DONOR-ACCEPTOR METHODS FOR BAND GAP REDUCTION IN CONJUGATED POLYMERS: THE ROLE OF ELECTRON RICH DONOR HETEROCYCLES

By

CHRISTOPHER A. THOMAS

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2002 Copyright  2002

by

Christopher A. Thomas

All rights reserved. ACKNOWLEDGMENTS

I thank my parents, Nancy and Larry, for their continuous support and attempts to understand and encourage me during what has been simultaneously the most enjoyable and most stressful part of my life. They made this document possible by encouraging and par- ticipating in experiences that ensured I would have the backgound and interest in trying to figure out how the world works.

The decision about whom to work for in graduate school is one of the events I have agonized the most about in my life. As promised, the graduate advisor-student relationship is a complicated entity consisting of advisor, boss, counselor and friend that can change its active role without warning. Despite this, I am positive that there is no other person that I would rather have worked with during this process. I especially appreciate being given an unusual amount of decision making and direction setting power in the projects I was involved with and the freedom to explore aspects of science that interested me even when they did not overlap cleanly with Professor Reynolds’ research interests or funding.

For their major contributions to my enjoyment and interest in chemistry, I thank Joe

Carolan for his contagious enthusiasm, and Joel Galanda, who immensely affected the three years of my life he was involved with my chemistry and physical science education.

I especially appreciate his role in enabling my employment at the Rockwell International

Science Center, one of the experiences I consider most valuable in my education. I thank my mentor at Rockwell, Les Warren, who, from high school through undergraduate and graduate school, has influenced my decisions on how to approach life, my education and

iii career more than any other person I have met. To Tim Parker, thanks go for largely direct- ing my research experiences as an undergraduate and showing me how to do synthesis even after being hit by my snowboard.

Several coworkers have had an important role both in this document and graduate school. Thanks go to Jacek Brzezinski and Kyukwan Zong for making the fluorenones and

XDOPs respectively, allowing me to spend my time doing things I found to be more inter- esting. To Roberta Hickman, thanks go for friendship, helping me, for making lab-life more fun, and for your contributions to the data which eventually became Figure 5-5. Thanks go to C. J. Du Bois, the reigning Reynolds’ group lab clown, for the Yuengling-Tücher inspired conversations at the Copper Monkey and always finding the reference that I knew existed but could not find. I also appreciate the rest of the members of the Reynolds group, especially Carl Gaupp and Irina Schwendenman, for being helpful when I needed it and generally fun to be around. I also appreciate the help provided by Lori Clark, Donna Bal- colm, and Professor James Deyrup for assitance in navigating through the UF bureaucracy,

Lorraine Williams for making life on the polymer floor easier and Khalil Abboud and Peter

Steel who solved the crystal structures included in this document.

Finally, my time here would not have been nearly as fun without the friends that provided the distractions necessary to have made this work possible to get through when it felt like work. Thanks go to Dominic Rice, who throws the best parties I have ever been to and was always up for hanging out when I needed to blow off steam; Kevin Boone for the football weekend fun and the good friends I have met through him; the KA’s for fun times tubing and at the lake house; Andrew Cottone, the Mobsters and the Bandits for the softball and baseball experiences; and Megan Odroniec for a fun couple of years.

iv TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... iii

ABSTRACT...... vii

CHAPTERS

1 BAND GAP: THE DEFINING QUANTITY FOR CONTROLLING THE PROPERTIES OF CONJUGATED POLYMERS ...... 1

Brief History of Conjugated Polymers ...... 1 Origin of Bands and Conductivity in π-Conjugated Organic Materials ...... 4 Why Band Energies Are Important ...... 12 Conductivity Mechanisms in Organic Materials ...... 15 Ionic Conductivity and Redox Conductivity ...... 17 Polarons, Bipolarons and p-Dimers ...... 20 Doping Induced Optical Transitions ...... 24 Conjugated Polymer Synthesis and Device Incorporation ...... 26

2 ELECTROCHEMICAL AND COUPLED IN SITU EXPERIMENTAL METHODS IN CONJUGATED POLYMER RESEARCH ...... 36

Introduction ...... 36 Monomer Crystallization ...... 37 Spectroelectrochemistry ...... 38 Colorimetry ...... 43 Cells and Reference Electrodes for Electrochemistry ...... 47 General Electrochemistry and Cyclic Voltammetry (CV) ...... 52 Differential Pulse Voltammetry (DPV) ...... 61 In Situ Conductivity ...... 62 Four-Point Conductivity on Free-Standing Films ...... 69 Electrochemical Quartz Crystal Microbalance (EQCM) ...... 72

3 DONOR MEDIATED BAND GAP REDUCTION IN A HOMOLOGOUS SERIES OF CONJUGATED POLYMERS ...... 76

Introduction and Literature Overview ...... 76 Synthesis and Monomer Properties ...... 81 v Monomer Structural Features and Crystallography ...... 85 Computational Explanations of Monomer Properties ...... 90 Polymer Optical Properties ...... 93 Polymer Electrochemistry ...... 106 Conclusions and Perspective ...... 115 Experimental Section ...... 122

4 THE ROLE OF INTERGAP REDOX STATES IN CONJUGATED POLYMER REDUCTION PROCESSES: THIENYL AND EDOT CONTAINING FLUORENONES ...... 129

Introduction ...... 129 Monomer Synthesis and Properties ...... 133 PBEDOT-DCF and PBTh-DCF Synthesis, Electrochemistry and Spectroelectrochemistry ...... 136 Monomer Electrochemistry ...... 148 Conclusions ...... 151 Experimental Section ...... 153

5 POLY(ALKYLENEDIOXYPYRROLES): AQUEOUS COMPATIBLE CONDUCTING POLYMERS WITH LOW FORMAL REDOX POTENTIALS ..160 Introduction ...... 160 Monomer Synthesis and Polymer Electrosynthesis ...... 165 Polymer Electrochemistry ...... 169 Polymer Spectral Characteristics ...... 174 Electrochemical Quartz Crystal Microbalance Studies on Ion Transfer ...... 180 Conclusions ...... 184 Experimental Section ...... 184

APPENDIX

A CRYSTALLOGRAPHIC INFORMATION FOR CYANOVINYLENE MONOMERS ...... 188

B CRYSTALLOGRAPHIC INFORMATION FOR FLUORENONE MONOMERS ...... 210

REFERENCES ...... 217

BIOGRAPHICAL SKETCH ...... 226

vi Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

DONOR-ACCEPTOR METHODS FOR BAND GAP REDUCTION IN CONJUGATED POLYMERS: THE ROLE OF ELECTRON RICH DONOR HETEROCYCLES

By

Christopher A. Thomas

May 2002

Chair: Professor John R. Reynolds Major Department: Chemistry

Keywords: conducting polymers, conjugated polymers, donor-acceptor, thiophene,

EDOT, EDT, electropolymerization, pyrrole, 3,4-alkylenedioxypyrrole, EDOP, ProDOP

Three families of conjugated polymers based on cyanovinylenes, fluorenones and

3,4-alkylenedioxypyrroles (XDOPs) were investigated to determine the role of the donor heterocycle in donor-acceptor (D-A) polymers (cyanovinylenes and fluorenones) and aqueous compatible conjugated polymers (PXDOPs) which are stable in the presence of strong biological reductants. When donors such as thiophene (Th), 3,4-ethylenedioxy- thiophene (EDOT, EDT) or 3,4-ethylenedioxypyrrole (EDOP) were used as donors and electropolymerized in a D-A-D configuration where the acceptor was cyanovinylene, spectroelectrochemistry confirmed band gaps ranging from 1.6 eV to 1.1 eV. Cyclic volta- mmetry confirmed p-type doping typical of conjugated polymers containing these hetero- cycles, and in situ conductivity indicated that a conductivity increase of at least 5 orders of magnitude occurred upon p-type doping. Spectroelectrochemistry and colorimetry experi-

vii ments establish a color change from a deep blue (neutral form) to a transmissive blue (p- type doped) form when switched in acetonitrile on indium tin oxide. The cyanovinylene family of polymers exhibit an uncharacteristically sharp reductive cyclic voltammetric response, but do not change colors upon reduction. In situ conductivity indicates that the change in conductivity on reduction is small, and the symmetrical shape is evidence it is predominantly due to redox conductivity and not electronic conductivity as is the case for p-type doping. These results combine to suggest that the strongest donor necessary for band gap reduction in these systems is EDOT as increasing the donor character by using

EDOP did not reduce the band gap further. They also suggest that acceptor units be present in the polymer backbone in a 1:1 ratio (to donors) and that the acceptor unit not have strong acceptors pendant to the main conjugation path for improved n-type conduc- tivity.

PEDOP and PXDOPs in general are uniquely suited for use as aqueous compatible conjugated polymers in applications where poly(pyrrole) (PPy) is now used. This interest is largely due to the PXDOPs having the lowest reduction potentials of any conjugated polymer reported to date, lending them improved stability in the p-type doped conducting state. This stability was tested with the biological reductants dithiothreitol and glutathione which were found not to reduce the PXDOPs but did reduce PPy. Colorimetry established color changes for the PXDOPs from red/orange (neutral) to transparent blue/gray (oxi- dized) with band gaps between 1.9 to 2.2 eV by spectroelectrochemistry.

viii CHAPTER 1 BAND GAP: THE DEFINING QUANTITY FOR CONTROLLING THE PROPERTIES OF CONJUGATED POLYMERS

1.1 Brief History of Conjugated Polymers

Research interest in conjugated and electroactive polymers has been invigorated with the discovery that these materials can be incorporated into flexible, ultimately low cost displays for a variety of applications. The field of conjugated polymers1 (CPs) originated near the end of the 1970’s with the discovery that poly(acetylene) (PA, 1) can be made highly conductive by doping.2 Subsequently, poly(aniline) (PANI, 2), poly(pyrrole) (PPy,

3), poly(thiophene) (PTh, 4) and more recently poly(3,4-ethylenedioxythiophene)

(PEDOT, 5) (Figure 1-1) have been extensively studied and the field itself is the subject of many reviews.3 Originally it was thought that the principal application for CPs would be to replace relatively dense metals such as copper and aluminum in weight sensitive applications such as air and space travel. While heavily doped poly(acetylene) (PA) can be made more highly conducting than copper on a conductivity per weight basis, its instability

1. While the terms conjugated polymer and conducting polymer are often used inter- changeably throughout the literature, “conjugated polymer” is preferred here since the materials made are often not highly conductive, especially when undoped, and many properties described herein make use of the non-conductive state.

2. Chiang, C. K.; Park, Y. W.; Heeger, A. J.; Shirakawa, H.; Louis, E. J.; MacDiarmid, A. G. Phys.Rev.Lett. 1977, 39, 1098.

3. (a) Roncali, J. Chem. Rev. 1992, 92, 711-738. (b) Chan, H. S. O.; Ng, S. C. Prog. Polym. Sci. 1998, 23, 1167-1231. (c) Roncali, J. J. Mater. Chem. 1999, 9, 1875-1893.

1 2 under ambient conditions which include water and oxygen precluded the practical use of

PA practically. Search for alternative CPs with superior stabilities led to the discoveries of

PANI, PPy, PTh, and PEDOT which, although blessed with improved stability in the doped, conducting form, are still much less conductive than PA and have thus reduced research efforts into CPs as light-weight metal replacement conductors. Far from marking the death of the field, the successors to PA have found use in a wide range of applications including thin film transistors (TFTs),4 flexible light emitting displays (OLEDs),5 chemical sensors for a wide array of analytes,6 and nerve cell guidance channels and biological substrates.7 Largely in recognition of the important role that these materials are expected to play in the near term technology infrastructure of the internet, portable electronic devices and computers, the Nobel Prize in Chemistry was awarded in 2000 to the discoverers of PA,

Drs. Heeger, MacDiarmid and Shirakawa.

Poly(acetylene) (PA), PANI, PPy, and PTh represent the first generation of CPs, and as such, are not the ideal materials for most applications for reasons that are detailed throughout this chapter. PEDOT has been the most successful of the 2nd generation CPs

4. Halls, J. J. M.; Walsh, C. A.; Greenham, N. C.; Marseglia, E. A.; Friend, R. H.; Mor- atti, S. C.; Holmes, A. B. Nature 1995, 376, 498-500.

5. (a) Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Marks, R. N.; Taliani, C.; Bradley, D. D. C.; Dos Santos, D. A.; Bredas, J. L.; Logdlund, M.; Salaneck,W.R.Nature 1999, 397, 121-128. (b) Kraft, A.; Grimsdale, A. C.; Holmes, A. B. Angew. Chem., Int. Ed. Engl. 1998, 37, 402-428.

6. (a) Albert, K. J.; Lewis, N. S.; Schauer, C. L.; Sotzing, G. A.; Stitzel, S. E.; Vaid, T. P.; Walt,D.R.Chem. Rev. 2000, 100(7), 2595-2626. (b) McQuade, D. T.; Pullen, A. E.; Swager, T. M. Chem. Rev. 2000, 100(7), 2537-2574. (c) Ellis, D. L.; Zakin, M. R.; Bernstein, L. W.; Rubner, M. F. Anal. Chem. 1996, 68(5), 816-822.

7. C. E. Scmidt, V. R. Shastri, J. P. Vacanti, R. Langer, R. Proc. Natl. Acad. Sci. USA 1997, 94, 8948. 3

H H N S S x N x x x x

OO

PA, 1 PANI, 2 PPy, 3 PTh, 4 PEDOT, 5

Figure 1-1. Common conjugated polymers

because of its reduced band gap compared to PTh and increased environmental stability as

a doped conductor. As such, it has been commercialized by Agfa and Bayer and found

applications in transparent static protection of electronics during shipping, improved

soldering processes for circuit boards, photographic film production and as a component in

organic light emitting devices (OLEDs). Because of PEDOT’s unique properties, research

interest and related publications have increased exponentially over the last several years

from the first one in 1991 to 65 in 1999.8

Simple heterocycles have largely been exhausted as possible candidates for 3rd

generation materials because a combination of lower band gap, aqueous compatibility and

the ability to be n-type doped without necessitating heroic efforts to exclude air and water

are needed for these applications. Because of these requirements, it seems likely that next

generation materials will come from more complicated monomers built up from smaller

functional segments to incorporate the electronic properties needed while maintaining the

ability to be effectively polymerized. Designing optimized materials requires a

fundamental understanding of how structural modification is related to the ultimate

material properties. Recent research efforts by several groups indicate that the quantities

8. Groenendaal, L. B.; Jonas, F., Freitag, D.; Pielartzik, H.; Reynolds, J. R. Adv. Mater. 2000, 12, 7, 481-494. 4 that are most significant to the control of CP properties are the band gap and position of the edges of the conduction band and valence band. This chapter establishes the molecular origin of band gap, how it manifests itself in current materials, what needs to be done to optimize some materials, and the impact that band gap has on materials in terms of conductivity.

1.2 Origin of Bands and Conductivity in π-Conjugated Organic Materials

While PA is essentially a defunct material in terms of current research interest, its structural simplicity provides a convenient entry into the band gap discussion. PA is composed of a chain of sp2 hybridized carbon atoms linked by alternating single and double bonds, (CH)x. The theoretical work was done long before the actual synthesis and PA was assumed to have a structure like that shown in Figure 1-2 A where resonance causes the double bonds and single bonds to have the same bond length on average and there is extensive delocalization along a chain. The communication along the main chain of the polymer is due to the overlapping p orbitals forming π bonds and this material was expected to be highly conductive as a result. The anticipated orbital diagram is shown in Figure 1-2

A, and when the electrons are added in, there should be no energy difference between the full VB arising from the HOMO of acetylene and the unfilled CB from the LUMO. This explains the expected high conductivity which requires a partially filled band. Physicists have an alternative model to chemists for describing the band structure of materials that adds insight into the behavior of these materials at the expense of being slightly harder to visualize. In this model the energy of the orbitals, and thus bands, are viewed in reciprocal k space used for crystallography. For hypothetical PA, the VB runs up in energy and the

CB runs down to meet at a point indicating no band gap. Chemists tend to be uncomfortable 5

A. HYPOTHETICAL POLY(ACETYLENE) BAND STRUCTURE

k(E) DOS(E) = E

k DOS B. POLY(ACETYLENE) AFTER PEIERLS DISTORTION

= E EF

k DOS

C. DEVELOPMENT OF THE BAND STRUCTURE OF POLY(THIOPHENE)

B3P86-30%/cep-31g* S 3.0 x

0.0 CB -3.0 Eg -6.0

Energy /eV VB -9.0

-12.0

-15.0 0123456∞ ~DOS Number of Rings (x)

Figure 1-2. Development of band structure from monomer to polymer. (A) Poly(acetylene) is composed of a series of p orbitals which were initially expected to have the band structure shown on the right. (B) PA actually undergoes a Peierls distortion causing a band gap to open with the approximate density of states shown. Adapted from Hoffman, R. Angew. Chem. Int. Ed. Eng. 1987 26, 846-878. (C) The PTh band structure is based on orbitals from monomer through oligomer overlapping as shown. Adapted from Salzner, U.; Lagowski, J. B.; Pickup, P. G.; Poirier, R. A. Synth. Met. 1998, 96, 177-189.

visualizing k space and as a result, density of states (DOS) is utilized frequently as a model

that incorporates the important information from k space while preserving the intelligibility 6 of real space for non-physicists. The derivation and methods of approximating the DOS of a material given the band structure are detailed by Hoffman.9

When PA was synthesized, it was found to be an insulator with a conductivity of ca.10-13 Scm-1 for the cis-PA form.10 Isomerization occurs incompletely by heating and

11 efficiently upon n-type doping and charge compensation with AsF5. trans-PA effectively dimerizes by a Peierls distortion,12 opening up a band gap (1.48 eV by photoconductivity for trans,~2eVforcis) at the Fermi level and dramatically reducing the conductivity because electrons have to be thermally excited across the band gap to delocalize in the partially filled CB. The band structure and approximate DOS results of this process are represented in Figure 1-2 B. Because of this, PA must be doped to either partially fill the

CB by adding electrons (n-type doping) or partially vacate the VB by oxidation (p-type doping). This doping process, while possible and quite effective at improving the conductivity to a maximum of 2(10)4 Scm-1, does not result in an air stable material and caused chemists to look elsewhere for other polymers that could be doped more easily.

In the early 1980’s it was discovered that electron rich poly(heterocycles) could be chemically or electrochemically oxidized to from CPs that have relatively high conductivities. Pyrrole and thiophene are two examples of these heterocycles. The

9. Hoffman, R. Angew. Chem. Int. Ed. Eng. 1987, 26, 846-878.

10. trans-PA has a neutral conductivity of 10-5 -10-6 Scm-1.

11.Chien,J.C.W.Polyacetylene Chemistry, Physics and Material Science; Academic Press, Harcourt Brace Jovanovich: Orlando, 1984; Chapter 5.

12. The bond length alternation has been experimentally determined to be ~0.08 A. (a) Fincher, C. R.; Chen, C. E.; Heeger, A. J.; Macdiarmid, A. G.; Hastings, J. B. Phys. Rev. Lett. 1982, 48, 100. (b) Yannoni, C. S.; Clarke, T. C. Phys. Rev. Lett. 1983, 51, 1191. 7

evolution of the band structure of PTh is shown in Figure 1-2 C for monomer through

hexamer and eventually polymer. Figure 1-3 shows the same formation of bands from

overlapping molecular orbitals for poly(pyrrole) PPy as determined by Density Functional

Theory (DFT) calculations. These calculations are becoming widely used in CP research

due to their accounting of electron correlation, accuracy in band gap prediction and

relatively light computational requirements.13

DEVELOPMENT OF THE BAND STRUCTURE OF POLY(PYRROLE)

B3P86-30%/cep-31g* H 3.0 N x

0.0 CB

-3.0 Eg

-6.0 VB

Energy /eV -9.0

-12.0

-15.0 0123456 ∞ Number of Rings (x)

Figure 1-3. Evolution of the band structure of PPy from monomer through hexamer extrapolated to polymer as determined by density functional theory. Adapted from Salzner, U.; Lagowski, J. B.; Pickup, P. G.; Poirier, R. A. Synth. Met. 1998, 96, 177-189.

There are five classes of conductivity that can be identified and are largely

differentiated by their conductivity magnitude and temperature dependence. Of these, four

can be easily explained by simple band diagrams (Figure 1-4) while the fifth,

13. Salzner, U.; Pickup P. G.; Poirier, R. A.; Lagowski., J. B. J. Phys. Chem. A, 1998, 102 (15), 2572 -2578. 8 superconductivity, has a more complicated mechanism. Metals are the most familiar conducting materials to most and are characterized by a zero band gap originating from a partially filled band or, alternatively, a contacting VB and CB. Semi-metals are relatively uncommon and examples are mostly limited to alkali and alkaline earth elements. The conductivity mechanism in semi-metals involves overlapping orbitals creating a partially filled band which enables electrical conductivity.

With the exceptions of PA, PPy, PEDOT (which when doped with specific dopants are metallic)14 and poly(sulfurnitride) (an inorganic semi-metal), the vast majority of CPs exhibit semiconductor or insulator-like conductivity. These operate under an identical mechanism where population of the CB is thermally activated and the conductivity magnitude is determined by the temperature and the size of the band gap. The difference between a semiconductor and an insulator is rather arbitrarily determined at a band gap of

3eVinmostcases.

The temperature dependence of conductivity is the primary indicator of conduction mechanism for newly discovered materials. Figure 1-4 B shows the temperature dependence for a prototypical material for each class of conductivity. Since metals and semi-metals have no band gap, the thermal component to metallic conductivity serves to decrease conductivity with increasing temperature by scattering electrons through increased phonon (or lattice vibration) intensity. This is observed for Cu and (SN)x.

Conductivity is described by the equation σ =neµ where σ is the conductivity, n is the number of charge carriers, e is the sign of the charge carrier and µ is the charge carrier

14. (PPy) Lee, K.; Miller, E. K.; Aleshin, A. N.; Menon, R.; Heeger, A. J.; Kim, J. H.; Yoon, C. O.; Lee, H. Adv. Mater. 1998, 10(6), 456-459. (PEDOT) Aleshin, A. N.; Kie- booms, R.; Heeger, A. J.; Synth. Met. 1999, 101, 369-370. 9

A. CLASSES OF CONDUCTIVITY NOTE: Superconductivity does not fit neatly into this model CB

CB

CB CB Eg > 3 Eg < 3 Eg = 0 Eg = 0

VB VB VB VB

METAL SEMIMETAL SEMICONDUCTOR INSULATOR

B. CONDUCTIVITY vs. TEMPERATURE FOR VARIOUS MECHANISMS 108 Copper (METAL) A

106

(SN)x (SEMIMETAL) 104

(TTF)(TCNQ) (TYPE III LINEAR CHAIN CONDUCTOR) -1 102

1 Mobility of Charge Carriers B α µ = T-

Conductivity 10-2 σ = neµ

Conductivity /S cm PPy-DBSA (SEMICONDUCTOR)

10-4 Concentration of Charge Carriers n = e-Eg/2kT 10-6

Temperature 10-8 050100 150 200 250 300 350 Temperature /K Figure 1-4. Conductivity temperature dependence. (A) Simple band diagrams showing the difference between metals, semi-metals, semiconductors and insulators. (B) Conductivity vs. temperature profiles for the four types of conductivities indicates metals and semi- metals have a negative correlation of conductivity vs. temperature where semiconductors and insulators (not shown) have a positive correlation. This results from the competing relationship of charge carrier mobility and concentration (B, inset) for semiconductors. For metals, only the mobility is important since there is no band gap. Adapted from Reference Epstein, A. J.; Miller, J. S. Sci. Am. 1979, 241, 4, 52-61. 10 mobility. All three quantities that determine conductivity have different temperature dependences (Figure 1-4 B inset) and it is the competition between these quantities that determines the overall relationship between σ and T. The number of charge carriers, n, follows an Arrhenius activation and depends on an exponential relationship of band gap and temperature. The charge carrier mobility is related to T and decreases with increasing temperature because of scattering. For PPy doped with dodecyl benzene sulfonic acid (PPy-

DBSA), the conductivity increases with temperature due to increasing thermal population of the CB (increase in n).15

An alternative visualization to the mechanism of conductivity differences between metals and semiconductors is shown in Figure 1-5. In metals, the mobile electron can be viewed as being scattered by a lattice vibration (phonon) that grows with increasing temperature. At low temperatures, the electron has enough energy to surmount the phonon barrier height where at high temperatures, the electron is scattered. In contrast, semiconductors use a hopping mechanism where the electron is initially trapped in a lattice defect. Increasing temperatures causes the lattice defect to deform and the electron can move under the influence of a potential field.

Minimizing the band gap is an important goal for maximizing the neutral conductivity of CPs. Most CPs synthesized to date have band gaps greater than 2 eV characterizing them as mid- to high band gap polymers. Polymers with band gaps lower than 1.5 eV are considered relatively low band gap materials and few confirmed examples are available of polymers with band gaps below 0.8 eV.16 PITN is a notable example of one

15. Aleshin, A. N.; Lee, K.; Lee, J. Y.; Kim, D. Y.; Kim, C. Y. Synth. Met. 1999, 99, 27- 33. 11

of the earliest low band gap polymers, made so because of the competition between

aromaticity and quinoid geometry in the polymer.17 For reference, silicon has a band gap

of 1.1 eV. A. BAND-TYPE CONDUCTION (METALS) LATTICE VIBRATION

DELOCALIZED ELECTRON

INCREASED TEMP

SCATTERED ELECTRON

B. CONDUCTION BY HOPPING (SEMICONDUCTORS) ENERGY ELECTRON LOCALIZED BY DEFECTS OR DISORDER

LATTICE VIBRATION

DISTANCE IN ELECTRIC FIELD

Figure 1-5. Conductivity mechanism (A) A delocalized electron is scattered by a lattice vibration in metals which increases at increasing temperatures. (B) In semiconductors, conduction by hopping is assisted by lattice vibrations and increases with increasing temperature. Adapted from Epstein, A. J.; Miller, J. S. Sci. Am. 1979, 241, 4, 52-61.

16. Some poorly defined materials with extremely low band gaps have been reported but further confirmations are necessary: (a) Huang, H; Pickup, P. G. Chem. Mater., 1998, 10(8), 2212 -2216. (b) Huang, H; Pickup, P. G. Chem. of Mater.; 1999; 11(6); 1541- 1545.

17. Wudl, F. J. Org. Chem 1984, 49, 3382. 12

1.3 Why Band Energies Are Important

While it is practically the band gap that determines the conductivity and color of the neutral polymer, the positions (energies) of the band edges determine how easily doped (p- type or n-type) a polymer is and how stable it is in the doped states compared to the neutral forms. Figure 1-6 describes the four ways in which most CPs can be classified. These are either high or low band gap with either a high or low VB or CB. The most common scenario by far is the polymer with a band gap of 2 eV or greater and a relatively high VB. Examples of polymers with these characteristics are PPy, PTh, PEDOT and poly(3,4- alkylenedioxypyrroles) (PXDOPs), an extreme example of a high VB whose properties are described in Chapter 5. These polymers are all prepared by oxidative polymerization routes and are easily p-type doped18 by virtue of their high lying VB. Because of their high band gap, the CB is generally out of reach for stable electrochemical n-type doping and PPy cannot be n-type doped. PTh does exhibit n-type doping at very negative potentials and

PEDOT is more difficult to n-type dope than PTh. Another situation is the high gap/low VB polymer such as PPV and its analogues. PPV is not easily doped electrochemically,19 but its bands are positioned so that charge carriers of either sign can be injected making this class of polymer useful in light emitting devices where the emission energy color is tuned by modifying the band gap.

Low gap polymers are a relatively recent topic of research interest and these types of polymers have many interesting applications. Regardless of how the band energies line

18. p-type doping is the creation of charge carriers by partial oxidation of a polymer or semiconductor creating a partially vacated VB. Conversely, n-type doping is the cre- ation of charge carriers by partial reduction creating a partially filled CB.

19. Tanaka, S.; Reynolds, J. R. J. Macro. Sci., Pure Appl. Chem. 1995, A32, 1049-1060 13 up, the low band gap often causes these polymers to be transparent in the doped state (p- or n-type) (see section 1-5 for a discussion of how band gap translates into colors). This makes them useful in a variety of devices such as electrodes and windows where this transparency can be exploited. As with the high gap polymers, the color of low gap polymers is dictated by the magnitude of the band gap. Since the distance between the edge of the VB and CB is smaller by definition in low gap systems, the CB is generally more accessible to n-type doping even in low gap polymers with relatively high VBs. This class of materials is perhaps the most intriguing, offering the transparency of low gap systems with the ability to be both n-type and p-type doped in the same polymer.

While it is obvious that the VB needs to be relatively low in energy for facile n-type doping, the question arises, how low does it need to be for the n-type, conductive state to be stable in devices? The answer to this question is somewhat complicated and lies in the available redox couples that can interfere with the n-type state described and is shown in

Figure 1-7. For a p-type polymer, the reaction of interest is the O2 to H2O redox couple where oxygen could oxidize the neutral form of a polymer to the p-type doped state. This redox couple is very near the Eº’ (formal redox potential, see Chapter 2) for several conducting polymers and p-type doped polymers with an Eº’ more anodic of the O2 to H2O redox couple are reduced by H2O and are stable in the neutral form. Polymers with a high enough VB to have an Eº’ cathodic of the O2 to H2O couple are stable in the p-type, oxidized state. For n-type conducting polymers the reaction of interest is the H2OtoH2 couple. A polymer with an Eº’ cathodic of this couple (every authentic n-dopable polymer to date) will be oxidized by H2O to the neutral form of the polymer. This couple appears at ca. -0.65 V vs. SCE. With typical overpotentials it is necessary to have a polymer with an 14

CB -2.0

-1.5

CB CB -1.0

CB -0.5 Potenial /V vs SCE 0 VB VB

Eg > 1.5 eV Eg < 1.5 eV +0.5 VB

+1.0 VB E E High g Polymers Low g Polymers

Low VB (HOMO) High VB (HOMO) High CB (LUMO) Low CB (LUMO) Easily p-type XXX doped Easily n-type XXX doped Example PPV Poly(pyrrole) Cyanovinylenes PCDM Poly(XDOPs) (Chapter 3) few others (Chapter 5) many others

Figure 1-6. Possibilities for positioning of band edges in both high and low band gap polymers. The common case is a high band gap polymer with a high VB. For many applications, a low band gap polymer with a low CB is desirable and few examples are available.

Eº’ more anodic of -0.5 V vs. SCE for the n-type doping process for the polymer to be stable

in the presence of water.20 This is a far more significant problem than the need to have an 15

Eº’ for p-type doping cathodic of ca. +0.5 V vs. SCE to be stable in the oxidized state since even moderately electron rich polymers are well cathodic of this value. To have a polymer that is stable to water in its n-type form and stable to oxygen and water in its p-type form, the polymer needs to have a band gap of less than 1 eV and be very easy to reduce compared to current materials. Identifying the structural factors that could enable this are the impetus for much of the research in this work.

1.4 Conductivity Mechanisms in Organic Materials

While the ability to dope a material at potentials that are compatible with ambient environmental conditions is important for electrochromic devices and sensors, the conductivity magnitude is an important property for microwave camouflage devices for military applications. CPs offer a unique property in that their conductivity can be tuned over 8 or more orders of magnitude in the same material. Several different conductivity mechanisms are available for electrically conducting CPs depending on their structure and hybrid materials can have multiple mechanisms. Figure 1-8 shows the typical conductivity ranges for materials commonly considered metals, semiconductors and insulators although the factor that determines this classification is really the temperature dependence of conductivity rather than the conductivity magnitude. The three most common conducting polymers shown in Figure 1-8 span the range of conductivities from metal to insulator depending on doping level. This property is unique to CPs. The mechanism for neutral polymer conductivity has been discussed in section 1-2. Neutral polymers tend to be semiconductors or, less frequently, insulators since the band gap is generally < 3 eV yet is

20. de Leeuw, D. M.; Simenon, M. M. J.; Brown, A. R.; Einerhand, R. E. F. Synth. Met. 1997, 53-59. 16

RED OX FORM FORM

-2.0 PTh- PTh° AC

-1.5

-1.0

H2O H2 -0.5

Potenial /V vs SCE PEDOT° PEDOT+ 0 {}PPy° PPy+

PANI° PANI+ +0.5 2H2O O2

B +1.0 PPV° PPV+

A p-type dopable polymers in the neutral state with E1/2(° to +) in this range are oxidized by O2 (p-doped form stable under ambient conditions). n-type dopable polymers in the reduced state with E1/2(° to -) in this range are oxidized by O2 (neutral form stable under ambient conditions). B p-type dopable polymers in the neutral state with E1/2(° to +) in this range are stable to O2 (neutral form stable under ambient conditions).

C n-type dopable polymers in the reduced state with E1/2(° to -) in this range are oxidized by H2O (neutral form stable under ambient conditions).

Figure 1-7. Stability of CPs in their oxidized or reduced forms depends on where the formal oxidation potential (Eº’) lies (a direct consequence of the VB (p-type) or CB (n- type). A p-type polymer with a low Eº’ is stable in its neutral state. Adapted from de Leeuw, D. M.; Simenon, M. M. J.; Brown, A. R.; Einerhand, R. E. F. Synth. Met. 1997, 53- 59. 17

too high for efficient population of the CB by kT. Doping by introducing charge carriers

(positive or negative) has a drastic effect on conductivity as noted in Figure 1-8 and the

details of this are important for designing new materials.

POLY- POLY- SILVER 106 ACETYLENE THIOPHENE COPPER σ > 2 X 104 S cm-1 σ = 2000 S cm-1 METALS IRON max max BISMUTH 104 (SN)x DOPED 2 TTF•TCNQ 10 NMP•TCNQ InSb KCP 1

GERMANIUM 10-2 SEMICONDUCTORS 10-4 TRANS (CH)x SILICON 10-6

SILICON -8 BROMIDE 10 UNDOPED CIS (CH) S x x 10-10 POLY- GLASS PYRROLE σ -1 max = 500 S cm INSULATORS DNA 10-12 MOST MOLECULAR H CRYSTALS -14 N DIAMOND 10 x

SULFUR 10-16 QUARTZ 10-18 Ω-1 cm-1

Figure 1-8. The conductivity range available to CPs spans those common for metals through insulators. Adapted from the Handbook of Conducting Polymers.

1.4.1 Ionic Conductivity and Redox Conductivity

A wide variety of conductivity mechanisms are seen in a variety of polymeric

materials. Non-conjugated organic polymers can display conductivity in certain situations.

The simplest form of conductivity in organic polymers is ionic conductivity (Figure 1-9 A).

Nafion,21 a polymer used in fuel cells, and poly(ethyleneglycol) (PEG),22 used in batteries,

are the canonical materials that exhibit ionic conductivity. The mechanism for this involves 18 ions (Li+,H+) moving in the presence of an electric field. In Nafion, phase segregation of the hydrophobic perfluorinated sections from the hydrophilic sulfonic acid groups forms cavities ca. 4.5 nm across which have highly concentrated acid groups. Proton conduction occurs through these channels. In PEG, an intercalated carbon electrode injects Li+ ions into the polymer which are then allowed to move by segmental motion of the PEG chains.

In some materials anions are the dominant mobile species and the transference number is the ratio of the mobility of the cation to that of the anion.

Another conductivity mechanism that is observed in organic materials is redox or self-exchange conductivity studied in detail by Royce Murray’s group (Figure 1-9 B).23

One class of material that exhibits this type of conductivity includes transition metal bis-α- diimine structures functionalized with PEG chains around them. This results in a matrix of easily oxidizable or reducible transition metal centers which are fixed in location between two electrodes. A maximum in conductivity is attained when there is a mixed valent condition, often with nearly the same concentration of neutral and charged sites. Since the redox centers cannot move throughout the film, electron or hole hopping between adjacent sites of different charge state allows conductivity to occur.

21.(a)Hsu,W.Y.;Gierke,T.D.J. Membr. Sci. 1983, 13, 307. (b) Srinivasan, S.; Manko, D. J.; Koch, H.; Enayetullah, M. A.; Appleby, J. J. Power Sources 1990, 29,367.(c) Heitner-Wirguin, C. J. Membr. Sci. 1996, 120,1.

22.Gray,F.M.Solid Polymer Electrolytes-Fundamentals and Technological Applica- tions; VCH: Weinheim, Germany, 1991.

23. (a) Williams, M. E.; Masui, H.; Long, J. W.; Malik, J.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 1997. (b) Long, J. W.; Kim, I. K.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 11510. (c) Williams, M. E.; Crooker, J. C.; Pyati, R.; Lyons, L. J.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 10249. 19

A. IONIC CONDUCTIVITY (Nafion 117 & PEG)

Nafion 117™ []() CF2CF2 mnCF2CF2 O CF m = 5 - 13.5 2 []CFCF3 k n = ca. 1000 O - k = 1, 2, 3, ... CF2CF2SO3

Nafion Tertiary Structure CONDUCTIVITY MECHANISM IN PEG Li+ Li° A-1 A-2

~ ~~ ~ ~~ ~ - ~ O O ~ ~ O O ~ ClO4 ~ O ~~ O ~~ O O O O O O + O O Li Li+ O O O O O O O - O

O O ClO4

~ ~ ~ O ~ O O O ~~ O ~~ O

+ + CsLi Li° CsLi Li°

B. REDOX (SELF-EXCHANGE) CONDUCTIVITY (Example: Polyether Hybrids)

O O ) ( O x O

3 X ( ( x = or O O N N Co(III) Co(II) + M = Fe, Co, Ni x = 3, ~7, ~11, ~15

B-1e- B-2 B-3

e- e- e-

Gate electrode maintains Electrodes are typically 3:17 Co3+ : Co2+ Pt IME or bands

Figure 1-9. (A) Ionic conductivity is the predominant mechanism of charge transport in Nafion fuel cells and polymer electrolytes for batteries. (B) In redox conductivity, fixed redox centers pass charge by electron hopping between adjacent sites of different oxidation state. 20

1.4.2 Polarons, Bipolarons and π-Dimers

In doped PA, a conductivity mechanism is evident that is different than ionic or redox conductivity (Figure 1-10 A). PA has a degenerate ground state, that is, there are two structures composed of exchanging carbon-carbon double bonds that have the exact same total energy. PA chains with an odd number of carbon atoms have an unpaired electron (a neutral soliton). Upon oxidation or reduction, a radical cation or anion is generated which moves along a polymer chain by the mechanism shown in Figure 1-10 A.24 This charge has the properties of a solitary wave, thus the soliton terminology from physics, and does not dissipate as it traverses the chain. Solitons have a width because the spin density (for a neutral soliton) or the charge density (cationic or anionic solitons) is not localized on one carbon but is rather spread out over several. Passing from one side of a soliton to another, the double bonds become gradually shorter, the single bonds gradually longer until they are equal in the center of the soliton.25 This double bond pattern is reversed on the other side of the soliton.

All CPs other than poly(acetylene) have non-degenerate ground states. In poly(heterocycles) there is an aromatic state and a quinoidal state of higher energy (except for PITN). Oxidation of a poly(heterocycle) creates a radical cation called a polaron where the delocalization is over ca. 4 to 5 rings. This radical cation exhibits an EPR signal since there is an unpaired spin and moves under the effect of the applied potential along the polymer main chain. Further oxidation at higher doping levels in poly(heterocycles) can

24.Su,W.P.;Schrieffer,J.R.;Heeger,A.J.Phys.Rev.Lett. 1979, 42, 1698.

25. (a) Rebbi, C. Sci. Am. 1979, 240, 92. (b) Boudreaux, D. S.; Chance, R. R.; Bredas, J. L.; Silbey, R. Phys Rev. B: Condens. Matter. 1983, 28, 6927. (c) Thomann, H.; Dalton, L. R.; Tomkiewicz, Y.; Shiren, N. S.; Clarke, T. C. Phys.Rev.Lett. 1983, 50, 533. 21

A. SOLITON CHARGE CARRIER IN POLY(ACETYLENE)

[OX] A-1 A-2 PA is degenerate A-3 B. POLARONS AND BIPOLARONS AS CHARGE CARRIERS B-1 S S S S S S S S S x Neutral Polymer 1 e- OX B-4 - S S B-2 ClO4 S S S S S S S S S x poly(heterocycles) are A Single Polaron non-degenerate 1 e- OX ClO - ClO - B-3 4 4 S S S S S S S S S x A Bipolaron

C. π-DIMERS AS AN ALTERNATIVE TO BIPOLARONS

- * A

O A-

x O

MEH-PPV

Figure 1-10. Charge carriers in conjugated polymers. (A) Solitons are the primary method of charge transport in doped PA. (B) Polarons and bipolarons are suggested as the charge carriers for doped poly(heterocycles) with non-degenerate ground states although π-dimers (C) are proposed as an alternative to polarons, especially in oligomeric materials.

create either a second polaron or a dication, called a bipolaron. A bipolaron is an EPR silent

dication delocalized over 4 to 5 rings. Considerable debate about the nature of charge

carriers in CPs has ensued but we know certain facts. First, at low doping levels, an EPR

active signal is observed consistent with polarons. Second, as doping level increases, the 22

EPR signal vanishes consistent with a spin paired conducting state. Third, oxidized PTh shows a two peaked absorption structure in the UV-Vis that is partially consistent with bipolarons as the charge carriers. The theory of electronic transitions for polarons and bipolarons was originated in the Su-Shrieffer-Heeger (SSH) model which predicted bipolarons as the stable charge carriers.26 This model was enhanced by a continuum electron-phonon coupled model proposed by Fesser, Bishop and Campbell (FBC).27 Early calculations have shown that the bipolaron is more stable than two polarons28 by 0.45 eV

(the difference between the bipolaron binding energy (0.69 eV) and the polaron binding energies (2 X 0.12 eV)).29 More recent calculations have shown that up to an oligomer with ten monomer units, a bipolaron is the stable charged species. Beyond a decamer, charge repulsion between the two cations sharing the same lattice distortion suggests that two polarons are the more favorable charged species.30 The conundrum is that if this two polaron state is indeed favored, it is difficult to explain the EPR data suggesting that there are no spins present in highly doped polymers. A proposal that provides a partial solution to this problem is the π-dimer (Figure 1-10 C) as the stable charge carrier in highly doped conducting polymers. This would allow the predicted two polaron state to exist but solves the EPR problem by pairing spins on two cofacially aligned chains. This has been demonstrated for oligomers of PPV31 and poly(thiophenes).32 The optical properties of

26.Su,W.P.;Schrieffer,J.R.;Heeger,A.J.Phys.Rev.Lett. 1979, 42, 1698.

27. Fesser, K.; Bishop, A. R.; Campbell, D. K. Phys. Rev. B. 1983, 27, 4804.

28. Heeger, A. J.; Kivelson, S.; Schrieffer J. R.; Su, W.-P. Rev. Mod. Phys. 1988, 60, 781.

29.Bredas,J.L.;Street,G.B.Acc. Chem. Res. 1985, 18, 309-315.

30. Sakamoto, A.; Furukawa, Y.; Tasumi, M. J. Phys. Chem. B 1997, 101, 1726-1732. 23 bipolarons are different than those of π-dimers as described in the next section and enable distinguishing between the two.

In most p-type doped CPs, the anions move in and out of the film on doping and dedoping and can be exchanged in electrolyte spontaneously, but are immobile in the film when dry once a constant doping level has been reached. Typically, the charge carriers must hop from one anion to the nearest anion along the chain as they move. Thus, there is a pinning potential for separating the polaron cation from its anion. At low doping levels, when the pinning potential is larger than the potential across the polymer, low conductivity results. As the doping level increases, charge carriers have to migrate much shorter distances through the film to find the next anion and the pinning potential is thus smaller.33

Effective conjugation length or coherence length is independent of molecular weight and is the number of repeat units that an electron or charge carrier can effectively access. There is no firm consensus in the literature on what the effective conjugation length of most conjugated materials is. Single-molecule spectroscopy on PPV indicates that this material can be treated as a polymer that has a coherence length of only about 10-15 rings despite molecular weights (Mw ~ 100,000) which would allow much greater effective conjugation lengths.34 In other materials this may be much shorter and even approach the

31. Sakamoto, A.; Furukawa, Y.; Tasumi, M. J. Phys. Chem. B 1997, 101, 1726-1732.

32. Hill, M. G.; Penneau, J.-F.; Zinger, B.; Mann, K. R.; Miller, L. L. Chem. Mater. 1992, 4, 1106-1113.

33. Reynolds, J. R.; Schlenoff, J. B.; Chien, J. C. W. J. Electrochem. Soc. 1985, 1131- 1135.

34. (a) Vanden Bout, D. A.; Yip, W.-T.; Hu, D.; Fu, D.-K.; Swager, T. M.; Barbara, P. F. Science 1997, 277, 1074-1077. (b) Hi, D.; Yu, J.; Barbara, P. F. J. Am. Chem. Soc. 1999, 121, 6936. 24 length where the bipolaron is the most stable charged species in highly doped materials. In electropolymerized materials where molecular weight information is difficult to obtain, there may only be oligomers in the sample. Regardless, it is clear that optical and electrical properties saturate very early in the formation of polymer from monomer, much earlier than the physical properties for most polymers.

