1 DIOXYPYRROLE BASED SUPERCAPACITORS FOR ENERGY STORAGE By MERVE ERTAS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Merve Ertas
3 To my mom and dad
4 ACKNOWLEDGMENTS I would lik e to express my sincere gratitude to my advisor Prof. John R. Reynolds. The time in graduate school, especially in a forei gn country, was one of th e most challenging, yet rewarding and life changing experiences I have yet to go through, and none of it would have been possible without the support and guidance of him. I admire him not only as a scientist but as a person. I feel very fortunate having him as my research advisor and de eply thankful for his attention on my scientific, prof essional and personal developm ent. I appreciate his complete confidence in my abilities as a chemist. He has al ways treated me with the utmost respect and I have learned plenty from him. I am grateful fo r his unique enthusiasm and optimism in science, his encouragement, his patience, his words of wisdom, and most importantly his genuine concern. I will forever be thankful for his presence in my life. I would like to give my special gratitude to Prof. Andrew G. Rinzler, who in reality has been a second advisor to me. For the last three years, he welcomed me in his laboratory, guided me, was always interested in and appreciative of my work and patiently answered all my questions. I have to admit that I feel extremel y privileged having coll aborated with such a passionate scientist. I would also like to express my appreciation to the rest of my supervisory committee members; Professors Kenneth B. Wage ner, David B. Tanner and Randolph S. Duran for the time they invested in read ing and discussing this document. Sincere gratitude goes to my undergraduate and masters thesis advisor Prof. Levent Toppare, who introduced me to the area of conducting polymers and encouraged me to pursue higher studies. He prepared me well for the cha llenge of studies abroad. Without the support and guidance of him, I would have not become a chemist. I would like to thank Cheryl Googins, Sara Klossner and Gena Borrero in the polymer office and Lori Clark in the gr aduate office for their help ge tting me through the complex
5 administrative work. They always made sure that everything ran smoothly and have taken care of many problems that never reach our ears. During my graduate career, I have had a pleas ure to meet and work side by side among some extraordinary people who had a majo r impact throughout the completion of this dissertation. I would like to th ank many past and current members of the Reynolds group for their time, advice, scientific help and laughs. I also thank the Rinzler group members for welcoming me to their lab and being always so he lpful. Many thanks go to Dr. Ryan Walczak for working closely with me on several of the projec ts presented in this dissertation, providing me Sticky-PF and all the scientific and non-scientific discussions and also fo r his friendship. It was always a pleasure to argue with him over our co ffee breaks. One person that I can definitely not omit is Rajib Das, who I thank for his resear ch collaboration, his contributions in collecting some of the data presented in Chapter 3 and the ti me he put on assisting me in preparation of the SWNT films. I would like to thank Tim Steck ler for providing me his precious DAD monomers presented in Chapter 4. I could not ask for a be tter colleague and friend to go together through the different times in graduate school. I also acknowledge Karen Kelly for her expertise and for assisting me through the SEM an alysis. Special thanks go to Dr. Avni Argun, who was my mentor in the lab during the first year in gra duate school. I extend my th anks to Dr. Christophe Grenier for his support during the first year and Dr. Jeremiah Mwaura for his help during my qualifying exam. I also offer thanks to Dr. Svetlana Vasilyeva, my successor as Faraday Cage Coordinator, for taking over all the lab duties and her immediate help when I needed it during the last year. My time in Gainesville would not have been the same without my dear friends Dr. Josh Mcclellan, Dr. Kornelia Matloka, Dr. Sophie Be rnard, Dr. Florence Courchay and Dr. Piotr
6 Matloka. They have accepted me for who I am, and have always been there for me through my ups and downs. I will never forget all the wonderfu l time we spent together. I also thank Jordan Mathias for all the good memories. I would like to give special thanks to Laura Moody, especially for her help in proof reading this document and for being such a great friend through the good and the difficult times. Enough thanks cannot be expressed to Dr. Olga Zolotarskaya for being such a wonderful friend whom I was very lucky to meet during the last chapter of Gainesville. I would also like to thank my friends from Turkey for being never more than a phone call away whenever I needed them. I am particularly thankful to Dr. Zeynep & Umut Sargut for their continuous friendship, being th e best support a person co uld ask, and making me feel close to home throughout all the years in Gainesville. They were always there for me in the worst and the best of times and truly hold a specia l place in my heart. I would like to thank Dr. Scott Selph, for his support no matter how unbearab le I get, especially during the incredibly stressful writing process. In th e last two years of my graduate career, his constant love and enthusiasm kept me going and got me through some rough times. Even far away he always managed to keep my spirit up when I needed it the mo st. I feel very lucky to have him in my life. Last and foremost, I would like to thank my pa rents Tomris and Prof. Arif Ertas for their eternal love, understanding, and support. They taught me how to work hard to pursue my dreams and have always supported all my decisions even at younger ages. Their courage and faith in me are the reasons that I became the person I am t oday. Words could never de scribe the respect, the love, and the gratitude I feel. I also thank my dear brother Emre Ertas for his support through the difficult times and cheerful conversations on the phone I would have never come this far without them.
7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF TABLES ...........................................................................................................................9LIST OF FIGURES .......................................................................................................................10ABSTRACT ...................................................................................................................... .............14CHAPTER 1 INTRODUCTION .................................................................................................................. 16Historical Background ............................................................................................................16Operation Fundamentals .........................................................................................................18Comparison of Energy Storage in Capacitors and Batteries ........................................... 19Electric Double-Layer Capacitance .................................................................................21Pseudocapacitance ........................................................................................................... 23Adsorption pseudocapacitance ................................................................................. 24Redox pseudocapacitance ........................................................................................26Electrode Materials .................................................................................................................27Carbon ........................................................................................................................ .....27Activated carbons .....................................................................................................29Carbon aerogels ........................................................................................................30Carbon nanotubes .....................................................................................................30Metal Oxides ...................................................................................................................31Conducting Polymers ...................................................................................................... 32Type I & II supercapacitors ...................................................................................... 35Type III & IV supercapacitors ................................................................................. 37Review of conducting polymer base d supercapacitor literature .............................. 37Applications and Current Industry ..........................................................................................39Structure of Dissertation .........................................................................................................412 EXPRIMENTAL TECHNIQUES ..........................................................................................43Chemicals and Materials .........................................................................................................43Inert Atmosphere Handling .................................................................................................... 44Electrochemical Methods ....................................................................................................... 45Cyclic Voltammetry ........................................................................................................46Constant Potential Method: Chronoc oulometry / Chronoamperometry ......................... 48Constant Current Method: Chronopotentiometry ............................................................ 49Electropolymerization ......................................................................................................... ....49Electrochemistry of Electroactive Film .................................................................................. 51Supercapacitor Device Fabrication .........................................................................................52Energy and Power Characteristics of a Device ....................................................................... 55
8 Surface Characterization Techniques .....................................................................................56Scanning Electron Microscopy ........................................................................................57Atomic Force Microscopy ............................................................................................... 573 PProDOP BASED TYPE I SUPERCAPACITORS ............................................................... 60PProDOP as a Charge Storage Electrode Material .................................................................60Electropolymerization of ProDOP ..........................................................................................62Electrochemistry of PProDOP Films ......................................................................................66Surface Analysis of PProDOP ................................................................................................69Capacitances of the PProDOP Films on Gold Electrodes ...................................................... 72Type I Supercapacitor Devices with Gold Substrate .............................................................. 76Stability of Type I Supercapacito r Devices with Gold Substrate ........................................... 80SWNT Films as Electrode Substrates ..................................................................................... 83SWNTs Film Preparation ....................................................................................................... 84Comparison of Gold and SWNTs Subs trates in Device Performances ..................................86Non-Covalent Modification of SW NT Surfaces with Sticky-PF ........................................... 89Electropolymerization of ProDOP on Bare and Sticky-PF coated SWNT Films ..................94Electrochemistry of PProDOP Film on Sticky-PF coated SWNTs Film ............................... 97Surface Analysis of PProDOP Film on Sticky-PF Coated SWNTs Film ............................. 102Type I Supercapacitor Devices with Sticky-PF Coated SWNT Substrates ..........................104Stability of Type I Super capacitors with S ticky-PF coated SWNTs Substrate .................... 105Conclusions and Perspective ................................................................................................1084 TYPE IV SUPERCAPACITORS: Donor-Acceptor-Donor System s ...................................110Introduction .................................................................................................................. .........110Cyclic Voltammetric Deposition of DAD Systems ..............................................................113Electrochemistry of P(DAD) Films ...................................................................................... 116Capacitances of P(DAD) Films ............................................................................................121Type IV Supercapacitors ......................................................................................................127Conclusions and Perspective ................................................................................................1365 HYBRID SUPERCAPACITORS: Ru thenium Oxides|PProDOP ........................................137Introduction .................................................................................................................. .........137Cyclic Voltammetric Deposition of Hydrous Ruthenium Oxide ......................................... 139Characterization of RuOxnH2O and Composite Films of RuOxnH2O with PProDOP ....... 142Electrochemical Characterization ..................................................................................142SEM Analysis ................................................................................................................151XPS Analysis .................................................................................................................152Characterization of the RuOxnH2O|PProDOP Supercapacitor ............................................154Conclusions and Perspective ................................................................................................157LIST OF REFERENCES .............................................................................................................159BIOGRAPHICAL SKETCH .......................................................................................................170
9 LIST OF TABLES Table page 3-1 Capacitances, energy and power densities of PProDOP f ilm s and devices. .................... 1054-1 Charge densities and capacita nces of PBEDOT-BBT, PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 with different deposition cycles. ........................................................ 1254-2 Charge densities and capacitances of devices prepared by PBEDOT-BBT, PBEDOTTQ-Me2 and PBEDOT-TQ-Hx2 with different amount of PProDOP films. .................... 1345-1 Capacitances, energy and power densities of the films and devices of RuOxnH2O|PProDOP composite. ................................................................................... 157
10 LIST OF FIGURES Figure page 1-1 Ragone plot of various energy storage devices. ................................................................. 181-2 Schematic of a electrica l double-layer supercapacitor. ......................................................221-3 Schematic diagram of a supercap acitor utilizing conducting polymers. ........................... 351-4 Classification of conjugated polymers in Type I to Type III suparcapacitor. ................... 363-1 Structure of poly(3,4-propylenedioxy pyrrole) with its electropolymerization mechanism. .................................................................................................................... ....623-2 Electrochemical polymerization of Pro DOP in TBAP/PC on Au Button, in LiBTI/PC on Au Button and in LiBTI/PC on Au/Kapton. .................................................................653-3 Cyclic voltammograms a nd capacitances as a function of applied potential of PProDOP films on Au/Kapton substrates .......................................................................... 683-4 Scanning electron micrographs of PProDOP films on Au/Kapton deposited in LiBTI/PC............................................................................................................................703-5 Scanning electron micrographs of PPro DOP films on Au/Kapton deposited in TBAP/PC. ...................................................................................................................... ....713-6 Calibration curves: Deposition charges and capacitances of PProDOP as a function of the mass of PProDOP on Au/Kapton substrates. ........................................................... 733-7 Schematic diagram of Type I supercapaci tor configuration utilizing PProDOP and photograph of Type I PProDOP supercap acitor with Au/Kap ton substrate. ..................... 773-8 Cyclic voltammograms and cap acitances as a function of ap plied potentials of Type I PProDOP supercapacitor with Au/Kapton substrates. ....................................................... 783-9 Stability of Type I PProDOP super capacitors with Au/Kapton substrates. ....................... 823-10 Fabrication of SWNT Film.. .............................................................................................. 853-11 Atomic force micrograph of SWNTs film ......................................................................... 863-12 Cyclic voltammograms and capacitances as a function of applied potential of Type I PProDOP supercapacitor with SWNTs substrates. ............................................................ 873-13 Constant current charging/discharging of Type I PProDOP supercapacitor with Au/Kapton and SWNTs film substrates. ............................................................................ 89
11 3-14 Structure of poly(9,9-dioctylflouorene) (Sticky-P F) and its proposed non-covalent association with a SWNT surface. ..................................................................................... 913-15 Atomic force micrographs of SWNTs film before and after Sticky-PF coating. .............. 933-16 Absorbance spectra of Sticky-PF coated SW NT film and solution of Sticky-PF in chloroform..................................................................................................................... .....933-17 Schematic representation of PProDOP electrodeposition onto Sticky-PF|SWNTs film. ......................................................................................................................... ...........943-18 Electrochemical polymerization of Pr oDOP in LiBTI/ACN on Sticky-PF|SWNTs and bare SWNTs ................................................................................................................953-19 Cyclic voltammograms of PProDOP on Sticky-PF|SWNTs and bare StickyPF|SWNTs film in LiBTI/ACN. ........................................................................................ 993-20 Calibration curves: Deposition charges and capacitances of PProDOP as a function of the mass of PProDOP on Sticky-PF|SWNTs substrates. .............................................1013-21 Cyclic voltammograms and cap acitances as a function of applied potential of PProDOP films on StickyPF|SWNTs in LiBTI/ACN ....................................................1023-22 Scanning electron micrographs of PProDOP films on Sticky-PF|SWNTs substrates. .... 1033-23 Cyclic voltammograms and capac itances as a function of applied potential of Type I PProDOP supercapacitor with StickyPF|SWNTs substrates using LiBTI gel electrolyte. .................................................................................................................. ......1043-24 Stability of Type I PProDOP supercap acitors with Sticky-PF|SWNTs substrates using LiBTI gel electrolyte ..............................................................................................1074-1 Structures of donor-acceptor-donor E DOT-benzobisthiadiazole and thiadiazolequinoxaline monomers ..................................................................................................... 1124-2 Electrochemical polymerization of BEDOT-BBT on Au button in TBAP/ACN and on Au/Kapton in TBAP/PC .............................................................................................1144-3 Electrochemical polymerization of BEDOT-TQ-Me2 and BEDOT-TQ-Hx2 on Au button in TBAP/ACN. ..................................................................................................... 1154-4 Anion radical structures demonstrati ng the hypothetical mechanism of the reductive doping of PBEDOT-BBT. ...............................................................................................1174-5 Reductive cyclic voltammograms of PBEDOT-BBT PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 on Au button. 1194-6 Cyclic voltammograms, capacita nces as a function of app lied potential and charge densities as a function of time of varying amounts of PBEDOT-BBT. ........................... 122
12 4-7 Cyclic voltammogra ms, capacita nces as a function of app lied potential and charge densities as a function of time of varying amounts of PBEDOT-TQ-Me2 ......................1234-8 Cyclic voltammograms, capacita nces as a function of app lied potential and charge densities as a function of time of varying amounts of PBEDOT-TQ-Hx2 ......................1244-9 Stability of n-doping of PBEDOT-BBT, PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 in TBAP/ACN ..................................................................................................................1274-10 Expanding cell voltage of Type IV supercapacitor by combining PProDOP with PBEDOT-TQ-Me2. .......................................................................................................... 1294-11 Cyclic voltammograms, cap acitances as a func tion of applied potential and charge densities as a function of time of PB EDOT-BBT and corresponding devices. ............... 1304-12 Cyclic voltammograms, cap acitances as a func tion of applied potential and charge densities as a function of time of PBEDOT-TQ-Me2 and corresponding devices .......... 1314-13 Cyclic voltammograms, cap acitances as a func tion of applied potential and charge densities as a function of time of PBEDOT-TQ-Hx2 and corresponding devices. .......... 1324-14 Stabilities of Type IV supercap acitors of PBEDOT-BBT, PBEDOT-TQ -Me2 and PBEDOT-TQ-Hx2 in TBAP/ACN. .................................................................................. 1355-1 Elecctrochemical deposition of RuOxnH2O on bare SWNTs film and PProDOP|Sticky-PF|SWNTs substrate ............................................................................ 1405-2 Peak current densities of RuOxnH2O deposition on bare SWNTs film and PProDOP|Sticky-PF|SWNTs substrate ............................................................................ 1425-3 Cyclic voltammograms a nd capacitances as a function of applied voltage of PProDOP on Sticky-PF|SWNTs in LiBTI/ACN and in 0.5M H2SO4 .............................1435-4 Cyclic voltammograms a nd capacitances as a function of applied potential of RuOxnH2O on bare SWNTs film and PProDOP|Sticky-PF|SWNTs substrates ............. 1445-5 Cyclic voltammograms and charge dens ities as a function of time of PProDOP before the RuOxnH2O deposition and the composite films ............................................1475-6 Capacitances of the composite as a f unction of the initial cap acitance of PProDOP ...... 1485-7 Capacitances of the RuOxnH2O as a function of the initia l capacitance of PProDOP ... 1495-8 Scanning electron micrographs of RuOxnH2O on bare SWNTs film and PProDOP|Sticky-PF|SWNTs substrates. ......................................................................... 1515-9 XPS survey scans of RuOxnH2O|PProDOP|Sticky-PF|SWNTs film and PProDOP|Sticky-PF|SWNTs films. .................................................................................153
13 5-10 Cyclic voltammograms and capacitances as a fu nction of applied potential of TYPE I Hybrid RuOxnH2O|PProDOP supercapacitor with Sticky-PF|SWNTs substrates ..........154
14 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DIOXYPYRROLE BASED SUPERCAPACITORS FOR ENERGY STORAGE By Merve Ertas August 2009 Chair: John R. Reynolds Major: Chemistry This work details the app lication of conjugated polymer s, particularly poly(3,4propylenedioxypyrrole) (PProDOP), to supercapacitor devices. The oxidation and reduction (redox) processes in electroactive polymers make it possible to use them as charge storage materials. Within this work, a detailed elec trochemical and morphologi cal characterization of PProDOP on varying electrode substrates was conducted. In the first study, devices based on PProDOP on Au electrodes were introduced with outstanding long-term stabilities (~20% loss after 37,000 switches). The use of single walled carbon nanotube (SWNT) films with porous 3D network structures were used, fo r the first time, to realize the interpenetration polymer into the SWNT matrix it, leading more material loading per unit area in the supercapacitor. A novel electrode design created by non-c ovalent modification of the su rface of SWNT films with a pyrene functionalized polyfluorene has been de monstrated as a compatiblizer between the nanotube film surfaces and conducting polymers. A significant areal capacitance improvement (3 fold) has been observed with electrodes c ontaining porous substrates over conventional nonporous flat metalized substrates Any type of supercapacitor de vice utilizing varying charge storage materials can benefit from this method.
