [Eq. (72)], and quinquethiophene EC5T 287 (64% yield) [Eq. (73)],

146

3 Oligothiophenes

respectively. EC5T 287 and the higher members 290-291 were synthesized by first reacting BrMg-T1-H 42 with 2-bromotetrahydrobeno[b]thiophene 283 under 'Kumada-conditions' to form the 'mono-capped' bithiophene 288 in 5 1% yield which was successivelybrominated with NBS to the other important key component 289 in 67% yield. Transformation of 289 into the Grignard compound and nickelcatalyzed coupling with Br-T1-Br 45, Br-T2-Br 46, and Br-T3-Br 47 gave EC5T 287 in 58% yield [Eq. (73)], the hexamer EC6T 290 in 58% yield [Eq. (74)], and the heptamer EC7T 291 in 44% yield [Eq. (75)], respectively.

284

285

\ Ji)

288

IowlBr iii)

286

O@p \ I

\ I

287

dii) n

(74) 290

289

291

(i) NBSIDMF125"C [89%]; (ii) BrMg-T,-H (42)/Ni(dppp)C12/Et20 [87%]; (iii) NBSIDMF125"C [67%]; (iv) 1 . Mg/Et20;2. Ni(dppp)C1,/5-bromo-2-(4,5,6,7-tetrahydrobenzo[b]thien-2-yl)thiophene 283 [47%]; (v) 1. Mg/Et20; 2. Ni(dppp)CI,/Br-T,-Br (45) [64%]; (vi) 1. Mg/Et20; 2. Ni(dppp)CI,I Br-T,-Br (46) [78%]; (vii) 1. Mg/Et,O; 2. Ni(dppp)CI,IBr-T,-Br (47) [64%]; (viii) 1. Mg/Et,Olbenzene; 2. Ni(dppp)CI,/Br-T1-Br (45) [58%]; (ix) 1. Mg/Et,O/benzene; 2. Ni(dppp)CI,IBr-T,-Br (46) [58%]; (x) 1. Mg/Et20/benzene;2. Ni(dppp)CI,/Br-T,-Br (47) [44%].

Purification of the oligomers was achieved by repeated chromatography and recrystallization. In the case of the higher members, extraction of the crude material and fractional sublimation gave the best results. The final purification of all compounds in this series was achieved by repeated fractional sublimation in a glass tube with temperature gradient.

3.2 Synthesis of Oligothiophenes

147

The enhanced solubility and the blocking of the reactive positions without perturbing the .rr-conjugation allowed, even in the case of the shorter oligomers, the precise determination of the electronic and structural features at various oxidation levels and their correlation with the chain length (see section 3.1.2.1.4). The longest wavelength absorptions and the emission maxima shift to lower energies as the extent of the 7r-system in the oligomer increases. An excellent correlation

I

0

0,2

0,4

0.6

0.8

1

Inverse chain length (lh) b

04

o

I

0.2

0.4

0,6

0.8

Inverse chain length (lh)

1

Figure 9. (a) Correlation of the absorption and emission energies of the ‘end-capped’ oligothiophenes EClT-EC6T 282, 284287, 290 with the inverse chain length (lln). (b) Correlation of the first and the second oxidation potentials of the ‘endcapped’ oligothiophenes EClT-EC6T 282, 284-287, 290 with the inverse chain length (lin) ~ 3 1 .

148

3 Oligothiophenes

of these transition energies with the inverse chain length is observed (Fig. 9, a). Extrapolation to a hypothetical infinite chain length models the properties of an = 538nm ‘ideal’ polymer (in solution) and gives a maximum absorption at, ,A (2.30eV) and a maximum emission at, , ,A = 704nm (1.76eV). These energies lie lower than those found experimentally for (solid) polythiophene films as a result of defects and interruptions of the conjugated backbone in the polymer. On the other hand, this correlation allows the estimation of the mean conjugation length in the ‘real’ polymer which is in this case for polythiophene about 10-11 a-linked thiophene units and thus differ dramatically from the mean chain length of the polymer. Except the monomer EClT 282 all compounds fluoresce strongly and = 7% (EC2T the quantum yield increases with increasing chain length 284) to 40% (EC6T 290)]. It was shown by cyclovoltammetry that even the trimer EC3T 285 is reversibly oxidized to the cation radical. Starting with the quaterthiophene EC4T 286 even stable dications can be created. In analogy to the spectroscopic results, the oxidation potentials of both the mono and the dications are gradually shifted to lower energies with increasing size of the r-system. Again for both redox potentials an excellent correlation of the energy levels versus the inverse chain length is obtained. The energy difference between the first and the second oxidation potential gradually decreases and finally both regression lines intersect as the ‘ideal’ infinite chain length is approached (Fig. 9b). This result clearly implies that in the case of very long chains a second and probably additional electrons can simultaneously be removed at the same energetic level as long as the charged defects can reside sufficiently separated on the conjugated 7r-system without interaction. In comparison to ‘real’ bulk polythiophene which exhibits a broad reversible redox wave ( E O M 0.3 V vs. Fc/Fc+) due to the inhomogeneity of the material, the redox potential of an ‘ideal’ polythiophene is estimated to be considerably lower (Eo NN 0.07 V vs. Fc/Fc+). Vice versa, the estimation of the mean conjugation length of the ‘real’polymer from this correlation results in about 5-10 correctly linked thiophene rings [13, 721. Due to the stability of the radical cations, it was now possible to investigate the reversible dimerization equilibrium of the ‘end-capped’ oligothiophenes by temperature-dependent in situ spectroelectrochemistry combined with EPR [223]. These studies revealed now a clear dependence of the dimerization tendency on the chain length. The dimerization enthalpy which was determined either by UV/VIS/NIR or EPR increases as the chain length increases. The experiments showed that radical cations of very long oligomers are almost completely dimerized at room temperature and show only weak EPR activity (see above) [169]. The correlation of the transition energies obtained for the ECnT monomeric and dimeric radical cations 283-285 also exhibit a linear dependence with the inverse chain length. Extrapolation of both sets to an infinite chain length reveals that at this extreme point, the regression lines for the pair of transitions almost intersect. This result clearly implies that the charge becomes more and more delocalized in very long chains and that Coulomb repulsion decreases with increasing chain length. The electronic structure of the dimeric cation radical thus approaches that of the monomeric cation. There is a considerable variance

3.2 Synthesis of Oligothiophenes

149

in experimental data for polarons and bipolarons in doped polythiophene (El M 1.3-1.4eV; E2 M 0.3-0.5 eV) and thus an unavoidable uncertainty in the estimation of the conjugation length from this diagram. The extrapolated transition energies for infinite chain length (El M 1.25-1.38 eV; E2 M 0.45-0.54eV) do basically correspond to the experimental values and are slightly lower for the high energy transition [223]. The data now permitted construction of an energy level diagram for monomeric and dimeric radical cations which is consistent with the observed transitions and explains their distinct blue-shift upon dimerization. This ‘Davidov blue shift’ [224] is coherent with a stack-like arrangement of the two oligomeric cations. Recently, from electrochemical measurements on ,8-alkyl substituted dodecithiophene 126 it was concluded that oxidized species form four-fold charged 7r-dimers [225]. In this context, the analogy between the 7r-dimerization of oligothiophene radical cations and highly conducting charge transfer salts is astonishing. Since increased crystallinity and orientation in e.g. polyacetylene leads to an increase in conductivity [226] and since doped microcrystalline (longer) oligothiophenes approach conductivities of the corresponding polymers (20 S cm-’) [156], it is reasonable that in analogy to the conducting crystalline charge-transfer salts, high conductivity in conducting polymers and even more in well-defined oligomers on the molecular level might be primarily due to charged (micro)crystalline stacks of conjugated segments. However, the measurable macroscopic conductivity is determined and diminished by the transfer of the charge carriers from stacks to stacks and to bigger aggregates. Since in most applications conjugated materials are used in the solid form, an important advantage of oligomers is therefore that the solid state properties can be investigated in crystals or in vapor-deposited (thin) films. However, in the solid state morphological and supramolecular effects may play an important role and lead to different properties than those found in solution. Investigations on the optical and transport properties of the ‘end-capped’ oligothiophenes ECnT 286, 287, 290, 291 in the solid state have been undertaken. The absorption and emission spectra of thin films (thickness 40-60 nm) show strongly structured bands including several vibronic couplings. The maxima gradually shift to lower energies with increasing chain length. Also for the solid state, excellent correlations of the transition energies with the inverse chain length are obtained. Nevertheless, the fluorescence quantum yields are diminished by three to four orders of magnitude in comparison to those in solution, but are still much higher than in unsubstituted oligothiophenes which form oriented monolayers when evaporated on fused silica [72, 2271. First EPR studies on photoexcited triplet states of oligothiophenes were performed using the ‘end-capped’ oligothiophenes EC2T to EC6T 284-287, 290 in frozen solutions at 4 K [228]. The characteristic lineshape of the EPR spectra provides evidence that photoexcitation leads to molecular triplet states in all compounds. The fine structure parameter D could be determined and decrease continuously with increasing oligomer chain length, i.e. the wavefunction becomes more extended the longer the oligomer chain is. The best correlation between D and the number of thiophene rings was found by plotting D against the inverse chain

1.50

3 Oligothiophenes

length (lln). The extrapolation to infinite chain length suggests that triplet excitation on an infinite one-dimensional oligothiophene would possess a finite extension. Solid-state in situ ATR, FTIR- and FT-Raman spectra of the whole series of ‘end-capped’ oligothiophenes ECnT have been studied experimentally and theoretically. The spectra in the neutral state show for some bands (C=C double bond vibration) a convergent behavior with increasing chain length and are shifted to lower energies. This clearly indicates that both the bond order and the bond fixation is slightly decreased. Finally the spectra approach the vibrational properties of polythiophene [229]. During doping with iodine vapor, films of EC3T to EC6T 285-287, 290 show broad doping induced bands in the region of 6500-5500cm-’ similar to polythiophene which are due to free charge carriers whereas the bands in the region from 1800-660 cm-’ are much narrower. In the doped state no convergence of the bands is found indicating that the oligomers are too short to allow a fully extended defect [230]. Electrical transport properties of non-doped and iodine doped ‘end-capped’ oligothiophenesin thin films were studied by current/voltage measurements between gold microcontacts [231, 2321. The experiments were performed as a function of dimension of the microstructures, film thicknesses, chain length of the oligomers, doping state and time. The devices show Ohmic behavior and a logarithmic dependence of the conductivities on the inverse chain length and furthermore a strong dependence on the in situ doping time and geometric parameters. The conductivities of non-doped devices are independent of the film thickness and electrode distance (d L. l00nm). Light-emitting diodes (LED) based on conjugated materials are at present possibly the most important application of conjugated materials. However, there are still some drawbacks, e.g. lifetimes necessary for industrial applications. The underlying and limiting physical and chemical processes such as charge carrier injection of holes and electrons at the electrodes or their recombination are therefore of considerable interest and mostly barely understood. In order to study the electroluminescent properties of conjugated systems systematically, the first organic LEDs based on defined conjugated oligomers were developed by Umbach et al. 1211. The series of ‘end-capped’ oligothiophenes EC4T to EC7T 286,287,290,291 was used to prepare LEDs by vacuum sublimation of the active organic material and allowed to investigate the dependence of their transport properties and spectral distributions on the chain length of the oligomers. The devices yield light emission in the yellow/orange color range at relatively low voltages (22.5 V) and moderate current densities. The electroluminescence spectra which possess the same shape and energy positions as the photoluminescence spectra indicate that the same radiative decay process in both cases is valid. The peak maxima are gradually shifted to lower energies with increasing chain length and an excellent correlation with the inverse chain length was obtained. Since these materials are easily oxidized or p-doped, respectively, further investigations on metal/ECnT/ITO LEDs showed that the corresponding current/voltage curves are due to the injection and transport of holes. The electroluminescence, however, is correlated to the injection of electrons at the cathode and light emission

3.2 Synthesis of Oligothiophenes

151

directly arises from a region close to the cathode [233]. Current and intensity of electroluminescence were also measured as a function of various metal cathodes in a wide range of temperature and thicknesses of the active EC6T film [234]. The to significantly depend on the metal external quantum efficiencies (q = contact (Ca, Mg, Al, In, Ag) and the device temperature (4-300K). At room temperature they are found to be in the same order as those reported for LEDs based on various polythiophene derivatives (q = 3 x lo-’ to 1 x [235]. In comparison with LEDs based on H-T6-H 6 [236] the efficiencies of the EC6T LEDs are about one order of magnitude higher, which is in agreement with the ratio of the respective photoluminescence yields [237]. Even much smaller efficiencies were reported for LEDs based on e.g. Me-T6-Me 248 (q = 3 x lop9) [238]. On the other hand, the efficiency of a two layer oligothiophene LED consisting of H-T6-H 6 and (iPr)3Si-T6-Si(iPr)3269 could be remarkably enhanced [218]. Controlled vacuum-deposition of the ‘end-capped’ oligothiophenes EC3T to EC6T 285-287, 290 in very thin films down to the monolayer regime on Ag( 111)-surfaces allowed to investigate their supramolecular behavior by means of STM [239]. The STM-images of the oligothiophenes showed in each case extremely large areas with highly ordered 2D crystalline monolayers in which a well-oriented and nearly defectfree stack-like arrangement of the oligomers is observable. This indicates a high purity of the oligomer materials (Fig. 10). Surprisingly and for the first time, images with a submolecular resolution of the oligothiophene units are obtained. Oligomers with an even number of thiophene rings (EC4T, EC6T) form structures with equal stacks and one molecule per unit cell, whereas an uneven number of thiophene rings (ECST) leads to the formation of two different stacks and two molecules per unit cell. The interpretation of the images which are underlined by LEED and theoretical investigations leads a

b

Figure 10. STM image of a monolayer of ECST 287 on Ag(l11). (a) Scan size 330 x 330A and (b) 80 x 80A [239].

152

3 Oligothiophenes

to the conclusion that each thiophene ring and the ‘end-caps’ are represented by white spots. The comparison of the distances in the STM-images and the calculated geometries in the case of EC5T monolayers revealed that the single molecules include the energetically favorable all-trans orientation of the thiophene rings. Conformations including cis arrangements are less probable, but cannot be excluded. With respect to their properties in solution and the solid state, the model character and their applications as new materials, the ‘end-capped’ oligothiophenes ECnT seem to be one of the best investigated series up to now. Due to their defined character, their effective purification methods and controllable processability by vacuum evaporation highly pure materials are obtained. Tour et al. have synthesized and characterized an a$-substituted oligothiophene series up to an octamer in which the reactive a-positions are blocked by trimethylsilyl (TMS) end groups and the solubility is retained by the alkyl substituents regioregularly attached to free ,&positions [240]. In contrast to the above mentioned ‘end-capped’ oligothiophenes, in this case the solubilizing alkyl side chains infer some distorsion of the oligothiophene 7r-system and thus induce reduction of the overall conjugation. By stepwise metal-catalyzed coupling of substituted (oligo-)thiophene building blocks the whole series was synthesized. Terthiophene 294 was obtained in 73 %O yield by ‘Suzuki-reaction’of 5-trimethylsilyl-2-thiopheneboronic acid 292 and 2,5-diiodo-3,4-dimethylthiophene 293 [Eq. (76)]; quaterthiophene 297 in 42% yield by ‘Stille-coupling’ of two bithiophene units: 5-iodo-3-methyl-5‘trimethylsilyl-bithiophene 296 and 5-tributylstannyl-3-methyl-5’-trimethylsilyl-2,2’bithiophene 295 [Eq. (77)]; pentamer 298 in 47% yield by ‘Stille-coupling’ of two equivalents of the latter tinorganic derivative 295 and I-T1-I 54 [Eq. (78)]; hexamer 303 in 52% yield by palladium-catalyzed coupling of two terthiophene units: 5-iodo-3,4’-”methyl-5’’-trimethylsilyl-terthiophene 301 and 5-tributylstannyl3,4’-dimethyl-5’-trimethylsilyl-a-terthiophene 302 [Eq. (SO)]; heptithiophene 304 from two equivalents of the latter stannylated terthiophene 302 and I-TI -I 54 in 64% [Eq. (Sl)] and the heptamer 305 in 58% yield by reacting the organotin reagent 302 with 3,4-dimethyl-2,5-diiodothiophene 293 [Eq. (SO)]. Finally, the longest defined oligomer in this series, the octithiophene 307 was obtained by the ‘Stillecoupling’ of the stannylated terthiophene 302 and bithiophene I-T2-I 306 in 52% yield [Eq. (82)].

295

296

297

3.2 Synthesis of Oligothiophenes

153

YC’

302

54 (R= H):293(R=Me)

304, R=H[64%] 305,R=CH3[58%]

302

306

307

The linear optical [240], nonlinear optical [241], and electronic properties [242], of these thiophene oligomers were studied. The absorption maxima increase with increasing chain length, but no saturation was reached. Due to the different number of methyl groups in the oligomers and therefore different influence on the conjugation, in this series only a rough correlation with the (inverse) chain length can be obtained. By comparison of the UVlVIS spectra with the spectra of the

154

3 Oligothiophenes

analogous polymers, it was concluded that electrochemically prepared poly(3= 430-440 nm) should have an effective conjugation length alkylthiophene),,A,( of 6-7 correctly linked thiophene units. The longest homolog, octamer 307, has = 458 nm. The third-order non-linear optical maximum absorption at , , ,A studies, determined by third-harmonic generation, on this series of oligomers corroborates well with the results obtained on polymeric systems while refuting data that had been obtained on the less soluble unsubstituted oligothiophenes. The soluble thiophene oligomers with three and more units can be electrooxidized stepwise to the radical cation and the dication. First and second oxidation potential and absorption energy of the radical cations as well as of the dications correlate well with the inverse chain length. The correlation of the electronic transitions of the oxidized oligomers permitted to estimate the delocalization length of the radical cation to 12 units and the dication to 10 units in the corresponding polymer [242]. In a related study directed toward the construction of molecular electronic devices [243] Tour et af. synthesized orthogonally fused oligothiophenes which might include potentially addressable ‘on’ and ‘off’ states [244]. First, the spiro core 309 was constructed from tetraalkyne 308 in 41% yield, transformed into the tetrabromo derivative 310 in 88% yield, and then the four thiophene ‘branching arms’ added at one time by metal-catalyzed coupling reactiom‘Stille-type’ coupling of the brominated spiro core 310 with excess 2-tributylstannyl-5-trimethylsilylthiophene 311 resulted in the orthogonally fused terthiophene 312 in 41% yield [Eq. (83)]. Reaction of the spiro compound 310 with excess of terthiophene 313 yielded the spiro-fused heptathiophene 314 in 86% yield [Eq. (84)]. x

’

1. EuU.

‘( 7‘

MqSi

2. &a, X

SiMe,

3oa

a

g +

Me$

~ _ _

M e S i T a .

141%I

X

Me,Si

SIMe,

A’, \ I

309.X = SiMe, (41%)

\ I

(83)

312

310. X = Er (88%)

A

[%%I

,Me

tsi)

Me,

(84)

Me

SiMe,

313

Me

Me

’

314

Each oligothiophene unit could be independently charged to the radical cation and dication by means of cyclovoltammetry, indicating that there is no

3.2 Synthesis of Oligothiophenes

155

cross-communication between the orthogonally fused segments which is certainly a prerequisite for the ‘molecular electronic’ device capability [245]. A number of a,P-methyl-substituted ‘end-capped’ bi-, ter- and quaterthiophenes were synthesized and characterized by Engelmann and Kossmehl in order to have model compounds for the closer elucidation of the polymerization mechanism and kinetics of thiophenes [246]. Nearly all compounds were synthesized by nickelcatalyzed ‘Kumada-coupling’ reactions in yields ranging from 7 to 69%. Investigations on the various radical cations by fast scan voltammetry, ring disc electrode kinetics, and EPR revealed for the first time that the ,&positions 3,3’ and 4,4’ in bithophenes give different contributions to the reactivity of electrochemically generated radical cations. The chemical oxidation of several ‘monocapped’ bi- and terthiophenes with aqueous FeC13 lead in good yields to permethylated dimerization products and related quater- and sexithiophenes were isolated. Furthermore, oxidation of compounds like 4,4/,5,5/-tetramethy1-2,2’bithiophene 315 now lead to novel reaction products in which two bithiophene units are linked via a methylene group. In t h s particular case (4,4/,5’-trimethyl-2,2’bithien-5-y1)(4,5,4’,5’-tetramethyl-2,2/-bithien-5-yl)methan 316 was isolated in 25% yield [Eq. (SS)].

Me Me Me

(85)

Me

FeCI,x6=

I=%]

Me

315

316

In their series of regioirregular P,@’-alkylsubstituted oligothiophenes (see section 3.2.2.1) Wynberg et al. also included a P-alkylated undecithiophene 317 which was synthesized similarly to the oligomers by ‘Stetter reaction’ and subsequent cyclization of the diketone with L.R. In this case, the terminal groups do practically not influence the properties in comparison to the P-alkylated undecithiophene 122. Conductivity, solubility, and absorptions are nearly identical [1561.

317

3.2.2.4 Functionalized Oligothiophenes Several series of oligothiophenes bearing functional groups in the /3- and aposition are summarized below as far as these compounds are relevant to conjugated materials. This includes functional groups which might be electrondonating or -accepting. For example, donor-substituted oligothiophenes represent

156

3 Oligothiophenes

ideal model compounds for the corresponding polythiophenes in order to better elucidate the steric and electronic influence of the substituent onto the properties. Especially, poly(a1koxythiophenes) are very promising materials because due to the electron-donating effect of the substituents they show an excellent environmental and electrochemical stability, high conductivity, and transparency in the conducting state. In this respect, doped poly[(3,4-dioxyethylen)thiophene] is one of the most stable polythiophenes known and in the meanwhile is commercialized in antistatic and electromagnetic shielding layers in photographic document films [247]. Donor- and acceptor-substituted oligothiophenes The synthesis of donor-substituted oligothiophenes as model compounds for the corresponding polymers and as starting monomers for polymerization has become attractive. Besides several examples of 3,3’- and 4,4’-dialkoxybithiophenes and parent mixed alkoxy,alkyl-substituted derivatives for subsequent polymerization [248, 2491, Gronowitz and Peters obtained 3’-methoxy-2,2’:5’,2’’-terthiophene 320, which also occurs naturally, in 54% yield by the exchange of the halogen function in 3’-iodo-2,2’:5’,2“-terthiophene 318 with sodium methoxide and cupric oxide in a nucleophilic substitution reaction. As a non-negligible side reaction dehalogenation takes place and the 30% H-T,-H 3 formed could be separated by column chromatography. Iodoterthiophene 319 itself is prepared by the Pd(0)-catalyzed reaction of 2,3,5-triiodothiophene 318 with 2-thiopheneboronic acid [Eq. (86)] [97].

.

‘I 318

319

.

0%

320

In order to evaluate the influence of the regiochemistry on polymer properties, Zotti et al. synthesized two series of donor-substituted oligothiophenes which bear pentoxy groups either in the 3-position of the terminal thiophene rings or in the 4-position [250]. 3-Pentoxythiophene 321 is converted by iodination with iodine and mercury oxide to 2-iodo-3-pentoxythiophene322 (91YOyield) which serves as starting material for the oligothiophenes 323-325. Thus, bithiophene 323 was prepared by Ni(0)-catalyzed homo-coupling in 83% yield. In contrast, terthiophene 324 was obtained by Pd(0)-catalyzed coupling of iodothiophene 322 with thiophene-2,5-diboronic acid in 50% yield. Quaterthiophene 325 was either isolated as by-product in the preparation of the trimer 324 or obtained in a ‘Kumada-type’ coupling of the Grignard reagent of iodopentoxythiophene 322 with I-T2-I 306. However, in this case no yields are given [Eq. (87)]. The regioisomeric 4,4‘-dipentoxy-2,2’-bithiophene327 bearing the pentoxy groups in the ‘outer’ ,&positions was synthesized in 70% yield by oxidative coupling of lithiated 3-pentoxythiophene 321 with copper chloride. 4-Pentoxy-2-thiopheneboronic acid 326 which was synthesized from 3-pentoxythiophene 321, lithiumdiisopropylamid and trimethylborate was coupled with I-T1-I 54 and I-T2-I 306,

3.2 Synthesis of Oligothiophenes

157

respectively, under [Pd(PPh3)4]-catalysisto result in the corresponding terthiophene 328 (67%) and quaterthiophene 329 (no yield given) [Eq. (SS)]. OR

322

(87’

OR

OR

321

326

I 323-325(n=O-2) R = C,H,

327-329(n=O-2) R = C,H,,

Due to the electron-donating character of the alkoxy groups, the redox potentials of these oligomers are diminished in comparison to the non-substituted parent oligomers and as expected decrease with increasing chain length. In the series of oligothiophenes 323-325 which bear the substituents at the ‘inner’ P-positions, reversible cyclovoltammograms were obtained (Ep = 0.38-0.26 V vs. Ag/AgCl). In contrast, the members of the other series 327-329 exhibit irreversible redox waves indicating the follow-up reaction of the radical cations to form higher oligomers or polymers (Epa= 0.67-0.40 V vs. Ag/AgCl). Miller et al. presented a series of structurally defined methoxy-substituted oligothiophenes, dimers through hexamers, symmetrically substituted at the ‘outer’ (330-332) or ‘inner’ P-positions (333-336) and with terminal methyl (337340) or carbonic acid groups (341) [251].The electron-donating methoxy groups and terminal alkyl groups stabilize cationic species and thus radical cations and protonated oligothiophenes could be investigated. Furthermore, these compounds serve as models for the interchain radical .ir-dimerswhich were found by Miller and others to be an important alternative to bipolarons in oxidized polythiophene [164]. The carboxylic acid endgroups provide the solubility of the hydrophobic oligomer in water which was found to be an ideal medium for the observation of .ir-stacking of radical cations [252]. The oligomers were built up by the cross-coupling of (oligo-)P-methoxy-a-iodothiophenesand (oligo-)a-stannylthiophenes, catalyzed by Pd(0)-complexes. Furthermore, the oxidative homo-coupling of a-lithiated thiophenes by Fe(acac)3 was used to prepare the dimers 330 and 333. Interestingly, OCH, H,CO,

330-332( n = 0-2)

333-336( n = 0-3;R = H) 337-340( n = 0-3:R = CH3) 341 ( n = 2;R = COOK)

158

3 Oligothiophenes

for these type of compounds ‘Kumada-type’ and ‘Suzuki-type’ couplings were not successful. Corresponding tetramethoxy-substituted pentamers 344,345 and a hexamer 347 were synthesized starting from 3,3”-dimethoxy-cr-terthiophene 342. Pd(0)-catalyzed coupling of the bis-stannylated terthiophene with 2-iodo-3-methoxythiophene343 results in quinquethiophene 344 (62% yield). This is subsequently lithiated and methylated in the terminal a-positions with n-BuLilDMS to form pentamer 345 in 80% yield [Eq. (SS)]. Monomethylated terthiophene 346 obtained in 81% yield from 342 with the same procedure was dimerized with the system n-BuLi/Fe(acac)3 to the tetramethoxy,dimethyl-substitutedhexamer 347 in 70% yield [Eq. (90)]. The quaterthiophene dicarbonic acid 341 was obtained from the tetramer 335 by reaction with n-BuLi/C02 in nearly quantitative yield. Oxidation of the dicarbonic acid 341 in water produces stable radical cations and aggregated 7r-stacks were demonstrated spectroscopically. lsolation of the latter results in an electrically conducting salt with a conductivity of D = 2x S cm-’. Partially oxidized mixed valence salts even exhibited a ten-fold higher conductivity [251b]. Absorption data of the ‘inside’ and ‘outside’-substituted oligothiophenes in comparison to the parent non-methoxylated a,@’-dimethyloligothiophenes showed the clear trend that adding two inside methoxy groups leads to an increase of the maximum absorption by about Ax,, = 17nm, adding two outside methoxy groups by about Ax,,, = 22 nm. It thus becomes clear that in contrast to alkyl chains these substituents do not much perturb the electronic structure and represent valuable models for non-methoxylated polythiophenes. X-ray crystallographic analysis of the tetramer 335 shows in coherence to the spectroscopic results that the molecule is nearly coplanar in a trans conformation [251a]. H, H,CO

-

1) fl-BuLi / CISnBu,

342

i

343

344 [R = H).345 ( R = CH,)

n-BuLi I DMS

346

347

From the ‘capped’ trimer 331 and tetramer 332 Miller et al. prepared the first protonated oligothiophenes 348 and 349 in solution by the treatment with trifluoroacetic acid (TFA) [Eq. (91)l [252]. Since a protonated species could not be obtained from Me-T3-Me 262, the methoxy groups clearly enhance the

3.2 Synthesis of Oligothiophenes

159

basicity and stabilize the positive charge. These compounds can also be regarded as a-complexes, which are normally reactive intermediates in electrophilic substitution reactions, and which were characterized by H-NMR and absorption spectroscopy. In comparison to the neutral parent compound, the protonated cationic species exhibit a distinct bathochromic shift of the longest wavelength absorptions (AA = 204, 216nm) which indicates a delocalization of the positive charge. Interestingly, the absorption maxima are located very similar to those of authentic = 590 nm; 3492’ terthiophene dications obtained by oxidation (34S2+, , ,A , , ,A = 570 nm). These data indicate that the oligothiophene .rr-dimers which absorb in the near IR are not a-dimers or dimeric a-complexes which e.g. can be isolated as follow-up products of radical cations formed in the oxidation of triaminobenzenes [2531.

348,349 (n=1,2)

331,332 (n=1.2)

A homologous series of donor-substituted oligothiophenes bearing somewhat weaker electron-donating methylmercapto groups in the ‘outer’ P-positions were synthesized up to a quaterthiophene by Bauerle et af [254]. The central building block for the synthesis of the P,P-disubstituted oligothiophenes was 2-bromo-4(methy1mercapto)thiophene 350 which in contrast to many other 2,4-disubstituted thiophenes could be synthesized selectively and free of isomers. It was reacted with magnesium to the corresponding Grignard reagent and homo-coupled under Ni(0)-catalysis to the 4,4’-di(methylmercapto)-2,2’-bithiophene 351 in 66% yield. Cross-coupling of the monothiophene with Br-T1-Br 45 led in 35% yield to the homologous P,P-disubstituted trimer 352 and cross-coupling with Br-T2-Br 46 in 63% yield to the corresponding tetramer 353 [Eq. (92)]. For comparison purposes, 3’-(methylmercapto)-2,2’:5’,”’-terthiophene 355 has been synthesized via ‘Kumada-coupling’ of 2,5-dibromo-3-(methylmercapto)thiophene354 and two equivalents of Grignard reagent BrMg-T,-H 42 in 65% yield [Eq. (93)].

SMe

(93)

4

\

354

355

SMe

160

3 Oligothiophenes

Physical properties are dependent on the conjugated chain length. With increasing chain length of the oligomers 351-353 the (irreversible) oxidation potentials decrease stepwise (Epa= 1.30V to 0.80V vs. Ag/AgCl). Simultaneously, the maximum absorption ,,A,( = 282-393 nm) and emission energies (A, = 392-484 nm) are gradually shifted to lower energies. Relative to unsubstituted oligothiophenes, the fluorescence quantum yield of the methylmercapto derivatives is slightly decreased. Due to the substitution pattern, which allows enough spin density in the ‘outer’ a-positions, even the longer oligomers exhibit good film forming properties and could be electropolymerized to the corresponding donor-substituted polythiophenes including a stereoregular structure. The electronic properties were found to be dominated by the donor strength of the substituents. The synthesis and structural characterization of all regioisomeric di(methylthi0)substituted bithiophenes 356-358 was reported by Folli et al. [255].Supported by force field calculations, crystal structure, absorption, and H NMR-NOE data, the conformational properties of the regioisomeric bithiophenes were investigated which are head-to-head, head-to-tail, and tail-to-tail repeating units of the corresponding polythiophenes. Despite the great differences in the electronic and steric properties of the methyl and thiomethyl groups, the conformational properties are very similar to each other. The regiochemistry dominates over the intrinsic properties of the substituent which is in fact different to the above mentioned methoxy substitutents. Experimental evidence suggests the fact that syn- or s-cis conformations play a role in solution.

356

357

358

The extension of this work led now to the synthesis of a series of soluble quater- and sexithiophenes, regioregularly substituted with the electron-donating thiomethyl groups in 0-positions [256].‘Stille-type’ cross-coupling of the dibromo derivative of bithiophene 356 with 3-methylthio-2-trimethylstannylthiophene 363 and a Pd(0)-catalyst lead to the tetrasubstituted quaterthiophene 360 whereas ferric chloride oxidation of the bithiophene 356 resulted in the regioisomeric quaterthiophene 359. Bis-bromination of both compounds gave the corresponding a,a’-dibromoquaterthiophenes 361 and 362, respectively, which were successively cross-coupled with 2-trimethylstannylthiophene under Pd(0)-catalysis to the regioisomeric tetrasubstituted sexithiophenes 364 and 365, respectively. Crosscoupling of bromoquaterthiophene 362 with 3-methylthio-2-trimethylstannylthiophene 363 under the same conditions give the hexasubstituted sexithiophene 366 [Eq. (94)].In this paper no yields are given. The maximum optical absorptions of the hexamers (Amax = 406-430 nm) are located close to these of the unsubstituted sexithiophene,,A,( = 432 nm) indicating that the loss of .rr-conjugation due to the steric effect of the P-substituents are nearly compensated by the mesomeric effect

3.2 Synthesis of Oligothiophenes

16 1

i.e. the delocalization of the electron lone-pairs of the methylthio group into the aromatic system. .~2

N B S I F

R

R 359 (R , R, = SMe. R, = H) 360 (R , R, = SMe. R, = H)

361 (R , R, = SMe. R, = H) 362 (R , R, = SMe. R, = H)

$

/

,,;Pd(PPn,w /

SnMe,;

,

(94)

w

364 (R , R, = SMe. R, = H. R3 = H) 365 (R , R p = SMe, R, = H, R, = H) 366 (R , R., R, = SMe, R, = H)

Among the electron-donating substituents, dialkylamino groups exhibit the most pronounced resonance effects and the strongest donor character which is reflected in very negative Hammett cT’-constants. While 3-di(alkylamino)thiophenes are long known, due to the excellent stabilization of the corresponding radical cation, poly[di(alkylamino)thiophenes] are not formed. Bauerle et al. now synthesized and characterized the first series of oligothiophenes 369-371 bearing the strongest known electron-donating substituent, the pyrrolidino group, in the ‘outer’ ppositions [257]. Despite the direct metallation of thiophenes with coordinating substituents in 3-position proceeds normally in the neighbouring 2-position, 4pyrrolidino-2-trimethylstannylthiophene367 could be synthesized as key building block in 76% yield free of isomers by direct metallation of 3-pyrrolidinothiophene Pd(0)-catalyzed homocoupling of the with n-butyllithium/trimethylstannylchloride. stannylthiophene 367 with 2-iodo-4-pyrrolidinothiophene 368 which was created in situ from 367 and iodine resulted in 4,4’-dipyrrolidino-2,2’-bithiophene369 in 82% yield [Eq. (95)]. Cross-coupling of the stannylthiophene 367 with I-TI-I 54 gave nevertheless a mixture of the 0,P-disubstituted terthiophene 370 (23YO)and the homo-coupling product 369 (55%) which could be separated by sublimation. ‘Stille-type coupling’ of the organostannyl compound 367 with I-T2-I 306 finally resulted in the largest homolog, P,P-dipyrrolidino-quaterthiophene371 in 58% yield [Eq. (96)]. Due to the strong electron donating effect of the pyrrolidino groups the three oligomers 369-371 are oxidized at very low potentials. Interestingly, bithiophene 369 exhibits the lowest oxidation potential (Epa= 0.08 V vs. Fc/Fc+) so far found for oligothiophenes whereas, surprisingly, it increases again and rests constant on

162

3 Oligothiophenes

tL

SnMe,

I

370,371 ( n = i , ~ )

going to terthiophene 370 (Epa= 0.11 V) and quaterthiophene 371 (Epa= 0.1 1V). Evidently, in bithiophene 369 already exists the full conjugation between the two donor substituted thiophene rings, whereas in the higher homologs 370 and 371, respectively, the terminal substituted rings are oxidized more or less independently at the same potential, irrespective of the size of the oligothiophene. The independent addressing of the thiophene units could also be observed for the corresponding poly(pyrro1idinothiophenes) which were obtained by electropolymerization of each of the three oligomers and which exhibit practically the same redox potential. An interesting series of cw’-bis(aminomethy1)-functionalized oligothiophenes 378-381 has been reported and the properties compared to the corresponding dimethyl derivatives [258]. The amino groups had to be protected before building up the oligomeric r-system. Thus, 2-(aminomethy1)thiophene 372 was reacted with tetramethyl-l,4-dichlorodisilyiethyleneto yield the corresponding protected monothiophene 373 in 96% yield. Transformation of the latter into the Grignard reagent and cross coupling with dibromo(o1igo)thiophenes under ‘Kumadaconditions’ yielded the silyl-protected oligothiophenes 374-377 (dimer to pentamer) in 38-72% yield. Deprotection with hydrochloric acid leads to water soluble dihydrochlorides which are transformed to the free amines 378-381 with hydroxide in 70-85% yield [Eq. (97)]. The electronic spectra were virtually the same compared to those of Me-T,-Me 261 (n = 2), 262 (n = 3), 247 (n = 4), 263 (n = 5), 248

374-377 ( n = 0-3)

378-381 ( n = 0-3)

3.2 Synthesis of Oligothiophenes

163

(n = 6 ) . The absorption maxima were only slightly red-shifted indicating a slightly higher electron-donating effect of the aminomethyl group in comparison to the methyl group. An extended analog of tetrathiafulvalene, a strong donor which is frequently used as a building block for conducting charge transfer salts [259],including an a-terthiophene spacer group was synthesized by Roncali et al. in order to get more insight into the influence of the structural parameters on the charge-transport mechanism and the (super)conducting properties [260]. The dilithiated 3’-substituted terthiophenes 13,172,175,183were formylated by the reaction with N-methylformanilide or acylated with acetic acid and phosphoric acid catalysis to the bisformylated terthiophenes 382-385 in 50-60% yield and the a,a’-diacetyl-terthiophene 386 in 35% yield, respectively. Wittig-Horner olefination of these terthiophenes with the 1,3-dithi01-2-ylidene-phosphonate anion gave the bis-donor-substituted terthiophenes 387-391 in 40-80% yield [Eq. (98)].

(-+p LDA

x

w

x

HCONPhMe

0

0

R 13,172,175,183

R 382-385 ((R = H. Me, C,H,,.

(CH2CH20)2Me);X = HI

386 (R = H; X = Me)

387-390 [(R= H.M e , C8H,,. (CH,CH,O),Me);

X = HI

391 (R = H; X = Me)

The donor-substitution in terthiophenes 387-391 clearly effects a decrease of the HOMOiLUMO gap. With respect to the corresponding terthiophenes 13, 171, 175, 183, the longest wavelength absorptions are distinctly red-shifted (Ax,,, = 108 nm). The oxidation potentials are decreased by about 500-600mV and the TTF-analogous terthiophenes 387-391 show two reversible and one irreversible redox waves. Production of charge transfer salts obtained by iodine doping yielded conductivities in the range of = 10-3-10-4 Scm-’ which is far lower than the usual TTF donor/acceptor complexes. Series of a,a’-disubstituted oligothiophenes bearing electron accepting groups have been developed recently [2611. Thus, a,a’-diformyl-oligothiophenes395-397 were prepared in a three-step synthesis up to a hexamer. OHC-T2-H 73 and OHC-T3-H 74 were obtained from H-T2-H 2 and H-T3-H 3 by formylation with phosphorous oxychloride and dimethylformamide in 91YO and 84% yield, respectively. Formylation thereby is one of the very few reactions of oligothiophenes which leads selectively to monosubstitution products which are deactivated for further electrophilic substitution. This is due to the electron-withdrawing effect of the carboxaldehyde group. Successivebromination of the latter compounds resulted

164

3 Oligothiophenes

in the unsymmetrical a-bromo-a'-oligothiophene carbaldehydes 393,394 [Eq. (99)]. 5-Bromothienyl-2-carboxaldehyde 382 is available commercially. The bis-formylated oligothiophenes 395-397 were finally obtained by symmetric coupling of bromoformyloligothiophenes 392-394 in the presence of zinc, NiC12 and triphenylphosphine in DMF (see above) in 70-85% yield [Eq. (loo)]. Interestingly, the attempted synthesis of the desired compounds via 'Kumada-coupling' of the acetal protected parent compounds failed according to the authors.

73,74 (n - 2.3)

2,3 ( n = 2.3)

Br W n C H O

- I Zn I PPh3 NiCle

393,394( n =2.3)

oHcqj-#LcHo 2n-2

392-394( n= 1-3)

395-397( n = 1-3)

OHC-T2-CH0 395 and OHC-T3-CHO 382 could also be obtained in 65 and 60% yield by dilithiation of H-T2-H 2 and H-T3-H 3, respectively, and trapping of the lithioorganic species with dimethylformamide. However, 'Vilsmeier-formylation' of H-T3-H 3 led to only 18% of OHC-T3-CHO 382 because OHC-T3-H 74 was formed as main product [262]. An interesting combination of oligothiophenes and electron accepting aromatic tropylium ions as capping substitutent was presented by Takahashi et al. [263]. Dilithiation of H-T2-H 2 and reaction with two equivalents of tropylium tetrafluoroborate gave the 5,5'-bis-( lH-~ycloheptatriene)-2,2'-bithiophene398 in 60% yield. Thermal isomerization by heating in refluxing xylene gave the isomeric 4H-cycloheptatriene 399 and successive hydride abstraction with trityl fluoroborate yielded the stable bis-dicationic bithiophenes 400 in 89% yield [Eq. (lol)]. n-BuLi

2

H

399

400

Monoreaction of H-T2-H 2 with tropylium tetrafluoroborate gave in 40% yield the monosubstituted bithiophene 401 which was isomerized to the corresponding4Hderivative402. Oxidative homo-coupling of lithiated bithophene 402 with CuC12 gave the disubstituted quaterthiophene 403 in a moderate yield (27%). Following hydride

3.2 Synthesis of Oligothiophenes

165

abstraction with tritylfluoroborate yielded the stable bis-dicationic quaterthiophene 404 in 73% yield [Eq. (102)l.

2

401

H

402

t i -

403

404

The spectroscopic data and the redox potentials exhibit the influence of the cationic withdrawing substituents. The longest wave length absorption of the dimer 400,,A,( = 568 nm) and the tetramer 404,,A,( = 652 nm) are considerably red-shifted in comparison to the non-substituted oligothiophenes. With increasing chain length, interestingly, both oligothiophenes exhibit an increased electron affinity and are irreversibly reduced at Epc = -0.15V and -0.09V vs. SCE. Semiempirical calculations reveal that the contribution of a quinoidal resonance structure (Q) increases with increasing chain length in comparison to the aromatic form (A). The quinoidal structure is similar to an oligothiophene dication or a bipolaron in conducting polythiophene [Eq. (103)l.

400 A

400 Q

Albers et al. synthesized a series of (4-pyridyl)-‘capped’ oligothiophenes [264]. Thus, the smallest homolog, 5,5’-di(4’-pyridyl)-2,2’-bithiophene 407 was prepared by homo-coupling from 5-iodo-2-(4’-pyridyl)thiophene 405 in 49% yield [Eq. (104)l. The synthesis of the corresponding terthiophene 408 was most successful by cross-coupling the organozinc derivative of 2-(4’-pyridy1)thiophene 406 and Br-T1 -Br 45 under Pd(dppf)C12-complexcatalysis (66% yield). As a by-product the mono-substituted product 411 was isolated. In contrast, oxidative coupling with CuClz or ‘Kumada-coupling’ were not successful and only led to very poor yields (<2%). The higher homologs, dipyridyl-substituted quaterthiophene 409 and quinquethiophene 410, formed by cross-coupling reaction of pyridylthiophene 405 and the Br-T,-Br however, could only be identified but not separated from

166

3 Oligothiophenes

mixtures with the mono-coupling products 412 and 413, respectively. The reaction to the corresponding hexamer did not proceed anymore [Eq. (lOS)].

/ q

4

-

x

407

\ WB, Pd(dPpf)CI,

N

406 (X 405 ( X ==I), fiQ)

&

N

+

N

~

:

(lo5) ;

Bf

n 408-410 ( n= 5 5 )

411-413 ( n = 3 - 5 )

a,a’-Functionalized oligothiophenes bearing both a donor and an acceptor substituent are very interesting compounds with regard to their strong solvatochromic and second-order nonlinear optical properties. In this context, Effenberger and Wiirthner reported on the synthesis of a series of donor/acceptor-substituted oligothiophenes [265].The oligomers 414-426 were built up either by cross-coupling of organozinc or organotin derivatives bearing the electron-donating endgroup and halogenated thiophenes bearing the accepting substitutent in 8-87% yield (Table 8). The ‘Stille-type’ coupling gave exceedingly higher yields. However in two cases, surprisingly, non-negligible amounts of a phenylated side-product (30% yield) which must stem from the triphenylphosphino ligands of the catalyst were found and had to be separated from the product. Additionally, several donor/acceptor-substituted ter- and quaterthiophenes were synthesized in order to gain series with varying chain lengths. Thus, MeO-T1-ZnC1 427 was cross-coupled with Br-Tz-H 50 and resulted in 5-methoxy terthiophene MeO-Tg-H 428 in 53% yield, the homo-coupling products MeO-Tz-OMe 429, and H-T4-H 4 [Eq. (106)l. Subsequent coupling of MeO-Tz-ZnC1 430 and MeO-T3-ZnC1 431 with 2-iodo-5-nitrothiophene Table 8. Donor/acceptor-substituted bithiophenes 414-426 [2651.

D A

414

415

416

417

418

H NO2

Me0 NO2

MeS NO,

Me2N Pyr NO2 NO2

421

422

423

424

419

420

Me0 CHO

Me2N CHO

425

426

~

D A Pyr = Pyrrolidino

Me0 H

Me2N Pyr H H

Me2N Me2N CN C=C(CN)*

Me2N S02Me

3.2 Synthesis of Oligothiophenes

167

I-T1 -NO2 432 gave the donor/acceptor-substituted oligothiophenes MeO-T3NOz 433 and MeO-T4-NOZ 434 in 73% and 74% yield, respectively [Eq. (107)]. Similarly, the corresponding a-pyrrolidino-a’-nitro derivatives 437 and 438 were obtained in somewhat lower yields (48Y0, 55%) by palladium-catalyzed coupling of 435 and I-T1 -NOz 432 or I-T2-NOZ metalated 5-pyrrolidino-2,2/-bithiophene 436, respectively [Eq. (lOS)].

427

50

428

430,431 ( n = 2,3)

432

433,434 ( n = 1.2)

435 ( n = 2)

432,436 ( n = 1.2)

437,438 ( n= 1.2)

The longest wavelength absorptions of the novel donor/acceptor oligothiophenes are polarized along the long axis of the molecules and therefore strongly influenced by the endgroups. The bathochromic shift and the increasing intensity with rising donor and acceptor strength is consistent with a charge-transfer character of this transition. All compounds show exceptionally strong solvatochromic properties which are useful for the empirical determination of solvent polarity. Thus e.g. pyrrolidino/nitro-substituted bithiophene 418 is readily soluble in all organic solvents and displays a solvatochromic shift which extends almost over the whole = 466 nm (n-hexane) to, , ,A visible range ,,A,[ = 597 nm (formamide/water)] indicating a large electronic interaction between the two endgroups and full electron delocalization. X-ray structure analysis shows a coplanar arrangement of the two thiophene rings and an equalization of the C-C bond lengths in the thiophene rings. Second-order non-linear optical properties were also found to be very high. According to EFISH measurements, the static hyperpolarizability of bithiophene 418 is twice as large as that of the corresponding biphenyls and agrees well with the value calculated from the solvatochromic data. This trend is also obtained for other donor/acceptor combinations and different chain length of the conjugated 7r-system [266]. Hutchings et al. also synthesized donor/acceptor-substituted bithiophenes and their measurement of the linear and non-linear optical properties corroborated by theoretical calculations confirm the results found by Effenberger et al. Thus e.g. bithiophene 441 substituted in the a,a’-positions by a dimethylhydrazono and a dicyanovinyl group showed a bathochromic shift of the longest wavelength absorption, a slightly reduced dipole moment, and a hyperpolarizability

168

3 Oligothiophenes

(Po = 100 x esu) which is twice as high in comparison to pyrrolidino/nitrobithiophene 418 (Po = 54 x lop3’esu) [267]. The synthesis was performed by reacting OHC-T2-H 73 with N,N-dimethylhydrazine to yield the corresponding hydrazone 439 in 85% yield. Formylation with n-BuLilDMF gave 5’-formyl-2,2’bithiophene-Scarboxaldehyde dimethylhydrazone 440 in 53% yield. The final transformation to donor/acceptor oligothiophene 441 was achieved by reaction of the hydrazone 440 with malodinitrile in 67% yield [Eq. (109)l.

441

440

A series of more complicated donor/acceptor-substituted oligothiophenes were synthesized by Effenberger et al. as model compounds for intramolecular photoinduced energy transfer [268]. The conjugated triad molecules consist of an oligothiophene as conjugated bridge which is terminally linked to the 9-position of anthracene acting as photoexcitable donor and a meso-position of a porphyrin as acceptor. The synthetic sequence starts from anthrone 442 whch is reacted with I-TI -I 54 and n-BuLi to form the donor-substituted 2-(9‘-anthryl)-5iodothiophene 443 in 77% yield. Elongation of the conjugated chain by one and two thiophene rings is achieved by palladium-catalyzed coupling of the latter with the zincorganic derivative of protected thiophene aldehydes 444 and 445, respectively. 5-(9”-Anthryl)-5’-formyl-2,2’-bithiophene 446 and the corresponding terthiophene 447 were formed in 42% and 30% yield, besides homo-coupling products. The final porphyrins 448 and 449 were prepared by condensation of pyrrole and a mixture of the anthryl-oligothienyl aldehydes 446, 447, and 1-hexanal. Subsequent oxidation with p-chloranil using ‘Lindsey’s method’ gave mixtures of porphyrins containing different ratios of pentyl and anthryl-oligothienyl meso substituents. MPLC-chromatographic purification allowed one to isolate the desired triad molecules, anthryl-bithienyl-porphyrin 448 and anthryl-terthienyl-porphyrin 449 in 9-10% yield [Eq. (1 lo)]. In addition, anthryl-oligothiophenes and oligothienylporphyrins representing models for the donor-chain and the chain-acceptor, respectively, were included and synthesized by the same methodology [269]. Despite direct connection of the endgroups with the oligothiophene bridge, the electronic interactions of the anthracene and porphyrin moieties are weak. Due to steric reasons, the individual subunits in the supermolecule can be recognized in the absorption spectra by separated bands. Thus, selective excitation of the anthryl unit is possible and a unidirectional and efficient energy transfer to the emitting porphyrin acceptor via the conjugated oligothiophene chain has been studied by

3.2 Synthesis of Oligothiophenes

169

R 448,449 (R = C5H,, , n = 1.2)

446,447 ( n= 1.2)

steady-state fluorescence spectra, fluorescence exitation spectra and picosecond time-resolved fluorescence measurements. Another combination of oligothiophenes with porphyrins was used by Shimidzu et al. to construct electrochemically one- and two-dimensional porphyrinoligothiophene copolymers as models for ultra-fine nanostructures [270]. The porphyrin moieties are linked either by oligothiophene bridges in the rneso-positions to result in two-dimensional net-works or linked in the axial direction via a central phosporous which forms two stable axial bonds with oligothienylalkoxy groups. Thus, rneso-tetrakis(bithieny1)porphyrin 450 and the terthiophene analog 451 were synthesized by refluxing the oligothiophene carboxaldehydes with pyrrole in propionic acid. The axially connected derivatives 452 and 453 were synthesized by the reaction of dichlorophosphorus(v)-tetraphenylporphyrin and the corresponding oligothienylmethylalcohols. H

@I"

Ph

H

@I"

bh

H

450,451 (n= 2.3)

452,453 ( n = 2.3)

170

3 Oligothiophenes

Several metal complexes of these porphyrinoligothiophene copolymers were used to construct sandwich cells which exhibit diode-like rectifying properties. 3.2.2.5 Amphiphilic Oligothiophenes Very recently, (oligo-)thiophenes were presented which are functionalized in such a way that they exhibit amphiphilic behavior and form ordered assemblies of conjugated molecules. Ideally, longer fatty acids or alkane thiols may be organized in Langmuir-Blodgett (LB) films or in self-assembled monolayers (SAM), respectively. The incorporation of thiophenes in these ordered systems would possibly allow polymerization in a more ordered way. Thus, the amphphilic 3thienylpentadecanoic acid forms stable films on the water subphase and can be transferred onto a variety of hydrophilic and hydrophobic substrates as monoand multilayers [271]. The inclusion of oligothiophenes in these assemblies, however, would result in a stack-like arrangement of the conjugated 7r-systems. Such supramolecular structures are very interesting with respect to the conduction mechanism in oxidized oligothiophenes, since conductivity along stacks might contribute to the overall conductivity. From several a-substituted terthiophenes 454-457, Nakahara et al. obtained LB-films with well-defined alignments of the conjugated molecules [272]. Alkylesters of these terthiophenes with a short alkyl chain form stable monolayers on water and can be deposited on substrates in such a way that the long axes of the oligothiophenes stand nearly vertical.

454-457 [(R = CH3.C,H,,C,,HH,. CH=C(CH,), ]

466-467 (R = H. Ac)

Amphiphilic bithiophenes 459, 460 and corresponding terthiophenes 463-465 which bear a longer alkyl chain and terminal polar head groups like carboxyalkyl, alkylthiol or viologen were recently synthesized and their film forming properties investigated. Starting from H-T2-H 2, 5-( 1l-bromoundecyl)-2,2/-bithiophene 458 was synthesized via acylation and by successive reduction of the carbonyl group and the terminal halogen function were replaced by a carbonic acid or a thiol group to yield the amphiphilic bithophenes 459 and 460, respectively. Elongation to the corresponding terthiophenes was achieved via bromination of bithiophene 458 with NBS to yield 461, nickel-catalyzed cross-coupling with BrMg-T1-H 42 gave the 5-( 11-bromoundecy1)-a-terthiophene462, and analogous exchange of the terminal bromo group by carboxylic acid 463, thiol 464 or 4,4/bipyridinium group 465 [Eq. (1 1l)] [72]. The viologen-functionalized terthiophene 465 represents a donor/spacer/acceptor-system which forms LB-mono- and multilayers [273]. In these structures, the electron donating units are separated from the electron accepting units by the alkyl chain in a defined manner giving access to a controllable vectorial electron transfer. Undecylthiol-substituted terthiophene 464 forms very stable SAMs on gold surfaces due to the additional 7r-7r-interaction of the conjugated moieties [274]. Tour et al. have reported that (2,2/-bithien-5yl)thiol466 and the corresponding thioacetyl compound 467 form also SAMs [275].

3.2 Synthesis of Oligothiophenes

J

461

462

463,464,465 (X = COOH, SH. Viologen)

3.2.2.6 Transition Metal Complexes of Oligothiophenes Mann et al. have reported the first metal 7r-complexes of oligothiophenes [276]. Fourteen different cyclopentadienyl (Cp) and pentamethylcyclopentadienyl (Cp*) ruthenium complexes of the general form [CpRu(oligothiophene)]PF6 and [Cp*Ru(oligothiophene)]PF6 were synthesized and characterized. Ruthenium is found to bind in the complexes 468, 469 q 5 to the outermost thiophene ring of H-T2-H 2, H-T3-H 3, H-T4-H 4, and Me-T3-Me 262, respectively. In the case of Ph-T3-Ph complexes (470, 471) the metal is bound q6 to the pendant phenyl groups. The latter complexes are very stable with respect to decomplexation, whereas complexes in which the ruthenium is bound to a thiophene ring are stable in the solid state and in dichloromethane solutions, whereas in acetone rearrangements occur producing equilibrium mixtures of free oligothophene, mono- and diruthenated species. By NMR spectroscopy it could be shown that the coordination of ruthenium affects the electronic structure of the complexed oligothiophene and extends over the bound ring and the nearest neighbor rings.

468 (n=O,l,2)

470 (n=3)

469 (n=0,1,2)

471 (n=3)

172

3 Oligothiophenes

3.3 Conclusions Are de$ned oligomers more than just model compounds for conducting polymers? This contribution intended to describe the development in oligothiophene chemistry up to mid 1996 and to cover those compounds which have relevance to the field of conjugated materials and conducting polymers, respectively. The strength of the ‘oligomeric approach’ is very evident for oligothiophenes, since they represent the class of defined model compounds which by far is the most frequently synthesized, characterized, and utilized in applications. A great variety of known synthetic methods have been used, but new strategies have also been developed to build up homologous series of oligothiophenes. Methods starting either from acyclic precursors, which are cyclized by specific ring closure reactions, or a variety of mostly transition metal-catalyzed arene/ arene coupling reactions lead to oligomeric systems with defined chain and conjugation length. A broad range and various series of oligothiophenes with varying chain length and substitution pattern have been developed and characterized in recent years, whose structure and consequently properties can be controlled and tailored, due to the introduction of substituents such as alkyl side chains, donors, acceptors or other functional groups. A valuable insight into the relationship between geometric and electronic structure and the resulting properties is thus provided. In some cases, structural studies by X-ray analysis have been possible and gave valuable information about the subtle influence of steric and electronic factors on conformational properties. Very recently, even the structural data of the very long member, H-T8-T 14 were determined [277]. In homologous series of oligomers, the properties can be followed as a function of chain length. Therefore, clear structure/property-relationshipscan be evaluated. Indeed, as can be shown experimentally and theoretically, nearly all of the (electronic) properties depend linearly on the inverse (conjugated) chain length. These correlations now allow the effective conjugation length of a corresponding polymer, and values for a (hypothetical) infinite chain length to be estimated by extrapolation. This information is not normally available from investigations on the corresponding polydisperse polymeric systems. For very long members in oligothiophene series, even some physical properties surpass those of the corresponding polymers which makes them attractive candidates as active materials in (opto)electronic applications. Do defined oligothiophenes represent more than just model compounds for conducting polymers? In fact, there are definitely much more exciting aspects - not known for the corresponding polymers - which entitle them to be an extremely interesting class of compounds in their own right. Thus, if conducting polymers are still considered as ‘molecular wires’, defined oligomers are certainly the much better candidates and components to built up future ‘molecular electronic devices’. Besides the fact that, for example with oligothiophenes a new species contributing to the charge transport in conjugated systems could be found, one of the most intriguing aspects is the possibility of arranging them in well-ordered supramolecular architectures and adsorbing them specifically on substrate surfaces. The

References

173

ordering of these self-assembled mono- and multilayers is also a function of chain length and substitution pattern of the oligomers. Optical and electric properties and the imaging on a submolecular level, provide totally new information and views about conjugated systems. The most challenging vision, which is to contact a single molecule with a STM-tip and to control the (unidirectional) charge transport along a single conjugated molecule or ‘wire’ is only possible with welldefined and well-ordered assemblies of oligomers. Furthermore, with the recently achieved synthesis of very long oligomers, dimensions are already approached which in the meanwhile can also be produced in nanostructures by special lithographic techniques. Therefore, we might already have the tools in hand to span two nanoelectrodes with a single conjugated molecule. If simultaneously the charge transport along the ‘molecular wire’ can be controlled via attached functional groups which can be addressed from the outside, the way to the realization of ‘molecular electronic devices’ like a diode or a transistor on the molecular level seems to be opened due to now available well-defined oligomeric systems. Acknowledgement First of all, I would like to thank my coworker Christiane Heim for helping to prepare the manuscript. It is a great pleasure to thank my students and coworkers in our group who did exciting research and contributed to the field of oligothiophenes: Dr. Kai-Uwe Gaudl, Dr. Stefan Scheib, Dr. Markus Hiller, Dr. Thomas Fischer, Ullrich Mitschke, Andreas Emge, Guido Rimmel, Jens Kromer, Alexander Meyer, Gerhard Gruner, Michael Schon, Christiane Heim and Jens Glaser. I’m very indebted to Prof. Eberhard Umbach and his group, Physics Department, University of Wiirzburg, for a very fruitful and intensive cooperation over many years. In many discussions, they opened our eyes and taught us about surfaces and their interaction with organic molecules. Finally, I’d like to thank the editor, Dr. Denis Fichou, C.N.R.S, Thiais, who gave us the opportunity to contribute to this book series.

References 1. (a) H. Shirakawa, E. J. Louis, A. G. MacDiarmid, C. K. Chiang, A. J. Heeger, J . Chem. SOC., Chem. Commun., 1977, 578; (b) C. K. Chiang, Y.W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis, A. G. MacDiarmid, J. Chem. Phys., 1978,69, 5098; (c) C. K. Chiang, C. R. Fincher, Y. W. Park et al., Phys. Rev. Lett., 1977, 39, 1098. 2. A. F. Diaz, J. I. Castillo, J . Chem. SOC.,Chem. Commun., 1980, 397. 3. D. M. Ivory, G. G. Miller, J. M. Sowa, L. W. Shacklette, R. R. Chance, R. H. Baughman, J. Chem. Phys., 1979, 71, 1506. 4. G. E. Wnek, J. C. W. Chien, F. E. Karasz, C. P. Lillja, Polymer, 1979, 20, 1441. 5. J. F. Rabolt, T. C. Clarke, K. K. Kanazawa, J. R. Reynolds, G. B. Street, J. Chem. SOC., Chem. Commun., 1980, 347. 6. R. R. Chance, L. W. Shacklette, G. G. Miller et al., J . Chem. SOC.,Chem. Commun., 1980, 348. 7. A. F. Diaz, J. A. Logan, J . Electroanal. Chem., 1980, I l l , 111. 8. G. Tourillon, F. Garnier, J . Electroanal. Chern., 1982, 135, 173.

174

3 Oligothiophenes

9. H. Guenther, M. D. Bezoari, P. Kovacic, S. Gronowitz, A.-B. Hoernfeldt, J. Pofym. Sci., Polym. Lett. Ed., 1984, 22, 65. 10. J. Bargon, S. Mohmand, R. J. Waltman, Mol. Cryst. Liq. Cryst., 1983, 93, 279. 11. P. Bauerle, Adv. Mater., 1993, 5, 879. 12. F. Martinez, R. Voelkel, D. Naegele, H. Naarmann, Mol. Cryst. Liq. Cryst., 1989, 167, 227. 13. P. Bauerle, Adv. Mater., 1992, 4, 102. 14. (a) F. Garnier, G. Horowitz, D. Fichou, Synth. Met., 1989, 28, C705; (b) D. Fichou, G. Horowitz, Y. Nishikitani, F. Garnier, ibid, C723; (c) F. Garnier, G. Horowitz, X. Peng, D. Fichou, Adv. Mater., 1990, 2, 592; (d) B. Xu, D. Fichou, G. Horowitz, F. Garnier, Adv. Mater., 1991, 3, 150. 15. K. Yoshino, Synth. Met., 1989, 28, C669. 16. D. Fichou, J.-M. Nunzi, F. Charra, N. Pfeffer, Adv. Mater., 1992, 4, 64. 17. M. Busch, W. Weber, J . prakt. Chem., 1936, 146, 146. 18. R. Kuhn, Ch. Grundmann, Chem. Ber., 1938, 71, 442. 19. (a) W. Steinkopf, W. Kohler, Lieb. Ann. Chem., 1936, 522, 17; (b) W. Steinkopf, H.-J. v. Petersdorf, R. Gording, Lieb. Ann. Chem., 1937, 527, 272; (c) W. Steinkopf, R. Leitsmann, K.-H. Hofmann, Lieb. Ann. Chem., 1941,546, 180. 20. R. Leitsmann, Dissertation Technische Hochschule Dresden, 1941. 21. F. Geiger, M. Stoldt, H. Schweizer, P. Bauerle, E. Umbach, Adv. Mater., 1993, 5 , 922. 22. L. Zechmeister, J. W. Sease, J. Am. Chem. SOC.,1947, 69, 273. 23. S. Gronowitz, H.-0. Karlsson, Arkiv Kemi, 1960, 17, 89. 24. (a) C. Soucy-Breau, A. MacEachern, L. C. Leitch, J. T. Arnason, P. Morand, J. Heterocycl. Chem., 1991,28,411; (b) A. MacEachern, C. Soucy, L. C. Leitch, J. T. Arnason, P. Morand, Tetrahedron, 1988,44, 2403; (c) R. Rossi, A. Carpita, M. Ciofalo, J. L. Houben, Gazz. Chim. Ital., 1990, 120, 793. 25. (a) K. R. Downum, D. Provost, L. Swain, Bioact. Mol., 1988, 7, 151; (b) R. Jente, G. A. Olatunji, F. Bosold, Phytochemistry, 1981, 20, 2169; (c) F. Bohlmann, C. Zdero, Chem. Ber., 1976, 109, 901. 26. (a) R. J. Marles, R. L. Compadre, C. M. Compadre et al., Pestic. Biochem. Physiol., 1991, 41, 89; (b) J. C. Scaiano, A. MacEachern, J. T. Arnason, P. Morand, D. Weir, Photochem. Photobiol., 1987,46, 193; (c) J. T. Arnason, B. J. R. Philoghe, C. Berg et al., Phytochemistry, 1986, 25, 1609; (d) J. Kagan, S. K. Arora, J. Org. Chem., 1983, 48, 4317. 27. (a) J. Bakker, F. J. Gommers, I. Nieuwenhuis, H. Wynberg, J . Biol. Chem., 1979, 254, 1841. (b) J. T. Arnason, G. F. Q. Chan, C. K. Wat, K. Downum, G. H. N. Towers, Photochem. Photobiol., 1981, 33, 821. 28. (a) J. Roncali, F. Garnier, M. Lemaire, R. Garreau, Synth. Met., 1986,15,323;(b) B. Krische, M. Zagorska, Synth. Met., 1989, 28, C263. 29. J. Heinze, J. Mortensen, K. Hinkelmann, Synth. Met., 1987, 21, 209. 30. (a) R. J. Waltman, J. Bargon, A. F. Diaz, J. Phys. Chem., 1983, 87, 1459; (b) Y. Yumoto, S. Yoshimura, Synth. Met., 1985, 13, 185. 31. 0. Inganas, B. Liedberg, C. R. Wu, H. Wynberg, Synth. Met., 1985, 11, 239. 32. (a) A. Alberti, L. Favaretto, G. Seconi, J . Chem. SOC.,Perkin Trans. 2, 1990, 931; (b) J. Caspar, V. Ramamurthy, D. R. Corbin, J . Am. Chem. SOC., 1991, 113, 600; (c) D. Fichou, G. Horowitz, F. Garnier, Synth. Met., 1990, 39, 125. 33. D. Fichou, G. Horowitz, B. Xu, F. Garnier, Springer Ser. Solid State Sci., 1989, 91, 386; (b) D. Fichou, G. Horowitz, B. Xu, F. Garnier, Synth. Met., 1990,39,243; (c) D. Fichou, B. Xu, G. Horowitz, F. Garnier, Synth. Met., 1991, 41, 463. 34. F. Garnier, G. Horowitz, X. Peng, D. Fichou, Adv. Mater., 1990, 2, 592. 35. U. Schoeler, K.-H. Tews, H. Kuhn, J. Chem. Phys., 1974,61, 5009. 36. K. Auwers, T. V. Bredt, Chem. Ber., 1894, 27, 1741. 37. (a) H. D. Hartough in The Chemistry of Heterocyclic Compounds, A. Weissberger, ed.: Thiophene and Its Derivatives, S. Gronowitz, ed., John Wiley, 1952, p.459; (b) J. Nakayama, T. Konishi, M. Hoshino, Heterocycles, 1988, 27, 1731; (c) R. Hakansson in The Chemistry of Heterocyclic Compounds, Volume 44, Part 5: Thiophene and Its Derivatives, S. Gronowitz, ed., John Wiley, 1992, p. 755. 38. Y. Ito, T. Konoike, T. Harada, T. Saegusa, J. Am. Chem. SOC.,1977, 99, 1487.

References

175

W. Steinkopf, J. Roch, Lieb. Ann. Chem., 1930, 482, 251. A. E. Lipkin, J. Gen. Chem. USSR, 1963, 33, 188. S. Gronowitz, H.-0. Karlson, Arkiv Kemi, 1961, 17, 89. T. Kauffmann, Angew. Chem., 1974, 86, 321. J. Kagan, S. K. Arora, Heterocycles, 1983, 20, 1937. J. P. Morizur, Bull. SOC.Chim. Fr., 1964, 1331. J. P. Morizur, R. Pallaud, Compt. Rend., 1962, 254, 1093. T. Kauffmann, Angew. Chem., 1979, 91, I . T. Kauffmann, H. Lexy, Chem. Ber., 1981, 114, 3674. F. Garnier, G. Horowitz, D. Fichou, Synth. Met., 1989, 28, C705. N. Jayasuriya, J. Kagan, D.-B. Huang, B. K. Teo, Heterocycles, 1988, 27, 1391. (a) W. S. Rapson, R. G. Shuttleworth, J. N. van Niekerk, J . Chem. SOC., 1943, 326; (b) G. Wittig, Q. Rev. Chem. SOC.,1966, 20, 205; (c) H. A. Staab, F. Binnig, Chem. Ber., 1967, 100, 293; ibid. 889. 51. (a) B. Greving, A. Woltermann, T. Kauffmann, Angew. Chem., 1974, 86, 475; Angew. Chem. Znt. Ed. Engl., 1974, 13, 467; (b) T. Kauffmann, B. Greving, J. Konig, A. Mitschker, A. Woltermann, Angew. Chem., 1975,87, 745; Angew. Chem. Znt. Ed. Engl., 1975,14,713; (c) T. Kauffmann, B. Greving, R. Kriegesmann, A. Mitschker, A. Woltermann, Chem. Ber., 1978, 111, 1330; (d) T. Kauffmann, H. P. Mackowiak, Chem. Ber., 1985, 118,2343. 52. C. Heim, Diplomarbeit Universitat Wiirzburg, 1995. 53. A. Berlin, G. A. Pagani, F. Sannicolb, J. Chern. SOC.,Chem. Commun., 1986, 1663. 54. D. Fichou, G. Horowitz, F. Garnier, Europ. Patent 402269, 12.12.1990 (Frankreich); Chem. Abstr., 1990, 114, 3863878. 55. N. Noma, T. Tsuzuki, Y. Shirota, Adv. Mater., 1995, 7, 647. 56. Z. Xu, D. Fichou, G. Horowitz, F. Garnier, J. Electroanal. Chem., 1989, 267, 339. 57. G. M. Davies, P. S. Davies, W. E. Paget, J. M. Wardleworth, Tetrahedron Lett., 1976, 795. 58. H. Wynberg, A. Logothetis, J . Am. Chem. SOC.,1956, 78, 1958. 59. R. E. Atkinson, R. F. Curtis, G. T. Phillips, J. Chem. SOC.C . , 1967, 2011. 60. M. Nilsson, Ch. Ullenius, Acta Chem. Scand., 1970, 24, 2379. 61. P. E. Fanta, Synthesis, 1974, 9. 62. L. Groenendaal, H. W. I. Peerlings, J. L. J. van Dongen, E. E. Havinga, J. A. J. M. Vekemans, E. W. Meijer, Macromolecules, 1995, 28, 116. 63. M. Nilsson, Tetrahedron Letters, 1966, 679. 64. M. S. Kharash, 0. Reinmuth, Grignard Reactions of Non-metallic Substances, Prentice-Hall, New York, 1954. 65. (a) M. Tamura, J. K. Kochi, Synthesis, 1971, 303; (b) M. Tamura, J. K. Kochi, J . Am. Chem. SOC.,1971, 93, 1483. 66. R. J. P. Corriu, J. P. Masse, J. Chem. SOC.,Chem. Commun., 1972, 144. 67. K. Tamao, K. Sumitani, M. Kumada, J. A m . Chem. SOC.,1972, 94,4374. 68. M. Yamamura, I. Moritani, S. Murahashi, J. Organornetal. Chem., 1975, 91, C39. 69. K. Tamao, K. Sumitani, Y. Kiso et al., Bull. Chem. Soc. Jpn., 1976, 49, 1958. 70. M. Kumada, Pure Appl. Chem., 1980, 52, 669. 71. K. Tamao, S. Kodama, I. Nakajima, M. Kumada, Tetrahedron, 1982, 38, 3347. 72. P. Bauerle, Habilitationsschrift Universitat Stuttgart, 1994, p. 90. 73. P. Bauerle, F. Wiirthner, G. Gotz, F. Effenberger, Synthesis, 1993, 1099. 74. (a) G. R. van Hecke, W. de W. Horrocks, Jr., Znorg. Chem., 1966, 5 , 1968; (b) S. S. Sandhu, M. Gupta, Chem. Ind. (London), 1967, 1876. 75. I. R. Butler, W. R. Cullen, T.-J. Kim, S. J. Rettig, J. Trotter, Organometallics, 1985, 4, 972. 76. A. Carpita, R. Rossi, C. A. Veracini, Tetrahedron, 1985, 41, 1919. 77. C. van Pham, A. Burkhardt, R. Shabana et al., Phosph. Sulf: S i l k , 1989, 46, 153. 78. R. Rossi, A. Carpita, A. Lezzi, Tetrahedron, 1984, 40, 2773. 79. I. Colon, D. R. Kelsey, J . Org. Chem., 1986, 51, 2627. 80. M. Kumada, M. Zembayas, Tetrahedron Lett., 1977, 4089. 81. J. Nakayama, T. Konishi, S. Murabayashi, M. Hoshino, Heterocycles, 1987, 26, 1793. 82. K. Yui, Y. Aso, T. Otsubo, F. Ogura, Bull. Chem. SOC.Jpn., 1989, 62, 1539.

39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50.

176

3 Oligothiophenes

83. (a) T. Yamamoto, K. Osakada, T. Wakabayashi, A. Yamamoto, Macromol. Chem. Rapid. Commun., 1985, 6, 671; (b) T. Yamamoto, K. Sanechika, J. Yamamoto, J. Polym. Sci., Polym. Lett. Ed., 1990, 18, 9. 84. S. Gronowitz, Chem. Scr., 1987, 27, 535. 85. (a) N. Miyaura, T. Yanagi, A. Suzuki, Synth. Commun., 1981, 11, 513; (b) T. Hayashi, M. Konishi, M. Fukushima et al.. J. Am. Chem. SOC..1982. 104. 180. 86. P. Bauerle, F. Wurthner, S. Heid, Angew. Chem., 1990,’102,414;Angew. Chem. Int. Ed. Engl., 1990, 29, 419. 87. J. K. Stille, Angew. Chem., 1986, 98, 504; Angew. Chem. Int. Ed. Engl., 1986, 26, 508. 88. V. N. Kalinin, Synthesis, 1992, 413. 89. (a) H. Azizian, C. E. Eaborn, A. Pidcock, J. Organomet. Chem., 1981, 215, 49; (b) A. N. Kashin, I. G. Bumagina, N. A. Bumagin, V. N. Bakunin, I. P. Beletskaya, J. Org. Chem. USSR, 1981, 17, 789. 90. V. Farina, B. Krishnan, J. Am. Chem. SOC.,1991, 113, 9585. 91. V. Farina, S. Kapadia, B. Krishnan, C. Wang, L. S. Liebeskind,J. Org. Chem., 1994,59, 5905. 92. J. Malm, P. Bjork, S. Gronowitz, A.-B. Hornfeldt, Tetrahedron Lett., 1992, 33, 2199. 93. G. T. Crisp, Synth. Commun., 1989, 19, 307. 94. D. K. Morita, J. K. Stille, J. R. Norton, J. Am. Chem. SOC.,1995, 117, 8576. 95. R. B. Miller, S. Dugar, Organometallics, 1984, 3, 1261. 96. S. Gronowitz, D. Peters, Heterocycles, 1990, 30, 645. 97. (a) S. Gronowitz, K. Lawitz, Chem. Scr., 1983, 22, 265. 98. (a) S. Gronowitz, V. Bobosik, K. Lawitz, Chem. Scr., 1984, 24, 5; (b) S. Gronowitz, A. Svensson, Isr. J. Chem., 1986, 27, 25. 99. (a) E. Negishi, F. T. Luo, R. Frisbee, H. Matsuchita, Heterocycles, 1982,18, 117; (b) T. Frejd, T. Klingstedt, Synthesis, 1987, 40. 100. A. Minato, K. Tamao, T. Hayashi, K. Suzuki,M. Kumada, TetruhedronLett., 1980,845. 101. L. Groenendaal, H. W. I. Peerlings, E. E. Havinga, J. A. J. M. Vekemans, E. W. Meijer, Synth. Met., 1995, 69, 467. 102. N. A. Bumagin, P. G. More, I. P. Beletskaya, J. Orgunomet. Chem., 1989, 364, 231. 103. P. Vincent, J.-P. Beaucourt, L. Pichat, Tetrahedron Lett., 1984, 25, 201. 104. S. Gronowitz, A.-B. Hornfeldt, V. Kristjansson, T. M u d , Chem. Scr., 1986, 26, 305. 105. S. Gronowitz, H.-0. Karlsson, Arkiv Kemi, 1960, 17, 89. 106. J. Kagan, S. K. Arora, I. Prakash, A. Uestuenol, Heterocycles, 1983, 20, 1341. 107. H. Wynberg, A. Logothetis, D. VerPloeg, J. Am. Chem. Soc., 1957, 79, 1972. 108. R. F. Curtis, D. M. Jones, G. Ferguson et al., J. Chem. SOC.,Chem. Commun., 1969, 165. 109. H. Stetter, B. Rajh, Chem. Ber., 1976, 109, 534. 110. H. Wynberg, J. Metselaar, Synth. Comm., 1984, 14, 1. 111. J. Kagan, S. K. Arora, Heterocycles, 1983, 20, 1941. 112. T. Asano, S. Ito, N. Saito, K. Hatakeda, Heterocycles, 1977, 6, 317. 113. R. M. Moriarty, 0. Prakash, M. P. Duncan, Synth. Commun., 1985, 15, 789. 114. D. R. Shridhar, M. Jogibhukta, P. S. Rao, V. K. Handa, Synthesis, 1982, 1061. 115. H. J. Kooreman, H. Wynberg, Rec. Trav. Chim. Pays-Bas, 1967,86, 37. 116. A. Merz, F. Ellinger, Synthesis, 1991, 462. 117. P. J. Stang, F. Diederich, Modern Acetylene Chemistry, Verlag Chemie, Weinheim, 1995. 118. D. M. Perrine, J. Kagan, Heterocycles, 1986, 24, 365. 119. K. E. Schulte, J. Reisch, L. Horner, Chem. Ber., 1962, 95, 1943. 120. (a) J. Nakayama, M. Shimomura, M. Iwamoto, M. Hoshino, Heterocycles, 1985, 23, 1907; (b) J. Nakayama, Y. Nakamura, T. Tajiri, M. Hoshino, Heterocycles, 1986, 24, 637; (c) J. Nakayama, Y. Nakamura, S. Murabayashi, M. Hoshino, Heterocycles, 1987, 26, 939. 121. J. Nakayama, S. Murabayashi, M. Hoshino, Heterocycles, 1987, 26, 2599. 122. J. L. BrCdas, G. B. Street, Acc. Chem. Res., 1985, 18, 309. 123. (a) R. Radeglia, S. Dahne, J. Molec. Struct., 1970, 5, 399; (b) S. Dahne, Chimia, 1991, 45, 288; (c) S. R. Marder, J. W. Perry, B. G. Tiemann et al., J. Am. Chem. SOC.,1993, 115, 2524; (d) S. R. Marder, C. B. Gorman, B. G. Tiemann, L.-T. Cheng, J. Am. Chem. SOC., 1993, 115, 3006. 124. R. S. Becker, J. S. de Melo, A. L. Macanita, F. Elisei, Pure & Appl. Chem., 1995, 67, 9.

References

177

125. P. M. Lahti, J. Obrzut, F. E. Karasz, Macromolecules, 1987, 20, 2023. 126. A. F. Diaz, J. Crowley, J. Bargon, G . P. Gardini, J. B. Torrance, J . Electroanal. Chem., 1981, 121, 355. 127. (a) D. Fichou, G. Horowitz, B. Xu, F. Garnier, Springer Ser. Solid State Sci., 1989, 91, 386; (b) D. Fichou, G. Horowitz, F. Garnier, Synth. Met., 1990, 39, 125; (c) D. Fichou, G. Horowitz, B. Xu, F. Garnier, Synth. Met., 1990, 39, 243; (d) D. Fichou, B. Xu, G. Horowitz, F. Garnier, Synth. Met., 1991, 41, 463. 128. (a) A. Alberti, L. Favaretto, G. Seconi, J. Chem. Soc., Perkin Trans. 2, 1990, 931; (b) C. E. Brillas, A. G. Davies, L. Fajari et al., J. Org. Chem., 1993, 58, 3091. 129. P. Enzel, T. Bein, Synth. Met., 1993, 55, 1238. 130. (a) G. Horowitz, X. Peng, D. Fichou, F. Garnier, J . Mol. Electron., 1991, 7, 85; (b) F. Garnier, F. Deloffre, G. Horowitz, R. Hajlaoui, Synth. Met., 1993, 57, 4747; (c) H. Akimichi, K. Waragai, S. Hotta, H. Kano, H. Sakaki, Appl. Phys. Lett., 1991,58, 1500. 131. J. Paloheimo, H. Stubb, L. Gronberg, Synth. Met., 1993, 57, 4198. . 132. (a) J. P. Reyftmann, J. Kagan, R. Santus, P. Morliere, Photochem. Photobiol., 1985,41, 1; (b) C. Evans, D. Weir, J. C. Scaiano et al., Photochem. Photobiol., 1986, 44, 441. 133. (a) H. Chosrovian, S. Rentsch, D. Grebner, D. U. Dahm, E. Birckner, H. Naarmann, Synth. Met., 1993, 60, 23; (b) S. Rentsch, H. Chosrovian, D. Grebner, H. Naarmann, Synth. Met., 1993, 57, 4740. 134. D. Oelkrug, H.-J. Egelhaaf, J. Fluorescence, 1995, 5, 165. 135. (a) D. Fichou, F. Garnier, F. Charra, F. Kajzar, J. Messier, Spec. Publ. Roy. Soc., 1989, 69, 176; (b) M. T. Zhao, M. Samoc, B. P. Singh, P. N. Prasad, J . Phys. Chem., 1989, 93, 7916. 136. E. Pellegrin, H. Fritzsche, N. Nucker et al., Synth. Met., 1991, 41, 1207. 137. (a) H. Fujimoto, U. Nagashima, H. Inokuchi et al., J. Chem. Phys., 1990, 92, 4077; (b) H. Fujimoto, U. Nagashima, H. Inokuchi et al., Phys. Scr., 1990, 41, 105; (c) D. Jones, M. Guerra, L. Faveretto, A. Modelli, M. Fabrizio, G. Distefano, J. Phys. Chem., 1990, 94, 5761. 138. R. Lazzaroni, A. J. Pal, S. Rossini, G . Ruani, R. Zamboni, C. Taliani, Synth. Met., 1991,42, 2359. 139. S. Destri, M. Mascherpa, W. Porzio, Adv. Muter., 1993, 5, 43. 140. S. Hotta, K. Waragai, Synth. Met., 1989, 32, 395. 141. B. Fabre, G. Bidan, Adv. Muter., 1993, 5, 646. 142. D. Oeter, C. Ziegler, W. Gopel, Synth. Met., 1993, 61, 231. 143. H. Nakahara, J. Nakayama, M. Hoshino, K. Fukuda, Thin Solid Films, 1988, 160, 87. 144. A. Carpita, R. Rossi, Gazz. Chim. Ztal., 1985, 115, 575. 145. N. Jayasuriya, J. Kagan, Heterocycles, 1986, 24, 2261. 146. N. Jayasuriya, J. Kagan, Heterocycles, 1986, 24, 2901. 147. J. Kagan, S. K. Arora, I. Prakash, A. Uestuenol, Heterocycles, 1983, 20, 1341. 148. J. H. Uhlenbroek, J. D. Bijloo, Rec. Trav. Chim. Pays-Bas, 1960, 79, 1181. 149. J.-P. Beny, S. N. Dhawan, J. Kagan, S. Sundlass, J . Org. Chem., 1982, 47, 2201. 150. H. Wynberg, A. Bantjes, J. Am. Chem. Soc., 1959, 84, 1421. 151. (a) K. Y. Yen, G. G. Miller, R. L. Elsenbaumer, J . Chem. SOC,Chem. Commun., 1986, 1346; (b) M. Sato, S. Tanaka, K. Kaeriyama, J. Chem. SOC,Chem. Commun., 1986, 873. 152. R. D. McCullough, R. D. Lowe, M. Jayaraman, P. C. Ewbank, D. L. Anderson, S. TristranNagle, Synth. Met., 1993, 55, 1198. 153. T.-A. Chen, R. D. Rieke, J . Am. Chem. Soc., 1992, 114, 10087. 154. T.-A. Chen, R. D. Rieke, Synth. Met., 1993, 60, 175. 155. D. Delabouglise, M. Hmyene, G. Horowitz, A. Yassar, F. Garnier, Adv. Muter., 1992,4, 107. 156. (a) W. ten Hoeve, H. Wynberg, E. E. Havinga, E. W. Meijer, J . Am. Chem. SOC.,1991, 113, 5887; (b) E. E. Havinga, I. Rotte, E. W. Meijer, W. ten Hoeve, H. Wynberg, Synth. Met., 1991, 41-43, 473. 157. R. L. Elsenbaumer, K. Y. Yen, R. Oboodi, Synth. Met., 1986, 15, 169. 158. A. Yassar, D. Delabouglise, M. Hmyene, B. Nessak, G. Horowitz, F. Garnier, Adv. Muter., 1992, 4, 490. 159. P. Bauerle, F. Pfau, H. Schlupp et al., J . Chem. SOC.Perkin Trans. 2, 1993, 489. 160. J. H. Uhlenbrock, J. D. Bijloo, Recl. Trav. Chim. Pays-Bas, 1960, 79, 1181.

178

3 Oligothiophenes

161. R. Kellogg, A. P. Schaap, H. Wynberg, J . Org. Chem., 1969, 34, 343. 162. R. F. Curtis, G. T. Phillips, J. Chem. Soc., 1965, 5134. 163. P. Bauerle, U. Segelbacher, K.-U. Gaudl, D. Huttenlocher, M. Mehring, Angew. Chem., 1993, 105, 125; Angew. Chem. Int. Ed. Engl., 1993, 32, 76. 164. (a) M. G. Hill, K. R. Mann, L. L. Miller, J.-F. Penneau, J . Am. Chem. SOC.,1992,114,2728; (b) M. G. Hill, J.-F. Penneau, B. Zinger, K.R. Mann, L.L. Miller, Chem. Muter., 1992, 4, 1106; (c) B. Zinger, K. R. Mann, M. G. Hill, L. L. Miller, Chem. Muter., 1992, 4 , 1113. 165. A. Stabel, J. P. Rabe, Synth. Met., 1994, 67, 47. 166. J. K. Herrema, H. Wildeman, F. van Bolhuis, G. Hadziioannou, Synth. Met., 1993,60,239. 167. (a) M. Sato, M. Hiroi, Chem. Lett., 1994, 745; (b) M. Sato, M. Hiroi, Chem. Lett., 1994,985; (c) M. Sato, M. Hiroi, Chem. Lett., 1994, 1649; (d) M. Sato, M. Hiroi, Synth. Met., 1995, 71, 2085. 168. G. Tourillon, in Handbook of Conducting Polymers, T. A. Skotheim (Ed.), Dekker, New York, 1986. 169. P. Bauerle, T. Fischer, B. Bidlingmaier, A. Stabel, J. P. Rabe, Angew. Chem., 1995,107, 335; Angew. Chem. Int. Ed. Engl., 1995, 34, 303. 170. A. Amer, A. Burkhardt, A. Nkansah et al., Phosph. Sul$ Silic., 1989, 42, 63. 171. J. L. Brkdas, R. Silbey, D. S . Boudreaux, R. R. Chance, J. Am. Chem. SOC., 1983, 105, 6555. 172. D. D. Cunningham, L. Laguren-Davidson, H. B. Mark, Jr., C. V. Pham, H. Zimmer, J . Chem. SOC.,Chem. Commun., 1987, 1021. 173. L. Laguren-Davidson, C. V. Pham, H. Zimmer, H. B. Mark, Jr., J . Electrochem. Soc., 1988, 135, 1406. 174. H. Matsuda, K. Kaeriyama, H. Suezawa, M. Hirota, J . Polym. Sci.: Part A , Polym. Chem., 1992, 30, 945. 175. P. Bauerle, M. Hiller, S. Scheib, M. Sokolowski, E. Umbach, Adv. Muter., 1996, 7, 214. 176. P. Bauerle, S . Scheib, Adv. Muter., 1993, 5 , 848. 177. C. Arbizzani, A. Bongini, M. Mastragostino, A. Zanelli, G. Barbarella, M. Zambianchi, Adv. Muter., 1995, 7, 571. 178. G. Barbarella, A. Bongini, M. Zambianchi, Macromolecules, 1994, 27, 3039. 179. H. Mao, B. Xu, S . Holdcroft, Macromolecules, 1993, 26, 1163. 180. G. Barbarella, M. Zambianchi, Tetrahedron, 1994, 50, 11249. 181. M. Schon, Diploma Thesis, Universitat Wurzburg, 1996. 182. A. Iraqi, J. A. Crayston, J. C. Walton, J. Muter. Chem., 1995, 5, 1831. 183. E. E. Havinga, L. W. van Horssen, Makromol.Chem., Makromol. Symp., 1989, 24, 67. 184. J. Roncali, A. Gorgues, M. Jubault, Chem. Muter., 1993, 5, 1456. 185. J. Kankare, J. Lukkari, P. Pasanen, R. Sillanpaa, H. Laine, K. Harrnaa, Macromolecules, 1994, 27, 4327. 186. T. Bennicori, E. Brenna, F. Sannicolo et al., J . Chem. SOC.Chem. Commun., 1995, 881. 187. J. C. Home, G. J. Blanchard, E. LeGoff, J. Am. Chem. Soc., 1995, 117, 9551. 188. L. DeWitt, G. J. Blanchard, E. LeGoff, M. E. Benz, J. H. Liao, M. G. Kanatzidis, J. Am. Chem. Soc., 1993, 115, 12158. 189. G. Barbarella, M. Zambianchi, A. Bongini, L. Antolini, Adv. Muter., 1994, 6, 561. 190. G. Barbarella, A. Bongini, M. Zambianchi, Adv. Muter., 1991, 3, 494. 191. G. Barbarella, M. Zambianchi, A. Bongini, L. Antolini, Adv. Muter., 1993, 5, 282. 192. G . Barbarella, M. Zambianchi, A. Bongini, L. Antolini, Adv. Muter., 1993, 5 , 834. 193. P. Audebert, J.-M. Catel, G. Le Coustumer, V. Duchenet, P. Hapiot, J. Phys. Chem., 1995, 99, 11923. 194. J.-H. Liao, M. Benz, E. LeGoff, M. G. Kanatzidis, Adv. Muter., 1994, 6, 135. 195. J. Roncali, M. Giffard, M. Jubault, A. Gorgues, J . Electroanal. Chem., 1993, 361, 185. 196. K. Faid, M. Leclerc, J . Chem. SOC.,Chem. Commun., 1993, 962. 197. G. Barbarella, A. Bongini, M. Zambianchi, Tetrahedron, 1992, 48, 6701. 198. K. A. Murray, S. C. Moratti, D. R. Baigent et al., Synth. Met., 1995, 69, 395. 199. A. Kraak, A. K. Wiersma, P. Jordens, H. Wynberg, Tetrahedron, 1968, 24, 3381. 200. G. Zerbi, R. Radaelli, M. Veronelli, E. Brenna, F. Sannicolo, G. Zotti, J. Chem. Phys., 1993, 98, 4531.

References

179

201. (a) G. Zotti, G. Schiavon, A. Berlin, G. Pagani, Macromolecules, 1994,27, 1938; (b) G. Zotti, A. Berlin, G. Pagani, G. Schiavon, S. Zecchin, Adv. Muter., 1994, 6, 231. 202. T. Pilati, Acta Cryst., 1995, C51, 690. 203. J. Roncali, C. Thobie-Gautier, Adv. Muter., 1994, 6, 846. 204. T. L. Lambert, J. P. Ferraris, J. Chem. SOC.Chem. Commun., 1991, 752. 205. (a) J. Roncali, H. Brisset, C. Thobie-Gautier, M. Jubault, A. Gorgues, J. Chim. Phys., 1995, 92,771; (b) H. Brisset, C. Thobier-Gautier, A. Gorgues, M. Jubault, J. Roncali, J . Chem. SOC. Chem. Commun., 1994, 1305. 206. J. P. Ferraris, T. L. Lambert, J. Chem. Soc., Chem. Commun., 1991, 1268. 207. J. P. Ferraris, J. Henderson, D. Torres, D. Meeker, Synth. Met., 1995, 72, 147. 208. G. Zotti, G. Schiavon, S. Zecchin, A. Berlin, G. Pagani, Synth. Met., 1994, 66, 149. 209. G. Zotti, A. Berlin, G. Pagani, G . Schiavon, S. Zecchin, Adv. Muter., 1995, 7, 48. 210. H. Brisset, C. Thobie-Gautier, M. Jubault, A. Gorgues, J. Roncali, J . Chem. SOC.,Chem. Commun., 1994, 765. 211. (a) G. Zotti, G. Schiavon, A. Berlin, G. Pagani, Chem. Muter., 1993, 5, 430; (b) G. Zotti, G. Schiavon, A. Berlin, G. Pagani, Chem. Muter., 1993, 5 , 620; (c) G. Zotti, G. Schiavon, A. Berlin, G. Pagani, Adv. Muter., 1993, 5, 551. 212. S. Hotta, K. Waragai, Adv. Muter., 1993, 5, 896. 213. S. Hotta, K. Waragai, J. Muter. Chem., 1991, I , 835. 214. S. Hotta, K. Waragai, J. Phys. Chem., 1993, 97, 7427. 215. K. Tanaka, Y. Matsuura, Y. Oshima, T. Yamabe, S. Hotta, Synth. Met., 1994, 66, 295. 216. F. Garnier, A. Yassar, R. Hajlaoui et al., J . Am. Chem. SOC.,1993, 115, 8716. 217. A. Yassar, F. Garnier, F. Deloffre, G. Horowitz, L. Ricard, Adv. Muter., 1994, 6, 660. 218. G. Horowitz, P. Delannoy, H. Bouchriha et ul., Adv. Mater., 1994, 6, 752. 219. J. P. Parakka, M. P. Cava, Tetrhedron, 1995, 51, 2229. 220. C. Aleman, E. Brillas, A. G. Davies et ul., J. Org. Chem., 1993, 58, 3091. 221. (a) J. L. Sauvaljol, C. Chorro, J.-P. L6re-Porte et al., Synth. Met., 1994,62,233; (b) P. Hapiot, L. Gaillon, P. Audebert, J. J. E. Moreau, J.-P. Ltke-Porte, M. Wong Chi Man, Synth. Met., 1995, 72, 129. 222. C. Soucy-Breau, A. MacEachern, L. C. Leitch, T. Arnason, P. Morand, J. Heterocyclic Chem., 1991, 28, 41 I . 223. P. Bauerle, U. Segelbacher, A. Maier, M. Mehring, J. Am. Chem. SOC.,1993, 115, 10217. 224. A. S . Davidov, Theory of Molecular Excitons, McGraw-Hill, New York, 1962. 225. B. Nessakh, G. Horowitz, F. Garnier, F. Deloffre, P. Srivastava, A. Yassar, J. Electrounal. Chem., 1995, 399, 97. 226. H. Haberkorn, H. Naarmann, K. Penzien, J. Schlag, P. Simak, Synth. Met., 1982, 5 , 51. 227. (a) H.-J. Egelhaaf, P. Bauerle, K. Rauer, V. Hoffmann, D. Oelkrug, J. Mol. Struct., 1993, 293, 249; (b) H.-J. Egelhaaf, P. Bauerle, K. Rauer, V. Hoffmann, D. Oelkrug, Synth. Met., 1993, 61, 143. 228. M. Bennati, A. Grupp, M. Mehring, P. Bauerle, J. Phys. Chem., 1996, 100, 2849. 229. (a) C. Ehrendorfer, H. Neugebauer, A. Neckel, P. Bauerle, Synth. Met., 1993, 55, 493; (b) C. Ehrendorfer, A. Karpfen, P. Bauerle, H. Neugebauer, A. Neckel, J. Molec. Struct., 1993, 298, 65. 230. C. Ehrendorfer, H. Neugebauer, P. Bauerle, A. Neckel, Synth. Met., 1995, 69, 393. 231. M. Stoldt, P. Bauerle, H. Schweizer, E. Umbach, Synth. Met., 1993, 57, 4059. 232. M. Stoldt, P. Bauerle, H. Schweizer, E. Umbach, Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A , 1994, 240, 127. 233. H. Neureiter, W. Gebauer, C. Vaterlein, M. Sokolowski,P. Bauerle, E. Umbach, Synth. Met., 1994, 67, 173. 234. C. Vaterlein, H. Neureiter, W. Gebauer et al., J. Appl. Phys., submitted. 235. (a) D. Braun, G. Gustafsson, D. McBranch, A. J. Heeger, J. Appl. Phys., 1992, 72, 546; (b) M. Bergreen, G. Gustafsson, 0. Inganas, M. R. Anderson, 0. Wennerstrom, T. Hjertberg, Adv. Muter., 1994, 6, 488. 236. G. Horowitz, P. Delannoy, H. Bouchriha et al., Adv. Muter., 1994, 6, 752. 237. H.-J. Egelhaaf, D. Oelkrug, J. ofSPZE, 1996, 2362, 398. 238. K. Uchiyama, H. Akimichi, S. Hotta, H. Noge, H. Sakaki, Synth. Met., 1994, 63, 57.

180

3 Oligothiophenes

239. A. Soukopp, K. Glockler, P. Bauerle, M. Sokolowski, E. Umbach, Adv. Muter., 1996, 7, 902. 240. (a) J. M. Tour, R. Wu, Macromolecules, 1992, 25, 1901; (b) J. M. Tour, Chem. Rev., 1996, 96, 537. 241. L.-T. Cheng, J. M. Tour, R. Wu, P. V. Bedworth, Nonlinear Opt., 1993, 6, 87. 242. (a) J. Guay, P. Kasai, A. Diaz, R. Wu, J. M. Tour, L. H. Dao, Chem. Muter., 1992, 4, 1097; (b) J. M. Tour, J. Guay, A. Diaz, R. Wu, L. H. Dao, Chem. Mater., 1992, 4, 254. 243. A. Aviram, J. Am. Chem. SOC.,1988, 110, 5687. 244. (a) J. M. Tour, R. Wu, J. S . Schumm, J . Am. Chem. SOC.,1990, 112, 5662; (b) J. M. Tour, R. Wu, J. S . Schumm, J. Am. Chem. SOC.,1991,113, 7064. 245. (a) J. M. Tour, J. Guay, A. Diaz, R. Wu, J. Am. Chem. Soc., 1993, 115, 1869; (b) J. R. Diers, M. K. DeArmond, J. Guay et al., Chem. Muter., 1994, 6, 327. 246. (a) B. Kirste, P. Tian, G. Kossmehl, G. Engelmann, W. Jugelt, Magn. Res. Chem., 1995, 33, 70; (b) G. Engelmann, Dissertation, FU Berlin 1995. 247. G. Heywang, F. Jonas, Adv. Muter., 1992, 4, 116. 248. M. Hasik, J. E. Laska, A. Pron, I. Kulszewicz-Bajer, K. Koziel, M. Lapkowski, J . Polym. Sci., A: Polym. Chem., 1992, 30, 1741. 249. R. Cloutier, M. Leclerc, Synth. Met., 1993, 55-57, 1272. 250. G. Zotti, M. C. Gallazzi, G. Zerbi, S. V. Meille, Synth. Met., 1995, 73, 217. 251. (a) L. L. Miller, Y . Yu, J. Org. Chem., 1995, 60, 6813; (b) L. L. Miller, Y. Yu, E. Gunic, R. Duan, Adv. Mater., 1995, 7, 547. 252. Y. Yu, E. Gunic, L. L. Miller, Chem. Muter., 1995, 7, 255. 253. F. Effenberger, Acc. Chem. Res., 1989, 22, 27. 254. (a) P. Bauerle, G. Gotz, A. Synowczyk, J. Heinze, Lieb. Ann., 1996, 279; (b) H. Theobald, V. Harries, U. Kardorff, P. Bauerle, G. Gotz, Ger. Offen. DE 4.201.305 (Chem. Abstr., 1993, 119, 225805s). 255. (a) U. Folli, D. Iarossi, M. Montorsi, A. Mucci, L. Schenetti, J . Chem. Soc., Perkin Trans. I , 1995, 537; (b) G. Barbarella, M. Zambianchi, L. Antolini et al., J. Chem. Soc., Perkin Trans. 2, 1995, 1869. 256. G. Barbarella, M. Zambianchi, R. DiToro, M. Colonna, L. Antolini, A. Bongini, Adv. Muter., 1996, 8, 327. 257. P. Bauerle, G. Gotz, S . Scheib, R. Klose, J. Heinze, Europ. J . Org. Chem., submitted. 258. H. Muguruma, T. Saito, S. Sasaki, S. Hotta, I. Karube, J. Heterocyclic Chem., 1996,33, 173. Chem. Commun., 1970, 1453. 259. F. Wudl, G. M. Smith, E. J. Hufnagel, J. Chem. SOC., 260. (a) J. Roncali, M. Giffard, P. Fr&e, M. Jubault, A. Gorgues, J. Chem. SOC.,Chem. Commun., 1993, 689; (b) J. Roncali, M. Giffard, M. Jubault, A. Gorgues, Synth. Met., 1993, 60, 163. 261. Y. Wie, B. Wang, W. Wang, J. Tian, Tetrahedron Letters, 1995, 36, 665. 262. J. Nakayama, T. Fujimori, Sulfur Letters, 1990, 11, 29. 263. T. Nihira, S . Tarutani, K. Takase, K. Takahashi, Heterocycles, 1995, 41, 2169. 264. W. M. Albers, G. W. Canters, J. Reedijk, Tetrahedron, 1995, 51, 3895. 265. (a) F. Effenberger, F. Wiirthner, Angew. Chem., 1993, 105, 142; Angew. Chem. h t . Ed. Engl., 1993, 32, 719; (b) F. Effenberger, F. Wiirthner, F. Steybe, J. Org. Chem., 1995, 60, 2082. 266. F. Wiirthner, F. Effenberger, R. Wortmann, P. Kramer, Chem. Phys., 1993, 173, 305. McGeein, J. 0. Morley, J. Zyss, I. Ledoux, J . Chem. 267. M. G. Hutchings, I. Ferguson, D. .I. SOC.,Perkin Trans. 2, 1995, 171. 268. F. Wiirthner, M. S . Vollmer, F. Effenberger et al., J. Am. Chem. SOC.,1995, 117, 8090. 269. P. Emele, D. U. Meyer, N. Holl et al., Chem. Phys., 1994, 181, 417. 270. H. Segawa, F.-P. Wu, N. Nakayama et al., Synth. Met., 1995, 71, 2151. 271. (a) M. Schmelzer, S. Roth, P. Bauerle, R. Li, Thin Solid Films,1993,229,255;(b) M. Schmelzer, M. Burghard, P. Bauerle, S. Roth, Synth. Met., 1993, 61, 97; (c) M. Schmelzer, M. Burghard, P. Bauerle, S . Roth, Thin Solid Films, 1994,243, 620. 272. H. Nakahara, J. Nakayama, M. Hoshino, K. Fukuda, Thin Solid Films, 1988, 160, 87. 273. M. Schmelzer, M. Burghard, P. Bauerle, S . Roth, Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A , 1994, 252-253, 465.

References 274. 275. 276. 277.

18 1

B. Liedberg, Z. Yang, I. Engquist et al., J. Chem. Phys., 1997, 101, 5951. J. M. Tour, L. Jones 11, D. L. Pearson et al., J. Am. Chem. Soc., 1995, 117, 9529. D.D. Graf, N.C. Day, K.R. Mann, Znorg. Chem., 1995, 34, 1562. D. Fichou, B. Bachet, F. Demanze, I. Billy, G. Horowitz, F. Garnier, Adv. Muter., 1996, 8, 500.

This Page Intentionally Left Blank

4 Structure and Properties of Oligothiophenes in the Solid State: Single Crystals and Thin Films Denis Fichou and Christiane Ziegler

4.1 Introduction Well-defined oligomers are often used as model compounds to rationalize or predict the structure and properties of polymers. An infinity of virtual oligomers with various chain lengths and conformations can thus be easily designed and computed. This is the case of 7r-conjugated polymers such as polyacetylene or polythiophene whose conducting and semiconducting properties have been modelled as a function of chain length and conformation on short finite oligomers [l, 21. In some cases, series of oligomers with various length and bearing various substituents have been successfully prepared by organic chemists. The largest oligomeric family so far among conjugated organic compounds is probably that of oligothiophenes. This is a direct consequence of the extremely rich chemistry of the aromatic thiophene ring which has been extensively used to synthesize a large variety of p-conjugated a-linked oligothiophenes (a-nT, with iz = number of thiophene rings in the oligomeric sequence, see Scheme 1) [3-51.

R

R

R

Scheme 1

The first driving force behind this enormous synthetic effort was the potential use of short oligothiophenes ( n = 2,3) as biologically active compounds. Many bi- and terthiophene derivatives exhibit biological activities such as phototoxicity against nematodes [6] and other microorganisms [7, 81, mammalian skin pigmentation [9], seed germination inhibition [ 101or antibiotic and antiviral properties in the presence of UVA light [ l l , 121. The second motivation of the recent development of oligothiophenes is related to their use as materials for electronic device applications. As early as 1974, a precursor article by H. Kuhn et al. [13] describes photocurrent measurements on LangmuirBlodgett films of a-quinquethiophene (a-5T). In the mid-l980s, oligothiophenes

184

4 Structure and Properties of Oligothiophenes in the Solid State

were used as model compounds as well as monomers toward electrically conducting polythiophenes [14-161. But the true starting point of this second generation of oligothiophenes is the discovery in 1988 of the charge transport properties of a-sexithiophene (a-6T), the conjugated linear all-a-linked hexamer of thiophene [17, 181. A year later, evaporated thin films of a-6T were used as the p-type semiconductor of an organic field-effect transistor (OFET) [19, 201. The mobilities of the majority of carriers measured in these devices were in the range p = 10-4cm2/Vs, i.e. one to two orders of magnitude higher than those of PT-based OFETs [21, 221 but still substantially lower than those of conventional a-SiH based MISFETs [23]. At that time, it was also demonstrated that an all-organic a-6T OFET could be fabricated on a flexible substrate, thus opening the era of 'plastic electronics' [24]. Since the early 1990s, a considerable number of studies aimed at investigating the properties of oligothiophenes have appeared in the literature. In particular, emphasis has been put on the electronic structure [25-281 and charge transport properties [29-321 of semiconducting a-nT thin films. These studies are essentially motivated by their implications on charge transport in thin film devices such as OFETs [33]. In particular, it has been demonstrated that a-6T based OFETs show an improved carrier mobility when long range molecular ordering is achieved [34]. The highest mobilities (pFE = 0.04cm2V-' s-l) are obtained in OFET devices using highly oriented thin films and are close to that measured on a-6T single crystals (pFE = 0.16 cm2V-' s-l [35]). This shows that charge transport between source and drain of an OFET occurs essentially through molecular channels of a-6T molecules oriented perpendicular to the substrate and having the herringbone arrangement found in the single crystal. Beside these fundamental studies, the search for efficient OFETs recently engaged a number of laboratories worldwide. Great progress has thus been accomplished in the understanding and fabrication of these devices resulting in higher mobilities, reduced size, and easy processability [36]. Finally, a-oligothiophenes have also been used in other devices such as light-emitting diodes [37, 381, spatial light modulators [39], electro-optical modulators [40-421 and photovoltaic cells [43] summarized by Granstrom et al. in Chapter 8 of this book [44]. The aim of this chapter is to review the structure and properties of solid state oligothiophenes, particularly single crystals and vacuum-deposited thin films. Particular attention is given to the influence of chain length and substitution keeping in mind their potential applications in electronic devices.

4.2 Single crystals 4.2.1 General description The crystal structure of most a-oligothiophenes has been investigated recently, in contrast to oligophenylenes for which a variety of structural data have been known

4.2 Single crystals

185

for two decades [45-481. With the exception of a-septithiophene (a-7T), single crystals of non-substituted a-oligothiophenes have been grown and characterized up to the octamer a-8T. The structure of various end- and side-substituted a-nT derivatives has also been determined, such as for example 2,5”’-dimethylquaterthiophene[49] or 3’,3””,4‘,4””-tetrabutylsexithiophene [50]. Beside experimental studies, computational procedures have been used in some cases to confirm, refine or predict the crystal structure of a-nTs. One important contribution of theoretical simulations is to identify the origin of crystal packing in terms of intra- and intermolecular forces. In the crystalline form, all non-substituted oligothiophenes are quasi-planar, i.e. the inter-ring torsion angle being below the measurement accuracy (
1.44 382.8

1.503 2195.2

8 15.225 5.635 25.848 98.15

m / c

~511 monoclinic

a-3T

1.50

[531 monoclinic P2l/a 4 30.52 7.86 6.09 91.8 72

a-4T

8.936 5.750 14.341 97.22 55.67 31.5 1.501 731.1

2

[61, 621 monoclinic P2 I /a

a-4T/HT

1.55

[531 monoclinic P21/a 4 39.00 7.77 6.00 97.7 65

a-5T

4,tilt angle of the long molecular axis with a. 'D, calculated density.

[59] monoclinic P2Ijc 2 7.873 5.771 8.813 107.07

a r,herringbone angle.

Reference System

a-2T

~

5.68 20.67 97.78 55 41.5 1.55 1064.2

1661 monoclinic

a-6T/HT

HT, high temperature polymorph.

7.85 6.03 90.8 66 23.5 1.553 21 16.5

1641 monoclinic

a-6T

Table 1. Comparative experimental structural data of non-substituted oligothiophenes single crystals.

90.3 65 24 1.578 2773.0

6.00

7.84

1651 monoclinic

a-8T

s

5

a;

$?

6

S'

0

5

4 k

2.

? 2

G

2J

4!a

4.2 Single crystals

187

Beside oligomers, a number of structural studies have been performed on polythiophene and its 3-alkylated derivatives, both in their neutral and doped forms, but the determination of the structure of PTs is limited by their inherent low crystallinity [67]. Nevertheless, two stable conformations, a nearly planar all-trans rod and a all-cis helical or coil geometries, can be predicted from theoretical calculations for polythiophene [68].

4.2.2 X-ray structures 4.2.2.1 Bithiophene (a-2T) and derivatives The structural parameters of the 2,2-bithiophene (a-2T) molecule (bond lengths and angles) were first reported in the early 1950's by Bastiansen and Almenningen in a series of articles [69, 701. These parameters were obtained from electron diffraction measurements on a-2T vapour and are in good agreement with values obtained on thiophene by microwave investigation [71]. In the gas phase, the molecule is not planar and has an anti conformation with a twist angle of 34" that the authors attribute to steric hindrance between the S atom and the nearest H atom. In 1968, after many attempts Visser et al. obtained suitable crystals of three bithiophene geometric isomers (Scheme 2) from solutions [%I. These authors were able to grow pale green crystals of a-2T in toluene, a,P-2T in light petroleum and /3,@-2Tin benzene by slow evaporation of the solute at room temperature. Crystals of a-2T and a,P-2T were obtained at room temperature and those of P,P-2T at -20°C. Since crystals of a-2T (mp = 32-33°C) decompose under X-ray exposure, crystallographic data have been collected at low temperature (- 140°C).

Scheme 2

In the crystaI, the a-2T molecules are planar in the anti conformation and situated in a twofold position so that there is only one independent thiophene ring per cell. The space group is P2& and the number of molecules per unit cell is Z = 2. We will see that most non-substituted oligothiophenes adopt the P2,/c space group. The projection of the a-2T structure along the short b axis is represented in Fig. 1 (Top). The central bond of the molecule lies approximately on parallel (202) planes. The angle between the plane of the molecule and the short b axis is 29.7" while the angle between two molecular planes is 59". All molecules have the same tilt angle (7.4") relative to the (101) crystallographic direction. The structure of a-2T has recently been re-investigated with a higher degree of accuracy by Chaloner et al. at 173K [59] and by Pelletier and Brisse at 133K [60].

188

4 Structure and Properties of Oligothiophenes in the Solid State

.............

................... ...................................

................. ’.( ....................

4

-6.3A-

-

I

1

1. . . . . . . . . . . . . . . . . . . . .

........

. . . . . . . . . . . . . . . . . .)... ..... ...........I ...................

-t 4

-i 4

Figure 1. Projections of the structure of 2,2‘-bithiophene (a-2T, top), 2,3’-bithiophene (a,P-2T, middle) and 3,3/-bithiophene u,P-2T, bottom) along the short crystal axis (b axis for a-2T and @2T; c axis for P,P-ZT). From ref. [58].

4.2 Single crystals

189

Figure 2. Stereographic view of the packing of the a-2T crystal with the ac face in the plane of the figure. From ref. [60].

The crystals are either grown by sublimation [59] or solvent recrystallization [60]. The molecules pack in the HB geometry with molecules alternately tilted along the diagonal of the ac face (see Fig. 2). If these structures are broadly similar to those found by Visser et al., some discrepancies appear between the three studies. The main one concerns the unit cell dimensions which vary significantly (up to 1.5%) depending on the crystal growth technique (see Table 2). The greater variation is observed for the c axis. As noted in the two most recent studies [59, 601, these differences may be explained by the coexistence of two a-2T phases. The first phase would be fully ordered and stable while the second would be partially disordered and unstable. Finally, crystals of the two geometrical isomers 2,3’-bithiophene and 3,3’-bithiophene show a high degree of disorder and no information on their molecular conformation could be obtained. Table 2. Crystal data of 2,2’-bithiophene (a-2T) single crystals.

Data taken from Reference:

a D,

calculated density.

58

59

60

133 monoclinic p2 1/ c 2 7.16 5.90 8.91 106.6

173 monoclinic P21,C 2 7.873 5.171 8.813 107.07 1.44 382.8

133 monoclinic P2l/c 2 1.134 5.729 8.933 106.72 1.451 379.0

190

4 Structure and Properties of Oligothiophenes in the Solid State

4,4’-dimethyl- and 4,4/-bis(methylsulfanyl)-2,2’-bithiophenes Recently, Barbarella et al. have studied the conformation of a series of 3,3’-, 3,4’and 4,4/-dimethyl-and 4,4/-bis(methylsulfanyl)-2,2/-bithiophenes [72]. Particularly, the X-ray crystal structure of the two tail-to-tail 4,4”-substituted derivatives (see Scheme 3) has been investigated. Single crystals are grown by slow evaporation of cyclohexane solution. Similarly to most 2,2’-bithiophenes, both compounds crystallize with an anti coplanar conformation in the monoclinic system (2 = 2). One of the most interesting observations by Barbarella et al. is the large dissymmetry of the thiophene rings. This demonstrates the ‘plasticity’of oligothiophenes, i.e. the capability of these compounds to minimize their conformational strain and adopt unusual (syn) conformations in the solid state. Further examples of this plasticity have also been observed with longer oligomers.