1.5 Doping Induced Optical Transitions

Poly(acetylene). Section 1.4 showed that PA is unique among CPs because it has a degenerate ground state. Neutral PA has one allowed optical transition, that from VB to CB as shown in Figure 1-11. Three options are then possible for creating charge carriers on a

PA chain which appear as a localized state at mid-gap. In trans-PA with an odd number of electrons, a neutral soliton forms from the odd p orbital and single π electron. Oxidation creates an empty state and reduction creates a doubly occupied state at mid-gap, all of which are spectroscopically observed.

Poly(pyrrole) and poly(thiophene). The nondegenerate ground state due to the energy difference between the aromatic form and quinoidal form of poly(heterocycles) necessitates different charge carriers and thus optical responses. The quinoid structure has a larger affinity for charges as suggested by calculations which show a lower ionization potential and larger electron affinity than the aromatic structure.35 PPy is perhaps the most well-studied of the poly(heterocycle) systems and provides a view that is consistent with other similar conjugated polymers such as PTh. Theoretical and optical studies place the

λ band gap of PPy at 2.7 to 2.8 eV with a max of 3.2 eV. Single electron oxidation creates a

35. Bredas, J. L.; Themans, B.; Fripiat, J. G.; Andre, J. M.; Chance, R. R. Phys.Rev.B: Condens. Matter. 1984, 29, 6761. 25 chain with a polaron on it and two intragap states that are roughly 0.5 eV inside the original band edges (Figure 1-11 B). According to Bredas and Street (Acc. Chem. Res. 1985), when a second electron is taken out of the chain, a bipolaron is formed, as is consistent with their calculations. More recently, Furukawa revisited the spectrochemical data for neutral and oxidized forms of PTh, PPV poly(2,5-thienylenevinylene) and polyaniline.36 The original papers on these polymers had identified the stable charge carriers as bipolarons in the heavily doped state based on the two bands observed in the UV-Vis-Near-IR of the p-type doped form. The FBC model predicts that the ratio of absorbance intensities Eb1/Eb2 should be 14. In all cases the observed ratios are much smaller, on the order of 1.2. This was termed the “intensity anomaly” and several theoretical approaches were pursued in attempt to explain it while still maintaining the bipolaron assignment. Furukawa reassigned the bands to correspond to polarons based on contemporary oligomer studies and reevaluated the peak intensities. Similar transitions for polarons (solid arrows in the polaron diagrams of

Figure 1-11 B) compared to bipolarons are allowed. Specifically, the VB to lowest polaron band is allowed for polarons as it is for bipolarons, but the lowest polaron to highest polaron band transition is allowed for polarons and is not for bipolarons. Additionally, the VB to highest bipolaron band (Eb2) is polarized in the direction perpendicular to the chain axis making it extremely weak and explaining the intensity anomaly. Consequently, reassignment of all the polymer’s charge carriers in the heavily doped state to arrays of polarons proves more consistent with theory than the assignments as bipolarons. The lack of EPR signal is explained by the polarons forming lattices or arrays (π-dimers) that attenuate spin density. Heavily doped PTh shows metallic like behavior such as a Pauli spin

36. Furukawa, Y. J. Phys. Chem. 1996, 100, 39, 15644-15653. 26 succeptibility, linear temperature dependence of the thermoelectric power and a negative correlation of conductivity with temperature. This has traditionally been explained by the lowest bipolaron band growing in size until it meets the VB allowing a partially filled band necessary for the metallic conductivity mechanism. Furukawa’s explanation is that the polaron lattice, built up of an array of singly occupied states, would form essentially a half filled band, enabling a metallic conductivity mechanism as well. Oligomers, and how their properties relate to polymers, continue to be an active research area and it appears that the idea of π-dimers, polarons and arrays or polarons as the dominant charge carriers in conjugated polymers is gaining acceptance. Due to scarcity of easily n-type doped materials, the optical properties of these polymers are only recently beginning to be understood. It is postulated that the charge carriers are the same regardless of the sign and this seems to be true, at least for p-sexiphenyl, for which polarons are the stable n-type charged species.37

1.6 Conjugated Polymer Synthesis and Device Incorporation

Electrochromic devices. The previous background has been provided to introduce the concepts that dictate the properties of devices fabricated from conjugated polymers, namely band gap, position of the band edges and conductivity. This section details how these design components are unified to make electrochromic and light emitting devices.

Electrochromics, driven by consumer interests in displays and military interests for absorbers, have also found use in commercial applications such as electrochromic mirrors in cars and eventually in dial-tint windows for residential and business uses. With the

37. Koch, N.; Rajagopal, A.; Ghijsen, J.; Johnson, R. L.; Leising, G.; Pireaux, J.-J. J. Phys Chem. B 2000, 104, 1434-1438. 27

A. POSSIBLE OPTICAL TRANSITIONS FOR PA

Neutral Neutral Positive Negative Poly(Acetylene) Soliton Soliton Soliton

B. POSSIBLE OPTICAL TRANSITIONS FOR A NON-DEGENERATE POLY(HETEROCYCLE)

Eg Eb2

Eb1

Neutral Polaron Bipolaron Bipolaron Metallic-Like Polymer States States Bands States at High Doping Level

Figure 1-11. Optical transitions for doped conducting polymers. (A) neutral PA, p-doped and n-doped PA all have a mid-gap optical transition. (Adapted from Bredas, J. L.; Street, G. B. Acc. Chem. Res. 1985, 18, 309-315) (B) The non degeneracy of poly(heterocycles) causes the charge carriers to be polarons or bipolarons at high doping levels. (Adapted from Fesser, K.; Bishop, A. R.; Campbell, D. K. Phys.Rev.B. 1983, 27, 4804)

exception of a Gentex product based on an organic electrochromic, consumer

electrochromics are based almost solely on inorganic tungsten trioxide which, while robust,

suffers from slow switching times and is not easily extendable to other colors. Organic

electrochromics, especially those based on conjugated polymers, offer faster switching

times38 and are much more easily modified to provide different colors based on research in

38. Welsh D. M.; Kumar A.; Meijer E. W.; Reynolds J. R. Adv. Mater. 1999, 11(16): 1379-1382. 28 band gap modification. Figure 1-12 shows a typical device used for testing organic electrochromic materials. In this device an indium tin oxide (ITO) electrode on either glass or a flexible substrate like Mylar or poly(ethylene) is to electropolymerize a conjugated polymer. Two different complimentary CPs are used on opposite electrodes with a transparent gel electrolyte between them.

Anodically and cathodically coloring polymers. The color of these devices is based on the band gap of the polymer as shown in Figure 1-12. Typically a high band gap and a low band gap polymer are used and are assembled in the device where one is in its neutral form while the other is in its doped form. Low band gap polymers tend to have the

π to π* transition for the neutral form in the visible region and are thus cathodically coloring, meaning that the neutral state is the colored state. High band gap polymers tend to have the π to π* transition shifted to higher energies into the UV and are thus near colorless in this state. When oxidized however, an absorption in the visible region becomes apparent making high band gap polymers anodically coloring. Low band gap polymers tend to have the absorbances that appear upon doping pushed to low energies in the NIR and are thus colorless or nearly so in the doped state. Thus, this complementarity works quite well for devices that include these materials since when the neutral polymer is oxidized at one electrode, the oxidized polymer is reduced at the other electrode and the effect is a pair of polymers that are either both nearly colorless or both deeply colored at the same time.

Single polymer devices can also be used when a specific color transition is required rather than the clear to opaque transition described above. Currently, no materials are available for the n-type to neutral polymer transition because materials with low enough CBs and high n-type doped conductivities that are stable to repeated switching do not exist. 29

A. ELECTROCHROMIC DEVICE SCHEMATIC ITO on Substrate Electroactive Polymer 1

Gel Electrolyte Electroactive Polymer 2 1.50 B. Colored State

1.25 Visible Light

1.00 Region B-1

0.75 Region B-3 Oxidatively Doped Absorbance 0.50 Region B-2 High Band Gap Polymer

0.25 Neutral Low Band Gap Polymer 0 02341 Energy /eV 1.50 C. Bleached State

1.25 Visible Light Neutral High Band Gap Polymer 1.00

0.75 Absorbance 0.50

0.25 Oxidatively Doped Low Band Gap Polymer 0 02341 Energy /eV Figure 1-12. Electrochromic Device Schematic. (A) typical makeup of a two polymer electrochromic device operating in either the opaque to transparent or conducting to insulating modes. (B) In the opaque to transparent mode, the colored state is derived from a doped high gap polymer and neutral low gap polymer combining to both contribute absorbance in the visible region. (C) The bleached state is formed by the compliment to the colored state - a neutral high gap polymer with transitions in the NIR and a doped low gap polymer with transitions in the UV. Adapted from Sapp, S. A.; Sotzing, G. A.; Reynolds J. R. Chem Mater. 1998, 10(8), 2101-2108. 30

Electrochemical polymerization. Several methods are available for the preparation of CPs. For preparation directly on an electrode, oxidative electrochemical polymerization is certainly the simplest alternative. The mechanism proceeds with oxidation of a heterocyclic monomer such as Py, Th or EDOT at an electrode. The highest

HOMO density is located on the heteroatom (S or N) but the highest spin density in the lowest energy form is found at the 2 position (See Figure 1-13). Initially in the polymerization there is only monomer and oxidized monomer radical cations available. In one possible mechanism, two radical cations dimerize to form the dication bis(heterocycle).

The lower energy of the aromatic form drives the rearomatization of the dimer to form the neutral species. The extended conjugation in the neutral dimer causes it to have a lower oxidation potential than the monomer and it is preferentially oxidized. The radical cation dimer is likely present at lower concentrations than the radical cation monomer and either meets with this monomer to form the trimer or dimerizes to form the tetramer (less likely).

Rearomatization proceeds as before to afford the neutral species. As chain length increases, solubility decreases and the oligomers precipitate out on the electrode. In order to continue depositing on the electrode, the neutral oligomers formed must be oxidized to the conducting form so that electrons can be transferred through the forming film to continue deposition. The films formed are thus insoluble and infusible.

Figure 1-13 B shows the possible ways of coupling radical cations of simple poly(heterocycles). Since the spin and charge density are not entirely on the 2 or 5 positions, coupling at the 3 or 4 positions (the β positions) is possible. These αβ and ββ coupling products can form in electrochemical polymerizations and are viewed as detrimental since they block conjugation. In EDOT, the 3 and 4 positions are blocked with 31

A. ELECTROCHEMICAL POLYMERIZATION MECHANISM (X = S, NH, O)

- e- XXEpa

H X X 2 + 2H+ X X H X dimer 4 3 - X - e X H 5 2 X X X 1 H X Numbering Key X H X X X X H

X + 2H+ X X trimer

B. ALTERNATIVE COUPLING POSSIBLITIES IN HETEROCYCLE POLYMERIZATION

β αα' X coupling XXXX α X

αα' coupling X

β αβ ' XXX coupling α XXX XX

Figure 1-13. Oxidative polymerization mechanism for a generalized poly(heterocycle). (A) Monomer is oxidized to the radical cation which then couples and rearomatizes releasing two protons. The dimer formed has a lower oxidation potential than monomer and either dimerizes or couples to another monomer to form trimer. (B) In simple poly(heterocycles), the 3 and 4 positions are unprotected making conjugation blocking αβ and ββ couplings possible.

a cyclic ether making polymerization proceed exclusively through the 2 and 5 positions and

simplifying structural assumptions.

Electrochemical polymerization is often maligned, especially by traditional

polymer chemists, who feel that the inability to characterize these films by even the most

standard polymer characterization techniques to determine molecular weight and structure

negates the usefulness of these polymers. While some of these arguments are valid, 32 electropolymerized polymers will remain quite useful for several applications for the following reasons: First, despite the difficulty in obtaining molecular weight by standard polymer solution methods such as gel permeation chromatography (GPC), vapor pressure osmometry (VPO) and light scattering, radio labelled end group analysis of tritiated derivatives has estimated the degree of polymerization to be about 20 in PPy and can be used when this information is necessary to gather for characterization.39 Grazing emission

XRF has also been developed to estimate molecular weight in other solid state polymer films.40 While this degree of polymerization is low by polymer standards, it is certainly past the point where the electrical properties saturate (vide supra). Second, when it is polymer film on a conducting surface that is required, there is no easier way to prepare this than by electropolymerization. There are no catalysts to prepare or complicated workups and purifications to perform on solution polymerized materials. It is also necessary to address electrode confinement issues and the redissolution of solution polymerized soluble materials from the electrodes after they are cast or spin-coated onto an electrode. Third, electropolymerized polymers are much more electroactive than even solution cast polymers with nearly identical repeat units. This is due to the memory of the film for the dopant ion used during electrosynthesis and the morphology adjustment that must be made on doping and undoping. Finally, it is generally easier to synthesize monomers for electropolymerized materials than for their soluble counterparts. Typically all that must be done is to attach electron rich heterocycles known to polymerize around a core. Since it is

39. Nazzal, Street, G. B. Chem. Comm. 1984.

40. Blockhuys, F.; Claes, M.; Van Grieken, R.; Geise, H. J. Anal. Chem. 2000, 72, 3366- 3368. 33 the radical that is coupling, it is the purity of the radical, not the purity of the monomer that determines molecular weight. This core can tolerate a variety of functionalities making it far more facile to build up libraries of compounds to test theories on band gap modification or transport properties. Electrochemically prepared materials are unsuitable in applications such as conductive fibers in which case traditional polymer processing techniques are necessary. Additionally, when more traditional properties such as glass transition and lack of crystallinity are more important than electroactivity, solution polymerized materials with much higher molecular weights are the obvious choice. Electropolymerized materials are also justly criticized for the difficulty that ensues when trying to convert their synthesis to continuous manufacturing processes and the yields from electropolymerization are quite low since the only monomer that is reactive is the fraction that diffuses into the electrochemical double layer.

Light emitting devices are quickly becoming the most important application in conjugated polymeric materials. As of mid 2001, these materials are emerging from the labs and finding their way into devices such as car stereo and cell phone color displays where wide viewing angle, low cost and intrinsic lighting are more important than resolution. While lifetimes are not yet comparable to LCDs or more traditional CRTs, they are improving rapidly. PPV and specifically, MEH-PPV is now the most widely used material in OLED research. It is synthesized (Figure 1-14) by either radical or anionic chain growth addition type polymerization of a symmetrically functionalized PPV with either a sulfonium, sulfone or xanthate salt. This chain growth method allows high molecular weights (100,000s) of a polymer precursor that can be cast, spin coated or spun into fibers.

Subsequent thermal elimination forms the active PPV light emitting material. Additionally, 34

Heck and other organometallic polymerizations are also used to produce functionalized

PPV derivatives.

L -HL Nu: Nu

L LL Monomer [X]

L = Cl, Br, sulfonium [X] sulfone, or xanthate

[X] -HL

L L x x Precursor Polymer PPV

e- CB E CB F

VB VB

EF h+

Metal Organic Organic Metal HTL ETL

Figure 1-14. Polymerization of PPVs. (A) Chain growth radical or anionic addition polymerization is used to make high molecular weight PPV precursor polymer which is then thermally eliminated to make PPV or a derivative. (B) PPVs are widely used in light emitting devices, the simplest of which have the setup shown above. From the left, holes (h+, hollow dots) are injected from a low work function electrode such as Al or Ca and migrate through a hole transporting layer where the recombine radiatively with electrons (solid dots) injected from the transparent electrode (ITO or PEDOT).

An energy diagram is included in Figure 1-14 showing the simplest of the light

emitting devices. In this setup, a hole injected from a low work function metal such as Ca

or Al meets an electron injected from a transparent electrode (EDOT or ITO) inside the

active layer. A device engineering contribution to this field is the recognition that the hole

moves at different speeds than the electron and hole transporting layers are now included 35 into devices to ensure productive charge carrier recombination inside the active area of the material.

To create the next generation of materials, the factors that influence band gap, the edges of the semi-metals VB and CB, and the conductivity must be understood. Currently, there are no materials that have high p-type and n-type conductivity in the same material, few materials that have a band gap under 1.5 eV and relatively few CPs that are suitable for biological applications. The subsequent chapters address the structural components of these issues and provide a foundation for the design of materials that meet the requirements for next generation materials. CHAPTER 2 ELECTROCHEMICAL AND COUPLED IN SITU EXPERIMENTAL METHODS IN CONJUGATED POLYMER RESEARCH

2.1 Introduction

Chapter 1 described the importance of band gap and the properties that arise from it, namely, doping (both p- and n-type), color and conductivity changes. The experimental methods used to quantitatively characterize these properties has largely been the realm of electrochemists who are interested in describing and optimizing charge storage capacity, dc electronic conduction and ion conduction in CP films and composites.

The doping of CPs affects the conductivity, electronic and vibrational spectra, spin concentration and physical structure. Most of these properties are best studied by in situ methods to avoid the effects of exposing the polymer to ambient conditions while in a metastable state. Coupled techniques to study the in situ changes in electronic spectra

(Section 2.3), conductivity (Section 2.8) and mass transport (Section 2.10) are described in this chapter. Despite several recent reviews related to CP synthesis,1 there is no review available which addresses the techniques common to CP characterization with the intended audience being the synthetic chemist who must characterize a newly prepared material.

Complete characterization of an electrochemically prepared CP includes determining the potential where polymer deposition occurs, the polymer’s band gap, formal potentials for

1. (a) Roncali, J. Chem. Rev. 1992, 92, 711-738. (b) Chan, H. S. O.; Ng, S. C. Prog. Polym. Sci. 1998, 23, 1167-1231. (c) Roncali, J. J. Mater. Chem. 1999, 9, 1875-1893.

36 37 p-type doping and n-type doping, doping level, conductivity onset and magnitude, mass changes upon doping, color transitions and changes in optical absorbances. These values and methods are referred to repeatedly in the subsequent chapters and the intent of this chapter is to provide the necessary references and background material to understand and perform these experiments. Where possible, examples using PEDOT data are shown.

2.2 Monomer Crystallization

Monomer crystal structures provide information on the bond length alternation, relative orientation of rings and degree of planarity, most of these quantities having some potential to be extrapolated to the polymer. Additionally, monomer crystal structures provide a starting point and a method of validation for computational probes of polymer electronic and optical properties. Crystallization of organic compounds remains somewhat of an art and several methods were used to obtain single crystals for x-ray diffraction in this work. Ideally monomers were left in capped NMR tubes to slowly evaporate over the course of several weeks although narrow bore glass tubing is a more economical way to achieve the same effect. In some cases this worked with NMR solvent, CDCl3 but in most cases several solvents were tried. The high surface area and low volume of an NMR tube makes this environment more suitable for crystallization than larger vessels with different aspect ratios. Typical solvents used for the planar aromatic compounds described in this document are chloroform, toluene, benzene, acetonitrile, ethanol, and dichloroethane. In all cases, the monomer solution was nearly saturated and was filtered before addition to the

NMR tube to avoid particulates that may cause nonproductive nucleation.

In cases where several vessels were used with the above solvents and no quality crystals developed, vapor diffusion was used. In this method, a small open vial containing 38 a nearly saturated solution of monomer in a good solvent was placed in a larger vial containing a non-solvent for the monomer with a high vapor pressure (usually pentane).

This larger vial was sealed so the contents of the two smaller vials could reach equilibrium over the course of several days. During this process the volume of the non solvent decreases and the volume of the good solvent, monomer and diffused poor solvent increases. It is important to leave enough free volume in the smaller vial so that the poor solvent can diffuse in without overflowing the monomer solution. It is also important that the two solvents be miscible and to dissolve the monomer in a solvent with a sufficiently high vapor pressure so that the process of solvent mixing equilibrium occurs mostly in the smaller vial.

Solvent diffusion through an inverted U-tube was also used successfully to obtain crystals of monomers. In this method, a 4-5 mm tube was bent in a U-shape and filled with a good solvent for the monomer with some means of sealing this U-tube to a vial (usually a vial cap with a hole drilled in it and the U-tube epoxied in place). While inverted, one vial is filled with saturated solution of monomer in good solvent and the other is filled with a poor solvent that is miscible with the monomer solution. The U-tube is mated with these two vials and the setup is allowed to reach equilibrium in the dark in a place where it is isolated from shock and most vibrations. Even though the color throughout the setup becomes homogenous quite quickly, it takes several days for crystals to appear in the top of the U-tube. Other methods of monomer crystallization are diffusion cooling and slow monomer sublimation for volatile monomer solids.

2.3 Spectroelectrochemistry

Spectroelectrochemistry is an important and routine method used in the characterization of CPs. It provides the band gap (Eg) of a polymer, information about the 39

intraband states that appear on doping, and can be used to establish the formal redox

potential through spectroelectrochemical titration. Spectroelectrochemistry is best

performed in a UV-Vis-NIR instrument (a Cary 5e was used throughout this work). 1 X 1

X 4.5 cm cells were used which were transparent from <190 nm (the UV-cutoff for ACN)

to >2600 nm although smaller ranges are typically used. Figure 2-1 shows the cell and

electrode arrangement used throughout this work for potential control in a cuvette vs.aAg

wire pseudo reference electrode. In this method, a Teflon cap is drilled with holes to accept

(1) an ITO glass slide through a slit shown in the front, (2) a tube to accept the Pt counter

electrode which has a rectangular hole cut in it slightly larger than the beam from the

spectrophotometer, (3) a hole to accept the Ag wire pseudo reference electrode which

passes through a tube that keeps it from shorting with either the ITO/polymer or the Pt

counter electrode and (4) a hole to accept a needle for degassing the solvent prior to

collection of spectra subsequently maintaining an inert gas blanket over the cell.

Figure 2-1. Spectroelectrochemical cell designshowingscaffoldingusedtokeepthe electrodes from moving within the cuvette. This cell utilizes a Pt counter electrode with a rectangular hole cut in it, a Ag wire pseudo reference electrode, a hole for degassing and a slit for the insertion of ITO/glass. 40

In a typical spectroelectrochemical experiment, solvent (ACN, [0.1 M] TBAP and blank ITO are placed in both the sample cuvette and the blank cuvette. Both cuvettes are degassed with argon for 5 minutes and are placed in the spectrophotometer for background collection. The background is collected over the range 1600 nm to 250 nm. While the spectrometer, sample and cuvettes can be scanned over a much larger range, at wavelengths longer than 1600 nm noise due to water is visible in the spectra and is not easily removed.

Additionally, data is collected every 1 nm but reported vs. eV where 1 eV = 1240 nm making the near IR region highly compressed (many points cover a small range in eV).

Thus, 1600 nm = 0.775 eV and 2600 nm = 0.477 eV, so collecting data in this noisy range amounts to only ca. 0.3 eV of extra data. After background collection, the polymer on ITO/ glass is placed in the sample cuvette and leads are attached. For most of the spectra in this work, the solution was degassed only prior to background collection. In all of the polymers studied, this was sufficient to limit degradation in the neutral state. Active bubbling of argon through the sample solution tends to cause the film to flake off throughout the experiment and does not appear to attenuate the noise at wavelengths longer than 1600 nm.

The polymer is then switched over several potentials while monitoring the absorbance to determine the potentials where the polymer is completely oxidized and completely charge compensated (neutral). With these potentials established, the polymer film is broken in with a potential square wave between the oxidized and neutral potentials. p-Type dopable polymers typically break in over roughly 20 double potential steps and this process is monitored by following the current and charge during application of the potential square wave. The majority of polymers in this work are stable in their oxidized state and undergo spontaneous reaction with oxygen in the neutral state to be reoxidized (see Figure 1-7), 41

sometimes with degradation of the polymer. In order to collect good data, the oxidized state

was probed first ensuring that the film would not be held in the usually unstable neutral

state for a long period of time before the spectra for the p-type doped state was recorded.

The potential was then poised where the polymer is completely oxidized and held under

potential control while the energy was scanned from 1600 nm (0.775 eV) to 250 nm (4.96

eV). For most polymers, the potential was then stepped in 100 mV increments until the

range of polymer states between oxidized and neutral had been examined. Example data for

PEDOT(collectedevery50mV)isshowninFigure2-2.

1.2 λ S max x π to π* 1.2

max 1.0 OO λ ~ E°' E Eb2 g 0.8 1.0 E +0.87 V b1 0.6

0.4

Absorbance at

Neutral Bipolaron 0.2 Polymer States 0.8 -1.0 -0.5 0.0 0.5 1.0 Potential /V vs SCE

e Eb1

0.6

Absorbanc

0.4

0.2

-0.93 V 0.0 Eg ~ 1.6 - 1.7 eV

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Energy /eV Figure 2-2. Spectroelectrochemistry for PEDOT. Film was deposited from PC [0.1 M] TBAP, [~300 mM] EDOT and switched in ACN [0.1 M] TBAP. Inset shows Absorbance vs. potential. Band gap is determined by extrapolating the onset of the π to π* absorbance to the background absorbance. The Eb1 transition is allowed and is visible at intermediate doping levels. 42

From the data shown in Figure 2-2, it is possible to determine the band gap by extrapolating the onset of the π to π* absorbance to the background yielding a value of 1.6 to 1.7 eV for PEDOT. In order for these results to be valid, it must be clear that the polymer is completely neutral, that is, there are no shoulders merged with the π to π* transition that may be taken as the band gap. Given the above limitation, this method of estimating band gap is the most reliable of those mentioned in this chapter because only one onset has to be estimated.

The positions of intragap states are also able to be determined from spectroelectrochemistry. At intermediate doping levels, there is a peak that appears that is consistent with Eb1. Since the corresponding absorbance for Eb2 is not allowed, it is not observed in the spectra but is equidistant from the CB edge as Eb1 is from the VB edge. λ From the absorbance vs. potential data at max, it is possible to determine the polymer reduction potential, E°’, where there is 50% neutral polymer and 50% oxidized polymer as shown in the inset of Figure 2-2.2 The assumptions made are that there is 100% neutral

λ polymer at max and 100% oxidized polymer when the absorbance stabilizes at low λ absorbance values at max. From the Nernst equation, vide infra, it is then possible to determine E°’ from the point where there is 50% oxidized and neutral polymer. This method can be useful when electrochemical methods are inconclusive due to highly capacitive films.

While the cell shown in Figure 2-1 proved adequate, several modifications are suggested to ease collection of spectra. First, the cell cap should be made of a polymer such as poly(etheretherketone) PEEK3 that is nearly as solvent resistant as Teflon but is more

2. Heineman,W.R.Denki Kagaku 1982, 50(2), 142-148. 43 physically robust, can hold screws without stripping, and can be wet by epoxy. It should also have a longer section that mates with the cuvette so the tendency to rock is reduced.

This cell cap should also have a wider opening for the ITO so that it is easier to avoid having the polymer touch the cell cap when inserting the polymer/ITO slide. Given the 1 cm2 opening at the top of the cuvette, it would be difficult to fit in even the small SCE electrodes used herein but a flared glass double potential junction could be used in the center of the cap to allow incorporation of the greater stability of the SCE electrode. This also alleviates the problem of having to make sure that the reference electrode contacts neither the film/

ITO or the counter electrode, the former case almost always results in destruction of the film. The Pt counter electrode with the rectangular hole cut into it to allow the passage of light is overkill for fast switching polymers containing EDOT although slower materials will require a counter electrode with matched area. Films switch reliably and homogeneously with Pt wires (straight or coiled) run down the side of the cuvette despite a lower surface area. This counter electrode assembly as shown has the tendency to twist in the cuvette sometimes coming in contact with the reference wire or working electrode.

Improvements to this cell are also useful in colorimetry experiments.

2.4 Colorimetry

Since markets that are emerging for CPs concentrate on displays, whether electrochromic, OLED, or OLEC, the color of light absorbed or emitted from polymers is quite important to quantify. Colorimetry, in its simplest form, is spectroscopy that corrects

3. PEEK (VictrexTM 450G, 381G or 450GL30) is a high temperature thermoplastic with excellent resistance to a wide range of solvents except strong acids and long term exposure in chlorinated solvent. It is finding wide use in medical applications as well as in HPLC fittings. Other options are Delrin, a polyacetal used in bearings spacers and gears and Duroid, a PTFE ceramic material. 44 for the nonlinearity of the human eye.4 Light entering the eye is imaged on the retina by two classes of photoreceptors called rods and cones. Rods are highly sensitive and are capable of detection down to the single photon level making them suited to detecting small amounts of light. Cones are less sensitive than rods and are the active detection mechanism in well-lit rooms and daylight situations. There is only one type of rod and three types of cones. Therefore, low light vision is in grey scale (using the single rod type), and well-lit situations appear in color due to the differing spectral sensitivities of the cones. These cones are dubbed L, M,andS because of their sensitivity to long, medium and short wavelength light respectively. The basis for colorimetry is the measured sensitivities of the three cones which overlap across the visible region and have a gaussian form. Signals that result in the same cone signals are perceived as the same color to the brain. Since there are many more

L and M cones compared to S cones, coupled with the fact that short wavelength light is scattered more in the optics of the eye, the eye is less sensitive to colors in the blue and purple region of the visible spectra.

While colorimetry can be performed on standard spectrophotometers with linear detectors and then deconvoluted in software, to date, colorimetry is best performed with a dedicated instrument using the same cell shown in Figure 2-1. Representing color using the three cone values is not convenient and several systems have evolved to specify colors in more useful ways. One of the most popular methods of determining whether two colors match is the color scale of the CIE 1931 standard observer based on data from seven to ten color-normal individuals in the 1920s in England. From these spectral sensitivities

4. Berns, R. S. Billmeyer and Saltzman’s Principles of Color Technology, Third Edition , Wiley & Sons: New York, 2000. 45 tristimulus values are available. Color is then described by multiplying the relative spectral power of a standard light source, the reflectance factor of an object and the tristimulus values of the standard observer. The tristimulus values of the object are then calculated.

These values are represented by chromaticity coordinates (x and y) and the luminance (Y) in a chromaticity diagram. The standard chromaticity diagram is a horseshoe shaped spectrum locus such as that shown in Figure 2-3. The colors are shown in this figure as a rough estimate of the actual colors represented. Inks used to print these diagrams cannot fully encompass the exact color since there is no space for black, gray or brown. Because of this, and since the luminance value (the z coordinate not shown) is absent, subsequent figures referring to color measurements will show only the horseshoe and not the approximate colors it represents.

The experimental setup for this experiment is nearly identical to spectroelectrochemistry. The polymer film is poised at a potential and held under active potential control while the colorimeter (in this work a Minolta CS 1000) is used to obtain the Luminance (Y), and CIE x and y coordinates as a complete description of color. Besides the Y, x and y data from the colorimeter for the polymer film, the values for an illuminating light source are also required to translate the above values into L*a*b values that can be input into image processing programs to conveniently represent real colors. The light source in these experiments is nominally a 5000 K source, but is measured through ITO/ glass and solution before or after each experiment to correct for drift or warm/cool spots in

λ the source. From the color values, it is possible to establish the dominant wavelength ( d), and the color purity of the polymer at a specific potential, usually at the end of the doping range. The dominant wavelength is the wavelength of color (on the spectrum locus) whose 46

y 0.9 S x

OO 0.8

0.7 Purity (+0.87 V) = 0.0588 Purity (-0.93 V) = 0.475 0.6

0.5 Purity(+0.87 V)

0.4

0.3

0.2

Purity(-0.93 V) 0.1

0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 x

Figure 2-3. CIE Colorimetry data for PEDOT-TBAP in ACN relative to the 1931 standard observer. Film was identical to that used for spectroelectrochemistry (Figure 2-2). Color purity is designated by the ratio of the distance from the light source to a point and the distance from the light source to the pure color on the edge of the locus.

chromaticity is on the straight line between the sample and the illuminant source. The color

purity is the distance between the illuminant and a point divided by the distance between

the illuminant and spectrum locus and indicates how close a sample point is to a pure 47 wavelength. Colorimetry data for PEDOT switched in ACN/TBAP is shown in Figure 2-3 and the indicated values for color purity are shown as well.

2.5 Cells and Reference Electrodes for Electrochemistry

Electrochemical methods have proven the most useful means of characterizing CPs throughout recent years. With the exception of the combination in situ methods

(spectroelectrochemistry, colorimetry and EQCM), most electrochemical experiments can be performed in the relatively simple cell shown in Figure 2-4. This cell consists of a glass vial and a Teflon cap milled at the University of Florida with several holes drilled into it.

Typically, a Pt flag electrode fabricated from a piece of Pt foil and wire spot welded together is used as an auxiliary electrode. This is placed through the Teflon cap which is then assembled with the body and taped together with electrical tape so torque applied from the leads does not move the cap. A Pt disc (source: BAS, often referred to as a button) working electrode is inserted through a hole in the cap and is secured a slight distance from the bottom of the cell by either friction or an o-ring placed on the outside of the cell. The hole for the working electrode has a small groove cut into it so that it can accept ITO/glass slides in the standard dimensions produced by Delta Technologies. This groove is cut so the ITO face will be perpendicular to the auxiliary electrode. The reference electrode is inserted through a similar hole and held in place by the same methods and there is a small hole for a piece of either Ag or Pt wire placed close to the larger reference hole so either of these pseudo reference electrodes could be used instead. Solution is added through one of the electrode holes and the cell is filled to the point where all of the electrodes are bathed in solution. Typically 10 mL of [10 mM] monomer solution is prepared and 2-4 mL is used at a time. Higher monomer concentrations (up to 0.1 M) are used for extensive film 48 deposition on ITO electrodes but for deposition on Pt, 10 mM is sufficient for several experiments. The leads for each of the above mentioned electrodes are connected to a potentiostat by either alligator clips or preferably micro clips since the current magnitude tends to be small. Electrochemistry throughout this work was performed on one of several

Perkin-Elmer (formerly EG&G) Princeton Applied Research (PAR) potentiostats models

263A, 273 or 273A controlled via GPIB (National Instruments) bus by a computer running

Windows 98 or 2000 and Scribner Associated CorrWare 2 electrochemical software.

Additionally, a Pine bipotentiostat (model ABP-AFCBP1) and a programmable constant current source was used for certain experiments (noted). Data is evaluated by plotting in either CorrView (Scribner) or Microcal’s Origin.

Several different types of reference electrodes were used depending on the circumstances; all with the objective of measuring the interfacial potential at the working electrode. The ideal reference electrode for most aqueous experiments is the saturated calomel electrode (SCE). The redox process governing the potential of this electrode is

- - Hg2Cl2 +2e ' 2Hg+2Cl. This electrode is available as a kit (BAS RE-2) and is assembled by adding mercury prior to use. It consists of an H-cell where one arm contains a pool of mercury with calomel paste (Hg2Cl2)followedbyasaturatedKClsolutionontop of this paste. A small amount of glass wool is placed on top of the calomel paste to keep it from moving if jarred and the electrode is filled with saturated KCl solution so that two halves of the H-cell are in contact with the same solution. A Vycor tip separates the KCl solution from the sample medium. This tip is kept immersed in saturated KCl(aq) when not in use and a U-tube was fabricated at the University of Florida glass shop to hold two SCE electrodes for comparison. These electrodes (four total were used) were compared two at a 49

Saturated KCl Solution 2 cm KCl Crystals 7.5 cm Calomel Paste 6.5 cm Hg

6 mm 6 mm 6 mm

Figure 2-4. Schematic of electrochemical cell suitable for most electrochemical techniques. Working and reference electrodes are obtained from BAS. Cell, cap and auxiliary materials are fabricated on-site. Both SCE electrodes (shown in cell) and Ag/Ag+ electrodes (top right) are shown for comparison.

time by connecting a multimeter to the Pt wire protruding from the electrode body and

measuring the potential difference between the two references. In no cases was this ever

more than 10 mV and was typically near 1 mV indicating excellent stability with regard to

potential drift. While prolonged use in aqueous solutions poses no problems for SCE

electrodes, organic solvent use can damage these electrodes after some time. Specifically,

KCl is poorly soluble as is KClO4 in ACN. This leads to plugging of the frit by precipitate 50 which eventually increases both the impedance and junction potential of the reference.

While frequent calibration was used here to ensure that the potentials measured were valid, prolonged use suggests a slightly modified SCE electrode. NaClO4 and tetraethyl ammonium chloride TEACl are much more soluble in ACN than the Li-salt and while reference electrodes containing these salts are not available commercially, they can be easily prepared.

All reference electrodes here were calibrated by the use of a 5 mM standard solution of ferrocene (Fc) in 0.2 M LiClO4 - ACN. A cyclic voltammetry experiment was conducted

+ to measure the E1/2 of the Fc/Fc redox couple. Typically, scanning over the potentials -

500 mV to +600 mV at 100 mV sec-1 included the peak associated with this transformation.

Averaging the peak potentials yielded the E1/2 and this was noted for comparison to other reference electrodes and to account for drift. This was consistently at +344 mV vs.SCE,a value consistent with current literature values of 356 mV.5

In experiments requiring the absence of air and water, SCE is unsatisfactory due to two factors. First, water leaks through the Vycor frit contaminating the solution. This can be alleviated by the use of a salt bridge or a double junction with a second Vycor frit separating the reference from the sample solution. The middle solution contains organic electrolyte. This works in principal, but this adds the uncertainty of another junction potential. The cell design used throughout does not adequately seal out air and these types of experiments are best performed in an Ar dry box. SCE electrodes, double junction or not are incompatible with the dry box atmosphere. In these types of situations two options are

5. Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications Wiley: New York. 2001; p 54. 51 available. First, a Ag/Ag+ electrode can be constructed from a silver wire immersed in a solutionof1-10mMAgNO3 in 0.1 M TBAP in ACN separated from the solution by a

Vycor frit. This electrode works reasonably well on the benchtop but the silver ions leak through the frit, are reduced by DMF and are photochemically unstable necessitating near daily preparation of the reference and reference solution. Additionally, the Vycor frits available from BAS seem to require some water in their preparation in order to have a low enough junction potential to be used as a reference. When prepared in rigorously dry conditions, they simply do not work. Thus in the dry box, a Ag wire pseudo reference is used and its potential is checked repeatedly since it depends on the sample solution composition. This potential is often -30 mV compared to SCE but can be as much as -300 mV vs. SCE when checked by the standard Fc solution above. The values of several available reference electrodes and their conversions to SCE are shown in Figure 2-5. The value for Ag° is shown as a range since this potential drifts. SCE appears to be the most popular electrode used throughout the literature and, thus, all potentials here are reported vs. SCE. Rigorous electrochemistry demands that when Ag wire pseudo references are used, Fc is added at the end of the experiment and the potential reported vs. Fc. However, since the standard Fc solution containing ACN and perchlorate is nearly identical in composition to most of the solutions used here to characterize polymers, the potential difference should be negligible thus removing the need to use Fc as an internal standard for most experiments.

The accuracy of electrochemical potentials reported in this work is affected by several factors. The absence of any temperature control adds an uncertainty of 5-10 mV to the potential values measured. Drift in the reference electrode adds another 10 mV of 52

-0.5 Ag/Ag+ E[Vacuum] = EAg/Ag+ + 4.66 V

-0.4

ESCE = EAg/Ag+ + 0.2578 V -0.3 Ag (Pseudo) SCE

ESCE = EAg/AgCl - 0.0442 V -0.2 Ag/AgCl

ENHE = ESCE + 0.2412 V E = E - 4.4 V -0.1 SCE [Vacuum]

EAg/AgCl = ENHE - 0.197 V 0 NHE E[Vacuum] = ESCE + 4.4 V + 4.16 [Vacuum]

Figure 2-5. Reference electrode conversion values. Note: Ag wire as a pseudo reference is typically converted to SCE by subtracting 30 mV but can vary by several hundred mV and needs to be calibrated with each use.

uncertainty and for those experiments conducted with Ag wire pseudo references, the

uncertainty is probably as much as 50 mV given the dependence of potential on solution

composition. Considering the comparatively larger effects of small changes in

polymerization conditions, these values are acceptable.

2.6 General Electrochemistry and Cyclic Voltammetry (CV)

Several electrochemical methods are used to characterize CPs. In general these can

be divided into controlled potential (potentiostatic) methods and controlled current

(galvanostatic) methods. The Nernst equation governs the applied potential E, the formal

redox potential E°’ and the concentrations of analyte at the electrode and has implications

for all of the electrochemical methods discussed here. The polymers in this work are

electrochemically prepared, a process that can be done by potentiostatic, galvanostatic or

CV methods. In a CV experiment, the potential is swept at a constant rate while the current 53 is measured.6 The data for the standard Fc solution described above is shown in Figure 2-

6 and the shape of the trace indicates several things about the kinetics of electron transfer to the electrode. The Fc/Fc+ couple is known as a reversible couple, a term that has very specific meanings to electrochemists that go beyond the simple similarity to other reversible systems. Reversible redox processes are those that have concentrations of oxidized and reduced species that are predicted by the Nernst equation. Additionally, the

∆ ∆ peak to peak separation ( Ep) at 25 °C is 58 mV although processes that have Ep of 60-

65 mV are usually considered reversible. The ratio of the peak currents, ip,a and ip,c are 1 and the peak current scales linearly with the square root of the scan rate. The shape is a reversible CV and is characterized by a generally increasing slope with a peak superimposed on it. Upon potential reversal, a similar shape is observed with the opposite current sign and slightly shifted potential at the peak. The peak is due to the diffusion of analyte through the solution and the limitations this places on effective mass transport to the electrode. The peak current is proportional to the rate of electrolysis at the electrode and higher rates of potential change correspond to higher electrolysis rates while the concentrations of electroactive species are adjusted to comply with the Nernst equation.

′ RT C EE= 0 + ------ln------O Equation 1 nF CR

6. More specifically, since digital instruments are used, the potential is stepped by very small amounts dictated by the resolution of the DAC in the potentiostat. This process is essentially analogous to an analog sweep. Bott, A. W. Current Separations, 1997, 16(1), 23-26. 54

Ferrocene electrochemistry is quite different than that of a CP because ferrocene is dissolved in solution and diffuses to the electrode where it is oxidized while an electropolymerized CP is usually electrode adhered. Figure 2-6 shows the waveforms for the CV experiment and data obtained for the Fc standard solution under these conditions.

Charging currents are responsible for most of the limitations of the CV experiment. These limitations are that there is a detection limit of ca. [10 µM], the charging current is directly proportional to the scan rate but the peak current is proportional to the square root of the scan rate, effectively limiting both the sensitivity and the scan rate. Figure 2-6 B shows the waveform for the common step method DPV that addresses some of these limitations and is discussed later. Since diffusion to the electrode is responsible for the offset in peak potential in CV traces, electroactive species that are electrode confined have nearly identical peak potentials on the forward and reverse scans. The redox processes for films of CPs in solution or on electrodes are rarely reversible because the extensive changes in polymer structure that occur on oxidative or reductive doping preclude compliance with the kinetic limitations of the Nernst equation.7

Figure 2-7 shows data from the electrochemical synthesis of PEDOT along with the

CV for the polymer on a Pt electrode. Figure 2-7 A shows the first scan of a cyclic voltammetric deposition at a scan rate of 50 mV sec-1. No electrochemical response is evident at potentials cathodic of +1 V vs. SCE. At ca. 1.2 V vs. SCE the current increases rapidly peaking at ca. 1.4 V. These values indicate the onset of polymerization (Eonset)and peak polymerization potential for the monomer (Ep,m) and are collected on the first scan

7. Monolayers and very thin layers of polymer can appear much more reversible that the systems treated here. 55

+0.385 V

· CV Conditions: Fe 20 mA · 100 mV/s · Scan Rate 1 Point/mV

Potential · E1/2(CV) = +0.335 - +0.344 Time

General Conditions: 5 mM Ferrocene 0.2 M LiClO4 Acetonitrile Pt electrode A = 0.02 cm2

+0.285 V +0.354 V +0.302 V Step Size + Amplitude Step Size

Step I2 Time · 5 mA

Potential · i = I2 - I1 I1 Time

E1/2(DPV) = +0.344

DPV Conditions: Step Time = 0.0167 sec Step Size = 2 mV Seg #1 Amplitude = 0 mV, 4 Points Seg #2 Amplitude = 100 mV, 1 Point

+0.333 V

-0.5 0.0 0.5 1.0 Potential /V vs SCE Figure 2-6. Comparison between CV and DPV methods and results for ferrocene.

and used as a metric to compare different monomers. The return scan after the peak in the

first scan crosses the anodic wave in a feature exclusive to CP polymerization and metal

deposition called a nucleation loop. This is due to the quantity of polymer put down due to

the first scan increasing the electrode area. Figure 2-7 B shows the remaining 5 scans of 56

PEDOT polymerization. The peak at 1.4 V remains and the peak current increases in value.