15 The key to achieving higher energy densities in an electrochemical supercapacitor is to expand the cell voltage as the ener gy stored scales with the square of this variable. For this purpose, a new family of n-dopable electroac tive polymers, based on a donor-acceptor-donor design along with PProDOP were util ized in Type IV supercapacitors. The operating voltage ranges of the devices were enhanced up to 3V. A novel device design utilizing a composite of conducting polymer and hydrous ruthenium oxide as the charge storage material in Type I supercapacitors was introduced. A two step all electrochemical method for preparation of in terpenetrating PProDOP and hydrous ruthenium oxide composite was demonstrated. This method allows 100% increase in capacitance per unit area, of ruthenium oxide deposited on PProDOP compared to bare SWNT films. Capacitance contributions of each of the composite material to the total capacitance were also confirmed.
16 CHAPTER 1 INTRODUCTION Historical Background Electrical ch arge storage is based upon principles that were discovered in the mideighteenth century, with th e observation of phenomena associated with static electricity and is hence about 200 years old. The discovery of the Leyden Jar by Musschenbroek in 1745 represented the first cap acitor and was of fundamental import ance in the study of electricity.1 Although electrostatic capacitors have been widely used in energy storage for nearly a century, their low capacitance values have traditionally li mited their applications. The concept of storing relatively large quantities of elect rical energy (compared to the energy density of batteries) at the interface between a metal and an electrolytic solution in reason ably small capacitors was first proposed in 1957 by Becker at General Electric when a patent was issued for an electrolytic capacitor using porous carbon electrodes.2 Although the mechanism was unknown at that time, the device exhibited exceptionally high capacitance and it was mentioned that the energy was being stored in the pores of the carbon. In 1966, researchers at The Standard Oil Com pany, (SOHIO) rediscovered the same effect using the double-layer capacitance of high area carbon materials in non-aqueous electrolyt e solution and patented a capacito r consisting of two layers of activated charcoal separated by a thin porous insulator where the energy stored in the double layer.3 This basic design remains the basis of today s modern electric doubl e-layer capacitors. In 1970, SOHIO patented a disc-shaped capacitor utili zing a carbon paste soaked in electrolyte; however, unsuccessful sales led SOHIO to license their technology to the Nippon Electric Company (NEC).4 Eight years later, NEC, produced th e first commercially successful doublelayer capacitors designed to provide backup power for maintaining computer memory. The term supercapacitor was the trade na me of this device, and conti nues to being common usage today.5
17 A different principle was proposed and de veloped between 1975 and 1981 by Conway in Ottowa.6 In this proposal, the co ncept of energy storage was based on pseudocapacitance associated by electrochemical adsorption of H or monolayer levels of electrodeposition of some base metals (Pb, Bi, Cu) at Pt or Au.7 Solid oxide redox systems (RuOx) were also utilized as pseudocapacitance materials and were found to approach almost ideal capacitive behavior.8,9 This work was continued by the Pinnacle Resear ch Institute (PRI) and the first high-power double-layer capacitors, known as ultracapacitors were pr oduced for the US military applications including laser weaponr y and missile guidance systems. These developments toward the use of hi gh surface area carbon, or metal oxide redox systems, continued as the market expanded, and by 1980s numerous manufacturers were commercially producing high-capacitance electrochem ical capacitors. Examples include The Gold Capacitor developed by Panasonic (Matsush ita Electric Industrial Co., Osaka, Japan) and the double-layer capacitors produced by ELNA under the name Dynacap. This was followed by a study by the United States Department of En ergy (DoE) to explore th e use of capacitors in the context of hybrid electric vehicles, and by 1992 the DoE Ultracapacitor Development Program was underway at Maxwell Laboratories.10 In 2005, the supercapacitor market was estimated between US $270 to $400 million11 and continues to grow through improved performance and drastic reduction in cost. Today, various supercapacitors are being manufactured worldwide for applications that require rapid rechargi ng, high power output, and repetitive cycling. The high capacitance systems described have be en labeled with a variety of names, many of which are unique to the particular manufacturer. Recently, however, the technical literature appears to have agreed upon a universal term pr oposed by Burke: electrochemical capacitors,
18 to refer to systems based on the double layer or pseudo-capacitances. Within this dissertation, the technology will also be referred as supercapacitor, a term will be used interchangeably with electrochemical capacitors. Operation Fundamentals Batteries, fuel cells and capacitors have many features in common, and are all based on electrochemical principles. The energy obtained from any energy storage and conversion device depends strongly not only on the device but also on the power output. The so-called Ragone plot shown in Fig 1-1, visuali zes the graphical comparison of the power density and energy density performance characteristics of various energy storage devices. 1, 12, 13 Figure 1-1. Ragone plot, in which the energy de nsity (Wh/kg) is plotted against the power density (W/kg), of various energy storage devices. [Reprinted from Electrochimica Acta, 45/15-16, R. Kotz and M. Carlen, Principles and applications of electrochemical capacitors, pages: 2483-2498, Copyright (2000), with permission from Elsevier] Batteries are typically high energy, low power devices, whereas conventional capacitors such as electrolytic capacitors or metalized film capacitors are higher power density devices with
19 limited energy storage capabilities. The quality of the energy obtai ned from conventional capacitors is generally poor due to variation of the voltage deliv ered with the state of discharge, whereas batteries tend to have a fairly constant output voltage. Fuel cells can have high energy storage, but their power output is limited. As demonstrated in Fig 1-1, electrochemical capacitors properties fill the gap in energy storag e devices capabilities that was previously vacant. This is important as, electrochemical cap acitors have the ability to store more energy than conventional capacitors; yet, they are able to deliver more power than batteries. When used in combination, these devices can improve battery performances, as well as capacitor performances, in terms of power dens ity or energy density, respectively. Comparison of Energy Storage in Capacitors and Batteries There are tw o fundamental ways electrical energy can be stor ed: (i) directly, as negative and positive electric charges on the plates of a capacitor in electrostatic fashion and (ii) indirectly, as in batteries by Faradaic oxidation and reduction of the electrochemically active species. When ions, or electrical char ges, are released by the electr ochemically active species going under redox process in a battery, th ey migrate from one electrode into the electrolyte, and a corresponding number of ions from the electrol yte are stored through another redox combination in the other electrode. The transfer and storage of ions does the electric work, while the current and resulting electrical potential produce the voltage difference across the poles of battery cells. Accordingly, the energy storage capability of a battery at a given electrochemical potential is directly proportional to the number of ions that can be absorbed in the electrodes. The useable energy stored electroche mically in a battery is given as QV where V is the voltage of the cell and Q is the cell charge capacity, the electri cal charge transferred to the load during the chemical reaction. During discharging of an ideal battery ce ll exhibiting a Nernstian
20 cell potential, progressively increasing charge is being added at a relatively constant voltage which remains ideally constant until all reac tant materials have been electrochemically consumed. On the other hand, the energy storage phenomen on in a capacitor is based on charge separation, which is quite different than in batteries. The simplest conventional capacitors store energy in the form of electrical charge in a thin layer of diel ectric material supported by metal plates that act as the terminals for the device. The current is a result of electron accumulation on one electrode and depletion on the complementary el ectrode rather than ion transfer. Positive and negative electrostatic charges physica lly reside at the surf ace of the plates. The electrical field is what generates the voltage, rather th an an electrochemical potential. The capacitance C is a measure of the energy stor age capability expressed in Farads (F). It is given by the following Equation 1-1, which relates the amount of charge th at can be stored in relation to the strength of the applied potential; C = Q/V = 0 A/d (1-1) where V (V) is the voltage difference between the terminal plates and Q (Coulombs) is the charge accumulated on each plate. As it is also shown in relation to the geometrical dimensions of a capacitor, the capacitance de pends on the dielectric constant, and the thickness, d of the dielectric material or the separation distance between parallel plates and its geometric area, A where 0 is the permittivity of free-space (8.854x10-12 F/m). Often in the cases of double-layer and pseudo-capacitance, C is not constant with changing V therefore a differential capacitance is defined as C = dq/dV. The charge density of a capacitor is described by Q/A and the maximum voltage that can be achieved by a dielect ric capacitor is dependent on the breakdown
21 characteristics of the dielectric material. The stor able energy in a capacitor charged to a potential difference of V can be calculated as E = CV2 = QV (1-2) for an accumulated charge Q residing on plates of the capacitor. During charging of a capacitor, work is being continuously done against the charges being accumulated on the electrodes, therefore the voltage progressively rises in th e charging process. This phenomenon explains the fact that the energy stored in the capacitor is half the value of the equivalent charge stored in a battery where energy scales as QV For a usual separation distance of 10-4m and a dielectric materials having 100 (water = 78 at 298K), the capacitances ar e very small, on the order of pF calculated due to the Equation 1-1.14 It is convenient to discuss the energy storage mechanism in electrochemical capacitor in terms of double-layer capacitance and pseudocapacitance separately due to the fundamental differen ces between these two principal types of supercapacitors. Electric Double-Layer Capacitance Sim ilar to conventional electr ostatic capacitors, charge st orage in electric double layer capacitors (EDLCs) is largely el ectrostatic in nature. Instead of charges accumulating on two conductors separated by a dielectri c as in conventional capacitors, the charge accumulates in the double-layer formed at the interface between th e solid electrode material surface and an electrolytic solution in the micropores of the electrodes.15 A schematic of an EDLC is shown in Fig 1-2. When an electric potential is applied acr oss the electrodes, one layer forms on the charged electrode, and the other la yer comprised of ions forms in the electrolyte giving rise to the double layer formation. The thickness of double la yer or the separation of charges is very
22 small, on the order of several Angstroms. An estimate of the specific capacitance of such a double-layer can be obtained using Equation 1-3. C/4 (1-3) where C is the capacitance, A is the surface area, is the relative dielectric constant of the medium between the two layers (the electrolyte), and is the distance between the two layers (the distance from the electrode surf ace to the centre of the ion layer).4 In terms of the charges q that are accumulating across electrode electrolyte interphase to extents the potential difference built up across the interphase, V, the double layer capacitance can be expressed as C=q/V or d(q)/d(V) Figure 1-2. Schematic of a electrical double-laye r supercapacitor. [Adapted from Journal of Power Sources, 91/1, A. Burke, Ultracapacitors: why, how, and where is the technology, pages 37-50, Copyright ( 2000), with permission from Elsevier] EDLCs are constructed by two electrodes immersed in an electrolyte with an ion permeable separator placed between the electrodes in order to prev ent electrical contact, but still allow ions from the electrolyte to pass through. The ions in the pores of the active material create the double layer by diffusion. As a result, at each electrode/electrolyte interface, there is
23 one double-layer present. In additi on to the capacitance that arises from the separation of charge, sometimes a small Faradaic contribution to th e double-layer capacitan ce can be made from electrochemical surface reactions (e.g. revers ible surface quinonoid-type redox reactions or chemisorption processes) that can oc cur on the surface of the carbon electrode.14 Since the thickness of the double layer is very small, a fraction of a nm, capacitance values of 15 F/cm2 are possible with the EDLCs. Moreover, by utilizing high surface area of materials (~1000 m2/g), large capacitances of 150 F/g are practically attainable.16-21 Since the double layer capacitance does not involve any phase transition mechanism as in batteries, charging/discharging is highly reve rsible, rapid and hundreds of thous ands of cycles are typically achievable with a given element of charge being ab le to be admitted or withdrawn at virtually the same potential.14 Pseudocapacitance Pseudocapacitance arises from Faradaic charge transfer process o ccurring at the surface or in the bu lk near the surface of the electrode materi al to an extent limited by a finite quantity of reagent or of available surface. The potential at wh ich charge is being passed is a function of the charge accumulated. Unlike ideal Nerstian behavior in batteries, where a constant electrode potential is independent of the ex tent of the reaction, in the pseudocapacitance process, the extent of Faradaically admitted charge depends linearly, or approxim ately linearly, on the applied voltage. In other words, the charge transfer leading to passage of extent of charge, Q is voltage dependent. Consequently, a capacitive phenom enon occu rs which is not electr ostatic in origin as in double-layer capacitance, and is denoted as pseudo in order to differentiate it from electrostatic capacitance. The sy stems utilizes pseudocapacitan ce provide a transition between electrochemical battery-cell a nd double layer electrostatic ch arge and energy storage.
