One-dimensional metals: physics and materials science

Denis Fichou Handbook of Oligo- and Polythiophenes @ WILEYWCH Related titles from WILEY-VCH K. Miillen / G. Wegner...

Author: Siegmar Roth

60 downloads 1121 Views 36MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

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


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


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


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).


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


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


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


[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


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


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%


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


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


%

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.


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


1.3.1.6 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


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


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


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


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.


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


‘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


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


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].


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. 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


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


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.


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


2.2 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


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


2.2.2 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.


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


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


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


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.


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


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.


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


[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.


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


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


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

+


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.


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)


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 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


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


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.


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


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.


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


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.


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


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


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


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


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,


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


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


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)


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


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 (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

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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


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


5.2 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


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


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


5.2.3.2 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).


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 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].


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.


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


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


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


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


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


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


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


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


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


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


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


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


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].


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


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


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


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].


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


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).


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


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)


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


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


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


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


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


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


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


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


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


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


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

----
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


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


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


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


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 (

One-dimensional metals: physics and materials science

Fundamentals of Radiation Materials Science: Metals and Alloys

Theory of Weldability of Metals and Alloys (Materials Science Monographs)

Physics of Amorphous Metals

Physics of Amorphous Metals

Chemical Thermodynamics for Metals and Materials

Materials science and engineering

Radiation Physics of Metals and Its Applications

Radiation Physics of Metals and Its Applications

Materials Science

Thin Film Growth: Physics, materials science and applications

Physics of Functional Materials

Physics of Functional Materials

Composite Materials: Science and Applications

Physics of functional materials

Composite Materials: Science and Applications

Nanoporous Materials: Science and Engineering

Ceramic Materials: Science and Engineering

Materials and man's needs: Materials science and engineering; summary report

Composite Materials: Science and Applications (Engineering Materials and Processes)

The Physics of Metals, Volume 2

Applied Materials Science

Materials Science Research Horizons

EMC 2008, Materials Science

Crosslinking In Materials Science

Computational materials science

Mechanical Behaviour of Engineering Materials: Metals, Ceramics, Polymers, and Composites

Electronic Materials Science

Investment Casting (Materials Science)

Engineering - Materials Science

Physics of Magnetism and Magnetic Materials

One-dimensional metals: physics and materials science

Fundamentals of Radiation Materials Science: Metals and Alloys

Theory of Weldability of Metals and Alloys (Materials Science Monographs)

Physics of Amorphous Metals

Physics of Amorphous Metals

Chemical Thermodynamics for Metals and Materials

Materials science and engineering

Radiation Physics of Metals and Its Applications

Radiation Physics of Metals and Its Applications

Materials Science

Thin Film Growth: Physics, materials science and applications

Physics of Functional Materials

Physics of Functional Materials

Composite Materials: Science and Applications

Physics of functional materials

Composite Materials: Science and Applications

Nanoporous Materials: Science and Engineering

Ceramic Materials: Science and Engineering

Materials and man's needs: Materials science and engineering; summary report

Composite Materials: Science and Applications (Engineering Materials and Processes)

The Physics of Metals, Volume 2

Applied Materials Science

Materials Science Research Horizons

EMC 2008, Materials Science

Crosslinking In Materials Science

Computational materials science

Mechanical Behaviour of Engineering Materials: Metals, Ceramics, Polymers, and Composites

Electronic Materials Science

Investment Casting (Materials Science)

Engineering - Materials Science

Physics of Magnetism and Magnetic Materials

Recommend Documents