The other feature that is evident is the increase in current over the potential region between

-1 V and ca. 0.4 V due to doping and undoping of the polymer. CPs can be electrochemically deposited by potentiostatic deposition above Eonset or by galvanostatic deposition. Figure 2-7 A-1 shows the galvanostatic deposition of PEDOT at a current of 10

µA (500 µAcm-2). In this experiment, the potential is measured at constant current and initially peaks then slowly decreases because PEDOT is highly conducting and can deposit more easily on previously grown PEDOT than on bare Pt. The potential at the end of this polymerization is labeled Elim and under these conditions, the value of Elim for all the polymers discussed herein is important for other experiments. Potentiostatic deposition is also an effective means of depositing a polymer film and is used mainly when a specific amount of charge must be passed while the potential is maintained below a certain value.

This is the primary method used to deposit films on ITO.

Upon completion of the electrochemical deposition, the polymer coated electrode is rinsed in the solvent used for deposition and is placed in a solution of solvent and electrolyte without monomer. Figure 2-7 C shows the CV obtained for the PEDOT polymerization described above. This spectrum is similar in shape to that obtained for ferrocene except that the peak separation is wider and the peak definition is not as sharp.

The peak values are typically reported for conjugated polymers and the average is taken as the E1/2. Frequently the redox process from the p-type doped state to the neutral state is poorly defined making determination of an accurate E1/2 difficult. For an electrode adhered film, the peak current has a linear dependence on scan rate which is routinely verified over the scan rates 10 mV sec-1 to 300 mV sec-1. Polymers often have a linear scan rate 57

14 A. First Scan from CV Deposition 12 S 1.25 Ep, m x 10 1.20 1.15 OO 8 1.10 A-1 1.05 Galvanostatic Elim 1.00 Deposition 10 µA 6 0.95

Potential /V vs SCE 0.90 4 0.85 0.80 0 60 120 180 240 300 2 Time

0 nucleation loop -2 Eonset

-2 12 B. Second Through Sixth Scans from CV Deposition cm 10 8 sity /mA 6 Polymer Doping Current 4 2

Current Den 0 -2

E 4 C. PEDOT CV p, an, polymer 3 2 1 0 -1 -2 -3 Ep, cath, polymer -4 -5 -1.0 -0.5 0.0 0.5 1.0 1.5 Potential /V vs SCE Figure 2-7. Some electrochemical data for PEDOT. (A) Deposition by CV (first scan only) at 50 mV sec-1 and by galvanostatic methods (A-1). (B) Second through sixth scans for PEDOT CV growth indicating increasing quantity of polymer on the electrode. (C) CV of PEDOT on A = 0.02 cm2 Pt from the above CV deposition in monomer free [0.1 M] TBAP ion ACN at 50 mV sec-1. 58 dependence between 10 and 100 mV sec-1 but only the fastest switching polymers remain in the diffusion controlled regime above 100 mV sec-1. This points out one of the difficulties of the CV experiment which is that it contains both kinetic (scan rate) and thermodynamic data (E1/2) in the same experiment. This is often difficult to separate and other experiments are employed to separate these energetic factors.

CV deposition for a number of polymers allows the comparison of how changing the monomer structure and electronic properties affects Ep,m and Eonset. Figure 2-8 shows the CV deposition for EDOT as above and shows the peak polymerization potentials for several other monomers. From this several observations are made. First, when extending the monomer conjugation by going from Th to BiTh, the peak polymerization potential is decreased. Second, appending electron rich units onto the monomer unit further decreases

Ep,m and the combination of increased monomer electron richness (EDOT) and extended conjugation (TerEDOT) results in a cathodic shift of Ep,m by nearly 2 V. Third, for a given monomer conjugation length of one ring, changing the heteroatom from S to N reduces

Ep,m by 1.2 V. Utilizing the above observations, the organic chemist can avoid the poly(thiophene) paradox where the oxidation of monomer to form polymer occurs at potential where the polymer is unstable. It is this decreased Ep,m for EDOT resulting in more facile electrochemical synthesis that has been the reason being much of the current research activity surrounding this molecule.

Galvanostatic, potentiostatic and coulometric methods are also utilized extensively to study polymer electrochemistry. Of particular importance are coulometric methods where the charge needed to p-type dope and neutralize a polymer is necessary to compare to the changes in optical absorbance during this transition to determine coloration 59

5 mA cm-2

S H S O O N H S S N S O O

CH3 O O O O OO S S S S S OO OO

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 Potential /V vs SCE

Figure 2-8. Peak polymerization potentials for common monomers. Potentials where overoxidation are likely to occur are shaded in red. Extended conjugation decreases peak oxidation potential as does electron rich character of the parent heterocycle.

efficiency. Figure 2-9 shows results from a potentiostatic deposition of EDOT at four

potentials between 1.2 V and 1.5 V vs. SCE followed by immediate charge neutralization

to -1 V. The experiment is started from open circuit potential which accounts for the

variations in the first 30 seconds of Figure 2-9 B. This experiment is useful for determining

the doping level of PEDOT by assuming a 100% deposition efficiency and using the

amount of charge passed on neutralization to determine how many sites are occupied.8

Most polymers have a doping level of between 0.25 and 0.3 when highly doped which

indicates a cation every three to four rings. 60

1.5 A. Potential 1.5 V S Profile 1.4 V x 1.3 V 1.2 V OO PEDOT Deposition 1.0

0.5

0.0

Potential /V vs SCE Open Circuit (No Deposition) -0.5

Neutral PEDOT -1.0

0.05 B. Potential Square Wave for PEDOT

Qgrowth 0.04 1.5 V Qdedoping 1.4 V 0.03

1.3 V 0.02

Charge /Coulombs 0.01

0.00 1.2 V

0 306090120150 Time /Seconds

Figure 2-9. Determination of doping level by chronocoulometric deposition. (A) Potential profile for determining charge passed during deposition and charge passed during de- doping. (B) Data for EDOT deposition and de-doping.

8. Randriamahazaka H.; Noel V.; Chevrot C.; J. Electroanal. Chem. 1999, 472(2), 103- 111. 61

2.7 Differential Pulse Voltammetry (DPV)

Differential pulse voltammetry (DPV) has several advantages over CV while allowing the determination of the same basic values for polymer films. First, it is several times faster than CV over the same potential range. Second, it more easily allows determination of E1/2 values for a polymer since for a reversible system, the peak potential is the same on the forward and reverse scans and corresponds to E1/2. Figure 2-6 (above) shows both the CV discussed in the previous section and the DPV for the ferrocene standard solution. Qualitative evaluation of reversibility involves comparing the peak heights and peak to peak separation. As mentioned previously, conjugated polymers do not undergo reversible doping processes and so the peaks for the anodic and cathodic scans are not symmetrical and do not occur at the same potential but the DPV results are more symmetrical because of the more effective rejection of charging current. Third, DPV factors out charging current by sampling the current twice at each potential. The current is sampled first at time τ’ immediately before the pulse and then again at τ immediately after the pulse. The current reported is the differential current δi = i(τ)-i(τ’) at each base potential. The pulse height (step size + step amplitude) is maintained at 100 mV at each potential in the experiment and the timing is chosen to minimize noise and maximize experiment speed. The shape of the DPV relative to the CV of Fc can be explained by considering the concentrations of species in the double layer during oxidation of Fc to Fc+.

At a potential cathodic of the formal redox potential of Fc (-0.2 V) early in the experiment no faradaic current (current associated with Fc oxidation) occurs before or after the 100 mV pulse because the pulse height is too small to cause oxidation. The differential current is thus near zero. At anodic potentials near the end of the experiment, Fc is oxidized at the 62 maximum rate and the step size is still too small to change this rate making the differential current again essentially zero. Therefore, only near the formal redox potential is any differential current observed because near this potential, Fc is oxidized at a rate that is sub- maximal and the step size is large enough to perturb this equilibrium and result in faradaic current. The discrimination against charging currents is due to the step size and from the pre-electrolysis (the three points collected and discarded before the step) establishing apparent bulk concentrations which are then observed by the pulse.

2.8 In Situ Conductivity

Conductivity is perhaps the most important physical property of conjugated polymers and surprisingly, few papers contain this data. The worst effect of this is the misleading conclusions drawn from the current debate on n-type dopable and low band gap materials where the sole method of characterization and proof of n-type doping is a cyclic voltammogram. While the equilibrium in situ conductivity experiment described below is non-routine in that it involves two potentiostats and requires somewhat lengthy data workup, it is certainly within the capabilities of most groups. The setup described here allows the determination of conductivity onset, comparison of the magnitudes of p-type and n-type doped conductivity magnitudes, and the conductivity profile at high doping levels.

Methods for determining absolute conductivity are not definitively described in this section although proposed methods for doing so are mentioned.

The first mention of conjugated polymer deposition on interdigitated microelectrodes (IMEs or IDEs) from Wrighton describes the fabrication of molecule- based transistors and is the basis for the conductivity method used throughout this work with few changes.9 Subsequent Wrighton publications10 refined the technique which is 63

1234 Pins 1 & 3: 10 µm line and gap widths 25 Pt digits on each bus

Chip

Pins 2 & 4: 5 µm line and gap widths 50 Pt digits on each bus Figure 2-10. IME photos and schematics. Photos courtesy of IMT, University of Neuchâtel

summarized as follows. A monomer is electropolymerized galvanostatically (10 µA, 0.5

mA cm-2) for 250 sec (2.5 mC) on a 0.02 cm2 Pt button as an initial guess for the ensuing

polymerization on an IME (Figure 2-10 A). The limiting potential at 250 seconds (Elim)is

then used as the initial potential for deposition on an IME.11 Pictures of the IMEs are shown

9. Kittleson, G. P.; Wight, H. S.; Wrighton, M. S. J. Am. Chem. Soc. 1984, 106, 7389- 7396.

10. (a) Paul, E. W.; Ricco, A. J.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 1441-1447. [polyaniline] (b) Thackeray, J. W.; White, H. S.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 5133-5140. [poly(3-methylthiophene)] (c) Ofer, D.; Crooks, R. M.; Wrighton, M. S. J. Am. Chem. Soc. 1990, 112, 7869-7879. [PTh, PPy, PANI].

11. IMEs were obtained from Dr. Giovanni C. Fiaccabrino, Institute of Microtechnology, (IMT), University of Neuchâtel, Rue Jaquet-Droz 1, 2007 Neuchâtel. email: Jean- [email protected]. www-samlab.unine.ch. The devices used herein contain 2 IMEs on one chip, the first with 5 µm Pt line and gap widths, the second with 10 µm Pt line and gap widths. Each IME is 1 µm in length and for most monomers and cleaning methods, superior results are obtained with PECVD nitride passivation on the surface of the Pt so that only the edges are exposed. 64 in Figure 2-10. Two potentiostats are setup as shown in Figure 2-11 B where one potentiostat acts as a gate potentiostat and the other as a drain potentiostat measuring the interelectrode current.12 The gate potentiostat can be inexpensive as it is used mainly to supply a constant potential for polymerization and throughout this work a Pine bipotentiostat model AFCBP1 was used. When configured as shown, Elim is applied from the gate potentiostat followed immediately by application of a 10 mV drain voltage across the IME. The current is monitored for ca. 300 seconds at which time, if there is no drain current flowing, the experiment is terminated. The gate potential is then raised by 50 mV and the above experiment is repeated until a drain current starts to flow within 300 seconds indicating that the CP has bridged the electrode gap. Successful polymerization results for

PEDOT-TBAP is shown in Figure 2-11 C. Polymer deposition is allowed to continue until the drain current reaches 65 µA. This value was arrived at empirically and assumes that the polymer is deposited in near its most conducting state. When the drain current is kept below these values, it ensures in most cases that the polymer is more resistive than the leads used to perform the experiment. In cases where the polymer in its most conducting state is more conductive than the leads, the conductivity trace plateaus at a resistance equivalent to ca.

160 Ω. In these cases a thinner polymer film must be redeposited.

The viability of the polymer film on the IME can be tested in several ways. In early attempts to perform this experiment, the individual IME leads were cycled and compared to the CV when the leads were shorted. If the peak currents are nearly identical, then there is electronic communication throughout the polymer and both IME leads. In every attempt at verifying this, the cyclics were indeed identical and this diagnostic was later dropped

12. Schiavon, G.; Sitran, S. Zotti, G. Synth. Met. 1989, 32, 209-217. 65

A. DETERMINING DEPOSITION POTENTIAL

1.3 20 µA E 1.2 (1 mA/cm2) lim 1.1 µ 1.0 10 A S 0.9 x 0.8 OO 0.7 Galvanostatic Deposition Gate Potential /V vs SCE 0.6 0.02 cm2 Pt electrode 0.5 0 50 100 150 200 250 Time /Seconds

B. EXPERIMENTAL SETUP C. POLYMER DEPOSITION ON IME 50 E AUX (A) GATE

REF (R) A removed EGATE = Elim

µ 40 CELL EGATE = 1.15 V vs SCE

ent / 30 VDRAIN = 10 mV V 20 GATE V (= 10 mV) Rlead = Drain Curr 10 imax R < R ≈ 160 Ω WWA lead V 0 R i ≤ 62 µA DRAIN max 0 20406080100 Time /Seconds

D. IN SITU CONDUCTIVITY ∆i 1 Conductivity DRAIN ∝ σ ∆ = VDRAIN RDRAIN

1 S 50 µA 1st Scan

2nd Scan Drain Current 1 mV/sec 4 Points/mV

Drain Current 0.975 1.000 1.025 100 µA Gate Potential /V vs SCE

-1.0 -0.5 0.0 0.5 1.0 Gate Potential /V vs SCE Figure 2-11. Procedure for In Situ Conductivity of PEDOT. 66 since appearance of a drain current during deposition of a well behaved CP is indicative of electronic communication across the previously insulating gap. Typically the polymer is broken in with the IME leads shorted in normal potentiostatic operation by repeated potential square waves encompassing the potentials over which conductivity data is to be collected. Typically 20 square waves are collected to ensure that peak current is reproducible. The gate-drain two potentiostat setup is then reconnected and polymer conductivity data is collected (Figure 2-2 D). A gate potential is applied so that the polymer is in its neutral, non-conducting state. When the current stabilizes and drops to near negligible values (< 1 µA although sometimes this value is not reached), a CV is run on the drain potentiostat between 25 mV around 0 V (0 V is effectively the gate potential value) at 1 mV sec-1 with 4 points collected every second. For n-type dopable polymers, where this data is to be collected, the polymer is scanned cathodically first. The gate potential is stepped in 50 mV increments and the drain CV is repeated after equilibration. The resulting data are several lines centered around each gate potential where the slope is the pseudo- conductivity and the reciprocal of the slope is the resistance. This data is worked up in a spreadsheet and the pseudo-conductivity value is reported vs. potential.

This data is called pseudo-conductivity (units: 1/Ω or S) because it does not carry the same meaning as volume conductivity (or just “conductivity”, units: S cm-1).

Conversion to volume conductivity from pseudo-conductivity requires knowledge of the film thickness, a troubling issue that, even with the sacrifice of an electrode for SEM analysis, tends not to yield useful data. The problem is even more difficult with highly conducting polymers such as PEDOT where the polymer does not make a uniform film across the electrode, but where fibrils are sufficiently conductive to reach the designated 65 67

µA. Polymer conductivity data in the literature is typically normalized to poly(3- methylthiophene) (P3MT) by assigning this polymer a conductivity of 60 S cm-1.13 This is anything but a fair assumption to make as film thickness, electrode geometry, dopant ion and several other variables affect the conductivity of polymer films. Cell constants have been determined by impedance methods with conductivity standards for IMEs and other electrodes and used to normalize conductivity but this method has not found widespread use among the chemists that are making and characterizing new materials probably due to its complexity and impedance requirements.14 It should be possible to normalize conductivity for the moderately conducting polymers proposed in this study by measuring the charge passed by the gate potentiostat during polymerization, assuming a 100% polymerization yield, polymer density of 1.5 g cm-3, 2.3 electrons per doped polymer repeat unit and uniform coverage of the electrode between the gaps (with known area A in cm2). However, polymers prepared throughout this work were polymerized using instrumentation lacking the capability (a coulometer) to measure charge passed and this was not attempted. Pickup has used a polymer film attached to a micrometer with thin film of evaporated gold to measure in situ conductivity.15 This method seems more difficult than the IME method. Several other groups use a banded electrode prepared from mylar Pt sandwiched between two sheets of Pt foil.16

13. Tourillon, G.; Garnier, F. J. Phys. Chem. 1983, 87, 2289.

14. (a) Sheppard, Jr., N. F.; Tucker, R. F.; Wu, C. Anal. Chem. 1993, 65(9), 1199-1202. (b) Paeschke, M.; Wollenberger, U.; Kohler, C.; Lisec, T.; Schnakenberg, U.; Hintsche, R. Anal. Chim. Acta. 1995, 305, 126-136.

15. Ochmanska, J.; Pickup, P. G. J. Electroanal. Chem. 1991, 297, 211-224.

16. Schiavon, G.; Sitran, S. Zotti, G. Synth. Met. 1989, 32, 209-217. 68

Cleaning of IMEs. Polymer coated IMEs were never subjected to physical cleaning (i.e. scraping or wiping) but were instead cleaned chemically. A solution of SC1

(5:1:1 H2O:NH4OH : H2O2) was prepared by adding the ammonia, then hydrogen peroxide to the water in a beaker. This cleaning solution was then heated to 50 °C under magnetic stirring and the tip of the IME was immersed in the cleaning solution with effort being made to keep the epoxied gold wires out of the SC1 as this is deemed a failure point for the Swiss electrodes used herein. The electrodes were kept in this stirred solution for 1 minute and then rinsed profusely with DI water (water pressure from the tap was used to attempt to loosen this film. For highly adherent films, this process was repeated but never more than a second time (total 2 minutes in SC1). The IMEs were shorted together and made the working electrode in a 3 electrode electrochemical cell and placed in a solution of 1 M H2SO4. The IMEs were then cycled at 20 mV/sec from + 1.15 V vs. SCE to -0.3 V vs. SCE; current passing at the negative extreme but not peaking. This was repeated 5 times consecutively and the scans compared. In cases where the scans were not equal, an additional 5 scans were performed until a consistent CV was observed. On particularly stubborn to remove films, the electrodes were cycled over -1 to -2 V vs. SCE in 1 M H2SO4 to evolve H2 from the electrodes followed by the oxidative cleaning/conditioning recipe shown above. This always resulted in electrodes that showed minimal current passed (10-8

A) when a potentiostat was connected across the IMEs and a 10 mV bias was applied.

Occasionally, one side of the IME would be unusable (rendering the entire IME unusable as well) presumably from breakage of the attachment from the IME chip to the epoxy header. This situation is detectable during the cleaning phase by proceeding with the cleaning as described above with all the IME leads shorted and treated in unison until a 69 consistent response is observed. The IME leads are then decoupled and cycled individually.

Overlay of the two corresponding IME CVs indicates that both leads are active.

2.9 Four-Point Conductivity on Free-Standing Films

The electrochemical in situ conductivity method described in the previous section is useful for investigating the changes in conductivity upon doping and undoping and the conductivity turn-on potential in solution. Characterization of non-solvated polymer samples of constant doping level requires the use of different electrodes and techniques.

Free standing films of CPs are prepared by either solution casting, spin coating or electrodepositing a material on a substrate, followed by rinsing, drying and removal of the

CP from the substrate. For the electrochemically prepared materials described herein, a thick polymer film is prepared electrochemically on ITO/glass, stainless steel, titanium or glassy carbon. The nature and morphology of this electrode exerts considerable influence on the properties of the resulting film, especially with regard to the surface reflectivity of the film. Free standing films of the highest conductivity are generally prepared by slow galvanostatic deposition at reduced temperatures on polished electrodes in a low vapor pressure solvent such as propylene carbonate (PC) or γ-butyrolactone (GBL) that can act as a plasticizer. At a sufficient film thickness, the electrode film coated electrode can be removed from the solvent and the film dried and removed by peeling from the electrode surface. The film can be cut with a razor or scissors to afford pieces that are of appropriate size for conductivity measurement. Factoring out the effects of lead and contact resistance precludes the use of two-probe methods. Three types of volume conductivity measurements are typically performed on free-standing polymer films depending on the 70 available instrumentation and film geometry. These are all variations of the four-point probe method and are described in the following sections.

Laboratories equipped for solid-state electronics research typically have a four- point probe device (such as the Signatone S-301-417 used in our laboratories) available with a predefined electrode geometry such as that shown in Figure 2-12. In this device, four equally spaced points on a line make electrical contact with a polymer film on an insulating surface. Between the outer pair of these points, a constant current (I) is applied while the potential difference arising (∆V) from this current is measured at the two inner contacts.

For typical conjugated polymers in this work, currents ranging from 10 µAto1mAwere injected through the films which produce potential differences between 10 µV and 10 V.

For any given film, ca. six currents within the above range were selected and the volume conductivities (ρ, S/cm) were found according to Equation 2 and averaged to give a single value for a polymer film. In this equation, derived by Valdes,18 t is the film thickness in cm.

Since the contacts are easily repositioned throughout the film, this process was repeated over several areas to gauge the homogeneity of the CP. This conductivity value is independent of electrode contact area as long as the distance between the points, d, is much greater than the film thickness, t.19 ∆V ρ = 2πt ------= 2πtR Equation 2 I

17. The particular probe is labeled as an 85 g weight, T.C. probe with a d spacing of 40 mils (40 mils X 0.0254 mm mils-1 = 1.016 mm)

18. (a) Blythe, A. R. Polymer Testing 1984, 4, 195-200. (b) Valdes, L. B. Proc. Inst. Radio. Engs. 1954, 42, 420. 71

+ 1 2 - I VI

w or b d° d d°

l

a

Figure 2-12. Four-point resistivity measurement on a free standing film. t, not shown, is the polymer film thickness. This method is applicable to the use of four-point probe devices (Signatone, Equation 2) as well as wires Ag -pasted across the film (Equation 3). Adapted from Smits, F. M. Bell. Sys. Tech. J. 1958, 37, 711-718.

Often, electrochemically prepared polymer films are brittle and not sufficiently

robust to handle the application of the probe points from the previous device to make

electrical contact. In these situations, and at times when the conductivity over a larger area

of film is to be measured as a function of time, or in laboratories lacking a four-point probe

device, leads may be attached directly to the film. This measurement is best performed on

a film that is much longer than it is wide and can be taped to a glass slide with double sided

tape. Four small wires (Ni or Cu was used throughout this work) are attached along the film

so that there is a much greater distance between the contacts measuring ∆V than there is

between the contacts injecting I (this can be visualized in Figure 2-12 where the outside d

values are small and the middle d value is comparably large. Current values similar to those

19. a t/d ratio of 0.1 is sufficient but correction factors need to be applied for ratios larger than 0.4. Since our probe has d = 1.016, a thickness of 102 µm is sufficient but care must be exercised for films 400 µm or thicker. 72 above are applied and the measured potential difference and Ohm’s law (V = IR)areused to calculate the resistance R. The volume conductivity is derived from Equation 3 where l is the length (in cm) between the potential measuring leads and w is the width of the film

(cm). l ρ = ------Equation 3 Rtw

Most CP systems can be treated by one of the two methods above. However, in devices or polymers where contacts must be irregularly attached, the equations and methodology for measuring conductivity in these systems have been described by van der

Pauw.20

2.10 Electrochemical Quartz Crystal Microbalance (EQCM)

During the deposition as well as oxidation and reduction of CP films, mass changes occur at the electrode. The coupling of electrochemical control to a quartz crystal microbalance is termed the electrochemical quartz crystal microbalance (EQCM) and is useful in elucidating some of the changes that occur in electroactive polymers.21 The

EQCM is simple in concept and is based on the piezoelectric behavior (physical deformation in a potential) of a quartz crystal. In the experiment, an alternating field is applied between two Pt electrodes evaporated on opposite faces of a quartz crystal driving the oscillation parallel to the crystal faces. The Saurbrey equation (Equation 4) describes

20. (a) van der Pauw, L. J. Philips. Res. Repts. 1958, 13, 1-9. (b) van der Pauw, L. J. Phil- ips, Res. Repts. 1961, 16, 187-195. (c) Montgomery, H. C. J. Appl. Phys. 1971, 42(7) 2971-2975.

21. (a) Deakin, M. R.; Buttry, D. A. Anal. Chem. 1989, 61, 1147A. (b) Buttry, D. A. Elec- troanal. Chem. Bard, A. J., Ed.; Dekker: New York, 1991; Vol. 17 p. 1. (c) Buttry, D. A.; Ward, M. D. Chem. Rev. 1992, 92, 1355. (d) Ward, M. D. Phys. Electrochem. (Rubinstein, I. ed.) Dekker, New York, 1995, p. 293. 73 ∆ the relationship between mass (m) and resonant frequency ( f)ofaquartzcrystalwheref0 is the fundamental frequency (ca. 8.9 MHz for experiments in this document but it varies),

ρ is the density of quartz (2.648 g cm-2), µ is the shear modulus of quartz (2.947(10)11 g cm-1 sec-2) and n is the harmonic number of the oscillation (typically 1). The constants are –2f 2mn Equation 4 ∆f = ------0 -= –C m ()ρµ 12⁄ f

2 µ -1 usually lumped together to provide a single sensitivity factor (Cf, units: Hz cm g ) that should be checked for each crystal and sample solution used since the Saurbrey equation strictly applies only in the vacuum or gas phase. Because viscoelastic issues affect the resonant frequency of quartz, the rigidity of any polymer film must be checked by crystal impedance or conductance analysis at every oxidation state of the polymer studied to ensure the validity of the results and interpretation using the Sauerbrey equation.

The most important use of the EQCM is as a probe of the often complex ion transport mechanisms in CPs which depend on the rate of electrosynthesis and this morphology, solvent viscosity and polarity, and size and charge of the counter ion. In general for a p-type CP, it is expected that the mass should increase as the polymer is oxidized and decrease as the polymer is re-neutralized, a condition called anion dominance since the anion and its associated solvent ions enter the film. This can be reversed in polymer composites where a dopant anion such as poly(styrenesulfonate) can be co- synthesized with the film and thus immobilized in it. Such films tend to be cation dominant at low potentials as cations from the solution enter the film to charge compensate the sulfonate anions which are available from the recently neutralized polymer. 74

In this work, EQCM results were obtained using a Seiko EG&G (now Perkin-

Elmer) QCA 917 connected by BNC cable to the auxiliary in of a Perkin-Elmer PAR 273A potentiostat. The QCA outputs a signal ranging from 20 kHz V-1 or 2 kHz V-1. The 20 kHz

V-1 range is used during potentiostatic or more typically galvanostatic polymerization to monitor the higher mass of polymer deposited compared to the smaller mass changes which occur on doping and dedoping (2 kHz V-1 range). Scribner Associates CorrWare 2 was used to control the 273A and monitor the aux channel which has a range of ± 10 V. For a typical deposition, the frequency at open circuit is followed for at least 10 minutes until it stabilizes followedbysettingthezerofrequencyoftheQCAtoca.+5Vasmeasuredontheauxin.

This allows a 15 V X 20 kHz V-1 = 300 kHz range to be measured during deposition. It is common practice to express EQCM results in units of ng cm2 Hz-1 from Suerbrey’s original work. However, if the crystal is evenly covered (as inspected visually), the sensitivity can be expressed as Hz ng-1 and in this work is corrected for the theoretical sensitivity of the crystal22 used 1.09 ng Hz-1 and expressed as ng.23

Polymer rigidity was determined on a Hewlett Packard LF Impedance Analyzer model 4192A having a range of 5 Hz to 13 MHz. This instrument is not computer controlled and either a strip chart recorder is used or the full width at half maximum

22. The sensitivity of the crystal is determined in 0.5 M AgNO3 in 0.5 M HNO3 to deposit Ag at a current density of 0.25 mA cm-2. The calibration factor is calculated from the slope of the plot of mass of metal (converted from charge assuming 100% deposition yield) vs. frequency. It has been reported as 1.03 ng Hz-1 on the 2 kHz scale and 1.11 ng Hz-1 on the 20 kHz scale compared to the theoretical value of 1.09 ng Hz-1. These results are from Cameron, D. A. Ph. D. Dissertation, University of Florida.

23. Bruckstein, S.; Shay, M. Electrochim. Acta. 1985, 20, 1295. (B) Skoog, D. A.; West, D. M. Fundamentals in Analytical Chemistry; 4th ed.; Saunders: Philadelphia, 1982, p. 586. 75

(FWHM) for the QC resonance is deduced manually. The procedure is as follows: While still in the EQCM cell, the resonant frequency of the system is measured. The cell is then transferred to the impedance analyzer and scanned from 100 kHz below this resonant frequency to 100 kHz above it. The maximum is noted as is the FWHM. If the FWHM does not change appreciably on doping and is comparable to a bare crystal in solvent and dry, the film is assumed to rigid and the Sauerbrey equation is valid. CHAPTER 3 DONOR MEDIATED BAND GAP REDUCTION IN A HOMOLOGOUS SERIES OF CONJUGATED POLYMERS

3.1 Introduction and Literature Overview

The evolution of the field of conjugated electroactive polymers1 (CEPs) as replacements for inorganic materials in electronics devices has been fueled by the flexibility in tailoring specific electronic properties by rational chemical synthesis. To this end, organic light emitting and electrochromic devices are making rapid progress toward commercialization.2 The unique properties of CEPs are largely based on the ability to tailor the energy levels of the valence band (VB) and conduction band (CB) both relative to each other (band gap control) or in an absolute sense (modifying the ease of oxidation or reduction). Tailoring the band gap (Eg) of CEPs allows variation in emission wavelength, absorptive colors in electrochromic devices and conductivity in the neutral state. Changes in band energy allow optimization of interfacial energy level alignment between the polymer and electrode contacts in organic light emitting devices (OLEDs).3 Low band gap

1. Conjugated electroactive polymers are differentiated from conjugated polymers in that their propensity to undergo electrochemical doping and undoping reactions can be leveraged.

2. (a) Groenendaal, L. B.; Jonas F.; Freitag D.; Pielartzik H.; Reynolds, J. R. Adv. Mater. 2000 12(7), 481. (b) D. M. de Leeuw Physics World 1999 31. (c) Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Marks, R. N.; Taliani, C.; Bradley, D. D. C.; Dos Santos, D. A.; Brédas, J. L.; Lögdlund, M.; Salaneck, W. R. Nature 1999 397, 121.

3. Ishii, H.; Sugitama, K.; Ito, E.; Seki, K. Adv. Mater. 1999, 11, No. 8, 605-625.

76 77

CEPs are particularly interesting because of the possibility of high intrinsic conductivity in the neutral state, transparency in the p-doped conducting state, and their tendency to achieve reduced band gaps through conduction band energy lowering making stable n-type doping possible.

Several factors affect the band gap in conducting polymers and the field in general has been well reviewed.4 These factors are detailed in Figure 3-1 and are: The stability of the quinoid form (reduction of bond length alternation reduces band gap) of a poly(heterocycle), planarity of the repeat unit, interchain effects, resonance contributions5 and donor-acceptor effects.6 While ensuring planarity of the repeat unit is difficult to achieve without compromising solubility,7 and interchain effects have been only sparsely described,8 most of the research effort has been devoted to increasing the stability of the quinoid character in the neutral polymer and capitalizing on donor-acceptor effects.

Despite some recent success for poly(acetylene), poly(thiophene) and poly(pyrrole),9 there exists no accurate and general computational tools to probe the relationship between

4. Roncali, J. Chem. Rev. 1997, 97, 173.

5. (a) Wudl, F.; Kobayashi, M.; Heeger, A. J. J. Org. Chem. 1984, 49, 3382. (b) Koba- yashi M.; Colaneri N.; Boysel M.; Wudl F.; Heeger A.J. J. Chem. Phys. 1985 82(12) 5717-5723. (c) Roncali, J. Chem. Rev. 1992, 4, 711-738.

6. Zhang, Q. T.; Tour, J. M. J. Am Chem. Soc. 1998 120, 5355-5362.

7. Yao,Y.;Lamba,J.J.S.;Tour,J.M.J. Am Chem. Soc. 1998, 120, 2805-2810.

8. (a) Smith, R. C.; Fischer, W. M.; Gin, D. L. J. Am Chem. Soc. 1997, 119, 4092-4093. (b) Hu, D.; Yu, J.; Barbara, P. F. J. Am. Chem. Soc. 1999, 121, 6936-6937.

9. (a) Brédas, J.-L.; Cornil, J.; Beljonne, D.; Dos Santos, D. A.; Shuai, Z. Acc Chem. Res., 1999, 32, 267-276. (b) Salzner, U.; Pickup, P. G.; Porer, R. A.; Lagowski, J. B. J. Phys. Chem. A 1998, 102, 2572-2578. 78 structure and electronic properties for materials that derive their properties from extended range inter- and intramolecular interactions. We have embarked upon a program to understand how substitution affects band gap in perfectly alternating Donor-Acceptor (D-

A) copolymers.

The quintessential low band gap conducting polymer is poly(isothianapthene), a thiophene ring fused to a benzene ring across the 3 and 4 positions of the thiophene ring.

Polymerization occurs through the 2 and 5 positions of the thiophene ring. This arrangement induces a competition for aromaticity in the monomer repeat since it is impossible for both thiophene and benzene to simultaneously be aromatic. Benzene itself has an energy of aromatization (Eres) of 1.56 eV while thiophene is 1.26 eV.10 These energetics predict that benzene will remain aromatic and force the thiophene to adopt a pseudo-diradical electronic state at the 2 and 5 positions. When polymerized, this amounts to the bond connecting the thiophenes existing in the quinoid state, which lowers the band gap of the system by decreasing bond length alternation. While this proves to be an effective method of lowering polymer band gap (PITN has a band gap of 1.1 eV, about 1 eV lower than PTh), This structural topology proved limited in scope.

Studies over the last decade have been geared almost exclusively to manipulating band gap through either planarization of the repeat unit by fusing heterocycles together, or by creation of polymers that have alternating donor and acceptor moieties or some combination of these two approaches. The donor-acceptor route has by far the most utility in terms of diversity in synthetic possibility while avoiding solubility problems that plague planar organic molecules. The route described in this chapter and the next chapter is more

10. March, J. Advanced Organic Chemistry; Wiley: New York, 1985. 79

accurately described as a donor-acceptor-donor route since the need to electropolymerize

these monomers enforces the condition that easily oxidized and electropolymerizable

heterocycles are at the perimeter of the molecule. The donor-acceptor approach can be

loosely classified into two distinct families; those monomers for which resonance

structures can be drawn that include the acceptor group and main chain polymer

conjugation together, and those that which cannot and use the acceptor to inductively

modify the backbone as shown in figure 3-1.

C12H25 C12H25 S N N x N PITN N N N N N C6H13 C6H13 nm Planar Poly(Pyridines) S S RESONANCE S PLANARITY 2 2 EFFECTS OO x PT S,S-Dioxides

BONDND LENGTH INTERCHAIN ALTERNATION EFFECTS C8H17 x S S PA S x

C8H17 C8H17 P3OT DONOR-ACCEPTOR EFFECTS

CN S O O CN CN S

O SSx O O PBEDOT-CNV PBEDOT-DCF D-A Included in Main Chain Resonance D-A Separate from Main Chain Resonance Chapter 3 (This Chapter) Chapter 4

Figure 3-1. Overview of methods for the modification of band gap. Cutout indicates methods discussed in this chapter and the following chapter.

The logic behind the donor-acceptor approach (hereafter D-A) is that the high

HOMO of the donor and the low level of the LUMO are incorporated into the resulting 80 monomer and thus polymer electronic structure.11 Figure 3-2 illustrates this concept with data for PEDOT,12 poly(cyanoacetylene) (PCA)13 and anticipated data for PBEDOT-CNV

(vide infra). In this example, the band gaps are established through the onset of the π to π* transition in the spectroelectrochemical series and the energies are estimated from the potentials for the electrochemical doping/undoping redox couples. Specifically, the HOMO for PEDOT was obtained by cyclic voltammetry, the band gap by spectroelectrochemistry, and the position of the LUMO was deduced by subtraction of the band gap from the HOMO position.14 PCA poses a similar set of problems as there is no observable oxidation but there is a reduction wave to which the band gap (UV-Vis of the neutral form) was added to arrive at the low lying HOMO energy. PBEDOT-CNV spectra is described in detail below and does not require similar compromises.

In this contribution, the synthesis of a matrix of structurally homologous conducting polymers differing systematically in the density of electron rich chromophore in direct conjugation with the polymer backbone is described. In addition, the effect of increasing the electron density along the polymer main chain is detailed. This approach is an effective method of decreasing the band gap along a polymer chain and lends insight into

11. (a) van Mullekom, H. A. M.; Vekemans, J.A.J.M.;Havinga,E.E.;Meijer,(Review not Released Yet). (b) E. W. Havinga, E. E.; ten Hoeve, W.; Wynberg, H. Synth. Met. 1993, 55-57, 299.

12. PEDOT data was obtained as described in the experimental section detailing spectro- electrochemistry.

13. Gorman, C. B.; West, R. C.; Palovich, T. U.; Serron, S. Macromolecules 1999, 32(13), 4157-4165.

14. Only recently has n-type doping of PEDOT been demonstrated. Even under rigor- ously controlled atmospheric conditions its stability is unremarkable. Ahonen, H. J.; Lukkari, J.; Kankare, J. Macromolecules 2000, 33(18), 6787-6793. 81

-2.5 CB

-2.0 + -1.5 Eg E1

-1.0 Eg E1 E 2 E -0.5 VB 2

0 E

Energy /V vs Ag/Ag g

+0.5

+1.0 x S CN S CN x S O

OO O x O O Figure 3-2. The Donor-Acceptor approach, Alternating donor and acceptor moieties results in a polymer that has the combined optical properties of the parent donor or acceptor monomers.

techniques for modifying the electronic properties of conjugated materials in general. In

this work, the generally held assumption that acceptor strength must continually increase

in order to decrease the band gap of a conjugated polymer is examined and found lacking.

Moreover, our approach to creating the acceptor unit through the high yielding

Knoevenagel condensation has general applicability to the synthesis of a variety of high

molecular weight, conjugated step growth condensation polymers which are soluble and

processible for use in OLEDs.15

3.2 Synthesis and Monomer Properties

Overall synthetic yields for D-A monomers are often hampered by the need to cross

couple two aromatic heterocycles with the concomitant loss in yield for the preparation of

15. (a) Greenham, N. C., Moriatti, S. C.; Bradley, D. D. C.; Friend, R. H.; Holmes, A. B. Nature 1993, 365, 628-630. (b) Hanack, M.; Segura, J. L.; Spreitzer, H. Adv. Mater. 1996, 8, 663-666. 82 organo-tin, -ZnCl, or -boronate reagents and during the coupling itself. The cyanovinylene acceptor core described here lessens some of the complications of combining pre-made D-

A units and eliminates the need for expensive and toxic tin reagents completely while placing the lower yielding synthetic steps early in the monomer synthesis.

In this work, the premonomers were prepared by using the reaction schemes outlined in Figure 3-3. Aromatic carboxaldehydes of thiophene (commercially available)

EDOT and EDOP were synthesized by the Vilsmeier formylation of the parent heterocycle.

The structural diversity of this family is due to its ability to react three unique aromatic carboxaldehydes with two unique aromatic acetonitriles to prepare six monomers. The inclusion of the recently rediscovered 3,4-ethylenedioxypyrrole (EDOP) unit and its ability to react by Vilsmeier chemistry to form EDOP carboxaldehyde increases the structural variety of monomers available.16 An aromatic acetonitrile acts as the second premonomer component and the two used in this work were prepared by a catalytic route shown in

Figure 3-3 (EDOT-ACN) or purchased (Th-ACN). The Knoevenagel condensation of an aromatic aldehyde with an aromatic acetonitrile is used as the last step in monomer synthesis with simultaneous preparation of the cyanovinylene core.17 The monomers available with this route are listed in figure 3-4 and are described by the matrix which combines the two types of premonomer to make six monomers. The Knoevenagel reaction

16. (a) Merz, A.; Schropp, R.; Dötterl, E. Synthesis 1995, 795-800. (b) Savage, D. J.; Schell, B. A.; Brady, B. K. Imaging Element Containing Poly(3,4-Ethylene Dioxypyr- role)/Styrene Sulfonate. U. S. Patent 5,665,498, September 9, 1997. (c) Schottland, P.; Zong, K.; Gaupp, C. L.; Thompson, B. C.; Thomas, C. A.; Giurgiu, I.; Hickman, R.; Abboud, K. A.; Reynolds, J. R. Macromolecules 2000, 33(19), 7051-7061.

17. Ho, H. A.; Brisset, H.; Elandaloussi, E. H.; Frére, P.; Roncali, J. Adv. Mater. 1996, 8, 990. 83

is carried out in ethanol or t-butanol with a slight excess of potassium t-butoxide and

routinely allows for unoptimized yields of 80 to 95 percent. Vilsmeier Formylation O S S a H

OO OO

Catalytic Alkylation S S b CN

OO OO

Vilsmeier Formylation and N-benzyl Deprotection BzBz O H O N N N cdH H

OO OO OO

Knoevenagel Condensation O CN S SS e H CN S

Figure 3-3. Reactions used to synthesize cyanovinylene monomers. a.POCl3, 1,2- dichloroethane. b.i.BuLi,THF,-78C,1hii.ZnCl2 iii. Ni(acac)2,PPh2Cy, BrCH2CN, THF. c.POCl3, 1,2-dichloroethane. d.cat.H2. e.KOtBu,EtOH,reflux,1h. 84

Aldehyde H Component S S N O O O O O

H HHO O Acetonitrile Th-CHO EDOT-CHO EDOP-CHO Component

H CN CN S CN S CN N O O

S S S O S O Th-ACN BTh-CNV Th-CNV-EDOT Th-CNV-EDOP

H CN CN S CN S CN N S S S O S O O O O O O O O O O O

EDOT-ACN EDOT-CNV-Th BEDOT-CNV EDOT-CNV-EDOP

Figure 3-4. Synthesis of Cyanovinylenes. Matrix describing synthetic access to the cyanovinylene family of monomers. Premonomers are condensed with KOtBu in EtOH (reflux).

Electron richness for the three unique heterocycle cores in this study follow the

trend EDOP > EDOT > Th as evidenced by monomer oxidation potential. Evidence that

EDOP is considerably more electron rich than EDOT is noticeable as solutions of EDOP

readily polymerize on standing to form deep blue/purple solutions while EDOT undergoes

some initial chemistry but passivates and does not completely polymerize to dark solutions.

λ Monomer UV-Vis spectra arranged in order of increasing max are shown in Figure 3-5.

These follow the expected trend where the less electron rich monomers have blue-shifted

λ max values compared to the more electron rich monomers.

There are some other notable features of the UV-Vis monomer spectra. In those

monomers where there is an EDOT that is directly adjacent to the cyano group (EDOT-

CNV-Th, BEDOT-CNV and EDOT-CNV-EDOP), a further transition at high energy 85 occurs for these monomers and appears above the 190 nm UV-cutoff for acetonitrile.

Additionally, the monomers that have thiophene in the position adjacent to the cyano group, the complement of the previous selection, BTh-CNV, Th-CNV-EDOT and Th-

λ CNV-EDOP, all have higher molar absorbances at max than monomers with EDOT immediately adjacent to the cyano group.