24 Pseudocapacitance occurs when the extent of reaction, Q is some continuous function of potential, so that a derivative dQ/dV arises that has the properties of a capacitance. There are two types of electrochemical reactions that can involve a charge tran sfer that is voltage dependent, giving rise to pseudocapacitance. These include adsorption pseudocapacitance, resulting from a process of progressive occupation of surface sites on an electrode by ions from the electrolyte, and also from redox pseudocapacitance associated with the redox processes involving ions from the electrolyte where the electrode potential E of a redox system is a logarithmic function of the ratio of activities of the oxidized and reduced species. Both of th ese two processes are principally surface mechanisms; hence are highly dependent on the surface area of th e electrode material. The redox pseudocapacitance can also arise from the oxidation or reduction followed by charge neutralization of active conducting polymer material in the el ectrode which is more of a bulk process. Although the surface area of these materi als is not that critical for high capacitance values, relatively high surface area with micropores is required for the insertion and expulsion of ions from the electrolyte to the matrix. All the processes leading to an experimentally accessible pseudocapacitance require high electronic conductivity of electr odes in order to distribute and collect the electron curr ent. Both adsorption and redox pseudocapacitances will be explained in more detail in the following sections. Adsorption pseudocapacitance Two-dimensional surface depositi on of ions to form a monolay er on the electrode substrate is the best example of reversible process that re sults in a Faradaic charge transfer, and is hence associated with pseudocapacitance. These proces ses can give rise to almost one hundred times greater capacitance than the double-layer capacitance at the same electrode. Reversible deposition/desorption of H at Pt, Rb and Ir, or deposition/desorption of Pb and Bi atoms on Au
25 or Ag are good examples of two-dimensional Fara daic electrode processes. These adsorption and desorption process can be expressed as follows: where S is the electrode substrate (usually a noble metal), C is the concentra tion of the ionic species A 1A is the fractional free surface area avai lable for adsorption at a coverage, A. 22 AA (1-4) The fractional coverage can be determined by using Equation 1-4, for an equilibrium situation at any potential where it assumed that sites are occupied randomly in a fixed lattice; where K is an electrochemical equilibrium constant. This relation shows the dependence of the extent of fractional coverage proportional to charge passed, on the electrode potential, V. Since is proportional to charge passed and is a function of V, then d /dV is proportional to corresponding capacitance. If Q1 is the amount of charge required for formation or dispersion a complete monolayer, then a ps eudocapacitance can be derived by differentiating Equation 4 with respect to V giving;14 (1-5) According to Equation 1-5 and cyclic vo ltammetry experimental results, these adsorption/desorption process are highly reversible with respect to changes of scan rate ( dV/dt ) from negative to positive values, giving mirror-image voltammograms. Since only twodimensional arrangement are involved in adso rption/desorption process without any phase change or reconstruction mechanisms, the capaci tance versus voltage pr ofile is stable and
26 repeatable over thousands of cycles in a clean solution and can be maintained up to quite high sweep rates of 100 V/s. Redox pseudocapacitance A general redox reaction involving a ny oxidized (ox) and reduced ( red ) s pecies can be simply expressed as ox + z ered and the equilibrium redox potential of this reaction is given by the Nernst equation, (1-6) where E is the standard potential, R is the gas constant, T is the absolute temperature, F is the Faraday constant, and ox and red are the concentrations or the activities of oxidized and reduced species, respectively. The concentrations of the active species can also be expressed as their relative fractions, [ox]/([ox]+[red]) and [red]/([ox]+[red]) or 1-[ox]/([ox]+[red]) so that Nernst equation can be rewritten which is also shown in Equation 6. When is defined as [ox]/([ox]+[red]) the analogy between (/1-) and ( A/1A) in Equation 4, is obvious. Thus, redox pseudocap acitance is formally analogous to an electrochemical adsorption capacita nce with charge transfer. The amount of charge required to convert a given quantity of ox to red or vice versa, is given by the product of z and F and as it can be clearly seen from the Nernst equation, it is a function of the potential E. If Equation 6 is rearranged as the following, it takes the form of Equation 1-4. (1-7) Therefore, differentiation of Nernst equation produces a pseudocapacitive expression, having the same relation to pot ential as in the Equation 1-5.14
27 Similar pseudocapacitive behavior arises in the case of intercalation processes (e.g., lithium intercalation into layer-lattice host mate rials). The equation for these reactions has the same form as in Eq 1-4 and Eq 1-7 with the conversion of the logar ithmic portion including the three-dimensional site fraction occupancy by the intercalated guest atom or ion (e.g., Li+). Li+ batteries can be regarded in some sense as pse udocapacitors since their ex tents of charge are a continuous function of voltage, although their re sponse times in a capacitative sense are longer.23 Electrode Materials Carbon Since the early years of the supercapacitor de velopment, a substantial fraction of EDLCs have relied on carbonaceous materials, which rema in the most popular electrode material today. Carbon has been an attractive choice due to its lo w cost, availability, and long term history of use. Although carbon has now been used in EDLC s for over 20 years, the mechanisms of energy storage and the relationships between the quantif ied specific capacitances and the morphological properties of the different carbon materials have b een studied in depth only relatively recently. It is necessary for the carbon materials to have high surface areas, on the order of 1000 m2/g and good intra and inter par ticle conductivity in porous matrices to be optimal for todays state of the art supercapacitors.12, 13, 24, 25 Although surface area and capacitance can generally be increased by increasing porosity of the car bon materials, the specific capacitance is not necessarily directly proportional to the surface area of the electrode. 26 High specific capacitances are achievable with EDLCs due to their high el ectrode/electrolyte in terface area and a small charge layer separation of atomic dimensions. Charge is stored in the pores at or near the interface between the active carbon material and the electrolyte. If electrolyte ions can not access the pores, the surface area will not contribute to the specific capacitance. Therefore, the
28 electrolyte accessibility to intr apore surface area is what ultima tely influences the capacitance, rather than simply a high surface area. As a consequence, the way the surface is developed greatly influences the specific cap acitance value that can be obtained from carbon materials. The mobility of ions in the electroly te solution within the pores is significantly influenced by pore size. The key factor in designing car bon electrodes is to attain the best pore size related to the electrolyte ion size, which will maximize the speci fic capacitance. In this regard, many research groups have studied the ways to control porosity in the mesoporous range (2 to 10 nm) to maximize the capacitance. 27-30 In contrast to the traditional view of linear dependence of the capacitance on the pore size, recently it has been demonstrated th at even micropores of less than 2 nm, which are smaller than the size of solvated electrolyte i ons can contribute to the charge storage mechanism.31-34 This phenomena is explained by the di stortion of the solvation shell of the ions as they enter to the narrower pores (<1 nm) leading to closer distance of the ion to the carbon surface so thus enhanced capacitances are achieved. In addition to surface area and porosity, the conductivity of carbon electrodes is of substantial importance to the power density of EDLCs. The conductivity of the electrodes is inversely proportional to the particle size of the carbon materials. The electrodes with high surface area that are made of sma ller particles usually show increas ed resistances, and thus lower power densities. However, using larger particle sized mate rials will limit the specific energy due to the reduced surface area. In conclusion, a vital aspect in the fabrication of carbon electrodes for supercapacitors is to compromise between th e pore-size distribution to ensure easy access of electrolyte ions, or the specific surface area to ensure high capacitance, and the particle size to achieve good conductivity.35 In its more dispersed and conduc ting forms, carbon electrodes can take many different manufactured forms that can be utilized in EDLCs, such as foams, fibers,
29 and nanotubes, all of whic h are nanoporous material.36 EDLCs fabricated using such electrodes in conjunction with both aqueous and organic electrolytes have demonstrated capacitances between a few Farads to seve ral thousand Farads per cell.37 Several representative technologies available today are br iefly explained below. Activated carbons Activated carbons are the most extensively used material in the fabrication of EDLCs due to their high surface area and relatively low cost Traditionally, these high specific area carbons are in powder or fiber form which gives rise to good mechanical reli ability and electrical conductivity, both of which are useful for EDLCs applications. Activated charcoal is a powder made up of extremely small and very rough partic les, which in bulk form a low-density volume of particles with holes between them resembling a sponge. Treatment of carbon materials has a significan t effect on the structure of the el ectrode morphology in te rms of surface size and porosity. The capacitance varies de pending on the process used to prepare the carbon electrode. High surface area electrodes with the randomly sized porous networks are obtained by activation processes. The resulting overall accessible surface area of even a thin layer of activated carbon is in the range of 1000-2000 m/g which is many times greater than the geometrical surface area of the electrode, allowing many more el ectrons to be stored in any given volume. Activated carbon powders usually require binders to be processed into electrode films. These inactive components occupy a significant fraction of the total electrod e weight which results in low energy densities. The capacitances of these materials are about 100 F/g and 50 F/cm3 using inorganic electrolytes. In contrast to powders, activated carbon fabr ics can be used wit hout any binder addition. Although they have high surface areas compared to activated carbons and high electrical conductivity (200-1000 S/cm), the high cost restricts their use in EDLCs to very specific applications.
30 Carbon aerogels Besides activated carbo ns, recently nanostructured carbons, such as aerogels38, nanotubes17 and nanotemplates39 have been the focus of attention to be used as electrode materials in EDLCs. Aerogels are low-density suspen sions of carbon nanoparticles with in a gel in which the liquid component of the gel has been replaced with gas. A pyrolysis treatment in an inert atmosphere leads to a controlled and uniform particle and pore sized (mesoporous between 2 and 50 nm) carbon aerogels with a good electrical conductiv ity (several S/cm).35 Although, the usable surface areas of carbon aerogels (400 -900 m/g) are usually lower th an the activated carbons, the advantages of these materials are mainly their lo w ionic and electronic char ging resistance due to the ordered and interconnected pore structure, which leads to high specific power. Specific capacitances in the range of 50 -100 F/ g have been repor ted in the literature for these materials.36 Since the use of binding materials in the carbon based electrodes greatly reduces the conductivities, the fact that ae rogels can be used without bi nding material makes them an attractive supercapacitor electrodes material. In addition to power performance, capacitance and cyclability of EDLCs are improved by replac ing the activated carbons with carbon aerogel electrodes. Carbon nanotubes Carbon nanotubes have recen tly become a new choi ce of electrode material to be used in EDLCs due to their unique architecture and rema rkable characteristics such as high specific surface area, unique pore structure, excellent electrical conductivity and interconnectivity, chemical stability and low mass density.17, 40 They can be produced as single walled (SWNTs) or multi walled carbon nanotubes (MWNTs) depending on the synthesis route, and both are actively being researched as supercapacito rs electrode materials. Capacitance values from 20 to 180 F/g have been reported depending on the purity of carbon nanotubes and the electrolyte.21, 36, 41-46 To
31 improve these values, significant efforts have been focused on surface functionalization of nanotubes; however, the results are still not impressive and cycling stability is a problem.35 Recent trends in nanotube-based supercapacitor research involve the nano-engineering of microscopic forests of dense, nano-ordered, and vertically ali gned nanotubes to the current collector. Such carbon naotube ba sed electrodes, analogous to a paintbrush, as opposed to a sponge as in activated carbon electrodes, could help increase capacita nce by maximizing the accessible surface area through modi fication inter-tube distances.47-49 Metal Oxides Metal oxides are attractive electrode materials, and are utilized in supercapacitors due to their high specific capacita nce and low resistance. 12, 13, 24, 50, 51 Application of metal oxides in pseudocapacitance research has been started with ruthenium oxides. However, this material is a noble metal and far too expensive for many commerci al applications. Most of the early work on ruthenium oxides was carried out for military applications, where cost was of less concern than military capability. The charge storage process in ruthenium dioxide electrodes is surface limited and progresses in several one electron steps, resulting in broad range of capacitance values. Thus, a constant pseudocapacitan ce occurs over the full operating voltage range. Subsequent to crystalline ruthenium oxide, the use of hydrous ru thenium oxide as an electrode material was investigated and specific capacitance of 750 F/g was found with an aqueous electrolyte sulfuric acid solution.14 The US Army Research Lab has assembled prototype cells with an impressive energy density of 8.5 Wh/kg and a power density of 6 kW/kg.15 Since, high capacitance and fast charging resu lt from H adsorption, metal oxides electrodes must be used with acidic aqueous electrolytes to provide good proton co nductivity. This narrows the achievable cell voltage and moreover, most metal-oxides break down rapidly in acidic solutions. Although the high cost of ruthenium prev ents its large-scale use, the best charge
32 storage capability of any capaci tor reported to date have b een 840 F/g (200Wh/kg) and 500 mF/cm2 with hydrous ruthenium oxides in an aqueous electrolyte sulfuric acid solution. In addition to the high power densities and the al most constant capacitance over a wide voltage range, it shows an excellent reversibility with a cycle life over several hundred-thousand cycles.16, 52-58 Current academic efforts are focused on other lower-cost metal oxides to be utilized in more practical supercapacitors. R ecently, some materials used in batteries such as manganeseoxides and vanadium oxidies have attracted attention for applica tion in pseudocapacitors due to their lower cost and encouraging specific capacitances. The interest in manganese oxides for electr ochemical energy storage has shown a steady and continuing growth since 2000. Specific cap acitances of amorphous manganese oxides powders were reported as 200 F/g in aqueous electrolytes.59, 60 Although these capacitance values are lower than ruthenium-oxides, the intere st in manganese oxides for energy storage applications is driven by their low cost and low toxicity as a practical alternative to other metal oxides. Moreover, it has been shown that milder aqueous electrolyte solutions such as potassium chloride can be used with manganese-oxides. Va nadium oxides are another class of metal oxides that draw attention as a low-cost alternative to ruthenium oxides for supe rcapacitor applications. They exhibit a broad range of morphologies, on which the electrochemical characteristics are strongly dependent and their disc harging curves demonstrate t ypical capacitive-like behavior.61, 62 These materials are likely contenders for a number of future commercial supercapacitor applications. Conducting Polymers Another interesting class of m aterials that demonstrate pseudocapacitance with highly reversible behavior is the family of conducting polymers.14, 63-65 Conducting polymers store and
33 release charge through the redox process. Typically oxidizing or reducing and neutralization of these compounds is referred to as doping a nd dedoping, respectively. Co nducting polymers can be doped and dedoped rapidly to high charge densities. During the oxidation and reduction process, charge is removed from or transfe rred to the polymer, resp ectively creating electron deficiency (p-doping) or exce ss of electrons (n-doping) on the backbone as delocalized electrons. Consequently, the appropriate counter ions from the electrolyte enter into the polymer to com pensate the charge created on the backbone which is referred as doping. These two types of doping processes are shown below: p-doping: P + yA[Py+ yA-] + yen-doping: P + yC+ + ye[PyyC+] When neutralization occurs, th e ions are released back in to the electrolyte solution, referred to as dedoping, or neut ralization. Doping of conducti ng polymers takes place throughout the bulk volume of the material, ra ther than just at the outer surf ace of particles as is the other cases for electrochemical capacitances. Since the entire volume of the polymers is involved in the charge storage process, high levels of capacitive energy densities can be achieved with polymer films. Conducting polymers can be tailored to prov ide specific properties, such as high conductivity, flexible morphologies, wide voltage windows, high storage capacity, porosity, and reversibility. Moreover, conducting polymer materials can be easily and economically made in large amounts. Although, conducting polymer systems are less stable than ruthenium dioxides, due to the phenomenon of swelling and shri nking during doping and dedoping processes12; they still demonstrate stabilities over thousands of cycles with wide voltage ranges66, and are much lower-cost than ruthenium oxides. Another adva ntage of conducting poly mer systems is that their redox capacitance are not diffusion limited in terms of the oxidized and reduced species
34 since they are not solution species but they are physically anchored on the electrode surface. Thus, rapid charging and discharg ing giving high power densities is possible. In terms of satisfactory energy and power densi ties, cycle life, low weight and thermal stabilities with the predicted low material cost relati ve to the other materials that were traditionally used, conducting polymers represent an attractive material curren tly available to be us ed in electrochemical capacitors. Characteristically, conducting po lymers offer specific capacitances in the range of about 100-250 F/g.16, 67-72 As initial work in this field, a pr ototype polymer film capacitor with an energy density of 39 Wh/kg and a power density of 35 kW/kg has been demonstrated by the Los Alamos National Laboratory.63 Over the years, several superc apacitors have been fabricated using polyaniline, polythiophine, polypyrrole, and th eir derivatives with different electrolyte and substrate systems to achieve high ener gy storages with high capacitances. Conducting polymers were originally generalized using three schemes in which they can be utilized in electrochemical capacitors by Rudge et al. These were referred as Types I, II and III.73 In Fig 1-3, a general diagram of a super capacitor utilizing conducti ng polymers as anode and cathode electrodes is shown. Among these, Type I and II supercapacitors utilize only pdoping conducting polymers for both positive and negative electrodes. A Type I device is based on a symmetric configuration with the same p-dopable polymer on both electrodes, whereas a Type II device uses an asymmetric configuratio n with two different p-dopable polymers on each electrode. A Type III supercapacitor consists of a p-dopable polymer and an n-dopable polymer as the positive electrode and the negative electrod e, respectively. This original classification of application of conducting polymers into superc apacitors is based on the increasing operating voltage range and charge storage capacity of the devices. However, there has been an ambiguity
35 about symmetric and asymmetric Type III supercap acitor since Type III is the only term that incorporates an n-doping polymer.74 Hence, the term Type IV supercapacitors was later proposed to distinguish between symmetric an d asymmetric devices that use an n-doping polymer. Type III refers to a symmetric configura tion utilizing the same polymer that both pand n-dopable, while Type IV super capacitors are based on an asymmetric configuration that utilize one polymer as the p-doping and a different polymer for the n-doping on each electrode. Figure 1-3. Schematic diagram of the genera l construction of a s upercapacitor utilizing conducting polymers as anode and cathode electrodes. Type I & II supercapacitors For the construction of these devices, polym ers are initially set in the fully oxidized (pdoped) state and laminated with the fully neutralized forms to establish the initial charge state of the capacitor. With the discharge of the supe rcapacitor, neut ralization of the p-doped films proceeds with concurrent oxidation of the cathode films in the device. In Type II supercapacitors, the polymer with the higher oxidation potential is used as the anode and the polymer with the lower oxidation potential serves as the cathode. Si nce oxidation and reduction potential of the anode and cathode are equal in Type I, at most only half of the total charge capacity of the polymer is used.14, 75, 76 Type I devices usually have approxi mately 1Voperating potential ranges. As a consequence of the difference between oxidation and reduction poten tial of the anode and cathode films, in Type II supercap acitors, the cell potenti al range is higher than Type I devices (around 1.0 to 1.25 V) and only about 75 % of the total charge capacity of the polymers can be
36 utilized. Type II supercapacitors benefit from the increased cell voltage, which allows them to deliver more charge. In both Type I and II cas es, the cell voltages are limited by the intrinsic over-oxidation resistance of the polymers. The representation of the separate cycling of each electrode during charging of a Type I and II supercapacitors is shown in Fig 1-4a and b, as the anode oxidizes and the cathode neutralizes. Since there is either complete or partial overlap between the voltage range during oxidation/neutralization of the a node/cathode electrodes, the cycl ic voltammogram of an ideal Type I or II supercapacitor woul d be a rectangular plot, showing a constant current within the operating voltage window of the device.14, 75-78 Furthermore, under constant current conditions, the voltage outcome of the both devices would show a linear decay in an idealized case. Although Type I supercapacitors can be reverse bi ased because there is no polarity between the electrodes, a CV plot would resemble the original one. However, the charge capacity of the Type II devices will be much lower when reverse biased, and stability could be a problem since there is a polarity between the electrodes. Figure 1-4. Classification of conjugated polymers in th ree generalized configurations in an increasing charge storage capacity and operating potential range in a) Type I, b) Type II and c) Type III supercapacitors. [Adapted from Journal of Power Sources, 47/1-2, A. Rudge, J. Davey, I. Raistrick, S. Go ttesfeld and J. P. Ferraris, Conducting polymers as active materials in electrochem ical capacitors, pages: 89-107, Copyright (1994), with permission from Elsevier].