Scheme 3

3,3/- and 4,4’-dipentoxy-2,2/-bithiophenes In a very similar way, Meille et al. report on the conformational flexibility of 2,2’bithiophene derivatives substituted by long alkoxy side chains [73]. According to (Scheme 4) crystallize in these authors, 3,3’- and 4,4/-dipentoxy-2,2/-bithiophenes the monoclinic system (space group P2&) with coplanar anti-parallel conformation. The head-to-head 3,3’-derivative packs in parallel layers, an unusual arrangement among 2,2/-bithiophenes, while the tail-to-tail one adopt the more conventional herringbone pattern (Fig. 3). Comparable geometries are observed for the crystals of 3,3/- and 4,4/-dimethoxy analogs [74, 751.

0‘ I

C5HI I Scheme 4

Other 2,2’-bithiophene derivatives Finally, the structure of various other substituted 2,2/-bithiophenes have also been investigated recently. This is particularly the case of 5,5‘-dimethyl-2,2’bithiophene and 5,5’-bis(trimethylsilyl)-2,2‘-bithiophene[76], (3’,4,4/,5/-tetramethyl2,2’-bithiophene-5-y1)-(3~,4,4~,5,5~-pentamethyl-2,2~-bithiophene-3-yl)methane [77], the solvatochromic 5-dimethylamino-5/-nitro-2,2/-bithiophene [78], and a bithiophene-derived annulene [79].

4.2 Single crystals

c

191

2

Figure 3. Crystal packing of (a) 3,3’-dipentoxy-2,2‘-bithiophenes and (b) 4,4’-dipentoxy-2,2’-bithiophenes viewed along respectively b and c. From [73].

4.2.2.2 a-Terthiophene ( a 3 T ) and derivatives Up to now, there have been very few experimental studies on the a-3T crystal structure as for other oligomers of odd number of thiophene units a-5T and a-7T. The X-ray structure of a-3T (Scheme 5 ) has been solved by Van Bolhuis et al. in 1989 using thin plates crystals grown by slow evaporation (one week) of an ether solution [5 11. It reveals two identical but crystallographically-independent molecules in the asymmetric unit cell which contains eight molecules ( Z = 8) as shown in Fig. 4.In contrast to other oligothiophenes, a-3T is not strictly planar with torsion

Scheme 5

192

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 4. Crystal packing of a-terthiophene (a-3T).From ref. [51].

angles of 6-9" but thiophene rings are in an anti-parallel conformation. The molecules are in the HB structure with an angle of 62". One essential parameter for charge transport properties in solid state oligothiophenes is the distance between molecules. In the a-3T crystal, the shortest S. . .S distances measured by Van Bolhuis et al. a;e those betwFen the outer sulfur atoms of adjacent identical molecules, i.e. 3.690A and 3.71 1 A respectively [51]. All other S. . .S distances are 4.1 or longer.

A

3',4'-dibutyl-2,2/:5',2''-terthiophene The X-ray crystal structure of 3',4'-dibutyl-2,2':5',2/'-terthiophene (Scheme 6), a model compound of soluble poly(alkylthiophenes), has been determined by DeWitt et al. [go]. It reveals an all-anti conformation with a 33" mean dihedral angle between adjacent rings. Non-planarity is ascribed to the presence of the n-butyl groups which are fully extended in a trans conformation in order to minimize steric hindrance.However, the authors indicate that the inter-ring C-C bond is significantly short indicating a relatively strong double bond character and .rr-delocalizationin spite of the large 33" torsion angle. The torsional energy barrier between thiophene rings has been measured by DeWitt et al. to be as high as 19.7kcalmole7' in the ground state using 'HNMR and 4.2kcalmole-' in the first excited singlet state as measured by fluorescence lifetime.

, C4H9

Scheme 6

. C4H9

4.2 Single crystals

193

3,4’,4’’-trimethy1-2,2’:5,2’’-terthiophene(a-3T(P-Me)3) One peculiarity of P-substituted oligothiophenes is that they can exist as non-planar conformational isomers depending upon the sign (positive or negative) of the interring torsion angle. To be observed experimentally, such a chiral oligomer must crystallize as a single enantiomer in a chiral space group. The first (and unique) example of such a chiral oligothiophene has been described by Barbarella et al. [81]. Indeed, 3,4”4”-trimethyl-2,2’:5’,2”-terthiophene, ~ - 3 T ( p - M e(Scheme )~ 7) crystallizes in the chiral P212121 space group as a single enantiomer with a non-planar (torsion angle = 173”) anti-conformation. This material being non-centrosymmetric may have some interesting second order optical nonlinearities.

CH3

Scheme 7

3’,4’-dibutyl-2,5’’-diphenyl-2,2’:5‘,2’’-terthiophene One of the only known example of a crystal structure obtained on a ‘doped’ oligothiophene has been described very recently by Graf et al. who crystallized the radical cation of 3’,4’-dibutyl-2,5’’-diphenyl-2,2’:5’,2”-terthiophene (Scheme 8) either chemically or electrochemically using PF; as the counter-anion [82, 831. This oxidized oligomer is highly conductive in powder form (a= lod3 S cm-’) or single crystal (a = lop3S cm-’) and provides a model for electrical conductivity in doped polythiophene. It crystallizes as dark purple needles with metallic reflectance. Surprisingly, the structure is temperature dependent. According to Graf et al., at 293K the crystal adopts the monoclinic system with a C2/c space group [82] while at 106K it has a monoclinic system with a P2’/n space group [83]. But both structures consists in columnar ‘slipped T-stacks’ of the oligomeric cation and channels containing the PF; anion parallel to the needle axis (Fig. 5). As expected, the crystal structure of the neutral (i.e. non-doped) terthiophene derivative (Fig. 6) is far different with an orthorhombic system ( Z = 4)and a P ~ a space 2 ~ group [83].

4.2.2.3 a-Quaterthiophene (a-4T)and derivatives The X-ray structure of a-4T single crystals grown from the vapor phase has been solved simultaneously by two groups during the course of this review [61, 621. Nevertheless, data had been previously predicted on the basis of calculations by Gavezzotti and Filippini [52] and further refined by Porzio et al. from X-ray powder diffraction data using the Rietveld full-profile analysis [53, 631.

194

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 5. Crystal packing of PF, doped 3’,4/-dibuty1-2,5”-diphenyl-2,2’:5‘,2’’-terthiophene (a) at 293K showing the channels formed by the butyl groups for the PF, anions, and (b) at 106K viewed parallel to the ac plane (top) and along the c axis. From ref. [82, 831.

The unit cell is monoclinic with a P21/a space group and the a-4T molecules (Scheme 9) are planar. Nevertheless, a slight deviation from co-planarity (10’) has been found by Porzio et al. using a crystalline powder obtained by recrystallization from chloroform-ethanol solutions. It is worth noting that crystals grown from the vapor phase are in the form of thin platelets while microcrystals obtained from solution are reported to be needles [53]. The number of molecules per unit cell is Z = 4 although the minimization of the packing potential energy leads Gavezzotti and Filippini to predict a 2 = 2 structure [52]. This choice was dictated by the wellknown principle of organic crystallography according to which organic molecules always make their centre of symmetry coincide with a crystal centre of symmetry.

4.2 Single crystals

Y

195

Yc

Figure 6. Unit cell of the non-doped 3’,4’-”buty1-2,5”-diphenyl-2,2’:5’,2’’-terthiophene crystal viewed along the c axis (top) and parallel to the ab plane. From ref. [83].

From the comparison between the above results and those simultaneously obtained by Antolini et al. [61] and Siegrist et al. [62] during the course of this review, it appears clearly that a-4T crystallizes in two slightly different structures depending on the crystal growth conditions. Such polymorphism has been observed previously on a-6T crystals grown either from the melt or by sublimation (see below). The two structures differ by the number of molecules in the unit cell ( Z = 4 or Z = 2). The experimental herringbone angle, i.e. the angle between mean planes of adjacent molecules, has been found to be of the order of 55” [61,62]for a-4T grown from vapor phase while a substantially greater value (70”) was found by Porzio et al. from X-ray diffraction data obtained with polycrystalline powder samples [63], but the most

Scheme 9

196

4 Structure and Properties of Oligothiophenes in the Solid State

striking difference is that the two polymorphs have different packing modes, neighboring molecules being more or less staggered relative to each other. Ferro et al. recently used a molecular mechanics computational procedure to refine and understand the crystal structure of a-4T, a-6T and PT [84]. For both oligomers, their calculations maintain the essential features of the original structure proposed by Porzio et al. [53], i.e. the small deviations from co-planarity of the four rings and the lack of elements of molecular symmetry (centre or twofold axis). Both a-4T and a-6T have, in fact, a molecular centre of inversion which does not act on the crystal packing.

a,a'-Dimethyl-quaterthiophene [a-4T(c~-Me)~] End-substitution of oligothiophenes by alkyl groups on the a-carbons of the terminal rings is an efficient method to prevent chemical reactivity of these sites. In addition to enhanced chemical stability, end-substitution does not induce inter-ring torsion (as often does side-substitution) thus allowing retention of high .rr-conjugation and electroactivity. Hotta and Waragai have applied this strategy to a-4T by grafting methyl groups on the two a-carbons and investigated the crystal structure of this compound (noted a-4T(~x-Me)~, see Scheme 10) in both its neutral [33, 491 and doped [85] forms.

Scheme 10

In contrast to non-substituted a-oligothiophenes, crystals of neutral a-4T(a-Meh are orthorhombic with a Pbca space group and 2 = 4. The molecules are almost coplanar, the two outer rings being slightly bent atcording to the mean plane of the two inner ones (maximum deviation is ca. 0.16A for CS). Although packed in the HB mode with an angle of about 59" between adjacent molecular planes, the molecules are aligned in an original zig-zag fashion (see Fig. 7), the molecular long axis being at an angle of 26" with the c axis. Hotta and Waragai have shown that ~ 4 T ( a - M ecrystals )~ can also be grown in their doped form using acceptors like iodine, NOBF4 or NOPF6 in organic solvents [49]. These authors also report on the growth of large charge transfer single crystals ( 5 x 0.2mm2) of a-4T(c~-Me)~ doped by TCNQF4. The X-ray 8-20 of these doped crystals shows a layered structure similar to that of the neutral compound although the profile of fine needles of the [ ~ - ~ T ( ( Y - M ~ ) ~ , T Ccomplex NQF~] is considerably more complicated than that of the iodine-doped crystal. Finally, Hotta and Waragai mention that the doping does not cause brittleness of the material or morphology change and conclude that the intercalated dopant anions induce a slight lattice expansion. Electrical conductivity of these crystals is briefly reported in section 4.2.3.2 of this review. The radical cation salts of a-4T(c~-Me)~ can also be produced by electrochemical oxidation of the neutral oligomer in the presence of large size anions [85]. The use of an end-substituted oligomer prevents coupling reactions and affords the formation of black-needle crystals with anions such as AsF;, TaF; and GaC1; . Nevertheless,

4.2 Single crystals

197

P Figure 7. Crystal structure of cy,c”y1-2,2’:5’,2’’:5’’,2”’-quaterthiophene a-axis. From ref. [33].

viewed along the

no X-ray structure of these doped crystals has been reported yet although Matsuura et al. describe their electrical conductivity ( lOP4-l OP6 S cm-’) and spectroscopic properties (see section 4.2.3.2) [85].

4,4/,3”,4”-Tetramethyl-quaterthiophene[cY-~T(P-M~)~] One of the most widely investigated conducting polythiophenes is poly(3-methylthiophene) P3MeT. As stated above, introduction of a substituent on the thiophene rings may induce conformational twist along the polymer chain and consequently decrease its mean conjugation length. It is then crucial to evaluate precisely the influence of substitution on the P3MeT polymer structure. An elegant way to achieve this goal is to synthesize a regioselective oligomer substituted by as many methyl groups as thiophene rings in the chain and investigate its crystal structure.

Scheme 11

198

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 8. Molecular (a) and crystal (b) structure of 4,4’,3”,4’”-tetramethyl-2,2’:5’,2~’:5’’,2”-quaterthiophene. From ref. [86].

4.2 Single crystals

199

The crystal structure of the model compound C X - ~ T ( P - M (Scheme ~ ) ~ 11) has been investigated by Barbarella et al. [86] in comparison with its conformation in solution [87]. Crystals of C X - ~ T ( P - M grown ~ ) ~ in hexane are monoclinic with a P2,c space group and Z = 2. As it can be seen in Figure 8, the molecules are stacked in parallel layers along the a axis similarly to unsubstituted oligothiophenes. Surprisingly, the strong steric interactions introduced by the four methyl groups on the P-carbons do not break the coplanarity of the ~-4T(,l?-Me)~ molecule in the crystal, or the all-trans conformation as it does in solution (cis-trans-cis conformation with an angle of 40” between the planes of the two internal rings). Nevertheless, methyl-substitution induces significant bond length and angle deformations in the solid state. Barbarella et al. attribute this effect more to the ability of the CX-~T(P-M molecule ~)~ to retain maximum r-conjugation through coplanarity rather than crystal packing forces. Furthermore, the mean plane of the two outer rings makes a dihedral angle of 3.4” with respect to the inner ones.

3,3’,4”~3’’’~Tetra~~ethylsulfanyl)-2,2’:5‘,2’’:5’’,2‘”-quaterthiophene [w~T(P-SM~)~] Another example of polymorphism in the oligothophene series, besides that of non-substituted C X - ~ T is, reported by Barbarella et al. on 3,3’,4”,3’”-tetra(methylsulfany”-2,2’:5’,2’’:5’’,2”’-quaterthiophene (noted C X - ~ T ( P - S Msee ~ ) ~Scheme , 12) [88, 891. The latter molecule has the particular ability of crystallizing in two distinct forms under the same experimental conditions. Both yellow, monoclinic and orange, triclinic crystals are obtained simultaneously during slow overnight evaporation of a petroleum ether/ethylacetate solution of a unique C X - ~ T ( P - S M compound. ~)~ This phemomenon, known as conformational polymorphism, in which a molecule adopts significantly different conformations in crystal polymorphs was elucidated by Bernstein and Hagler twenty years ago [90]. A recent review by Dunitz and Bernstein on ‘disappearing polymorphs’ illustrates the difficulty in obtaining a particular conformer under controlled and reproducible conditions [911.

Scheme 12

The crystal structures of the triclinic and monoclinic forms of C X - ~ T ( P - S M are~ ) ~ given in Table 3. In the orange triclinic crystal, the inner thiophene rings are coplanar while the two outer ones are moderately twisted by 27.4” according to the plane of the central rings [MI. But the most interesting feature is that the triclinic form has a sandwich-type molecular arrangement instead of the usual herringbone packing (see Fig. 9). Barbarella et al. tentatively attribute this quasi-r-stack to the partial loss of coplanarity of the molecule. The relative orientation of the external rings would result from steric interactions between the sulfur atoms of thiophene on one side and the methylsulfanyl group on the other side.

(4)

a

1.50 1470.5

-

91.8

-

P31 monoclinic P21/a 4 30.52 7.86 6.09

1.444 1649.8

1.457 379.0

-

-

-

1.501 731.1

106.72

-

90

-

1861 monoclinic p21/c 2 7.734 5.729 8.933

WI

orthorhombic Pbca 4 7.707 5.941 36.031

4T(P-Me)4

cu4T(c~-Me)~

97.22

-

[61, 621 monoclinic P21ja 2 8.936 5.750 14.341

a-4T/HT

D, calculated density. HT, high temperature polymorph.

Da ( ~ g . q - ~ ) Volume (A3)

Y

P (deg)

ff

a b (4) c (A)

Z

Reference System Space group

a-4T

Table 3. Comparative crystal data of 0-4T derivatives.

P1 1 7.567 7.601 11.193 93.99 100.88 116.75 1.538 555.80

WI triclinic

4T(P-SMe)4

1.443 184.7

-

97.14

-

~391 monoclinic p2 I /c 2 12.345 7.565 12.784

4T(,L?-SMe)4

$!

5

eK

E S' E rp

G.

B5

0 =

%

0,

2.

2

G

2

9rL

0 0

N

4.2 Single crystals

20 1

Figure 9. ORTEP (Top) and crystal (Bottom) structures of tetrakis(methylsulfany1)- 2,2’:5‘,2”:5”,2”’quaterthiophene (a) in its triclinic orange form (from ref. [SS]) and (b) in its monoclinic yellow form. From ref. [89].

Concerning the yellow monoclinic form of the ~x-4T(,f3-SMe)~ crystal, the two outer rings are even more twisted (55.0’) than in the orange triclinic form. This strong geometry distorsion is at the origin of the substantial 7r-delocalization decrease in the yellow form and hence of its yellow color. Another direct consequence of this loss of molecular planarity is that the outer rings of the monoclinic

202

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 9 (continued).

form are in the syn conformation while they still are in the anti conformation in the triclinic form. Finally, in the triclinic form the SCH3 substituents are roughly coplanar while they strongly deviate from coplanarity in the monoclinic form.

4.2.2.4 a-Quinquethiophene (a-5T) and derivatives The structure of a-5T single crystals has not been elucidated yet and the only studies have been performed on sublimed powder samples (together with a-4T and a-6T) by Porzio et al. using the Rietveld full-profile analysis [53]. Considering that a-5T is isomorphous of a-4T and a-6T, these authors found that the dimensions of the a-5T unit cell are intermediate between those of a-4T and a-6T (see Table 1) but mention that this was not expected due to the odd number of thiophene rings which induces a different internal symmetry.

Scheme 13

4.2 Single crystals

203

3”,4”-dibutyl-a-quinquethiophene[ ~ - ~ T ( P - B U ) ~ ] The X-ray crystal structure of ~ - ~ T ( P - B(Scheme U)~ 14) has been recently investigated by Liao et al. together with that of a P-tetra-alkylated sexithiophene derivative [50]. In both cases, the P-alkyl chains induce structural and conformational deviations from the standard non-substituted homologs. First, the ~ - S T ( P - B U ) ~ unit cell ( Z = 8) contains four independent molecules, one of them having its two terminal thiophene rings in the syn conformation (see Fig. lo), similarly to the yellow monoclinic form of the ~ - 4 T ( @ - s M ecrystal )~ [89]. The syn conformation of this oligomer is reminiscent of the coil geometry claimed for polythiophene.

Scheme 14

4.2.2.5 a-Sexithiophene (a-6T) and derivatives The first crystal structure of a-6T (Scheme 15) has been reported by Porzio and co-workers by using the Rietvield full-profile analysis on polycrystalline powder samples [16]. In agreement with the theoretical predictions of Gavezzotti and Filippini [52], these authors found a P21/a space group with four molecules in the unit cell ( Z = 4). Note that beside solid state crystals, a nematic liquid crystal mesophase has been observed with a-6T by Taliani et al. [92]. This liquid crystal phase is reported to appear at 3 12°C as measured by differential scanning calorimetry. Cooling the samples below the melting point leads to a freezing of the a-6T nematic phase which may thus be studied at room temperature. The formation of a a-6T mesophase above 305°C has also been observed by Destri et al. but could not be assigned to either a nematic or smectic phase [93].

Scheme 15

The first a-6T single crystals were obtained in 1995 simultaneously by Siegrist et al. from the melt at high temperature (a-6T/HT) [66] and by Horowitz et al. using a sublimation technique [64] adapted from that of Lipsett [94]. Surprisingly, these two kinds of crystals have two different crystal structures which essentially differ by the number Z of molecules per unit cell [66,95]. While Z = 4 is found in crystals grown from the vapor phase, crystals grown from the melt (the so-called ‘high temperature polymorph’ a-6T/HT) are characterized by Z = 2 (see Table 1). Both structures are monoclinic with flat molecules in the usual herringbone packing but they differ slightly in the position and orientation of the molecules in the unit cell (Fig. 11). A similar polymorphism has been observed with a-4T depending on the sample quality, i.e. single crystals grown by sublimation [61, 621 or polycrystalline powder [63]. Note that the structure of a-6T crystals obtained from the vapor phase is very similar to that of a-8T ones grown by the same technique [65].

204

Q

4 Structure and Properties of Oligothiophenes in the Solid State

C

3

cx

Figure 10. Molecular (top) and crystal (bottom) structure of 3”,4”-dibutylquinquethiophene [~-ST(P-BU)~]. From ref. [50].

Alkylated-a-sexithiophenes Before the elucidation of the crystal structure of non-substituted a-6T, a few alkylated sexithiophene a-6T derivatives have been studied. In particular, Herrema et al. [96]have synthesized two stereoregular dialkyl sexithiophenesand investigated their X-ray structure. Both 4’,3””‘-dibutyl-~exithiophene (Scheme 16) and bistrimethylsilyldioctyl-sexithiophene(Scheme 17) crystals are grown from solution. They crystallize with the monoclinic system in respectively the C2/c and P21/n

4.2 Single crystals

a

205

a

C

Figure 11. Crystal structures of the two a-sexithiophene polymorphs (a) obtained at high temperature a-6T/HT (from ref. [66]) and (b) from the vapor phase (from ref. [64]).

206

4 Structure and Properties of Oligothiophenes in the Solid State

space group. In spite of the bulky substituents on end- and side-positions, the molecules are almost planar (torsion angles between 5 and 11”) with the anti-parallel conformation.

C4H9

Scheme 16

These results support the idea that the syn or anti conformations are essentially determined by packing forces at the detriment of intramolecular forces. This is nicely illustrated by the results obtained by Liao et al. on 3”,4”-dibutyl-quinquethiophene and 3’,3””,4’,4””-tetrabutyl-sexithiophene(Scheme 18) [50]. In the unit cell (2 = 4), three molecules have an all-anti conformation while in the fourth molecule the external rings have a syn conformation relative to their neighbor. Furthermore, the external rings substantially deviate from the coplanarity of the molecule whatever their syn or anti conformation.

C4H9

C4H9

Scheme 18

Finally, Yassar et al. have determined the crystal structure of a,w-bis(triisopr0pylsily1)-sexithiophene(Scheme 19) [97]. This end-substituted compound crystallizes from a heptane solution as blade-shaped brownish crystals in the unusual triclinic system with only one molecule per unit cell ( Z = 1). The most stricking feature of t h s compound is that, although the molecules have the anti-parallel conformation with no substituent on the P-positions, all the rings are strongly twisted around the

Scheme 19

4.2 Single crystals

207

C i6

c2eCB

Figure 12. Molecular and (a) and crystal (b) structure of a,w-bis(triisopropylsily1)-a-sexithiophene [a-6T(a-Si-iPr&]. From ref. [97].

central axis (from 21 to 37"). This is particularly the case of the two external rings whose plane makes an angle of 37" with that of the closest ring (Fig. 12). Finally, the two bulky triisopropylsilyl groups induceaa crystal packing with staggered parallel arrays of molecules (stagger length = 4.5 A), an arrangement also observed by Herrema on stereoregular dialkyl-sexithiophenes [96]. For comparison, the essential crystal data of sexithiophene derivatives described in the litterature up to now are given in Table 4.

4.2.2.6 a-Octithiophene (a-8T) a-8T is the longest non-substituted oligothiophene that can be isolated as a pure compound [98]. It possesses four additional conjugated double bonds as compare to a-6T and is a priori susceptible to have better transport properties. But only a few reports deal with a-8T due to the difficulty in purifying it. Nevertheless, a-8T has been used as a p-type organic semiconductor in field-effect transistors [99]

Scheme 20

28.7 1.325 3050.8

55 41.5 1.55 1064.2

66 23.5 1.553 21 16.5

angle. 4, tilt angle of the long molecular axis with a. D, claculated density.

a 7,herringbone

131.5

97.78

90.8 38 1.26 2284.7

93.20

1.262 1892.1

100.79

12.403 8.690 17.871

[501 monoclinic

[961 monoclinic P21jn 2 17.071 6.011 22.299

[961 monoclinic c2/c 4 32.532 5.651 22.105

[661 monoclinic P21ja 2 9.140 5.684 20.672

1641 monoclinic P21/n 4 44.71 7.85 6.03

c~-6T(a-SiMe~)~(,L?-Oct)~ a - 6 T ( p - B ~ ) ~

a-6T(P-Bu)z

a-6T/HT

a-6T

Table 4. Comparative crystal data of a-6T derivatives.

67 1.291 1038.9

1971 triclinic P-I 1 7.537 8.358 17.006 89.70 89.68 75.90

a-6T(a-SiPr3)2

a

3 2

Fm

22

4.2 Single crystals

209

Figure 13. Fluorescence under UV excitation of the a-nT microcrystals obtained by slow vacuum sublimation of crude a-8T powder. Bottom up: a-8T (red), a-6T (orange), 0-4T (yellow) and a-3T (blue). From ref. [98].

and photovoltaic devices [43,100] and its linear and nonlinear optical properties have been investigated in polycrystalline thin films [27,101,102]. Because it is virtually insoluble in any organic solvent, a-8T crystals cannot be grown from solution. Nevertheless, in spite of a very high melting point (m.p. = 370°C), a-8T crystals can be obtained by slow vacuum sublimation of the powder material at 300°C over 24 hours (Fig. 13) [65]. Figure 14 shows two optical micrographs of a 2 mm long a-8T crystal illuminated by polarized white light and photographed with crossed-polarization using angles of 0" and 90" between the two polarizers respectively. The concentric colored fringes arise from optical interferences due to thickness effects. Extinction of the absorption of the crystal is obtained by rotating the sample by 90" in its surface plane. Dichroism and absence of facetting are macroscopic evidence for the molecular ordering and the net alignment of the unique optical axis b in a crystal lattice. The a-8T crystallographic data are gathered in Table 1. a-8T crystallizes in the monoclinic system although it is very close to the orthorhombic one (p = 90.31" instead of 90"). The space group P21/a was finally chosen after trying P2& and P2Jn. The b and c lattice parameters of a-8T are almost similar to those of

210

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 14. Optical micrographs of a 2 mm long a-octithiophene single crystal under polarized white light. The angle between polarizers are 0" (a) and 90" (b). From ref. [65].

a-6T wgle, as expected, the a-axis is much longer for a-8T (58.916A) than for a-6T (44.708 A) and explains the record structural anisotropy of a-8T crystals (a/c= 10). Importantly, the calculated density of a-8T ( D = 1.578 g ~ m - ~is )slightly higher than that of a-6T ( D = 1.553gcmP3). Such an increase of the packing coefficients with the chain length have been theoretically predicted by Gavezzotti and Filippini [52]. These authors attributed this dependence to the fact that the molecules are roughly brick-shaped and the larger the block the more efficient is the space occupation. The atomic numbering of the a-8T molecule is shown in Fig. 15 where it can be observed that the molecule is rigorously planar, the dihedral inter-ring torsion

2

(b) Figure 15. ORTEP front view and atomic numbering of the a-8T molecule (a). Side view of the planar a-8T molecule in its crystal (b). From ref. [65].

.-......-..._..

C

Figure 16. Crystal packing of a-8T viewed with the a axis vertical (a) and along the main molecular axis (b). Note that the four molecules of a neighboring cell are also represented. From ref. [65].

212

4 Structure and Properties of Oligothiophenes in the Solid State

Table 5. Minimum distances between atoms of two neighbouring molecules A and B situated (a) face-to-face and (b) aligned along the same line. From ref. 1651

angle being smaller than 1". Two typical views of the a-8T unit cell are represented in Fig. 16. The angle between the planes of two neighboring molecules is 65" and that between the long molecular axis and the a-axis of the cell is 24". At the macroscopic scale, the surface of the crystal corresponds to the bc plane. Table 5 gives the minimum distances between atoms of two face-to-face molecules A and B in the a8T unit cell and the distances between the nearest ends of two molecules A and B in the same alignment. It can be observed that these distances are shorter by 3-6 x for a-8T and could result in slightly greater intermolecular interactions. Such a crystal packing could in turn improve the charge transport perpendicular to the main molecular axis as compare to a-6T. This assumption has been verified with a-8T based thin film transistors (TFTs) during the preparation of this review. As a matter of fact, hole mobilities in the range 0.06-0.09 cm2/Vs, i.e. three times higher than in a-6T based TFTs, have been measured with high purity oriented a-8T-TFTs [103]. However, these high mobilities may also find their origin at the molecular level, a-8T having a longer 7r-conjugated chain than a-6T. Typical values of interatomic distances and bond angles of an outer (S,) and an inner (S5) thiophene rings in the molecule are selected in Fig. 17. These values are very similar to those found for a-3T and a-6T which shows that they are not chain-length dependent.

A

1.41

3

Figure 17. Bond lengths and angles of an outer (left) and inner (right) thiophene ring of the a-8T molecule. From ref. [65].

4.2 Single crystals

213

4.2.2.7 Polythiophene and 3-alkylated derivatives Although bulk crystals of polythiophenes have never been prepared, a number of structural studies have appeared on their crystallinity in powder form and thin films. Two conformations of the PT chain are possible: a linear and nearly planar all-trans structure (called the rod conformation) and a helical all-cis one (called the coil conformation). Both the rod and coil geometries, which are separated by an energy barrier of about 3-6kcalmole-', have been proposed on the basis of X-ray diffraction studies depending upon the substitution and doping state of polythiophene [67,104]. But for a crystal model of a polymer to be coherent, the chain conformation must be compatible with the chain packing. Mo et al. [67] investigated the X-ray scattering of chemically coupled PT and found that neutral PT can fit either an orthorhombic or a monoclinic unit cell. The main structural data of PT in both systems are gathered in Table 6. Later on, further experimental studies including scanning tunneling microscopy brought visual evidence of microcrystalline and helical polymer chain growth [105, 1061. Inganas et al. also report on the syn conformation in some segments of polythiophene [107, 1081. Other experimental studies on the crystal structure of polythiophenes have been reported in the literature [109-llla]. From a theoretical point of view, Cui and Kertesz have studied the geometrical and electronic structures of PT and PMeT by energy-band theory where a screw axis of symmetry has been taken into account [68]. These authors confirm that the anti (rod) structure is slighltly more stable for PT while the syn (coil) conformation is preferred for PMeT. Finally, Ferro et al. applied molecular mechanics to obtain a monoclinic model of crystalline PT [84]. Their calculations show the weak dependence of packing energy on the unit cell parameter p, i.e. on the relative length of adjacent PT chains, in agreement with centro-symmetric and translationally disordered planar chains.

Table 6. The various experimental and calculated crystal structures of non-doped polythiophene PT reported in the litterature. PT Reference System Space group Z a (A, b (4) c (A) P (deg) Db (Mg.mp3)

~ 7 1

PT ~ 7 1

PT" ~341

monoclinic

orthorhombic

monoclinic p2 1/"

2 7.83 5.55 8.20 96 1.537

2 7.80 5.55 8.03 90 1.567

4 7.807 5.526 7.803 105.3

a Calculated structure.

214

4 Structure and Properties of Oligothiophenes in the Solid State

4.2.3 Optical and electrical properties 4.2.3.1 General remarks

Because most oligothiophenes crystals grow as ultra-thin and brittle lamellae of small size (i.e. usually less than a millimeter), they require extremely delicate handling, which restricts their physical study. Nevertheless, a few optical and electrical properties of non-substituted a-6T and a-8T crystals grown by sublimation have been investigated over the last two years. As we pointed out above, these asgrown platelike crystals have poor mechanical resistance and cannot be cut or polished, for device fabrication. Consequently, they are interesting essentially for fundamental studies, particularly optical studies due to the rather good optical quality of their plane-parallel faces. Frequent macroscopic defects of these crystals as revealed by simple observation are cleavage (or twinning association) and fractures. These two defects are easily identified by color changes under polarized light due to the slight dipole misalignment that they induce. Consequently, selection of an oligothiophene single crystal adapted to refined physical studies should be done under polarized light and must possess none of these defects. The most attractive feature of a-6T and a-8T crystals is undoubtedly the access that they provide to optical dichroism (absorption and luminescence) and charge transport anisotropy (carrier mobility) so that comparison can be made with what is currently observed in polycrystalline thin films and disordered polythiophenes. Beside a-6T and a-8T, a few studies have also been done on doped a 4 T ( c ~ - M esingle ) ~ crystals.

4.2.3.2 Dimethylquaterthiophene

Hotta and Waragai performed two-probe measurements of the electrical conductivity of blade-shape single crystals of a-4T(a-Meh doped with iodine [33, 491. They found conductivities of 3.4 x lop5S cm-' along the c-axis, i.e. vertical to the crystal plane, and 1.2 x lo-' S cm-' and 2.9 x 1 0-2 S cm-' for the directions longitudinal and transverse to the crystal plane respectively. The latter directions, however, do not coincide with either the a- or b-axis of the molecules. This also argues in favor of a preferred charge transport along the .rr-stacks rather than along the molecular axis. Nevertheless, Hotta and Waragai explain this as the effect of the iodine anions which are located between the ends of the molecules, thus preventing charge carriers from hopping from one site to the next. Unfortunately, no value of the conductivity anisotropy of the TCNQF4 complexes of ~ - 4 T ( a - M ehave ) ~ been reported because in these doped crystals the counter-anions are sandwiched between molecular planes. The electrical conductivity within the blade plane have also been measured for various dopants, i.e. 12, NOBF4 and NOPF6. Values of 1.0 x lO-'Scm-', 4.4 x lop2S cm-' and 2.1 x lop2Scm-' have been found respectively [49]. These conductivity results together with X-ray structures (see section 4.2) show that 0!-4T(a-Me)~is a quasi two-dimensional conductor in which transport essentially occurs through face-to-face molecular arrays.

4.2 Single crystals

2 15

More recently, Matsuura et al. report on the electrochemical growth of a-4T (a-Me)z doped by PF,, AsF;, TaF;, BF, and GaC1; anions. At room temperature, all these salts have electrical conductivities in the range 10-4-10-6 Scm-' with an activation energy of 0.1-0.2 eV [85]. These conductivities are substantially lower than that of the iodine-doped compound suggesting that the layered crystal structure which favors charge transport is not retained when doping by large-size anions.

4.2.3.3 a-Sexithiophene (a-6T) Optical properties The polarization of vibration modes and coupling due to crystal-field effects of a-6T single crystals have been studied using polarized IR and Raman spectroscopies [64]. In agreement with previous studies on powders and thin films, Horowitz et al. confirm that the bands at 1491cm-' (C=C antisymmetric strectching), 1372cm-' (C-C intra-ring stretching) and 1205cm-' (C-C inter-ring stretching) are polarized along the long axis L of the molecule. Beside, some IR bands split into doublets, each of the two components being polarized either parallel or perpendicular to the b-axis. In particular, a large band splitting of the terminal C-H bond has been observed at 686 and 698 cm-l which was not resolved on the spectra of a-6T in the powder form. More recently, Horowitz et al. studied some optoelectronic properties of a-6T single crystals [112]. The UV-visible absorption spectra under light polarized parallel and perpendicular to the b-axis confirm that the fundamental T-T* transition is L-polarized although a substantial absorption persists in the 2.5-3.5 eV region. In the transparency region of the crystal, i.e. below 2.2 eV, well-defined interference patterns allow to determine the refractive indices nil and nl and their spectral dependences. Although nil is practically independent of the wavelength (nil = 1.656 over the 500-800 nm range), n1 fits the Sellmeier equation (nl = 1.95 at 545 nm) the best fit leading to n, = 1.867. The photoluminescence (PL) emission spectrum of the a-6T single crystal exhibits three vibronic components separated by ca. 0.19 eV, the most intense peaking around 640 nm [112]. The PL quantum yield of has not been measured for single crystals but has been estimated between and lop3 for polycrystalline a-6T thin films. These values are much lower than those obtained in solution (0.35 to 0.40). Transport properties In spite of their brittleness and the difficulty of growing them as large crystals, the transport properties of a-6T single crystals (a few mm2) have been investigated. Horowitz et al. report on field-effect transistors (FET) built with a-6T crystals playing the role of the semiconductor [l 1lb]. In this device, a a-6T crystal (5 microns thick) is carefully deposited on top of a PMMA layer (the insulator) which is itself spincoated on an aluminum gate electrode. The device is completed by two gold source and drain electrodes. A field-effect hole mobility of 0.075 cm2/Vs is then recorded at room temperature, i.e. roughly three times greater than that of a-6T polycrystalline thin films. The on-off ratio above lo4is mainly limited by leaks through the insulator. The current-voltage characteristics of this FET allows to estimate the dopant concentration in the a-6T crystal to 0.2 ppm and a free carrier density of 10" ~ m - These ~ .

216

4 Structure and Properties of Oligothiophenes in the Solid State

results are consistent with the high degree of purity of the material. An even higher field-effect mobility (0.16 cm2V-' s-I) has been reported very recently by Horowitz et al. for a-6T crystal [35]. According to the authors, this could be due to the use of a very small crystal as well as its low doping level. Nevertheless, it is not possible in this device geometry to measure the anisotropy of the mobility in the a-6T crystal but it can be anticipated that it is much higher in the direction perpendicular to the long axis of the molecule, i.e. between the source and drain electrodes of the above device. An alternative method to measure the field-effectmobility of oligothiophene single crystalsis to grow them directly on top of a transistor device by slow vacuumevaporation. One of the advantages of this method is that the metallic contacts as well as the interface between the semiconducting crystal and the insulator are of better quality compared to what they currently are, in devices using a macroscopiccrystal deposited on top of the insulator. Another advantage of controlling the growth of the crystal directly on the FET device is that it allows to check the transport properties along the various crystal axes. This technique has been recently used by Vrijmoeth et al. to study the intrinsic charge transport properties of a-4T crystals with diameters up to 20 microns [l 1lc] and is currently under way for a-6T crystals. Furthermore, these authors have developed a method to determine the crystal axis direction and thickness of the individual crystal by polarization spectroscopy. Photoconductivity (PC) measurements on a-6T single crystals have also described and the influence of the illumination wavelength (action spectrum), time and magnetic field on the PC signal of a 5 pm thin crystal mounted on an alumina substrate has been reported [112]. In particular, the PC spectrum evidences two PC regimes at low and high energy that the authors attribute to respectively generation of exciton and a direct ionization process. Finally, Klein et al. report on the femtosecond time-resolved spectroscopy of a-6T single crystals and explain the induced absorption and stimulated emission by the decay of the singlet excitons to the lower Davydov band [113]. 4.2.3.4 a-Octithiophene (a-8T) UV-visible absorption The absorption spectra of an a-8T single crystal (2 x 1 x 0.05mm3) have been recorded under polarized light at oblique incidence (Fig. 18) [99]. As expected, a strong dichroism is observed over the whole visible range when the light is polarized with the E vector perpendicular (Elb) and parallel (Ellb) to the crystallographic b-axis. These absorption spectra confirm that the fundamental T-T* transition of a-8T, which extends over the 350-600 nm wavelength region, is essentially polarized parallel to the long axis L of the molecules. Stimulated emission Recently, stimulated emission (SE) has been observed in single crystals of a-8T under intense photoexcitation at normal incidence (Fig. 19) [114]. SE occurs at a pump energy threshold of -0.1 mJ per pulse (33 ps, 10 Hz) to yield an intense and

4.2 Single crystals

300

400

500 600 Wavelength (nm)

217

700

Figure 18. (a) Scheme of a a-8T single crystal submitted to a polarized light beam at normal incidence. E is polarized either perpendicular (Elb) or parallel (Ellb)to the b axis. (b) Dichroic spectra of a a-8T single crystal under polarized light at normal incidence with Ellb (-) and ELb (- - -). From ref. [98].

narrow emission line peaking at 700nm (Fig. 20). At pump energies higher than 15 pJ per pulse, a second narrow line of weaker intensity emerges at 640 nm. Gain narrowing of these two lines can be easily described in the usual framework of stimulated emission (SE). Light amplification originates from the combination of a net dipole alignment and efficient waveguiding towards the edges of the crystal. In contrast to dye-doped crystals, e.g. anthracene in a fluorene matrix [115, 1161, the a-8T molecules provide both the emitter and the optical cavity due to their quasi-2D dimensions. Spectral SE selection of the two SE lines at 640 and 700 nm can be monitored by simply scanning the spatial position of the pump beam onto the surface of the a-8T crystal. The pump beam (energy=4.0pJ per pulse) is focused to a diameter of -0.3mm and its position is scanned along the x-direction, i.e. the crystal length.

218

4 Structure and Properties of Oligothiophenes in the Solid State

24O'i... ... .. , . Laser 4........

*

Crystal thickness (3.5 Pm)

Figure 19. Experimental arrangement for observation of stimulated emission in a a-8T single crystal. All molecules are aligned parallel to each other and make a 24" tilt angle with the (b,c) plane of the crystal. Detection is performed through an optical fiber positioned close to the upper edge of the crystal at an angle of 30". From ref. [114].

As expected for this pump energy, the SE line at 700 nm dominates over most of the crystal surface while that at 640nm is almost inactive (Fig. 21a). Figure 21b shows that the situation is totally reversed in a small region located close to the crystal origin, the dominating SE line being that at 640nm although residual activity of the line at 700nm persists. These results suggest that spectral SE selection can be related either to a thickness effect or to the presence of structural defects whose inhomogeneous distribution could influence the emission process. Similar SE phenomena are also observed in a-4T and a-6T crystals [I 17, 1181.

Emission Wavelength (nm) Figure 20. Normalized stimulated emission spectra of a a-8T single crystal under pulsed laser pumping (& = 532nm, 33ps) at variousenergies 1/0.1pJ, Z/l.OpJ, 3/13.0pJ, 4/27.0pJ, 5/50.0pJ.From ref. [114].

4.2 Single crystals

219

-4

=!I

m

v

>

c1 .-

v)

C

a)

CI

C

-

0

C

a,

+-I

% t

-

0

I

400

600

500

400

,

,

500

700

l L ,

600

It,,

800

.

700

,,

,

800

900

,

b),

900

Emission Wavelength (nm) Figure 21. Spatial dependence of stimulated emission in a a-8T single crystal when the pump beam (4= 0.3 mm) is scanned along the x-direction (see Figure 3) from the edge toward the inner part of the crystal. The two SE spectra shown here are recorded in the outer (a) and inner (b) regions of the crystal. From ref. [114].

Two-photon SE can be easily realized in plane parallel single crystals of oligothiophenes (a-4T7a-6T and a-8T)by pumping with a near-IR laser beam in-between the lower and upper faces of the crystal [ 1 181. Since the crystals are almost transparent in this region, it allows the pump beam to propagate through the crystal, thus ‘seeding’ the up-converted SE beam on longer distances. Excitation and up-converted detection are performed along the same direction perpendicular to the a-axis of the crystals. Under pumping in the 750-900 nm range, the two-photon excitation spectrum of a a-8T crystal peaks at 850nm. Then, when an optical pump at 850 nm is used, the energy threshold for spectral narrowing in a a-8T crystal is estimated to be around 3 2 ~ J c m - Such ~ . an ultra-low up-converted SE threshold is the result of the particular gain-guiding geometry achieved in the strongly anisotropic crystal. Another important consequence of gain-guiding is the highly directive nature of the two-photon SE beam inside the crystal. Since oligothiophenes are efficient organic semiconductors (with mobilities close to 0.1 cmP2V-’ s-’), these single and two-photon SE phenomena bridge the gap between conventional organic laser dyes and luminescent conjugated polymers [I 19, 1201 on the way towards the organic laser diode. The fabrication of such an organic laser diode requires a material combining efficient charge injection from the electrodes, good transport properties and reasonable luminescence efficiency in the solid state. This is obviously the case of a-6T and a-8T and laser diodes based on oligothiophenes thin films are currently under way.

220

4 Structure and Properties of Oligothiophenes in the Solid State

Photoconductivity Finally, photoconductivity (PC) studies have been performed by Moses et al. [121] on a-8T single crystals to study the intrinsic properties of photoexcitation and transport in such a model molecular crystal systems and to determine the role of structural defects on these properties. Picosecond transient PC measurements over a wide range of temperatures (10-300K) demonstrate that the dependence of the transient photocurrent on light intensity and electric field in the single crystal a-8T is radically different from that in vacuum-deposited polycrystalline films. The photocurrent lifetime in the a-8T crystal is of the order of a nanosecond whereas in the film it is less than loops. These observations indicate a bi-molecular carrier recombination component prevailing in the a-8T single crystals, whereas a monomolecular mechanism operates in polycrystalline films.

4.3 Thin films 4.3.1 Deposition techniques 4.3.1.1 Vacuum deposition techniques Gas phase (ultra high) vacuum deposition is expensive and time-consuming but clean. It can be employed for most oligothiophene oligomers up to octithiophene (a-8T) and also for polythiophene with a polymerization degree of 20-25 monomeric units. An exception are molecules with thermally unstable substituents. The molecules are filled into Knudsen- or Langmuir-type evaporation cells and heated up to their sublimation temperature. Mass spectra show that the materials are usually clean or can be cleaned by sublimation and sublime without destruction, even above the temperature at which the maximum evaporation rate is obtained. The sublimation conditions remain constant and reproducible in contrast to many other ‘stable’ organic materials like phthalocyanines. Usually no residuents remain in the cell. This shows the excellent processability of oligothiophenes for thin film formation from the gas phase. Typical evaporation conditions will be summarized in Tables 7 and 8 for a-5T and a-6T, respectively (see below).

4.3.1.2 Preparation from solution There are several wet deposition procedures, including sedimentation, spin coating, electrochemical deposition, self assembly, and the Langmuir-Blodgett technique. Common for all these methods is that no ordered films of unsubstituted pure oligothiophenes can be prepared. The easiest (and most unreproducible) way is to elutriate the oligomer in an organic solvent, put a substrate in, and evaporate the organic solvent. Usually these films are not uniform in thickness, include remnants of solvent, and are

4.3 ThinJilrns

221

often mechanically unstable. Similarly solvents with temperature dependent solubility can be used, from which the molecules precipitate at lower temperatures. On some substrates the molecules adsorb spontaneously. This leads to more reproducible films and their structures are often thermodynamically controlled. Spin coating is a good alternative to produce uniform films on large substrates. However, this method is seldomly used for oligothiophenes, but will become of increasing importance if technical devices have to be produced by low-cost-methods. Electrochemical deposition of polythiophene films from solutions of mono- or bithiophenes is one of the commonly used methods and can also be applied to prepare oligothiophene films. The films are formed in situ during the polymerization process. Due to the oxidative polymerization only ‘doped’ (oxidized) materials are obtained and appropriate counter-ions have to be added. This method can only be applied to conductive substrates. Self-assembly of molecules relies on the self-organization properties of the oligomers. In particular oligothiophenes with long alkyl chains as substituents are able to organize onto a substrate. This leads to highly ordered monolayers of flat lying molecules on highly oriented pyrolytic graphite (HOPG) and other ultra-flat surfaces. Experiments to use thiol-substituted thiophenes to be deposited onto gold showed excellent ordering of the molecules (compare section 4.3.2.3). A related method is the Langmuir-Blodgett technique in which amphiphilic molecules like carbonic acid substituted thiophenes self-organize in a compressed layer on water which can be transferred onto a solid substrate [122]. An alternative way is to use unsubstituted molecules which are mixed with arachidic acid [123, 1241. Another attempt is to prepare films with isolated oligothiophene molecules in uniaxially oriented matrices. Oligothiophenes diffuse into the matrix which consists of cyclic siloxanes [125], dextrines, nematic liquid crystals [ 1261, or zeolites [127, 1281 which all possess cavities of defined diameter and length.

4.3.2 Morphology 4.3.2.1 General remarks In this section the morphology of oligothiophene thin films will be reviewed. In particular the orientation of the molecules with respect to the substrate normal and the long-range order will be discussed. In the following an organic film is denoted as highly ordered, if there is at least one perfect monolayer with the molecules oriented in the same direction, i.e. having the same tilt angle against the surface normal (‘uniaxial orientation’). Crystalline films have a three-dimensional order with sharp reflexes in diffraction measurements, whereas epitaxial films are crystalline with a respective orientation to a single-crystalline substrate surface. Large differences in the geometric structure occur between unsubstituted and substituted oligothiophenes. As substituents mainly alkyl-chains or alkyl-rings are used. The alkyl-substituted molecules can be divided into three groups, which behave differently in their film formation properties:

222

4 Structure and Properties of Oligothiophenes in the Solid State

Alkyl-chains at the a and w position (end-substitution,Scheme lc) generally lead to films with a higher degree of order as unsubstituted molecules but similar (herringbone) structure. Polar groups at the end of one of the alkyl-chains lead to molecules with amphiphilic character which can be used for the formation of Langmuir-Blodgett or self-assembly films. Sidechain-substituted oligothiophenes (Scheme 1b) only form ordered layers by casting from solution [ 129- 1311. The evaporation onto silica surfaces gives amorphous films [ 1321. The observed structures can vary significantly depending on the number and position of the alkyl-chains. In regioselectively substituted oligomers often the alkyl-chains determine the packing of the molecules rather than the oligothiophene backbone if the number of sidechains is sufficient to allow for a densly packed structure. So-called ‘endcupped’ oligothiophenes (Scheme 21) could only be prepared in ordered layers by UHV evaporation onto Ag(ll1) [133, 1341, whereas on silica only disordered films were obtained [135, 1361.

Scheme 21

There are only very few investigations about the influence of dopants on the geometric structure and in particular on the exact position of the counterions in the doped materials. Most of these investigations are of theoretical nature with all known uncertainties.

4.3.2.2 Monothiophene (1T) and derivatives Monothiophene thin films were prepared by gas phase deposition onto metal single crystals, i.e. Pt(ll1) [137,138]and Cu(100) [139] at low temperatures. As determined by NEXAFS (near edge X-ray absorption fine structure) and vibrational HREELS (high resolution electron energy loss spectroscopy) the molecules lie flat on the surface up to the coverage of one monolayer. If the coverage is increased at 150K, rather than a second layer being built up, the molecules in the first layer turn upright leading to a higher packing density. If this compressed monolayer is then heated to 180K, only flat lying molecules could be observed. 3-Thienylpentadecanoic acid forms Langmuir-Blodgett films on various substrates like Suprasil glass, Si02 on Si(lll), and Ag layers on Si(ll1). This was tested by UV-Vis-, FTIR- and SPR measurements [140-1431. The molecules form densly packed films with the carbon chains oriented 10.6” against the surface normal. Multilayer formation up to 140 layers was possible, in which the thiophene moieties partially interpenetrate. Mixtures of this molecule with distearylviologene also form stable Langmuir-Blodgett films, in which electron transfer from the thiophene unit to the acceptor-type viologene unit should be possible through the alkylchains [142, 1431.

4.3 Thinfilms

223

4.3.2.3 Small oligomers (a-2T-a-4T) and derivatives

There are only few investigations on thin films of small oligomers. On metals, vacuum evaporated quaterthiophene (a-4T) films on Ag(ll1) and solution deposited films of a-2T-a-4T on Ag sols have been studied. Quaterthiophene on Ag( 1 1I) forms commensurate epitaxial films with flat-lying molecules, even for coverages above one monolayer, in contrast to monothiophene. This could be proven by LEED (low energy electron diffraction) measurements. The first layer is covalently bound to the Ag surface and cannot be desorbed intactly by heating. The second layer also shows a larger binding energy if compared to the third and higher layers. For layers of and above three monolayers the strong parallel order gradually decreases as determined by NEXAFS and IR data [144-1491. However, if such films are annealed at 300K a complete change in the LEED patterns is observed. NEXAFS data show that the molecules are oriented with their molecular plane parallel to the substrate surface, independent of the film thickness [148]. This behavior is quite similar to that of end-capped quaterthiophene [150]. Two different monolayer phases are found, a compressed saturated monolayer below 494K and a relaxed monolayer at low coverages and/or after heating the sample above 494K. This structural phase transition is reversible. On Ag sols molecules lie mainly flat on the surface, as determined from SERS (surface enhanced Raman spectroscopy) measurements. However, the molecules are twisted and there is evidence for a small inclination against the surface. The twisting decreases and the inclination increases with increasing chain length [151- 1531. The observed twisting of the molecules is in line with crystal powder X-ray diffraction data (see section 4.2.2). Completely different behavior is observed for evaporated films on oxidic substrates at room temperature [135, 1361. UV-Vis, IR, and fluorescence spectroscopy show, that the long axes of a-4T molecules are oriented mainly perpendicular to the (silica) surface normal for monolayer coverages. This can be deduced from a dichroism D = - =Ez3 EX,Y

with E as molar absorption coefficient of the first UV absorption band in the direction of the film normal (z) and within the film plane (x,y),respectively. The angle 0 between surface normal and molecules can be derived as 0 = arcosd= 39", because the transition dipole moment of the first absorption band and the long molecule axis run parallel [135]. This orientation was corroborated by X-ray diffraction measurements of 50nm thin films on silicon wafers [ 1541. At higher coverages this order decreases. Furthermore these films are not totally stable but tend to crystallize as their smaller counterparts do spontaneously [ 1551. This crystallization can be derived from light scattering as well as from a factor group splitting of peaks in the IR spectrum due to the symmetry in the crystalline state [156]. IR absorption and resonance Raman scattering on vacuum evaporated l p m thick films of a-4T on glass and KBr point at a twisting of the molecules also on these oxidic substrates [157].

224

4 Structure and Properties of Oligothiophenes in the Solid State

As for the longer chain lengths (see below) it was shown in a recent study by X-Ray diffraction and SFM measurements that increasing the substrate temperature (here: 373-393K) leads to larger crystals of 6pm2 [158]. By using a nonconstant evaporation rate, increasing from 0 to 600 pm/s, these authors were able to obtain crystals as large as 75 pm2. The angle between long molecular axis and surface normal was derived as 22". Submonolayers of a-2T-a-4T adsorbed from CH2C12solution onto silica plates show small crystallites which can be formed due to the high surface mobility of the molecules [136, 1591. Although their structure has not been investigated it is most likely that the molecules in the crystallites pack in a herringbone structure with an inclination angle between two units of 60" to 70" as it is determined for crystal powders (compare section 4.2.2 [51-531).

Substitution effects As mentioned above, alkyl-chains at the a and w position (end-substitution) generally lead to films with a higher degree of order as unsubstituted molecules but similar (herringbone) structure. This is shown for dimethyl-quaterthiophene [33,49, 160, 1611. End-substitution of quaterthiophene with two amino-groups also leads to films with the molecular axis nearly parallel to the surface normal, despite the dipole moment of the substituent [ 1621. The dihydrochloride, however, only forms amorphous films. Side-chain-substituted dodecyl-quaterthiophene on HOPG packs in lamellar structures where the oligothiophene units overlap slightly [1311. Dimethylquaterthiophene crystallizes epitaxially on PTFE, but is polymorphlc. Needle-like crystals are oriented with the c-axis of the molecules lying in the PTFE plane. Also platelike crystals with the c-axis perpendicular to the PTFE plane are observed [163]. The formation of Langmuir-Blodgett films by mixing a-3T or a-4T with arachidic acid in the molar ratio 1: 2 failed, in contrast to longer oligothiophenes (see next sections) [ 1221. Furthermore, terthiophenes with different polar head groups were tested as LB films [122]. Whereas Xa-3T with X=CHO and CHzOH do not form stable monolayers, LB films on quartz plates could be prepared from molecules with X = CH20COCH3, CH20COCH=C(CH3)2, CH20CO(CH&CH3, and CH20CO(CH2)16CH3.Within the fYms close packing with nearly vertical terthiophene units could be observed. Furthermore terthiophene with a C11H22viologene-substituent forms LB-films. This molecule allows for electron transfer within one molecule [1431. It could also be shown recently, that a-hexadecylterthiophene forms LB films on hydrophilic substrates [164]. Self-assembly of quaterthiophene diphosphonic acid in alternating layers with tetravalent zirconium on phosphorylated (silane) substrates leading to densly packed films at temperatures of 353K is reported by Katz et al. [165]. These molecules were also used to self-assemble onto Zr-phosphonate LB-films which act as template layers. In comparison to the films assembled on the phosphorylated silane surface, these layers do not need heating to give highly ordered films [166, 1671. Alkylthiol- (HS(CH2)1*-)substituted terthiophene as well as terthiophene-disulfides show highly ordered self-assembled films prepared on flat Au-substrates (Fig. 22) [168]. This could be shown by a combined ellipsometry, contact angle,

4.3 Thin-films

Gold

225

I

Figure 22. Proposed molecular arrangement and defect structure of HS-(CHZ)11-T3-H on gold (lower panel). The tilt and rotation angles for the alkyl chains are 35" and 45", respectively. The tilt and rotation angles for the T3-unit are -14" and 33", respectively. The top panels show the space-filling models of the crystalline parts of the assembly [168].

IR, X-ray photoelectron spectroscopy (XPS), and cyclic voltammetry study. The kinetics of monolayer formation vary dramatically for the two compounds. The alkanethiol assembles rapidly, within a few minutes, and forms a densely packed, highly organized monolayer with the alkylchains in an almost perfect all-trans conformation and the C,-C, axis of the thiophene moiety tilted about 14" away from the surface normal. The disulfide assembles much slowlier within at least one day. It forms an organized monolayer with a thiophene chain tilt of about 33" against the surface normal. Alkylthiol- (HS(CH&-)substituted terthiophene also self-assembles on Pt [ 1691. The maximum coverage of 99.5% percent is reached within 1 hour, as deduced

226

4 Structure and Properties of Oligothiophenes in the Solid State

from electrochemicalmeasurements. From those it could also be concluded that the molecules do not lie flat on the substrate, but are oriented almost perpendicular on the surface. Dopant effects Hotta and Waragai conclude from their diffraction measurements, that strong acceptors like iodine and nitrosyl intercalate between the molecular layers of oligothiophenes, at least for dimethyl-quaterthiophene [33,49]. The doping does not cause brittleness and the morphological features remain unchanged [ 1701. Doping with weaker acceptors like TCNQ brings about mixed stacks comprising the oligothiophene and the acceptor [33].

4.3.2.4 Quinquethiophene (a-5T) and derivatives a-5T behaves quite similar if compared to the smaller molecules stressed above, but tends to form layers on oxides with a higher degree of order. Casting from solution onto silica leads to small aggregates in coexistence with single molecules. The latter can only be detected by their fluorescence due to their small concentration, which does not allow for UV absorption measurements. In contrast to the aggregates, in which the molecules are oriented with their long axes mainly parallel to the surface normal (see below), single molecules lie parallel to the substrate surface with the angle between surface normal and long molecular axis 0 = 70" as determined from IHH/IVV = 0.5 with I as fluorescence intensity under perpendicular (VV) and parallel (HH) polarized excitation and detection with respect to the plane of incidence and observation [136]. Deposition of a-5T from a Knudsen cell in high vacuum (HV) or ultra-high vacuum (UHV) at room temperature leads to highly ordered layers on SiOz and other oxidic substrates. The a-5T molecules are oriented with their long axes mainly perpendicular to the substrate plane. This can be shown with polarized UV-Vis absorption and FT-IR spectroscopy, NEXAFS, and SFM (scanning force microscopy) measurements. Figure 23 shows the polarized UV absorption spectra of a 5 nm thick a-5T film and a comparison between solution and film spectra [171]. The most pronounced difference between solution and film spectra is the large blue shift of the first absorption band maximum. This indicates aggregates of molecules with their long axes oriented parallel to each other (H-aggregates) because the excitonic coupling between such molecules leads to forbidden low-energy transitions resulting in a highly asymmetric peak with the maximum intensity shifted to higher energy. As described in section 4.3.2.3 the angle 8 between long molecular axis and surface normal can be determined from the dichroism of the first absorption band. Good quality films show 0 < 20". This means that the molecules form a 'pin cushion' type of aggregate. Also from fluorescence anisotropy measurements a high degree of order can be evaluated [172]. The orientation can be proved indirectly by comparison with the polarized spectra of Langmuir-Blodgett films discussed below [ 13][1731, where pin cushons are commonly accepted, and more directly by the dichroism in polarized IR

4.3 Thinfilms

(cm-’)

20000

20000 v

30000

227

30000

40000

(cm-’)

Figure 23. (a) UV absorption spectra of a5T in different media: region of first electronic transition. (b) p-polarized W absorption spectra of a 5 nm thick a5T-film on silica under the indicated angles of incidence [ 1711

transmission and IR-ATR spectra of the asymmetric C=C stretch vibration around 1495cmpl and the C-H stretch vibration around 3060 cmpl shown in Fig. 24. Both vibrations are polarized along the molecular axes of a-nT-molecules [135][1741. The macroscopic film quality can be proven by SFM measurements [175]. In the monolayer range a nearly complete film with many pin-holes can be detected. With increasing coverage two complete pin-hole free layers are observed, before the growth mode changes to a simultaneous multilayer growth (Fig. 25). The good agreement of the molecule length of a-5T (2.2 nm) with the measured layer heights confirms, that the molecules in the film are standing with their long axes mainly perpendicular to the substrate plane, although the absolute tilt angle of the molecules to the surface normal cannot be deduced. These findings are also in line with comparative SFM and X-ray diffraction measurements of Hotta and Waragai on unsubstituted and alkyl-substituted a-5T [33][49].

4 Structure and Properties of Oligothiophenes in the Solid State

4

I

921 90

-

1450

1500

1550

1400

1350

wavenumber

9 4

V

92 I

3200

-

3150

3100 3050 wavenum ber

3000

Figure 24. IR-ATR measurements of a 15nm thick a-5T film on ZnSe (similar spectra are obtained on Si02): regions of vibrations oriented along the long molecular axis are marked with arrows [174]. (a) asymmetric C=C stretch vibration at 1495cm-'. (b) C-H stretch vibration a t 3060cmp'. The TE mode excites components of the transition dipole moment in the substrate plane, the TM mode components in as well as out of the substrate plane.

For thicker films the growth mode changes from layer to layer growth to a simultaneous multilayer growth for room temperature substrates, although a two-dimensional expansion is still dominating the third layer growth. Although the molecules pertain their orientation these films exhibit a very rough surface (Fig. 26a). A uniform width of the multilayer terraces is caused by an equal growth velocity of the multilayer islands perpendicular and parallel to the substrate plane, i.e. in thicker a-5T films a three-dimensional growth behavior is dominating. Also polycrystalline regions can be detected which coexist with ordered multilayers. This is also in line with the above-mentioned optical measurements which show a reduced order in thicker films. The quality of a-5T films, however, can vary significantly upon changing the deposition parameters. For the growth of the above-mentioned uniaxially oriented films in ultra-high vacuum (UHV) an optimum deposition rate of several hundreds of pm/min was determined. For these conditions an increase in substrate temperature to around

4.3 ThinJilms

229

Figure 25. SFM images of a nominally 6.1 nm thick (1.5-T film on SOz: (a) grey scale image, (b) cross-section through one of the multi-layer islands. The base-level in the cross-section measurement as well as the homogeneous grey value between the layered islands in the grey scale image correspond to two monolayers of a-5T. The height differences of the multilayer terraces are about 3.2nm and multiples thereof [175].

373K leads to the growth of crystalline films with crystal sizes above several pm at least up to thicknesses of 20nm (Fig. 26b,c) [176]. The change in structure in comparison to oriented films can also be deduced from shifts and new splittings in the IR-spectra (Fig. 27) [177]. The optimum temperature could be deduced from thermal desorption measurements in which a-5T films were heated up to 474K. They show a main peak with a sharp cut-off pointing at a zero-order desorption kinetic with a desorption energy of 210 kJ mol-'. A secondary peak at lower temperature can be attributed to a temperature-induced less ordered and probably more mobile phase [178].

230

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 26. Influence of the substrate temperature on the film growth [176, 177: SFM images of 20nm thick a-5T films on SiOzdeposited at (a) Tsub = 298K, (b) Tsub = 343K, and (c) TSub = 363K.

Therefore a deposition temperature should be chosen which is high enough for molecular diffusion but low enough that the molecules are not evaporated, i.e. around 363-383K. On the other hand Egelhaaf et al. found, that the orientation for high vacuum deposition (lop6mbar) and deposition rates of 500pms-' is decreased for increased substrate temperature. However, slightly reduced pressure and deposition rates of l00pms-' with room temperature substrates result in lOOnm thick oriented films, in which the above-mentioned dichroism is perfectly preserved [1591. These findings can be summarized to the following film growth behavior on completely amorphous oxidic substrates like Si02(Fig. 28): Single a-5T molecules lie flat [136]and diffuse on the substrate surface. As soon as two or more molecules meet on the surface they form aggregates in which the molecules align with their long axes parallel to each other and stand nearly upright on the surface. This leads to two nearly defect-free layers with axially oriented molecules which most probably

4.3 Thinjlrns

I

,

740

-

720

231

I

700 680 wavenurnber

660

640

(b) 0.020-’

0.018-

.-0

3

-

0.016-

crystalline oriented

1

0.0140.012-

:::;::

0.006

-

-

0.004 0.002 0.000 I

740

720 700 680 +waven urnber

660

1

640

Figure 27. IR transmission spectra in the region of the CH,,,-,f,lane-vibration [177]: (a) 2.67 x lop3 molar solution of (2.5-T in CS2. The spectrum is compensated for the CS2 absorption. (b) 70 nm thick (2-5T films on SiOZ deposited at Tsub= 363K (‘crystalline’) and at Tsub= 300K (‘oriented’).

form a smectic (liquid-crystalline) arrangement. If higher substrate temperatures are used, large crystallites can form in which the molecular orientation is preserved (compare also Table 7). On oxidic substrates like ordered layers of ‘SiO’ on top of Si(100) prepared by N20 decomposition on clean Si substrates [ 1781, single-crystalline A1203(1102) [ 1761 or single crystalline Ti02(1 10) and TiOz(100) (Fig. 29) [1791 with pinning centers as steps, kinks or perhaps also electronically active defects a slightly other film growth behavior is found (Fig. 30). Here, the diffusing single molecules can be pinned at these defects, if low deposition rates are used. The molecules then form nucleation centers at which crystallites are formed. There is no evidence that the molecules in these crystallites have a preferred orientation with respect to the substrate surface. If the deposition rate is enhanced, aggregation is kinetically favored if compared to diffusion. Therefore closed layers are formed like on amorphous substrates in which the molecules

232

4 Structure and Properties of Oligothiophenes in the Solid State

I

I Figure 28. Schematic representation of the proposed model describing the different steps of film formation on amorphous, flat surfaces like SOz. The homogeneous layers consist of oriented molecules in liquid crystalline (smectic) arrangement if rsub = 298K, and of crystallites of several micrometer diameter with similarly oriented moelcules if rsub N 373K.

again stand nearly upright on the substrate surface. They cover even very rough surfaces like a carpet. Raising of the temperature leads to an enhanced diffusion velocity and therefore again to a rough clustered surface. However, it may be possible to form also crystalline films with large crystallites on carefully prepared (nearly defectfree) Ti02(100) and probably also other surfaces [179]. It is still not clear whether on oxides the substrate interaction plays no role at all or whether dipole forces between the oxide surface and the molecules are important. Force field calculations, which neglect these interactions, can fully explain the orientation of the molecules, just taking the intermolecular forces into account, but they are still too inaccurate to lead to a final conclusion [ 1801. On non-oxidic substrates, however, the interaction energy between a-5T and the substrate can play a significant role. PTFE and metal substrates allow for a strong interaction with the molecules. For PTFE mainly the overall van der Waals interaction and therefore also the shape of the molecules determine the adsorption layer structure. The molecules are therefore oriented with their axes perpendicular and parallel to the substrate surface with the latter molecules aligned with the PTFE strands [181]. Metals should allow for additional strong interaction of metal d-orbitals with the 7r-system of the molecules or even for charge-transfer as

150

188

50

E

E

588

Distance:

l0EB 1149.89

1588

A

2088

2580

3088

Tracm Dlbtanoe tfi) Height: 23 46 I?

3588

4888

Angle(deg):

1.17

Figure 29. SFM image of a 72nm thick a-5T layer on Ti02(100) [179].

I

R low and/or T,,

high

/

high 1

( / 7 /

I I I I

I

Figure 30. Schematic representation of the proposed model describing the different steps of film formation on crystalline substrates with surface defects. R is the evaporation rate, Tsubthe substrate temperature.

234

4 Structure and Properties of Oligothiophmes in the Solid State

it is proposed for a-6T [153]. This would lead to the energetically preferred orientation of flat lying molecules with maximum geometric overlap between molecules and substrate. However, there are no published data on the orientation of a-5T molecules on metals, but there are hints that molecules lie also flat on the substrate as their smaller counterparts (section 4.3.2.3). Furthermore it can be observed on clean Au surfaces, that a-5T crystallizes under all (UHV) deposition conditions and no complete coverage of the substrate can be reached, even for layer thicknesses of several hundred nanometers. As soon as the pressure is only slightly increased and most probably a hydrocarbon layer is formed at the surface, a-5T layer formation is possible. This indicates that the surface free energy of a-5T is higher than that of pure Au leading to a cluster growth mode [182]. In Table 7 a summary of the evaporation conditions, substrates, and observed film structures is given. Due to the very different parameter sets and too many ‘white spots’, the best conditions to optimize the wanted structure cannot be determined exactly. Nevertheless, the trends are clear. Substitution effects End-substituted dimethyl-quinquethiophene shows a higher degree of order if compared to unsubstituted a-5T. Side-chain-substituted dimethylquinquethiophene crystallizes epitaxially on PTFE, but is polymorphic. Needle-like crystals are oriented with the c-axis of the molecules lying in the PTFE plane [163]. So-called ‘endcapped’ quinquethiophenes could only be prepared in ordered layers by UHV evaporation onto Ag(ll1) (Fig. 31) [133, 1341. The five thiophene rings are clearly resolved in the STM spectra whereas the exact geometry of the alkyl-rings cannot be determined. The slightly bent form of the Czy symmetry a-5T molecules is obvious and leads to a structure with two molecules per unit cell. This is in contrast to linear C2h molecules with only one molecule per unit cell. On Si02endcapped oligothiophenes (also those with other chain-lengths between IZ = 2-7) grow almost completely disordered, probably due to their reduced intermolecular interaction [135, 1361. A completely different system are Langmuir-Blodgett films of a-5T. Whereas the molecules themselves are not amphiphilic and therefore do not form films on a water surface, a mixture of Cd arachidate, arachidic acid, or other long alkyl-chain carboxylic acids and a-5T can be transformed into films in which the a-5T molecules are oriented perpendicular to the Si02 (glass, quartz, or wafer) substrate surface [13, 16, 122, 124, 173, 1831. No clear statement on the distribution of the molecules within the carboxylic acid layer is given except in [16]. Whereas the scetches of the film structure imply a statistical distribution of non-interacting molecules, the absorption spectra show the same blue shift as found in aggregates pointing at a phase separation of the two components.

4.3.2.5 Sexithiophene (a-6T) and derivatives From the point of applications a-6T is by far the most important oligothiophene. Like a-5T sexithiophene forms highly ordered or crystalline films on oxidic surfaces,

RT*

plane fused silica

** Around

Layer thickness

evaporated at 7 x IOp5mbar 0.005 nrnjs

RT?

RT

glass slides

Si(100) +

0.2-0.6 nmjs

300-325K

PTFE

mbar.

high vacuum

0.0005 nmjs 0.01 nmjs 5 x 10-I'mbar 0.02 nm/s 5 x 1OpI0mbar 0.001-0.01 nrnjs 5 x lO-"mbar

RT RT RT

A1203( 1 T02) Ti02(100) Au, Au on SiOz

5 x 10-"mbar

up to 70nm

5 x 10-"mbar

0.001 nmjs

373K

om201

two layers

5 x lo-'' mbar

<0.001 nmjs

5 nm

0.2-4nm at least 1 layer a t least 1 layer up to 850nm

a t least 1 layer

400 nm

three and more

two layers

RT

mbar

Si(100) with native oxide SiOz

2x

up to 100 nm

high vacuum 0.001 nmjs

up to 30nm

26 nm

high vacuum

high vacuum

high vacuum** 2.5 nm, 10 nm and higher

Base pressure

RT

0.2-0.6nmjs

Evaporation rate

Si(100) with native oxide

RT Si with native oxide planar fused silica, R T sapphire s.c., Si with native oxide RT fused silica

Tsub

Substrate

* Room temperature, i.e. 300K.

Literature

Table 7. Evaporation conditions, substrates, and substrate temperatures for a-5T deposition.

perpendicular, closed layers perpendicular, multilayer growth perpendicular, closed layers, spacing 2.3 nm crystals with diameters >5 pm layered structure, spacing 1.945 nm perpendicular, "closed" cluster perpendicular, closed perpendicular, closed crystallites, no closed film parallel, oriented along PTFE strands

perpendicular (18"-21")

perpendicular (15"f3)

fluorescence, UV-Vis absorption

SFM SFM XPS, SEM

SFM

XD, SFM

SFM

SFM

SFM

fluorescence, fluorescence excitation, UV-Vis absorption IR-ATR, IR transmission fluorescence, fluorescence excitation, UV-Vis absorption, IR-ATR fluorescence

perpendicular

perpendicular

Method

Observed structure

236

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 31. STM images of end-capped a-ST (ECST) on Ag(ll1) (occupied states) [134]. (a) Large area scan (35.2nm)’, (b) small area scan (4.4nm)2.

4.3 Thinjlms

237

depending on the preparation conditions. The trend of forming films with higher crystallinity by increasing the substrate temperature is even more pronounced. On the other hand there are no reports on completely closed uniaxially oriented films. The dichroism of the first UV-Vis absorption band of HV evaporated films in the monolayer range is only slightly lower if compared to a-5T7leading to 0 = 22" [ 18I]. This orientation, however, deteriorates much faster by increasing the layer thickness. Hamano et al. [184] find an even better orientation (@ = 4.3") for films evaporated in UHV. In SFM most of the films appear not completely uniform in their height level (but see below), but show crystallites. This is in line with several other publications in which a-6T polycrystalline films were grown. X-ray diffraction data of these films show, that different orientations of a-6T molecules can appear in the crystallites, depending on the preparation conditions. At room temperature, deposition rates of 10nm/s lead to two monoclinic phases with the molecules oriented parallel and perpendicular to the substrate surface. At lower temperatures and higher deposition rates the metastable phase with flat lying molecules should be preferred whereas at higher temperature and/or lower deposition rate the molecules have energy and time enough to stand up on the substrate and crystallize in t h s most stable form [55]. This could be underlined by experiments with deposition rates of 0.5 nm s-l at Tsub= 300K, where the stable form was found in experiments at 77K, where mainly flat lying molecules were determined. At higher substrate temperatures, different layer spacings could be observed if compared to the perpendicularly oriented molecules deposited at room temperature: the deviation between molecular axes and surface normal is 2 l o for room temperature substrates, 23" for Tsub= 463K, and near 0" for Tsub= 533K [56]. Also the crystal size is influenced by increasing the substrate temperature, leading to elongated grains of 30 x 200nm2 at Tsub= 533K, which is much smaller if compared to a-5T, but therefore leads to a high optical quality of the films [185]. In a series of papers Biscarini et al. showed the influence of the substrate temperature not only on the crystal size but also on the smoothness of the thiophene film surface [ 186- 1891. Whereas films grown at room temperature exhibit a granular morphology, above 423K strongly anisotropic lamellae are formed. However, the transition between both morphologies appears to be continuous. Increasing substrate temperature leads to larger crystals up to Tsub= 448K, but only at temperatures well above this temperature a smooth surface could be obtained. Also annealing of films grown at room temperature above T,,, = 433K gave smooth films. From temperature-dependent SFM data and theoretical evaluations they could conclude that the transition to the smooth surfaces represents the onset of fast diffusion of a-6T molecules on the surface. There are few measurements on non-oxidic substrates, although more attention has been paid to this topic recently. On polar zinc chalcogenides a-6T behaves like on oxidic substrates [156, 1901. In recent papers it has been shown, that a-6T orients mainly parallel to a hydrophobic SiHjSi surface, whereas under the same deposition conditions the molecules stand upright on Si02/Si [191-1931.

238

4 Structure and Properties of Oligothiophenes in the Solid State

The orientation of a-6T on PTFE is similar to a-5T, i.e. the molecules orient parallel to the substrate surface and along the polymer strands [194, 1951. This is also true for a-6T evaporated onto stretched polyethylene [1901. Adsorption on Ag sols from solution leads also to flat lying molecules with slightly tilted molecules, as described above for bithiophene [ 1531. However, on Ag and Cu films exposed to air a-6T forms highly oriented films with an inclination of 20" against the surface normal as determined by NEXAFS and IR-RAS. Whether this orientation is induced by an oxide layer is not clear [196]. The same group finds a random orientation on evaporated Au film surfaces. However, on single crystal Au(llO)(l x 2) surfaces in UHV, the molecules are oriented parallel to the Au rows, i.e. lie flat on the surface. In the first monolayer they form a (T x 4) superstructure where T = 2.74/0.28 denotes the lattice vector along the Au rows, where 2.74nm is close to the sexithiophene length and 0.288 nm corresponds to the Au-Au distance in the row. Possible superstructures are shown in Fig. 32. For the second and following layers a decreasing order was found by He diffraction [197, 1981.

&n T6

16.32 A

0

s

Figure 32. Sketches of two possible models describing the T x 4 structure of sexithiophene on Au(llO)(l x 2) surfaces [197, 1981.

4.3 Thinjlms

239

Figure 33. Projection of a-6T molecules close to parallel to their molecular axes, showing their relative dispositions to a substrate (top-view) as well as the origin of their internal fracture. The (020) planes are much more facile cleavage planes due to their poorer interdigitation if compared to the (4,0,1) planes [199].

Lovinger et al. [ 154, 1991 find that films sublimed onto carbon-coated electronmicroscope grids with thicknesses below 200 nm form very small (< 100 nm) crystals and have no regular crystallographic boundaries. With growing film thickness lamella are formed on these crystallites in which the molecules usually have an orientation perpendicular to the surface, but also parallel orientation eventually appears. A major characteristic of these lamella is their spontaneous fracture along (020) planes (Fig. 33). In very thick films the lamella splay and warp about their growth axes which leads to bad transport properties. This lamellar growth behavior seems to be independent of the substrate (see results in [186] above). As for a-5T Table 8 summarizes the preparation conditions for evaporated a-6T films. The trends are the same with a higher probability for crystallization. Substitution effects End-substituted dimethyl-sexithiophene [ 160, 1611 and dihexyl-sexithiophenes [56][132][200]exhibit a larger molecular order than unsubstituted a-6T. Vacuumevaporated a,w-dihexyl-sexithiophene(a,wDH6T)on SiOz gives films with a much higher degree of order if compared to a-6T, as determined from the narrowing of the X-ray diffraction patterns (Fig. 34) and polarized UV-Vis measurements. As

Literature

RT

RT

RT

plane fused silica

planar fused silica, sapphire s.c., Si with native oxide

fused silica

?

?

10nmjs

463K

533K

RT

RT

Si (with oxide)

Si with native oxide

Si (with oxide)

mbar ?

high vacuum

2-5 x

crystallites with parallel and perpendicular molecules perpendicular

25 nm

spacing 2.37 nm spacing 2.44 nm parallel, spacing 2.44 nm perpendicular (23"), spacing 2.242 nm perpendicular (0') 2-3 pm

few pm

perpendicular (4.3") better than at RT, worse than lower deposition rate

100 nrn

UHV

IR-ATR, IR transmission

XD

XD, W-Vis

UV-Vis

decreased order

thicker layers 0.01 nm/s 0.002 nmjs 0.01 nmjs

fluorescence, fluorescence excitation, UV-Vis absorption, IR-ATR

perpendicular (15" f3)