3.3 Monomer Structural Features and Crystallography

The x-ray crystal structure of BTh-CNV was determined at -100(2) °C. It crystallizes in the orthorhombic space group Pbca. The structure determination confirms the cis relationship of the thiophene not adjacent to the cyano group (ring B) and the cyano group. The sulfur atoms are cisoid to the carbon-carbon double bond. While there is considerable disorder in the crystal structure, it is due to the entire molecule being flipped end for end at certain sites. In contrast to EDOT-CNV-Th, which also contains a Th in the ring B position, there is no disorder about the thiophene-vinyl linkage with the thiophene adopting the configuration of the major contributor in EDOT-CNV-Th. The molecule is close to planar with the angle between ring A and the cyano group being 5(1)° and the angle between ring B and the cyano group being 3(1)°. The angle made between the planes of the two heterocycle rings is 5(1)°.

The x-ray crystal structure of EDOT-CNV-Th was determined at -120 °C. It crystallizes in the orthorhombic space group Pna21, with a single molecule in the asymmetric unit, and has the thiophene ring disordered over two unequally occupied (79%:

21%) orientations. The structure determination confirms the cis relationship of the cyano and thiophene substituents. The whole molecule is close to planar. Specifically, the plane of the double bond is inclined to the plane of the EDOT thiophene ring at an angle of only 86

Molar Absorptivity 30,000

λ S max 364 nm onset 418 nm ε 30500 CN S max BTh-CNV

O O λ max 378 nm onset 430 nm S ε max 21600 CN S EDOT-CNV-Th

λ S O max 383 nm onset 442 nm O ε 28300 CN S max Th-CNV-EDOT

O O λ max 394 nm O onset 449 nm S ε O max 22300 CN S BEDOT-CNV

λ S O max 394 nm onset 454 nm O ε CN N max 29100 H Th-CNV-EDOP

O O λ max 401 nm O onset 458 nm S O ε 27600 CN N max H EDOT-CNV-EDOP

200 300 400 500 600 700 800 Wavelength /nm Figure 3-5. Monomer molar absorbance (M-1cm-1) for UV-Vis region in acetonitrile.

3.4(3)° and to the plane of the other thiophene ring at an angle of 4.3(3) °, while the planes

of the EDOT and thiophene rings make an angle of 7.2(3)°. In contrast to the isomeric

compound Th-CNV-EDOT18, the conformation of the EDOT substituent is such that the 87 sulfur is cisoid with respect to the geminal nitrile substituent, while the major contributor to the disorder has the sulfur of the thiophene ring cisoid to the vicinal nitrile group, as observed in Th-CNV-EDOT, but with the minor contributor having a transoid orientation of these groups.

The x-ray crystal structure of Th-CNV-EDOT was determined at -125 °C. It crystallizes in the monoclinic space group P21/n with a single molecule in the asymmetric unit. The structure determination confirms the cis relationship of the cyano and EDOT substituents. The conformations of the two thiophene rings are such that their sulfur atoms are cisoid to the carbon-carbon double bond. Most significantly, the whole molecule is close to planar. Specifically, the plane of the double bond is inclined to the plane of the

EDOT thiophene ring at an angle of 6.6(3)° and to the plane of the other thiophene ring at an angle of 8.7(3)°, while the planes of the two thiophene rings make an angle of only

2.8(3)°. The cyano substituent points slightly out of the plane of the double bond. The bond lengths and angles are similar to those in other structurally related compounds and there are no unusually short intermolecular interactions.

The x-ray crystal structure of BEDOT-CNV was unable to be obtained due to lack of suitable quality single crystals. All methods described in Chapter 2 were attempted for the purpose of obtaining an x-ray crystal structure using acetonitrile, ethanol, chloroform, methylene chloride, benzene and various other chlorinated solvents in different crystallization environments.

18. G. A. Sotzing, C. A. Thomas, J. R. Reynolds and P. J. Steel. Macromolecules, 1998, 31, 3750-3752. 88

BTh-CNV EDOT-CNV-Th

Crystals Growing

Th-CNV-EDOT BEDOT-CNV

Th-CNV-EDOP EDOT-CNV-EDOP Figure 3-6. Perspective and top views of all the monomers in the cyanovinylene family. 89

The x-ray crystal structure for Th-CNV-EDOP was partially established. However, due to instability of the monomer, the crystal decomposed before complete data collection was possible. The partially resolved structure is shown in the figure but angles were unable to be obtained.

Table 3-1. Incline angles for the planes comprised of the thiophene, EDOT, EDOP and cyanovinylene units in the monomer crystal structures.

Name Ring Aa-CNVb Ring Bc-CNV Ring A-Ring B BTh-CNV 5.(1) 3.(1) 5.(1) EDOT-CNV-Th 3.4(3) 4.3(3) 7.2(3) Th-CNV-EDOT 8.7(3) 6.6(3) 2.8(3) BEDOT-CNV Th-CNV-EDOP EDOT-CNV-EDOP 4.3(1) 21.9(1) 24.5(1) a. Ring-A is defined by the four carbon atoms that make up the heterocycle nearest the cyano unit (left-most heterocycle in the structures in Figure 3-5). b. CNV is defined by the three carbon atoms in the ethylene and cyano group. c. Ring-B is defined by the four carbon atoms in the heterocycle β to the cyano group (right most heterocycle in structures throughout this chapter).

The x-ray crystal structure for EDOT-CNV-EDOP was determined at -100(2) °C.

It crystallizes in the triclinic space group P-1 with X molecules in the asymmetric unit. The conformation of the EDOT is identical to the EDOT conformation in EDOT-CNV-Th and the conformation of EDOP is identical to that in Th-CNV-EDOP. Specifically, the sulfur of EDOT is transoid to the carbon-carbon double bond while the nitrogen of EDOP is cisoid to the carbon-carbon double bond. The molecule deviates considerably from planarity compared to the other monomers in this family. Ring A (EDOT) is nearly planar to the cyano group with an angle of 4.3(1)°. The angle between ring B (EDOP) and the cyano 90 group is 21.9(1)°, the largest deviation from planarity for any of the monomers in this family. Consequently, the angle formed between the planes of EDOT and EDOP is

24.5(1)°.

3.4 Computational Explanations of Monomer Properties

The interesting behavior elicited by the EDOT-CNV-Th monomer bears further scrutiny. It is the only monomer in this family which crystallizes in a form where the thiophene unit is found to be in two different orientations relative to the cyano group. These orientations are related by single bond rotation about the ethylene-thiophene bond and this type of rotation is not observed for BTh-CNV, the other monomer containing a thiophene in the Ring B position.

The crystal structure of BTh-CNV, however, is not without its own disorder arising from what effectively amounts to translation of the cyano group relative to other molecules in the crystal rather than the rotation of individual bonds. Calculations were undertaken on structures19 of EDOT-CNV-Th and Th-CNV-EDOT in order to determine whether the cause of the disordered crystal structure is due simply to a lower barrier of rotation for the thiophene when it is β to the cyano unit in EDOT-CNV-Th compared to the case when it is

α as in Th-CNV-EDOT. The crystal structure coordinates for both Th-CNV-EDOT and

EDOT-CNV-Th were modified to vary the dihedral angle between the thiophene ring and the cyanovinylene unit. Semiempirical single point calculations at the PM3 level were performed on the unrelaxed monomer geometries and the total energy, HOMO and LUMO energies were extracted. PM3 is a semiempirical force field that accurately reproduces the

19. x-ray coordinates were modified to vary the dihedral angle between the Th ring and the remaining section of the molecule. 91

A

X-ray Structure Total Energy /kcal/mol Total

0.5 kcal/mol

X-ray Structure

8.2 B

8.0

/eV

7.8

MO Difference 7.6 Th-CNV-EDOT

HOMO-LU 7.4 EDOT-CNV-Th 7.2 0 30 60 90 120 150 180 Thiophene Twist Angle /degrees Figure 3-7. Single Point PM3 calculations on un-relaxed x-ray data for Th-CNV-EDOT (blue) and EDOT-CNV-Th (red) with specified twist angle about the thiophene ring. Part A: Total molecular energy as a function of twist angle. Part B: HOMO-LUMO difference as a function of twist angle.

bond lengths and angles observed in the crystal structures for this family of molecules and

was therefore deemed to accurately model the energetics of these systems. Additionally, 92 these calculations were performed at the semiempirical level to minimize computational times and because of the recognition that lower levels of theory are acceptable when the trends are more important than quantitative values. Figure 3-7 shows these results plotted as a function of thiophene twist angle for these two monomers.

From the results in Figure 3.7 A, several things are apparent. First, the lowest energy conformation appears to be at a twist angle that is remote from the two structures as determined by the crystallographic data for EDOT-CNV-Th and the single rotational structure found for Th-CNV-EDOT in the x-ray results. This minimum is found between

135 to 145 degrees and is presumably absent from the crystal structure due to crystal packing forces not accounted for in the calculations. Second, For Th-CNV-EDOT, the total energy. vs. twist angle data correctly predict the result observed in the x-ray data to be the most stable by roughly 0.7 kcal/mol, a number that is larger than kT at room temperature

(ca. 0.5 kcal/mol), where both monomers were crystallized. For EDOT-CNV-Th, the data is slightly more ambiguous. It predicts the minor conformer observed in the x-ray results to be the most stable, albeit by a value that is quite a bit smaller than kT. This implies that relatively free rotation about the cyanovinylene-heterocycle single bond is allowed while the molecule is crystallizing and becomes locked into the nearest planar conformation upon crystallization. The inability to clearly predict the most stable rotational contributor to the

EDOT-CNV-Th crystal data is likely due to lack of correction for relaxation of other bond distances and angles upon rotation.

Additionally, from the relationship between the thiophene torsional angle and the

HOMO-LUMO difference (Figure 3.7 B), it is apparent that the energy difference approximates the expected cosine function predicted for rotation about these bonds. 93

Furthermore, it appears that below a twist angle of 10 degrees on either side of perfectly planar, there seems to be minimal effect on the HOMO-LUMO difference for either material. This result has implications in the comparison of data derived from these monomers which, with the exception of EDOT-CNV-EDOP, are all planar within this range. Further discussion will assume that the differences between these monomers and their resulting polymers (except EDOT-CNV-EDOP and its polymer) are not due to the slight variation in ring twist angles.

3.5 Polymer Optical Properties

Monomers in this family are easily polymerized by galvanostatic, potentiostatic or scanning voltammetric methods to yield electroactive films on Pt or ITO. Polymerization proceeds with the same general mechanism outlined in Chapter 1 with minor changes.

Since the monomers in this study are all unsymmetrical about the polymerization sites, the resulting polymers formed are likely prepared with a mixture of head to tail (H-T), head to head (H-H) or tail to tail (T-T) topologies where the head of the polymer is defined somewhat arbitrarily as the heterocycle nearest the cyano group so as to ensure consistency among the other visual aids in this work as well as the monomer naming convention. Figure

3-8 shows the possible coupling geometries for these monomers. Determining the actual amounts of H-T, H-H or T-T couplings in electropolymerized polymers such as these has not been demonstrated experimentally, however some hypotheses exist. It is anticipated that one electron oxidation of a monomer in this family removes an electron from the

HOMO which is located on the heteroatom (S or N) of the most electron rich heterocycle.

This cation radical then rearranges to place the highest spin density predominantly at the 5 position (the site immediately adjacent to the heteroatom but not connected to the 94

remainder of the molecule) of the heterocycle which then undergoes dimerization as

described for symmetrical monomers such as thiophene, pyrrole, EDOT and EDOP. This

dimerization occurs presumably with another radical cation where the spin is similarly

located primarily on the 5 position of the most electron rich heterocyclic moiety since this

is the radical species that is anticipated to occur in the highest concentration early in the

electropolymerization. Assuming that the radical can be located on the most electron rich

heterocycle with some certainty, this produces a H-H or T-T dimer depending on whether

the electron rich heterocycle is nearest the cyano group or not. CN CN T S -e- S H S S x

O O O O

Regioregular (H-T) Polymer H-T CN CN CN S S S S S S x

O O O O O O

Polymer Containing H-H, T-T and H-T Couplings BiEDOT-(TMS)2 O O CN CN S S S S S S x CN O O O O T- T H-H Figure 3-8. Regiochemistry of cyanovinylene polymerization showing head to tail (H-T), head to head (H-H) and tail to tail (T-T) couplings and planarity of T-T couplings from the crystal structure of polymer subunit BiEDOT-(TMS)2.

From here multiple options exist. First, extended conjugation monomers have lower

oxidation potentials than their conjugation-limited predecessors and would possibly

oxidize next, encounter additional molecules of the initially described radical cation and 95 form the first H-T coupling in the ensuing trimer. The electrochemical deposition of Th-

CNV-EDOP is slightly different from the other monomers in that preparation at low current densities leads to nodular films rather than the smooth deposits observed for the other monomers under identical conditions. Due to the poor slow-growth film forming properties of Th-CNV-EDOP, the monomer with the greatest difference in electron rich character between the two heterocycles, we surmised that there was a potential floor, below which, the T-T dimer (thiophenes on the perimeter of the dimer) would not be oxidized and only this dimer would be formed. Semiempirical calculations on the monomer, H-H, T-T and H-

T dimers were used to test this hypothesis. Using orbital energies from the PM3

ε 20 calculations on all likely dimers ( HOMO), Koopmans’ theorem and a correction factor from vacuum to SCE which includes a solid state polarization parameter,21 the ionization potentials (IP) were estimated for these dimers and the monomer. Koopmans’ theorem ε Koopmans theorem: IPcalc = – HOMO simply states that the ionization potential can be derived from the HOMO energy while a solid state polarization correction is necessary to first convert the potential reference from the vacuum level to SCE, and second to take into account the nonspecific solid interactions that differentiate the actual dimer from the vacuum calculations. The actual values (as compared to the Koopmans’ estimates values) are off for the monomer by nearly 1 V but if the trends are to be believed, the calculations indicate that every possible dimer formed has a lower oxidation potential than the monomer. The most surprising result of these calculations is that for all the dimers of Th-CNV-EDOP, the H-H dimer (with EDOPs on

20. Koopmans Physica, 1933, 1, 104.

21.Brédas,J.L.;Heeger,A.J.Macromolecules 1990, 23, 1150-1156. 96

the periphery) has the highest oxidation potential. This could be the result of facile

oxidation of the electron rich (EDOP)2 unit at the core of the T-T dimer (Figure 3-9) rather

than the terminal heterocycle as proposed in the mechanism outlined above. This may result

in spin localization on the interior of the monomer distant to where it needs to be to

productively from polymer. These results suggest that the poor slow deposition film quality

is due to poorly matched oligomer solubility with solvent conditions or poor coupling of

the T-T dimer from spin localization rather than the existence of a potential floor which

limits oligomer growth to dimer. Since the other monomers are all more closely matched

in terms of flanking heterocycle electron rich character, the idea that all dimers have lower

oxidation potentials than the monomers appears to be a generally valid assumption for this

monomer family. Th-CNV-EDOP 2.2 MONOMER H O O CN N H CN O N S S N S O CN H 2.0 OO H-H DIMER

1.8

T-T DIMER O O 1.6 H-T DIMER CN H S N N S OOO O H CN CN CN OO Ionization Potential (Calculated) /V vs SCE Ionization Potential S S 1.4 N N HH Figure 3-9. Calculated ionization potentials for various dimers of Th-CNV-EDOP.

Figure 3-10 shows data for the polymerization of Th-CNV-EDOT that is

representative of all the monomers in this family. The scanning voltammetric deposition

curve (main part) displays the expected qualities for the deposition of a conducting polymer 97 film with continuously increasing surface area from scan 1 to scan 3 as evidenced by the increase in peak current on subsequent scans. Notable features of this polymerization include the nucleation loop observed during the first scan which is the crossover of the current trace on the cathodic scan immediately following the peak labeled ip,a and is indicative of an increase in the electrode surface area due to electroactive polymer deposition. While this deposition method is commonly used for the preparation of polymer films for electrochemistry on Pt buttons, it is not well suited to the electrosynthesis of films on ITO with reproducible qualities or for the preparation of films on IMEs. The inset of this figure shows the galvanostatic growth of Th-CNV-EDOT on a 0.02 cm2 Pt button (same as above) at 20 µA(1mA/cm2) for 250 sec. Despite using slow-growth conditions as described above to arrive at a 25 mC/cm2 film, the limiting potential at the end of this experiment (Elim) is a good starting point for establishing the ideal potential to grow films for in situ conductivity experiments on IME electrodes of entirely different geometries.

The optical properties of conducting polymers have been treated in a number of research papers and reviews.22 Films for spectroelectrochemistry on ITO were grown from

ACN/TBAP solutions at a potential that is adjusted empirically to deposit a film at ca. 1 mC s-1 cm-2 (25 mC/cm2 total). An initial guess of the correct potential is generated based on Elim from the galvanostatic experiment on Pt. If the film on ITO takes longer than 25 seconds to reach 25 mC/cm2, the film is removed (see below) and the potential is increased by 20 mV until the entire deposition takes near 25 seconds. The trial films are removed with a lint free wipe soaked in acetone, resulting in multiple depositions on the same ITO before

22. (a) Patil, A. O.; Heeger, A. J.; Wudl; F. Chem. Rev. 1988, 88, 183-200. (b) Furukawa, Y. J. Phys. Chem. 1996, 100, 15644-15653. (c) Cornil, J.; Beljonne, D.; Brédas, J. L. J. Chem. Phys. 1995, 103(2) 834-841. 98

3rd Scan CN S S O O 2nd Scan 1.4 i p,a 1st Scan 1.2

1.0 Elim

ial /V vs SCE 0.8

Potent 0.6

0.4 0 50 100 150 200 250 Time /seconds 0.5 mA

nucleation loop

0.5 1.0 Potential /V vs SCE Figure 3-10. Polymer deposition by repeated potential scanning of Th-CNV-EDOT. Inset: galvanostatic growth at 0.05 mA on a 0.02 cm2 Pt button electrode for 250 sec.

obtaining a spectra quality film. It has been established that the removal of first growth

polymer films leaves small nodules which act as nucleation points for subsequent polymer

depositions causing them to occur in a more homogenous fashion.

Figure 3-11 shows the spectroelectrochemical series for monomers that are related

by complete substitution of thiophene by EDOT. PBTh-CNV (top) has a band gap23 of 1.5

eV (as measured in our labs), defined as the onset of the absorbance for the π to π*

transition of the neutral form. Contention about this polymer exists in the literature as its

23. Please see the spectroelectrochemistry section in Chapter 2 for definition and signifi- cance of this and related terms such as switching depth, carrier tail and isosbestic point. 99 band gap was initially reported to be 0.8 eV,24 the event which initially drew our attention to this family. The subsequent resynthesis and characterization for this work confirms that the polymer in the original study (Ho, Brisset, Elandaloussi, et. al.) was not completely reduced prior to estimation of the band gap resulting in a shoulder at lower energy being taken for the band gap. As this polymer was the first material synthesized in this family

(and does not reflect the improvements evident here), it reflects some of the liabilities of this system, namely, PBTh-CNV has the highest oxidation potential for all of the monomers in this family and is the most difficult material to reduce completely. The spectrum for the neutral polymer (Figure 3-11 (top), curve h) represents a polymer that has been electrochemically reduced in the presence of ca. 0.25 mL hydrazine. Despite these rather forcing reducing conditions, PBTh-CNV still shows evidence of charge carriers present in the form of the bipolaron peak E1 at 1.2 eV. In contrast to the behavior of PBTh-

CNV, PBEDOT-CNV demonstrates much cleaner spectral properties (Figure 3-11 bottom). It is easily reduced electrochemically and upon oxidation, a highly absorbing tail forms at low energy indicative of charge carrier formation at the expense of the π to π*peak in the neutral form. A well defined isosbestic point is also present which is taken as evidence that the polymer undergoes a state to state transition between p-type doped and neutral. The band gap for PBEDOT-CNV is ca. 1.1 eV, 0.4 eV lower than PBTh-CNV,

λ resulting in a reduction in the energy of max from 512 nm to 695 nm. PBEDOT-CNV also undergoes a much more significant optical switching depth in the visible and NIR spectral regions compared to BTh-CNV.

24. Ho, H. A.; Brisset, H.; Elandaloussi, E. H.; Frére, P.; Roncali, J. Adv. Mater. 1996, 8, 990. 100

Figure 3-12 shows the spectroelectrochemistry results for the two monomers containing one cyanovinylene, one EDOT and one thiophene representing the effect of switching the position of the EDOT and thiophene units about the cyanovinylene group.

This exchange leads to a decrease in band gap of ca. 0.2 eV from PEDOT-CNV-Th to PTh-

CNV-EDOT. Similar to the polymers PBTh-CNV and PBEDOT-CNV above, the polymer with an EDOT group not directly adjacent to the cyano group shows improved spectral properties in the sense that PTh-CNV-EDOT undergoes greater optical switching depth, and displays a more well defined band in the NIR that is attributed to E1.InFigure3-12top ∆ λ representing PTh-CNV-EDOT, the Aat max is 0.6 absorbance units where in the bottom spectra for PEDOT-CNV-Th, the ∆A is only 0.1 absorbance units. Because of the above properties, and the well defined electrochemistry, facile monomer synthesis and good stability, PTh-CNV-EDOT represents one of the most useful polymers in this family.

Figure 3-13 details the effects of further increasing the electron rich character of the heterocycle flanking the cyanovinylene group. The band gaps of PTh-CNV-EDOP and

PEDOT-CNV-EDOP saturate at 1.1 eV and are lower than the non-pyrrole containing polymers discussed previously with the exception of PBEDOT-CNV. PEDOT-CNV-

EDOP has the cleanest isosbestic point of all the polymers described here and is the most easily oxidized. The spectroelectrochemical series for PTh-CNV-EDOP shows a broad π to π* transition and very little switching depth through the visible region, possibly due to film morphology. Suggestions for the non-standard film forming characteristics of PTh-

CNV-EDOP have been mentioned previously and additional comment is due about the monomer stability of the pyrrole containing members of the family. Th-CNV-EDOP is the monomer with the greatest differential in electron rich character of the heterocycles around 101 cyanovinylene. While solutions of the Th-CNV-EDOP monomer in ACN/TBAP tend to be moderately stable, there is slight discoloration after standing capped for several days but the solution does not become completely dark as is the case for EDOT-CNV-EDOP monomer after standing for only hours. This observation lends further credence to the theory regarding stability of the T-T coupled dimer in Figure 3-9.

A summary of the spectral features of these systems in presented in Table 3-2. From this table, which is in the same order as the peak absorbance values for the monomers, it is

λ evident that the band gap decreases upon increasing electron richness. The max also increases with the exception of PTh-CNV-EDOP and the bipolaron transition E1 remains relatively constant.

Although the band gaps for this family of polymers changes over a range of 0.4 to

0.5 eV, they are quite similar in terms of the qualitative color transitions that they undergo.

Films on ITO change from a moderate, absorbing blue (undoped state) to a transparent light blue (p-type conducting). The transparency of the light blue undoped state appears greater to the eye for PTh-CNV-EDOT and PBEDOT-CNV than for the other polymers. The color changes during the p-type to neutral transition are illustrated quantitatively in Figure 3-14 for PTh-CNV-EDOT. Figure 3-14 part A shows the luminance as a function of potential and indicates that to the eye, the polymer undergoes a fifty percent change in luminance over its doping range. Figure 3-14 part B shows a section of the xy color space representing hue and saturation and the trace that PTh-CNV-EDOT undergoes as it is doped and undoped. The area surrounding the 1.53 and 1.38 V designations is near the white point, which, combined with the high luminance at this potential, causes the film to be highly transmissive. 102

x λ A max CN S 1100 nm 512 nm 1.0 S PBTh-CNV Eg ~ 1.5 eV 0.8

0.6 rbance a Abso 0.4

0.2

h Visible 0.0 λ B 1100 nm max x 695 nm CN S a 1.0 S O O O O

0.8 PBEDOT-CNV Eg ~ 1.1 eV

0.6 rbance Abso 0.4

0.2

m 0.0 0 1 2 3 4 Energy /eV Figure 3-11. Spectroelectrochemical series of PBTh-CNV and PBEDOT-CNV representing the effect of complete substitution of EDOT for thiophene on polymer optical properties. BTh-CNV (top): a = 1.53 V vs SCE to h = -0.47 V vs SCE in 0.5 V increments. PBEDOT-CNV (bottom): a = 0.98 V vs SCE to m = -0.77 V vs SCE in 0.5 V increments. 103

1.6 l A max x 598 nm CN S 1.4 O S O

1.2 962 nm PTh-CNV-EDOT a Eg ~ 1.2 eV

1.0 rbance 0.8 Abso

0.6

0.4

0.2 j Visible 0.0 x B lmax 531 nm CN S S

0.4 O O

1020 nm PEDOT-CNV-Th Eg ~ 1.4 eV 0.3 rbance

Abso 0.2

a 0.1

0.0 k

0 1 2 3 4 Energy /eV Figure 3-12. Spectroelectrochemical series of PTh-CNV-EDOT and PEDOT-CNV-Th (top): a = +1.33 V vs SCE to j = +0.43 V vs SCE in 0.5 V increments and PEDOT-CNV-Th (bottom): a = +1.03 V vs SCE to k = +0.03 V vs SCE in 0.5 V increments. 104

1.4 λ x A ~1100 nm max H 642 nm CN N O 1.2 S O PTh-CNV-EDOP E ~ 1.1 eV 1.0 g a

nce 0.8

Absorba 0.6

0.4

0.2 r Visible 2.2 1400 nm λ H x B max CN N 2.0 756 nm S O 1.8 O O a O 1.6 PEDOT-CNV-EDOP Eg ~ 1.1 eV 1.4

nce 1.2

1.0 Absorba 0.8

0.6

0.4

0.2 s

0.0 0 1 2 3 4 Energy /eV Figure 3-13. Spectroelectrochemical series for PTh-CNV-EDOP and PEDOT-CNV-EDOP (top): a =+XVvs SCE to r =-XVvs SCE in 0.5 V increments and PTh-CNV-EDOP (bottom): a = +1.23 V vs SCE to s = -0.77 V vs SCE in 0.5 V increments. 105

Table 3-2. Polymer optical properties obtained from spectroelectrochemical series.

Optical Band Gapb λ a Name max /nm (neutral form.) (Eg) E1 /nm PBTh-CNV 512 1.5 1100 PEDOT-CNV-Th 531 1.4 1020 PTh-CNV-EDOT 598 1.2 962 PBEDOT-CNV 695 1.1 1100 PTh-CNV-EDOP 642 1.1 ~1100 PEDOT-CNV-EDOP 756 1.1 1396 λ a. max, E1 obtained directly from spectral peaks. b. Eg obtained by extrapolating the absorption onset to zero absorbance.

While the next section (Section 3.6) discusses the electrochemical details of the p- type and n-typing doping processes, only the p-type spectroelectrochemistry is reported in this section. Spectroelectrochemical series were attempted in the region comprising the neutral to n-type doping region. These results were unsuccessful in that no observed change in spectral signature was observed before polymer degradation was evident. This is presumably due to one of two possibilities. First, sealing of the cell used to perform p-type to neutral spectroelectrochemistry is likely inadequate for the more rigorous demands imposed by reduced polymers. Even so, changes in optical characteristics in the dry box were not evident to the eye for these polymers even though the corresponding electrochemical processes are easily documented and are quite stable. A full explanation for this observation await the electrochemical results described next but this suggests the possibility that even though charges are created on the polymer chain with the concomitant 106 ion transport associated with doping, charge carriers are not produced and thus no color change occurs.

3.6 Polymer Electrochemistry

In summary of the previous section, incorporation of increasingly electron rich moieties around the cyanovinylene acceptor results in band gap reduction which eventually saturates at 1.1 eV. What is not clear is the specific mechanism of band gap reduction. Is this due to increasing the HOMO energy, reducing the LUMO energy or some combination of the two (See Figure 3-2)? Electrochemical deposition on Pt button electrodes and IMEs followed by characterization with cyclic voltammetry (CV), differential pulse voltammetry

(DPV) and In Situ Conductivity (σ) are used to explain these phenomena in greater detail.

The voltammetric results for all six polymers are shown in Figure 3-15. These results indicate that there is a broad p-type doping process which occurs at potentials in the range of 0 V to + 0.8 V vs. SCE and is generally described by capacitive charging overlaying the faradaic polymer doping process. This broad oxidative process is coupled to an even broader charge neutralization process. Scanning cathodically of this oxidation/ neutralization, a plateau is reached where very little current passage occurs. Even further scanning results in a new set of peaks attributed to n-type doping of the polymer at potentials between -1.0 V to -1.7 V vs. SCE. Charge neutralization from the reduced form then occurs upon reversal of the scan direction. Voltammetric studies were performed by poising the polymer at a potential in the plateau region corresponding to the neutral form and scanning anodically to access the p-type electrochemistry. Scan direction reversal then neutralizes the polymer and accesses the n-type region. This process was repeated until the correct potential limits were determined to observe the polymer electrochemistry for stable 107

AB x x CN S CN S S

S O O

50 µA 50 µA

-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 PBTh-CNV Potential /V vs SCE PEDOT-CNV-Th Potential /V vs SCE x CDCN S S O O O O

x CN S O

S O 40 µA 50 µA

-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 PTh-CNV-EDOT Potential /V vs SCE PBEDOT-CNV Potential /V vs SCE

x E H F H x CN N CN N O S O S O O O O

40 µA 40 µA

-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 PTh-CNV-EDOP Potential /V vs SCE PEDOT-CNV-EDOP Potential /V vs SCE

Figure 3-14. Polymer cyclic voltammetry on Pt button electrodes A = 0.02 cm2. Polymer deposition was by galvanostatic growth (Figure 3-10) to create a 25 mC/cm2 film. All electrochemistry is in a H2O and O2 free dry box. 108 processes. This has the added utility of breaking in the polymer morphology to support efficient ion transport. p-Type breakin for these polymers ranges from 10 to 20 scans at which point the polymer CV does not undergo any marked changes in line shape of peak current magnitude. n-Type doping occurs over 20-50 scans; this longer time frame likely corresponding to the larger dopant ion used for reduction (assuming oxidation results in anion insertion and reduction leads to cation insertion). Exceptions to this are found in the two polymers that contain a thiophene in the ring B position, namely PBTh-CNV and

PEDOT-CNV-Th. The oxidation peak for these polymers is quite sharp and occurs near 0.5

V vs. SCE.

PTh-CNV-EDOP (Figure 3-15 part E) is the exception to the above procedure where the polymer oxidative and reductive voltammetric processes are separated to illustrate the instability of the cathodic process in PTh-CNV-EDOP. On the first cathodic scan for this polymer, a small peak is observed at -1.4 V attributed to reduction of the polymer. On the subsequent return scan, this peak disappears owing to the instability of the reduced polymer. This type of heightened sensitivity in the reduced form and decreased current responses in the reductive regions of the polymers which contain EDOP (PTh-

CNV-EDOP and PEDOT-CNV-EDOP, E and F respectively) is typical of most pyrrole containing polymers. Despite the relatively weak acidity of the pyrrollic proton, it is likely succeptible to nucleophilic attack from other reduced pyrroles. There is no analogous process in thiophene or EDOT containing polymers.

Stability to redox switching is excellent in these polymers as determined by the magnitude of the n-type peak current after several scans. For PTh-CNV-EDOT, which was studied for extended periodof time, the peak currents decrease ca. 10 percent over then first 109

20 scans and then decrease only ca. 5 percent over the next 200 scans. Additionally, the shape of the reductive processes in PTh-CNV-EDOT and PBEDOT-CNV are the most symmetrical of all the polymers in this study which indicates that there is little preference for ion transport in the doping or dedoping processes.

In addition to the two main redox processes available in this polymer family, there are several smaller peaks that occur inside the main oxidation process and the main reductive process. These peaks are more pronounced for samples of polymer that have been deposited under slow, more controlled conditions and as such, are presumed to be related to order in the polymer secondary structure. Figure 3-16 shows the cyclic voltammogram for PTh-CNV-EDOT deposited under these controlled conditions at a current of 10 µA

(500 µA/cm2). This polymer displays two sets of two peaks, dubbed prepeaks in the literature,25,26 inside of the main voltammetric processes. Some question about the exact nature of these prepeaks appears in the literature.25 What is apparent from this work is that first, prepeaks only appear after a reductive potential excursion. Repeated scanning over the p-type region does not result in the observance of prepeaks. Second, these prepeaks near the main anodic and cathodic voltammetric processes are correlated to each other. That is, the outer two peaks, labeled bn and bp in Figure 3-16 and the inner two peaks, labeled an and ap are coupled and do not exist unless a potential more cathodic or more anodic of the peak is accessed during cycling. Figure 3-16 part C illustrates the coupled nature of the inner set of prepeaks. Part B partially illustrates the coupled nature of the outer set of prepeaks but appeal to the callouts must be made for definitive proof. The callout on the left (red) shows three sequential cyclic voltammograms over the region of bn, an and ap.On

25. Zotti, G.; Schiavon, G.; Zecchin, S. Synth. Met. 1996, 72, 275-281. 110

x CN S O -0.94 V

S O

bp ap 20 µA

an A b n +0.98 V

-1.14 V

increasing increasing scan # B (x2) scan #

C (x2) 5 µA 5 µA

-0.8 -0.4 0.0 0.4 -0.4 0.0 0.4 0.8 Potential /V vs SCE Potential /V vs SCE

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Potential /V vs SCE Figure 3-15. Explanation for prepeaks observed in PTh-CNV-EDOT films. A, B, and C illustrate the effect of narrowing the potential window scanned over while the callouts, red and blue represent the effect of scanning over a moving window of 3 prepeaks.

the first scan bn is visible but is absent from the second two scans indicating that it is

necessary to access bp in order for bn to appear on subsequent scans. The complement to

these results is shown in the blue callout of Figure 3-16. In this series of voltammograms,

the potentials where an, ap,andbp appear are accessed and while bp appears on the first

scan, it does not appear on the second or subsequent scans when bn is not accessed. The

physical intuition for prepeaks is that they are due to trapped charge in the polymer.26

When PTh-CNV-EDOT is oxidized (ca. + 0.98 V), then charge compensated to the neutral

form, there remain sites that are still oxidized where the dopant is either trapped in polymer 111 due to some dense morphology or there is a surface or near electrode interaction that prohibits this section of polymer from being charge neutralized. As the potential is scanned more negative, these sites an and bn are neutralized. The polymer n-type doping neutral to anionic transition is observed on scanning more cathodic of the prepeaks. On the return anionic-to-neutral transition, trapped anionic charge states are not neutralized until ap and bp are scanned. The prepeaks can be quite large compared to the actual polymer doping process as seen in Figure 3-15 D for PBEDOT-CNV. This implies that for this polymer where large prepeaks are observed and other polymers deposited with ordered tendencies, there is a considerable volume of polymer that is electrochemically inaccessible to charge compensation compared to the subset of polymer that is electrochemically active.

The DPV and in situ conductivity results for PBEDOT-CNV are shown in Figure

3-17. Increases in the conductivity occur for both oxidative (p-type) and reductive (n-type) doping. The p-type conductivity onset for PBEDOT-CNV is ca.0Vvs. SCE and occurs slightly later than the extrapolated onset in the CV although the latter must be extrapolated due to the large prepeak obscuring the p-type onset. The p-type conductivity exhibits the typical S-shape expected for conductivity results.27 The conductivity is not volume normalized but is rather reported as the reciprocal of the resistance since the methods of reporting volume conductivities on this type of electrode are of unsatisfactory precision without accurate film thickness measurements. The onset for n-type conductivity occurs

26. (a) Zotti, G.; Schiavon, G.; Zecchin, S. Synth. Met. 1996, 72, 275-281. (b) Crooks, R. M.; Chyan, O. M.; Wrighton, M. S. Chem. Mater. 1989 1,2.(c)Borjas,R.;Buttry,D. A. Chem. Mater. 1991 3, 872.

27. (a) H. S.; Wrighton, M. S. J. Am. Chem. Soc. 1984, 106, 7389. (b) Paul, E. W.; Ricco, A. J.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 1441. (c) Thackeray, J. W.; White, H. S.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 5133. 112 around -1.2 V vs SCE giving this polymer a conductivity derived band gap of 1.2 eV. This compares to the optical band gap of 1.1 eV and the DPV-derived band gap of 0.9 - 1.0 eV.

The optical band gap requires the accurate estimation of only one onset and should for this reason be more accurate. The DPV results for PBEDOT-CNV mirrors the CV results in that the p-type doping processes appear quite broad in the CV and do not peak in the DPV. The causes of this will be discussed below when the question of charge transport mechanisms is addressed. Similarly, the n-type doping electrochemical processes match in terms of peak potential for both the CV and DPV results. This reaffirms the intuitive answer that polymer doping, both reductive and oxidative, is not a reversible process since there should be a peak to peak separation of 0 for both the CV and DPV results for an electrode adhered film.

This is due simply to the fact that the transport of ions and resulting film swelling that occurs upon polymer doping and undoping is slower than the electrochemical time scale used to probe it.

Scan rate dependence experiments for PTh-CNV-EDOT shown in Figure 3-18 exhibit the typical linear dependence on scan rate expected for an electrode adhered film over the scan rates where doping is not diffusion limited. This occurs over 100 to 500 mV sec-1 for PTh-CNV-EDOT and deviates from perfect linearity only due to polymer breakin over the first 30 scans, an observation that is consistent with reports that n-type doping breaking takes more scans than p-type doping breakin at approx. 40-50 scans and 10-20 scans respectively.28 This scan rate linearity occurs over a much broader range of scan rates than the typical homo polymers of pyrrole, thiophene and EDOP that make up these polymers excluding EDOT which demonstrates exceptional linear scan rate dependence.

28. (a) Zotti, G.; Schiavon, G.; Zecchin, S. Synth. Met. 1996, 72, 275-281. 113

7 x 151 Ω CN S O 6 S O O O 5

4

3

2

Pseudo Conductivity /mS Conductivity Pseudo X 1000 1

0 66 MΩ

20

15 -1.3 V

10

5 A µ

0

Current / Current -5

-10

-15 -1.46 V -20 Electrochemical Eg ~ 0.9 eV

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 Potential (V vs SCE) Figure 3-16. PBEDOT-CNV in situ conductivity and DPV results in 0.1 M TBAP/ACN (drybox). n-type conductivity is multiplied by 1000.

Figure 3-19 shows the CV on an IME, the DPV on a Pt button and the In Situ

Conductivity on an IME for PTh-CNV-EDOT. The CV results on the IME are similar to

the CV results in Figure 3-15 obtained on a Pt button which indicates our ability to obtain 114

-300 x j CN S O

-200 S O

-100

A a

0

200

Current /m

100 100 0

Current /uA -100

200 -200

-300 0 100 200 300 400 500 Scan Rate /mV/sec 300 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4

Potential /V vs SCE Figure 3-17. PTh-CNV-EDOT reductive electrochemistry and scan rate dependence. Cyclic voltammetry for the n-type doping of PTh-CNV-EDOT at scan rates a =50mV/sec to j = 500 mV/sec in 50 mV increments. Inset: peak current vs. scan rate for cathodic and anodic peaks.

useful cyclic data on the IME which is useful for direct comparison to the conductivity data.

The positions of the peaks are labeled for comparison to the DPV and conductivity results.

The DPV results are similar to those obtained for BEDOT-CNV where a broad p-doping

wave is observed that does not peak like the comparatively sharper n-type DPV process.

The DPV n-type to neutral transition (Figure 3-19 part B top left) wave peaks at nearly the

exact potential as the corresponding CV process. In contrast, the neutral to n-type wave in

the DPV experiment peaks at -0.98 V compared to the much more cathodic -1.26 V in the

CV. For the in situ conductivity plot (Figure 3-19 part C) a comparatively thick film was 115 deposited in order to observe the n-type conductivity more accurately. This results in a bell shaped conductivity trace that is nearly symmetrical about the peak at -1.03 V vs. SCE.

Vertical lines drawn from the peaks in the CV trace indicate that the peak in the n-type conductivity and indeed, all of the observable conductivity in this range occur between the peaks in the CV spectrum.

From the plot of n-type conductivity it is not clear whether this particular conductivity profile is due to an intrinsic property of the polymer or is due to irreversible degradation of the polymer as the potential is scanned more cathodic of the peak conductivity. The experiment was duplicated after 72 hours of storage in an Ar dry box to determine the extent of change in the conductivity plot. The result is plotted in Figure 3-20 and indicates that very little degradation is evident in the polymer over this time scale and even less is to be attributed to the initial conductivity experiment itself.

Experiments to probe the countercation tolerance were also performed. A film of

PTh-CNV-EDOT was deposited from a solution of TBAP. Assuming anion dominant ion transport for p-type doping, this resulted in a film containing perchlorate. CV data was then collected over 10 scans with TBAP, TEAP, and finally LiClO4. The polymer stability in

TBAP has been discussed. Peaks associated with n-type doping when TEAP or LiClO4 were involved decreased rapidly and were not observed again even upon return to TBAP.

3.7 Conclusions and Perspective

The above results indicate the validity of concentrating on D-A effects as a method of band gap reduction in conjugated polymers. This study establishes that for a fixed acceptor, cyanovinylene, increasing the electron density around the acceptor reduces the resulting polymer band gap. Several more conclusions can be drawn from relationships 116

0.15 A. CV x CN S - 0.88 V O

0.10 S O

0.77 V 1.09 V 0.05

0.00 Current /mA

-0.05 0.96 V

-0.10 - 1.25 V

- 0.86 V 300 B. DPV

200 A µ 100

0

-100

-200

-300 - 0.98 V

0.007 C. In Situ Conductivity

0.006

0.005 - 1.03 V

0.004

0.003

0.002 Conductivity /S Differential Current / 0.001

0.000

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Potential /V vs SCE Figure 3-18. PTh-CNV-EDOT electrochemistry in 0.1 M TBAP/ACN. A. CV at 100 mV/ sec on 10 µm gap and line width masked Pt IME. B. DPV on 0.02 cm2 Pt button. C. In Situ conductance on 10 µm gap and line with masked Pt IME. 117

5.0 A. B. x CN S 4.0 O S O

3.0

2.0

Conductivity /mS 1.0

0.0

-1.50 -1.25 -1.00 -0.75 -0.50 -1.25 -1.00 -0.75 -0.50 Potential /V vs SCE Potential /V vs SCE Figure 3-19. PTh-CNV-EDOT conductance after growth (a)and 72 h (b).

between the electrochemical onsets for doping vs. degree of electron rich character around

cyanovinylene data arranged in order of increasing electron density. Figure 3-21 shows this

type of plot where the primary energy of the band onsets are derived from CV and are

checked against in situ conductivity results when available. Despite the inclusion of

thermodynamic and kinetic data in the CV results possibly obscuring the desired

thermodynamic information, in this particular case the CV correlates with the

spectroscopic results more closely. Higher priority is placed on conductivity results than

CV onsets and absolute determination of band gap is left to spectroelectrochemical results.

When spectroelectrochemical results contradicted CV results, the baricenter of the band

gap was determined by CV and the band gap magnitude was superimposed upon this to get

the valence band (VB) and conduction band (CB) onsets. In all cases, CV slightly

overestimates the band gap when the onsets are not obscured by prepeaks. DPV results

slightly underestimate the band gap and conductivity results come the closest to

corroborating spectroelectrochemical results while providing added information about 118 band edges in energy space. DPV results are quite sensitive to any sort of ion transport process resulting from redox chemistry29 and in general the quantities change over two orders of magnitude. Conductance results often change over 5 to 7 (or greater) orders of magnitude and measure the contribution from redox chemistry to conductivity only. This allows a partial explanation of why the DPV and conductivity results do not match exactly.

While the DPV trace is beginning to rise, there is very little contribution to the conductivity that is visible on a linear scale and thus the conductivity appears to lag the DPV. From the plot in Figure 3-21 it is apparent that the reduction in band gap is entirely the result of the

VB being raised in energy while the CB remains fixed. This follows from the premise of this chapter (Figure 3-2), that the VB energy is set by the identity of the donor while the CB energy is set by the identity of the acceptor.

Given the scope of donor and acceptor moieties in this work, there appears to be a lower limit to the band gap that can be achieved by simply increasing the electron rich character about an acceptor. Ultimately, a band gap of 1.1 eV was achieved whether EDOT was used alone or a combination of thiophene, EDOT and the much more electron rich

EDOP. An argument can be made that the ultimate example of electron rich heterocycle around the cyanovinylene core (EDOT-CNV-EDOP) is not truly the lowest band gap that can be demonstrated by this system due to the considerable deviation from planarity (ca.