37 Type III & IV supercapacitors In Type III and IV configurati ons, polym ers are in the fully oxidized (p-doped) state on the anode and in the fully reduced (n-doped) form on the cathode electrode in the charged state of the device, while in the discharged state both polymers are in their neutral state. In Type III & IV configuration, the full to tal charge capacity of both anode and cathode polymers is used. Among the remaining polymer supercapacitor configurati ons, only the p-/n-type device configuration has the ability to outperform the conventional doubl e-layer carbon supercap acitor. The key to achieving higher energy density in an electroc hemical supercapacitor is to expand the cell voltage the energy of a capacitor scales with the square of the cell voltage (E=CV2) as has been described in the previous secti ons. With the p-/n-type design, the operating voltage range of the devices can be increased up to 3 to 4 V, as it is shown in the repr esentation of the charging cycle in Fig 1-4c. Hence the highest energy density can be obtained. Besides providing the broadest cell operating voltages, Type III&IV supercapacitors can de liver all the doping charge during discharge at high potentials; consequently high power densities can be demonstrated. Compared to Type I and II supercapacitors, under constant current conditions, the charge is delivered at a much higher voltage and voltage drops very rapidly after discharge. Review of conducting polymer based supercapacitor literature Over the years, the bulk of conducting polymer based supercapacitor research has focused on polyaniline, polypyrrole, polythiophine and their derivatives with diffe rent electrolyte and substrate systems, directed toward achievinghi gh energy storage with high capacitances. Most of this work have been evaluated for use in Type I supercapacitors.67, 68 69-72 Polyaniline: In addition to its usage in batteries79-81 polyaniline (PANI) is one of the most studied conducting polymers which is utilized in supercapacitors.82-87 Most PANI electrodes are prepared with electrochemically grown films, as chemically synthesized PANI presents
38 solubility and processability limitations. These polymers operate well in both aqueous and nonaqueous electrolytes.88 Very high doping levels of one el ectron per two monomeric units are achievable with PANI.89 Supercapacitors exhibiting speci fic capacitances from 107 F/g 85, 86 to 250 F/g84 have been reported for PANI. The PANI electrodes also demons trated stabilities over 20,000 cycles with 5% loss in electroactivity.82 Currently, the development of soluble polyaniline based supercapacitors were being inves tigated for military applications, with a focus on fast switching, improved cycle ti me and reduced production costs.90 Polypyrrole: Another conducting polymer; polypyrrole (PPy) which is also widely used for battery research, has received considerab le attention for supercapacitor applications.65, 73, 91-94 Specific capacitance values rangi ng from 40-84 F/g (100-200 F/cm3) were exhibited by PPy electrodes, using diffe rent electrolytes.82, 95-97 Similar to electrochemi cally deposited PANI, PPy works well in aqueous and non-aqueous electrolytes. PolyThiophene: Although very high capacitance valu es of 250 F/g were reported for polythiophene (PTh)87, because of its instability, only a lim ited number of research studies have focused on PThs for the energy storage applications. However, a derivative of PTh; poly(3,4ethylenedioxythiophene) (PEDOT) which is one class of PTh derivative, has been extensively studied for use in supercapacitor electrode, due to its excellent chemical stability and fast switching speeds. 78 PEDOT offers specific capacitances of 210 F/g98. Energy densities of 1-4 Wh/kg with power densities of 35-2500 W/ kg have been reported for PEDOTs based supercapacitors.78, 99 Moreover, devices that exhibit high char ge storage capabil ity with potential window of 2.15 V have been demonstrated with PEDOT derivatives.98 Relative to Type I supercapacitors, there ha ve been far fewer publications on Type II devices. The device performances are often reported as similar to Type I devices based on the
39 same polymers. The best systems reported include PPy/poly(3-methylthiophene)95, PPy/PANI devices with specific capacita nce values as high as 25 F/g100, and (PEDOT)/poly(3,4propylenedioxythiophene)77, 101. In attempts to obtain high energy and power densities, a number of research groups have focu sed on n-doping polymers for use in Type III and Type IV supercapacitors, allowing a wide voltage range for development of pseudocapacitance.73, 75, 76, 95, 98, 102-105 Although specific capacitances of 70-180 F/g ha ve been reached for n-doping of PTh derivatives unfortunately, many n-doping polymers suffer from poor cycle stability105. Energy and power densities in the ra nge of 6-39 Wh/kg and 1600-3500W/kg, respectively, were reported with these type of devices with cell voltages exce eding 3 V; however in every case cycle life was a problem. In order to eliminate the need for a stable n-doping polymer, several groups106-108 have tried to use carbonaceous electrodes as the cathode electr ode, referring these devices as hybrid supercapacitors. Another approach towards improved capacitance, energy, power and better stability is to use a combination of electroactive polymers with carboneous materials. Several groups have investigated electr oactive polymer based composites based on PPy85, 104, 109, 110, PANI85, 111-114, and PTh.87, 107 Applications and Current Industry Commercial activity in electrochemical capacitor technology development was quite limited until 1995, due to limitations in fabrication equipment and the lack of testing procedures and standards for these devices. However, the U.S. Department of Energy (DoE) has since stimulated the development of electrochemical capacitors for commercial and other purposes through research funding and sta ndardization of performance goa ls. With the increasing demand for new manufacturing techniques, novel applications with cha llenging power requirements, and the requirement for materials of high capacitance s and low resistance, world-wide academic and commercial efforts have been focused on the ma ny aspects of supercapacitors development.
40 As intermediate energy and power sources betw een batteries and conve ntional electrolytic capacitors, supercapacitors fill the gap in the energy storage technology. Their range of applications is broad, with pot ential uses varying from altern ative power sources in diverse electronic applications such as computer power backup, medical equipment, electronic fuses, burst power for military systems, to high power applications incl uding load leveling, electrical vehicles, space crafts, etc.12, 13, 15, 24, 97 One of the most important and large-scale appl ications of supercapacitors is load leveling in capacitor-battery hybridization for electric vehicle drive systems.115 This application arises from the recent push toward the production of el ectric cars, with current rechargeable battery systems unable to satisfy increased power demand s, faster charging time and lower costs. The high power capability of supercapacitors is predic ted to take the main load from the battery component during acceleration, wh ile regenerative braking systems are envisoned for partial recharging and overall energy economy. In addition to automotive app lications, industrial equipment such as cranes, fork-lifts, and elevator s, can gain energy efficiency from recovery of energy normally wasted during braking of repe titive motion by using supercapacitor hybridelectric power systems. Since the Li-ion and Nihydride batteries are approaching their limits for the higher power density demands, supercapacitors al so present as new charge storage systems in applications where weight is a concern, such as portable consumer electronic devices. 116, 117 Moreover, supercapacitors are becoming important energy and power sources in military applications such as aircrafts, missiles or the portable devices such as GPS locators and night vision goggles. There are a wide range of co mmercial electrochemical capacitor research and development activities in progress with different energy and power density requirements; from small
41 millifarad size devices up to several kilofarads size devices. When designing, fabricating and testing electrochemical supercapacitors, these power and energy density requirements of the intended application are important to take into account. Devices will be different designed for high power pulse delivery purposes su ch as electric vehicles, starti ng assist hybrid systems, than for small less demanding applications as co mputer memory backup, domestic electronic applications, the telecommunication systems. Parameters like power, energy, weight, volume, cycle life, cost, safety must be distinguished and assessed for diffe rent type of applications. For instance, heat management for high power chargi ng and discharging applications or provision for overcharge protection are important concerns for larger units.118 Numerous companies around the world curren tly manufacture supe rcapacitors in a commercial capacity. The major commercial research and development activities in this field are under progress mainly in the United States, Japan, and Europe. Structure of Dissertation The m ain characteristics of this work are the utilization of conjugate d polymers in different supercapacitor device designs. Among conducting polymers, poly(3,4-propylenedioxypyrrole) (ProDOP) has been the focus of this disserta tion. Chapter 2 mainly summarizes the background information of the experimental methods used throughout the dissertation and also details the supercapacitor device fabrication. The Type I supercapacitors prepared using PPr oDOP were investigated in Chapter 3.Two different current collector substr ates have been used in the de vice construction. A new concept to utilize porous 3D network stru ctured single walled nanotube (SWN T) film as the substrate in order to achieve higher amount of capacitances per unit volume, by more material loading, is demonstrated. Moreover, the novel technique of non-covalent modification of SWNT surfaces with a pyrene functionalized polyfluorene (Stick y-PF) is introduced. By this means, carbon
42 naotube film electrodes provides a significant areal capacitance improvement over conventional non-porous flat metalized substr ates. Any supercapacitor utiliz ing varying charge storage materials can benefit from this method. Chapter 4 details the Type IV supercapacitors utilizing PProDOP and a family of ndopable donor-acceptor-donor systems. It evaluate s the importance of n-dopa bility in order to increase the conjugated polymer ba sed supercapacitor cell voltages. Application of a composite of PProDOP and hydrous ruthenium oxide in a Type I supercapacitor configuration is demonstrated in Chapter 5. The unique electrode design allows interpenetration of metal oxide particles through PProDOP matrix. Ruthenium oxide loading per unit area were increased, accordingly, improved en ergy storage capabilities were realized with contributions from both PProDOP and ruthenium oxide.
43 CHAPTER 2 EXPRIMENTAL TECHNIQUES This chapter provides necessary background info rm ation of the experimental methods used throughout this dissertation. These te chniques will be frequently referred to in the subsequent chapters. Chemicals and Materials HPLC grade propylene carbonate (P C) and reag ent grade acetonitrile (A CN) in Sure Seal were purchased from Aldrich. ACN was distilled over calcium hydride before use. PC was percolated through type 3A activated molecular sieves (Aldrich), followed by fractional vacuum distillation. All solvents used on the bench were purged with Argon prior to use and were stored in Schlenk flasks under Argon. Lithium bis(trifluoromethanesulfonyl) imide Li(CF3SO2)2N) or (Li-BTI), was purchased from Aldrich, and dried under v acuum at 150C for 24 hours prior to use. Tetrabutylammonium perchlorate (TBAP) was either purchased from Aldrich or synthe sized by mixing a 1:1 mole ratio of tetrabutylammonium bromide dissolved in a minimal amount of water with perchloric acid. The resulting white precipitate was filtered, recr ystallized from a 1:1 molar ratio mixture of ethanol and water and dried under vacuum at 60C for 48 hours. Poly(methylmethacrylate) (PMMA) (Mw:996,000 g/mol) was purchased from Aldrich and was dried under vacuum at 50C for 12 h and stored under argon prior to use. The monomers 3,4-propylenedioxypyrrole (ProDOP),119 120 4,8-bis(2,3-dihydrothieno[3,4b][1,4]dioxin-5-yl)benzo[1,2c ;4,5c ]bis[1,2,5]thiadiazole (BEDOT-BBT)121, 4,9-bis(2,3dihydrothieno[3,4b][1,4]dioxin-5-yl)-6,7-dimethyl-[1,2,5]thiadiazolo[3,4-g]quinoxaline (BEDOT-TQ-Me2), 4,9-bis(2,3-d ihydrothieno[3,4b][1,4]dioxin-5-yl)-6,7-dihexyl[1,2,5]thiadiazolo[3,4g]quinoxaline (BEDOT-TQ-Hx2) whose structures are shown in Chapter 4
44 were synthesized according to th e previously reported literature. Poly (9, 9-dioctylflouorene) (Sticky-PF) was synthesized specifically for this study.122 Gold coated Kapton (100 nm gold layer on 1 mm Kapton sheets) used as an electrode substrate was purchased from Astral Technologies (product no: ATU1310) and cut into the desired size. Adhesive conducting copper tape was used to make electrical contacts to the gold electrodes and was purchased from 3M Elect ronics Department (3M EMI Copper Foil Shielding Tape 1194). Platinum wires and sheets and silver wires were purchased from Alfa Aesar. Gold and platinum button electrodes were purchased from Bioanalytical Systems, Inc. (BAS). Single walled nanotubes films (SWNT) were prepared according to methodologies previously reported41 and details will be explained in Chapter 3. Inert Atmosphere Handling Since n-dop ing of donor-acceptor-donor systems is extremely sensitive to water and O2 in ambient conditions, all of the Type IV devices were handled in a dr y box (VAC Omni Lab) providing a working area hermeti cally sealed from the ambient environment, consisting of an inert atmosphere of Argon, nearly free of moisture (~1ppm) and oxygen (~0.15ppm). An antechamber mounted on the side of the dry box is used for passing materi als in and out without disturbing the inert atmosphere of the box. Any time the antechamber is exposed to ambient atmosphere for the insertion of items into the dry box, three cycles of evacuation and filling of the antechamber with inert gas before any of th e objects can be safely br ought inside the dry box, without contaminating th e inert atmosphere. The solvents used in the glovebox were deoxygenated by freeze-pump-thaw for three cycles before being transferred into the glovebox This process is as follows: The distilled solvents in a sealed Schlenk flasks were placed in liquid nitrogen filled dewar. Once the solvent solidified, the flasks were exposed to vacuum fo r 30 minutes. Then the flasks were sealed again
45 and solvents were allowed to melt by removing away from liquid nitrogen dewar. After transferring into the glov ebox, activated molecular sieves were again added to the solvents to remove any residual water. Electrochemical Methods All the electrochem ical measurements were performed using an EG &G PAR model 273A potentiostat which utilizes a th ree electrode configuration, cont rolled using CorrWare software (Scribner Associates). A typica l one compartment three electrod e cell setup is comprised of a working electrode, an auxiliary (counter) electrode, and a re ference electrode in a glass container. A number of different materials used as working elect rodes in this work include 0.02 cm2 diameter gold and platinum electrodes encaps ulated in a Teflon casing (BAS), gold coated Kapton sheets, gold/chromium coated glass slides SWNTs film on glass and SWNTs film on polyethylene terephthalate (PET). Elect rical contact to gold coated Kapton, gold/chromium coated glass slides electrodes were made with copper tape and to SWNTs on glass or SWNTs on PET were made with palladium layer in contac t with nanotubes. A piece of Pt foil welded together with a Pt wire was used as the counter electrode, providing the required current to sustain the developing processes at the worki ng electrode. The size of the foil Pt counter electrode is always chosen as la rger than the area of the work ing electrode to be used. This arrangement prevents large currents from passi ng through the reference electrode and changes its potential. Several different types of reference electrodes were us ed in this work. In aqueous solvents, a silver-silver chloride (Ag/AgCl) sa turated in KCl electrode, commercially available from BAS, was used. For non-aqueous reaction me diums, a silver wire in contact with 0.1M AgNO3 dissolved in a particular solvent (e .g., ACN) is used and denoted as Ag/Ag+ reference. A silver wire (Ag) pseudo-referen ce is used by directly immersing into the reaction medium. Since the potential of silver wire pseudo-reference electrodes rigorously depends on the instant
46 conditions, it was frequently calibrated versus a standard ferrocene solution, which consist of 10 mM ferrocene in an electrolyte solution of interest, and the pote ntials are reported due to the E1/2 of the ferrocene/ferrocinium (Fc/Fc+) against the silver wire. All solutions used for electrochemical techni ques discussed within th is dissertation were carefully prepared from dry solvents as well as pure electrolytes and monomers. Electrolyte and monomer purity is a significant contributor to maximize reproducibility in electrochemistry. Therefore, fresh solutions were used unless otherwise noted, due to the highly reactive nature of electron-rich heterocyclic monomers. Furthermore, the solutions were car efully bubbled with Ar to remove incipient oxygen prior to any experiment. The argon lin e is positioned just above the surface of the solution to mainta in an inert atmosphere blanket during the experiments. This procedure helped to eliminate or at l east minimized any experimental variable. Cyclic Voltammetry Cyclic voltammetry (CV) is one of the m ost extensively used electrochemical techniques to study electroactive and conjuga ted polymers as well as metal oxides due to its simplicity and versatility. It provides both quantitative and qualitative information about the system under study. In this technique, the current density at the working electrode/s olution interface is monitored as a function of the potential, while the potential is being swept over a specified voltage range at a constant rate This dynamic parameter, rate of the potential cycling can be variable accordingly for different reactions and is expressed in mV/s units. The obtained voltammograms reveal information regarding th e electrochemical potentials at which the oxidation and reduction processes occur, how fast these processes occur, the potential range over which the electrochemical system is stable a nd the degree of reversib ility of the electrode reactions under study. The scan rate, switching po tentials, as well as the magnitudes of the anodic peak current (ipa), cathodic peak current ( ipc), anodic peak potential ( Epa) and cathodic
47 peak potential (Epc) are the most important parameters of cyclic voltammetry. Furthermore, CV reveals information regarding the stability of the product during multiple redox cycles. In this dissertation, CV is used for thr ee different purposes; first, to prepare the electroactive films via potentiodyna mic electrodeposition; second, to study electroactive films of polymers and metal oxides; and finally, to characterize the supercapacitor devices which are fabricated by combining the two electrode film s together by a conducting electrolytic media. Cyclic voltammetry measures current with regard to applied voltage at a constant sweep rate, ( =dV/dt ), and is therefore a means of evaluating ca pacitance. As explained in Chapter I of this dissertation, the capacitance of a capacitor is defined by the relation C = dQ/dV, where V the voltage difference between the plates a ssociated with accommodation of charge Q on the each plate. Since the charge actually is the integrati on of current over time range and can be calculated with the following equation; Q = I dt (2-1) where the current I is expressed in amperes (A),the charge Q is in coulombs (C or As) and time t is in seconds. The capacitance can then be calculated through incorporati ng Equation (2-1) into the capacitance equation as follows; C = dQ/dV = I dt / dV = I t/V = I / (2-2) where I is the average current density (A/cm2) and is the scan rate in V/s.123 Throughout this dissertation the voltammogr ams are often graphed as capacitance vs. voltage as well as current vs. voltage. Ideally, a perfect capacito r response would be a rectangular shaped voltammogram. However, the resistance is unavoidable in real systems so most experimental data take the shape of a parallelogram with irregular peaks. The plots obtained at
48 different scan rates are often displayed on the same graph to demonstrate the rate of charging and discharging characteristics which corresponds to different power levels Faster sweep rates indicate higher power levels. As expected, capacitance decreases with higher discharging frequencies. Voltammograms that depicts mirrorimages represents reversible charging and discharging profile while an irreversible proce ss will have two separate charge and discharge profiles. Constant Potential Method: Chrono coulometry / Chronoamperometry In the con trolled potential experi ment, the potential is set at a constant value sufficient to cause the rapid oxidation or reduction of the species and is maintained at this value until only the oxidized or reduced species is present. During potential step methods, the potential of the working electrode is switched in stantaneously between two potential s, either as a single step or repeated steps. Both chronoamperometry and chronocoulometry have the same potential wave form. In chronocoulometry, the charge is m onitored as a function of time, whereas in chronoamperometry, the current is monitored as a function of time. The data obtained during the chronoamperometry experiment is based on the Cottrell equation, which defines the current-time dependence for linear diffusion control: i = nFACD-t (2-3) where, n is the number of electrons transferred per molecule, F is Faraday's constant (96,485 C/mol), A is the surface area of th e working electrode (cm2), D is the diffusion coefficient of the redox active sites (cm2/s), C is the bulk concentration (mol/cm3), and t is the time in seconds. The analysis of chronocoulometry data is base d on the charge calculated by integrating the Cottrell equation, which defines the charge-time dependence for linear diffusion control: Q = 2nFACD-t (2-4)
49 Chronocoulometry has an advantage over chr onoamperometry that the integration of current smoothes random noise and eliminates tim e independent current. In this dissertation, constant and step potential methods are used fo r to prepare homogeneous electroactive films and to examine these electroactive f ilms of polymers and metal oxides as well as supercapacitor devices in terms of switching charges prior to device construction for balanced switching, stability of the films and devices, etc. Constant Current Method: Chronopotentiometry In chronopotentiom etry, a current step is appl ied across an electrochemical cell, between the counter and working electrodes, and the potentia l of the working electrode is monitored with respect to the reference electrode. The data obtained is based on the rate of change in potential versus time. In order to support the applied curre nt, a redox reaction occurs at the surface of the working electrode. In this technique, the ohmic dr op results from solution resistance are constant and equal to the product of the current and the solution resi stance. The ohmic distortion can be simply corrected by a constant po tential offset in contrast to potentiostatic experiments. To analyze the supercapacitor devices as well as active films, cyclic chronopotentiometry, in which instantaneous current reversal is enforc ed on the working electrode, is used. The films or devices were charged and disc harged repeatedly; and the potent ial chance is monitored versus time. From the slope of the linear part of the discharging curves, the vo ltage change over the time interval; V/t can be calculated. Since the current is constant, by incor porating the slope of the discharging curve into Eq 2-2, the discharge capacitance Cd can be evaluated. Electropolymerization Electrochemical d eposition of th e polymer films was carried out in solutions composed of 10 mM monomer in 0.1 M electrolyte solution of interest unless otherwise noted. The electrodeposition is perfor med via potentiodynamic or potentiostatic methods.