2 nm

high vacuum

0.2-0.6 n m j ~

fluorescence, fluorescence excitation, UV-Vis absorption

perpendicular decreased order

2.5 nm 10 nm

high vacuum

Method

Observed structure

Layer thickness

Base pressure

~

Evaporation rate

0.5 nmjs 10nmjs 10 nmjs

~~~

RT RT 71K

373K

Tsub

Substrate

~~

Table 8. Evaporation conditions, substrates, and substrate temperatures for or-6T deposition.

373K

RT

RT

RT-373K

PTFE

BK7 glass slides with 50 nm Ag

Ag, Cu films (with oxide?) on Si(100)

Au films on Si(100)

Au(l10)(1 x 2 )

carbon-coated el.-microsc. grid

[194][195]

[1851

11961

[197][198]

[I 54][ 1991

11961

77K

SiHjSi

[ 191-1931

RT?

423-448K >448K

RT

ruby mica

[186-1891

RT

SiOZjSi

[191-1931

1 O-* mbar

lo-* mbar

HV

“slow sublimation”

1.33 x 1.33 x

mbar

10-6-10-7 mbar

mbar

0.005-0.05MLj~ UHV

0.033 nmjs

0.033 nmjs

1-2 nmjs

0.01-1 nmjs

“high rate”

0.02-0.1 nmjs

“high rate”

perpendicular, small crystals

<200 nm

+E D

thermal He atom scattering

parallel, T x 4 superstructure (Fig. 32) 1 ML

TEM

NEXAFS, IRRAS

NEXAFS, IRRAS

SEM

random orientation

crystallites

20 nm1.1 pm

IRRAS, UV-Vis transmission, fluorescence, TEM + E D

55 nm

parallel, oriented along PTFE strands

30-60 nm

FTIR-ATR, UV-Vis, XD

perpendicular (20”)

parallel

15-40 nm

SFM

FTIR-ATR, UV-Vis, XD

55 nm

perpendicular, granular morphology larger crystals smooth surface

perpendicular

lOOnm

15-40 nm

L

N P

%’

Y

2

242

4 Structure and Properties of Oligothiophenes in the Solid State

a

3

8

004-

W z

A

I-

z

-?

00

10 0

20 0

30 0

40 0

50 0

60 0

SUBSTRATE

Figure 34. X-ray diffractogram of a vacuum evaporated film of (a) a-6T and (b) a,w-dihexyl-a-6T. (c) Schematic representation of a a,w-dihexyl-a-6T monolayer on a substrate surface, showing two adjacent rows of oligomers [200].

unsubstituted a-6T a,wDH6T crystallizes in a monoclinic arrangement. The molecules stand up on the substrate with the (a,b) plane as contact plane. The second orientation found for a-6T with flat lying molecules could not be determined. The observed structure results from the packing effects of the rigid thiophene core and the flexible alkyl chains. Maximum packing density leads to a tilt angle between the conjugated core and the chains of 16". The same molecules lie flat on HOPG when adsorbed from the liquid phase [129].

4.3 Thinfilms

243

Figure 35. (a) STM image of a non-regioselective 2,5-didodecyl-sexithiopheneon highly oriented pyrolytic graphite (HOPG) (occupied states). (b) STM image of a regio-selective isomer of 2,5didodecyl-sexithiophene on HOPG (occupied states). (c) Molecular model for the non-regioselective isomer on HOPG. The variations in the intermolecular spacing along the lamella axis in (a) can be attributed to this non-regioselectivity [ 1291.

244

4 Structure and Properties of Oligothiophenes in the Solid State

Side-chain-substituted dodecyl-sexithiophene on HOPG packs in lamellar structures where the oligothiophene units overlap slightly (Fig. 35) [129]. Side-chain-substituted dimethyl-sexithiophene crystallizes epitaxially on PTFE, but is polymorphic. Needle-like crystals are oriented with the c-axis of the molecules lying in the PTFE plane [163]. As reported for a-5T, the formation of Langmuir-Blodgett films by mixing a-6T with arachidic acid was possible [122]. Dopant effects p-doping of sexithiophene can lead to a polymerization of the oligomeric units because of the formation of reactive radical cations. From Raman spectra indirect evidence is given on a distorted conformation around the inter-ring single-bond [157]. This could be explained by model calculations which show, that although a planar, rigid-rod like doped polythiophene is built from chains with an all-anti conformation of the thiophene units, the syn conformation is not able to form a coplanar chinoid (doped) form, leading to a helical structure of the polymer [68].

4.3.2.6 Longer oligothiophenes (a-7T, a 3 T ) and derivatives

a-5T and a-6T represent the oligothiophenes with the highest degree of order observed in thin films. a-7T and a-8T show also a dichroism of the first UV-Vis absorption band, but much smaller (best monolayer values 0 = 27”). Like a-6T the orientation gets lost even more rapidly with increasing film thickness, and higher substrate temperatures lead to a better orientation [181]. At least a-8T also tends to form crystallites on silica surfaces [185]. Substitution effects Side-chain-substituted dodecyl-octi- and hexadeci-thiophene on HOPG pack in lamellar structures where the oligothiophene units overlap slightly [130, 1311. This overlap decreases with increasing chain length and is only fix for regioselectively substituted molecules. The alkyl-sidechains fill the space between two lamellas and show a high mobility. They are oriented parallel to one of the crystallographic axes of graphite. a-7T can also be incorporated into arachidic acid to form Langmuir-Blodgett films with perpendicularly oriented molecules. Even with one or more p-linkages film formation is possible, but then at least a molar ratio of 1: 5 between septithiophene and arachidic acid has to be chosen [122]. In LB-films of oligoalkylthiophenes (chainlength 7,10, and 11) mixed with arachidic acid a phase separation exists, which is more pronounced for small amounts of thiophene. No preferential orientation of the oligoalkylthiophene domains could be observed by polarized UV-Vis spectra [123]. Oligo-(heptadecyl 3-thiopheneacetate) (mixture of a-2T-a-9T) form stable LB films on hydrophobic substrates. The thiophene backbone lies on the substrate plane and the alkyl-chains are oriented perpendicular to the surface [2011.

4.3 Thinfilms

245

4.3.2.7 Polythiophene and derivatives

In general there are many different structures observed in polythiophenes, depending on the monomer unit, the polymerization method, and, for electrochemical preparation, the counterion and solvent. Polythiophenes usually form disordered layers, in particular if electrochemical deposition is used [15,202-2041. Preparation of plasma-polymerized polythiophene leads to disordered layers with comparable structure to electrochemicallyprepared films [205]. Yamamoto et al. [l 1 1, 206, 2071 and Kurata et al. [208] prepared polythiophene layers by vacuum evaporation. To ensure the deposition of poly- and not oligothiophenes, they removed low molecular weight material (n < 15-17) by Soxhlet extraction with CHC13. The evaporated material is assumed to consist of polythiophene with a molecular weight of about 1500-2000 and 20-25 thiophene units. The X-ray and electron diffraction data of Yamamoto et al. point to a (partially) crystalline film with the polymer axes oriented perpendicular to a carbon or gold substrate plane if Tsub= 423K. At lower substrate temperatures the orientation and crystallinity are worse. This behavior is not only true for the first deposited layer, but at least up to 10 layers, i.e. 100nm. On polyimide substrates, however, the molecules lie on the substrate plane and orient along the rubbing direction of the polyimide [206]. Under the assumption that the first hyperpolarizability ,B (compare section 4.3.3.1) is along the chain axis of polythiophene, Kurata et al. also found by SHG (second harmonic generation) measurements, that the polymer strands stand upright on a glass substrate surface. The angle between surface normal and the chain axes is decreased by increasing the substrate temperature with 0 = 14" at Tsub= 473K. Substitution effects Spinning of end-substituted poly(3-hexylthiophene) and poly(3-octylthiophene) on HOPG leads to aggregates and to single molecules with different conformations with anti and syn arrangements of the thiophene units leading to flat lying loops, rods, and coils, as determined by STM measurements [130]. Side-chain substituted polythiophenes in stretched films show a lamellar structure whereby the alkyl-chains act as spacers between two-dimensional sheets formed by stacking the polymer main chains one above another. The alkyl-chains have zig-zag conformation and their planes are slightly tilted against the thiophene backbone [209]. Precipitation of poly(3-octylthiophene) from solvents with strong temperature dependent solubility leads to an orientation of the alkyl-chains perpendicular to a glass surface [2lo]. Poly(3-hexadecylthiophene)/behenic acid mixtures [211] as well as pure poly(heptadecyl 3-thiopheneacetate) [2011 and poly(3-decylethoxythiophene) [2121 form stable LB films. Dopant effects Doping of perpendicularly oriented, partially crystalline polythiophene vacuum evaporated onto carbon substrates with iodine leads to reversible structural

246

4 Structure and Properties of Oligothiophenes in the Solid State

changes. At low doping levels only a broadening of the X-ray diffraction patterns could be observed, but at saturation doping a new structure is formed which is shown in Fig. 36 [206]. There are different results on stretched films of poly(3-alkylthiophene) doped with iodine, although the same pristine structures were observed for polyoctyl- and polydodecylthiophene. In one study they show continuous structural evolution upon iodine uptake [209]. There is a crossover from a packing configuration that strongly suppresses alkyl interdigitation into one in which interdigitation is pronounced. In the other study no interdigitation is discussed [213]. In both cases the iodine (I;) lies off to the side of the polythiophene main chains while closely nested by the surrounding alkyl chains (Fig. 37). The iodine atom closest to the backbone is situated approximately 0.3 nm away from the thiophene sulfur atom [209, 213, 2141. Electrochemicallyprepared polythiophene films often consist of an open microlamellar structure, which is not distorted by diffusion of counterions in oxidation/ reduction cycles. In contrast to the counterions the solvent molecules play a critical role and lead to a swelling of the polymer film, as can be determined by cyclovoltammetry [203], SIMS and XPS [215], and SEM measurements [216]. In films of 3,4disubstituted polythiophenes steric factors become determining and lead to a twist between the thiophene units [203]. Changing the counterion during the electrochemical preparation, however, leads to films with different morphology. Whereas a gelatinous polymer gel was grown in an H2S04/CH3CNelectrolyte solution, which shows a microfibrilar structure after drying, a smooth, flat film was obtained by changing the solvent to N-methyl-2-pyrrolidone [ 151. Electron diffraction measurements point at an intercalation of the counterions between the chains at empty places within the planes rather than between the basal plane [163]. This is supported by X-ray diffraction measurements on poly(3-decylthiophene) films doped with

\

H 1.0 nrn

Figure 36. Top-view of the proposed structure of polythiophene (PTh) saturately doped with iodine (I; chains) [206].

4.3 Thinfilms

241

2 09 nm ’-

undoped (or lightly doped)

moderately doped

1.98 n

m

d

heavily doped

Figure 37. Schematic drawing of the projected equatorial structure of poly(3-octylthiophene) at various doping levels. The circles indicate the iodine atoms in 1; [209].

[FeC14]- and examined before and after stretching [217]. The dopant anions prefer to locate in the crystalline regions of the polymer. The authors suggest that the anions are located between the planes of the polymer chains and act as ‘lubricant’, because the doped polymers can more easily be streched. Furthermore, the possibility to undergo conformational changes during drawing is higher in the doped materials. A more detailed description of the iodine position in poly(3-alkylthiophene) is given in [209]. STM measurements on electrochemically prepared polythiophene (in (C4H9)4NBF4/propylenecarbonate) show a lamella-like structure with steps of about 5 nm. Different regions could be distinguished by STM. From the height of the corrugations the authors determine from the images, that disordered regions are highly doped whereas ordered regions are undoped or only slightly doped [218].

4.3.3 Optical characterization 4.3.3.1 Undoped oligothiophenes

The electronic structure of oligothiophenes is characterized by UV photoelectron spectroscopy (UPS) and molecular orbital (MO) calculations for the occupied states and by a variety of optical techniques for the transitions between occupied and unoccupied levels. Compared to oligothiophene monomers in solution, the photophysics in the solid state is much more complicated. The ground state absorption is affected by intermolecular interactions, the fluorescence quantum yield and the excited state lifetime are dramatically reduced, the recombination dynamics are highly non-exponential and depend on the excitation density, and charge carriers are among the long-lived byproducts of photogenerations. Therefore many

248

4 Structure and Properties of Oligothiophenes in the Solid State

of the optical features are still a matter of discussion. Because of the vast number of papers on optical measurements only a selection demonstrating the most important features can be presented here. A more detailed description of the electronic structure is given in Chapter 6 of this book [28]. Occupied states From the comparison of UPS measurements in thin films and in the gas phase of monothiophene it can be concluded that the molecular properties are conserved in the solid state. There are several calculations (and UPS studies) on the molecular orbitals of these systems, which come to similar results [219-2241. Ab initio as well as semi-empirical calculations attribute the HOMO of the monomer to the (laz) x-orbital with contributions from the carbon pz orbitals. This 7r-orbital splits into n orbitals for a-nT and forms a wide 7r-band for polythiophene. The HOMO-1 (3bJ orbital of the monomer has no pz contributions and forms in an-T a non-bonding narrow x-band with little dispersion. At around 4 eV a broad band results from nonbonding and bonding 7r-orbitals, whereas the broad emissions at 8 eV and 11.5eV result from c7 orbitals. The UPS data on longer oligomers (n 2 6 ) exhibit mainly the features of polythiophene [222, 2231. Ionization potentials as determined by MNDO or VEH calculations are varying in value, even in different papers of the same author, but most are located around 7eV for (hypothetic perfect gas phase) polythiophene and around 9eV for the monomer with the calculated values varying within 1 eV around these values (see, e.g. [221][225-2301). The experimentally (UPS) determined gas phase values for the small oligomers are in the same range, whereas the solid state ionization energies are about 2 eV lower. Electron affinities are around 5.5 eV for polythiophene and 3 eV for monothiophene The introduction of defects like /3-links alters the electronic structure substantially, in particular concerning the 7r-orbitals because of the limited delocalization. The x-bands in the MNDO simulated UPS spectrum of a3pa3,e.g. is very similar to adding the spectra of two trimers and a monomer [222, 223, 231, 2321. This is also underlined by studies of the thermochromic effects in polyalkylthiophene which can be explained by conformational changes which lead to segments with shorter conjugation length [ 108, 2281. Absorption spectra Typical UV-Vis spectra of a-5T in different media were given in Fig. 23 above [171]. In Fig. 38 HREELS and UV-Vis spectra of a-nTs with n = 4-8 are shown together with the peak positions of solvent and film spectra data and HMO calculated band positions [27,233,234]. Similar results are also published by many other groups, see, e.g. “4, 235-2391. The HREEL spectra are independent of the film thickness. The loss peaks at 0.38 eV and 0.76 eV are attributed to the CH stretch vibration and its overtone. The region of electronic loss starts with a steep onset which shows a characteristic shoulder Ao, labelled by a triangle. Four (five in the spectra of a-7T and a-8T) clearly discernible loss bands, A,, through E, follow on the high energy side.

4.3 ThinJilms

(4

249

Energy Loss (cm -l) loo00

0.0

I.o

30000

2ooM)

2.0

3.0

4.0

40000

5.0

6.0

Energy Loss (eV)

20000

3 0 m

7 (cm-’) Figure 38. (a) HREEL and UV-Vis absorption spectra of thick a-nT films ( d > 1Onm) [27]. (b) Transition energies vs. reciprocal chain-length. Symbols: experimental data from (a) and HMO calculated values. Lines: linear regression curves [234].

250

4 Structure and Properties of Oligothiophenes in the Solid State

00

0.1 ,n 0.2

00

01

1I n

02

0.0

0.1

1/ n

02

Figure 38. (continued).

Plots of the band maxima against the reciprocal chain-length of the molecules, l/n, yields straight lines with a common intercept at v = 17000 cm-' (2.1 1 eV), but with different slopes. The ratio of the slopes of bands A,,, :B: C : D is found to be 1 : 1.8: 2.7:4.2. The slope of the shoulder A0 runs parallel to that of A,,, while the position of loss band E is almost chain-length independent. In typical UV-Vis absorption spectra of thick films (d = lO-IOOnm), three principal regions of absorption A, C, and E can be distinguished. Absorption band A consists of a high energetic main part forming a broad continuum with the maximum A,,, (circles), which shifts to shorter wave length with decreasing film thickness, and of a low energetic edge revealing a structure (Ao, A,, A2, . . . ) with spacings of =1750cm-', i.e. 0.22eV (Ao-A1) and of =1500cm-', i.e. 0.19eV (A,-A2), independent of chain length. The main peaks can be related to singlet transitions by simple HMO calculations (Fig. 38b). More (and less) sophisticated calculations lead to the same principal assignments with 1 'B(,):symmetry for the lowest and 2 'ACg)for the next higher excitation [240-2441. Beljonne et al. find the 2 'A state either below or above the I 'B state, depending on the number of levels taken into account for the configuration interaction [245]. (The lower case terms in brackets refer to molecules with C2h symmetry.) According to the HMO calculations the five absorption bands A through E are assigned to the transitions S1,1* (A), S2,1* and Sl,2* (B), S2,2* (C), S3,2* and S2,3*(D), and the intra-ring transition S , , * (E), respectively. Here, the orbistarting with 1 for the HOMO and tals are numbered consecutively from I(*)to d*), I* for the LUMO. In UV-Vis spectra only bands A, C, and E are visible, because bands B and D are optically parity forbidden in molecules with C2h symmetry or

4.3 ThinJilms

251

have an extremely low probability in Czv molecules [181]. In HREELS these transitions can appear because a finite momentum is imparted to the molecules ('impact scattering') [246]. Fig. 39 summarizes the electron configurations and the singlet states including electron correlation for CZh molecules. Longer oligothiophenes as a-6T or higher give spectra which are very similar to (real) polythiophenes, e.g. exhbit a band-gap of around 2.2-2.3eV. This can either be due to a short effective conjugation length of polythiophene or due to defect induced gap-states (X-traps, see below) [247]. The linear extrapolation of the plot of peak positions against reciprocal chain length to infinite chain-length, i.e. to a perfect, defect-free polythiophene, leads to a minimal band-gap of 17000cm-', i.e. 2.11 eV. The linear extrapolation to very large chain-length is not straightforward because the HMO-model, a model based on classical oscillators, and the exciton model predict a non-linear behavior in that region [181], leading to an even larger predicted band-gap. The theoretically (VEH) calculated band-gap lies between 1.6 eV [226]-1.7eV 1225, 227, 2281-1.8 eV [229] for polythiophene and at 5.5 eV for monothiophene [225, 2281. The ab initio HOMO-LUMO gaps are always substantially too large and are determined, e.g. to 5.9 eV [221] or 6.862 eV [230] for polythiophene. Also semi-empiricalcalculations neglecting electron correlation by Jurimae et al. result in too large HOMO-LUMO distances, e.g. 6.72 eV for a-6T [248]. For a comparison between experimentally observed band-gaps E, and HOMO-LUMO 'gap' values AE(HOM0-LUMO), compare [249]. The appearance of the UV-Vis spectra changes considerably with decreasing layer thickness (compare Fig. 22 and 38) (see, e.g. [181, 2491). The loss of oscillator strength on the lower energy side of absorption band A, the blue shift of the band maximum A,,,, and the growing dichroism are caused by the formation of H-aggregates, i.e. molecules oriented perpendicular to the substrate and parallel to each other as discussed in section 4.3.2.4. In the exciton model this leads to a

43000 cm 37600 cm - I

32900 cm

0

CI configuration

--A electronic state

Figure 39. Electron configuration and singlet states taking into account electronic correlation effects (CI). Also indicated are experimentally determined transition energies for a-2T in solution [181].

252

4 Structure and Properties of Oligothiophenes in the Solid State

more effective intermolecular coupling of transition moments in thin layers (‘excitonic coupling’). On the other hand spectra simulation using the Fresnel model with anisotropic dielectric functions shows clearly that the blue shift and the dichroism can also be satisfactorily explained without necessarily introducing an increase of exciton splitting in the H-aggregates [190]. In HREELS collective excitations are not possible due to the short wave length of electrons with a kinetic energy of 15eV, i.e. 0.32 nm, which is the range of the size of single molecules. The spectra of thin and thick film are therefore identical [27]. The nature of the lowest excited state A. is still subject of discussion. The position of this band is independent of the preparation conditions and the film thickness and therefore independent of the form of aggregates. The most probable explanation attributes A. to the 0-0 vibronic band of the 1 ‘A, --t 1 ‘B, transition. For a more detailed discussion compare [250] and Chapter 6 of this book [28]. Alkyl-substitution influences the absorption spectra in thin films mainly by their influence on the coplanarity of the molecules, particularly in the excited state, and only to a minor extent by their electron pushing (inductive) effect [251]. The latter should result in a red shift which is not observed in absorption spectra of thin films but can be detected in the fluorescence spectra. In absorption spectra major blue shifts of the absorption peaks occur if larger differences in the torsion angle between different rings are induced by the substituents. As long as the coplanarity is not distorted quite similar absorption spectra are observed if compared to unsubstitued oligothiophenes [33, 2521. As expected, directly electron donating or accepting groups have a much higher influence on optical spectra if compared to simple alkyl substituents. The refractive index and the rugosity, i.e. the ‘roughness’ of the sample, were studied by optical transmission in the transparent region of unsubstituted and 0-alkyl substituted a-6T [253, 2541. The refractive index is n = 1.904 for a-6T, n = 1.763 for diethyl-substituted a-6T, and n = 1.688 for didecyl-substituted a-6T, the rugosity of a-6T is determined as 44.1 nm which is much higher than in amorphous silicon with 16.9nm. The latter is attributed to the large thickness of the film (1.4 pm) and to the microscopic molecular structure in the film. The differences of the refractive index can be simply explained by the lower density of the alkyl-substituted films. This is also a valid explanation for the lower refractive index of polythiophenes and the higher one of amorphous Si. The refractive index was studied as a function of chain-length by evanescent wave spectroscopy [102, 1851. The results are n = 1.835-1.857 for a-4T, n = 1.950-1.966 for a-6T, and n = 2.076 for a-8T measured at X = 632.8 nm, i.e. a linear increase of n with the chain-length is observed. The different values are obtained for different film thicknesses and different waves (surface plasmons, waveguide modes). Much lower values are reported by Zhao et al. from quasiwaveguide measurements [255] which reveal n = 1.562 for a-3T, n = 1.581 for a-4T, n = 1.600 for a-5T, and 1.623 for a-6T. Egelhaaf et al. evaluated the anisotropy of the refractive index n and the absorption coefficient K and find for perpendicularly oriented a-5T films (0 = 15’) n, = 0.70, n, = 1 . 7 0 , = ~ ~1.14,and K~ = 0.04for a wavelength X = 345 nm. The index n denotes the value normal to the sample surface, t the value in the film plane [ 1811.

4.3 ThinJilms

253

Emission spectra Fluorescence spectra of oligothiophene thin films on glass are summarized in Fig. 40 [181, 2491 and are in line with most of the literature spectra (see, e.g. [256-2581). At their high energy side the fluorescence spectra exhibit a shoulder Fsh which is followed by two bands F1and F2. The fluorescence spectra are better resolved than the corresponding absorption spectra which indicates a narrow distribution of excited-state molecular geometries, characterized by a more rigid planar conformation of the thiophene rings if compared to the ground-state configuration. In the latter the rotational barrier is rather small and therefore a torsional disorder can arise. This is not possible in the quinoid-like excited state. The peak positions plotted against the reciprocal chain length were given above and show straight lines which run parallel to those of Ao, A,,,, and the respective fluorescence peaks in solution [1811. Therefore the fluorescence results from the S,,1* singlet state. The Stokes shift between absorption peak A. and Fshin these films is 1150cm-', i.e. 0.14eV and 2500cm-', i.e. 0.31eV between A. and F1 and can be explained by torsional disorder which induces inhomogeneous broadening [257,259]. Highly ordered films on HOPG [260] and Ag(ll1) [149], however, do not show a Stokes shift. From this and the fact that the peak position vs. reciprocal chain-length shows the same slope as F1, F2, and the first absorption peak it seems appropriate to identify the shoulder as the O-O-transition. A more detailed discussion is given in [250]. Fluorescence quantum yields are also summarized in [250]. a-5T-a-8T show fluorin highly ordered thin films (d = 3-5 nm) and escence quantum yields of QF 5 around 1 order of magnitude higher values in thicker films [171, 181, 2491. For the same thickness the yield is smaller for more ordered films which is expected for molecules with strong excitonic coupling. Microcrystalline a-4T has a yield of aF= 7 x lop3 even in thin films, and of aF= 2 x in thick films. The yields of end-capped oligothiophenes are around 3 x for all layer thicknesses due to their lower tendency to form ordered films. Furthermore the fluorescence yields of a-substituted (highly ordered) films are much lower if compared to /?-substituted (disordered) films [261, 2621. On cooling to 77K the quantum yields are increased by a factor of 3-4 [181]. The fluorescence excitation spectra are almost identical to the absorption spectra pointing at a constant fluorescence quantum yield over the whole range of the spectrum [ 18I]. Dippel et al. find in low temperature measurements on a-6T an increase of the fluorescence quantum yield below 2.3 eV [257]. The group around Taliani (see, e.g. [25,237,247,263,264]) find a large photoexcitation signal at around 2.17 eV for a-6T thin films measured at 4.2K which is attributed to the excitation into the 'B, state (compare Chapter 6 of this book [28]). In many fluorescence as well as in some absorption spectra on films on glass also a low energy component can be identified. This may be attributed to gap states due to physical defects. These states are called X-traps and show a spectral distribution for more than 2000 cm-', i.e. 0.25 eV (compare Fig. 41 below) [237]. They are absent on highly ordered films of a-4T on Ag(ll1) [149]. The fluorescence decay is strongly non-exponential and can be fitted satisfactorily by a sum of three exponentials with T < 100 ps, T M 250 ps, and T M 1ns for a-4T,

254

4 Structure and Properties of Oligothiophenes in the Solid State

Figure 40. (a) Fluorescence (F) and absorption (Abs) or fluorescence excitation (FE) spectra of a-nT thin films on quartz glass. Excitation in the maximum of absorption band A, angle of incidence 60". (b) UV-Vis transition energies vs. reciprocal chain-length for the fluorescence peaks and the first absorption peak of a-nT thin films. The lower index A indicates room temperature spectra, the lower index B spectra taken at 77 [181].

4.3 Thin3lrns

0.0

0.1

1/ "

0.2

255

0.3

Figure 40 (continued).

a-5T, and endcapped a-5T [159]. Also other techniques probing the excited states dynamics show several decaying processes with different time constants (see, e.g. for a-6T [265, 2661). Probably a distribution of decay times exists, which can be ascribed to the effect of disorder on the diffusion-limited recombination by dispersive diffusion (i.e. time dependent hopping rate) to recombination centers or recombination center local density fluctuations [265].

Figure 41. Schematic presentation of probable decay mechanisms in oligothiophene thin films. X: X-traps; k,, k,,, kf, kCT:rate constants for radiative, non-radiative, and fluorescence decay and the transition into the charge-transfer state, respectively; kT: thermal energy; E: external or internal electric field [18 11.

256

4 Structure and Properties of Oligothiophenes in the Solid State

The low fluorescence quantum yield and the fact that no phosphorescence could be observed in oligothiophenes leads to the conclusion that most of the electronic excitation energy decays radiationless. There are different views of the participation of triplet states in this decay process in thin films [39, 236, 247, 249, 257, 259, 267-2741. In solution, however, the relaxation via triplet states is well agreed. From the broad distribution of fluorescence decay times and the broad line-width for fluorescence bands of films on quartz glass substrates at 4K (see, e.g. [275]), it can be concluded that emission results not from a defined S1,l*state but from a variety of states derived from Sl,l* by intermolecular interaction. This is also in line with the two orders of magnitude smaller oscillator strength of the emission if compared to the first absorption band. Figure 41 summarizes probable decay mechanisms in oligothiophene aggregates without triplet state contributions [181]. After excitation into the S 1 , ~ *state the system relaxes to the lowest exciton band state from which optical transitions are dipole forbidden. One decay channel is the trapping of electrons in X-traps from which radiative transmissions are allowed due to the lowered symmetry near the defects. As competitive processes trapping in non-radiative traps and the formation of charge transfer excitons, i.e. charges on two adjacent molecules, occur. In presence of internal or external fields these charges can be separated and contribute to the photocurrent. The rate constants are kcT x >> For a model taking triplet states into account see, e.g. [273].

er e.

Nonlinear optical properties There are several studies on the cubic susceptibilitiesx ( ~This ) . value is derived from the field dependence of the (macroscopic) polarization

x is the linear, x(2)the quadratic, and x(3)the cubic susceptibility. The quadratic susceptibility is zero for an overall centrosymmetry. The macroscopic values arise from molecular properties, i.e. the dipole moment, which is defined as p=

+ a E + i P E 2 + 1/6yE3

(2)

Here po is the permanent dipole moment, Q the (linear) polarizability and and y are the first and second hyperpolarizability ,respectively. The first hyperpolarizability is zero for centrosymmetric molecules. Whereas in theoretical papers the hyperpolarizabilities are calculated, in real devices always susceptibilities are measured from which the hyperpolarizabilities can be deduced. As the oligothiophenes with even ring number are centrosymmetric only third order effects are important in these systems and no report about second order effects in odd numbered thiophenes are reported so far. In general conjugated polymers are of interest for non-linear optics due to their one-dimensional delocalization and correlation of their 7r-electrons which leads to relatively large third-order optical non-linearities. Of principal interest is the

4.3 Thinfilms

257

development of the second hyperpolarizability with increasing conjugation length and a possible saturation due to the finite delocalization length of the 7r-electrons. There are several theoretical predictions concerning the influence of the electronic structure on the second order hyperpolarizability which are summarized in [276][277], see also Chapter 5 of this book. The susceptibilities, the hyperpolarizabilities, and the linear polarizabilities of several oligothiophenes in different matrices are summarized in [250]. The values deviate sometimes substantially from each other, but there are no data available on exactly the same molecule in the same medium, measured with the same method. Therefore differences due to density changes and, in particular, dispersion and resonance enhancement near absorption regions can be responsible for this. Furthermore the experiments show different degeneracy factors leading to systematic deviations of 37THG = y4wM,6yTHG= YEFISH (THG = third harmonic generation, 4WM = four wave mixing, EFISH = electric field induced second harmonic generation) [278]. The only thin film data are given by Fichou et al. [loll with x ( ~=) 1.88 x lopi2esu esu for a-6T, both measured with THG at for a-5T and x ( ~=) 2.38 x X = 1907 nm.

4.3.3.2 Charges in oligothiophenes Upon the formation of positive and negative charges new electronic states are created. These charges can either result from chemical or electrochemical doping upon which positively or negatively charged molecules (with counterions) are produced. Alternatively photoexcitation is possible provided that the photon energy is above a minimum threshold generally at an energy higher than the first excited singlet state. Upon photoexcitation always a positive and a negative charge are created simultaneously within a neutral system. Because the counterions influence the local electric field no exact agreement can be expected for the energetic position of the states produced by the two methods [279, 2801. Electric field excitation will be discussed in Chapter 9 in view of applications and will not be stressed here. The new states can either be interpreted in the band picture as polarons and/or bipolarons which occur due to electron-phonon coupling after ionization (see below) (electron-electron interaction is neglected) or as the new HOMO and LUMO states in charged particles if localized molecular orbitals are assumed (electron-electron correlation is taken into account, electron-phonon coupling neglected). Horowitz et al. [281] discuss a transition between short oligomers which are better described in terms of molecular orbitals whereas the one-electron band model of conjugated polymers can be applied for longer oligothiophenes and the polymer. The transition between these two regions is assumed to be between 9-1 1 rings. In shorter oligomers the polaron state, in longer ones the bipolaron state would be more stable. This is in line with ab initio studies by Ehrendorfer and Karpfen which state that the spatial extent of a bipolaron is 9-1 1 thiophene rings (in unsubstituted molecules) with the distinct quinoid structure extending over 5-7 rings [282] whereas BrCdas et al. calculate the extension of the polaron over six monomeric units [283]. Irle and

258

4 Structure and Properties of Oligothiophenes in the Solid State

Lischka showed in a recent paper that the extension of a bipolaron state decreased from 11 thiophene units without to 2 units in the presence of the counter-ion (C1- for p-, Li' for n-doping) [280]. Both models will be discussed in parallel in the following subsections. However, in general discussions the band structure terms will be used as they are more common in literature. In non-degenerate groundstate conjugated polymers two new gap-states evolve upon charge generation (see, e.g. [284-2861). In weakly doped materials often so-called polaron states are created, whereas at higher doping level either bipolarons, bipolaron bands, or polaron bands can be formed due to the interaction between polarons. Bipolarons are spinless states and can therefore in principle be distinguished from polarons and polaron bands by ESR (electron spin resonance) measurements. Nevertheless the problem whether polarons or bipolarons are formed in poly- and oligothiophene thin films is still under discussion (see below). In any case defects arise as quinoid structural elements within the aromatic system (Fig. 42). However, as mentioned above, the defects are not as localized as Fig. 42 implies, but are extended over several double bonds. Figure 43 summarizes schematically the different energetic situations for symmetric gap-states. Symmetric gap-states are proposed by several calculations from Bredas [287-2891 and Heeger [285], whereas Springborg [290,291] finds asymmetrically lying states. Bertho and Jouanin [292] find nearly symmetric polaron and asymmetric bipolaron states.

r

1

S

I

J

I

Relaxation

r

1

-e' r

1

I 1

r

+el

Figure 42. Schematic representation of a n T in the undoped, singly oxidized, and doubly oxidized (left) and reduced (right) state.

(

,

Ev

I -+-I--

+-+-

Figure 43. Schematic representation of successive (a) p-doping and (b) n-doping in a band model. From left to right: undoped state, polaron states (here: symmetric) for lightly doped an-T, bipolaron states (above, here: symmetric) or polaron bands (below) for intermediate to strongly doped a-nT, bipolaron bands for strongly doped a-nT. The polaron and bipolaron states originate from the valence and conduction band near edge states of the undoped material. The dashed areas mark occupied bands.

260

4 Structure and Properties of Oligothiophenes in the Solid State

p-type doping It is commonly accepted that in solution @-6Tis oxidized in two steps, first to the radical cation, then to the dication. The cationic state is believed to exist as a .rr-dimer in which two radical cations couple via their r-systems and which has a nearly spinless state [292-2971. This 7r-dimer forms as easier as longer the chain length and as lower the temperature is due to its exothermic building process. ESR measurements show that the mono-oxidized oligothiophene is a (nearly) free radical with an intense and narrow Lorentzian signal centered at g = 2.0023-2.0025 for unsubstituted [298] as well as alkyl-substituted a-nT 12931. The dication formation is only possible for @6T and longer oligomers, a-4T and a-5T either do not react or dimerize to octi- and decithiophene, respectively [299]. (For a review on results obtained in solution, compare [250].) In the solid state, however, the nature of the charge carriers remains unclear. The main reason is the very few investigations made with different materials, dopants, and characterization methods which do not allow for an easy interpretation. The radical cation precipitates as a salt from CH2C12solutions as a black powder. The disadvantage of this method for thin film formation is the simultaneous precipitation of FeC12, solvent molecules, and unreacted species. The black solid exhibits a sharp (4 G) Dysonian ESR signal if HC104 is used as oxidant. (FeC13 as oxidant masks the ESR spectrum by the high concentration of Fe3+ions [300,301].) A Dysonian line-shape shows a high asymmetry and is characteristic for metallic behavior [302]. From the weak intensity of the IR ring modes if compared to translational modes it can be concluded that the sulfur atom does not strongly contribute to the electronic structure of a-6T'+ [300]. This is also in line with ESR data on powder samples of dimethyloligothiophenes [303], the free radical character observed in solution (see above), and with results on monothiophene in solution which show that the single electron occupies an 'a2 orbital with a node at the sulfur atom [304, 3051. Below it will be pointed out that the anionic state shows a quite different behavior. There are only few published results on intentionally doped oligothiophene thin films. Doping of films can in principle be established by exposure to iodine vapor, doping with acceptors in solution in which the oligothiophene is insoluble, or by electrochemical doping. Alternatively thin films can be doped in situ by cosublimation of the oligothiophene and FeCI3 under UHV conditions whch leads to ultraclean materials [306-3081. In Fig. 44 the UPS spectra of pristine a-6T, FeC13, and doped a-6T with different doping levels are shown. The high binding energy region of the spectra of the doped material is basically determined by the superposition of the spectra of pure a-6T and pure FeC13. But also a shift of the whole spectrum towards the Fermi level EF is observed. This proves the success of the p-type solid state doping process. The plot of changes in the work function, core level shifts, valence level shifts, and the energetic difference between valence band and Fermi level against the dopant concentration reveals two distinct regions. The first reveals pronounced changes, in the second a saturation of all energy level shifts and a linear change of the work function can be observed. This can be interpreted as there is no further doping in the second region with an intermixing of FeC13 and the doped

4.3 Thin$lms

(a)

261

/UPS H e l l

I

l

14

~

l

10

12

t

I

2.0

1.5

'

i

'

8 6 Binding Energy (ev)

I

1.o

l

4

~

2

l

~

l

~

0

1

I

0.5

0.0

Binding Energy (eV)

Figure 44. (a) UPS He I1 spectra of undoped a-6T, pure FeC13, and two differently doped a-6T films. The percentages refer to the reaction a-6T 12 FeCI3+ c~6-T6~'~++ 6 FeC1;+ 6 FeCI2, i.e. 100% correlates with one charge and radical per thiophene ring to allow for comparison with polymer data. The vertical dotted lines indicate the positions of the lower a-band which was used as stable, i.e. dopant independent, reference to determine peak shifts. (b) UPS He I spectra of undoped a-6T and for a-6T with increasing doping levels. The spectra were calibrated as indicated in (a). Also indicated is the Fermi level EF as determined from a clean Au foil [306].

+

a-6T. The second region starts at a 1 : 1 molar ratio of a-6T and FeC13.This can be interpreted in two ways: firstly, only 1 mole FeC13 is needed for the charge transfer of one charge to the molecule. In contrast to solution the counterion in the solid state would then be C1- instead of [FeClJ which seems to be a reasonable

262

4 Structure and Properties of Oligothiophenes in the Solid State

assumption. Alternatively only 50% of the molecules can be doped by this solid state reaction which would be in line with the interpretation by Hotta et al. that dimers of a cation radical and a neutral molecule are formed (see below). The same situation occurs for a-5T and a-7T which can also be doped to the same extent. The Fermi level shift is accompanied by changes in the HREELS spectra shown in Fig. 45. In comparison to the spectra of undoped a-6T (compare also Fig. 38) two additional loss structures are observed at 0.5 eV and 1.1 eV with a third very weak structure at 1.7eV. These spectra indicate polaron formation with the two expected transitions. On the other hand these energies differ from those in solution and from those obtained by photoexcitation in thin films (see below). In principle this is not surprising and a red-shift of the absorption peaks of doped (substituted) oligothiophene species in the solid state is also observed by Hotta et al. [252], but at different absolute positions (see below). But it is in contrast to results on tetradecyl-a12T which shows completely similar spectra of electrochemically doped thin films and chemically doped molecules in CH2C12solution with two absorption peaks at 0.85eV and 1.7eV [253, 2541. The authors attribute these transitions to bipolarons but do not explain why they find two and not only the one expected transition. Cornil et al. try to explain this discrepancy by the formation of two interacting bipolarons but state themselves that the mild applied conditions should not lead to such a highly oxidized species [309] so that the origin of these transitions remains unclear. Absorption peaks at 0.55 eV and 1.01 eV are also found by Harrison et al. in spectra of field-induced charges [41, 3101 and were attributed to 7r-dimers in highly ordered films. Because Oeter et al. did not study the geometric structure of their

5

Energy Loss (ev) Figure 45. HREELS spectra of undoped a-6T, pure FeCI3,and doped a-6T (9%) (for definition compare Fig. 44). The dotted lines indicate the positions of the polaron levels. The inset shows these polaron states, the optically allowed (0.5eV, 1.1 eV) and the forbiddem (1.7 eV) transitions [306].

4.3 Thinfilms

263

films this interpretation could not be proved so far although in thick films intermixed with FeC13 on Au substrates a highly disordered film is expected. Different absorption peaks are found by Hotta et al. by doping vacuum evaporated dimethyl-a-6T in solution with iodine, NOPF6, or NOBF4 although the undoped species resemble the same ‘band-gap’ transition of 2.3 eV as unsubstituted a-6T [252]. For iodine doping and low concentrations of the nitrosyl salts two features at 0.71 eV and 1.44eV were found whereas higher concentrations of NO’ result in one absorption peak at around 1.15eV. In contrast to solution spectra the former HOMO-LUMO (7r-n*) transition of the neutral molecule is not bleached but even enhanced at the low energy side. Hotta et al. explain the intense 7r-r*transition by forming n-dimers which are only singly charged, i.e. consist of a radical cation and a neutral molecule, in contrast to the 7r-dimers discussed in solution. Furthermore sometimes additional weak structures at 0.84 eV and 1.66 eV arise in the solid state upon doping which are not present in solution. They are explained by vibronic transitions. The higher oxidation state with only one transition is explained by either a dicationic 7r-dimer or by a bipolaronic state. Bromine-doping of condensed bithiophene layers is reported by Ramsey et al. [311]. In the UPS spectra a shift towards the Fermi level, a new 4.3 eV emission, and some smaller changes in the intensities of the upper two 7r-bands are observed. The ELS (electron energy loss) spectra reveal two transitions at 2 eV and 1.2eV attributed to absorptions due to the presence of polaron states and a gap state excitation at 2 eV for higher dopant concentrations due to bipolaron formation. In summary, one probable explanation of the given results could be the formation of either 7r-dimers as in solution for highly ordered films or the formation of a different kind of 7r-dimers in which a neutral and a radical cation are combined. Due to the lack of ESR measurements of such film structures no final conclusions can be drawn. The formation of bipolarons in thin films seems to be only possible, if at all possible, if very strong oxidizing conditions are applied. n-type doping There is even less known about the n-doping of oligothiophenes. In solution the electrochemical reduction or the reaction of alkali metals in THF leads to mono- and dianions without any evidence for 7r-dimer formation. The ESR signal of the monoanion reveals g = 2.0046 for didodecylsexithiophene which is substantially different from the cation and the free electron. It can be concluded that in the case of the anion the sulfur orbitals contribute to a large extent to the singly occupied molecular orbital of the anion [293]. n-doping of thin films is mainly established by alkali metal deposition onto the asprepared thiophene films, which leads to a complete intermixing even at temperatures as low as 100K. Also reported is the n-type electrical behavior for sexithiophene annealed at 423K in air [18, 3121. Published data so far concern mainly the electronic structure in the valence band region as determined by UPS, ELS, and comparative theoretical calculations [313-3171. Recently also HREELS measurements were performed [318, 3191.

264

4 Structure and Properties of Oligothiophenes in the Solid State

The first published data concern the n-doping of condensed bithiophene films [314, 3151. Dosing with Cs initially results in two sharp loss features at 1.5eV and 2.7eV in the ELS spectra which grow with increasing Cs exposure at the expense of the original T-T* transition. At concentrations of about one Cs atom per two bithiophene molecules these two states in the gap dominate the loss spectrum. On further exposure the 1.5 eV feature wanes while the 2.7 eV loss suffers a gradual shift to lower loss energy until it settles at 2.3eV as a broad feature for more than one Cs atom per bithiophene molecule. The observed energies are different to those observed on p-type doping of the same molecules by bromine exposure (see above [3111) which is explained by asymmetric gap states as proposed by Springborg [290, 2911. In UPS for initial Cs exposures two states in the band gap are observed with binding energies of 1.1 eV and 3.5 eV referenced to the Fermi level. They grow in intensity and are replaced by peaks at 2.1 eV and 4 eV for increasing Cs concentrations. Both results are interpreted as a polaron to bipolaron transition. From UPS cross sections, XPS, and NEXAFS results a high localization of the negative charge at the sulfur is deduced. Calculations by Irle and Lischka [313] and Brtdas et al. [317] also indicate a considerable amount of negative charge transfer to the sulfur atoms which is more pronounced for the inner rings. (The total charge transfer as calculated by Irle and Lischka is between 0.6 and 1 electron per Li atom.) This localization is also in line with measurements on Na [318] and Cs doped sexithiophene [319] in which new S2s and S2p components appear upon doping which are shifted to lower binding energies by 3.7 eV. Such a shift is expected if a high negative charge density is located at the sulfur atom. Furthermore Logdlund et al. [316] conclude from the sodium and sulfur XPS intensities that the saturation doping is about one sodium, i.e. one charge per thiophene ring which would be impossible for largely extended polaron or bipolaron states. In other measurements an even higher saturation concentration of Na is found (see below) [318]. Upon Na doping a shift of the spectra away from the Fermi level is observed as expected for n-type doping and a new structure appears in the UPS spectra which is located at 0.55eV [318] respectively 0.60eV [316] away from the valence band edge, i.e. 1.13eV from the HOMO peak maximum in both papers. Logdlund et al. could explain all observed shifts by ab initio and local spin density approximation calculations in which the alkali metal was explicitly taken into account whereas VEH calculations without the counterion failed to explain some of the observed results. At a concentration of 1.3 Na/a-6T HREELS spectra of Na-doped sexithiophene [318] reveal a new, very broad loss-structure at 1.7eV (Fig. 46) (for bare sexithiophene compare also Fig. 38). With further Na-deposition the intensity of this loss peak increases very strongly and dominates the spectrum at the highest concentration of 8.3 Na/a-6T. Simultaneously a loss peak at 3.9 eV arises. Both effects lead to a less clear peak structure at energies between 1.O and 4.5 eV due to the broadness of the peaks. Depending on the Na concentration there may also maxima be detected at 1.3 and 2.1 eV which arise within the 1.7eV peak structure as shoulders. There may also be a new peak at 0.95 eV for 3.2 Na/a-6T accompanied by a slight increase of the peak intensity at 0.75 eV. This intensity increase can only be explained by an additional electronic energy loss at 0.75 eV because the intensity of the overtone of the C-H stretch vibration is constant. The new peaks can either be explained as

4.3 Thinjilrns

0.0

1.0

2.0 3.0 Energy Loss (eV)

265

4.0

Figure 46. HREELS spectrum of a 70 nm thick a-6T film with increasing Na concentration. The inset shows a possible energy scheme with polaron levels, the arrows indicate transitions discussed in the text [318].

polaron levels or as the energies of the HOMO and LUMO of a new Na-sexithiophene compound. Cs doping leads to qualitatively similar results, but the new peaks are located at different positions [319]. This favors the localized picture which is in line with all experiments, in particular with the low conductivity (compare section 4.3.4.7). For a detailed discussion, compare [318]. Photoinduced charge carriers The photogeneration of charged excitations in oligothiophene thin films was only studied for a-6T in which the threshold for photoexcitation of charges has been located at around 2.2eV by a comparison of the one photon excitation curve for radiative recombination and the photoconduction action spectrum [257]. Poplawski et al. [271, 2721 find in their photoinduced absorption (PA) spectrum two strong bands at 0.80eV and 1.54eV and smaller absorption peaks located at 0.96eV, 1.08eV, and 1.68eV. Similar results were also published by the group around Taliani [270, 320, 3211 which find two intense photoinduced bands at 0.74eV and 1.43eV as well as two minor features at 0.92eV and 1.085eV. The peaks at 0.92 eV and 0.74 eV show a characteristic decay time of 3 ms, the weak structure at 1.085eV one of 500 ps. Lane et al. [322, 3231 detect three dominant PA bands at 10K at 0.8, 1.1, and 1.54 eV and side-bands at 0.97, 1.27, and 1.7 eV. All groups correlate the two strong absorptions at 0.74/0.8 and at 1.43/1.54eV with positive polaron states due to their similarity with Fichou et al.’s results on a-6T which was chemically p-doped in solution [300, 3241. This is corroborated by the measurement of PA-detected magnetic resonance (PADMR) where these excitations show a strong negative PADMR signal indicating a spin l/2 excitation [322, 3231. The peak at 0.97/0.96/0.92 eV is associated to the same charged state and assigned to a vibronic side-band because it shows the same temperature- and frequency-dependence as the other states and the spacing is similar to the vibronic splitting in undoped a-6T (compare section 4.4.1.2). The same holds for those

266

4 Structure and Properties of Oligothiophenes in the Solid State

peaks at 1.27 and 1.7 detected by Lane el al. The absorption at 1.08/1.1eV may be explained as the absorption to a bipolaron (dication) state due to its similarity to bipolaron absorption in solution. This assignment is also corroborated by the PADMR measurements by Lane et al. On the other hand this absorption is also found under very dilute conditions where bipolarons should not be formed. Therefore Taliani et al. suggest as alternative explanation a triplet-triplet transition. Different explanations for the absence of additional absorption features for the also formed negative charges are given by the two groups of Poplawski and Taliani. Poplawski et al. suggest that the negative charges are trapped outside the oligomer molecules and do not contribute to the spectrum. These negative charges are not able to recombine with the positive charges situated on the molecule, thus giving rise to the very long life time of the polaron states. Also the positive polarons seem to be bound to traps with a thermal activation energy of 49.5 meV. Another explanation for the absence of other absorptions is given by Taliani et al. which suggest that the spectra of positive and negative polarons do not d a e r in energy for more than 0.04eV. In fact this is in contrast to the results by Bauerle et al. in solution [293] which find 0.87eV and 1.60eV for p-doping and 0.72eV and 1.58eV for n-doping and with the findings by our group (compare results by Oeter et al. [306-3081 and by Murr et al. [318, 3191 given above). The long life time is explained by a separation of the charges into different molecular layers which are weakly correlated [257,320]. The charge separation follows the formation of a neutral Frenkel exciton with binding energy 0.4eV which thermalizes into a charge-transfer exciton which then eventually dissociates [257, 259, 3201. Recently Egelhaaf et al. produced oligothiophene radical cations on silica gel by two-photon ionization with laser intensities Z 2 5 x lo6cmP2 and laser flash energies above 15mJ [181]. The electron is excited into a continuum state and no anion is formed. The absorption spectra are very similar to those observed for radical cations in solution. The ESR signal exhibits a signal with g = 2.0028 at room temperature. At T = 77K an anisotropic ESR spectrum is found. There is no evidence of T-dimer formation or dimerization to an oligomer with double chain length as in solution, even not for a-2T and a-3T. After one day the original spectra of the undoped materials are revealed. Apparently the radical cations remain nearly immobile on the surface.

4.3.4 Electrical characterization In this section only experimental results are summarized. For theoretical considerations compare Chapter 7 of this book by Horowitz and Delannoy [32].

4.3.4.1 General remarks

Absolute values of the electrical conductivity vary from group to group due to the often undefined preparation conditions and hence doping levels and structural

4.3 Thinjilms

267

order. The latter is also influenced by the choice of the substrate material, as outlined in section 4.3.2. Furthermore two-point and four-point measurements are realized under varying gas exposure or vacuum conditions and often undefined light exposure. Therefore a comparison between results of different groups is extremely difficult because in many papers the experimental conditions are not even mentioned. Dark conductivities of pure (UHV-prepared) a-6T thin films and most probably of all other oligothiophenes lie below lop9S/cm-' due to their large band-gap or, in the molecular picture, HOMO-LUMO energetic difference. Air-exposure or light-exposure increase the specific conductivity to values in the range between 10-6-10-9 Scm-', even hours after switching off the light or pumping down the system. Such low conductivities of pure materials were also found for a-6T single crystals for which the diffusion coefficient of possible dopants like oxygen is extremely low [240] and for compressed powders measured under vacuum conditions. Most of the data presented below in Tables 9-1 1 are therefore data of unintentionally (oxygen) doped compounds rather than intrinsic properties. From the results on the electronic structure of doped materials and results on ultraclean anthracene crystals from Karl et al. (see, e.g. [325] and references therein) Garnier et al. speculate on the mechanism of the charge transport [132]: In this model a positive charge is injected into the organic molecular material from the electrode, a polaronic type radical cation is created, and (variable range) hopping of this charge between adjacent molecules generates the overall charge transport. This charge hopping is assumed by most groups (but see the results from Vaterlein et al. [326] presented below or by Zotti et al. [327, 3281) although no direct proof exists. However, many results as those on the structure dependent conductivity presented below and on field effect transport between metal islands [329] strongly support the model of hopping transport. 4.3.4.2 Contacts, I/V-curves, carrier injection

Several different principal types of electrical contacts to thin films are possible. They can be devided into two categories concerning the direction of the measured charge transport with respect to the substrate surface. Most often devices which measure transport in the direction along the substrate surface are used, because the contacts can be prepared onto the (insulating) substrates before the organic film is prepared and they are often commercially available. Typical contact materials for such devices comprise Au or Pt, typical substrate materials include silica-coated Si-wafers, quartz, and sapphire. Much more complicated is the preparation of top-contacts to measure transport perpendicular to the (conducting) substrate surface. If the typical evaporation is used for the preparation a large amount of heat is transferred to the organic film during deposition. Due to the lack of strong intermolecular bonding the hot metals can easily diffuse into the film and can even form short-circuits. (For metal reactions, compare [250] and references therein.) In case of pure electrical measurements as substrates and contact materials Au, Ag, or Pt are preferred due to the p-type behavior of as-prepared oligothiophenes.

268

4 Structure and Properties of Oligothiophenes in the Solid State

If simple band models are assumed for an-T and the contacts, materials like the noble metals with a workfunction of 5.3 eV (Au) or 5.6 eV (Pt) lead to ohmic contacts whereas materials with low workfunctions as A1 (4.28 eV), Mg (3.66 eV), or Ca (2.87 eV) form Schottky barriers. The rectification ratio I(+U)/I(-U) was determined for endcapped a-6T in a LED device to be 240 for Ca, 7 for Mg, and 40 for A1 [330]. This shows that the work function is not the only factor influencing the Schottky barrier height, but that also trap states or an interfacial layer due to reactions between metal and thiophene may play a role. The influence of interface layers on Schottky barriers is also shown for In [331] and eutectic Ga,In [332] on p-doped dodecathiophene. For other Schottky diodes, compare [250] and references therein. Both types of contacts are necessary for electro-optical measurements. Here also one electrode has to be optically transparent. The most common material for the latter purpose is indium-oxide doped tin-oxide (ITO). This material is highly transparent and highly conductive but has the problem that the substrates always exhibit several ‘spikes’ standing out of the surface. The other type of semitransparent electrodes are ultrathin metal films evaporated onto the organic film. Only few attempts have been made so far to use conducting polymers or other organic materials as contacts. However, TCNQ has been used recently to optimize the contact resistance between a-4T and Au contacts [333, 3341. The influence of TCNQ is assumed to make a layer in which a-4T is doped by TCNQ. Due to the fact that TCNQ is a larger molecule with extended 7r-system it does not diffuse in high electrical fields. (For the influence of mobile dopants on FET characteristics, compare [335]). With ohmic contacts, the current-voltage relation is often ohmic in nature up to a certain value and then becomes space-charge limited. This can be seen by a linear relation between current I (or current density j ) and voltage V at low voltages and a quadratic dependence for higher voltages [336]:

jscLc = 9 / 8 d ? p V 2 / d 3

(3) with E as dielectric constant, 0 the fraction of the total charges free to move, which depends on the trap density, the activation energy to move a charge from the trap to the band, and the temperature, p the mobility, V the voltage, and d the thickness of the sample. Space-charge limited currents (SCLC) also lead to a field independence of the slope of Arrhenius ( I vs. 1/T) plots. Both behaviors are generally found (but see below) for not intentionally doped and therefore p-type oligothiophenes contacted by noble metals like Au [18,337,338]. The transition from the linear to the quadratic dependence of the I/V-curve is often found around 1 V [337, 3381. However, the linear dependence at low voltages does not in all cases imply ohmic behavior, as is pointed out by Horowitz et al. [338]. Additionally the linear slope of the I/V-curve and hence the specific conductivity has to be independent of the film thickness which is not found for Au/a-6T/Au sandwich structures with a-6T layer thicknesses below 2 pm. On the other hand not only field injected carriers are possible as in the case of the normal SCLC regime but also thermally injected carriers without any applied voltage. This leads also to a linear I/V-curve at low applied

4.3 Thinfilms

269

voltages. At low voltages therefore two types of carriers contribute to the overall current, the thermally injected charges and the non-injected bulk free carriers which generate the ohmic conductivity. This model also implies the independence of the transition voltage between the linear and quadratic regime from the layer thickness, the transport properties, or the trapping parameters in the film because always injected charges play the major role. This is in line with the above-mentioned results that 1 V is found for different film thicknesses [338] and even for compressed powder samples [337]. Egelhaaf et al. found SCLC behavior for a-5T on Au comb structures even for voltages as low as 100 meV, whereas on Pt comb structures no SCLC could be detected up to 100 V [181]. This is in line with results from Vaterlein et al. [326] which do not find any evidence of SCLC in their a-6T thin films on Pt comb structures measured in high vacuum. The density of acceptors in nonintentionally doped films varies quite largely from 3 x lOI3cmP3 for a-6T derived from I/V-measurements in the SCLC regime [338] over 5 x 1015cmP3 for unintentionally doped side-chain substituted dodecathiophene [339] to 2 x 1017cmP3for a-6T [17, 19,340,3411 and EC6T [326] determined from capacitance vs. voltage measurements at a Schottky junction.

4.3.4.3 Influence of the structure on conductivity data

Charge transport by hopping conduction will depend on the intermolecular distance and on the .rr-orbitaloverlap and hence on the film structure. To prove this assumption the group around Garnier systematically investigated the influence of the film structure on the value and the anisotropy of the conductivity. For this purpose a-6T, end-substituted a-6T, and side-chain-substituted a-6T were deposited onto different substrates [132, 342, 3431. To measure the conductivities in the direction of the substrate plane, films were evaporated onto a planar geometry of evaporated Au electrodes. The substrate material is probably an insulating oxide like SiOz or A1203. Conductivities measured perpendicular to the substrate surface were performed with the film sandwiched between two vacuum-evaporated Au contacts. The results are summarized in Table 9 together with results from other groups on similar materials.

Table 9. Conductivity data of unintentionally doped a-6T: influence of substitution. Molecule* 0-6T ~,u-DHcx-~T EC6T P,P’-DHa-6T P,P’-DDa-6T

(S/cm) 1x 6x

10-l~

4.9

crL (Sjcm)

Reference

1 x 10-~ 5 10-~

11321 ~321

W I

10-~ 10-l~

10-l~

~321 [243, 3431

270

4 Structure and Properties of Oligothiophenes in the Solid State

From this it can be concluded that the better the structural order in the film, i.e. best for end-substituted, worst for side-chain substituted a-6T, the higher the conductivity and the higher its anisotropy. The absolute values and ratios, however, are critical to interpret due to several reasons: The films were evaporated onto two different substrates which can lead to a completely different film structure, as discussed in section 4.3.2. The orientation of the molecules in the measurements parallel and perpendicular to the surface is therefore not necessarily the same. Furthermore the orientation of the end-substituted molecules is assumed to have the alkyl-substituents oriented perpendicular to the substrate surface. Therefore in the measurement of the conductivity the charges have to be injected through saturated alkyl-spacers which also separate each oligothiophene layer from the next. Even if the transport would be better along the long molecular axes, in a,wsubstituted molecules this would be covered by the influence of the substituents. Last but not least the indiffusion of Au into the layer could be different for the three different materials, leading to different specific conductivities. This, however, should lead to formally higher conductivities for less ordered films due to the higher diffusion coefficient in such structures which was not found in these measurements. The difficulty of such measurements is also seen in the investigations by Servet et al. [56] which correlate their X-ray data on differently prepared a-6T films with conductivity data. Although the structural order increases with increasing substrate temperature a decrease in the absolute value is obtained from 6 x S cm-' for Tsub= 77K to 1.2 x 1 0 - ~Scm-' for Tsub = 553K (conductivity measured parallel to the substrate surface) which is attributed to the desorption of impurities from the substrate which can act as dopants. The anisotropy, however, is increased as expected and the field effect mobility remains nearly constant. These results are in line with results obtained on single crystals (compare section 4.2.3).

4.3.4.4 Influence of the structure on mobility data

That the main effect of the enhanced conductivity by higher structural order is the enhanced mobility is corroborated by measurements of field effect mobilities (Table 10) [loo, 132, 333, 334, 342-3461, although field effect mobilities are lower if compared to intrinsic mobilities due to parasitic series resistances in short channel devices [347][348] and channel length shortening [348, 3491. Today thin films can be prepared in which the mobility of a-6T films nearly meets that of single crystals and those of amorphous silicon (10-3-1 cm2V-' s-I). Furthermore newer measurements on highly ordered a-4T and a-5T films [333, 3341 exhibit 4 respectively 2 orders of magnitude higher mobilities which are similar to the values of a-6T, although 'leaky' PMMA gates were used instead of SiOz. Also better purification of a-8T leads to higher mobilities [loo]. The field effect mobility may also be decreased by interface states. This is supported by the increase of the field effect mobility by using different insulators [250, 3501. Also drift mobilities extracted from electroluminescence measurements show a thickness dependence pointing to an influence of the interface [351].

4.3 ThinJilms

271

Table 10. Field effect (hole) mobilities, measured in FETs with Si02 gate for unintentionally doped a-nT. (cm2 v-’ s-’)

Molecule*

pFET

a-3T a,w-DE a-3T a-4T (higher order, PMMA gate) a,w-DE a-4T a-5T

not measurable 1.9 2.2 x lo-’ 2.5 lo-’ 9 x 10-~ 1 10-~ 5 1.5 9 IO-~ 2 10-~ 1-3 x lop2 7.5 x lo-* 1 x 10-2 1.5 x 5 x lo-* not measurable not measurable (< lo-’) 2 IO-~ 1-3 x lop2 1 x lo-* 1 lop5 6 x lop5 10-9-10-*0

(higher order, PMMA gate) CY.,U-DE a-5T a-6T (single crystal) a,w-DM a-6T a.w-DH a-6T P,P’-DH a-6T P,P’-DD a-6T a-8T (higher purity) a,w-DH a-8T polythiophene

Reference

4.3.4.5 Temperature dependence

Temperature-dependent conductivity data of a-6T are also inconsistent so far. Vaterlein et al. [326] find an exponential behavior with temperature T ( a = a. exp(aT)) for undoped and doped a-6T as well as for EC6T even at low temperatures whereas an Arrhenius fit ( a = a. exp(-EA/kT) with EA as activation energy) reveals a large error. This exponential relation does neither fit into the above-mentioned model of hopping conduction nor in the SCLC model with shallow traps. For SCLC from equation 3 and = (N(v)V/gN(v)T) exp[-(ET

-

EV)/kT)I

(4)

with N ( V ) as ~ ,density ~ of states on top of the valence band and of trapping levels, respectively, EV,Tas the energies of the valence band edge and the trapping level, respectively, and g as the degeneracy factor, the energetic position of the trap levels can be derived from Arrhenius plots of temperature dependent measurements. Often a shallow trap at around 0.3 eV above the valence band edge is derived for thin films [338]. For powder samples an activation energy of 0.73 eV is found [337]. The temperature-dependence of the field effect mobility was studied on a-6T by Horowitz et al. [29] and Torsi et al. [352] and on dimethyl-a-6T by Waragai et al. [30]. Horowitz et al. discuss their results above 150K by free carriers which undergo

272

4 Structure and Properties of Oligothiophenes in the Solid State

multiple thermal trapping in and release from shallow traps and below 150K by free carriers hopping among deep traps. Waragai et al. attribute the transport to thermally activated hopping of polarons between thiophene molecules. Torsi et al. could explain the increase of mobility with decreasing temperature below 50K by Holstein's small polaron theory. In a recent paper Wu and Conwell give a refined model of polaron transport which also fits the data by Torsi et al. [353].

4.3.4.6 Conjugation length influence Conductivities of unsubstituted molecules are summarized in Table 11, field effect mobilities can be taken from Table 10. In older papers a strong dependence of the mobility on the chain-length was found [342, 343, 354, 3551 although a nearly independent mobility is expected. From newer data [loo, 333, 334, 3561 thls independence can really be determined. Therefore the older data are masked by structural disorder and/or impurities. The decrease in mobility observed for a-8T and polythiophene may also stem from an increase of conjugation defects. The lack of field effect in a-3T can be explained because this chain is regarded to be too short to bear a radical cation but extrapolation of data obtained in a-3Tlpolycarbonate mixtures with different a-3T concentrations to the pure oligomer yield p = 3.3 x 10-9cm2V-'s-1 [357]. For the longer oligomers the conductivity increases with increasing chain-length. If only the values from one reference are taken, the conductivities reveal a logarithmic decrease with inverse chain-length.

4.3.4.7 Influence of dopants Oeter et al. studied in situ FeCL p-doped a-6T prepared by UHV coevaporation [308]. In line with the observation of new states in the band-gap (compare section

Table 11. Conductivity of unintentionally doped anT: influence of chain-length. Molecules

(S/cm) ~

a-3T

~

lo-" 10-'0

a4T

a-5T a-6T

a-8T

Polythlophene

10-~ 10-8 4 x 10-6 lo-' 10-6 lop7 lop6 10-6-10-8

Reference

4.3 Thin.films

273

4.3.3.2) the conductivity increases with the amount of dopant from non-measurable values (< lop9S cm-') for pristine a-6T up to 0.1 S cm-' for concentrations of FeC13 above the molar ratio 1:1 where it saturates. However, the environmental stablity is very low due to the reduction product FeClz which remains in the sample and leads to a 'dedoping' of the sample [307]. However also [FeClJ itself shows a high photolability which leads to a dedoping at least for poly(3-alkylthiophene) films [358, 3591. On the other hand comparative studies of doping polyalkylthiophenes with different dopants as FeC13, 12, and [PF& showed the high stability of FeC13 doped samples if compared to the others [360]. Also [S03CF3]- as counterion leads to stable doped polythiophenes [3611. de Leeuw studied the influence of iron(III)toluenesulphonate, 2,3-dichloro-5,6dicyano- 1,6benzoquinone, and FeC13 on the conductivity of spin-coated films of side-chain-substituted dodecathiophene [362]. He finds values of 4 x S cm-' for unintentionally doped films and up to 20 S cm-' for dopant concentrations of 4 molecules dopant per oligothiophene molecule where the conductivity reaches a plateau. The stability of the doped films depends on the amount of dopant. Whereas films doped up to the saturation level and unintentionally doped films show a high stability over 120 days, films with intermediate dopant concentration show an exponential decay of the conductivity. The iodine doping process of end-capped oligothiophenes was studied by Stoldt et al. [363, 3641. Iodine increases the conductivity of evaporated EC4T up to values of 1 S cm-' ,but the absolute values obtained were not reproducible and also yielded values of 2.5 x S cm-' for the same preparation conditions. For preparation from solution by evaporation of the solvent much lower conductivities were obtained. Iodine doping also leads to drastic changes of the conductivity with time [364]. Within the first minutes of iodine exposure a logarithmic increase of the conductivity is found, reaching a maximum after about 8 minutes, followed by a slight decrease for the next 100 minutes and then a subsequent increase of the conductivity up to 10 days. The maximum conductivity obtained after 10 days was 6 x lo-' Scm-' for EC7T, 2.5 x lo-' Scm-' for EC6T, and 5 x lo-* Scm-' for EC5T. If the iodine vapour is removed, however, a dedoping process with exponentially decreasing conductivities is observed [364]. The influence of oxygen doping on EC6T was studied by the same group [326] whereas the aging of a-6T in air was studied by Horowitz et al. [365]. In both cases 'doping' by oxygen or air exposure leads to an increase of the conductivity to values as reported in Table 9 or 11. This doping is even increased by illumination of the sample due to a photochemical reaction of oxygen with thiophenes yielding charge carriers [366]. The conductivity decreases after evacuation with a drop of one order of magnitude during the first few hours and a very slow decrease during the following 30 days. This shows that unintentional doping of air-exposed samples leads to 'wrong' conductivity data if they are attributed to intrinsic properties. By Horowitz et al. [338] also the role of oxygen as trap-killer is discussed. Because the SCLC (see section about I/V-curves) is also much lower in vacuum not only a mere dedoping can be discussed because SCLC is independent of the number of bulk charge carriers. On the other hand the conductivity can also be decreased by an increase of the trap level density.

274

4 Structure and Properties of Oligothiophenes in the Solid State

n-doping with alkali metals also influences the conductivity [318]. However, even for saturation doping, i.e. one charge per thiophene monomeric unit, the conductivity of a-6T does not exceed lop6 S cm-' . Apparently the mobility of negative charge carriers is very low. From the spectroscopic results mentioned in section 4.3.3.2 it can be concluded that the negative charge is highly localized at the sulfur atoms. 4.3.4.8 Photoconductivity

The photoconduction action spectra are usually similar to the absorption spectra. Illumination leads to two to three orders of magnitude higher conductivity if compared to dark conduction [13,18 13. In presence of oxygen the conductivity increases another order of magnitude whereas nearly no influence of oxygen on the dark conductivity is found [181, 3671. The photoconductivity raises proportional to the square root of the illumination intensity [181]. If the long axes of the molecules are perpendicular to the applied field a square root dependence of the logarithm of the photocurrent with the applied field is found according to the expected Poole-Frenkel behavior, whereas molecules oriented with their long axes along the field exhibit a linear field dependence [13].

References 1. Handbook of Conducting Polymers (Eds. T. A. Skotheim, R. L. Elsenbaumer and J. R. Reynolds), Marcel Dekker, New York, 1998, 2nd edition. 2. D. Beljonne, J. Cornil, D. A. dos Santos, 2. Shuai and J.-L. Brtdas, in Primary Photoexcitations in Conjugated Polymers: Molecular Exciton versus Semiconductor Band Model (Ed.: N. S. Sariciftci), World Scientific, Singapore, 1998. 3. J. Nakayama, T. Konishi and M. Hoshino, Heterocycles, 1988, 27, 1731. 4. R. Hakansson in Thiophene and its Derivatives (Ed.: S. Gronowitz), John Wiley, 1992, Vol. 44, Part 5, Chapter 111. 5. For a recent review on the chemistry of oligothiophenes, see P. Bauerle in Chapter 111 of this book. 6. F. J. Gommers, Nematologica, 1972, 18, 458. 7. T. Arnason, T. Swain, C.-K. Wat et al., J. Biochem. Syst. Ecol., 1981,9, 63. 8. J. Kagan and G. Chan, Experientia 1983, 39, 402. 9. G. H. N. Towers, T. Arnason, C.-K. Wat, E. A. Graham, J. Lam and J. C. Mitchell, Contact Dermatitis, 1979, 5, 140. 10. G. Campbell, J. D. H. Lambert, T. Arnason and G. H. N. Towers, J. Chem. Ecol. 1982,8,961. 11. J. B. Hudson Antiviral Research, 1989, 12, 55. 12. R. Rossi, A. Carpita, M. Ciofalo and J. L. Houben, Gazz. Chim. Ztal., 1990, 120, 793. 13. U. Schoeler, K. H. Tews and H. Kuhn, J. Chem. Phys., 1974,61, 5009. 14. M. Akimoto, Y. Furukawa, H. Takeuchi, I. Harada, Y. Soma and M. Soma, Synth. Met., 1986, IS, 353. 15. Y. Yumoto and S. Yoshimura, Synth. Met., 1986, 13, 185. 16. S. Tasaka, H. E. Katz, R. S. Hutton, J. Orenstein, G. H. Fredrickson and T. T. Wang, Synth. Met., 1986, 16, 17. 17. D. Fichou, G. Horowitz, Y. Nishikitani and F. Gamier, Chemtronics, 1988, 3, 176.

References

275

G. Horowitz, D. Fichou and F. Garnier, Solid State Commun.,1989, 70, 385. G. Horowitz, D. Fichou, X. Peng, Z. Xu and F. Garnier, Solid Stste Commun., 1989, 72, 381. X. Peng, G. Horowitz, D. Fichou and F. Garnier, Appl. Phys. Lett., 1990,57,2013. A. Tsumura, H. Koezuka and T. Ando, Synth. Met., 1988,25, 11. A. Assadi, S. Svensson, M. Wilader and 0. Inganas, Appl. Phys. Lett., 1988, 53, 195. Sze, F. Garnier, G. Horowitz, D. Fichou and X. Peng, Adv. muter., 1990, 2, 592. R. Zamboni, N. Periasamy, G. Ruani and C. Taliani, Synth. Met., 1993, 54, 57. C. Taliani and L. M.Blinov, Adv. Muter., 1996, 8, 353. D. Oeter, H.-J. Egelhaaf, Ch. Ziegler, D. Oelkrug and W. Gopel, J . Chem. Phys., 1994, 101, 6344. 28. For a review, see W. Gebauer and C. Taliani in Chapter 6 of this book. 29. G. Horowitz, R. Hajlaoui and P. Delannoy, J . Phys. ZZZFrunce, 1995, 5, 355. 30. K. Waragai, H. Akamichi, S. Hotta, H. Kano and H. Sakaki, Phys. Rev. B, 1995,52, 1786. 31. L. Torsi, A. Dodabalapur, L. J. Rothberg, A. W. P. Fung and H. E. Katz, Phys. Rev. B, 1998, 57, 227 1. 32. For a review, see G. Horowitz and P. Delannoy, Chapter 7 of this book. 33. S. Hotta and K. Waragai, Adv. Muter., 1993, 5 , 896. 34. F. Garnier, G. Horowitz, D. Fichou and A. Yassar, Suprumoleculur Science, 1997, 4, 155. 35. G. Horowitz et al., European J . Phys., Appl. Phys., 1998, 1, 361. 36. For a survey of oligothiophene-based OFETs, see A. Dodabalapur, H. E. Katz and Bao in Chapter 8 of this book. 37. F. Geiger, M. Stoldt, H. Schweizer, P. Bauerle and E. Umbach, Adv. Muter., 1993, 5 , 922. 38. G. Horowitz, P. Delannoy, H. Bouchriha et ul., Adv. Muter., 1994, 6, 752. 39. D. Fichou, J.-M. Nunzi, F. Charra and N. Pfeffer, Adv. Muter., 1994, 6, 64. 40. F. Charra, M-P. Lavie, A. Lorin and D. Fichou, Synth. Met., 1994, 65, 13. 41. M. G. Harrison, R. H. Friend, F. Garnier and A. Yassar, Synth. Met., 1994, 67, 215. 42. D. Fichou and F. Charra, Synth. M e f . , 1996, 76, 11. 43. N. Noma, T. Tsuzuki and Y. Shirota, Adv. Muter., 1995, 7, 647. 44. M. Grandstrom, M. G. Harrison and R. H. Friend in Chapter 8 of this book. 45. Y. Delugeard, J. Desuche and J. L. Baudour, Actu Crystullogr. B, 1976, 32, 702. 46. J. L. Baudour, H. Cailleau and W. B. Yelon Rivet, Actu Crsytullogr. B, 1977, 33, 1773. 47. J. L. Baudour, Y. Delugeard and P. Rivet, Actu Crsytullogr. B, 1978, 33, 625. 48. K. N. Baker, A. V. Fratini, T. Resch et ul., Polymer, 1993, 34, 1571. 49. S. Hotta and K. Waragai, J . Muter. Chem., 1991, 1 , 835 50. J.-H. Liao, M. Benz, E. LeGoff and M. G. Kanatzidis, Adv. Muter., 1994, 6, 135. 51. F. van Bolhuis, H. Winberg, E. E. Havinga, E. W. Meijer and E. G. J. Staring, Synth. Met., 1989, 30, 381. 52. A. Gavezzotti and G. Filippini, Synth. Met., 1991, 40, 257. 53. W. Porzio, S. Destri, M. Mascherpa and S . Briickner, Actu Polymer, 1993, 44, 266. 54. P. Ostoja, S. Guerri, S. Rossini, M. Servidori, C. Taliani and R. Zamboni, Synth. Met., 1993, 54, 447. 55. B. Servet, S. Ries, M. Trotel, P. Alnot, G. Horowitz and F. Garnier, Adv. Muter., 1993,5,461. 56. B. Servet, G. Horowitz, S. Ries et al., Chem. Muter., 1994, 6, 1809. 57. Y. Kanemitsu, N. Shimizu, K. Suzuki, Y. Shiraishi and M. Kuroda, Phys. Rev. B, 1996, 54, 2198. 58. G. J. Visser, G. J. Heeres, J. Wolters and A. Vos, Actu Crystullogr., 1968, B24, 467. 59. P. A. Chaloner, S. R. Gunatunga and P. B. Hitchcock, Actu Cryst., 1994, C50, 1941. 60. M. Pelletier and F. Brisse, Actu Cryst., 1994, C50, 1942. 61. L. Antolini, G. Horowitz, F. Kouki and F. Garnier, Adv. Muter., 1998, 10, 382. 62. T. Siegrist, Ch. Kloc, R. A. Laudise, H. E. Katz and R. C. Haddon, Adv. Muter., 1998, in press. 63. W. Porzio, S. Destri, M. Mascherpa, S. Rossini and S . Bruckner, Synth. Met., 1993, 55, 408. 64. G. Horowitz, B. Bachet, A. Yassar et ul., Chem. Muter., 1995, 7, 1337. 65. D. Fichou, B. Bachet, F. Demanze, I. Billy, G. Horowitz and F. Garnier, Adv. Muter., 1996, 6, 500. 66. T. Siegrist, R. M. Fleming, R. C. Haddon et ul., J. Muter. R e x , 1995, 10, 2170.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

276

4 Structure and Properties of Oligothiophenes in the Solid State

67. Z. Mo, K.-B. Lee, Y. B. Moon, M. Kobayashi, A. J. Heeger and F. Wudl, Macromol., 1985, 18, 1972. 68. C.’X. Cui and M. Kertesz, Phys. Rev. B, 1989, 40,9661. 69. 0. Bastiansen, Acta Chem. Scand., 1949, 3, 408; ibid., 1950, 4, 926; ibid., 1952, 6, 205; ibid, 1954,8, 1593. 70. A. Almenningen, 0.Bastiansen and P. Svendsas, Acta Chem. Scand., 1958,12,1671 and references therein. 71. B. Bak, C. Christensen, J. Rastrup-Andersen and E. Tannenebaum, J. Chem. Phys., 1956, 25. 892. 72. G.’Barbarella, M. Zambianchi, L. Antolini et al., J. Chem. SOC,Perkins Trans., 1995,2, 1869. 73. S. V. Meille, A. Farina, F. Bezzicheri and M. C. Gallazzi, Adv. Muter., 1994, 6, 848. 74. E. F. Paulus, R. Dammel, G. Kampf and P. Wegener, Acta Cryst., 1988, B44, 509. 75. E. F. Paulus, K. Sam, K. Wolinski and L. Schafer, J. Mol. Struct., 1989, 196, 171. 76. C. Aleman, E. Brillas, A. G. Davies et al., J. Org. Chem., 1993, 58, 3091. 77. G. Engelmann, G. Kossmehl, J. Heinze, P. Tschunky, W. Jugelt and H.-P. Welzel, J. Chem. SOC.,Perkin Trans. 2, 1998, 169. 78. F. Effenberger and F. Wiirthner, Angew. Chem. Znt. Ed. Engl., 1993,5, 719. 79. Z. Hu, J. L. Atwood and M . P. Cava, J. Org. Chem., 1994, 59, 8071. 80. L. DeWitt, G. J. Blanchard, E. LeGoff, M. E. Benz, J. H. Liao and M. G. Kanatzidis, J. Am. Chem. SOC.,1993, 115, 12158. 81. G. Barbarella, M. Zambianchi, A. Bongini and L. Antolini, Adv. Muter., 1994, 6, 561. 82. D. D. Graf, J. P. Campbell, L. L. Miller and K. R. Mann, J. Am. Chem. SOC.,1996,118,5480. 83. D. D. Graf, R. G. Duan, J. P. Campbell, L. L. Miller and K. R. Mann, J. Am. Chem. SOC., 1997, 119, 5888. 84. D. R. Ferro, W. Porzio, S. Destri, M. Ragazzi and S. Briickner, Macromol. Theory Simul., 1997, 6, 713. 85. Y. Matsuura, Y. Oshima, Y. Misaki et al., Synth. Met., 1996, 82, 155. 86. G. Barbarella, M. Zambianchi, A. Bongini and L. Antonili, Adv. Muter., 1992, 4, 282. 87. G. Barbarella, A. Bongini and M. Zambianchi, Adv. Muter., 1991,3,494. 88. G. Barbarella, M. Zambianchi, R. Di Toro, M. Colonna, L. Antolini and A. Bongini, Adv. Muter., 1996, 8, 327. 89. G. Barbarella, M. Zambianchi, M. del Fresno I Marimon, L. Antolini and A. Bongini Adv. Mater., 1997, 9, 484. 90. J. Bernstein and A. T. Hagler, J. Am. Chern. Sac., 1978, 100, 673. 91. J. D. Dunitz and J. Bernstein, Ace. Chem. Res., 1995, 28, 193. 92. C. Taliani, R. Zamboni, G. Ruani, S. Rossini and R. Lazzaroni, J. Mol. Electron., 1990, 6, 225. 93. S. Destri, M. Mascherpa and W. Porzio, Adv. Muter., 1993, 535, 43. 94. F. R. Lipsett, Can. J. Phys., 1957,5, 284. 95. R. A. Laudise, P.M. Bridenbaugh, T. Siegrist, R. M. Fleming, H. E. Katz and A. J. Lovinger, J. Crystal Growth, 1995, 152, 241. 96. J. K. Herrema, J. Wilderman, F. Van Bolhuis, G. Hadziioannou, Synth. Met., 1993, 60, 239. 97. A. Yassar, F. Gamier, F. Deloffre, G.Horowitz and L. Ricard, Adv. Muter., 1994, 6, 660. 98. D. Fichou, M.-P. Teulade-Fichou, G. Horowitz and F. Demanze, Adv. Muter., 1997, 9, 75. 99. R. Hajlaoui, D. Fichou, G . Horowitz, B. Nessakh, M. Constant and F. Garnier, Adv. Muter., 1997,9,557. 100. a. D. Fichou, M.-P. Teulade-Fichou, F. Demanze, G. Horowitz and F. Garnier, 2nd JapanFrance Joint Forum (JFJF’2) on Organic Materials and Devices for Optoelectronics, Paris, November 1995; b. V. Videlot, D. Fichou and F. Garnier, J. Chim. Phys., 1998, 95, 1335. 101. D. Fichou, F. Garnier, F. Charra, F. Kajzar and J. Messier, Organic Materials for Nonlinear Optics, Eds. R. Hahn and D. Bloor, Royal SOC.Chem., London, 1989, p. 176. 102. H. Knobloch, D. Fichou, W. Knoll and H. Sasabe, Adv. Muter., 1993, 5, 570. 103. G. Horowitz, R. Hajlaoui, D. Fichou and A. El Kassmi, J. Appl. Phys., submitted. 104. F. Gamier, G. Tourillon, J.Y. Barraud and H. Dexpert, J. Muter. Sci., 1985, 20, 2687. 105. R. Yang, K. M. Dalsin, D. F. Evans, L. Christensen and W. A. Hendrickson, J. Phys. Chem., 1989, 65, 23; ibid., 1989, 93, 51 1.

References

277

106. J.-P. Aime, F. Bargain, M. Schott, H. Eckhardt, G. G. Miller and R. L. Elsenbaumer, Phys. Rev. Lett., 1989, 62, 55. 107. 0. Inganas, W. R.Salaneck, J. E. Osterholm, J . Laasko, Synth. Met., 1988, 22, 395. 108. W. RSalaneck, 0. Inganas, B. Theman et al., J. Chem. Phys., 1988,89,4613. 109. S. Bruckner and W. Porzio, Makromol. Chem., 1988, 189, 961. 110. M. J. Winokur, D. Spiegel, Y. Kim, S. Hotta and A. J. Heeger, Synth. Met., 1989,28, C419. 111. a. T. Yamamoto, T. Kanbara and C. Mori Synth. Met., 1990, 38, 399. b. G. Horowitz, F. Garnier, A. Yassar, R. Hajlaoui and F. Kouki, Adv. Muter., 1996,8, 52; c. W. A. Schoonveld, R. W. Stok, J. W. Weijtmans, J. Vrijmoeth, J. Wildeman and T. M. Klapwijk, Synth. Met., 1997, 84, 583. 112. G. Horowitz, S. Romdhane, H. Bouchriha et al., Synth. Met., 1997, 90, 187. 113. G. Klein, S. Petit, C. Hirlimann and A. Boeglin, to be published. 114. D. Fichou, S. Delysse and J.-M. Nunzi, Adv. Muter., 1997, 9, 1178. 115. N. Karl, Phys. Status Solidi, 1972, 23, 651. 116. N. Karl, J. Lumin., 1976, 12/13, 851. 117. G. Horowitz, P. Valat, F. Garnier, F. Kouki, V. Wintgens, Opt. Muter., 1998, 9, 46. 118. D. Fichou, V. Dumarcher, S. Delysse and J.-M. Nunzi, Proc. SPZE Conf., San Jose, USA, 1998, 3281, 202. 119. N. Tessler, G. J. Denton and R. H. Friend, Nature, 1996, 382, 695. 120. F. Hide, M. A. Diaz-Garcia, B. J. Schwartz, M. R. Anderson, Q. Pei and A. J. Heeger, Science, 1996,273, 1833. 121. D. Moses, J. Wang, D. Fichou and C. Videlot, International Conference on Synthetic Metals (ZCSM’98), Montpellier, France, 1998. 122. H. Nakahara, J. Nakayama, M. Hoshino and K. Fukuda, Thin Solid Films, 1988, 160, 87. 123. W. Porzio, A. Bolognesi, S. Destri, M. Catellani and G. Bajo, Synth. Met., 1991,41-43, 537. 124. J. Paloheimo, P. Kuivalainen, H. Stubb, E. Vuorimaa and P. Yli-Lahti, Appl. Phys. Lett., 1990, 56, 1157. 125. G. Gauglitz, V. Hoffmann, S. Kohlhage and M. Kramer, J. Mol. Struct., 1995, 349, 65. 126. N. Serdar Sariciftci, U. Lemmer, D. Vacar, A. J. Heeger and R. A. J. Janssen, Synth. Met., 1997, 84, 609. 127. J. V. Caspar, V. Ramamurthy and D. R. Corbin, J . Am. Chem. SOC.,1991, 113, 600. 128. P. Enzel and T. Bein, Synth. Met., 1993, 55-57, 1238. 129. A. Stabel and J. P. Rabe, Synth. Met., 1994, 67, 47. 130. A. Stabel, R. Heinz, F. C. De Schryver and J. P. Rabe, J. Phys. Chem., 1995,99, 505. 131. P. Bauerle, T. Fischer, B. Bidlingmeier, A. Stabel and J. P. Rabe, Adv. Mat., 1995, 107, 335. 132. F. Garnier, A. Yassar, R. Hajlaoui et a/., J . Am. Chem. SOC.,1993, 115, 8716. 133. C. Seidel, Ph.D. thesis, University of Stuttgart (FRG) 1993. 134. A. Soukopp, K. Glockler, P. Bauerle, M. Sokolowski and E. Umbach, Adv. Mat., 1996,8,902. 135. H.-J. Egelhaaf, P. Bauerle, K. Rauer, V. Hoffmann and D. Oelkrug, J. Mol. Struct., 1993, 293, 249. 136. H.-J. Egelhaaf, P. Bauerle, K. Rauer, V. Hoffmann and D. Oelkrug, Synth. Met., 1993, 61, 143. 137. J. Stohr, J. L. Gland, E. B. Kollin et al., Phys. Rev. Lett., 1984, 53,2161. 138. A. P. Hitchcock, J. A. Horsley and J. Stohr, J . Chem. Phys., 1986,85,4835. 139. B. A. Sexton, Surf. Sci., 1985, 163, 99. 140. M. Schmelzer, S. Roth, P. Bauerle and R. Li, Thin Solid Films, 1993, 229, 255. 141. M. Schmelzer, M. Burghard, P. Bauerle and S. Roth, Synth. Met., 1993, 61, 97. 142. M. Schmelzer, M. Burghard, P. Bauerle and S. Roth, Thin Solid Films, 1994, 243, 620. 143. M. Schmelzer, M. Burghard, S. Roth and P. Bauerle, Mol. Cryst. Liq. Cryst., 1994, 253, 173. 144. A. Soukopp, diploma thesis, University of Stuttgart (FRG) 1993. 145. A. Soukopp, C. Seidel, R. Li, M. Bassler, M. Sokolowski and E. Umbach, Thin Solid Films, 1996,284-285, 343. 146. R. Li, P. Bauerle and E. Umbach, Surf. Sci., 1995,331-333, 100. 147. K. M. Baumgartner, M. Volmer-Uebing, J. Taborski, P. Bauerle and E. Umbach, Ber. Bunsenges. Phys. Chem., 1991, 95, 1488. 148. W. Gebauer, M. BaBler, A. Soukopp et al., Synth. Met., 1996,83,227.

278

4 Structure and Properties of Oligothiophenes in the Solid State

149. W. Gebauer, M. BaBler, R. Fink, M. Sokolowski and E. Umbach, Chem. Phys. Lett., 1997, 266, 177. 150. C. Seidel, A. Soukopp, R. Li, P. Bauerle and E. Umbach, Surf. Sci., 1997, 374, 17. 151. U. K. Sarkar, S. Chakrabarti, A. J. Pal and T. N. Misra, Spectrochim. Acta, 1992, 48A, 1625. 152. U. K. Sarkar, A. J. Pal, S. Chakrabarti and T. N. Misra, Chem. Phys. Lett., 1992, 190, 59. 153. U. K. Sarkar, S. Chakrabarti and T. N. Misra, Chem. Phys. Lett., 1992,200,55. 154. A. J. Lovinger, D. D. Davis, A. Dodabalapur and H. E. Katz, Chem. Muter., 1996,8,2836. 155. D. Oelkrug, J. Haiber, R. Lege, H. Stauch and H.-J. Egelhaaf, Thin Solid Films, 1996, 284-285, 58 1. 156. V. Hoffmann, Tubingen, personal communication 1995. 157. G. Louarn, J. P. Buisson, S . Lefrant and D. Fichou, J. Phys. Chem., 1995, 99, 11399. 158. W. A. Schoonveld, R. W. Stok, J.W. Weijtmans, J. Vrijmoeth, J. Wildeman and T. M. Klapwijk, Synth. Met., 1997, 84, 583. 159. D. Oelkrug, H.-J. Egelhaaf, J. Gierschner and A. Tompert, Synth. Met., 1996, 76, 249. 160. J. Fink, N. Niicker, S. Scheerer and H. Neugebauer, Synth. Met., 1987, 18, 163. 161. K. Uchiyama, H. Akimichi, S. Hotta, H. Noge and H. Sakaki, Synth. Met., 1994, 63, 57. 162. H. Mugumura, T. Sato, A. Hiratsuka and I. Karube, Langmuir, 1996, 12, 5451. 163. C. Y. Yang, Y. Yang, S. Hotta, Synth. Met., 1995, 69, 303. 164. Y. Q. Liu, Y. Xu and D. B. Zhu, Synth. Met., 1997,84, 197. 165. H. E. Katz, M. L. Schilling, C. E. D. Chidsey, T. M. Putvinski and R. S. Hutton, Chem. Muter., 1991, 3, 699. 166. H. Byrd, S. Whipps, J. K. Pike, J. Ma, S. E. Nagler and D. R. Talham, J. Am. Chem. SOC., 1994, 116, 295. 167. H. Byrd, S. Whipps, J. K. Pike and D. R, Talham, Thin Solid Films, 1994, 244,768. 168. B. Liedberg, Z. Yang, I. Engquist et al., 1.Phys. Chem. B., 1997, 101, 5951. 169. R. Michalitsch, P. Lang, A. Yassar, G. Nauer and F. Garnier, Adv. Mat., 1997, 9, 321. 170. K. Waragai and S . Hotta, Synth. Met., 1991, 41-43, 519. 171. H.-J. Egelhaaf and D. Oelkrug, SPZE, 1995,2362, 398. 172. J. Gierschner, H.-J. Egelhaaf and D. Oelkrug, Synth. Met., 1997, 84, 529. 173. E. Vuorimaa, P. Yli-Lahti, M. Ikonen and H. Lemmetyinen, Thin Solid FiZms, 1990,190,175. 174. M. Kramer, 0. Bohme, V. Hoffmann and C. Ziegler, to be published 175. 0. Bohme, C. Ziegler and W. Gopel, Adv. Mat., 1994, 6, 587. 176. 0. Bohme, Ph. D. thesis, University of Tiibingen 1996. 177. M. Kramer, 0. Bohme, C. Ziegler and V. Hoffmann, in prep. 178. 0. Bohme, C . Ziegler and W. Gopel, Synth. Met., 1994,67, 87. 179. E. Muller, C. Ziegler, unpublished results 180. E. Hadicke, Ludwigshafen, personal communication 1995. 181. H.-J. Egelhaaf, Ph.D. thesis, University of Tiibingen (FRG) 1995. 182. E. Miiller, diploma thesis, University of Tiibingen (FRG) 1995. 183. G. Marowsky, R. Steinhoff, L. F. Chi, J. Hutter and G. Wagnikre, Phys. Rev. B., 1988,38,6274. 184. K. Hamano, T. Kurata, S. Kubota and H. Koezuka, Jpn. J. Appl. Phys., 1994, 33, L1031. 185. H. Knobloch, W. Knoll, D. Fichou and H. Sasabe, Mol. Cryst. Liq. Cryst., 1994,252, 269. 186. F. Biscarini, R. Zarnboni, P. Samori, P. Ostoja and C. Taliani, Phys. Rev. B., 1995,52, 14868. 187. F. Biscarini, P. Samori, A. Lauria et al., Thin Solid Films, 1996,284-285,439. 188. F. Biscarini, P. Samori, 0. Greco and R. Zamboni, Phys. Rev. Lett., 1997, 78, 2389. 189. P. Viville, R. Lazzaroni, J. L. Brkdas, P. Moretti, P. Samori and F. Biscarini, Adv. Mat., 1998, 10, 57. 190. D. Oelkrug, H.-J. Egelhaaf and J. Haiber, Thin Solid Films,1996, 284-285, 267. 191. P. Lang, R. Hajlaoui, F. Garnier et al., J. Phys. Chem., 1995, 99, 5492. 192. P. Lang, R. Hajlaoui, J. P. Dallas, F. Garnier, A. Yassar and G. Horowitz, J . Chim. Phys., 1995, 92, 967. 193. P. Lang, M. El Ardhaoui, J . C. Wittmann et al., Synth. Met., 1997, 84, 605. 194. P. Lang, P. Valat, G. Horowitz et al., J . Chim. Phys., 1995, 92, 963. 195. P. Lang, G. Horowitz, P. Valat, F. Garnier, J. C. Wittmann and B. Lotz, J . Phys. Chem. B., 1997, 101, 8204.

References

279

196. T. Okajima, S. Narioka, S. Tanimura et al., J . Electron Spectrosc. Relat. Phenom., 1996, 78, 379. 197. M. Buongiorno Nardelli, D. Cvetko, V. De Renzi et al., Synth. Met., 1996, 76, 173. 198. M. Buongiorno Nardelli, D. Cvetko, V. De Renzi, et al., Phys. Rev. B., 1996, 53, 1095. 199. A. J. Lovinger, D. D. Davis, A. Dodabalapur, H. E. Katz and L. Torsi, Macromolecules, 1996, 29, 4952. 200. F. Garnier, A. Yassar, R. Hajlaoui, G. Horowitz and F. Deloffre, Electrochim. Acta, 1994, 39, 1339. 201. S. Sagisaka, M. Ando, T. Iyoda and T. Shimidzu, Thin Solid Films, 1993, 230, 65. 202. G. Tourillon and F. Garnier, J. Electroanal. Chem., 1982, 135, 173. I Electroanal. . Chem., 1984, 161, 51. 203. G. Tourillon and F. Garnier, . 204. R. J. Waltman, J. Bargon and A. F. Diaz, J . Phys. Chern., 1983,87, 1459. 205. R. K. Sadhir and K. F. Schoch, Thin Solid Films,1993, 223, 154. 206. T. Yamamoto, A. Morita, Y. Miyazala et al., Macromolecules, 1992,25, 1214. 207. T. Kanbara, C. Mori, H. Wakayama et al., Solid. State Comm., 1992, 82, 771. 208. T. Kurata, H. Fuchigami, H. Koezuka, T. Yamamoto and T. Fukuda, Jpn. J. Appl. Phys., 1992,31, 3869. 209. M. J. Winokur, P. Wamsley, J. Moulton, P. Smith and A. J. Heeger, Macromolecules, 1991, 24, 3812. 210. J. Mardalen, E. J. Samuelsen and A. 0. Pedersen, Synth. Met., 1993, 55-57, 378. 211. C. G. dos Santos, C. P. de Melo and R. Souto Maior, Synth. Met., 1995, 71,2083. 212. G. Bajo, A. Bolognesi, S. Destri, Z . Geng and W. Porzio, Mol. Cryst. Liq. Cryst., 1993, 229, 91. 213. T. J. Prosa, M. J. Winokur, J. Moulton, P. Smith and A. J. Heeger, Synth. Met., 1993, 55-57, 370. 214. J. T. Lopez Navarrete and G. Zerbi, Synth. Met., 1989, 28, C15. 215. G. Morea, L. Sabbatini, R. H. West and J. C Vickerman, Surf. Interface Analysis, 1992, 18, 421. 216. G. Tourillon in Handbook of Conducting Polymers (Ed. T. A Skotheim), Dekker, New York, 1986. 217. W. Luzny, S. Niziol, G. Straczynski and A. Pron, Synth. Met., 1993, 62, 273. 218. S. J. Kamrava, M. Zagorska, B. Krische and S. Soderholm, Phys. Scripta , 1991,44, 112. 219. U. Gelius, C. J. Allan, G. Johansson, H. Siegbahn, D. A. Allison and R. Siegbahn, Phys. Scripta, 1971, 3, 237. 220. C. R. Wu, J. 0. Nilsson, 0. Inganas, W. R. Salaneck, J.-E. Osterholm and J. L. BrCdas, Synth. Met., 1987, 21, 197. 221. D. Jones, M. Guerra, L. Favaretto, A. Modelli, M. Fabrizio and G. Distefano, J. Phys. Chem., 1990,94, 5761. 222. H. Fujimoto, U. Nagashima, H. Inokuchi et al., J . Chem. Phys., 1990,92,4077. 223. H. Fujimoto, U. Nagashima, H. Inokuchi et al., Phys. Scripta, 1990,41, 105. 224. D. Oeter, Ph.D thesis, University of Tubingen, 1994. 225. J. L. BrCdas, R. Silbey, D. S. Boudreaux and R. R. Chance, J . Am. Chem. SOC.,1983, 105, 6555. 226. J. L. Bredas, R. L. Elsenbaumer, R. R. Chance and R. Silbey, J . Chem. Phys., 1983, 78,5656. 221. B. Themans, J. M. Andre and J. L. Bredas, Synth. Met., 1987, 21, 149. 228. B. ThCmans, W. R. Salaneck and J. L. BrCdas, Synth. Met., 1989,28, C359. 229. J. L. BrCdas, M. Dory, B. ThCmans, J. Delhalle and J. M. And& Synth. Met., 1989,28, D533. 230. H. 0. Villar, P. Otto, M. Dupuis and J. Ladik, Synth. Met., 1993, 59, 97. 231. H. Fujimoto, U. Nagashima, H. Inokuchi et al., J . Chem. Phys., 1988,89, 1198. 232. U. Nagashima, H. Fujimoto, H. Inokuchi and K. Seki, J . Mol. Struct., 1989, 197, 265. 233. D. Oeter, C. Ziegler and W. Gopel, Fres. Z . Analyt. Chem., 1995, 351. 234. H.-J. Egelhaaf, D. Oelkrug, D. Oeter, C. Ziegler and W. Gopel, J . Mol. Struct., 1995, 348, 405. 235. R. Lazzaroni, A. J. Pal, S . Rossini, G. Ruani, R. Zamboni and C. Taliani, Synth. Met., 1991, 41-43,2359. 236. R. A. J Janssen, L. Smilowitz, N. S. Sariciftci and D. Moses, J. Chem. Phys., 1994, 101, 1787.

280

4 Structure and Properties of Oligothiophenes in the Solid State

237. C. Taliani, R. Danieli, R. Lazzaroni, N. Periasamy, G. Ruani and R. Zamboni, Synth. Met., 1993,55-57,4714. 238. D. Fichou, G. Horowitz, B. Xu and F. Garnier, Synth. Met., 1992,48, 167. 239. J.-M. Nunzi, F. Charra and N. Pfeffer, J. Phys. III France, 1993, 3, 1401. 240. Z. G. Soos and D. S. Galvao, J. Phys. Chem., 1994,98, 1029. 241. R. Colditz, D. Grebner, M. Helbig and S. Rentsch, Chem. Phys., submitted 242. F. Negri and M. Z. Zgierski, J. Chem. Phys., 1994, 100, 2571. 243. R. S. Becker, J. Seixas de Melo, A. L Macanita and F. Elisei, Pure Appl. Chem., 1995,67, 9. 244. J. Karlsson, “examensarbete” 931020, LiTH-IFM-EX-574, Linkoping (S) 1993. 245. D. Beljonne, Z. Shuai and J.-L. Bridas, J. Chem. Phys., 1993, 98, 8819. 246. E. Pellegrin, H. Fritzsche, N. Nucker et al., Synth. Met., 1991, 41-43, 1207. 247. N. Periasamy, C. Taliani, G. Ruani and R. Zamboni, Synth. Met., 1993, 55-57, 4991. 248. T. Jiirimae, M. Strandberg, A. M. Karelson and J.-L. Calais, Znt. J. Quunt. Chem., 1995, 54, 369. 249. D. Oelkrug, H.-J. Egelhaaf, D. R. Worrall and F. Wilkinson, J. Fluoresc., 1995, 5, 165. 250. C. Ziegler, in Handbook of Organic Conductive Molecules and Polymers (Ed.: H. Nalwa), Wiley 1997, Vol. 3, pp. 677. 251. P. F. van Hutten, R. E. Gill, J. K. Herrema and G. Hadziioannou, J. Phys. Chem., 1995,99, 3218. 252. S . Hotta and K. Waragai, J. Phys. Chem., 1993, 97, 7427. 253. A. Yassar, D. Delabouglise, M. Hmyene, B. Nessak, G. Horowitz and F. Garnier, Adv. Mat., 1992, 4, 490. 254. A. Yassar, A. Bennouna, M. Khaidar et al., J. Appl. Phys., 1992, 72, 4873. 255. M.-T. Zhao, B. P. Singh and P. N. Prasad, J. Chem. Phys., 1988, 89, 5535. 256. S. L. Bondarev, I. I. Ivanov, Y. N. Romashin, 0. G. Kulinkovich, J. Appl. Spectrosc., 1992, 56, 440. 257. 0. Dippel, V. Brandl, H. Bassler, R. Danieli, R. Zamboni and C. Taliani, Chem. Phys. Lett., 1993,216, 418. 258. F. Deloffre, F. Garnier, P. Srivastava, A. Yassar and J.-L. Fave, Synth. Met., 1994, 67, 223. 259. L. M. Blinov, S. P. Palto, G. Ruani et al., Chem. Phys. Lett., 1995,232,401. 260. W. Gebauer, C. Vaterlein, A. Soukopp. M. Sokolowski and E. Umbach, Thin Solid Films, 1996,284-285,576. 261. A. Yassar, P. Valat, V. Wintgens et al., Synth. Met., 1994, 67, 277. 262. A. Yassar, G. Horowitz, P. Valat et al., J. Phys. Chem., 1995,99, 9155. 263. C. Taliani, R. Danieli, R. Lazzaroni, N. Periasamy, G. Ruam and R. Zamboni, Mol. Cryst. Liq. Cryst., 1992, 217, 101. 264. N. Periasamy, R. Danieli, G. Ruani, R. Zamboni and C. Taliani, Phys. Rev. Lett., 1992,68,919. 265. G. Lanzani, M. Nisoli, S. De Silvestri and F. Abbate, Chem. Phys. Lett., 1997, 264, 667. 266. G. Klein, C. Jundt, B. Sipp et al., Chem. Phys., 1997, 215, 131. 267. X. Cheng, W. G. Herkstroeter, J. Perlstein, K.-Y. Law and D. G. Whitten, J. Phys. Chem., 1991, 98, 5138. 268. X. Cheng, K. Ichimura, D. Fichou and T. Kobayashi, Chem. Phys. Lett., 1991, 185,286. 269. G. Lanzani, R. Danieli, M. Muccini and C. Taliani, Phys. Rev. B., 1993, 48, 15326. 270. G. Lanzani, C. Taliani, L. Rossi and A. Piaggio, Mol. Cryst. Liq. Crys., 1994. 271. J. Poplawski, E. Ehrenfreund, J. Cornil et al., Mol. Cryst. Liq. Cryst., 1994, 256, 407. 272. J. Poplawski, E. Ehrenfreund, J. Cornil et al., Synth. Met., 1995, 69,401. 273. K. Watanabe, T. Asahi, H. Fukumura, H. Masuhara, K. Hamano and T. Kurata, J . Phys. Chem. B., 1997, 101, 1510. 274. X. Wei, P. A. Lane, M. Liess et al., Synth. Met., 1997, 84, 565. 275. W. Gebauer, C. Vaterlein, A. Soukopp et al., Synth. Met., 1997, 87, 127. 276. G. R. J. Williams, J. Mol. Electron., 1990, 6, 99. 277. Y. Verbandt, H. Thienpont, I. Veretennicoff and G. L. J. A. Rikken, Phys. Rev. B., 1993, 48, 8651. 278. B. F. Levine and C. G. Bethea, J. Chem. Phys., 1975,63,2666. 279. D. Grebner, H. Chosrovian, S. Rentsch and H. Naarmann, Inst. Phys. Con$ Ser., No. 126, Section IV, 1991, p. 509.

References

28 1

S. Irle and H. Lischka, J. Chem. Phys., 1997, 107, 3021. G. Horowitz, A. Yassar and H. J. von Bardeleben, Synth. Met., 1994, 62, 245. Ch. Ehrendorfer and A. Karpfen, J . Phys. Chem., 1994, 98, 7492. J. L. Bredas, R. R. Chance and R. Silbey, Phys. Rev. B., 1982,26, 5843. S. Roth and H. Bleier, Adv. Phys., 1987, 36, 385. A. J. Heeger, S. Kivelson, J. R. Schrieffer and W.-P. Su, Rev. Mod. Phys., 1988,60, 781. S. Roth, Mat. Sci. Forum, 1989, 42, 1. W. R. Salaneck and J. L. Bredas, Synth. Met., 1994, 67, 15. S. Stafstrom and J. L. Bredas, Mol. Cryst. Liq. Cryst., 1988, 160, 405. S. Stafstrom and J. L. Bredas, J . Mol. Struct., 1989, 188, 393. M. Springborg, J. Phys. Condens. Matter, 1992, 4, 101. M. Springborg, Synth. Met., 1997, 85, 1037. D. Bertho and C. Jouanin, Synth. Met., 1988, 24, 179. P. Bauerle, U. Segelbacher, K.-U. Gaudl, D. Huttenlocher and M. Mehring, Angew. Chem., 1993, 105, 125 294. M. G. Hill, J.-F. Penneau, B. Zinger, K. R. Mann and L. L. Miller, Chem. Muter., 1992, 4, 1106. 295. P. Hapiot, P. Audebert, K. Monnier, J.-M. Pernaut and P. Garcia, Chern. Muter., 1994, 6, 1549. 296. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Chem. Muter., 1993,5, 620. 297. P. Audebert, P. Garcia, P. Hapiot, K . Monnier and J.-M. Pernaut, J . Chim. Phys., 1995, 92, 827. 298. A. Alberti, L. Favaretto, G. Seconi and G. F. Pedulli, J . Chem. SOC.Perkin Trans., 1990, II, 931. 299. Z. Xu, D. Fichou, G. Horowitz and F. Garnier, J . Electroanal. Chem., 1980,267, 339. 300. D. Fichou, G. Horowitz, B. Xu and F. Garnier, Synth. Met., 1990, 39, 243. 301. D. Fichou, B. Xu, G. Horowitz and F. Garnier, Synth. Met., 1991,4143, 463. 302. F. J. Dyson, Phys. Rev, 1955,98, 349 303. K. Tanaka, Y. Matsuura, Y. Oshima, T. Yamabe and S . Hotta, Synth. Met., 1994,66, 295. 304. D. N. R. Rao and M. C. R. Symons, J . Chem. Soc. Perkin Trans., 1983, IZ, 135. 305. M. Shiotani, Y. Nagata, M. Tasaki, J. Sohma and T. Shida, J . Phys. Chem., 1983, 87, 1170. 306. D. Oeter, C. Ziegler, W. Gopel and H. Naarmann, Ber. Bunsenges. Phys. Chem., 1994, 97, 448. 307. D. Oeter, C. Ziegler and W. Gopel, Synth. Met., 1993, 61, 147. 308. D. Oeter, C. Ziegler, W. Gopel and H. Naarmann, Synlh. Met., 1994, 67, 267. 309. J. Cornil, D. Beljonne and J. L. Bredas, J . Chem. Phys., 1995, 103, 842. 310. M. G. Harrison, R. H. Friend, F. Garnier and A. Yassar, Mol. Cryst. Liq. Cryst., 1994,252, 165. 31 1. M. G . Ramsey, D. Steinmiiller and F. P. Netzer, Synth. Met., 1993,54, 209. 312. D. Fichou, G. Horowitz, Y. Nishikitani, J. Roncali and F. Garnier, Synth. Met., 1989, 28, C729. 313. S. Irle and H. Lischka, J. Chem. Phys., 1995, 103, 1508. 314. M. G. Ramsey, F. P. Netzer, D. Steinmiiller, D. Steinmuller-Nethl and D. R. Lloyd, J . Chem. Phys., 1993, 97, 4489. 315. D. Steinmiiller, M. G. Ramsey and F. P. Netzer, Phys. Rev. B., 1992, 47, 13 323. 316. M. Logdlund, P. Dannetun, C. Fredriksson, W. R. Salaneck and J. L. Bredas, Phys. Rev. B., 1996,53, 16327. 317. J. L. Bredas, B. Themans, J. G. Fripiat, J. M. Andre and R. R. Chance, Phys. Rev. B., 1984, 29, 6761. 318. J. Murr and C. Ziegler, Phys. Rev. B, 1998, 57, 7299. 319. J. Murr and C. Ziegler, to be published 320. G. Lanzani, L. Rossi, A. Piaggi, A. J. Pal and C. Taliani, Chem. Phys. Lett., 1994, 226, 547. 321. R. Zamboni, G. Ruani, C. Taliani and A. J. Pal, Mol. Cryst. Liq. Cryst. 322. P. A. Lane, X. Wei, Z. V. Vardeny et al., Chem. Phys., 1996,210, 229. 323. P. A. Lane, X. Wei and Z. V. Vardeny, Synth. Met., 1997,84, 521. 324. D. Fichou, G. Horowitz and F. Gamier, Synth. Met., 1990, 39, 125.

280. 281. 282. 283. 284. 285. 286. 287. 288. 289. 290. 291. 292. 293.

282

4 Structure and Properties of Oligothiophenes in the Solid State

N. Karl, Mol. Cryst. Liq. Cryst., 1989, 171, 31. C. Vaterlein, B. Ziegler, W. Gebauer et al., Synth. Met., 1996, 76, 133. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Adv. Mat., 1993,5, 551. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Synth. Met., 1993, 61, 81. J. Paloheimo, H. Stubb and L. Gronberg, Synth. Met., 1993, 57, 4198. C. Vaterlein, H. Neureiter, W. Gebauer et al., J . Appl. Phys., 1997,82, 3003. E. J. Lous, P. W. M. Blom, L. W. Molenkamp and D. M. de Leeuw, Phys. Rev. B., 1995, 51, 17251. 332. E. J. Lous, P. W. M. Blom, L. W. Molenkamp and D. M. de Leeuw, J . Appl. Phys., 1997, 81, 3537. 333. F. Garnier, F. Kouki, R. Hajlaoui and G. Horowitz, Mat. Res. Bull., 1997, 22, 52. 334. R. Hajlaoui, G. Horowitz, F. Garnier et al., Adv. Mat., 1997, 9, 389. 335. G. Horowitz, Adv. Mat., 1996,8, 177. 336. M. Pope and Ch. E. Swenberg, Electronic Processes in Organic Solids (Monographs on the Physics and Chemistry of Materials 39), Clarendon, Oxford, 1982 337. S. Sen, P. Pal, S . Rossini, T. N. Misra, J . Phys. Chem. Solids, 1994, 55, 17. 338. G. Horowitz, D. Fichou, X. Peng and P. Delannoy, J. Phys. France, 1990,51, 1489. 339. D. M. de Leeuw and E. J. Lous, Synth. Met., 1994,65,45. 340. D. Fichou, G. Horowitz, Y. Nishikitani and F. Garnier, Synth. Met., 1989, 28, C723. 341. F. Gamier, G. Horowitz and D. Fichou, Synth. Met., 1989,28, C705. 342. F. Garnier, F. Deloffre, G. Horowitz and R. Hajlaoui, Synth. Met., 1993, 55-57,4747. 343. F. Gamier, F. Deloffre, A. Yassar, G. Horowitz and R. Hajlaoui in Intrinsically Conducting Polymers: An Emerging Technology (Ed.: M. Aldissi) pp. 107, Kluwer Academic Press, Amsterdam, 1993. 344. H. Akimichi, K. Waragai, S. Hotta, H. Kano and H. Sakaki, Appl. Phys. Lett., 1991,58,1500. 345. K. Waragai, H. Akimichi, S. Hotta, H. Kano and H. Sasaki, Synth. Met., 1993,55-57,4053. 346. K. Waragai, H. Akimichi, T. Inoshita, S. Hotta and H. Sakaki, ANTEC’93, p.2331 347. A. Dodabalapur, L. Torsi and H. E. Katz, Science, 1995, 268, 270. 348. L. Torsi, A. Dodabalapur, H. E. Katz, J. Appl. Phys., 1995, 78, 1088. 349. M. Shur, M. Hack, J. G. Shaw, J . Appl. Phys., 1989,66,3371. 350. G. Horowitz, X.-Z. Peng, D. Fichou and F. Garnier, Synth. Met., 1992,51,419. 351. A. J., 19, Pal, R. Osterbacka, K.-M. Kallman and H. Stubb, Appl. Phys. Lett., 1997, 71,228. 352. L. Torsi, A. Dodabalapur, L. J. Rothberg, A. W. P. Fung and H. E. Katz, Science, 1996,272, 1462. 353. M. W. Wu and E. M. Conwell, Chem. Phys. Lett., 1997,266,363. 354. F. Garnier, G. Horowitz, X. Z. Peng and D. Fichou, Synth. Met., 1991,45, 163. 355. F. Garnier, A. Yassar, G. Horowitz and F. Deloffre, Mol. Cryst. Liq. Cryst., 1993, 228, 81. 356. G. Horowitz, F. Garnier, A. Yassar, R. Hajlaoui and F. Kouki, Adv. Mat., 1996,8, 52. 357. A. R. Hepburn, D. M. Goldie, J. M. Marshall, J. M. Maud and D. M. Haynes, J. Non- Cryst. Solids, 1993, 164-166, 1263. 358. M. S. A. Abdou and S . Holdcroft, Chem. Mater., 1994,6, 962. 359. T. Taka, M. T. Loponen, J. Laakso, K. Suuronen, P. Valkeinen and J.-E. Osterholm, Synth. Met., 1991, 41-43, 567. 360. M. T. Loponen, T. Taka, J. Laakso et al., Synth. Met., 1991, 41-43, 479. 361. G. Tourillon and F. Garnier, J. Electrochem. SOC.:Electrochem. Sci. Technol., 1983,2043. 362. D. M. de Leeuw, Synth. Met., 1993,55-57, 3597. 363. M. Stoldt, P. Bauerle, H. Schweizer and E. Umbach, Synth. Met., 1993,55-57, 4059. 364. M. Stoldt, P. Bauerle, H. Schweizer and E. Umbach, Mol. Cryst. Liq. Cryst., 1994, 240, 127. 365. G. Horowitz, X. Peng, D. Fichou and F. Garnier, Appl. Phys. Lett., 1990,67,528. 366. M. S. A. Abdou, F. P Orfino, Z. W. Xie, M. J. Deen and S . Holdcroft, Adv. Mat., 1994,6,838. 367. H.-J. Egelhaaf, L. Luer, D. Oelkrug, G. Winter, P. Haisch and M. Hanack, Synth. Met., 1997, 84, 897 (1997)

325. 326. 327. 328. 329. 330. 331.

5 Charge Transport in Semiconducting Oligothiophenes Gilles Horowitz and Phillippe Delannoy

The concept of ‘semiconductor’ has largely evolved since the first appearance of the word, by the end of the 19th century. In most dictionaries, semiconductors are defined as ‘non-metallic materials the conductivity of which lies between that of an insulator and that of a metal’. As a matter of fact, although the band theory yields a clear distinction between metals and insulators, it does not delineate a decisive border line between semiconductors and insulators. A fundamental difference can be made between two sets of semiconductors, the intrinsic and extrinsic ones. In the former, electrons are thermally excited from the valence to the conduction band, which results in a thermally activated conductivity, with an activation energy that equals half the band gap. Because the concentration of intrinsic carriers at a given temperature also decreases exponentially with the band gap, intrinsic semiconductivity is only observed in low band gap materials, the prototype of which is germanium (E, = 0.66eV). (It must be pointed out that even ultra pure silicon (E, = 1.12eV) is not intrinsic at room temperature.) Extrinsic semiconductivity results of adding to an insulating material minute amounts o f a doping impurity. (Here, ‘insulating’must be taken under the meaning of ‘non metallic’, and therefore also includes the so-called semiconductors.) Because the latter kind of semiconductivity requires ultra pure materials, it has only been brought to light at the aftermath of the second world war, well after intrinsic semiconductivity. Its technological importance is however crucial, since it launched the development of modern solid-state microelectronics. Owing to the fact that even high band gap materials, such as diamond (E, = 5.47 eV) or boron nitride (E, = 7.5 eV), are currently studied in view of their use in electronic devices (particularly in view of realizing blue laser diodes), ‘extrinsic semiconductors’ could also be narned ‘doped insulators’. ‘Organic semiconductors’ were first referred to in 1948 [l], when it was found that phthalocyanine presented a thermally stimulated conductivity, a behavior typical of an intrinsic semiconductor. It was shown later on that organic compounds actually behave mainly as extrinsic semiconductors. Photovoltaic cells [2], light emitting diodes (LEDs) [3], and more recently field-effect transistors (FETs) [4] have now been realized with organic materials. Nevertheless, organic semiconductors present fundamental differences with their inorganic counterparts. These differences are readily seen from the conductivity, which is much lower in organic semiconductors than that in inorganic ones. This may come from two sources: a low concentration of charge carriers, and a low mobility. In fact, both explanations are relevant. The low carrier density in organic compound comes from the difficulty to dope them in a controlled way, and represents a major drawback in the development of organic electronic devices. The origin of low mobility constitutes the subject of this chapter.

284

5 Charge Transport in Semiconducting Oligothiophenes

5.1 Basic models Charge transport in a solid is measured at a macroscopic scale by its conductivity CJ

= nqp

(1)

where n is the density, q the charge, and p the mobility of the carriers. The latter is defined by

v = pF

(2)

which assumes that the mean velocity of the carriers v is proportional to the applied electric field F (exceptions to this rule, leading to a field dependent mobility, will be discussed in section 5.1.2). The high conductivity of metals is essentially due to a very high charge (i.e. free electron) density. Yet, charge mobility in metals is rather low (typically between 10 and 100cm2V-' s-'), being limited by a high rate of collisions. On the other hand, the charge density in conventional inorganic semiconductors is lo4 to lo8 lower than in metals, but their mobility can be up to lo3 higher.

5.1.1 The band model The band model derives directly from quantum mechanics. It has proven highly successful to explain satisfactorily both the high conductivity of metals, and the properties of high mobility conventional inorganic semiconductors. Although it is far less appropriate to the case of organic compounds, it is worth giving here a concise overview of this basic model. The concept of energy bands in a solid can be physically understood by considering a one dimensional crystal, with a lattice constant a, and a nearly free electron [5], for which the bands are treated as a weak perturbation. The energy and wavefunction of a free electron are respectively of the form

ttk2

Ek = -

2m

(3)

exp ikr

(4)

(see left-hand side of Fig. l), and

$+(r)

0;

which is the equation of a plane wave. As any other plane wave (one can think to the electromagnetic wave associated with an X-ray), the propagation of the wave function of the nearly free electron in a crystal is exposed to Bragg reflection. The condition for reflection in a one dimensional lattice occurs when k = f G / 2 , where G = n r / a is a vector of the reciprocal lattice. The region where - r / a 5 k 5 + r / a defines the first Brillouin zone (BZ). Because they are reflected there, the wavefunctions at the BZ boundaries are no longer the traveling waves exp(&irx/a) of the free electron, but standing waves, made up equally of waves traveling to the right and to

5.1 Basic models

285

Figure 1. Energy dispersion law of (left) a free electron, and (right) of a nearly free electron in a one dimensional periodical potential. The hatched area corresponds to the forbidden gap.

the left. These standing waves can be either a symmetric, $Js cx cos kx, or antisymmetric $J, o( sin kx, linear combination of traveling waves. Unlike that of a traveling wave, the probability density (defined as of these standing waves varies as a function of the position. If we assume that the origin of x is at a center of lattice site, and owing to the fact that the potential energy of an electron in the field of a positive ion is negative, we expect to find the potential energy of $Js lower, and that of q!Ja higher than that of a traveling wave. The difference between these potential energies defines the energy gap Eg (see right-hand side of Fig. 1). The model above describes a free electron in a one dimensional crystal at the BZ boundaries. In the more general case, the solution of the time independent Schrodinger equation ( 5 ) in a three dimensional crystal can be obtained with the help of the Bloch theorem, which states that if the potential energy V(r) is periodic, the solutions $Jk(r)of

are of the form q!Jk(r)= exp(ik . r)Un(k,r)

(6) where Un(k,r) is periodic with the same periodicity as V(r), that is the periodicity

of the direct lattice. It can then be shown that the energy Ek has the periodicity of the reciprocal lattice, that is, Ek = Ei(+G(again, G is a vector of the reciprocal lattice). In other words, for a given band, it suffices to use k's in a primitive cell of the reciprocal lattice. In the standard conventions, this primitive cell is the first BZ. Calculated band structures show a series of allowed and forbidden bands. The highest energy occupied allowed band is called the valence band (VB), and the lowest energy unoccupied one the conduction band (CB). Between these bands lies a forbidden band, the width of which defines the energy gap of the solid. Note that in a molecular solid, a one to one correspondence can be made between the VB and the CB on the one hand, and the Highest Occupied Molecular Orbital (HOMO) and Lowest Unoccupied Molecular Orbital (LUMO) of the isolated molecule on the other hand.

286

5 Charge Transport in Semiconducting Oligothiophenes

Another important concept in band theory is that of effective mass. In a semiconductor, most of the charges reside at the edge of the conduction or valence band. Band edges can be approximated to parabolic bands, by analogy with the free electron dispersion law, Eq. (3) E=E,,+-

h2k2 2m*

(7)

Accordingly, the effective mass m* is defined as

_1 -m*

1 $E(k) h2 dk2

--

(In practice, as the energy dispersion curves depend on the crystal direction, m* is tensorial.) As a general rule, it can be stated that the stronger the binding between the elemental components of the crystal, the wider the allowed energy bands. Accordingly, in covalently bound solids, allowed bands have a width of several tens of eV, and effective masses are not very different from (and most often lighter than) the electron mass. States in these bands are thus delocalized, and electrons in the CB and holes in the VB free to move, their mobility being only limited by scattering with phonons or ionized impurities. 7r-conjugated polymers also present strong coupling in the direction of the chain, which should in principle suggests that electron are free to move, and that band theory may apply in that direction. This view is however far from being verified by experiment, mainly because isolated chains cannot be practically accessed, and hence macroscopic electrical behaviors are limited by interchain mechanisms. In molecular solids, all atomic bonds are already engaged within the molecules. Therefore, intermolecular forces - generally van der Walls forces - are weak and short-ranged. This results in narrow bands, which in turn lead to high effective mass and low mobility. ‘Narrow’ band means that the band width compares with kT at room temperature. A criterion for the suitability of the band model in the case of narrow bands is that the mean free path X of charge carriers is larger than the lattice constant a. If the converse is true, a transport through delocalized states would be physically meaningless. In that case, other transport mechanisms must be invoked, as will be seen in the next section. We now introduce the tight-binding approximation, which is very useful in the case of narrow bands. It assumes that the effect of the potential at a given site of the crystal is limited to its nearest neighbors. In that case, the energy dispersion is given by

Ek = Eo - 2Jcos(ka)

(9)

where J i s the overlap (or transfer) integral and a a lattice vector. (Jis defined by the matrix element (hI Vi I$,), where and $, are the wave function at two neighboring sites, and Vi the potential at site i.) In that case the effective mass is m* = h2/2Ja2. The carrier mobility p can now be estimated as follows. The mobility is connected through Einstein’s equation (10) to the diffusion coefficient D = (u2). = X(u), where

5.1 Basic models

(u)is the mean velocity of the charge, above defined mean free path.

7

p=-

287

a relaxation constant, and X = ( u ) the ~ eD kT

Here, e is the electron charge, k is the Boltzmann constant and T the absolute temperature. The mean velocity, averaged over the states within the band, is estimated to (v) M Ja/A. We finally get p=-

eX Ja kT A

Equation (1 1) predicts a T -I dependence of the mobility, plus any additional temperature dependence of J , a and A, all of which usually follow a power law. The overall dependence of the mobility will hence be of the form pCXT-"

(12 )

In fact, Eq. (12) describes the general temperature dependence of the mobility in most conventional inorganic semiconductors. It is generally not observed in organic

L

0.3'

4

'

'

30

'

'

' '

" '100 -TT[Kl

I 300

Figure 2. Temperature dependence of the electron ( p - ) and hole (p') mobility of anthracene (from Ref. 6).

288

5 Charge Transport in Semiconducting Oligothiophenes

compounds, except in highly pure molecular crystals, where n was found to vary between 1 and 2, as shown in Fig. 2 [6]. But even for these molecular crystals, the use of band theory is not really satisfactory, because the calculated mean free path generally compares with the lattice parameter.

5.1.2 Hopping 5.1.2.1 Localization

The band theory applies to the perfectly regular organization of a crystal, leading to delocalized Bloch wave functions, Eq. (6).In a now classical paper, Anderson [7] has shown that disorder may result in a localization of the states. In that case, the oneelectron wave function takes an exponential form

where a is the size of the localized state. In contrast to the traveling wave function of the free electron, which is delocalised, Eq. (13) is the wave function of an electron localized on a particular site. Charge transport now occurs via tunneling between these localize states. The difference between the delocalized and localized transport can be illustrated by the schematic representation given in Fig. 3. The charge transport in delocalized bands is only limited by the diffusion of charges on lattices vibrations (phonons). As a result, the mobility decreases as the temperature increases (see Fig. 2). When states are localized, the tunneling of charges from one site to the next one may be assisted by the phonons. The mobility is therefore thermally activated (it increases when the temperature increases). This mechanism is called phonon assisted hopping. It is worth pointing out that hopping transport is not restricted to disordered materials. Charges can also be localized in well-ordered crystalline materials via electron-phonon interaction, forming small polarons. This issue will be dealt with in section 5.1.3.

Figure 3. Schematic representation of the charge transport in the delocalised band model (A), and in a localized state material (B). In (A), the conductivity is limited by the diffusion on lattice vibrations (phonons), whereas in (B), the tunneling of charge from one site to the next one is phonon assisted.

5.1 Basic models

289

5.1.2.2 Temperature dependence The temperature dependence of the conductivity in hopping models is obtained by evaluating the hopping rate rij,that is, the probability that an electron at site i will hop to sitej. A well known derivation of this rate was made by Miller and Abrahams [8], in the case of a lightly doped semiconductor at very low temperature. The localized sites are the impurity shallow levels, the energy of which stands within a narrow range, so that even at low temperature, an electron at site i will easily find a phonon to jump to the nearest sitej, The hopping rate is then given by

wheref(E) is the Fermi function, and "io a constant that contains an electronphonon coupling term and the phonon density of states. Ei and Ej are the energy of sites i and j , respectively, and Ri, the distance between the sites. The first exponential term at right hand side represents the density of phonon of energy Ei - Ej, and the last one the tunneling factor. To get the conductivity, the hopping rate has to be modified by an applied electric field F, which will give rise to a net current in the field direction [9]. The modified equation appears as a voltage drop times a factor that can be regarded as the ohmic conductance connecting sites i and j . The macroscopic conductivity can therefore be obtained by constructing an equivalent network and calculating its conductivity from Kirschoff's equations. The calculation will not be developed here. The global result is that in the case of low doping, the average distance between hopping sites will be large and the exponential form of the tunneling factor (last term in Eq. (14)) prohibits hopping to more distant sites: only hops to the nearest neighbor site will occur ('fixed range hopping'). The temperature dependence in that case is governed by that of the first exponential in Eq. (14): The conductivity is thermally activated. Another famous hopping model is that of the Mott's 'variable range hopping' [lo]. It is assumed in that case that the localized sites are distributed in energy over the entire gap. At low temperature, the probability to find a phonon of sufficient energy to hop to the nearest site will be low. Consequently, hop over a large distance can be the most favorable. Mott's derivation leads to

(-g)

114

u = uoexp

where

Here, NF is the density of states at the Fermi level. It was pointed out later by David Emin [Ill that hopping transport could be divided into four categories, which result from two couples of jump regimes. First, a jump can be either strongly coupled (Emin uses the term 'small-polaronic') or weakly coupled [12]. Last, a hop is either adiabatic or nonadiabatic [13].

290

5 Charge Transport in Semiconducting Oligothiophenes

10”-

Id0

-

3-

-L lo9 -

.$

i!! lo8U a

$jlo7 7

lo6 -

lo5 -

Figure 4. Logarithm of the jump rate as a function of the reciprocal temperature (in units of the Debye temperature @) for an electron strongly coupled to the lattice (from Ref. 14).