25 degrees) in this particular monomer. If one is simply trying to obtain the lowest band gap material, attention to the details that cause EDOT-CNV-EDOP and future monomers basedonEDOPtodivergefromplanarityisnecessary. However, given the synthetic

29. DPV is able to observe electrochemical changes that are not productive for increasing conductivity. 119 complexity of EDOP synthesis and the incorporation of EDOP into other systems due to the difficulty in lithiating EDOP and preparing protected analogues of EDOP, it is worth asking whether much is gained from this effort. Indeed, a fully substituted EDOT system appears to be the best compromise in terms of maximum donor strength and ease of synthesis worth investing into a conducting polymer system. EDOT is far more stable as a monomer than EDOP, is easily lithiated and stannylated for use in Stille coupling reactions, and is easily reducible when incorporated into monomers with an acceptor.

The issue of reduction bears further comment. The main impetus of low band gap

CEP research has been to make available materials that n-type dope in an accessible potential region. Taking into account the reduction potentials for oxygen and water while allowing for common overpotentials places the need to have a polymer n-type dope at a currently inaccessible -0.5 V vs SCEtobestableinair.30 It appears that a reduced polymer containing EDOP would not be stable in devices regardless of this presently out of reach potential target.

Failing to absolve the all-thiophene and all-EDOT containing polymers from liability is the dismal n-type conductivities observed in these systems. This situation arises presumably from the necessity to use TBA+ as a dopant cation rather than the smaller and more n-type compatible TEA+ or Li+. Use of either of these alternative cations leads to rapid polymer degradation under cycling and little utility for devices containing the above polymers and TEA+ or Li+. The use of TBA+ electrolytes creates a scenario where the conductivity rises and falls to nearly zero within the confines of the doping processes

30. de Leeuw, D. M.; Simenon, M. M. J.; Brown, A. R.; Einerhand, R. E. F. Synth. Met. 1997, 53-59. 120 O x O N H 1.1 eV CN O S O O x O N 1.1 eV H CN S O x O S 1.1 eV CN O S O O x O VALENCE BAND VALENCE S CONDUCTION BAND 1.2 eV CN S x S CN 1.4 eV O S O x S CN 1.6 eV S PBTh-CNV PEDOT-CNV-Th PTh-CNV-EDOT PBEDOT-CNV PTh-CNV-EDOP PEDOT-CNV-EDOP Figure 3-20. HOMO-LUMO band edge progression in cyanovinylene monomers 0 -2.0 -1.5 -1.0 -0.5 +0.5 +1.0 121 determined by CV. This strongly implies that the total conductivity, which is the sum of the electrical conductivity and the redox conductivity, is dominated by the redox conductivity,31 that is, the conductivity derived from the gate potential forcing the polymer to be partially doped. In a polymer system suitable for n-type devices, the polymer architecture must be engineered to tolerate smaller, more mobile dopant ions so that the electrical conductivity dwarfs the conductivity derived from mixed-valent states along the polymer backbone. A fundamental requirement of conductivity matching also appears necessary where the n-type and p-type conductivities are within an order of magnitude of each other. In these systems, the n-type conductivity is generally four to five orders of magnitude less than the p-type conductivity.

Aside from the above mentioned liabilities, the series of polymers discussed herein demonstrates some important concepts that should affect the strategies used to discover low band gap and n-type dopable polymers. Specifically, while EDOT has been identified as the ideal donor in terms of ease of synthesis and adequate donor strength, increased acceptor strength over the systems in this study appears necessary to decrease the band gap below 1.1 eV. In these systems, there are always twice the number of donor molecules as there are acceptors. Thus the donor should exert much more of an influence than the acceptor, an observation reinforced by Figure 3-2. For next generation low band gap systems, increased acceptor strength and number will be necessary while maintaining planarity. Additionally, the use of symmetrical monomers drastically decreases system complexity.

31. See Chapter 2 122

3.8 Experimental Section

General methods.NMRspectra(1Hand13C) were obtained on a QE-300 spectrometer. Chemical shifts are reported in ppm relative to residual protio solvent

1 13 1 (CHCl3 7.24 ppm ( H), 77.0 ppm ( C); DMSO 2.49 ppm ( H) (HOD). FAB-HRMS were obtained at the Mass Spectrometry Core Laboratory at the University of Florida. Elemental

Analysis (CHNS) were obtained from Robertson Microlit Laboratories. UV-Vis-NIR spectra were recorded on a Cary 5E spectrophotometer in acetonitrile (ACN). IR spectra were recorded on a Biorad FTS-40A spectrophotometer on NaCl plates. Ethanol (Aaper,

200 proof) was used as received. ACN was distilled from CaH2 prior to use and stored over

3A molecular sieves in a dry box for electrochemical experiments in the reductive regime.

DMF (anhydrous) was used as received. BTh-CNV,32 EDOP33 and EDOP-CHO34 were prepared as previously reported. Synthetic manipulations were performed under an argon atmosphere in flame or oven dried glassware using standard Schlenk techniques.

Electrochemistry was performed in an argon filled dry box with either a PAR 273,

273A or 263A potentiostat/galvanostat controlled with Scribner Associates Corrware software. Platinum working electrodes were used in three electrode cells with Pt flags as counter electrodes and either Ag wire (dry box), SCE or Ag/Ag+ (bench top). Potentials are reported vs SCE and are calibrated vs. internal ferrocene standard. DPV results were

32. Roncali BTh-CNV

33. Merz, A.; Schropp, R.; Dötterl, E. Synthesis 1995, 795.

34. Synthetic procedures for the ethylenedioxypyrrole monomers and premonomers were developed by Kyukwan Zong and are denoted by a KZ after the compound name. 123 obtained using the waveform described in Chapter 2 using a step size of 2 mV, step time of

0.0167 sec and an amplitude of 100 mV.

In situ conductivity data was obtained using the Wrighton method35 on 10 µmgap and line width (1 mm length, 25 pairs of bands) or 5 µm gap and line width (1 mm length,

50 pairs of bands) passivated Pt IME electrodes supplied by the Institute of

Microtechnology, Neuchatel, Switzerland. The passivation layer was silicon nitride over Pt prepared by plasma enhanced chemical vapor deposited (PECVD) on Corning #7740 borosilicate glass. Polymer films were deposited potentiostatically on the above IMEs with a 10 mV drain potential applied between the IMEs to detect when the film bridged the gap.

Deposition was continued until the drain current reached approximately 50 µA to ensure the polymer resistance in the conducting state was more than the lead resistance.

Equilibrium data were collected every 50 mV (gate potential) by a +25 mV to -25 mV

(drain potential) sweep at 1 mV/sec (4 points/sec). The resulting trace was fitted by linear regression for each gate potential and the conductivity at Vgate is reported as the slope of this line. Conductivities are reported directly, uncorrected for polymer film thickness. The experimental apparatus and data treatment are described in detail in Chapter 2.

X-ray crystallography. Intensity data were collected with a Siemens SMART

λ CCD area detector using monochromatized Mo Ka ( = 0.71073Å) radiation. Crystal dimensions and atom coordinates, bond lengths and angles are available in Appendix A.

Structures were solved by direct methods using SHELXS9036,andrefinedonF2 by full-

35. Schiavon, G.; Sitran, S. Zotti, G. Synth. Met. 1989, 32(2), 209. Kittlesen, G. P.; White, H. S.; Wrighton, M. S. J. Am. Chem. Soc. 1984, 106, 7389. Paul, E. W.; Ricco, A. J.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 1441. Thackeray, J. W.; White, H. S.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 5133. 124 matrix least-squares procedures using SHELXL9637. All non-hydrogen atoms were refined with anisotropic displacement coefficients.

2-(3,4-ethylenedioxythienyl)carboxaldehyde (EDOT-CHO).Toathreenecked

RBF fitted with a reflux condenser, pressure equalizing addition funnel and a nitrogen purge was added 10 mL of 1,2-dichloroethane, 1.0 g (7.04 mmol) of EDOT, and 0.52 g

(7.12 mmol) POCl3. The flask was cooled via an ice bath for 20 minutes before POCl3 addition over the course of 15 minutes. The solution was then heated to 90 °C for two hours, cooled, poured into ice, and the aqueous layer was washed with ether (2 X 50 mL). Light yellow crystals separated after 24 hours, yield 1.1 g (92%). mp 135-136 °C. 1HNMR

(CDCl3) 3.93 (m, 2H), 4.63 (m, 2H), 6.55 (s, 1H), 10.21 (s, 1H). Elemental analysis calculated for C7H6SO3: C, 49.41; H, 3.53; S, 18.10. Found: C, 49.16; H, 3.49; S, 18.92.

FAB-HRMS expected 170.0038, found: 170.0014.

2-(3,4-ethylenedioxythienyl)acetonitrile (EDOT-ACN). BuLi (14.1 mL, 35 mmol, 1 eq) was added to EDOT (5g, 1 eq) and 200 mL THF at -78 °C in a single necked flask fitted with a septum. ZnCl2 (4.8 g, 1 eq) was added and the reaction was stirred at room temperature for an hour. This solution was transferred to a pressure equalizing addition funnel and added slowly to a 500 mL 3-necked flask fitted with a reflux condenser and an Ar inlet containing Ni(acac)2 (0.90 g, 10 mol%), BrCH2CN (4.2 g, 1 eq), and cyclohexyldiphenylphosphine (0.94 g, 10 mol%) in 100 mL THF. After addition was complete, the flask was heated to 60 °C for 1 h. The flask was then cooled and cold 1 N

HCl(100mL)wasaddedwith50mLether.The organic phase was washed with water,

36. Sheldrick, G. M., Acta Crystallogr., Sect. A, 1990, 46, 467.

37. Sheldrick, G. M., SHELXL-96, University of Gottingen, 1996. 125 dried with MgSO4 and the solvent was removed in vacuo. The oily residue was chromatographed in the dark on silica gel (20 X 5 cm) with 15% EtOAc in hexane. The product turns darks when standing neat in a freezer in the dark and was not submitted for elemental analysis. Yield, 33%. mp 52 (some clearing) - 72 (flow) °C. FAB/HRMS calculated for C8H7NO2S: 181.02, found: 181.0167.

N-benzyl-1-cyano-2-(2-(3,4-ethylenedioxypyrrolyl))-1-(2-thienyl)vinylene.

(KZ) To a solution of EDOP-CHO (0.60 g, 2.47 mmol) and 2-thiopheneacetonitrile (0.34 g, 2.72 mmol) in tert-butanol was added potassium t-butoxide (0.61 g, 5.43 mmol) at room temperature. The reaction mixture was stirred for 3 h at 50 °C. After cooled to room temperature, t-butanol was removed in vacuo and the residue was diluted with dichloromethane, washed with water and dried over MgSO4. The residue was purified by chromatography on silica gel using the eluent (hexane/ethyl acetate = 3: 1) to give the product as a yellow solid and the product was subjected to debenzylation in next step. 1H

NMR (300 MHz, CDCl3) (7.18 (m, 3H), 7.19 (m, 2H), 7.13 (m, 2H), 6.99 (m, 1H), 6.96 (s,

1H), 6.44 (s, 1H), 4.97 (s, 2H), 4.35 (m, 2H), 4.27 (m, 2H).

General condensation procedure for the preparation of the cyanovinylene monomers BTh-CNV, EDOT-CNV-Th, Th-CNV-EDOT, BEDOT-CNV, Th-CNV-

EDOP and EDOT-CNV-EDOP. 1-cyano-2-(2-(3,4-ethylenedioxythienyl))-1-(2- thienyl)vinylene (Th-CNV-EDOT). To a three necked round bottom flask equipped with a reflux condenser and argon inlet was added 50 mL absolute ethanol, 0.73 g (5.9 mmol) of 2-thiopheneacetonitrile, 1.0 g (5.9 mmol) of 2-(3,4- ethylenedioxythienyl)carboxaldehyde and 1.0 g (8.8 mmol) of potassium t-butoxide. The 126 solution was heated to 70 °C and allowed to stir for three hours, afterwhich, the solution was poured into water.

λ 1-cyano-1,2-bis(2-thienylene)vinylene (BTh-CNV). UV-Vis (ACN) max 364 nm

(30500 M-1 cm-2). FT-IR 3100, 2213 (CN), 1584, 1430, 1326, 1243, 1049, 850, 828, 713,

-1 700, 578, 507 cm . Elemental analysis calc. for C11H7NS2: C, 60.80; H, 3.25; N, 6.45; S,

29.51. Found: C, 60.62; H, 3.11; N, 6.45; S, 29.22.

1-cyano-1-(2-(3,4-ethylenedioxythienyl))-2-(2-thienyl)vinylene (EDOT-CNV-

Th). Prepared via same method as Th-CNV-EDOT. Yield: 62%. mp 117-118 °C. 1HNMR

(CDCl3):? 7.81 (s, 1H), 7.62 (d, 1H), 7.50 (d, 1H), 7.11 (dd, 1H), 6.39 (s, 1H), 4.35 (m, 4H).

13C NMR:? 142.10, 139.58, 138.07, 132.53, 130.97, 129.05, 127.66, 99.68, 64.97, 64.87,

λ -1 -2 64.27. UV-Vis (ACN) max 378 nm (21600 M cm ). FT-IR 3107, 2211 (CN), 1574,

1488, 1436, 1419, 1362, 1324, 1208, 1170, 1019, 920, 897, 856, 709, 617, 594, 463 cm-1.

FAB/HRMS calc'd for C13H9NO2S2: 275.01, found: 275.0073. Elemental anal. calc. for

C13H9NO2S2: C, 56.71; H, 3.29; N, 5.09; S, 23.29. Found: C, 56.67; H, 3.29; N, 5.01; S,

22.84.

Th-CNV-EDOT. The orange solid was collected via filtration and recrystallized

1 from ethanol/H2O (90:10) to give 1.5 g (93%) of Th-CNV-EDOT. (mp 167-168 °C) H

NMR (DMSO-d6) 4.25 (m, 2H), 4.36 (m, 2H), 6.99 (s, 1H), 7.10 (t, 1H), 7.25 (d, 1H), 7.50

13 (s, 1H), 7.76 (d, 1H); C NMR with APT (DMSO-d6) 64.31 (down), 65.34 (down, 97.87

(down), 105.64 (up), 111.89 (down), 117.15 (down), 125.71 (up), 127.81 (up), 128.53 (up),

λ -1 128.36 (down), 141.46 (down), 144.98 (down). UV-Vis (ACN) max 383 nm (28300 M cm-2). FT-IR 3093, 2203 (CN), 1583, 1558, 1519, 1477, 1442, 1329, 1269, 1188, 1064, 127

1010, 959, 910, 844, 825, 744, 697, 625, 600, 526, 478 cm-1. FAB-HRMS calculated for

C13H9NS2O2: 275.0075, found 275.0067.

1-cyano-1,2-bis(2-(3,4-ethylenedioxythienyl))vinylene (BEDOT-CNV). 0.497 g

(1) (2.9 mmol, 1 eq), 0.531 g (2) (1 eq) and 0.358 g KOtBu (1.1 eq) were combined in a 50 mL flask. 25 mL ethanol was added and the solution was refluxed for 2 h. At this time the reaction was cooled, 10 mL water was added and the product was filtered. Analytical samples for of BEDOT-CNV were obtained by soxhlet extraction into ethanol to afford the product in 87% yield. The other monomers are more easily purified by chromatography on

1 a silica gel plug with CH2Cl2. mp 203-204 °C. HNMR(CDCl3): d 7.79 (s, 1H), 6.90 (s, 13 λ - 1H), 6.65 (s, 1H), 4.25 (pair of t, 8H). C NMR: UV-Vis (ACN) max 394 nm (22300 M

1 cm-2). FT-IR 3101, 2877, 2207 (CN), 1580, 1558, 1486, 1452, 1433, 1368, 1336, 1273,

1181, 1070, 1009, 959, 905, 730, 681, 631, 610, 462 cm-1. FAB/HRMS calculated for

C15H11NO4S2: 333.01, found: 333.0157. Elemental anal. calc. for XXX: C, 54.04; H, 3.33;

N, 4.20; S, 19.24. Found: C, 53.47; H, 3.15; N, 4.10; S, 18.54.

1-cyano-2-(2-(3,4-ethylenedioxypyrrolyl))-1-(2-thienyl)vinylene (Th-CNV-

EDOP). (KZ) The solution of the compound (?) (0.9 g, 2.6 mmol) in THF was very slowly added to the solution of sodium (0.15 g. 6.5 mmol) in NH3 (30 mL) at -78 °C. The reaction mixture was stirred for 3 h and 1.0 M NH4Cl aqueous solution (20 mL) was carefully added. The stopper on the reaction vessel was removed and allowed to ambient temperature. After evaporation of NH3, the aqueous phase was extracted with dichloromethane and dried over MgSO4. The residue was purified by chromatography on silica gel using the eluent (hexane/ethyl acetate = 3: 1) to give the product as a yellow

1 crystal (0.45 g, 50%): mp 164-165 °C; H NMR (300 MHz, CDCl3) (8.60 (br, 1H), 7.30 128

(s, 1H), 7.16 (m, 2H), 7.02 (dd, J = 4.9, 3.8 Hz, 1H), 6.55 (d, J = 3.3 Hz, 1H), 4.30 (m, 2H),

λ -1 -2 4.22 (m, 2H); UV-Vis (ACN) max 394 nm (29100 M cm ). FT-IR (CDCl3) 3460, 2989,

-1 + 2930, 2202 (CN), 1575, 1538, 1343 cm ; HRMS (FAB) (M ) calcd for C13H10N2O2S

258.0463, found 258.0457; Anal. Calculated for C13H10N2O2S: C, 60.45; H, 3.90; N,

10.85. Found: C, 60.35; H, 3.89; N, 10.91.

1-cyano-2-(2-(3,4-ethylenedioxypyrrolyl))-1-(2-(3,4- ethylenedioxythienyl))vinylene (EDOT-CNV-EDOP). (KZ) This compound was prepared in analogy to Th-CNV-EDOP except the BOC group was advantageously removed during the condensation. Yellow crystals were obtained after chromatography

1 (1.20 g, 75%); mp 148 °C (decomp); H NMR (300 MHz, CDCl3) (8.60 (br, 1H), 7.34 (s, λ 1H), 6.49 (d, J = 3.8 Hz, 1H), 6.25 (s, 1H), 4.40-4.15 (m, 8H); UV-Vis (ACN) max 401 nm

-1 -2 (27600 M cm ). FT-IR (CDCl3) 3459, 2989, 2934, 2859, 2202 (CN), 1569, 1531, 1459,

-1 + 1344 cm ; HRMS (FAB) (MH )calcdforC15H13N2O4S 317.0596, found 317.0601;

Elemental analysis calculated for C15H12N2O4S: C, 56.95; H, 3.82, N, 8.86. Found: C,

56.44; H, 3.89; N, 8.61. CHAPTER 4 THE ROLE OF INTERGAP REDOX STATES IN CONJUGATED POLYMER REDUCTION PROCESSES: THIENYL AND EDOT CONTAINING FLUORENONES

4.1 Introduction

Fluorenes, and their polymeric counterparts and derivatives have become an integral aspect of organic polymer LED and LEC construction1 due to their high photoluminescence and electroluminescence quantum yields and thermal stability.2

Poly(fluorenes) have been used as electron transport materials in light emitting devices and fluorene derivatives have been useful as the active layer in blue emitting LEDs due their high band gap (ca. 3 eV) which arises from simultaneously exhibiting a high LUMO and a low HOMO. Structurally, fluorenes and their derivatives offer facile functionalization about the 2 and 7 positions which can be used as sites for conjugation extension3 (Figure

4-1 BPyrA-FC6) and fixture of an electroactive (Figure 4-1 BiPyrA-FFc) or inductive group (Figure 4-1 A, B) effectively makes the fluorene a tunable planarized biphenyl.4

Figure 4-1 shows fluorenone derivatives with nitro and cyano substituents at the 9 positions

1. A review of polymers for OLED applications: Kraft, A.; Grimsdale, A. C.; Holmes, A. B. Angew. Chem. Intl. Ed. 1998, 37, 402.

2. (a) Pei, Q.; Yang, Y. J. Am. Chem. Soc. 1996, 118, 7416. (b) Yoshida, M.; Fujii, A.; Ohmori, Y. Yoshino, K. Appl. Phys. Lett. 1996, 69, 734. (c) Grell, M.; Long, X.; Bradeley,D.D.C.;Inbasekara,M.Woo,E.P.Adv. Mater. 1997, 9, 798.

3. (a) Justin Thomas, K. R.; Lin, J. T.; Lin, Y.-Y.; Tsai, C.; Sun, S.-S. Organometallics (ASAP Received January 18, 2001). (b) Justin Thomas, K. R.; Lin, J. T.; Lin, H.-M.; Chang, C. P. Chuen, C.-H. Organometallics 2001, 20, 557-563.

129 130

O O H17C8 C8H17 S S O O S S 6 x

BBTPE-FC8 RR

S S NC CN H17C8 C8H17 O2N NO2 NC CN x O H ABCO NO2 NO2 NO2 NO2

N N N N Fe N H21C10 C10H21 H13C6 C6H13 S

2PA NLO BPyrA-FC6 BiPyrA-FFc Figure 4-1. A structural survey of existing aromatic dicyanomethylidene and related fluorene derivative molecules and polymers.

4 having multiple reversible electrochemical half waves (E1/2) that are accessible in air.

Fluorene derivatives are also able to be difunctionalized at the 9 position with either

solubilizing groups or inductive groups capable of tuning the biphenyl electronic character

but not participating in effective conjugation through the 2,7 positions. The sum of these

functionalization sites makes fluorenes able to be highly electron accepting with suitably

electronegative side groups making them desirable as acceptor moieties in donor-acceptor

molecules, polymers and complexes. Both thiophene and EDOT fluorene monomers have

been chemically polymerized to form soluble materials with either di-ethyl, -hexyl and -

4. (a) Perepichka, I. F.; Popov, A. F.; Orekhova, T. V.; Bryce, M. R.; Andrievskii, A. M.; Batsanov, A. S.; Howard, J. A. K.; Sokolov, N. I. J. Org. Chem. 2000, 3053-3063.(b) Mysyk, D. D.; Perepichka, I. F.; Perepichka, D. F.; Bryce, M. R.; Popov, A. F.; Gold- enberg, L. M.; Moore, A. J. J. Org. Chem. 1999, 64, 19 6937-6950. (c) Perepichka, I. F.; Kuz’mina, L. G.; Perepichka, D. F.; Bryce, M. R.; Goldenberg, L. M.; Popov, A. F.; Howard,J.A.K.J. Org. Chem. 1998, 63, 6484-6493. 131

octyl side chains at the 9 position (Figure 4-1 BBTPE-FC8 and Figure 4-2 EDOT-FC8 and

5 PBEDOT-FC8). The inclusion of fluorene containing materials in the increasingly

important area of two-photon absorbing NLO materials (Figure 4-1 2PA NLO) underscores

this chromophores utility in light harvesting and generating applications.6

O O O O

H C C H H17C8 C8H17 17 8 8 17 S S S x OO EDOT-FC8 PBEDOT-FC8

NC CN NC CN

S SSz SSx y x OO

PCDM PCDM-co-EDOT Figure 4-2. EDOT-Fluorene model and polymer along with PCDM and PCDM-co-EDOT

In light of the ease and variety of functionalization possibilities and our experience

modifying chemically rich multifunctional molecular fragments with terminal EDOT

moieties to aid electropolymerization, it seemed natural to explore the use of 2,7-

fluorenone EDOT and thiophene derivatives as electropolymerizable monomers for our

multicolor electrochromic, n-type dopable and band gap control goals. Continuing with the

theme of electropolymerizable cyano-containing groups started in Chapter 3 and absent the

5. (a) Larmat, F.; Reynolds, J. R.; Reinhardt, B. A.; Brott, L. L.; Clarson, S. J. 1997, 3627- 3636. (b) Donat-Bouillud, A.; Lévesque, I.; Tao, Y.; D’Iorio, Beaupré, S.; Blondin, P.; Ranger, M.; Bouchard, J.; Leclerc, M. Chem. Mater. 2000, 12, 1931-1936. (c) Bel- letête, M. Morin, J.-F.; Beaupré, S.; Ranger, M.; Leclerc, M.; Durocher, G. Macromol- ecules, 2001, 34, 2288-2297.

6. (a) Belfield, K. D.; Hagan, D. J.; Van Stryland, E. W.; Schafer, K. J.; Negres, R. A. Org. Lett. 1999, 1, 10, 1575-1578. (b) Belfield, K. D.; Scafer, K. J.; Mourad, W.; Reinhardt, B. A. J. Org. Chem. 2000, 65, 15, 4475-4481. and references therein. 132 constraint to make soluble polymers to test the viability of these ideas, the natural acceptor for the 9 position of fluorenone was the planar, highly electron withdrawing dicyanomethylene group. This group has been used successfully in both the PCDM family

7 of polymers (Eg ~ 0.8 eV, Figure 4-2) which has been reported to copolymerize with

EDOT to afford co-polymers with a band gap near 0.1 eV (Figure 4-2 PCDM-co-EDOT).8

The CDM core itself is reminiscent of an electron rich fluorenone because the conjugation path has the same resonance relationship to the tuning dicyanomethylene group. Alternate choices include the ketone precursor (Figure 4-3 molecules 7 and 8), or an ethyl or propyl ketal derivative of this which has found success in PCDM applications.9 While the fluorenone derivatives proposed here are not expected to be superior (in the sense of low band gap) to the previously reported CDM containing polymers due to the more extreme frontier orbital positions of poly(fluorenone) materials (HOMO in excess of 1 V, LUMO inexcessof-2V)comparedtoPCDM(HOMOca. 0.3 V, LUMO ca.-0.5V),bandmixing is expected to bring the valence and conduction bands (polymer HOMO and LUMO) within an electrochemically accessible range for device inclusion. Despite the observation that sulfur containing heterocycles tend to quench photo- and electroluminescence yields,10

7. (a) Lambert, T. L.; Ferraris, J. P.; Chem Commun. 1991, 752-754 also, 1268-1270. (b) Neef, C. J.; Brotherston, I. D.; Ferraris, J. P. Chem. Mater. 1999, 11, 1957-1958. (c) Huang, H.; Pickup, P. G. Chem. Mater. 1999, 11, 1541-1545. (d) Salzner, U.; Kiziltepe, T. J. Org. Chem. 1999, 64, 764-769.

8. Huang, H.; Pickup, P. G. Chem. Mater. 1998 10, 2212-2216.

9. Kitamura, C.; Tanaka, S.; Yamashita, Y. J. Chem. Soc., Chem. Commun. 1994, 1585.

10. (a) Belletête, M.; Beaupré, S.; Beuchard, J.; Blondin, P.; Leclerc, M.; Durocher, G. J. Phys. Chem. B 2000, 104, 9118-9125. (b) Biczók, L.; Bérces, T.; Inoue, H. J. Phys. Chem. A 1999 3837-3842. 133

these monomers are expected to play a didactic role in our attempts to optimize new

polymer’s redox states and controllable energy levels.

4.2 Monomer Synthesis and Properties

RR1 RR1 RR1 a, b a, c

S S Si(Me)3 (Me)3Sn S Si(Me)3 1, 2 3, 80 % 5, 84 % 4, 90 % 6, 86 %

1,3,5,7,9: R = R = H (Thiophene) 1 d 2,4,6,8,10: R, R1 = -OCH2CH2O- (EDOT)

R1 RRRNC CN 1 R1 RRR1 O e (Me)3SiS S Si(Me)3 (Me)3SiS S Si(Me)3

9 = BTh-DCF (89 %) 7, 86 % 10 = BEDOT-DCF (92 %) 8, 78 % Figure 4-3. Synthesis of fluorenone monomers. a. n-BuLi, THF/hexanes, -78 0C,2h.b. Me3SiCl,-78ºC1htor.t,2h.c.Me3SnCl,THF,-78ºC,1htor.t.,2h.d.2,7- Dibromofluoren-9-one, DMF or PhMe. Pd(Cl2(PPh3)2 (cat.), 100 ºC, 36 h or 4 h. e. CH2(CN)2, PyH, r.t., 24 h.

The synthesis of BTh-DCF (9) and BEDOT-DCF (10) monomers is depicted in

Figure 4-3. In this synthetic sequence, thiophene and EDOT are lithiated at the 2-position

with n-butyllithium. This anion is trapped with chlorotrimethylsilane to afford the TMS

heterocycles 3 and 4. These intermediates need not be isolated and the 5-position can be

subsequently lithiated in one pot and trapped with chlorotrimethylstannane to afford the

compounds 5 (thiophene) and 6 (EDOT) which are suitable for Stille coupling. Vacuum

distillation affords the products which are colorless crystalline solids. The trimethylsilyl

group was introduced to counteract the efficient solid state packing and thus insolubility of

the target molecules in this work and is known to be both compatible with

electropolymerization and have beneficial effects for film deposition on ITO glass.11 The

11. Sotzing, G. A., Reynolds, J. R. Adv. Mater. 1997, 9, 795-798. 134 key step in this synthesis is the bis(triphenylphosphine)palladium(II)dichloride catalyzed

Stille coupling reaction joining two equivalents of either 5 or 6 with 2,7- dibromofluorenone. The products of these reactions are orange to red and are subsequently reacted with malononitrile under Doebner modified conditions of the Knoevenagel condensation to afford the target compounds 9 (88%) and 10 (82% last step). With the presence of the solubilizing TMS groups, these compounds are soluble in nonpolar solvents such as chloroform, methylene chloride and toluene. The monomer colors are purple (BTh-

DCF, 9)12 and light blue (BEDOT-DCF, 10)13 in dilute solution and have high molar extinction coefficients (ca. 700,000 M-1 cm-1 range) making them deeply colored when concentrated.1Hand13C NMR spectra of compounds 3-10 are consistent with the proposed structures. In these spectra, satellites due to coupling between hydrogen and carbon atoms and 29Si, 117Sn and 119Sn isotopes are visible.14 The values of these coupling constants were especially helpful in the assignment of the signals to the hydrogen and carbon atoms in compounds 5 and 6.

Monomer UV-Vis characterization was performed in methylene chloride and the relevant spectral information is reported in Figure 4-4. The differences in monomer color

12. CIE color coordinates for the monomers in methylene chloride solution. BTh-DCF [132 mM]: Luminance = 377, x = 0.257, y = 0.250.

13. BEDOT-DCF [78 mM]: Luminance = 580, x = 0.257, y = 0.329.

14. (a) Harris, R. K.; Kennedy, J. D.; McFarlane, W. Group IV-Silicon, Germanium, Tin and Lead. In NMR and the Periodic Table; Harris, R. K. and Mann, B. E. Ed.; Aca- demic Press: London, New York, San Francisco, 1978; pp. 310-330, 342-366 and ref- erences within. (b) Harrison, P. G. Investigating Tin Compounds Using Spectroscopy. In Chemistry of Tin; Harrison, P. G. Ed.; Blackie & Son Ltd.: Glasgow, London, 1989; pp. 71-81 and references within. (c) Davies, A. G. Organotin Chemistry; VCH: Wein- heim, New York, Basel, Cambridge, Tokyo, 1997; pp. 18-24 and references within. 135

are due to a roughly 10 nm shift in absorption onset when increasing the electron rich

character of the perimeter heterocycle. Additionally, the spectra for BEDOT-DCF displays

three discernible peaks. Despite the high molar absorptivity of ca. 750,000 mole-1 cm-1 of

the main peaks, a broad band is observed for each monomer near 600 nm

800000 NC CN

(Me)3SiS S Si(Me)3 700000 BTh-DCF λ max = 335 nm, 755,000 600000 onset = 400 nm

500000 OOO NC CN O

(Me)3SiS S Si(Me)3 400000 BEDOT-DCF

Molar Absorptivity λ max = 339 nm, 736,000 300000 367 nm, 721,000 386 nm, 597,000 onset = 410 nm 200000

100000

0

300 400 500 600 700 800 Wavelength /nm Figure 4-4. Monomer UV-Vis spectra in methylene chloride for BEDOT-DCF and BTh- DCF.

X-ray quality crystals of BEDOT-DCF were obtained by vapor diffusion of pentane

into a small vial of monomer dissolved in benzene. After several days large blocks of dark

diffracting BEDOT-DCF were observed. The x-ray crystal structure indicates that the

molecule is close to planar with a molecule of benzene incorporated into the unit cell.

BEDOT-DCF is slightly bowed and the conformation of the EDOT rings relative to the

fluorenone are in the anticipated orientation. Coordinates are available in Appendix B. 136

BTh-DCF does not crystallize under these conditions or similar conditions. Much like

BEDOT-CNV in the previous chapter, all the methods in Chapter 2 were attempted using

similar extensive solvents and solvent combinations.

O O NC CN O O

TMS S S TMS

BEDOT-DCF Figure 4-5. X-ray crystal structure for BEDOT-DCF crystallized by vapor diffusion from benzene/pentane.

4.3 PBEDOT-DCF and PBTh-DCF Synthesis, Electrochemistry and Spectroelectrochemistry

The poor solubility of these monomers in ACN suggested that initial attempts to

electrochemically deposit films of PBTh-DCF and PBEDOT-DCF be attempted in

methylene chloride and binary combinations of methylene chloride and ACN. Attempts at

deposition by cyclic voltammetric, galvanostatic or potentiostatic methods in ACN were

unsuccessful owing to the high potential necessary to oxidize the monomer (>1.5 to 2 V vs.

SCE in all cases) rendering any polymer formed at this potential electrochemically inactive. 137

A survey of available solvents suggested that toluene be combined with ACN to form a mixture that could simultaneously solvate the monomer and electrolyte while not being prone to oxidation under the electrochemical conditions necessary to effect electropolymerization.

This was indeed the case for BTh-DCF and BEDOT-DCF which can be deposited optimally onto a Pt button (A = 0.02 cm2) from a solution of 4:6 toluene:ACN [0.1 M

TBAP] (BTh-DCF) or 6:4 toluene:ACN [0.1 M TBAP] (BEDOT-DCF).15 Given the scarce use of toluene in electrochemical systems and near absence from conjugated polymer systems, no explanation for the utility of toluene and failure of methylene chloride is immediately obvious. Deposition proceeds by cyclic voltammetry, galvanostatic or potentiostatic growth with galvanostatic growth used to deposit all of the films on Pt button electrodes and potentiostatic growth used to form films on both ITO and Pt/IME electrodes.

Galvanostatic electrochemical deposition of PBEDOT-DCF on a Pt button at 10 µA

(0.5 mA cm-2) (Figure 4-6 B) is typical of conjugated polymer depositions at constant current. The potential profile observed is that of a potential peak to 1.1 V vs. SCE within the first 10 seconds followed by a rapid drop to < 1 V and a final settling near 1 V after 250 seconds. This correlates to a 2.5 mC film (0.125 mC cm-2) which typically results in a polymer that is both thick enough to have currents in the several tens of microamps (tens of mA cm-2 on 0.02 cm2 Pt buttons) during doping/undoping and thin enough to enable facile ion transport. The limiting potential (Elim) at the end of this polymerization is used throughout this work as a metric for both comparing deposition parameters and

15. BTh-DCF can also be deposited from the 6:4 toluene:ACN solution but is not as elec- troactive as that derived from the 4:6 ratio. 138 determining starting points for further electrochemical experiments such as IME deposition for in situ conductance. After electrochemical deposition, the polymer film is washed with toluenethenACN(~2mL)andplacedinasolutionofACN0.1MTBAP.Figure4-6A shows the CV obtained from the galvanostatic polymerization of BEDOT-DCF described above. The polymer undergoes the typical broad oxidation waves (approximate E1/2 = 0.65

V vs. SCE, current density multiplied by 5 for comparison to reductive scan rate dependence data) observed for nearly every electroactive polymer. The CV displayed in

Figure 4-6 A is created by first finding the peaks in the CV indicating complete doping, then applying a 20 wave double potential step between the p-type doped and neutral potentials in order to break in the polymer. The p-type doped polymer adopts an open circuit potential (OCP, Voc)ofca. +0.25 to +0.35 V indicating the neutral form is the preferred form. Scanning cathodically of +0.1 V vs. SCE, at least two reductions are observed, accessing the second of which rapidly degrades peak current during CV (a measure of polymer electroactivity) and polymer life. The first reduction (E1/2 = -0.6 V vs.

SCE) can be repeatedly accessed electrochemically without significant effects on polymer switching lifetime. Scan rate dependence (Figure 4-6 C) of this reductive electroactive process indicates that the peak currents scale linearly (10 mV sec-1 to 200 mV sec-1)with the scan rate, a result consistent with an electrode adhered film with diffusion parameters suitable for ion flux at these speeds (i.e. the redox process is not diffusion limited over this regime). Electrochemical results for PBTh-DCF are nearly identical with a slightly anodic p-type doping E1/2 (~0.7 vs. SCE) and identical shape in the reductive region.

DPV results (Figure 4-6 D) are also typical of most conjugated polymers and appear similar in form to the cyanovinylenes discussed in Chapter 3 displaying a plateau near the 139

8 A 200 mV/sec 1.2 6 150 mV/sec 1.0 100 mV/sec 0.8 B 4 0.6 E /V vs SCE 0.4

-2 0.2 2 0 50 100 150 200 250 Time /Seconds 10 mV/sec X5

/mA cm /mA 0

J 8 6

-2 -2 4 2 C /mA cm /mA 0 -1 peak -2 20 mV sec -4 J -4

TBAP 0 20 40 60 80 100 120 mV sec-1

D O O NC CN O O

S S x 5 PBEDOT-DCF

-2

0

/mA cm /mA

Diff

J

-5

LiClO4 TBAP -10 -1.0 -0.5 0.0 0.5 1.0 Potential /V vs SCE Figure 4-6. Electrochemistry (CV) for PBEDOT-DCF (A) on a Pt button (A = 0.02 cm-2) when deposited by galvanostatic deposition as in (B). (C) Scan rate dependence for reduction and (D) DPV for PBEDOT-CNV in TBAP (line) and LiClO4 (dash). OCP which slopes sharply but does not peak on the anodic side of the OCP and slopes

sharply and peaks on the cathodic side of the OCP. DPV traces are shown in Figure 4-6 D

for PBEDOT-DCF in ACN with either 0.1 M TBAP (solid line) or 0.1 M LiClO4 (dashed 140 line). When changing electrolytes, the polymer film was rinsed with the second electrolyte and switched 20 times between the potentials outside of the CV limits to ensure complete transformation to the new dopant ions. The reductive side of the TBAP-switched film has a peak at -0.46 V vs. SCE scanning cathodically and a peak at -0.7 V scanning anodically.

The LiClO4-switched films have peaks at -0.60 V and -0.89 V scanning cathodically and anodically respectively. This yields an estimated E1/2 for the PBEDOT-DCF film switched in TBAP of -0.6 V and -0.75 V for the film switched in LiClO4 which is comparable, but not identical, to the CV-determined E1/2 of -0.6 V in TBAP. DPV is a simple electrochemical technique to estimate band gap because of the definition in the onsets compared to CV (see Chapter 2.7) and these DPV results suggest a band gap of ca. 0.55 eV if the difference between the onsets of the DPV traces are considered and the reduction is assumed to be due to n-type doping.

PBTh-DCF electrochemistry in ACN/TBAP indicates the same relatively anodic

(ca. -0.5 V) reductive cyclic voltammetric process as PBEDOT-DCF. Figure 4-7 shows these CV results and the DPV results. Qualitatively, the ratio between the p-type doping process in the CV compared to the n-type doping process in the CV is reversed compared to the analogous regions of in the DPV. This polymer is generally unstable and even though the oxidative process can be pushed to more anodic potentials, the response quickly degrades.

Spectroelectrochemistry is often the most reliable determinant of band gap because only one peak onset must be estimated and it is not necessary to determine whether a particular peak is due to the band gap or a mid-gap state. Figure 4-7 shows the spectroelectrochemistry for PBEDOT-DCF in ACN/[0.1 M TBAP] degassed with Ar. This 141

6 A. PBTh-DCF CV NC CN

4 S S x

2 -2

0 J mA cm -2

-4

15 B. PBTh-DCF DPV

10

5 -2

0 mA cm diff

J -5

-10

-15

-1.0 -0.5 0.0 0.5 1.0 1.5 Potential /V vs SCE

Figure 4-7. PBTh-DCF electrochemistry in TBAP/ACN. (A) CV of PBTh-DCF as deposited from ACN/TBAP galvanostatically (10 µA, 250 sec). (B) DPV results.

film was prepared by potentiostatic deposition of BEDOT-DCF at 1.17 V vs. SCE until a

charge of 25 mC cm-2 had passed (in this case 50 to 60 seconds). The film was rinsed well

in both toluene and ACN. At +0.37 V vs. SCE or below, the polymer adopts a reproducible

spectrum characteristic of the neutral form where a gently sloping trace to near 2 eV (620 142 nm) followed by a rapid rise to a peak at 2.7 eV (460 nm) and another rise to a peak at 3.7 eV (335 nm) is evident. As the potential is stepped up in 0.5 V increments, the peak at 2.7 eV attributed to the π to π* transition decreases and eventually becomes a trough at the same energy while a peak at 2.2 eV (560 nm) grows in. This can also be visualized as a shifting of what was originally the π to π* to lower energy and absorbance. The peak at 3.7 eV (335 nm) undergoes almost no change in absorbance during potential switching and is

λ attributed to monomer absorbance ( max = 339 nm) remaining in the polymer. This can be due to one of several things. The most likely explanations for this are absorbance due to trapped monomer in the film or secondly, to states or chromophores that do not delocalize in the polymer and disappear. The rinsing in a good solvent for the monomer (toluene) presumably takes care of the trapped monomer in a film this thin (as no absorbances due to monomer have been observed in other films in this work following the same rinsing procedure) but this possibility cannot be entirely ruled out. Duplication of this spectroelectrochemical experiment produced identical results. Attempts to electrosynthesize redox active films of PBTh-DCF on ITO were unsuccessful as an optically inactive brown film similar in color to PBEDOT-DCF was created at long deposition times where it was necessary to raise the potential more anodic of 1.5 V in order to get any polymer to deposit in less than 20 minutes.

Colorimetry on the same PBEDOT-DCF film used for spectroelectrochemistry indicates that the film switches from a brown neutral state 0.17 V (near the Voc) to a purple oxidized state at 1.17 V. Figure 4-8 shows this data for two PBEDOT-DCF films, Film 1 being thicker than Film 2 (25 mC cm-2) since Film 1 was deposited to a charge density of nearly 35 mC cm-2. A line drawn from the color coordinates of the light source to the locus 143

1.4

O O NC CN O O 1.2 S S x

PBEDOT-DCF 1.0

0.8

0.6

Absorbance 0.4

1.27 V 0.2

0.0 0.37 V 0.5 V increments E vs SCE

-0.2 1234 Energy /eV Figure 4-8. PBEDOT-DCF spectroelectrochemistry in ACN/[0.1 M TBAP] degassed in Ar. PBEDOT-DCF was stepped from 0.37 V vs. SCE to 1.27 V vs. SCE in 0.5 V increments.

of the CIE 1931 color scale indicates that the color purity of the neutral polymer is

qualitatively greater than that of the doped form as the neutral form is nearer to the locus

traced in the figure. Spectroelectrochemical or colorimetry measurements are unavailable

for the reduced forms of this polymer due to lack of a sufficient cell to exclude O2 and H2O.

Qualitative observation of a polymer film on ITO viewed through the dry box case indicates 144

520 0.8 530 O O NC CN O O 510 540 S S x 550 PBEDOT-DCF

560 0.6 570 500 0.17 V vs SCE 580 5000 K 590 0.4 Source Film 2 600

CIE y Film 1 610 490 780 1.17 V vs SCE 0.2

480

470 0.0 380

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 CIE x Figure 4-9. Colorimetry of PBEDOT-DCF in ACN/[0.1 M TBAP]. Film 1 was deposited to 35 mC cm-2, Film 2 was deposited to 25 mC cm-2 from 4:6 ACN:toluene [0.1 M TBAP]. Colors are shown for each terminal film state for Film 1.

that there is no color change in this polymer even when poising the polymer at a potential

consistent with the unstable second reduction.