50 The potentiodynamic polymerization of a m onomer is typically performed by cyclic voltammetry. The anodic scanning usually starts at low potentials where no redox reactions occur. Once the electrode potential reaches to a sufficient value, the monomer is oxidized to its radical cation, which is determined by an a nodic peak appearance. Monomer oxidation is immediately followed by coupling of radicals to form oligomers which precipitate onto the electrode surface as polymers. The electroactivit y of the polymers that was deposited onto the working electrode can be instantly monitored by the appearance of a reduction peak of the oxidized polymer while scanning in the cathodic direction. Since the current is directly proportional with the electrode area, as shown in th e Randles-Sevcik equation124 (Equation 2-1), the increase in the peak currents is attributed to an increase of the working electrode area. ip = (2.69 105) n3/2AD1/2Cb1/2 at 25C (2-5) where n is the number of electrons, A is the surface area of th e working electrode (cm2), D is the diffusion coefficient (cm2/s), Cb is the bulk concentration of th e electroactive species (mol/cm3), and is the scan rate (V/s). Continuous cycling yields more polymer de position on the electr ode, revealed by the increasing anodic and cathodic peak current densities of the polymer oxidation/reduction. Besides the polymers reduction, usually the re turn scan right afte r the monomer oxidation crosses the anodic wave and is called as the nuc leation loop which is ex clusive to conjugated polymers. This feature also arises due to the increased conducting surf ace area of the working electrode by polymer (or metal) deposition. Furthermore, the Ra ndles-Sevick equation dictates that the peak current is proportional to the square root of the scan rate for diffusion-controlled solution-based electrochemical systems.
51 The potential for potentiostatic electropolymeri zation of a monomer either via single or repeated steps, is typically determined from cyclic voltammograms of the monomer. When the peak current for monomer oxidation is determin ed, the corresponding potentia l is applied to the working electrode during subsequent potentio static electropolymerization experiments. Potentiostatic deposition is a very effective method to deposit smooth and homogeneous conjugated polymer film. The polym erization can be terminated once certain amount of charge has passed (determined by chronocoulometry), so the desired mass of the polymer has reached. By using a calibration plot of charge versus pol ymer mass, determined by performing separate experiments for different monomers, the result ing polymer mass can be chosen by setting the potential to be applied until the necessary charge has passed. Once the polymer films are prepared they were rinsed thoroughly with fresh solvent or monomer-free electrolyte solution to wash away unreacted monomers tr apped on the film prior to elec trochemically switching in a monomer-free electrolyte. Electrochemistry of Electroactive Film Electrochemistry of activ e f ilms was performed again in a one compartment threeelectrode cell as described above. Only this time the polymer-coated (or metal oxide coated) electrode were used as the working electrode. The working potential rang e of a polymer includes a region from the neutral form of the polymer to a value sufficient to oxidize or reduce the polymer. Electroactive films were re versibly cycled in this range between their neutral and doped forms in a monomer free 0.1 M electrolyte solution using cyclic voltammetry. As the electroactive polymer films are steppe d between their fully neutral and fully doped states, an increase or a decrease in the curren t response is observed, which is attributed to polymer oxidation or reduction and incorporation or expulsion of the charge compensating counter-ions throughout the polymer film, respectiv ely. Prior to electrochemical study, all the
52 films were switched between their oxidized and reduced states 3 to 8 times to condition the polymer by stabilizing ion and solvent diffusion in and out of the polymer film. This is called break-in period required to obt ain more reproducible experi mental observations. Important parameters of a polymer CV ar e; the half-wave potentials E1/2, (the potential where the concentrations of the oxidized and reduced species are equal), determined by taking the average of the peak anodic and cathodic cu rrent potentials; scan rate depe ndence of the peak current; and the reversibility (shape) of the potential wave. In those cases for electrode adsorbed species such as electroactive polymers adhered to the working electrode, the Randles-Sevick equation (Eq 2-5) does not apply, because the redox processes of electrode-bound conjugated polymers are not diffusion controlled. Therefore a different approach must be taken. The following Eq 2-6, for surface-confined species dictates that both the anodic and cathodic current res ponses will scale linearly with scan rate. ip = n2F2A / 4 RT (2-6) where n is the number of electrons, F is Faradays constant (96,485 C/mol), A is the surface area of the working electrode (cm2), is the concentration of surface bound electroactive centers (mol/cm3) and is the scan rate (V/s). In scan rate de pendence experiments, a linear relationship between the current responses and scan rate indicates a non-di ffusion controlled process where the electroactive polymer is we ll adhered to the working elect rode surface. Although conjugated polymers can switch rapidly withou t loss of current response, th e rate of switc hing is still dependent upon film thickness, electrolyte, and the morphology of the polymer. Supercapacitor Device Fabrication The construction of a supercapacitor device co n sists of combining together positive and negative electrodes containing activ e films with a conducting electr olytic media. The polymeric
53 gel electrolyte used as the conductive media in Chapter 3 an d regular liquid base electrolyte solutions were used in the studies pr esented in Chapter 4 and Chapter 5. The uniform eletroactive films used as electrodes of the device were obtained by electrodeposition of the polymers or metal oxide s onto the substrates, where the details are specified in each chapter. The amount of comp lementary polymers or metal oxides on each electrode can influence the device operation to a la rge extent. Therefore, it is important to match redox switching charges of the complementar y polymers for balanced switching in a supercapacitor device fabrication. As it was me ntion earlier in the previous section, the deposition amounts can be c ontrolled by monitoring the total charge density passed during electrosynthesis. However, it is not so accurate to get same amount of active material deposition in each deposition due to the unavoidable experimental errors. For that reason, it is essential to check the charge density of the electroactive films wh ile switching between it is doped (por n-) and neutral states prior to device assembly. By using CV or simply by applying an oxidizing or reducing potentials, respectively, conjugated polymers can be switched between their doped and neutral states and the total charge passed during the experiments can be ca lculated as explained in the previous sections; (Equa tion 2-1). This total charge, Q is related to the number of electrons transferred per molecule, n and the number of moles of the oxidized species initially present, N through Faraday's law: Q = nFN (2-7) where F is Faraday's constant (96,485 C/mol), n = number of electrons transferred/molecule, and N = amount of material (mol). When dealing wi th conjugated polymers, the doping level of the films should be included in this equation as well.
54 After the electropolymerization, the films were set to the desired doping state of fully oxidized, fully reduced or neutra l by applying a consequent potent ial for 15s to ensure charge balance level prior to device asse mbly. In Type I and hybrid devi ces as described in Chapter 3 and Chapter 5 respectively, anode films were oxidatively doped (fully oxidized) while cathode films were fully neutralized in a monomer-free so lution in order to establish the initial charge balance state of the device. In Type IV devices described in Chapter 4, both anode and cathode films were fully neutralized prior to device cons truction and a special Teflon cell was used that allows insertion of two Au-Button electrodes sepa rated by 0.5cm with the electrolyte solution in between them. In Chapter 5, the positive and nega tive electrode films were placed 1.5 cm away from each other in a glass contai ner with the electrolyte solution. In Chapter 3, multiple layers including tw o electroactive films deposited on a conducting substrate either onto Gold/ Kapton or onto SWNTs were sandwiched together to form the supercapacitor device. The films were coated via syringe with a viscous gel electrolyte until the entire polymer surface was uniformly covered to ensure adequate swelling of the polymer. These substrates were then carefully applied to face to face and allowed to stabilize for 24 hours. To prevent electrodes from shorti ng out that might occur while applying two electrodes to one another, a porous separator paper (Gore Excellerator) were used between the electrodes. In case of flexible Gold/Kapton and SWNT film s which are not on glass slides, acetate (3M transparency film, PP2500) were used as a support and cover for th e substrates. Besides acting as an ion transport material, the gel electrolyte provides a relative encapsulation at the edges of the device as the PMMA becomes insoluble. This mi nimizes further solvent evaporation, prevents leaking, and allows for long-term testing at am bient conditions. Although the devices are selfencapsulated by a seal formed around the edges by the gel electrolyte, the entire device still
55 carefully sealed on all four edges using transparent tape to prevent a possible leakage. The device schematic and photographs can be s een in the subsequent chapters. The polymeric gel electrolyte used as the c onductive media was based of a salt plasticized in solvent that has a high boiling point. Th e optimized composition of the gel electrolytes includes 10% salt, 20% ultrahigh molecular wei ght PMMA and 70% solvent by weight ratio in order to get both substantially c onducting and mechanically stable and viscous gel. The gel was prepared by dissolving Li[N(CF3SO3)2] in tetraethylene glycol dime thyl ether or PC, followed by very slowly adding PMMA while vigorous stirring and mild heati ng (60C) for a period of about two hours or until all components were incorpor ated to form a highly viscous, a honey like consistency. The performance of the capacitors was charac terized using linear sweep voltammetry and galvanostatic charge/discharge techniques as described in the pr evious sections. Energy and Power Characteristics of a Device Supercapacitors are energy storage and conve rsion devices. Besides the charge storage ability, the energy and power densities ar e am ong the most important metrics for supercapacitors. The performances of these device s are characterized using their energy storage and conversion features as demonstrated in Ragone plots shown in Fi g 1.1 (Chapter I). The energy density of a supercapacito r charged to a potential difference of V, for an accumulated charge Q (Coulombs) residing on electrodes of the cap acitor, can be calculat ed due to Eq 1-2: E = CV2 = QV where C is the capacitance in Farads (F). Althoug h the voltage is defined as the potential difference between the cathode and the anode; th e voltage of a superc apacitor progressively decays during the discharging process since work is being continuously done against the charges being accumulated/released on the electrode. Thus an average voltage of
56 V = (Vcharged + Vdischarged)/2 (2-8) is realized in calculati on of the energy density of a superc apacitor device. This equation assumes a ideal linear voltage decay at constant curren t. The energy density is usually expressed in (Wh/kg) as the specific (or the gravimetric) ener gy density and can also be stated as the areal (Wh/m2) and the volumetric (Wh/L) energy density depending on how the capacitance is expressed. Power is the average amount of energy deliver ed per unit time, in essence, it is the discharge rate of the supercapac itor device. The power density va lues are evaluated by dividing the energy density values by the discharge time of the cell. Correspondin gly, the average power delivered by a supercapacitor is defined as: Power = iV (2-9) where i is the average current and V is the average voltage described above. In a chronopotentiometry, discharge performed at cons tant current, thus determining the average power requires only monitoring of voltage during discharge. In a CV experiment, the average current density during the discha rge of the cell should be dete rmined. The power density is typically expressed in W/kg as the specific (or th e gravimetric) power density and can also be stated as the areal (Wh/m2) and the volumetric (W/L) energy density depending on how the capacitance is expressed. Surface Characterization Techniques In this dissertation, both atom ic force microscopy (AFM) and scanning electron microscopy (SEM) used to characterize surface a nd cross section morphologies of films. Both techniques are the foremost tools for im aging and offers high-resolution at the nanometer scale.
57 Scanning Electron Microscopy In SEM, images are form ed by scanning the sample surface with a high-energy beam of electrons in a raster scan pattern. The electr ons in teract with atoms at or near the surface of the sample resulting in low energy secondary electro ns being emitted from the sample surface. These emitted electrons are detected by a photomultipli er tube to form the topographic image. The other types of imaging with the SEM includ es backscatter imaging that can provide compositional information, charac teristic x-ray imaging providi ng information on the elements present in the sample and specimen current im aging that can be used to show sub-surface defects. Depending on the instrument, the SEM can produce very high-resolution images with the resolution between less than 1 nm and 20 nm and also with a three-dimensional appearance useful for understanding the su rface structure of a sample. For all SEM studies, an Hitachi S-4000 FE-S EM was utilized for obtaining topographic images throughout this dissertation. The car bon double side conductive tabs (PELCO Tabs) were used to place the films onto the sample holder and the conducting carbon paint (EM Graphite Conductive Adhesive 154) is applied to the edges of th e films to enhance the adhesion. Prior to imaging, the samples were sputter coat ed a thin layer (about 100A thickness) of gold and palladium alloy in a Denton Vacuum sputter coater in order to prevent surface charging, minimize radiation damage and to increase the electron emission from the films. A 45 tilted angle sample holder is used unless otherwise is in dicated, in order to improve the quality of the images obtained. An acceleration voltage of 6 kV was selected for examining the samples. Atomic Force Microscopy This contact technique has an operation princi ple involving a stylus or cantilever that moves acros s a sample surface. The sample surface is scanned with a shar p tip, with nominal tip radius on the order of 10 nm, located at the end of the cantilever. AFM tips and cantilevers are
58 typically microfabricated from Si or Si3N4. This flexible cantilever with a very low spring constant induces forces smaller than the interatomic forces between the tip and the sample, thus the topography of the sample can be monitored without displa cing the atoms. AFM uses an optical lever detection system involving a focused beam from a laser diode onto the back of the cantilever, to sense the deflection of the cantilever as a response of surface height variation. The beam reflects off the back of the cantilever onto a segmented position sensitive photodiode. The amplified differential signal between the uppe r and lower photodiodes provides a sensitive measure of the cantilever deflecti on. Finally, this response signal is converted to a digital signal. AFM has the advantage of imaging almost any t ype of surface, including polymers, ceramics, composites, glass, and biological samples, whether it is conducting, semi conducting or insulating surface. Three imaging modes; contact mode, non-contact mode and tapping mode can be used to produce topographic images of sample surfaces. Contact Mode AFM: The contact mode AFM operate s by dragging the tip over the surface corrugation, while monitoring the change in cantilever deflection with a split photodiode detector. The tip is in contact with the surface through the adhesive forces between the tip and the surface, and therefore it is adjusted to maintain a consta nt height above the surface, so a constant deflection. The dragging motion of the probe tip can caus e substantial damage to both sample and the probe and create artifacts in the image data. Non-Contact Mode AFM: This mode uses an oscillating stiff cantilever located quite close to the sample surface in th e attractive regime, but not touching it. Attractive Van der Waals forces between the tip and the sample lowers the cantilevers resonant frequency, and topographic images are monitored based on measur ing changes to the resonant frequency or
59 amplitude of the cantilever. Since the forces betw een the tip and sample ar e substantially low, on the order of pN (10-12 N), this mode has a very low resolution. Dynamic Force/Tapping Mode AFM: A very stiff cantilever is oscillated closer to the sample than in noncontact mode and close to it s resonance frequency (normally on the order of 100 kHz) so that the tip intermittently touches to the sample only for a short duration in each oscillation cycle. The amplitude at the operating frequency is maintained at a constant level, called the set-point amplitude, by adjusting the rela tive position of the tip with respect to the sample during scanning. As the tip touches the sa mple, the interactions between the tip and the surface molecules alter the amplitude, resonance frequency and the phase angle of the oscillating cantilever. The changes in the amplitude and phase angle of the cantilever probe are monitored as the topographic (height) and the phase image, respectively. The phase image often provides significantly more contrast than the topographic image and has been shown to be sensitive to the material surface properties such as stiffness, viscoelasticity and chemical composition. This method of operation results in high resolution imag es especially on soft samples, while virtually eliminating the lateral forces such as dr ag, so less surface damage is inflicted. All the AFM experiments were carried ou t using the Digital Instruments MultiModeTM Nanoscope III in this dissertation. Tapping mode AFM utilized in all of the AFM images unless otherwise is indicated.