The strength of the electron lattice coupling corresponds to the square of the ratio of the length of the localized state to that of an atomic vibration, or alternatively, to the ratio of the localized state binding energy Eb to the quantum of vibrational energy Awo. The equation of Miller and Abrahams, which was calculated in the case of shallow impurity states, only applies to weakly coupled localized states, where only hops resulting from the absorption of one phonon are taken into account. In the case of strongly coupled states (which is generally the case in organic compounds), multiphonon processes have to be considered. The resulting temperature dependence of the jump rate is given in Fig. 4 (in units of Debye temperature) [14]. The thermally activated behavior is found at high temperature, whereas the low temperature regime is obviously not thermally activated. The second alternative pointed out by Emin opposes adiabatic to non adiabatic processes. An adiabatic hop occurs when the electron transfer energy between two sites is sufficientlylarge so that the charge carrier can follow the atomic motion. For a strongly coupled jump, the energy required must exceed a characteristic phonon energy. The adiabaticity criterion is in fact correlated to the intersite separation R. A hop is adiabatic when R 5 3a, where a is a lattice constant. One sees therefore that the calculation of Miller and Abraham (which assumes low doping, and hence large intersite distances) correspond to a non adiabatic regime. In the adiabatic regime, the jump rate becomes insensitive to the intersite distance.

5.1 Basic models

29 1

5.1.2.3 Field dependent mobility A field dependence of the mobility in polyvinylcarbazole (PVK) was reported from time of flight measurements (a technique that will be described in section 5.2.2) by Pai in 1970 [15]. The result was explained b y a trapped controlled transport (more details on this kind of charge transport can be found in section 5.1.4). The field dependence occurs through a Poole-Frenkel mechanism, in which the coulomb potential near the trap centers is modified by the applied field in such a way as to increase the tunnel transfer rate (the mechanism is analogous to that occurring in the image force lowering of the potential barrier at a semiconductor-metal interface.) The general dependence of the mobility is given by Eq. (16)

where p(0) is the mobility at zero field, /3 = TEE,,)'/^ the Poole-Frenkel factor and F the magnitude of the electric field. It was pointed out later by Gill [16] that, although the model was able to predict the correct field dependence, and even its correct magnitude, it could be objected that the use of a trapped controlled mechanism assumes a transport in delocalized bands (see section 5.1.4). Experimental values of the microscopic mobility, which range from lop2to do not agree with such a transport. However, this objection could be removed by the model developed by Jonscher and Ansari [17], which assumes a thermally stimulated hopping transport in an energy distribution of localized states. Field dependent mobility is observed at high fields ( E > lo5V cm-') in disordered (mainly polymeric) materials. It is generally associated with dispersive transport, an issue which shall be shortly dealt with in section 5.2.2. However, field dependent mobility has also been reported in oligothiophenes, as will be shown in section 5.3.

5.1.3 Polarons A polaron results from the combination of a charge carrier with the lattice distortion induced by this charge. A distinction is usually made between large and small polarons. In the former, the lattice distortion extends over distances large as compared to the lattice constants, while the converse is true for the latter. Energetically, one has to introduce the polaron binding energy Eb, which is defined as the energy gain of an infinitely slow carrier owing to the polarization and distortion of the lattice. For a large and small polaron, the polaron binding energy will be smaller and larger, respectively, than the conduction (or valence) bandwidth. We shall restrict here to small polarons, which are those most generally encountered in organic compounds. 5.1.3.1 Small polaron The theory of the small polaron was first introduced by Yamashita and Kurosawa [18] to account for the very small mobility found in transition metal oxides. It has been extensively developed later by Holstein [19]. A review of the small polaron

292

5 Charge Transport in Semiconducting Oligothiophenes

theory was made in 1970 by Bosman an van Daal[20], and more recently by Shluger and Stoneham [21]. A small polaron can be viewed as an electron self-trapped in its own polarization field. Because it is so strongly coupled to the lattice, hopping transport is generally involved. (A schematic view of the hopping transport of polarons is actually given in Fig. 3, where the self-trapping mechanism is illustrated by a deepening of the potential at the very site where the charge resides.) The Holstein small polaron model is a one-electron, one dimensional model. The Hamiltonian is composed of three terms. The first one concerns the lattice, which is assumed to comprise Nmolecules of reduce mass M that oscillate at a unique frequency wo.The term corresponding to the electrons is that of the above mentioned tight-binding approximation, and is hence characterize by a transfer energy J. The last term accounts for the electron-lattice interaction, and has the general form E, = -Au,, where A is a constant and u, the relative coordinate of the nth molecule. The small polaron mobility was calculated by assuming that the electron transfer energy J can be considered as a small perturbation. The equation arrived at by Holstein is

sinh

"1

fiW0

[

~

2kT exp -27 tanh 4kT

Y

(17)

Here, y is the electron-lattice coupling constant, the calculation of which is both crucial and complex. In the continuum approximation, Yamashita and Kurosawa [181 arrive at a coupling constant e2

where E, is and E~ are the optical and static dielectric constants, respectively. At high temperature ( T > 0, where the Debye temperature 0 is defined by kQ = hw0) Eq. (17) simplifies to

or simply p

0: (kT)-3/2 exp

(-%)

Eb = yAw0 is the polaron binding energy. The condition of validity of Eqs (19) were specified by Holstein. First, the existence of the small polaron requires that J < Eb/2. Moreover, the use of perturbation theory restricts the formula to J < hwo. This upper limit is in fact that for a nonadiabatic process. The adiabatic process, for which J > fiwo, has been studied by Emin and Holstein [22]. The high temperature mobility in that case is given by p =3 ea2 t?wo exp

47r fi kT

J)

5.1 Basic models

_..-

293

Hopping

-H a ping (a - PoiLon b a n ! ) I-...

10-5

." 0

1

2

3

4

5

@IT

Figure 5. Mobility of a small polaron plotted against the reciprocal temperature (in units of the Debye temperature 0). The transition between polaron band (low T ) and polaron hopping (high T ) modes occurs at about 0.40 (adaptated from Ref 19).

It is stressed that, since the overlap integral J is treated as a perturbation, the condition J < Eb/2 should still hold. The low temperature regime has been extensively dealt with by Holstein [19]. He arrives at the result that, under given circumstances, a band regime can occur when T is lower than about 0.40. The mobility is now given by

The temperature variation of the mobility in the polaron hopping and band regimes is shown in Fig. 5. It is worth pointing out that, till now, definitive evidence for a small polaron band conduction has not yet been reported. It has been argued by Yamashita and Kurosawa that, because the small polaron band width is extremely narrow, the small polaron band model cannot be regarded as suitable when impurities are present to a sufficient degree.

5.1.3.2 Molecular 'nearly small' polaron

An attractive model, specifically designed for the case of molecular semiconductors is that of the molecular 'nearly srnall'polaron (MP) developed by Silinsh [23]. Initially, the model was developed to resolve the contradiction between a power law temperature dependence of the mobility (Eq. 12) reported in naphthalene and perylene, which is typical of band theory, and a mean free path of the order of the lattice constant, which would lead to hopping transport. As no transport theory was able to give a satisfactory quantitative description of the experimental facts, a phenomenological approach was adopted. In the model, the carrier is considered as a polarontype particle resulting of interactions of the carrier with vibrations of the lattice.

294

5 Charge Transport in Semiconducting Oligothiophenes

In order to give a physical meaning to the concept of molecular polaron, Silinsh introduces an interaction time T which characterizes the time needed for the formation of the polarization cloud around the carrier [24]. In Fig. 6, the value of T is given for various kinds of interactions. The 'electron polaron' corresponds to the nearly free electron of the conventional band model, whereas the 'lattice polaron' is more or less the above discussed small polaron. By using the Heisenberg uncertainty principle, the interaction time T may be evaluated from the overlap integral J a s T = h/J. In a molecular crystal, the overlap integral J is of the order of 0.01 eV. T is also connected to the mobility through Eq. (22).

Here, r,j is the distance between nearest neighbor molecules. A typical mobility in molecular crystals such as na hthalene, anthracene, tetracene or perylene, ranges between 1 and 10 cm2V-' s-' at room temperature. Silinsh traces a border line between the electronic polaron (i.e. the nearly free electron) and the molecular polaron at 100 cm2V-' s-' . Above this value, the charge carrier moves faster than the time necessary for the formation of the polarization cloud. A very interesting issue of the theory of the molecular polaron is that, unlike the small polaron, its transport occurs via stepping by tunneling without activation energy. Accordingly, the mobility is not thermally activated; rather, it follows a power law in T -n, like in the band theory. The temperature dependence of the mobility can hence be used as a criterion to discriminate between the lattice (small) polaron and the MP.

Mobility

Interaction time-scale

t(s)

cmWs 10-16

0.1

-

)

Interaction energy (eV)

Polaron tYPe

electronic polarization

1.0- 1.5

electronic polaron

molecular polarization

0.1 -0.2

molecular polaron

lattice polarization

10.1

lattice polaron

shallow trapping

20.03

trapped polaron

10-12

Figure 6. Types of interactions between a charge carrier and the nuclear subsystem in a molecular crystal. Typical mobility, interaction time scale and interaction energy are given, from left to right, for each corresponding polaron quasi particule (adaptated from Ref 24).

5.1 Basic models

295

5.1.3.3 Polarons in n--conjugated polymers and oligomers

A linear polymer consists of the repetition along the direction of the polymer chain of a small monomeric subunit. As such, it can be viewed as a one dimensional lattice. Accordingly, the electron energy level can be obtained by using Bloch functions, which results in one dimensional energy bands along the direction of the polymer chain. A 7r-conjugated polymer is characterized by a regular alternation of single and double carbon-carbon bonds. The presence of double bonds results in 7r electrons that are delocalized all over the whole polymer chain. This 7r delocalization gives rise to relatively low band gaps (conjugated polymers are colored, whereas non conjugated ones are transparent), and the possibility of doping the polymer through charge transfer from electron donor or acceptor species. Owing to the low dimensional nature of 7r-conjugated polymer, adding a charge on a polymer chain will result in a profound change of its geometrical structure. By analogy with the above described case of ionic crystals, the association of the charge with the induced deformation of the chain is also called a polaron. The now classical image of a polaron in a n--conjugatedpolymer is depicted in Fig. 7, in the case of polythiophene. The geometrical change is characterized by a permutation of the bond alternation over a certain number of monomeric subunits. Apart from this geometrical change, the polaron can also been visualized as a self localized state, which manifests itself by localized levels within the forbidden gap, which results in a change of the optical absorption (and hence of the color) of the material. The spatial extension of polarons has been calculated for various n--conjugated polymer [25-271. In the case of polythiophene, it corresponds to five monomer units. In a short oligomer, the polaron can no longer be considered as a charge free to move along the polymer chain. Hence, the concept of polaron has to be changed to that of radial cation, in which the geometrical structure modification extends all over the molecule. Radical cations in oligothiophene have been extensively studied during the past years [28-361. They are quite readily obtained in solution by adding an oxidizing agent such as ferric chloride [28-311, or by photochemical methods [33-3.51. The

Figure 7. Polaron in a polythiophene chain. Top: geometrical change of the molecular structure. Bottom: associated localized levels in the forbidden gap.

296

5 Charge Transport in Semiconducting Oligothiophenes

radical cations have been characterized by UV-vis spectroscopy and electron paramagnetic resonance (EPR). Figure 8 shows how the addition of ferric chloride to a solution of sexithiophene leads to the successive formation of the paramagnetic radical cation, and the double charged dication, which is diamagnetic (i.e. does not give rise to an EPR signal). The UV-vis spectrum of the former presents two transitions, whereas the latter has only one. A third species, which presents also two transistions but is diamagnetic, was later recognized as the dication dimer [37], formed by the agregation of two radical cations. It usually forms when the temperature of the solution is lowered. All of these species have also been identified in the solid state [38 -421, and it is currently recognized that the radical cation constitute the charge carrier in undoped or moderately doped oligothiophenes. Charged species in solids can be formed either through addition of a dopant [38, 39, 41, 421, or by injecting charges in a metalinsulator-semiconductor (MIS) structure [40]. The first kind of doping leads to a substantial increase of the conductivity. However, this conductivity (around lop3 to Scm) remains much lower than that of a doped conjugated polymer, and it has been recognized that the hgher conductivities claimed in the first report of doped oligothiophenes [38] was probably due to phenomena of polymerization. The low conductivity of doped oligothiophenes can be related to that since they are strongly bound to the molecule on whch they reside, isolated radical cations behave much like polarons in solids, so that their transport occurs via hopping from one molecule to the next. Although such a description corresponds well to the general definition of the polaron given in the remainder of this section, it differs from that of the polaron in a conjugated polymer, which is assumed to stay on the same polymer chain on which it can move freely. Accordingly, the polaron in an oligomeric crystal could be termed a transverse polaron. We note that transverse polarons and bipolarons, i.e. polarons (bipolarons) that hop from one polymer chain to the next, are also sometimes referred to as the limiting step in the charge transport in conjugated polymers [43].

c I\

1-

0.5

1

-

.........Neutral

1.5

2 2.5 Energy (eV)

3

3.5

4

Figure 8. Electronic (W-vis) absorption spectra of neutral sexithiophene, and of its radical cation and dication in solution. The charged species are obtained by adding F e Q to the solution.

5. I Basic models

297

5.1.4 Multiple trapping Traps are levels localized at impurities or lattice defects that are prone to immobilize carriers. One can differentiate between deep traps, which are localized near the center of the band gap, and shallow traps, which stand near the conduction or valence band. Shallow traps are either donor states localized near the valence band (hole traps) or acceptor states near the conduction band (electron traps). Traps may play an important role in the charge transport in low conductivity materials through the mechanism of Multiple Trapping and Release [44],which is the most recognized model to account for charge transport in hydrogenated amorphous silicon (a-Si : H). In this model, one can visualize the charges as moving in a narrow transport (conduction or valence) band. During their transit through these delocalized states, they interact with the localized states through trapping and thermal release. The following assumption are usually made: First, the carriers arriving at a trap are assumed to be trapped instantaneously with a probability close to one. Second, the release of trapped carriers is controlled by a thermally activated process. The resulting drift mobility pD is related to the mobility po at the delocalized band edge by an expression of the form

In the case of a single level trapping state, Et is the difference between the energy of the trap and the transport band edge, and Q = Nc,v/Nt,where NC,,,is the density of state at the transport (conduction of valence) band edge, and Nt the density of traps. In the case of an energy distribution of traps, effective values of Et and a can be calculated, depending on the form of the distribution of localized states.

5.1.5 Summary It is worthwhile at this point to draw some guidelines among the models developed so far. Firstly, two main categories of transport mechanisms have been identified: transport may occur via delocalized or localized states. From theoretical arguments based on the small polaron theory, Glarum [45]established the frontier between band (delocalized) and hopping (localized) conduction at around 1 cm2V-' s-l. Secondly the most confident way to differentiate these transport mechanisms is to measure the temperature variation of the mobility. In the band model, a dependence as T -n is expected. Insofar as hopping models are concerned, we have seen that a thermally activated mobility is far from being universal. However, a positive temperature coefficient can be taken as an indication of a hopping process. A transport via localized states may also be characterized by a field dependent mobility. Thirdly, the presence of traps will generally induce an additional activation energy, whether the transport proceeds via delocalized or localized states. As a consequence, the measurement of the temperature dependence of the mobility may not be sufficient to discriminate between the hopping and multiple trapping mechanisms.

298

5 Charge Transport in Semiconducting Oligothiophenes

5.2 Measurement of the mobility It results from what came before that the key parameter for studying charge transport is the mobility p. Several techniques have been developed to measure the mobility of free charges in organic materials. A crucial issue for having access to the mobility is the generation of charge carriers. The latter may occur from three different processes: (i) thermal generation, which results either from direct excitation from the valence band to the conduction band (intrinsic process), or from ionization of impurity levels (extrinsic process). As stated in our introduction, only the latter process is of interest in real materials; (ii) charge injection from metal electrodes, and (iii) photogeneration. The first process is involved in direct conductivity measurements, the second one in the space-charge limited and field-effect current, and the third one in the time of flight technique.

5.2.1 Conductivity The conductivity of a solid is given by Eq. (1). If we assume that, as in conventional extrinsic semiconductors, the concentration of free carrier is simply given by the density of dope, the mobility would be proportional to the conductivity, and the measurement of the conductivity would be a very simple way to have access to the mobility of an organic material. Unfortunately, this is rarely the case, and an independent measurement of the free carrier density would be most often required. As reliable determinations of the latter is generally not available, a direct measurement of the mobility is generally the necessary counterpart to conductivity measurements. The easiest way to measure the conductivity is the so-called two probe technique, in which the conductivity is estimate from the current voltage characteristic of an organic film or crystal equipped with two similar metal electrodes. This technique assumes that the contact resistance is much lower than that of the bulk material, which is most often the case in organic semiconductors. However, a distinction must be made between blocking and injecting contacts. This point will be dealt with in more details in section 5.2.3. The starting point is the formation of the interface between a semiconductor and a metal. When two materials are put into intimate contact, electron transfers occur in order to equalize the Fermi level at both sides of the interface. The resulting surface charge gives rise to an electric field, a band curvature, and eventually the formation of a potential barrier. The process is depicted in Fig. 9 in the case of a p-type semiconductor (p-type semiconductivity is that encountered in oligothiophenes). The barrier height is given by the difference between the ionization potential (that is, the distance from the top from the valence band to the vacuum level) of the semiconductor and the work function of the metal. Accordingly, low work function metals will give high barrier height (blocking contact), whereas high work function metals result in low barrier height (ohmic, low resistance contact). (The converse would be true for a n-type semiconductor.) In practice, gold (work function

5.2 Measurement of the mobility

A 4n

299

A EC ‘P

-

EF EV

Figure 9. Formation of a potential barrier at the semiconductor-metal interface. Left: before contact; right: after contact. Ec, Ev and EF are the energy of the bottom of the conduction band, the top of the valence band and the Fermi energy, respectively. c,?Im, is the work function of the metal and Zp the ionization potential of the semiconductor. The barrier height $, is given by c,?I, = Zp - c,?Im.

5.1 eV) has been recognized to give good ohmic contacts to most oligothiophenes, whereas aluminum (work function 4.2 eV) generally forms a blocking contact. Although conductivity cannot be used by itself to estimate the charge mobility, it may be useful to study the transport anisotropy of organic materials. Its temperature dependence may also be of interest, insofar as the free carrier density remains constant, which is indeed the case in extrinsic semiconductors in certain temperature ranges.

5.2.2 Time of flight Time of flight (TOF) is undoubtedly the technique of choice for measuring the mobility in low mobility materials. The principle of the TOF experiment, which was first described by Le Blanc in 1960 [46], is depicted in Fig. 10.

Time

Figure 10. Principle of the time of flight measurement. Top: schematic view of the carrier generation and transport. Bottom: resulting time dependent current.

300

5 Charge Transport in Semiconducting Oligothiophenes

The sample consists of an organic film or crystal sandwiched between two conducting electrodes. Charge generation occurs via a short pulse of light, which creates electron-hole pairs in the immediate vicinity of the front electrode. For this reason, the latter is most often constituted by a transparent conductor, such as indium doped tin oxide (ITO), but semitransparent metal electrodes are also often used. The thickness of the charge generation layer, which is defined by the penetration of the light in the material, must be much lower than the total thickness of the organic film or crystal. If this requirement is fulfilled, the initial carrier distribution may be regarded as a two-dimensional sheet. Concurrently with the light pulse, a voltage is applied between both electrodes. Depending on the polarity of the resulting electric field, one carrier species will readily discharge at the front electrode, whereas the other one will have to travel across the film to reach the rear electrode. This charge transport gives rise to a displacement current which can be recorded in the external circuit (bottom of Fig.10). This current is constant, and then falls down to nothmg at the time T~ at which the charge sheet arrives at the rear electrode. The transit time is related to the mobility through Eq. (24). T

r2

Here, L is the distance between the electrodes, P the electric in the organic layer, and V the external voltage across the sample. In principle, the TOF signal should present a step shape (see Fig. lo), where the falloff of the current would correspond to the arrival of the charged sheet. Unfortunately, life is not so simple, and the charges may experience several features during they travel from the front to the rear electrode. One of these is diffusion, which will result in a spreading of the charges, and lead to a smoothing of the falloff of the signal. Another one is trapping. If charge trapping occurs, an exponential decay, due to a reducing of the number of charges, is superimposed to the step curve. If trapping is too strong, this decay may eventually hide the final falloff that indicates the arrival of the charges. For this reason, TOF measurements on organic crystals are only achievable with highly pure and flawless samples. In the case of highly disordered materials, such as amorphous solid, or molecules dispersed in a polymer matrix, the TOF signal is generally not seen on a linear current vs. time curve. However, plotting the same curve on a log log plot shows a change of slope at a certain time. This behavior can be explained by the theory of dispersive transport, which was first initiated by Scher and Montroll [47]. In their model, the transport in a disordered medium is depicted through a concept of time random walk, based on an anomalous time-dependent dispersion law that differs from the classical Gaussian law. As such a transport is usually not encountered in well ordered oligomers, we shall not detail here the theory of dispersive transport.

5.2.3 Space-charge-limited current Space-charge limited current (SCLC) occurs in materials with very low concentrations of free carriers (ideally, perfect insulators). When such materials are put into

5.2 Measurement of the mobility

301

intimate contact with a metal, which can be viewed as a reservoir of free carriers, these free carriers will tend to diffuse into the insulator, provided the Fermi level of the metal is located at (or close to) allowed levels of the insulator, i.e. in the case of an injecting contact. This mechanism is called charge injection. It may concern holes or electrons, depending on whether the Fermi level is placed near the valence or conduction band edge, respectively. The former mechanism occurs with high work function metals, whereas the latter pertains to low work function metals.

5.2.3.1 Profile of injected charges When charges diffuse into an insulating material, they give rise to an electrostatic field that acts against the penetration of additional charges into the material. The resulting carrier profile can be calculated from Maxwell's equations, the more pertinent of which is Poisson's equation:

dV

dF

qn(x)

z=dx== Here, E is the dielectric constant of the insulator, and E~ the permittivity of free space. The electric field created by the injected charges generates a current density J , which is the sum of a drift and a diffusion current dn dx where D is the diffusion coefficient. Combining (25) and (26),and making use of Einstein's equation (10), we obtain J

= nqpF

-

qD

-

This second order differential equation cannot be analytically solved in the general case. In 1940, Mott and Gurney derived an exact solution in the particular case where J = 0, for a semi infinite material, and assuming that the electric field is zero at x = cc [48]. The charge density and potential are then given by n(x) =

2~~0kT

q2(x + x0l2

and 4

Here

{ z 2~~0kT

XO =

can be regarded as the spatial extension of the injected charge. no is charge density at x = 0. The total injected charge is Qo = qnoxo.

302

5 Charge Transport in Semiconducting Oligothiophenes

5.2.3.2 Estimation of the space-charge limited current Putting a metal electrode at both faces of an insulating film results in making a capacitor. When a voltage Vis applied between both electrodes, charges are injected into the bulk of the linsulator. The total induced charge is limited to Q = CV, where C = E E ~is/ L the capacitance per unit area of the insulating film, and L its thickness. The resulting drift current is J = (Q/L)pLF = EEO,UV~/L~. Here, F = V / Lis the mean electric field. We note that the typical V 2 dependence results of the voltage dependent charge density, which contrasts with the ohmic conduction where the bulk carrier density is inde endent of the applied voltage. The thickness dependence in LP3results of the L- dependence of the injected charge density, as opposed to the ohmic carriers density, which is independent of the thickness. The derivation of the exact expression of the space charge limited current would require the resolution of Eq. (27) in the case of a non zero current. Unfortunately, such a resolution cannot be performed analytically. The standard expression of the space charge limited current (Eq. 30), as derived by Lampert [49], was obtained by neglecting the diffusion current (second term of the right hand side of Eq. (26)). It is in fact very close to that deduced from the phenomenological model developed above.

P

a

The actual current-voltage characteristic of an insulator (or organic semiconductor) film usually presents two regimes. The quadratic SCLC is observed at higher voltages, whereas a linear regime is recorded at low voltages. This is most often attributed to an ohmic behavior, due to a low, but not zero, concentration of extrinsic free carriers in the material. (The presence of intrinsic carriers is generally ruled out owing to the large band gap of insulators.) However, it has been pointed out that a linear regime also would also exist in an ideal insulator (with no charge at all), which would originate from carriers injected by the electrodes at both edges of

L3

xo

L2

Figure 11. Charge profile in an insulator film with symmetrical injecting contacts at both sides of the film, for various film thickness L 1 ,Lz and L3. xo is the characteristic length given by Eq. (29) (adaptated from Ref. 50).

5.2 Measurement of the mobility

303

the insulating film. If the film is thin enough, the concentration of injected carriers near the electrodes (Eq. 27) cannot be neglected any longer. The resulting profile for two symmetrical contacts is given in Fig. 1 1 [50]. As the thickness L of the insulating film decreases, the concentration of injected carrier increases, and may become important when L is lower that xo.If the minimum concentration at x = L/2 exceeds that of the extrinsic carriers, the linear current will only results from injected charges. This is likely to occur when L < xo.Such a space-charge limited linear current can be distinguished from the conventional ohmic current from its thickness and temperature dependence. We shall come back to this linear regime of the SCLC in the next section.

5.2.3.3 Effect of traps The effect of traps on the SCLC has been widely studied, both theoretically and experimentally. In the case of a single shallow trap level of density Nt lying at an energy Et below the conduction band (or above the valence band), the current is simply multiplied by a factor 0 = nf/(nf nt), where nf and nt are the density of free and trapped carriers, respectively. In the case where nf << nt, and assuming that the carrier distribution follows a simple Boltzmann statistic, 0 is given by

+

The case of an exponential distribution of traps has been extensively developed by Mark and Helfrich [51]. Such a distribution is characterized by the function h ( E ) = H exp(--&) k T, ~

where E is the energy measured from the nearest band edge, H the total trap density and T, a characteristic temperature that defines the slope of the distribution. Equation (32) only holds for temperatures lower than the characteristic temperature. 0 is now given by

and the SCLC by Eq. (34)

where l = T J T . In the case of a single trap level, & = 1 (T, = T ) ,and the trap-free limit corresponds to 4 = 1 and 0 = 1. One can verify that in this case, Eq. (34) reduces to (30). As stated in the previous section, the standard SCLC theory, as developed by Lampert and Mark and Helfrich, neglects the carrier diffusion. A more comprehensive theory, also including an exponential distribution of traps, has been elaborated

304

5 Charge Transport in Semiconducting Oligothiophenes

by Bonham [52]. Two limiting cases can be discerned. At low voltages, the current is linear (‘ohmic’), with a conductance G =j / V

r(e+ 1) P+I

r(1/2)r(c

(35)

+ 1/2)

+

where r(x) = (x l)! is the gamma function. In the trap free case (l = 1 , O T, = T), (35) reduces to

=

1 and

It must be stressed again that this linear current does not correspond to the conventional ohmic current; it appears in materials completely devoid of free carriers, and is only due to carriers injected from the metal electrodes at both sides of the insulator. In other words, the linear current has the same origin as the more classical space charge limited current observed at higher voltages. We note once more that the space charge limited linear current differ from the conventional ohmic current through its temperature and film thickness dependence. At high voltages, the current voltage curve merges with that predicted by the standard SCLC theory, Eqs (30) and (34). The log-log plot of a typical current voltage curve is shown in Fig. 12 [53]. At low voltages, the curve is always linear (slope 1). This linear regime may be due to either

Figure 12. Log log plot of a typical space charge limited current (from Ref. 53).

5.2 Measurement of the mobility

305

extrinsic carriers, or carriers diffused from the electrodes (linear space charge limited regime). At higher voltages, the log-log plot presents a slope of I + 1, which corresponds to the standard SCLC regime. When traps are present, the current may show a sharp rise at a potential VTFL. This is the so-called trap filled limit, which corresponds to the filling of all traps by the injected charges. Above VTFL, the current follows the trap free regime, with a slope of 2. In the case of a single shallow trap, the trap filled limit occurs at

5.2.4 Field-effect A field-effect transistor (FET) is a three terminal device, equipped with three electrodes named source, drain and gate. The device architecture mostly used with organic semiconductors is the thin film FET (TFT), which is schematized in Fig. 13. TFTs were first realized with polycrystalline semiconductors [54],and showed very well suited to low conductivity materials. They have been used for long for determining the mobility, and especially the distribution of localized states in hydrogenated amorphous silicon [%], and occasionally in organic semiconductors [56]. It was only later that their use in practical devices has been envisioned: They are currently employed in active matrix liquid crystal displays [57]. A TFT operates as follows. Owing to both the low mobility and thickness of the semiconducting film, a very low current flows between source and drain when a bias Vd is applied between these electrodes. Applying a second bias Vg between the source and the insulated gate results in charging the insulating film: Opposite charges form at the gate and source. If the latter is injecting to the organic semiconductor (see preceding section), the charge will spread over the semiconductor insulator interface. In turn, this charge can be drifted by the source drain voltage.

lrain semiconductor

insulator substrate

Figure 13. Schematic view of a thin film field effect transistor. W and L are the channel width and length, respectively. V, and Vg are the source-drain and source-gate voltages, and Id the drain current.

306

5 Charge Transport in Semiconducting Oligothiophenes

In other words, the source drain current is modulated by the source gate bias. At low v d , the current is proportional to the drain bias: This defines the linear regime. As the drain bias is increased, while keeping the gate bias constant, the drain-gate voltage decreases. When eventually v d = Vg, the drain-gate bias vanishes, and the drain current does not depend any more on the drain voltage: The so-called saturation regime has been reached. The drain current I d in both regimes is given by Eqs (38) [57, 581.

Here, Wand L are the channel width and length, respectively, Ci the insulator capacitance per unit area, and V , the threshold voltage, above which field effect occurs. Equations (38) show that the drain and gate bias dependence of the drain current allows a direct determination of the carrier mobility. A classical way to estimate p is the plot of the square root of the saturation current as a function of the gate bias. The saturation current is generally measured at equal gate and drain voltages. The plot should give a straight line, which would intersect the voltage axis at the threshold voltage, and the slope of which can be used to calculate the mobility. An alternative way is the use of the linear regime; by differentiating Eq. (38a) with respect to the gate voltage, we get the so-calledtransconductance gm of the TFT [59]:

In this case, the mobility is obtained by a numerical derivation of the drain current measured at a low drain voltage (at least, lower than the gate voltage) as a function of the gate voltage. One advantage of this second method is the access to an eventual gate voltage dependence of the mobility.

5.3 Transport properties of oligothiophenes Though oligothiophenes have been first isolated and synthesized in 1947 [60], their electrical properties have been studied only very recently. They were first investigated as models of their parent polymer, polythiophene, which has been previously recognised as a p-type conducting polymer (i.e., an organic polymer the conductivity of which can be considerably increased by doping with an electron acceptor), but soon appeared as interesting materials by themselves. As stated in section 5.1.3.3, charges in oligothiophenes are radical cations. Their transport is usually described with the help of polaron models. The determination of the carrier mobility, and of its dependence as a function of the temperature is therefore of crucial importance.

5.3 Transport properties of oligothiophenes

307

5.3.1 Conductivity, mobility and carrier density Owing to the poor solubility of the oligomers longer than five thiophene units (quinquethiophene), most electrical measurements on oligothiophenes are carried out on films vacuum evaporated on insulating substrates of on substrates already equipped with a conducting electrode (generally gold or indium tin oxide). The conductivity of S cm-' [61]. The non intentionally doped oligothiophenes ranges from lo-'' to latter figure, which compares with that of undoped polythiophenes, is obtained with the longer chains, sexithiophene and octithiophene. Non-substituted oligomers may present a weak anisotropy, the conductivity along the film being roughly five times higher than the conductivity perpendicular to the film [62]. This anisotropy can be related to the structure established from X-ray diffraction and UV-vis spectroscopy, where the molecules in an evaporated film orient primarily with their long axis perpendicular to the film [63]. Accordingly, the charge transport appears easier in the direction perpendicular to the long axis of the molecules. That is, the charge transport is two dimensional. This is in contrast with conducting polymers, where the most favorable transport direction is parallel to the polymer chains. The anisotropy of the conductivity is substantially increased when the molecules are substituted at each end by alkyl chains. X-ray measurements have shown that such a substitution tends to improve the orientation of the molecules, hence increasing the two dimensional character of the material [62]. Furthermore, the alkyl chains at each end of the substituted molecule tends to decrease the conductivity in the direction perpendicular to the film, which is also the direction parallel to the molecules.

5.3.1.1 Variation with chain length The charge mobility in sexithiophene was originally estimated from the SCLC of a symmetric sandwich configuration (see section 5.3.1.2). However, the most widely used technique now is that of field-effect. The first FET based on an oligothiophene (sexithiophene) was reported in 1989 by Horowitz and coworkers [64], with a mobility of a few lop4cm2 sC1V-' . Since then, the performance of oligothiophene FETs Table 1. Dependence of the field-effect mobility of oligothiophenes as a function of chain length and alkyl substitution at end position. Compound

Mobility (cm2V-' sC1)

References

Non substituted oligomers Quaterthiopene Quinquethiophene Sexithiophene Octithiophene

2 x lo-' (1 -- 3) x 5 x 1 0 - ~- 2 2x

[651 [61, 65, 661 [61, 62, 671 1611

Alkyl substituted oligomers Terthiophene Quaterthiophene Quinquethiophene Sexithiophene

2x 9 (2 - 5) x

[651 ~ 5 1 1651 [62, 65, 681

308

5 Charge Transport in Semiconducting Oligothiophenes

Table 2. Recent data on the field effect mobility of unsubstituted and alkyl substituted oligothiophenes. Compound

Mobility (cm2V-' s-')

Non substituted oligomers Quaterthiopene Quinquethiophene Sexithiophene Octithiophene

- 6 x lop3 1.5 (1 - 3) x (1 - 3) x

Alkyl substituted oligomers Quaterthiopene Sexithiophene Octithiophene

(2 - 4) x 10-2 (3 - 5 ) x (1 -2) x lo-*

References

[731 [62, 68, 741 [721

has steadily improved, due to progress in both the purity of the materials and technique of vacuum deposition. A first breakthrough occurred with the synthesis of oligothiophenes substituted by alkyl groups at end position [62,65]. As can be seen in Table 1, the field-effect mobility of oligthiophenes increases as the chain length increases. Importantly, the field-effect mobility of the substituted oligomer is ten to one hundred times higher than is non substituted homologue. We also note that the mobility of the octamer is lower than that of the hexamer. This was first attributed to the difficulty in purifying this quasi insoluble compound. It appeared later that the mobility of the unsubstituted sexithiophene could reach values close to that of the end substituted compound by evaporating the material on a heated substrate [63], or by using low evaporation rates [69].As stated in the previous section, it has been recognised by X-ray measurements that such circumstances lead to films where the molecules tend to orient the molecules parallel to the normal of the film, an arrangement that is naturally occurring with the substituted oligomer. Moreover, recent measurements on shorter oligothiophenes (quater- and quinquethiophene) gave a mobility very to that of the sexi- and octithiophene [70, 711. Significant recent results are gathered in Table 2. These results would indicate an ultimate value of a few 10-2cm2s-'V-'. The question of whether this limit can be overcome is still open. Recent measurements on sexithiophene single crystals gave a mobility of 0.075 cm2s-' V-' [75]. Though this figure is significantly higher than that reported for polycrystalline films, it does not indicate the possibility of a real improvement of charge transport through an improvement of the crystallinity.

5.3.1.2 Carrier density The carrier density in oligothiophenes can be estimated from independent measurements of the conductivity and field-effect mobility. With a conductivity between lop7 and lop6S cm-' and a mobility of about cm2V-' s-l, the carrier density in sexithiophene ranges from lo'* and l O I 5 ~ m - More ~ . recent measurements on highly pure sexithiophene, as well as single crystals [75], indicate an even lower conductivity of lo-'' to lop9S cm-', whereas the mobility has increased to a few lop2cm2V-' s-'

5.3 Transport properties of oligothiophenes

309

(see Table 2), which means that the carrier concentration has fallen down to 10" ~ m - Obviously, ~. earlier works dealt with rather impure materials. An independent way to estimate the carriers density in a semiconductor is the so-called capacitance-voltage (CV) method, which consists in measuring the capacitance of a Schottky diode as a function of the reverse voltage. We shall not detail here the principle of the method, which can be found in various textbooks [57, 581. It is based on the fact that the thickness of the depletion layer that forms under reverse polarization depends on the applied voltage. The carrier density is given by the slope of the plot of the squared inverse capacitance as a function of voltage. Carrier densities found by this method usually ranges between lOI7and 1019cm-3 [76],which is significantly higher than that estimated from independent measurements of the conductivity and mobility. Such a discrepancy has not been the subject of extensive discussions in the literature so far. It has however been suggested that it could result from the presence of traps [77]. CV measurements are in fact sensitive to both free and trapped charges, whereas only the former are seen in conductivity measurement. The conductivity of oligothiophenes can be increased by several orders of magnitude by a deliberate doping with an electron acceptor (p-type doping). Most of the reports implicate doping with inorganic impurities, such as FeC13,which results in a considerable increase of the conductivity, up to 0.01 S cm-' [39]. We note that the achievement of low intentional doping levels, which is crucial in the development of electronic devices, are generally reported not to give any detectable change of the electronic properties of conjugated molecules. It appears in fact that the attainment of low impurity levels is very difficult to achieve in a controlled way in these materials. A possible solution to that problem could the use of organic dopants, which would be more compliant with organic semiconductors [71]. 5.3.1.3 Variation with temperature

Although the temperature dependence of the mobility is crucial for determining the variety of polaron transport that occurs in oligothiophenes, temperature dependent measurements are scarce. A first example of a determination of the temperature dependent conductivity and mobility of sexithiophene was performed by Horowitz and coworkers [77], from the measurement of the current voltage curve of films sandwiched between two gold contacts. The curves appeared to deviate from linearity at high voltages, which was interpreted as evidence for SCLC. The space-charge limited current, which presented a quadratic voltage dependence, was used to estimate the mobility with the help of Eq. (30). The conductivity deduced from the linear current was thickness dependent in thin films, and became independent of the thickness for thicker films. This was interpreted as evidence for a linear (low voltages) space-charge limited current (Eq. 36) at low thickness, the true ohmic current only occurring at high thickness. The conductivity at room temperature was estimated to 1.3 x lop7Scm-*. The conductivity and mobility was determined from the current voltage curves at various temperature in the linear and quadratic regime, respectively. Both the conductivity and the mobility were thermally activated, with an activation energy of

3 10

5 Charge Transport in Semiconducting Oligothiophenes

0.29 and 0.28 eV, respectively. Clearly, the temperature variation of the conductivity could be assigned to that of the mobility alone, which would indicate that the density of free carriers was temperature independent. The later was estimated to about 3 x 1013crnp3. The temperature dependence of the mobility was attributed to a single shallow trap level, located above the valence band at an energy equal to the activation energy. It appeared later that the low purity and high disorder of these early sexithiophene films could indeed generate a high concentration of traps (see section 5.3.2). Later, a set of temperature dependent measurements was conducted by Waragai and coworkers [78] on ethyl substituted oligomers. Again, the mobility was found to decrease with decreasing temperature, which was interpreted as the indication of a hopping transport. Unfortunately, the temperature range covered by the data (90 to 279K) was not large enough to establish a clear dependence that could allow to discern between the various hopping models. However, the saturation current appeared to vary as the square root of the source drain field F, with a slope increasing with decreasing temperature, which was attributed to a field dependent mobility. The data were explained on the basis of thermally activated transport of small polarons, with a Poole-Frenkel like electric field dependence. The mobility is then given by p = po exp(-E,,,/kT), where Etot= Eb/2- e312E~- 1 1 2 ~ 1 1 2. Here, E b is the polaron binding energy, and cp is defined by l/cp = l / c m - 1/c0, where E~ and E~ are the high frequency and static dielectric constants, respectively. The polaron binding energy was estimated by extrapolating E,,, to zero electric field, and was compared to that obtained from the absorption spectra of oxidized oligothiophene in the solid state (see Figure 8), where the lower energy peak is associated with half the polaron binding energy. The agreement seems good, the respective energy being 0.24, 0.18 and 0.16 eV for the quater-, quinque- and sexithiophene, respectively, obtained from the field dependent drain current, as compared to 0.27, 0.21 and 0.18 estimated from the spectroscopic data. More recently, Torsi and coworkers [79] conducted temperature dependent measurements over a wider range (4 to 350K). A remarkable feature found by these authors is a transition temperature TT.At high temperatures, the mobility is roughly thermally activated, with an activation energy of about 50meV. Below TT, there is first a steep increase of the mobility, by two orders of magnitude, followed by a levering off. The mobility at low temperature nearly equals that at room temperature. The transition temperature is 45K for sexithiophene, and 60K for its dihexyl derivative. Torsi et al. analyze their results within the framework of the small polaron theory. The transition temperature would correspond to the transition from the hopping to the band transport. A fit of the data to Eq. (17) and (21) gives estimated values for the polaron binding energy Eb, the vibrational energy hwo, and the overlap integral J of around 100meV, 10meV and lOmeV, respectively. These values satisfy the small polaron condition J < Eb/2. We note that this polaron binding energy is significantly lower than that estimated by Waragai et al. (160meV). This discrepancy, which could be attributed to a higher purity and ordering of the materials, could rise questions regarding the equivalence between the polaron binding energy and the energy levels of the polaron that are measured by optical absorption spectroscopy. Finally, the leveling off of the

5.3 Transport properties of oligothiophenes

3 1I

mobility at low temperatures to 0.01 cm2V-' s-', which is too low to be consistent with delocalized band transport, is accounted for by noting that because of the very narrow width of the polaron band ( lop6 to lop7eV), the delocalization of charges is prevented by scattering by defects, a feature already predicted by Yamashita and Kurosawa (see section 5.1.3.1).

5.3.2 Traps As seen above, the temperature behavior of the charge mobility in oligothiophenes is essentially explained on the basis of the small polaron model. An alternative explanation to the thermally activated mobility has been proposed by Horowitz and coworkers [59, SO], in which a major role is devolved upon traps. An analytical model of the field effect transistor with a single shallow trap level was first developed [SO]. One important conclusion of this modeling was that the charge trapping, and hence the field-effect mobility, is dependent upon the gate voltage. At low gate voltage, the density of trapped charges equals the equilibrium density; that is, that occurring when no voltage is applied at the gate electrode. As the gate voltage is increased, the concentration of trapped charged decreases, up to a voltage where all charges are freed. This feature, which can be compared to that occurring at the trap filled limit in the space charge limited current (see section 5.2.3.3), can be visualized from Fig. 14. In a TFT operating in the enrichment mode, the band curvature at the insulator-semiconductor interface tends to move the majority carrier band (the conduction band in Fig. 14, drawn in the case of a n-type semiconductor) closer to the Fermi level EF. As the gate voltage is increased, the Fermi level eventually crosses the trapping level, which means that this level is emptied, resulting in freeing all charge carriers. The energy of the trap level deduced from this modeling was in good agreement with that deduced from the measurement of the temperature dependent conductivity [76]. The field-effect mobility of sexithiophene and dihexyl sexithiophene have then been measured as a function of both the temperature and the gate voltage [59]. This was achieved by estimating /LFE in the linear regime, with the help of Eq. (39), instead of the conventional estimation from the variation of the saturation drain current, Eq. (3Sa). It appeared that, although the mobility indeed decreases as the

Figure 14. Energy diagram of the insulator-semiconductor interface in the accumulation regime, in the case of a n-type semiconductor with a single shallow trap level Et. At low gate bias (left), all the traps are empty and any new injected charge is trapped. At higher gate voltages (right), the Fermi level EF crosses the trap level. Traps near the interface are filled, and additional injected charges can move like in a trap free material.

3 12

5 Charge Transport in Semiconducting Oligothiophenes

temperature decrease, it is not really thermally activated, that is, its Arrhenius plot (log&) as a function of 1/T)did not actually give an adequate straight line. Furthermore, the activation energy was found to be strongly gate voltage dependent. These behaviors were rationalized by assuming an exponential distribution of trap, similar to that given by Eq. (32). A very attractive conclusion of this modeling is that the field-effect mobility of dihexyl-sexithiophene, 0.04 cm2V-ls-' at room temperature - the highest value of the oligothiophene series - corresponds to a trap free regime (in other words, it is an intrinsic arameter of the material), whereas that of unsubstituted sexithiophene, 0.003 cm'V-l s-' , is limited by traps, with a trap free limit practically equal to that of the dihexyl substituted molecule. Therefore, it arises that the lower mobility of the unsubstituted molecules (and this is also probably true for the other oligomers) is directly connected to a higher concentration of traps, originating from a poorer structural organization. Interestingly, recent results (see Table 2 above) tend to show that the field effect mobility of oligothiophenes is little dependent on the chain length and substitution (provided the oligomer is substituted at end position), which would confirm that the physical limitation of charge transport is traps. However, the fact that the trap free mobility is still lower than the theoretical limit between localized and delocalised transport, together with its decrease with decreasing temperatures, are clear indications that the ultimate trap free charge transport mechanism is polaronic.

5.4 Concluding remarks Charge transport in organic materials was once studied for itself, mainly because the mechanisms involved differ from those encountered in their inorganic counterpart. In particular, the phenomenon of localization is more pronounced, and organic compounds are good models for polaron theories. Since the late eighties, the possibility of making operative devices with organic semiconductors has emerged. Oligothiophenes have now proved to be among the best candidates for fabricating field effect transistors. The prospect of potential applications has given rise to an outburst of both theoretical and experimental studies on the charge transport of these materials. With the help of the models already developed in the early age of organic semiconductors, some general properties of the charge transport mechanisms in oligothiophenes can now be outlined. (i) There is a general agreement on that charge carriers are radical cations, or polarons. The energy levels of this charge species are now well established, both in the isolated molecule and in the solid state. Moreover, these levels can be theoretically calculated with reasonable agreement with experimental data. Nevertheless, it is not yet clear how these levels are connected to the basic parameters of polaron transport, particularly the polaron binding energy. (ii) While the purity of materials was improving, the initial large scattering of the reported field-effect mobility of oligothiophenes has converged to a chain length independent value of a few 0.01 cm2V-' s-I. Moreover, sexithiophene single crystals do not exhibit a significantly higher mobility than polycrystalline layers

References

3 13

(0.075 cm2V-' s-', against 0.04cm2V-' s-' for the polycristalline dihexyl substituted sexithiophene). This feature can be attributed to the particular operating mode of the thin film field-effect transistor - the enrichment mode - which corresponds to a trap filled limit, where all charge carriers move like in trap free material. Similarly, the low mobility of earlier reports could be attributed to poorly ordered devices, where charge transport is strongly limited by traps. (iii) The trap free mobility presents a positive temperature coefficient (it increases with increasing temperature), which is typical of a polaronic transport. Moreover, a steep increase of the mobility of sexithiophene and of its dihexyl substituted derivative at low temperature has been reported, which was identified as a transition between the hopping and band transports, as predicted by the theory of the small polaron. This behavior is in fact consistent with the nature of charge (self localized cation radicals).

References 1. D. D. Eley, Phthalocyanines as semiconductors, Nature, 1948, 162, 819; D. D. Eley, G. D.

Parfitt, M. J. Perry and D. H. Taysum, Semiconductivity of organic substances I., Trans. Faraday SOC.,1953, 49, 79. 2. D. Kearns and M. Calvin, Photovoltaic effect and photoconductivity in laminated organic systems, J. Chem. Phys., 1958, 29, 950. 3. W. Helfrich and W. G. Schneider, Recombination radiation in anthracene crystals, Phys. Rev. Lett., 1965, 14, 229. 4. F. Ebisawa, T. Kurokawa and S. Nara, Electrical properties of polyacetylene-polysiloxane interface, J. Appl. Phys., 1983, 54, 3255. 5. C. Kittel, Introduction to Solid State Physics, John Wiley, New York, 1976. 6. N. Karl, J. Marktanner, R. Stehle and W. Warta, High-field saturation of charge carrier drift velocities in ultrapurified organic photoconductors, Synth. Met., 1991, 42, 2473. 7. P. W. Anderson, Absence of diffusion in certain random lattice, Phys. Rev., 1958, 109,1492. 8. A. Miller and E. Abrahams, Impurity conduction at low concentrations, Phys. Rev., 1960, 120, 745. 9. H. Overhof, Hopping conductivity in disordered solids, Festkorper., 1976, 14, 239. 10. N. F. Mott and E. A. Davis, Electronic processes in non-crystalline materials, Clarendon Press, Oxford, 1971. 11. D. Emin, Low-temperature AC conductivity of adiabatic small-polaronic hopping in disordered systems, Phys. Rev. B, 1992, 46, 9419. 12. D. Emin, Phonon-assisted jump rate in noncrystalline solids, Phys. Rev. Lett., 1974, 32, 303. 13. D. Emin and T. Holstein, Adiabatic theory of electron in a deformable continuum, Phys. Rev. Lett., 1976, 36, 323. 14. D. Emin, Basic issues of electronic transport in insulating polymers, in Handbook of Conducting Polymers (Ed.: T. A. Skotheim), Marcel Dekker, New York, 1985, Chapter 26. 15. D. M. Pal, Transient photoconductivity in poly(N- vinylcarbazole), J. Chem. Phys., 1970, 52, 2285. 16. W. D. Gill, Drift mobilities in amorphous charge-transfer complexes of trinitrofluorrenone and poly-N-vinylcarbazole, J. Appl. Phys., 1972, 43, 5033. 17. A. K. Jonscher and A. A. Ansari, Photo-currents in silicon monoxide films, Philos. Mag., 1971, 23, 205. 18. J. Yamashita and T. Kurosawa, On electronic current in NiO, J. Phys. Chem. Solids, 1958,5,34. 19. T. Holstein, Studies of polaron motion Part 11. The small polaron, Ann. Phys., 1959, 8, 343.

3 14

5 Charge Transport in Semiconducting Oligothiophenes

20. A. J. Bosman and H. J. van Daal, Small polaron versus band conduction in some transitionmetal oxides, Adv. Phys., 1970, 19, 1. 21. A. L. Shluger and A. M. Stoneham, Small polarons in real crystals: concepts and problems, J. Phys. Condes. Matter, 1993, 5 , 3049. 22. D. Emin and T. Holstein, Studies of small polaron motion. IV: Adiabatic theory of the Hall effect, Ann. Phys., 1969, 145, 645. 23. E. A. Silinsh and V. Capeck, Organic molecular crystals. Interaction, localization and transport phenomena, AIP Press, New York, 1994. 24. E. A. Silinsh and S. Nespurek, On the nature charge-carriers in low mobility solids, Chem. Listy 1996, 90, 43. 25. J. C o r d and J. L. Bredas, Nature of the optical transitions in charged oligothiophenes,Advan. Muter., 1995, 7 , 295;J. Cornil, D. Beljonne and J. L. Bredas, Nature of optical transitions in conjugated oligomers. 2.Theoretical characterization of neutral and doped oligothiophenes, J . Chem. Phys., 1995, 103, 842. 26. C. Ehrendorfer and A. Karpfen, Spatial extension of a bipolaronic defect in oligothiophenes and in polythiophene: A combined semiempirical and ab initio study, J. Phys. Chem., 1994, 98, 7492. 27. A. J. W.Tol, The electronic and geometric structure of dications of oligo-thiophenes, Chem. Phys., 1996, 208, 73. 28. D. Fichou, G. Horowitz and F. Gamier, Polaron and Bipolaron formation on isolated model thiophene oligomers in solution, Synth. Met., 1990,39, 125;D. Fichou, G. Horowitz, B. Xu and F. Garnier, Stoichiometriccontrol of the successive generation of radical-cation and dication of extended alpha-conjugated oligothiophenes: a quantitative model for doped polythiophene, Synth. Met., 1990, 39, 243. 29. G. Horowitz, A. Yassar and H. J. von Bardeleben, ESR and optical spectroscopy evidence for a chain-length dependence of the charged states of thiophene oligomer - Extrapolation to polythiophene, Synth. Met., 1994, 62, 245; B. Nessakh, G. Horowitz, F. Garnier, F. Deloffre, P. Srivastava and A. Yassar, Cyclic voltammetry and differential cyclic voltabsorptometry of soluble oligothiophenes - evidence for a 4-fold charged pi-dimer in duodecithiophene, J. Electroanal. Chem., 1995, 399, 97. 30. S. Hotta and K. Waragai, Alkyl-substituted oligothiophenes - crystallographic and spectroscopic studies of neutral and doped forms, J. Muter. Chem., 1991, I , 835. 31. J. Guay, A. Diaz, R. L. Wu, J. M. Tour and L. H. Dao, Electrooxidation of soluble alpha, alpha-coupled thiophene oligomers, Chem. Muter., 1992, 4 , 254; J. Guay, P. Kasai, A. Diaz, R. L. Wu, J. M. Tour and L. H. Dao, Chain-length dependence of electrochemical and electronic properties of neutral and oxidized soluble alpha,alpha-coupled thiophene oligomers, Chem. Mater., 1992, 4, 1097. 32. J. V. Caspar, V. Ramamurthy and D. R. Corbin, Preparation and spectroscopic characterization of polarons and bipolarons of thiophene oligomers within the channels of pentazil zeolithes: the evolution of organic radical ions into conducting polymers, J. Am. Chem. Soc., 1991, 113, 600. 33. C. H. Evans and J. C. Scaiano, Photochemical generation of radical cations from alphaterthienyl and related thiophenes: kinetic behavior and magnetic field effects on radical ion pairs in micellar solution, J. Am. Chem. SOC.,1990, 112, 2694. 34. V. Wintgens, P. Valat and F. Garnier, Photochemicalgeneration of radical cations from thiophene oligomers, J . Phys. Chem., 1994, 98, 228. 35. B. Zinger, K. R. Mann, M. G. Hill and L. L. Miller, Photochemical formation of oligothiophene cation radicals in acidic solution and nafion, Chem. Muter., 1992, 4 , 1113. 36. D. M.Deleeuw, Stable solutions of doped thophene oligomers, Synth. Met., 1993,57, 3597. 37. M. G. Hill, K. R. Mann, L. L. Miller and J. F. Penneau, Oligothiophene cation radical dimers - An alternative to bipolarons in oxidized polythiophene, J. Amer. Chem. SOC., 1992,114,2728;M. G. Hill, J. F. Penneau, B. Zinger, K. R. Mann and L. L. Miller, Oligothiophene cation radicals - pi-dimers as alternatives to bipolarons in oxidized polythiophenes, Chem. Mater., 1992, 4, 1106. 38. Y. Cao, D.Guo, M. Pang and R. Qian, Studies on iodine doped thiophene oligomers, Synth. Met., 1987, 18, 189.

References

3 15

39. K. Tanaka, Y. Matsuura, Y. Oshima, T. Yamabe and S. Hotta, Electronic properties of p-type doped oligothiophenes, Synth. Met. 1994, 66, 295; K. Waragai, H. Akimichi, S. Hotta, H. Kano and H. Sakaki, Charge transport in thin films of semiconducting oligothiophenes, Phys. Rev. B, 1995, 52, 1786. 40. M. G. Harrison, R. H. Friend, F. Garnier and A. Yassar, A study of the charged excitations in thin-films of alpha-sexithiophene by voltage-modulation spectroscopy and photoimpedance measurements, Molec. Cryst. Liq. Cryst. Sci. Technology Section A - Molec. Cryst. Liq. Cryst., 1994, 252, 165. 41. P. A. Lane, X. Wei, Z. V. Vardeny, J. Poplawski, E. Ehrenfreund, M. lbrahim and A. J. Frank, Spin signature of photoexcitations in sexithiophene, Synth. Met., 1996, 76, 57. 42. E. J. Lous and M. P. Creusen, Electrical characterization of gap states in a highly doped conducting oligomer, Synth. Met., 1996, 76, 233. 43. M. N. Bussac and L. Zuppiroli, Stability of transverse bipolarons in conducting polymers, Phys. Rev. B, 1994,49, 5876. 44. P. G. Le Comber and W. E. Spear, Electronic transport in amorphous silicon films., Phys. Rev. Lett., 1970, 25, 509. 45. S. H. Glarum, Electron mobilities in organic semiconductors, J . Phys. Chem. Solids, 1963, 24, 1577. 46. 0. H. Le Blanc Jr., Hole and electron drift mobilities in anthracene, J. Chem. Phys., 1960, 33, 626. 47. H. Scher and E. W. Montroll, Anomalous transi-time dispersion in amorphous solides., Phys. Rev. B. Solid-State, 1975, 12, 2455. 48. N. F. Mott and R. W. Gurney, Electronic Processes in Ionic Crystals, Clarendon Press, Oxford, 1940. 49. M. A. Lampert and P. Mark, Current Injector in Solids, Academic Press, New York, 1970. 50. P. Delannoy, Dark conductivity, photoconductivity and photovolatic conversion in organic molecular solids, Muter. Sci., 1981, 7 , 13. 51. P. Mark and W. Helfrich, Space-charge-limited currents in organic crystals, J. Appl. Phys., 1962, 33, 205. 52. J. S. Bonham and D. H. Jarvis, A new approach to space- charge-limited conduction theory, Aust. J. Chem., 1977, 30, 705; J. S. Bonham and D. H. Jarvis, Theory of space-charge-limited current with one blocking electrode, Aust. J . Chem., 1978, 31, 2103; J. S. Bonham, Spacecharge-limited current theory for a spatially non-uniform trap distribution, Aust. J. Chem., 1978,31,2117. 53. M. Campos, Space charge limited current in naphthalene single crystals, Mol. Cryst. Liq. Cryst., 1972, 18, 105. 54. P. K. Weimer, The TFT - A new thin-film transistor, Proc. IRE, 1962, 50, 1462. 55. W. E. Spear and P. G. Le Comber, Investigation of the localised state distribution in amorphous Si films, J. Non-Cryst. Solids, 1972, 8-10, 727. 56. D. F. Barbe and C. R. Westgate, Surface state parameters of metal-free phthalocyanine single crystals, J . Phys. Chem. Solids, 1970, 31, 2679. 57. M. Shur, Physics of semiconductor devices, Prentice-Hall, Englewood Cliffs, 1990. 58. S. M. Sze, Physics of Semiconductor Devices, John Wiley, New York, 1981. 59. G. Horowitz, R. Hajlaoui and P. Delannoy, Temperature dependence of the field-effect mobility of sexithiophene. Determination of the density of traps, J . Phys. IZZ France, 1995, 5 , 355. 60. J. W. Sease and L. Zechmeister, Chromatographic and spectral characteristics of some polythienyls, J. Am. Chem. Soc., 1947,69,270; L. Zechmeister and J. W. Sease, A blue-fluorescing compound, terthienyl, isolated from marigolds, J. Am. Chem. SOC.,1947, 69, 273. 61. G. Horowitz, X. Z. Peng, D. Fichou and F. Garnier, Organic thin-film transistors using 7r-conjugated oligomers. Influence of the chain length, J . Mol. Electron., 1991, 7, 85. 62. F. Garnier, A. Yassar, R. Hajlaoui, G. Horowitz, F. Deloffre, B. Servet, S. Ries and P. Alnot, Molecular engineering of organic semiconductors - design of self-assembly properties in conjugated thiophene oligomers, J . Am. Chem. SOC.,1993, 115, 8716. 63. B. Servet, G. Horowitz, S. Ries, 0. Lagorsse et al., Polymorphism and charge transport in vacuum-evaporated sexithiophene films, Chem. Muter., 1994,6, 1809.

3 16

5 Charge Transport in Semiconducting Oligothiophenes

64. G. Horowitz, D. Fichou, X. Z. Peng, Z. G. Xu and F. Gamier, A field-effect transitor based on conjugated alpha-sexithienyl,Solid State Commun.,1989,72,381;G .Horowitz, X. Z. Peng, D. Fichou and F. Garnier, The oligothiophene-based field-effect transistor: How it works and how to improve it, J. Appl. Phys., 1990,67,528. 65. H. Akimichi, K. Waragai, S. Hotta, H. Kano and H. Sakati, Field-effect transistors using alkyl substituted oligothiophenes, Appl. Physl Lett., 1991,58, 1500. 66. J. Paloheimo, P. Kuivalainen, H. Stubb, E. Vuorimaa and P. Yli-Lahti, Molecular fieldeffect transistors using conducting polymer Langmuir-Blodgett films, Appl. Phys. Lett., 1990, 56, 1157. 67. P. Ostoja, S . Guerri, M. Impronta et al., Instability in electrical performance of organic semiconductor devices, Adv. Muter. Optics Electron., 1992,1, 127; P. Ostoja, S. Guerri, s. Rossini et al., Electrical characteristics of field-effect transistors formed with ordered alpha-sexithienyl, Synth. Met. 1993, 54, 447. 68. F. Garnier, R. Hajlaoui, A. Yassar and P. Srivastava, All polymer field-effect transistor realized by printing techniques, Science, 1994, 265, 1684. 69. A. Dodabalapur, L. Torsi and H. E. Katz, Organic transistors: Two-dmensional transport and improved electrical characteristics, Science, 1995,268, 270. 70. H. E. Katz, L. Torsi and A. Dodabalapur, Synthesis, material properties, and transistor performance of highly pure thiophene oligomers, Chem. Muter., 1995, 7, 2235. 71. R. Hajlaoui, G. Horowitz, F. Garnier et al., Improved field-effect mobility in short oligothiophenes: quaterthiophene and quinquethiophene, Adv. Muter., 1997, 9, 389. 72. R. Hajlaoui, D. Fichou, G. Horowitz, B. Nessakh, M. Constant and F. Garnier, Organic Influence of transistors using alpha-octithiophene and alpha,omega-dihexyl-octithiophene: oligomer length versus molecular ordering on mobility, Adv. Muter., 1997, 9, 557. 73. F. Garnier, R. Hajlaous, A. El Kassmi, G. Horowitz, L. Laigre, W. Porziu, M. Armanini and F. Provasoli, Dihexylquaterthiophene, a two dimensional liquid crystal-like organic semiconductor with high transport properties, Chem. Mat. 1998, in press. 74. H. E. Katz, A. Dodabalapur, L. Torsi and D. Elder, Precursor synthesis, coupling, and TFT evaluation of end-substituted thiophene hexamers, Chem. Muter., 1995, 7, 2238. 75. G. Horowitz, F. Gamier, A. Yassar, R. Hajlaoui and F. Kouki, Field-effect transistor made with a sexithiophene single-crystal, Adv. Muter., 1996, 8, 52. 76. D. Fichou, G. Horowitz, Y. Nishikitani and F. Garnier, Conjugated oligomers for molecular electronics: Schottky diodes on vacuum evaporated films of alpha-sexithienyl, Chemtronics, 1988,3,176; D. Fichou, G. Horowitz, Y. Nishikitani, J. Roncali and F. Garnier, Schottkyjunctions based on vacuum evaporated Urns of thiophene oligomers, Synth. Met., 1989,28, C729. 77. G. Horowitz, D. Fichou, X. Z. Peng and P. Delannoy, Evidence for a linear low-voltage spacecharge-limited current in organic thin films. Film thickness and temperature dependence in alpha-conjugated sexithienyl, J. Phys. France, 1990, 51, 1489. 78. K. Waragai, H. Akimichi, S. Hotta, H. Kano and H. Sakaki, FET characteristics of substituted oligothiophenes with a series of polymerization degrees, Synth. Metal., 1993, 57, 4053. 79. L. Torsi, A. Dodabalapur, L. J. Rothberg, A. W. P. Fung and H. E. Katz, Intrinsic transportproperties and performance limits of organic field-effect transistors, Science, 1996,272, 1462. 80. G. Horowitz and P. Delannoy, An analytical model for organic-based thin-film transistors, J. Appl. Phys., 1991, 70, 469.

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes: Relation to Their Use in Electro-optic and Photonic Devices J. Cornil, D. Beljonne, V. Parente, R. Lazzaroni, and J. L. Brkdas

6.1 Introduction The use of organic conjugated systems in new technological applications is a challenging task [l]. Such conjugated compounds are very attractive due to their ease of synthesis, processability, and the tuning of their properties for a given purpose. Since the pioneering works of the seventies essentially dealing with trans-polyacetylene [2, 31, many efforts have been devoted to the characterization of other conjugated chains with enhanced environmental stabilities in both the neutral and charged states. In this context, polythiophene and its substituted derivatives have received particular attention, namely since their electronic properties can be easily modulated upon substitution in the p positions along the conjugated backbone [4]. In the ground state, the electronic and optical properties of polythiophene chains are governed by a filled 7r-electron band (7r-valence band) and an empty 7r*-electron band (r-conduction band). Valence Effective Hamiltonian (VEH) calculations, performed on stereoregular infinite single chains on the basis of ab initio geometric parameters provide a bandgap of about 2.2 eV for the planar conformation while the 7r-bandwidths are calculated to be on the order of 4.5 eV [5]. However, such calculations, based on the use of periodic conditions, do not truly reflect the actual chemical nature of a typical polymer chain. Topological defects breaking the conjugation along the chains are indeed formed in many instances during the chemical synthesis. As a result, the description of a polymer chain is usually more coherent when expressed in terms of a Gaussian distribution of conjugated segments with finite lengths; the average conjugation length is estimated to be around ten repeat units in the case of polythiophene [6]. Due to the molecular nature of most polymer samples, the study of oligomer analogs has recently emerged has a very convenient tool to shed light on the properties of the corresponding polymers [7]. Oligomer compounds offer the advantage of presenting a well-defined chemical structure and higher processability/solubility relative to those of the parent polymers; they are thus candidates of choice for a wide range of experimental spectroscopic investigations. From a theory standpoint, finite-size systems allow for sophisticated theoretical treatments taking electron correlation effects into account. The validity of the oligomer approach is supported

3 18

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

by our calculations performed on conjugated oligomers, which have provided in many instances a detailed interpretation of the experimental data collected on polymers [8]. In the early days, conjugated polymers were mostly investigated for their high electrical conductivity upon chemical or electrochemical doping [I]. More recently, the focus has largely shifted towards the design of electro-optic devices involving the conjugated system as the active element. Among these, organic light-emitting diodes (LEDs) constitute the most promising application, since nearly all the requirements in view of commercialization are being fulfilled [9]. The first electroluminescence signal from a conjugated polymer was reported by the Cambridge group in 1990 following the fabrication of a single layer LED device based on polyparaphenylene vinylene, PPV [lo]. The typical architecture of an LED device is composed of an organic layer sandwiched between two metallic electrodes, i.e., a high workfunction metal such as indium-tin oxide or emeraldine salt [l 11 that serves to inject holes into the valence band of the polymer and a low workfunction metal such as calcium or aluminum to pour electrons into the conduction band, see Fig. 1. The positive and negative charges injected into the polymer matrix relax under the form of polarons and drift in opposite directions; when they meet, they form singlet and triplet polaron-excitons in a 1/3 statistical ratio, which sets the upper bound for intrinsic quantum efficiency to 25%. Note that the polaron-exciton terminology refers to a bound electron-hole pair coupled to a local lattice distortion reminiscent to that observed for a charged polaron [12]. It is the radiative decay of the singlet polaron-excitons that gives rise to light emission; the color is directly dependent on the bandgap of the emitter. It must be emphasized that the luminescence process competes with several nonradiative decay routes, such as intersystem crossing between the singlet and triplet manifolds, multiphonon internal conversion, and trapping mechanisms at quenching sites. Over the past few years, several strategies have been set up to increase significantly the electroluminescence quantum efficiency (the first PPV-based LEDs had

Anode

I Cathode

,1 Figure 1. Schematic representation of a single-layer light-emitting diode polarised under forward bias. The electrons (holes) are injected in the matrix at the interface between the polymer and the low &gh) workfunction metal. The charge carriers then recombine under the form of singlet and triplet excitors and light emission is obtained from the radiative decay of the singlet species.

6.1 Introduction

319

an efficiency on the order of only 0.01%, i.e. 1 emitted photon per 10000 injected electrons). Successful routes consist for instance in confining the polaron-excitons in quantum wells (either by introducing non-conjugated segments along the conjugated backbone [13, 141 or by accumulating the charge carriers at the interface between two organic layers [15]) and/or in matching at best the energy of the frontier levels of the polymer and the Fermi energy of the metallic contacts (in order to balance the injection rate of electrons and holes in the organic matrix and avoid the quenching of polaron-excitons by charged defects [16]). As a result of such improvements, the highest quantum efficiency reported to date stands around 12% [17]. Although much interest has been devoted to the use of PPV and related compounds in light-emitting diodes, polythiophene and its substituted derivatives have also emerged as very attractive candidates to be exploited in the fabrication of such devices. As a matter of fact, it was shown for instance that the color of the emitted light in an LED based on polythiophene could span the whole UV-vis range by simply tuning the bandgap of the polymer via the steric hindrance induced by /?-substitutions on the thiophene rings [18]. A similar result was achieved by controlling the effective conjugation length in the samples through the design of block copolymers [ 191 or regioregular polyalkylthiophene chains [20]. Furthermore, Inganas and collaborators demonstrated that the emitted color of organic diodes involving a blend of various substituted polythiophene chains could be modulated by the strength of the applied driving voltage [21]; other remarkable features reported by the same group were the possibility to produce white light from an LED device incorporating a substituted polybithiophene chain [22] and to make polarized light sources by stretch orientation of the polymers [23]. It is worth mentioning that oligomers, in particular sexithiophene, have also been successfully incorporated as active materials in LEDs [6, 24-26], as well as in a wide range of devices such as field-effecttransistors (FETs) [27,28] and optical modulators [29,30]. In the search for new routes towards the development of electro-optic devices with enhanced efficiencies, the methods of quantum-chemistry and solid-state physics can play a significant role by providing an in-depth understanding of the fundamental mechanisms governing the operation of the devices. To illustrate this aspect, we review in this chapter the results of recent theoretical calculations on oligothiophenes, performed by means of sophisticated treatments that allow to apprehend reliably their intrinsic geometric, electronic, and optical properties and the relation to their use in electro-optic devices. Most of the theoretical data presented in this contribution address the properties of isolated oligomers; however, it is important to stress that we are currently developing tools to take explicit account of the effects of intermolecular interactions, which can sometimes alter drastically the properties of a single molecule [31, 321. After a brief methodology section, the third section focusses on the characterization of the lowest singlet and triplet excited states of oligothiophenes; we describe their relative energy positions and the nature of the lattice distortions taking place in the excited state upon photoexcitation. We put special emphasis on the lowest-lying one-photon allowed excited state (S1) since it plays a pivotal role in light-emitting diodes; this state indeed acts simultaneously as the source of

320

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

light emission upon radiative decay as well as the origin of intersystem crossing processes towards the triplet manifold that contribute to a decrease in fluorescence yield and can lead to the appearance of a weak phosphorescence signal [33]. We also discuss the impact of substitution of the conjugated backbone with 7r-acceptor and 7r-donor groups on the relaxation phenomena in the excited states and on the rate of the main nonradiative decay route; we also show that such substitutions are useful to modulate the positions of the ./r-frontierlevels and hence, to match these at best with the Fermi energy of metals used in the fabrication of LEDs. In section 6.4, we investigate the changes occurring in the geometric, electronic, and optical properties of oligothiophenes when charged species are generated upon chemical doping (following a redox process), charge injection (as in a lightemitting diode), electrochemical doping (for instance in a light electrochemical cell [34]), or photoexcitation. We provide a unified picture of the charged defects created in conjugated oligomers and polymers and their associated optical signature. This discussion allows us to critically analyze the traditional assignment of the optical transitions observed in the spectra of polymers supporting charged defects and to question the relative stabilities of the different charged species. Section 6.5 is dedicated to a theoretical characterization of the nature of the interfaces between a conjugated polymer and a metal as typically encountered within the architecture of an organic light-emitting diode. In the present case, we pay particular attention to the interface formed between model compounds of polythiophene and aluminum, which is a metal widely used as the LED cathode due to its high environmental stability. It is shown that covalent bonds are created between the conjugated backbone and the aluminum atoms, in contrast to the situation observed with calcium atoms for which charge transfer processes take place [35]; the formation of these covalent bonds is accompanied by the appearance of new active vibrational modes, as seen from theoretical simulations of vibrational spectra. Many efforts also currently focus on the gain of a better understanding of the optical nonlinearities in conjugated systems [36-401. The fast and intense nonlinear responses of organic compounds make them very exciting candidates for the field of photonics, i.e. for an all-optical treatment of information. We discuss in the last section the calculated frequency dispersion of the third-order optical nonlinearities in oligothiophenes. We also describe the nature of the essential states contributing to the nonlinear response and analyze the chain-size evolution of the calculated hyperpolarizabilities.

6.2 Theoretical methodology Several quantum-chemical methods with varying degrees of sophistication are used in our theoretical investigations. We usually optimize the geometry of oligothiophenes by means of the Hartree-Fock semiempirical Austin Model 1 (AM1) [41]or Modified Neglect of Differential Overlap (MNDO) [42] methods, that allow one to deal

6.2 Theoretical methodology

32 1

explicitly with rather large systems, i.e., up to twelve thiophene rings for the results presented in the next section. In our case, the MNDO approach is used with neutral oligomers due to its ability to provide C-C bond lengths in good agreement with X-ray data [43]; the AM1 Hamiltonian is chosen in the study of charged oligomers (note that singly charged systems are treated within the Restricted Open Shell Hartree-Fock (ROHF) formalism) and in situations where the knowledge of the torsion angles between adjacent thiophene rings is of relevance. We also refer to ab initio calculations performed at the 3-21G* level for a detailed analysis of the geometry of short oligothiophenes and substituted derivatives. The geometry optimizations related to the polymer/metal interfaces presented in section 5, are carried out by means of the Density Functional Theory (DFT) methodology [44] (using the DMol package [45-461) with a numerical basis set of split-valence quality including polarization functions on the heavy atoms; in all cases, we use the local density approximation scheme and the Volko-Wilk-Nusair expression for the exchange-correlation term [47]. On the basis of the equilibrium geometries, the vibrational frequencies are calculated as the second derivative of the total energy with respect to the atomic coordinates while the intensities are estimated from the square of the first derivative of the dipole moment versus the atomic coordinates (the vibrational spectra thus relate to infrared absorption data). The charge distribution are calculated via a Mulliken population analysis. A proper characterization of the excited states is essential when dealing with electro-optic or photonic aspects. It requires to go beyond the one-electron picture provided by the Hartree-Fock calculations. To do so, we couple the semiempirical Hartree-Fock Intermediate Neglect of Differential Overlap (INDO) method [48] to a configuration interaction scheme. Only singly excited configurations (SC1) are involved in the CI expansion to simulate the optical absorption spectra of neutral molecules, i.e. to deal with the lowest B, excited states; the number of configurations is chosen in a way ensuring the absence of significantchanges in the calculated properties when increasing the size of the CI basis. In the case of singly and doubly charged oligomers. the singlet transition energies and their relative intensities are often estimated with the help of the Valence Effective Hamiltonian (VEH) approach [49]; this method indeed proves to be well-adapted to reproduce the location of the defect levels within the gap (in each case, the use of such an one-electron picture is initially validated by CI calculations ascertaining that the optical transitions of interest in the spectra of the charged compounds are dominated by a single dominant configuration [SO]). In order to obtain an overall description of the lowest singlet and triplet excited states in oligothiophenes, further sophistication of the theoretical methods is required. We then make use of the INDO Hamiltonian coupled to a MultiReference Double-Configuration Interaction (MRD-CI) scheme [51], in which the CI expansion is built with configurations corresponding to single and double excitations with respect to a restricted number of reference determinants (including the SelfConsistent Field (SCF) determinant itself and configurations built by promoting one or two electrons from the highest occupied levels to the lowest unoccupied levels [52]). In this way, triple and quadruple excitations are effectively taken into account at a cost largely inferior to that of a conventional quadruple CI calculation.

322

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

6.3 Electronic and linear optical properties of neutral oligothiophenes 6 3.1 Nature of the lowest excited states Within CZhsymmetry, the lowest singlet excited state S1 belongs to the B, representation and is symmetry allowed from the ground state (1 Ag). In short oligothiophenes, in contrast to the case of polyenes, the lowest two-photon excited state (2Ag) is located above the 1 B, state [53, 541, as has been confirmed by high level ab initio CASSCF/CASPT2 calculations on the thiophene dimer and trimer 1551. Within the INDO/MRD-CI approach, the evolution with chain length can be properly reproduced only when including an increasing number of electronic levels with the number of repeat units [52]. In that case, we calculate a decrease in the 1B,-2Ag energy difference when the chain grows, in good agreement with experiment (the 1 BU-2A, energy separation has been estimated to be on the order of 1 eV in bithiophene [53] and 0.2eV in the hexamer [54]. In Fig. 2, we compare to the measured So -+ S, transition energies, the INDO/MRD-CI values calculated on the basis of the MNDO-optimized geometries [43]. The theoretical results compare very well with the experimental values extracted from optical absorption measurements in solution [56-591, as shown in Fig. 2. Note that a complete theoretical treatment of absorption in solution would require taking account of the possible salvation

5’0 4.5

I 5

4.0

5

3.5

P

3.0

6 4

2 2.5 E

l-

2.0

1.5

0.0

0.1

0.3

0.2

0.4

0.5

1 /n

Figure 2. Evolution of the INDO/MRD-CI-calculated So+ S and So -.+TItransition energies (full circles) as a function of the inverse number of thiophene units (l/n).We also plot the corresponding experimental data (open squares) extracted from Refs. [56, 571 and [59] for the singlet-singlet transition and from Ref. [62] for the singlet-triplet excitation.

6.3 Electronic and linear optical properties of neutral oligothiophenes

323

effects; such an approach is, however, beyond our current capabilities. A linear dependence, typical of conjugated compounds, appears between the So --t S1 transition energies and the inverse number of repeat units, see Fig. 2; this significant redshift of the lowest electronic transition as the chain grows, is related to the progressive extension of the .rr-delocalized system. Figure 2 also includes the energy difference between the ground state SOand the lowest triplet excited state T1,as calculated at the INDO/MRD-CI level on the basis of the So equilibrium geometry. We find the evolution with chain size of the SO-+ T1 transition energies to be much slower than for the So -+ S1 excitation: the singlettriplet excitation is only lowered by -0.2eV when going from the dimer to the hexamer while a bathochromic shift of -1.4 eV is observed for the singlet-singlet transition. Such behavior actually reflects the stronger confinement of the triplet exciton with respect to the singlet; this trend is consistent with Optically-Detected Magnetic Resonance (ODMR) data for polythiophene that indicate that the T1 triplet state hardly extends over more than a single thiophene unit [60]. It should be stressed, however, that the spin-spin separation parameter, deduced from ODMR measurements on a series of oligothiophenes slightly raises with chain length [61], a feature that has been associated to an increase in the spatial extent of TI;this experimental observation has been confirmed by recent theoretical calculations, showing a distribution of the spin density over at most four thiophene rings in longer oligomers [611 (although the inner two rings receive, by far, the largest contribution). Optical absorption spectra measured in a solvent containing heavy atoms have located the position of the TI state in terthiophene [62]; in this case, a non vanishing intensity is observed for the So + TI transition due to spin-orbit coupling related to ‘heavy-atom effects’ induced by the solvent; this leads to the appearance of a broad and weak absorption band around 1.71eV, in very good agreement with the 1.68eV calculated value. Janssen and co-workers have also recently reported that addition of C60to solutions of oligothiophenes ranging in size from the hexamer to the undecamer results in a quenching of the triplet state of the oligothiophenes to produce a C60triplet state via energy transfer [57]; the T1state in these oligothiophenes is therefore expected to lie between the energies of the C60triplet state and of the TI state in the trimer, which are estimated at 1.57 eV [63] and 1.71 eV [62], respectively. Finally, we mention that Xu and Holdcroft have detected a phosphorescence peak at -1.5 eV in polythiophene [33], a value close to that obtained by extrapolating the So t TI excitation energies at the scale of an infinite polymer chain. Singlet excitons formed upon photoexcitation or electron-hole recombination can decay nonradiatively via intersystem crossings to lead to the creation of metastable triplet excitons; such species give rise to a single, intense and long-lived triplet-triplet T1+ T, electronic transition seen in photoinduced absorption experiments on oligothiophenes in solution [57, 641. These observations are consistent with the results of MRD-CI calculations where we find the lowest-energy triplet state to be strongly coupled to a single higher-lying triplet excited state whose wavefunction corresponds to a complex mixing of singly and doubly excited configurations. The evolution with chain length of the theoretical T1-+ T, transition energies are reported in Fig. 3 together with the experimental values derived from photoinduced absorption experiments. Despite a systematic overestimation of the theoretical values, a similar

324

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

5.0

4.0

z-

6 3.0 m

m

Bp 2.0 e

k-

1.0

0.0 0.0

0.1

0.2

0.3

0.4

0.5

1In

Figure 3. Evolution of the INDO/MRD-CI-calculated (full circles) and experimental(open squares) TI 4 T, transition energies as a function of the inverse number of thiophene units (]in).The expenmental results are taken from Ref. [57] .

chain-length dependence of the calculated and experimental transitions is obtained and points to a significant red-shift of the TI+ T, transition as the chain grows. It thus appears that the higher-lying triplet state T, has a less confined wavefunction than the TI state, which is consistent with its smaller binding energy.

6.3.2 Intersystem crossing processes As emphasized in the introductory section, the emission of light in electroluminescent devices competes with numerous nonradiative decay routes [65]; these processes include interchain effects related for instance to the formation of excimers, quenching of the singlet excitons by extrinsic or conformational defects as well as by low-lying two-photon states [66], singlet fission into two triplets [67], or intersystem crossing (ISC) from the singlet to the triplet manifolds [68]. Note that the possibility of quenching of the excitons by a lower-lying two-photon state is ruled out in oligothiophenes up to the hexamer since both experimental [54] and theoretical [69] studies have located the 2A state above the 1B state. Time-resolved fluorescence measurements on unsubstituted thiophene oligomers in solution indicate a sharp increase of the fluorescence quantum yield q& when the number of thiophene units is increased from two to seven [56, 701; in such experiments, we expect the migration of the excitons towards trapping centers to be minimized due to the finite size of the systems and the absence of interchain effects. The evolution of q 4 ~with chain size has been related to a decrease in nonradiative decay rate kNR, since the radiative decay constant kR is observed to be almost unaffected

6.3 Electronic and linear optical properties of neutral oligothiophenes

325

when going from one oligomer to the next [70]. Among the various nonradiative processes, the singlet-to-triplet intersystem crossing has been found to provide the R most significant contribution to ~ N [70]. We have tried to provide a coherent picture of the ISC processes in oligothiophenes in order to rationalize the trends observed experimentally [43]. As first suggested by Rossi et al. in the case of terthiophene, kNRcan be expressed as a sum of two contributions, kl and k2:

""sc> kT

Here, kl includes various nonactivated nonradiative phenomena while the second term corresponds to an activated intersystem crossing process. The amplitude of k2 depends both on the spin-orbit interaction through the pre-exponential A2 factor and on the singlet-triplet energy difference through the AErsc value. Although an accurate description of the ISC processes would require to take the spin-orbit coupling interaction into account, the strong evolution with chain length that is seen, is unlikely to be based on significant modifications in the strength of this coupling. We thus relate the observed decrease with chain length in the k N R decay rate to the evolution of the energy difference AEIsc between the singlet and triplet states involved in the crossing. The S1-Tl energy differences calculated at the INDO/MRD-CI level in oligothiophenes are much too large to lead to an efficient singlet-triplet overlap, and hence to a significant probability of intersystem crossing. However, the calculations indicate that one higher-lying triplet excited state (T4) lies within the same energy range as S1. We plot in Fig. 4 the evolution of the So -+ S1and So + T4 excitation energies versus 4.5

4.0

5 Y

x

e t ," 3.5 0 ..+A

In

F

I-

3.0

2.5 0.1

0.2

0.3

0.4

0.5

1 /n

Figure 4. Evolution of the INDO/MRD-CI-calculated SO --+ S1 (open squares) and SO--+ T4 (full circles) excitation energies as a function of the inverse number of thiophene rings (l/n).

326

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

the inverse number of thiophene rings. Starting in bithiophene from a situation where the triplet T4 is located below the singlet S1, we observe a reversal in the ordering of these two states as the chain size is increased; the crossing between the S1 and T4 states occurs for a chain length corresponding to the trimer. Note that the position of the T4 excited state is overestimated for the trimer due to the absence in the calculations of spin-coupling interactions (these are expected to mix efficiently the singlet and triplet wavefunctions). The trends derived from the consideration of the T4 triplet excited state as the essential state leading to intersystem crossing in oligothiophenes, are consistent with the experimental measurements. In bithiophene, the fact that the T4 state lies below the S1 state gives rise to a nonactived and very efficient ISC process, and hence to a very low fluorescence yield. The activation energy, however, increases when the chain elongates making the probability of ISC processes lower and the fluorescence yield & substantially higher.

6.3.3 Lattice relaxation phenomena The MNDO equilibrium geometries in the S1 and T1 states of the dimer, trimer, and tetramer are reported in Table 1, together with the ground-state geometry and the Table 1. C-C and C-S bond lengths (in A) in: (i) the ground state So;the lowest singlet excited state S1; and (iii) the lowest triplet excited state TI of the dimer (Th2), trimer (Th3) and tetramer (Th4) of oligothiophenes, as optimized at the MNDO level. We include the MNDO-calculated relaxation energies (Ere,, in eV) with respect to vertical excitations. The atoms are labelled according to the Figure; we prevent any redundancies by taking explicit account of the symmetry of the systems. s2 C

6

7

SO

s1

TI Th3

Bond

Th2

Th3

Th4

Th2

Th3

Th4

Th2

1-2 2-3 3-4 I-s1 4-S1 4-5 5-6 6-7 5-S2 7-8 8-S2 8-9

1.374 1.447 1.388 1.674 1.694 1.447

1.374 1.447 1.388 1.674 1.694 1.447 1.388 1.442 1.689

1.380 1.447 1.388 1.674 1.694 1.447 1.388 1.441 1.689 1.388 1.689 1.446

1.396 1.421 1.427 1.671 1.701 1.415

1.387 1.431 1.414 1.670 1.692 1.420 1.424 1.410 1.696

1397 1.408 1.457 1.677 1.699 1.382

1.382 1.434 1.416 1.673 1.692 1.405 1.457 1.376 1.697

0.25

0.15

1.384 1.437 1.405 1.670 1.690 1.425 1.421 1.406 1.692 1.428 1.696 1.415 0.16

0.39

0.34

Ere1

Th4 1.378 1.441 1.397 1.670 1.689 1.431 1.429 1.394 1.687 1.451 1.699 1.383 0.33

6.3 Electronic and linear optical properties of neutral oligothiophenes 2

3

10

6

321

11

7

15

14

.oa .06 .04 0 ._ I

-E

.02

I

5

0.00

m -

+ 2 -.02 -.04 -.06

-.oa 2

4

6

a

10

12

14

16

Site

Figure 5. Evolution with carbon site position i, of the bond length alternation in quaterthiophene, Ar (calculated at the MNDO level as the difference between the lengths of the (i, i 1) and (i, i - 1) carbon-carbon bonds), in the So (solid line), S, (dashed line), and TI (dotted line) states. Atoms 1 and 16 correspond to the extremities of the molecule, as shown in the molecular sketch.

+

MNDO/CI-calculated relaxation energies with respect to vertical transitions from the ground state. The evolution with site position i of the bond-length alternation (defined as the difference between the lengths of the (i, i 1) and (i, i - 1) C-C bonds) is illustrated in Fig. 5 for the So, S1 and TI states of the tetramer. Analysis of the geometry deformations occurring in the S1 state reveals lattice distortions that are small and characterized by the appearance of a quinoid character withip the central rings; the C-C bond length alternati?n is found to drop to 0.04-0.05 A in the external rings and to a value close to 0.02 A in the central part of the system. Such geometry changes are reminiscent of those associated to the creation of a soliton-antisoliton pair in polyacetylene [7 11, except that the electron-hole pairs in oligothiophenes are bound due to both the Coulombic attraction between the charge carriers and the nondegenerate ground state nature of polythiophene. On the basis of the MNDO-calculated bond length deformations and additional theoretical calculations [43], we estimate the singlet exciton to extend over 3-4 repeat units.

+

328

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

In contrast, much more pronounced lattice distortions are observed for the lowest triplet excited state, as illustrated by the amplitude of the relaxation energy that is almost twice as big as in the singlet state (0.33 eV vs 0.16 eV) in the tetramer. The formation of a (bound) soliton-antisoliton pair clearly emerges when looking at the geometric deformations along the chain axis; indeed, in the tetramer, going from the end towards the center of the chain, the C-C bond length alternation first decreases (as a consequence of the shortening of the single bonds and elongation of the double bonds), then vanishes at the connection between the first and second rings (the C-C bond lengths are there equal) then reverses sign and peaks in absolute value at the center of the oligomer (where the maximum value of the bond-length alternation is recovered). Due to the exchange potential term, a stronger confinement is thus obtained for the triplet with respect to the singlet; the results indicate that 1 or 2 thiophene rings are needed for a proper accommodation of the triplet defect, as also suggested by ODMR data [60]. The knowledge of the relative locations of the lowest singlet and triplet excited states is a valuable information for the design of materials where nonradiative decay processes such as intersystem crossing are prevented at best. The theoretical insight we are gaining into the excited-state electronic properties of polythiophene and corresponding oligomers can thus prove useful to achieve improved efficiencies in LED devices.

6.3.4 Effects of substitution We now illustrate the impact of cyano functionalization on the electronic properties of oligothiophenes. This issue is of prime interest in the context of organic lightemitting diodes, as demonstrated before in the case of PPV [72-741; the goal is to uncover the substitution patterns that lead to devices with enhanced efficiencies. In the following, we focus on cyano-substituted bi- and terthiophene derivatives that have been synthesized and characterized both electrochemically and optically by the Thiais group, see Fig. 6 [75]. On the basis of the optimized geometry provided by ab initio 3-21G* calculations, we have first calculated the INDO locations of the HOMO and LUMO levels of substituted coplanar oligomers, see Table 2; note that planarity of the conjugated backbone is expected to occur in the solid state due to paclung effects [76-791. These calculations demonstrate that substitution with acceptor groups leads to an overall stabilization of the frontiers levels, whose amplitude depends on the amount and position of the cyano groups; note that this stabilization is asymmetric in the sense that the energy of the LUMO level is more affected than that of the HOMO level. We also note that the shifts of the frontier levels are much larger in the disubstituted compounds. Despite the fairly good agreement observed between the theoretical data and the electrochemical measurements (see Table 2), a detailed analysis of the theoretical results indicates that calculations performed on planar conformations fail to rationalize simultaneously the experimental evolutions of the reduction and oxidation

6.3 Electronic and linear optical properties of neutral oligothiophenes B1

329

82

B3

T1

T3

'CN

m'

Figure 6. Chemical structure of the cyano-substituted bi and terthiophene series. The AMIcalculated dihedral torsion angles are indicated above each inter-ring bond.

330

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

Table 2. Energy (in eV) of the HOMO and LUMO levels, lowest transition energy (E*, in eV) and related oscillator strength (OS, in arbitrary units) of the coplanar cyano-substituted oligothiophenes in a planar conformation, as provided by INDOjSCI calculations. We present in parentheses the INDO shifts of the frontier levels (with respect to a reference compound in each group) that are compared to the corresponding experimental values appearing in bold [80]. Compound

HOMO

LUMO

B2 B3 B4 TI T2 T3 T4 T5 T6 T7

-7.69 -7.94 -7.92 -7.02 -7.25 -7.19 -7.27 -7.47 -7.36 -7.48

-0.86 -1.30 -1.27 -0.61 -1.03 -0.99 -0.99 -1.35 -1.27 -1.27

(-0.25/-0.32) (-0.23) (-0.23/-0.19) (-0.17/-0.20) (-0.25/-0.27) (-0.45/-0.41) (-0.34/-0.42) (-0.46/-0.55)

(-0.44-0.36) (-0.41/-0.32) (-0.42/-0.43) (-0.38/-0.39) (-0.38/-0.39) (-0.74/-0.65) (-0.66/-0.65) (-0.66/-0.65)

3.59 (0.86) 3.48 (1.04) 3.45 (0.67) 3.26 (1.04) 3.17 (1.25) 3.12 (0.98) 3.22 (0.96) 3.10 (1.44) 3.03 (0.96) 3.15 (0.89)

potentials among the series of oligomers [75].The oxidation potential is found to increase with: (i) the lowering of the oligomer chain length; (ii) the number of substituted sites; and (iii) a substitution on a central unit compared to a terminal one. The reduction potential is characterized by a similar behavior, although the evolution with chain size and with the position of the cyano groups is much less marked. The electrochemical gap decreases as the number of thiophene subunits increases, and is larger with central rather than terminal substitution. The fact that the oxidation potential is larger when the cyano group occupies a more central position in the trimer (as observed in the T5 + T6 -+ T7 series, see Fig. 6 for notation) is not accounted for in the theoretical data. However, we observe that the relative positions of the frontier orbitals are well reproduced when addressing separately the group of derivatives T3, T4, T6 and T7, where the cyano groups are located in the inner part of the molecule, and the group of derivatives T2 and T5 where the substituents are found on the outer carbons; furthermore, we note that a larger discrepancy between theory and experiment is obtained for derivatives T6 and T7, which have two substituents at the @-position of the thiophene rings. This suggests that the substituted molecules of a given group are characterized by similar torsion angles between adjacent thiophene units in solution and that the absolute values of these angles differ within the two groups. Since it is rather difficult to treat accurately the solvent effects, we have adopted a simple approach to better apprehend the relation between a given substitution pattern and the amplitude of the torsion angles between adjacent thiophene rings. It consists in optimizing the torsion angles with the help of the semiempirical Hartree-Fock Austin Model 1 1411 calculations performed within a rigid-rotor approximation, i.e. by keeping the bond lengths and bond angles frozen at the ab initio values [80] (the choice of the AM1 formalism is validated in the case of bithiophene by the close agreement obtained between the AM1 -optimized torsion angle and the values obtained following ab initio and density functional calculations, around 35" [Sl])

6.3 Electronic and linear optical properties of neutral oligothiophenes

33 1

The AMl-geometry optimizations on the isolated molecules show that the torsion angle between two adjacent thiophene units is on the order of 35" in situations where the rings are unsubstituted or the cyano groups are connected in a positions, see Fig. 6. In contrast, the theoretical results establish that derivatives with cyano groups in p positions are characterized by larger dihedral angles between the substituted ring and the adjacent interacting unit. Since the measurements are carried out in solution, it appears necessary to deal explicitly with twisted conformations to rationalize the experimental data. We have therefore calculated the INDO positions of the frontier levels in the trimer derivatives within a simple model: we assume that the torsion angle between unsubstituted rings in weakly polar solvents is as large as that calculated for the isolated molecule (-35"); this is further validated by the results of recent self-consistent reaction field (SCRF) calculations taking explicit account of the dielectric constant of the medium during the optimization procedure [82]. A similar value is chosen in the case of thiophene units substituted by cyano groups in the a position. We have then determined the optimal value of the torsion angle between an unsubstituted ring and a ring involving a cyano substituent in the p position, in order to match at best the experimental evolution of the oxidation potential; we find this value to be close to 45". In this context, where we use just two different values for the torsion angles, we obtain an excellent agreement between the theoretical and experimental evolutions of the oxidation potentials, see Table 3 and Fig. 7a; this emphasizes the need for torsion angles to be considered in our model. Note that the ionization potential values are governed by the position and amount of substituents and by the amplitude of the torsion angles, which all lead to a stabilization of the HOMO level. In contrast, the position of the LUMO level results from a compromise between the stabilization induced by the cyano substituents and the destabilization associated to larger torsion angles. Despite the fact that the quantitative agreement is less satisfactory for the LUMO level evolution vs. reduction potential data, both theory and experiment indicate that the stabilization of the LUMO level is stronger for the monosubstituted terthiophenes when the cyano group is at the terminal position, see Fig. 7b; although a similar behavior is expected for the disubstituted compounds

Table 3. Energy (in eV) of the HOMO and LUMO levels, lowest transition energy (E,,, in eV) and related oscillator strength (OS, in arbitrary units) of the trimers with torsion angles defined by our model, as provided by INDOjSCI calculations. We present in parentheses the INDO shifts of the frontier levels (with respect to compound T1) that are compared to the corresponding experimental values appearing in bold [SO].

Compound

HOMO

LUMO

Etr

T1 T2 T3 T4 T5 T6 T7

-7.17 -7.40 -7.37 -7.55 -7.61 -7.63 -7.73

-0.39 -0.84 -0.72 -0.68 -1.14 -0.94 -1.00

3.50 (0.95) 3.03 (1.17) 3.41 (0.85) 3.58 (0.79) 3.32 (1.31) 3.41 (0.82) 3.48 (0.72)

(-0.23/-0.19) (-0.20/-0.20) (-0.38/-0.27) (-0.44/-0.41) (-0.46/-0.42) (-0.56/-0.55)

(-0.45/-0.43) (-0.33/-0.39) (-0.29/-0.39) (-0.75,'-0.65) (-0.55/-0.65) (-0.61/-0.65)

(0.S.)

332

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

1

2

3

5

4

7

6

Compound

-0.3

2.1

-0.4

2.0

-0.5

1.9

-0.6

-

-0.7

4

-0.8

1.a U

1.7

0

1.6 -0.9

-1 .o

1.5

-1.1

1.4 1.3

-1.2 1

2

3

4

5

6

7

Compound

Figure 7. Measured (open symbols) and INDO-calculated (filled symbols) positions of the: (a) HOMO and (b) LUMO levels in the terthiophene series. The experimental data correspond to the (inverse) oxidation and reduction potentials, respectively.

on the basis of the calculations, this is not the case experimentally as the reduction potential is observed to be nearly independent from the substitution pattern. As a crucial point for electron transporting materials is the position of the LUMO level, this discrepancy requires further studies in order to determine the most favorable position of the substituents.

6.4 Electronic and linear optical properties of charged oligothiophenes

333

The changes occurring in the lowest transition energy and related intensity calculated within the same model are fully consistent with the experimental data [SO]. Note that significant changes in the emitted color is expected upon cyano functionalization since a red-shift (with respect to the planar unsubstituted trimer) of the lowest transition energy as high as 0.2eV is expected in the case of derivative T6 in its planar conformation. We can thus conclude from this discussion that the substitution of the conjugated backbone by 7r-acceptor (and 7r-donor by extension) substituents is a convenient tool to modulate the position of the frontier levels of the active element used in LEDs and to sweep the color of the emitted light over the whole UV-vis range. Derivatization affects in a specific way the position of the singlet and triplet excited states, with important consequences on the intersystem crossing rate [43].

6.4 Electronic and linear optical properties of charged oligothiophenes In this section, our goal is to investigate the optical properties of charged oligothiophenes and to establish whether they are consistent with those of the corresponding polymers; we have chosen to focus on the properties of sexithiophene, which is one of the most promising materials for molecular electronics [83], and on those of the twelve-ring oligomer that better depicts the longer conjugated segments encountered in polymer samples. Upon doping (i.e. oxidation or reduction), charges removed from or injected into the conjugated backbone give rise to the appearance of spatially localized geometric defects, as a result of the strong electron-phonon coupling characteristic of conjugated chains [84]. In terms of condensed-matter physics, such charges coupled to a local lattice distortion of the backbone are described as positive (negative) polarons or bipolarons upon single or double oxidation (reduction) processes; note that, in the following, we will also make use of the chemical terminology by referring to radical-cations and di-cations (anions) for singly and doubly oxidized (reduced) chains, respectively. The formation of polarons and bipolarons also induces a strong modification of the electronic structure of the polymer: two new localized one-electron levels appear within the original gap, as shown in Fig. 8. In the case of oxidation (reduction), the lowest (highest) polaron level is singly occupied (which leads to a magnetic signature) whereas the bipolaron levels are both empty (occupied). According to the one-electron band-structure model developed by Fesser et al. [85], three new subgap optical transitions are expected following the formation of polarons while only two are predicted in the presence of bipolarons (see Fig. 8). It is worth stressing that this model appears to be consistent with most of the experimental absorption reported to date on doped polymers [86, 871. We now consider the theoretical results obtained on the sexithiophene molecule. The AM1-optimized geometries of the doubly charged oligomers show that the

334

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

a

b

C

Figure 8. Sketch of the one-electron band-structure model for polythiophene: (a) in the neutral state; (b) in the presence of a positively charged polaron; and (c) in the presence of a positively charged bipolaron. The possible subgap optical transitions induced upon doping are represented by arrows.

defect (bipolaron) is localized at the center of the molecule and is characterized by a reversal of the single and double character of the C-C bonds; the formation of such a bipolaron thus induces the appearance of a strong quinoid character within the rings [88]. On the other hand, the geometries of the radical-cations exhibit weaker structural deformations; as the AM1 results provide C-C bond lengths intermediate between those obtained for the neutral and doubly oxidized systems, the formation of polarons leads to the appearance of a semiquinoid character along the chain [88]. We further notice that the amplitude of the geometric deformations are different upon p and n doping, due to a breakdown of the electron-hole symmetry induced by the incorporation of heteroatoms within the conjugated backbone, and are found to be weaker in the reduction case. As the size of the oligomer increases, the amplitude of the geometric deformations is found to diminish when going from the center to the end of the molecule. We describe in Fig. 9 the typical evolution of the C-C bond lengths upon single and double oxidation and reduction; this plot indicates that bipolarons extend over nine repeat units, as also suggested by recent theoretical calculations [89], while a weaker spatial extension of five rings is expected for polarons. However, these estimates have to be considered as upper limits due to the fact that the influence of counter-ions is neglected in the present calculations. As mentioned above, the geometric relaxations taking place upon oxidation or reduction are accompanied by a modification of the one-electron structure of the oligomers: two molecular orbitals move inside the original gap to give rise to new subgap features in the optical absorption spectra [84]. In the case of radicalcations, VEH calculations show the appearance of two new subgap features that, in ascending order, originate from an electron transition between the HOMO level and the lower polaronic level (H-POL1) and between the two polaron levels (POL1-+ POL2). Two optical transitions (POL2 -+ L and POL1 POL2) with slightly different energies are also calculated for the radical-anion, as observed experimentally [90]. We report in Table 4 the transition energies and intensities -+

6.4 Electronic and linear optical properties of charged oligothiophenes

i

+Positive polaron

0.06

0.04

335

-? Negative polaron

5

P

2

0.02

:

r)

2

0.00

C .-

v1

g -0.02 .c m 0 U

-0.04

-0.06 -10

-8

-6 -4

(4

2

0

-2

4

6

8

10

Bond number

0.10 -0-

Positive bipolaron -V-Negative bipolaron

0.08 0.06 fp 0.04 0 -

l

8

0.02

0.00

C ._

3 -0.02 P 2 -0.04 u -0.06 t

V

I

I

(

W

U

-0.08 -0.10~, I -10 -8

(b)

,

1

I

-6

4

-2

, , 0

2

I

I

I

,

4

6

8

10

Bond number

Figure 9. AMI-calculated changes in the C-C bond length of sexithiophene upon: (a) single oxidation and reduction; and (b) double oxidation and reduction. The bond labeled 0 is located in the middle of the oligomer.

obtained on both the positively and negatively charged systems. The VEHcalculated transitions of oxidized sexithiophene are in very good agreement with absorption data measured upon oxidation [58, 59, 91, 921 as well as with photoinduced absorption [93, 941 and voltage-modulation spectroscopy data [95].

336

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

Table 4. VEH-calculated polaron and bipolaron transition energies and related oscillator strengths (in arbitrary units) of the sexithiophene molecule upon single and double reduction and oxidation. Nature of the transition

Transition energy (eV)

Oscillator strength

Singly oxidized H + POLl POL1 -t POL2

0.93 1.48

1.34

Singly reduced POL2 -+ L POL1 + POL2

0.76 1.41

1.17 1.16

Doubly oxidized H -t BIPl

1.10

1.48

Doubly reduced BIP2 + L

0.97

1.31

1.18

The absence of any transition between the HOMO level and the upper polaronic level in singly-oxidized sexithiophene (or between the lower polaronic level and the LUMO level in the radical-anion) is related to the selection rules imposed by the symmetry of the oligomers. Taking into account CZhsymmetry, this transition is forbidden since the two levels that are involved belong to the same irreducible representation (H and POL2 have a, symmetry whereas POLl and L are characterized by b, symmetry) 1961. On the other hand, electron transitions from a, to b, levels (or vice versa) give rise to excitations that are polarized within the molecular plane and thus present significant intensities, such as those observed in the spectra of charged oligothiophenes. We note that a third symmetry-allowed transition is predicted by our calculations [88]; this weak feature is mainly described by electron transitions between the lower polaronic level and the LUMO 1 level (POL1 L 1) and between the HOMO-1 level and the upper polaronic level (H-1 +POL2). This third absorption peak is calculated to lie within the same energy range as the first excitation of the neutral system (it thus tends to be overshadowed by the presence of any remaining neutral molecules). The formation of bipolarons leads to the appearance of a single subgap absorption peak that originates from an electron transition between the HOMO level and the lower bipolaronic level in doubly-oxidized sexithiophene (and between the upper bipolaronic level and the LUMO level in the dianion). The symmetry considerations invoked above result in a vanishing intensity for the transition between the HOMO level and the upper bipolaronic level in the dication (between the lower bipolaronic level and the LUMO level in the negatively charged compound). There is once again excellent agreement between the theoretical values and the spectroscopic data [58, 59, 91, 921. We display in Fig. 10 the spectra of sexithiophene in various oxidation states simulated on the basis of VEH-calculated transition energies and intensities estimated from the INDO/SCI formalism. (It should be noted that the intensities of the two polaronic transitions, as obtained in the framework of the VEH one-electron picture, are significantly higher than those provided at the correlated level [88]).

+

+

6.4 Electronic and linear optical properties of charged oligothiophenes

337

100 90

80 70 .-In ..

Fa

60

$p

50

22

-

40

30 20 10

0 0.5

1.o

1.5

2.0

2.5

Energy (eV)

Figure 10. VEH-INDO/SCI calculated absorption spectra of sexithiophene in the neutral (solid line), singly oxidized (dotted line), and doubly oxidized (dashed line) states. The spectra are simulated by means of Gaussian functions whose full width at half maximum is set to 0.2 eV.

These theoretical results allow us to rationalize a wide range of experimental observations. However, an additional type of charged species, referred to as r-dimers, has been recently isolated; the r-dimers correspond to complexes formed upon interaction of two polaron-carrying oligomers [97, 981. Such defects, which are spinless, are invoked when two subgap absorption features are observed in the optical absorption spectra of lightly doped oligomers where no paramagnetic signal is detected, thus excluding the formation of polarons; the differentiation with respect to isolated bipolarons is straightforward since the latter would lead to a single subgap feature. The fact that a single subgap absorption feature can be observed in the optical absorption spectra of doubly charged oligomers is in marked contrast to the typical optical properties of doped conjugated polymers where the formation of charged bipolarons is traditionally assumed: two intense subgap peaks are then observed. A first reason that can be invoked to remove this discrepancy is that the symmetry rules governing the nature of the optical transitions in the oligomers do not have to hold true at the scale of a long disordered polymer chain; however, calculations where we impose the bipolarons to be localized near one end of the chain reveal that such a symmetry breakdown in the relaxation of the defects does not give any significant changes in the aspect of the optical spectra [88]. As a consequence, the spectra of polymers supporting isolated bipolarons should be characterized by a single dominant subgap feature at low doping level. Another explanation relies on the fact that the lineshape of the spectra is strongly affected as soon as interaction between the charged defects takes place [88]; this could then lead to the appearance of two intense subgap features such as those observed in the spectra

338

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

of conjugated polymers at high doping level [86]. On the other hand, Furukawa and collaborators have assigned the optical transitions in the spectra of doped polymers by assuming that charged polarons are the dominant species generated upon doping of long conjugated chains [99, 1001. To get a better understanding, it is of interest to address the optical absorption spectra of a doubly oxidized twelve-ring oligomer reported by Garnier and co-workers [loll. In this case, it is established that two charges are removed per chain, two strong subgap features are observed in the spectrum, and no paramagnetic signal is detected. We can therefore argue that: (i) the generated species are not isolated bipolarons (as initially assumed [loll) since a single subgap feature would then be observed, as indicated by our theoretical calculations; (ii) the formation of two interacting bipolarons is incompatible with the level of doping reached in the experiment; (iii) a four-fold oxidized entity corresponding to a double r-dimer could be created, as suggested by photovoltammetry measurements [1021. The latter point is, however, inconsistent with very recent mass spectrometry measurements that allow for the unambiguous detection of such complexes [103]. The most reasonable conclusion is that the two subgap features likely originate from the formation of two polarons. This then has the important consequence that, in the experimental conditions of Ref. 101, two polarons are more stable than a single bipolaron on the twelve-ring oligomer (this is also suggested by recent theoretical calculations [104, 1051). Another consequence is that bipolarons might be observed in shorter conjugated segments, such as in sexithiophene, because the polarons are then forced to interact strongly. However, there remains an unresolved issue: the absence of magnetic signal from the two polarons has still to be understood and further experimental studies are thus required. The absence of a magnetic signal has been so far explained by a model based on a weak interaction between the two polarons, leading to a splitting of the polaron levels in the gap, and hence to a closed shell configuration [1001. These considerations illustrate the very fine balance that exists, in the case of doubly-oxidized compounds, between the energetics of formation of two polarons (whose wavefunctions can interact) and that of a bipolaron [106]. We have also to keep in mind that it is very difficult to model accurately the two situations at the quantum-chemical level since not only are highly correlated calculations requested but explicit consideration of the medium (solvent) effects is also required. The signature of isolated defects should in principle be detected in photoinduced absorption experiments, since the concentration of the photogenerated charged species is expected to be very weak. The two long-lived subgap absorption features observed in the polymer photoinduced absorption spectra have been originally assigned to bipolaron transitions [ 1071. Thls interpretation is in contrast to the results of our calculations on isolated bipolarons. This discrepancy can be explained in several ways: (i) either 7r-dimers are formed; (ii) or long-lived bipolarons might be trapped in more ordered regions where they could interact; (iii) or the two subgap features observed in the photoinduced spectra correspond to the optical signature of polarons, as suggested by recent optical modulation experiments [1081. The analysis presented in this Section is widely applicable to the nondegenerate ground-state conjugated systems with some degree of symmetry in the geometric