As discussed in Chapter 2, spectroelectrochemical results coupled with in situ

conductance data are the most reliable way to establish the energy levels and band gap of

a polymer electrochemically. Figure 4-10 shows the electrochemical results obtained on a

5 µm Pt IME electrode. PBEDOT-DCF was deposited potentiostatically at the IME 145

8 80 Deposition on IME A 70

A µ 60 B 6 50

rrent / 40 30 Gate Off 4 20 Drain Cu 10 0 EGate = 1.27 V vs SCE -10 0 200 400 600 800 1000

A 2 Time /Seconds

µ 0

Current / -2

O O NC CN O O -4 S S x

-6 PBEDOT-DCF

TBAP 8 TEAP C

y /mS 6 C 1 C2

4

onductivit 40 µS 2

Pseudo C 0 TEAP C3 2 µS

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Potential /V vs SCE Figure 4-10. PBEDOT-DCF electrochemistry on an IME and in situ conductance in ACN. (A) Deposition on Pt-IME (5 mm) at a gate potential of 1.27 V vs. SCE as shown in (B). (C) In situ conductance results for reduction in TBAP (C1), p-type doping in TBAP (C2) and reduction in TEAP (C3). Experimental conditions are described in Chapter 2. working electrode analogous to the gate electrode in Figure 2-10 at a polymerization

potential of 1.27 V vs. SCE in Figure 4-6 B. The polymerization potential was established

from Elim in the Pt button galvanostatic growth data (ca. 1.0 V vs. SCE). This potential was

then applied for 300 seconds, if no drain current was observed between the two IMEs 146 having an offset potential of 10 mV, the polymerization potential was increased by 50 mV and the 300 second trial deposition was repeated. This process was continued until drain

µ current (Edrain = 10 mV) was detected and reached approximately 50 A indicating the film had shorted the electrodes to a sufficient conductance to avoid lead resistance. The gate potentialwasthenallowedtoreachopencircuit while the drain current was measured for another 200 seconds. This drain current typically jumps when the gate potential is removed and then quickly plateaus. It should be noted that this behavior occurs in only this family of monomers of all the monomers studied in this work. Figure 4-10 A shows the CV obtained on the IME at a scan rate of 20 mV sec-1. The CV was started at a potential where the polymer was neutral and scanned cathodically first which is the reason for the pre-peak in the p-type doping wave (See Chapter 3). This CV correlates quite well with the electrochemical results obtained on a Pt button (Figure 4-6) both in the shape and intensity of the electrochemical waves. Figure 4-10 C show several in situ conductance results. The order these results are presented in is described by the arrows at the top of Figure 4-10 C.

The polymer as deposited was broken in over 20 double potential square wave potential steps encompassing both the p-type doped and reduced forms. The polymer was then returned to a potential where the neutral state predominates (-0.2 V) and in situ conductance data was collected first for the reductive process in TBAP, then the p-type doping process in TBAP, followed by the reductive process in TEAP. TEAP was selected as an alternate electrolyte since it contains a smaller cation that maintains the same low electrophilicity as

TBA+. Additionally, CV switching and DPV results in TEAP are comparable to those in

TBAP while the LiClO4 results show a substantial decrease in the current and differential current compared to the TBAP standard. The initial TBAP results for reduction (Figure 4- 147

10 C1) indicate a small relative peak in conductance near -0.6 V and a large conductance peak near -1.15 V. Both of these peaks are symmetrical and quite different in shape than typical oxidative p-type doping waves. The p-type doping conductance profile for

PBEDOT-DCF (Figure 4-10 C2) mirrors the typical sigmoidal shape seen in other conducting polymers throughout this work where there is a sharp onset and rise to peak after which the conductance slowly drops. In order to avoid over-oxidation of the polymer film, the experiment is reversed when this drop in conductance is evident from the decreasing slope of the i-V line. The return path for PBEDOT-DCF nearly exactly mirrors the forward path. Once the conductance returned to that of the neutral form, the polymer was rinsed with ACN/[0.1 M TEAP] and cycled ten times (neutral form to reduced form) in this electrolyte to remove any remaining TBAP. conductance results in TEAP (Figure 4-

10 C3) indicate substantially lower conductance than seen in the TBAP results over the same potential regime. However, similarly shaped symmetrical peaks at identical potentials to those in TBAP are observed, the first of these corresponding almost exactly to the stable polymer reduction observed in Figure 4-6. The reduction in conductance magnitude may be due to polymer breakdown or may be the result of electron-counter ion effects that limit charge carrier mobility. Previous results on cyano-containing polymers (Chapter 3) indicate that the peak magnitude of the reductive event on the forward scan decreases slightly in magnitude on subsequent experiment repetition. This is not likely to be the cause

µ µ of the scale change from 40 S (Figure 4-10 C1)to2 S (Figure 4-10 C3) observed on the change from TBAP to TEAP. Additionally, it should be noted that the method for determining conductance used throughout this work is excellent for establishing relative polymer conductance at equilibrium states but does not contain information on absolute 148 conductance (S cm-1) because the data is reported as conductance in Siemens (S). However, the amount of charge passed during the double potential square wave break-in and a visible estimate of the amount of polymer on the electrode allows an estimation of the relative conductance of the polymer compared to others in this work. Using such a qualitative approach it is apparent this family of polymers has the lowest conductance in the p-type doped state of any polymer in this work. This is especially true of PBTh-DCF (data not shown) for which it is difficult to deposit films with suitably high conductance for this experiment. Despite this, the data for PBTh-DCF indicates a remarkable similarity to

PBEDOT-DCF in conductance onset (ca. 0.6 V vs. SCE compared to 0.5 V for PBEDOT-

DCF).

What is clear from the polymer electrochemistry is that it is very un-polymerlike.

Specifically, the peaks are more similar to those of a quasi-reversible monomer diffusing from solution to an electrode rather than an electroactive polymer undergoing structural changes concomitant to charge carrier formation. An investigation of the monomer electrochemistry was undertaken to understand the differences.

4.4 Monomer Electrochemistry

The discrepancies between the spectroelectrochemically deduced band gap and the electrochemically determined band gap led to an investigation of the monomer electrochemical properties. The edge of the HOMO was accurately established from the onset for conductance in Figure 4-9. The edge of the LUMO should then be Eσ onset -Eg =

0.5 V - 1.8 V = -1.3 V vs. SCE. However, there are two electrochemical couples that occur more anodic of this region with the more cathodic of these two couples leading to polymer degradation. Scanning more cathodic of these two couples leads to an inactive polymer. 149

Background scans on the electrolyte used indicate no redox activity and the redox couple

at -0.6 V can be repeatedly accessed as described in Section 4-2 and has a scan rate

dependence consistent with an electrode adhered redox couple. These data suggest that

there is an electroactive component from the monomer remains in the polymer upon

polymerization. 0.8

0.8 OOO O B NC CN 0.6 0.6 0.4 (Me)3SiS S Si(Me)3 -2 0.2 0.0 BEDOT-DCF 0.4 -0.2

J /mA cm J /mA -0.4 A -0.6 0.2 -0.8 2 4 6 8 10 12 14 -2 (Scan Rate /mV sec-1)1/2 0.0 a1

J /mA cm J /mA

-0.2 a2

-0.4

-0.6

-0.8 -1.5 -1.0 -0.5 0.0 Potential /V vs SCE

Figure 4-11. Monomer electrochemistry (CV) and scan rate dependence for BEDOT-DCF. (A)CVat20mVsec-1startingat-0.2Vsavingsecondoftwoscans.(B.)Scanrate dependence for redox couple a1 over the range 10 mV sec-1 to 200 mV sec-1 viewed as the square root of scan rate. 4:6 ACN:toluene [0.1 M TBAP], 10 mM BEDOT-DCF.

Analysis of the scan rate dependence for monomer reduction can establish whether

the electroactive species is electrode confined. The CV for the monomer BEDOT-DCF

with TMS solubilizing groups does indeed show two nearly reversible redox couples as 150

shown in Figure 4-11 A. The scan rate dependence for the more anodic of the two redox

couples (Figure 4-11 a1) in this figure is shown in Figure 4-11 B. The linear dependence of

the peak currents for these processes vs. the square root of scan rate indicate that the redox

active species is diffusing to the electrode rather than being confined at the electrode

surface. Similar attempts to repeatedly scan over the range encompassing a2 in Figure 4-11

results in a fouled electrode. There is no color observed on the electrode surface at the

conclusion of this experiment but the electrode activity is drastically suppressed for all

∆ redox processes. Figure 4-10 a1 has an E1/2 of -0.59 V and a E of 150 mV while a2 has an ∆ E1/2 of -1.22 V and E =160mV. 1.0

NC CN A (Me)3SiS S Si(Me)3 0.5 BTh-DCF

a1 0.0

a2

-2 -0.5 200 mV/sec 150 mV/sec B 100 mV/sec

J /mA cm J /mA 2.0 C 1.5 -1.0 1.0 -2 0.5 10 mV/sec 0.0 -0.5

J /mA cm J /mA -1.5 -1.0 -1.5 -2.0 2 4 6 8 10 12 14 (Scan Rate /mV sec-1)1/2 -2.0 -2.0 -1.5 -1.0 -0.5 0.0 Potential /V vs SCE Figure 4-12. Monomer electrochemistry (CV) and scan rate dependence for BTh-DCF 151

The electrochemical characteristics of the BTh-DCF monomer are simultaneously superior in quality (∆E, linear scan rate dependence vs. the square root of scan rate) and nearly identical in terms of redox potentials and stability of the second redox couple to

BEDOT-DCF. Figure 4-12 A shows the CV of these two redox couples accessed

∆ sequentially. Redox couple a1 has an E1/2 of -0.54 V and E of 68 mV, exactly that of the theoretical peak-to-peak separation for a reversible system. a2 has an E1/2 of -1.17 V and ∆ E = 98 mV. Additionally, for a2, the oxidation back to the mono-anion shows some flattening that indicates the BTh-DCF-2 anion may be the species involved in the degradation and electrode fouling. Figure 4-12 B shows the scan rate dependence graphically while Figure 4-12 C shows the linearity of the peak current vs. square root of scan rate.

4.5 Conclusions

The monomers and polymers described in this chapter combine several molecular fragments which are excellent in particular applications of polymer optics and electronics because of their redox properties. Namely, the fluorenone core has excellent photo- and electroluminescent properties, EDOT has excellent electrochemical polymerization properties and the dicyanomethylene group is a highly effective electron poor fragment used to tune similar materials for low band gap applications. However, the sum of these parts does not make a family of materials that excels in these areas. Fluorescence measurements on the monomer and qualitative observations on the neutral polymer indicate ineffective light fluorescence for both. The characteristic ease of polymerization typical of EDOT containing molecules is not evident for these systems. 152

-2.0 PBEDOT-DCF O O NC CN O O

S S x -1.5 Hypothetical LUMO/Conduction Band 2nd CN Reduction

-1.0 SCE

vs 1st CN Reduction -0.5 1.8 V (Spectral Band Gap) Potential /V 0

+0.5 HOMO/Valence Band

+1.0 Figure 4-13. Energy level diagram for PBEDOT-DCF

Additionally, it appears that the band gap mixing described in Figure 3-2 is

occurring in these systems but is not sufficient to bring the LUMO low enough in energy

to cause the non-delocalized dicyanomethylene chromophore higher in energy than the two

CN reductions. This, coupled with the high band gap of PBEDOT-DCF, is likely the reason

why only redox conductance is observed in this polymer rather than a conductance profile

characteristic of electronic type conductivity and doping. Figure 4-13 shows schematically

the energy levels of the HOMO and LUMO relative to the two CN reductions observed in

both the monomer and polymer. This indicates that, as the potential is scanned cathodically

of open circuit, an electron is injected into each CN group on every molecule in the film.

Upon further cathodic scanning, the hypothetical electrons that would delocalize and

populate the conduction band would then be injected forcing there to be two to three 153 electrons on every repeat unit. While this is feasible and fairly common for solution electrode reactions, it seems highly unlikely that such a high charge concentration in a film would give rise to any modicum of stability.

The conclusions this leaves with the designer of conjugated polymeric materials is that while it is possible to drastically shift the bands using alternating donors and acceptors, there needs to be some attempt made to match the energy levels. This concept seems to work with the cyanovinylenes in Chapter 3 because at least one band from the mixing polymer can be drawn inside the band gap of the original polymer. In the fluoreneones, the

LUMOs for PEDOT and poly(fluorenones) are nearly the same (more cathodic than -2 V) and the HOMO of the original poly(fluorenone) would be more anodic of PEDOT.

4.6 Experimental Section

General. Melting points were determined using a capillary apparatus and are uncorrected. The yields are given on pure products and were not optimized. The 1HNMR spectra were recorded at 300 MHz and the 13C NMR spectra were recorded at 75 MHz on a Varian VXR-300S and Varian Gemini-300 spectrometers respectively. If not otherwise stated CDCl3 wasusedasasolventandMe4Si (δ=0 ppm) as internal standard. Elemental analyses were performed by Robertson Microlit Laboratories, Inc., Madison, New Jersey.

Analytical thin layer chromatography tests (TLC) were carried out on Whatman silica gel plates (silica gel 60Å, fluorescent indicator UV254, layer thickness 250 µm) and visualized by UV lamp at 254 nm and in an iodine chamber. n-Butyllithium (1.6 M solution in hexanes), thiophene (99+%), trimethylsilyl chloride (99+%), trimethyltin chloride (1.0 M solution in tetrahydrofuran), 2,7-dibromofluoren-9-one (96%), dichlorobis(triphenyl- phosphine)palladium(II) (99.99%), malononitrile (99%) and cellulose microcrystalline 154 powder were from Aldrich and were used as received. 3,4-Ethylenedioxythiophene

(EDOT) from Aldrich was distilled under reduced pressure prior to use. Pyridine from

Fisher Scientific (certified A.C.S.) was used as received. Tetrahydrofuran (THF) used was freshly distilled from potassium-sodium benzophenone ketyl. Toluene used was freshly distilled from sodium.

General procedure for the trimethylsilylation reaction of thiophene and EDOT to prepare 3 and 4. n-Butyllithium solution in hexanes (62.5 mL, 0.100 mole) was added via syringe to a stirred solution of thiophene derivative 1 or 2 (0.100 mole) in tetrahydrofuran (100 mL) at -78ºC under argon. Stirring was continued at this temperature for 2 h when trimethylsilyl chloride (13.3 mL, 0.105 mole) was added via syringe. The reaction mixture was stirred at -78ºC for 1 h, was allowed to reach room temperature and then stirred for 2 h at room temperature. All volatile materials were distilled off (in the case of product 3 under atmospheric pressure and in the case of product 4 under reduced pressure) and water (100 mL) was added to the residue. See below for individual differences in the isolation and purification of the products.

Trimethyl(2-thienyl)silane (3). Prepared from thiophene (1) (8.41 g, 0.100 mole).

The whole lot was obtained after addition of water to the crude mixture was extracted with diethyl ether (3 x 40 mL), the combined extracts were washed with water (40 mL portions) until neutral and dried over anhydrous magnesium sulfate. From the etheral extracts diethyl ether was distilled off under atmospheric pressure and the product was fractionated under reduced pressure giving a colorless liquid. Yield 12.48 g (80%), b.p. 68-69ºC/33 mm Hg

1 (lit. 159-160ºC/748 mm Hg). HNMR[δ, ppm; J, Hz]: 7.58 (1H, dd, J5-4=4.6, J5-3=0.8, 5-

1 H), 7.26 ( H, dd, J3-4=3.3, J3-5=0.8, 3-H), 7.17 (1H,dd, J4-5=4.6, J4-3=3.3, 4-H), 0.32 (9H, 155

2 29 1 13 s, with satellites J( Si- H)=6.7, CH3Si-H). C NMR [reference CDCl3 (δ=77.0 ppm); δ, ppm; J, Hz]: 140.0 (2-C), 133.9, 130.3, 128.1, 0.0 (with satellites 1J(29Si-13C)=53.5,

CH3Si-C).

[3,4-(Ethylenedioxy)-2-thienyl]trimethylsilane (4). Prepared from 3,4-

(ethylenedioxy)thiophene (2) (14.22 g, 0.100 mole). From the suspension obtained after addition of water an insoluble solid was filtered, washed on a filtration funnel with water

(3 x 30 mL), dried at room temperature under reduced pressure in a dessicator over anhydrous calcium chloride and fractionated under reduced pressure giving a colorless solid. Yield 19.35 g (90%), b.p. 97-101ºC/2.2 mm Hg, m.p. 48.5-52ºC. For C9H14O2SSi calculated C 50.43%, H 6.58%, S 14.96%; found C 50.39%, H 6.57%, S 14.90%. 1HNMR

[δ, ppm; J, Hz]: 6.53 (1H, s, 5-H), 4.12-4.20 (4H, m, CH2-H), 0.28 (9H, s, with satellites

2 29 1 13 J( Si- H)=6.9, CH3Si-H). C NMR [reference CDCl3 (δ=77.0 ppm); δ, ppm; J, Hz]:

147.2 (3- or 4-C), 142.6 (3- or 4-C), 111.2 (2-C), 104.7 (5-C), 64.5 (CH2-C), 64.4 (CH2-

1 29 13 C), -0.8 (with satellites J( Si- C)=53.8, CH3Si-C).

General procedure for the trimethylstannylation reaction of the trimethylsilyl derivatives 3 and 4. n-Butyllithium in hexanes (31.3 mL, 0.0500 mole) was added via syringe to a stirred under argon solution of trimethylsilyl derivative 3 or 4 (0.0500 mole) in tetrahydrofuran (100 mL) at -78ºC. Stirring at this temperature was continued for 2 h and after that trimethyltin chloride solution in tetrahydrofuran (52.5 mL, 0.0525 mole) was added via syringe. The reaction mixture was stirred at -78ºC for 1 h, then it was allowed to reach room temperature and stirred for further 2 h. From this reaction mixture, all volatile materials were removed under reduced pressure and water (50 mL) was added to the obtained residue. From this suspension, an insoluble solid was filtered, washed with water 156

(3 x 15 mL), dried at room temperature under reduced pressure in a dessicator over anhydrous calcium chloride and fractionated under reduced pressure giving a colorless solid.

Trimethyl(5-trimethylstannyl-2-thienyl)silane (5). Prepared from trimethyl(2- thienyl)silane (3) (7.62 g, 0.0500 m). Yield 13.41 g (84%), b.p. 87-88ºC/0.6 mm Hg, m.p.

52-54ºC. For C10H20SSiSn calculated C 37.64%, H 6.32%, S 10.05%; found C 37.61%, H

1 6.25%, S 9.83%. HNMR[δ, ppm; J, Hz]: 7.40 (1H, d, J3-4=3.1, with satellites

4 117,119 1 J( Sn- H)=4.2 appeared as doublets coupled by J=3.1, 3-H), 7.29 (1H, d, J4-3=3.0, with satellites 3J(117Sn-1H)=23.0 and 3J(119Sn-1H)=23.9 appeared as doublets coupled by

2 117 1 2 119 1 J=3.1, 4-H), 0.37 (9H, s, with satellites J( Sn- H)=55.0 and J( Sn- H)=57.6, CH3Sn-

2 29 1 13 H), 0.32 (9H, s, with satellites J( Si- H)=6.8, CH3Si-H). C NMR [reference CDCl3

(δ=77.0 ppm); δ, ppm; J, Hz]: 146.0 (with satellites 1J(29Si-13C)=65.3, 2-C), 143.0 (with satellites 1J(117Sn-13C)=363.0 and 1J(119Sn-13C)=380.2, 5-C), 136.1 (with satellites

2J(117,119Sn-13C)=35.5, 4-C), 134.9 (with satellites 3J(117,119Sn-13C)=48.1, 3-C), 0.1 (with

1 29 13 1 117 13 satellites J( Si- C)=53.8, CH3Si-C), -8.2 (with satellites J( Sn- C)=356.2 and

1 119 13 J( Sn- C)=372.2, CH3Sn-C).

[3,4-(Ethylenedioxy)-5-trimethylstannyl-2-thienyl]trimethylsilane (6).

Prepared from [3,4-(ethylenedioxy)-2-thienyl]trimethylsilane (4) (10.72 g, 0.0500 mole).

Yield 16.13 g (86%), b.p. 121-124ºC/1.2 mm Hg, m.p. 66-72ºC. For C12H22O2SSiSn calculated C 38.22%, H 5.88%, S 8.50%; found C 38.24%, H 5.68%, S 8.23%. 1HNMR

2 117 1 [δ, ppm; J, Hz]: 4.11-4.19 (4H, m, CH2-H), 0.34 (9H, s, with satellites J( Sn- H)=55.8

2 119 1 2 29 1 and J( Sn- H)=58.5, CH3Sn-H), 0.28 (9H, s, with satellites J( Si- H)=6.9, CH3Si-H).

13 C NMR [reference CDCl3 (δ=77.0 ppm); δ, ppm; J, Hz]: 148.7 (with satellites 157

2J(117,119Sn-13C)=14.9, 4-C), 147.9 (with satellites 3J(117,119Sn-13C)=35.5, 3-C), 118.0 (2-

1 117 13 1 119 13 C), 114.8 (with satellites J( Sn- C)=349.3 and J( Sn- C)=366.9, 5-C), 64.5 (CH2-

1 29 13 1 117 C) -0.7 (with satellites J( Si- C)=53.8, CH3Si-C), -8.7 (with satellites, J( Sn-

13 1 119 13 C)=359.6 and J( Sn- C)=375.6, CH3Sn-C).

General procedure for the preparation of

Bis(trimethylsilylthienyl)fluorenones 7 and 8. A solution of the thiophene under argon,

5 or EDOT, 6 (0.01050 mole) in toluene (10 mL) was added via syringe to a stirred suspension of 2,7-dibromofluoren-9-one (1.69 g, 0.00500 mole) and dichlorobis(triphenyl- phosphine)palladium(II) (0.175 g, 0.00025 mole) in toluene (40 mL). The resulting suspension was stirred at room temperature for 10 min. and then heated at 100ºC. Progress of the reaction was monitored by thin layer chromatography (chloroform–hexane 3:1 v/v).

After the reaction was completed the reaction mixture was cooled to room temperature, filtered through cellulose powder to remove catalyst and from the filtrate toluene was removed under reduced pressure. To the obtained residue a mixture toluene–hexane 1:4 v/ v (20 mL) was added and an insoluble solid was filtered and recrystallized yielding the product as a red solid. See below for individual differences in time of reaction and recrystallization solvent.

2,7-Bis(5-trimethylsilyl-2-thienyl)fluoren-9-one (7). Prepared from trimethyl(5- trimethylstannyl-2-thienyl)-silane (5) (3.35 g, 0.01050 mole)) by heating for 5 h. Yield 2.10 g (86%) after recrystallization from toluene, m.p. 223-225ºC. For C27H28OS2Si2 calculated

C 66.34%, H 5.77%, S 13.12%; found C 66.22%, H 5.87%, S 12.99%. 1HNMR[δ, ppm;

J, Hz]: 7.81(2H, d, J1(8)-3(6)=1.4, 1- and 8-H), 7.61 (2H, dd, J3(6)-4(5)=7.8, J3(6)-1(8)=1.7, 3- and 6-H), 7.31-7.38 (4H, m, 4-, 5- and thienyl 3- or thienyl 4-H), 7.18 (2H, d, J3-4 or J4- 158

2 29 1 3=3.5, thienyl 3- or thienyl 4-H), 0.35 (18H, s, with satellites J( Si- H)=6.8, CH3Si-H).

13 C NMR [reference CDCl3 (δ=77.0 ppm); δ, ppm; J, Hz]: 193.1 (C=O), 148.1, 142.6,

140.8, 135.2, 135.1, 134.9, 131.6, 124.9, 121.4, 120.6, -0.2 (with satellites 1J(29Si-

13 C)=53.8, CH3Si-C).

2,7-Bis[5-trimethylsilyl-3,4-(ethylenedioxy)-2-thienyl]fluoren-9-one (8).

Prepared from [3,4-(ethylene-dioxy)-5-trimethylstannyl-2-thienyl]trimethylsilane (6)

(3.96 g, 0.01050 mole) by heating for 4 h. Yield 2.36 g (78%) after recrystallization from a mixture toluene-hexane 1:1 v/v, m.p. 207.5-209ºC. For C31H32O5S2Si2 calculated C

61.56%, H 5.33%, S 10.60%; found C 61.36%, H 5.20%, S 10.61%. 1HNMR[δ, ppm; J,

Hz]: 8.05 (2H, d, J1(8)-3(6)=1.5, 1- and 8-H), 7.78 (2H, dd, J3(6)-4(5)=8.1, J3(6)-1(8)=1.7, 3- and 6-H), 7.42 (2H, d, J4(5)-3(6)=8.1, 4- and 5-H), 4.27-4.35 (4H, m, CH2-H), 4.19-4.27

2 29 1 13 (4H, m, CH2-H), 0.32 (18H, s, with satellites J( Si- H)=6.8, CH3Si-H). CNMR

[reference CDCl3 (δ=77.0 ppm); δ, ppm; J, Hz]: 193.8 (C=O), 147.6, 141.9, 139.5, 134.7,

134.1, 131.3, 121.5, 121.0, 120.2, 109.7 (thienyl 5-C), 64.5 (CH2-C), 64.2 (CH2-C), -0.7

1 29 13 (with satellites J( Si- C)=53.9, CH3Si-C).

General procedure for the preparation of

Dicyanomethylenebis(trimethylsilylthienyl)fluorenes 9 and 10.Asolutionof malononitrile (0.13 g, 0.002 mole) in pyridine (2 mL) was added dropwise to a stirred solution of the ketone 7 or 8 (0.001 mole) in pyridine (30 mL) at room temperature. Stirring at this temperature was continued for 24 h and after pouring of the reaction mixture into water (100 mL) a solid precipitated. It was filtered, washed on a filtration funnel with water

(3 × 10 mL), dried at room temperature under reduced pressure in a dessicator over 159 anhydrous calcium chloride and recrystallized yielding the product as a purple solid. See below for individual differences in recrystallization solvent.

9-Dicyanomethylene-2,7-bis(5-trimethylsilyl-2-thienyl)fluorene (9). Prepared from 2,7-bis(5-trimethylsilyl-2-thienyl)fluoren-9-one (7) (0.49 g, 0.001 mole). Yield 0.48 g (89%) after recrystallization from ethyl acetate, m.p. 267-269ºC. For C30H28N2S2Si2 calculated C 67.12%, H 5.26%, N 5.22%, S 11.94%; found C 66.90%, H 5.33%, N 5.20%,

1 S 11.90%. HNMR[δ, ppm; J, Hz]: 8.42 (2H, s, 1- and 8-H), 7.58 (2H, dd, J3(6)-4(5)=8.0,

J3(6)-1(8)=1.4, 3- and 6-H), 7.27-7.41 (4H, m, 4-, 5- and thienyl 3- or thienyl 4-H), 7.17 (2H,

13 d, J3-4 or J4-3=3.5, thienyl 3- or thienyl 4-H), 0.35 (18H, s, CH3Si-H). C NMR [reference

CDCl3 (δ=77.0 ppm); δ, ppm; J, Hz]: 160.7 (9-C), 147.4, 141.6, 140.4, 135.4, 135.2, 134.9,

131.6, 125.3, 123.7, 120.9, 113.1 (CN-C), 76.3 (dicyanomethylene-C), -0.1 (with satellites

1 29 13 J( Si- C)=53.8, CH3Si-C).

9-Dicyanomethylene-2,7-bis[5-trimethylsilyl-3,4-(ethylenedioxy)-2- thienyl]fluorene (10). Prepared from 2,7-bis[5-trimethylsilyl-3,4-(ethylenedioxy)-2- thienyl]fluoren-9-one (8) (0.60 g, 0.001 m). Yield 0.53 g (82%) after recrystallization from a mixture toluene-hexane 1:1 v/v, m.p. 260.5-262.5ºC. For C34H32N2O4S2Si2 calculated C

62.54%, H 4.94%, N 4.29%, S 9.82%; found C 62.53%, H 4.97%, N 4.29%, S 9.87%. 1H

NMR [δ, ppm; J, Hz]: 8.83 (2H, d, J(1(8)-3(6)=1.2, 1- and 8-H), 7.56 (2H, dd, J3(6)-4(5)=7.9,

J3(6)-1(8)=1.6, 3- and 6-H), 7.25 (2H, d, J4(5)-3(6)=8.1, 4- and 5-H), 4.27-4.35 (4H, m, CH2-

13 H), 4.18-4.27 (4H, m, CH2-H), 0.33 (18H, s, CH3Si-H). C NMR [reference CDCl3

(δ=77.0 ppm); δ, ppm; J, Hz]: 160.9 (9-C), 147.8, 140.0, 139.3, 134.5, 133.9, 130.2, 123.5,

120.2, 120.1, 113.2 (CN-C), 110.0 (thienyl 5-C), 75.6 (dicyanomethylene-C), 64.4 (CH2-

1 29 13 C), 64.2 (CH2-C), -0.7 (with satellites J( Si- C)=53.9, CH3Si-C). CHAPTER 5 POLY(ALKYLENEDIOXYPYRROLES): AQUEOUS COMPATIBLE CONDUCTING POLYMERS WITH LOW FORMAL REDOX POTENTIALS

5.1 Introduction

Chapters 3 and 4 detail the importance of VB and CB control for impacting band gap and n-type doping properties. In that work, substantial benefit was realized by appending a 3,4-ethylenedioxy group to a simple heterocycle such as thiophene to form

EDOT. This change shifts the reduction potential to more cathodic potentials making these polymers easier to oxidize and thus more difficult to reduce. To this point there have been only cursory examinations of the effect the 3,4-ethylenedioxy addition has on poly(pyrrole’s) electronic properties. Poly(pyrroles) have been extensively studied1 and are interesting for a variety of applications. The most notable of these are applications that involve PPy’s high conductivity2 and ability to be electrodeposited and switched extensively in aqueous environments. The aqueous compatibility opens up an array of interesting biological applications involved in human health ranging from detection of

DNA mismatches3 and human lymphocyte antigen groups, to the tissue engineering of nerve cell guidance channels. Korri-Youssoufi and Garnier appended an oligonucleotide

1. Street,G.B.Handbook of Conducting Polymers, Skotheim, T. A., Ed.; Dekker: New York, 1966; Vol 1. pp. 265-290.

2. Diaz, A. F.; Bargon, J. Handbook of Conducting Polymers, 1986, Skotheim, T. A., Ed.; Dekker: New York, Vol. 1; pp. 81-115.

3. Schafer, A. J.; Hawkins, J. R. Nature Biotech. 1998, 16, 33-39.

160 161

(5’CCT AAG AGG GAG TG3’) to a spacer attached to the three position of PPy (Figure 5-

1 A) and found that it was electroactive in aqueous solution containing a variety of biological buffers including non-complimentary salmon DNA.4 Incubation with varying concentrations of the complementary strand (5’CAC TCC CTC TTA GG3’) and a noncomplimentary strand (5’GGT GAT AGA AGT ATC3’) of DNA indicates the oligonucleotide hybridization of the complimentary strand causes an easily observable change in the CV. This effect is based on the conformational modifications required upon doping while the noncomplimentary strand, which is incapable of hybridization, results in no change in the polymer electrochemistry. Saint-Aman and coworkers synthesized poly(pyrroles) functionalized at the N position with glucose (Figure 5-1 B) with the goal of preparing materials with enantioselective recognition properties that are capable of performing chiral electrosynthesis.5 While only transiently stable in aqueous perchlorate systems, this polymer is enantioselective to (1S)-(+)-10- and (1R)-(-)-10-camphorsulfonic acid and polymerization is inhibited in the presence of the (S) enantiomer. Also from

Garnier’s laboratories, PPy functionalized in the 3 position with a chiral dipeptide (Figure

5-1C,glycyl-D-phenylalanine, Gly-D-Phe) that was selected for its specificity and binding capacity toward carboxypetidase A and trypsin was used for enantioselective switching.

The presence of either of these enzymes during electrochemical switching (CV) leads to an

6 increase in polymer oxidation potential, Ep,m. Tripathy and colleagues found that the third major class of biomolecule, DNA, can be incorporated into PPy films and have performed

4. Korri-Youssoufi, H.; Garnier, F.; Srivastava, P.; Godillot, P.; Yassar, A. J. Am. Chem. Soc. 1997, 119, 7388-7389.

5. Moutet, J.-C.; Saint-Aman, E.; Tran-Van, F.; Angibeaud, P.; Utille, J.-P. Adv. Mater. 1992, 4, No. 7/8, 511-513. 162

imaging experiments on these films.7 From these works, it is clear that even rather delicate

biomolecules including enzymes8 (horseradishperoxidase, glucose oxidase), histone

proteins9 and photodynamic proteins (bacteriorhodopsin) are stable to some degree in a

PPy matrix. A. Specific DNA Recognition Based on Oligonucleotide-PPy

O N O O O OH HN CCT-AAG-AGG-GAG-TG O O O O

1) Electropolymerization 0.6 + 0.4 2) H2N-CCT-AAG-AGG-GAG-TG N N N 0.6N 0.4 H H H H

B. Glucose-Pyrrole Chiral Electrodes C. Enzyme Recognition with for Enantioselective Recognition PPy-Bioactive Peptides

OH O * OCOCH H3COCO 3 OCOCH3 HN O O OCOCH 3 NH O O O

N N x x H Figure 5-1. Examples of biologically relevant pyrrole modifications. (A) Garnier’s specific recognition of DNA by CV detection in the presence of various concentrations of complimentary and noncomplimentary hybrids. (B) N-substituted glucose pyrrole is enantioselective for camphorsulfonic acid. (C) dipeptide modified Py is capable of binding and recognizing carboxypeptidase A.

6. Garnier, F.; Korri-Youssoufi, H.; Srivastava, P.; Yassar, A. J. Am. Chem. Soc. 1994, 116, 8813-8814.

7. Pande, R.; Ruben, G. C.; Lim, J. O.; Tripathy, S.; Marx, K. A. Biomaterials, 1998, 19, 1657-1667.

8. Selampinar, F.; Akbulut, U.; Ozden, M. Y.; Toppare, L. Biomaterials, 1997, 18(17), 1163-1168

9. Prezyna, L. A.; Qiu, Y.-J.; Reynolds, J. R.; Wnek, G. E. Macromolecules, 1991, 24, 5283-5287. 163

Langer has found that PPy can noninvasively control the shape and growth of mammalian cells10 and advanced the applications of PPy to tissue engineering in his seminal paper on the stimulation of neurite outgrowth using PPy in an applied electric field.11 These studies suggest the ability to apply a field to a suitable cell substrate, while simultaneously being able to control hydrophilicity, will revolutionize tissue engineering.

Subsequent studies have extended this work to endothelial cell attachment and growth using a PPy-heparin composite.12

The change from the general use of PTh to PEDOT when a stable CP is desired has been gradually occurring since about 1996. As of 2001, researchers active in developing new conjugated materials recognize that EDOT is a superior monomer in almost every way

(except cost) compared to Th. It easily electropolymerizes but is relatively stable as a monomer, and the polymers are remarkably stable and structurally homogenous due to the blocked 3 and 4 positions which preclude β couplings. Switching is more facile in PEDOT than PTh in terms of both low polymer oxidation potential and switching speed. Finally,

PEDOT is stable to air and water in its oxidized, conducting state where PTh is more stable in the neutral, insulating state (Figure 1-7). PPy based materials, however, even when prepared using optimized conditions, are poorly defined in that there is a significant amount of α-β coupling.13 The presence of these defect sites along the polymer backbone decreases

10. Wong, J. Y.; Langer, R.; Ingber, D. E. Proc. Natl. Acad. Sci. USA 1994, 91, 3201- 3204.

11. Schmidt, C. E.; Shastri, V. R.; Vacanti, J. P.; Langer, R. Proc. Natl. Acad. Sci. USA 1997, 94, 8948-8953.

12. Garner, B.; Hodgson, A. J.; Wallace, G. G.; Underwood, P. A. J. Mater. Sci.: Mater. Med. 1999, 10, 19-27. 164 its effective conjugation length, induces structural disorder, limits the electrochemical response, and is implicated as the primary site of polymer breakdown due to over- oxidation, attenuating the electrochemical switching life-time.14 Additionally, oxidized

PPyisunstabletoreductionbyevenrelatively weak reducing agents. Regardless of the polymer eventually used, it is clear that future advances in CP based biosensors, controlled release devices15 and eventually implantable devices16 require a material that is redox active, stable as a conductor, compatible in aqueous environments and biocompatible.

Figure 5-2 illustrates this as a puzzle that is simultaneously multi-tiered where the redox active, aqueous compatible polymer must be well understood for the advanced applications to be possible.

The dioxy substitution pattern adds electron density to the aromatic heterocycle, reduces both the monomer and polymer oxidation potential (thus reducing propensity towards over-oxidation), and tends to lower the electronic band gap of the π system.

Specifically, the reduction potential of PTh (E1/2 > +0.5 V vs. SCE) is reduced to approximately 0.0 V in poly(3,4-ethylenedioxythiophene) (PEDOT), and the band gap is concurrently lowered from 2.0 to 1.6 eV. We hypothesized that if this principle of a bridged

3,4-dioxy substituent was applied to form a poly(3,4-alkylenedioxypyrrole), the expected decrease in the polymer's E1/2 would yield an increased stability in the doped (conducting)

13. Pfluger, P.; Street, G. B. J. Chem. Phys. 1984, 80, 544.

14. Zotti, G.; Schiavon, G.; Zecchin, S. Synth. Met. 1995, 72, 275.

15. Pyo, M.; Maeder, G.; Kennedy, R. T.; Reynolds, J. R. J. Electroanal. Chem. 1994, 368, 329.

16. Peppas, N. A.; Langer, R. Science 1994, 263, 1715-1720. 165

TISSUE ENGINEERING & IMPLANTABLE DEVICES

DEVICES: BIOSENSORS &CONTROLLED RELEASE

STABLE CONDUCTORS AQUEOUS REDOX ACTIVITY COMPATIBILITY

BIOCOMPATIBILITY

Figure 5-2. Requirements for tissue engineering and implantable devices. Biologically relevant devices for biosensors and controlled release must be based on a foundation of redox active, stable conductors with aqueous and biological compatibility.

form. Relative to pyrrole, the monomer's oxidation potential would decrease, and the

blocked 3- and 4-positions would preclude α-β coupling leading to a polymer repeat unit

with fewer defects. This chapter addresses the synthesis, electrochemical characterization

as a measure redox activity, and stability in aqueous media of the poly(3-4-

alkylenedioxypyrroles) PXDOPs.

5.2 Monomer Synthesis and Polymer Electrosynthesis

The synthetic conditions necessary to prepare the simplest monomer (Figure 5-3),

17 3,4-ethylenedioxypyrrole (EDOP, n = 0, thus R1,R2 n.a.) have been reported with the

17. Merz, A.; Schropp, R.; Dötterl, E. Synthesis 1995, 795. 166

only mention of polymer properties appearing in the context of a photographic imaging

agent.18 Treatment of dimethyl-N-benzyl-3,4-dihydroxypyrrole-2,5-dicarboxylate with a

variety of dibromo- or dimesylalkanes in the presence of base gave the corresponding

dicarboxyl-3,4-alkylenedioxypyrrole intermediates. After subsequent deprotection of the

benzyl group in the presence of H2/Pd(C), hydrolysis in 1 N NaOH and decarboxylation in

triethanolamine at 180 °C, a series of parent 3,4-alkylenedioxypyrrole monomers was

obtained.19 N-substitution was easily effected by treatment of the parent pyrrole with NaH

in THF followed by alkylation. In addition to EDOP, ProDOP (n = 1, R1,R2 =H)isthe

other standard parent monomer in this study and will be compared to EDOP as a potential

replacement for Py as the monomer of choice for future bio-oriented experiments. EtO OEt OH OH K2CO3, CH3OH O O NaOCH3 H CO C CO CH CH3OH, reflux 3 2 N 2 3 NH2 MeO2C N CO2Me 8h, 56 % Bz OMe Bz Br 2 R R O 1 2 n X X

K2CO3 dry DMF 110 ºC 10 h

R R R R R1 R2 1 2 1 2 (1) H2, Pd/C AcOH, 70 ºC n n N(CH CH OH) n 48 h, 95 % O O 2 2 3 O O O O 180 ºC, 5-10 min. (2) 2 N NaOH (aq.) 60-65 % RT, 12 h 95 % H CO C CO CH N HO2C N CO2H 3 2 N 2 3 H H Bz

Figure 5-3. Synthesis of PXDOP monomers.

Single crystals of some of the XDOPs were obtained and the crystal structures for

EDOP and ProDOP are depicted in Figure 5-3. Densities in the single crystal are 1.384 g

18. Savage, D. J.; Schell, B. A.; Brady, B. K. U.S. Patent 5 665 498, 1997

19. Thomas, C. A.; Zong, K.; Schottland, P.; Reynolds, J. R. Adv. Mater. 2000 12(3), 222- 225. 167 cm-3 and 1.405 g cm-3 respectively. The pyrrole ring fragment of EDOP is more planar than that of ProDOP with the C2-N-C5-C4 torsional angle (numbered as described in Figure 1-

13) being 0.4° for EDOP and 1.1° for ProDOP. It has been postulated that the slight strain in the propylenedioxy ring vs. the ethylenedioxy ring is responsible for the slightly less electron rich character observed in ProDOP (126.6°) compared to EDOP (123.7°) (the ideal being 120°).20 Whatever the reason, neither monomer is stable in the crystalline state at room temperature, the solids turning dark brown and eventually black over the course of days to weeks. ProDOP is considerably more stable than EDOP but both monomers require storage in the freezer and sublimation (0.01 mm Hg) before use when purity was critical.

Solution stability is decreased to days at the most when any oxygen is present, the solutions turning blue then a deep purple as the monomer is oxidized to form oligomers. In contrast to Chapters 3 and 4 where planarity was critical, the torque in the seven membered propylenedioxy bridge of ProDOP when viewed from the side should not be considered as a disadvantage in this monomer over EDOP. Recent studies from the Reynolds group have indicated that the open morphology provided by propyl and gem-dimethylpropyl linkers in

PProDOTs facilitates switching.21 Propyl linkers also impart a unique spectral feature in the UV-Vis-NIR spectra of conjugated polymers known as vibronic bands where multiple peaks are present and taken to be a sign of order in the polymer.

Despite the propensity to spontaneously polymerize in air, a variety of electrochemical methods are routinely used to polymerize XDOPs on Pt, ITO and glassy

20. Schottland, P.; Zong, K.; Gaupp, C. L.; Thomas, C. A.; Giurgiu, I.; Hickman, R.; Abboud, K. A.; Reynolds, J. R. Macromolecules 2000, 33(19) 7051-7061.

21. Kumar, A.; Welsh, D. M.; Morvant, M. C.; Abboud, K.; Reynolds, J. R. Chem. Mater. 1998, 10, 896-902. 168

A. EDOP top and side (below) views B. ProDOP top and side (below) views

Figure 5-4. Crystal structures of EDOP and ProDOP when viewed from above and from the side.

carbon electrodes. Compared to PPy, it is easier to obtain smooth, homogenous films of

PEDOP or PProDOP on ITO, even to the naked eye. Propylene carbonate (PC), a viscous,

high dielectric solvent widely used in the battery industry and a supporting electrolyte of

tetraethylammonium p-toluenesulfonic acid (tosylate, OTs) (TEAOTs) was found to be the

ideal solvent-electrolyte combination for galvanostatic, potentiostatic or scanning potential

deposition. Other solvents tried were acetontrile with several dopant ions ranging from

perchlorate, hexafluorophosphate, tetrafluorborate to triflate and

bis(trifluoromethane)sulfonylimide (3M salt) and γ-butyrolactone as solvent, like PC, also

common in the battery industry. This ideal solvent finding is no surprise given that high

quality films are also observed with PPy under these same conditions and it appears that

there is a special plasticizing effect in tosylate-pyrroles systems. The notable difference 169 between the XDOPs studied here and PPy are that no conditions have been found yet that allow the polymerization of an XDOP to proceed electrochemically in water using any of the above dopant anions in addition to dodecylbenzene sulfonate and Na+ poly(styrenesulfonate) (PSS).

To standardize experiments comparing the qualities of PPy, PEDOP and PProDOP based on film thickness, the film thickness vs. deposition charge profiles for several different dry films were investigated as prepared in PC based electrolytes (Figure 5-5). The results are that for a given charge density, PProDOP is thicker than PEDOP which is thicker than PPy. PPy and PEDOP are linear over the entire charge density range collected but

PProDOP has at least two different linear regimes. The first of these spans the thin film range until the deposition charge density is ca. 100 mC cm-2 where the membrane thickness is comparable to that of PEDOP. Between ca. 100 mC cm-2 and 300 mC cm-2 the film deposited is thicker per given coulomb than any of the other polymers. It appears there may be an even faster increase in film thickness above this charge density but data is limited.