60 CHAPTER 3 PPRODOP BASED TYPE I SUPERCAPACITORS PProDOP as a Charge Storage Electrode Material The use of conjugated polymers (CPs) as ps eudocapacitor electrode materials has been widely investigated.125, 126 As explained in Chapter I, their en ergy storage is attributed to redox pand n-doping Faradaic reactions. The redox process of conjugated polymers is fast, and occurs efficiently since films are highly porous and accessi ble, allowing fast perc olation of electrolyte through the matrix. Thus, fast disc harging gives rise to the high power densities that are possible with these materials. CPs have the ability to store charge throughout their entire volume; therefore, elevated levels of high charge densities can be achie ved. One of the most attractive aspects of CPs is the ability to tailor specifi c properties such as c onductivity, morphology, etc. with straight forward synthetic modifications without drastic adjust ments to the polymer backbone. In addition to fast redox electrochemistry with suitable morphology, CPs lightweight nature, synthetic flexibility and ease of prepar ation in the form of extended surface films make them attractive for supercapacitor applications. Moreover, CPs can be easily and inexpensively made in large amounts. Since the entire volume of polymer is involved in the charge storage process, supercapacitors that use CPs have high capacitive energy densities. Furthermore, their fast doping-dedoping processes allow for devices w ith low equivalent se ries resistance (ESR) and high specific powers. In terms of satisfact ory energy and power dens ities, cycle life, low weight and thermal stabilities with the predicted low material cost relative to other materials, conducting polymers represent an attractive alternative to ma terials currently used in electrochemical capacitors. Characteristically, conduc ting polymers offer specific capa citances in the range of about 100-250 F/g. Over the years, the bulk of conducting polymer based supercapacitor research has
61 focused on polyaniline, polypyrrole, pol ythiophine and their derivativ es with different kind of electrolyte and substrate systems to achieve high energy storages with high capacitances.4, 15, 127 Most of this work has been directed toward use in Type I supercapacitors. Being one of the most commercially succe ssful polymers used for supercapacitor applications to date, poly(3,4ethylenedioxythiophene) (PEDOT), exhibits higher chemical and electrochemical stability than most of the other CPs. 69, 72, 128 Introduction of the ethylenedioxy bridge with electron-donating oxygen atoms onto the 3and 4-positions of the thiophene ring increases the electr on density on the -system, which raises the highest occupied molecular orbital (HOMO).129, 130 As a result, PEDOT can be easil y oxidized, the polymer has a lower bandgap (1.6 eV) than unsubstituted PTh, po ssesses lower oxidation potentials, and higher stability upon oxidation. A fam ily of poly(3,4-alkylenedioxyth iophene) derivatives (PXDOTs) with varying size and composition of the alkylene bri dges has been synthesized by the Reynolds group, and shown to have a variety of the electronic and optical properties.131-135 Moreover, through the same approach by intr oducing alkylenedioxy br idge as in the PX DOTs, a series of poly-(3,4-alkylenedioxypyrrole)s (PXDOPs) have been synthesized in the Reynolds group to obtain polymers with very low oxidation potentials.119, 120, 136Analogous to PXDOTs, introduction of the alkylenedioxy moiety in the 3and 4positions of the pyrrole ring increases the electron-rich character of the monomer, and raises the HOMO level of the molecule.137-139 Consequently, this new family of monomers a nd polymers have a relatively lower bandgap (~2.2 eV), exhibit outstanding stabilities as analogues to PXDOTs display very low oxidation potentials that allows milder polymerization conditions yielding polymers with much fewer structural defects.140 Moreover, some of them possess aqueous compatibility similar to the intrinsic property of polypyrroles.141
62 Electropolymerization of ProDOP Among the PXDOP derivatives, poly(3,4-pr opylenedioxypyrrole) (PProDOP), whose structure is illustrated in Fig 31, has started to attract attention in the academic world recently, due to its multitude of interesting and unique ma terials properties. PPr oDOP is highly electron rich in character, easily p-t ype doped, and exhibits highly cap acitive behavior in its redox electrochemistry. Besides its high doping capacit y, it provides a long-term chemical stability, high electrical conductivity, along with fast and efficient switching between redox states. This chapter will focus on the behavior of PProDOP in greater detail, specifically in the realm of supercapacitors. Figure 3-1. Structure of poly(3,4-propylenedioxypyrrole) (PPr oDOP) shown along with the electropolymerization mechanism and th e doping/dedoping processes of PProDOP, where Arepresent dopant ion.
63 Electrochemical formation of PProDOP proceed s via a step-wise chain growth mechanism by oxidation of the neutral monomer, shown in Fig 3-1, that yields the radi cal cation formation at the anode surface. Then coupling of two radical ca tions (dimerization reac tion) proceed. Further, this dihydro dication dimer spec ies loses two electrons and rear omatizes to form the neutral dimer. Consequently the dimer oxidizes again to form the radical cat ion. Through coupling of two radical cations or a radical cation with a neutral monomer, the polymerization proceeds and the chain length increases. As the oligomers beco me insoluble in the el ectrolyte solution, they start precipitating on the anode. The polymeri zation proceeds through coupling at the 2, 5 positions of the ProDOP monomer resulting in a linear backbone polymer with an enhanced degree of order. As was mentioned above, the substitution of the PPy monomer with electron donating propylenedioxy group increases the el ectron-rich character of the monomer and decreases the monomer oxidation pote ntial. This results in one of the most important features of PProDOP that the electropolymeri zation occurs under milder conditions, reducing the possibility of obtaining overoxidize d material with structural defects. Cyclic voltammetry offers a powerful method fo r the characterization of the monomer and polymer redox processes. Oxidative electrochemi cal polymerization of ProDOP was carried out using multiple scan cycling in propylene carbonat e (PC) with 0.1 M electrolyte solution on gold substrates as illustrated in Fig 3.2. Tetrabut ylammonium perchlorate (TBAP) and lithium bis(trifluoromethanesulfonyl) imide (LiBTI) were us ed as electrolytes, as shown in Fig 3.2A and Fig 3-2B, respectively. Electroact ive films were easily prepared from a 10mM solution of monomer in each electrolyte system by potentiodynamic sweeping from a potential of negative to that of the polymer reduction, to a potential approximately 50mV pa st the peak monomer oxidation potential. The p eak oxidation potentials ( Ep,m) obtained during the first scan were
64 found to be +0.65 vs. Fc/Fc+ and were all in agreement with each other for different electrolyte systems. The lower oxidation potential compared to pyrrole (+0.9 vs. Fc/Fc+) can be attributed to ease of formation of the radical cation at the electrode due to the electron rich character of ProDOP monomer. The changing shape of the cu rrent-voltage relationshi p with the progression of the cycles in Fig 3-2A and B is most pr obably due to the deposition of polymer on the electrode with an increased resistance. ProDOP exhibits the fastes t nucleation and the most e fficient deposition rate among several other XDOPs.119 ProDOP electropolymerized quite efficiently in each electrolyte medium; only few potential sweeps were enough fo r the formation of an adhering, adsorbed electroactive film onto the electrode surface. This efficient film deposition is attributed to fast coupling of the reactive intermediate radical catio ns with instant polymer ization on the electrode surface. No color is evident in the electrolyte solution and, thus; soluble oligomer formation does not occur. The coupling of radicals to form oligomers results in precipitation of them on the electrode surface. The electroactiv ity of polymer on the gold electrode can be easily seen from the voltammograms evidenced by the redox peak a pprearance by the as made polymer film. The redox peak potentials of oxidative doping of CPs to their conducti ng forms are functions of the dopant ion, solvent and supporting electrolyte used.130 There were not any sharp faradaic peaks from polymer redox processes observed for PProDOP films. This result is probably due to the polymers capacitive behavior, es pecially in the presence of LiBTI electrolyte. PProDOP exhibits a half-wave potential ( E1/2,p) approximately of -0.3V vs Fc/Fc+ and is among the polymers with the lowest oxidation potentials fo r p-type doping. As an example, PEDOT has an half-wave potential of around -0.2V vs Fc/Fc+ in regular electrolyte so lutions, which shows its more resistive character to oxidaditive doping compared to ProDOP.
65 -1.0 -0.5 0.0 0.5 1.0 -5 0 5 10 15 E(V) vs. Fc/Fc+AI (mA/cm2) -1.0 -0.5 0.0 0.5 1.0 -5 0 5 10 15 BI (mA/cm2)E(V) vs. Fc/Fc+ -0.6-0.4-0.20.00.20.40.60.8 -1 0 1 2 3 4 5 CI (mA/cm2)E(V) vs. Fc/Fc+ Figure 3-2. Electrochemical polyme rization (first 5 cycles) of Pr oDOP by potential scanning A) in TBAP/PC on Au Button, B) in LiBTI/PC on Au Button and C) in LiBTI/PC on Au/Kapton at 50mV/s.
66 In Fig 3-2C, cyclic voltammogram of ProDOP in a same electrolyte solution (LiBTI/PC) as in Fig 3-2B, however on a larger area gold substrate (1cm2) is shown. The broader monomer oxidation peak, compared to the same syst em on gold button electrode which is 0.02cm2, is typically observed for larger area electrodes wi th conducting polymer systems. Moreover, the appearance of a shift in the monomer oxidation potential was observed with the progression of each cycle, and may indicate a growth involving the coupling of soluble oligomers, which are probably more reactive than the monomer its elf, so have a reduced oxidation potential. 142, 143 Therefore, the combination of th e oxidation of monome r species with the oxidation of oligomer species results in a broader and shifted peak. Electrochemistry of PProDOP Films As was introduced in Chapter I, conducting polymers store and release charge through their redox doping and dedoping processes. Fig 3-1 illustrates the mechanism of this Faradaic charge storage in PProDOP. It can be seen th at the neutral polymer in its uncharged state becomes positively charged (p-type doping) upon the removal of an electron from the polymer to form a delocalized radical cation (a polaron) The polaron can under go a second oxidation to form the dication (a bipolaron). Cons equently, the dopant anions, shown as Ain the figure, from the electrolyte enter into the polymer to comp ensate the charge created on the backbone. When neutralization occurs the anions are re leas ed back into the electrolyte solution. Introduction of charge carriers in the form of polarons and bipolarons into the polymer matrix occurs at the polymer/electrode interface and th ese charged states are quickly dispersed through the polymer matrix due to migration and diffu sion via a hopping mechanism. PProDOP can be doped and dedoped rapidly to high charge densities. Electropolymerization is a suitable synt hetic method to prepare electrodes for supercapacitor applications, sin ce it gives rise to polymers that are more open and porous
67 structure compared to the packed morphologies that result from spin coating deposition.144 While cyclic voltammetry offers a powerful method for the characterization of the monomer and polymer redox processes, deposition of the polym er has been carried out by potentiostatic electrochemical polymerization in order to get higher current e fficiencies, more homogeneous (uniform) and smooth films of PProDOP.119 Polymer films were deposited at an applied potential equal to the monomer peak potential of +0.65V vs. Fc/Fc+ plus 0.05V on Gold/Kapton electrodes (1cm2) in PC using 0.01M monomer and 0.1M TBAP and LiBTI as electrolytes. PProDOP is stable to higher potentials, si nce the presence of 3,4-alkylenedioxy bridge substituent at 3 and 4 position of the pyrrole ring prevents overoxi dation of the polymer thus the degradation of the conjugated backbone. The prepared films were washed in monomer free electrolyte media and characterized by cyclic voltammetry. Cyclic voltammograms of PPro DOP films with the scan rate dependence for the redox switching are shown in Fig 3-3a&c. Cyclic voltammograms were carried out at different scan rates from 50 to 500 mV/s between .4 and +0.4 V vs Fc/Fc+. Well-defined voltammograms with reversible broad peaks, which are typical of large surface area (1cm2) electrodes, were observed for both electrolyte systems. The peak currents in the cyclic voltammograms in the presence of LiBTI electrol yte increase linearly up to 500 mV/s scan rate, indicating that the elec troactive sites are surf ace bound to the electrode and that the oxidation and reduction processes are not diffusion limited. The variation of the voltammograms in the case of TBAP electrolyte shows almost linear dependence up to 250 mV/s scan rate. However, at higher scan rates from 250 to 500mV/s, a s light deviation has been observed for each voltammogram, which is attributed to resistive effect on the kinetic s of the slow diffusion of ions in and out of the polymer with respect to the fast potential change and also to a probable
68 resistance occurs during redox processes. In other words, the diffusion of ions can be considered as the rate-determining step, as it is much slower than the electron-transfer reaction. An important aspect is that PProDOP presents a highly capacit ive like voltammogram shape, especially in the presence of LiBTI electrolyte, with fast and effi cient switching. This is likely due to the size effect of the ions, especially during the electrochemical synthesis, and how the size of the ions influences the morphology of the polymer. -0.4 -0.2 0.0 0.2 0.4 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 A 500mV/s E(V) vs. Fc/Fc+I (mA/cm2) 50mV/s-0.4 -0.2 0.0 0.2 0.4 -6 -3 0 3 6 C (mF/cm2)E(V) vs. Fc/Fc+500mV/s 50mV/s B-0.4-0.20.0 0.2 0.4 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 I (mA/cm2)E(V) vs. Fc/Fc+500mV/s 50mV/s C-0.4-0.2 0.0 0.2 0.4 -9 -6 -3 0 3 6 C (mF/cm2)E(V) vs. Fc/Fc+500mV/s 50mV/s D Figure 3-3. A&C) Cyclic voltammograms and B& D) capacitances as a function of applied potential of PProDOP films on Au/Kapton substrates in A&B) LiBTI/PC and C&D) TBAP/PC at 50 to 500mV/s w ith 50mV/s intervals. Fig. 3-3B&D shows the capacitance as a functio n of potential at scan rates from 50 to 500mV/s. In Fig 3-3B&D, the capac itances of the films were obtai ned by using the relationship; C=i/s (Equation 2-2), where i is the current density (A/cm2) and s is the sweep rate (mV/s). Moreover, the average ca pacitance of each polymer film, determined as the ratio of the
69 average charge density, integrated from the cycl ic voltammograms, involved in the charging and discharging process over the pot ential range. The equation dictates that the capacitances calculated for each voltammogram to be independ ent of the scan rates. In the Figure 3-3, a decrease in the capacitances w ith increasing scan rates can be attributed to the unavoidable resistance occurrence during redox switching of the polymer. Overall, the results show efficient switching with reasonable fluctuations of the valu es in the limits of experimental errors. PProDOPs present an almost ideal ratio of 1 of the anodic to cat hodic peak currents ( ipa/ ipc) even at a scan rates as high as 500mV/s, whic h illustrates the outstanding reversibility of PProDOPs redox processes. Its capacity to be sw itched in highly reversible manner at high scan rates is rather unusual for conducting polymers. This phenomenon might be stem from the high conductivity of PProDOP films. Electrochemically prepared f ilms of PProDOP have been reported to exhibit conduc tivities of 95 S/cm with a variability of about one order of magnitude depending on the polymerization conditi ons, dopant ion, and film thickness.120 This fine conductivity of PProDOP is probably due to a relatively planar polymer backbone with no substituent units, along with a negligible steric hindrance between adjacent units.145 This allows for better packing of the monome r chains, allowing efficient -stacking and consequently enhancing the materials conductivity via in terchain hopping (the mechanism of charge percolation). This also suggests that PProDOP has a long effective conjugation length with limited defects that could localize charge carriers. Surface Analysis of PProDOP The structures of the PProDOP film s on Kapton/Au electrodes (1cm2) were examined using a scanning electron microscope (SEM). SEM images of polymer films deposited under constant potential polymerization conditions in the presence of el ectrolytes LiBTI and TBAP in PC are shown in Fig 3-4 and Fig 3-5 respectively. The prepared f ilms were reasonably smooth as
70 viewed by the eye. The SEM studies were aimed to gain insight into the fine structure of the films. A 45 tilted angle sample holder was us ed in order to improve the quality of the topographic images obtained. Figure 3-4. Scanning electron micrographs of PPro DOP films on Au/Kapton, deposited at potentials A) before, B) at and C) after the Eox of PProDOP in LiBTI/PC from low to high (A1 to A3, etc.) magnification. The films prepared with LiBTI were found to be highly porous with a large interconnecting network of polymers, as revealed by SEM images shown in Fig 3-4. In order to observe the influence of the polymerization po tential on the morphology of the resulting film, samples were deposited using va rious potentials. PProDOP films were deposited at the monomer oxidation peak potential ( Ep,m = +0.65 vs Fc/Fc+), plus 0.05V and minus 0.05V. Fig 3-4 presents SEM images for each polymerization potential from low magnification to high magnification 30 m 30 m 30 m 1.5 m 1.5 m 1.5 m 600nm 857nm 600nm A1 B1 C1 A2 A3 B3 C3 B2 C2
71 (A1 to A3, etc.). In the low magnification SEM images, the film resembles a folded blanket covering the electrode surface. For each polymerization potential the films were overlaid into ridges and valleys to varying degrees as s een in Fig 3-4 A1, B1 & C1. Increasing the magnification, the films prepared at the potenti als below and above the monomer oxidation peak potential were found to be highly porous with a large interconnecting network of polymers, as revealed by SEM images shown in Fig. 3-4 A1&A 2 and C1&C2, respectively. Clearly, there are a large number of micro and nano sized pores pr esent in the network of PProDOP. This large surface area, highly porous structure provides ea sy access to the dopant ions. In Figure 3-4 B2&B3, the high magnification SEM images of the films prepared at the monomer oxidation peak potential are shown. In contrast with the other films, these films show less porous morphology, however, a large continuous networ k still exists. There was not any major difference observed between the cyclic voltamm ograms of these three PP roDOP films prepared with LiBTI at different potentials. Figure 3-5. Scanning electron micrographs of PProDOP films on Au/Kapton, deposited in TBAP/PC at varied magnification. Furthermore, the films prepared in the presence of TBAP are shown in Fig 3-6. These films are polymerized at a potential of the m onomer oxidation peak poten tial plus 0.05V. SEM images of the resulting films revealed comp letely different morphology than PProDOP film prepared with LiBTI. The nature of the electrolyte used during the deposit ion strongly influences 5 m 6 m 3.33mA B C
72 the morphology of the resulting film as shown in Fig 3-4 & Fig 3-5. A more packed, nodular structured continuous layer of PProDOP-TBAP films relative to PProDOP-LiBTI films are immediately evident in Fig 3-6 A&B. One of the most likely causes for formation of this packed structure is the ion size and the interaction of these ions with the solvent swollen polymer during electropolymerization. Cracks that developed throughout the film during the drying process are evident. The splits in film reveal the smooth gold substrate surface underneath, which shows that the polymer film was not directly attached to the electrode surface at all points. The SEM image in Fig. 3-5C shows a magnification of the rising broken piece of ri dge of the film. This packed texture, along with the poor attachment to the electrode surface, correlates with the electrochemical response of PPr oDOP-TBAP films shown in Fig 3-3. The resistant behavior seen on the cyclic voltammograms of the films with TBAP compared to LiBTI salt, especially at higher scan rates, is possibly due to this packed structure of the film preventing easy access of the ions, unlike the accessible porous morphology, and also likely influenced by the decreased electrical conductivity of the film. As a result, SEM analys es of the films prepared at various conditions indicate that the conditions during electrochemical polymerization method and the nature of the electrolyte strongly influence morphology. Capacitances of the PProDOP Films on Gold Electrodes Besides obtaining smooth and homogeneous polym er films, one of the advantages of potentiostatic deposition techniqu e is the control of the amount of charge passed during the deposition process. The charge density applied to the system during polymerization, given the electrochemical systems, reproduc ed with identical concentrati ons and compositions, determines the amount of PProDOP film deposited on the electrode. As menti oned earlier, the film deposition can be terminated once a certain amount of charge has passed, so the desired mass of the polymer can be conveniently controlled. In order to relate the depos ition charge passed with
73 the resulting mass of the polymer, a simple calibra tion graph was prepared, as shown in Fig 3-6a. Moreover, another calibration plot that relates the capacitance obt ained from varying amount (or thicknesses) of PProDOP with th e corresponding mass of the films was prepared as shown in Fig 3-6B. The latter graph is more convenient for this work since the charge density monitored during the deposition process depends on several parameters like concentrations of species, temperature, etc. therefore, experimental e rrors are unavoidable. However, once the polymer films were obtained, the amount of capacitance is definite for the amount of material for various films. So when a certain capaci tance value was obtained from any random PProDOP film, the amount of PProDOP would be easily determined from the calibration plot. Figure 3-6. Calibration curves: A) Deposition char ges of ProDOP and B) capacitances of PProDOP as a function of the mass of PProDOP on Au/Kapton substrates. 0 0.05 0.1 0.15 0.2 0.25 010203040506070Q (C/cm2)mass (g/cm2)A 0 1 2 3 4 5 6 7 8 9 10 010203040506070C (mF/cm2)mass (g/cm2)B
74 In theory, calibration curves would be exp ected to be linear and the capacitance (mF/cm2) would increase linearly with increasing mass of charge storage material. A linear relation was observed between the charge consumed for elec tropolymerization and also the capacitances of the resulting films and the mass of them. However, after a ce rtain amount of material was deposited, the films became less stable and tended to delaminate from the gold surface after several switching cycles between oxidized and reduced states. This was not unexpected, as charging and discharging proce sses in redox reactions based on pseudocapacitive materials involve ion intercalation and depletion due to the electric neut rality requirement, which results in volume expansion and contraction of the conducting pol ymer systems. As a re sult, the film starts to delaminate from the electrode substrate as it gets progressively thicke r. In conclusion, only relatively thin PProDOP films c ould be grown coherently, as fu rther electropolymerization lead to poorly or non-adhered polymer s films on the substrate. In Fig 3-6B the capacitances (mF/cm2) obtained from the gold electrodes holding varying amount of PProDOP were demonstrated as a function of mass of PProDOP. A very small intercept of the plot arises from the double-layer charging of the bare gold electrode. The highest capacitance values obtained from PProDOP films on gold electrodes were found to have capacitance of 8.1mF/cm2. From the slope of the graph, th e overall specific capacitance of PProDOP on gold electrodes was calculated as 141F/g. Although PProDOP films displaying smaller capacitances than 4mF/cm2 probably holding less than ~30 g/cm2 were prepared, the mass values of those films could not be meas ured accurately due to the very small mass deposition. On the other hand, the films having larger amount of PProDOP were observed to delaminate from the electrode surface during the washing processes or upon several cycle switching.