6.5 Characterization of metallpolymer interfaces

339

structure. The same trends prevail when looking at the absorption spectra of oxidized or reduced oligopyrroles [1091, oligophenylenes [ 1lo], and oligo(pheny1eneviny1ene)s [ 1 1I].

6.5 Characterization of metal/polymer interfaces As presented in the Introduction, one major aspect of conjugated polymer-based LEDs is the presence of interfaces between the metal electrodes and the active polymer layer. In particular, the electron-injecting contact is made of a metal with a low workfunction such as aluminum. As this metal is known to be reactive, it is important to determine to which extent the phenomena occurring at the interface with the polymer can affect the geometric and electronic structure of the 7r-conjugatedsystem and how the interface interactions can influence the operation of the device. Theoretical studies based on Hartree-Fock quantum-chemical calculations [ 1 12- 1141have shown that aluminum tends to interact strongly with the conjugated system, through the formation of Al-C covalent bonds. This reaction leads to dramatic changes in the electron distribution and in the n-electron conjugation along the polymer chain. In good agreement with those theoretical results, photoelectron spectroscopy experiments [114-1161 carried out during the initial stages of interface formation point to the appearance of new carbon species with higher electron density (as expected upon A1 bonding) and to major modifications to the uppermost occupied 7r-levels. We have recently completed this study by performing a series of density functional theory (DFT)-based calculations on molecular model systems for the aluminum/ polythiophene interface. We have chosen the DFT approach for two reasons. First, with respect to the Hartree-Fock methodology we had used in earlier works [l 12-1 141, the DFT approach has the advantage of taking electron correlation into account, which is essential when dealing with metal atoms. We have found that DFT geometry optimizations of aluminum/polyene molecular models can lead to bonding configurations different from those obtained at the Hartree-Fock level [117, 1181; the DFT geometries (obtained within the local spin density approximation, LSDA) for those systems are actually very similar to those obtained at a postHartree-Fock level [ 1181 (such as second-order Moller-Plesset perturbation theory, MP2), which is a clear indication of the better quality of the DFT geometries (at least for this kind of organo-metallic systems). The methodology used here has provided geometries of polyenes [118] and oligothiophenes [119] in very good agreement with previous theoretical and experimental data. The second reason for using the DFT method is related to the simulation of the vibrational spectra of the species appearing at the metal/polymer interface. The DFT techniques provide accurate vibrational frequencies, with no need for rescaling; in this work, the maximum error is on the order of f 3 % between theory and experiment for polyenes [ 1 181. Note also that for such vibrational calculations, a high-quality geometry optimization is required in order to determine properly the energy minimum and to derive reliable values for the frequencies.

340

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

The Al/polythiophene interface has been modeled by having one or two A1 atoms interact with thiophene or terthiophene. Since it is known [112-116, 1201 that the interaction between A1 atoms and the organic molecule is very localized, the complex with the thiophene molecule (the basic building block of a polythiophene chain) can provide important information, at a high theoretical level, about the interface phenomena. Geometry optimizations have been performed on different starting configurations of the metal atoms around the molecules, in order to determine the most stable configurations for the complexes between the A1 atoms and the thiophene oligomers. We have analyzed the charge density redistribution occurring upon A1 bonding in order to assess the consequences of the interaction on the electronic structure of the oligomers and to interpret the photoelectron spectroscopy data. Finally, we have calculated the vibrational frequencies of the molecular complexes, to determine the expected vibrational signature of the A1 species thought to be present at the Al-polythiophene interfaces. Compared to the results of photoelectron spectroscopy, which are very sensitive to changes in charge distribution and electronic structure, we believe that the examination of the vibrational properties of such complexes offers a more direct probe to the actual chemical structure at the interface. In recent works [118, 1201, we have described the evolution of the vibrational spectrum calculated for a polyene molecule, octatetraene, upon bonding of two A1 atoms, in order to model the Al/polyacetylene interface formation. These theoretical results indicate that important changes can be expected in the experimental infrared spectrum as a consequence of: (i) the formation of A1-C covalent bonds; and (ii) strong modifications in the bond pattern along the chain.

6.5.1 Geometric structures The most stable configurations of the Al/thiophene complexes are presented in Fig. 11; the geometric parameters of these complexes are reported in Table 5, together with those of the thiophene molecule. When one aluminum atom interacts with thiophene, two configurations have been found (Figs 1l a and b): either the metal atom lies approximately above the center of the ring, or it remains located above a carbon atom in the alpha position relative to the sulfur atom (Ca). The energy difference between these two complexes is very small; it is evaluated to be only 4.4kcal/mol at the LSDA level and decreases to only 1.4 kcal/mol when nonlocal corrections are included in the DFT calculations [121]. In the case of the thiophene molecule interacting with two A1 atoms (Fig. llc), each metal atom forms a single A1-C bond on C a positions. In these three complexes, we observedhe formation of A1-C covalent bonds: the Al-C bond lengths lie in the 2.10-2.32 A range, mainly with the Ca of the thiophene ring. To our knowledge, the only experimental data available for such b?nds, obtained on trimethylaluminum and its dimer, indicate values around 2.0 A for the A1-C bond length [122]. Even in the complex of Fig. l l a , where the aluminum atom is located almost above the center of the ring, the A1-Ca bonds are significantly shorter than the Al-CP distances. A1 bonding to thiophene drives the

6.5 Characterization of metallpolymer interfaces

341

b

a

Figure 11. DFT-optimized chemical structure of the Allthiophene complexes with: a single aluminum atom above (a) the center of the ring, and (b) an 01 carbon atom; (c) two aluminum atoms located above different a carbon atoms.

C

carbon involved towards sp3 hybridization; this is clearly shown by the fact that the sulfur atom in complex 1l a and the hydrogen atoms bonded to the reacting Ca in complexes I Ib and 1lc have'drifted away from the thiophene molecular plane by about 20°, 35", and 50°, respectively. Moreover, as reported in Table 5, the interaction Table 5. DFT-optimized geometric parameters (in phene complexes presented in Fig. 11.

A) of the thiophene molecule and the Allthio-

Thiophene

Complex 1l a

Complex 11b

Complex I lcb

cff-s

1.711 1.711

1.765 1.765

1.785" 1.752

1.834 1.835

Ca-cp

1.370 1.370

1.421 1.421

1,433" 1.375

1.445 1.445

CP-CP

1.412

1.390

1.400

1.365

2.313 2.313

2.157a 3.159

2.08612.086" 3.50013.504

2.373 2.372

2.451a 2.965

2.161/2.163" 3.40113.406

-

2.844

2.716

2.63312.633

-

-

-

2.897

A1-S AI-A1

" For atoms closest to the metal atoms. bThe values on the left-hand side refer to one aluminum atom and those on the right-hand side, to

342

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

with the metal atoms leads to a reversal in the single/double bond character of the carbon-carbon bonds in the thiophene ring. In complexes 1 l a and 1lc, the CP-CP bond length decreases to a value typical of a double carbon-carbon bond (C=C) while in pristine thiophene, it possesses a single-bond character (C-C); concomitantly, the Ca-CP bonds switch from a double to a single carbon-carbon bond character. In complex llb, the evolution is somewhat different due to the loss of symmetry. The Ca-CP bond closer to the A1 atom presents a single C-C bond character (the other one remaining unaffected) while the CP-CP bond adopts a length intermediate between single- and double-bond character. Finally, it is worth mentioning that the Al-+l distance in complex l l c is calculated to be 2.87 8, a value close to the 2.70 A experimental value reported for the aluminum dimer [123]. In order to investigate the effect of longer conjugation on the interaction, we have also investigated Al/terthiophene complexes; the results are illustrated in Fig. 12. When considering two A1 atoms, two possible configurations are obtained. In system 12a, the A1 atoms locate in a symmetric position around the central ring; each metal atom interacts through the formation of covalent bonds with two Ca of two adjacent rings (C(2), C(7) and C(5), C(12)) and, to a lesser extent, with one CP of the terminal ring (C(8) and C(13)). In the other configuration (12b), only two rings are strongly affected by theocomplexation,the third one remaining unaffected. The shortest Al-C bond (2.26A) involves two beta carbons (C(3) and C(8)), the Al-Ca(2) and Al-Ca(7) distances being of 2.44 and 2.48 respectively. System 12b is more stable at the LSDA level than complex 12a by 16.3kcal/mol; this higher stability likely comes from a shorter A1-A1 distance: 2.78 in system 12b

A,

A

a

b Figure 12. DFT-optimized chemical structure of the Al/terthiophene complexes with two aluminum atoms located above: (a) different inter-ring bonds; and (b) the same inter-ring bond.

6.5 Characterization of metal/polymer interjaces

343

A

vs. 3.02 in system 12a. In both cases, sp3 rehybridization appears due to the complexation, as shown by a slight loss of planarity of the organic molecule: the central sulfur atom in system 12a and the hydrogen atoms bonded to Cp(3) and Cp(8) in system 12b leave the molecular plane. As in the thiophene case, the T-conjugated system is strongly affected by the metal deposition. The switch in single/double bond character observed in complex 12a is similar to that of system 1lc and extends to the Ca-CP bonds of the peripheric rings involved in the A1 bonding. As a consequence, conjugation is lost in the C( 13)-C( 12)-C(5)-C(4) and C(3)-C(2)-C(7)C(8) segments. A similar evolution is observed in system 12b, but only in one half of the molecule. It is informative to compare these DFT results with the previous Hartree-Fock ab inito results obtained on the Al*/terthiophene complex [113]. We find that the evolution of the geometry is basically the same, namely formation of new A1-C bonds, sp3 rehybridization of the carbon atoms involved in the complexation, and a strong perturbation in the 7r-conjugated system with a switch in single/double bond character. However, Hartree-Fock predicts the formation of only one A1-Ca covalent bond per aluminum atom; morepver, no interaction is found between the two aluminum atoms (Al-Al=3.91A). In contrast, the DFT approach leads to a multiple bonding pattern with formation of three new bonds between the aluminum atom and Ca, CP, or the other A1 atom; this bonding scheme appears to be more sound from the chemical point of view. This multiple Al-C bonding character maintains the central part of the molecule almost planar, preventing a strong distortion of the terthiophene molecule as obtained with the Hartree-Fock methodology [ 1131.

6.5.2 Electronic structure In simple chemical terms, the Al-C bonds can be considered as formed by one valence electron of aluminum and one C2p ( T ) electron formally belonging to the a carbon. In the complexes, the HOMO is a combination of the LUMO of the pristine thiophene oligomer with the A13p atomic orbitals. We note that the A13s atomic orbitals interact with other T molecular orbitals of the thiophene units to form occupied molecular orbitals of the considered configuration. As a consequence, the electron system is severely perturbed by A1 bonding, which translates into the large geometric changes described above. Significant modifications in the charge density distribution of the conjugated system, as calculated with a Mulliken population analysis, are also induced by the formation of the A1-C bonds. In the complexes with the thiophene molecule, a small charge transfer is observed from the aluminum atom towards the organic molecule (0.03 lei, 0.07 lei and 0.13 lei per A1 atom in complexes l l a , l l b and 1 lc, respectively). Moreover, a reorganization of the electronic distribution in the thiophene unit also occurs, with an important increase of the electronic population on the Ca, (only the one involved in the A1 bonding in system 1lb): -0.16 [el, -0.26 [el, and -0.341el for systems 1la, 1l b and 1lc, respectively. This charge distribution is

344

6 Geometric and Electronic Structure and Optical Response of Oligo-and Polythiophenes

consistent with the photoelectron spectroscopy data, that point to the appearance of a new carbon species with higher electronic density in the first stage of the interface formation [116]. The change in the electronic density on the sulfur atom is vanishingly small. However, the presence of a second metal layer is expected to lead to an increase in the electronic population on the sulfur atom [121].Jndeed, a second A1 atom placed above the first ope in complex 1l a (@-A1 = 2.74 A) induces a shortening of the Al-C bond (2.11 A with C a and 2.22 A with Cp), stronger charge transfer from the closest A1 (0.12 [el),larger increase of the electronic population on the C a (-0.19 [el),and a slight increase of the sulfur electronic population. T h s shows the possible influence of aluminum adlayers. For system 12a, the charge transfer is evaluated to be 0.08 [el from each A1 atom. In this case, both Ca(2) and Ca(5) of the central ring and Cp(8) and Cp(13) of the peripheric rings gain a significant amount of charge (more than -0.1 [el).Note that Ca(7) and Ca(12), which are involved in the A1 bonding, and the sulfur atoms remain almost unaffected. In complex 12b, the changes in electron density are mainly localized on the two thiophene units involved in the A1 bonding, with the larger increase (-0.22 [el)occurring on the Cp(3) and Cp(8) carbon atoms involved in the bonding. The DFT calculations thus indicate that the interface formation leads to a small charge transfer from the metal to the organic system. In all cases, electron density increases on the carbon atoms involved in the A1 bonding, in good agreement with photoelectron spectroscopy data [114-1161. The evolution of the electronic densities on the sulfur atoms is more difficult to rationalize, the changes being very small and not always in the same direction.

6.5.3 Vibrational signature We now focus on the consequences of the Al-C bonding upon the vibrational properties of the conjugated molecules. Figure 13 shows the calculated infrared absorption spectra of the thiophene complexes presented in Fig. 11 together with the spectrum of the pristine thiophene molecule. The spectra of both the pristine terthiophene molecule and system 12a are presented in Fig. 14; only the most important features are discussed below. The spectrum of thiophene is dominated by a band at 670 cm-' corresponding to out-of-plane bending of the C-H bonds; in the present context, other important features are the bands typical of C=C stretching around 1500cm-' and the C-S stretching vibration around 850 cm-' . The in-plane C-H bending and the C-C stretching modes lie in the 1000-1250~m-~region. Finally, the peaks related to the stretching of the C-H bonds are located above 3000cm-' (for the sake of clarity, they are not shown on the spectra). Dramatic differences appear upon A1 bonding to thiophene; the frequency region below ~OOCII-' is now populated by a series of strong peaks, whatever the considered bonding configuration. These peaks correspond to C- Al bending and stretching vibrations, appearing below 200 cm-' and between 300 and 450 cm-'

6.5 Characterization of metallpolymer interfaces

345

Al,/thiophene: complex c

Alkhiophene: complex b

I

I

0

200

Allthiophene: complex a

400

600

800

1000 1200 1400 1600 1800

Wnve numbers (cm")

Figure 13. DFT-calculated infrared absorption spectra of the thiophene moIecule and its complexes with aluminum.

(312cm-' in system l l a , 386cm-' in system Ilb, and 343 and 458cm-' in system 1lc, respectively). The remarkable intensity of these peaks is related to the strong dipole moment of the Al/C bonds. At the actual interface, the corresponding dipole moment may be attenuated, for instance by the presence of small A1 clusters instead of single A1 atoms. Nevertheless, the appearance of vibrational bands in the 200-500 cm-' spectral region would constitute a clear signature of the formation of covalent A1-C bonds. As a consequence of the geometric changes induced in the thiophene ring by the A1 bonding, its vibrational frequencies are also significantly modified. In complex 1 la, the strong out-of-plane C-H bending is split into two new bands located at 700 and 740cm-I; they are related to the C-H bonds on the a and ,B carbons. The symmetric C-S stretching mode is shifted down by -70 cm-' . The region of the C-C conjugated backbone stretching is also deeply affected: the peak appearing above 1500 cm-' (antisymmetric stretching of both Ca=CP bonds) in the spectrum of thiophene is shifted downwards at 1338cm-' , which is probably due to the change in the single/double bond character of the Ca=CP bonds. The changes relative to the out-of-plane C-H bending and C-S

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

a -

A1,lterthiophene: complex a

Terthiophene

0

I

I

200

400

I

1.

.

600

800

1000

1200

1400

1600

1800

Wave numbers (an-')

Figure 14. DFT-calculated infrared absorption spectra of the terthiophene molecule and its complex with aluminum (complex 12a).

stretching are similar in complex 1lc. No peak is observed above 1350cm-' in this A12-thiophenecomplex; in fact, the CP=CP double bond stretching mode is located at 1535cm-' but it is not active in infrared. The very weak peak at 1315cm-' arises from the stretching of the Ca-CP bonds, which are of single bond character after the interaction. The vibrational spectrum of complex 1l b is more complex due to the loss of symmetry. Several vibrational modes appear in the region between 600 and 900 cm-'; they present both C-H bending and C-S stretching characteristics. The peaks located at 1164 and at 1432cm-' are both related to the stretching of the Ca-CP bonds, the first one characteristic of a single C-C bond (coming from the interaction) and the second one characteristic of a double C-C bond (unaffected by the interaction). The stretching of the Cp-CP bond, which is intermediate between a single and a double C-C bond, appears at 1402m-'.Note that in systems 1l a and 1lc, the relative intensities of the modes located above 1000cm-' are very low, which has to be related to the small change in the dipole moment inherent to those vibrations. The spectrum of terthiophene (Fig. 14) is obviously more complex than that of thophene; nevertheless, its major characteristics are similar: (i) there are very few bands below 500 cm-', with the most intense one located at 457 cm-'; (ii) the strongest peaks, located between 600 and 850 cm-', correspond to the out-of-plane C-H bending and the C-S stretching modes; and (iii) the conjugated C-C stretching modes are easily distinguished between 1300 and 1600cm-' . As it clearly appears from Fig. 14, the terthiophene spectrum is strongly affected by the complexation, the interaction being characterized by numerous new peaks. We again observe the

6.6 Nonlinear optical properties of neutral oligothiophenes

347

appearance of the new modes related to the A1-C bonds below 500 cm-' ; in particular, strong A1-C stretching vibrations are located around 300 cm-*. The evolution above 1300cm-' has to be related to C-C stretching modes, localized at different frequencies according to their bond order. The upper last peaks (1575 and 1608cm-I) represent the vibrations of the remaining, now isolated, C=C double bonds: Ca( 14)=C,B(1 9 , Ccr(9)=C,B( 10) and C/3(3)=C,B(4). The three modes located around 1530cm-' involve mainly the inter-ring C-C bond coupled together with the vibrations of other bonds along the carbon backbone. Two new peaks also appear at 1424 and 1369cm-', corresponding to the stretching of the bonds involving the carbons linked to the A1 atoms and which tend to reach a single-bond character. The C-S stretching, standing around 750 cm-' in the pristine molecule, is shifted downwards by over 50 cm-' . Finally, the spectrum now exhibits a series of peaks in the 500-650 cm-' range, related to ring deformation and out-ofplane C-H bending. It thus appears that interaction with A1 should produce a clear vibrational signature in the case of the Al/polythiophene interface. In particular the formation of A1-C covalent bonds should be evidenced following the appearance of new peaks between 300 and 500 cm-' . We stress that preliminary infrared absorption measurements on aluminum/sexithiophene systems have revealed new bands around 360 cm-' [124], a feature which is consistent with the theoretical results presented in this section. To conclude, the DFT results confirm that aluminum and polythiophene oligomers interact preferentially through the formation of covalent bonds between A1 atoms and carbon atoms of the thiophene rings. In all cases, the interaction induces strong geometric and electronic modifications; the .rr-conjugated system is strongly perturbed and may even be disrupted. The formation of the A1-C bonds and the geometric changes induced in the organic molecules lead to important changes in the vibrational spectra, which should allow the experimental detection of the vibrational signatures of the species formed at the Al-polythiophene interface. Finally, it is interesting to note that the same kind of theoretical investigation, but this time modeling the deposition of a thiophene molecule on an aluminum surface, indicates that similar interactions take place, i.e. the formation of A1-C covalent bonds [125].

6.6 Nonlinear optical properties of neutral oligothiophenes Due to their structural simplicity, polyenes and polyacetylene have been the focus of many experimental and theoretical studies dealing with the nonlinear optical response of conjugated systems [ 126-1 351. Experimentally, third-harmonic generation (THG) measurements in polyacetylene indicate the appearance of two distinct peaks attributed to two- and three-photon resonances [127, 1281. Calculations carried out on short polyenes by Garito and co-workers in the framework of a Complete Neglect of Differential Overlap/Configuration Interaction (CNDO/CI)

348

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

approach, have shown very large transition moments between the lA, state (the ground state) and the lB, state (the first one-photon allowed excited state) and between the lB, state and a higher-lying A, state (denoted the mA, state) 11311. The two peaks observed in the THG spectrum of polyacetylene have thus been often attributed to resonances to the lB, and mA, states. (Note that, in the following, for the sake of simplicity, the excited states are denoted by the symbols A and B corresponding to A, and B, [A, and B2] within C2h [C2,] symmetry). In the case of ,&carotene and alkyl-substituted polythiophenes (PT), very good fits to frequency-dependent THG measurements have been obtained by Torruellas et al. when considering a four-state model including (in order of increasing energy) the 1A 2A, lB, and mA states [136]. The inclusion of the 2A state marginally improves the fits over those obtained with just the three other excited states. However, in the case of the 4BCMU polydiacetylene, the agreement with the experimental THG spectrum afforded by the four-state model was poor [136]. Spurred on by this discrepancy, Mazumdar and co-workers have reconsidered the nature of the important states needed to describe the third-order nonlinearity in conjugated polymers; on the basis of calculations performed on short polyenes, they have proposed an 'essential-state' model where the relevant states are: the 1A ground state, the 1B lowest one-photon excited state, a set of mA states (which in long chains are predicted to appear close to the 1B state), and a set of higher-lying nB states (that are said to correspond to the bottom of the conduction band) [137-1391. The essential-state model predicts that two intense three-photon resonances and a rather weak two-photon resonance should appear in the THG dispersion curve of conjugated polymers [ 138, 1391. Many theoretical studies have also focused on the scaling of the cubic nonlinearity, y, with the extent of the 7r-system [52, 140-1471. In those studies, they values are usually expressed as a function of a power, z, of the chain length: y cx N'") where N corresponds to the number of repeat units in the chain. From simple tightbinding calculations, Flytzanis and co-workers found a power value equal to 5 [140]. Actually, the power value z ( N ) is itself dependent on the chain length and tends to one when a purely additive regime is entered, i.e. when saturation is reached. In the case of polyenes, a number of theoretical calculations concur to indicate that the saturation sets in after about 50-60 carbons [131, 148, 1491. Thienpont et al. have reported the observation of saturation with chain length of the third-order polarizability measured on a series of conjugated oligothiophenes [150]. They observed a strong chain-length dependence of y with N up to N 7, followed by a much weaker one, more or less linear for longer chains. Prasad and coworkers have also reported third-order susceptibility measurements for oligothiophenes ranging in length from one to six thiophene rings. In that range of oligomer size, the evolution of y with the number of repeat units shows a supralinear dependence [151]. More recently, saturation of the cubic response with effective conjugation length has been reported by Bjornholm et al. from THG measurements on regio-regular and regio-irregular poly(3-dodecylthiophenes) [ 1521.

-

6.6 Nonlinear optical properties of neutral oligothiophenes

349

In this section, we first analyze the chain-length dependence of the static thirdorder polarizabilities in oligothiophenes ranging in size from two to eight rings. We then examine the dynamic response and simulate the two-photon absorption (TPA) spectrum and the third-harmonic generation (THG) dispersion curve of a long thiophene oligomer, Th7, the seven-ring oligomer. For that purpose, we apply the Sum-Over-States formalism [ 1531 on the basis of the description of the excited states provided by the INDO/MRD-CI scheme (on the basis of A M l optimized geometries). By considering a power expansion for the Stark energy, the ijkl component of the cubic polarwability tensor writes: yijkl(-Wa; w l , w27 w3)

where P(i,j , k, 1; -wu; w1,w2, w3)is a permutation operator defined in such a way that for any permutation of (i J, k, l ) , an equivalent permutation of (-wc; w l ,w2,w3) is made simultaneously; w, = w1 + w2 + w3 is the polarization response frequency; w l , w2, w3 indicate the frequencies of the perturbing radiation fields (wl = w2 = w3 = w in the case of the THG process); i, j , k, and 1 correspond to the Cartesian axes x,y , and z; rn, n, and p denote the INDO/MRD-CI excited states and 0,the ground state; pi is the i component of the dipole operator; w,, is the transition is the damping factor associated to frequency between the m and o states; and rmo excited state m.The average value of y is defined as: (7) = 1/5[Yxxxx+ "iyvvv

+

"izzzz

+ 2("ix*yr + yxxzz + r,,zz)l

(4)

For both two-photon absorption (TPA)and third-harmonic generation (THG), we consider a damping factor of 0.1 eV for all the resonances.

6.6.1 Chain length dependence of the third-order polarizabilities in thiophene oligomers In Table 6 , we present the T~~~~ chain-axis components and the (y) average values of the third-order polarizabilities in thiophene oligomers ranging in size from 1 to 8 monomer repeat units. The evolution of ( y )with chain length, illustrated in Fig. 15, can be divided into three parts [52]:

(i) the y response first picks up significantly with extension of the chain length; the power value z ( N )defined in Eq. 2 strongly increases;

350

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

(ii) for a number of thiophene rings between 3 and 6-7, the power value stays nearly constant; in these first two parts of the curve, polarization and delocalization of the 7r-electron cloud over the whole chain plays an essential role; from N = 1 to N = 7, the average power value ((~((y)))) has been calculated by a least square fit to be 3.9; (iii) finally, the power value starts decreasing for N equal to 7, which indicates the beginning of the saturation regime. Table 6. INDOIMRD-CI SOS average and chain-axis component values (in esu) of the static third-order polarizability of thiophene oligomers. We also indicate the evolution of the power law and yxxxx0: Nz(yxxxx)) together with the dependence of y as a function of chain length ((7) c( NZ((?)) average values of the exponent calculated from N = 1 to N = 7 ((~((y))) and (z(yxxxx))). Thx

(4

Z((Y))

Thl Th2 Th3 Th4 Th5 Th6 Th7 Th8

0.39 3.11 20.14 76.98 181.0 349.8 520.6 658.5

2.99 4.61 4.66 3.83 3.61 2.58 1.76

I

(Z((Y)))

TXXXX

z(,Yxxxx)

1.43 19.93 106.9 386.8 912.2 1750. 2610. 3293.

I

= 3.9

3.80 4.14 4.47 3.84 3.58 2.59 1.74 (Z(-Yxxx*N

= 4.0

~

'm/ 80

P

t 0 1

2

3

5

4

6

7

0

N

Figure 15. INDO/MRD-CI/SOS chain-length dependence of the average third-order polarizability, ( 7 ) .in oligothiophenes.

6.6 Nonlinear optical properties of neutral oligothiophenes

35 1

The evolution with chain length of the chain axis component yxxxxis very similar to that of the average value; the contributions arising from electronic excitations along the chain axis actually dominate the third-order response. These results are in very good agreement with the available experimental data. The DFWM (Degenerate Four Wave Mixing) measurements of Zhao et al. [151] recorded at 602 nm (i.e. off-resonance) show a power dependence of y with respect to N , N varying from 1 to 6 , characterized by a power value that amounts to 4.05. Thienpont et al. [150], who measured the (y) values by the EFISHG (Electric Field Induced Second Harmonic Generation) technique at a wavelength of 1064nm, found a power exponent of 4.6. The larger value of the exponent measured experimentally in that case can be explained by the fact that part of the increase in (7)with increasing chain length is due to an increasing resonance enhancement as the second harmonic photon energy approaches the band gap. In accord to the data of Thienpont and co-workers, we observe the beginning of the saturation regime of (y) for N = 7, a chain length after which the band gap becomes nearly constant. In fact, recent THG measurements (near resonance) on regio-regular and regio-irregular poly(dodecy1thiophenes) indicate that complete saturation of the NLO response is only achieved after about 15-20 thiophene units [152]. In Table 7, we list the chain-axis components of the main electronic dipole transition moments among the states of Th7. In long conjugated chains, the excited states get close to one another in energy and start forming 7r-bands of A or B symmetry (states of the same symmetry, belonging to the same band, then share the overall oscillator strength related to that band). In the following, we therefore associate to the term ‘excited state’, a band in which excited states of the same symmetry are characterized by a similar description of the wavefunction (i.e. similar CI expansion coefficients) and are separated in energy by less than 0.2 eV. For instance, what is referred to in Table 7 as the nlB excited state, in fact corresponds to ‘band’ formed by two states (the 5B and 6B states) separated by 0.05 eV and highly coupled to the 2A state. Using that convention, we can draw the following conclusions from Table 7 [52]: (i) Only the first one-photon allowed excited state (1B state) has a large transition dipole moment with the ground state (1A state). (ii) The 1B state is strongly coupled to different states of A symmetry, which are by order of decreasing coupling; the 6A, 5A, and 2A states. The 6A state (hereafter Table 7. INDO/MRD-CI chain-axis components of the transition dipole moments (in D) between the most important B and A states of Th7. The state energies (in eV) are given between parentheses.

1B (2.46) nl(= 5)B (4.56)

nl(=6)B (4.61) n2(= 12)B (5.81) n2(= 13)B (5.95) n3(= 16)B (6.22) n3(= 17)B (6.30)

1A (0.0)

ml(= 2)A (2.91)

m2(= 5)A (4.05)

m3(= 6)A (4.43)

- 15.9

6.29 -11.2 -18.8 2.19 3.64 -1.53 3.33

-11.7 2.21 4.52 8.54 18.8 -2.22 4.45

24.0 -7.82 -12.1 -4.69 -5.99 10.8 -25.2

-0.83 -0.36 -0.63 -0.25 0.35 -1.12

352

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

denoted m3A) corresponds to the dominant two-photon state which has been the focus of a great amount of theoretical works dealing with the third-order nonlinear response of polyenes [130-1351. Another A excited state, the 5A state (hereafter denoted m2A), which is close in energy to the m3A state, is also strongly coupled to the 1B state. This feature is typical of longer chains and has also been observed previously in polyenes by Garito and co-workers [131], Mazumdar and co-workers [137-1391, and our group [135,52]. The coupling between the 1B and 2A (or mlA) states is of intermediate strength. (iii) The 2A, m3(= 6)A, and m2(= 5)A states have strong transition dipole moments with B-symmetry bands of higher energy. These higher-lying B states (denoted nlB, n2B, and n3B, in order of increasing energy) are only very weakly coupled to the ground state, which means they hardly contribute to the one-photon absorption spectrum. We stress that this description is in overall agreement with the general model of Mazumdar and co-workers [ 137- 1391. Three excitation pathways dominate the expression of the static third-order susceptibility: two positive contributions from the 1A + 1B -r m3A + 1B + 1A and 1A + 1B -r m2A + 1B + 1A channels; and a negative contribution from the 1A -+ 1B-+ 1A + 1B+ 1A channel. The y value mainly results from the incomplete cancellation between these terms. The 2A state is involved in two types of pathways: 1A + 1B -+ 2A 1B + 1A and 1A + 1B+ 2A+ nlB + 1A. As these two contributions have opposite signs, the resulting role played by the 2A state in the description of the third-order nonlinearity is strongly reduced. We note, however, that the cancellation is only partial (in contrast to what is assumed in the essential-state model of Mazumdar and co-workers [137-1391) and that a quantitative estimate of the static y values requires to include in the SOS expression the following states: lA, lB, 2A, m2A, m3A, nlB, n2B, and n3B. These states form our model for the NLO response of thiophene oligomers.

6.6.2 Dynamic third-order response of Th7: Two-photon absorption and third-harmonic generation Pfeffer et al. [154] have recorded the TPA spectrum of poly(3-octylthiophene) in solution. This compound presents a one-photon absorption peak at a very high energy, 2.9eV; such a high energy value is due to torsions around the inter-ring bonds and, as a result, lower degree of conjugation. The TPA spectrum shows an intense peak around 1.75eV and a weaker one that parallels the one-photon absorption (around 1.45eV). According to Pfeffer et al., the peak at 1.75eV corresponds to a two-photon resonance from the ground state to the mA state, while the weak signal around 1.45eV has two possible origins: (i) one-photon absorption to the 1B state which becomes allowed by symmetry breaking; or (ii) two-photon absorption to the 2A state. In Ths, Taliani and co-workers [54] have observed a TPA peak at 1.13 eV which they attribute to the origin of the 2A state (as well as a much weaker response at

6.6 Nonlinear optical properties of neutral oligothiophenes

353

1.08eV assigned to the 1B state made partially allowed by disorder and reduction of the molecular symmetry). As Th6 is a good model for polythiophene, these authors state that a TPA response associated to the 2A state is expected to remain in the polymer. On that basis, we expect two two-photon resonances in the THG dispersion curve of long thiophene oligomers and polythiophene: an intense one due to the mA state and a weaker one associated to the 2A state. To clarify this point, we have simulated the TPA spectrum of the Th7 oligomer, see Fig. 16. To allow for a direct comparison to experiment, we have considered in the calculations the two-photon excited-state energies provided by the TPA measurements on the polymer (i.e. 2.9eV for 2A and 3.5 eV for mA) [154]. By considering this approach, the shape of the TPA curve is calculated to be in good agreement with the experimental data from Pfeffer et al.: the spectrum is dominated by absorption into the mA band while the 2A state leads to a weaker feature at lower energy (the ratio between the intensities of the two peaks is approximately 15). Furthermore, by fitting the experimental TPA data with a four-level model (lA, lB, 2A, and mA), Pfeffer et al. have estimated the transition dipole moments among excited states: (1Blp12A) = 7 D and (1BIplmA) = 18 D; these values are in good agreement with the results of the INDO/MRD-CI calculations (6.3 D for the 1B-2A transition; 11.7 and 24.0 D for the 1B-m2A and 1B-m3A excitations, respectively). THG measurements on alkyl-substituted polythiophenes performed in a range of energies going from 0.6 to 1.3eV show an intense peak at 0.8 eV with a shoulder at about 1.15 eV [136]. The 1.15eV peak cannot be accounted for in a THG spectrum 70000 L

60000

1

50000

I

J

t

I

i

1 .oo

1.25

1.50

1.75

2.00

Energy (eV)

Figure 16. INDO/MRD-CI/SOS simulation of the two-photon absorption spectrum of Th7, as a function of fundamental energy.

354

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

simulated on the basis of a model considering only the lA, lB, m2A, and m3A states (since the dominant two-photon states have been located at much higher energy by TPA measurements [ 1361); such a model thus fails for the description of the dynamic properties of the third-order polarizabilities in long thiophene oligomers, whereas it was found to be adequate in the description of the static response. In order to obtain better agreement with the experimental THG data, more states (i.e. the 2A state as well as high-lying B states) have to be included in the SOS formalism. In Fig. 17, we compare the THG frequency-dependent energy curve calculated for Th7 on the basis of various simplified models to the spectrum measured in the polymer [136] (note that the much lower transition energy, 2.4eV, found experimentally for the 1A-1B vertical transition energy in this case most probably results from a larger degree of conjugation). Here also, we make use of the experimental excitation energies for the 2A (2.16eV [54]) and nlB (3.3eV [138]) states. We only find good agreement with the experimental data when both the 2A and n l B states are included in the SOS expression, see Fig. 17. Two resonance peaks dominate the low-energy domain of the THG dispersion curve: (i) a three-photon resonance to the 1B state; and (ii) superimposed resonances to the nlB and 2A states. Analysis of the NLO response shows that the 1A-t 1B +m3A+ 1B + 1A and 1A+ 1B-t m2A -+ 1B -+1A pathways dominate the description of the NLO response in the

0' 0.0

I

I

I

I

0.3

0.6

0.9

1.2

1

1.5

Energy (eV)

Figure 17. Comparison between THG experimental data (squares) for polythiophene (from Ref. 138) and the lNDO/MRD-CI/SOS simulated spectrum of Th7 versus fundamental photon energy obtained on the basis of models including: (a) the IA, lB, ml(=2)A, m2A, m3A, nlB, n2B, and n3B states (solid curve); (b) the lA, lB, m2A, m3A, and nlB states (dashed curve); and (c) the lA, IB, ml(=2)A, m2A, and m3A states (dotted curve).

6.7 Synopsis

355

region of the first resonance while channel 1A 3 1B + 2A -r nlB + 1A provides the largest contribution to the nonlinearity in the energy domain close to the second resonance feature. We conclude that both the 2A and nlB states play an important role in the resonant regime of the THG process. At higher energies, resonances to the m2A, m3A, n2B, and n3B states are also calculated; their actual intensity strongly depends on whether double resonances phenomena occur or not [155]. In conclusion, a proper description of the dynamic NLO response of long conjugated oligothiophene chains requires to include in the Sum-Over-States expression: (i) the ground state (IA); (ii) the first one-photon allowed excited state (IB); (iii) the first two-photon allowed excited state (2A); (iv) a band of two-photon states (mA), strongly coupled to 1B; and (v) bands of higher-lying B states (nB) with large transition moments with both the 2A and mA states.

6.7 Synopsis We have illustrated in this review that quantum-chemical calculations are helpful to provide a basic understanding of the intrinsic properties of oligothiophenes and polythiophenes and to better apprehend the relation to their use as active element in electro-optic and photonic applications. The theoretical calculations have been shown to be particularly useful in the context of engineering new compounds with enhanced characteristics to be exploited in such devices. We have exemplified in this Chapter that several strategies can be developed to tailor a given property, such as modifying the topology of the conjugated backbone, changing the size of the molecules, or grafting electroactive moieties onto the conjugated path. It is worth stressing that such a control is expected to be much more difficult handle for a polymer material formed by an inhomogeneous distribution of segments of different effective conjugation lengths. The next step to accomplish in order to improve the degree of sophistication of the theoretical models we are using, is to allow for a fine understanding of the solid-state effects encountered in the devices. This of course requires our theoretical modeling to go beyond the consideration of isolated molecules and to deal with several molecules in interaction. This is the reason why a large number of theoretical studies recently initiated in our group is entirely dedicated to this challenging task.

Acknowledgements This work is partly supported by the Belgian Prime Minister’s Office of Science Policy ‘P6le d’Attraction Interuniversitaire en Chimie Supramoleculaire et Catalyse’; the Belgian National Fund for Scientific Research (FNRS); the Training and Mobility of Researchers Network ‘SELOA’ and an IBM Academic Joint Study. JC is Aspirant, DB Charge de Recherches, and RL Maitre de Recherches of the

356

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

FNRS. We are grateful to H. Bassler, R. H. Friend, G. Horowitz, R. A. J. Janssen, S. Mazumdar, W. R. Salaneck, Z. Shuai, Z. G. Soos, C. Taliani, W. Torruellas, and R. Zamboni for stimulating collaborations or discussions.

References 1. Handbook of Conducting Polymers (ed.: T.A. Skotheim),Marcel Dekker, New York, 1986 Conjugated Polvmers: The Novel Science and Technology of Conducting and Nonlinear Optically Active Materials (eds.: J. L. Bredas and R. Silbey),Kluwer, Dordrecht, 1991; Handbook of Conductive Polymers (2nd Edn) (eds.: T. A. Skotheim, J. R. Reynolds and R. L. Elsenbaumer), Marcel Dekker, New York, 1998. 2. H. Shirakawa, E. J . Louis, A. G. MacDiarmid, C. K. Chiang and A. J. Heeger, J . Chem. SOC. Chem. Commun., 1977,578. 3. W. P. Su, J. R. Schrieffer and A. J. Heeger, Phys. Rev. Lett., 1979,42, 1698. 4. J. Roncali, Chem. Rev., 1992, 92, 71 1. 5. J. L. BrCdas, G. B. Street, B. Themans and J. M. AndrC, J . Chem. Phys., 1986,83, 1323. 6. F. Geiger, M. Stoldt, H. Schweizer, P. Bauerle and E. Umbach, Adv. Muter., 1993, 5, 922. 7. Electronic Materials: The Oligomer Approach (Eds.: G. Wegner and K. Mullen), Wiley-VCH, Weinheim, 1998. 8. D. Beljonne, J. Cornil, D. A. dos Santos, Z. Shuai and J. L. BrCdas, in ‘Primary Photocxcitations’ in Conjugated Polymers: Molecular Exciton versus Semiconductor Band Model (Ed.: N. S. Sariciftci), World Scientific, Singapore, in press. 9. J. Bell, Opto&Laser Europe, 1996, 34, 19. 10. J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H. Friend, P. L. Burn and A. B. Holmes, Nature, 1990, 347, 539. 11. G. Gustafsson, Y. Cao, G. M. Treacy, F. Klavetter, N. Colaneri and A. J. Heeger, Nature, 1992, 357, 477. 12. J. L. BrCdas, J. Cornil and A. J. Heeger, Adv. Muter., 1996,8,447. 13. P. L. Burn, A. Kraft, D. R. Baigent et al., J . Am. Chem. SOC.,1993,115, 10 117. 14. D. Braun, E. G. J. Staring, R. C. J. E. Demandt, G. L. J. Rikken, Y. A. R. R. Kessener and A. J. H. Venhuizen, Synth. Met., 1994, 66, 75, 15. A. R. Brown, D. D. C. Bradley, J. H. Burroughes et al., Appl. Phys. Lett., 1991, 61, 2793. 16. L. S. Swanson, J. Shinar, A. R. Brown et al., Phys. Rev. B, 1992, 46, 15072. 17. D. R. Baigent, N. C. Greenham, J. Gruner et al., Synth. Met., 1994,67, 3. 18. M. Berggren, G. Gustafsson, 0. Inganas, M. R. Anderson, 0. Wennerstrom and T. Hjertberg, Adv. Mater., 1994, 6, 488. 19. G. G. Malliaras, J. K. Herrema, J. Wildeman et al., Adv. Muter., 1993, 5, 721. 20. R. E. Gill, G. G. Malliaras, J. Wildeman and G. Hadziioannou, Adv. Muter., 1994, 6, 132. 21. M. Berggren, 0. Inganas, G. Gustafsson et al., Nature, 1994, 372, 444. 22. M. Berggren, G. Gustafsson, 0. Inganas, M. R.Anderson, 0. Wennerstrom and T. Hjertberg, J . Appl. Phys., 1994, 76, 7530. 23. P. Dyreklev, M. Berggren, 0. Inganas, M. R. Anderson, 0.Wennerstrom and T. Hjertberg, Adv. Mater., 1995, 7, 43. 24. K. Uchiyama, H. Akamichi, S. Hotta, H. Noge and H. Sasaki, Synth. Met., 1994,57-59,63. 25. G. Horowitz, P. Delannoy, H. Bouchriha et al., Adv. Muter., 1994, 6, 752. 26. R. N. Marks, F. Biscarini, R. Zamboni and C. Taliani, Europhysics Lett., 1995, 32, 523. 27. F. Garnier, R. Hajlaoui, A. Yassar and P. Srivastava, Science, 1994, 265, 1684. 28. A. Dodabalapur, L. Torsi and H. E. Katz, Science, 1994, 268, 270. 29. D. D. C. Bradley, Physics World, 1994, April, p. 29. 30. D. Fichou, J. M. Nunzi, F. Charra and N. Pfeffer, Adv. Muter., 1994,6,64. 31. M. Yan, L. J. Rothterg, F. Papadimitrakopoulos, M E. Calvin and T. M. Miller, Phys. Rev. Lett., 1994, 72, 1104.

References

357

J. Cornil, A. J. Heeger and J. L. Bredas, Chem. Phys. Lett. 1997, 272,463. B. Xu and S. Holdcroft, J . Am. Chem. SOC.,1993,115, 8447. Q . Pei, G . Yu, C. Zhang, Y. Yang and A. J. Heeger, Science, 1995, 269, 1086. W. R. Salaneck, M. Logdlund, J. Birgersson, P. Barta, R. Lazzaroni and J. L. Brkdas, Synth. Met., 1997, 85, 1219. 36. J. P. Hermann and J. P. Ducuing, J . Appl. Phys., 1974, 4.5, 5100. 37. Introduction to Nonlinear Optical Effects in Molecules and Polymers (Eds.: P. N. Prasad and D. J. Williams), Wiley Interscience, New York, 1991. 38. Nonlinear Optical Properties of Organic Molecules and Crystals (Eds.: D. S. Chemla and J. Zyss), Academic, New York, 1987. 39. Nonlinear Optical Properties of Polymers (Eds.: A. J. Heeger, J. Orenstein and D. R. Ulrich), Materials Research Society Symposium Proceedings, Vol. 109, Materials Research Society, Pittsburgh, 1988. 40. Conjugated Polymeric Materials: Opportunities in Electronics, Optoelectronics and Molecular Electronics (Eds.: J. L. Brtdas and R. R. Chance), NATO-AS1 Series, Vol. E182, Kluwer, Dordrecht, 1990; J. L. BrCdas, C. Adant, P. Tackx, A. Persoons and B. M. Pierce, Chem. Rev., 1994, 94, 243. 41. M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart, J. Am. Chem. SOC.,1985, 107, 3902. 42. M. J. S . Dewar and W. Thiel, J . Am. Chem. SOC.,1977,99,4899. 43. D. Beljonne, J. Cornil, R. H. Friend, R. A. J. Janssen and J. L. BrCdas, J . Am. Chem. SOC., 1996,118, 6453. 44. Density-Functional Theory of Atoms and Molecules (Eds.: R. G. Parr and W. Yang), Oxford University Press, New York, 1989. 45. B. Delley, J . Chem. Phys., 1990, 92, 508; ibidem, 1991, 94, 7245. 46. DMOL User Guide, Version2.3.5, December, 1993, San 72 Diego: Biosym Technologies,1993; Release Notes 2.3.5, June, 1994, Sun Diego: Biosym Technologies;DMOL 95.0/3.0.0 Quantum Chemistry User Guide, October, 1995, San Diego: Biosym/MSI. 47. S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200. 48. J. Ridley and M. Zerner, Theoret. Chim. Actu, 1973,32, I l l . 49. G. Nicolas and Ph. Durand, J. Chem. Phys., 1980, 72, 453; J. L. BrCdas, R. R. Chance, R. Silbey, G. Nicolas and Ph. Durand, J . Chem. Phys., 1981, 75, 255. 50. J. Cornil, D. Beljonne and J. L. BrCdas, J . Chem. Phys., 1995,103,834. 51. R. J. Buenker and S. D. Peyerimhoff, Theoret. Chim. Acta, 1974,35, 33. 52. D. Beljonne, Z. Shuai and J. L. Brtdas, J . Chem. Phys., 1993,98, 8819. 53. D. Birnbaum and B. E. Kohler, J. Chem. Phys., 1991, 95, 4783; ibidem, 1992, 96, 2492. 54. N. Periasamy, R. Danieli, G . Ruani, R. Zamboni and C. Taliani, Phys. Rev. Lett., 1992,68,919. 55. M. Rubio, M. Merchan, E. Orti and B. 0. Roos, J. Chem. Phys., 1995,102,3580; L. SerranoAndres, M. Merchan, M. Fulscher and B. 0. Roos, Chem. Phys. Lett., 1993, 211, 125. 56. H. Chosrovian, S. Rentsch, D. Grebner, D. U. Dahm and E. Birckner, Synth.Met., 1993,60,23. 57. R. A. J. Janssen, L. Smilowitz, N. S. Saricifici and D. Moses, J . Chem. Phys., 1994, 101, 1787; R. A. J. Janssen, D. Moses and N. S. Sariciftci, J . Chem. Phys., 1994, 101, 9519. 58. J. Guay, P. Kasai, A. Diaz, R. Wu, J. M. Tour and L. H. Dao, Chem. Muter., 1992, 4 , 107. 59. G. Horowitz, A. Yassar and H. J. von Bardeleben, Synth. Met., 1994, 62, 245. 60. L. S. Swanson, J. Shinar and K. Yoshino, Phys. Rev. Lett., 1990, 65, 1140. 61. M. Bennati, A. Grupp, P. Bauerle and M. Mehring, Mol. Cryst. Liq. Cryst., 1994, 256, 751; ibidem, J. Phys. Chem., 1996, 100,2849. 62. J. C. Scaiano, R. W. Redmond, B. Mehta and J. T. Arnason, Photochem. Photobiol., 1990, 52, 655. 63. Y. Zeng, L. Biczok and H. Linschitz, J . Phys. Chem., 1992,96, 5237. 64. J. P. Reyftmann, J. Kagan, R. Santris and P. Morliere, Photochem. Photobiol., 1985, 41, 1. 65. N. C. Greenham, I. D. W. Samuel, G. R. Hayes et al., Chem. Phys. Lett., 1995,241, 89. 66. B. E. Kohler, Chem. Rev., 1993, 93, 41. 67. R. H. Austin, G. L. Baker, S. Etemad and R. Thompson, J. Chem. Phys., 1989, YO, 6642. 68. R. Rossi, M. Ciofalo, A. Carpita and G . Ponterini, J . Photochem. Photobiol. A: Chem., 1993, 70, 59. 32. 33. 34. 35.

358

6 Geometric and Electronic Structure and Optical Response of Oligo- and Polythiophenes

Z. G. Soos, D. S. Galvao and S . Etemad, Adv. Muter., 1994,6, 280. R. S. Becker, J. S. de Melo, A.L. Maganita and F. Elisei, PurediAppl. Chem., 1995,67,9. W. P. Su, J. R. Schrieffer and A. J. Heeger, Phys. Rev. B, 1979,22, 2099. J. L. BrCdas and A. J. Heeger, Chem. Phys. Lett., 1994, 21 7, 507. J. Cornil, D. A. dos Santos, D. Beljonne and J. L. BrCdas, J . Phys. Chem., 1995,99, 5604. N. C. Greenham, S . C. Moratti, D. D. C. Bradley, R. H. Friend and A. B. Holmes, Nature, 1993, 365, 628. 75. P. Hapiot, F. Demanze, A. Yassar, F. Garnier, J. Phys. Chem., 1996, 100, 8397. 76. F. van Bolhuis, H. Wynberg, E. E. Havinga, E. W. Meijer and E. G. J. Staring, Synth. Met., 1989, 30, 381. 77. S. Hotta and K.Waragai, Adv. Muter., 1993, 5 , 896. 78. G. Horowitz, B. Bachet, A. Yassar, P. Lang, F. Demanze, J. L. Fave and F. Garnier, Chem. Muter., 1995, 7, 1337. 79. T. Siegrist, R. M. Fleming, R. C. Haddon et al., J. Muter. Research, 1995, 10, 2170. 80. F. Demanze, J. Cornil, F. Gamier, G . Horowitz, P. Valat, A Yassar, R. Lazzaroni and J. L. Bredas, J. Phys. Chem. B, 1997,101, 4553. 81. P. M. Viruela, R. Viruela, E. Orti and J. L. BrCdas, J. Am. Chem. SOC.,1997, 119, 1360. 82. V. Hernandez and J. T. Lopez Navarette, Synth. Met., 1996, 76, 221. 83. C. Taliani and L. M. Blinov, Adv. Muter., 1996, 8, 353. 84. J. L. Bridas and G Steet, Acc. Chem. Res., 1985, 18, 309. 85. K. Fesser, A. R. Bi p and D. K. Campbell, Phys. Rev. B, 1983,27,4804. 86. T. C. Chung, J. H. Kaufman, A. J. Heeger and F. Wudl, Phys. Rev B, 1984,30, 702. 87. G. Harbeke, E. Meier, W. Kobel, M. Egli, H. Kiess and E. Tosatti, Solidstate Commun., 1985, 55, 419. 88. J. Cornil, D. Beljonne and J. L. Bredas, J. Chem. Phys., 1995,103, 842. 89. C. Ehrendorfer and A. Karpfen, J. Phys. Chem., 1994, 98, 7492; ibidem, 1995,99, 5345. 90. P. Bauerle, U. Segelbacher,K. U. Gaudl, D. Huttenlocher and M. Mehring, Angew. Chem. Znt. Ed. Eng., 1993, 32, 76. 91. D. Fichou, G. Horowitz, B. Xu and F. Garnier, Synth. Met., 1993, 39, 243. 92. S. Hotta and K. Waragai, J. Phys. Chem., 1993,97, 7427. 93. G. Lanzani, L. Rossi, A. Piaggi, A. J. Pal and C. Taliani, Chem. Phys. Lett., 1994, 226, 547. 94. J. Poplawski, E. Ehrenfreund, J. Cornil et al., Mol. Cryst. Liq. Cryst., 1994, 256, 407. 95. M. G. Harrison, R. H. Friend, F. Garnier and A. Yassar, Synth. Met., 1994, 67, 215. 96. J. Cornil and J. L. Brkdas, Adv. Muter., 1995, 7, 295. 97. P. Bauerle, U. Segelbacher, A. Maier, M. Mehring, J. Am. Chem. SOC.,1993, 115, 10217. 98. M. G. Hill, J. F. Penneau, B. Zinger, K. R. Mann and L. L. Miller, Chem. Muter., 1992, 4, 1106. 99. N. Yokonuma, Y. Furukawa, M. Tasumi, M. Kuroda and J. Nakayama, Chem. Phys. Lett., 1996, 225, 43 1. 100. Y. Furukawa, J. Phys. Chem., 1996, 100, 15644. 101. A. Yassar, D. Delabouglise, M. Hmyene, B. Nessakh, G. Horowitz and F. Garnier, Adv. Muter., 1992, 4,490. 102. B. Nessakh, G. Horowitz, F. Garnier, F. Deloffre, P. Srivastava and A. Yassar, J . Electroanal. Chem., 1995,399, 97. 103. J. A. E. H. van Haare, E. E. Havinga, J. L. J. van Dohgen, R. A. J. Janssen, J. Cornil and J. L. BrCdas, Chem. Eur. J., in press. 104. Y. Shimoi and S . Abe, Phys. Rev. B, 1994,50, 14781. 105. A. J. W. Tol, Chem. Phys., 1996,208, 73. 106. A. 0. Patil, A. J. Heeger and F. Wudl, Chem. Rev., 1988,88, 183. 107. Z. Vardeny, E. Ehrenfreund, 0. Brafman, M. Nowak, H. Schaffer, A. J. Heeger and F. Wudl, Phys. Rev. Lett., 1986, 56, 671. 108. P. A. Lane, X. Wei and Z. V. Vardeny, Phys. Rev. Lett., 1996, 77, 1544. 109. G. Zotti, S. Martha, G. Wegner and A. D. Schluter, Adv. Muter., 1992,4, 798. 110. R. K. Khanna, Y. M. Jiang, B. Srinivas,C. B. Smithhart and D. L. Wertz, Chem. Muter., 1993, 5, 1792; H. Gregorius, W. Heitz and K. Mullen, Adv. Muter., 1993, 5, 279. 111. R. Schenk, H. Gregorius and K. Miillen, Adv. Muter., 1991, 3, 492. 69. 70. 71. 72. 73. 74.

References

359

112. S. Stafstrom, M. Boman, R. Lazzaroni and J. L. Bredas, in Polymer-Solid Interfaces (Eds.: J. J. Pireaux, P. Bertrand and J. L. BrCdas), Institute of Physics, Bristol, 1992, p. 189. 113. M. Boman, S. Stafstrom and J. L. BrCdas, J . Chem. Phys., 1992, 97, 9144. 114. W. R. Salaneck, S. Stafstrom and J. L. Bredas, Conjugated Polymer Surfaces and Interfaces, Cambridge University Press, Cambridge, 1996. 115. P. Dannetun, M. Logdlund, C. Fredriksson et al., in Polymer-Solid Interfaces (Eds.: J. J. Pireaux, P. Bertrand and J. L. Bredas), Institute of Physics, Bristol, 1992, p. 201. 116. P. Dannetun, M. Boman, S. Stafstrom et al., J . Chem. Phys., 1993, 99, 664. 117. C. Fredriksson, R. Lazzaroni, J. L. BrCdas, A. Ouhlal and A. Selmani, J . Chem. Phys., 1994, 100,9258. 118. V. Parente, C. Fredriksson, A. Selmani, R. Lazzaroni and J. L. Brkdas, J . Phys. Chem. B, 1997, 101, 4193 119. V. Parente, unpublished data. 120. R. Lazzaroni, V. Parente, C. Fredriksson and J. L. BrCdas, Synth. Met., 1996, 76, 225. 121. V. Parente, R. Lazzaroni, A. Selmani and J. L. Bredas, Synth. Met., 1995, 67, 147. 122. A. Almenningen, S. Havorsen and A. Haaland, Acta Chem. Scand., 1971,25, 1937. 123. 0. Jones, J . Chem. Phys., 1993, 99, 1194. 124. V. Parente, G. Pountois, R. Lazzaroni, J. L. BrCdas, G. Ruani, M. Murgia and R. Zamboni, Adv. Muter., 1998, 10, 319. 125. V. Parente, G. Pourtois, R. Lazzaroni and J. L. Bredas, Synth. Met., 1997, 85, 1031. 126. B. Hudson, B. E. Kohler and K. Schulten, in Excited States (Ed.: E. C. Lim), Academic, New York, 1982, Vol. 6, p. 1. 127. W.-S. Fann, S. Benson, J. M. J. Madey, S. Etemad, G. L. Baker and F. Kajzar, Phys. Rev. Lett., 1989, 62, 1492. 128. C. Halvorson, T. W. Hagler, D. Moses, Y. Cao and A. J. Heeger, Chem. Phys. Lett., 1992, 200, 364. 129. P. Tavan and K. Schulten, J . Chem. Phys., 1986, 85, 6602; ibidem, Phys. Rev. B, 1987, 36, 4337. 130. J. L. BrCdas and A. J. Heeger, Chem. Phys. Lett., 1989, 154, 56. 131. A. F. Garito, J. R. Heflin, K. Y. Wong and 0. Zamani Khamiri, in Ref. 4, p. 91. 132. B. M. Pierce, J . Chem. Phys., 1989, 91, 791. 133. S. Mazumdar and S. N. Dixit, Phys. Rev. Lett., 1983, 51, 292; S. Mazumdar and D. K. Campbell, ibidem, 1985, 55, 2067; S. N. Dixit, D. Guo, and S. Mazumdar, Phys. Rev. B, 1991, 43, 6781. 134. Z. G. Soos and S. Ramasesha, Phys. Rev. B, 1984, 29, 5410; S. Ramasesha and Z. G. Soos, J . Chem. Phys., 1984, 80, 3278. 135. Z. Shuai, D. Beljonne and J. L. BrCdas, J . Chem. Phvs., 1992,97, 1132. 136. W. E. Torruellas, D. Neher, R. Zanoni, G. I. Stegeman, F. Kajzar and M. Leclerc, Chem. Phys. Lett., 1990, 175, 11; W.E. Torruellas, K. B. Rochford, R. Zanoni, S. Aramaki and G. I. Stegeman, Opt. Commun., 1991, 82, 94. 137. D. Guo and S. Mazumdar, J. Chem. Phys., 1992,97,2170; D. Guo, S. Mazumdar and S. N. Dixit, Synth. Met., 1992, 49, 1 . 138. D. Guo, S. Mazumdar, G. I. Stegeman et al., in Electrical, Optical and Magnetic Properties of Organic Solid Staie Materials (Eds.: L. Y. Chiang, A. F. Garito and D. J. Sandman), Materials Research Society Symposium Proceedings, Vol. 109, Materials Research Society, Pittsburgh, 1991, p. 151. 139. D. Guo, S. Mazumdar, S. N. Dixit, Y. Kawabe, F. Jarka and N. Peyghambrian, Phys. Rev. B 1993,48, 1; S. Mazumdar, D. Guo and S. N. Dixit, SPZE, 1991, 1436, 136. 140. G. P. Agrawal, C. Cojan and C. Flytzanis, Phys. Rev. B, 1978, 17, 776. 141. E. F. McIntyre and H. F. Hameka, J . Chem. Phys., 1978,68,3481. 142. S. C. Mehendale and K. C. Rustagi, Opt. Commun., 1979,28, 259. 143. M. G. Papadopoulos, J. Waite and C. A. Nicolaides, J . Chem. Phvs., 1982, 77, 2257. 144. C. P. de Melo and R. Silbey, Chem. Phys. Lett., 1987, 140, 537; ibidem, J . Chem. Phys., 1988, 88, 2567. 145. D. N. Beratan, J. N. Onuchic and J. W. Perry, J. Phys. Chem., 1987, 91, 2697. 146. J. R. Heflin, K. Y. Wong, 0.Zamani-Khamiri and A. F. Garito, Phys. Rev. B, 1988,38,1573.

360

6 Geometric and Electronic Structure and Optical Response of Oligo-and Polythiophenes

147. D. Fichou, F. Garnier, F. Charra, F. Kajzar and J. Messier, Spec. Pub1.-R. SOC.Chem., 1989, 69, 176. 148. (a) Z. Shuai and J. L. Brkdas, Phys. Rev. B, 1991,44,5962;(b) Z. Shuai and J. L. Brkdas, Phys. Rev. B, 1992,46,4395. 149. S. Mukamel, A. Takahashi, H. X. Wang and G . Chen, Science, 1994,265,250. 150. H. Thienpont, G. L. J. A. Rikken, E. W. Meijer, W. ten Hoeve and H. Wynberg, Phys. Rev. Lett., 1990, 65, 2141. 151. M. T. Zhao, B. P. Singh and P. N. Prasad, J. Chem. Phys., 1988, 89, 5535. 152. T. Bjornholm, D. R. Greve, T. Geisler, J. C. Petersen, M. Jayaraman and R. D. McCullough, Adv. Mater., 1996, 8, 920. 153. B. J. Orr and J. F. Ward, Molec. Phys., 1971, 20, 513. 154. N. Pfeffer, P. Raimond, F. Charra and J. M. Nunzi, Chem. Phys. Lett., 1993,201, 357. 155. D. Beljonne and J. L. Brtdas, J . Opt. SOC.Am. B, 1994, 11, 1380.

7 Electronic Excited States of Conjugated Oligothiophenes c

7.1 Introduction The excited electronic states of conjugated organic molecules have grown in interest since the early 1930s, when the first theoretical approaches [l, 21 towards an understanding of the photophysical properties of large 7r-electron systems were established. Besides the electronic levels of the isolated molecules, molecular solid state properties have attracted many researchers and molecular crystals have been topics of numerous experimental and theoretical efforts [3-51. Over the last few years thin$Zms of conjugated molecular systems have drawn considerable attention to polymers as well as oligomeric materials, e.g. oligothiophenes. These investigations were motivated by the possible technological applications, such as organic light emitting devices (OLEDs) [6-91, field effect transistors (FETs) [lo-131, and photovoltaic cells [14]. Thin films of conjugated materials are much favored for this kind of applications since they can be easily prepared compared to the sophisticated processes of organic single crystal growth [15] and can be tailored for special purposes in thickness and geometry in a well-defined way, e.g. multilayer preparation [16, 171 or optical microcavities [18, 191. In addition to the technological chances, there is still a need for a deeper understanding of the thin film properties and among the various scientific questions on conjugated electronic systems, the changes of the molecular electronic and hence the optical properties in the solid state is of fundamental interest. However, elucidation of the solid state changes with respect to the isolated molecules is not straightforward since chemical as well as structural defects may prohibit thorough spectroscopic investigations on thin films. Characteristic model systems are therefore necessary to optimize the growth of thin organic films and to acquire detailed information on well-defined and ordered films, to establish a deeper understanding of the electronic properties. Within this framework oligothiophenes are one of the most studied r-conjugated oligomeric model systems. This class of molecules has been a target for chemical tailoring in different groups worldwide [20-231 to provide a huge variation in welldefined chain lengths from 1 to 16 thiophene rings [24] by using different end- or sidechain-substituents to keep the longer oligomers soluble. Since the main interest is in the basic properties of the oligothiophenes and not in the substitutional effects we will concentrate in this article on the basic model class of unsubstituted oligothiophenes nT [25], with II being the number of thiophene units. In fact, soluble oligothiophenes are obtained by substitution with long alkyl chains, which in turn have a strong influence on the intermolecular packing due to the van der Waals interaction among adjacent alkyl chains [24, 261. This may even lead to modifications

362

7 Electronic Excited States of Conjugated Oligothiophenes

of the electronic properties, e.g. thermochromic effects [27, 281. Due to the thermal stability of the nT, evaporation procedures for purification can be applied, yielding ultrapure materials with a minimum amount of chemical impurities. The nT series offers the possibility of examining the polymer properties of polythiophene with well-defined conjugated backbones. Furthermore, due to the highly-ordered structure, the oligomeric solid state reveals exciting and unusual properties which would not be expected for polymers due to their disordered nature. Although the electronic properties of oligothiophenes have already been reviewed [29, 301, the picture of the nature and ordering of the electronic states in the solid state is still diffuse and incoherent. Therefore we will cover recent results on the optical properties especially in the solid state and the outcome may lead to a clearer picture of the excited state ordering. This may even contribute to a better understanding of the electronic states for other rigid-rod-like conjugated materials. The review is organized as follows: first, a short summary of the polythiophene properties will be given; then we concentrate on the properties of the oligothiophene model molecules in solution. In the next section the changes of the molecular properties from an isolated molecule to the solid phase will be discussed, focussing on the influence of the geometric structure and the morphology. A brief report on the applied features of polarized electroluminescent devices based upon oligothiophene films will follow, before we finally review the nonlinear optical properties.

7.2 Electronic structure of conjugated polymers 7.2.1 General concept Even before electroluminescent devices, based on organic molecules [311 or polymers [32] brought conjugated materials one step ahead to possible technical application, there had been interest in the electronic properties and structures of conjugated materials. This interest was mainly based on polymers resembling one-dimensional systems and extrapolation and experimental verification of the ‘infinite’conjugated chain played a major role. The theoretical approach taken by Su, Schrieffer and Heeger (SSH-model) to describe the ideal polymer was quite similar to earlier considerations of Longuet-Higgins and Salem [33] for short alternating hydrocarbons. The main assumption was to neglect electron-electron correlations and to derive a description for the conjugated 7r-electron system within a one electron picture [34361. In the SSH-model lattice relaxation due to electron-phonon interaction was taken into account to explain the observed optical energy gap in conjugated polymers as a result of the bond alternation due to a Peierls distortion. An overview of the numerous theoretical approaches and refinements concerning ideal polymers can be found in Ref. [37]. One of the important results within the SSH-model and the neglect of electron correlation effects was the identification of the optical gap to be the energy gap for the generation of separated electron-hole pairs, and thus the exciton binding energy was thought to be negligible. However it was shown

7.2 Electronic structure of conjugatedpolymers

363

for the prototypical polyacetylene that the exciton binding energy is in the order of 0.5 eV and thus the electron-hole or electron-electron correlation play an important role in the description of conjugated systems [37, 381. For an understanding of the particular properties of polyacetylene well-defined oligomeric model molecules turned out to give very helpful experimental contributions. The model oligoenes investigated by Hudson et al. [39] and Kohler et al. [40,41] revealed quite clearly that the lowest excited singlet state for the oligoenes is of gerade symmetry. This finding could not be described in simple one-electron pictures and underlined that electron correlation effects are very important for conjugated electron systems. The occurrence of a splitting between the singlet and the triplet excited states is further evidence for the essential role of electronic correlations. Furthermore the disordered nature of most of the real polymer systems has very important consequences on their macroscopic optical and electrical properties and may lead to deviations from the theoretical expectations on an ideal infinite chain length polymer.

7.2.2 Polythiophene Due to the absence of luminescence in degenerate ground state polymers with low band gaps ( e g polyacetylene), which is understandable if the lowest excited state has dipole-forbidden gerade symmetry, the interest in optical properties aimed at non-degenerate ground state materials, e.g. polyphenylenevinylene (PPV) [42, 431 and polythiophene (PT). For the latter, photoexcitations have frequently been studied. Generally only very broad absorption profiles could be found even at low temperatures [44, 451 and a typical spectrum is displayed in Fig. 1 [45].The almost featureless absorption band was originally interpreted as electronic interband transitions which show an onset at about 2.0eV (16 000cm-’) [46]. The idea

pdy thiophene

1.1

1.3

1.5

1.7

PHOTON

1.9

2.1

2.3

2.5

2.7

ENERGY ( e V )

Figure 1. Typical absorption and PL spectra of polythiophene. Reproduced from Ref. [45] by kind permission of Elsevier.

364

7 Electronic Excited States of Conjugated Oligothiophenes

of separated electron-hole pairs was corroborated by the fact that charged excitations could be found by different experimental techniques, e.g optical detected magnetic resonance [45] and electron spin resonance [47]. However, it was well known from photoconductivity investigations in molecular crystals f3] that exciton dissociation at defect sites can even occur in highly ordered systems and hence generation of charged species has to be expected even if the optical absorption would be primarily due to neutral excitations. Turning to the photoluminescence (PL) [48] spectra in Fig. 1, structured emission contrasts with the broad absorption band profile. This progression is attributed to the coupling of a carbon-carbon stretching mode (see section 7.3.2.4) to the electronic origin. Since vibronic structure was expected to be absent for interband transitions, the excitonic nature of the PL was acknowledged [49,50], although attributed to the effects of interchain excitons on adjacent polymer backbones [45]. One of the main problems from the experimental view point was the poor sample quality of PT. Disorder, ageing effects as well as polymerisation route induced changes, played a major role for the spectroscopic features [51]. Yoshino et al. [52] tried to improve the order within the PT by the fabrication of oriented PT, where the orientation of the polymer backbone led to a high dichroism (>15) of the absorption as well as the PL spectra showing the T-T* nature of this optical transition. The correlation between disorder and the broadening of optical features in polymers was the topic of many experimental [43, 531 and theoretical considerations [43, 54, 551. New insights into the excitonic nature of the optical excitations were achieved using time-resolved spectroscopy to study the energy transfer mechanisms [43, 53,561. The resulting picture turned out to be in contrast to the original opinion that optical absorption in polymers is due to interband transitions of free charges. Within the exciton picture polymers are regarded as an assembly of conjugated segments with a certain distribution of conjugation lengths [43], which correspond to undistorted parts of the conjugated chain. Optical transitions occur within these conjugated segments and the absorption bands are thus broadened due to the distribution of transition energies as well as the inhomogenous broadening of the poorly defined surrounding. The PL, however, can be quite well resolved, because energy relaxation within the broad density of available states will occur, down to a mobility edge, which is then a fairly defined level for the emission origin. However, different experimental reports on PT have not revealed a coherent picture of the role of the intermolecular (interchain) coupling onto the optical properties up to now. This may be mainly due to the low structural order and hence the missing fine structure details in the optical spectra, which prohibit accurate interpretation.

7.3 Oligothiophene model molecules 7.3.1 Molecular structure Within the framework of conjugated materials, the studies on different polymeric systems led to the conclusion that a great step forward in the understanding

7.3 Oligothiophene model molecules

365

could be made, if it were possible to use well-defined conjugated materials and hence oligomeric systems. The advantages of oligomers are quite obvious. On the one hand, the variation of the chain length can be studied in detail. On the other hand, even the conformation of the molecules (cis, trans, distorted) is much better defined than in the polymer so that a higher resolution within the optical spectra of the oligomers with respect to the polymer (Fig. 1) can be achieved. Despite the above mentioned variety of substituted oligomers we will describe the data on the prototypical nT whose molecular structures (n = 4-6) are shown in Fig. 2. At first glance it becomes apparent that even and odd numbered molecules have different point symmetry namely CZhand C2",respectively. However, the symmetry assignments are only valid for planar molecules and in fact the planarity is also important for photophysics. Although certain torsion angles between adjacent thiophene rings were found in ab initio caculations on isolated molecules [57, 581, the nonplanarity decreases with the number of thiophene units (2T < 30°, 6T < 10'). Numerical results for the ring twisting depend strongly on the level of approximations within the calculations and sophisticated approaches to take electron-electron interactions into account give rise to a further planarisation of the molecules [58, 591, but within the ground state a certain twisting between adjacent thiophene rings seems possible. In the solid state, the planarity was investigated by X-ray diffraction (XRD) techniques. Rietveld refinement methods investigating the molecular structures gave evidence for planar molecules with adjacent thiophene rings in trans-position (torsion angle < lo') within the solid state [60, 611. The planarity of the molecules in the solid state was also confirmed for a-substituted oligothiophenes by X-ray investigation [22, 621.

4T

5T

6T Figure 2. Schematic molecular structures of the model compounds 4T, 5T and 6T

I

20000

30000

v/crn”

400

Figure 3. Absorption and PL spectra of 2T-6T in dichloromethane at room temperature; reproduced by kind permission of the author [63].

7.3 Oligothiophene model molecules

367

7.3.2 Singlet states 7.3.2.1 Assignments The basic information about the electronic properties of the isolated organic molecules can be achieved by studying the optical spectra in solution. Figure 3 shows typical absorption and PL spectra for the n T series with n = 2-6 in dichloromethane at room temperature [63]. For all five molecular compounds at least two absorption bands of different electronic origin can be found. The lowest absorption band (I) shifts to lower energy with increasing chain length due to the increase in conjugation (section 7.3.2.2). The higher energy band (IV) around 40 000 cm-' shows no spectral shifts with varying chain length and was therefore attributed to a transition localized on a single thiophene ring [64]. For the longer molecules (n = 4-6) two additional excited states (11, 111) can be seen as weak shoulders in the absorption spectra. The assignment was further based on fluorescence anisotropy measurements which revealed changes in polarization behavior exciting at the respective spectral positions [63, 641. The PL spectra in Fig. 3 are obviously much better resolved than the absorption spectra and show vibronic progressions, which are quite similar to those observed in PT (Fig. 1). The relative intensity of the highest energy peak increases upon growing chain length. These changes in the vibronic progression are well understood in terms of the Franck-Condon principle [3]. The intensity distribution among the vibronic transitions for an allowed electronic transition is determined by the change of the molecular geometry (normal coordinate) upon excitation, which is characterized by a dimensionless parameter (often termed as Huang-Rhys parameter). For longer oligomers the geometrical relaxation upon excitation may take place over a larger number of carbon bonds and hence the relative change per bond decreases, thus leading to a weaker intensity of the higher vibronic transitions. The better resolution of the PL spectra with respect to the absorption spectra was attributed to a planarisation of the molecular structure in the excited state and hence a well-defined initial state for the PL, which was predicted by quantum chemical calculations [65], whereas the broad absorption was interpreted as due to the slight twisting of adjacent thiophene rings in the ground state [58, 631. Upon introducing the nTs into a rigid surrounding (e.g. frozen solution, matrix isolation), the torsional degrees of freedom become frustrated and hence the spectral resolution increases, especially for the absorption band (section 7.3.2.4). One of the interesting results comparing solution at room temperature and matrix isolation is the fact that absorption maximum at room temperature does not represent the 0-0 transition, but can be at considerably higher energy (7~2000cm-') [63, 661. Within the given molecular geometries the polarization properties of the observed transitions are of great importance. The directions of the optical transition moments were determined experimentally by using stretched polyethylene sheets for orienting the molecules [67, 681. The spectra gave evidence for the polarization of the lowest optical transition along the stretching and thus along the long molecular axis with high dichroic ratios of more than 10 for 3T-6T [63, 671. It is worth noting that although the C2h-symmetryof the even-numbered nT does not distinguish between

368

7 Electronic Excited States of Conjugated Oligothiophenes

a polarization along the short and the long in-plane molecular axis for a B,-A, transition a priori, all the experimental studies confirmed the long axis polarization of the transition moment for the lowest excited state [63, 67, 691. 7.3.2.2 Chain length dependence In the following we discuss the chain length dependence of the first optical transitions (Fig. 3) for the nT in solution, which was a topic of numerous investigations from experimental and theoretical points of view. Figure 4 shows the transition energies [70] of the unsubstituted nT as well as P-substituted oligothiophenes in solution as a function of chain-length [71]. The studies of the groups of Oelkrug et al. [71,72], Rentsch et al. [73], Becker et al. [58] and the highly-resolved spectroscopy on matrix-isolated oligothiophenes, carried out by Kohler et al. (2T [74], 3T [75], and 4T [76]) all revealed a very similar behavior for the energy of the lowest lying electronic transition. For the shorter oligomers (n < 6) a rather linear dependence on the reciprocal chain-length (l/n) is observed, whereas the higher oligomers tend to depend on the chain-length more weakly. Recent experiments on alkyl-substituted oligomers with up to 16 thiophene rings (Fig. 4) [24] confirmed that a limit of approximately 18 000 cm-I can be expected by extrapolating to the infinite chain in solution [71]. The expected value for solid polythiophenes will thus be around 16 500cm-', since the screening of the excitation in the solid state leads to a red shift with respect to the solution (section 7.4.2). This value is also in good agreement with the experimentally deduced transition energies of PT (Fig. 1).

classical model 0 0 0 2.

F a, a,

c

.c

1 4 4 , 0,o

, 0.1

,

, 0,2

,

I

03

.

I

0,4

,

,

,

0, 5

reciprocal chainlength ( l h ) Figure 4. Optical transition energies [70] of oligothiophenes in dichloromethanesolution as a function of reciprocal chain-length (l/n) [71]. In addition to the unsubstituted nT (Fig. 2), the data for dodecyl-substituted oligothiophenes are included (DDnT) [71]. The dotted line shows a linear extrapolation (l/n -+ 0), whereas the solid line corresponds to a model calculation according to [79]. Note the difference for small lin values.

7.3 Oligothiophene model molecules

369

Nevertheless, it is important to mention that the behavior of the transition energy as a function of chain-length is not linear in l/n as is typically expected as a first approximation [58, 771. The linear behavior would be expected according to simple free molecular orbital calculations [78]. However, the non-linearity is wellknown for other linear conjugated systems with alternating single and double bonds [78] and different classical [2, 791, semiclassical [41], and quantum mechanical descriptions [SO] have been derived for this functional behavior. This nonlinearity in 1/ n reflects the polaronic character of the strongly bound electron-hole pair, which limits the delocalization of the excitation. This is especially important for longer oligomers, and obviously linear extrapolations of the transitions energies will lead to underestimations for the transition energy, which do not describe the electronic properties of PT properly. From Fig. 4 it becomes apparent that the optical transition in T6 is already very close to the values extrapolated for PT (compare with Fig. 1) and thus it is generally acknowledged that the electronic excitation in 6T is already very similar to that in PT [58, 811.

7.3.2.3 Nature of the lowest singlet transition Theoretical considerations on the first electronic singlet transition started from the knowledge on the analogous class of molecules, the polyene oligomers (oligoenes). The most interesting feature in the optical properties of the oligoenes is a crossing of states representing different symmetries as a function of chain length. The states of interest within C2h symmetry are of 1'B, and 2'Ag symmetry. As the number of double bonds increases a crossover of the lowest transition occurs from an allowed (B,) to a forbidden transition (Ag). This 21Ag state was mainly attributed to a doubly excited electron configuration with two electrons in the lowest unoccupied

LLMO-

+ -.

- --t

HOMO+-

-

++

1Ag

Ag2

Ag3

1Bu

Figure 5. Model for the electronic structure of linear conjugated systems. Simple descriptions for excited molecular states have to include doubly excitations (A,2) as well as higher excitations (A,3) to reproduce the transition energies as a function of chainlength [41]. Reproduced from Ref. [41] by kind permission of the American Institute of Physics.

370

7 Electronic Excited States of Conjugated Oligothiophenes

molecular orbital, as illustrated in Fig. 5 [41]. Due to the similaritiesin the molecular structures between the oligoenes and the nT, if sulfur atoms are neglected, theoretical approaches on oligothiophenes also predicted a crossing of states of different symmetries for oligothiophenes for n > 6 [74] or even at shorter chain lengths [82]. Calculating optical transition energies is generally a complex task, since electron correlation plays an important role and simple one electron pictures (e.g. Huckel methods) do not describe the excited states properly [78]. From the investigations on the oligoenes it turned out that the crucial point concerning the level crossing is the correct description of the configuration interaction and that the inclusion of multiple excitations is very important (Fig. 5) [41]. Very recently different groups tried to improve their theoretical approximations to include the latter properly [58, 73, 831. However, the agreement of calculations, taking account of multiple excitations, with experimental data is still insufficient to predict the level crossing very exactly [73]. From the experimental point of view it is well established that no level crossing occurs for oligothiophenes in solution and that the lowest excited singlet state corresponds to B, symmetry [58, 711. This was deduced from large extinction coefficients (Fig. 3) yielding high oscillator strengths ( f ) for the longer nT ( f M 0.6 for 6T), the smooth behavior of the transition energies (data in Fig. 3), and was further corroborated by studies on the fluorescence yields and fluorescence lifetimes [58, 84, 851 of the nT (2 < n < 8). The experimental data on alkylated thiophenes even show that the assignment holds for longer oligomers (16 thiophene units) [71], so that up to now there is no evidence for a level crossing in oligothiophenes.

7.3.2.4 Franck- Condon coupling In the optical spectra of Figs 1 and 3 it is apparent that Franck-Condon (FC) vibronic coupling plays an important role in the photophysics of the lowest singlet absorption. All PL spectra show a vibronic structure with a spacing of about 1470cm-' which is fairly independent of the chain length. Vibrational analyses of nTs have been carried out [57, 861 and the vibrational energies and assignments are well understood. The ~ 1 4 7cm-' 0 vibration is a symmetric carbon-carbon stretching mode which involves both the nominal single and the nominal double bonds [57] and can thus be termed as a ring breathing mode. Although this is the prominent vibronic mode in the PL spectra, even further insight into other coupling modes and the strength of the FC vibronic couplings could be found for the nT in matrix isolation at very low temperatures (4.2K) [74-761. Figure 6 shows an example of the PL and PL excitation spectra for 4T molecules in tetradecane matrix [76]. Birnbaum et al. found five fundamental modes coupling noticeable to the electronic transition which are exemplarily listed for the 4T molecule in Table 1. From these data two important vibronic features could be deduced. On the one hand it is quite remarkable that the prominent ring-breathing mode in emission changes its frequency from 1478cm-' to 1235cm-' in absorption [76]. This is due to the different (more quinoid) electronic structure and hence the different bonding

7.3 Oligothiophene model molecules

37 1

+

+

+ +

10

5

I

+

!9 4588

4688

4788

4088

4880

Angst r o m t

+

+

15

4200

4388

+

+

+

I

4400

4508

Rngst roms Figure 6. Highly-resolved PL (a) and PL excitation spectra (b) of 4T in tetradecane at 4.2K. The spectra. The vibrational finestructure was assigned to the coupling of five fundamental Vibrational modes, which are listed in Table 1. Reproduced from Ref. [76] by kind permission of the American Institute of Physics.

situation in the excited state where even charge transfer to the sulfur atoms may play a major role [58, 761. On the other hand the relative intensities of the vibronic lines change from the emission to the absorption spectrum so that the often seen mirror relationship between the emission and the absorption spectra is not quite fulfilled for the nT.

372

7 Electronic Excited States of Conjugated Oligothiophenes

Table 1. Assignments and vibrational frequencies (in cm-') of the strong coupling vibrational modes for 4T in tetradecane according to [76]. v1

v2

v3

v4

v5

deformation

deformation

deformation

ring-breathing

C=C stretch

327 333

703 688

1478 1235 !

1531 1551

Emission 162 Absorption 161

7.3.3 Triplet states Although efforts have been made on the triplet states for the nT in solution, clear evidence for the location of the lowest triplet states through phosphorescence has only been reported for 1T with a O-O-transition at 27 600 cm-' [58]. Longer oligomers have also been investigated, but no phosphorescence was observed, either for the nT [58, 63, 851 or for alkylated oligothiophenes [87]. Scaiano et al. used solvents containing heavy atoms and reported the value for 3T to be around 13900cm-' [88], whereas Becker et al. deduced a value of 12700cm-' for 3T from photoacoustical investigations [SS]. The energetic separations between the first excited singlet and tri let states were deduced from very few data to be about 10 000 cm-' to 12 000 cm- [58], which is in good agreement with the high level configuration interaction calculations of Beljonne et al. [83, 891. These calculations predict a very weak dependence of the triplet energy on the chain length due to a stronger confinement of the lowest triplet state in the nT compared to the lowest singlet state. The value for the lowest triplet in the case of 6T was predicted at about 13000 cm-' . However, clear assignments about the position of the lowest triplets in nT were found recently in the solid state by measuring the delayed fluorescence in nT crystals [90] and will be described below (section 7.4.3.6). Higher triplet states may be even more important for the nT, since the main nonradiative channel in solution was found to be the intersystem crossing (ISC) to the triplet. Since the lowest triplet is well separated from the lowest singlet state the ISC is most probably due to some higher triplet level T, which is energeticallyclose to the S1 state. The lifetime of the lowest triplet states were measured by transient absorption in different solutions yielding lifetimes between 20 ps for 7T up to 104 ps for 2T in benzene [58]. The efficient triplet population was determined by measuring the singlet oxygen yield in solution [58, 841, where a high singlet oxygen efficiency correlated very well with the low fluorescence quantum yields for the shorter oligomers (n < 4), and vice versa for the longer oligomers (n > 4). Summarizing all the optical spectroscopy on isolated molecules, 6T can well be assumed to be a model compound for longer oligothiophenes and especially for polythiophene from all the experimental data as well as theoretical approaches. The similarity between 6T and polythiophene was not only revealed by the transition energies, but also appeared from many other spectroscopic experiments, e.g. photoinduced absorption (PA) spectra [911, or photoemission spectroscopy on the valence levels [92], and have been discussed elsewhere [29, 301.

P

7.4 Solid state properties

373

7.4 Solid state properties 7.4.1 Molecular packing Although we pointed out in the last section, the similarity between long oligomers as 6T and the polymer there is a main difference in the solid state due to the different degree of order and the difference in molecular packing. As described above, the scenario of polymers can be well understood by regarding them as an assembly of weakly interacting molecular units, corresponding to undistorted parts of the polymer chain with appropriate conjugation lengths [43]. This weak interaction is due to the van der Waals forces governing the molecular packing if there are no polar substituents to the polymer. Naturally, these van der Waals forces also determine the packing of the nT [93], but due to their well-defined molecular shape they form densely-packed and highly-ordered structures. This self-aggregating and self-ordering tendency of organic molecules is well known [94] and has been the basis of the huge efforts which have been made towards a controlled growth of various organic materials in organic single crystals [15] or for epitaxial layers on inorganic substrates [95]. Such efforts have also been undertaken in the case of the nT and resulted in the growth of single crystals of 6T [61] and 8T [96] and epitaxial thin films of 4T on Ag(ll1) [97-991. Furthermore, even a layer by layer growth has been observed for 5T on SiOz [30, 1001. Due to the low solubility of the longer nTs, single crystals could only be successfully grown by using a sublimation technique [15], the so called Lipsett technique. For 6T [61] and 8T [96] single crystals (plates) of macroscopic dimensions were obtained with a length of a few millimeters and a thickness of some tens of microns. a

L

Figure 7. Structure of the 6T single crystal unit cell [61]. The a, b and c-axes are indicated as well as the molecular geometry: long in-plane axis (L), short in-plane axis (M) and (N) perpendicular to L and M.

374

7 Electronic Excited States of Conjugated Oligothiophenes

The structural analysis of the crystal packing reveals a herringbone arrangement as a basic feature which is shown for the example of 6T in Fig. 7 1611. The intermolecular interaction of the rigid-rod-like oligomers leads to a herringbone structure with a parallel arrangement of the long molecular axis within a monoclinic unit cell of P2,,, symmetry. The herringbone angle of about 66" can be interpreted as a repulsion of the n-electron clouds of adjacent molecules and thus only a small overlap of the n-orbitals can be expected in the solid state. Nevertheless the n-overlap withn the bc-plane will be much larger than along the a axis, so that any transport (excitation or charges) will show almost a two-dimensional character. The packing scheme of the single crystal is similar to that which was determined previously by X-ray studies on polycrystalline films of 4T, 5T, and 6T [12, 60, 1011. The unit cell parameters of different samples (polycrystalline film [60] or single crystal [61]) are nearly identical, but the planarity of the 6T molecule seems to be more pronounced in the single crystal [61]. However, some differences in the structures may occur, depending much on the preparation conditions, e.g. rate of growth [lo21 or temperature [103]. Polymorphic structures were commonly observed for the organic crystals [104], and polymorphism may be taken into account for an understanding of spectroscopic features. A further, more detailed analysis of the crystal structures will be presented in another chapter within this book. Finally we want to note that the crystal structures of the smaller oligomers , e.g. 3T [lo51 and 2T [106], differ from those observed for the longer nT with n = 4-6 and n = 8.

7.4.2 Theoretical approach 7.4.2.1 The exciton concept and the lowest excited state in 6T

Before we focus on the experimental results on the solid state properties of the nT it is worth summarizing the basic ideas concerning molecular excitations in the solid state. At least four major differences must be expected for the spectral properties of the molecule in a crystalline surrounding [3]: (i) shift of the optical transition relative to the isolated molecule (solvent shift); (ii) splitting of spectral lines with a corresponding change in the polarisation properties; (iii) variation of the oscillator strengths and the selection rules; (iv) change in the intermolecular vibrational frequencies and appearance of intermolecular lattice modes (frustrated rotation, translation). Optical excitations in molecular crystals are well known as Frenkel excitons and the detailed descriptions have been derived by Davydov [4] and Craig and Walmsley [5]. Molecular excitons resemble very much the optical properties of the isolated molecules, since the exciton is confined on one molecule and only the weak interaction with the surrounding molecules leads to the formation of a collective excitation. This is contrary to the large radius Mott-Wannier excitons in conventional semiconductors, where the electron and the hole are typically loosely bound with

7.4 Solid state properties

375

energies in the order of tens of meV [107], and the strong binding between the electron and the hole is one of the main characteristics of the Frenkel exciton. One of the important conclusions from the concept of molecular excitons is that the optical band gap only corresponds to neutral excitations and does not reflect the energy levels of charged excitations. An extended literature can be found on the prototypical studies on molecular excitons in anthracene crystals (see [3, 108, 1091 and refs. therein), and in the following we will present a brief summary of the main features (i-iii) of the exciton concept, which may be important for the interpretation of the experimental results on the nT. Assuming weak interactions between N identical molecules leads to a Hamiltosimply consisting of the sum of the individual nian operator for the crystal (H,,,,,,) molecular Hamiltonians (H,) and the sum of the weak intermolecular interactions ( Vnm)between molecule n and m, where we consider only electrostatic interactions in the following [3]:

Following the tight-binding approximation a good description of the ground state crystal wavefunctions Qcrystal is then given by the product of the single molecule wavefunctions (P:, N

n= 1

In this representation the wave functions of a molecular excitation at site i can be written as: nfi

but since all N molecules are identical, there are N degenerate excited states if the intermolecular interaction is zero. Since the Hamiltonian includes the interaction, this degeneracy in the crystal can be lifted by using linear combinations of the degenerate wavefunctions. For a further simplification it is important to consider the translational symmetry properties of the crystal. Since the molecules at sites R, underly translational symmetry in the ordered solid, the electronic wavefunctions will also be symmetric with respect to translation (periodic boundary conditions). These symmetry requirements are solved by constructing Bloch-wavefunctions which are then characterized by a wavevector k in the reciprocal space, where p = hk corresponds to the momentum of this exciton [5]:

Since the lattice spacing is small compared to the optical wavelength, adjacent molecules can only be excited in phase, which means in terms of the wavevector k that an additional selection rule has to be considered, the conservation of momentum: k = 0 (for a further discussion of the selection rule see: [lOS]).

376

7 Electronic Excited States of Conjugated Oligothiophenes

In cases where there is more than one molecule per unit cell (e.g. anthracene: 2 molecules, 6T: 4 molecules) and the molecules are related by symmetry operations, the crystal wavefunctions are constructed out of the subsets of non-equivalent molecules which leads to representations of the crystal states which are the symmetric or the antisymmetric combinations of the subset wavefunctions [3]. These wavefunctions do then represent different irreducible representations of the crystal space group and in the following we will sketch the theoretical ideas for the example of 6T. The lowest molecular singlet transition for 6T is from a state of A, symmetry to a Bu-state. Due to the C2hsymmetry of the crystal point group the symmetry representations are the same as in the molecular symmetry frame and for clarification we will use lower case letters with respect to the crystal framework. Since the site occupied by a 6T molecule has no special symmetry operations, the site symmetry is GI. Withm the unit cell (space group P2,,,,) there are 4 molecular sites which are related to each other by different symmetry operations: 1: identity, 2: inversion, 3: glide plane, 4: two-fold axis. These symmetry operations correspond to the CShfactor group whose irreducible representations are displayed in Table 2 [110]. The molecular wavefunctions have thus to be written as follows, to require the symmetry properties of the crystal (see characters in Table 2) [l lo].

;

@ag(o) = (@Subset I f @'Subset2 f @Subset3

-k

+ @Subset41

@Subset2

-

@Subset3 - @Subset41

= (@Subset1 - @Subset2

-

@Subset3

+ @Subset41

@bu(O)= ;(@Subset1 - @Subset2 f

@Subset3

-

@bg(O) = ;(@Subset1

;

aSubset4)

The upper two wavefunctions belong to gerade crystal symmetries a, and b, and are thus forbidden transitions, whereas the lower two states are of ungerade symmetry, optically allowed. However, the a, component will be visible with a polarization parallel to the b-crystal axis whereas the bu component will be polarized within the ac-plane of the crystal. The corresponding energies can be then written as following:

+ + 1 1 1 + 112 + 1 1 3 + 114

Eag(0) = EO D

+ + + 112 1 1 3 Ea,(O) = Eo + D + 111 - - + Eb,(O) = EO+ D + 112 + Ebg(0)= EO D

111

-

112

111 -

114

113

11,

113

114

Table 2. Characters of the irreducible representations of the CZhgroup. Irreducible representations

E

A,

c::

1

1

AU

1

1

B,

1 1

BU

-1 -1

I

Real space coordinates

U

1

1 b

-1

-1

1 -1

-1

1

a,c

7.4 Solid state properties

377

The transition energies are determined by the isolated molecular transition energy Eo, corresponding to the unperturbed Hamitonian. D represents the difference in the van der Waals interactions in the ground state and the excited state (solvent shift (i)). Typical solvent shifts are in the order of 1OOOcm-' (naphtalene: D = 3 4 0 0 ~ m - lanthracene: ~ D = 2300cm-') [109]. The excitonic bandstructure is constructed upon the interactions of the translationally equivalent molecules (II1) and the non-translationally equivalent molecules (ZI2, Z13, II4). The observable excitonic optical transitions have to be taken at k = 0 and thus the differences in the transition energies (the so called Davydov-splitting (ii)) differ due to the non-translationally equivalent interactions. Since the different components correspond to electronic states of different crystal symmetry, the polarization properties for the transitions are different as well (iii). It is worth noting that in a molecular crystal the transition moments of the individual molecules are no longer the principal directions for the optical transitions, but the crystal symmetry leads to a polarization with respect to the symmetry axes of the crystal. Any numerical treatment which could be applied to molecular crystals needs to take account of the intermolecular potential, which governs the resonance interaction. For a singlet transition in a molecule, the excitation can be well described by the transition dipole moment M . Therefore the first order approximation to the intermolecular potential Vn, is through dipole-dipole interaction: eL (xnxrn + YnYm- 2 z n z m ) . R3 The z-axis is connecting the centers of the molecules n and m which are at a distance R (x, y , z are orthogonal coordinates of the molecular sites). The non-equivalent interactions thus resemble dipole sums and can be written as: vnm

M-

Lm(0) =

c

(cp,*P*I~nrnIcpncp%

nfm

where the molecules n and m belong to the different subsets. The Znm(0)term is proportional to the square of the transition moment: M 2 and thus large splittings have to be expected for allowed transitions with high oscillator strength. Unfortunately these dipole sums do not converge rapidly, since the interaction depends on the distance with RP3while the number of molecules within the surface of a sphere increase with R2 so that a detailed understanding of the shape of the integration volume is required to obtain a numerical calculation for the excitonic bands [4]. Although Philpott and Lee [1 111 developed methods to obtain very reasonable results for the anthracene splittings, the general validity of their assumptions are not established yet. Furthermore, the interaction potential Vn, described above corresponds to a sum of point dipoles, which is obviousby a very rough approximatiqn when rigid rod like oligomers with a length of 20 A are at a distance of about 5 A. Further corrections by including extended dipoles [112], higher multipoles [5], conduction band effects have been suggested [ 1 131. The calculations and experimental verifications are much easier for triplet excitons where the intermolecular interaction is due to electron exchange, since due to its short range the calculations can be restricted to nearest neighbour interactions [114]. However, this is not the case for

378

7 Electronic Excited States of Conjugated Oligothiophenes

a,

Molecule

Crystal

Figure 8. Davydov-splitting for the lowest singlet state (L-polarized) in 6T according to simple dipole calculations (Table 2). Note the degeneracy of the gerade and ungerade states [110].

long range interacting singlet transitions and simple nearest neighbour considerations will be not sufficient [4]. A detailed description of the numerous experimental investigations on molecular crystals and their comparability to theory can be found in the references [3, 1081. Within the framework of point dipole calculations one is now able to calculate the expected splittings by using the geometric arrangement of the single crystal. The calculations were carried out using Ewald sums [115] and the crystal structure of Fig. 7. The main results are displayed in Fig. 8 and Table 3 [110]. The molecular transitions were taken to be long axis polarized (L), in plane short axis polarized (M) and normally polarized (N) with a transition dipole moment of M = 1 debye. For the lowest lying excitation, which is L-polarized, a huge splitting is calculated between the a, and b, crystal-components with a very large polarizaton ratio (infinite). The expected splitting for 6T may be even larger than 10 000 cm-' (Table 3), since the deduced transition dipole moment seems to be more than 1 debye [63, 1101. Furthermore, it is very interesting that the gerade and ungerade states are degenerate, which may give rise to some peculiarities within the optical spectra (section 7.4.3.4). Although the numerical values should not be taken too literally at this stage, it is important that the calculations predict a large splitting (lowest singlet in anthracene: 200cm-I) with an ordering of the states where the lower allowed Davydov component is of a, character and hence polarized along the monoclinic b-axis and that the higher component is of b, symmetry, polarized within the ac-plane. This has been pointed out earlier to be a direct consequence of the exact parallel arrangement of the 6T molecules within the crystal [97, 1161. Table 3. Results of the Ewald sum point-dipole calculations, assuming a transition dipole of 1 debye for molecular transitions along the L, N, and M axis according to Fig. 7. The dipole sums were carried out for the different polarizations (energies in cm-]) and the polarization ratios were also calculated for the bc face [l lo]. Molecular transition

ag

a, (b-pol.)

b,

b, (a,c pol.)

pol.-ratio

/I L

-2149 -2173 -3896

-2149 -2173 -3896

6369 -564 1847

6369 -564 1847

Ic/Ib + cc Ic/Ib = 2.26 Ib/I, = 3.20

11 M // N

7.4 Solid state properties

379

7.4.2.2 Higher transitions - extended states

The above summarized concept on collective excitations in molecular crystals does also apply for higher excited states. Very detailed informations about the excitonic properties of higher states were derived for anthracene [5] and other polyacenes [5,108]. Especially the 40000 cm-' transition in anthracene is an interesting example for the Davydov splitting, since a very large splitting of more than 10 000 cm-' can be observed for that very intense transition [5]. Up to this point the discussion has been focussing on neutral excitations where the charges are not separated, but localized on the same molecular site. Excitations where a separation of charges occurs, also exist and are termed as charge transfer excitons (CTE). However, due to their weak optical absorbance, which is caused by the small overlap of the localized ground state wave function with the delocalized excited state wave function, the experimental identification of CTE is complicated. The manifold of singlet absorption bands with large extinction coefficients usually prevents a direct observations of CTE so that electric field modulated spectroscopy has to be used [117]. Nevertheless, CTE have been identified for a large number of conjugated molecules and the corresponding energies were in all the cases much higher (anthracene: 0.8eV, or even more ) [lo91 than the respective first singlet absorption bands. The charge transfer states are in quite close relationship to the Mott-Wannier excitons within conventional semiconductors. This becomes quite obvious by looking at the energy levels of CTE which can be written as following: ECT

= EBand

eL

Gap - ;

where E , c0 are dielectric constants. 47r~~~r

EBandCaprefers to the energy of totally separated charges (of charge e) in the conduction and valence band. In both cases (CT as well as Mott-Wannier excitons), the energy of the band gap is lowered by the coulombic attraction of the opposite charges. In the case of CT excitons the distance r corresponds to the separation between two distinct molecular sites, whereas in the Mott-Wannier case r corresponds to the radius of the hydrogen-like bound electron-hole pair. For a large distance between the two charges, CT excitons can be treated similar to Mott-Wannier excitons [3, 1181. If the separation between the charges further increases, CTE should approach the states corresponding to the conduction and valence band in organic crystals. Therefore the conduction band can be located by an extrapolation of CTE energies to large separation distances. However, due to the nature of organic solids with a small overlap of the electronic wavefunctions of neighbouring molecules, a band-like description of the conduction levels may not be valid especially at elevated temperatures. Deeper discussions of the conduction and valence band properties and the validity of band descriptions in real crystals at finite temperatures where vibrations play a major role, can be found in the references [3, 1091. In the case of 6T little efforts have been done towards a band like description of the excitonic or the charge carriers, because in most cases disorder dominated the experimental data [30] (see below) and hence a band like description did not seem applicable. However, Siegrist et al. [lo31 applied an extended Huckel theory to

380

7 Electronic Excited States of Conjugated Oligothiophenes

calculate the band structure for the four highest occupied and the two lowest unoccupied molecular orbitals of 6T. The crystal structure used for the calculations is slightly different from that in Fig. 7 but the herringbone arrangement and the layered stacking of parallel 6T molecules is very similar. From these calculations it was derived that the system only shows dispersion in two dimensions corresponding to the poor .rr-overlapof 6T molecules in the direction of the long molecular axis (a-axis in Fig. 4). The main result is the assignment for 6T to have an indirect energy gap between the conduction and valence bands of 1.95 eV (15 600 cm-'). The valence band maximum is located at the ??-point within the reciprocal space whereas the minimum of the conduction band is not. Optical investigations, however, did not show any band gaps as small as 15600cm-' (section 7.4.3). Furthermore, the authors point out that the structure of the uppermost valence band is quite similar to that in certain organic superconductors [119]. Nevertheless, charge transfer in the ground state is of basic importance in organic superconductors, which is missing in the case of pure 6T. By introducing electron accepting materials, the structural properties may change with respect to the pure material and therefore lead to a different valence band structure. The majority of the investigations on the transport properties of oligothiophene films, however, do not show any effects which could be related to a band-like conduction mechanism up to now. Different groups reported very low dark conductivities [120, 1211, low field effect mobilities [30] and a strong thermal activation behavior [120, 1221 for transport processes. Furthermore a strong influence of the structure and the morphology of the polycrystalline oligothiophene films establish a picture of noncoherent hopping transport where different trapping levels dominate the motion of the charges [ 1221. Although the mobilities of the charge carriers as well as the conductivity showed a strong increase with increasing chain-length of the oligothiophenes [121,123],the main rise for the observation of conduction phenomena is the unintentional doping of samples by exposure to ambient atmosphere [30,120]. In contrast to that, Torsi et al. [124] reported recently on the temperature dependence of field effect mobilities, which were attributed to a coherent carrier motion at very low temperatures ( T < 50K) [125]. In summary the experimental observations do not allow up to now to draw a clear picture about the conduction band levels for oligothiophenes. The various experimental investigations on organic molecular crystals [1091, and the recent experiments on ultrapure FETs of 6T [11,126,1271showed that it seems necessary to improve the sample quality in terms of order and cleanliness to study the influence of the intermolecular coupling on the motion of charges and hence on the conduction levels.

7.4.3 Experimental evidence for the nature of the lowest excited states 7.4.3.1 Structural and morphological aspects of polycrystalline thin films The preparation of thin organic films offers easy access to the solid state properties of molecular materials. Although many different film preparation procedures are

7.4 Solid state properties

38 1

possible, e.g. spin casting from solution or Langmuir-Blodgett preparation [ 1281, vapor deposition in high or even ultra high vacuum is one of the most favorable and commonly used techniques. The great advantages of the sublimation with respect to the preparation from solution are the superior cleanliness, reproducability, and the possibility to use various, even reactive, substrates. Due to the simplicity of the preparation most of the samples were prepared in high vacuum (HV) on poorly defined substrates as glass or quartz, since the use of a transparent substrate is highly desirable for optical spectroscopy. Nevertheless, in cases where superior film properties should be achieved, metal surfaces are especially wellsuited for film preparations in ultra high vacuum (UHV), since the adsorbatesubstrate interaction may lead to a higher structural definition of the films and the UHV conditions to a further reduction of chemical impurities [116, 129-1311.

Orientation It has been known for a long time that oligothiophene films can be prepared with a preferential molecular orientation with respect to the substrate [1281. However, this was primarily true for films prepared by the Langmuir-Blodgett technique. Since oriented thin films of the nT were considered as interesting models for nT crystals, many groups undertook efforts to produce oriented films by vapor deposition in order to extract their basic optical and electrical properties. Orientation, structure and morphology of thin vapor deposited nT films were therefore studied by various techniques, e.g. optical and infrared absorption [67, 1321, X-ray diffraction (XRD) [12, 101, 1331, atomic force microscopy (AFM) [loo, 134-1361. The orientational behavior of the nT, especially on dielectric substrates, was discussed in detail earlier [30] and therefore only a brief summary will be given here. On polar substrates, e.g. oxides such as glass, quartz and sapphire, the nT are preferentially aligned with the long molecular axis perpendicular to the substrate. Detailed XRD analysis allows the conclusion that the molecules show an angle of 32" with respect to the surface normal which corresponds to the relative orientation of a 6T molecule with respect to the a-axis in the unit cell and thus the b,c-crystal plane is parallel to the substrate [12, 61, 1331. The structure is polycrystalline and the crystallinity as well as the orientational perfectness are highly influenced by the preparation conditions, e.g. high evaporation rates may lead to disorder or parallel molecules with respect to the surface [133]. For the 5T molecule a perfect perpendicular alignment on S O z substrates was reported for the first few layers [137]. The situation was however quite different when metallic substrates were used [130, 1311. For 4T films on Ag(ll1) the covalent bonding of the .rr-system led to a commensurate structure in the first molecular layer [131, 1381. Within this bonding the .rr-system and hence the 4T molecule arranged parallel to the surface. This preferendially 'lying' orientation can be maintained also for thicker films of some tens of A [98,99]. A parallel orientation was also found in the case of 6T deposited onto a 2 x 1 reconstructed Au(ll0) surface [139].Other studies on 2T adsorbed on a more reactive surface, namely Ni(l1 I), showed that a strong bonding occurs which may even lead to a strong distortion of the 2T molecule [140].

382

7 Electronic Excited States of Conjugated Oligothiophenes

In general it may be summarized that on oxidic surfaces the nT prefer a 'standing' orientation whereas metallic surfaces lead to a 'lying' orientation, if the interaction between the molecules and the metal substrate is not too strong. The general relevance of the molecular orientation for the analysis of absorption spectra was discussed recently [141]. An increasing film thickness lead to a decrease in the degree of the orientational order in most of the experiments. Structure The XRD measurements also revealed the crystal structures [12, 60, 101, 1331 (discussed in section 7.4.1). It is worth noting that the nT for n = 4-6 crystallize with a very similar structure (Fig. 7), where only the a axis of the unit cell (related to the length of the molecules) changes. The herringbone angle is almost the same for all the investigated nT (= 60") [60, 1011 so that the application of the excitonic concept leads to analogous expectations for the Davydov splittings and polarization properties of the transitions (Fig. 8, section 7.4.2). This is very important, since experiments on different oligomers can then be compared quite well. Since XRD is not applicable for ultrathin films, low energy electron diffraction (LEED) is a method well-suited to investigate the geometric structure of those films and a typical LEED pattern for a 4T film on Ag(ll1) is shown in Fig. 9. The clearly resolved diffraction peaks show a commensurate structure [97] with respect to the substrate lattice and the 4T films are thus among the few examples of an epitaxial organic film preparation [99]. Although the interaction of the .rr-electronswith the surface leads to a covalent bonding of the molecules, there is still enough mobility of the molecule to find different adsorption sites and hence

-12

-08

-04

DO

04

08

12 12

04

0.0

9c

-0.4

-0'1

. ,b.

~

;>

@,

-*

-1 2 -12

-08

oQpo

B*

-04

00

0

04

;[-"

'(3

I3 08

-12 12

Figure 9. High resolution LEED pattern of an epitaxial 4T film on Ag(ll1) (nominal thickness N 20 A) at an electronenergy of 20 eV. The sixfold symmetry is due to the Ag( 11 1) substrate. The reciprocal lattice vectors a* and b* indicate one of the six symmetry equivalent domains [99].

7.4 Solid state properties

383

to arrange in a stable well-ordered configuration [98, 1311. The sixfold symmetry of the LEED pattern reflects the substrate symmetry which leads to six symmetry equivalent domains [99]. The domain sizes are mainly limited by the substrate surface quality, which was deduced from the spot profiles (Fig. 9). The diffraction pattern is attributed to a molecular arrangement which is very close to the parameters found in polycrystalline 4T films, although the structure displayed in Fig. 9 is commensurate (in second order) to the Ag( 111) substrate [97,99, 1161. Due to the close relationship of the geometric structure within the epitaxial 4T films on the Ag( 111) substrate to the polycrystalline 4T structure [97, 99, 1161 their optical spectra could also contribute to the discussion of the optical features within the solid state.

Morphology Electron microscopy [ 142, 1431, light scattering [143], and especially AFM investigations [loo, 134- 1361 revealed that the polycrystalline textures depend strongly on the preparation conditions. Biscarini et al. [134, 1351 studied these morphological aspects and the influence of the preparation conditions, e.g. growth temperature, for polycrystalline 6T films in detail. Figure 10 shows typical AFM images of 6T films deposited on freshly cleaved mica [134]. The images show an arrangement of tightly packed grains as long as the deposition temperature stays below 150°C [134]. These grains are quite isotropic and the average grain size also depends on the deposition temperature, where room temperature yields typical diameters of

Figure 10. AFM images of polycrystalline 6T films deposited on mica as a function of growth temperature. 10 pm x 10 ,urn frames show topographical images a t substrate temperatures of: (a) 22°C; (b) 75°C; (c) lOo"C, (d) 150°C [134].

384

7 Electronic Excited States of Conjugated Oligothiophenes

300 nm [134]. Above the threshold deposition temperature of 150°Cthe shape of the grains changes drastically into anisotropic lamellae [ 1341. From a detailed analysis of the film roughnesses as well as the grain sizes, it was suggested that diffusion of the molecules limits the growth at lower temperatures, whereas the number of available adsorption sites limits the growth at higher temperatures [135]. The crystallite sizes further depend on the film thicknesses [30,142,144], however, all the morphological studies show a polycrystalline nature with crystallite dimensions on the submicrometer scale. From that finding we deduce that grain boundaries may still play a significant role for the macroscopic properties of all the oligothiophene films. 7.4.3.2 Optical properties of thin polycrystalline films

The optical properties of thin nT films reflect very well the orientational aspects. Oelkrug's group determined the dichroic behavior [63, 67, 771 of the optical transitions within vapour-deposited films, and typical spectra of a 2.5 nm thin 6T film on quartz are displayed in Fig. 11 [63]. For the thin film the bands labelled I, 111, and IV can be well distinguished (assignment analogous to Fig. 3). By using s- and p- polarized light it becomes apparent that the bands labelled I and I11 are polarized perpendicular, whereas band IV is polarized parallel to the surface. Despite of the dichroism the overall shape of the lowest energy band changes drastically from solution to thin films. However, even if the most intense feature in orien-ted films is peaked around 28 000 cm-' the onset of the absorption band is located around 18 500 cm-' . If the sample thickness increases (Fig. 11b), the degree of orientational order decreases and the shape of the optical spectra change drastically for the lowest transition between 18 000 and 30 000 cm-' . Further investigations on the electronic transitions have used high resolution energy electron loss spectroscopy (HREELS), which is attractive, because dipole selection rules can be overcome [64, 771. These spectra also confirmed that the lowest energy transition is below 20 000 cm-' and interestingly the band shape in the HREELS spectra was not affected by the molecular orientation. In contrast to the low resolution found for the absorption spectra within most of the literature, Fichou et a f . [145] succeeded in a much better resolution of the vibronic structure for thin nT films. From the spectra it became quite clear that the electronic origin of the optical transition for 6T films must be located lower than 18 500cm-' (Figs 11 and 14). The spectral features for different oligomers (n = 4-6) were found to be very similar as expected from the spectra in matrixisolation [76], although the vibronic fine structures could not be fully assigned [145] at that time. For the very unusual bandshape of the lowest transition band (Fig. 11) different interpretations were offered. By applying the excitonic molecule model [146] (a twodimensional dipole picture), Oelkrug and his coworkers pointed first to an excitonic band structure [64, 671 with a width of lOOOOcm-' where the k = 0 states correspond to the highest energy levels within the excitonic band, whereas the bottom of the band is located at the zone boundary with k # 0. The apparent 'blue-shift' in the spectra would then correspond to the width of the excitonic band. A very similar interpretation was published later by Kanemitsu et af. [147], again neglecting the

7.4 Solid state properties

385

0 20000

30000 "lcrn-1 40000

'

i

I

(b) 20000

30000

v/c,n-l

40000

Figure 11. (a) Polarized absorption spectra of a 2.5 nm thick 6T film on fused silca. The spectra were taken under angle of 50" with respect to the surface normal in s- and p-polarization. (b) Average extinction coefficient of 6T films as a function of film thickness [63]. Note the enormous change for the first absorption band with increasing thickness. (a) and (b) reproduced by kind permission of the author [63].

three-dimensional crystal structure. However, in a later work Oelkrug et al. pointed out that the dielectric function has to be considered in more detail [141], since transmission spectra do not necessarily resemble only the absorption coefficient. By using the Fresnel equations for an anisotropic system they showed that the high degree of anisotropy, in combination with a typical anomalous dispersion above the transition, may lead to a blue-shifted band-shape without any change in the transition energies for the respective electronic levels [63, 1411. Fave et al. [66, 1481 attributed the band shape to the observation of a Davydovsplitting with a huge splitting of about 10 000 cm-' where the higher energy transition (bu)is much stronger than the lower one (a,). They further stated that the observed PL originates from some higher energy level of molecular A,-symmetry and not from the lowest a,-crystal component. However, this conclusion could up to now not explain the drastic changes upon variation of the film thickness and hence the molecular orientation. As stated above, the X-ray data on thin 6T films did not show a

386

7 Electronic Excited States of Conjugated Oligothiophenes

change of the geometric structure upon a variation of film thickness. If, however, the structure does not change, the Davydov-splitting should be independent of film thickness, which is in quite contrast to the optical data displayed in Fig. 11. In addition to the absorption, PL spectra were also analysed to disentangle the nature of the lowest excited state. The general luminescenceproperties of thin T6 films were quite poor and a typical example is shown in Fig. 12 [149, 1501. By going from solution to thin films the PL quantum yields decreases by three orders of magnitude [84] and apparently broad emission lines dominate the spectra. Even at very low temperatures the resolution of the optical spectra is rather poor (several 100cm-I) and spectroscopic details are smeared out. In most cases a considerable red-shift between the absorption and PL onsets and multiple different PL-components could be found within the spectra (Fig.12) [149, 1511. The main radiative decay channels were attributed to deep trap levels [149] or aggregates [S], which are strongly depending on the preparation conditions and film thicknesses (Fig. 12) [149]. By site selective PL-spectroscopy the positions of at least three trap levels could be located which are up to 2 000 cm-' lower than the absorption onset Wavelength (nm)

12

14

16

18

20

22

24

Wavenumbers ( 1000 cm-')

Figure 12. PL- and absorption spectra of 6T films on glass substrates at T < 10K. The lowest ~ . PL in thick films (>lo0 nm) is mainly based absorption band is located around 1 9 0 0 0 ~ m - The on a deep trap level (labeled ao) which is located at -17350cm-' [116]. (b) With decreasing film thickness a second component (arrows, bo located at =17 900 cm-') increases in the relative intensity with respect to component a. Both components are redshifted with respect to the absorption onset and show the same vibronic progression [149].

7.4 Solid state properties

387

[150]. For an elucidation of the optical solid state features a variety of experiments made it quite clear that, unless the poor sample quality could have been overcome, it would not have been possible to reveal the detailed information which are necessary. 7.4.3.3 Highly ordered systems However, very recently two different approaches towards highly-ordered oligothiophene systems had been undertaken, namely the growth of single crystals by Garnier’s group [61, 961 and the thin film growth by epitaxial preparation routines which was established by Umbach’s group [98, 99, 1521. In section 7.4.3.1 it was pointed out that the molecule-substrate interaction at the inorganic-organic interface is very important and that it may even lead to epitaxial preparation procedures for large organic molecules [131,1521. Based on that knowledge, two substrates were used frequently, namely highly oriented pyrolitic graphite (HOPG) [97, 116, 1531 and Ag(ll1) [98, 991. Cleaving of the HOPG leads to very inert, clean and flat surfaces which give rise to a very undisturbed growth of the probably polycrystalline material and due to its superior surface quality HOPG is most commonly used for scanning probe microscopies [154]. In the case of 4-6T the high structural film quality by using HOPG substrates led to highly-resolved optical spectra, which gave new insight into the optical fine structures [153]. Figure 13 shows the absorption and the PL spectrum of a 2.5 nm thick 4T film on HOPG [155]. The vibrational fine structure in PL turned out to be similar I