Assuming comparable polymerization yields, polymer density is the most obvious factor affecting these values where PProDOP would be the least dense of the polymers and PPy thedensest.MostCPsareassumedtohaveadensitynear1.5gcm-3 but crude float tests were unable to pin down any differences in polymer film density between the above polymers.

5.3 Polymer Electrochemistry

The polymer electrodepositions by CV for PEDOP and PProDOP in PC are shown in Figures 5-6 and 5-7 respectively. These processes are typical of CPs where as the potential is scanned anodically, a peak corresponding to monomer oxidation is observed. 170

5

PProDOP PEDOP PPy 4

H N x

m OO

µ 3

/

H N x

OO 2

Film Thickness

H N x 1

0 0 100 200 300 400 500 Charge / mC cm-2

Figure 5-5. Film thickness vs. deposition charge for PPy, PEDOP and PProDOP. Approximately 0.9 cm2 films were deposited on ITO from [10 mM] monomer [0.1 M] TEAOTs in PC at 0.1 mA cm-2 for a varying amount of time until a desired charge was reached. Thickness is reported as measured across a scratch on a profilometer.

Upon scan direction reversal, no current peak is seen indicating that the oxidized monomer

has coupled to form dimer, trimer, and eventually oligomer and polymer. As the potential

is scanned cathodically, the current arising from the reduction of the oxidized polymer to

the neutral polymer is observed. Upon reversal of the potential scan, the polymer p-type

doping process is observed followed by subsequent monomer oxidation, all at higher

currents than in previous scans. This indicates oxidation on an electrode with a higher

surface area. The alternative electrochemical deposition techniques (galvanostatic and

potentiostatic) used in Chapters 3 and 4 are also applicable here and are preferable to CV 171

deposition because they are simpler and it is easier to measure the charge passed during

polymerization and the resulting film thickness.

A. PEDOP Deposition H N x

OO

0.2 mA

B. PEDOP CV 1st and 3000th Scan (Inset)

0.5 mA

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Potential /V vs SCE

Figure 5-6. PEDOP deposition and aqueous electrochemistry. (A) Electrochemical deposition from PC showing even cycles scans 4 through 20. (B) PEDOP switching in aqueous buffer initially and after 3000 scans indicates that it retains ca. 95% of its charge storage capacity.

After electrosynthesis in PC, films were washed with deionized water and [0.1 M]

KCl-H2O buffered to pH 7.2 with Na2HPO4 (phosphate buffered KCl). They were then

broken in over 20 double potential step cycles in Ar degassed solvent with an Ar blanket to

remove PC and OTs and to acclimate the polymer to chloride influx and egress. The CVs 172

A. PProDOP Deposition H N x

OO

0.5 mA

B. PProDOP CV 1st and 3000th Scan (Inset)

0.5 mA

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Potential /V vs SCE

Figure 5-7. PProDOP deposition and aqueous electrochemistry. (A) Electrochemical deposition from PC showing scans 2 through 21. (B) Cyclic voltammetry as a measure of stability indicates that PPy retains much (90%) of its charge storage capability after 3000 scans in aqueous buffer.

in aqueous KCl buffer for PEDOP and PProDOP are shown in part B of Figures 5-6 and 5-

7 respectively. They indicate that the switch from organic solvent and electrolyte to the

aqueous system cathodically shifts the E1/2 of each polymer in a significant manner. In fact,

the E1/2s observed for these polymers are among the lowest ever observed for any CP.

Repeated cycling of each of these polymers indicates they are extremely stable to

switching. Each figure shows the polymer shortly after it was deposited when it had less 173 than a 10 scan history on the outside. The inner CVs show the behavior of the polymers

3000 scans after the scan shown. Analysis of the charge stored in each cycle indicates that

PEDOP retains 96% of its electroactivity while PProDOP retains 90% of its electroactivity after 3000 cycles. By contrast, PPy under the same conditions retains a lower but still respectable 80% of its charge storage capability.22 Because of the similarity in polymer oxidation potentials and optical properties (Section 5.4) of PEDOP and PProDOP, and because of the relative difficulty in handling EDOP as a monomer, PProDOP is likely to be the candidate for next generation materials and its properties will be exclusively discussed below. Additionally, the ability to symmetrically modify the propyl bridge in PProDOP lends this polymer to more facile functionalization such as shown in Figure 5-1. Unless noted, all properties investigated for PEDOP that rely on polymer redox potential are identical to PProDOP further rationalizing this decision.

The in situ conductance of PProDOP (Figure 5-8), synthesized and characterized under identical electrochemical conditions to the CVs above, indicates that the conductance onset for p-type doping is quite cathodic at -0.65 V vs. SCE compared to any other CP as shown in Figure 5-8. Furthermore, unlike other pyrrole containing polymers studied by other groups, the polymer remains conducting even at potentials very anodic of the onset for conductance. In these other materials, often a window of conductivity is observed followed by a decay to low conductivities. While the method of in situ conductance measurement used in this work is unable to determine absolute conductivities, free standing films of PProDOP are routinely in the range of 50-80 S cm-1 (relative to PPy under the same

22. (a) Yamato, H.; Ohwa, M.; Wernet, W. J. Electroanal. Chem. 1995, 397, 163-170. (b) Pyo, M.; Reynolds, J. R.; Warren, L. F.; Marcy, H. O. Synth. Met. 1994, 68, 71-77. 174 conditions which is 250 S cm-1) at low current densities on glassy carbon electrodes.

Despite free standing film conductivities representing an entirely different aspect of conductivity than these solution measurements, the trend is probably preserved and

PProDOP films are slightly less conducting than PPy and perhaps an order of magnitude less than PEDOT. Of note also is the extremely low potential needed to polymerize the monomer used for conductance measurements, +0.63 V. This affords the possibility of using many more compounds as pendant groups attached to ProDOP monomers or including interesting easily oxidized dopant ions while retaining compatibility with the low oxidation potential.

5.4 Polymer Spectral Characteristics

Spectroelectrochemistry of PEDOP and PProDOP in aqueous KCl buffer are shown in Figure 5-9. Both polymers switch repeatedly in aqueous buffer solution and undergo clean state to state transitions as observed by the relatively sharp isosbestic points. In both part A and B of Figure 5-9, it is clear that complete reduction of these polymers is difficult even at potentials nearing water reduction (ca. -1 V) as spectral signatures of doping are evident near 1.5 eV for PEDOP and 1 eV for PProDOP. Films showing distinct electrochromic response (<1 µm thick) attain about 90% of their full color switch within

0.25 s compared to PPy, which takes over 1 s under similar conditions.23 The full electroactivity of PEDOP at extremely low potentials is evident by the nearly complete loss of the π to π* transition at 2.5 eV when the polymer is held at -0.33 V.24 Neutral PEDOP shows a band gap (defined as the onset of the π to π* transition) of 1.9 eV, while the band

23. Switching speeds are largely controlled by the supramolecular structure of the poly- mer film and are not suitably correlated to molecular structure. 175

6 H N x 5 OO

4

Gate Potential Removed

3 60

50 A µ 40 2 30 Psuedo ConductivityPsuedo /mS 20 Drain Current / 1 10 0 Egate = 0.63 V vs. SCE

0 30 60 90 120 150 0 Time /Seconds

-1.0 -0.5 0.0 0.5 1.0 Potential /V vs SCE

Figure 5-8. In situ conductance if PProDOP in aqueous buffer. PProDOP was deposited on 10 mm IME at a gate potential of +0.63 V vs. SCE in PC-OTs and bridged the electrode in ca. 55 seconds. Equilibrium conductance measurements were obtained every 0.5 V starting from neutral polymer at -0.9 V vs. SCE and scanning anodically from this potential.

gap for PProDOP is slightly higher at 2.2 eV as expected for the increased steric distortion

caused by the seven-membered ring. Both polymers show the typical loss of the interband

transition and evolution of low energy absorptions attributed to the charge carriers in the

conducting form.

24. Despite attempts made to standardize potentials, those derived from spectroelectro- chemical experiments frequently do not compare to those from standard Ag wire. 176

H 1.0 A. PEDOP 495 nm N x

OO 0.8 Eg ~ 1.9

ce 0.6

-0.33 V 0.4

Absorban

0.2

0.0 -0.93 V visible

B. PProDOP 520 nm 480 nm H 2.5 N x

OO 2.0 Eg ~ 2.25

ce 1.5 -0.13 V

1.0

Absorban

0.5

0.0 -0.73 V 01234 Energy /eV Figure 5-9. Spectroelectrochemistry of PEDOP and PProDOP in aqueous buffer. (A) PEDOP spectroelectrochemistry in buffer but deposited galvanostatically from PC. (B) PProDOP spectroelectrochemistry under same conditions. Spectra are recorded in 0.5 V increments from the oxidized form of the polymer to the neutral form. Potentials are referenced to Ag wire and corrected to an external SCE reference electrode and are not congruent with SCE potentials reported elsewhere in this work.

This stability in the oxidized state is greater than that exhibited by PPy and is

brought out in an experiment with the strongest biological reductants commonly available. 177

Dithiothreitol and glutathione are reducing agents which are themselves oxidized by the formation of a disulfide bond. Glutathione (GSH) is a tripeptide having the structure β-Glu-

Cys-Gly and dimerizes with the formation of a disulfide bond on the internal Cys while dithiothreitol (DTT) is a simple dithiol and undergoes internal disulfide formation (Figure

5-10). The oxidation potentials for GSH25 and DTT26 are ca. -0.5 V vs. SCE and -0.58 V vs. SCE respectively and are measured indirectly since sulfides react to foul Pt and Au electrodes. PEDOP and PProDOP-OTs films were prepared by galvanostatic electrochemical polymerization at 0.1 mA cm-2 to a charge density of -25 mC cm-2 on ITO.

These are roughly the same thickness as PPy films prepared at the same current density to

50 mC cm-2 (ca.1µm) and have comparable optical densities. After rinsing in acetonitrile anddrying,PEDOP-OTs,PProDOP-OTsandPPy-OTsfilmswereplacedinanaqueous reducingsolutionof30mMDTTred for 30 min. Upon removal from the medium, the PPy film had changed from its original black to a lighter rose color indicating reduction to the neutral form. The PEDOP film did not change from its original blue even after treatment with con-centrated DTTred solution (up to 2 M). When compared with the standard reduction potentials of DTT and GSH in water, the oxidized and conducting forms of

PEDOP and PProDOP are found to be stable, while oxidized PPy is unstable to these reductants as illustrated by the potential ladder shown in Figure 5-10. For PPy, the reaction leading to reduction is PPy+ +DTTred -> PPy0+DTTox (Figure 5-10). This reaction does not proceed for oxidized PEDOP. Identical results are obtained by reaction with GSH.

25. (a) Lamoureux, G. V.; Whitesides, G. M. J. Org. Chem. 1993, 58, 633. (b) Lees, W. J.; Whitesides, G. M. J. Org. Chem. 1993, 58, 642.

26. Millis, K. K.; Weaver, K. H.; Rabenstein, D. L. J. Org. Chem. 1993, 58, 4144. 178

A. Potential Ladder for Postulated Pyrrole Reduction B. Summary of Observed Reactions H H -0.8 RED OX N N x x

OO OO -0.7 SCE + PProDOP PProDOP vs. PProDOP PProDOP+ -0.6 DTTred DTTox X

HO HO -0.5 GSH GSSG SH S Potential /V SH S PPy PPy+ HO HO DTTred DTTox -0.4

C. Glutathione Structures OOH OOH H H H O H O N N N OH N OH x x H2N N H2N N O H O O H O PPy+ SH S PPy S O H O N NH HO N 2 O H HO O GSH GSSG

Figure 5-10. Explanation for the reaction with PPy+ and DTT. (A) Potential ladder indicating that PPy+ reactswithbothGSHandDTTred to form GSSG or DTTox and PPy. (B) Summary of observed reaction with DTT indicating that PProDOP does not react. (C) Structures of oxidized and reduced glutathione.

Smooth films of PEDOP and PProDOP are easily deposited (when compared to the

deposition characteristics of PPy) on ITO and are electrochromic, switching between

shades of highly transmissive blue/gray in the oxidized p-doped states (E = +0.85 V, x =

0.34, y = 0.38) and red/orange in the neutral state (E = -0.80 V, x = 0.52, y = 0.45).27 Figure

5-11 shows the colorimetry results relative to the CIE 1931 standard observer for PProDOP

undergoing p-type doping-dedoping in aqueous buffer. The colors exhibited are identical

27. x and y represent true color values as referenced to the CIE 1931 color table. True color values are reported as measured with a Minolta CS-100 Chromameter and refer- enced to the CIE color table. See Granström, M.; Berggren, M.; Pede,D.; Inganäs, O.; Andersson, M. R.; Hjertberg, T.; Wenneström, O. Supramol. Sci. 1997, 4, 27. 179

to those observed in PC-OTs solvent-electrolyte and are very similar to PEDOP, which has

a more red-colored reduced state. y 0.9 H N 520 x 525 0.8 515 530 OO 535 510 540 545 0.7 550 505 555 560 0.6 565 570 500 575 0.5 580 585 590 495 0.4 595 λ d, p-type = 481 600 λ = 584 +0.07 V vs. SCE d, neutral -0.93 V vs. SCE 610 620 0.3 490 630 650

0.2 485

480 0.1 475 470 460 0.0 380 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 x

Figure 5-11. Colorimetry of PProDOP in aqueous buffer. The star represents the CIE 1931 coordinates of the light source through ITO and aqueous KCl buffer (pH 7.2). The color transitions from a highly pure orange in the neutral state at -0.93 V to a highly transparent blue/gray at +0.07 V. 180

5.5 Electrochemical Quartz Crystal Microbalance Studies on Ion Transfer

Much of the interest in aqueous compatible CPs that are able to be repeatedly switched is for applications in controlled molecular release where the rate and period of bioactive molecules supplied to a reservoir can be controlled.28 This interest is largely due to the need to target increasingly potent drugs to the area of the body where they are physiologically needed, dodging possible side effects due to unintended systemic reactions.

CPs are interesting for these applications because of the ability to alter their redox states in small increments to allow controlled ion transport, all while preserving the ease of structural modification and molecular tuning inherent in organic materials. An alternative to polymeric active release is the system developed by Langer utilizing an array of wells containing a pharmacophore held in place by a thin Au film which can be electrochemically removed.29 Alternatives to active release are either the cleavage of a biologically inert polymer containing a drug molecule (poly(lactic acid-co-glycolic acid), PLGA),30 or the swelling of an inert polymer such as PEO/PEG containing water soluble drug molecules.31

An elegant alternative to the PLGA approach is to incorporate a difunctional drug such as

28. (a) Park, K., Ed. Controlled Drug Delivery, Challenges and Strategies;American Chemical Society: Washington, DC, 1997. (b) Langer, R. Science 1990, 249, 1527.

29. Langer, R. Acc. Chem. Res. 2000, 33, 94-101

30. (a) Miyajima, M.; Koshika, A.; Okada, J.; Ikeda, M. J. Control. Rel. 1999, 60, 199- 209. (b) Langer, R. J. Control. Rel. 1999, 62,7-11.

31. (A) Chandra, R.; Rustgi, R. Prog. Polym. Sci. 1998, 23, 1273-1335. (B) Moriyama, K.; Ooya, T.; Yui, N. J. Control. Rel. 1999, 59, 77-86. (C) Peppas, N. A.; Keys, K. B.; Torres, Lugo, M.; Lowman, A. M. J. Control. Rel. 1999, 62, 81-87. (D) Breitenbach, A.; Li, Y. X.; Kissel, T. J. Control. Rel. 2000, 64, 167-178. 181

salicylate into a polymer pro-drug where the degradation process itself releases the active

drug, a process Uhrich has used to implement anesthetic sutures.32

A. Cationic Pharmacophores for Controlled Release

HO HO HO OH OH HO HO HO NH3 NH2 NH2 CH3

DOPAMINE EPINEPHRINE METAPROTERENOL

B. Anionic Pharmacophores for Controlled Release O NH2 H CH3 N N - N CO2 O N OH O O O N N O NaO P O P O P O O NaO P O O ONa OH OH ONa OH OH N3 SALICYLATE ATP AZTMP Figure 5-12. Pharmacophores for controlled release. (A) cationic and anionic (B) molecules of pharmaceutical interest which are capable of being released from a conjugated polymer.

The relative simplicity and generality of using CPs to release charged

pharmacophores still remains an attractive option for actives release and the use of PPy to

this end has been well established for salicylate and ferrocyanide,33 glutamate,34 and

ATP.35 Figure 5-12 shows several molecules which are interesting as dopants in CPs. A

32. Uhrich, K. E.; Cannizzaro, S. M.; Langer, R. S.; Shakesheff, K. M. Chem. Rev. 1999, 99, 3181-3198.

33. Chang, A. C.; Miller, L. L. J. Electroanal. Chem., 1988, 247, 173.

34. Zinger, B.; Miller, L. L. J. Am. Chem. Soc. 1984, 106, 6861.

35. (a) Pyo, M.; Maeder, G.; Kennedy, R. T.; Reynolds, J. R. J. Electroanal. Chem. 1994, 368, 329. (b) Pyo, M.; Reynolds, J. R. Chem. Mater. 1996, 8, 128. 182 complete understanding of the mass transport properties of a CP release candidate must be understood before a device is designed around it. The mass transport characteristics for the

PPy p-type to neutral transition are anion dominant. That is, upon charge neutralization the compensating anions are released from the film. The mass transport dominance of PPy can be reversed by polymerization using an anionic polymeric counterion such as PSS. In this composite, the polymer is deposited in the oxidized state with sulfonate ions trapped in the film acting as counterions. Charge neutralization of the polymer leaves vacant anionic sites that are charge compensated by the diffusion of solution cations in the polymer membrane making the polymer cation dominant. EQCM is typically used to probe these mass transport dependencies as it allows measurement of mass changes on a surface while an electrochemical potential is applied. EQCM data for PProDOP deposited from PC-OTs and switched in aqueous KCl buffer is shown in Figure 5-13. Similar to PPy, the mass transport for PProDOP-Cl is anion dominant. Starting from p-type doped polymer at 0 V vs.SCE, the polymer film contains chloride ions. A cathodic potential scan neutralizes the polymer and the anions formerly used for charge compensation are released into solution with a concomitant mass decrease of the polymer. Because the redox potential of PProDOP is near where water is reduced, ca. -1 V vs. SCE, the polymer cannot be scanned very cathodic of its reduction potential and the mass transfer is not observed to plateau. On scan reversal,

- - the polymer is still neutral and Cl is still expelled until near the E1/2,whereCl ions are taken up again.

Attempts to switch the mass transport dominance of PProDOP by electrochemical deposition in the presence of PSS were unsuccessful. Since ProDOP cannot be deposited from aqueous solution and NaPSS is not soluble in PC, a soluble form of PSS had be 183

SOLUTION SOLUTION

O O

O H H O

N N O O

O H H O N

N N N H H O O O O

O O H O O H

N

N N N H H O O O O

O O H O O H

N

N N N H H O O O O

O

O H O O H N

POLYMER POLYMER N

N N

H H

O

O O O

O O O

H O H

O O O

O H O O H N

H N H N N

O O O O N N H

N O H N O

O

H O H

O

O O H O O O H N

H N N H N N N

N N

O O

O O O O

O

O N N H N N H

H H O H H O H H O O O O O O O

H O H

O H x N O O H x N O

N

H O N H O

O O N N

O O O

O O O

O O N

N N x N x

H

N H N

O O

O

O O OO

H H O

O H O O H O x O H O O H x

N N N N N

N O H O H

O O

N

O N O O O O O

N O N N N

H H

H H

H H O

O N N O O

H H H H

O O O

O HH O O O O HH O O O O

N N

N

N O H N O H N O O

O

O O N O N

O

N O O N O

O O

H H H H

O O N N N N

O

O H O O O O H O O N N

H H

O O

N

O H N O H

N

O N O H O O H

O O N

N N N

O

O O O

H

H N N H H

H H

O

O O N O N

H H

O

H H O O H H H O H

N

O N O O

O O O

N

O N H N O N H

O

O N N O N N O O

O O O H H

O O

H

O N H O N

O O N

O O O O N

N

O N O

N H N H H x H x O

O O H O O O H

N

N O O

O O

H H

H H O N N O N N O O

H

H O O H H O H O O H O N

O O N O

H N O H N

O

O N O O N O O

H H O N O N

N

O O N O O

N

O N O

H H O O O H O O O H O O N N

H H

O

O N N H H O O H H H O O H

O N O N O O O

O N O N O

O O O O

N

O N O

N

N O O

O O N N

H H O O H O O H H H

H H H H

H O O N H O O N

N N O O N N

O O O O O O O O O O O O O O O O O O O O O O O O

O O O O

O O

N N N O N O

H H H H H H H x H H H H H H H x

H H H

H N N N N N N N O H N N N N N N N O H N N N N N N N N N N N

H O N O N H O O N

N N H H H H H H H H H H H H x x O O O O O O O O O O O O O O O OO x O O O O O O O O O O O O O OO x O O O O

Pt on QUARTZ Pt on QUARTZ

H H N N Mass Increases x x

OO OO

J /mA cm-2 Mass /ng

0.1 mA cm-2 4000 ng

-1.0 -0.5 0.0 0.5

Potential /V vs SCE

Figure 5-13. EQCM and CV in aqueous buffer for PProDOP on Pt quartz crystal. Exposed crystal area was approximately 0.96 cm2. PProDOP was deposited to a charge density of 100 mC cm-2 in PC as described previously. After rinsing in aqueous buffer, and potential square wave break in, CV mass transport data was collected at 25 mV sec-1.

synthesized. This is accomplished by exchange of Na+ for TEA+ cations through a

membrane. TEAPSS is indeed PC soluble and electroactive films are achieved by

electrochemical deposition as above for PProDOP-Cl. The CV for this composite looks

similar to that of PProDOP-OTs. However, these films did not undergo mass changes in the

potential range over which the polymer dopes. 184

5.6 Conclusions

PProDOP was investigated as a replacement candidate for PPy in aqueous systems of biological interest. It excels compared to PPy in all areas examined except for cost and compatibility with water during electrosynthesis. The electrochemical stability and charge storage capacity for PProDOP and PEDOP under repeated potential cycling in aqueous solution indicates that both of these polymers retain higher electroactivity than PPy.

PProDOP and PEDOP undergo similar color transitions and are both nearly colorless in the p-type doped state making them ideal for optical observation of film confined species of biological interest. The high stability in the conducting state precludes the strongest biological reductants from neutralizing either PProDOP or PEDOP and the conductivity is sufficient and constant over the entire stable p-type doped range. Furthermore, the mass transfer characteristics with standard anionic dopants are similar to PPy. Since the ProDOP monomer is more stable than EDOP, it is suggested as the starting point for development of aqueous compatible systems.

5.7 Experimental Section

General.PC,Py,KCl,TEABr,NaPSSandTEAOTswerepurchasedfromAldrich.

All were used as received except for Py which was stored over KOH and passed through a short alumina (Brockman Activity I) column (gravity eluted in a disposable pipet) immediately prior to use. Millipore 18 MΩ water was used throughout. ProDOP and EDOP were prepared by Kyukwan Zong using the method described in the literature and were vacuum sublimed to a white powder, and used even if slightly gray. Monomer was stored in a freezer (vials degassed before closure) in the dark. Solutions of EDOP and ProDOP were prepared by degassing PC [0.1 M] TEAOTs prior to adding it to weighed monomer 185 in a volumetric flask. Flasks were stored in a refrigerator or freezer immediately upon use and were discarded after a week of use or when the color became more than slightly yellow

(almost always less than a week). Electrochemistry, spectroelectrochemistry, colorimetry, and in situ conductance were performed as described in Chapters 2 and 3.

Profilometry. A Dektak 3030 profilometer was used for film thickness measurements in collaboration with Roberta Hickman. Films were prepared from [10 mM] monomer [0.1 M] PC-OTs solutions and deposited at 0.1 mA until a desired charge density was reached. The films were rinsed in acetone and allowed to air dry, or if not measured immediately, stored in a vacuum dessicator. The film was scored in the middle with a piece of broken ITO36 and the profilometry measurement spanned the gap. The data was leveled across the gap and the surface roughness was calculated on the polymer while the film thickness was taken from the average film height to the average depth of the score.

Aqueous NaCl buffer preparation. 1.46 g NaCl, and 3.0 g NaH2PO4 were added to a 100mL volumetric flask. The flask was filled with 18 MΩ water and the solution was adjusted to pH 7.2 with ca. 10 % w/w NaOH. KCl buffer was prepared in an analogous fashion.

Reduction experiments. PEDOP and PProDOP were galvanostatically electrodeposited from [10 mM] monomer [0.1 M] PC at 0.1 mA cm-2 on an ITO slide until the current density reached 25 mC cm-2 (250 seconds for this ITO size). The films were rinsed in acetone and allowed to air dry. PPy was deposited under identical conditions except to 50 mC cm-2 charge density. 10 mM DTT (7.7 mg in 5 mL water) and GSH (15.3

36. Several scoring implements were used including those ranging in hardness from wooden dowels to razor blades but the broken ITO was found to give flattest region in the score. 186 mg in 5 mL) solutions were prepared. Polymer films (duplicates) were dipped in each solution for 30 minutes and the color observed and compared to a third film which was dipped in water.

EQCM. The UF (Don Cameron) designed cell was used on an EG&G (now Perkin-

Elmer) model QCA 917 as described in Chapter 2. The QCA probe and crystal were mounted vertically and the crystal sealed with poly(butadiene) o-rings. Monomer solution

(standard [10 mM] monomer, [0.1 M] TEAOTs) in PC was added and leads attached. An open circuit baseline was observed for 10 minutes in both the 20 kHz V-1 (used for deposition) and 2 kHz V-1 (used for observing polymer) ranges of the EQCM. When the crystal stabilized, as determined by minimal change during an open circuit experiment, the voltage output of the QCA was normalized to ca. +5 V so that a decrease in frequency (mass increase) range of 15 V X 20 kHz V-1 = 300 kHz total was available for observation.

Polymer was deposited galvanostatically at 100 µA for 196 seconds (A = 0.196 cm2) making a 100 mC cm-2 film (the frequency usually changing by 200 kHz). The monomer solution was then removed and the electrode was rinsed in the cell with 5 X 2 mL portions of Ar degassed KCl buffer. If the water solution leaked, the crystal was removed from the cell and the solvent swollen o-rings were replaced with fresh ones upon remounting. The polymer was then subjected to a CV to determine the potential limits for switching (current autoranging was turned off) and broken in by 20 double potential square waves of 20 seconds each over the CV determined limits. The mass changes were measured on the 2 kHz V-1 range which was zeroed prior to collection. Electrochemical data was then collected and the polymer treated as if it was broken in. 187

TEAPSS. (Effective molecular weight, 313.5 g/eq.) 10 g NaPSS (Mw 70,000) and

30 g TEABr were dissolved in 250 mL water with heating and stirring. This solution was loaded into a wet, knotted on one end, spectra/Por dialysis membrane (MWCO 3500, ca.9 inches before filling) and the other end knotted. This tube was then treated as follows at room temperature in a bucket. 48 h in distilled water (tube became taut). 48 h in a 500 mL water solution containing 25 g TEABr (done in a narrow vacuum dewar with stirring to limit solvent usage). 48 h in distilled water changed every 12 h on average. A small portion

- was then tested for Br content with AgNO3. No precipitate was observed and the control

TEABr solution tested positive. The water solution bumped excessively on the rotovap and was instead dried in a crystallization dish in a fume hood, redissolved in EtOH and dried in vacuo overnight to a yellow solid to yield 9.8 g isolated and dried. APPENDIX A CRYSTALLOGRAPHIC INFORMATION FOR CYANOVINYLENE MONOMERS

X = N or S N1 X2 C12 S1 C11

C6 C7 C10 O4 C5 C8 C2 C9 C15 C3 C4 O3 O1 C16 O2 C13 C14

Note: In thiophene containing monomers, O1, O2, C13, C14 and/or O3, O4, C15, C16 may be skipped.

Figure A-1. Numbering system for cyanovinylene monomer crystal structures. BTh- CNV, Th-CNV-EDOP and EDOT-CNV-EDOP were determined by Khalil Abboud. Th- CNV-EDOT and EDOT-CNV-Th were determined by Peter Steel.

() ∑ F0 – Fc R1 = ------∑ F0

⁄ []()2 2 2 12 ∑ wF0 – Fc 1 wR2 = ------, w = ------[]()2 2 []σ2()F2 ++()0.0370 × p 2 0.31 × p ∑ wF0 0

2 12⁄ ()2 2 2 2 ∑wF0 – Fc []max() F , 0 + 2 × F S = ------, p = ------0 c ()np– 3

188 189

Crystal data and structure refinement for BTh-CNV.

Empirical formula C11H7NS2 Formula weight 217.30 Temperature 173(2) K Wavelength 0.71073 Å Crystal system Orthorhombic Space group Pbca Unit cell dimensions a = 9.2545(7) Å α = 90° b = 18.090(2) Å β = 90° c = 23.917(2) Å γ = 90° Volume 4004.2(6) Å3 Z16 Density (calculated) 1.442 Mg/m3 Absorption coefficient 0.485 mm-1 F(000) 1792 Crystal size 0.40 x 0.15 x 0.04 mm3 Theta range for data collection 1.70 to 27.49° Index ranges –7 ≤≤h 11, – 23 ≤≤k 10, – 30 ≤≤l 10 Reflections collected 11624 Independent reflections 4224 [R(int) = 0.0545] Completeness to theta = 27.49° 92.0 % Absorption correction Integration Max. and min. transmission 0.9869 and 0.9571 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 4224 / 68 / 343 Goodness-of-fit on F2 1.014 Final R indices [I>2sigma(I)] R1 = 0.0560, wR2 = 0.1262 [2600] R indices (all data) R1 = 0.1026, wR2 = 0.1531 Largest diff. peak and hole 0.484 and -0.354 e.Å-3 190

Table A-1. Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2 x103) for BTh-CNV. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

xyzU(eq) S1 4489(2) 1302(1) 7015(1) 32(1) C2 4922(5) 1009(3) 7693(2) 27(1) S2 8379(2) -927(1) 8028(1) 38(1) C3 4160(18) 1371(8) 8110(7) 129(6) C4 3180(20) 1903(9) 7832(6) 50(4) C5 3248(11) 1924(5) 7266(4) 31(3) C6 5978(4) 432(2) 7790(2) 24(1) C7 6672(4) 46(2) 7386(2) 27(1) C8 7701(5) -542(3) 7416(2) 27(1) C9 8353(9) -918(4) 6945(3) 41(2) C10 9277(19) -1458(10) 7131(6) 53(5) C11 9449(14) -1501(6) 7692(5) 46(3) N1 6567(13) 216(6) 8838(2) 45(2) C12 6269(7) 297(4) 8372(2) 33(1) S1A 8101(7) -800(3) 6917(2) 48(2) C2A 7748(19) -574(9) 7604(5) 26(5) S2A 4185(5) 1353(2) 8104(2) 27(1) C3A 8470(40) -1063(18) 7961(11) 129(6) C4A 9300(40) -1605(16) 7629(12) 34(8) C5A 9350(50) -1460(20) 7074(12) 32(9) C6A 6698(15) -10(7) 7760(6) 45(4) C7A 5930(16) 422(7) 7421(7) 48(4) C8A 4882(17) 1001(9) 7491(5) 25(4) C9A 4250(30) 1428(14) 7040(9) 41(2) C10 3310(50) 1950(20) 7249(12) 80(18) C11A 3170(50) 1970(20) 7812(11) 21(7) 191

Table A-1. (Continued) Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2 x103) for BTh-CNV. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

xyzU(eq) N1A 6350(40) 69(17) 8832(7) 34(6) C12A 6560(30) 102(11) 8360(6) 35(6) S21 680(2) 236(1) 9250(1) 43(1) C22 1416(8) -622(3) 9404(3) 26(2) S22 -1353(3) -2875(2) 9309(2) 36(1) C23 2729(11) -612(5) 9555(4) 43(3) C24 3336(11) 180(5) 9535(4) 46(2) C25 2310(10) 664(4) 9389(3) 39(2) C26 577(5) -1290(2) 9366(2) 15(1) C27 -918(5) -1319(2) 9384(2) 17(1) C28 -1889(7) -1918(3) 9349(3) 24(1) C29 -3480(20) -1880(13) 9345(9) 187(9) C30 -4062(13) -2532(6) 9311(5) 42(2) C31 -3124(15) -3126(8) 9285(10) 38(4) N21 2179(11) -2443(5) 9222(5) 30(2) C32 1477(8) -1930(4) 9294(3) 29(2) S21A -3411(4) -1823(2) 9341(2) 44(1) C22A -1623(12) -2123(6) 9328(6) 35(4) S22A 2991(5) -528(2) 9577(2) 43(1) C23A -1530(30) -2812(11) 9324(15) 43(3) C24A -2990(30) -3204(14) 9310(20) 53(10) C25A -4050(20) -2719(9) 9298(11) 36(5) C26A -324(14) -1676(8) 9346(7) 91(5) C27A -109(17) -993(8) 9378(7) 90(5) C28A 1078(13) -476(7) 9381(7) 32(4) C29A 720(30) 298(15) 9273(15) 187(9) 192

Table A-1. (Continued) Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2 x103) for BTh-CNV. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

xyzU(eq) C30A 1814(16) 700(9) 9286(6) 40(4) C31A 3120(20) 388(8) 9488(7) 30(4) N21A 2050(30) -2472(15) 9214(15) 85(10) C32A 1130(13) -2056(7) 9282(7) 30(4) 193

Table A-2. Bond lengths [Å] for BTh-CNV.

Bond Bond Distance (Å) Bond Bond Distance (Å) S1-C5 1.716(7) S21-C25 1.727(9) S1-C2 1.752(6) S21-C22 1.736(7) C2-C3 1.386(15) C22-C23 1.268(12) C2-C6 1.450(6) C22-C26 1.438(7) S2-C11 1.644(11) S22-C31 1.702(12) S2-C8 1.738(6) S22-C28 1.805(7) C3-C4 1.483(17) C23-C24 1.540(13) C4-C5 1.356(12) C24-C25 1.337(14) C6-C7 1.355(5) C26-C27 1.385(6) C6-C12 1.439(6) C26-C32 1.437(7) C7-C8 1.430(6) C27-C28 1.409(7) C8-C9 1.446(9) C28-C29 1.47(2) C9-C10 1.373(12) C29-C30 1.30(2) C10-C11 1.354(14) C30-C31 1.382(16) N1-C12 1.157(7) N21-C32 1.145(7) S1A-C5A 1.71(2) S21A-C25A 1.728(15) S1A-C2A 1.725(13) S21A-C22A 1.742(12) C2A-C3A 1.40(2) C22A-C23A 1.248(18) C2A-C6A 1.457(15) C22A-C26A 1.450(14) S2A-C11A 1.62(2) S22A-C31A 1.675(14) S2A-C8A 1.725(13) S22A-C28A 1.834(13) C3A-C4A 1.48(3) C23A-C24A 1.52(2) C4A-C5A 1.35(2) C24A-C25A 1.31(2) C6A-C7A 1.331(14) C26A-C27A 1.253(13) C6A-C12A 1.454(15) C26A-C32A 1.519(14) C7A-C8A 1.438(15) C27A-C28A 1.443(15) C8A-C9A 1.451(18) C28A-C29A 1.46(2) 194

Table A-2. (Continued) Bond lengths [Å] for BTh-CNV.

Bond Bond Distance (Å) Bond Bond Distance (Å) C9A-C10 1.37(2) C29A-C30A 1.25(3) C10-C11A 1.36(2) C30A-C31A 1.415(18) N1A-C12A 1.149(16) N21A-C32A 1.148(14) 195

Table A-3. Bond angles[°] for BTh-CNV.

Atoms Bond Angle (°) Atoms Bond Angle (°) C5-S1-C2 91.6(3) C25-S21-C22 91.0(4) C3-C2-C6 124.6(8) C23-C22-C26 123.2(7) C3-C2-S1 114.0(8) C23-C22-S21 115.1(6) C6-C2-S1 121.4(4) C26-C22-S21 121.7(5) C11-S2-C8 93.3(4) C31-S22-C28 89.6(6) C2-C3-C4 107.2(12) C22-C23-C24 110.7(8) C5-C4-C3 115.8(11) C25-C24-C23 111.0(9) C4-C5-S1 111.4(7) C24-C25-S21 112.2(6) C7-C6-C12 121.1(4) C27-C26-C32 123.5(5) C7-C6-C2 125.2(4) C27-C26-C22 124.7(5) C12-C6-C2 113.8(5) C32-C26-C22 111.8(5) C6-C7-C8 131.5(4) C26-C27-C28 131.7(4) C7-C8-C9 126.0(6) C27-C28-C29 127.0(10) C7-C8-S2 125.5(4) C27-C28-S22 124.4(5) C9-C8-S2 108.5(4) C29-C28-S22 108.6(10) C10-C9-C8 110.1(8) C30-C29-C28 111.8(17) C11-C10-C9 115.7(9) C29-C30-C31 116.6(14) C10-C11-S2 112.2(7) C30-C31-S22 113.4(9) N1-C12-C6 176.2(8) N21-C32-C26 178.0(9) C5A-S1A-C2A 94.9(11) C25A-S21A-C22A 91.8(9) C3A-C2A-C6A 127.2(15) C23A-C22A-C26A 120.2(15) C3A-C2A-S1A 110.0(13) C23A-C22A-S21A 112.0(13) C6A-C2A-S1A 122.4(10) C26A-C22A-S21A 127.8(9) C11A-S2A-C8A 96.1(11) C31A-S22A-C28A 89.0(8) C2A-C3A-C4A 110.(2) C22A-C23A-C24A 113.9(18) C5A-C4A-C3A 115.(2) C25A-C24A-C23A 110.(2) C4A-C5A-S1A 109.1(18) C24A-C25A-S21A 111.8(16) 196

Table A-3. (Continued) Bond angles[°] for BTh-CNV.

Atoms Bond Angle (°) Atoms Bond Angle (°) C7A-C6A-C12A 118.3(13) C27A-C26A-C22A 133.1(14) C7A-C6A-C2A 127.6(13) C27A-C26A-C32A 108.2(13) C12A-C6A-C2A 114.0(12) C22A-C26A-C32A 118.6(11) C6A-C7A-C8A 135.8(14) C26A-C27A-C28A 139.5(16) C7A-C8A-C9A 125.1(14) C27A-C28A-C29A 116.7(15) C7A-C8A-S2A 128.4(11) C27A-C28A-S22A 134.6(10) C9A-C8A-S2A 106.5(11) C29A-C28A-S22A 108.1(13) C10-C9A-C8A 110.4(18) C30A-C29A-C28A 112.(2) C11A-C10-C9A 117.(2) C29A-C30A-C31A 117.8(19) C10-C11A-S2A 110.2(19) C30A-C31A-S22A 112.3(13) N1A-C12A-C6A 168.(3) N21A-C32A-C26A 166.(2) 197

Crystal data and structure refinement for EDOT-CNV-Th.

Empirical formula C13H9NO2S2 Formula weight 275.33 Temperature 153(2) K Wavelength 0.71073 Å Crystal system Orthrhombic Space group Pna2(1) Unit cell dimensions a = 18.606(3) Å α= 90° b = 12.0690(18) Å β= 90° c = 5.5511(8) Å γ = 90° Volume 1246.6(3) Å3 Z4 Density (calculated) 1.467 Mg/m3 Absorption coefficient 0.418 mm-1 F(000) 568 Crystal size 0.71 x 0.09 x 0.05 mm3 Theta range for data collection 2.76 to 26.41° Index ranges –22 ≤≤h 23, – 14 ≤≤k 15, – 4 ≤≤l 6 Reflections collected 13685 Independent reflections 2126 [R(int) = 0.0389] Completeness to theta = 26.41° 98.9 % Max. and min. transmission 0.9794 and 0.7555 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 2126 / 17 / 182 Goodness-of-fit on F2 1.051 Final R indices [I>2sigma(I)] R1 = 0.0318, wR2 = 0.0698 R indices (all data) R1 = 0.0420, wR2 = 0.0738 Absolute structure parameter 0.02(8) Largest diff. peak and hole 0.208 and -0.199 e.Å-3 198

Table A-4. Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2x103) for EDOT-CNV-Th. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

xyzU(eq) S(1) 6914(1) 8914(1) 5305(2) 38(1) S(2) 5490(1) 6791(1) 13531(3) 40(1) S(2') 6505(4) 4941(5) 13610(14) 29(1) O(1) 8059(1) 6429(1) 7572(4) 42(1) O(2) 8720(1) 7547(1) 3494(4) 51(1) N(1) 5508(1) 8770(2) 9486(5) 52(1) C(2) 7111(1) 7794(2) 7153(5) 31(1) C(3) 7753(1) 7330(2) 6468(5) 34(1) C(4) 8073(1) 7870(2) 4467(5) 37(1) C(5) 7685(1) 8740(2) 3656(6) 42(1) C(6) 6612(1) 7471(2) 9058(5) 31(1) C(7) 6700(1) 6575(2) 10498(6) 32(1) C(8) 6246(1) 6178(2) 12403(5) 32(1) C(9) 6359(4) 5195(5) 13749(14) 28(1) C(9') 5618(8) 6498(11) 13670(30) 42(1) C(10) 5840(1) 4976(2) 15535(6) 43(1) C(11) 5357(1) 5800(2) 15508(6) 47(1) C(12) 5992(1) 8183(2) 9328(5) 36(1) C(13) 8817(2) 6329(2) 6939(7) 53(1) C(14) 8920(2) 6450(2) 4291(7) 58(1) 199

Table A-5. Bond lengths [Å] for EDOT-CNV-Th.

Bond Bond Distance (Å) Bond Bond Distance (Å) S(1)-C(5) 1.715(3) C(2)-C(6) 1.461(4) S(1)-C(2) 1.736(2) C(3)-C(4) 1.418(4) S(2)-C(11) 1.642(3) C(4)-C(5) 1.351(4) S(2)-C(8) 1.710(2) C(6)-C(7) 1.354(4) S(2')-C(10) 1.635(4) C(6)-C(12) 1.446(3) S(2')-C(8) 1.705(3) C(7)-C(8) 1.435(4) O(1)-C(3) 1.372(3) C(8)-C(9) 1.417(6) O(1)-C(13) 1.458(3) C(8)-C(9') 1.417(6) O(2)-C(4) 1.376(3) C(9)-C(10) 1.408(7) O(2)-C(14) 1.445(3) C(9')-C(11) 1.409(7) N(1)-C(12) 1.149(3) C(10)-C(11) 1.341(4) C(2)-C(3) 1.373(3) C(13)-C(14) 1.489(5) 200

Table A-6. Bond angles[°] for EDOT-CNV-Th.

Atoms Bond Angle (°) Atoms Bond Angle (°) C(5)-S(1)-C(2) 92.47(13) C(9)-C(8)-C(7) 125.6(3) C(11)-S(2)-C(8) 93.03(13) C(9')-C(8)-C(7) 139.4(4) C(10)-S(2')-C(8) 91.2(2) C(9)-C(8)-S(2') 11.1(3) C(3)-O(1)-C(13) 111.1(2) C(9')-C(8)-S(2') 106.0(3) C(4)-O(2)-C(14) 111.4(2) C(7)-C(8)-S(2') 114.6(2) C(3)-C(2)-C(6) 130.2(2) C(9)-C(8)-S(2) 106.9(3) C(3)-C(2)-S(1) 109.75(19) C(9')-C(8)-S(2) 11.9(3) C(6)-C(2)-S(1) 120.04(18) C(7)-C(8)-S(2) 127.54(17) O(1)-C(3)-C(2) 124.1(2) S(2')-C(8)-S(2) 117.9(2) O(1)-C(3)-C(4) 122.7(2) C(10)-C(9)-C(8) 115.3(4) C(2)-C(3)-C(4) 113.2(2) C(11)-C(9')-C(8) 118.8(5) C(5)-C(4)-O(2) 123.8(2) C(11)-C(10)-C(9) 108.2(3) C(5)-C(4)-C(3) 113.2(2) C(11)-C(10)-S(2') 121.2(3) O(2)-C(4)-C(3) 123.0(2) C(9)-C(10)-S(2') 13.3(3) C(4)-C(5)-S(1) 111.3(2) C(10)-C(11)-C(9') 102.7(4) C(7)-C(6)-C(12) 120.6(2) C(10)-C(11)-S(2) 116.6(2) C(7)-C(6)-C(2) 124.3(2) C(9')-C(11)-S(2) 13.9(3) C(12)-C(6)-C(2) 115.1(2) N(1)-C(12)-C(6) 177.8(3) C(6)-C(7)-C(8) 129.1(2) O(1)-C(13)-C(14) 110.7(3) C(9)-C(8)-C(9') 95.0(5) O(2)-C(14)-C(13) 111.0(3) 201

Crystal data and structure refinement for Th-CNV-EDOT.