75 The capacitance value of supercapacitor material s are usually reported in the literature as specific capacitance which is the capacitance per mass of the material (F/g). However, focusing only on the mass specific capacitance in the ev aluation of supercapacito r electrode materials might be misleading.146 As mentioned above, in th eory the capacitance (mF/cm2) of an electrode would increase linearly with increasing the amount of the charge storage material. The charging and discharging processes of conducting polymers involve the in sertion and expulsion of ion during the doping and dedoping processes. Besides the mass of the material, the thickness, the conductivity, the porosity of the polymer and the diffusion of the ions also play an important role in the kinetics of the charging and discharging pr ocess. Therefore, the switching process would become restricted by kinetics after certain amount (or thickness; the thickn ess of the material on an electrode substrate of a given surface area is di rectly proportional to the mass) of polymer has obtained. As a result, it is not always possible to gain higher ca pacitance values by just simply increasing the mass of the material after th e thickness has reached the kinetic limit. The specific capacitance of a certain mass of PProDOP with the ideal morphology, which is expected to exhibit doping and dedoping redox processes without any kinetic limitations, is calculated by using Equation 2-7. The theoretical specific capacitance of PProDOP is found as 173 and 208 F/g for the doping levels of 0.25 and 0.3, respectively. Accordi ngly the calculated theoretical capacitance of PProDOP was found to be somewhat la rger than the experimentally obtained capacitance of 141 F/g, sugge sting the presence of a portion of the polymer film that is incapable of undergoing the redox re actions, due probably to the pa rtial lack of uniformity, open morphology or the presence of defects. Ideally, the calibration curve rela tes the area capacitance (F/cm2) and the amount of the polymer would be expected to in itially increase and after reaching the kinetics limits, to level out
76 (stabilized) at certain capacitance. On the ot her hand, the graph relate s the mass capacitance (or specific capacitance) (F/g) and the amount of the polymer w ould be expected to be stable and decrease at the kinetic limit of the electrode Consequently, evaluati on of capacitance of a supercapacitor material must be done using by both mass and areal capacitance values. While mass capacitance gives intrinsic information about the material, the area l capacitance measures the practically accessible capacitance over a unit geometric area of the electrode substrate.147 Hence, for device fabrication issu es, the areal capacitance is of greater practical importance since its value reflects the uti lization of the polymer material effectively on a certain electrode substrate. Type I Supercapacitor Devices with Gold Substrate Fig 3-7 illustrates the Type I supercapac itor device design which utilizes the same pdoping conducting polymer in a symmetric conf iguration on both positive and negative electrodes.74 It can be seen that Kapton-Au sheets (1cm2) were used as the current collector material and two of these substrates were co ated with a film of PProDOP and sandwiched together using a gel electrolyte (LiBTI based) as the conductive media. A porous separator paper was used to prevent possible shorts between th e electrodes in the devices. PProDOP films were made by electrochemical polymerization. One polyme r film was initially set in the fully oxidized state and laminated with a fully neutralized film to establish the initial charge state of the capacitor. Subsequent discharging of the device causes half neutralization of the p-doped films proceeding with concurrent half oxidation of th e cathode films in the de vice. For two-electrode systems, the quantity of charge th at leaves one electrode is equal to that injected onto the other electrode charge. Therefore, it is essential to match charge capac ity of each electrode, thus the amount of deposited polymers on each side prior to device construction in order to achieve
77 optimal and balanced switching, so no over accumulation of charges would occur when the device was switched between its charged and uncharged states. Figure 3-7. A) Schematic diagram of Type I supercapacitor configuration utilizing PProDOP switching between charged and discharged states and B) Photograph of Type I PProDOP supercapacitor with Au/Kapton substrate. Fig 3-8 shows the linear sweep voltammograms and capacitance as a function of applied potential for the capacitor cells made using LiBTI gel electrolyte at different scan rates from 50 to 250mV/s with 50 mV interval s. The capacitance of the devi ces was calculated using the relationship: C=i/s (Equation 2-2). In the Type I device, th ere is a complete overlap between the voltage ranges of oxidation/ne utralization of anode/cathode electrodes since both electrodes utilizes the same polymer. Thus, the cyclic vol tammogram of an ideal Type I supercapacitors would typically have a rectangular or parallelogram shape, showi ng a constant current within the operating voltage window of the device. In th e PProDOP based superc apacitors, the cyclic voltammograms taken at the sweep rates from 50 to 250mV/s are almost close to an ideal shape of rectangle as shown in Fig 3-8A. Such rect angular shape at a constant scan rates is a gel electrolyte / separator Pol y me r 0 Substrate Pol y me r + Substrate gel electrolyte / separator Pol y me r 1/2+ Substrate Pol y me r 1/2+ Substrate discharging charging B A
78 characteristic of a potential independent constant capacitive be havior. The linear scan rate dependence is indicative of non diffusion lim ited redox switching and the good conduction with the gold substrate. Even for higher scan rates of 250mV/s as shown in Fig 3-8A, the rectangular behavior can be observed, which is the indicative of fast switching rate of ions at the sites of electrode/electro lyte interfaces. This result also shows that LiBTI based polymeric gel electrolyte act as good conductive media between the electrodes without any ma jor resistive effect on the kinetics of the diffusion of ions. Moreover, the capacitance response of the device has been found to be independent of scan rates, as can be clearly seen from the Fig 3-6B, which is expected due to the equation of capacitance In conclusion, these voltammograms demonstrate that PProDOP devices show excellent capacitiv e behavior with fast and efficient switching. -1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0 250mV/sI (mA/cm2)E(V) 50mV/s A -1.0 -0.5 0.0 0.5 1.0 -6 -4 -2 0 2 4 6 C (mF/cm2)E (V) 250mV/s 50mV/s B Figure 3-8. A) Cyclic voltammograms and B) capacita nces as a function of applied potentials of Type I PProDOP supercapacitor with Au /Kapton substrates using LiBTI gel electrolyte at 50 to 250mV/s with 50mV/s intervals. The capacitance obtained from a PProDOP supercapacitor was calculated as 3.2mF/cm2. The film capacitances are measured as 8.1mF/cm2 and the correspond ing mass value was calculated from the calibration plot. Finally the specific capacitance of the device was calculated by dividing the capacitance of 3.2mF/cm2 to the total mass of PProDOP on both positive and negative electrodes. The specific capacitance of the device shown in Fig 3-8 was as 30F/g.
79 The equivalent circuit of a two-electrode supercapacitor consists of two electrodes connected in series, so the equi valent capacitance value is given by the reciprocal of the sum of the reciprocals of the individual capacitance values of the two electrodes. This relationship is shown in Eq 3-1 as follows: device 1 2 (3-1) where Cdevice is the capacitance of the device, C1 and C2 are capacitances of the positive and the negative electrodes, respectively. Since the system s presented in this work contain the same polymer on each side, the equation predicts that the specific capacitance of the device, Cdevice will be one half of the capacitance of the individual film electr odes. The measured capacitance (3.2mF/cm2) of the device was found to be in agre ement with the expected value (4 mF/cm2). The difference between the measured and the calc ulated capacitance of th e device was attributed to a substantial value of equivalent series resist ance (ESR), present in real capacitors, especially using gel electrolyte systems as the conducting media. Moreover, PProDOP supercapacitors can be biased in the reverse direction with respect to the positive electrode. As was mentioned in Chap ter I, there is almost a complete overlap between the redox potentials of the anode and cathode redox proc esses in Type I devices. Therefore, the cyclic voltammogram of an ideal Type I supercapacitors would be a rectangular plot, so Type I supercapacitors can be reverse biased because there is no polarity between the electrodes. The voltammograms shown in Fig 3-8 demonstrates this phenomenon since the device was scanned between +0.8V and -0.8V, showing no difference in voltammograms between the anodic or cathodic scanning. The vol tammograms show almost mirror images of each other in the positively and negatively biased directions.
80 The energy density of the device, shown in Fig 3-8, was calcu lated using the relationship; E = QV (Eq 1-2). The average charge density was determined from the cyclic voltammograms and the average voltage of the de vice was taken as the half of the potential difference between the electrodes at the initial charged state of the device (Equation 2-8) since this difference decays to 0V as the device disc harges. The total mass of films on both electrodes were calculated from the calibration plot (Fig 3-6B). The energy density was found as 2.6Wh/kg. Moreover, the corresponding power density wa s determined using the relationship: P= iV (Equation 2-9). The average current density was determined form the cyclic voltammograms. The power density of the supercapacitor was ca lculated as 584W/kg. Since conducting polymer based supercapacitors are analogous of battery systems, the energy and power densities calculated for PProDOP devices, match in the battery vicinity in the electrochemical supercapacitor region on th e Ragone Plot (Fig 1-1). Stability of Type I Supercapacito r Devices w ith Gold Substrate In a supercapacitor device, stab ility is of fundamental impor tance since one of the most important aspects of supercapacitors is the high number of reversible charging and discharging cycles. Some common applications require freq uent cycling, with minimal change in the performance, including hybrid-e lectric vehicles for transient load leveling or capturing the energy used in braking, or star ting diesel vehicles. The char ging/discharging process in supercapacitors move charge and ions only and does not make or break chemical bonds unlike in a battery. Conducting polymer systems undergo vol ume expansion and contraction (swelling and shrinking phenomenon) when the oxidation level is changed due to the inse rtion and expulsion of ion during the redox doping and dedoping processes, respectively. The stability of the PProDOP based supercapacitor devices was examined by exposing the devices to continuous doping and dedoping processes by cyclic voltammetry in a 1V ra nge at a scan rate of 200mV/s, as shown in
81 Fig 3-9A. A cyclic voltammogram of the device switching was recorded in every 100 switches up to a total of 32,767 switches. The average cap acitance for each voltammogram was calculated and plotted as a function of number of cycles. In Fig 3-9B, capacitances are shown as a function of applied potential; very stable and almost c onstant values of capacitances have been observed for the overall voltage range of the device. Fig 3-9A shows a fast de crease (fluctuation) in anodic and cathodic current responses during the first several hundred cycles. This initial 20% of the capacitance loss is attributed to the so called break-in or conditioning period, which is typically observed for conducting polymers films, permitting permeation of the electrolyte ions into the polymer, and may be due to the some unavoidable irreversible charge consumption reactions of possible impurities or side reactions of some species. This break-in period is followed by a very slow decrease of capacitanc e by less than 2% loss of the total capacitance value with over 32,000 cycles. Consequently, th e stabilities of the PProDOP devices were outstanding since the total capac itance decreased by le ss than 20% after few hundred cycles and less than 2% thereafter for over 32,000 cycles. Over all, the devices retained almost 80% of their electroactivity, even after 30000 switches, which is quite attractive for supercapacitors. It should be noted that the all the experiments were conducted on a bench top at ambient conditions, and as can be seen from the photogra ph of a PProDOP supercapacitor in Figure 3-7B, no special procedures were applie d to seal the device. Only regul ar adhesive tape was used to cover the corners of the transpar ency sheet that covered the electrodes. Since all components of the device system, from electrodes materials to separator papers, contri bute to the overall stability, this high cyclic stability of the PPr oDOP supercapacitors even without an air-free sealing procedure indicate that a further optimi zation of each component of the device will allow even more stable switching with less loss of capacitance.
82 0.00.20.40.60.81.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 32767I (mA/cm2)E(V) A 1 0.00.20.40.60.81.0 -4 -3 -2 -1 0 1 2 3 4 C(mF/cm2)E(V) 32767 1 B 05000100001500020000250003000035000 0.0 0.5 1.0 1.5 2.0 2.5 3.0 C(mF/cm2)number of cycles C Figure 3-9. Stability of Type I PProDOP supercapacitors with Au /Kapton substrates using LiBTI gel electrolyte by non-stop cycling at 200m V/s; A) cyclic voltammograms, B) capacitances as a function of applied potentials and C) cap acitances as a function of number of switching cycles (1st, 100th 500th, 1000th and every 5000th cycle).