I

I

I

I

I

I

I

I

v4 I

.0 C v)

a

-c c

PL Excitation Reflection

I

19

20

I

I

I

I

21

22

Wavenumbers ( 1000 cm-’)

Figure 13. PL and absorption [155] spectra of a 2.5 nm thick 4T film on HOPG at 20K. Note the high vibrational resolution and the coincidence of the absorption and the emission onsets [153].

388

7 Electronic Excited States of Conjugated Oligothiophenes

to the one observed for matrix-isolated 4T [76, 1531, and corroborated the picture of almost undistorted molecules in the thin film. The 0-0 transition for 4T could thus be assigned at 21 025 cm-’, and was found in coincidence with the weak but clearly resolved lowest energy peak in absorption. Remarkably, no Stokes-shift (Fig. 13) was observed within the experimental resolution (=lo cm-’). This was interpreted as evidence for an allowed but weak transition, because for a forbidden transition one might expect a shift between the onsets of the absorption and PL spectra originating from the Herzberg-Teller coupling. The spectral shift would then correspond to the energies of the involved modes [40]. However, the fine structure which can be seen in absorption could not be analysed straightforwardly. Further insight was given by comparing the fine structures of different nT (n = 4-6) [116, 1531, whose spectra are shown in Fig. 14. It could be derived that the fine structure in absorption can be assigned to a coupling of the same modes which are present in the matrix-isolation (Table 1) [76, 1531, if it is assumed that in the case of 4T and 6T the progression starts not at the lowest energy peak (the 0-0’ transition in Fig. 14), but on the second one (0-0). The first two peaks in the absorption thus were attributed to different electronic origins. From these absorption spectra the pure electronic transitions and hence the lowest energy exciton states are clearly resolved at 21 025 cm-’ and 18 350 cm-’ for 4T and 6T, respectively. In both cases the PL originates from the same onset, if the films on HOPG are sufficientlythin [97, 1161. 7

J\ A

2 3

L

0

700 1000

2000

A E (cm-’)

Figure 14. Absorption comparison of 4T, 5T, 6T. All spectra are displayed with respect to the given energies, to emphasize the analogy within the fine structures. The positions of the vibrational modes are indicated and the first three fundamental modes are labelled (Table 1).

7.4 Solid state properties

389

Since no unusual features could be found in the absorption spectra at higher energies, the appearance of a doublet in the absorption spectrum (Figs 13 and 14) was tentatively assigned to the two permitted Davydov components where the smaller low energy peak resembles the a, component whereas the upper component is assigned to be the b, component [116]. The splitting of the first two lines were found to be 160cm-' and 120cm-' for 4T and 6T, respectively. Surprisingly, 5T did not show a similar feature in the absorption spectra (Fig. 14). This may be due to an effect of the different molecular symmetry or simply due to a very weak intensity of the lowest component which is beyond the experimental limits, but up to now the situation for 5T is not quite clear. Additionally to the here presented framework, the epitaxial 4T films on Ag(ll1) gave additional evidence for the assignment of the lowest electronic states. Although the PL and absorption spectra on Ag(ll1) appeared to be very similar to the ones for 4T/HOPG the relative intensities within the progression showed a strong dependence on the film thickness, which is displayed in Fig. 15 [99]. For very thin films (3 nominal layers) the fine structure can be assigned to a Franck-Condon progression for the 1470cm-' ring breathing mode (v4 in Fig. 15) in analogy to the isolated molecules. With increasing film thickness, the intensity of the first PL band (21 02518 700 cm-') decreases very rapidly and the vibronic coupling can then no more be described as a Franck-Condon progression [99]. Since no features of a trap emission could be found within these spectra, the unusual vibronic coupling was attributed to

16

I

I

18

20

22

Wavenumbers ( 1000 cm-') Figure 15. PL spectra of 4T on Ag(ll1) for two different thicknesses [99]. For comparison the PL of a 4T film on glass and a 6T single crystal are shown (Fig. 17). The upper three curves were measured at x40K, whereas the 6T spectrum was taken at 4.2K. Note the small intensity of the 0-0 transition with respect to the vibronic transition of the ring breathing mode v4 unless the film is very thin.

390

7 Electronic Excited States of Conjugated Oligothiophenes

a weakening of the pure electronic transition with increasing film thickness. For thicker films more and more molecular layers contribute to the collective excitation and hence the coupling becomes more important. Since for thicker films the oscillator strength decreases, Herzberg-Teller coupling becomes more important and therefore the relative intensities within the progression of the strongly involved ring breathing mode changes. This finding is important, because due to the very low intensity of the highest energy band in PL the 0-0 transition may be overlooked easily. A weak intensity for the excitonic emission origin was also found in the case of 6T single crystals (Fig. 1 9 , if the deep trap emission did not rule out the observation of the excitonic PL component [110,150]. This finding further corroborates the interpretation that the intermolecular coupling leads to the lowest energy transition for the nT films which is a very weak transition and can be identified with the a, component of the Davydov split bands [99]. Unfortunately the polarization properties of the transitions could not be determined due to the probable polycrystalline nature of the nT films on HOPG and due to the non-transparency of the HOPG as well as the Ag( 111) substrates. Therefore we turn now to the optical absorption spectra of the 6T single crystals, whch are still under investigation. Due to the high order within the molecular surrounding the spectra show again well-defined fine structures, as expected. The most intriguing information has been obtained on the polarization properties of the absorption. Figure 16 shows the polarized absorption spectra of a 6T single crystal which were taken with a wavevector parallel to the a-axis [150]. The polarization of light was then chosen to be parallel to the b-axis and perpendicular, respectively. At first glance the structures within the spectra are very similar to the ones obtained for the polycrystalline films in Fig. 14, but obviously the absorption spectra show strong differences between the two polarizations. The lowest energy peak is again found

polarizedftc

8000

-

A

1

17500

18000

18500

19000

polarized Ilb

19500

20000

Wavenumbers (cm-') Figure 16. Polarized absorption spectra of a 6T single crystal ( T = 30K) measured with the light

passing parallel to the crystal a-axis. The polarization was chosen parallel to the monoclinic band perpendicular (almost //c). axis (/b)

7.4 Solid state properties

39 1

at 18350cm-' and is strongly polarized along the b-crystal axis, as it would be expected for an a,-transition from the above presented considerations (section 7.4.2). The higher energy pattern was attributed to vibronic activity of totally symmetric modes. However this vibronic progression is obviously preferentially polarized along the c-axis, which means within the ac-plane of the crystal. In addition to the previous interpretation of the 18 350 cm-' and the 18470 cm-' spectral lines as a Davydov doublet which is in full agreement with the polarization properties, the same polarization behavior could be expected, if the line at 18 470 cm-' were due to a Herzberg-Teller coupled vibrational mode with an energy of 120cm-'. The further progression would then be built on the 18 470 cm-' line. In summary, it is not possible up to now, to decide whether the 18 470 cm-' line is of electronic or vibronic origin, but its polarization property clearly shows a strong c-polarization which would be expected for a b,-transition. Further interpretation of the single crystal absortion spectra is on its way, which focusses on the full range of the absorption spectrum to assure an assignment of all the features observed there (see notes added in proofs at Sec. 7.4.3.7). The PL spectra were also taken on the single crystal of 6T [l lo], and again a huge number of fine structured lines appeared, which are displayed in Fig. 17. The spectra were also taken in different polarizations (b, c) and at low temperature (5K). The highly-resolved peaks were assigned by a careful line position analysis [110]. Most importantly the spectra revealed that the prominent component in b-polarization origins at 18 330 cm-', which is very close to the lowest energy origin in absorption. 1,o-

098 -

03

0,7-

0,o

1

I

I

1

I

1

15000

16000

17000

18000

19000

Wavenumbers (cm-') Figure 17. Polarized PL spectra of a 6T single crystal (T = 5K).The polarizations are similar to Fig. 16 (//b and / / c ) .

392

7 Electronic Excited States of Conjugated Oligothiophenes

The b-polarization is again in agreement with an assignment of an excitonic a,-crystal component and similar to the spectra of 4T on Ag( 111) the hghest energy band (18 330-1 7 000 cm-') is weak. The vibronic fine structure is rather similar to the one observed for 4T films (Fig. 13) and the vibrational modes were further compared to off-resonant Raman-spectra which showed good agreement. In addition to this bpolarized component three (false) origins can be found in the crossed polarization which show the same vibronic fingerprint as the one described above. The origins of these emission components are located at 18 165cm-', 18085cm-' and 18025 cm-' (A, B, and C in Fig. 17), which is very close to the absorption onset and hence these features are either due to Herzberg-Teller coupling or may be attributed to well-defined and distinct X-traps [110].

7.4.3.4 Two-photon excitation Up to now we treated the one-photon allowed processes. However, the location of the excited molecular A, states is only possible by using two-photon excitation to conserve the parity of the states. Two-photon excitation (TPE) spectra were taken on isolated 2T molecules [74] where the molecular A, state was found to be more than 6500cm-' higher than the first excited B, level. In the solid state, TPE was only performed on thin film samples which did not show a very high structural quality. Periasamy et al. found a very sharp two-photon signal with an onset at 18 350cm-' [156]. In that work the one-photon PL spectra were quite poor in resolution and due to the deep trap levels which give rise to the PL in 6T films (Fig. 12) [149, 1501. Due to the deep lying emitting level the authors believed that the molecular A, state is located well above the B, state [156]. However, this interpretation has to be modified in the light of the highly resolved spectra of 6T on HOPG and of the 6T single crystal. Since it is now well established that the lowest allowed component is located at 18350cm-' [97, 116, 1501 the onset of the one-photon allowed and the two-photon allowed components are coincident. A comprehensive interpretation of the recent single crystal data and the TPE spectra is not straightforward. We assign the TPE signal in T6 not to the molecular 2'A,, but to the Davydov split a, state, which is expected to be degenerate with the one-photon allowed but weak a, state (section 7.4.2).

7.4.3.5 Extended states

The extended states were also studied in 6T thin films by electroabsorption (EA) [ 1571. This technique is well suited to derive the charge transfer states, since the separated charges are supposed to have a large static dipole moment and therefore should give rise to a strong first order Stark effect. By applying an electric field one should then be able to differ neutral from excitations where a charge separation occurs, since neutral excitations only show a second order and hence quadratic Stark effect [117,157]. The neutral excitations found in the EA spectrum correspond very well with the maxima of the absorption spectrum in the low energy range up

7.4 Solid state properties

393

to approximately 22 000 cm-' . The higher energy features had been attributed to charge transfer excitons and although it is not possible to determine the band gap quite exactly the lower limit of 2.78 eV (22 250 cm-') could be derived from the energy of the CTE [29,157]. This means that the lowest singlet state is well separated from the band states by more than 3200 cm-' .

7.4.3.6 Triplet states In section 7.3.3 we stated that phosphorescence in solution was only reported for 1T. In the solid state Xu et al. reported a phosphorescence for 3T (presumably in thin films) at 12 100cm-' [158, 1591, but since they found a similar feature for a film of poly-(3-hexylthiophene) there was considerable doubt about the assignement since the first triplet-state is also expected to vary with chainlength [58]. Clear evidence for the location of the lowest triplet states in the nT could be found recently by high-resolution spectroscopy on the delayed fluorescence of single crystals (n = 2,3,5) and polycrystalline films (4T) [90]. Figure 18 shows the triplet excitation spectra in the region of the 0-0 transitions at a temperature of 6K. The nT show apparently a large red shift with increasing chain length and a summary of the 0-0 transitions is given in Table 4.The red shift with increasing n was comparable to the one found for the singlet states in solution and was also attributed to an increase in conjugation. In the case of 3T two electronic origins were found, which were not attributed to a Davydov splitting but to two different sites (subgroups) which are present in the crystallographic unit cell of 3T [105], which differs

12000

14OOO

-

16OOO

v i cm-'

Figure 18. Low energy region of the triplet excitation spectra of nT in the solid state (n = 2-5). The 0-0 positions are listed in Table 4. Reproduced from Ref. [90] by kind permission of Elsevier.

Table 4. Location of the lowest triplet in solid nT (energies in cm-'); from Ref. [90]. 2T [90]

3T [90]

4T [90]

5T [90]

6T presumably

18224

15 170/15090

13445

12 750

12 200

394

7 Electronic Excited States of Conjugated Oligothiophenes

from the structure found for 4-6T [60]. Up to now, no reports are known on the Davydov splittings in triplets, which are generally much smaller than the singlet splittings due to the much weaker interaction [114].

7.4.3.7 Excited states ordering Summarizing all the information on the optical properties in 6T a tentative energy level diagram can be drawn, which is shown in Fig. 19. The lowest excited triplet state will be situated around 12 200 cm-' (sections 7.4.3.6, 7.3.3). The first allowed excited state is the (weak) a,-Davydov component of l'B, at 18 35Ocm-' [97, 110, 1161. At the same energy the two-photon signal (section 7.4.3.5) showed a state of gerade symmetry, which we now interpret as the lower a,-Davydov component of 1'B,. The charge transfer states and the conduction band are at least 0.4eV (3200cm-') higher in energy, so that the lowest excitations should not be influenced by the extended states. The intense photocurrent action, which was found to be in analogy to the absorption spectra was related to the breaking of exciton pairs due to disorder and traps and does not correspond to the conduction band [160]. The transition at 18 470 cm-' shows the character of a b,-transition although it seems not clear if this is due to an unexpected small Davydov-splitting [116, 1531 or due to vibronic (Herzberg-Teller) coupling [16 11. The most interesting question concerning the Davydov splitting is the location of the allowed b,-component in the solid state, which has already been assigned between 18 470 cm-' [ 116,1531and 28 000 cm-' [148]. Further experimental evidences can be found by investigations on high quality single crystals which are now available. Theoretical considerations will also be very important, however simple point-dipole calculations may eventually not be able to provide good agreement, since the strict Electronic states in solid 6T

4ooo01

Triplet 30000

-

Singlet

Extended

.., ..... ? b, (28,000)[I481 optical effect [I411

-

- Conduction level [157] CTE (22,250) 11571

'E

......,. -bu(18,470)[116]

2 20000* P

a, (1 8,350);ag (18,350)TPE

a?

5

T,

1

-

10000

Figure 19. Schematic ordering of the electronic states in 6T summarizing the present datasets.

7.5 Polarized electroluminescence

395

parallel arrangement of dipoles within the crystal structure of the nT resembles a very extreme situation which did not occur in the former studies on the polyacenes. One should be aware of the fact that in the point-dipole approximation all the dipole sums in the direction perpendicular to the long axis will be strictly zero. Since the excited state involves the whole conjugated backbone [89] a promising theoretical approach should be the use of extended dipoles or even more sophisticated calculations based upon model type excited state wavefunctions to calculate the interaction integrals. This approach has indeed being taken by M. Muccini et al. [161]. As a second point we may refer to the ideas of Rice and Jortner [113] that the conduction levels (band) can influence the Davydov splitting. This seems especially important in the case of 6T since the dipole calculations (section 7.4.2) predict a very large splitting, leading to an upper Davydov component which is high above the charge transfer states. One might speculate that the electronic excitation avoids the high energy splitting by mixing with states of separated charges and that therefore the huge splitting is hardly observed or not present. A third interesting question which is still open, is the location of the molecular excited A, states. Up to now, only very few data are available and for a full assignment of the higher extended states the position of the molecular 21Ag state as a function of chain length would help the experimentalists as well as the theoretical procedures to develop a deeper understanding of the nature of the electronic states of conjugated molecules. At this point it is worth noting that the 2'Ag state has been investigated for polythiophene in solution by Nunzi et al. using nonlinear optical methods (Kerr ellipsometry; section 7.6) [162]. They reported the 2lA, state to be about 5000 cm-' higher in energy than the first allowed absorption band thus indicating again that the 2'A, state is not crossing the l'B, state in oligothiophenes. Note added in proofs: The b, upper Davydov component with ac polarization has been observed at 20 945 cm-l by studying the single crystal absorption spectrum at 4.2K [161]. The Davydov splitting is therefore of about 2600cm-'.

7.5 Polarized electroluminescence In this chapter we want to focus on the applicational visions which might be developed further in the future. Different oligothiophene derivatives have been used for the production of light emitting devices (LEDs) and turned out to be valuable model molecules for a basic understanding of LED properties, e.g. the I-V characteristics [122], the role of substituents on quantum yields [23], or the influence of interfacial layers [144]. In the following we want to summarize the application of 6T for polarized electroluminescence, which was the first oligomeric material to be used for polarized LEDs [8]. Since the nT show a preferential perpendicular orientation with respect to polar substrates the emission properties of 6T films were expected to be highly anisotropic. As stated above (section 7.4.3.1) the grain size and hence the area of undistorted crystals are strongly dependent on the preparation conditions and especially on

396

7 Electronic Excited States of Conjugated Oligothiophenes

the growth temperature. Therefore Marks et al. [144] varied the growth conditions to analyze the resulting film properties as a function of growth with a geometric configuration which is shown in Fig 20. Although the luminescence is due to trap emission as shown in section 7.4.3, it can be clearly seen from Fig. 21 that the nluminiiim

\

aT6 niolecules

detccior

Figure 20. Sketch of a LED and the experimental setup of measuring the angularly resolved polarized electroluminescence [1441.

viewing angle (deg)

Figure 21. Angular dependence of electroluminescence in 6T based LEDs for devices made at 155°C (top), 104°C (centre), and 55°C (bottom) substrate temperature. Open circles correspond to p-polarized, triangles to s-polarized light [1441.

7.6 Nonlinear optical properties of polythiophene and thiophene oligomers

397

expected polarized emission occurs if the growth temperature and hence the size of the polycrystalline grains is larger than 600 nm. Although the angular dependent polarizational behavior may also be influenced by the anisotropic index of refraction, the observed emission behavior was attributed to a well-defined surrounding for the emitting trap site in the case where the fraction of volume to grain boundaries is optimized [8]. Unfortunately an emission of preferentially ‘standing’ molecules only leads to a high degree of polarization at high viewing angles (Fig. 21). Therefore the application of ‘lying’ molecules in combination with azimuthal ordering is a more desirable configuration for a high polarization ratio in all viewing angles. One of the promising approaches may be to use metallic substrates or thin metallic films. Recent work on 6T on the anisotropic Au(ll0) surface [1631 or 4T derivatives on Ag( 110) [1521 showed that at least very thin films of a few molecular layers can be prepared with a parallel adsorption geometry and a preferential azimuthal orientation. Otherwise the use of ordered matrices [1641may also be used for the tayloring of molecular based LEDs.

7.6 Nonlinear optical properties of polythiophene and thiophene oligomers Nonlinear optical properties of conjugated polymers have been extensively studied in the last decade [165-1671. Experimentally it has been shown that third-order nonlinearities of conjugated polymers are very fast and reach very high magnitudes opening the perspective of the possible development of organic based all-optical switches and modulators. Polyacetylene has been the first material to be studied and soon after other conjugated systems have been investigated among which polythophene and modified polythiophenes can be found. The nonlinear response of polythiophene, 3-alkyl substituted derivatives and thiophene oligomers have been measured by using thirdharmonic generation (THG) and degenerate four-wave mixing (DFWM). Thirdharmonic generation, which is the process of generation of a photon at three times the energy of the impinging photon at frequency w, is represented in the solid by the susceptibility tensor x ( ~(-3w; ) w, w, w). A series of experimental papers using this technique on polythiophenes and thiophene oligomers have been published in recent years [ 168- 1741. Degenerate four-wave mixing, indicated ) w, w, -w) identifies the process of interaction by the susceptibility tensor x ( ~(-w; of three beams at frequency w and the generation of a fourth beam at w. This technique has been applied to polythiophenes by several authors [ 175-1 781. While third-harmonic generation probes the intrinsic ultra-fast electronic response of the system, degenerate four-wave mixing, by measuring the nonlinear diffraction efficiency of the medium, is monitoring not only the instantaneous electronic response but also the nonlinear response of the photo-induced excitations (intra and intermolecular vibrations, acoustic phonons etc.) in which the system decays in time once the excited states have been prepared (i.e. in resonance). The

398

7 Electronic Excited States of Conjugated Oligothiophenes

response time of these excitations is longer and may extend by more than six orders of magnitude compared to the electronic response. For this reason a direct comparison of the two non-linearities could be misleading if the intrinsic differences of the two tensors were not taken into account. Nevertheless it has been shown that in both cases the cubic optical susceptibility x(3)is wavelength dependent and does not exceed lo-' e.s.u.. In Table 5 we summarize the literature data on nonlinear optical properties of polythiophenes and thiophene oligomers. The nonlinear response of polythiophene is strongly wavelength dependent, and the effect has been attributed to a population grating effect which becomes prominent at resonance [176,179,180]. The response gets larger at shorter wavelength but absorption losses are also larger. Since the perspective of making practical use of the nonlinearity requires that the system should be transparent at the working wavelength, the resonant increase cannot be exploited in fast all-optical devices. Furthermore in the case of substituted polythiophenes, the longer is the side chain of the substituent, the weaker is the response [169, 172, 1731 indicating that, while the side chain is a requisite for improving solubility, its net effect is to dilute the nonlinear response by increasing the total mass of the material without contributing to the nonlinearity. The rather weak response of 5T and 6T [174]is due to the unfavorable orientation of the long molecular axis with respect to the normal of the textured thin films. Since the long molecular axis makes an angle of 32" with respect to the normal [12,133](section 7.4.3.l), the lowest B, state contributes accidentally very little to the actual threephoton resonance enhancement (vide infra). A marked increase of the three-photon resonance would be observed if the long axis would lie in the substrate plane. We should point out that the off-resonance response of polythiophene (about lo-'' e.s.u.) is of the same order of magnitude of other conjugated polymers indicating that the nonlinear optical response of a conjugated chain does not depend Table 5. Summary of the literature reporting on the cubic susceptibility of polythiophene (PT), substituted polythiophenes and thiophene oligomers (nT). 3 alkyl-substituted thiophenes are labelled by P3C,T where n indicates the alkyl chain length and P3C120MeT indicates an alkylmethoxy derivative. _ _ _ _ _ _

Ref.

x(3)(x lo-'

Material

PT P3ClT P3C12T PT P3C 120MT PT P3CIOT P3C4T P3C6T 5T 6T

DFWM DFWM DFWM THG THG THG THG THG THG THG THG

~

e.s.u.) at different wavelengths (nm)

532

585

595

6.6 4.6

5

3.8

602

630

705

1907

0.03

0.7

0.3-0.5

1064

0.03-0.05 0.35 0.005 0.35 0.01 0.01

0.01 0.0019 0.0024

7.6 Nonlinear optical properties of polythiophene and thiophene oligomers

399

markedly on the conjugated structure but mainly on the number of conjugated double bonds. Theoretical investigations of the underlying microscopic mechanism have also attracted a large attention. A recent review by Bredas et al. covers most of the theoretical aspects [181]. The conjugation length dependence of the third-order optical nonlinearity has been the focus of the attention of several theoretical papers [182-1891. The dependence of the magnitude of the cubic nonlinearity on the number of conjugated double bonds N is found to follow a power-law dependence y = kN' where a ranges from 3 to 6 for small Nand then becomes linear in N for larger N. By means of systematic studies of the series of thiophene oligomers up to 6T, Prasad et al. [190] have shown that there is no levelling effect in the cubic nonlinearity and the exponent cy is about 4. Beljonne et al. [82] have evaluated the static third-order cubic hyperpolarizability of a series of oligomers finding a saturation at about seven repeat units and a power law dependence which reproduce quite closely the experimental results of Prasad et al. [190]. By investigating soluble oligomers Thienpont et al. [191] have been able to extend the investigation to longer oligomers containing up to 22 double bonds and showing a saturation effect at about 14 double bonds (7 repeat thiophene units). For the series of thiophene oligomers the heptamer or the octamer mark therefore the limit for effective enhancement of the effective third-order nonlinearity. T h s is in contrast to the case of soluble polyene oligomers [ 1921in which it was shown that saturation occurs at more than 100 double bonds. The experimental determination of the wavelength dispersion of the cubic susceptibility has opened a new possibility to check the validity of the microscopic nonlinear mechanism. In this respect the knowledge of the ordering of the electronic excited states is essential. Considering the simple case of a polyene, Hudson et al. [39] showed that the lowest excited electronic state is a gerade state contrary to the expectations based on simple one electron theories (sections 7.2.1 and 7.3.2.3). Gerade states can be treated only by considering the electron-electron correlation effects [188, 193-1951. Furthermore, only by talung account of the resonance with gerade states it is possible to account for the dispersion of the third harmonic generation signal. Since the quantum-chemical calculation of the ordering and the nature of the excited states depend on the adequate treatment of electron-electron correlation, the application to large oligomers has posed a large demand in the computing capabilities and only recently [82, 1961 a detailed description of the energies and transition strength of the lowest excited states has been given. A detailed discussion of the excited state ordering in nT is given in section 7.4.3.7. Fichou et al. have measured the dispersion of xFAGof 3T, 5T and 6T thin films showing a maximum at about 1.91 pm which has been interpreted in terms of threephoton resonance with the lowest electronic excited state [174]. Torruellas et al. [171] by measuring the dispersion of x!AG of spin coated films of alkylsubstituted polythiophenes have fitted the results with a four-level model in which there is a gerade state just below the lowest B, and an upper gerade state at 5000cm-' above the B, giving the largest contribution to the xTHG (3) dispersion. Nonlinear spectroscopy of alkylsubstituted polythiophene reveals in fact a two-photon state with a large cross section at 5000 cm-' above the lowest singlet [197].

400

7 Electronic Excited States of Conjugated Oligothiophenes

Acknowledgements This review is the result of the work of a large number of investigators, and it will be impossible to thank all of them individually. We would like to thank H. Baessler, D. Beljonne, J.L. Brkdas, J. Cornil, H.-J. Egelhaaf, R. Mahrt, D. Oelkrug and G. Weiser for numerous illuminating discussions. We would also like to thank E. Umbach and M. Sokolowski and their coworkers of the Wurzburg group with whom one of the authors [W.G.]worked for several years during the graduate studies. For the considerable part of the experimental work conducted at the Institute of Molecular Spectroscopy, we would like to acknowledge F. Biscarini, A. Degli Esposti, E. Lunedei, R. Marks, R. Michel, M. Muccini, M. Murgia, T. Virgili and R. Zamboni.

References 1. E. Hiickel, Z . Phys., 1931, 70, 204. 2. H. Kuhn, J. Chem. Phys., 1499,17, 1189. 3. M. Pope and C. E. Swenberg, Electronic Processes In Organic Crystals, Clarendon Press, Oxford, 1982. 4. A. S. Davydov, Theory Of Molecular Excitons, Plenum Press, New York, 1971. 5. D. P. Craig and S . H. Walmsley, Excitons In Molecular Crystals, Benjamin, New York, 1968. 6. F. Geiger, M. Stoldt, H. Schweizer, P. Bauerle and E. Umbach, Adv. Muter., 1993,5, 1684. 7. K. Uchiyama, H. Akimichi, S . Hotta, H. Noge and H. Sakaki, Synth. Met., 1994, 63, 57. 8. R. N. Marks, F. Biscarini, R. Zamboni and C. Taliani, Europhys. Lett., 1995,32, 523. 9. G. Horowitz, P. Delannoy, H. Bouchriha et aL, Adv. Mater., 1994,6, 752. 10. F. Gamier, F. Z. Peng, G. Horowitz and D. Fichou, Adv. Muter., 1990,2, 592. 11. F. Gamier, R. Hajlaoui, A. Yassar and P. Srivastava, Science, 1994, 265, 1684. 12. P. Ostoja, S. Guerri, S. Rossini, M. Servidori, C. Taliani and R. Zamboni, Synth. Met., 1993, 54,447. 13. A. Dodabalapur, L. Torsi and H. E. Katz, Science, 1995, 268, 270. 14. J. Simon and J.-J. Andrk, Molecular Semiconductors, Springer, Berlin, 1985. 15. N. Karl, in Crystal Growth And Applications, Springer, Heidelberg, 1980, Vol. 4, p. 1. 16. M. Muccini, R. F. Mahrt, R. Hennig, et al., Chem. Phys. Lett., 1995, 242,201. 17. C. Adachi, S. Tokito, T. Tsutsui and S . Saito, Jpn. J. Appl. Phys., 1988,27, 269. 18. A. Dodabalapur, L. J. Rothberg, T. M. Miller and E. W. Kwock, Appl. Phys. Lett., 1994,64, 2486. 19. N. Tessler, G. J. Denton and R. H. Friend, Nature, 1996,382, 695-697. 20. P. Bauerle, Adv. Muter., 1992, 4, 102. 21. P. Bauerle, G. Gotz, M. Hiller, S . Scheib et al., Synth. Met., 1993, 61, 1-2. 22. S. Hotta and K. Waragai, J . Muter. Chem., 1991, I , 835. 23. A. Yassar, F. Gamier, F. Deloffre, G. Horowitz and L. Ricard, Adv. Mater., 1994,6, 660. 24. P. Bauerle, T. Fischer, B. Bindlingmeier, A. Stabel and J. P. Rabe, Angew. Chemie, 1995, 107, 335. 25. Different nomenclatures may be found in the literature for the unsubstituted oligothiophenes (anT, T,, Tn, nT) referring to the same class of molecules. 26. J. P. Rabe and S . Buchholz, Phys. Rev. Lett., 1991, 66, 2096. 27. G. Zerbi, B. Chierichetti and I. Inganas, J. Chem. Phys., 1991, 94, 4646. 28. S. A. Chen and J . M. Ni, Synth. Met., 1993, 55, 576.

References

40 1

29. C. Taliani and L. M. Blinov, Adv. Mater., 1996, 8, 353. 30. C. Ziegler, in Handbook of Conductive Molecules and Polymers:, Vol. 3 Conductive Polymers: Spectroscopy and Physical Properties, John Wiley & Sons Ltd, 1997, pp. 677-743. 31. C. W. Tang and S. A. van Slyke, Appl. Phys. Lett., 1987, 51, 913. 32. J. H. Burroughes, D. D. C . Bradley, A. R. Brown et al., Nature, 1990, 347, 539. 33. H. C. Longuet-Higgins and L. Salem, Proc. Roy. SOC.(London), 1959, A251, 172. 34. W. P. Su, J. R. Schrieffer and A. J. Heeger, Phys. Rev. Lett., 1979, 42, 1698. 35. W. P. Su, J. R. Schrieffer and A. J. Heeger, Phys. Rev. B 1980, 22, 2099. 36. A. J. Heeger, S. Kievelson, J. R. Schrieffer and W. P. Su, Rev. Mod. Phys., 1988, 60, 781. 37. H. Kiess (ed.), Conjugated Conducting Polymers, Springer, Berlin, 1992. 38. G. Weiser, Phys. Rev., 1992, B 45, 14076. 39. B. S. Hudson, B. E. Kohler and K. Schulten, Excited States, 1982, 6, 1. 40. B. E. Kohler and J. B. Snow, J. Chem. Phys., 1983, 79, 2134. 41. B. E. Kohler, J. Chem. Phys., 1990, 93, 5838. 42. R. H. Friend, D. D. C. Bradley and P. D. Townsend, J . Phys. D: Appl. Phys., 1987,20, 1367. 43. U. Rauscher, H. Bassler, D. D. C. H. Bradley and M. Hennecke, Phys. Rev. B, 1990,42,9830. 44. K. S. Wong, W. Hayes, T. Hattori et al., J . Phys. C: Solid State Phys., 1985, 18, L843. 45. J. Shinar, Z. Vardeny, E. Ehrenfreund and 0. Brafman, Synth. Met., 1987, 18, 199. 46. Z. Vardeny, E. Ehrenfreund, 0. Brafman et al., Phys. Rev. Lett., 1986,56,671. 47. F. Moraes, H. Schaffer, M. Kobayashi, A. J. Heeger and F. Wudl, Phys. Rev. B, 1984,30,2948. 48. PL, Within this article the general term photoluminescence is used, although the authors are quite aware of the fact that the more specific term ‘fluorescence’is used in some of the original publications. 49. Z. Vardeny, E. Ehrenfreund, J. Shinar and F. Wudl, Phys. Rev. B, 1987, 35, 2498. 50. Z. V. Vardeny and W. Xing, Mol. Cryst. Liq. Cryst., 1994, 256, 465-72. 51. B. C. Hess, J. Shinar, Q. X. Ni, Z. Vardeny and F. Wudl, Synth. Met., 1989, 28, C365. 52. K. Kaneto, F. Uesugi and K. Yoshino, Solid State Comm., 1988, 65, 783. 53. H. Bassler, M. Gailberger, R. F. Mahrt, J. M. Oberski and G. Weiser, Synth. Met., 1992,49,341. 54. R. Richert and A. Blumen (eds.), Disorder Effects on Relaxational Processes, Springer, Berlin, 1994. 55. T. Kobayashi (ed.), Relaxation in Polymers, World Scientific, Singapore, 1993. 56. G. S. Kenner, X. Wei, B. C. Hess, L. R. Chen and Z. V. Vardeny, Phys. Rev. Lett., 1992, 69, 538. 57. A. Degli Esposti, 0. Moze, C. Taliani, J. T. Tomkinson, R. Zamboni and F. Zerbetto, J . Chem. Phys., 1996, 104, 9704. 58. R. S. Becker, J. Seixas de Melo, A. L. Maqanita and F. Elisei, J . Phys. Chem., 1996, 100, 18683. 59. A. Degli Esposti and F. Zerbetto, 1997, private communication. 60. W. Porzio, S. Destri, M. Mascherpa and S. Brueckner, Acta Polymer., 1993, 44, 266. 61. G. Horowitz, B. Bachet, A. Yassar, P. Lang, F. Demanze, J.-L. Fave and F. Garnier, Chem. Mater., 1995, 7, 1337. 62. S. Hotta and K . Waragai, Adv. Muter., 1993, 6, 896. 63. H.-J. Egelhaaf, PhD Thesis (University of Tuebingen, Department of Physical Chemistry, 1996). 64. H.-J. Egelhaaf, D. Oelkrug, D. Oeter, C. Ziegler and W. Gopel, J . Mol. Struct., 1995,348,405. 65. G. Lanzani, M. Nisoli, V. Magni, S. De Silvestri, G. Barbarella, M. Zambianchi and R. Tubino, Phvs. Rev. B, 1995,51, 13 770. 66. F.‘Garnier, G. Horowitz, P. Valat, F. Kouki and V. Wintgens, Appl. Phys. Lett., 1998, 72, 2087. 67. H.-J. Egelhaaf, P. Bauerle, K. Rauer, V. Hoffmann and D. Oelkrug, J . Mol. Struct., 1993, 293, 249. 68. J. Michl and E. W. Thulstrup, Elementary Polarization Spectroscopy, VCH, New York, 1989. 69, R. Lazzaroni, A. J. Pal, S. Rossini, G. Ruani, R. Zamboni and C. Taliani, Synth. Metals, 1991, 42, 2359. 70. Transition energies were determined by extracting the value of the intersection between absorption and emission curves if both are scaled to equal maximum height (Fig. 1).

402

7 Electronic Excited States of Conjugated Oligothiophenes

71. H. J. Egelhaaf, D. Oelkrug, W. Gebauer et al., Optical Mat., 1998,9, 59. 72. H. J. Egelhaaf and D. Oelkrug, Proc. SPIE The International Society for Optical Engmeering, 1995,5, 19-22. 73. R. Colditz, D. Grebner, M. Helbig and S. Rentsch, Chem. Phys., 1995, 201, 309. 74. D. Birnbaum and B. E. Kohler, J. Chem. Phys., 1992,96,2492. 75. D. Birnbaum and B. E. Kohler, J. Chem. Phys., 1989,90,3506. 76. D. Birnbaum, D. Fichou and B. E. Kohler, J. Chem. Phys., 1992,96, 165. 77. D. Oeter, H. J . Egelhaaf, C. Ziegler, D. Oelkrug and W. Gopel, J . Chem. Phys., 1994, 101, 6344-52. 78. M. Klessigner and J. Michl, Light Absorption And Photochemistry Of Organic Molecules, VCH, Weinheim, 1989. 79. W. Kuhn, Helv. Chim. Acta 1948, 31, 1780. 80. R. Pariser and R. G. Parr, J . Chem. Phys., 1953,21,466 and 767. 81. H. Chosrovian, S. Rentsch, D. Grebner, D. U. Dahm, E. Bircher and H. Naarmann, Synth. Metals, 1993, 60, 23. 82. D. Beljonne, Z. Shuai and J. L. Bredas, J. Chem.Phys., 1993,48,8819. 83. D. Beljonne, J. Cornil, J. L. Brtdas and R. H. Friend, Synth. Met., 1996, 76, 61. 84. D. Oelkrug, H.-J. Egelhaaf, D. R. Worrall and F. Wilkinson, J. Fluorescence, 1995, 5, 165. 85. R. S. Becker, J. Seixas de Melo, A. L. MaGanita and F. Elisei, Pure Appl. Chem., 1995,67,9. 86. J. T. Lopez Navarrete and G. Zerbi, J . Chem. Phys., 1991,94,957 and 965. 87. R. A. Janssen, L. Smilowitz, N. S . Sariciftci and D. Moses, J. Chern. Phys., 1995, 101, 1787. 88. J. C. Scaiano, R. W. Redmond, B. Mehta and J. T. Arnason, Photochem. Photobiol., 1990,52, 655. 89. D. Beljonne, J. Cornil, R. H. Friend, R. A. J. Janssen and J. L. Bredas, J. Am. Chem. Soc., 1996, 118, 6453-6461. 90. P. Landwehr, H. Port and H. C. Wolf, Chem. Phys. Lett., 1996,260, 125. 91. G. Lanzani, L. Rossi, A. Piaggi, A. J. Pal and C . Taliani, Chem. Phys. Lett., 1994,226, 547. 92. H. Fujimoto, U. Nagashima, H. Inokuchi et al., J. Chem. Phys., 1990, 92,4077. 93. A. Gavezzotti and G . Filippini, Synth. Met., 1991, 40, 1684. 94. M. Pierrot (ed.), Structure And Properties Of Molecular Crystals, Elsevier, Amsterdam, 1990. 95. M. Moebus, N. Karl and T. Kobayashi, J. of Cryst. Growth, 1992, 116,495. 96. D. Fichou, B. Bachet, F. Demanze, I. Billy, G. Horowitz and F. Garnier, Adv. Muter., 1996, 8, 500. 97. W. Gebauer, PhD Thesis, Physics Department, University of Wiirzburg, 1997. 98. W. Gebauer, M. Bassler, A. Soukopp et al., Synth. Met., 1996, 83, 227. 99. W. Gebauer, M. Bassler, R. Fink, M. Sokolowski and E. Umbach, Chem. Phys. Lett., 1997, 266, 177. 100. 0. Bohme, C. Ziegler and W. Gopel, Adv. Muter., 1994, 6, 587. 101. W. Porzio, S. Destri, M. Marschepa, S. Rossini and S . Brueckner, Synth. Met., 1993,54,447. 102. B. Servet, G. Horowitz, S . Ries et af., Chem. Mater., 1994,6, 1809. 103. T. Siegrist, R. M. Fleming, R. C. Haddon et al., J. Muter. Res., 1995, 10, 2170. 104. J. Bernstein, J. R. P. A. Sarma and A. Gavezotti, Chem. Phys. Lett., 1990, 174, 361. 105. F. van Bolhuis, H. Wynberg, E. E. Havinga, E. W. Meijer and E. G. J. Staring, Synth. Metals, 1989, 30, 381. 106. G. J. Visser, G. J. Heeres, J. Wolters and A. Vos, Acta Crystallogr., B, 1968, 24, 467. 107. J. I. Pankove, Optical Processes In Semiconductors, Prentice-Hall, 1971. 108. V. L. Broude, E. I. Rashba and E. F. Sheka, Spectroscopy Of Molecular Excitons, Springer, Berlin, 1985. 109. E. A. Silinsh and V. Capek, Organic Molecular Crystals, Interaction, Localization and Transport Phenomena, AIP Press, Woodbury, 1994. 110. M. Muccim, E. Lunedei, C. Taliani, G. Horowitz, F. Garnier and A. Bree, J. Chem. Phys., 1998, 108, 7327. 1 1 I. M. R. Philpott and J. W. Lee, J . Chem. Phys., 1973,58, 595. 112. V. Czikkely, H. D. Foersterling and H. Kuhn, Chem. Phys. Lett., 1970, 6, 1 1 and 270. 113. S. A. Rice and J. Jortner, in Physics And Chemistry Of The Organic Solid State (Ed. D. Fox, M. M. Labes and A. Weissberger), Interscience, New York, 1967.

References

403

114. H. Port, in Organic Molecular Aggregates (Ed.: P. Reinecker, H. Haken and H. C. Wolf), Springer, Berlin, 1983. 115. P. P. Ewald, Ann. d. Phys., 1921, 64, 253. 116. W. Gebauer, M. Baessler, R. Fink, M. Sokolowski and E. Umbach, Chem. Phys. Lett., 1997, 266, 117. 117. L. Sebastian, G. Weiser and H. Bassler, J . Chem. Phys., 1981, 61, 125. 118. P. J. Bounds and W. Siebrand, Chem. Phys. Lett., 1980, 75, 144. 119. R. C. Haddon, A. P. Ramirez and S. H. Glarum, Adv. Muter., 1994,6, 316. 120. C. Vaterlein, B. Ziegler, W. Gebauer et al., Synth. Met., 1996, 76, 133. 121. F. Garnier, F. Deloffre, G. Horowitz and R. Hajlaoui, Synth. Met., 1993, 55-57, 4747. 122. C. Vaterlein, H. Neureiter, W. Gebauer, et a/., J . Appl. Phys., 1997, 82, 3003. 123. F. Garnier, G. Horowitz, X. Z. Peng and D. Fichou, Synth. Met., 1993, 45, 163. 124. L. Torsi, A. Dodabalapur, L. J. Rothberg, A. W. P. Fung and H. E. Katz, Science, 1996, 272, 1462. 125. M. Wu and E. M. Conwell, Chem. Phys. Lett., 1997, 266, 363. 126. A. Dodabalapur, L. Torsi and H. E. Katz, Science, 1995,268,270-271. 127. A. J. Lovinger, D. D. Davis, A. Dodabalapur, H. E. Katz and L. Torsi, Macromolecules, 1996, 29,4952-4957. 128. U. Schoeler, K. H. Thews and H. Kuhn, J . Chem. Phys., 1995,61, 5009. 129. P. Yannoulis, R. Dudde, K. H. Frank and E. E. Koch, Surf. Sci., 1987, 1891190, 519. 130. E. Umbach, Prog. Surf. Sci., 1991, 35, 113. 131. E. Umbach, C. Seidel, J. Taborski, R. Li and A. Soukopp, Phys. Stat. Sol. B, 1995,192,389. 132. R. Li, P. Bauerle and E. Umbach, Surf. Sci., 1995, 331-333, 100. 133. B. Servet, S. Ries, M. Trotel, P. Alnot, G. Horowitz and F. Garnier, Adv. Muter., 1993,5,541. 134. F. Biscarini, R. Zamboni, P. Samori, P. Ostoja and C. Taliani, Phys. Rev. B, 1995, 52, 14868. 135. F. Biscarini, P. Samori, 0. Greco and R. Zamboni, Phys. Rev. Lett., 1997, 78, 2389. 136. W. A. Schoonveld, R. W. Stok, J. W. Weijtmans et al., Synth. Met., 1997, 84, 583. 137. 0. Bohme, C. Ziegler and W. Gopel, Synth. Met., 1994, 67, 87. 138. A. Soukopp, C. Seidel, R. Li, M. Bassler, M. Sokolowski and E. Umbach, Thin Solid Films, 1996.2841285. 343. 139. M. B. Nardelli, D. Cvetko, V. De Renzi et al., Phys. Rev. B, 1996, 53, 1095. 140. M. G. Ramsey, G. Koller, I. Kardinal and F. P. Netzer, Surf. Sci., 1996, 352-254, 128. 141. D. Oelkrug, H. J. Egelhaaf and J. Haiber, Thin Solid Films, 1996, 2841285, 267. 142. M. Stoldt, PhD Thesis, University of Stuttgart, Physics Department, 1994. 143. D. Oelkrug, J. Haiber, R. Lege, H. Stauch and H. J. Egelhaaf, Thin Solid Films, 1996,2841285, 581. 144. R. N. Marks, F. Biscarini, T. Virgili, M. Muccini, R. Zamboni and C. Taliani, Phil. Trans. R. SOC.Lond. A, 1997, 355, 763. 145. D. Fichou, G. Horowitz, B. Xu and F. Garnier, Synth. Met., 1992, 48, 167. 146. R. M. Hochstrasser and M. Kasha, Photochem. Photobiol., 1964, 3, 317. 147. Y. Kanemitsu, N. Shimizu, K. Suzuki, Y. Shiraishi and M. Kuroda, Phys. Rev. B, 1996, 54, 2 198-2204. 148. F. Deloffre, F. Garnier, P. Srivastava, A. Yassar and J. L. Fave, Synth. Met., 1994, 67, 223. 149. W. Gebauer, C. Vaeterlein, A. Soukopp, M. Sokolowski,H. Port, P. Bauerle and E. Umbach, Synth. Met., 1997, 87, 127. 150. R. N. Marks, E. Lunedei, M. Muccini et al., Chem. Phys., 1997, 227, 49. 151. R. N. Marks et al., to appear in J. Chem. Phys. 152. E. Umbach, R. Fink and M. Sokolowski,Appl. Phys. A , 1996,63, 565. 153. W. Gebauer, C. Vaeterlein, A. Soukopp, M. Sokolowski and E. Umbach, Thin Solid Film, 1996, 2841285, 576. 154. S. N. Magonov and M.-H. Whangbo, Surface Analysis With S T M And AFM, VCH, Weinheim, 1996. 155. Although the spectra were measured in reflection, we term the spectra as ‘absorption’ spectra, since highly resolved PL excitation spectra agreed very nicely with the reflection spectra. 156. N. Periasamy,D. Danieli, G. Ruani, R. Zamboni and C. Taliani, Phys. Rev. Lett., 1992,68,919. 157. L. M. Blinov, S. P. Palto, G. Ruani et ul., Chem. Phys. Lett., 1993,232,401.

404

7 Electronic Excited States of Conjugated Oligothiophenes

158. B. Xu and S. Holdcroft, Adv. Muter., 1994, 6, 325. 159. B. Xu and S. Holdcroft, J. Am. Chem. SOC.,1993, 115, 8447. 160. 0. Dippel, V. Brandl, H. Bassler, R. Danieli, R. Zamboni and C. Taliani, Chem. Phys. Lett., 1993,216,418. 161. M. Muccini, E. Lunedei, D. Beljonne, J. Cornil, J. L. Bredas and C. Taliani, J. Chem. Phys., submitted. 162. N. Pfeffer, P. Raimond, F. Charra and J.-M. Nunzi, Chem. Phys. Lett., 1993,201, 357. 163. F. Biscarini and e. al., Phys. Rev. B 1997, submitted. 164. N. S. Sariciftci, U. Lemmer, D. Vacar, A. J. Heeger and R. A. J. Janssen, Adv. Muter., 1996,8, 651. 165. Nonlinear Optical Effects in Organic Polymers, Vol. 162 (Ed.: J. Messier, F. Kajzar, P. Prasad and D. Ulrich), Kluwer Academic Publisher, The Netherlands, 1989. 166. Conjugated Polymeric Materials: Opportunities in Electronics, Optoelectronics and Molecular Electronics, Vol. ? (Ed.: J. L. Bredas and R. R. Chance), Kluver Academic Publishers, The Netherlands, 1990. 167. Photoactive Organic Materials, Vol. ? (Ed.: F. Kajzar, V. M. Agranovich and C. Y. Lee), Kluver Academic Publisher, The Netherlands, 1996. 168. T. Sugiyama, Synth. Met., 1989,28, C323. 169. H. Sasabe, T. Wada, T. Sugiyama, H. Ohkawa, A. Yamada and F. Garito, in Conjugated Polymeric Materials: Opportunities in Electronics, Optoelectronics and Molecular Electronics (Ed.: J. L. Bredas and R. R. Chance), Kluwer Academic Publishers, The Netherlands, 1990, v. 399. 170. D. Neher, A. Wolf, M. Leclerc, A. Kaltbeitzel, C. Bubeck and G. Wegner, Synth. Met., 1990, 37. 249. 171. W. E. Torruellas, D. Neher, R. Zanoni, G. I. Stegeman, F. Kajzar and M. Leclerc, Chem. Phys. Lett., 1990, 175, 11. 172. T. M. Lee, S. Mittler-Neher, D. Neher, et al., Optical Mat., 1992, I , 65. 173. C. L. Callender, S. J. Karnas, J. Albert, C. Roux and M. Leclerc, Optical Mat., 1992, 1, 125. 174. D. Fichou, F. Garnier, F. Charra, F. Kajzar and J. Messier, in Organic Materials for Nonlinear Optics (Ed.: R. Hahn and D. Bloor), Royal SOC.Chem., London, 1989, p. 176. 175. H. J. Byrne, W. Blau and K. Y. Jen, Synth. Met., 1989, 32, 229. 176. B. P. Singh, M. Samoc, H. S. Naiwa and P. N. Prasad, J . Chem. Phys., 1990,92,2756. 177. S. A. Jenekhe, S. K. Lo and S. R. Flom, Appl. Phys. Lett., 1989,54,2524. 178. C. Bubeck, A. Kaltbeitzel, A. Grund and M. LeClerc, Chem. Phys., 1991, 154,343. 179. R. Dorsinville, L. Yang, R. R. Alfano et al., Optics Lett., 1989, 14, 1321. 180. L. Yang, R. Dorsinville, Q. Z. Wang et al., J.Opt. SOC.Am. B, 1989, 6, 753. 181. J. L. Bredas, C. Adant, P. Tackx, A. Persoons and B. M. Pierce, Chem. Rev., 1994,94,243. 182. K. C. Rustagi and J. Ducuing, Opt. Commun., 1974, 10, 258. 183. G. P. Agrawal, C. Cojan and C. Flytzanis, Phys. Rev. B, 1978, 17, 776. 184. C. P. De Melo and R. Silbey, Chem. Phys. Lett., 1987, 140, 537. 185. B. M. Pierce, J.Chem. Phys., 1989, 91, 791. 186. Z. Shuai and J. L. Bredas, Phys. Rev. B, 1991,44,5962. 187. S. Mukamel and H. X. Wang, Phys. Rev. Lett., 1992, 69, 65. 188. Z. Soos and S. Ramasesha, J.Chem. Phys., 1989,90, 1067. 189. G. D. Mahan and A. G. Rojo, Phys. Rev. B, 1994,50,2642-2644. 190. M. T. Zhao, B. P. Singh and P. N. Prasad, J.Chem. Phys., 1988,89, 5535. 191. H. Thienpont, G. L. J. A. Rikken, E. W. Meijer, W. ter Hoeve and H. Wynberg, Phys. Rev. Lett., 1990, 65, 2141. 192. I. D. W. Samuel, I. Ledoux, C. Dhenaut and J. Zyss, Science, 1994,5175, 1070. 193. S. N. Dixit, D. Guo and S. Mazumdar, Phys. Rev.& 1991,43,6781. 194. S . Mazumdar and F. Guo, J. Chem. Phys., 1994, 100, 1665. 195. Z. Shuai, D. Beljonne and J. L. Bredas, J. Chem. Phys., 1992,97, 1132. 196. J. Cornil, D. Beljonne and J. L. Bredas, J.Chem. Phys., 1995,103,842. 197. N. Pfeffer, P. Raimond, F. Charra and J. M. Nunzi, Chem. Phys. Lett., 1993, 201, 357.

8 Electro-optical Polythiophene Devices Magnus Granstrom, Mark G. Harrison, and Richard H. Friend

8.1 Overview In this chapter, we discuss a number of optical and optoelectronic device applications such as light-emitting diodes (LEDs), photovoltaic and photoconductive devices, field-effect optical modulator devices and all-optical modulator devices made from thiophenes. Following an introduction to the basic operating principles of each device, we will assess progress in the development of each type of device and focus on the underlying semiconductor physics issues. The field of organic semiconductors, has existed for several decades. Molecular crystals of acenes (Fig. la), phthalocyanines, small molecules and metal-organic complexes such as Alq, (Fig. lb) were studied because of their photoconductive [l, 21 and semiconducting [3, 41 properties and also as an approach to probe the optoelectronic properties of biological membranes. Small organic semiconductor molecules blended in polymer hosts have already found a major application in xerography [ 5 ] . Recently, there has been renewed commercial interest and research activity into organic semiconductors with the development of conducting and semiconducting conjugated polymers. These offer scope for preparing large area conducting films for lightweight conductors, electromagnetic shielding and large area semiconducting films for displays, solar cells and transistor arrays. Since this book deals with thiophenes, we will not dwell on the more established small molecular organic semiconductors but will focus on oligo(thiophenes) and poly(thiophenes). Oligo(thiophenes) can be viewed both as materials with great potentials for devices, partly because of their high field-effect mobilities, and also as finite model systems for the poly(thiophenes). Structures of some oligo(thiophenes) are shown in Fig. 2.

8.1.1 Relationship between polymers and oligomers As mentioned above, we identify two main reasons for studying oligomers of conjugated polymers: Firstly, oligomers represent model systems for understanding the fundamental electronic properties of the corresponding polymer. Oligomers can be synthesised with a well-defined molecular length, as shown in the structural formula of the extensively studied oligomer, a-sexithiophene (a-6T) (Fig. 2a). Oligomers have therefore been recognised for some time as model systems for theoretical [6] and experimental [7] investigations aimed at extrapolating physical properties of

406

8 Electro-optical Polythiophene Devices

Anthracene

Tetracene

Pentacene

t-Be

0

y

J

y

-

. j

PBD

Figure 1. Structural formulae: (a) The acene oligomers: Anthracene, Tetracene, Pentacene, Perylene, Coronene, (b) The metal-organic complex Alq3, and the electron transport material PBD.

finite oligomers to the corresponding ideal polymer of infinite length. In marked contrast, real conjugated polymers exhibit a distribution of lengths, along which r-conjugation is effective. The coherent conjugated segments of the polymer chain are interrupted by defects, which may be of a conformational nature (e.g. twisting of the chain so that it is no longer planar) or of a chemical nature, such as a saturated sp3-hybridisedcarbon atom located somewhere along the chain. Extrapolations of quantitative characteristics from studies of oligomers can therefore also yield estimates of the effective conjugation length in real polymers. Oligomers are well-defined systems of monodisperse (uniform-length) molecules, with greatly reduced occurrence of defects within the molecular chains, in comparison with polymers. They therefore offer the possibility of better ordering of the molecules and consequently more well-defined optical properties. This renders them particularly appealing for both theoretical and experimental investigations

8.1 Overview

407

FlOH21

I

GoHz1

Figure 2. Structural formulae of a-sexithiophene and derivatives: (a) unsubstituted sexithiophene [cu-6T],(b) end-capped sexithiophene [EC6T], (c) regio-random @-substituteddidodecyl-6T [2D6T], (d) derivative of 6T with bulky triisopropylsilyl end-groups [DPS6T], (e) oligothiophene with two triphenyleneamine end-groups [BMA-nT].

into a number of issues, which cannot be so readily assessed in polymeric systems. These include the following: (i)

dependence of the energies and equilibria of neutral and charged excitations as a function of the coherence (molecular) length; (ii) substitution of oligomers, either with electro-active groups or with the aim of inducing order or disorder; (iii) the role of intermolecular processes. When trying to understand the behavior of oligomers, it is helpful to consider concepts employed for conjugated polymers and also those from the more established field of molecular semiconductors and charge-transfer salts, since interchain processes can be more easily observed in thin films of oligomers. Secondly, in some cases, oligomers have already been shown to exhibit characteristics complementary with those currently found in many conjugated polymers. For instance, since the discovery of blue electroluminescence from anthracene [8, 91,

408

8 Electro-optical Polythiophene Devices

there has been interest in using short oligomers, particularly to achieve the blue emission [ 10-171 required for full-color displays. The energies of optical transitions of oligomers often vary linearly with the reciprocal of the oligomer length [6, 71, since the length of the molecule confines the spatial extent of many of the charged and neutral excitations of the oligomer. Therefore, in short oligomers the lowest excited state of the singlet exciton is more confined than in long polymers, so higher excitation energies can be achieved, leading to blue emission. We also include discussion of an all-optical spatial light modulator [18] prototype, which could have major applications in rapid image-processing. Recently, solid state ultra-fast reversible photoswitches have been demonstrated, in which the active material contains oligothiophenes as chromophore. Such materials show great promise in all-optical data storage devices with non-destructive readout.

8.2 Preparation of thin film devices 8.2.1 Introduction Thin film opto-electronic devices, such as LEDs, modulators and photocells all require the deposition of one or more thin semiconductor layers onto a substrate. Such films are usually of thickness in the range between 50 nm and 10 pm. A perceived advantage of organic semiconductors over traditional, crystalline semiconductor materials is that thin films can be deposited over large areas on a variety of substrates, including glass and flexible substrates [19-221 such as Mylar films, without the concerns about crystalline defects and matching of lattice periodicities during film growth. Thin films of organic semiconductors have been deposited onto substrates either from solution or from the vapor phase. The method of choice depends mainly on the molecular weight and solubility of the material.

8.2.2 Polymers In order to manufacture devices made with conjugated polymers, the polymers must be processible into thin films or coatings. The earliest and structurally most simple conjugated polymers, e.g. polyacetylene, polypyrrole, polythiophene, and poly(phenylenevinylene), do not posses this property. Looking at neighboring areas, such as liquid crystalline polymers, this problem had been solved by the use of alkyl or alkoxy side chains [23]. If these side chains were of sufficient length, the polymers became soluble in ordinary organic solvents. This knowledge was then transferred to the conjugated polymers, and soluble and fusible poly-(3-alkylthiophenes) were obtained by adding alkyl side chains at the 3-position of thiophene monomers 124-281. For the thiophene system, the side chain must have at least

8.2 Preparation of thinjlm devices

409

four carbons to induce the solubility. The monomers can be polymerised both chemically (Grignard coupling or FeC13-coupling) [26, 271 and electrochemically [24], with considerable molecular weights obtained using chemical polymerisation ( M , = 30 000, Mw= 150 000) [27]. Organic solvents, such as toluene and chloroform, can then be used to dissolve these substituted thiophenes, and it is possible to form thin, uniform films by casting or spinning from solution. By addition of side chains or by using precursor methods [29-331, similar properties can also be obtained for poly(phenyleneviny1enes) (PPV)s, the extensively studied class of the electroluminescent polymers. The thiophene system has an advantage in the fact that it is possible to put side groups on either the 3- or 4-position or on both, which gives a large flexibility in the choice of chemical and geometrical structure.

8.2.3 Oligomers Many small molecular semiconductors and unsubstituted oligomers take the form of rather rigid, planar molecules, which are generally insoluble in many organic solvents. Thin films of these low weight materials are generally achieved by sublimation. Conversely, films of soluble conjugated polymers or soluble partially-conjugated precursor polymers can be deposited from solution by casting and spin-coating as discussed above. However, most unsubstituted oligomers lack the advantage of solution-processing which can be achieved with polymers, since they are synthesised as fully conjugated rigid-rod molecules. These show a strong tendency to aggregate in solution, particularly at the high molecular concentrations required for film deposition. Unsubstituted oligomers are therefore usually deposited from the vapor phase, by sublimation under vacuum.

Figure 3. Structural formula of a copolymer containing pendant oxadiazole segments for chargetransport and distyrylbenzene (PPV oligomers) as blue emitters. Adapted from Li et al.

410

8 Electro-optical Polythiophene Devices

In order to transfer the attractive solution-processing properties of polymers to oligomers, the smaller oligomers can be rendered soluble either by suitable chemical modifications, such as addition of alkyl side-chains in a similar fashion as the poly(3-alkylthiophenes) [34-371 (see Fig. 2b,c) or else they can be blended within a soluble polymer [lo-12, 14, 151 or chemically grafted as pendant side-chains on a polymer backbone [16, 171, as shown in Fig. 3.

8.2.4 Relative merits of the different methods to achieve solubility 8.2.4.1 Substitution with side-chains

These are generally long flexible alkyl chains, which give rise to entropic stabilisation of the polymer chain in solution. Soluble derivatives of oligomers have been synthesised, so that they can be deposited from the solution phase by spin-coating or dip-coating. Flexible side-chains [36, 371 and cycloalkane end caps [38, 391 have been used, as shown in Fig. 2b-c, though solubility is sometimes achieved at the expense of the electrical transport in the films [40]. Cycloalkane end-caps have the additional advantage of inhibiting further polymerisation of the oligomers, by blocking the reactive a-carbons of the outermost thiophene rings. Substitution with side-chains can have its disadvantages if the structural regioregularity of the oligomer or polymer is not carefully controlled by the chemical

Figure 4. Schematic view of intermeshed stacks of alkylthiophene chains.

8.2 Preparation of thinJilm devices

41 I

synthesis. In the solid state, crystallographic studies [41-441 and images obtained by scanning tunnelling microscopy [45] indicate that alkyl-substituted oligothiophenes and polythiophenes have a tendency to aggregate or self-assemble in a stacked interlocking comb-like structure as shown in Fig. 4; the aromatic backbones rich in .rr-electron density tends to stack cofacially, while the alkyl side chains align perpendicular to the main chain and are attracted to alkyl substituents on adjacent oligomers lying above or below. In alkyl-substituted oligothiophenes, the a-carbons nearest to the sulphur atom are involved in the bonding of the oligomer backbone, leaving either one of the P-carbons (two positions away from the sulphur atom) available for substitution. In Fig. 5a, we show the structural formula of poly(3-alkylthiophene), in which side-chains are attached to either one of the P-carbons (furthest from the sulphur atom), giving head-to-head, head-to-tail and tail-to-tail interactions between adjacent thiophene rings.

-

Head-to-head

Tail-to-tail

----<

Head-to-tail

Figure 5. (a) Regio-random alkylthiophene chain, showing Head-to-Head and Tail-to-Tail interactions resulting in twisting of the backbone. (b) Regioregular alkylthiophene chain, showing exclusively Head-to-Tail interactions, resulting in a planar backbone.

412

8 Electro-optical Polythiophene Devices

Regioregular alkyl-substituted poly(thiophenes) [46, 471 and oligo(thiophenes) [48] have also been synthesized, in which the alkyl chains are always substituted to the same type of P-carbon atom, thus producing predominantly head-to-tail interactions, as shown in Fig. 5b, and therefore reducing steric interactions between the alkyl chains. In those alkyl-substituted oligo(thi0phenes) and poly(thiophenes), where the synthesis is not controlled to yield regioregular substitution, head-to-head and tail-totail interactions give rise to a twisted backbone, in which the thiophene rings are not coplanar. This results in a reduction of the intramolecular 7r-conjugation along the chain and also a slight increase in the separation between 7r-conjugated chains when stacked in the solid state, leading to reduced intermolecular n-stacking. Additionally, regiorandom ,&substitution can cause spatial disorder in the wavefunction overlap of aromatic rings involved in cofacial 7r-stacking, which is detrimental to aggregation and intermolecular electrical conduction mediated by n-stacking. Regioregularity or lack of it also has implications for the quantum efficiencies of fluorescence since torsional oscillations are considered to be involved in some nonradiative decay channels [49].

8.2.4.2 Using a soluble partially-conjugatedprecursor polymer This is subsequently converted to the fully conjugated material after film deposition, usually by heating under vacuum or in acidic vapor. As an example of this, we show in Fig. 6 , the tetrahydrothiophene precursor route [50] to poly( para-phenylenevinylene) (PPV) and note that the final stage of the reaction (indicated by the bold arrows) is carried out once the thin film of the precursor polymer has been deposited. The precursor route strategy has not been extensively applied to oligomers,

(NaOH MeOH / HzO

vacuum

Im

22o'c

\

Figure 6. The tetrahydrothiophene precursor route to PPV. The bold arrows indicate the final thermal conversion performed after deposition of films of the partially-conjugated precursor polymer.

8.3 Electronic excitations in oligothiophenes

41 3

although very recently, Miillen and co-workers have developed precursor routes for oligoacenes [51], opening up the possibility of preparing thin films of pentacene and longer oligoacenes by solution-coating methods. Pentacene prepared from such a precursor route has been used in field effect transistors, for which mobilities as cm2V-' s-l have been measured [52]. high as

8.2.5 Blends between polymer and oligomers Whereas polymers are generally immiscible unless there is a specific advantageous chemical interaction, it is possible to incorporate small molecular semiconductors and oligomers within soluble polymeric hosts [lo-12, 14, 151, such as polystyrene and poly(9-vinylene carbazole) (PVK). A further development of this technique is to chemically graft the oligomers as pendant side-chains of a polymer [16, 171. These methods retain the advantages of solution-processing properties of polymers and are discussed later in more detail in the section on LED applications.

8.3 Electronic excitations in oligothiophenes 8.3.1 Introduction In oligomers and polymers, neutral excited states, or excitons, can be produced by photo-excitation or charge recombination (capture of electrons and holes in LEDs). These can either decay radiatively, as desired for light-emitting diodes or nonradiatively, with the possibility of yielding mobile charge carriers, for photoconductive and photovoltaic cells. We discuss here some of the physical issues involving organic LEDs and photocells, so that with this background, we can better appreciate the technological strategies for optimising device performance in the later sections on LEDs and photocells . In the following discussion, we consider first the intramolecular nonradiative decay channels, which can occur for isolated oligomers, then intermolecular nonradiative decay channels, which may also operate in solid state thin films, where the oligomers or polymers are densely packed. We also consider the effects of interring torsion and coplanarity of the n-conjugated chains, which give rise to both intramolecular and intermolecular effects.

8.3.2 Intra-molecular non-radiative decay channels Figure 7 provides a schematic overview of the intramolecular decay processes which we will now discuss. These include internal conversion, intersystem crossing, fission of singlet excitons.

4 14

8 Electro-optical Polythiophene Devices

-

-

-

SINGLET MANIFOLD

TRIPLET MANIFOLD

S* PhotoInternal Internal Range of

a energies

T1 Non-radiative decay by emission of

LL

-’

I---.

Weak Dhosnhorescence

.

.

Figure 7. Schematic energy level diagram showing singlet and triplet manifolds and intramolecular decay channels (internal conversion, intersystem crossing, singlet fission, etc.).

In the following sections, we give a brief introduction to each of the processes and also refer to predictions made on the basis of recent quantum chemical calculations on the geometries and energies of the singlet and triplet excited states of isolated thiophene oligomers by Beljonne et al. [53]. Assuming planar molecules, they used Hartree-Fock semi-empirical modified neglect of differential overlap (MNDO) calculations to optimise the geometry, followed by intermediate neglect of differential overlap/multi-reference double-configuration interaction (INDO/ MRD-CI) calculations to describe the singlet and triplet excited states. Their calculations indicate that the lowest triplet state (T1)is strongly confined and extends over approximately one thiophene ring, while the lowest singlet excited state (S1) is much more extended, the So H S1transition showing a much larger red-shift with increasing chain length. The calculated energy (1.57 eV) of the Sot+ T1 transition for terthiophene (3T) is in good agreement with experimental values determined by optical absorption in a solvent containing heavy bromine atoms [54] and by energy transfer from Cs0 [55, 561. Extrapolation to infinite chain length gives a value of the So H T1 transition of poly(thiophene) as 1.49 eV, in very good agreement with the energy of the phosphorescence peak [57] in polythiophene (1.5 eV). Time-resolved PL measurements on dilute solutions of oligothiophenes [58-611 indicate that as the length increases, the radiative decay rate does not change significantly, although the quantum yield increases. This indicates a decrease in the nonradiative decay rate as the length of the oligomers increases [58,59,61]. Referring to the calculations of Beljonne et al. [53], we include a discussion of how the rates of each of the intramolecular nonradiative decay processes, (internal conversion, intersystem crossing and singlet fission) depend on oligomer length.

8.3 Electronic excitations in oligothiophenes

415

8.3.2.1 Internal conversion Internal conversion refers to relaxation from a highly excited state to a state of lower energy, within the same manifold, either singlet or triplet, by emission of phonons. This process can be rapid, on the femtosecond timescale. The excess energy is usually lost as phonons. Internal conversion can provide a nonradiative decay channel if the transition from the relaxed state to the ground state is forbidden by the symmetry of the wavefunction of the relaxed state. The 2A, and lB, states have been located in low-temperature n-alkane solutions of a,w-dithienyloligoenes [62] with up to four double bonds between the terminal thiophene rings. In these oligomers, the lowest singlet excited state is the 2A, singlet state, from which radiative transitions to the ground state are symmetry forbidden, but which can be coupled to the ground state by phonon emission. The difference in energy between the 2A, and IB, singlet states appears to increase as the oligomer length increases from apdithienylbutadiene to a,w-dithienyloctatetraene. The question of whether or not the A, state falls below the B, state as the oligomer length increases towards an infinite polyene is still the subject of debate. A single-particle description, such as the Hiickel and Hartree-Fock theories, predicts that the the 2A, state lies above the lowest optically allowed lB, state [63] and therefore does not inhibit fluorescence of the molecule. Conversely, if manyparticle electron-electron interactions are important, the 2A, state would lie below the lB, state and fluorescence transition would be forbidden by symmetry considerations. In sexithiophene (a-6T), the 2A, singlet state has been located by two-photon spectroscopy [58] as being 0.1 eV higher in energy than the lowest (allowed) singlet excited state (1BJ. Therefore, internal conversion to a 2A, state does not represent a nonradiative decay channel for sexithiophene. However, the separation between 2A, and IB, states is less than the vibrational energy of the C=C stretch mode, so some coupling may be possible. On the basis of a l / n extrapolation of the energies of the 2A, state and lB, states of bithiophene (a-2T)and sexithiophene (a-6T), it had been suggested [64] that the 2A, state would lie below the lB, state for oligothiophenes with more than six rings. More recent photophysical measurements on oligomers up to seven rings [61] show that this is not the case and estimate the crossover to be nearer nine rings. It has been argued [65] that extrapolations from oligomers of finite length to infinite polymers based solely on a l/n dependence on the conjugation length, i.e. a 1-dimensional particle-in-a-box model, are unreliable and that for the higher energy excited states, the energy may depend non-linearly on the reciprocal conjugation length, l/n, e.g.

E = A + B(l/n)

+ C(l/n2).

For short oligomers, the contribution from non-linear terms could be rather large, so that predictions of convergence or crossover of the A, and B, states based on I/n-type extrapolations from short oligomers should be treated with caution. The theoretical work of Mazumdar et al. [65] and also experimental studies on carotenes [66, 671 suggest that the 2A, state may only be weakly coupled to the lB, state and

416

8 Electro-optical Polythiophene Devices

that a higher A, state is responsible for two-photon absorption. The higher A, state is considered to be above the B, state, although the energy separation between these states decreases with increasing conjugation length. 8.3.2.2 Intersystem crossing

This refers to efficient energy transfer between the singlet and triplet manifolds. Subsequent internal conversion usually leads to the lowest triplet state (TI), which does not fluoresce. The rate of intersystem crossing is governed by the spin-orbit coupling term, which is enhanced by heavy atoms, such as those of the sulphur atoms in oligothiophenes.

Intersystem crossing enhanced

L

1/n @)

5.0

_7

I

ZT

4.0

5

E(So-S,)22E(S

3.0 6T

x

2

8

I

2.0

o-Tl)

Singlet fssion requires no thermal activation

v

W

1 .o

0.0

~

"

"

"

"

"

"

"

"

"

"

"

,

"

"

'

"

'

" " "

55

Figure 8. (a) A comparison of the calculated energies of the (SO+ S1) and (SO4 T4) transitions of oligothiophenes, as a function of the number of rings, n. Adapted from Beljonne et al. The rate of intersystem crossing is reduced for the longer oligomers. (b) A comparison of the calculated energies of the (So + S1) and (So + TI)transitions of oligothiophenes, as a function of the number of rings, n. Adapted from Beljonne et al. The probability of singlet fission is reduced for longer oligomers.

8.3 Electronic excitations in oligothiophenes

417

According to the calculations of Beljonne et al. [53],the energy difference between S1 and T1 is too large to give efficient singlet-triplet overlap. However, the INDO/

MRD-CI calculations predict that there is one triplet excited state (T4)with energy close to that of the lowest singlet excited state (S1). For bithiophene (2T), T4 lies below S1,so intersystem crossing occurs readily, resulting in a very low fluorescence yield for short chains. In Fig. 8a, we see that as the oligomer length increases, S1 falls below T4, so intersystem crossing becomes increasingly unlikely, resulting in higher PL quantum efficiencies for longer oligomers.

8.3.2.3 Singlet fission Singlet fission describes a process in which two triplet excitons are produced from the fission of one singlet excited state. Such a process has been observed in tetracene, anthracene [68, 691 and polydiacetylene [70]. The excited singlet state may be the lowest excited singlet state (denoted S,) or in the studies cited above, a higher excited singlet state, produced by singlet-singlet fusion under intense irradiation. In the following discussion, we consider only the possible fission of the S1 state, which may play a r61e in the nonradiative decay of oligomers in light-emitting diodes. If singlet fission is to be an exothermic (energetically allowed) process, the energy of the singlet excited state undergoing fission must be at least twice as large as the energy of the lowest triplet excited state, in this case, E(S0 H Sl) 2 2E(So c-) T1).As shown in Fig. 8b, the calculations of Beljonne et al. indicate that this requirement is satisfied for short oligomers, although for longer oligomers the S1 level falls below twice the T1energy, so singlet fission can no longer contribute to nonradiative decay for long oligomers.

8.3.3 Inter-molecular non-radiative decay channels in thin films Photoluminescence efficiencies in solid films of oligomers are generally much lower than those of isolated oligomers in solution [49]. Also, in the solid state, the substituted oligomers show higher photoluminescence yield than the unsubstituted oligomers. Below, we discuss the effects of aggregation and formation of chargetransfer excitons which constitute additional nonradiative decay channels when oligomers are densely packed in the condensed phase. 8.3.3.1 Aggregation and Davydov splitting

In polycrystalline thin solid films, the close packing of the oligomers often gives rise to aggregates, in which the exciton wavefunction is delocalised over two or more adjacent oligomers in close proximity, leading to a broad emission spectrum, shifted to lower energies than the fluorescence spectrum of the isolated oligomers, often accompanied by a reduction in the photoluminescence (PL) quantum efficiency. Many conjugated oligomers crystallise with a herringbone structure [71-751, in which there are two translationally inequivalent molecules per unit cell, as depicted

418

8 Electro-optical Polythiophene Devices

Isolated Oligomer

Crystal

Figure 9. (a) Herringbone arrangement of oligomers found within close-packed layers. (b) The Davydov splitting which results from having two translationally inequivalent molecules per unit cell.

in Fig. 9a. The coupling of the dipole moments between such molecules gives rise to Davydov splitting; in the crystal, the excited states of the single molecule are split into pairs of levels. The optical transition between the ground state and the lower crystalline excited level is completely forbidden, as indicated in Fig. 9b. The magnitude of the Davydov splitting decreases as r P 3 ,where r is the separation between molecular centres. The dipole selection rules apply for an ideally infinite single crystal and are relaxed when the crystallite size is reduced or when introducing disorder.

8.3.3.2 Charge-transfer excitons In addition to intramolecular Frenkel excitons which are confined to one oligomer unit and give rise to the photoluminescence of isolated oligomers in dilute solution, charge-transfer excitons can also be generated by photo-excitation of oligomers in the condensed phase. Charge-transfer excitons consist of a weakly bound electron-hole pair, which is localised over two or more adjacent oligomers.

8.3 Electronic excitations in oligothiophenes

419

Charge-transfer excitons are not readily observed in the linear absorption spectrum because there is usually poor overlap between the wavefunctions of the ground state and excited state, since the ground state is localised on a single oligomer, while the excited state is distributed over two or more adjacent oligomers. However, they can be detected in electroabsorption, since the excited state has a large dipole moment and results in an electroabsorption signal which follows the second derivative of the absorption spectrum. The energy separation of intramolecular Frenkel excitons and charge-transfer excitons depends on the first ionisation energy, the electron affinity and the intermolecular distance. As the length of an oligomer increases, so does the spatial extent of its delocalised electronic 7r-system, leading to stronger 7r-7r van der Waals forces and lower intermolecular distances, as well as lower first ionisation potentials and larger electron affinities. Therefore, as the length of an oligomer increases, all of these factors tend to reduce the energy separation of the intermolecular charge-transfer excitons and intramolecular Frenkel excitons and open up a competing nonradiative decay channel, lowering quantum yields for fluorescence in the solid state, while increasing quantum efficiencies for photoconductivity. However, charge-transfer excitons are only an intermediate species to photocarrier generation. If the weakly bound electron-hole pair can recombine sufficiently rapidly, they can form the intramolecular singlet exciton before charge separation takes place. For thin films of a-sexithiophene (a-6T), crystallographic studies have determined that the oligomers are arranged in a stacked layer structure, in which conductivity parallel to the stacks (perpendicular to the long oligomer axis) is much greater than that between layers [40].Dippel et al. [76] therefore explained the photoconductivity as resulting from rapid charge separation of longitudinal charge-transfer excitons which extend across stacks. Therefore, close packing of oligomers within the layers should favour efficient charge separation and photoconductivity rather than fluorescence. In order to favour fluorescence, efficient charge transport can be suppressed by reducing the degree of order within the film, either by using polycrystalline films with small crystallite size or by increasing the separation of oligomers by the addition of bulky substituent groups.

8.3.4 Effects of inter-ring torsion and coplanarity of oligomers The issue of planarity and inter-ring torsion is relevant to oligomers in which the rings are linked by carbon-carbon single bonds, such as the a-oligothiophenes. It is especially relevant to oligomers substituted with alkyl side-chains for solubility and also to rigid bridged ladder-type oligophenyls synthesised for blue electroluminescence. It is perhaps worth emphasising from the beginning, that the trends observed in dilute solution are in marked contrast with those in the solid state, relevant to optoelectronic device applications. We therefore decided to discuss this topic after consideration of purely intramolecular and intermolecular decay mechanisms. While planarity may favour higher PL efficiencies in isolated molecules or dilute, well-dispersed blends, planarity also favours aggregation and hence lower PL efficiencies in the solid state.

420

8 Electro-optical Polythiophene Devices

8.3.4.1 Solution

Nonradiative decay channels are influenced by low frequency inter-ring torsional oscillations of the oligomer backbone. Berlman [49] noted that rigidity in the first excited state was important for fluorescence. Nijegorodov et al. [77] have studied photophysics of dilute solutions of several series of oligomers, (e.g. biphenyl, terphenyl) and have observed a systematic decrease in the intersystem crossing rate decreasing rapidly as the rigidity and planarity increase and conclude that high PL efficiencies can be achieved in solution if the ground state is non-planar and of low symmetry, while the excited state should be approximately planar and of higher symmetry, as is the case for many oligothiophenes and oligophenyls upon formation of the quinoid geometry in the singlet excite state. Becker et al. [61] have suggested that torsional oscillations may give rise to highly efficient nonradiative decay from the lowest triplet excited state to the ground state. In dilute solution or the gas phase (in the case of oligomers), the ground state of oligothiophenes and polythiophenes is often more twisted than the relaxed singlet excited state. This is particularly true at higher temperatures [78] and thermochromism and solvatochromism in poly(alky1thiophenes) is well known [79-831. At low temperatures, a red-shift of the absorption spectrum is observed [60, 611, while the fluorescence spectrum remains essentially unchanged. This indicates that at low temperatures, the ground state is more planar. Photoluminescence efficiencies in solution are generally much higher for planar molecules because the geometries in the ground and excited states are similarly coplanar and do not allow torsional oscillations to couple so effectively [49]. Xu and Holdcroft [84] reported that in the case of polythiophenes substituted with alkyl side-chains, head-to-head and tail-totail interactions (see Fig. 5a) result in increased twisting of the ring and lower PL efficiencies in solution, while regioregular poly(alky1thiophenes) which have predominantly head-to-tail interactions exhibit higher PL efficiencies in solution. Time-resolved fluorescence studies [60, 61, 851 of dilute solutions of alkyl-substituted oligomers up to seven thiophene rings in length show that as the length of the oligomer increases, the extent of the r-conjugation also increases and is accompanied by increased PL quantum yields, due primarily to a rapid decrease in the nonradiative decay rate and a decrease in the yield of triplets [61]. 8.3.4.2 Solid state

In the solid state, quantum yields of fluorescence for thiophenes are generally lower than in solution by one or two orders of magnitude. This has been attributed to quenching of fluorescence by aggregation effects and is more pronounced for planar molecules [49]. Moreover, in the solid state, the twisted alkylthiophene chains tend to show higher PL efficiencies. X-ray diffraction studies [42] of regiorandom alkylthiophenes (50% head-to-tail) show very little long-range crystalline order, while films of regioregular alkylthiophenes with 80% head-to-tail content are semicrystallin? with cofacial packing of the more planar aromatic chains, stacked with 3.8 A separation between 7r-conjugated backbones. Although the chains of regiorandom polyhexylthiophene (with 50% head-to-tail interactions)

8.4 Electroluminescent devices

42 1

are insufficiently planar for aggregation effects to be particularly strong, the decreased PL efficiency in the solid state is attributed to limited excimer formation in the excited state, which is usually more planar than the ground state.

8.3.5 Concluding remarks We have discussed a number of the many nonradiative decay mechanisms considered to be active in oligomers. In the solid state environment which applies to optoelectronic devices, the inter-oligomer separation is reduced and the torsion of the oligomer backbone is also reduced, compared to isolated oligomers in dilute solution or the gas phase. Both of these factors favour aggregation of oligomers. This generally results in lower efficiencies for electroluminescence and improved charge transport and charge-separation efficiencies. Trends in solution are sometimes reversed in the solid state and there are clearly many factors and compromises which must be borne in mind when formulating strategies for synthesis of new oligomers optimised for both ease of fabrication (film deposition) and efficient operation in devices. However, as a general rule of thumb, it may be advantageous to employ relatively disordered films within EL devices, while better ordered films are more suitable for field-effect transistors and photovoltaic applications. We have discussed a number of factors that contribute to the lowering of the PL efficiency, particularly in the solid state, but have not been able to present quantitative measures of the effects of these processes. This is because there are relatively few measurements of absolute PL efficiency reported, but in view of the importance of PL efficiencyin setting limits to achievable LED efficiency, such values are now becoming available. Thin-film structures are best measured using an ‘integrating sphere’ to collect emitted light from all directions; this is described in [86, 871. There are relatively few values reported for thiophenes, largely because these values are low; Greenham et al. [87] report a value of around 2% for poly(3-hexylthiophene). This value is higher than that for most of the thiophene oligomers, but lower than the blue-shifted polymers discussed in the following section and also lower than another red-orange emitting polythiophene derivative, specifically designed for higher PL quantum efficiency [88].

8.4 Electroluminescent devices 8.4.1 Introduction In organic electroluminescent devices, the semiconductor layer is sandwiched between two electrodes, as shown in Fig. 1Oa. One electrode, such as gold (Au) or indium tin oxide (ITO), is chosen to have a high work function, for injection of positive charges (holes). The other electrode, often aluminium (Al), calcium (Ca) or

422

8 Electro-optical Polythiophene Devices

Cathode wlth low wwkfunction (e g. AI,Ca,Mg) Conjugated oligomer or pdymer Anode with high workfunction (e.g ITO, Au)

(1) Injection of charges, holes from the cathode, electrons from the anode

(2)

(3)

(4)

Electron-hole capture t o form excitons, both singlet and triplet

Radiative recombination of singlet excitons:

No fluorescence from

s,

+,so

triplet excitons

+

Figure 10. (a) Schematic structure of a single-layer organic LED. (b) Operation of a single-layer organic LED.

magnesium (Mg), is chosen to have a low work function for injection of negatively charged electrons. The operation of a typical single-layer organic electroluminescent device is summarised in Fig. lob. Once the charges are injected, electrons and holes combine to form excitons, in which the spins may either be opposed (singlet) or aligned (triplet). Electroluminescence is due to radiative decay of singlet excitons, from the first excited electronic state (SJ to the ground state (So). However, as discussed in the previous section, several nonradiative mechanisms also exist for de-excitation of singlet excitons, which have the effect of reducing the quantum efficiency for electroluminescence (EL) and photoluminescence (PL), by competing against the radiative decay of singlet excitons. Additional semiconductor layers (charge-transport layers) may be included between the emissive layer and the electrode to facilitate transport of charges of one polarity, while impeding charges of the opposite polarity, thereby encouraging radiative recombination within the emissive layer, as shown in Fig. 11, for a polymer heterojunction device [30]. The efficiency of EL diodes made with organic semiconductors is critical to their usefulness in display (or possibly lighting) applications. The overall quantum efficiency for generation of light within the diode (internal quantum efficiency), qnt. of a diode that operates according to the scheme shown in Fig. 10b has been summarised as the product of three terms: Vint = T

stq

8.4 Electroluminescent devices

423

Emissive layer Electrontransport

2-(4-biphenylyl)-5-(4-tert-butyIphenyl) 1,3,4-oxadiazole

PPV

[PBDI

PBD fin PMMA

Figure 11. Schematic energy level diagram of a double-layer polymer LED in which the thin PBD blend layer transports electrons, while blocking holes, thus enhancing radiative recombination in the PPV layer. Adapted from Brown et al.

where y is the ratio of exciton formation events within the device to the number of electrons flowing in the external circuit, rst is the fraction of excitons formed as singlets, and q is the efficiency of radiative decay of these singlet excitons [89]. In the absence of evidence to the contrary, the value of vst is generally taken to be 0.25 on the assumption of a 1: 3 ratio for sing1et:triplet generation. We note that there is direct evidence for the formation of triplet excitons in PPV-based LEDs [90]. A high value of y requires that there is good balancing of electron and hole currents, hence the use of multiple-layer devices designed to trap electrons and holes at suitably-designed heterojunctions. The value of q is related both to the properties of the emissive semiconductor and also to the ‘photonic’ structure of the device. The emissive properties of the organic semiconductor are best assessed from the solid-state photoluminescence efficiency; many of the more efficient EL devices use emissive layers with PL efficienciesof at least lo%, often higher than 50%. In this context, we note that the thiophene oligomers and polymers show lower PL efficiencies than commonly-used emissive layers such Alq3 [91] or PPV and its derivatives [92]. The role of the optical structure of the device in controlling the effectiveness of light emission has been increasingly recognised. There are a number of factors which need to be considered. Firstly, refraction of the emerging light reduces the fraction of emitted light that can leave the device in the forward direction (the rest being internally-reflected and thus wave-guided within the glass substrate/polymer layers) [93]. Secondly, the pattern of standing wave electromagnetic photon modes imposed by the presence of at least one reflective electrode (usually the metallic cathode) modifies the coupling between the exciton and final photon state, and thirdly, coupling between excitons in the emissive semiconductor and excited states in the metal layers can result in energy transfer and reduced emission (transfer of energy to surface plasmon modes). The effects of these latter two

424

8 Electro-optical Polythiophene Devices

factors cause strong suppression of emission when excitons are formed too close to the cathode, typically at distances less than 50 nm. A summary of these effects in presented in Ref. [94].

8.4.2 Historical survey of organic LEDs 8.4.2.1 LEDs based on molecular semiconductors In the mid1950s, electroluminescence was reported from cellulose films doped with organic dye molecules [95, 961. Ten years later, the first report of electroluminescence from organic semiconductors was the observation of blue emission from molecular crystals of anthracene. Pope et al. [8] used crystals 10-20 pm thick, prepared by sublimation and from solution. As electrodes, they used either silver paste electrodes of unequal area on opposite sides of the crystal or else liquid electrodes with aqueous sodium chloride solution as electrolyte. Large voltages, typically 400V were required and the solid state devices could be operated under either DC or AC voltage, while the LEDs with solution electrodes emitted light only under square-wave AC operation, during the parts of the square-wave cycle where the voltage was changing most rapidly. Helfrich et al. [9] used glass tubes filled with liquid electrodes, which were cemented to either side of a single crystal, typically 5mm thick. Large operating voltages, of the order of several kV were required, though the applied electric fields were in fact comparable with those currently employed in polymer LEDs (-lo7 Vm-'). However, the large drive voltages and use of liquid electrodes (usually solutions of anthracene cations and anions) made the devices commercially unattractive. One motivation for investigating oligomers and small molecules of relatively short conjugation lengths and hence high T-T* energy gaps is to achieve blue electroluminescence, particularly since there are few blue phosphors to be found amongst inorganic semiconductors, unless multiple-quantum-well structures are employed. For a review of the early work on organic electroluminescence, see also the review by Kalinowski [97]. By using evaporated thin films rather than single crystals of anthracene and perylene, Vincett et al. I981 reported drive voltages as low as 12V, using solid state device structures in which the film is sublimed onto an oxidised aluminium electrode, followed by a semi-transparent gold electrode. External quantum efficiencies in the range 0.03-0.06% were calculated for sublimed film anthracene devices. Since then, significant progress has been achieved, first by researchers at Kodak laboratories [91, 991, then in Japan [loo-1031, using small molecules, such as Alq3 and oxadiazoles, both as charge-transporting layers and as emissive layers, either blended within a polymeric host or deposited as sublimed films. Tang et al. [913 reported external quantum efficienciesup to 1YOfor a bilayer LED of the structure: Glass/ITO/Diamine/Alq3/Mg: Ag, with luminous efficiencies of 1.5 lumen/Watt and brightness exceeding 1000cd m-* at operating voltages less than IOV. The aromatic diamine layer served as a hole-transporting layer, while electroluminescence originated from the Alq3 layer. Improved quantum efficiencies

8.4 Electroluminescent devices

425

(2.5% photons/electrons) were reported [99] for a multilayer device of the form: Glass/ITO/diamine/Alq3/doped Alq3/Alq3/Mg: Ag. The amorphous aromatic diamine layer served to transport holes, while the undoped Alq, layers transported electrons. The doped Alq3 layer contained fluorescent dye molecules, such as DCM and Coumarin, allowing energy transfer from the excited Alq, host molecules to the fluorescent dye guest molecules. It was determined that a narrow recoombination zone was located at the diamine/Alq3 interface, extending up to 50A into the Alq3 layer. In undoped Alq3, a broad emis$on zone was observed, due to diffusion of excitons on a lengthscale of around 200 A. In the LEDs fabricated with the doped Alq3 layer close to the interface, efficient energy transfer to the fluorescent dye guest molecules resulted in a much narrower emission zone. Fluorescent guest species are now found to improve both efficiency and also the operating lifetime of molecular organic EL diodes [ 1041. Adachi et al. [loo] also reported a three-layer LED of the structure Au/TPD/ Acene/PV/Mg, where TPD is an amorphous aromatic diamine for hole transport, PV is an amorphous layer of a perylene tetracarboxylic acid derivative and the polycrystalline acene layer consisted of either anthracene, perylene and coronene. Much higher efficiencies (up to 0.04% photons/electron external quantum efficiency, 0.001% W/W power conversion efficiencies) were obtained [loll by using a similar structure but a 1Zphthalo perinone derivative rather than the acenes. By using a bilayer device structure using an oxadiazole (PBD) as an electron-transporting layer, Adachi et al. [ 1021 obtained electroluminescence from a hole-transporting material, in this case a triphenylamine derivative, NSD. Green emission was obtained, with luminescence efficiencies of 0.2 lumen/Watt and a turn-on voltage as low as 4.0 V. Intensities up to 1000cd m-2 were reported for drive voltages of 16 V and current densities of 100 mA cm-2. Adachi et al. [lo31 also fabricated a triple layer device of the structure: ITO/TPD/ NSD/PBD/Mg : Ag, which showed luminance efficiencies as high as for the previous ITO/NSD/PBD/Mg : Ag bilayer device even when the thickness of the emitter layer (NSD) was as low as 50 This clearly demonstrated confinement of excitons within the emitter layer, by insertion between two materials of larger T-T* energy gaps which allowed only unipolar charge transport into the emissive layer. Adachi et al. also emphasised the important role played by the charge-transport layers in providing a barrier layer between the emissive layer and the metallic electrodes, at which excitons are considered to be effectively quenched [105, 1061, see also more recent investigations [94].

A.

8.4.2.2 Polymeric LEDs

The developments in the field of molecular and oligomeric LEDs naturally spurred interest in polymeric materials as well, both out of pure curiosity and also because polymers often show better long term stability and processing advantages compared to smaller molecules. The first polymer to be reported to act as a charge-transport medium in EL diodes was poly(vinylcarbazo1e); in these devices, emission is due to a dye dopant, such as perylene, as Partridge described in a series of articles in early 1980s [107-1101. EL from conjugated polymers was first reported in 1990 by

426

8 Electro-optical Polythiophene Devices

Burroughes et al. using poly(pheny1enevinylene), PPV, as the light emitting semiconducting layer [l 111. This report became the starting point for an intense activity in this new area in the field of conjugated polymers. Soon afterwards, a number of other conjugated polymers were successfullyused in electroluminescentdevices,confirming the Cambridge results. Among these we find MEH-PPV (alkoxy substituted PPV) [29], poly(para-phenylene) (PPP) [112], and poly(3-alkylthiophene) [113]. The levels of efficiency for the first generation of polymer LEDs were low, of order 0.01% in the first PPV LEDs, which were fabricated as single layer devices with aluminium as negative electrodes. Two of the major factors influencing the external quantum efficiency are the balance between the injection conditions for holes and electrons, and the combination of the refractive indices of the layer that comprise the device. The latter has been discussed by Greenham et al. [93]. The injection conditions are important because the generation of light requires that both types of carrier are injected into the semiconducting layer, so that a recombination can occur, forming bound excitons. These excitons can then decay radiatively. In the first devices, the injection conditions were far from balanced, resulting in a larger hole current than electron current and, subsequently, a low electroluminescence quantum efficiency of about 0.05%, compared to a photoluminescence quantum efficiency of about 8% for the materials used in those devices. The injection properties can easily be more balanced by changing to other metals for use as negative electrodes. Braun et al. showed this for MEH-PPV, where the efficiency could be raised to above 1% simply by using calcium instead of aluminium as electron injecting contact. As mentioned earlier, an alternative method (borrowed from the molecular devices) is to make heterojunction devices in order to confine the otherwise favored carrier at an internal interface in the device, thereby increasing the probability for recombination. This was demonstrated using PPV as the hole transporting layer and an electronegative, cyano-substituted PPV derivative, as the electron transporting layer. Such a double layered device, with aluminium as negative electrode, showed a remarkable increase in efficiency (above 4% internal quantum efficiency) compared to the single layer device [32]. From then on, many conjugated polymers have been shown to exhibit electroluminescence and a number of different device design improvements have been suggested. Recently, the positive electrode IT0 has become the target of an increased research effort. This follows the reports that the I T 0 can be one of the causes for the limited lifetime shown by many polymer light emitting diodes [I 14-1 191. Because I T 0 is an oxygen rich compound, there is for instance a danger that oxygen can diffuse into the polymer layer, causing a degradation of the luminescent properties since most conjugated polymers are sensitive to photo-oxidation. To avoid this, it is possible to use doped conjugated polymers as a conducting, but chemically blocking, layer between the IT0 and the luminescent polymer. The first report of the use of a conjugated polymer in such a role is due to Hayashi et al. who formed an electrochemically-synthesised layer of poly(3-methyl thiophene) formed on an I T 0 substrate, and completed the EL diode with a sublimed molecular film (perylene) and an evaporated metal cathode [ 1201. Both polyaniline and a polythiophene derivative, poly(3,4-ethylene-dioxy-thiophene)have been used as such a layer, resulting in dramatically improved device characteristics: the injection voltage drops, the luminous

8.4 Electroluminescent devices

427

efficiency increases and, most significantly, the rate of decay of the luminance decreases by up to two orders of magnitude [121-1251. The polymeric anodes can also be used as complete replacement for the I T 0 layer, enabling the use of flexible substrates for the devices [20, 221.

8.4.3 LEDs based on oligothiophenes As mentioned above, the thiophene oligomers show low PL efficiency as solid films and are not therefore obvious candidates for use as emissive layers in organic LEDs. However, the high field-effect mobilities reported for sexithiophene and related oligomers have generated interest in their application to LEDs, particularly for their use as charge transport layers. The reason for this is that high current densities are required in LEDs, and low carrier mobilities result in problems with space-charge build-up, so that currents are limited. This is well-demonstrated for PPV-based LEDs [126]. LEDs have been fabricated with oligothiophenes and oligophenylenes, as singlelayer devices and in bilayer and multilayer devices. However, electroluminescence efficiencies generally remain much lower than those of polymeric devices. The first reported single-layer thin-film oligothiophene LEDs were fabricated by Geiger et al. [39] using spin-coated films of cycloalkane end-capped oligothiophenes (ECnT), as shown in Fig. 2b, in a structure ITO/ECnT/Al, where IZ is the number of a-conjugated thiophene rings per oligomer. The efficiency of these LEDs was low (10-3-10-2%) and decreased for longer chain lengths. ECST gave devices with the highest efficiencies and this appears to be correlated with the presence of smaller microcrystals within the film and hence increased disorder/reduced aggregation effects. Uchiyama et al. [1271later reported electroluminescencefrom single layer devices with highly crystalline films of a-w-dimethylsexithiophene (Me-6T-Me) deposited under ultra-high vacuum in the structure ITO/Me-6T-Me/Al. Although high rectification ratios were obtained (1500at f l O V ) and onset voltages for EL were low (4V), very low quantum efficiencies ("3 x were obtained, presumably as a result of the high crystallinity resulting in the transition from the lowest excited state to the ground state being dipole-forbidden in the crystal. Uchiyama et al. [1281have reported an increase of the EL efficiencyby a factor of up to 1000 by fabricating a bilayer device in which a shorter oligomer of higher energy gap is inserted between the 6T oligomer and the aluminium electrode in order to block holes and enhance recombination at the heterojunction, following the strategy of Adachi et al. [ 1021This strategy has also been u!ed by Muccini et al. [12?] in a multilayer oligomeric LED of the structure IT0/400 A a-4T/500 a-6T/100 A (*.-4T/Al. Horowitz et al. have prepared bilayer LEDs [1301 using unsubstituted sexithiophene (6T) and substituted derivatives, either substituted with two decyl sidechains, (2D6T), as shown in Fig. 2c or with triisopropylsilyl end-groups, (DPS6T), as in Fig. 2d. Compared to single-layer IT0/6T/AI devices, the quantum yield of bilayer LEDs can be increased by three to four orders of magnitude, from

428

8 Electro-optical Polythiophene Devices

4x for IT0/6T/Al to 2 x lop3%for the bilayer device IT0/6T/2D6T/Al and 1.5 x lop2% for the bilayer device IT0/6T/DPS6T/Al. There are two main effects which contribute to t h s improvement: Disorder and increased separation between oligomers Substitution with bulky tri-isopropyl end-groups is considered to increase separation of chains, as evidenced by X-ray diffraction [131].This disorder and increased separation applies both longitudinally and laterally, and thus reduces the occurrence of charge-transfer exciton formation (considered to provide a nonradiative decay channel involved in photoconduction) and also lateral aggregation of chains (n-stacking), which can result in Davydov splitting, so that in the perfect crystal, the optical transition from the lowest lying excited state to the ground state is dipole-forbidden. The bulky side-groups therefore render the fluorescence partially allowed. More efficient electron injection The conductivity of films of the randomly P-substituted oligomers is usually lower than that of unsubstituted 6T [40]. Many conjugated polymers and oligomers are p-type materials, in which the holes have a much higher mobility than the electrons. Also, since the electron injection barrier from cathode to the LUMO level is usually much greater than the hole injection barrier from I T 0 to the HOMO level, EL quantum efficiency is usually limited by electron-injection from the cathode. This appears also to be the case for 6T and its derivatives; a device of the form IT0/6T/DPS6T/Al shows increased efficiency compared to a single-layer device, while exchanging the order of the oligomer films (i.e. constructing ITO/DPS6T/6T/Al) reduces the efficiency. An explanation for the higher efficiency in the IT0/6T/DPS6T/Al device is that the voltage is not dropped uniformly across both layers but that the field is much higher in the more resistive substituted DPS6T layer, thus lowering the tunnelling barrier for injection of electrons, while holes are still transported relatively easily through both layers, as shown schematically in Fig. 12. Oligomers have also found applications as charge-transport and charge-blocking layers. Hosokawa et al. [ 1321have fabricatedmultilayer oligomericLEDs of the structure IT0/6T/TBS/Alq/Mg : Ag, where the emissive layer is Alq. TBS, a sexiphenyl (6P) derivative is used for electron confinement in the Alq emissive layer. a-6T is used as a hole-transporting layer, rather than triphenylenediamine derivatives previously used in bright organic LEDs. Bright green emission of 2300cdm-’ was obtained at a voltage of 13 V and current density of 313 mAcmp2. Recently, Noda et al. [1331 reported much improved electroluminescence quantum efficienciesfor oligo(thiophene) derivatives with bulky end-substituents consisting of triphenylenediamines, as shown in Fig. 2e. These materials readily form stable amorphous glasses with relatively high glass transition temperatures, around 90°C. They form uniform amorphous films by vacuum deposition and emission colors from light blue to orange have been obtained by increasing the number of thiophene rings from one to four. Recently, the authors reported a luminous efficiency of 1 lumen/Watt and maximum luminance of 13 000 cd m2 (at 18 V drive voltage) for one of these oligothiophene derivatives in a bilayer device [134], sandwiched between an

8.4 Electroluminescent devices

6T

DPS6T

DPS6T

429

6T

Higher EL efficiency due to narrower tunnelling barrier for electrons

Figure 12. Schematic diagram showing enhanced electron injection (reduced width of barrier) in double layer devices when the highly resistive DPS6T layer is inserted between the 6T layer and cathode. Adapted from Horowitz et al. [130].

I T 0 anode and a Mg/Ag cathode, with an Alq3 layer next to the cathode, as electrontransport layer. The bright yellow emission was observed to originate from the oligothiophene layer, also in the double-layer device. In addition to their potential use as blue-emitting materials, polarised electroluminescence [135,136]has recently been achieved in a multilayer LED in which oriented para-sexiphenyl(6P) oligomers form the emissive layer and also in LEDs fabricated with oriented sublimed films of a-sexithiophene (a-6T). Era et al. [135]oriented the sexiphenyl oligomers by sublimation onto a template consisting of a thin film of sexiphenyl oligomers, which had been rubbed in one direction, in order to induce alignment parallel to the substrate. Marks et al. [ 1361 reported polarised emission from sublimed films in which sexithiophene oligomers stand almost perpendicular to the substrate. The anisotropy was found to increase as the size of the crystallite grains was increased and the proportion of grain boundaries was reduced. The relative ease of inducing alignment of discrete oligomer chains, rather than their polymeric counterparts may yet render the use of oligomer films particularly advantageous in certain applications, such as waveguided or microcavity electroluminescent devices.

8.4.4 LEDs based on polythiophenes Polythiophenes can show good red emission - a color that 'has been hard to produce with other polymeric or molecular semiconductors. Combined with convenient solution-processing, e.g. for the poly(3-alkylthiophenes), this provides a good rationale for their use in EL diodes, in spite of relatively indifferent PL efficiencies [87].

430

8 Electro-optical Polythiophene Devices

Polythiophene light emitting diodes were first reported by Ohmori et al. in 1991 [1131. These devices used poly(3-alkylthophenes) as emitting materials, with different lengths of the side chains. The device design resembled in many ways the one used for the PPV LEDs, with the difference that a magnesium-containing indium electrode was used as electron injector instead of aluminium or calcium. The report also showed that the luminescence output is strongly dependent on the length of the side chains, with four times more light coming out of a diode with 22 carbon atoms in the side chain than with 12. No efficiency numbers were given in this first report, but Braun et al. shortly afterwards reported a room temperature efficiency of 0.003% when using poly(3-octylthiophene) and Ca as negative eletrode [ 1371. At low temperatures (80K) t h s increased to about 0.01%. Since then, a large number of polymer LEDs using different polythiophenes have been presented, both as ‘pure’ poly(thiophene) [138-1441 devices and in co-polymers [145, 1461. Diodes made with poly(3-alkylthiophenes) can show efficiencies as high as 0.2% [147], and enhanced electroluminescence has been reported from blends of poly(3-hexylthiophene) (P3HT) in poly(2-methoxy-5-(2’-ethylhexoxy)-1 ,Cphenylene vinylene) (MEH-PPV). The external EL quantum efficiency 77 of Ca/P3HT :MEH-PPV/ITO diodes initially increases with P3HT content and goes through a maximum at 71 = 1.7% photons/electron with only 1wt.% P3HT [148]. Below we will take a closer look at some of the most interesting polythiophene LED designs. 8.4.4.1 Polythiophene LEDs covering the whole visible spectrum and a bit more

In order to explain the wide range of colors available from polythiophene light emitting diodes, it is necessary to look at some of the underlying phenomena. In 1988 it was found that thin films of poly(3-alkylthiophenes) changed color when the substrate was heated [149]. This color change is always a shift towards the blue region (shorter wavelengths, hgher energy), which indicates an increase of the band gap. In addition to this, the temperature range in which the color change occurs, is shifted towards lower temperatures when the length of the side chains is increased [150]. This can be understood by considering that thermal movements of the side chains, shifting from a predominantly planar structure at low temperatures to a random coil configuration when the temperature is increased, force the chain out of planarity, thereby decreasing the conjugation length of the polymer. This model is strongly supported by vibrational spectroscopy measurements [151- 1531. Longer side chains would then give larger influence, as observed. By this decrease of the conjugation length, which means that the overlap of the pz-orbitals decreases, the band gap is increased and the absorption shifts to make the film look more purple or blue. When the temperature is decreased to the initial level, the mechanism is reversed and the color is shifted towards red again. The same mechanism can also be used to tune the band gap at fixed temperatures, which is what can been used in the design of polythiophenes for use in polymer tight emitting diodes, where the band gap governs the color of the emitted light. Shorter conjugation length gives a blue-shifted emission and long conjugation length results in emission in the red region [154]. This variety in conjugation lengths is accomplished by using substituents that are bulkier than the alkyl chains. Doing

8.4 Electroluminescent devices

43 1

so, the consecutive thiophene rings are forced out of planarity even at room temperature, resulting in the desired change in band gap energy. To increase the ease of synthesis, it is favourable to have small repeating units, something that is offered by the thiophene system, where substituents can be attached either at the 3-position or at both the 3- and 4-positions. Using different amounts and sizes of the substituents on the repeating unit, the amount of twisting of the main chain can be controlled. This approach has been developed by a number of groups, including those at Groningen [ 1551 and at Linkoping/Gothenburg. A further consideration in this discussion is the regioregularity of the polymers. When polymerising monosubstituted thiophenes, they can link in both head-to-tail and head-to-head configuration, see Fig. 5. Large amount of head-to-head coupling gives shorter conjugation length, and blue shift, compared to a regioregular head-to-tail coupled polymer [84]. By building devices with the type of polymers described, and for the high band gap polymers including a hole blocking layer 2-{4-biphenylyl}-5-{4-tert-butylphenyl}1,3,4-oxadiazole, PBD (structure is shown in Fig. lb) to increase the efficiency, it

IV.PCHMT III. PCHT

I

300

111 I1

IV

400

11. PTOPT

500

600

1. POPT

I

700

800

900

Wavelength (nrn) Figure 13. Molecular structure and electroluminescencespectra of the polythiophenes covering the visible spectrum. I. POPT, poly({4-octylphenyl}thiophene). II. PTOPT, poly(3-{4-octylphenyl}2,2'-bithiophene). 111. PCHT, poly(3-cyclohexylthiophene). IV. PCHMT, poly(3-cyclohexyl-4methylthiophene).

432

300

8 Electro-optical Polythiophene Devices

350

400

450

500

550

600

650

700

Wavelength (nm)

Figure 14. Emission spectrum of the UV-diode. The emission peak has its maximum at 394nm.

has been possible to cover the whole visible spectrum. The chemical structures for the polymers employed are shown in Fig. 13, together with the electroluminescence spectra of diodes made from these polymers. The synthesis of these polythiophene derivatives is described in [154]. Polymer I, POPT, requires some special attention because it exists in two different forms. The pristine form, labelled I, is upon thermal treatment or in the presence of chloroform vapor irreversibly converted into a more crystalline form, labelled POPT* or I*. This increased ordering is made possible by a high regioregularity (94 f2%, as determined by 'H NMR) [140, 1561, and allows for an optical signature previously found only in highly regioregular poly(3-octylthiophene) [157]. Using another aspect of the hole-blocking material PBD, namely its luminescence properties, it is also possible to extend the emission from polymer LEDs into the ultra-violet region [158]. PBD shows photoluminescence in the UV region, but normally no electroluminescence. However, this can be accomplished by blending polymer I1 (PTOPT) with PBD, in a ratio of 1:50 (PTOPT : PDB). The conjugated polymer is then used as hole injecting material, and it is necessary to keep the amount of PTOPT low to get at least some dispersion of the polymer into the PBD. This increases the contact area drastically, making it possible to inject enough holes into the PBD layer to get radiative recombination and light emission from the device. The emission spectrum from such a device is shown in Fig. 14. Regarding other UV emitting devices with possibilities for large area device fabrication, one must consider hybrid organic-inorganic materials such as polysilanes [ 1591.

8.4.4.2 Intrinsically-polarised polymer LEDs Apart from covering the visible spectrum, there are a number of other aspects that can be utilised when making polymer LEDs. For instance, the intrinsic one-dimensionality of the polymer chains can lead to anisotropy in terms of macroscopic

8.4 Electroluminescent devices

433

properties, such as electrical transport [160]. It has also been shown that oriented films of conjugated polymers give highly polarised photoluminescence [ 161, 1621, and by stretch alignment of an electroluminescent polythiophene film, an intrinsically polarised light emitter can be manufactured [ 1631. After mechanical stretching, the chains and their electrical dipoles will be preferentially oriented in the stretching direction, and this produces photoluminescence polarised mainly in the orientation direction. If the electrons of the polymer are instead excited electrically to give electroluminescence, we would expect a similar result. To get a film that is thin enough to be used in a device, the electroluminescent film (PTOPT, polymer I1 in Fig. 13) is spin coated on top of a polymeric carrier substrate. This bi-layer is then thick enough to be handled, and can be stretched mechanically. The thin luminescent film ( 4 0nm) is then transferred to the device substrate, which has been prepared by spin coating of a non-oriented film of the same polymer to enhance adhesion. The device can be made with different processible polythiophene derivatives, nonoriented or oriented, suggesting that this method is a general way of depositing a polymer layer on a substrate. The anisotropy in the emission is large enough (up to 3.1 :l), as shown in Fig. 15, to tell us that the emission takes place mainly in the oriented film closest to the negative electrode, rather than in the non-oriented film close to the positive electrode. The emission anisotropy is also roughly the same as the anisotropy in absorption. Polarised polymer LEDs have also been made using other techniques and a device using another polythiophene derivative was recently presented by Bolognesi et al. [164]. In this case oriented poly(3-decylmethoxythiophene), PDMT, films were produced using the Langmuir-Blodgett technique. Normally, the polymer would be mixed with, for instance, arachaic acid to facilitate the LB preparation, but as this drastically decreases the lifetime of the devices, a homogeneous film is preferable. The oxygen containing side-chains in PDMT makes it possible to make such homogeneous LB-films [ 165,1661.These films will then have the polymer backbones preferentially oriented along the dipping direction, as shown by a ratio of 9 : 5 between the

550

600

650

700

750

Wavelength (nm)

Figure 15. Absolute intensity emission spectra of the polarised LED parallel and perpendicular to the orientation direction.

434

8 Electro-optical Polythiophene Devices

absorption intensities parallel and perpendicular to the dipping direction. The ratio between parallel and perpendicular emission is however smaller, being about 4 : 3, possibly explained by a deorientation of the polymer backbones either during the metal evaporation process or because of heat evolved when a voltage is applied across the film. The devices show an external quantum efficiency of 0.05%. Polarised polymer LEDs have also been made from other conjugated polymers, and using different techniques. Some examples are rigid rod polymers [167], rubbing aligned PPV-derivatives [I 681, and tensile drawn PPV [ 1691.

8.4.4.3 Polytbiophenes in microcavity structures Recently, the field of microcavities in combination with polymer LEDs has attracted a great deal of interest [170-1831. There are different types of microcavities, where the simplest one is a single reflecting surface. The next step would be to combine two parallel mirrors at a small distance. If an emitting dipole is put at a distance comparable to the wavelength or shorter to such a mirror, or combination of mirrors, the so-called microcavity effect can be observed. What happens is that the dipole starts to interact with its mirror image, causing changes in the decay rates of the optical processes involved [184]. At different positions inside the cavity, the decay processes will be enhanced or suppressed, depending on the allowed modes. The type of structure normally used when making polymer LEDs is in itself resembling the microcavity structures, being planar with a thickness of the active polymer layer smaller than the wavelength of the emitted light and very often having at least one metallic electrode that can act as a mirror. By the addition of a spacer layer, the spacing between the mirrors can be controlled and the desired modes achieved. A typical device fabrication sequence is as follows: An aluminium layer is deposited on a silicon substrate. On top of the aluminium surface a transparent spacer layer of the thermosetting polymer benzocyclobutene (BCB) is spin coated, and by changing the spin rate, different thicknesses can be obtained. This makes it possible to control the distance between:mitting tipoles and the aluminium mirror. A thin layer of chromium and gold, 5 A and 90 A respectively, is then evaporated on the BCB surface to form a transparent positive electrode. On top of this, the

Figure 16. Chemical structure of PDOPT.

8.4 Electroluminescent devices

43 5

Figure 17. Device structure of the microcavity polythiophene LED.

polymer LED is built by spin coating an electroluminescent polythiophene film (in this example PDOPT, see Fig. 16), evaporFting a hole blocking layer (PBD) if needed and a thin layer (in total about 300A) of calcium/aluminum as negative electrode and semitransparent mirror. This somewhat complicated device structure is shown in Fig. 17. By changing the BCB thickness, the resonance within the poly(thiophene) film will occur for different energies. In Fig. 18 the emission spectra from PDOPT-LEDs in cavities with different BCB-thicknesses can be compared to the emission spectrum of a non-cavity ITO/PDOPT/PBD/Ca/Al device. From the figure, we conclude that the cavity can be used to enhance the emission from a specific position within the polymer film, where the excitons couple with the allowed modes for the specific thickness of the microcavity. In this way, by changing the BCB thickness, different parts of the emission spectrum can be chosen. Another

300

400

500

600

700

800

900

Wavelength (nm)

Figure 18. Emission spectra of the microcavity LEDs with different BCB thicknesses, compared with the emission spectrum of a non-cavity LED (top).

436

8 Electro-optical Polythiophene Devices

important aspect is that the bandwidth of emission is considerably reduced, in this case to linewidths as narrow as 20 nm full width at half-maximum, compared with 110 nm for the non-cavity device. 8.4.4.4 Sub-wavelength size polymer LEDs

For certain applications, such as scanning near field optical microscopy (SNOM) [185-1871 and nano-lithography, it is important to have light sources with extremely small dimensions. In the SNOM case, it is necessary that either the light source or the light detector is smaller than the wavelength of the emitted light. Two different structures with such small light sources have been made using polythlophene LEDs. In these devices, radically different approaches have been used to reach the small size. In the first [188], the electroluminescent polythiophene film is spin coated on top of a patterned contact structure with circular contacts, l00nm in diameter, made from a porous micro-filtration membrane. The pores are filled with another polythiophene derivative, PEDOT (poly{3,4-ethylene-dioxythiophene}) [189, 1901 in its doped form before spin coating with the luminescent polymer. This is accomplished by attaching a membrane with open pores to a gold surface. PEDOT is then electropolymerised in the pores from a water solution using sodium-poly(styrenesu1fonate)as counter ion. The polymer filled membrane then acts as a patterned electrode with randomly distributed contacts. On top of the luminescent film, semi-transparent Ca/Al electrodes are evaporated through which the light is emitted. The resulting structure is shown in Fig. 19. In order to address some of the drawbacks with the type of device described above, such as a heat expansion of the PEDOT polymer in the pores causing damages to the luminescent layer and thereby device failure, and the fact that the light is emitted through a metal layer, decreasing the light output, a second type of nanometer size LED has been developed [22], using the fact that polymers tend to phase separate when mixed. By having a low fraction of luminescent polymer in a matrix polymer, say two or five percent PTOPT (polymer 11)in poly(methy1 methacrylate) (PMMA), the result is a polymer blend with PTOPT as the minority phase. The PTOPT phases are spread out in the PMMA matrix, and from scanning force microscopy investigations it is shown that the luminescent polymer forms domains with a size distribution of 50 to 200 nm. Substituting the luminescent layer in a polymer LED structure with the polymer blend (spin coated in the same fashion) results in an array of randomly distributed light sources smaller than 300nm. A photograph of such a device, taken in an optical microscope, is shown in Fig. 20. It should

Figure 19. Device structure for nanoLEDs using microfiltration membranes.

8.4 Electroluminescent devices

43 7

Figure 20. Photo of emitting nanoLED array. Magnification 1000 X, N/A 1.25.

be remembered that although the microscope limits the resolution to about 300 nm, the individual light sources are probably smaller than that. Using this approach, it is easy to fabricate simultaneously a large number of light sources smaller than the emitted wavelength. Depending on the choice of electroluminescent polymer, any desirable color of the emitted light could be achieved/obtained. The phaseseparation method can of course also be extended to other types of luminescent polymers besides the thiophenes, as was recently shown by Adachi et al. [191].