Empirical formula C13H9NO2S2 Formula weight 275.33 Temperature 148(2) K Wavelength 0.71073 Å Crystal system Monoclinic Space group P2(1)/n Unit cell dimensions a = 4.942(1) Å α= 90° b = 10.082(2) Å β= 90.430(10)° c = 23.879(3) Å γ = 90° Volume 1189.7(4) Å3 Z4 Density (calculated) 1.537 Mg/m3 Absorption coefficient 0.438 mm-1 F(000) 568 Crystal size 0.88 x 0.12 x 0.08 mm3 Theta range for data collection 2.19 to 23.98° Index ranges 0 ≤≤h 50, – ≤≤k 11, – 27 ≤≤l 27 Reflections collected 2134 Independent reflections 1872 [R(int) = 0.0391] Completeness to theta = 23.98° 99.9 % Max. and min. transmission 0.9658 and 0.6990 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 1872 / 0 / 163 Goodness-of-fit on F2 0.757 Final R indices [I>2sigma(I)] R1 = 0.0394, wR2 = 0.0682 R indices (all data) R1 = 0.0829, wR2 = 0.0738 Largest diff. peak and hole 0.237 and -0.258 e.Å-3 202

Table A-7. Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2x103) for Th-CNV-EDOT. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

xyzU(eq) S(1) -3185(2) 2138(1) 2415(1) 39(1) S(2) 3516(2) 1093(1) 427(1) 35(1) O(3) 4139(4) 4375(2) 1261(1) 27(1) O(4) 7707(4) 4317(2) 307(1) 30(1) N(1) -111(7) -1355(3) 1023(1) 48(1) C(2) -2879(7) 880(3) 1934(1) 24(1) C(3) -4678(6) -113(3) 2043(1) 28(1) C(4) -6283(7) 146(4) 2517(1) 35(1) C(5) -5683(7) 1321(4) 2756(2) 39(1) C(6) -843(7) 937(3) 1489(1) 24(1) C(7) 674(6) 2002(3) 1366(1) 24(1) C(8) 2708(6) 2202(3) 957(1) 22(1) C(9) 4267(6) 3322(3) 902(1) 22(1) C(10) 6038(6) 3292(3) 438(1) 21(1) C(11) 5862(7) 2144(3) 152(1) 33(1) C(12) -438(7) -321(4) 1211(1) 29(1) C(15) 5286(7) 5550(3) 1013(1) 33(1) C(16) 7937(7) 5280(3) 746(1) 34(1) 203

Table A-8. Bond lengths [Å] for Th-CNV-EDOT.

Bond Bond Distance (Å) Bond Bond Distance (Å) S(1)-C(5) 1.698(4) C(2)-C(6) 1.471(4) S(1)-C(2) 1.718(3) C(3)-C(4) 1.411(4) S(2)-C(11) 1.705(3) C(4)-C(5) 1.348(5) S(2)-C(8) 1.738(3) C(6)-C(7) 1.343(4) O(3)-C(9) 1.367(3) C(6)-C(12) 1.446(5) O(3)-C(15) 1.443(4) C(7)-C(8) 1.421(4) O(4)-C(10) 1.360(4) C(8)-C(9) 1.373(4) O(4)-C(16) 1.432(4) C(9)-C(10) 1.417(4) N(1)-C(12) 1.147(4) C(10)-C(11) 1.347(4) C(2)-C(3) 1.365(4) C(15)-C(16) 1.487(4)

. 204

Table A-9. Bond angles[°] for Th-CNV-EDOT

Atoms Bond Angle (°) Atoms Bond Angle (°) C(5)-S(1)-C(2) 91.76(18) C(9)-C(8)-C(7) 125.7(3) C(11)-S(2)-C(8) 92.35(16) C(9)-C(8)-S(2) 109.1(2) C(9)-O(3)-C(15) 111.1(2) C(7)-C(8)-S(2) 125.2(3) C(10)-O(4)-C(16) 113.0(2) O(3)-C(9)-C(8) 123.4(3) C(3)-C(2)-C(6) 128.0(3) O(3)-C(9)-C(10) 122.6(3) C(3)-C(2)-S(1) 110.7(2) C(8)-C(9)-C(10) 114.1(3) C(6)-C(2)-S(1) 121.2(3) C(11)-C(10)-O(4) 125.0(3) C(2)-C(3)-C(4) 112.8(3) C(11)-C(10)-C(9) 112.1(3) C(5)-C(4)-C(3) 112.3(3) O(4)-C(10)-C(9) 122.9(3) C(4)-C(5)-S(1) 112.4(3) C(10)-C(11)-S(2) 112.4(3) C(7)-C(6)-C(12) 121.5(3) N(1)-C(12)-C(6) 175.9(4) C(7)-C(6)-C(2) 125.1(3) O(3)-C(15)-C(16) 112.1(3) C(12)-C(6)-C(2) 113.3(3) O(4)-C(16)-C(15) 111.9(3) C(6)-C(7)-C(8) 131.5(3) 205

Crystal data and structure refinement for EDOT-CNV-EDOP.

Empirical formula C15H12N2O4S Formula weight 316.33 Temperature 173(2) K Wavelength 0.71073 Å Crystal system Triclinic Space group P-1 Unit cell dimensions a = 8.0521(4) Å α= 107.617(1)° b = 9.5708(5) Å β= 94.428(1)° c = 10.3402(5) Å γ = 113.677(1)° Volume 677.29(6) Å3 Z2 Density (calculated) 1.551 Mg/m3 Absorption coefficient 0.260 mm-1 F(000) 328 Crystal size 0.49 x 0.19 x 0.18 mm3 Theta range for data collection 2.12 to 27.49° Index ranges –10 ≤≤h 10, – 12 ≤≤k 11, – 9 ≤≤l 13 Reflections collected 4519 Independent reflections 3002 [R(int) = 0.0241] Completeness to theta = 27.49° 96.0 % Absorption correction Integration Max. and min. transmission 0.9626 and 0.9092 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 3002 / 0 / 200 Goodness-of-fit on F2 1.032 Final R indices [I>2sigma(I)] R1 = 0.0352, wR2 = 0.0949 [2716] R indices (all data) R1 = 0.0386, wR2 = 0.0979 Extinction coefficient 0.003(2) Largest diff. peak and hole 0.316 and -0.248 e.Å-3 206

Table A-10. Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2 x103) for EDOT-CNV-EDOP. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

xyzU(eq) S1 2340(1) 4027(1) 4923(1) 30(1) O1 96(2) 4963(1) 8063(1) 29(1) O2 -1033(2) 1476(1) 6711(1) 36(1) O3 2515(2) 10477(1) 10477(1) 34(1) O4 3281(2) 13652(1) 10247(1) 29(1) N1 4687(2) 8143(2) 4999(2) 36(1) N2 3528(2) 10737(1) 7263(1) 26(1) C2 1900(2) 5279(2) 6312(1) 23(1) C3 707(2) 4340(2) 6930(1) 23(1) C4 154(2) 2632(2) 6269(2) 26(1) C5 916(2) 2277(2) 5161(2) 32(1) C6 2709(2) 7042(2) 6640(1) 23(1) C7 2550(2) 8130(2) 7772(2) 26(1) C8 3041(2) 9817(2) 8105(2) 25(1) C9 2934(2) 10856(2) 9338(2) 25(1) C10 3329(2) 12368(2) 9223(2) 24(1) C11 3694(2) 12271(2) 7932(2) 27(1) C12 3771(2) 7647(2) 5706(1) 26(1) C13 -1536(2) 3698(2) 8220(2) 32(1) C14 -1279(2) 2199(2) 8080(2) 37(1) C15 2142(3) 11714(2) 11426(2) 37(1) C16 3471(3) 13421(2) 11547(2) 34(1) 207

Table A-11. Bond lengths [Å] for EDOT-CNV-EDOP.

Bond Bond Distance (Å) Bond Bond Distance (Å) S1-C5 1.7126(16) C6-C7 1.3650(19) S1-C2 1.7325(13) C6-C12 1.4382(19) O1-C3 1.3682(16) C7-C8 1.4182(19) O1-C13 1.4468(17) C7-H7A 0.9500 O2-C4 1.3684(17) C8-C9 1.3927(19) O2-C14 1.443(2) C9-C10 1.3965(19) O3-C9 1.3614(17) C10-C11 1.372(2) O3-C15 1.4434(17) C11-H11A 0.9500 O4-C10 1.3744(16) C13-C14 1.497(2) O4-C16 1.4342(19) C13-H13A 0.9900 N1-C12 1.1472(19) C13-H13B 0.9900 N2-C11 1.3658(18) C14-H14A 0.9900 N2-C8 1.3892(17) C14-H14B 0.9900 N2-H2A 0.8800 C15-C16 1.510(2) C2-C3 1.3739(19) C15-H15A 0.9900 C2-C6 1.4571(18) C15-H15B 0.9900 C3-C4 1.4243(19) C16-H16A 0.9900 C4-C5 1.358(2) C16-H16B 0.9900 C5-H5A 0.9500 208

Table A-12. Bond angles[°] for EDOT-CNV-EDOP.

Atoms Bond Angle (°) Atoms Bond Angle (°) C5-S1-C2 93.12(7) C11-C10-O4 128.99(13) C3-O1-C13 111.22(11) C11-C10-C9 107.48(12) C4-O2-C14 111.95(11) O4-C10-C9 123.51(13) C9-O3-C15 110.34(11) N2-C11-C10 107.99(12) C10-O4-C16 108.64(11) N2-C11-H11A 126.0 C11-N2-C8 110.19(11) C10-C11-H11A 126.0 C11-N2-H2A 124.9 N1-C12-C6 177.03(16) C8-N2-H2A 124.9 O1-C13-C14 111.21(13) C3-C2-C6 129.76(12) O1-C13-H13A 109.4 C3-C2-S1 109.60(10) C14-C13-H13A 109.4 C6-C2-S1 120.60(10) O1-C13-H13B 109.4 O1-C3-C2 124.05(12) C14-C13-H13B 109.4 O1-C3-C4 122.77(12) H13A-C13-H13B 108.0 C2-C3-C4 113.18(12) O2-C14-C13 111.46(13) C5-C4-O2 123.81(13) O2-C14-H14A 109.3 C5-C4-C3 113.25(13) C13-C14-H14A 109.3 O2-C4-C3 122.93(13) O2-C14-H14B 109.3 C4-C5-S1 110.84(11) C13-C14-H14B 109.3 C4-C5-H5A 124.6 H14A-C14-H14B 108.0 S1-C5-H5A 124.6 O3-C15-C16 111.98(13) C7-C6-C12 119.07(12) O3-C15-H15A 109.2 C7-C6-C2 124.42(12) C16-C15-H15A 109.2 C12-C6-C2 116.50(12) O3-C15-H15B 109.2 C6-C7-C8 129.76(13) C16-C15-H15B 109.2 C6-C7-H7A 115.1 H15A-C15-H15B 107.9 C8-C7-H7A 115.1 O4-C16-C15 111.78(13) N2-C8-C9 105.46(12) O4-C16-H16A 109.3 209

Table A-12. (Continued) Bond angles[°] for EDOT-CNV-EDOP.

Atoms Bond Angle (°) Atoms Bond Angle (°) N2-C8-C7 128.76(13) C15-C16-H16A 109.3 C9-C8-C7 125.48(13) O4-C16-H16B 109.3 O3-C9-C8 126.10(12) C15-C16-H16B 109.3 O3-C9-C10 125.01(12) H16A-C16-H16B 107.9 C8-C9-C10 108.88(12) APPENDIX B CRYSTALLOGRAPHIC INFORMATION FOR FLUORENONE MONOMERS

Figure B-1. Numbering system for fluorenone monomer crystal structures.

() ∑ F0 – Fc R1 = ------∑ F0

⁄ []()2 2 2 12 ∑ wF0 – Fc 1 wR2 = ------, w = ------[]()2 2 []σ2()F2 ++()0.0370 × p 2 0.31 × p ∑ wF0 0

2 12⁄ ()2 2 2 2 ∑wF0 – Fc []max() F , 0 + 2 × F S = ------, p = ------0 c ()np– 3

210 211

Crystal data and structure refinement for BEDOT-DCF.

Empirical formula C40H38N2O4S2Si2 Formula weight 731.02 Temperature 173(2) K Wavelength 0.71073 Å Crystal system Monoclinic Space group C2/c Unit cell dimensions a = 41.234(2) Å α= 90° b = 9.4131(4) Å β= 109.159(1)° c = 20.0598(8) Å γ = 90° Volume 7354.9(5) Å3 Z8 Density (calculated) 1.320 Mg/m3 Absorption coefficient 0.254 mm-1 F(000) 3072 Crystal size 0.21 x 0.18 x .07 mm3 Theta range for data collection 2.06 to 27.50° Index ranges –41 ≤≤h 53, – 10 ≤≤k 12, – 25 ≤≤l 25 Reflections collected 27056 Independent reflections 8381 [R(int) = 0.0890] Completeness to theta = 27.50° 99.2 % Absorption correction Integration Max. and min. transmission 0.9833 and 0.9537 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 8381 / 0 / 458 Goodness-of-fit on F2 1.041 Final R indices [I>2sigma(I)] R1 = 0.0509, wR2 = 0.1234 [5750] R indices (all data) R1 = 0.0812, wR2 = 0.1423 Extinction coefficient 0.00031(7) Largest diff. peak and hole 0.472 and -0.411 e.Å-3 212

Table B-2. Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2 x103) for BEDOT-DCF. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor. xyzU(eq)

S1 -946(1) 12243(1) 4782(1) 28(1) Si1 -1582(1) 12576(1) 5259(1) 27(1) O1 -806(1) 8180(2) 4900(1) 32(1) O2 -1384(1) 9025(2) 5324(1) 33(1) N1 -328(1) 5720(2) 4428(1) 36(1) C2 -746(1) 10675(2) 4688(1) 24(1) C3 -906(1) 9571(2) 4899(1) 25(1) C4 -1184(1) 9995(2) 5120(1) 25(1) C5 -1247(1) 11431(2) 5089(1) 26(1) C6 -469(1) 10668(2) 4383(1) 23(1) C7 -334(1) 9374(2) 4234(1) 25(1) C8 -83(1) 9410(2) 3914(1) 23(1) C9 38(1) 10706(2) 3734(1) 24(1) C10 -86(1) 11980(3) 3898(1) 29(1) C11 -337(1) 11951(3) 4222(1) 29(1) C12 102(1) 8241(2) 3701(1) 23(1) C13 75(1) 6828(3) 3821(1) 28(1) C14 -153(1) 6245(2) 4161(1) 29(1) C15 -1068(1) 7204(3) 4936(2) 34(1) C16 -1222(1) 7652(3) 5486(2) 34(1) C17 -1775(1) 13713(3) 4468(2) 45(1) C18 -1372(1) 13771(3) 6012(2) 42(1) C19 -1908(1) 11443(3) 5444(2) 56(1) S1' 1320(1) 9702(1) 2361(1) 31(1) Si1' 1955(1) 8652(1) 1994(1) 30(1) O1' 807(1) 6214(2) 2226(1) 32(1) 213

Table B-2. (Continued) Atomic coordinates (x 104) and equivalent isotropic displacement parameters (Å2 x103) for BEDOT-DCF. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor. xyzU(eq)

O2' 1418(1) 5823(2) 1836(1) 33(1) N1' 437(1) 4891(3) 3482(2) 73(1) C2' 986(1) 8662(3) 2433(1) 26(1) C3' 1030(1) 7302(3) 2232(1) 25(1) C4' 1329(1) 7121(3) 2035(1) 26(1) C5' 1521(1) 8320(3) 2080(1) 30(1) C6' 735(1) 9237(2) 2733(1) 25(1) C7' 551(1) 8343(3) 3036(1) 26(1) C8' 329(1) 8914(2) 3355(1) 24(1) C9' 286(1) 10403(2) 3374(1) 23(1) C10' 464(1) 11296(2) 3068(1) 28(1) C11' 684(1) 10712(3) 2749(1) 29(1) C14' 276(1) 5766(3) 3624(2) 45(1) C15' 955(1) 4840(3) 2187(2) 36(1) C16' 1135(1) 4844(3) 1641(2) 35(1) C17' 2285(1) 8607(4) 2886(2) 51(1) C18' 1944(1) 10463(3) 1628(2) 48(1) C19' 2055(1) 7314(4) 1411(2) 56(1) C20 -1877(1) 12466(3) 1325(2) 52(1) C21 -2060(1) 12840(3) 651(2) 48(1) C22 -2246(1) 14079(3) 516(2) 45(1) C23 -2247(1) 14957(3) 1066(2) 46(1) C24 -2065(1) 14587(4) 1744(2) 55(1) C25 -1877(1) 13334(4) 1878(2) 55(1) 214

Table B-3. Bond lengths [Å] for BEDOT-DCF.

Bond Bond Distance (Å) Bond Bond Distance (Å) S1-C2 1.729(2) S1'-C2' 1.732(2) S1-C5 1.737(2) S1'-C5' 1.734(3) Si1-C19 1.848(3) Si1'-C18' 1.851(3) Si1-C18 1.854(3) Si1'-C19' 1.856(3) Si1-C17 1.863(3) Si1'-C17' 1.858(3) Si1-C5 1.868(2) Si1'-C5' 1.878(3) O1-C3 1.372(3) O1'-C3' 1.376(3) O1-C15 1.439(3) O1'-C15' 1.444(3) O2-C4 1.379(3) O2'-C4' 1.373(3) O2-C16 1.442(3) O2'-C16' 1.439(3) N1-C14 1.144(3) N1'-C14' 1.151(4) C2-C3 1.369(3) C2'-C3' 1.372(3) C2-C6 1.464(3) C2'-C6' 1.461(3) C3-C4 1.418(3) C3'-C4' 1.421(3) C4-C5 1.373(3) C4'-C5' 1.366(4) C6-C11 1.404(3) C6'-C7' 1.400(3) C6-C7 1.411(3) C6'-C11' 1.406(3) C7-C8 1.388(3) C7'-C8' 1.385(3) C8-C9 1.408(3) C8'-C9' 1.415(3) C8-C12 1.479(3) C9'-C10' 1.382(3) C9-C10 1.386(3) C10'-C11' 1.386(3) C9-C9' 1.461(3) C15'-C16' 1.511(4) C10-C11 1.392(4) C20-C21 1.361(5) C12-C13 1.362(3) C20-C25 1.378(5) C12-C8' 1.479(3) C21-C22 1.372(4) C13-C14' 1.434(4) C22-C23 1.379(4) C13-C14 1.439(4) C23-C24 1.366(5) 215

Table B-4. Bond angles[°] for BEDOT-DCF.

Atoms Bond Angle (°) Atoms Bond Angle (°) C2-S1-C5 94.85(11) C10-C11-C6 121.8(2) C19-Si1-C18 111.30(16) C13-C12-C8’ '127.0(2) C19-Si1-C17 111.44(16) C13-C12-C8 126.6(2) C18-Si1-C17 107.49(14) C8'-C12-C8 106.37(19) C19-Si1-C5 109.49(12) C12-C13-C14’ 123.0(2) C18-Si1-C5 108.86(12) C12-C13-C14 124.1(2) C17-Si1-C5 108.17(12) C14'-C13-C14 113.0(2) C3-O1-C15 112.28(19) N1-C14-C13 176.8(3) C4-O2-C16 112.48(19) O1-C15-C16 111.3(2) C3-C2-C6 130.3(2) O2-C16-C15 111.2(2) C3-C2-S1 108.71(18) C2'-S1'-C5’ 94.69(12) C6-C2-S1 120.92(18) C18'-Si1'-C19’ 110.84(17) C2-C3-O1 123.4(2) C18'-Si1'-C17’ 108.55(16) C2-C3-C4 113.8(2) C19'-Si1'-C17’ 110.27(17) O1-C3-C4 122.8(2) C18'-Si1'-C5’ 106.83(13) C5-C4-O2 122.9(2) C19'-Si1'-C5’ 111.19(13) C5-C4-C3 115.1(2) C17'-Si1'-C5’ 109.06(14) O2-C4-C3 122.0(2) C3'-O1'-C15’ 111.90(19) C4-C5-S1 107.57(18) C4'-O2'-C16’ 112.08(19) C4-C5-Si1 134.43(19) C3'-C2'-C6’ 130.3(2) S1-C5-Si1 117.93(13) C3'-C2'-S1’ 108.49(18) C11-C6-C7 119.0(2) C6'-C2'-S1’ 120.96(18) C11-C6-C2 120.4(2) C2'-C3'-O1’ 123.0(2) C7-C6-C2 120.6(2) C2'-C3'-C4’ 113.9(2) C8-C7-C6 118.9(2) O1'-C3'-C4’ 123.2(2) C7-C8-C9 121.4(2) C5'-C4'-O2’ 123.3(2) C7-C8-C12 130.5(2) C5'-C4'-C3’ 114.8(2) 216

Table B-4. (Continued) Bond angles[°] for BEDOT-DCF.

Atoms Bond Angle (°) Atoms Bond Angle (°) C9-C8-C12 108.2(2) O2'-C4'-C3’ 121.9(2) C10-C9-C8 119.9(2) C4'-C5'-S1’ 108.14(19) C10-C9-C9’ 131.3(2) C4'-C5'-Si1’ 133.0(2) C8-C9-C9’ 108.8(2) S1'-C5'-Si1’ '118.53(15) C9-C10-C11 118.9(2) C7'-C6'-C11 '118.5(2) C7'-C6'-C2’ 121.0(2) C10'-C11'-C6’ 121.9(2) C11'-C6'-C2’ 120.4(2) N1'-C14'-C13 178.1(3) C8'-C7'-C6’ 120.1(2) O1'-C15'-C16’ 110.8(2) C7'-C8'-C9’ 120.3(2) O2'-C16'-C15’ 110.7(2) C7'-C8'-C12 131.8(2) C21-C20-C25 120.0(3) C9'-C8'-C12 107.9(2) C20-C21-C22 120.5(3) C10'-C9'-C8’ 120.1(2) C21-C22-C23 119.9(3) C10'-C9'-C9 131.2(2) C24-C23-C22 120.0(3) C8'-C9'-C9 108.7(2) C23-C24-C25 119.9(3) C9'-C10'-C11’ 119.1(2) C20-C25-C24 119.7(3) REFERENCES

Ahonen, H. J.; Lukkari, J.; Kankare, J. Macromolecules 2000, 33(18), 6787-6793

Albert, K. J.; Lewis, N. S.; Schauer, C. L.; Sotzing, G. A.; Stitzel, S. E.; Vaid, T. P.; Walt, D. R. Chem. Rev. 2000, 100(7), 2595-2626

Aleshin, A. N.; Kiebooms, R.; Heeger, A. J.; Synth. Met. 1999, 101, 369-370

Aleshin, A. N.; Lee, K.; Lee, J. Y.; Kim, D. Y.; Kim, C. Y. Synth. Met. 1999, 99, 27-33

Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications Wiley: New York. 2001; p 54

Belfield, K. D.; Hagan, D. J.; Van Stryland, E. W.; Schafer, K. J.; Negres, R. A. Org. Lett. 1999, 1, 10, 1575-1578

Belfield, K. D.; Scafer, K. J.; Mourad, W.; Reinhardt, B. A. J. Org. Chem. 2000, 65, 15, 4475-4481

Belletête, M.; Beaupré, S.; Beuchard, J.; Blondin, P.; Leclerc, M.; Durocher, G. J. Phys. Chem. B 2000, 104, 9118-9125

Belletête, M.; Morin, J.-F.; Beaupré, S.; Ranger, M.; Leclerc, M.; Durocher, G. Macromol- ecules, 2001, 34, 2288-2297

Berns, R. S. Billmeyer and Saltzman’s Principles of Color Technology, Third Edition, Wiley & Sons: New York, 2000

Biczók, L.; Bérces, T.; Inoue, H. J. Phys. Chem. A 1999 3837-3842

Blockhuys, F.; Claes, M.; Van Grieken, R.; Geise, H. J. Anal. Chem. 2000, 72, 3366-3368

Blythe, A. R. Polymer Testing 1984, 4, 195-200

Borjas,R.;Buttry,D.A.Chem. Mater. 1991 3, 872

Bott, A. W. Current Separations, 1997, 16(1), 23-26

Boudreaux, D. S.; Chance, R. R.; Bredas, J. L.; Silbey, R. Phys Rev. B: Condens. Matter. 1983, 28, 6927

217 218

Brédas, J.-L.; Cornil, J.; Beljonne, D.; Dos Santos, D. A.; Shuai, Z. Acc Chem. Res., 1999, 32, 267-276

Brédas, J. L.; Heeger, A. J. Macromolecules 1990, 23, 1150-1156

Brédas,J.L.;Street,G.B.Acc. Chem. Res. 1985, 18, 309-315

Brédas, J. L.; Themans, B.; Fripiat, J. G.; Andre, J. M.; Chance, R. R. Phys.Rev.B:Con- dens. Matter. 1984, 29, 6761

Breitenbach, A.; Li, Y. X.; Kissel, T. J. Control. Rel. 2000, 64, 167-178

Bruckstein, S.; Shay, M. Electrochim. Acta. 1985, 20, 1295. (B) Skoog, D. A.; West, D. M. Fundamentals in Analytical Chemistry; 4th ed.; Saunders: Philadelphia, 1982, p. 586

Buttry, D. A. Electroanal. Chem. Bard, A. J., Ed.; Dekker: New York, 1991; Vol. 17 p. 1

Buttry, D. A.; Ward, M. D. Chem. Rev. 1992, 92, 1355

Chan,H.S.O.;Ng,S.C.Prog. Polym. Sci. 1998, 23, 1167-1231

Chandra, R.; Rustgi, R. Prog. Polym. Sci. 1998, 23, 1273-1335

Chang, A. C.; Miller, L. L. J. Electroanal. Chem., 1988, 247, 173

Chiang, C. K.; Park, Y. W.; Heeger, A. J.; Shirakawa, H.; Louis, E. J.; MacDiarmid, A. G. Phys.Rev.Lett. 1977, 39, 1098

Chien, J. C. W. Polyacetylene Chemistry, Physics and Material Science; Academic Press, Harcourt Brace Jovanovich: Orlando, 1984; Chapter 5

Cornil, J.; Beljonne, D.; Brédas, J. L. J. Chem. Phys. 1995, 103(2) 834-841

Crooks, R. M.; Chyan, O. M.; Wrighton, M. S. Chem. Mater. 1989 1,2

Davies, A. G. Organotin Chemistry; VCH: Weinheim, New York, Basel, Cambridge, Tokyo, 1997; pp. 18-24

Deakin, M. R.; Buttry, D. A. Anal. Chem. 1989, 61, 1147A de Leeuw, D. M. Physics World 1999 31 de Leeuw, D. M.; Simenon, M. M. J.; Brown, A. R.; Einerhand, R. E. F. Synth. Met. 1997, 53-59 219

Diaz, A. F.; Bargon, J. Handbook of Conducting Polymers, 1986, Skotheim, T. A., Ed.; Dekker: New York, Vol. 1; pp. 81-115

Donat-Bouillud, A.; Lévesque, I.; Tao, Y.; D’Iorio, Beaupré, S.; Blondin, P.; Ranger, M.; Bouchard, J.; Leclerc, M. Chem. Mater. 2000, 12, 1931-1936

Ellis, D. L.; Zakin, M. R.; Bernstein, L. W.; Rubner, M. F. Anal. Chem. 1996, 68(5), 816- 822

Fesser, K.; Bishop, A. R.; Campbell, D. K. Phys. Rev. B. 1983, 27, 4804

Fincher, C. R.; Chen, C. E.; Heeger, A. J.; Macdiarmid, A. G.; Hastings, J. B. Phys. Rev. Lett. 1982, 48, 100

Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Marks, R. N.; Taliani, C.; Bradley, D. D. C.; Dos Santos, D. A.; Bredas, J. L.; Logdlund, M.; Salaneck, W. R. Nature 1999, 397, 121-128

Furukawa, Y. J. Phys. Chem. 1996, 100, 39, 15644-15653

Garner, B.; Hodgson, A. J.; Wallace, G. G.; Underwood, P. A. J. Mater. Sci.: Mater. Med. 1999, 10, 19-27

Garnier, F.; Korri-Youssoufi, H.;Srivastava,P.;Yassar,A.J. Am. Chem. Soc. 1994, 116, 8813-8814

Gorman, C. B.; West, R. C.; Palovich, T. U.; Serron, S. Macromolecules 1999, 32(13), 4157-4165

Granström, M.; Berggren, M.; Pede,D.; Inganäs, O.; Andersson, M. R.; Hjertberg, T.; Wen- neström, O. Supramol. Sci. 1997, 4,27

Gray,F.M.Solid Polymer Electrolytes-Fundamentals and Technological Applications; VCH: Weinheim, Germany, 1991

Greenham, N. C., Moriatti, S. C.; Bradley, D. D. C.; Friend, R. H.; Holmes, A. B. Nature 1993, 365, 628-630

Grell, M.; Long, X.; Bradeley, D. D. C.; Inbasekara, M. Woo, E. P. Adv. Mater. 1997, 9, 798

Groenendaal, L. B.; Jonas, F., Freitag, D.; Pielartzik, H.; Reynolds, J. R. Adv. Mater. 2000, 12, 7, 481-494

Halls, J. J. M.; Walsh, C. A.; Greenham, N. C.; Marseglia, E. A.; Friend, R. H.; Moratti, S. C.; Holmes, A. B. Nature 1995, 376, 498-500 220

Hanack, M.; Segura, J. L.; Spreitzer, H. Adv. Mater. 1996, 8, 663-666

Harris, R. K.; Kennedy, J. D.; McFarlane, W. Group IV-Silicon, Germanium, Tin and Lead. In NMR and the Periodic Table; Harris, R. K. and Mann, B. E. Ed.; Academic Press: Lon- don, New York, San Francisco, 1978; pp. 310-330, 342-366

Harrison, P. G. Investigating Tin Compounds Using Spectroscopy. In Chemistry of Tin; Harrison, P. G. Ed.; Blackie & Son Ltd.: Glasgow, London, 1989; pp. 71-81

Havinga, E. W.; ten Hoeve, E. E.; Wynberg, H. Synth. Met. 1993, 55-57, 299

Heeger, A. J.; Kivelson, S.; Schrieffer J. R.; Su, W.-P. Rev. Mod. Phys. 1988, 60,781

Heineman,W.R.Denki Kagaku 1982, 50(2), 142-148

Heitner-Wirguin, C. J. Membr. Sci. 1996, 120,1

Hi, D.; Yu, J.; Barbara, P. F. J. Am. Chem. Soc. 1999, 121, 6936

Hill, M. G.; Penneau, J.-F.; Zinger, B.; Mann, K. R.; Miller, L. L. Chem. Mater. 1992, 4, 1106-1113

Ho, H. A.; Brisset, H.; Elandaloussi, E. H.; Frére, P.; Roncali, J. Adv. Mater. 1996, 8, 990

Hoffman, R. Angew. Chem. Int. Ed. Eng. 1987, 26, 846-878

Hsu, W. Y.; Gierke, T. D. J. Membr. Sci. 1983, 13, 307

Hu, D.; Yu, J.; Barbara, P. F. J. Am. Chem. Soc. 1999, 121, 6936-6937

Huang, H; Pickup, P. G. Chem. Mater., 1998, 10(8), 2212 -2216

Huang, H; Pickup, P. G. Chem. of Mater.; 1999; 11(6); 1541-1545

Ishii, H.; Sugitama, K.; Ito, E.; Seki, K. Adv. Mater. 1999, 11, No. 8, 605-625

Justin Thomas, K. R.; Lin, J. T.; Lin, H.-M.; Chang, C. P. Chuen, C.-H. Organometallics 2001, 20, 557-563

Justin Thomas, K. R.; Lin, J. T.; Lin, Y.-Y.; Tsai, C.; Sun, S.-S. Organometallics (ASAP Received January 18, 2001)

Kitamura, C.; Tanaka, S.; Yamashita, Y. J. Chem. Soc., Chem. Commun. 1994, 1585

Kittleson, G. P.; Wight, H. S.; Wrighton, M. S. J. Am. Chem. Soc. 1984, 106, 7389-7396 221

Kobayashi M.; Colaneri N.; Boysel M.; Wudl F.; Heeger A.J. J. Chem. Phys. 1985 82(12) 5717-5723

Koch, N.; Rajagopal, A.; Ghijsen, J.; Johnson, R. L.; Leising, G.; Pireaux, J.-J. J. Phys Chem. B 2000, 104, 1434-1438

Koopmans Physica, 1933, 1, 104

Korri-Youssoufi, H.; Garnier, F.; Srivastava, P.; Godillot, P.; Yassar, A. J. Am. Chem. Soc. 1997, 119, 7388-7389

Kraft, A.; Grimsdale, A. C.; Holmes, A. B. Angew. Chem., Int. Ed. Engl. 1998, 37, 402-428

Kumar, A.; Welsh, D. M.; Morvant, M. C.; Abboud, K.; Reynolds, J. R. Chem. Mater. 1998, 10, 896-902

Lambert, T. L.; Ferraris, J. P.; Chem Commun. 1991, 752-754 also, 1268-1270

Lamoureux, G. V.; Whitesides, G. M. J. Org. Chem. 1993, 58, 633

Langer, R. Acc. Chem. Res. 2000, 33, 94-101

Langer, R. J. Control. Rel. 1999, 62,7-11

Langer, R. Science 1990, 249, 1527

Larmat, F.; Reynolds, J. R.; Reinhardt, B. A.; Brott, L. L.; Clarson, S. J. 1997, 3627-3636

Lee, K.; Miller, E. K.; Aleshin, A. N.; Menon, R.; Heeger, A. J.; Kim, J. H.; Yoon, C. O.; Lee, H. Adv. Mater. 1998, 10(6), 456-459

Lees, W. J.; Whitesides, G. M. J. Org. Chem. 1993, 58, 642

Long, J. W.; Kim, I. K.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 11510

March, J. Advanced Organic Chemistry; Wiley: New York, 1985

McQuade, D. T.; Pullen, A. E.; Swager, T. M. Chem. Rev. 2000, 100(7), 2537-2574

Merz, A.; Schropp, R.; Dötterl, E. Synthesis 1995, 795

Millis, K. K.; Weaver, K. H.; Rabenstein, D. L. J. Org. Chem. 1993, 58, 4144

Miyajima, M.; Koshika, A.; Okada, J.; Ikeda, M. J. Control. Rel. 1999, 60, 199-209

Montgomery, H. C. J. Appl. Phys. 1971, 42(7) 2971-2975 222

Moriyama, K.; Ooya, T.; Yui, N. J. Control. Rel. 1999, 59, 77-86

Moutet, J.-C.; Saint-Aman, E.; Tran-Van, F.; Angibeaud, P.; Utille, J.-P. Adv. Mater. 1992, 4, No. 7/8, 511-513

Mysyk, D. D.; Perepichka, I. F.; Perepichka, D. F.; Bryce, M. R.; Popov, A. F.; Goldenberg, L. M.; Moore, A. J. J. Org. Chem. 1999, 64, 19 6937-6950

Nazzal,A.;Street,G.B.Chem. Comm. 1984

Neef, C. J.; Brotherston, I. D.; Ferraris, J. P. Chem. Mater. 1999, 11, 1957-1958

Ochmanska, J.; Pickup, P. G. J. Electroanal. Chem. 1991, 297, 211-224

Ofer, D.; Crooks, R. M.; Wrighton, M. S. J. Am. Chem. Soc. 1990, 112, 7869-7879

Paeschke, M.; Wollenberger, U.; Kohler, C.; Lisec, T.; Schnakenberg, U.; Hintsche, R. Anal. Chim. Acta. 1995, 305, 126-136

Pande, R.; Ruben, G. C.; Lim, J. O.; Tripathy, S.; Marx, K. A. Biomaterials, 1998, 19, 1657-1667

Park, K., Ed. Controlled Drug Delivery, Challenges and Strategies; American Chemical Society: Washington, DC, 1997

Patil, A. O.; Heeger, A. J.; Wudl; F. Chem. Rev. 1988, 88, 183-200

Paul, E. W.; Ricco, A. J.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 1441-1447

Pei, Q.; Yang, Y. J. Am. Chem. Soc. 1996, 118, 7416

Peppas, N. A.; Keys, K. B.; Torres, Lugo, M.; Lowman, A. M. J. Control. Rel. 1999, 62, 81-87

Peppas, N. A.; Langer, R. Science 1994, 263, 1715-1720

Perepichka, I. F.; Kuz’mina, L. G.; Perepichka, D. F.; Bryce, M. R.; Goldenberg, L. M.; Popov, A. F.; Howard, J. A. K. J. Org. Chem. 1998, 63, 6484-6493

Perepichka, I. F.; Popov, A. F.; Orekhova, T. V.; Bryce, M. R.; Andrievskii, A. M.; Bat- sanov, A. S.; Howard, J. A. K.; Sokolov, N. I. J. Org. Chem. 2000, 3053-3063

Pfluger, P.; Street, G. B. J. Chem. Phys. 1984, 80, 544

Prezyna, L. A.; Qiu, Y.-J.; Reynolds, J. R.; Wnek, G. E. Macromolecules, 1991, 24, 5283- 5287 223

Pyo, M.; Maeder, G.; Kennedy, R. T.; Reynolds, J. R. J. Electroanal. Chem. 1994, 368, 329

Pyo, M.; Reynolds, J. R. Chem. Mater. 1996, 8, 128

Pyo, M.; Reynolds, J. R.; Warren, L. F.; Marcy, H. O. Synth. Met. 1994, 68, 71-77

Randriamahazaka H.; Noel V.; Chevrot C.; J. Electroanal. Chem. 1999, 472(2), 103-111

Rebbi, C. Sci. Am. 1979, 240,92

Reynolds, J. R.; Schlenoff, J. B.; Chien, J. C. W. J. Electrochem. Soc. 1985, 1131-1135

Roncali, J. Chem. Rev. 1992, 92, 711-738

Roncali, J. Chem. Rev. 1997, 97, 173

Roncali, J. J. Mater. Chem. 1999, 9, 1875-1893

Sakamoto, A.; Furukawa, Y.; Tasumi, M. J. Phys. Chem. B 1997, 101, 1726-1732

Salzner, U.; Kiziltepe, T. J. Org. Chem. 1999, 64, 764-769

Salzner, U.; Pickup P. G.; Poirier, R. A.; Lagowski., J. B. J. Phys. Chem. A, 1998, 102 (15), 2572 -2578

Savage, D. J.; Schell, B. A.; Brady, B. K. Imaging Element Containing Poly(3,4-Ethylene Dioxypyrrole)/Styrene Sulfonate. U. S. Patent 5,665,498, September 9, 1997

Schafer, A. J.; Hawkins, J. R. Nature Biotech. 1998, 16, 33-39

Schiavon, G.; Sitran, S. Zotti, G. Synth. Met. 1989, 32, 209-217

Schmidt, C. E.; Shastri, V. R.; Vacanti, J. P.; Langer, R. Proc. Natl. Acad. Sci. USA 1997, 94, 8948-8953

Schottland, P.; Zong, K.; Gaupp, C. L.; Thompson, B. C.; Thomas, C. A.; Giurgiu, I.; Hick- man, R.; Abboud, K. A.; Reynolds, J. R. Macromolecules 2000, 33(19), 7051-7061

Selampinar, F.; Akbulut, U.; Ozden, M. Y.; Toppare, L. Biomaterials, 1997, 18(17), 1163- 1168

Sheldrick, G. M., Acta Crystallogr., Sect. A, 1990, 46, 467

Sheldrick, G. M., SHELXL-96, University of Gottingen, 1996

Sheppard,Jr.,N.F.;Tucker,R.F.;Wu,C.Anal. Chem. 1993, 65(9), 1199-1202 224

Smith, R. C.; Fischer, W. M.; Gin, D. L. J. Am Chem. Soc. 1997, 119, 4092-4093

Sotzing, G. A., Reynolds, J. R. Adv. Mater. 1997, 9, 795-798

Sotzing, G. A.; Thomas, C. A.; Reynolds, J. R.; Steel, P. J. Macromolecules, 1998, 31, 3750-3752

Srinivasan, S.; Manko, D. J.; Koch, H.; Enayetullah, M. A.; Appleby, J. J. Power Sources 1990

Street, G. B. Handbook of Conducting Polymers, Skotheim, T. A., Ed.; Dekker: New York, 1966; Vol 1. pp. 265-290

Su,W.P.;Schrieffer,J.R.;Heeger,A.J.Phys.Rev.Lett. 1979, 42, 1698

Tanaka, S.; Reynolds, J. R. J. Macro. Sci., Pure Appl. Chem. 1995, A32, 1049-1060

Thackeray, J. W.; White, H. S.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 5133-5140

Thomann, H.; Dalton, L. R.; Tomkiewicz, Y.; Shiren, N. S.; Clarke, T. C. Phys. Rev. Lett. 1983, 50, 533

Thomas, C. A.; Zong, K.; Schottland, P.; Reynolds, J. R. Adv. Mater. 2000 12(3), 222-225

Tourillon, G.; Garnier, F. J. Phys. Chem. 1983, 87, 2289

Uhrich, K. E.; Cannizzaro, S. M.; Langer, R. S.; Shakesheff, K. M. Chem. Rev. 1999, 99, 3181-3198

Valdes,L.B.Proc. Inst. Radio. Engs. 1954, 42, 420 vanderPauw,L.J.Philips, Res. Repts. 1961, 16, 187-195 vanderPauw,L.J.Philips. Res. Repts. 1958, 13,1-9

Vanden Bout, D. A.; Yip, W.-T.; Hu, D.; Fu, D.-K.; Swager, T. M.; Barbara, P. F. Science 1997, 277, 1074-1077

Ward,M.D.Phys. Electrochem. (Rubinstein, I. ed.) Dekker, New York, 1995,p.293

Welsh D. M.; Kumar A.; Meijer E. W.; Reynolds J. R. Adv. Mater. 1999, 11(16): 1379-1382

White, H. S.; Wrighton, M. S. J. Am. Chem. Soc. 1984, 106, 7389

Williams, M. E.; Crooker, J. C.; Pyati, R.; Lyons, L. J.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 10249 225

Williams, M. E.; Masui, H.; Long, J. W.; Malik, J.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 1997

Wong, J. Y.; Langer, R.; Ingber, D. E. Proc.Natl.Acad.Sci.USA1994, 91, 3201-3204

Wudl, F. J. Org. Chem 1984, 49, 3382

Wudl, F.; Kobayashi, M.; Heeger, A. J. J. Org. Chem. 1984, 49, 3382

Yamato, H.; Ohwa, M.; Wernet, W. J. Electroanal. Chem. 1995, 397, 163-170

Yannoni, C. S.; Clarke, T. C. Phys. Rev. Lett. 1983, 51, 1191

Yao,Y.;Lamba,J.J.S.;Tour,J.M.J. Am Chem. Soc. 1998, 120, 2805-2810

Yoshida, M.; Fujii, A.; Ohmori, Y. Yoshino, K. Appl. Phys. Lett. 1996, 69,734

Zhang, Q. T.; Tour, J. M. J. Am Chem. Soc. 1998 120, 5355-5362

Zinger, B.; Miller, L. L. J. Am. Chem. Soc. 1984, 106, 6861

Zotti, G.; Schiavon, G.; Zecchin, S. Synth. Met. 1996, 72, 275-281 BIOGRAPHICAL SKETCH

Christopher A. Thomas was born in Santa Barbara county, California in 1974 where he lived for four years. After an eight year hiatus in Austin, Texas, he returned to Southern

California to attend high school in Thousand Oaks and undergraduate school at UCLA.

After obtaining a B.S. in chemistry in 1996 he moved to North Central Florida to study with

John Reynolds at the University of Florida.

226