8.4.4.5 Voltage-controlled colours The phase separation used in the nanometer size LEDs can also be utilised to get light with voltage controlled colors from polythiophene LEDs [192]. In this case, different electroluminescent polythiophenes are mixed with each other. Because the color of the emitted light is a function of band gap, the polymers will have different turn-on voltages, where they start to emit light. With the different polymers in different phases in the film, we have a great number of parallel diodes. When a voltage is applied between the electrodes, the low-bandgap material will start to emit light first, followed by the others when the voltage is increased. Emission spectra at two different voltages from such a device is shown in Fig. 21. By choosing the voltage drive scheme, it is possible to compensate the hgher intensity found at higher voltage by decreasing the fraction of time during which voltage is applied. In this device, it is necessary to have a much larger amount of the high band gap polymer compared to the low band gap, because it is energetically favourable for the radiative emission to occur in the latter. To avoid this, at least partially, a non-emissive matrix polymer can be included in the blend in order to suppress the energy transfer. For some combinations of polythiophenes, PMMA was found to work as such a blocking matrix, without simultaneously blocking the electron and hole transport at the interfaces between the polymer blend and the electrodes. Including such a matrix makes it easier to balance the ratio between the luminescent polymers to reach desired colors [193]. In Fig. 22, emission spectra from a device with three polythiophenes in combination with PMMA and a hole blocking PBD layer are shown.

438

8 Electro-optical Polythiophene Devices

350

400

450

500

600

550

650

700

750

Wavelength (nm)

Figure 21. Voltage dependence of electroluminescence from a 50 : 1 blend of PCHMT and PTOPT. Top: 21 V, bottom: 28V.

As can be seen, the red/orange emitters start to give light first, followed by the yellow/green and the blue when the voltage is increased. A problem that remains to be solved is to find a good green emitter, in order to get a more saturated green color from the voltage controlled device. With the polymers available today, the light emission varies from red via a broad band emission in the ‘beige’

.--

.,.

,

,--. I

j 350

\ \

-.

,’

_--._

--..

400

450

500

550

600

650

700

750

Wavelength (nm)

Figure 22. Electroluminescence spectra of a variable color device based on polymer II, III and IV, and using PMMA as energy transfer blocking matrix. In the 21 V curve, some of the emission also comes from the PBD layer. Applied voltages are indicated in the plot.

8.5 Photoconductive and photovoltaic devices

439

region towards the blue. However, this is not entirely a drawback, as it opens up the possibility to tune the device in such a way as to get white light emission [194], something that is desirable in a great number of applications.

8.5 Photoconductive and photovoltaic devices 8.5.1 Introduction Photoconductivity was first observed in anthracene by Pochettino [l] and Volmer [2] almost a century ago. Interest in using organic materials for xerography began in the 1950s and 1960s. Now small organic semiconductors have found widespread applications in xerography and laser printing. The principal advantages of using organic materials are that they can be prepared as flexible layers, suitable for coating cylindrical drums and belts, at relatively low cost, with spectral sensitivity throughout the visible region and into the near infra-red by the use of a mixture of photoreceptors of different semiconductor gaps, and with lower toxicity than the chalcogenide alloys, which they have largely replaced. Flexible layers for photogeneration and charge transport are prepared by dispersing the small organic semiconductors within a host polymer (blending) and depositing layers from solution. Because thermal charge generation rates are low, organic photoconductors can withstand the high voltages used in xerography. The disadvantages are that organic materials typically are not as resistant to abrasion as inorganic materials and have low mobilities for charge transport. Organic photoconductivity is generally explained in terms of photogeneration not of free electrons and holes, but rather of bound electron-hole pairs (excitons) and their subsequent dissociation into charge carriers which must then be transported to the electrodes: (i) Absorption of light produces excitons, which will not readily ionise because their binding energies are greater than the energy available from thermal fluctuations, kBT. (ii) Exciton ionisation and subsequent dissociation into charge carriers rarely occurs within the bulk of the film, but more usually when excitons interact with interfaces (with electrodes and other organic layers), trapped charges and regions of high electric field or suitably arranged discontinuities between respective HOMO and LUMO levels across a heterojunction. Hence, the heterojunction structure of Tang [ 1951 results in efficient dissociation, though only of those excitons generated within a diffusion length of the heterojunction. Oxygen is also considered to play an important role in the dissociation of excitons at surfaces, by acting as deep traps for electrons [196, 1971. (iii) When charge pairs are generated, the less mobile carrier can be deeply trapped within the bulk of the film, leading to a build-up of space charge and hindered charge separation. Most organic photoconductors support unipolar charge transport, in which one polarity of charge carrier is much more mobile than

440

8 Electro-optical Polythiophene Devices

the other. It is therefore important to carefully select materials with high carrier mobilities for (separately) electrons and holes. The use of a bi-layer device can circumvent this problem. In xerography, a thin layer for charge generation and a relatively thick layer for charge transport. Under applied fields, this has yielded very high xerographic gains (charges generated per incident photon), approaching unity and avoids potential problems of changes in sensitivity with time. In solar cells, the use of a p-n heterojunction device efficiently transports the electrons and holes away from the interface, so that they do not hinder further charge separation. For a much more detailed review of the organic photoconductivity, the interested reader is referred to the excellent books by Kao and Hwang [lo61 and (particularly for xerographic applications) Borsenberger and Weiss [5] and references therein. Although the small organic semiconductors used in xerography can often be viewed as oligomers, e.g. tetracene, pentacene, derivatives of distyrylbenzene, there have to date been relatively few reported investigations of the photoconductive and photovoltaic properties of oligomers of conjugated polymers. Having said that, the mechanism of photoconductivity in sexithiophene has been intensively studied, and is described in the following section.

8.5.2 Mechanism of photoconductivity in sexithiophene The photoconductivity action spectrum of sexithiophene is observed to be antibatic (complementary) with the photoluminescence excitation spectrum; at excitation energies higher than 2.2 eV, the fluorescence yield of sexithiophene thin films decreases rapidly, just as the material starts to become photoconductive. This indicates a new nonradiative decay channel opens up at higher excitation energies than that of the 1 'B, exciton which is considered to be responsible for photoluminescence. Zamboni et al. [ 1981 noted that the crossover between photoluminescence and photoconductivity occurs close to the energy of the 2 'A, exciton band, determined by two photon excitation spectroscopy [58] which might then result in singlet fission into two long-lived triplets or else offer an efficient crossing to extended ionised states, such as charge-transfer excitons in which a weakly bound electron-hole pair is localised over two or more adjacent oligomers. Crystallographic studies of sexithiophene thin films [74, 1991 conclude that oligomers are arranged in stacked layers. Within the layers, the oligomers are arranged with their long molecular axes oriented at 32" to the plane normal, with their broad planes arranged in a herringbone pattern, in common with many molecular crystals. Conductivity is highly anisotropic; charge carriers are much more mobile within layers than between stacks. Dippel et al. [76] proposed that in films of sexithophene, the stacked layer structure is favourable for photoconductivity. At energies higher than the lowest singlet excited state (1 'BJ, a charge-transfer exciton is formed in which the weakly bound electron-hole pair is distributed over two or more adjacent molecules. Of those charge-transfer excitons which are generated longitudinally, i.e. along the long oligomer axis and perpendicular to the planes,

8.5 Photoconductive and photovoltaic devices

441

the carrier mobility within the planes is sufficiently high to allow effective charge separation before the geminate charges recombine. Conversely, charge-transfer excitons generated parallel to the planes undergo rapid recombination of geminate electron-hole pairs and therefore do not contribute to the photoconductivity. Therefore, on a nanometre scale, the intrinsic layered structure of close-packed monolayers in sublimed films of a-6T has advantages for charge generation, in a way somewhat analogous to the use of several distinct microscopic physical layers to enhance interfacial charge separation in photocells. It is also worth noting that evolution has resulted in the use of layers of close-packed antennae as photoreceptors in nature.

8.5.3 Photovoltaic applications (solar cells) A related application is in the use of organic semiconductors as solar cells. Once again, the advantages of organic semiconductors are the possibility of depositing large area films from solution at low cost, with the possibility of selecting molecules of a variety of energy gaps and chemical modifications in order to make most efficient use of the solar spectrum. For a comprehensive, albeit not completely up-to-date, review of organic solar cells, the reader is referred to Chamberlain [200]. Important parameters used to assess the quality of the device are the short-circuit current, Jsc, open-circuit voltage, VOCand fill factor, FF, the power conversion efficiency, r]. Figure 23 shows a schematic I-V curve, similar to that published by Tang. The fill factor is a measure of the ability of the cell to do useful work. It is defined as the ratio of the maximum electrical power which could be obtained (the shaded rectangle within the I-V curve in Fig. 23) to the product of JsCVoc.The power conversion efficiency, r], is the ratio of the maximum electrical power (shaded area) to the power of the radiation which produced it.

I

"0,

Figure 23. Schematic photocurrent-voltage curve, indicating important parameters, such as open circuit voltage, Voc, short-circuit current, Jsc . The fill factor, FF, is the ratio of the shaded area divided by the product Voc JSC.Adapted from Tang [195].

442

8 Electro-opticalPolythiophene Devices

A major breakthrough in the use of organic semiconductors in solar cells was the report by Tang [195] and Panayotatos [201,202] of efficient two-layer photovoltaic cells, in which charge generation occurs primarily at the interface between the two organic layers rather than at the Schottky barrier or interface with the electrodes, a: is often the case for single-layerdevices. Using device of the form Glass/IT0/300 A sublimed copper phthalocyanine (CuPc)/500 A sublimed perylene tetracarboxylic derivative (PV)/Ag, power conversion efficiencies up to 1% were achieved, with the charge-generation efficiency relatively independent of bias. The open circuit voltage (Voc = 0.4 V) was found to be relatively insensitive to the choice of metal as the uppermost electrode (instead of Ag). The double-layer structure also drastically lowers the series resistance of the cells, resulting in fill factors up to 0.65, much higher than that which had been achieved with single-layer cells. By comparison with the absorption spectra, both layers were found to contribute to the spectral response. The operation of the device was attributed to the absorption of light in both layers, resulting in the creation of excitons which can diffuse to the interface between the two layers. At the interface, the excitons are efficiently split because the offset between the corresponding LUMO levels and HOMO levels across the interface (‘band offsets’) exceed the exciton binding energy, or for a less abrupt interface, because the local electric field in this region is expected to be particularly high. The holes are then transported through the p-type CuPc layer, to the I T 0 electrode, while electrons are transported through the n-type perylene derivative layer, to the Ag electrode. The stability of the device was also impressive, with both the open circuit voltage and short circuit current decreasing by less than 2% after five days of continuous operation in ambient atmosphere under illumination from a tungsten and under short-circuit conditions (2 mA cmP2), white light source (75 mW a?) although the fill factor decreased by about 30%. Very recently, Noma et al. [203] have used a similar bi-layer strategy with the oligomer a-octithiophene (~-8T)~used in place of copper phthalocyanine, in a device of the structure: Glass/IT0/400 A sublimed PV/300 A sublimed a-8T/Au. Under white light illumination at 105mW cmP2,a fill factor of 0.5 and a conversion efficiency of ca. 0.6% were achieved, comparable with the figures of merit reported by Tang. The open-circuit voltage (Voc = 0.42 V) and short-circuit current (Jsc = 2.9 mA cmP2) were also similar. The short-circuit photocurrent was linear with light intensity, indicating negligible recombination of photogenerated electrons and holes. The fill factors and conversion efficiencies also remained approximately independent of light intensity up to 200mWcm-2. Since octithiophene absorbs in the region 300-500 nm, while the PV perylene derivative absorbs in the 500-800 nm region, the spectral response of the oligomer-based cell responds to a wider spectral range than that of Tang. In an earlier study, Kuwabara et al. [204] reported much lower fill factors (0.29) and conversion efficiencies (0.02%) for a similar bi-layer cell using a-quinquethiophene (a-5T) The high conversion efficiency when a-8T is used rather than a-5T is attributed to higher dark conductivity, higher charge-carrier mobility (pFET), improved film quality and absorption at longer wavelengths, where there is a greater proportion of solar radiation to be harvested.

8.5 Photoconductive and photovoltaic devices

443

8.5.4 Photovoltaic devices based on polythiophenes In contrast with the oligothiophenes, the polythiophenes have been successfully used in photovoltaic devices. The first devices were presented already in 1984 by Glenis et al. [205]. These photovoltaic cells used electrochemically grown polythiophene and poly(3-methylthiophene) sandwiched between aluminium and gold or platinum electrodes. Since the electrochemical polymerisation produces a film in which the polymer is in its conducting (in this case oxidised) state, it then had to be electrochemically reduced to the p-type semiconducting state prior to evaporation of the top electrode. When illuminated with monochromatic light (470 nm) at low input power (P < 1 pm W cmP2),a power efficiency of 4% was attained. The same group later showed that doping (to a doping level of about 3-5%) the film resulted in a decreased quantum yield but an increased conversion yield [206]. This is explained by the increase in free carrier density and widened absorption spectrum that results from doping. Most of the work with polythiophenes in photovoltaic devices has been pursued by Yoshino’s group at Osaka University [207-2101. Different techniques have been used, such as single layer devices utilising poly(3-alkylthiophenes) mixed with CN-PPV [207] or C60[210], heterojunction devices [208], and multilayer structures [209]. The heterojunction devices poly(3-hexylthiophene) and poly(p-pyridyl vinylene) were used, which is possible since they can be dissolved in incompatible solvents, such as chloroform and formic acid, respectively. Such devices showed strong photoluminescence quenching and increased short-circuit current, opencircuit voltage and fill factor compared to single layer poly(3-hexylthiophene) structures. Since poly(p-pyridyl vinylene) is an n-type conducting polymer and the polythiophene ap-type, the photoinduced carriers that are generated at the interface can be efficiently collected. Similar behavior is found in the multilayer devices, where the structure is built up from the negatively charged poly(3-a-carboxylmethyl thiophene), PCMT, and the positively charged poly(dihexyldipropargy1 ammonium bromide), PDHDPAB. Starting with a positively charged substrate (using an amino silane covered I T 0 surface), a multilayer structure is built up by repeatedly dipping it in aqueous PCMT solution and PDHDPAB in methanol. This procedure allows for very accurate thickness determination of the active layer in the device, as each PCMT layer is about 1 nm thick. Similar to the heterojunction device described above, the multi-layer device also exhibits PL quenching and increased quantum efficiency. A unique feature of this type of device is the small thickness of the excited PCMT layer, which means that electron transfer can occur not only to the nearest neighbouring PDHPAB layer, but also further on, which might increase the probability of exciton dissociation. Combinations of polythiophenes and C60 have also been utilised by the Santa Barbara group for more fundamental studies [211, 2121, and by the Linkoping group for devices [213]. These devices actually comprise two different polythiophenes, poly(3-(4-octylphenyl)-2,2’-bithiophene), polymer I1 in Fig. 13, and poly(3-f2’-methoxy-5’-octylphenyl)thiophene),POMeOPT. The latter is used because it has a sidegroup reminiscent of anisole, which is known to interact with

444

8 Electro-optical Polythiophene Devices

c60, manifesting in a solvatochromic shift. POMeOPT is therefore believed to increase the interaction between the polymeric phases and c60 in the blends. An Al/POMeOPT PTOPT C60 (1 : 1:2)/ITO device shows a conversion efficiency of 15% with zero bias (A = 500 nm, 1.5mW cm2) and 60% when a reverse bias of 2 V is applied.

+

+

8.6 Electro-optical modulator devices As discussed in a previous chapter, charging of oligomers is accompanied by a distortion of the molecular geometry, giving rise to new energy levels within the former HOMO-LUMO gap, as shown schematically in Fig. 24 for the formation of the radical cation (polaron) and dication (intramolecular bipolaron) of an isolated a-6T oligomer. Theoretically, although simple one-electron Huckel models can often qualitatively predict the position of new molecular levels within the gap associated with charged excitations, electron-electron interactions are also important and have been successfully included in Valence Effective Hamiltonian methods of BrCdas et al. Furthermore, in the solid-state, the effects of aggregation and a generally more coplanar structure result in further modifications to the energy levels, compared to those of isolated oligomers. Returning to the schematic diagram of Fig. 24, we now discuss the new optical transitions expected for the charged oligomers. When an electron is removed from a neutral oligomer, the distortion of molecular geometry (lattice relaxation) results in a raising of the former HOMO level (now denoted SOMO = singly occupied molecular orbital) and lowering of the LUMO level. The level below the former HOMO level (HOMO-1) forms the HOMO of the radical cation and dication. Two new optical transitions are allowed for the radical cation (polaron), from the new HOMO to SOMO and SOMO to LUMO. LUMOtl LUMO

LUMO+I

LUMO+P

LUMO+l

+HOMO

&HOMO-1 Neutral 6T

HOMO

HOMO

Radical Monocation 6T"

Dication 6T 2+

Figure 24. Simplified energy diagram for the charged excitatitations of sexithiophene, showing the new levels within the energy gap and new optical transitions associated with the radical cation (monopolaron) [ST'+] and dication (intramolecular bipolaron) [6T2+].

8.6 Electro-optical modulator devices

445

Upon further oxidation to form the dication, the unpaired electron is removed from the SOMO and the former SOMO and LUMO move further towards the centre of the gap as the geometry distortion becomes even stronger. Now, one new optical transition is expected for the dication, from the HOMO to new LUMO (former SOMO), at a higher energy than the lower energy transition of the radical cation. In the solid state and aggregates, there is evidence from electron spin resonance [214-2191 and optical spectroscopy [214,220] that the radical cations can aggregate and that the energy levels are subject to Davydov splitting. Aggregated cations are often referred to as charged 7r-dimers or 7r-stacks. Simultaneous measurement of the electron spin resonance signal during cyclic voltammetry of an oligomer film in an electrochemical cell show that the number of spins (radicals) is observed to decrease long before the dication (second oxidation peak) is reached, indicating possible pairing of spins in aggregates. This effect is particularly marked at low temperatures. This may be due to increased screening by solvent molecules, as the dielectric constant rises, and perhaps also due to lower torsion and hence closer intermolecular packing at low temperatures. Hence, charging of oligomers gives rise to a rich variety of new optical transitions below the fundamental absorption, indicative of the charged species concerned. The equilibria between the different charged species is a function of both the charge density and the structural order within the material. Probing changes in the optical absorption or reflection of an MIS device is very useful as a spectroscopic tool to determine the charged excitations in the solid state. The technique has already been used for investigation of conjugated polymers [221,222] and demonstrated as the basis of an electro-optic modulator device [223]. When light is passed perpendicularly through an MIS diode, the modulation of the transmission is very low (typically 0.001%-0.01%). This is quite adequate for spectroscopic research, since lock-in amplification techniques can be used to detect such small changes. However, for a viable modulator device, a much larger modulation is required. This requires that the light to be modulated is coupled much more effectively to the region where the charge density is being modulated. This has been achieved for a conjugated polymer [223] by employing a waveguide geometry within a field-effect transistor structure. The light is effectively passed along the source-drain channel of the FET and hence much longer modulation path can be used, typically 2 mm, rather than the 1000A thickness of the oligomer layer. Consequently, modulation up to 20% can be achieved, although modulation frequency is limited by the mobility of the conjugated semiconductor. Recently, a very similar device has been reported [224], using an oligothiophene layer, presumably with the hope of increasing the modulation frequency, although to date the operational range reported for the oligomeric device is also limited to frequencies lower than a few kHz. Returning now to the MIS diode and optical probing of the charged states in oligomers, Fig. 25 shows a schematic view of the device structure and experimental configuration used in the following case study of sexithiophene, a-6T. The MIS diode is held under vacuum (-2 x torr) in an optical access cryostat, aligned so that a monochromatic probe beam passes through the device and is detected by an appropriate optical detector.

446

8 Electro-optical Polythiophene Devices Monochromatic light

Semi-transparent Gold Gate electrodes I 5

AC Modulation voltaae amlied to

Figure 25. Schematic diagram of the experimental configuration for optical probing of field-induced charge within semi-transparent MIS diodes based on oligomers.

By superimposing a small oscillatory (AC) voltage on the DC bias, it is possible to modulate the charge density within the device, either to probe the accumulation region, for negative DC bias or the edge of the depletion layer (i.e. the bulk of the oligomer layer) for positive DC bias. The optical transitions of the charged species, are detected as a modulation of the probe beam at the same frequency as the AC modulation voltage, which is detected using a lock-in amplifier. It should be noted that this method provides a very direct means of probing the charged excitations within an opto-electronic device environment, in which charge is induced by purely electrostatic means, rather than by charge-transfer doping or photo-excitation.

8.6.1 Optical probing of field-induced charge in a-sexithiophene Because oligomers can be synthesised and purified as monodisperse molecules and prepared as thin films in which the degree of order can be systematically controlled, we can expect better resolution of the optical properties. The technique has therefore been used to study charged excitations in oligothiophenes [225-2271, with the aim of investigatingwhich charged excitations were involved in charge-transport within the high mobility organic field-effect transistors and also trying to extrapolate towards polythiophenes by studying charged excitations in longer oligothiophenes [228]. Electro-modulated absorption and fluorescence measurements have also been reported [229, 2301, yielding complementary information about quenching of the fluorescence by field induced charges. Semi-transparent metal-insulator-semiconductor (MIS) diodes were fabricated, in which a thin film of a-6T of thickness around 100nm forms the semiconductor

8.6 Electro-optical modulator devices

447

layer. The MIS diode can indeed be viewed as an inverted thin-film transistor, with the source and drain electrodes shorted together, or rather replaced by a continuous IT0 electrode, while a semi-transparent gold film serves as the gate electrode, separated from the a-6T layer by a sublimed film of silicon monoxide, also around 1000 thick. Figure 26 shows the optical spectrum of modulated charge, taken at 298K, with a) 0 V DC bias (onset of depletion) and b) -8 V DC bias (accumulation). The spectrum at 0 V shows two new optical absorptions, at 0.83 eV and 1.70 eV, similar to the

A

a) Gate Voltage: OVflV (ac) Onset of depletion

p 1 ; Energy (eV) b) Gate Voltage: -4VflV (ac)

Accumulation

0.5 .

1.34eV ....... .... .

. .,

...

. . . . . . .........

.... I

0.5

1 .o

1.5

2.0

2.5

Energy (eV) Figure 26. Optical spectra of field-induced charge in sexithiophene [cv-~T]: (a) a t the onset of depletion (OV gate voltage). (b) in accumulation (-4V gate voltage).

448

8 Electro-optical Polythiophene Devices

signature of radical cations produced by charge-transfer doping in dilute solution [231-2331 or photo-excitation [234,235]. In the spectrum at -8 V bias, an additional spectral feature emerges at 1.34eV. This continues to increase in strong accumulation. This transition is attributed to a spinless dication. Figure 27 shows spectra taken with OV DC bias at a) 298K and b) 170K. The spectrum at room temperature shows the optical absorptions, at 0.70-0.83 eV and 1.70eV, while the low temperature spectrum shows additional transitions at 1.01eV and 2.05 eV. Similar transitions have also been observed in aggregates

1-./.]! ...

.~~XIR

b)

Gate voltage: OVSV (ac) Temperature: 170K

0.2

t

5

*0 Y

0.0 -0.2-

2.05eV

”--\. 1.70eV

-0.4

-

-0.6

-

-0.0 -1 .o

Energy (eV) Figure 27. Optical spectra of field-inducedcharge in sexithiophene [a-6T]: (a) at room temperature (300K). (b) at low temperature (170K).

8.7 All-optical modulator and memory devices

449

6 -

t 5 B T-

-4 -6

1.2

1.4

1.6

1.8

2.0

2.2

2.4

Energy (eV) Figure 28. Bias-dependence of the optical spectra of field-induced charge in sexithiophene [(Y-~T], showing evidence for three species of charged excitations.

of the oligomers in poor solvents, at low temperatures or high concentrations [214, 2201. A particularly narrow feature at 0.60eV is attributed to the additional transitions to aggregated states of radical cations (also known as 7r-stacks or 7r-dimers), corresponding to the charge-transfer band. These transitions were not observed in the amorphous films of spin-coated regiorandom side-chain substituted 6T [236]. From the bias dependence spectra of Fig. 28, it can be seen that typically three charged excitations may be simultaneously present within an oligomer film. Dications are formed at high charge density, in the accumulation layer, while the feature at 2.05 eV is strongest in depletion, indicating that 7r-stacking plays an important r61e in intermolecular charge transport in the relatively well-ordered bulk of the film, though not at the more disordered interface probed in accumulation. Radical cations are usually present under all bias conditions, even in films which have not been intentionally doped.

8.7 All-optical modulator and memory devices A promising optical application of oligomers is their potential use in opticallyaddressed spatial light modulators (OASLM). OASLMs could be used for real-time parallel processing of images and switching of light beams for telecommunications. Unlike most of the optical devices which we have discussed so far, all-optical SLMs do not need pixelated electrodes and therefore do not have the problems of interconnections and addressing of pixels, which are associated with flat-panel displays. Therefore, SLMs can use very small pixel dimensions, limited by optical

450

8 Electro-optical Polythiophene Devices

resolution to dimensions of around 1 pm and can therefore handle several channels of data in parallel, including graphic images. SLMs are required as input stages of optical computers and for image-processing (e.g. spatial filtering to eliminate periodic patterns from raster scanning or to enhance contrast) because the subsequent Fourier optics processing requires the use of coherent monochromatic light. An incoherent to coherent OASLM operates in two steps, as shown schematically in Fig. 29. Firstly, an image of an object is formed within a photosensitive film by exposure to light, which could even be incoherent and of a rather broad spectral range (e.g. white light). The image is usually stored as a change in absorption or refractive index within the film. Secondly, the stored image thus formed serves as a template, so that when the OASLM is later probed by a coherent probe beam (e.g. laser), a coherent image can be generated. Previous SLMs have used photoconductors coupled to liquid crystals [237,238], multiple-quantum-well semiconductors [239] and silver halide films [240] to exploit such effects as photorefraction [241], photochromism [242, 2431 and saturable absorption, etc. The non-linear optical properties of oligothiophenes make them promising candidates for active photochromic materials in optically-addressed spatial light modulator devices. In particular, speed and spatial resolution can be improved by two orders of magnitude, compared [244] with devices based on liquid crystals or semiconductor heterostructures. Fichou et al. [ 181 have reported such a device, based on sublimed films of sexithiophene (a-6T). The optical switching mechanism is due to absorption by a long-lived (T = 5 ns) triplet state formed only 10 picoseconds [245] after photo-excitation in the Incoherent

Laser

___--_-_---SiO, a-6T \,

Coherent image ,/"\TOai plane (coroon), I'

Figure 29. Diagram showing the operation of an all-optical spatial light modulator acting as an incoherent-to-coherent optical converter. Adapted from Fichou et al.

8.7 All-optical modulator and memory devices

45 1

absorption band of sexithiophene, as a result of rapid inter-system crossing from the lowest singlet excited state. In the areas initially exposed to the incoherent white light ‘write’ pattern, triplet excitons are thus formed. The probe beam can then detect an induced TI -+ T, absorption, which has a remarkably narrow linewidth (FWHM = 0.1 eV) compared to the linear absorption band (FWHM about 1.0eV) and shows none of the vibronic replicas which are present in the linear absorption spectrum. The particularly narrow linewidth of the triplet-triplet absorption is attributed principally to three causes [244]: i) damping of the vibronic side-bands, ii) conservation of oscillator strength, iii) homogeneity of the molecules. The need for homogeneity of the molecules is satisfied particularly well by conjugated oligomers of uniform length and is obviously not satisfied by the corresponding conjugated polymers, for which there is always a distribution of conjugation lengths. Similar photo-induced transient spectra are observed in dilute solutions of the oligomers, indicating that the fast optical transition is intramolecular. OASLMs based on oligo(thi0phenes) are therefore not limited either by slow collective effects, as for liquid crystals or polarisation effects in photorefractive materials and multiple quantum wells, nor by RC time constants, as for hybrid photoconductor/liquidcrystal devices. Fichou has suggested [18] that by dispersing a-6T with an outgassed polymer matrix, the triplet lifetime could be increased from around 5 ns in the evaporated film to the microsecond regime. Although this would lower the operating speed from around 200 MHz to the 1 MHz regime, the ‘written’ image would persist for longer, therefore reducing the energy required for exposing (‘writing’) the image to a reasonable value of 1 kW cm-2. It is estimated [246] that the isolated oligomers can withstand around lo6 excitation-relaxation cycles. Since the all-optical OASLM does not require any electrical connections, it would be possible to prepare uniform films of a-6T on continuous tape, so that after several exposures, the tape can be wound on, much like a photographic film, when the film degrades and needs to be replaced. Recently, oligothiophene derivatives have been synthesised by Irie and coworkers and also by Lehn and co-workers, which exhibit photochromism even in single-crystal phases [247] with switching speeds as fast at 10ps [248] (or 100 fs in solution [249]), with remarkably high stability - up to 70 000 cycles [250]. A typical structural formula of these materials is shown in Fig. 30. The materials are often based on a cyclopentene ring, to which are attached aromatic substituents. Thiophene oligomers are often used because they result in thermally stable photoisomers due to the low aromatic stabilisation energy of the thiophene ring, compared with, for example, a phenylene ring [250]. Because the short oligothiophene chromophores absorb high energy light, in the green to ultraviolet results in a photoisomerism of the perfluorocyclopentene, from an open to a closed form, with the effect of changing its bond alternation, so that the oligothiophene substituents at either side are now linked by a continuous conjugation pathway in the close form, whereas

452

8 Electro-optical Polythiophene Devices

open form (broken conjugation pathway)

closed form (continuous conjugation pathway)

Figure 30. Open and closed photoisomers of the photochromic materials based on oligothiophene derivatives. Adapted from Lehn et al. [251], Irie et al. [250].

the conjugation pathway is broken in the open form. The reverse isomerisation (ring closure) is usually induced by irradiation with light of lower energy (typically red to yellow), since the closed form resembles a longer oligo(thiophene), which absorbs light at lower energy. For certain derivatives, there are also profound changes in the fluorescencequantum yield between the open and closed forms [251]. The oligothiophene chain length can be varied to alter the spectral window of the absorption or fluorescence changes. Such oligo(thi0phene) derivatives clearly have potential applications in all-optical switching and memory devices.

References A. Pochettino, Acud. Lincei Rediconti, 1906, 15, 355. M. Volmer, Ann. Physik, 1913, 40, 775. D. D. Eley, Nature, 1948, 162, 819. A. T. Vartanyan, Zh. Fiz. Khim., 1948, 22, 769. P. M. Borsenberger and D. S. Weiss, Organic Receptors for Imaging Systems. B. J. Thompson, Eds., Optical Engineering Marcel Dekker, Inc., New York, 1993, vol. 39. 6. A. S. Davydov, Zhur. Eksptl. i Teoret. Fiz., 1948, 18, 515. 7. A. E. Gillam and D. H. Hey, J. Chem. SOC.,1939, 1170. 8. M. Pope, H. P. Kallmann and P. Magnante, J . Chem. Phys., 1963,38,2042-2043. 9. W. Helfrich and W. G. Schneider, Phys. Rev. Lett., 1965, 14, 229-231. 10. P. Frederiksen, T. Bjerrnholm, H. G. Madsen and K. Bechgaard, J. Muter. Chem., 1994,4, 675-678. 11. J. Griiner, H. F. Wittmann, P. J. Hamer et al., Synth. Met., 1994, 67, 181-185. 12. W. Tachelet, S. Jacobs, H. Ndayikengurukiye, H. J. Geise and J. Griiner, Appl. Phys. Lett., 1994,64,2364-2366. 13. G. Grem, V. Martin, F. Meghdadi, C. Paar, J. Stampfl, J. Sturm, S. Tasch and G. Leising, Synth. Met., 1995, 71, 2193-2194. 14. J. Kido, H. Shionoya and K. Nagai, Appl. Phys. Lett., 1995,67,2281-2283. 15. H. S. Woo, J. G. Lee, H. K. Min et al., Synth. Met., 1995, 71, 2173-2174. 16. X. C. Li, F. Cacialli, M. Giles et al., Adv. Muter., 1995, 7, 898. 17. P. Hesemann, H. Vestweber, J. Pommerehne, R. F. Mahrt and A. Greiner, Adv. Muter., 1995, 7, 388-390. 18. D. Fichou, J. M. Nunzi, F. Charra and N. Pfeffer, A h . Muter., 1994, 6, 64-67. 1. 2. 3. 4. 5.

References

453

19. F. Garnier, G. Horowitz, X. H. Peng and D. Fichou, Adv. Muter., 1990,2, 592-594. 20. G. Gustafsson, Y. Cao, C. M. Treacy, F. Klavetter, N. Colaneri and A. J. Heeger, Nature, 1992,357,477-479. 21. F. Garnier, R. Hajlaoui, A. Yassar and P. Srivastava, Science, 1994, 265, 1684-1686. 22. M. Granstr-m and 0. Inganas, Adv. Muter., 1995, 7, 1012-1014. 23. A. M. Donald and A. H. Windle, Liquid Crystalline Polymers, Cambridge Solid State Science Series Cambridge University Press, Cambridge, 1992. 24. M. Sato, S. Tanaka and K. Kaeriyama, J . Chem. SOC.Chem. Commun., 1986, 873. 25. M. A. Sato, S. Tanaka and K. Kaeriyama, Synth. Met., 1987, 18,229-232. 26. R. Su,gimoto,S. Takeda, H. B. Su and K. Yoshino, Chem. Express, 1986, I, 635. 27. J.-E.Osterholm, J. Laakso, P. Nyholm et al., Synth. Met., 1989, 28, C435. 28. M. R. Bryce, A. Chissel, P. Kathirgamanathan, D. Parker and N. R. M. Smith, J . Chem. SOC.- Chem. Comm., 1987,466-467. 29. D. Braun and A. J. Heeger, Appl. Phys. Lett., 1991,58, 1982. 30. A. R. Brown, D. D. C. Bradley, J. H. Burroughes et al., Appl. Phys. Lett., 1992,61,2793. 31. P. L. Burn, A. B. Holmes, A. Kraft, D. D. C. Bradley, A. R. Brown and R. H. Friend, J. Chem. SOC.Chem. Commun., 1992,32. 32. N. C. Greenham, S. C. Moratti, D. D. C. Bradley, R. H. Friend and A. B. Holmes, Nature, 1993, 365, 628. 33. A. Kraft, P. L. Burn, A. B. Holmes et al., Synth. Met., 1993, 55, 936. 34. K. Y. Jen, R. Oboodi and R. L. Elsenbaumer, Polym. Muter. Sci. Eng., 1985,53, 79. 35. R. L. Elsenbaumer, K. Y. Jen and R. Oboodi, Synth. Met., 1986, 15, 169. 36. E. E. Havinga, I. Rotte, E. W. Meijer, W. Tenhoeve and H. Wynberg, Synth. Mer., 1991, 41, 473-478. 37. D. Delabouglise, M. Hymene, G. Horowitz, A. Yassar and F. Garnier, Adv. Muter., 1992, 4, 107. 38. P. Bauerle, Adv. Muter., 1992, 4, 102-106. 39. F. Geiger, M. Stoldt, H. Schweizer, P. Bauerle and E. Umbach, Adv. Muter.. 1993, 5, 922-925. 40. F. Garnier, A. Yassar, R. Hajlaoui et al., J. Amer. Chem. SOC.,1993, 115, 8716-8721. 41. J. K. Herrema, J. Wildeman, F. Vanbolhuis and G. Hadziioannou, Synth. Met., 1993, 60, 239-248. 42. H. Mao, B. Xu and S. Holdcroft, Macromolecules, 1993, 26, 1163. 43. M. Leclerc, F. M. Diaz and G. Wegner, Makromol. Chem., 1989, 190,3105. 44. M. J. Winokur, D. Spiegel, Y. Kim, S. Hotta and A. J. Heeger, Synth. Met., 1989, 28, C419. 45. A. Stabel and J. P. Rabe, Synth. Met., 1994, 67, 47-53. 46. R. D. McCullough and R. D. Lowe, J. Chern. SOC.,Chem. Commun., 1992, 70. 47. R. D. McCullough, R. D. Lowe, M. Jayaraman and D. L. Anderson, J . Urg. Chem., 1993, 58, 904. 48. P. Bauerle, F. Pfau, H. Schlupp, F. Wiirthner, K.-U. Gaudl, M. B. Car0 and P. Fischer, J. Chem. SOC.Perkin Trans. II, 1993, 489-493. 49. I. B. Berlman, J . Phys. Chem., 1970, 74, 3085-3093. 50. P. L. Burn, D. D. C. Bradley, R. H. Friend, D. A. Halliday, A. B. Holmes, R. W. Jackson and A. Kraft, J . Chem. SOC.Perkin Trans. I, 1992,3225-3231. 51. T. Horn, S. Wegener and K. Mullen, Macromolecular Chem. & Phys., 1995,196,2463-2474. 52. A. R. Brown, A. Pomp, C. M. Hart and D. M. Deleeuw, Science, 1995, 270, 972-974. 53. D. Beljonne, J. Cornil, R. H. Friend, R. A. J. Janssen and J. L. Bredas, J. Amer. Chem. SOC., 1996, 118,6453-6461. 54. J. C. Scaiano, R. W. Redmond, B. Mehta and J. T. Arnason, Photochem. Photobiol., 1990, 52, 655. 55. R. A. J. Janssen, D. Moses and N. S. Sariciftci, J. Chem. Phys., 1994, 101, 9519-9527. 56. R. A. J. Janssen, M. P. T. Christiaans, K. Pakbaz, D. Moses, J. C. Hummelen and N. S. Sariciftci, J. Chem. Phys., 1995, 102, 2628-2635. 57. B. Xu and S. Holdcroft, J. Am. Chem. Soc., 1993, 115, 8447. 58. N. Periasamy, R. Danieli, G. Ruani, R. Zamboni and C. Taliani, Phys. Rev. Lett., 1992, 68, 919-922.

454

8 Electro-optical Polythiophene Devices

59. R. Rossi, M. Ciofalo, A. Carpita and G. Ponterini, J . Photochem. Photobiol. A: Chem., 1993, 70, 59. 60. H. Chosrovian, S. Rentsch, D. Grebner, D. U. Dahm, E. Birckner and H. Naarmann, Synth. Met., 1993, 60, 23-26. 61. R. S . Becker, J. S. de Melo, A. L. Maqanita and F. Elisei, Pure & Appl. Chem., 1995, 67, 9-16. 62. D. Birnbaum, B. E. Kohler and C. W. Spangler, J. Chem. Phys., 1991,94, 1684-1691. 63. K. Schulten, I. Ohmine and M. Karplus, J . Chem. Phys., 1976, 64,4422. 64. D. Birnbaum and B. E. Kohler, J . Chem. Phys., 1992,96,2492-2495. 65. S . Mazumdar, D. Guo and S. N. Dixit, J. Chem. Phys., 1992,96,6862-6867. 66. J. B. Vanbeek, F. Kajzar and A. C. Albrecht, J. Chem. Phys., 1991,95, 6400-6412. 67. J. B. Vanbeek and A. C. Albrecht, Chem. Phys. Lett., 1991, 187,269-276. 68. G. Klein, R. Voltz and M. Schott, Chem. Phys. Lett., 1973, 19, 391. 69. R. Katoh and M. Kotani, Chem. Phys. Lett., 1992,196, 108-112. 70. R. H. Austin, G. L. Baker, S. Etemad and R. Thompson, J. Chem. Phys., 1989,90, 6642. 71. D. P. Craig and S . H. Walmsley, Excitons in Molecular Crystals, Benjamin, New York, 1968. 72. M. Pope and C. E. Swenberg, Electronic Processes in Organic Crystals, Oxford University Press, New York, 1982. 73. J. Riihe, N. F. Colaneri, D. D. C. Bradley, R. H. Friend and G. Wegner, J . Phys. Cond. Matter., 1990, 2, 5495-5477. 74. W. Porzio, S. Destri, M. Mascherpa, S. Rossini and S . Bruckner,Synth. Met., 1993,55,408-413. 75. S . Hotta and K. Waragai, Adv. Muter., 1993, 5 , 896-908. 76. 0. Dippel, V. Brandl, H. Bassler, R. Danieli, R. Zamboni and C. Taliani, Chem. Phys. Lett., 1993,216,418-423. 77. N. I. Nijegorodov and W. S . Downey, J . Phys. Chem., 1994, 98, 5639-5643. 78. R. A. J. Janssen, L. Smilowitz, N. S. Sariciftci and D. Moses, J . Chem. Phys., 1994, 101, 1787-1798. 79. D. A. Dossantos, D. S. Galvao, B. Laks and M. C . Dossantos, Synth. Met., 1992,51,203-209. 80. Q . Pei, 0. Inganas, G. Gustafsson et al., Synth. Met., 1993, 55, 1221-1226. 81. M. Granstrom and 0. Inganas, Synth. Met., 1992,48,21-31. 82. K. Tashiro, Y. Minagawa, M. Kobayashi, S. Morita, T. Kawai and K. Yoshino, Synth. Met., 1993,55, 321-328. 83. K. Iwasaki, H. Fujimoto and S . Matsuzaki, Synth. Met., 1994, 63, 101-108. 84. B. Xu and S . Holdcroft, Macromolecules, 1993, 26, 4457-4460. 85. M. Belletete, L. Mazerolle, N. Desrosiers, M. Leclerc and G. Durocher, Macromolecules, 1995, 1995, 858778597, 86. J. C. de Mello, H. F. Wittmann and R. H. Friend, Adv. Muter., 1997, 9, 230. 87. N. C. Greenham, I. D. W. Samuel, G. R. Hayes et al., Chem. Phys. Lett., 1995,241, 89-96. 88. M. R. Anderson, M. Berggren, T. Olinga, T. Hjertberg, 0. Inganas and 0. Wennerstrom, Synth. Met., 1997, 85, 1383-1384. 89. N. C. Greenham and R. H. Friend, in Solid State Physics, H. Ehrenreich, F. A. Spaepen, Eds. Academic Press, 1995, vol. 49, pp. 2-150. 90. A. R. Brown, K. Pichler, N. C. Greenham, D. D. C. Bradley, R. H. Friend and A. B. Holmes, Chem. Phys. Lett., 1993, 210, 61-66. 91. C. W. Tang and S. A. Van Slyke, Appl. Phys. Lett., 1987,51,913-915. 92. J. C. Carter, I. Grizzi, S . K. Heeks et al., Appl. Phys. Lett., 1997, 71, 34-36. 93. N. C. Greenham, R. H. Friend and D. D. C. Bradley, Adv. Muter., 1994,6,491-494. 94. H. Becker, S. E. Burns and R. H. Friend, Phys. Rev. B, 1997,56, 1893-1905. 95. A. Bernanose, J. de Chim. Phys., 1955,52, 396-400. 96. A. Bernanose and P. Vouaux, J . de Chim. Phys., 1955,52, 509-510. 97. J. Kalinowski, Muter. Sci., 1981, 7 , 44. 98. P. S. Vincett, W. A. Barlow, R. A. Hann and G. G. Roberts, Thin Solid Films, 1982, 94, 171 -1 83. 99. C. W. Tang, S. A. Van Slyke and C. H. Chen, J. Appl. Phys., 1989,65,3610-3616. 100. C. Adachi, S. Tokito, T. Tsutsui and S. Saito, Jap. J. Appl. Phys. Pt. 2 - Letters, 1988, 27, L713-L715.

References

455

101. C. Adachi, S. Tokito, T. Tsutsui and S. Saito, Jap. J. Appl. Phys. Pt. 2 - Letters, 1988, 27,

L269-L271. 102. C. Adachi, T. Tsutsui and S. Saito, Appl. Phys. Lett., 1989, 55, 1489-1491. 103. C. Adachi, T. Tsutsui and S. Saito, Appl. Phys. Lett., 1990,57, 531-533. 104. J. Shi and C. W. Tang, Appl. Phys. Lett., 1997, 70, 1665-1667. 105. H. Kurczewski and H. Bassler, J. Lumin., 1977, 15, 261. 106. K. C. Kao and W. Hwang, Electrical transport in solids: with particular reference to organic semiconductors, International series in the science of the solid state, Pergamon, Oxford, 1981, vol. 14. 107. R. H. Partridge, Polymer, 1983, 24, 733-738. 108. R. H. Partridge, Polymer, 1983, 24, 739-747. 109. R. H. Partridge, Polymer, 1983,24, 748-754. 110. R. H. Partridge, Polymer, 1983, 24, 755-762. 111. J. H. Burroughes, D. D. C. Bradley, A. R. Brown et al., Nature, 1990,347, 539. 112. G. Grem, G. Leditzky, B. Ullrich and G. Leising, Adv. Mater., 1992, 4 , 36-37. 113. Y. Ohmori, M. Uchida, K. Muro and K. Yoshino, Jap. J. Appl. Phys. Part 2 Lett., 1991,30, L1938-Ll940. 114. B. H. Cumpston, I. D. Parker and K. F. Jensen, J. Appl. Phys., 1997,81, 3716-3720. 115. S. Fujita, T. Sakamoto, K. Ueda, K. Ohta and S. Fujita, Jap. J. Appl. Phys. Part 1 - Regular Papers Short Notes & Rev. Papers, 1997, 36, 350-353. 116. S. Honda, K. Chihara, M. Watamori and K. Oura, Appl. Surf. Sci., 1997, 114, 408-411. 117. F. Li, H. Tang, J. Shinar, 0. Resto and S. Z. Weisz, Appl. Phys. Lett., 1997, 70, 2741-2743. 118. A. R. Schlattman, D. W. Floet, A. Hilberer et al., Appl. Phys. Lett., 1996, 69, 1764. 119. C. C. Wu, C. I. Wu, J. C. Sturm and A. Kahn, Appl. Phys. Lett., 1997, 70, 1348-1350. 120. S. Hayashi, H. Etoh and S. Saito, Jap. Journ. Appl. Phys., 1986, 25, 773-775. 121. Y. Yang and A. J. Heeger, Appl. Phys. Lett., 1994,64, pp. 1245-1247. 122. Y. Yang, E. Westenveele,C. Zhang, P. Smith and A. J. Heeger,J . Appl. Phys., 1995,77,694-698. 123. J. C. Scott, S. A. Carter, S. Karg and M. Angelopoulos, Synth. Met., 1997, 85, 1197-1200. 124. S. A. Carter, M. Angelopoulos, S. Karg, P. J. Brock and J. C. Scott, Appl. Phys. Lett., 1997, 70,2067-2069. 125. Y. Cao, G. Yu, C. Zhang, R. Menon and A. J. Heeger, Synth. Met., 1997,87, 171-174. 126. P. W. M. Blom, M. J. M. Dejong and J. J . M. Vleggaar, Appl. Phys. Lett., 1996,68,3308-3310. 127. K. Uchiyama, H. Akimichi, S. Hotta, H. Noge and H. Sakaki, Synth. Met., 1994, 63, 57-59. 128. K. Uchiyama, H. Akimichi, S. Hotta, H. Noge and H. Sakaki, Mat. Res. SOC.Symp. Proc., 1994,328, 389. 129. M. Muccini, R. F. Mahrt, U. Lemmer et al., Chem. Phys. Lett., 1995, 242, 207-211. 130. G. Horowitz, P. Delannoy, H. Bouchriha et al., Adv. Muter., 1994, 6, 752-755. 131. A. Yassar, F. Garnier, F. Deloffre, G. Horowitz and L. Ricard, Adv. Muter., 1994,6,660-663. 132. C . Hosokawa, H. Higashi and T. Kusumoto, Appl. Phys. Lett., 1993, 62, 3238-3240. 133. T. Noda, H. Ogawa, N. Noma and Y. Shirota, Adv. Mater., 1997, 9, 720-722. 134. T. Noda, H. Ogawa, N. Noma and Y. Shirota, Appl. Phys. Lett., 1997, 70, 699-701. 135. M. Era, T. Tsutsui and S. Saito, Appl. Phys. Lett., 1995,67, 2436-2438. 136. R. N. Marks, F. Biscarini, R. Zamboni and C. Taliani, Europhys. Lett., 1995,32, 523-528. 137. D. Braun, G. Gustafsson, D. McBranch and A. J. Heeger, J . Appl. Phys., 1992, 72, 564-568. 138. P. Barta, W. R. Salaneck, M. Zagorska, A. Pron and S. Niziol, Adv. Mater. For Optic. & Electronics, 1996, 6, 406-408. 139. P. Barta, J. Birgersson, S. Guo, H. Arwin, W. R. Salaneck and M. Zagorska, Adv. Muter., 1997, 9, 135. 140. M. Berggren, G. Gustafsson, 0. Inganas, M. R. Andersson, 0. Wennerstrom and T. Hjertberg, Appl. Phys. Lett., 1994, 65, 1489-1491. 141. M. Berggren, G. Gustafsson, 0. Inganas, M. R. Andersson, T. Hjertberg and 0. Wennerstrom, J. Appl. Phys., 1994, 76, 7530-7534. 142. F. Garten, J. Vrijmoeth, A. R. Schlatmann, R. E. Gill, T. M. Klapwijk and G. Hadziioannou, Synth. Met., 1996, 76, 85-89. 143. T. Granlund, L. A. A. Pettersson, M. R. Anderson and 0. Inganas, J. Appl. Phys., 1997,81, 8097-8104. -

456

8 Electro-optical Polythiophene Devices

144. T. Ostergard, J. Paloheimo, A. J. Pal and H. Stubb, Synth. Met., 1997, 88, 171-177. 145. J. Birgerson, K. Kaeriyama, P. Barta, P. Broms, M. Fahlman, T. Granlund and W. R. Salaneck, Adv. Muter., 1996, 8, 982-985. 146. H. Saadeh, T. Goodson and L. P. Yu, Macromolecules, 1997,30,4608-4612. 147. N. C. Greenham, A. R. Brown, D. D. C. Bradley and R. H. Friend, Synthetic Metals, 1993, 55-57,4134-4138. 148. G. Yu, H. Nishino, A. J. Heeger, T. A. Chen and R. D. Rieke, Synthetic Metals, 1995, 72, 249-252. 149. 0. Inganas, W. R. Salaneck, J.-E. Osterholm and,J. Laakso, Synth. Met., 1988, 22, 395. 150. 0.Inganas, G. Gustafsson, W. R. Salaneck,J.-E. Osterholm and J. Laakso, Synth. Met., 1989, 28, 377. 151. G. Zerbi, L. Castellani, B. Chierichetti, G. Gallazi and 0. Inganas, Chem. Phys. Lett., 1990, 172, 143. 152. G. Zerbi, B. Chierichetti and 0. Inganas, J . Chem. Phys., 1991,94,4637. 153. G. Zerbi, B. Chierichetti and 0. Inganas, J. Chem. Phys., 1991,94,4646. 154. M. R. Anderson, M. Berggren, 0. Inganas, et al., Macromolecules, 1995,28, 7525-7529. 155. R. E. Gill, G. G. Malliaras, J. Wildeman and G. Hadziioannou, Adv. Materials 1994, 6, 132-135. 156. M. R.,,Andersson, D. Selse, H. Jarvinen, T. Hjertberg, 0. Inganas, 0. Wennerstrom and J.-E. Osterholm, Macromolecules, 1994, 27, 6503. 157. T. A. Chen and R. D. Rieke, Synth. Met., 1993,60, 175. 158. M. Berggren, M. Granstrom, 0. Inganas and M. R. Anderson, Adv. Muter., 1995, 7,900-903. 159. K. Ebihara, S. Koshihara, T. Miyazawa and M. Kira, Jap. J. Appl. Phys. Part 2 - Lett., 1996, 35, L1278-L1280. 160. P. Dyreklev and 0. Inganas, J. Appl. Phys., 1994, 76, 7915. 161. R. H. Friend, D. D. C. Bradley and P. D. Townsend, J . Phys. D (Applied Physics), 1987,20, 1367- 1384. 162. T. W. Hagler, K. Pakbaz, K. F. Voss and A. J. Heeger, Phys. Rev. B., 1991,44, 8652. 163. P. Dyreklev, M. Berggren, 0. Inganas, M. R. Anderson, T. Hjertberg and 0. Wennerstrom, Adv. Muter., 1994, 7, 43. 164. A. Bolognesi, G. Bajo, J. Paloheimo, T. OstergPrd and H. Stubb, Adv. Muter., 1997, 9, 121. 165. F. W. Embs, G. Wegner, D. Neher et al., Macromolecules, 1991,24, 5068-5075. 166. C. L. Callender, C. A. Carere, G. Daoust and M. Leclerc, Thin Solid Films, 1991,204,451-457. 167. V. Cimrova, M. Remmers, D. Neher and G. Wegner, Adv. Muter., 1996, 8, 146. 168. M. Hamaguchi and K. Yoshino, Appl. Phys. Lett., 1995,67, 3381-3383. 169. U. Lemmer, D. Vacar, D. Moses, A. J. Heeger, T. Ohnishi and T. Noguchi, Appl. Phys. Lett., 1996,68, 3007-3009. 170. A. Ochse, U. Lemmer, M. Deussen et al., Molecular Cryst. & Liq. Cryst. Sci. & Technology Section A - Molecular Cryst. & Liq. Cryst., 1994, 256, 335-342. 171. T. A. Fisher, D. G. Lidzey, M. A. Pate et al., Appl. Phys. Lett., 1995,67, 1355-1357. 172. U. Lemmer, R. Hennig, W. Cuss et al., Appl. Phys. Lett., 1995, 66, 1301-1303. 173. J. Griiner, F. Cacialli, I. D. W. Samuel and R. H. Friend, Synth. Met., 1996, 76, 137-140. 174. J. Gruner, F. Cacialli and R. H. Friend, J. Appl. Phys., 1996, 80,207-215. 175. V. Cimrova, U. Scherf and D. Neher, Appl. Phys. Lett., 1996,69, 608-610. 176. M. Berggren, 0. Inganas, T. Granlund, S. Guo, G. Gustafsson and M. R. Anderson, Synth. Met., 1996, 76, 121-123. 177. N. Tessler, G. J. Denton and R. H. Friend, Nature, 1996,382, 695-697. 178. N. Tessler, S. Burns, H. Becker and R. H. Friend, Appl. Phys. Lett., 1997, 70, 556-558. 179. G. J. Denton, N. Tessler, N. T. Harrison and R. H. Friend, Phys. Rev. Lett., 1997,78,733-736. 180. S . E. Burns, N. Pfeffer, J. Griiner, M. Remmers, T. Javoreck, D. Neher and R. H. Friend, Adv. Muter., 1997, 9, 395. 181. H. Becker, S. E. Burns, N. Tessler and R. H. Friend, J. Appl. Phys., 1997,81, 2825-2829. 182. M. A. Diazgarcia, F. Hide, B. J. Schwartz,M. D. Mcgehee, M. R. Anderson and A. J. Heeger, Appl. Phys. Lett., 1997, 70,3191-3193. 183. D. G. Lidzey, D. D. C. Bradley, M. A. Pate, J. P. R. David, D. M. Whittaker, T. A. Fisher and M. S. Skolnick, Appl. Phys. Lett., 1997, 71, 744-746.

References

457

184. P. R. Berman, Cavity Quantum Electrodynamics, Academic Press Inc., San Diego, 1994. 185. A. Lewis and K. Lieberman, Nature, 1991,354, 214-216. 186. N. F. van Hulst, M. H. P. Moers, R. G. Tack and B. Bolger, Micr. Res. Tech., 1993, 25, 177-178. 187. N. F. van Hulst, M. H. P. Moers and B. Bolger, J. Microscopy, 1993, 171, 95-105. 188. M. Granstrom, M. Berggren and 0. Inganas, Science, 1995,267, 1479. 189. G. Heywang and F. Jonas, Adv. Mat., 1992, 4 , 116-1 18. 190. Q. Pei, G. Zuccarello, M. Ahlskog and 0. Inganas, Polymer, 1994, 35, 134771351, 191. C. Adachi, S. Hibino, T. Koyama and Y. Taniguchi, Jap. J. Appl. Phys. Part 2 - Lett., 1997, 36, L827-L830. 192. M. Berggren, 0. Inganas, G. Gustafsson et al., Nature, 1994,372, 444-446. 193. M. Granstrom, M. Berggren, D. Pede et al., Supramol. Sci., 1997, 4, 27. 194. M. Granstrom and 0. Inganas, Appl. Phys. Lett., 1996, 68, 147. 195. C. Tang, Appl. Phys. Lett., 1986, 48, 183-185. 196. R. F. Chaiken and D. R. Kearns, J . Chem. Phys., 1966,45, 3966. 197. G. R. Johnston and L. E. Lyons, Aust. J. Chem., 1970,23, 1571. 198. R. Zamboni, N. Periasamy, G. Ruani and C. Taliani, Synth. Met., 1993, 54, 57-66. 199. P. Ostoja, S. Guerri, S. Rossini, M. Servidori, C. Taliani and R. Zamboni, Synth. Met., 1993, 54,441-452. 200. G. A. Chamberlain, Solar Cells, 1983, 8, 41. 201. P. Panayotatos, D. Parikh, R. Sauers, G. Bird, A. Piechowski and S. Husain, Solar Cells, 1986, 18, 71-84. 202. P. Panayotatos, G. Bird, R. Sauers, A. Piechowski and S. Husain, Solar Cells, 1987, 21, 301-31 1. 203. N. Noma, T. Tsuzuki and Y. Shirota, Adv. Muter., 1995, 7, 647-648. 204. K. Kuwabara, K. Miyawaki, K. Nawa, N. Noma and Y. Shiroto, Nippon Kagakukaishi, 1992, 1168- 1173. 205. S. Glenis, G. Horowitz, G. Tourillon and F. Garnier, Thin Solid Films, 1984, I l l , 93-103. 206. S. Glenis, G. Tourillon and F. Garnier, Thin Solid FiZms, 1986, 139, 221-231. 207. K. Yoshino, K. Tada, A. Fuji, E. M. Conwell and A. A. Zakhidov, IEEE Trans. Electr. Dev., 1997,44, 1315. 208. K. Tada, M. Onoda, A. A. Zakhidov and K. Yoshino, Jap. J. Appl. Phys. Part 2 - Lett., 1997, 36, L306-L309. 209. T. Kawai, T. Yamaue, K. Tada, M. Onoda, S. H. Jin, S. K. Choi and K. Yoshino, Jap. J. Appl. Phys. Part 2 - Lett., 1996, 35, L741LL744. 210. K. Yoshino, X. H. Yin, S. Morita, T. Kawai and A. A. Zakhidov, Sol. State Commun., 1993, 85, 85. 21 1. N. S. Sariciftci and A. J. Heeger, Int. J . Modern Phys. B, 1994, 8, 237-274. 212. K. H. Lee, R. A. J. Janssen, N. S. Sariciftci and A. J. Heeger, Phys. Rev. B - Cond. Matt., 1994, 49, 5781-5784. 213. L. S. Roman, M. R. Anderson, T. Yohannes and 0. Inganas, Adv. Muter., 1997, 9, 1164. 214. P. Bauerle, U. Segelbacher, A. Maier and M. Mehring, J. Am. Chem. Soc., 1993, 115, 1021 7-10 223. 215. M. G. Hill, J. F. Penneau, B. Zinger, K . R. Mann and L. L. Miller, Chem. Muter., 1992, 4 , 1 106-1 1 13. 216. M. G. Hill, K. R. Mann, L. L. Miller and J. F. Penneau, J . Am. Chem. SOC.,1992, 114, 2728-2730. 217. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Chem. Mater., 1993, 5, 620-624. 218. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Adv. Muter., 1993, 5, 551-554. 219. G. Zotti, G. Schiavon, A. Berlin and G. Pagani, Synth. Met., 1993, 61, 81-87. 220. U. Segelbacher, N. S. Sariciftci, A. Grupp, P. Bauerle and M. Mehring, Synth. Met., 1993,57, 4728-4733. 221. K. E. Ziemelis, A. T. Hussain, D. D. C. Bradley, R. H. Friend, J. Ruhe and G. Wegner, Phys. Rev. Lett., 1991, 66, 2231-2234. 222. M. G. Harrison, K. E. Ziemelis, R. H. Friend, P. L. Burn and A. B. Holmes, Synth. Met., 1993, 55, 218-223.

458

8 Electro-optical Polythiophene Devices

223. I. D. Parker, R. W. Gymer, M. G. Harrison, R. H. Friend and H. Ahmed, Appl. Phys. Lett., 1993,62, 1519-1521. 224. T. Kurata, C. Fukada, H. Fuchigami, K. Hamano and S . Tsunoda, Jpn. J. Appl. Phys. Pt. 2 Lett., 1995, 34, L1464-L1466. 225. M. G. Harrison, R. H. Friend, F. Garnier and A. Yassar, Molec. Cryst. Liq. Cryst. Sci. Technology Section A-Molec. Cryst. Liq. Cryst., 1994, 252, 165-174. 226. M. G. Harrison, R. H. Friend, F. Garnier and A. Yassar, Synth. Met., 1994, 67, 215-221. 227. F. Charra, M. P. Lavie, A. Lorin and D. Fichou, Synth. Met., 1994, 65, 13-17. 228. M. G. Harrison, D. Fichou, F. Garnier and A. Yassar, Optic. Muter., 1997, in press. 229. D. Fichou and F. Charra, Synth. Met., 1996, 76, 11-14. 230. F. Charra, D. Fichou, P. A. Chollet and D. Paquet, Synth. Met., 1996,81, 173-177. 231. D. Fichou, G. Horovitz and F. Garnier, Synthetic Metals, 1990, 39, 125-131. 232. D. Fichou, G. Horowitz, B. Xu and F. Garnier, Synth. Met., 1990, 1990, 243-259. 233. D. Fichou, B. Xu, G. Horowitz and F. Garnier, Synth. Met., 1991,41,463-469. 234. G. Lanzani, L. Rossi, A. Piaggi, R. Zamboni, A. J. Pal and C. Taliani, Chem. Phys. Lett., 1994, 226, 547. 235. J. Poplawski, E. Ehrenfreund, J. C o r d et al., Molec. Cryst. Liq. Cryst., 1994,256,407-413. 236. M. G. Harrison, Ph.D Thesis, University of Cambridge, 1994, 147. 237. J. Grinberg, A. Jacobson, W. Bleha, L. Miller, L. Fraas, D. Boswell and G. Myer, Opt. Eng., 1975, 14, 217. 238. K. Johnson, Phys. World, 1992, 5, 37. 239. B. G. Sfez, E. V. K. Rao, Y. I. Nissim and J. L. Oudar, Appl. Phys. Lett., 1992,60, 607. 240. K. Biedermann,in Holographic recording materials, H. M. Smit, Eds. Springer Verlag, Berlin, 1977, pp. 21. 241. J. W. Yu, D. Psaltis, A. Marrakchi, A. R. J. Tanguay and R. V. Johnson, in Photorefractive materials and their applications, P. Gunter, J. P. Huignard, Eds. Springer Verlag, Berlin, 1989 p. 275. 242. C. J. G. Kirkby and I. Bennion, IEE Proc., 1986,133,98. 243. T. Moriyama, J. Kajita, Y. Takanishi, K. Ishikawa, H. Takezoe and A. Fukuda, Jpn. J. Appl. Phys., 1993, 32, L589. 244. J.-M. Nunzi, F. Charra and N. Pfeffer, J. de. Phys. IZI France, 1993,3, 1401-1411. 245. J. M. Nunzi, N. Pfeffer, F. Charra and D. Fichou, Chem. Phys. Lett., 1993, 215, 114-119. 246. R. Raue and H. Harnisch, Heterocycles, 1984, 21, 167. 247. M. Irie, K. Uchida, T. Eriguchi and H. Tsuzuki, Chem. Lett., 1995,899-900. 248. H. Miyasaka, T. Nobuto, A. Itaya, N. Tamai and M. Irie, Chem. Phys. Lett., 1997, 269, 28 1-285. 249. N. Tamai, T. Saika, T. Shimidzu and M. Irie, J . Phys. Chem., 1996, 100,4689-4692. 250. M. Irie, Pure & Appl. Chem., 1996, 68, 136771371, 251. G. M. Tsivgoulis and J. M. Lehn, Adv. Muter, 1997, 9, 627-630.

9 Oligo- and Polythiophene Field Effect Transistors H. E. Katz, A. Dodabalapur and Z. Bao

9.1 Introduction In recent years, interest in transistors and other electronic and optical devices with organic/polymeric active materials has increased significantly. There are many reasons for this increase, and perhaps the most important ones have to do with the inherent compatibility of systems made from such devices with low-cost, large area processing techniques [ 1,2]. Improvements in organic transistor performance have led to performance levels that approach those of amorphous silicon field-effect transistors (FETs), and the recent increased effort has led to significant progress being made in understanding the fundamental physics and chemistry of electrically active organic materials. Recent reviews have summarized the progress made in the field [3, 41. The earliest organic FETs had polymeric active layers. Ebisawa et a/. observed electric field dependent conductivity in metal-insulator-semiconductor (MIS) structures with polyacetylene as the semiconductor [5]. Koezuka et al. [6] and Assadi et al. [7] demonstrated FETs with polythiophene, and Burroughes et al. reported work on polyacetylene FETs [8]. The use of smaller organic molecules as semiconductors was pioneered by researchers from CNRS in Thiais, France. In a series of articles the CNRS group reported comparatively high mobilities (>lo cm2V-' s-' with ordered oligomers of thiophene such as a-sexithiophene (a-6T) [9- 111, and pointed out the importance of molecular ordering in achieving high mobilities in organic FETs. Research on organic/polymeric FETs has not been confined to oligothiophene/ polythiophene materials. Important advances have been made with other classes of materials such as fullerenes [12, 131, fused ring compounds such as pentacene [14, 151, and the phthalocyanines. While most of this chapter is focused on oligoand poly-thiophene FETs, there are a few brief discussions of FETs with other organic materials to illustrate some key points. A significant part of current research is aimed at understanding the operation of organic transistors in greater detail [ 16, 171. However, some important strides have been made in developing logic families and circuits with organic FETs [18, 191. This chapter will not discuss logic families and circuits, but will concentrate on the operation and modeling of organic FETs, interface effects, short-channel effects, sub-threshold conduction, and the influence of energy levels. Section 9.4 will discuss synthesis procedures for oligothiophene and oligothiophene derivative materials and document the key figures of merit of FETs made with such materials by various groups worldwide. Section 9.5 is devoted to polythiophene FETs, and the significant

460

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophenefield effect

advantages of regio-regular polythiophene in comparison with regio-random polymers. Section 9.5.3 also describes the characteristics of a screen printed FET with polymer components.

9.2 Operation of a field-effect transistor The schematic structure of a typical organic thin-film transistor (TFT) is shown in Fig. 1. The substrate is doped Si, which also functions as the gate, and the thermally grown SiOzfunctions as the gate dielectric. Immediately above the gate dielectric are thin metallic source/drain electrodes. In most oligothiophene TFTs, Au is employed as the source (S) and drain (D) contacts, although other metals such as Al, Pt. etc. have also been successfully used. The active semiconducting material is deposited on top of the S/D electrodes, and it is necessary to keep the thickness of the S and D contacts sufficiently small so that there are no discontinuities between the electrode and the organic semiconductor. The device structure shown in Fig. 1 is used by most groups to evaluate different organic semiconductors for TFT applications, estimate field-effect mobilities, and study charge transport. In a simple variation to the structure shown in Fig. 1, the S/D contacts are defined above the organic semiconductor. Other device conflgurations, including those with polymer gate insulators and plastic substrates have been implemented and some of these are discussed later in the chapter. Most oligothiophene TFTs operate in the enhancement mode, which means that the transistor is normally off, when there is no gate bias. In order to have good enhancement-mode transistor operation the organic semiconductor must be substantially free of residual charges caused by adventitious dopants. If there are free charges due to dopants in the material, the transistor is not really ‘off’ under zero gate bias and the non-negligible channel conductivity can influence the performance characteristics of circuits and systems involving organic TFTs. Techniques to Organic Semiconductor

Source

\

Drain

Doped Si Gate

\ Gate Contact

Figure 1. Schematic structure of an organic FET formed on a Si substrate. The doped Si functions as the gate and thermal oxide as the gate insulator. The source and drain contacts in this particular structure are formed directly on the gate insulator. It is also possible to form the source/drain contacts above the organic semiconductor.

9.3 Modeling of oligothiophene TFTs

461

synthesize and purify oligothiophenes so that the residual conductivity is low are described in the following section. The basic operation of an enhancement mode p-channel TFT is described below, and in the following subsection, the operation is examined in more detail. When the gate bias is zero, the gate and source are both at zero potential (or grounded). Ideally, under such bias conditions there are no carriers in the channel region between the source and the drain. Under these conditions, the application of a drain bias (VDs < 0) does not lead to significant currents flowing between the source and drain. The current that does flow is called the off current. When the (negative) gate bias is increased in magnitude, positive charges are induced in the channel through the source electrode, which is always grounded. This charge induction is simply an electrostatic effect, and the density of the total induced charge is given by C,VGswhere Ci the gate insulator capacitance per unit area. In some organic semiconductors, some of the gate field-induced charge is trapped and does not contribute to the conductivity of the channel. The free charges induced by the gate field increase the conductivity of the channel leading to an increase in the drainsource current. For relatively small drain-source voltages, the drain-source current increases linearly with voltage for a fixed gate voltage. This regime, in which the channel behaves as if it were a resistor, is called the linear region. Further increase of the drain-source voltage causes the field-induced charge to be forced away from the drain electrode, and a depletion region forms between the source and the fieldinduced charge. This is called the saturation region, and the drain-source current ideally remains constant even when the drain-source voltage is increased.

9.3 Modeling of oligothiophene TFTs 9.3.1 Analytical modeling In analytical modeling, we attempt to reproduce the characteristics of the transistor with a few relatively simple equations derived from the basic physics of operation of the field-effect transistor. It is necessary to make some approximations in order to obtain closed form expressions for the transistor drain current. Such modeling quite accurately reproduces the on-current characteristics in both the linear and saturation regions. It is not very good at modeling the sub-threshold characteristics, the charge and electric field distributions. For these parameters, a 2-dimensional numerical model is useful and such modeling is described in the following section. Analytical modeling of a-6T TFTs by Torsi et al. [20] and by Horowitz et aE. [21] using approaches analogous to that employed in the modeling of amorphous silicon thin film transistors led to results which agreed quite well with experiment. Torsi et al. calculated the current-voltage characteristics of short-channel organic TFTs starting from the solution of the drain current equation for an enhancement mode p-channel MOSFET [22]:

462

9 Oligo- and Polythiophene Field EfSect Transistors Oligo- and polythiophene field efSect

where:

Z is the channel width and L is the channel length. ZD and VD are, respectively, source-drain current and voltage, n,,d is the induced surface charge concentration in the channel per unit area, VGand VT are gate and threshold voltages. q is the elemental charge, C, is the gate insulator capacitance per unit area, and ,LA is the fieldeffect mobility. VT is positive in the case of lightly doped a-6T since at zero gate bias the device is not totally insulating due to the p-type conductivity of the organic layer. It is also assumed that none of the induced charge is trapped. Short channel conditions are present when the channel length, L, is comparable to the sum of the source and drain depletion layer widths. When this happens the potential distribution in the channel region, as well as the electric field (Ex controlled mainly by VG and Ey controlled by VD), becomes two-dimensional and the gradual channel approximation (Ex >> Ey) is no longer valid. The TFT is then equivalent to an intrinsic TFT (TFTI) of channel length L1 = L - AL in series with an effective p-i-p diode of variable length AL [20, 221. A model of I-V characteristics for shortchannel TFTs has to take into account both the effect caused by the shortening of the channel length (AL) as well as by the associated parasitic resistance (&). This resistance is in series with the intrinsic TFT channel resistance. Since AL is the pinched-off portion of the channel, both of these effects have to be considered only at large drain voltages ( VD > Vgt). In the AL channel region, for V D > Vgt, a regime of space-charge-limited current is established and carrier transport in such regime may be different from the rest of the channel. If the parasitic resistance (Rp) becomes comparable to the channel on-resistance it will affect the transconductance of the TFT. In the saturation region ( VD > Vgt) the integration of Eq. (1) in the short-channel approximation gives:

For long channel FETs, AL = 0 and Eq. ( 3 ) reduces to the familiar expression for drain current in the saturation region. In the linear region ( VD 5 VEt) the integration of Eq. (l), gives:

In Fig. 2 are shown the I-V characteristics of a 12 pm channel length a-6T FET together with calculated characteristics using Eqs (3) and (4) and it can be observed that the agreement between data and the simulation is good. At such channel lengths, short-channel effects are negligible.

9.3 Modeling of oligothiophene TFTs

463

-2.7

E -1.3

0

-20

-40

-60

-80

a

-100

Drain Voltage (V) Figure 2. Current-voltage characteristics of an a-6T FET with channel length L = 12 pm and width 250 pm. The lines are experiemental data and the solid circles are the simulated characteristics. Reprinted with permission from Ref. 20. Copyright 1995 American Institute of Physics.

9.3.2 Numerical modeling Numerical modeling is far more powerful than analytical modeling in attaining a detailed understanding of the transistor characteristics above and below threshold [23]. There are many numerical simulation programs available, including some that are 2-dimensional. In 2-dimensional simulators, the fundamental device equations are solved self consistently along the channel length (first dimension) and perpendicular to the channel (second dimension). The parameters are assumed not to vary along the third dimension (along the channel width). This assumption is valid for large Z/L transistors. The fundamental equations used in numerical modeling are:

W V P ) = -4(P

dn

1

at

4

-= -

-

(5)

VJ,+(G-R)

”- -1-V J p + ( G - R ) at 4

(7)

where

J,

=

-qpnVp

+ qD,Vn

-w&

-

(8)

9DpVP (9) where Eq. (5) is Poisson’s equation, Eqs. (6) and (7) are known as continuity equations in which G and R are the charge generation and recombination rates respectively (both assumed to be zero in organic FETs), and Eqs. (8) and (9) are current density equations. p is the total hole/positive charge density, and n is the total Jp =

464

9 Oligo- and Polythiophene Field EfSect Transistors Oligo- and polythiophene field efect

electron/negative charge density. E is the dielectric constant, q is the elemental charge, cp is the potential, J, and J, are the electron and hole current densities, and pn and p, are the electron and hole mobilities. D, and D, are the diffusion coefficients which are assumed to be related to the respective mobilities through the Einstein relation D / p = k T / q . Equations (5)-(9) are solved numerically after including the appropriate boundary conditions ( S , D, and G potentials) and dimensions. The significance of Eqs. (5)-(9) is explained in Ref. 24.

9.3.3 Interface effects The location of field-induced charge can be calculated with the help of the model outlined above [23]. In Fig. 3 the distribution of charge normal to the plane of the semiconductor-dielectric interface is shown. It can be seen that most of the field-induced charges are located very close to this interface at moderately high gate fields. This finding is also supported by experimental data gathered from a series of a-6T TFTs with differing active layer thickesses. The field-induced conductivity is approximately the same for active thicknesses ranging from 5 nm to 150nm [25]. The a-6T molecules are oriented such that the long axis is approximately perpendicular to the plane of the interface with the gate insulator. Thus, a thickness 1oZ0

I

I

0.00

0.01

0.02

0.03

0.04

0.05

position (pn) Figure 3. Variation in the field-induced charge density perpendicular to the semiconductor-insulator interface which is at a position 0.05 pm. It can be seen that most of the charges reside within the first 5 mm from the interface. This profile was calculated using the numerical model and assuming that the thickness of the a-6T film is 50 nm. Reprinted with permission from Ref. 23. Copyright 1997 IEEE.

9.3 Modeling of oligothiophene TFTs

465

of 5 nm corresponds to only two molecular layers. The location of most of the fieldinduced charge emphasizes the importance of the semiconductor/insulator interface and the molecular ordering near at this interface. The importance of the nature of the semiconductor/gate insulator on organic transistor performance was first noted by Haddon et al. in C60FETs [13]. In these FETs the measured mobility was 0.08 cm2V-' s-', when the C60 was deposited directly on the gate insulator (SOz). In devices in which the insulator surface was treated with tetrakis(dimethylamin0) ethylene before depositing the C60,the mobility increased by a factor of nearly 4 (to 0.3 cm2V-' s-I), and the trap density and threshold voltage were reduced. Qualitatively similar improvements in field-effect mobility were reported by Lin et al. in pentacene FETs in which the interface between the organic semiconductor and the gate dielectric was treated with a selforganizing monolayer of octadecyltrichlorosilane (OTS) [141. This treatment was found to result in a factor of 2 improvement in mobility and substantial improvements in subthreshold characteristics. These examples illustrate the importance of interfaces and interface treatment in improving the characteristics of organic TFTs.

9.3.4 Short-channel effects Knowledge of the distribution of the electric field in the channel is very useful for a number of reasons. In many organic materials, the mobility is electric-field dependent, and an accurate model of the current-voltage characteristics must consider the influence of the electric field on the mobility. The numerically computed potential profile for an L = 1.5 pm channel length a-6T TFT is shown in Fig. 4.In this device the gate dielectric thickness is 300 nm, and the I-V characteristics exhibit pronounced short-channel effects. As noted in the previous sub-section, short-channel effects are a set of effects which become prominent when the electric field along the channel becomes comparable to the gate field (perpendicular to the channel). The main effects on the transistor characteristics are (1) the saturation characteristics are adversely affected, and, (2) the subthreshold current is increased. One way to mitigate the magnitude of short-channel effects (for a given channel length) is to lower the gate dielectric thickness. This lowers the drain voltage Organic Semiconductor Source

1

Gate

Drain

Gate insulator

Figure 4. Potential profile of a L = 1.5 pm a-6T FET for VDS= -lOOV and VGs = -8OV. The gate insulator thickness is 300nm. Reprinted with permission from Ref. 23. Copyright 1997 IEEE.

466

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiopheneJield effect

t

I

-3.0~10~

- 80

E

-20

0 0

-20

-40

-60

-80

-100

Drain Voltage (V) Figure 5. Current-voltage Characteristics of an a-6T FET with L = 1.5pm and gate insulator thickness 300 nm. Reprinted with permission from Ref. 20. Copyright 1995 American Institute of Physics.

required for saturation and therefore reduces the extent of short channel effects. Figure 5 shows the characteristics of a FETs with L = 1.5pm in which the gate dielectric thickness is 300 nm. The figure shows pronounced non-idealities in the electrical characteristics due to short-channel effects. Reducing the gate dielectric thickness to -50 nm will improve the I-V characteristics.

9.3.5 Sub-threshold characteristics The sub-threshold characteristics of a FET are important in many circuit applications. The behavior of the drain-source current below threshold depends on a number of factors such as the channel potential profile, interface states and morphology, doping and bulk trap density in the organic semiconductor. In highly purified a-6T synthesised by Katz et al., the threshold voltage was almost zero [26]. More generally, most oligothiophenes and polythiophenes are slightly doped and the threshold voltage is positive. The sub-threshold swing, which is a measure of how rapidly the drain current increases with gate voltage just below threshold, can be computed numerically. This is illustrated in Fig. 6(a) where the drain current variation with gate voltage has been plotted for two a-6T FETs with 4 pm and 25 pm channel lengths and 300 nm thick Si02gate dielectric thickness. The corresponding experimental data have been reported in Ref. 20 and are shown in Fig. 6(b) . The sub-threshold swing (indicated by the arrows) extracted from the calculations are 2.7 V/decade and 4.6 V/decade for the 4 pm and 25 pm channel length devices respectively. The measured sub-threshold swings are -1 0 V/decade, considerably different from the calculated values. This difference was attributed by Alam et al. to interface effects which were not included in the calculation but which are likely to be present in an actual device [23]. A small sub-threshold swing value is very desirable, and Lin et al. have reported on being able to control the sub-threshold slope in pentacene FETs by the use of interfacial layers such as OTS [14].

9.3 Modeling of oligothiophene TFTs

1o-’2

2

1o

467

, 0.0

1

-~

10.0

20.0 Applied Voltage (Volts)

30.0

.O

V, = - 5 0 V

L=15um

,

0

.-r

Gate Voltage (V)

Figure 6 Sub-threshold characteristics of a-6T FETs (a) From numerical modeling for two channel lengths L = 4 p m and 25pm, (b) Measured characteristics for L = 1.5, 4, and 25pm.

9.3.6 Energy levels Oligo- and poly-thiophene FETs are p-channel devices and the energy level of the highest occupied molecular orbital (HOMO) are generally in the range 4.9-5.5 eV below vacuum. Most p-channel organic FET materials have a HOMO level in this range. If the HOMO energy is too far from the vacuum level (>5.5 eV), trap states caused by impurities and defects compete effectively for the field-induced charge. In many such cases the formation of an accumulation layer is prevented by traps.

468

9 Oligo- and Potythiophene Field Effect Transistors Oligo- and polythiophene field effect

On the other hand, if the HOMO energy level is too close to vacuum, impurities tend to dope the organic semiconductor resulting in a large positively charged carrier density. The range 4.9-5.5eV appears to be optimal from t h s standpoint. Some groups have attempted to add electron withdrawing groups to oligothiophene molecules in order to move the lowest unoccupied molecular orbital (LUMO) level away from the vacuum level and achieve n-channel transistor operation. For n-channel operation, LUMO levels in the range 3.9-4.7 have been found to be optimal [ 151. No group has reported n-channel FET operation with oligothiophenes; however, the approach of adding electron withdrawing groups to alter the LUMO level has succeeded in the case of copper phthalocyanine (CuPc). CuPc ordinarily results in p-channel FETs [27]; however, Bao et al. have found that if 16 fluorine atoms are added to the CuPc molecule, the LUMO level is shifted by a sufficient amount to result in excellent n-channel operation [28].

9.4 Oligothiophene FETs 9.4.1 Synthesis and purification The key preparative step common to all syntheses of thiophene oligomers greater than three is the formation of the thiophene-thiophene bonds. There are three principal ways to create the necessary thiophene links, illustrated in Fig. 7. The most convenient in terms of reagents and equipment is to employ a neutral oxidizing agent such as ferric chloride in an apolar organic solvent [29] to convert a neutral, short thiophene oligomer to a radical cation-like species that couples to a second thiophene moiety at ambient temperature. This method has the least regiochemical selectivity, with side products bonded at positions other than a to the sulfur [30]. Significant quantities of dopant derived from the oxidizing agent are also incorporated. [261 The second and most commonly used method is to deprotonate the a carbon of the short precursor with an organolithium reagent or the lithium salt of a secondary amine, and then oxidatively dimerize the anionic precursors below room temperature [311. A milder oxidizing agent can be employed than for the neutral precursors, and cupric chloride is a frequent choice [32]. Little if any bonding to p carbons is observed with this method, but inorganic impurities are still retained to some degree. An additional side reaction is nucleophilic attack by the thienyllithium intermediate on groups such as chloro bound to the oxidizing metal, to give chlorinated thiophene impurities [26]. This side reaction is somewhat suppressed if the thienyllithium is prepared as its TMEDA complex, and is not observed at all if the extremely mild oxidizing agent, ferric acetylacetonate, is used [26]. There is a tradeoff between minimizing oxidizer-derived side products and avoiding decomposition of the organolithium species that can occur if the low reactivity of the oxidizer requires that the desired coupling occur at a higher temperature [4]. The third and most elaborate procedure is to separately prepare a-bromo and a-stannyl precursors and couple them at elevated temperature using palladium-based catalysts, the ‘Stille’ coupling [33,34]. Extra synthetic steps are required to obtain the precursors, although

9.4 Oligothiophene FETs

469

1

FeCI,

CuCI, or Fe(acac),

R

R

Figure 7. The three principal thiophene-thiophene bond forming reactions used in the synthesis of unsubstituted thiophene oligomers. The third reaction employs a palladium complex as a catalyst.

the procedures are facile and high-yielding. The coupling is selective and reliable, and can be used for oligomers with odd numbers of thiophene rings as well as for unsymmetrically substituted oligomers. Other, more elaborate routes to thiophene oligomers that involve construction of one of the rings as a final step have also been devised [35]. The degree of purity of the oligomers affects the electrical behavior of TFTs made from the oligomers in both the 'off' and 'on' states [26]. Any doping impurities can induce extrinsic charge carriers that greatly raise the off-current and lower the on/off ratio. This is illustrated by the I-V curves in Fig. 8. The thicker the oligomer film in the TFT, the greater the dopant-induced off-current. Extensively purified oligomers can give devices whose off-current at zero gate voltage is virtually undetectable, as shown in Fig. 9. Impurities have a lesser effect on the mobility than on the off-current. Conceivably, structural isomers or oligomers of mismatched lengths or

410

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophenefield effect

h

-4-

n -90

0

0

-10

-20

-12.5

-25

-2

h

v 5L

-0

-1

-

n 0

,v,

(V)

Figure 8. Current-voltage plots for transistors fabricated from highly impure and substantially purified a-6T.

-m -a,

E -a,

-PI P OJ

0 8

8

QQ

-20

-40

40 Drain Volbge (V)

-80

-100

Figare 9. Current-voltage plot for a transistor fabricated from a-6T purified to an ‘intrinsic’, or dopant-free level. Reprinted with permission from Ref. 26. Copyright 1995 American Chemical Society.

9.4 Oligothiophene FETs

471

orbital energies can cause localized morphological and electronic defects that would create traps or barriers to charge transport and lower the mobility. No dramatic effects of impurities on macroscopic film crystallinity or smoothness have been reported. Some impurities can actually enhance the apparent mobility in a device by lowering the barrier to charge injection from the source or by filling traps [36]. However, these effects are usually outweighed by the increased off-current. There has been much discussion of the dependence of mobility on the temperature at which TFTs are fabricated or annealed, and this dependence is often correlated with thermally induced morphological differences, to be discussed below. However, the effects of impurities are also fabrication-temperature dependent in that some impurities are desorbed from films at high temperatures or excluded from the near-interface region that controls the channel current.

9.4.2 Morphology Single crystals [37-411 and crystalline powders [42-441 of unsubstituted a-4T, 5T, 6T, and 8T share a common layered herringbone packing motif, with the molecular planes, in which to a very good approximation the thiophene rings lie, rotated about 60" relative to those of nearest neighbor molecules in the same layer. An example is shown in Fig. 10. The molecular long axes are slightly tilted with respect to the layer

Figure 10. Crystal packing diagram for the 'high-temperature' polymorph of a-6T illustrating the herringbone structure. From Ref. 39 (MRS).

472

9 Oligo- and Polythiophene FieId Effect Transistors Oligo- and polythiophenefield effect

boundary planes. Minor variations in these angles lead to polymorphs, which are present at least in the tetramer and hexamer, although at least one of these polymorphs is common to the tetramer, hexamer, and octamer, normalised for the three different molecular lengths [42]. In addition, this motif is shared by dimethylquaterthiophene [37, 451, and to a considerable degree, other elongated conjugated polycyclic compounds. In many cases, evaporated thin films of these oligomers on a range of substrates and some polymer surfaces display the same packing, with the layer planes parallel to the substrate and the long axes nearly perpendicular to the substrate [46]. The preferred perpendicular orientation is observed in films of oligomers of all four lengths, deposited both by conventional sublimation and molecular beam deposition, although molecular beam deposition can lead to enhancement of the orientational order [47,48]. The tendency toward perpendicular molecular orientation, producing layers parallel to the substrate, is most fortunate for oligothiophene applicability to TFTs since the hole mobilities are significant and most likely isotropic in the directions parallel to the planes, enabling the formation of a channel in a horizontally configured device with little or no dependence on the in-plane crystallographic axes relative to the source-drain direction. Excellent device characteristics have in fact been correlated with this expected favorable orientation. Beyond these generalizations, substrate temperature and polarity, rate of deposition, and thermal annealing all affect the orientational homogeneity and morphology of vacuum-deposited thiophene oligomer films. While in some cases, the ‘kinetic’, or low-temperature/fast deposition orientation may be to have the chains parallel to the substrate [49], the thermodynamically equilibrated orientation is to have the chains perpendicular to the substrate, and the layer planes parallel. Besides determining the size of ordered domains, the substrate temperature and deposition rate also affect the polymorphic identity (defined especially by the long axis tilt angle and the angle between the two shorter axes) of the film being grown, especially for a-6T [49]. One extraordinary example of a substrate effect is the induction of 88% parallel-tosubstrate orientation of a-6T chains evaporated onto friction-oriented teflon films [50, 511; both the extreme nonpolarity and specific topography of the substrate differentiate this system from amorphous and more polar substrate surfaces. A related effect has been observed on stretched polyethylene [52]. The most typical morphology for evaporated films of a-6T and related compounds is interconnected grains of sizes on the order of lOOnm [49, 531, implying that hole transport across grain boundaries is an important if not the dominant factor limiting TFT mobility. A micrograph displaying this morphology is shown in Fig. 11. Higher substrate temperatures can lead to larger grains, sometimes accompanied by larger and deleterious spaces between them. Very low substrate temperatures, on the other hand, produce extremely fine grains with little coherence. In such cases, although the intergrain spacing may be vanishingly small, the mobility is low because there are so many grain boundaries and the orientational discontinuities may be more abrupt. The grains are in the form of platelets, a manifestation of the preferred growth of crystallites along the 7r-stacking, in-plane directions. An additional way to obtain larger domains is by rapid thermal annealing of a polycrystalline film [53, 541. With this procedure, grains with lateral dimensions of many pm may be grown, as exemplified in Fig. 12. Unfortunately, these domains,

9.4 Oligothiophene FETs

473

Figure 11. Morphology of a-6T sublimed onto silicon/silicon dioxide showing tenth-micron domains. From Ref. 54 (ACS).

Figure 12. Morphology of rapid-thermal-annealed a-6T showing large, flat domains. Reprinted with permission from Ref. 54. Copyright 1995 American Chemical Society.

nearly featureless at first glance, are replete with microcracks that seem to limit the mobility just as the more obvious grain boundaries do in unannealed samples. Such cracks have also been observed in single crystals of dimethylquaterthiophene [45]. One beneficial effect of the annealing is to lower the off-current and thereby raise the on/off ratio of TFTs that have been subject to the treatment. However, this effect is most likely due to semiconductor purification via impurity desorbtion, rather than a morphological change.

9.4.3 Substituted oligothiophenes The liquid-phase processability [34,55-571 and thin film ordering [55,57-601 of thiophene oligomers is greatly enhanced by appropriate substitution at the terminal (Y carbons. On the other hand, substitution at internal positions imparts considerable

414

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophene field effect

distortions of the inter-ring conjugation [61, 621 and is less useful in designing thiophene oligomer semiconductors [58]. Terminal substitution with trialkylsilyl[63] or tetrahydrobenzo [64] groups is an intermediate case; while the resulting oligomers probably form organized films, the bulkiness of the substituents would be expected to diminish the intermolecular .Ir-overlap. Medium-sized alkyl and alkoxyalkyl groups appear to be the most useful of the end substituents considered, with the alkoxyalkyl groups conferring a somewhat greater solubility at the expense of slight disordering of the films [56]. Small substituents such as ethyl and methyl have little solubilizing effect, while very long chains such as octadecyl increase the intermolecular packing energy, decrease oligomer solubility, and increase the temperatures needed to anneal films to their most ordered state. The solid state structures of all the end-substituted oligomers feature segregation of the substituents from the conjugated cores, and perpendicular orientations of thin films. Either the substituents or the core may be at close to a 90" angle with respect to the substrate, while the other molecular component is more significantly tilted. Two synthetic approaches are available for end-substituted oligothiophenes, delineated in Fig. 13. One is to take advantage of the commercial availability of 2-alkylthiophenes and couple them via nonpolar organometallic intermediates to additional thiophene rings 1581. The second is to affix single polar substituents to bi- and terthiophene, transform the substituents to the desired nonpolar ones, and then couple the intermediates [56, 591. The first strategy avoids a moderate yield monosubstitution step, while the second affords the opportunity to rigorously and conveniently purify the precursor oligomers on the basis of polarity differences among starting materials, products, and oversubstituted byproducts. The Stille coupling, being fairly tolerant of functional groups, has been used in syntheses of thiophene chains with highly dipolar groups appended to their a carbons [34]. When this

Figore 13. Two synthetic routes to end-substituted thiophene hexarners.

9.4 Oligothiophene FETs I

I

I

t"'"""'"

475

r ' v_' l "-100' l P

104

- 4. 0-

r

"

."

in0

0

0

l I

-20

,

.

,

l

, I

-40

,

,

1

,

I

-60

,

,

1

, I

-80

,

,

.loo

Drain- Source Voltage (V) Figure 14. Current-voltage plot for TFTs fabricated by solution deposition of dihexyl-a-6T. Reprinted with permission from Ref. 56. Copyright 1998 American Chemical Society.

group is aminomethyl, further modification of the oligomer by attachment of additional molecular subunits, including biologically active ones, after deposition as a film may be accomplished [65]. This may have implications for the use of thiophene oligomer devices as biosensors. The effects of hexyl substituents on a-4T and a-6T are particularly pronounced [57-601. The domain sizes, dichroic ratios, and tendencies to form mesophases are all greater for thin films of dihexyl oligomers than for films of the corresponding unsubstituted oligomers prepared under similar conditions. Device-sized single crystal films of a-4T can be routinely obtained by evaporation, leading to astonishingly high mobilities and highly anisotropic charge transport properties favoring the intraplanar directions 11601.TFTs with mobilities >0.02 cm2V-' s-' can be fabricated by casting solutions of dihexyl-a-6T [56]. Characteristics of such devices are graphed in Fig. 14. These benefits may be attributed to the greater repulsion of these molecules by polar surfaces and the availability of the mesophases as intermediates in orientational annealing during film deposition at elevated temperatures. The influence of the hexyl group is less pronounced in the already well-ordered a-8T [66].

9.4.4 Fused ring materiels A conceivable modification to the oligothiophene motif is to embed the thiophenes in fused rings. The resulting structures are even more planar and rigid than the catenated single thiophene rings, and approach the polyacenes in their ribbon-like shapes. Benzo- [67], anthra- [68], and thienodithiophenes [69] have been investigated. Structures of these fused heterocycles are drawn in Fig. 15. The influence of deposition temperature on grain size is even greater than for the conventional thiophene oligomers, with in-plane dimensions ranging from tens to thousands of nm. The effect of end substitution is dramatic in that dihexylanthradithiophene,

476

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophene field effect

Figure 15. Alternative chemical structures for semiconductors incorporating thiophene rings. From the top: bis(benzodithiophene), anthradithiophene (ADT), bis(dithienothiophene), and bis (bithienyletheny1)thophene.

melting at 395"C, is nevertheless sufficiently soluble for solution fabrication of a TFT with mobility >0.01 cm2V sC1 [68]. The thienodithiophene is noteworthy for its face-to-face 7r-stacks in the solid state. A final, recently reported variation on the thiophene oligomer theme is to insert double bonds between some of the thiophene rings, specifically to give 2,5-bis(bithienylviny1)thiophene [70] (Fig. 15). Because the resulting molecule is less rigd than the all-heteroaromatic compounds, it is more soluble and can be deposited by spin-coating from N-methylpyrrolidone solutions. No orientational or morphological data were reported for the material in thin film form. The mobilities of various oligothophene FETs reported by several groups have been summarized in Table 1.

9.5 FETs based on polythiophenes Liquid phase processable materials are important in order to fabricate devices using low-cost processes such as spin-coating, casting, or printing. Three methods have been used to fabricate polymer TFT devices from the liquid phase. In the first method, a semiconducting polymer layer is formed directly on the electrodes by electrochemical polymerization, and these electrodes are used subsequently as drain and source electrodes [5, 77-79]. The first polythiophene TFT was fabricated in 1986 by this method and ossessed a mobility, p = 1.2 x cm2V-' sC1 [5, 771 Polypyrrole (p = 1.2 x 10- to 1.77cm2V-' s-') and poly(N-alkylpyrro1e)s (p = 6.3 x to 1.74crn2V-' s-') have also been prepared by electrochemical synthesis and their TFT properties were studied for these essentially doped films [79]. The second technique involves the use of a soluble precursor polymer which

P

9.5 FETs based on polythiophenes

417

Table 1. Field-effect mobilities of oligothiophenes. Compound

Mobility (cm2V-’ s)

a-6T derivs: unsubst. unsubst. dimethyl dihexyl dihexyl dihexyl didodecyl dioctadecyl bis(C40C3) bis(C40C3) bis(CsOC3) a-8T a-8T dihexyl-a-8T a-4T diethyl-a-4T dihexyl-a-4T dihexyl-a-4T bis(C40C3)-a-4T bis(C40C3)-cr-4T a-5T dimethyl-a-5T ADT dihexyl-ADT dihexyl-ADT didodecyl-ADT dioctadecyl-ADT bisbenzodithiophene bisdithienothiophene bisbithienylethenylthiophene

Dep Temp (“C)

0.025 0.02 0.006 0.05 0.04 0.03 0.016 0.001 3 0.033 0.008 0.009 0.02 0.012 0.015 0.005 5x 0.2 0.03

280

9 8x 0.0015 1 x 10-~ 0.09 0.17 0.015 0.16 0.13 0.04 >0.02 0.01 0.0014

80 70

70

70

30-80

85 85 100 85 85 100 100

Remarks

on polyimide anneal >100°C cast from C6H3CI3

cast from PhCl anneal 125°C on PMMA on PMMA TCNQ at S-D contacts cast from CHCI, single crystal morph. spun from CHC13, PMMA cast from PhCl

cast from PhCl

spun from NMP

Ref.

49 25 75 58 59 56 56 56 56 56 56 26 66 66 76 55 60 57 56 56 76 75 68 68 68 68 68 67 69 70 70

can undergo subsequent chemical reactions to give the desired conjugated polymer such as poly(thieny1ene vinylene) ( p = 0.22 cm2V-’ s-’) [SO]. In these two methods, low field-effectmobilities have been reported except for poly(thieny1enevinylene) and polypyrroles, which were doped to achieve high mobility [Sl, 821. The low mobility in most of these materials is probably due to poor ordering and the amorphous nature of the thin films. The third technique utilizes soluble conjugated polymers and they are fabricated by spin-coating, casting, or printing techniques. Different conjugated polymers have been studied. Examples including poly(2,5-dialkylpheneylene-co-phenylene)s, poly(2,5-dialkylphenylene-co-thiophene)s,poly(2,5-dialkylphenyleneviny1ene)s and the dialkox l derivatives of the above polymers [83]. However, very low (less than cm2V- s ’) or no field-effect mobilities have been found. More extensive effort has been directed towards soluble polythiophene derivatives since they are widely used as conducting and semiconducting materials. Among them, regioregular poly(3-alkylthiophene)~have been found to have the highest field-effect

Y-

478

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophene field effect

mobilities (ca. 0.05 cm2V-' s-') [84]. In this section, the transistor properties for different poly(3-alkylthiophene)~will be reviewed with special focus on regioregular poly(3-alk ylthiophene)~.

9.5.1 Regiorandom polythiophene FETS The regiorandom poly(3-alkylthiophene)~used for transistor studies are usually prepared using FeC13 oxidative polymerizations from 3-alkylthiophenes. Inganas et al. have investigated MOSFETs of re iorandom poly(3-hexylthiophene) [7]. The field-effectmobility was found to be 10- -10 -4 cm2V-' s-'at room temperature and decreased with increased temperature. Yoshino et al. have fabricated Schottky gated FETs as shown in Fig. 16 using regiorandom poly(3-alkylthiophene) freestanding films [85, 861. A carrier mobility of 3 x 10-3cm2V-' sC1 was reported for such a device. The low mobilities reported for regiorandom poly(3-alkylthiophene)s are attributed to their random structures and amorphous morphology in the solid-state. Attempts have been made to improve the molecular ordering of regiorandom poly(3-alkylthiophene)~by Paloheimo et al. to fabricate FETs using Langmuir-Blodgett (LB) techniques [87]. However, a large amount of arachidic acid (40 mol%) has to be incorporated in order to form stable LB films and the cm2V-' s-I). resulting devices showed low field-effect mobilities (ca. 6 x

9

9.5.2 Regioregular polythiophene FETS Regioregular poly(3-alkylthiophene)~have been shown to have very different properties from their corresponding regiorandom polymers, such as smaller bandgaps, better ordering and crystallinity in their solid states, and substantially improved electroconductivities [88]. The transistor properties of poly(3-alkylthiophene)~ have been studied by Bao et al. and field-effectmobility as high as 0.05 cm2V-' s-l and on/off ratios close to lo4 have been reported [84]. Figure 17(a) is an I-V curve for a regioregular poly(3-hexylthiophene) (PHT) p-channel device operating in the accumulation-mode in which the drain-source currents (IDs) of negative signs scale up with negative gate voltages (VG).The fieldeffect mobility calculated for the device shown in Fig. 17(a) is 0.045cm2V-' s-'. This is one of the highest values achieved for polymer TFTs. In addition, the devices operating in the depletion-modes also show high field-effect mobilities in the order of to cm2V-' s-' (see Fig. 17(b)).

F-I poly(3-dkylthiophene

Figure 16. Schottky gated poly(3-alkylthiophene) field-effect transistors fabricated by Yoshino et al. [85].

9.5 FETs based on polythiophenes

479

Drain-source voltage (V) Figure 17. The current-voltage characteristics of a FET prepared by Bao et at. with a channel length of 12pm operated in the accumulation mode (a) and depletion mode (b) at different gate voltages 1841.Reprinted with permission. Copyright 1996 American Institute of Physics.

The high mobilities in regioregular P3HT may be related to better ordering in these films [84]. A very strong, sharp diffraction peak at 5.4"is observed for all samples in the reflection X-rFy geometry (see Fig. 18), corresponding to an intermolecular spacing of 16.36A of the well-organized lamellar structure. However, the major peak visible under electron diffraction is around 3.7-3.81 A, which corresponds to the distance of thiophene rings in their stacks formed between adjacent chains (see Fig. 18 inset). This indicates preferred orientation, such that the hexyl side chains may be close to normal to the substrate and the backbone is essentially parallel to the substrate (Fig. 19). Such preferred orientation might account for the relatively high mobilities in these polythiophenes, since it would place the transport direction (i.e. that between thienyl rings) parallel to the substrate. The performance of these regioregular poly(3-alkylthiophene) devices varies significantly with solvents used for film preparations (Table 2). It might be due to the difference in film forming quality when different solvents are used. It was shown that when THF was the solvent, P3HT precipitated out during solvent evaporation, resulting in a non-uniform and discontinuous film. In the case of other solvents, different degrees of film discontinuity occur depending on the solubility of the polymer and nature of the solvent. Devices have been prepared by both casting and spin-coating and the off-currents of cast devices usually tend to be higher than those made from spin-coating, probably due to the higher thickness of the former [84]. However, the mobilities obtained

480

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophenefield effect

~

3"

10" 20" Diffraction Angle 20

30'

Figure 18. Transmission electron diffraction pattern obtained by Bao et al. of a thin film of regioregular goly(3-hexylthiophene) east from chloroform solution. The strong diffraction at around 3.7-3.8 A corresponds to the distance of thiophene rings in their stacks formed between adjacent chains [84].Reprinted with permission. Copyright 1996. American Institute of Physics.

from cast films are normally higher, possibly because slow evaporation of solvent enables slower growth of films and therefore allows ordering. Two methods have been shown useful in lowering the off-current while keeping the field-effect mobility almost unchanged [84]. In the first method, the films were treated with ammonia by bubbling N2 through ammonium hydroxide aqueous solution (see Table 2 entries 7 and 14). It was also found that thermal treatments such as heating the samples under N2 at 100°C for 5min can lower off-currents (see Table 2 entry 11). However, heating to 150°Clowers the mobility dramatically. Transistors made from regioregular poly(3-alkylthiophene)~with octyl (POT) and dodecyl (PDT) substituents have also been fabricated and their electrical characteristics have been studied [89]. POT has a similar field-effect mobility as PHT, while PDT has a much lower mobility of the order of lop6cm2V-' s-*: Both PHT and PDT have very strong, sharp diffraction peaks at 5.4" 28 (16.36A) and

interchain hopping

FFigure 19. Possible conformation of regioregular PHT in cast films.

48 1

9.5 FETs based on polythiophenes

Table 2. Field-effect mobilities and on/off ratios of regioregular PHT transistors preapred from different conditions. Condition 1, cast, vacuum pumped for 24 hours; Condition 2, spin-coated; Condition 3, treated with NH3 for 10 h; Condition 4, heated to 100°C under N2 for 5 min; Condition 5, heated to 150°C under N. for 35 min [84]. Reprinted with permission from Ref. 84. Copyright 1996. American Institute of Physics. ~~~

~

Entry ~~~

Solvent

Condition

Mobility (cm2V-' s-'la

1 1 2

6.2 x 1 0 - ~ 1.9 10-~ 1.9 x lo-' 3.6 1 0 - ~ 3.2 1 0 - ~ 4.7 1 0 - ~ 4.7 x 6.9 x 1 0 - ~ 6.8 x 2.4 x 1.4 x lo-* 3.3 x 9.2 1 0 - ~ 4.5 x 2.1 x

On/off ratiob

~

1

2 3 4 5 6 7 8 9 10

11 12 13 14 15

THE p-xylene toluene chlorobenzene 1,1,2,2-tetrachloroethylene 1,1,2,2-tetrachloroethane chloroform

1

2 1 entry 6 condition 3 2 1

1 entry 10 condition 4 entry 11 condition 5 2 1 entry 14 condition 3

10

40 2 10

25 10

80 72 35 6 35 15 80 340 9000

Field-effect mobility for the accumulation mode operation. bOn/off ratio is calculated for enhancement mode operation only and it is ten times higher for enhancement-depletion operation.

a

4.4" 28 (20.10A), respectively. These spacings correspond to the chain distances of a well-organized molecular laye: structure. PDT showed a much weaker diffraction peak at 3.2" 28 (27.10 A), indicating lower crystallinity or orientation. In addition, its higher volume fraction of insulating side chains may also contribute to its low field-effect mobility.

9.5.3 All-printed plastic FETs The soluble high field-effect mobility regioregular poly(3-alkylthiophene)~allow them to be processed by printing techniques. Screen printing is a simple and environment-friendly way to produce electronic circuitry and interconnections [90]. In this method, patterns are generated by using a 'doctor blade' to squeeze ink through predefined screen masks. It is a purely additive method in which an ink is added where needed. Therefore, patterns can be formed in a single step. With a pitch of printed lines as fine as 250 pm, the printing process can significantly reduce the time and cost associated with photolithography. The first printed transistor was demonstrated by Garnier et al. [l] In these transistors, however, only the gate electrode and a pair of drain and source electrodes, respectively, were printed separately on each side of a sheet of polyester film (1.5 pm thick) which acts as the dielectric layer. This film with electrodes was then taped to a

482

9 Oligo- and Polythiophene Field EfSect Transistors Oligo- and polythiophene jield ejgect

electrodes

Figure 20. Structure of an all-printed plastic transistors by Bao et al. G: gate; S : source; D: drain [89]. Reprinted with permission. Copyright 1997. American Chemical Society.

plastic substrate followed by vacuum deposition of an organic semiconductor layer of insoluble dihexyl-a-hexathienylene (DH-a-6T). For practical applications, it is desirable that all the necessary components may be printed in a continuous process. Therefore, liquid-phase processable organic semiconductors need to be used so that low-cost large area electronics with flexible plastic substrates for display or data storage can be realized by using printing techniques [91]. Recently, Bao et al. have demonstrated the first high performance plastic transistor in which all the essential components are printed directly onto plastic substrates [89]. A scheme of the printed plastic transistor is shown in Fig. 20. An ITO-coated poly(ethy1ene terephthalate) film is used as the plastic substrate in which the IT0 layer acts as the gate electrode. A polyimide layer is printed through a screen mask onto the I T 0 surface. An organic semiconductor layer consisting of regioregular poly(3-hexylthiophene)~is then put down by spin-coating, casting, or printing using chloroform as the solvent. Finally, the device is completed by printing the drain and source electrodes using a conductive ink through another screen mask. By using this procedure, many devices with different shapes or geometries can be easily obtained in large quantities simply by printing through suitable screen masks.

V,= 8 x

-50

V

6x

-40

V

-30

V

-20

v v

Ee

-10

ov 0

-20 -30 -40 Drain-source voltage (V)

-10

-50

Figure 21. The current voltage characteristics of a printed plastic transistor operated in the accumulation mode at different gate voltages.

9.6 Heterojunction FETs

483

Figure 21 shows the transistor characteristics of a typical printed device fabricated with regioregular PHT [89]. All the transistors are p-channel devices and can operate both in accumulation-mode and depletion-mode. The field-effect mobility was found to be between 0.01 and 0.03 cm2V-' s-*. This is one of the highest values achieved for polymer FETs. It is comparable to the results obtained for regioregular poly(3-hexylthiophene) by using a Si substrate and Si02 as the dielectric layer with lithographically defined electrodes [84]. The field-effect mobilities of printed plastic transistors are slightly higher (about Torr) two times) when measured in air compared to under vacuum (lop2or [89]. Spin-coated films tend to have lower mobility than cast or printed films possibly because the latter films have better ordering resulting from slower solvent evaporation and consequent slower crystal growth. The feature size which can be obtained by screen-printing is still relatively large in the range of 75-100pm. Other printing techniques, such as inkjet printing [92, 931 and microcontact printing [94], are possible alternative methods to pattern smaller transistors. Rogers et al. have fabricated regioregular poly(3-hexylthiophene)-based transistors with channel length as small as 25 pm using microcontact printing to form devices similar to those shown in Fig. 20 [95].

9.6 Heterojunction FETs Heterojunctions are widely used in a variety of organic devices including light-emitting diodes and photodiodes [96]. Heterojunctions can enhance the functionality of an organic FET, and an example of the use of heterojunctions in FETs is described by Dodabalapur et al. [97, 981. The FETs in refs. 98 and 97 possess two active materials, including one oligothiophene. The materials were chosen so as to possess different charge transport properties and energy levels. The schematic layer structure of the device is similar to the configuration illustrated in Fig. 1 except that the organic semiconductor consists of two active layers. The first (adjacent to the gate dielectric) active layer is made up of a-6T, and is typically 10 to 20nm thick. The second active material (chosen for n-channel operation) is c60 and is about 20 to 40nm thick. In many cases, a third electrically inactive organic layer such a-6T was deposited on top of c60 to protect it from the ambient. The two active materials were chosen not only because good p-channel and n-channel transistors have been demonstrated with them but also because of the favorable energy levels of HOMOS and LUMOs. These energy of the two materials are such that when the gate is biased negatively with respect to the source, the p-channel material (a-6T) is filled with holes and when the gate is biased positively, the n-channel material (c6,)) is filled with electrons. The energy band diagrams for the two modes of operation are shown in Fig. 22. Hence, with these two active materials, the same transistor can be used as either an n-channel or a p-channel device as shown in Fig. 23.

AuFrAuE,i

484

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophene field effect

p-channel enhancement

n-channel enhancement

SiQ,

Figure 22. Schematicenergy level alignment for p-channel and n-channel operation of a heterojunction FET with a-6T and C60active layers. Reprinted with permission from Ref. 97. Copyright AAAS.

-

The characteristics of the device under p-channel operation are similar to those of TFTs with only a-6T active layers. The field-effect mobility is 4 x cm2V-' s-' and the threshold voltage is -0 V. The devices under n-channel operation have slightly different characteristics. The threshold voltage is significantly larger (+40 V) due to traps in the c60. For small gate voltages there is what appears to be a large leakage component to the source-drain current (shown in Fig. 23 by dotted lines), which disappears as the gate voltage increases beyond the threshold voltage. This 'leakage current' is actually a hole current which flows in the a-6T layer under certain bias conditions. For positive VDs and VGsand with VDs > VGs,an accumulation layer of holes is formed near the drain in a-6T and an accumulation layer of electrons is formed near the source in c60. The reason this happens is that for VDs > V G ~the , gate is effectively negatively biased with respect to the drain at/near the drain contact leading to an accumulation of holes. The charge density profile (for electrons and holes) becomes very complicated. The power and versatility of numerical modeling is particularly helpful in understanding the operation of such FETs. The calculated charge density profile (for electrons) is shown in Fig. 24. The electron density varies along the channel, being highest near the source. The density also varies along the direction perpendicular to the plane of the gate insulator-a-6T interface. The modeling of such FETs has been discussed in detail by Alam et al. in Ref. 23 . The ability of a FET to operate in either the n-channel or thep-channel mode would simplify the construction of complementary circuits (which require both n-channel and p-channel FETs). Complementary circuits dissipate much less power and have higher noise margins compared to circuits with only p-channel FETs or only n-channel FETs.

9.7 Summary

3.3

-10 -4

3

-a

I c

5

-6

-27

F

5 *

Err 0

ii

0

s - 4

e

485

-2.3 -2

-2

-1.7 -1.3

n

"0

t

0

-10

-20 -30 -40 Drain Voltage (V)

-50

7 3

2.6

s

.r

0

2.3

e

G

9

0 2

"0

10

20 30 Drain Wtage M

40

1.7 13

50

Figure 23. Current-voltage characteristics of an a-6T/Csoheterojunction FET operating as (a) a pchannel FET with negative VGs and VDs, and (b) operating as an n-channel FET with positive VGS and V D S . In the n-channel mode of operation there are hole currents flowing in the a-6T layer for V D ~ >VGs and are shown by dotted lines. Reprinted with permission from Ref. 97. Copyright 1995 M A S .

9.7 Summary In this review we have covered the important characteristics of oligothiophene and polythiophene field-effect transistors. The physics of such transistors was explained and the utility of analytical and numerical modeling in understanding the operation of a FET was emphasized. The nature of charge accumulation, potential distribution, short-channel and interface effects were examined. The importance of the synthetic route and purification was highlighted with particular reference to asexithiophene, a representative oligomer. Film morphology plays an important if not crucial role in determining the magnitude of the field-effect mobility and the highest mobilities are obtained when the molecular ordering and the intergrain coupling are good. The family of oligothiophene-based materials used as active materials in FETs includes many end-substituted and fused ring compounds. Solution-based deposition of organic semiconductors is being eagerly pursued on

486

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophene field effect

1o'*

7-

5

1

'

/

.S 1014 -

/

E

/

c

c

/

/

/'

/'

/*

/

/

/#

i

C C

-

5 nm

i

a-6T

1 nm

Oxide

i #i

lo8

'

0.0

1.o

2.0

3.0

4.0

position (pm)

Figure 24. Electron concentration in Cm in an a-6T/Gj0 heterojunction FET. The legends refer to the distance from the a-6T/Cm inerface. Reprinted with permission from Ref. 23. Copyright 1997 IEEE.

account of perceived cost advantages of such fabrication methods. Some soluble end-substituted oligomers and polythiophenes with side chains have been very successfully used in FETs. Regioregular polythiophene FETs have achieved mobilities close to that of the best vacuum sublimed oligothiophene FETs due to the excellent ordering that is possible. The choice of solvent and deposition conditions are important in determining the film morphology and hence the mobility of FETs. Regioregular polythiophene has been used as the active semiconducting material in the first transistor in which all key layers are formed by printing, resulting in mobilities > cm2V-' s-l. Oligothiophenes have also been incorporated into multi-layer heterojunction FETs which can function as both n-channel and p-channel transistors.

Acknowledgments The authors thank A. Alam, J. Laquindanum, R. A. Laudise, A. J. Lovinger, T. Siegrist, and L. Torsi for helpful assistance.

References

487

References 1. F. Garnier, R. Hajloui, A. Yassar and P. Srivatsava, Science, 1994, 265, 684. 2. Z. Bao, Y. Feng, A. Dodabalapur, V. R. Raju and A. J. Lovinger, Chem. Mat., 1997,9,1299. 3. A. J. Lovinger and L. J. Rothberg, J . Muter. Res., 1996, 11, 1581. 4. H. E. Katz, J . Muter. Chem., 1997, 7, 369. 5. F. Ebisawa, T. Kurokawa and S. Nara, J . Appl. Phys., 1983,54, 3255. 6. H. Koezuka, A. Tsumura and T. Ando, Synth. Met., 1987,18, 699. 7. M. Assadi, C. Svensson, M. Wilander and 0. Inganas, Appl. Phys. Lett., 1988,53, 195. 8. J. H. Burroughes, C. A. Jones and R. H. Friend, Nature, 1988,335, 137. 9. Horowitz, D. Fichou and F. Garnier, Solid State Comm., 1989, 70, 385. 10. G. Garnier, G. Horowitz, X. Z. Peng and D. Fichou, Adv. Mat., 1990, 2, 592. 11. F. Horowitz, X. Z. Peng, D. Fichou and F. Garnier, J . Appl. Phys., 1990,67, 528. 12. G. Kastner, J. Paloheimo and H. Kuzmany, in Springer Series in Solid State Sciences (Eds. H. Kuzmany, M. Mehring and J. Fjink), Springer-Verlag, Berlin, 1993, 113, 512. 13. R. C. Haddon, A. S. Perel, R. C. Morris, T. T. M. Palstra, A. F. Hebard and R. M. Fleming, Appl. Phys. Lett., 1995, 67, 121. 14. Y.-Y.Lin, D. J. Gundlach, S. F. Nelson and T. N. Jackson, ZEEE Trans. Elec. Dev., 1997,44, 1325. 15. J. G. Laquindanum, H. E. Katz, A. Dodabalapur and A. J. Lovinger, J. Am. Chem. SOC.,1996, 118, 11 331. 16. P. Ostoja et al., Adv. Mat. for Opt. and Elec., 1992, I, 127. 17. W. A. Schoonveld et al., Synth. Met., 1997, 84, 583. 18. A. Dodabalapur, J. Laquindanum, H. E. Katz and Z. Bao, Appl. Phys. Lett.,, 1996, 69, 4227. 19. A. R. Brown, A. Pomp, C . M. Hart and D. M. deleeuw, Science, 1995, 270,972. 20. L. Torsi, A. Dodabalapur and H. E. Katz, J . Appl. Phys., 1995, 78, 1088. 21. G. Horowitz and P. Delannoy, J . Appl. Phys., 1991, 70,469. 22. M. Shur, M. Hack and J. G. Shaw, J . Appl. Phys., 1989, 66. 23. M. AIarn, A. Dodabalapur and M. Pinto, IEEE Trans. Elec. Dev., 1997, 44, 1332. 24. S. M. Sze, Physics of Semiconductor Devices, John Wiley and Sons, New York, 1981, p. 50. 25. A. Dodabalapur, L. Torsi and H. E. Katz, Science, 1995, 268, 270. 26. H. E. Katz, L. Torsi and A. Dodabalapur, Chem. Mat., 1995, 7, 2235. 27. Z. Bao, A. J. Lovinger and A. Dodabalapur, Adv. Mat., 1997, 9,42. 28. Z. Bao, A. J. Lovinger and J. Brown, J . Am. Chem Soc., in press. 29. D. Fichou, G . G. Horowitz and F. Garnier, Eur. Pat. Appl., EP 402,269, 1990; FR Appl. 89/7,610, 1989. 30. M. S. A. Abdou, X. Lu, Z. W. Xie, F. Orfino, M. J. Deen and S. Holdcroft, Chem. Muter., 1995, 7, 631. 31. J. Kagan and S. K. Arora, Heterocycles, 1983, 20(10), 1937. 32. D. Fichou, M-P. Teulade-Fichou, G. Horowitz and F. Demanze, Adv. Muter., 1997, 9, 5. 33. R. Wu, J. S. Schumm, D. L. Pearson and J. M. Tour, J . Org. Chem., 1996, 61, 6906. 34. Y. Wei, Y. Yang and J-M. Yeh, Chem. Muter., 1996, 8, 2659. 35. A. Merz and F. Ellinger, Synthesis, 1991, 462. 36. Y. Y. Lin, D. J. Gundlach and T. N. Jackson, Materials Research Society SymposiumProceedings Vol. 413, 1995, 413. 37. S. Hotta and K. Waragai, Adv. Muter., 1993, 5, 12, 896. 38. G. Horowitz, B. Bachet, A. Yassar, P. Lang, F. Demanze, J-L. Fave and F. Garnier, Chem. Mater., 1995, 7, 1337. 39. T. Siegrist, R. M. Fleming, R. C. Haddon et al., J. Muter. Res., 1995, 10, 2170. 40. A. J. Lovinger, D. D. Davis, A. Dodabalapur, H. E. Katz and L. Torsi, Macromolecules, 1996, 29,4952-4957. 41. D. Fichou, B. Bachet, F. Demanze, I. Billy, G. Horowitz and F. Garnier, Adv. Muter., 1996, 8, 500.

488

9 Oligo- and Polythiophene Field Effect Transistors Oligo- and polythiophene Jield effect

42. A. J. Lovinger, D. D. Davis, A. Dodabalapur and H. E. Katz, Chem. Muter., 1996, 8, 836. 43. W. Porzio, S. Destri, M. Mascherpa, S. Rossini, S . Briickner, Synthetic Metals, 1993, 55-57, 408-413. 44. S. Destri, M. Mascherpa and W. Porzio, Adv. Muter., 1993, 5 , 43. 45. S. Hotta and K. Waragai, J. Muter. Chem., 1991, I , 835-842. 46. K. Uchiyarna, H. Akimichi, S. Hotta, H. Noge and H. Sakaki, Synthetic Metals, 1994, 63, 57-59. 47. K. Hamano. T. Kurata, S. Kubota and H. Koezuka, Jpn. J . Appl. Phys., 1994, 33, L1031-Ll034. 48. T. Kurata, K. Hamano, S. Kubota and H. Koezuka, Organic Thin Films for Photonic Applications, 1993, Toronto. 49. B. Servet, G. Horowitz, S . Ries et al., Chem. Muter., 1994, 6, 1809-1815. 50. P. Lang, G. Horowitz, P. Valat, F. Gamier, J. C. Wittmann and B. Lotz, J. Phys. Chem., 1997, 101, 8204. 51. C. Y. Yang, Y. Yang and S. Hotta, Synthetic Metals, 1995, 69, 303-304. 52. D. Oelkrug, H-J. Egelhaaf and J. Haiber, Thin Solid Films, 1996, 284-285, 267-270. 53. A. J. Lovinger, D. D. Davis, R. Ruel, L. Torsi, A. Dodabalapur and H. E. Katz, J. Muter., 1995, 10, 2958. 54. L. Torsi, A. Dodabalapur, A. J. Lovinger et al., Chem. Muter., 1995, 7, 2247-2251. 55. H. Akimichi, K. Waragi, S. Hotta, H. Kano and H. Sakaki, Appl. Phys. Lett., 1991,8, 1500. 56. H. E. Katz, J. G. Laquindanum and A. J. Lovinger, Chem. Muter., 1998, accepted for publication. c7 2 1 . F. Garnier, presented at the Materials Research Society Spring Meeting, 1997, San Francisco, CA. 58. F. Garnier, A. Yassar, R. Hajlaoui et al., J. A m . Chem. Soc., 1993, 15, 8716. 59. H. E. Katz, A. Dodabalapur, L. Torsi and D. Elder, Chem. Muter., 1995, 7, 2238-2240. 60. H. E. Katz, A. J. Lovinger and J. G. Laquindanum, Chem. Muter., 1998, accepted for publication. 61. J-H. Liao, M. Benz, E. LeGoff and M. G. Kanatzidis, Adv. Muter., 1994,6(2), 135. 62. J. K. Herrema, J. Wildeman, F. van Bolhuis and G. Hadziioannou, Synthetic Metals, 1993, 60,239-248. 63. A. Yassar, F. Gamier, F. Deloffre, G. Horowitz and L. Ricard, Adv. Muter., 1994, 6, 60. 64. P. Bauerle, Adv. Muter., 1992, 4, 102. 65. H. Muguruma, T. Saito, A. Hiratsuka, I. Karube and S. Hotta, Langmair, 1996, 12, 5451. 66. R. Hajlaoui, D. Fichou, G. Horowitz, B. Nessakh, M. Constant and F. Garnier, Adv. Muter., 1997, 9, 557. 67. J. G. Laquindanum, H. E. Katz, A. J. Lovinger and A. Dodabalapur, Adv. Muter., 1997, 9, 36. 68. J. G. Laquindanum, H. E. Katz and A. I. Lovinger, J. Am. Chem. Soc., 1998, accepted for publication. 69. A. B. Holmes, Proceedings of 4th European Conference on Molecular Electronics, September, 1997, Cambridge, United Kingdom. 70. C. D. Dimitrakopoulos, A. Afzali-Ardakani, B. Furman, J. Kymissis and S. Purushothaman, Synthetic Metals, 1997,89, 193-197. 71. G. Horowitz, F. Gamier, A. Yassar, R. Hajlaoui and F. Kouki, Adv. Muter., 1996,8, 52. 72. B. Batlogg, presented at the Materials Research Society Full Meeting, 1997, Boston, MA. 73. J. G. Laquindanum, H. E. Katz, A. J. Lovingerand A. Dodabalapur, Chem. Muter., 1997,8, 2542-2544. 74. C. P. Jarrett, A. R. Brown, R. H. Friend et al., Synthetic Metals, 1997, 85, 1403-1404. 75. K. Waragai, H. Akimichi, S. Hotta, H. Kano and H. Sakaki, Synthetic Metals, 1993,55-57, 4053-4058. 76. R. Hajlaoui, G. Horowitz, F. Garnier et ul., Adv. Muter., 1997, 9, 389. 77. A. Tsumura, H. Koozuka and T. Ando, Appl. Phys. Lett., 1986,49, 1210-1212. 7 8 . D. M. Taylor, H. L. Gomes, A. E. Underhill, S. Edge and P. I. Clemenson, J. Phys. D: Appl. Phys., 1991,24,2032-2038.

References

79. 80. 81. 82. 83. 84. 85.

86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98.

489

C.-T. Kou and T.-R. Liou, Syn. Met., 1996,82, 167-173. H. Fuchigami, A. Tsumura and H. Koezuka, Appl. Phys. Lett., 1993,63, 1372-1374. A. R. Brown, D. M. de Leeuw, E. E. Havinga and A. Pomp, Syn. Met., 1994,68, 650. C. P. Jarrett, R. H. Friend, A. R. Brown and D. M. de Leeuw, J . Appl. Phys., 1995, 7, 6289-6294. Z. Bao, A .J. Lovinger and A. Dodabalapur, unpublished results. Z. Bao, A. Dodabalapur and A. J. Lovinger, Appl. Phys. Lett., 1996,69,4108-4110. K. Yoshino, H. Takahashi, K. Muro, Y. Ohmori and R. Sugimoto, J. Appl. Phys., 1991, 70, 5035. Y. Ohmori, K. Muro, M. Uchida, T. Kawai and K. Yoshino, Jpn. J. Appl. Phys., 1991, 30, L610. J. Paloheimo, P. Kuivalainen, H. Stubb, E. Vuorimaa and P. Y. Lahti, Appl. Phys. Lett., 1990, 56, 157. Chapter I, I1 and V of this book Z. Bao, Y. Feng, A. Dodabalapur, V. R. Raju and A. J. Lovinger, Chem Mater., 1997, 9, 1299- 1301. K. Gilleo, Polymer Thick Films, Van Nostrand Reinhold, New York, 1996. R. F. Service, Science, 1996, 273, 879. C. C. Wu and J. Sturm, presented the Electronic Materials Conference, June, 1997, Ft. Collins, CO. X. D. Xiang and P. G. Shultz, Adv. Mater., 1997, 9, 1046. Y . Xia and G. M. Whitesides, Angew. Chem., in press. J. Rogers, Z. Bao and V. R. Raju, Appl. Phys. Lett., submitted. N. C. Greenham and R. H. Friend in Solid State Physics (Eds. H. Ehrenreich and F. Spaepen), 1995, Vol. 49, p. 2. A. Dodabalapur, H. E. Katz, L. Torsi and R. C. Haddon, Science, 1995, 269, 1560. A. Dodabalapur, H. E. Katz, L. Torsi and R. C. Haddon, Appl. Phys. Lett.. 1996,8, 1108.

This Page Intentionally Left Blank

10 Application of Electrically Conductive Polythiophenes Gerhard Kossmehl and Gunnar Engelmann

10.1 Introduction The first electrically conductive organic material was described in 1948 independently by D. D. Eley [l] and A. T. Vartanyan [2]. They found that copper phthalo cyanine has electrical conductivity and photoconductivity higher than for typical organic materials. A short time later further organic compounds and polymeric materials like polyacetylene [3] and polyphthalocyanines [4-61 were prepared and studied in relation to their electrical properties and were suggested for various applications. Many donor-acceptor complexes with enhanced electrical conductivities [7] and polymers with tetracyano-p-quinodimethane (TCNQ) [8, 91 and other acceptors and donors have been widely studied. The concept of doping polymers with conjugated systems of 7r bonds [lo] made it possible to increase electrical conductivity up to values characteristic for metals, and was realized by the action of oxidants like iodine, bromine, arsenic pentafluoride or by reductants like alkali metals on polyacetylene. The next idea in the synthesis of electrically conducting polymers was the oxidative polymerization of heteroaromatic compounds like pyrrole and thiophene. Conjugated polymers with different structures became avail able, that can be doped (oxidized or reduced) to form materials described as polymeric metals with electrical conductivities up to 100 S cm-' and more. Also polyaniLine - well known for more than hundred years - is under research in relation to its interesting electrical properties. Most of the materials prepared in the past have interesting electrophysical properties but little practical application. The idea that interesting electrical properties can be combined with the applicable properties of polymeric materials cannot be realized for most of these materials. Nearly all polymers with conjugated 7r electron systems and therefore stiff structures which are available by chemical synthesis as insoluble and infusible powders, cannot be processed themselves as electrically conductive materials for technical purposes (polyacetylene, polyaniline, polypyrrole, polyphthalocyanines, polythiophene). In addition, some of them are not stable in air (polyacetylene, polypyrrole) and can only find applications in vacuo. Some of them (donor-acceptor complexes containing halogens, and polymeric TCNQ complexes at elevated temperatures) are not stable because the complexforming components evaporate. The highly conducting low molecular weight and polymeric TCNQ complexes have the disadvantage that they generally form very fine and long needles instead of compact solid state materials. If these materials are highly electrically conductive and stable in relation to their properties they can be processed in the form of powders, as fillers for commercial polymers like

492

10 Application of Electrically Conductive Polythiophenes

polystyrene, polyesters or polycarbonates in order to realize a distinct electrical conductivity as well as dielectric behavior, to prevent electrostatic charges of plastic materials or as components for shielding. The best condition for broad applications are soluble or fusible electrically conductive polymeric materials that can be processed from solution or from the melt in order to form compact electrically conductive parts. Thin layers - possible by casting from solution - can also be prepared electrochemically on surfaces that can be handled as electrodes in order to create electrically conductive coatings. Generally highly electrically conductive materials arising from polythiophene are available from the appropriate polythiophene with processable properties after oxidation with an oxidant in order to produce the oxidized form (p-type doped form) or after reduction with a reductant to produce the reduced form (n-type doped form) of the polymer in the case of chemical preparation. In the case of electrochemical formation of polythiophene or its derivatives starting from thiophene or suitable derivatives automatically the oxidized, neutral or reduced form of the electrically conductive material can be built up. In 1991 the first electrically conductive material, poly(3,4-ethylenedioxythiophene2,5-diyl) itself or blends of this material and polystyrenesulfonic acid were used to produce transparent, abrasion-resistant, non-corrosive coatings for photographic films and other materials with controlled antistatic properties [l 11; they came onto the market and can now be used on a commercial scale (see section 10.3). Many applications for polythiophene have been claimed for different purposes. They are described in scientific papers or patents; some are in use for special technical applications or will come on the market in the near future. The multiple potential technological applications of polythiophenes can be divided into three main groups: - the electrical conductivity of the doped conducting state; - the electronic properties of the neutral semiconducting state; - the electrochemical reversibility of the changes between the oxidized or reduced (doped) and the neutral (undoped) state.

10.2 Conducting materials Many products simply arising from thiophene or its derivatives have been claimed to be useful as electrically conducting materials. In this section a short overview will be given about these materials, their structures, their preparation and the values of their electrical conductivity. It is essential, that such polymers are oxidized (p-type doped) or reduced (n-type doped) for enhanced electrical conductivity (for reviews see [12-141. In all cases the electrical conductivity of the neutral form is about lop6 to 1OPs cm-' . Generally, polythiophene (PT) itself and most of its derivatives are available as insoluble and unmeltable powders by oxidative polymerization because of crosslinking. Some alkylated PTs are interesting because they can be prepared as soluble materials, that can be processed from solution to form films. Also water-soluble PTs

10.2 Conducting materials

493

have been described [15]. Such films, and especially free-standing films can also be prepared electrochemically. Among the large number of oxidants (dopants) for the polymerization of thiophene or its derivatives the most effective are ferric chloride and iodine. Iodine has the disadvantage that it partly evaporates from the electrically conductive material under environmental conditions, changing the electrical properties. Powderous PT can be doped in the gas phase by iodine, reaching values of the electrical conductivity of 6 S cm-' [16], 8 S cm-' [17]; alkylated PTs have electrical conductivities only as high as 0.2 S cm-' [18]. Alkylated PTs cast as films from solution and then doped by iodine have values of the electrical conductivity up to 12 S cm-' because of a higher solid state order than in powderous precipitates [19]. The highly regular structure of alkylated PTs is very important; a head-to-tail structure in the polymer main chain is responsible for a value of 1000S cm-'after doping with iodine [20]. The synthetic route to form PT and its derivatives plays an important role. A special polycondensation reaction by dehalogenation with zero-valent nickel complexes yields samples of PT after FeC13 treatment with 0.5 S cm-' and after iodine doping with 8 Scm-' [16, 17,21,22]. 3-MethylPT with 170Scm-' can be prepared electrochemically by an oxidation potential controlled technique [23]. 3-ButylPT electrochemically produced from 3-butylthiophene with tetrabutylammonium perchlorate as electrolyte salt at a potential of 1.O V has an electrical conductivity of 8 to 10 Scm-'; overoxidation at 1.6V gives materials with 0.5 to 1 Scm-' [24]. PT prepared by the oxidant and dopant cupric perchlorate starting from bithiophene yields PT with values of the electrical conductivity up to 4.5 S cm-' [25]. FeC13 is a good dopant for PT but not the best oxidant for the preparation of PT from thiophene: the doped PT prepared from thiophene with an excess of FeC13 has only 0.5 Scm-' [17]. 3-Alkylated PT films (with decyl substitution) have an electrical conductivity of up to 5 S cm-' [26]; 3P-dibutoxyPT doped with FeC13 reaches only 1 Scm-' [27-301. Other 3-alkylated PT materials (octyl, decyl) doped with FeC13are claimed to be stable for 10 to 100 years (by extrapolation!) [31-321. Increasing molecular weight (22 000- 130 000) in the case of 3-octylPT raises the electrical conductivity from 1.9 to 10.4 Scm-' [33]. Electrical conductivities between 7 and 30 S cm-' are available for 3-alkylated PTs with different long side chains by electrochemical preparation [34]; the longer the side chain is (hexyl, octyl, dodecyl, octadecyl) the lower is the electrical conductivity (dilution effect). Very important is a complete head-to-tail/head-to-tail coupling during the synthesis of 3-alkylated thiophene derivatives forming 3-alkylPTs as films with highly ordered structures. With increasing side chain length, an increasing electrical conductivity has been found for dodecyl 1OOOScm-', for octyl 200 S cm-' and for hexyl60 S cm-' [20]. Stretching of films (3-octy1PT) increases the electrical conductivity from 5 to 20 S cm-' for materials doped with FeC13, and up to 180 S cm-' for such materials doped with FeC13.6H20[35,36]. The electrical conductivity of iodine doped PT films increases along the drawing direction and decreases perpendicular to the drawing direction [37]. 1

494

10 Application of Electrically Conductive Polythiophenes

Electrical conductivities up to 6600 S cm-' have been created for fibrils in 3-methylPT, in which the polymeric chains are preferentially oriented parallel to the axes of these fibrils; this material shows higher electrical conductivities than conventional 3-methylPT [38, 391. Sodium poly(3-thienylethanesulfonate) and (-butanesulfonate) have been prepared electrochemically via the methyl sulfonates; the watersoluble polymers have been cast to freestanding films with electrical conductivities of up to 1OScm-' (in the oxidized form) and 3 to 8 orders of magnitude less in the neutral form [15]. These polymers are claimed as 'fast selfdoping materials' because of the proton spending character of the sulfonic acid functional group. Copolymers prepared from 3-octyl- and 3-methylthiophene and then doped with iodine have electrical conductivities up to 26 S cm-' (in the case of 25 mol% of 3-methylthiophene) [40]. Electrical conductivities as high as 40 S cm-' can be realized for a reaction product of poly(2,2'-bithiopheneylmethyl methacrylate) arising by electrochemical and chemical oxidative polymerization of the side chain thiophene systems resulting in cross-links between bithiophene units [41]. A unique electrically conductive material, distinguished by a long term stability and fast switchable electrophysical properties, is poly(3,4-ethylenedioxythiophene) with specific electrical conductivities up to 30 S cm-' and more. It can be used for antistatic coatings (see section 10.3), as electrode material for solid state capacitors (see section 10.5), for electrochromic devices (see section 10.8) and as IR absorber [ll]. Blends of 3-octylPT and poly(pheny1ene oxide) have electrical conductivities up to 3 S cm-' [42]. Shaped articles, e.g. rods, fibers and films, which are electrically conductive, can be produced from composite materials containing a nonconductive flexible chaincarrier polymer and an electrically conductive polymer [e.g. 3-octylPT, 3-dodecylPT or poly(2,5-thiophenediyl vinylene)] [43, 441. Electrically conductive fibers can manufactured from polyesters blended with PT [45]. Electrically conductive resin adhesives, that are heat, stock and moisture resistant, contain bifunctional epoxy methacrylates as adhesives and solders together with conductive polymers as filler materials (e.g. 3-dodecylPT, 3-octadecylPT) [46, 471. Composites having electrical conductivities up to about 5 S cm-' can be synthesized starting from porous cross-linked polystyrene imbibed with a bithiophene solution; after partial drying of the saturated host polymer imbibing with a FeC13 solution PT is produced by oxidative polymerization, and the PT network is additionally doped by FeC13 [25,481. An interesting electrically conductive material is poly(2,5-thiophenediyl vinylene) (PTV) because of its processability via a precursor route. Iodine doped material reaches 315 S cm-' [49]. PTV fibers possess values of the electrical conductivity up to 2000 S cm-' with increasing draw rate [50, 51, 521. Free-standing films of PTV with high electrical conductivities after doping with iodine (315 S cm-') and FeC13 (110 Scm-') can be synthesized via a precursor polymer bearing hydroxy and pyridinium substituents [49]. Such films can also be prepared via an alkyloxy substituted precursor polymer; after doping with iodine an electrical conductivity of

10.3 Antistatic coatings

495

100 S cm-' can be reached and increased up to 1100S cm-' by stretching these films [51-57]. Via poly[2,5-thiophenediyl(1-methoxyethylene)] as a soluble precursor polymer, oriented PTV films with up to 1000 S cm-' can be prepared following two routes: the more common simultaneous tensile deformation and conversion method to PTV and - more effective - the sequential conversion and drawing method to form PTV [57]. A copolymer having 20 to 230 S cm-' can be synthesized electrochemically with the electrolyte salt tetraethylammonium perchlorate in propylene carbonate starting from thiophene and 1,2-di(2-thienyl)ethylene forming a polymer with 2,2'-bithiophenediyl vinylene units [58].

10.3 Antistatic coatings The thiophene polymer with the best combination of technological properties is poly(3,4-ethylenedioxy-2,5-thiophenediyl)(PEDT). The monomer 3,4-ethylenedioxythiophene (EDT) forms PEDT by oxidative polymerization, a polymer without branching and cross-linking. PEDT can be cast from solution forming electrically conductive films or coatings on the surface of many materials like plastics, metals, glass or ceramics. PEDT has an excellent long time stability in regard of its electrical conductivity. The oxidative polymerization made with FeC13 in boiling benzonitrile (188°C) for 2 h yields PEDT having an electrical conductivity of 19 Scm-'; reaction for 6 h yields a product with 31 Scm-' [59]. These electrically conductive polymeric materials are formed in the doped state, meaning in the oxized form bearing FeC1; anions as counterions for the thiophenium radical cations in the polymeric chain. Other monomers like thiophene or alkylated thiophenes form polymers with very low electrical conductivities under these conditions [59,601. PEDT is used for antistatic coatings. The 1997 production of this product was in the order of about 1000kg per annum. As a transparent polymer it substitutes the classical materials carbon black for compoundings, ionic compounds as layers on different materials or metallic layers on the surfaces of such materials with a very high electrical resistivity. When EDT is polymerized in the presence of polystyrenesulfonic acid, a material is produced that can be processed from aqueous solution and yields effective antistatic coatings on films and other surfaces [61]. The antistatic effect can be achieved with very small quantities of the electrically conductive PEDT; just a few mg of PEDT per m2 are necessary in order to obtain films with a surface resistivity of lo6 R/square, enough for an effective antistatic treatment. Ths electrically conductive polymeric material (Baytron P, Bayer AG) [62] is used for antistatic coatings of photographic films made from cellulose triacetate or polyethylene terephthalate (PET, having a surface resistivity of 10l2 cR/square) and also for X-ray films. The electrically conducting layer is situated between the polymeric film material and the light sensitive layer and prevents discharge sparks arising from static charge reducing the quality of the photographic material. PEDT substitutes and surpasses

496

10 Application of Electrically Conductive Polythiophenes

with its properties ionic antistatic materials like watersoluble sodium polystyrenesulfonate that can only act as antistatic material if the atmospheric humidity is higher than 50% and will be dissolved during the developing process. For easy handling in technical applications a colloidal system for coating film materials has been developed [l 11: oxidative polymerization of EDT with potassium peroxydisulfate in the presence of sodium polystyrenesulfonate gives a colloidal solution of PEDT-polystyrenesulfonate. Such solutions are stable up to a solid content of 2%. From such solutions the polymer is precipitated on the surface of the material requiring an antistatic surface. In this way surface coatings of parts of machines, which may have complex geometric structures, for clean room outfits and for packaging films of high value for high quality electronic devices can be easily produced. PEDT is also used in electronic technology (DMS-E process) for the production of electrically conductive layers of conductive plates. The first conductive layer in this process is prepared from EDT forming PEDT with a lower resistivity than usual in the present normally used technical process. This new process yields in electrically conductive layers of higher quality than by the other used technical processes in shorter time combined with easier handling [63]. PEDT has qualified itself especially as an excellent material that can be used for through-hole electroplating by electroless deposition. In detail the electrically conductive layer of the conductive plates is produced in the technological process as follows: (i) the conditioning step of the plastic board (the ground material, practically all are plastics) is treated with special polymeric detergents; (ii) the loading step with potassium permanganate forms a dense layer of manganese dioxide; (iii) the 'catalytic' step forms PEDT (<1 pm) from EDT (as an aqueous microemulsion from 1% EDT and polystyrenesulfonic acid) by oxidation of manganese dioxide in acidic solution while the oxidant manganese(1V)salt is reduced and dissolved quantitatively, and (iv) this dense electrically conductive layer is used technologically for the electrolytic formation of the copper layer. The transparent microemulsion employed is long-term stable and consists of organic particles of 5-100 nm [63]. Under special conditions it is possible to prepare PEDT films of more than 20 m length [64]. For such films suitable polyanions, e.g. sulfated poly(P-hydroxyether)s or poly(hydr0xyether)s bearing trifluoromethyl groups and sulfated poly(butadiene)s are necessary. The preparation can be achieved on cylindrical electrodes of a computer controlled system. The electrical conductivity of such PEDT materials with the above mentioned polyanions composites exceed 150 S cm-', and reaches as high as 400 S cm-' if the polyester with trifluoromethyl groups is used. Such films are used for the production of batteries (see section 10.5) [64].

10.4 Electromagnetic shielding materials Electrically conductive polymers can be used also as shielding materials against electromagnetic interference (EMI), if they are incorporated as homogeneous fillers

10.5 Muteriuls,for rechargeable butteries, capacitors

497

in composites of common plastic materials. On the other hand it is possible to measure the shielding efficiency against electromagnetic interference for the characterization of homogeneity and stability of the electrical conductivity of electrically conductive plastic materials [65]. Blends containing up to 20% 3-octylPT are claimed to be useful as shielding materials in electronic techniques owing to their moderate electrical conductivity and the unintentional ‘holes’ and ’slits’ in these materials consisting of insulating matrix polymer, e.g. polystyrene, poly(viny1 chloride) or poly(ethy1 vinyl acetate) and the mentioned PT derivatives 1651. The use of unblended 3-octylPT as a material of higher electrical conductivity improves the shielding effect [65]. Technical electromagnetic shields contain compositions arising from polyheteroaromatic substances, poly(bithi0phene) and polymeric sulfates [66], or from an electrically non-conductive matrix polymer and 3-octylPT [67]. Microwave shieldings are claimed to be built up from one or several layers of doped electrically conductive polymers like PT or poly(bithi0phene) and a second organic insulating polymer forming altogether an interpenetrating network [68]. 3-AlkylPTs have been claimed to have pronounced magnetic properties depending on the length of the alkyl chain. Extensive studies have been made with PTs bearing dodecyl and hexadecyl side chains. The anomaly in the magnetic properties is discussed within the scope of differences in the crystalline phases of the alkyl substituted polymers brought about by the different length of the alkyl side-chains [69]. Polymers also containing organosilicone moieties have been claimed to have typical ferromagnetic properties [70].

10.5 Materials for rechargeable batteries, capacitors The most widely discussed application of electrically conductive polymers, that can be switched reversibly between two redox states, is their use for the storage of electrical energy in rechargeable batteries. Initial enthusiasm based on optimistic ideas has been quenched by several not fully solvable circumstances such as self-discharge and unsuficient cycle stability [12,13]. A loss of 10% in 1000cycles for instance is an admirable scientific result, but death for a technical application. In a 1997 review, electrochemically active polymers - including also PT - for rechargeable batteries are given [71]. In this review the first two really produced batteries containing polymeric materials are described intensively with their special problems [72]: (i) the German Li/polypyrrole battery (Varta-BASF battery) [73751 and (ii) the Japanese Li/polyaniline battery [75]. Early publications describe rechargeable batteries made from iodine doped PT combined with zinc or lithium electrodes; for the construction of primary and secondary electrochemical cells PT samples has been named to be prepared chemically or leading to better performances - electrochemically [76-781. Further studies have confirmed high voltage and good energy density of PT in electrochemical cells, but on the other hand the self-dischargingproperties as a serious problem for the technical application of this material was also named [79-831. Battery assemblies with PT as

498

10 Application of Electrically Conductive Polythiophenes

cathode material combined to solid-state lithium cells containing poly(ethy1ene oxide) lithium perchlorate electrolytes working only beyond the phase transition temperature of the electrolyte of 70°C have also be claimed 1841. The p-doped state of PT has been found to be subject to relatively rapid selfdischarge in electrochemicalcells. Reactions with the electrolytic medium have been suggested, although not demonstrated, as the reason for self-discharge meaning spontaneous dedoping because of redox reactions with the electrolyte system [85]. Self-discharge was found to be negligible in solid state cells during storage at room temperature rather than they are operating at temperatures of 70°C; under these conditions it is thought that the electrolyte takes part in the process of the self-discharge step [86]. An improvement of performance has been obtained with the use of bithiophene as a starting material instead of thiophene for the synthesis of PT. An increase in the specific charge was observed and attributed to the higher regularity in the PT structure [87]. A poly(bithiophene)/lithium solid state cell with poly(ethyleneoxide)/ lithium perchlorate electrolyte was found to have a remarkable stability of 300 cycles at a working temperature of 70°C [88]; an initial activation phase was required during cycling, afterwards the coulombic efficiency was 98%. Also substituted PTs have been claimed to be useful for application as energy storing materials [89-9 11; electrochemical studies have shown the relationship between the polymer structure and the electrochemical behavior of these substituted PTs as materials for rechargeable batteries having lower oxidation potentials than unsubstituted PT. Rechargeable lithium/sulfur dioxide batteries with the Li(S02)3AlC14electrolyte contain normally porous carbon cathodes or alternatively electrically conducting polymers especially alkyl and alkyloxy substituted PTs like 3-methylPT as thin films with a high surface area. Such films are also qualified to have a high discharge capacity, the recharge potential has to be lower than 3.9 V, which should avoid the formation of corrosive chlorine from the electrolyte salt [92-941. A rechargeable lithium battery with a 3-methylPT cathode has been produced with lithium hexafluoroarsenate-dimethyl carbonate electrolyte [95]. 3-MethylPT for the use in rechargeable batteries can be produced by polymerization of 3-methylthiophene under controlled conditions to polymeric materials with structures sustaining a stable doping degree of up to 0.38 e- per molecular unit, corresponding to almost l00Ah per kg [96, 97, 981. Insoluble 3-methylPT with incorporated copper(I1)salt arising from CuC12 used during chemical polymerization as co-promoter pressed as pellets has been used as both anode and cathode of a very stable battery cell with 0.1 M aqueous sodium perchlorate solution with perchloric acid as supporting electrolyte at pH 1.5. The charge-discharge process has an efficiency of 0.3 1, but it is unknown if the charge capacity is within the copper redox system or within the PT redox system [99]. PT prepared chemically has been studied intensively also as cathode material in secondary batteries [100- 1031. Noteworthy is a battery fabricated by an electrochemically prepared PT coated iron cathode and a lithium anode [104]. Cathodes consisting of an electrochemically prepared PT membrane on a carbon-fiber current collector are also named in literature [105].

10.5 Materials ,for rechargeable batteries, capacitors

499

Lithium batteries having cathodes consisting of electrically conducting polymers like 3-methyl- or 3-hexylPT and graphite are characterized by a high capacity and a remarkable long cycle life [1061. Composites produced from polyaromatics, 3-butylPT and a polymeric sulfate or from an electrically nonconducting matrix polymer and 3-octylPT may also be used as cathodes in batteries [67]. Secondary batteries with a long life cycle can be fitted with at first a protective layer (salts, oxides, hydroxides) and an electrically conductive polymer, e.g. PT [107]. A composite electrode produced from an electrically conductive polymeric material and an organic compound is described not only for the use of rechargeable batteries, but also for sensors (see section 10.9) as well as for electrochromic displays (see section 10.8) [108, 1091. Secondary batteries with excellent discharge properties and electrochromic display devices (see section 10.8) contain poly(vinyl-2,2';5/,2''-terthiophene) as an organic conductive material with redox properties [I 101. Poly(2,5-thiophenediyl vinylene) (PTV, structure see Scheme 1) possess good redox behavior and excellent charge-discharge characteristics such as a Coulombic efficiency higher than 99% and a stable electrode potential even under high chargedischarge conditions and has been claimed for the use in secondary batteries [I 113. PT also has been described for the control of charging in batteries [112].

n Polythiophene (PT)

Poly(3-alkylthiophene)

R=akyl n

@$

n Poly(2,5-thiophenediyhinylene)(PTV)

Poly(3,4-ethylenedioxy-2,5thiophenediyl) (PEDT)

Poly[ 1,2-di(2-thienyl)-

Poly(2,T-bithiophene-5-yl-

ethylene]

methylmethacrylate

Scheme 1

500

10 Application of Electrically Conductive Polythiophenes

Poly(3,4-ethylenedioxythiophene) (PEDT, structure see Scheme 1 and also section 10.3) can be prepared as a film of a length of up to 20m using a cylindrical rotating electrode and special polyanions as electrolyte salts, e.g. sulfated poly(P-hydroxyether)s, poly(P-hydroxyether)s bearing trifluoromethyl groups and sulfated polybutadiene forming the counterions for the polymer. These films are claimed to be useful for rechargeable batteries [64]. In future PT will be used in techniques as a material for rechargeable batteries, when it will be possible to overcome the self-dischargingprocess. In the charged battery PT is present in its oxidized form, which is an oxidant and may oxidize oxidizable components in the electrolyte solution while it itself is reduced. It is thought that a possible way to decrease the self-discharging process may be to increase the density of the electrically conductive polymeric material on the electrode [113]. Capacitors need materials with high dielectric constants; PT and its derivatives having suitable electrical conductivities are qualified for this application. Poly(3,4ethylenedioxythiophene) (PEDT) is claimed to be useful as electrode material for solid state electrolyte capacitors [114]; a typical arangement for this purpose is: one electrode is made from aluminum, the other from PEDT, and between both electrodes a layer of alumina as dielectricum is located. Electrically conductive polymers represent an interesting class of materials for use in electrochemical capacitors owing to the combination of high capacitive energy density and low material costs [115, 1161 High power densities in electrochemical capacitors can be obtained for n- and p-type doped poly[3-(4-fluorophenyl)thiophene] as has been shown by impedance studies [115, 1161. PT itself and substituted PTs are claimed to be useful as solid electrolytes in solid electrolytic capacitors: PT [117, 1181, 3-methylPT [119, 1201, 3-ethylPT [120], 3-propylPT [120], 3-butylPT [121], 3-hexylPT [122], 3-octylPT [121], 3-dodecylPT [121], 3-docosylPT [121], and PEDT [59, 1231. A composite film of 3-dodecylPT and poly(ethy1ene terephthalate) is resistant to migration and can be used as a film capacitor [ 1241. Growing interest in supercapacitors is stimulated by their potential application as power storage devices operating in parallel with batteries in electrical vehicles to provide pulses of peak power [125]. Two types of supercapacitors are under research: (i) the ‘double layer supercapacitor’ is electrostatic in origin, i.e. it arises from the separation of electron and ionic charges at the interphase between a high-specific electrode and an aqueous or organic electrolyte; (ii) in the ‘redox supercapacitor’ fast Faradaic charge-transfer takes place at the electrode materials as in a battery. For both types of supercapacitors are required electrode materials having high specific capacitances and quick charge-discharge characteristics. Beside of polyaniline, polypyrrole, also the redoxable PT and its derivatives can be used as efficient materials because of their dopability chemically and electrochemically. Prototypes of such supercapacitors of small geometric areas have been developed [126, 1271. 3-MethylPT and poly(dithieno[3,4-b:3’,4’-d]thiophene) have been grown galvanostatically on carbon paper electrodes by electrochemical polymerization of the appropriate monomers. The properties of studied supercapacitors with these materials are comparable to common carbon supercapacitors in the case of the first mentioned polymer and appear to be better in the case of the second polymer [125].

10.7 Resists, recording materials and fabrication of patterns

501

10.6 Junction devices and rectifying bilayer electrodes Ap-n-heterojunction type device can be fabricated on the basis of the junction formed between polypyrrole and PT derivatives, e.g. PT itself or 3-methylPT [128]. A 3-octadecylPT/ fullerene c60 junction device with a photo-induced charge transfer between 3-octadecylPT and fullerene c60 is prepared in the following way: a quartz plate is coated with a gold electrode, a layer of 3-alkylPT is brought up, followed by a layer of fullerene C60and at the end a layer of alumina as the second electrode [129]. Rectifying bilayer electrodes with sequential bilayer structures are prepared from any pair of 3-butylPT/polypyrrole and 3-bromoPT/polypyrrole by anodic electropolymerization on platinum electrodes [ 1301.

10.7 Resists, recording materials and fabrication of patterns Unsubstituted as well as alkyl substituted PT can be etched and cut by energetic irradiation to record optical images in the layer of the conducting polymer [131]: 3-alkylPTs, e.g. 3-hexylPT can be solubilized by energetic light irradiation (argon ion laser and a halogen lamp as light source) in organic solvents such as chloroform in the presence of oxygen. Solubilization occurs on irradiation with light of a photon energy higher than the band gap energy of the mentioned polymeric material, e.g. 2.2 eV (560 nm). In contrast to the above described process, thin films of soluble 3-hexylPT can be cross-linked and thus be made insoluble by radiation with UV/visible light [ 132, 1331. Irradiation of thin polymeric films through a photomask and subsequent development with a suitable solvent leaves a polymer image of the mask. The resulting polymer pattern can be made electrically conductive by a chemical oxidation process. Such electronically conductive ‘channels’ can be fabricated using conventional photolithographic techniques. Polysilane films treated with thiophene can be polymerized electrochemically on an indium tin oxide (ITO) glass electrode as an insoluble film under a pattern in such areas without UV light irradiation, whereas the polysilane film in such areas irradiated with UV light can be dissolved in a polar solvent such as propylene carbonate [134]. 3-HexylPT films containing diphenyliodonium hexafluoroarsenate are used for optical recording by light irradiation followed by washing with chloroform [135]. Poly(2,5-thiophenediyl vinylene) (PTV) as a material for patterns can reach an electrical conductivity of 0.1 S cm-’ after iodine doping; such layers can be prepared from thin films of the polymer layers to patterned UV irradiation and subsequent development with a suitable solvent [136]. Microlithographic patterning of micrometer scale structures in PT films can be prepared by two methods [137]: (i) trenches in oxidized silicon wafers are created

502

10 Application of Electrically Conductive Polythiophenes

and filled with PT using the flowing afterglow synthesis technique; (ii) PT films are coated with a low temperature spin-on glass; a photoresist is then applied, exposed and developed; a tetrafluoromethane/oxygen plasma transfers the lithographc pattern to the glass layer; a final oxygen plasma etch processes the film and removes the photoresist at the same time. Photoreceptors for electrophotography have interlayers containing alkylated PTs like 3-methyPT between an electrically conductive substrate and a light-sensitive layer [138-1401. Also other substituted thiophenes can be used for the electropolymerization of electrophotographic photoreceptors [141-1431. The problem in electron-beam lithography is that image distortion and pattern placement errors arise by resist charging during electron-beam exposure. 3-Alkylated PTs, e.g. a copolymer made from 3-methyl- and 3-butylthiophene or watersoluble poly(3-thienyl-2-ethanesulfonate)as discharge interlayer and top layer eliminate the charging problem during the electron-beam exposure [ 1441. Resist materials consist of electrically conductive and UV-absorbing polymers or of organic materials containing polymers such as 3-methoxyPT. When such a resist is used as the lower layer of bilevel resists, deformation of pattern and decrease of resolution are suppressed during the UV or electron-beam exposure [145]. A polystyrene composite with heterocyclic polymers (e.g. 3-methoxyPT) is claimed to be useful as the lower layer of bilevel electron-beam resists to prevent charge accumulation and to achieve accurate patterning [146]. A resist composition containing a soluble electrically conductive polymer like 3-dodecylPT together with a radiation sensitive (UV, deep UV, electron beam) acidor base-generating agent is described to be of high sensitivity and resolution [147]. PT and 3-butylPT as electrically conductive materials are useful as electron beam resists [148]. A negative electron beam resist can be produced by an electron beam to form functionalization and cross-linking in 3-octylPT in a single step creating submicron scale polymer structures carrying functionalized groups [149, 1501. A negative deep UV resist can be produced from a composition of 3-octylPT with [1501. Grid the cross-linker ethylene [1,2-bis(4-azido-2,3,5,6-tetrafluorobenzoate)] patterns can be reproduced without distortion on an electrically conductive electron beam resist material containing substituted PT like 3-dodecyloxyPT [151,1521. Composites from novolac and conductive polymers (e.g. PTV, 3,CdibutylPT, 3-propylPT) can be used as radiation resists; they are decomposed by irradiation with charged beams forming fine patterns [I 531. Photoresists and materials for nonlinear optics as electrically conducting materials can contain poly(2,5-thiophenediyl-tetramethyldisilanylene)[ 1541 or 3hexylPT as semiconducting material dispersed in an insulating polymer [ 1551. An electrically conductive pattern film can be produced by electrochemical polymerization of thiophene or its derivatives on a gold film that can act as an electrode in a desired pattern on a changable substrate: by electrochemical oxidation the gold is then dissolved to form the pattern film [ 1561. Printing on insulating material can be realized with a lipophilic ink on an isolating material covered with a layer of electrically conductive PT [157]. Electrically conductive powdered inks for color electrothermal-transfer printing use ink nuclei and electrically conductive polymers like 3-methylPT on the surface

10.8 Electrochromic devices

503

[ 1581. A laminate combination useful for an imaging process working with seamless polymeric belts comprise a laminate of a host polymer layer and a laminate of an electrically conductive polymer layer [PT, poly(bithiophene)] [1591. An electrothermal imaging device contains an array of pyroelectrical sensor elements supported by a pillar of a semiconductor like PT [160]. PT incorporated in a homogeneous dispersion of permanent magnetic particles linked either chemically or electrochemically to the polymer or as a dopant can be used for magnetic printing [161]. Poly(bithiophene) layers are claimed for use as imaging systems for the offset printing process, because this process is based on the different wettability of printing and nonprinting areas. The neutral hydrophobic polymer can be oxidized electrochemically in an electrolyte solution (tetrabutylammonium perchlorate in acetonitrile) to form the oxidized (doped) hydrophilic form of the polymer. This reaction can be conducted reversibly for many thousand cycles changing between hydrophilic and hydrophobic behavior of the polymeric surface [162- 1661. A new type of patterning is described for a polymer system of ultrathin cellulose ether networks as a host matrix for electrochemical polymerization of alkylthiophenes. An oriented multilayer assembly (typically 80 monolayers prepared by Langmuir-Blodgett technique) of butylated cinnamoylcellulose was cross-linked under a pattern by UV irradiation. After dissolving the unchanged material on an indium tin oxide (ITO) glass plate, this patterned plate was taken as an anode for the electropolymerization of 3-pentylthiophene. The areas coated with the photopolymerized material develop quickly and more intensively blue colored 3-pentylPT by electropolymerization [ 1671.

10.8 Electrochromic devices The typical spectral changes in the visible region accompanied with structural changes of the polymeric chain by oxidizing or reducing ( p - or n-doping) of PT and its derivatives have led to several proposals of electro-optical systems such as display devices or electrochromic windows (‘smart windows’) [ 168- 1751 In first studies revealed good optical contrast and response times ranging from 10 to 300ms have been found [176, 1771 as well as up to 1.2 x lo5 charge-discharge cycles for 3-methylPT films on bulky platinum electrodes [178]. Typical electro-optical cells consist of (i) an electrically conductive indium tin oxide (ITO) glass as electrode, (ii) a thin film of the electrochromic polymer in the electrolyte (e.g. lithium perchlorate in acetonitrile), and (iii) hold by a spacer a nickel gauze on a back glass plate [179]. Poly(is0thianaphthene) films on semitransparent gold electrodes can retain about 75% of their initial optical density after 6 x lo3 cycles [180j. The conditions of the electrosynthesis are important for the electrochromic properties under many repeated redox cycles [I8 11. A comparison of different materials brought the result that 3-methylPT is a better material for electrochromic applications than

504

10 Application of Electrically Conductive Polythiophenes

poly(bithiophene) and poly(dithieno[3,2-b:2',3'-d]thiophene) concerning the performance and durability of the devices [182]. The long term stability is limited by the doping level and the switching potential in order to prevent overoxidation, meaning destruction of the polymer; lifetimes with lo7 cycles can be reached for alkyl and polyether substituted PT films operating at 10% of their maximum capacity [183, 1841. Multicolor electrochromic devices consist of different electrochromic active conducting polymers showing specific colors at quite different operation potentials. Polyaniline-poly(sodium acry1ate)-PT and polyaniline-poly(sodium acry1ate)-3methylPT films are constructed by combining polyaniline-poly(sodium acrylate) with its specific color in the oxidized state with PT and 3-methylPT films showing their specific color in the reduced state [168]. An electroactive polymeric laminate suitable for an electrochromic display device comprises a conductive substrate, a first layer of an electroactive polymer, and a second layer of another electroactive polymer prepared from a monomer with a higher oxidation potential than that of the first monomer and adhering to the first layer [174]. A solid state electrochromic device can be constructed by combining a thin layer of an electrically conductive polymer like 3-octylPT produced by spin coating from solution with a solid polymer electrolyte and a metal oxide [175]. Substrates coated with PT have been claimed for use in electrochromic displays, in solar cells and for corrosion protection [185]. Poly(3,4-ethylenedioxythiophene)(PEDT) with its good electrochromic properties and long term stability can be used as an electrode in a solid state electrochromic cell [186]. The optical properties of PT can be changed by attaching substituents to the thiophene system because of influencing the band gap. Electron donating substituents lower the band gap, while steric interaction of side groups on adjacent thiophene rings forces the conjugated backbone to twist and increases the band gap. By choosing suitable combinations, polymers with specific optical properties can be designed for electrochromic applications. Electrofluorescence from these polymers cover the full visible spectrum, from the blue into the near infra-red. PEDT has shown to possess interesting optical properties, increased ionic conductivity in its oxidized form and a high thermal stability useful for application in electrochromic devices and windows [ 1871. Starting from the structure of PEDT the design of polymeric materials with tailormade bulk properties at the molecular level has resulted in the availability of new materials with a wide range of color transitions useful for electrochromic applications [188]. In a dual polymer electrochromic cell (see Fig. 1) PEDT (I) is working as a cathodically coloring electrochromic material in combination with the anodically coloring polyaniline or poly(3,6-carbazolediyl-3,4-ethylenedioxythiophene-3,4-ethylenedioxythiophene-3,6-carbazolediyl)(11) between indium tin oxide (ITO) glass electrodes. The special combination of I and I1 together with a gel electrolyte has the advantage that switching between -0.8 and +0.8 V produces changes from bleached to the colored state. The controllable transition goes from a pale transparent green to a dark opaque violet, because of the color changes during the electrochemical redox reactions (see Fig. 2). This device switches from 43 to 65% transmittance in 100 ms [188].

10.8 Electrochrornic devices IT0 Layer (Transparent condoctor)

505

Insulating Snbstrate

Gel Electrolyte Figure 1. Dual polymer electrochromic cell with anodically and cathodically coloring polymers

separated by a gel electrode (taken from [409b]).

Poly(is0thianaphthene) can be reversibly cation- and anion-doped forming two stable redox states of different color and is a potential candidate for electrochromic displays [1891. Poly{5-(cyano-nonafluorobutylsulfonyl)cyclopenta[2,l-b:3,4-b’]bit~ophene} has been prepared electrochemicallyforming a p - and n-doping material for electrochromic devices. In a model device incorporated this polymer as both the anode and the cathode electrochromic behavior is demonstrated [ 1901. a) Cathodically coloring polymer ECM

PEDOT(1) n

0

n

0

0

0

ox red 0

0

u

0

0-0

0

0

0

u

U

( deep blue - violet )

0

0

u

( transparent pale blue )

b) Anodically coloring polymer ECM

PBEDOT-CZ (11)

(deep blue-violet)

(transparent pale yellow)

Figure 2. Electrochemical oxidation and reduction of the electroactive polymers I and I1 (see text)

(taken from [409b]).

506

10 Application of Electrically Conductive Polythiophenes I

I

It

A polymethacrylate with pendant terthiophene groups, poly[(2,2 :5.2 -terthiophen-5-y1)methyl methacrylate], has been polymerized radically by the action of azobisisobutyronitrile. Electrochemical oxidation (no cross-linking between the thiophene systems in the terthiophene units occurs because of the high stability of the radical cations formed by oxidation!) of polymeric films in acetonitrile with tetrabutylammonium perchlorate as electrolyte salt produced a purple-colored polymer (doped form) with pendant terthiophene radical salt groups. The polymer undergoes a reversible clear color change from pale yellow to purple and vice versa on electrochemical oxidation and reduction (doping and dedoping) and is expected to be a potential electrochromic material [ 1911. Besides of electrochromic effects PT and its derivatives show changes in color, not only by applying an electrochemical redox reaction, but also on heating, on solvating, on applying of pressure or electricity or in contact with different counterions. It is thought that these effects may find technical interest; therefore one should also study the reviewed chapters on thermochromism, solvatochromism and ionochromism in Lit. [13].

10.9 Sensors Electrochemical sensors can be created by producing modified electrodes with selective functions. Based on selective recognition special electrodes can be used in various fields in analytical chemistry. Such devices are very easy to handle and therefore one should expect many applications in this field. Sensors can be used in the gas phase and also in solution to identify ions and also organic compounds. Also electrocatalytic activity of such materials can be seen in this connection. Modified electrodes for electrocatalytic processes have been prepared from PT/metal hybrid materials [ 192, 1931 and 3-methylPT [ 1941. Ionic and molecular recognition has been found with some functionalized PTs in selectively modified electrodes for electrocatalytic or electroanalytic applications [ 195, 1961. Organic electrochemical transistors have been realized with a 3-methylPT film deposited between the source and the drain; the drain current is driven by the electrochemically controlled conductivity of the 3-methylPT film [197]. This device has been shown to be sensitive to oxidants and reductants and to amplify signals up to kilohertz frequencies [ 1981.

10.9.1 Sensors for gases The change (mostly increase) of electrical conductivity of conductive polymers by applying gases like NOz has been studied. PT and its derivatives have also been included in this research area. The NO2 sensitivity of PT has been reported and the mechanism ofN02 detection is well known as a doping effect of NO2 to PT [199].

10.9 Sensors

507

The gas sensitivity of 3-butylPT towards air, water vapor, ethanol and chloroform in the gas phase has been investigated; enhanced source-drain currents have been found in comparison to that in vucuo [200]. 3-OctylPT, poly[3-(2-propanoyloxyethyl)thiophene] and 3-(2-decanoyloxy)PT have been degassed by NO2 during weighing by a quartz crystal microbalance; this very sensitive method gives a clear correlation between the NO2 concentration and the weight increase of the polymeric thiophene derivatives applied as multilayer thin film material mixed with octadecylamine [201]. A gas sensor is described by a two-stage reversible reaction of bromine with poly[1,2-di(2-thienyl)ethylene] [202].

10.9.2 Sensors for ions in aqueous solution Pt-electrodes coated with an electrochemically prepared 3-methylPT layer have enhanced anodic current response in cyclic voltammetry in nitrate and perchlorate aqueous solutions compared to the response in other salt solutions because of the oxidizing character of the two mentioned ions; therefore 3-methylPT has been claimed as a material for sensors for oxidizing ions in aqueous solutions [203, 2041. Glassy carbon electrodes coated with an electrochemically prepared layer of 3-methylPT show selective response to anions with different anion radii depending on the thickness of the 3-methylPT layer; thus nitrates, tetrafluoroborates, perchlorates, fluorides, primary phosphates, sulfates and tosylates can be detected [205]. Chemically prepared copper(I1)-containing 3-methylPT pressed powder used as working electrode of a three-electrode system in a thin layer amperometric cell unit can detect ionic analytes (fluoride, chloride, nitrate, sulfate, ascorbate, oxalate, citrate, each as sodium salts) in an aqueous stream by flow-injection analysis. At applied voltages between 0.2 and 3.0 V (against Ag/AgCl) this coated electrode possesses favorable sensitivity and stability in comparison with metallic electrodes normally under use. This electrode shows no deterioration in its performance for more than a year [206]. Graphite electrodes were coated electrochemically with 3-methylPT, then iodine was incorporated into the polymer phase by electrochemical oxidation of iodide at a potential of +0.70 V. This modified electrode as a potentiometric iodide ion sensor was found to be suitable for the measurement of iodide concentrations down to lop6M; its selectivity coefficients for most of the potential interferents (15 anions have been studied) have been estimated to be of the order of [207]. Platinum electrodes coated with electrochemically polymerized thiophene, 3methyl-, 3-octylthiophene, 2,2'-bithiophene as well as 4,4'-dioctyl-2,2'-bithiophene can be used as indicator electrodes giving a cationic response to monovalent cations such as H', Li', Na', K+, NH; and showing some sensitivity to divalent cations such as Mg2+ and Ca2+ [208]. Solid substrate electrodes (platinum, gold, vitreous carbon) with electrochemically deposited 3-octylthiophene is seen as a coated electrode with an ion-selective membrane to produce a solid-contact ion-selective electrode for several ions (e.g. Li+, Ca2+,CI- [2091.

508

10 Application of Electrically Conductive Polythiophenes

The incorporation of mercury into 3-methylPT results in an effective electrode for the analyses of lead(I1) ions in aqueous solutions down to concentrations of 0.05 ppm. The ‘mercury lims’ are deposited electrochemically after the electropolymerization step [210].

10.9.3 Sensors for organic materials Electrodes modified by electrodeposited 3-methylPT can be used as chemical sensors for organic and especially biological interesting molecules. The modified surfaces catalyze the oxidation of several compounds (e.g. ferrocyanide, catechol, ascorbic acid, hydroquinone, dopamine, epinephrine, acetaminophene, p-aminophenol, and reduced nicotinamide adenine dinucleotide). The sensitivity of these polymercoated electrodes is 4 to 10 times higher than pure electrodes of platinum. Binary and also ternary mixtures can also be analyzed; these electrodes are also suitable as an amperometric detector for flow injection analysis [211, 2121. A chemical detector for ascorbic acid is a polybithiophene-modified indium tin oxide (ITO) glass electrode. Reduction of the polymeric layer occurs when doped polymeric films are brought into contact with an aqueous solution of ascorbic acid. This redox reaction is clearly indicated by the change in the chromatographic property of the film [213]. Platinum electrodes coated with PT and 3-methylPT are claimed to work as sensor electrodes for N-phenyl-p-phenylenediamine;the mechanism of this anodic oxidation reaction is discussed [214, 2151. A solid support containing PT is described for the detection of pesticide residues in food and soil samples as well as in ground water [216].

10.9.4 Sensors for bio-organic materials Polymer-coated microelectrodes containing one or several monolayers fabricated by electropolymerization of thiophene, bithiophene and 3-methylthiophene at low potential with a low counterion content or with low affinity ions, can be readily exchanged by ion-exchange technique by proteins, antibodies, antigens and drugs in order to immobilize these molecules [217]. Biotinylated copolymers of 3-undecylthiophene and 3-thienylmethanol using streptaviodin as a cross-linker protein is an electroactive matrix for the attachment of the photoactive protein phycoerythrin. The biotinylation of the copolymer improves the film formability and results in a stable monolayer. The phycoerythrin binding to this polymeric layer can be monitored through fluorescence microscopy at the air-water interface [218]. Spectroscopic studies confirm, that the molecular properties of biologically active proteins, e.g. glucose oxidase (GOD) and peroxidase as guests in a polymer host matrix are not changed and keep their biological activitieswhen they are (i) embedded

10.9 Sensors

509

in PT derivatives during the electropolymerization of the appropriate monomers [1,2-di(2-thienyl)ethylene,3-pentyloxythiophene]on platinum electrodes from an acetonitrile/water suspension or (ii) immobilized in the polymer matrix of the chemically prepared PT derivatives by FeC13 in methanol on porous polycarbonate membranes and during changing the PT derivatives from the pristine to the doped state (neutral to the oxidized state) [219]. A very attractive area of research concerns the immobilization of biologically active proteins, especially enzymes, on electricallyconductive polymers, also including PT for the preparation of amperometric enzyme electrodes. Such biomolecules can be embedded into the electricallyconductive polymeric gel phase during the preparation of the matrix polymer, they are statistically distributed in the polymer phase; these immobilized biomolecules react by diffusion controlled mechanisms with the aqueous solution of the substrate [220, 2221. The second possibility to immobilize biomolecules to electrically conducting polymers is their covalent binding to the surface of the mostly dense polymer phases by polymer analogous reaction of functional groups of the polymer under the mild conditions necessary for biomolecules; in this case the biomolecules hang covalently bonded as pendant substituents on the outer polymer surface and are components in the surrounding aqueous phase because of their hydrophilic character [222]. It must be noted, that their concentrations in total are lower than in the above described case, but experience shows higher activities for the immobilized biomolecules concentrated in the outer surface of the polymer film on the electrode [222]. For covalently bonded immobilizates functionalized polymer structures have to be designed for the preparation of functionalized polymeric films on electrodes (‘functionalized polymer-covered electrodes’). Some 3-functionalized thiophene derivatives have been prepared, bearing e.g. carboxy, activated ester, hydroxy, amino groups, ready for reaction with bioorganic compounds via amino, carboxy or hydroxy groups of the biomolecules as well known in peptide chemistry. An amperometric biosensor produced by immobilization of redox enzymes at PT-modified electrode surfaces has been prepared by electropolymerization of the activated ester N-succinimidyl ~-(2,2’-bithiophen-3-yl)hexanoatein methylene chloride with tetrabutylammonium hexafluorophosphate as electrolyte salt on platinum [221] and following immobilization of glucose oxidase (GOD) by the reaction of the activated ester groups with the free amino groups of lysine units in aqueous buffer solution at pH 7.4 forming amide bonds to GOD pendant as substituent to the electrode (‘enzyme electrode’) (see Fig. 3) [222]. Calibration measurements show, that a polymer deposition of seven scans gives the highest current for such enzyme sensors having a sensitivity of less than 1 mM GOD solution and a good stability for 90 h in an electrochemical flow-through cell. Electrodes coated with a copolymer prepared from pyrrole-modified GOD [223] and bithiophene in aqueous acetonitrile (‘entrapped GOD electrode’) lead under the same conditions to considerable lower current responses at equal glucose concentrations [222]. Starting from commercially available compounds the copolymerization of 2thienylacetic acid with 3-methylthiophene can be performed easily forming ‘polymer covered electrodes with functional groups’ containing different concentrations of carboxy groups [224]. By activation of the carboxy groups with dicyclohexyl-

510

10 Application of Electrically Conductive Polythiophenes

Figure 3. Two-stepmodification of a PT-based enzyme electrode. (a) Electrochemicalformation of the functionalized activated ester PT. (b) Covalent binding of GOD via its lysine residues to the activated carboxy groups at the surface of the polymer (taken from [554]).

carbodiimide enzymes (see Scheme 2) like lactate oxidase (LOD) can be bonded covalently to the surface of such electrode forming an ‘enzyme electrode’ for the detection of lactic acid, which is oxidized by the action of LOD to pyruvate and hydrogen peroxide (see Fig. 4). Hydroperoxide is detected on free parts of the surface (or by diffusion to the surface?) of the platinum electrode against a standard

COOH

COOH

d $ +

s

R = alkyl

-2P-

- 2H’ R

+

Scheme 2

Dl Enzyme

+DCC

10.9 Sensors

”\.

/

Lactateoxidase

I

H-$-OH

-

0

0’

51 1

I

y=O

+

HzOz

CH,

CH,

Pyruvate

Lactate

Figure 4. Oxidation of lactate by the enzyme lactate oxidase (LOD) to pyruvate and hydrogen peroxide.

Ag/AgCl electrode [225]. Calibration graphs for two coated electrodes with LOD demonstrate the practicability as sensor for lactate (see Fig. 5). As can be seen from the graph for a sample that was prepared by galvanostatic electrocopolymerization of 2-thienylacetic acid together with 3-methylthiophene in the molar ratio 2 :1 (System 11, in Fig. 5), the detection limit of lactic acid is about 5 pmol per 1 [225]. For polymer analogous reactions it is interesting to start with activated esters (‘protected carboxy groups’); benzyl esters of 2-thienylacetic acid are easily available, can be electropolymerized and the benzyl ester group can be removed easily forming the free carboxy group [226]. Pentafluorophenyl and succinimidyl esters of 2-thienylacetic acid prepared in an one-step-synthesis from the acid can be electropolymerized or -copolymerized with 3-methylthiophene forming polymer coated electrodes with activated ester groups that can be reacted with amines (butylamine, 2-(6-aminohexylamino)-3-chloronaphthoquinone) as model compounds for enzymes [227]. An alternative route for binding enzymes to a functionalized PT starts with 3(2-hydroxyethyl)-thiophene (easily available from methyl 3-thienylcarboxylate by q

500 -

C

.-

C

;LOO -

a

5

u

300 -

0

100 200 300 Concentration of lactate in pmol/L

Figure 5. Calibration graphs of lactate oxidase (LOD) sensor for lactate.

512

10 Application of Electrically Conductive Polythiophenes OH

OX

Scheme 3

reduction with LiA1H4) [228]. This compound cannot be polymerized electrochemically, but in a protected form (as methyl ether, that can be cleaved after polymerization; also other ethers and esters as protected forms are helpful [229]). The polymer analogous prepared poly[3-(2-hydroxyethyl)thiophene] is reacted with triphosgene forming an oxycarbonylchloride that form - as well known from protein chemistry - with proteins, e.g. enzymes ‘enzyme electrodes’. In this way the immobilization of alcohol dehydrogenase was realized (see Scheme 3, route a). By hydrolysis and chromatography an amount of about 5 nmol enzyme was estimated for 1cm2 of the platinum electrode [230]. A second route for the preparation of ‘enzyme electrodes’ starting with poly[3-(2hydroxyethy1)thiophenel is the bromocyane method - also well known from protein chemistry. PT with pendant hydroxy groups is treated with bromocyane forming the reactive cyanate (see also Scheme 3, route b) which can be taken for the immobilization of enzymes, e.g. alcohol dehydsogenase [228, 2301. An interesting question is the interaction between two redox systems in polymers containing units with two different redox functions. In the case of viologen functionalized poly(3-alkylthiophene), EPR studies indicate a rapid motional averaging of anisotropic electron-nuclear couplings that can be explained by a counterion complexation [231].

10.10 Other applications The term ‘polymer electronics’ describes a field of properties and applications of electrically conductive polymers that arise besides of only electrical conductivity. These applications derive from a combination of properties of these materials especially at the interfaces of electrically conducting organic materials and metals or

10.10 Other applications

513

semiconductors [232]. Examples are: Schottky diodes, field-effect transistors and photovoltaic devices (see chapter 8). Besides the applications described in sections 10.2 to 10.9 further applications are summarized arising from special properties of PT and its derivatives. Possible applications are also added that are claimed for other electrically conducting polymers but may be possible for PT.

10.10.1 General consideration The redox system PT may reversibly oxidized or reduced chemically by oxidants or reductants or electrochemically (doping/dedoping reaction). So one can switch between the neutral and the oxidized form of PT: the nonpolar form changes to the polar form (salt form of PT); one produces an ion exchanger from an electrically neutral material and one can eliminate the ion exchanger structure by switching a neutral material; the neutral, less conductive form can be changed to the oxidized, loaded, highly conductive form having different electrical conductivities depending on the degree of oxidation (doping), the higher the degree of oxidation the higher is the electrical conductivity; changing the unpolar material with its hydrophobic character (lyophilic character) toward the loaded, polar material creates hydrophilicity. The solid state structure of the nonpolar, neutral, unloaded PT changes dramatically to the polar, electronically changed radical cations in the polymeric chains bearing and counterions attracting material; because of different chemical structures also different spectra in the UV/visible region, meaning different colors for the two forms of PT can be seen (electrochromic effect, see section 10.8); these two different forms of PT with their different polarities possess different interactions to other molecules as solvent molecules or gas molecules concerning their adsorption properties. PT for example swells slightly in aromatic solvents and also diffusion of such solvents will be small whereas swelling of the oxidized form of PT in aqueous solution is stronger and depends also on the structure and electronic properties of the electrolyte salt in the aqueous solution. So one can understand the different wettability of the two forms of PT and the potential application in offset printing (see section 10.7). Electrically controlled oxidation (doping) results in a controlled swelling of electrically conductive materials, generally. The counterions of the salt units in the polymeric chain push the polymer chains apart from each other and the volume expands. This controlled expansion can be used in ‘artificial muscles’, the so named ‘actuators’. During swelling surprisingly high forces elongate the material as large as 10%; this is two orders of magnitude higher than for piezoelectric polymers [233]. Membranes produced from 3-hexylPT show different permeabilities for drugs [234]; the transport of dopamine through membranes of this material is found to depend on the oxidation state of the polymeric sample. In the reduced state membranes are more permeable for dopamine, in the oxidized form a decrease of 40% of permeability is observed because of changes in the polymer film morphology caused by structural changes by the oxidation state.

5 14

10 Application of Electrically Conductive Polythiophenes

Membranes for loudspeakers depend on the piezoelectric effect. In a typical example a large 6pm polyester film (up to 2m long) is coated with a 0.1 pm layer of a conducting polymer and suspended between two metal grids. High voltage is applied to the grid and the speaker membrane, and an AC audio signal is given onto the grids via a transformator. The high voltage field keeps the membrane centered between the grids and the AC field moves it with acoustic frequency. The advantage of conducting polymer materials as compared to metals is the better adhesion. In addition such electrostatic loudspeakers have low inertia [235]. Corrosion protection by electrically conductive polymers has been claimed in several reports by coating metals by electrochemically generated polymers - including PT and its derivatives - in thin layers [236, 237, 2381. Electrical conductors with good heat and moisture resistance (no discoloration) are claimed to contain copper powder and 0.5% PT powder [239]. A German Company plans to put a polyaniline-based ‘rust prevention kit’ on the market and a Swedish group investigates polypyrrole as corrosion protection for dental implants [240]. For this purpose the polymer must be produced on the surface of the metal in its neutral, electrically less conductive form. Corrosion protection can only be effective if no oxidative salts or acids attack the protected material and oxidize the polymer layer forming a polar, hydrophilic material that allows diffusion of the agressive chemicals to the metalpolymer interface. Normally, the solid structure of conjugated polymers is porous, or a porous structure in a generally compact structure (without voids between particles or single condensed regions) that can be produced tailormade (‘molecular pores’). Such polymers like polyaniline become porous by repeating oxidation and reduction (doping/ dedoping). These pores are considerably smaller than the normal voids between condensed regions; they can be used for gas separation. In the case of polyaniline membranes have been cast, oxidized, reduced and oxidized again; such material has a high separation factor for H2 /N2 (1 1:1, after 3 :1 for the cast material). Electrically conductive membranes can be optimized also for other separation problems [241]. Organic electrically conductive membranes such as 3-methylPT synthesized on porous polycarbonate microfilter membranes can be used for the separation of neutral solutes [242]. Such superthin membranes can also consist of a heterolayer polymer, 3-methylPT and poly(pyrro1e) [243]. The reversible electrochemical oxidation of PT and its derivatives may be exploited for drug release devices with a neurotransmitter [244]. Catalytical activities have been claimed for copolymers arising from pyrrole and bithiophene with a small amount of dispersed platinum to be effective in the electrochemical oxidation of hydrogen, formic acid and methanol [245]; the same catalytic effect was found for electrodes of 3-methylPT with electrochemically deposited platinum and tin [246]. The reduction of tetracyano-p-quinodimethane and chloranil can be achieved at a glassy carbon electrode coated with 3-methylPT [247]. Phenol derivatives and carbon dioxide can be converted to salicylic acid derivatives on 3-hexylPT electrodes under irradiation in ethanolic solution [248]. This electrode material acts as a photocatalyst on irradiation with visible light, the luminescence is quenched by carbon dioxide as well as phenol.

10.10 Other applications

515

Electrochemical memory devices can be constructed on the basis of 3-alkylPT (containing 1 to 8 carbon atoms in the side chain) [249-2501; these materials are used as memory channels whose chemical conductivities are varied by 3 to 4 orders of magnitude by the degree of electrochemical oxidation. The writing and erasing times of the memory states between minimum and maximum conductivities are 4-5 sand 10-20 s [250,251].The memory states or the channel conductivities are controlled by the number of applied writing and erasing pulses [249, 2501. Memory devices can be realized by PT as an electron-conjugated semiconductor for electrical controlling systems [252,253]. Polyisothianaphthene as a reversible redox system is claimed to be a candidate for applications in electronic devices exploiting the learning effect; because of the good stability of partly charged states (positive and negative) this material can be used in writing-reading systems [254]. Transparent IR-absorbing polycarbonate films containing poly(3,4-ethylenedioxythiophene) (PEDT) and polystyrenesulfonic acid can be used for the fabrication of greenhouse and building glass [l 1, 1141. This blended material is transparent to visible light and has a high absorption at wavelength higher than 800 nm. Passivation of the surface of n-GaAs is possible with thin films of plasma polymerized thiophene [255]. A composition containing an electrically nonconductive polymer matrix and 3-octylPT is claimed as a material for parabolic antennas, reflectors for radar, heating systems, photoelectric devices, and electric circuits and apparatus [256]. PTs are used for manufacturing a nonlinear two-terminal device; this device is not asymmetrical, gives stable electrical characteristics, and is useful as a display device [257, 2581. PT and 3-octylPT as electrically conductive polymers are claimed to be useful for liquid crystalline display devices [259-2611. PTs are also used for the production of color filters for liquid crystal displays [262]. Electrically conductive polymeric gels can be combined with other materials to establish bimorphs. A bimorph may be produced by successive electrochemical polymerization of pyrrole and then 3-hexylthiophene forming a double-film with two layers consisting of polypyrrole and 3-hexylPT. During swelling in chloroform only the 3-hexylPT layer expands; the bimorph becomes curled (see Fig. 6). The curled bimorph again becomes straight on after placing it in ethanol, shrinking the 3-hexylPT. The curvature can be controlled by changing the degree of oxidation [263, 2641. The response time of the bimorph can be controlled by the oxidation potential and the electrolyte concentration [264]. Sexithiophene and other oligothiophenes are claimed to be ideal photochromic materials for efficient and ultrafast incoherent-to-coherent optical converters [265].

Figure 6. Bimorph consisting of polypyrrole (0) and 3-hexylPT (a).(a) in the dry state, (b) after swelling in chloroform.

5 16

10 Application of Electrically Conductive Polythiophenes

10.11 Summary, conclusions and future trends The combination of interesting or extraordinary electrophysical properties with the processability of polymeric materials (plastics) [266] is a challenge in techniques. The outstanding properties of polymeric materials with their special solid state order in contrast to inorganic as well as organic crystalline materials with applicable electrophysical properties will give rise to many new fields in material science. The development of new materials by research will facilitate new applications by techniques as yet unknown. An additional advantage for polymers lies in the generally lower costs for the production of plastics than for crystalline materials. Imagine this future scenario for possibilities for the application of electrically conductive polymers in our homes [267]: 0 Roofs covered with electrically conducting polymers for photovoltaic conversion. 0 Walls covered with conducting polymers for electrochemical heating pumps. 0 Optical fibers covered with conducting polymers for many sensors. Antistatic painting and varnish to avoid the deposition of dust. 0 Filter with conducting polymers to eliminate dust from air. 0 Electrostatic devices for ion injection in air. Smart windows. 0 Large screens on walls for TV and video observations. 0 Antennas for satellites reception. Antistatic textiles on walls. Textiles for heating systems. 0 Large electrostatic loudspeakers. 0 Regeneration of air and water. 0 Conducting polymers in the wall for many kinds of electrical contacts. 0 Electroluminescence diodes for IR beam and sensors and trigger all over the house. Generally, manipulation of the properties of PTs are possible by changing the primary structure (e.g. substitution of thiophene, comonomers), the secondary structure (e.g. type of coupling between the monomeric units), and the tertiary structure (e.g. surface morphology). So different PTs can be designed tailormade to possess different properties for a wide variability of applications such as differences in solubility, wettability, absorption behavior, electrical conductivity and electrophysical properties. However, systematic changing of properties by systematic changing of the structure is rather problematic owing to the complex structure of the polymeric material. The investigation into the relationship between the properties and the alkyl side chain length in 3-alkylated PTs and between the properties and the type of coupling between monomeric units shows promising results. Another problem is the synthesis of polymers with desirable structures. The synthesis of structurally homogeneous 3-alkylated PTs has achieved noteworthy progress since about 1990. Considering the various properties of PTs including also their good mechanical properties, their processability by conventional polymer techniques and their environmental stability, it is not surprising that their applications are so abundant (antistatic and shielding coatings, coatings with special properties,

10.11 Summary, conclusions and future trends

517

batteries, capacitors, field effect transistors, Schottky diodes, solar cells, sensors, catalysts, electrochromic and electroluminescentproperties, swelling properties leading to changes in volume, also bimorphs). These electrically conductive polymers with 7r-conjugated systems (also named ‘organic metals’, ‘synthetic metals’) will play an interesting and increasing role in fundamental research in chemistry, physics, material science and also in applied science in future. Further knowledge of the synthesis-structure-properties relationship will improve the directed synthesis of new required polymeric structures for special applications in future. The actual and potential applications on the basis of rather good long term stability of PT and its derivatives as a unique class of materials - in contrast to related structures such as polyacetylene, polypyrrole and also polyaniline - indicate convincingly that they will play an important role in near future. It is thought that this class of advanced molecular materials allows exciting sophisticated applications especially in the field of selectivesensors and molecular electronics (‘molecular wires’). One important point concerns the exertions in increasing the electrical conductivity of PT and its derivatives; until now, the most significant progress in the improvement of the molecular order and electrical conductivity has been realized for the simplest thiophene derivatives, and the obtained values of the electrical conductivity are still far below the expected intrinsic conductivity of 7r-conjugated polymers [268]. Further progress will be possible with better understanding of the electropolymerization reaction mechanism and the structure-property relationship for the PTs in order to design tailormade materials for special hightec applications. One point has been cleared up: electropolymerization in contrast to chemically conducted oxidative polymerization leads to PTs with the highest electrical conductivities and also to original structures such as helical PTs. . One of the applicationsfor redoxable polymers most discussed is its use as electrode material for rechargeable batteries [7 11. Such ion-inserting polymers must possess an extremely high long term stability, a high degree of exchangable charge carriers, a fast insertion or repulsion of ions and a low specific weight. Up to now no better material than lead is to be seen: four exchangable charge carriers (instead of one in the case of PT) and practical 100% of the lead atoms/ions react (instead of about 30% of the thiophene systems). The specific weight is lower for PT than for lead, but the box for the battery and the electrolyte solution are comparable. Summarizing: the good old lead battery cannot been beaten! Perhaps the solid state order of PT may be more stable for PT than the crystalline structure of the lead plates in relation to their changes in crystallinity (propagation of dendrites). It seems to be hard work to find a satisfactory solution of this problem.

5 18

10 Application of Electrically Conductive Polythiophenes

References In these references, patents are abbreviated as follows: EP USP JP DDR CanP DEP FRP GBP

European Patent United States Patent Japanese Patent German Democratic Republic Patent Canadian Patent German Patent French Patent British Patent

WO

World (International) Patent Application

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

D. D. Eley, Nature (London), 1948, 162, 189. A. T. Vartanyan, J. Physik. Chem. UdSSR, 1948,22, 769. M. A. Geiderich, E. E. Davydov, B. A. Krentsel et al., J. Polym. Sci., 1961, 54, 624. W. Felmayer, I. Wolf, J. Electrochem. SOC.,1958, 105, 141. C. S. Marvel, J. H. Rassweiler, J. Amer. Chem. Soc.1958,80, 1197. A. A. Berlin, L. G. Cherkashina, E. I. Balabanov, Makromol. Verbind., 1962, 4 , 376. R. Foster, Organic Charge-Transfer Complexes, Academic Press, London, 1969. P. L. Kronick, H. Scott, M. M. Labes, J . Chem. Phys., 1964,40, 890. J. H. Lupinski, K. D. Kopple, Science, 1964, 146, 1038. C. K. Chiang, C. R. Fincher, Jr., Y. W. Parket al., Phys. Rev. Lett., 1977, 39, 1098. F. Jonas, W. Krafft, Bayer AG, 1991, Eur. Pat. Appl., EP 440957, DEP 90-4003720. J. Roncali, Chem. Rev., 1992, 92, 711-738. G. Schopf, G. Kossmehl, Adv. Polym. Sci., 1997,129, 1-166. J. Roncali, Chem. Rev., 1997,97, 173-205. A. 0. Patil, Y. Ikenoue, F. Wudl, A. J. Heeger, J . Amer. Chem. Soc., 1987, 109, 1858. T. Yamamoto, A. Morita, T. Maruyama, Z. H. Zhou, T. Kanbara, K. Sanechika, Polym. (Tokyo), 1990,22, 187. T. Yamamoto, T. Maruyama, Z. H. Zhou, Y. Miyazaki, T. Kanbara, K. Sanechika, Synth. Met., 1991, 41, 345. J. Bartus, J. Macromol Sci. Chem., 1991, A 28, 917. H. S. Nalwa, Angew. Makromol. Chem., 1991, 188, 105. R. D. McCullough, R. D. Lowe, M. Jaryarman, P. C. Ewbank, D. L. Anderson, S. Tristamnagle, Synth. Met., 1993, 55, 1198. T. Yamamoto, Y. Miyazaki, T. Fukada et al., Synth. Met., 1993, 55, 1214. T. Yamamoto, Prog. Polym. Sci., 1993, 17, 1153. Y. E. Whang, J. H. Han, H. S. Nalwa, T. Watanabe, S. Miyata, Synth. Met., 1991,43, 3043. G. Kossmehl, M. Niemitz, Synth. Met., 1991, 41, 1065. E. Ruckenstein, J. S . Park, Synth. Met., 1991, 44, 293. M. Pomerantz, J. J. Tseng, H. Zhu et al., Synth. Met., 1991, 41, 825. P. C. van Dort, J. E. Pickett, M. L. Blohm, Synth. Met., 1991, 42, 2305. M. L. Blohm, J. E. Pickett, P. C. Vandort, Macromolecules, 1993, 26, 1704. M. L. Blohm, P. C. van Dort, J. E. Pickett, Polym. Mater. Sci. Eng., 1991, 64,210. M. L. Blohm, P. C. van Dort, J. E. Pickett, CHEMTECH., 1992, 22, 105. M. T. Loponen, T. Taka, J. Laakso et al., Synth. Met., 1991, 41, 479. Y. Wang, M. F. Rubner, Polym. Prep. (Am. Chem. SOC.,Div. Polym. Chem.), 1990,31,402. H. Ishikawa, X. Xu, A. Kobayashi, M. Satoh, M. Suzuki, E. Hasegawa, J. Phys., 1992,025, 897. S. Wang, H. Takahashi, K. Yoshino, K. Tanaka, T. Yamabe, Jpn. J . Appl. Phys., Part 1,1990, 29, 772. J. Laakso, J. E. Osterholm, P. Nyholm, H. Stubb, E. Punkka, Synth. Met., 1990, 37, 145.

References

5 19

36. J. Moulton, P. Smith, Synth. Met., 1991, 40, 13. 37. S. Osawa, M. Ito, R. Endo, K. Tanaka, Polym. Commun., 1990, 31, 92. 38. Z. Cai, J. Lei, W. Liang, C. R. Martin, Report TR-62- ONR; order No. AD-A 238841 Avail. NTIS; from Gov. Rep. Announce Index (U.S.) 91; Abstr. No. 164, 370, 51 p. 39. Z. Cai, J Lei, W. Liang, V. Menon, C. R. Martin, Chem. Muter., 1991, 3, 960. 40. X. Xu, H. Ishikawa, A. Kabayashi, M. Satoh, E. Hasegawa, Synth. Met., 1991, 57, 4973. 41. Y. Wei, R. Hariharan, R. Bakthavatchalam, J . Chem. SOC.,Chem. Commun., 1993, 1160. 42. N. Ljungqvist, T. Hjertberg, Synth. Met., 1993, 57, 4944. 43. A. J. Heeger, P. Smith, S. Tokito, J. D. Moulton, PCT h t . Appl. WO 9111325, 1991, USP 90-468737, USP 90-635455. 44. A. J. Heeger, P. Smith, PCT Int. Appl. WO 9220072, 1992, USP 91-697316. 45. I. Mizogushi, M. Ito, Jpn. Kokai Tokkyo Koho, JP 03137121, 1991. 46. M. Kobayashi, R. Otsuka, Y. Nagae, Jpn. Kokai Tokkyo Koho, JP 03137121, 1991, JP 89-158026. 47. R. Otsuka, M. Kobayashi, Jpn. Kokai Tokkyo Koho, JP 02308810,1990. 48. J. S . Park, E. Ruckenstein, Electron Mater., 1992, 21, 205. 49. K.Y. Jen, C. C. Han, R. L. Elsenbaumer, Mol. Cryst. Liq. Cryst., 1990, 186, 211. 50. I. Murase, T. Ohnishi, T. Nogushi, M. Hirooka, Polym. Commun., 1987, 28, 229. 51. A. Andreatta, S. Tokito, P. Smith, A. J. Heeger, Mol. Cryst. Liq. Cryst., 1990, 189, 169. 52. S. Tokito, P. Smith, A. J. Heeger, Synth. Mef., 1990, 36, 183. 53. W. Eevers D. de Schrijver, T. Dierick, C. Peten, J. van der Looy, H. J. Geise, Synth. Met., 1992, 51, 329. 54. S. Tokito, T. Momii, H. Murata, T. Tsutsui, S. Saito, Polymer, 1990, 31, 1137. 55. H. Fuchigami, A. Tsumura, H. Koezuka, Appl. Phys. Lett., 1993,63, 1372. 56. H. Kishiki, Jpn. Kokai Tokkyo Koho, 1992, JP 04057823. 57. C. Zhang, P. Smith, Synth. Met., 1992,46, 235. 58. K. Yamada, I. Nagano, Jpn. Kokai Tokkyo Koho, 1991, JP 03009955. 59. G. Heywang, F. Jonas, Adv. Muter., 1992, 4, 116. 60. M. Dietrich, J. Heintze, G. Heywang, F. Jonas, J. Electrounul. Chem., 1994, 369, 87. 61. F. Jonas, W. Krafft, B. Muys, Macromol. Symposia 1995, 100, 169. 62. F. Jonas, K. Lerch, Kunststofe, 1997, 87, 1401. 63. J. Hupe, G. D. Wolf, F. Jonas, Gulvanotechnik,1995, 86, 3404. 64. H. Yamato, K. Kai, M. Ohwa, T. Asahura, T. Kosh, W. Wernet, Synth. Met., 1996,83, 125. 65. T. Taka, Synth. Met., 1991,41, 1177. 66. W. Wernet, J. Stoffer, EP 358 188, CH 88-3374, 1990. 67. J. E. Osterholm, J. Laakso, P. Nyholm, T. Taka, S. Karjalainen, P. Mononen, PCT Int. Appl. WO 9 009 027, 1990, F189-427, 15 pp. 68. L. Olmedo, P. Raimond, J. Tanguy, EP 473 515, FR 90-10821, 1992. 69. G. Cik, F. Sersen, L. Dlham, L. Szabo, J. Bartus, Synth. Met., 1995, 75, 43. 70. K. Tanaka, H. Ago, T. Yamabe, Synth. Met. 1885, 72, 225. 71. P. Novak, K. Miiller, K. S. V. Santhanam, 0. Haas, Chem. Rev., 1997, 97, 207. 72. Lit. [148a], p.274. 73. H. Miinstedt, E. Voss, Progr. Butt. Sol. Cells, 1987, 6 , 250. 74. R. Bittihn, G. Ely, F. Woeffler, H. Miinstedt, H. Naarmann, D. Naegele, Mukromol. Chem., Makromol. Symp., 1987,8, 51. 75. D. Naegele, R. Bittihn, Solid State lonics, 1988, 28-30, 983. 76. T. Yamamoto, J . Chem. SOC.,Chem. Commun., 1981, 187. 77. T. Yamamoto, M. Zama, A. Yamamoto, Chem. Lett., 1984, 1577. 78. T. Yamamoto, M. Zama, A. Yamamoto, Chem. Lett., 1985, 563. 79. K. Kaneto, K. Yoshino, Y. Inuishi, Jpn. J. Appl. Phys., 1983,22, L 567. 80. R. J. Waltman, A. F. Diaz, J. Bargon, J . Efectrochem. Soc. 1984, 131, 1452. 81. J. H. Kaufman, T.-C. Chung, A. J. Heeger, F. Wudl, J. Electrochem. SOC.,1984, 131, 2092. 82. M. Biserni, A. Marinangeli, M. Mastragostino, J. Electrochem. SOC.,1985, 132, 1597. 83. S. Panero, P. Prosperi, P. Klaptse, B. Scrosati, Electrochim. Acta, 1986, 31,1597. 84. C. Arbizzani, M. Mastragostini, S. Panero, P. Prosperi, B. Scrosati, Synlh. Met., 1989, 28, C 663.

520

10 Application of Electrically Conductive Polythiophenes

J. Guay, A. F. Diaz, J. Y. Bergeron, J . Electroanal. Chem., 1993, 361, 85. C. Arbizzani, M. Mastragostino, Electrochim. Acta, 1990, 35, 251. S. Panero, P. Prosperi, D. Zane, B. Scrosati, J. Appl. Electrochem., 1992,22, 189. G. Mengoli, M. Musiani, D. Pletcher, S. Valcher, J. Appl. Electrochem., 1987, 17,525. T. Kawai, T. Kuwabara, S. Wang, K. Yoshino, J. Electrochem. SOC.,1990, 137, 3793. T. Kawai, T. Kuwabara, S. Wang, K. Yoshino, Jpn. J. Appl. Phys., 1990, 29, 602. F. Li, W. J. Albery, J. Electroanal. Chem., 1991, 302, 279. C. W. Walker, jr., J . Electrochem. SOC.,1991, 138, 1559. C. W. Walker, jr., J . Power Sources, 1991,36, 175. C. W. Walker, jr., Statutory Invent. Regist US 1054, 1992, US 1054. C. W. Walker, S. M. Slane, J . Power Sources, 1991, 36, 561. T. Nagatomo, M. Mitsui, K. Matsutani, 0. Omoto, Electrochem. SOC.Ext. Abstr., 1987,87, 252. 97. T. Nagatomo, M. Mitsui, K. Matsutani, 0. Omoto, Trans ZEZCE, 1987, 70, 346. 98. T. Novak-Nagatomo, 0. Omoto, J . Electrochem. SOC.,1988, 135,2124. 99. A. Czerwinski, C. V. Pham, D. D. Cunningham, L. Laguren-Davidsen, H. Zimmer, B. Mark, Jr., J. Electrochem. SOC.,1986, 133, 576. 100. K. Nishio, M. Fujimoto, N. Yoshinaga, N. Furukawa, 0. Ando, H. Ono, T. Suzuki, Power Sources, 1991, 34, 153. 101. S. Furukawa, N. Yoshinaga, M. Fujimoto, Jpn. Kokai Tokkyo Koho, JP 04137362, 1992. 102. N. Yoshinaga, M. Fujimoto, S. Furukawa, Jpn. Kokai Tokkyo Koho, JP 04104477, 1992. 103. Y. Tsujimoto, M. Ikeda, K. Tsukano, H. Kasama, T. Nishibe, Jpn. Kokai Koho, 03149766, 1991. 104. A. P. Chattaraj, I. N. Basumallick, J. Power Sources, 1991, 36, 557. 105. T. Tajiri, T. Tajiri, Jpn. Kokai Tokkyo Koho, JP 04332466,1992. 106. H. Ueno, Jpn. Kokai Tokkyo Koho, JP 04237948,1992. 107. I. Nakane, Y. Fujita, S. Furukawa, Jpn. Kokai Tokkyo Koho, JP 04028172,1992. 108. T. Kanbara, Y. Sato, Y. Uemachi, T. Tonomura, K. Takejama, Jpn. Kokai Tokkyo Koho, JP 0426363, 1992. 109. K. Tokeda, S. Ido, T. Noda, Jpn. Kokai Tokkyo Koho, JP 04253170,1992. 110. K. Nawa, Jpn. Kokai Tokkyo Koho, JP 04108889,1992. 111. T. Kawai, T. Iwasa, M. Onada, T. Sakamoto, K. Yoshino, J. Electrochem. SOC.,1992, 139, 3404. 112. H. Koshina, H. Okuno, T. Morita, Jpn. Kokai Tokkyo Koho, JP 04002065,1992. 113. K. Shinozaki, A. Kabumoto, H. Sato, K. Watanabe, H. Umemura, S. Tanemura, Synth. Met., 1990,38, 135. 114. Bayer AG, Product Information, Product AZ 4060, 1995, 02/95. 115. A. Rudge, I. Raistrick, S. Gottesfeld, J. P. Ferraris, Electrochim. Acta, 1994, 39, 273. 116. A. Rudge, J. Davey, J. Raistrick, S. Gottesfeld, J. P. Ferraris, J. Power Sources, 1994, 47, 89. 117. L. Libera, K. Eulenberger, G. Hahn, DDR Patent DD 291659,1991. 118. G. Schopf, G. Kossmehl, Adv. Polymer Sci., 1997, 129, 124, lit. 845-879. 119. M. Kobayashi, N. Naito, S. Kawakami, Jpn. Kokai Tokkyo Koho, JP 031 14213,1991. 120. M. Kono, I. Isa, Jpn. Kokai Tokkyo Koho, JP 02238613,1990. 121. Y. Kudo, T. Kojima, S. Tsuchiya, S. Yoshimura, Jpn. Kokai Tokkyo Koho, JP 02074017, 1990. 122. K. Kuranuki, J. Ozaki, Y. Oashima, Jpn. Kokai Tokkyo Koho, JP 02066921,1990. 123. F. Jonas, G. Heywang, Electrochim. Acta, 1994, 39, 1345. 124. A. Yasuda, Jpn. Kokai Tokkyo Koho, JP 04109613,1992. 125. C . Arbizzani, M. Mastragostino, L. Meneghello, R. Paraventi, Adv. Muter., 1996,8, 331. 126. A. Rudge, I. Raistrick, S. Gottesfeld, J. P. Ferraris, Electrochim. Acta, 1994, 39, 273. 127. C. Arbizzani, M. Mastragostino, L. Meneghello, Electrochim. Acta, 1996, 41, 21. 128. S. Miyauchi, Y. Goto, I. Tsubata, Y. Sorimachi, Synth. Met., 1991,41, 1051. 129. S. Morita, A. A. Zakhidov, K. Yoshino, Jpn. J . Appl. Phys., Part 2, 1993, 32, L 873. 130. W. Torres, M. A. Fox, Chem. Muter., 1990, 2, 306. 131. K. Yoshino, T. Kuwabara, Y. Manda, S. Nakajima, T. Kawai, Jpn. J. Appl. Phys., Part 2, 1990,29, L 1716.

85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96.

References

52 1

M. S. A. Abdou, G. A. Diaz-Guijada, M. I. Arroyo, S. Holdcroft, Chem. Mafer.,1991,3,1003. S. Holdcroft, M. S. A. Abdou, Can. P 2070043,1992; USP 91-707815. Y. Tachibana, Y. Sakurai, M. Yokojama, Chem. Lett. ( 6 ) , 1994, 1119. K. Yoshino, H. Takahashi, R. Sugimoto, Jpn. J. Appl. Phys. Part 2, 1991, 30, L 657. T. Tanaka, H. Doi, Jpn. Kokai Tokkyo Koho, JP 02127642, 1990. C. Kutsche, J. Targove, P. Haaland, J. Appl. Phys., 1993, 73, 2602. H. Anayama, T. Koyama, N. Fujimura, Jpn. Kokai Tokkyo Koho, JP 03039753, 1991. H. Komon, J. Hanna, K. Hoshino, J. Nakazawa, Y. Nakazawa, Jpn. Kokai Tokkyo Koho, JP 02269360,1990. 140. H. Anayama, N. Fujimura, T. Koyama, Jpn. Kokai Tokkyo Koho, JP 03080256, 1991. 141. H. Ono, A. Saida, Jpn. Kokai Tokkyo Koho, JP 04254860,1992. 142. M. Fuse, S. Otsuka, Jpn. Kokai Tokkyo Koho, JP 04098263,1992. 143. 0. Nabeta, N. Kosho, Jpn. Kokai Tokkyo Koho, JP 03192365,1991. 144. W. S. Huang, Polymer, 1994, 35, 4057. 145. T. Nishida, T. Tamamura, Jpn. Kokai Tokkyo Koho, JP 03216655, 1991. 146. Y. Kawasaki, K. Saito, S. Shiba, K. Watabe, Y.Yoneda, Jpn. Kokai, Tokkyo Koho, JP 02039156, 1990. 147. H. Horibe, H. Fichigami, S. Tanaka, T. Kumada, S. Kubota, Y. Hizuka, Jpn. Kokai Tokkyo Koho, JP 04212962, 1992, USP 90-310375. 148. M. Angelopoulos, W. S. Huang, R. D. Kaplan, M. A. le Corre, S. E. Perreault, J. M. Shaw, M. R. Tissier, G. F. Walker, EP 399299, 1990; USP 89-357565. 149. S. X. Cai, M. Kanskar, J. C. Nabity, J. F. W. Keana, M. N. Wybourne, J . Vac.Sci. Technol. B, 1992, 10,2589. 150. S. X. Cai, J.F.W. Keana, J. C. Nabity, M. N. Wybourne, J. Mol. Electron, 1991, 7, 63. 151. G. Kaempf, M. Feldhues, U. Scheunemann, J. Lingnau, R. Dammel, DEP 3903421, 1990. 152. K. Hshimoto, T. Koizumi, K. Kawakita, N. Nomura, EP 348961, 1990, JP 88-161640, JP 88-161644. 153. K. Hashimoto, N. Nomura, Jpn. Kokai tokkyo Koho, JP 03137649, 1991. 154. T. Asuke, K. Takemura, Jpn. Kokai Tokkyo Koho, JP 04202434,1992. 155. S. Miyashita, Jpn. Kokai Tokkyo Koho, JP 03218068,1991. 156. M. Koga, F. Ebisawa, Jpn. Kokai Tokkyo Koho, JP 04323217,1992. 157. M. Mogi, T. Nagase, Jpn. Kokai Tokkyo Koho, JP 03112683, 1991. 158. H. Ito, Jpn. Kokai Tokkyo Koho, J P 03294364,1991. 159. J. S. Facci, H. J. Yuh, W. W. Limburg, S. S. Badesha, EP 435592, 1991. 160. A. Sibbalt, S. Taylor, G. K. Chandler, EP 371657, 1990; GBP 88-27661. 161. G. Bidan, S. Chardon, E. Hannecart, R. Massart, EP 47351 1,1992, FRP 90-10883. 162. M. Niemitz, G. Kossmehl, Angew. Makromol. Chem., 1991, 1851186, 147. 163. G. Kossmehl, 1994, 5Ih European Polymer Federation Symposium on Polymeric Materials, Basle, Switzerland. 164. G. Kossmehl, D. Kabbeck-Kupijai, M. Niemitz, Chiu Z, 1990, 24, 106. 165. W. Rauh, PhD Thesis, Freie Universitaet Berlin, 1993. 166. W. Scheer, H. Fuhrmann, G. Kossmehl, M. Niemitz, D. Kabbeck-Kupijai, MAN Technologie GmbH, DEP 3705439 A l , 1988. 167. C. D. Henry, H. Tebbe, G. Wegner, F. Armand, A. Ruaudel-Teixier, Adv. Muter., 1997,9,805. 168. M. Morita, Makromol. Chem. , Macromol, Chem. Phys., 1993, 194, 2361. 169. D. E. Singleton, K. C. Ho, C. B. Greenberg, T. G. Rukavina, J. B. Slobodnik, 1991, EP 417624, USP 89-406477, USP 89-415001, USP 89-440832. 170. S. Panero, S. Passenni, B. Scrosati, Mol. Cryst. Liq. Cryst. 1993, 230, 337. 17 1. M. Mastragostini, C. Arbizzani, A. Bongini, G. Barbarella, M. Zambianchi, Electrochim. Acta. 1993, 38, 135. 172. T. Kanbara, Y. Sato, Y. Uemachi, T. Tonomura, K. Takeyama, 1992, Jpn. Kokai Tokkyo Koho, JP 0426363. 173. K. Nawa, 1992, Jpn. Kokai Tokkyo Koho, J P 04108889. 174. J. H. Helms, 1992, USP 5128013. 175. J. C. Gustafson, 0. Inganaes, a. M. Anderson, Proc. SPZE-Znt. SOC.Opt. Eng., 1992, 1728, 142.

132. 133. 134. 135. 136. 137. 138. 139.

522

I 0 Application of Electrically Conductive Polythiophenes

K. Yoshino,K. Kaneto, Y. Inuishi, Jpn. J. Appl. Phys., 1983, 22, L157. K. Kaneto, K. Yoshino, Y. Inuishi, Jpn. J. Appl. Phys., 1983,22, L 412. F. Garnier, G. Tourillon, M. Gazard, J. C. Dubois, J. Electroanal. Chem., 1983, 148, 299. K. Hyodo, Electrochim. Acta, 1994, 39, 265. H. Yashima, M. Kobayashi, K. B. Lee, D. Chung, A. J. Heeger, F. Wudl, J . Electrochem. SOC. 1987, 134,46. 181. P.Marque, J. Roncali, P. Garnier, Extended abstracts of the 3gfhMeeting of the Znternational Society f o r Electrochemistry, 14- 18 Sept. 1987, Maastricht, Netherlands. 182. M. Mastragostino, A. M. Marinangeli,A. Corradini, B. Giacobbe, Synth. Met., 1989,28, C501. 183. J. Roncali, R. Garreau, A. Yassar, P. Marque, F. Garnier, M. Lemaire, J. Phys. Chem., 1987, 91, 6706. 184. R. L. Blankespoor, L. L. Miller, J. Chem. SOC.,Chem. Commun., 1985, 90. 185. B. Wessling, 1990, Ger. O f e n . DEP 3834526. 186. J. C. Gustafsson, B. Liedberg, 0. Inganaes, 1994, Solid State Ionics 69, 145. 187. J. C. Gustafsson-Carlberg, 0.Inganaes, M. R. Anderson, C. Booth, A. Azens, G. Granqvist, Electrochim. Acta, 1995, 40, 2233. 188. S. A. Sapp, G. A. Sotzing, J. L. Reddinger, J. R. Reynolds, Adv. Mat., 1996,8, 808. 189. C. Botta, G. Zhuo, 0. M. Gelsen, D. D. C. Bradley, A. Musco, Synth. Met., 1993, 55, 85. 190. J. P. Ferraris, C. Henderson, D. Torres, D. Meeker, Synth. Met., 1995, 72, 147. 191. Y. Ohseda, I. Imae, N. Noma, Y. Shirota, Synth. Met., 1996, 81, 157. 192. G. Tourillon, J. Garnier, J. Phys. Chem., 1984, 88, 5281. 193. A. Yassar, J. Roncali, F. Garnier, J . Electroanal. Chem., 1988, 255,53. 194. N. F. Atta, A. Gala, A. E. Karagozler, H. Zimmer, J. F. Rubinson, H. B. Mark, Jr., Chem. SOC.,Chem. Commun., 1990, 1347. 195. M. Lemaire, D. Delabouglise, R. Garreau, A. Guy, J. Roncali, J. Chem. SOC., Chem. Commun., 1988, 658. 196. J. Roncali, R. Garreau, D. Delabouglise, F. Garnier, M. Lemaire, J. Chem. SOC.,Chem. Commun., 1989, 679. 197. J. W. Thackeray, H. S . White, M. S. Wrighton, J. Phys. Chem., 1985, 89, 5133. 198. E. P. Lofton, J. W. Thackeray, M. S . Wrighton, J. Phys. Chem., 1986,90,6080. 199. T. Hanawa, S. Kuwabata, H. Hashimoto, H. Yoneyama, Synth. Met., 1989,30, 173. 200. Y. Ohmori, M. Uchida, K. Muro, C. Morishima, K. Yoshino, Mol. Cryst. Liq. Cryst., 1993, 227, 285. 201. S.-R. Kim, %-A. Choi, J.-D. Kim, K. J. Kim, C. Lee, S. B. Rhee, Synth. Met., 1995, 71,2027. 202. G. Zotti, G. Schiavon, A. Berlin, G. Pagani, Synth. Met., 1993, 60, 179. 203. R. Oedegaard, S. M. Hesjevik, G. Hagen, P C T I n t . Appl. WO 920219, No. 90-3011,1992, 204. G. Hagen, A. H. Toresen, S. Sunde, S . M. Hesjevik, R. Oedegaard, Mol. Cryst. Liq. Cryst.1990, 189, 213. 205. W. Zhang, S. Dong, J. Electroanal. Chem. Interfacial Electrochem., 1990, 284, 517. 206. Z. L. Xue, A. E. Karagozler, 0. Y. Ataman et al., Electroanalysis, 1990, 2, 1. 207. A. E. Karagozler, 0. Y. Ataman, A. Galal, Z.-L. Xue, H. Zimmer, H. B. Mark, Jr., Anal. Chim. Acta, 1991,248, 163. 208. J. Bobacka, A. Levenstam, A. Ivaska, Talanta, 1993,40, 1437. 209. J. Bobacka, M. McCarrick, A. Levenstam, A. Ivaska, Analyst, 1994, 119, 1985. 210. Z. Wang, A. Galal, H. Zimmer, H. B. Mark, Jr., Electroanalysis, 1992, 4 , 77. 21 I . N. F. Atta, A. Galal, A. E. Karagozler, H. Zimmer, J. F. Rubinson, H. B. Mark, Jr., J. Chem. Soc., Chem. Commun., 1990, 1347. 212. N. F. Atta, A. Galal, A. E. Karagozler, G. C. Russel, H. Zimmer, H. B. Mark, Jr., Biosens. Bioelectron, 1991, 6, 333. 213. H. S. 0. Chen, S. C. Ng, S. H. Seow, Synth. Met., 1994, 66, 177. 214. M. Lapkowski, M. D. Levi, Synth. Met., 1992,51, 75. 215. M. D. Levi, M. Lapkowski, Electrochim. Acta, 1993, 38, 271. 216. R. G. Sandberg, L. J. van Houten, J. L. Schwartz et al.,A C S Symp. Ser. 511, 1992, 81. 217. G. G. Wallace, P C T Znt. Appl. WO 9002829, A U 88-294, 1990. 218. L. A. Samuelson,D. L. Kaplan, J. 0.Lim, M. Kamath, K. A. Marx, S. K. Tripathy, Thin Solid Films, 1994, 242, 50. 176. 177. 178. 179. 180.

References

523

219. E. Agosti, G. Zerbi, Synth. Met., 1996, 79, 107. 220. I. Willner, E. Katz, N. Lapidot, P. Bauerle, J . Electroanal. Chem. Bioelectrochem. Bioenerg., 1992, 29, 29. 221. P. Bauerle, M. Hiller, S. Scheib, M. Sokolowski, E. Umbach, Adv. Muter., 1996, 8, 214. 222. M. Hiller, C. Kranz, J. Huber, P. Bauerle, W. Schuhmann, A h . Muter., 1996, 8, 219. 223. W. Schuhmann, in: Diagnostic Biosensor Polymers (Eds.: A. M. Usmani, N. Akmal), Symp. Ser., 1994, 556, 110-123. 224. H.-P. Welzel, G. Kossmehl, H.-J. Stein, J. Schneider, W. Plieth, Electrochimica Acta, 1995,40, 577. 225. H.-P. Welzel, G. Kossmehl, G. Engelmann, B. Neumann, U. Wollenberger, F. Scheller, Macromol. Chem. Phys., 1996, 197, 3355. 226. G. Li, G. Kossmehl, H.-P. Welzel, G. Engelmann et al.. Macromol. Chem. Phys., 1998, 199, in press. 227. G. Li, G. Kossmehl, H.-P. Welzel, G. Engelmann et al., Macromol. Chem. Phys., 1998, 199, in press. 228. H.-P. Welzel, G. Kossmehl, J. Schneider, W. Plieth, Macromolecules, 1995, 28, 5575. 229. H.-P. Welzel, G. Kossmehl, H. Boettcher, G. Engelmann, W.-D. Hunnius, Macromolecules 1997,30, 7419. 230. H.-P. Welzel, G. Kossmehl, G. Engelmann et al., Macromol. Symp., 1997, 126, 283. 231. N. S. Sariciftci, M. Mehring, K. U. Gdudi, P. Bauerle, H. Neugebauer, A. Neckel, J . Chem. Phys., 1992, 96, 7164. 232. P. Bauerle, Adv. Mater., 1993, 5 , 879. 233. S. Roth, One-dimensional Metals, VCH, Weinheim, 1995, p. 225. 234. I. Stassen, T. Sloboda, G. Hambitzer, Synth. Met., 1995, 71, 2243. 235. S. Roth, One-dimensional Metals, VCH, Weinheim, 1995, p. 219. 236. D. W. DeBerry, J. Electrochem. Soc., 1985, 132, 1022. 237. B. Wessling, Adv. Muter., 1994, 6, 226. 238. H. Hosekawa, K. Kamada, EP 473515,1992, FRP 90-10821. 239. K. Mito, S. Hirayama, T. Motomura, M. Tobitaka, Jpn. Kokai Tokkyo Koho, JP 02218757, 1990. 240. S. Roth, One-dimensional metals, VCH, Weinheim, 1995, p. 228. 241. M. R. Anderson, B. R. Mattes, H. Reiss, R. B. Kaner, Science, 1991, 252, 1412. 242. D. L. Feldheim, C. M. Elliott, J. Membr. Sci., 1992, 70, 9. 243. T. Shimizu, K. Honda, T. Yada, H. Toyoda, Jpn. Kokai Tokkyo Koho, JP 02127472, 1990. 244. R. L. Blankespoor, L. L. Miller, J . Chem. SOC.,Chem. Commun., 1985, 90. 245. H. Laborde, J. M. Leger, C. Lami, F. Gamier, A. Yassar, J. Appl. Electrochem., 1990, 20,524. 246. S. Swathirajan, Y. M. Mikhail, J . Electrochem. Soc., 1992, 139,2105. 247. M. D. Levi, N. M. Alpatova, E. V. Ovsyannikova, M. A. Vorotyntsev, J . Electroanal. Chem., 1993,351, 271. 248. T. Kawai, T. Kuwabara, K. Yoshino, J . Chem. Soc., Faraday Trans., 1992,88, 2041. 249. T. Kawai, H. Motobayashi, T. Kuwabara, K. Yoshino, Technol. Rep. Osaka Univ., 1991, 41 (2053-2075): 325. 250. K. Kaneto, T. Asano, W. Takashima, Jpn. Appl. Phys., Part 2, 1991,30, L 215. 251. W. Takashima, K. Sasano, T. Asano, K. Kaneto, Polym. Znt., 1992, 27, 249. 252. Y. Kishimoto, S. Soga, Jpn. Kokai Tokkyo Koho, JP 03222476,1991. 253. Y. Kishimoto, Jpn. Kokai Tokkyo Koho, JP 03055878., 1991. 254. M. Onoda, H. Nakayama, S. Morita, K.Yoshino, J. Electrochem. Soc., 1994, 141, 338. 255. M. Vardhireddy, R. S. Bhide, S. V. Bhoraskar, J. Appl. Polym. Sci., Appl. Polym Symp. (Plasma Polym. Plasma Interact. Polym. Mater.), 1990, 46, 193. 256. J. E. Osterholm, J. Laakso, P. Nyholm, T. Taka, S. Karjalainen, P. Mononen, PCTZnt. Appl. WO 9009027, 1990, FI. Patent 89-427. 257. S. Miyashita, Jpn. Kokai Tokkyo Koho, JP 03126275, 1991. 258. S. Miyashita, Jpn. Kokai Tokkyo Koho, JP 03135079, 1991. 259. A. Moshizuki, T. Yoshihara, Jpn. Kokai Tokkyo Koho, JP 03235919,1991. 260. S. Ono, E. Funatsu, N. Imamura, Jpn. Kokai Tokkyo Koho, JP 04367823,1992.

524

10 Application of Electrically Conductive Polythiophenes

261. H. Takanashi, K. Akimoto, K. Hayashifuji, A. Yoshimatsu, A. Kondo, Jpn. Kokai Tokkyo Koho, JP 04136815,1992. 262. K. Itani, Jpn. Kokai Tokkyo Koho, JP 04301604,1992. 263. K. Yoshino, S. Morita, K. Nakao, Synth. Met., 1991, 41, 1039. 264. S. Shakuda, S. Morita, T. Kawai, K. Yoshino, Jpn. J. Appl. Phys., Part 1, 1993, 32, 5143. 265. D. Fichou, J. M. Nunzi, F. Charra, N. Pfeffer, Adv. Mat., 1994,6, 64. 266. A. T. Ponomarenko, V. G. Shevshenko, N. S. Enikolopyan, Adv. Polym. Sci., 1990,96, 125. 267. E. Genies, Intrinsically Conducting Polymers - An Emerging Technology, (Ed. M. Aldissi), Kluwer Academic Publishers, Dordrecht, 1993, p. 75 . 268. S. Kivelson, A. J. Heeger, Synth. Met., 1988, 22, 371

Index

Absorption spectra - poly(3-hexylthiophene) 52, 53, 54, 56, 69 - poly(3-methylthiophene) 68 - polythiophene 54 Aggregation 417f Alkoxy polythiophenes 25 - chemical synthesis 25ff - as conducting polymer sensors 27ff All-optical modulator and memory devices 449ff All-printed plastic field-effect transistors 481ff Amphiphilic oligothiophenes 170f Anthracene 406 Antistatic coatings 495f Applications 491ff - antistatic coatings 495f - conducting materials 492ff - electrochromic devices 503ff - electromagnetic shielding materials 496f junction devices and rectifying bilayer electrodes 501 - rechargeable batteries and capacitors 497ff resists, recording materials and fabrication of patterns 501ff - sensors 506ff - - for bio-organic materials 508ff _ _ for gases 506f - - for ions in aqueous solution 507f _ - for organic materials 508 - see also Electro-optical polythiophene devices; Field-effect transistors; Sensors Arene/arene coupling 93ff chemical/electrochemical oxidative coupling 96 - copper(I1)-promoted oxidative coupling 93ff - nickel- and palladium-catalyzed cross coupling 98ff 96 - oxidative coupling of organoboranes - Stille reaction 102f Suzuki reaction 103 - transition metal catalyzed coupling 97f Aromatization of tetrahydrothiophenes 104f

Atomic force microscopy 48 Austin Model 1 320 Band model of charge transport 284ff Bimorphs 515 Bio-organic material sensors 508ff a-Bithiophenes - P-substituted 128ff X-ray structure 187ff - crystal data 189 - - crystal packing 191, 192, 194 - - 4,4’-dimethyl and 4,4’bis(methylsulphanyl)-2,2’-bithiophenes 190 - - 3,3’- and 4,4’-dipentoxy-2,2’-bithiophenes 190 Branched alkyl polyalkylthiophenes 21ff Bridged cyclopentathiophenes 137 Bridged oligothiophenes 137ff ~

~

~

~

~

~

Cadio-Chodkiewicz procedure 107, 108 Capacitors 500 Carboxylic acid derivatives 33f Carrier density 308f Carriers - injection 267ff - photoinduced 265f - recombination 63ff Catalysts 15 Chain length - and charge mobility 307f - singlet states 368f and third-order polarizability of thiophene oligomers 349ff Charged oligothiophenes 257ff - electronic and linear optical properties 333ff Charge excitations in polythiophenes 57ff Charge storage configurations 65ff Charge transport - in polythiophenes 60ff in semiconducting oligothiophenes 283ff - - basicmodels 284ff - - bandmodel 284f ~

~

526

Index

Charge transport (continued) - - hopping 288ff - - multiple trapping 297 - - polarons 29% - measurement of mobility 298ff - - conductivity 298f - - field-effect 305f - - space-charge-limited current 300ff - - time of flight 299f - transport properties 306ff - - conductivity, mobility and carrier density 307ff - - traps 311f - see also Conductivity Chemical doping 60 Chemical oxidative coupling 99 Chemical synthesis - alkoxy polythiophenes 25ff - heteroatomic functionalized substitutents 24ff - - alkoxy polythiophenes 25ff - - carboxylic acid derivatives 33f - - chiral substituents 32f - - conducting polymer sensors 27ff - polyalkylthiophenes 5ff branched alkyl polyalkylthiophenes 21f 6ff - ferric chloride method - - metal catalyzed cross-coupling polymerizations 5f - phenyl sidechains 24 - regioregular, head-to-tail coupled polyalkylthiophenes l2ff - - IR and LJV-Vis 16ff - - McCullough method 12f - - mechanism and catalyst choice 15 - - NMR characterization 15f - - random copolymers of alkyl thiophenes 20f - - Riekemethod 13f - - self-assembly, X-ray and electrical conductivity 18f - regioregular polyalkylthiophenes 9ff - - HH-TTandTT-HHPATs 10ff - unsubstituted polythiophene 2ff Chemistry of oligothiophenes 89ff - synthesis - - first oligothiophenes and derivatives 90ff - - substituted oligothiophenes 118ff - - unsubstituted oligothiophenes 93ff Chemistry of polythiophenes Iff - chemical synthesis - - heteroatomic functionalized substituents 24ff - - polyalkylthiophenes 5ff - - unsubstituted polythiophene 2ff - fused ring systems 38f -

-

Chemoselective sensors 28, 3 1 Chiral alkyl polythiophenes 23 Chiral substituents, polythiophenes 32f Conducting materials 4928 Conducting polymers 44ff - carrier recombination 63ff - charge excitations in polythiophenes 57ff - charge transport in polythiophenes 60ff - spectroscopic studies of charged state 65ff - - charge storage configurations 65ff properties in solutions 73ff - structure and conformation 48ff - - conformational features 51ff - - morphology and crystal structure 48ff Conductivity 29Sf, 307ff - carrier density 308f - head-tail polyalkylthiophenes 18f head-to-tail coupled polyalkylthiophenes 18f - influence of structure on 269f - photoconductivity 220, 274 - variation with chain length 307f - variation with temperature 309ff - see also Charge transport Conformation of polythiophenes 51ff Conjugated oligothiophenes, electronic excited states 361ff - electronic structure 362f - - polythiophene 363f - model molecules 364ff - - molecular structure 364ff - - singlet states 367ff - - triplet states 372 - polarized electroluminescence 395ff - solid state properties 373ff - - lowest excited states 380ff - - molecular packing 373f - - theoretical approach 374ff Copper(I1)-promoted oxidative coupling 93ff Coronene 406 Crystal structure of polythiophenes 488 Current-voltage curves 267f Cyclization diacetylenes 107ff - 1,4-diketones 105ff Cyclopolyarenes 95 -

-

-

Davidov blue shift 149 Davydov splitting 417f Dehalogenation polymerization 6 Demercuration polymerization 6 Density Functional Theory method 321 Diacetylenes, cyclization 107ff Dichroic IR spectra, poly(3-hexylthiophene) 71, 72

Index Dielectric relaxation spectra, poly(3-hexylthiophene) 77 1,4-Diketones, cyclization of 105ff 1,4-Dithiins 11Off Domain-wall concept 45 Donor/acceptor-substituted oligothiophenes 166ff Dopant effects 272ff - on conductivity 272ff polythiophenes 245ff - sexithiophenes 244 - small oligomer thin films 226 Doped insulators 283 Doping 51, 60 - chemical 60 electrochemical 60 - n-type 263ff p-type 260ff -

-

-

molecular packing 373f theoretical approach 374ff Electronic properties of oligothiophenes 3l7ff - charged oligothiophenes 333ff - metal/polymer interfaces 339ff - - electronic structure 343f - - geometric structures 340ff - - vibrational signature 344ff - nature of lowest excited states 322ff neutral oligothiophenes 322f - - effects of substitution 328ff _ _ intersystem crossing processes 324ff _ _ lattice relaxation phenomena 326ff nature of lowest excited states 322ff _ _ nonlinear optical properties 347ff - theoretical methodology 286f Electronic properties of polythiophenes 46ff carrier recombination 63ff charge excitations 57ff - charge transport 60ff conducting polymers 44ff electroluminescence 64f photoluminescence 64f spectroscopic studies 65ff charge storage configurations 65ff properties in solutions 73ff spectroscopic studies of charged states 65ff - - charge storage configurations 65ff _ _ properties in solutions 73ff structure and conformation 48ff - - conformational features 51ff - - morphology and crystal structure 48ff Electronic structure 343f Electron polaron 294 Electro-optical modulator devices 444ff optical probing of field-induced charge 446ff Electro-optical polythiophene devices 405ff - all-optical modulator and memory devices 449ff - electroluminescent devices 421ff - - history of organic LEDs 424ff _ _ LEDs based on oligo(thiophenes) 427ff - - LEDs based on poly(thiophenes) 429ff electronic excitations in oligo-(thiophenes) 413ff _ _ inter-molecular non-radiative decay channels 413ff - - inter-ring torsion and coplanarity of oligomers 419ff electro-optical modulator devices 444ff _ _ optical probing of field-induced charge 446ff - - -

-

-

-

EDOT 27 Eglington coupling 107 Electroabsorption spectroscopy 56 Electrochemical doping 60 Electrochemical oxidative coupling 96 Electrochemical voltage spectroscopy 66 Electrochromic devices 503ff Electroluminescence 63ff - polarized 395ff - see also Electroluminescent devices Electroluminescent devices 421ff - history of organic LEDs 424ff LEDs based on molecular semiconductors 390ff _ _ polymeric LEDs 425ff - LEDs based on oligo(thiophenes) 427ff LEDs based on poly(thiophenes) 429ff - - intrinsically-polarized polymer LEDs 432ff _ _ poly(thiophenes) in microcavity structures 434ff - - sub-wavelength size polymer LEDs 436ff _ _ visible spectrum 430ff _ _ voltage-controlled colours 437ff Electromagnetic shielding materials 496ff Electronic excited states 361ff electronic structure 362ff _ _ polythiophenes 363f - model molecules 364ff - - molecular structure 364ff _ _ singlet states 367ff - - triplet states 372 polarized electroluminescence 395ff - solid state properties 373ff - - lowest excited state 380ff -

-

-

-

-

527

-

-

-

-

-

-

~

-

-

-

-

528

Index

Electro-optical polythiophene devices (continued) - photoconductive/photovoltaic devices 439ff - - photoconductivity in sexithiophene 440f - preparation of thin f ilm devices 408ff - - blends between polymers and oligomers 413 - - oligomers 409f - - polymers 408f - - solubility 410ff - relationship between polymers and oligomers 405ff - see also Applications; Sensors End-capped oligothiophenes 145, 147ff Enzyme electrodes 512 3,4Ethylenedioxythiophene 27 Excitons 374ff, 41 8f Extended states 379f, 392ff Extrinsic semiconductors 283 Ferric chloride polymerization 6ff, 20 - PEDOT 27 - polythiophene zipper sensors 35 Field-dependent mobility 291 Field-effect 305f Field-effect transistors 3 19,459ff - heterojunction 483ff - oligothiophene 468ff - - fused ring materials 475f - - modelling of 461ff - - morphology 471ff - - substituted oligothiophenes 473ff - - synthesis and purification 468ff - operation of 460f - polythiophene 4768 - - all-printed plastic FETs 481ff - - regiorandomFETs 478 - - regioregular FETs 478ff Fluoroalkyl substituted polythiophenes 23 Franck-Condon vibronic coupling 370ff Franz-Keldysh effect 56 Frequency-domain electric birefringence spectroscopy 75, 76, 77 FriedelLCrafts reaction 107f Fused ring materials 575f Fused ring systems 38f Gas sensors 506f Geometric properties of oligothiophenes 317ff - charged oligothiophenes 333ff - metal/polymer interfaces 339ff electronic structure 343f -

- _ -

geometric structures 340ff

- vibrational signature 344ff

nature of lowest excited states 322ff neutral oligothiophenes 322ff - - effects of substitution 328ff - - intersystem crossing processes 324ff - - lattice relaxation phenomena 326ff - - nature of lowest excited states 322ff - - nonlinear optical properties 357ff - - theoretical methodology 320f Geometric structures 340ff Glaser coupling 107, 108

-

Hartree-Fock semiempirical Austin Model 1 320 Hay coupling 107 Head-to-head couding 9, lOff IR and UV-Vis- 1Zff McCullough method 12f mechanism and catalyst choice 15 NMR characterization 15f polythiophenes 29 random copolymers 20 - head-to-tail coupled 20f Rieke method 13f self-assembly, X-ray and electrical conductivitv 18f Head-to-tail coupling 9 Heteroatomic functionalized substituents 24ff - alkoxy polythiophenes 25ff - - as conducting polymer sensors 27ff - carboxylic acid derivatives 33f - chiral substituents 32f Heterojunction field-effect transistors 483ff Highly ordered systems 387ff Hopping 288ff - field dependent mobility 291 - localization 288 - temperature dependence 289f Huang-Rhys parameter 367 Induced-rigidity concept 57 Infrared characterization 16ff Interchain electronic processes 47, 62, 64 Interface effects 464f Internal conversion 415f Inter-ring torsion 419ff - solid state 420f solution 420 Intersystem crossing processes 324ff, 416f Intrachain electronic processes 47, 62 Intrinsically-polarized polymer LEDs 432ff Ion sensors 507f -

Index

IR-active vibrational modes 71 Irregular polyalkylthiophenes 8, 9 Junction devices 501 Kharash reaction 98 Kumada reaction 5, 98ff, 124, 131, 141, 145, 158, 159 Lattice polarons 294 Lattice relaxation 326ff Lawesson’s reagent 106 Light-emitting diodes 318ff - oligo(thiophene) based 427ff - organic, history of 424ff LEDs based on molecular semiconductors 424f - - polymeric LEDs 425ff - poly(thiophene) based 429ff - - intrinsically-polarized polymer LEDs 4328 - - microcavity structures 434ff - - sub-wavelength size polymer LEDs 436f - - visible spectrum 430ff - - voltage-controlled colours 437ff - see also Electronic properties of oligothiophenes; Electro-optical polythiophene devices Lin/Dudek synthesis 2 Lithium batteries 499 Lithium/sulphur dioxide batteries 498 Lowest excited states 322ff, 380ff - extended states 392f - highly ordered systems 387ff - optical properties of polycrystalline thin films 384f - state ordering 394f - structure/morphology of polycrystalline thin films 380ff - - morphology 383ff - - orientation 381f - structure 382f triplet states 393f - two photon excitation 392 Lowest singlet transition 369f -

-

McCullough method 12f Mannich reaction 105 Metal catalyzed cross-coupling polymerizations 5f Metal-insulator-semiconductor devices 69 Metal/polymer interfaces 339ff - electronic structure 343f geometric structures 340ff - vibrational signature 34ff -

3-Methylthiophene 20 Microcavity structures 434ff Mobility 307ff - carrier density 308f - field dependent 291 measurement 298ff - - conductivity 298ff - - space-charge-limited current 300ff - - time of flight 299f - structure and 270f - variation with chain length 307f variation with temperature 309ff Model systems char& transport in semiconducting oligothiophenes 284ff bandmodel 284ff - hopping 288ff - multiple trapping 297 - polarons 291ff oligothiophene field-effect transistors 461ff - analytical modelling 461ff - energy levels 467f - interface effects 464f numerical modelling 463f - short-channel effects 465f - sub-threshold characteristics 466ff oligothiophenes 364ff - molecular structure 364ff - singlet states 367ff ModifiedY Neglect of Differential Overlap method 51, 320f Molecular ‘nearly small’ polarons 258f Molecular packing 373f Molecular wires 5 17 Monothiophene thin films 222 Morphology of polythiophenes 48ff Multiple trapping 297 MultiReference Double-Configuration Interaction scheme 321 -

-

-

-

Neutral oligothiophenes electronic and linear optical properties 322ff - - effects of substitution 328ff - - intersystem crossing processes 324ff - lattice relaxation phenomena 326ff - - nature of lowest excited states 322ff - nonlinear optical properties 347ff - - chain length dependence 349ff - - dynamic third-order response of Th7 352ff - see also Oligothiophenes Nickel-catalyzed cross coupling 98ff Nomenclature of polythiophenes 91ff -

-

529

530

Index

Non-radiative decay channels - inter-molecular 4178 - - aggregation and Davydov splitting 4 17f - - charge-transfer excitons 418f - intra-molecular 413ff - - internal conversion 415f - - intersystem crossing 416f - - singlet fission 417 n-type doping 263ff Nuclear magnetic resonance characterization 15f a-Octithiophenes optical and electrical properties 216ff - - photoconductivity 230 - - stimulated emission 216ff - - UV-visible absorption 216 - X-ray structure 2078 - - optical micrographs 210 3-Octylthiophene 20 Oligo-p-phenylenes, electronic properties 93 Oligothiophenes 89ff - amphiphilic 170ff - bridged 137 - charged 257ff - - electronic and linear optical properties 333ff - conjugated, electronic excited states of 361ff - - electronic structure 362ff - - modelmolecules 364ff - - polarized electroluminescence 395ff - - solid state properties 373ff - correlated electrochemical potentials 93 - donor/acceptor-substituted 1668 - electron affinities 93 - electronic excitations in 413ff - - inter-molecular non-radiative decay channels 4 17ff - - inter-ring torsion and coplanarity of oligomers 419ff - - intra-molecular non-radiative decay channels 413ff - end-capped 145, 148ff - field-effect transistors 468ff - - fused ring materials 475f modellingof 461ff - - morphology 471ff - - substituted oligothiophenes 473ff - - synthesis and purification 468ff - geometric/electronic structure 3 17ff - - charged oligothiophenes 333ff - - metal/polymer interfaces 339ff - - neutral oligothiophenes 322ff - - nonlinear optical properties of neutral oligothiophenes 347E - theoretical methodology 320f -

-

-

-

-

ionization energies 93 melting point 127f - mesomeric resonance forms 93 - neutral - - electronic and linear optical properties 322ff - - nonlinear optical properties 347ff - a-oligothiophenes, preparation and physical properties 114f - oxidation potentials 94 regioisomeric/cyclic, preparation and 117 physical properties - - semiconducting, charge transport in 283ff - - basic models 284ff - - measurement of mobility 298ff - - transport properties 306ff - single crystals 184ff - - description 184ff - - optical and electrical properties 2 14ff - - structural data 186 - _ X-ray structures 187ff - substituted see Substituted oligothiophenes - synthesis 89ff - - first oligothiophenes and derivatives 90ff - - ring closure reactions from acyclic precursors 104ff - - substituted oligothiophenes 118ff - - unsubstituted oligothiophenes 93ff - thin films 220ff - - deposition techniques 220ff - electrical characterization 266ff - - morphology221ff - - optical characterization 247ff - transition energies 93 - transition metal complexes 171 - transport properties 306ff conductivity, mobility and carrier density 307ff - - traps 311f - undoped 247ff absorption spectra 248ff - - emission spectra 253ff - - nonlinear optical properties 256f - - occupied states 248 - unsubstituted see Unsubstituted oligothiophenes Optical properties - dimethylquaterthiophene 21 3ff - a-octithiophenes 216ff oligothiophenes 214ff - - charged 333ff - - neutral 322ff - - thinfilms 247ff - polycrystalline thin films 384ff - a-sexithiophenes 215f -

-

-

-

-

-

-

Index

Organic material sensors 508 Organic metals 5 17 Organic semiconductors 283 Organoboranes, oxidative coupling 96 Organometallic intermediates 13 Oxidative coupling - chemical and electrochemical 96 - organoboranes 96

regioregular, head-to-tail coupled polyalkylthiophenes 12-21 - - regioregular polyalkylthiophenes 9ff - morphology and crystal structure 48 - properties in solution 73ff - X-ray diffraction profiles 49 Polycondensation dehalogenation synthesis 3 Polycrystalline thin films - structure/morphology 380ff morphology 383f _ - orientation 381f - - structure 382f Polyenes, electronic properties 92 Poly(is0thianapthene) 38 Polymer electronics 512f Polymer light emitting diodes 64f Polyparaphenylene vinylene 3 18f Polythiophenes applications of 49 1ff - - antistatic coatings 495f - conducting materials 492ff - - electrochromic devices 503ff - - electromagnetic shielding materials 496f junction devices and rectifying bilayer electrodes 501 - - rechargeable batteries and capacitors 497ff resists, recording materials and fabrication of patterns 501ff - _ sensors 506ff electronic structure 363f - electro-optical devices 405ff - - all-optical modulator and memory devices 449ff - - electroluminescent devices 421ff - - electronic excitations in oligo(thiophenes) 413ff - - electro-optical modulator devices 444ff - - photoconductive and photovoltaic devices 439ff - - polymers and oligomers 40% - - preparation of thin film devices 408ff - field-effect transistors 476ff - - all-printed plastic FETs 481ff _ - regiorandom FETs 478 - - regioregular FETs 478ff - thin films 24% dopant effects 24% - - substitution effects 245 X-ray structure 213 Polythiophene zipper sensors 33 p-type doping 260ff PVT 499. 501 - -

-

Palladium-catalyzed cross coupling 98ff Pattern fabrication 57ff PEDOT 26,27 - electrochromic materials 505 - ferric chloride synthesis 27 PEDT 495f, 500, 504, 515 Pentacene 406 Perylene 406 Phenyl substituted polythiophenes 24 Photoconductive devices 439ff - photoconductivity in sexithiophene 440f - solar cells 441f - _ based on poly(thiophenes) 443f Photoconductivity 220, 274 Photoinduced charge carriers 265f Photoluminescence 63ff Photovoltaic devices 441f - poly(thiophene) based 443f Polarized electroluminescence 395ff Polarons 291ff electron 294 - lattice 294 - molecular ‘nearly small’ polaron 293f in r-conjugated polymers and oligomers 295f small polaron 291ff Poly(2,5-thiophenediylvinylene) 499, 501 -

-

-

Poly(3,4-ethylenedioxy-2,5-thiophenediyl) 495f, 500, 504, 515 Poly(3,4-ethylene-dioxythiophene)see PEDOT Poly(3-hexylthiophene) - absorption spectra 52, 53, 54, 56, 69 - dichroic IR spectra 71, 72 - dielectric relaxation spectra 77 - X-ray diffraction profiles 50 Poly(3-methylthiophene) absorption spectra 68, 70 cell voltage versus charge 67 - magnetic susceptibility 66 Polyalkylthiophenes chemical synthesis 5ff branched alkyl polyalkylthiophenes _ _ ferric chloride method 6ff metal catalyzed cross-coupling polymerizations 5f - - phenyl sidechains 24 -

-

-

-

-

-

-

-

-

-

-

-

-

21f

53 1

-

a-Quaterthiophenes P-substituted 134

-

532

Index

a-Quaterthiophenes (continued) - X-ray structure 193ff - - comparative crystal data 200 - - crystal packing 197, 198 - - a,a‘-dimethyl-quaterthiophene 196f - - 4,4’,3”,4”-tetramethyl-quaterthiophene 197ff

3,3’,4’’,3’’’-tetra(rnethyls~lphanyl)2,2‘:5’,2”:5’’,2’’-quaterthiophene 199ff Quenching reactions 13 - -

a-Quinquethiophenes @-substituted 135 - thinfilms 226ff - - evaporation conditions, substrates and substrate temperatures 235 IR-ATR measurements 228 _ _ IR transmission spectra 23 1 _ _ SFM images 229, 233 - - substrate temperature and film growth 230 - - UV absorption spectra 227 - X-ray structure 202-3 -

-

-

Random copolymers 20 - head-to-tail coupled 20f Rechargeable batteries 497ff Recording materials 501ff Rectifying bilayer electrodes 50 1 Regioregular polyalkylthiophenes 9ff - head-to-head couplings 1Off - head-to-tail couplings 9, 12ff - - IR and UV-Vis 16ff - - McCulloughmethod 12f - - mechanism and catalyst choice 15 - - NMR characterization 15f - - random copolymers 20f - - Riekemethod 13f - - self-assembly, X-ray and electrical conductivity 18f - - tail-to-tail couplings IOff Resists 501ff Restricted Open Shell Hartree-Fock method 321 Rieke method 13f Rieke zinc 13 Ring closure reactions 104ff - aromatization of tetrahydrothiophenes 104f - cyclization of 1,4-diketones 105ff - cyclization of diacetylenes 107ff oligothiophenes from 1,4 dithiins 1l O f f -

S-anti conformation of polythiophene 45, 46 Self-assembly of head-tail polythiophenes 18f Sensors 506ff - for bio-organic materials 508ff

chemoselective 28, 31 conducting polymer 27ff - for gases 506f - for ions in aqueous solution 507f - for organic materials 508 - polythiophene zipper 33 - _ synthesis 35 a-Sexithiophenes - @-substituted 136 - optical and electrical properties 215f - optical probing of field-induced charge in 446ff - photoconductivity in 440f - structural formulae 407 - thin films 234ff - - dopant effects 244 - - evaporation conditions, substrates and substrate temperatures 240f - - model structures 238 - - STM images 236,243 - - substitution effects 239ff - - X-ray diffractogram 242 - X-ray structure 203ff - - alkylated-sexithiophenes 204ff - - comparative crystal data 208 Short-channel effects 465f Single crystals of oligothiophenes 184ff - description 184ff - optical and electrical properties 214ff - - dimethylquaterthiophene 214f - - a-octithiophene 216ff - - a-sexithiophene 215f - structural data 186 - X-ray structures 187ff - - bithiophene 187ff - a-octithiophene 207ff - - polythiophene 213 - - a-quaterthiophene 193ff - - a-quinquethiophene 202f - - a-sexithiophene 203ff - - a-terthiophene 191ff Singlet fission 417 Singlet states 367ff - assignments 367f - chain length dependence 368f - lowest singlet state 369f - vibronic coupling 370ff Small oligomer thin films 223ff - dopant effects 226 - substitution effects 224ff Small polarons 291ff Solar cells 441f - poly(thiophene) based 443f Solid state properties 3738 - molecular packing 373f - theoreticafappro&h 374R -

-

Index

exciton concept 374ff higher transitions 379f Solutions - inter-ring torsion and coplanarity of oligomers in 420 - polythiophene properties in 73ff - preparation of thin films from 220f - sensors for ions in 507f Solvatochromism 53 Space-charge-limited current 300ff - effect of traps 303ff - estimation of 302f - profile of injected charges 301 Spectroscopic studies 65ff - charge storage configurations 6% properties in solution 73ff State ordering 394f Sterically encumbered polythiophenes 22 Stetter reaction 105, 119, 145, 155 Stille reaction l02f, 103, 155, 163 Structure - conducting polymers 48ff - and conductivity 269f - electronic 343f geometric 340ff - and mobility 270f - polythiophenes 48ff - and temperature 271f - X-ray see X-ray structure Substituted oligothiophenes 119ff, 473ff - functionalized oligothiophenes 15% - - acceptor-substituted 167ff - - amphiphilic 170f - donor-acceptor-substituted 168ff - - donor-substituted 156ff - - transition metal complexes 171 @,@‘-substituted 119ff - a,a‘-substituted 139ff - a,@-substituted 14% - @,@’-substituted examples synthesized for subsequent polymerization 128ff - - homologous series 12% - see also Oligothiophenes Substitution effects longer oligothiophenes 244 - polythiophenes 245 - a-sexithiophenes 239ff - small oligomer thin films 234ff Sugimoto/Yoshino synthesis 3, 6 Supercapacitors 500 Suzuki reaction 107, 152 - -

- -

-

-

-

-

-

~

-

Tail-to-tail coupling 9, 10ff Temperature, and conductivity - hopping models 289f

271f, 309ff

533

a-Terthiophene, X-ray structure 291ff - crystal packing 192, 193, 194

3’,4’-dibuty1-2,2’:5’,2’’-terthiophene192 infrared spectra 346 - 3,4’,4’’-trimethyl-2,2‘:5,2‘’-terthiophene 29 3,4’,4’’-trirnethyl-2, 3’,4’-dibutyl-2, 5”-diphenyl-2,2’:5’2’’-terthiophene 193 Tetracene 406 Tetrahydrothiophenes, aromatization 104f Thin film devices 408ff - blends between polymer and oligomers 413 - oligomers 409f - polymers 408f - solubility 410ff - - substitution with side-chains 410ff - - use of soluble partially-conjugated precursor polymer 412f - see also Thin films Thin films 220ff - deposition techniques 220f - - preparation from solution 220f - - vacuum deposition 220 - electrical characterization 266ff - - conjugation length influence 272 contacts, I/V-curves, carrier injection 267ff - - influence of dopants 272ff - - influence of structure on conductivity data 269f - influence of structure on mobility data 270f - photoconductivity 274 - - temperature dependence 271f inter-molecular non-radiative decay channels 417ff - aggregation and Davydov splitting 417f charge-transfer excitons 418f - morphology 221ff - - longer oligothiophenes 244 - - monothiophene 222 - - polythiophenes 24% - - quinquethiophene 226ff - - sexithiophene 234ff - - small oligomers 223ff - optical characterization 247ff - - charges in oligothiophenes 257ff _ - undoped oligothiophenes 247ff - polycrystalline - - optical properties 384ff - - structure/morphology 380ff - see also Thin film devices Thioether side chains 38 Thiophanes, aromatization 104f Time of flight 299f Trans-gauchetransformation 50 -

-

-

-

-

-

-

-

-

-

534

Index

Transient electric birefringence 75 Transition metal catalyzed coupling 34, 35,97f Transition metal complexes of oligothiophenes 171 Transport properties of oligothiophenes 306ff - conductivity, mobility and carrier density 307ff - - carrier density 308f - - variation with chain length 273f - - variation with temperature 275ff - traps 311f Transport relaxation time 79 Trapping 3 1 1f - multiple 297 - and space-charge-limited current 269ff Triplet states 372, 393f Two photon excitation 392 Ullmann synthesis 34, 35, 90, 97f Undoped oligothiophenes 247ff - absorption spectra 248ff - emission spectra 253ff - nonlinear optical properties 256f - occupied states 248 Unsubstituted oligothiophenes - synthesis 93ff - - arene/arene-coupling methods 93ff - - ring closure reactions from acyclic precursors 93

Unsubstituted polythiophene, chemical synthesis 2ff UV-Vis characterization 16ff Vacuum deposition 308 Valence Effective Hamiltonian approach 317, 321 Variable range hopping 292 Vibrational signature 244ff Vibronic coupling 370ff Vilsmeier formylation 164 Vilsmeyer-Haack conditions 119 Voltage-controlled colours 437ff Water-soluble, self-doped polythiophenes 36 Wittig reaction 107 Wudl synthesis 3 X-ray diffraction profiles - poly(3-hexylthiophene) 50 - polyalkylthiophenes 49 X-ray of head-tail polythiophenes X-ray structures 187ff - bithiophene 187ff - a-octithiophene 207ff - a-quaterthiophene 193ff - a-quinquethiophene 202ff - a-sexithiophene 203ff - a-terthiophene 191ff Yamamoto synthesis 2, 3 Zipper sensors 33, 35

18f