83 SWNT Films as Electrode Substrates Conventional conducting polym er based super capacitor research has widely focused on metal substrates, for example; gold electrodes. As mentioned throughout the text, the long term stability of conducting polymers can be compro mised by the delamination of the conjugated polymer from the metal surfaces as a consequence of strain cycling.148 Swelling and shrinking of a conjugated polymer causes the polymer to pull its elf from the surface, which leads to reduced electrochemical responses and ev entually delamination and degradation of the polymer film during cycling.12 Also, gold itself can easily delaminate from the surface of plastic substrates through solvation or stress. In a supercapacitor device, a higher the amount of material leads to a higher capacitance for the device. Therefore, in addition to a gold electrode, using a high surface area electrode with a porous structure seemed to be an effective approach toward deposition of higher amount of polymer material pe r unit area, thus improved capacitance. As such, single wall carbon nanotube films have the po tential to replace and possibl y outperform gold electrodes in supercapacitors due to their hi gh conductivity, strong affinity fo r plastic and glass substrates, large surface area for charge injection into the polymer matrix, mechanical flexibility, chemical inertness, and large electr ochemical potential windows.149, 150 Since the polymer itself is a porous material, the three-dimensional SWNTs netw ork would offer more active sites for the electroactive polymer matrix to penetrate, possibly leading to fa ster switching and more stable devices. The major motivation of this approach was to design optimal electrode material for supercapacitors, taking advantage of the porosity of SWNT netw ork electrodes to obtain higher capacitances. The use of SWNT films as current collectors would allow effective switching of higher amount of CPs by increasing the adhesion of PProDOPs to th e substrate within which the polymer is mechanically interlocked, as well as enhancing the stability of device. Previous research has involved the coati ng/wrapping of nanotubes with c onjugated polymers to increase
84 the capacitances of SWNTs.16 In this work, SWNT films were conceived to provi de structural support (much as soil serves as a support to an chor plants via their r oots) to deposit higher amount of PProDOP. Moreover, the supercapacitor electrodes are ge nerally used as coatings of active material mixed with binders. In our a pproach of using porous network of SWNTs as current collector, the binding element is elimin ated which has a negative effect on the overall capacitance due to the increased contact resistance between the active material and the current collector and also the increased total mass of th e cell. In this work, SWNTs films are used as supportive electrodes and current collectors for the conductive polymer matrix for the first time. SWNTs Film Preparation Single wall carbon nanotube film s were prepared via a method developed in the group of Prof. Rinzler (UF Physics).41, 151 SWNTs grown by laser ablation were used in this study. Purified SWNTs were suspended in an aqueous 1% Triton-X-100 solution with a nanotube density of 0.001 mg/mL by ultra-sonication. Triton X-100 solution ac ts as stabilizing agent to suspend the nanotubes. Then this dilute surfacta nt based well dispersed suspension of purified nanotube solution vacuum-filtered through a memb rane (Millipore, VCWP, 100 nm pore) as illustrated in Fig 3-10. After the nanotube films were deposited evenly on the membrane surface, the films were washed away with DI water several times to remove the residual surfactants. The amount and thickness of the nanotub e films were controlled by the density and the amount of the nanotube suspension solution used. Unless otherwise is indicated, the nanotubes used for this study were 80nm in thickness. Once these highly uniform films on the filtration membrane were dried, they were cut into de sired shapes (area of 1cm2 in this case) and th e electrodes prepared by transferring these films to glass or plastic substrates. The ends of the substrates were previously coated with Cr/Pd as contact pa ds by either sputtering or therma l evaporation. This procedure is as follows: the membranes were applied to the subs trates face to face in contact with Cr/Pd from
85 the edge of the nanotube films and then were pl aced onto a previously wetted Teflon block. They were sandwiched in between several filter papers for cushioning and pressed between two metal bricks by spring clamping and left in 80C oven over night. As the water evaporates, the nanotube films firmly adhere onto the substrate. Finally, the dried substrates were removed away from press and the cellulose ester membranes were completely removed by dissolving in acetone vapor (or methanol) bath followed by washing rep eatedly in 8 fresh acetone baths to ensure the removal of any residual cellulose ester on the film surfaces. The resul ting films were highly uniform and electrically conduc tive, ranging from 600 to 6,000 S/ cm. Moreover they were highly thermally and chemically stable. To minimize contamination from extraneous dust, this procedure was performed entire ly in a Class 100 clean room. Figure 3-10. Fabrication of SWNT Film. Vacuum filtration of a surfactant based suspension of SWNTs, forming the film as a cake on the porous filtration membrane. As can be clearly seen from the AFM image of the obtained nanotube film shown in the Fig 3-11, they were highly porous and had a larg e surface area (high specif ic surface area). The randomly but homogeneously distributed na notube bundles throughout the film provide sufficient overlapping of nanotube s to result in good electric conductivity. This highly uniform morphology of the films is ideal for use as the electric current collector in polymer based
86 supercapacitor devices. The porous nature of the films was expected to allow interpenetration of active polymer into the matrix, enhance the amount of deposited polymer and improve the adhesion of the polymer by interlocking the polymer within its matrix. Thus, deeper and faster switching of the devices would be possible. The low mass density of the films would allow higher specific surface capacitances by lowering th e mass contribution of the electrode substrates of the device. Figure 3-11. Atomic force micrograph of SWNTs film mounted on sapphire wafer. Comparison of Gold and SWNTs Subs trates in Device Performances After preparation of SWNTs fil m, ProDOP wa s electrochemically polymerized onto these substrates. As mentioned before, in order to ob tain smooth uniform films of PProDOP as on flat gold substrates, a constant potential electrochemi cal polymerization method was applied. Moreover, different polymerization techniques of cyclic voltammetry and potentiostatic pulse electropolymerization (for the detailed explanati on of this technique go to section 3.13) were applied as well. Once formed on the electrode, the polymer films were thoroughly soaked and rinsed with ACN and equilibrated with monomer free electrolyte solution. As was in gold electrodes, the capacitances obtai ned from SWNTs based electrodes were found to be in a linear
87 relation with increasing the PProDOP deposition charge (and numb er of deposition steps, and mass of the PProDOP). However, very similar to gold electrodes, as the PProDOP film became thicker, the film delaminated from the nanotube surfaces upon washing process. This was surprising since the PProDOP layer was expected to adhere onto the SWNTs substrate better than on the gold electrodes. In Fig 3-12, cyclic voltammogram and the capac itance graph of the device prepared with SWNTs electrodes using LiBTI ge l electrolyte, at scan rates from 50 to 250mV/s with 50 mV intervals are shown. The capacitance obtained wi th PProDOP on SWNTs electrode devices was found to be 3.8mF/cm2, only slightly higher than the gold electrode devices (3.2mF/cm2). The voltammograms of the device show typical rectangul ar shape as was in gold substrate devices. At higher scan rates, more than 100mV/s, the device did not show an efficient switching and a slight deviation from the rect angular shape of the voltammogram s has been observed, which is attributed to a substantial value of equivalent series resistance (ESR). These results show that PProDOP probably does not diffuse into the poro us network of SWNTs and probably interacts very poorly with the nanotube surfaces. 0.00.20.40.60.81.0 -1.0 -0.5 0.0 0.5 1.0 250mV/sI (mA/cm2)E(V) 50mV/s A 0.00.20.40.60.81.0 -4 -2 0 2 4 C (mF/cm2)E(V) 250mV/s 50mV/s B Figure 3-12. A) Cyclic voltammograms and B) capaci tances as a function of applied potential of Type I PProDOP supercapacitor with SWNTs substrates in LiBTI gel electrolyte at 50 to 250 with 50mV/s intervals.
88 Besides linear sweep voltammetry, a constant current charge/discharge technique (chronopotentiometry) was also used to character ize the device performances. The devices were subjected to a cyclic square-wave galvanostati c current, and the voltage response was measured as a function of time. Th e discharge capacitance, Cd was evaluated from the linear part of the discharge curves us ing the relation: Cd = i t/ V (Equation 2-2), where i is the constant current and t is the time interval for the voltage change of V Typical charge/discharge plot of a capacitor usually demonstrates a symmetric linear increase and decrease as a function of time. Fig 3-13A&B, show the galvanostatic chargi ng and discharging cycling of both gold and SWNTs film based devices, respectively.The devices were charged to 1.0V and discharged at a constant current density of 0.5mA/cm2. The IR drops upon discharging in both devices is due to the internal resistance of the supercapacitors. Af ter the IR drop, the discharge behaviors of the devices are linear with respect to time. Th e capacitance values are calculated as 2.9mF/cm2 and 3.5mF/cm2 for gold and SWNTs based devices, respectiv ely. These results are in agreement with the capacitances values calculated from the cycl ic voltammograms of the devices. The IR drop observed for gold substrate devices, as shown in Fig 3-13, arises probably from the interfacial resistance between the electrode/electrolyte cont acts and penetration of electrolyte into the polymer matrix. The doping and dedoping process is a relatively slow reaction compared to electronic charging. As a result the IR drop app ears upon switching of the device, which is also seen in the cyclic voltammograms as the deviatio n of the shape with the increased scan rates. However in Fig 3-13B, the IR drop observed in th e SWNTs based device is higher than the gold device, showing the higher intern al resistance in the device. This internal resistance most probably arises because of the interfacial re sistance between the charge storage material (PProDOP) and the electrode substrate (SWNTs film). Moreover, a decrease in the total
89 discharge time of the devices with the SWNTs substrate devices, compared to gold substrate devices, was observed for the same current density application. This feature can be attributed to increased accessible surface area of the substrate so the total electrode area in contact with PProDOP. As a result, SWNTs based devices res pond faster to discharging, however create a high resistance inside the cell. Figure 3-13. Constant current (0.5mA/cm2) charging/discharging of Type I PProDOP supercapacitor with A) Au/Kapton and B) SWNTs film substrates. Non-Covalent Modification of SWNT Surfaces w ith Sticky-PF The delamination of the PProDOPs upon electropolymerization suggests that the coordinative interaction of the polym er with the electrode surface is poor, therefore it is surmised that the surface of carbon nanotubes, due to their low energy graphene nature, would only weakly interact with co njugated polymers through -stacking interactions. Some recent research studies have also shown that conducting pol ymers exhibit poor adhesion and mixing with nanotube materials due to nanotubes low energy graphene surfaces.152 Furthermore, PProDOP is polar and mildly hydrophilic, whereas SWNTs are hydrophobic, and PProDOP is sterically encumbered due to the bulkiness of the alkylene bridge, so its interfaci al adhesion via van der Waals interactions with the surface of a SWNT is poor. Due to this lack of affinity, shear stress, bending, other mechanical deformations, as well as electrochemically induced morphological -0.2 0.2 0.6 1.0 020406080E (V)t (s)A -0.2 0.2 0.6 1.0 020406080E (V)t (s)B
90 changes (which has been previously observed for polypyrrole and PProDOP)153 induces poor adhesion of the polymer to the surfaces of nanotubes and finally delamination of the polymer from the nanotube film substrates, thus destroying the device. The major motivation for using SWNT films as substrate was to increase the adhesion of higher amounts of PProDOP to the substrate w ithin which the polymer is mechanically interlocked by taking advantage of the mesoporosity of SWNT network electrodes. In order to fully utilize the advantages of porous network of SWNTs, and also to realize the benefit of SWNTs as current collectors with high charge dy namics and long term stability in supercapacitor performance, it was necessary to electrically c ouple the active polymer strongly to the nanotube surface. Such electrical coupling requires intimate proximity between the organic layer and the nanotube surface. To accomplish this, we presen t a novel method of introducing an interfacial compatiblizer (a buffer layer) between the PProDOP and the SWNT surfaces, which would overcome the aforementioned interf acial challenges, and signifi cantly improve polymer adhesion to the substrate and thus improve the device performance. This method involves non-covalent functionalization of carbon nanotube surfaces through polycyclic aromatic hydrocarbons, specifically pyrene derivatives to improve the interface between nanotube surfaces and organic molecule.154 Such non-covalent interaction with th e nanotubes permits association of other molecules with the nanotubes while minimally im pacting their intrinsic electronic transport properties.155 A polyfluorene derivative; poly (9,9-dioctylflouorene) with pyrene pendant groups (Sticky-PF)122, whose structure is illu strated in Fig 3-14 with the proposed non-covalent association on the surface of na notube, was synthesized by Dr. Ryan Walzcak in the Reynolds group. Thereby this compatiblizer la yer would allow an increased a ssociation of the electroactive PProDOP to the SWNT electrode surface.
91 Figure 3-14. Structure of poly( 9,9-dioctylflouorene) (Sticky-PF) and its proposed non-covalent association with a SWNT surface. The utility of non-covalent functionalization, as opposed to irreversible covalent functionalization, is that small molecules or po lymers can be associated with the nanotube surfaces without significantly altering the nanot ubes electronic propertie s. This non-covalent functionalization can be achieved by the attachment of a polycyc lic aromatic hydrocarbon such as pyrene to the molecule desired for attachment.156 Furthermore, Sticky-PF were designed to have sufficient plurality of polycyclic aromatic hydrocarbons as pendant groups along a polymer chain, so the association/dissociati on kinetics attributed to single molecules would be sufficiently attenuated as to render the polymer association w ith the carbon nanotube side wall, for all intents and purposes, permanent. The octyl functionali ties were coupled to the polyflouorene backbond to increase the solubility of the polymer in co mmon solvents for ease of application processes. The deposition of Sticky-PF was achieved by immersion of the bare SWNT films into a solution of Sticky-PF/Chloroform (0.5mg/ mL) overnight, where the polymer would O O C8H17 C8H17 Sticky-PF
92 spontaneously assemble onto the nanotubes, follo wed by subsequent soaking of the film in pure chloroform for several hours to remove unads orbed material, then repetition of the same procedure one more time, followed by washing with methanol. These films were then used as electrodes for electrochemical deposition of PProDOP. To minimize contamination from extraneous dust, this procedure was performed entirely in a Class 100 clean room. To examine the attachment of Sticky-PF layer on the na notubes, the surface topography of 7-nm thin nanotube film networks deposited on silicon wafe rs (witness chips) were measured with AFM. While the nanotubes appeared as thin and smooth bundles before the Sticky-PF coverage as can be clearly seen in the Fig 3-15A, rougher surface of bundles were obser ved after the solution dipping process of the film to Sticky-PF soluti on. As shown by AFM image in Figure 3-15B, the adsorption of Sticky-PF onto the SWNT surface was evident by the widened girth of the nanotube bundles, as well as some attached spheroid nanostructures. It is believed that this image demonstrates that a fairly homogeneous monol ayer coating was deposited during the selfassembly process. It is also important that th e Sticky-PF were adsorbed on the SWNTs only as a thin layer and did not fill up the porous struct ure between the bundles which would prevent the ProDOP diffusion into the porous network of SWNTs film. For the remainder of research presented in this chapter, all of the SWNT films used were coated with Sticky-PF unless otherwise specified. In addition to AFM images, the absorbance sp ectra of self assembled Sticky-PF covered SWNT films are presented in Fig 3-16 with comp arison to the solution sp ectra of Sticky-PF in chloroform. It is obvious that there is a self-a ssembled layer of StickyPF present on the naotubes film since both spectra show the same peaks belonging to th e pyrene subunits.
93 Figure 3-15. Atomic force micrographs of A) SWNTs film on Si wafers before Sticky-PF coating and B) SWNTs film after Sticky-PF coating. 280300320340360380400420440460 0.0 0.2 0.4 0.6 0.8 1.0 BNormalized AbsorbanceWavelength (nm) A Figure 3-16. Absorbance spectra of A) Sticky-PF coated SWNT film and B) solution of StickyPF in chloroform. The Sticky-PF would be expected to have a ne gative effect on the conductivity of the nanotubes film, since it is moderately insulating, however the sheet resistances of SWNTs film were found to be very similar as 92 and 89 / before and after the Sticky-PF coverage, B A
94 respectively. These results show that coating SWNTs with Sticky-PF does not render the film electronically passivated and charge carriers can be tunnel through th is very thin adsorbed layer of Sticky-PF. Accordingly, the non-covalent functionalization of carbon nanotube surfaces through pyrene units have minimal impact on the intrinsic electronic transport properties. Electropolymerization of ProDOP on Bare and Sticky-PF coated SWNT Films Cyclic voltammetry offe rs a powerful method for the characterization of the monomer and polymer redox processes. Since the electropolym erization of ProDOP on gold substrates was already examined, it was necessary to observe this process on the Sticky-PF coated SWNTs electrodes. After coating of the SWNTs film wi th Sticky-PF using a solution-dipping process, ProDOP was electrochemically polymerized onto these substrates. Figure 3-17 demonstrates the representation of this process. Figure 3-17. Schematic representation of PProDOP electrodeposition onto Sticky-PF|SWNTs film. (grey: SWNTs, orange: Sticky-PF, blue: PProDOP)
95 PProDOP films were synthesized electroche mically on both bare SWNT film and StickyPF coated SWNT film by linear sweep cyclic voltammetry. Fig 3-18 shows the voltammograms at a rate of 25mV/s of these processes on both substrates. Electrochemical polymerization processes are expected to differ slightly from each other when any of the parameters of the system, such as substrate electrodes or solution media are changed. As illustrated in the cyclic voltammograms in Fig 3-18, ProDOP behaves quite differently on a Sticky-PF coated SWNTs film than on a bare SWNTs film. Cyclic vo ltammogram of ProDOP on Sticky-PF coated SWNTs film show consistent behavior with the ProDOP deposited on gold (Fig 3-2), most notably the irreversible monomer oxidation peak with the absence of a reductiv e wave which is indicative of radical-radical coupling, and form ation of polymeric material. -0.4-0.20.00.20.220.127.116.11 0.0 0.3 0.6 0.9 0.0 0.3 0.6 0.9 BE(V) vs. Fc/Fc+I (mA/cm2)A Figure 3-18. Electrochemical polym erization (first 5 cycles) of ProDOP by potential scanning in LiBTI/ACN on A) Sticky-PF|SWNTs and B) bare SWNTs at 25mV/s.
96 As was observed in electrodeposition of Pr oDOP on gold electrodes, a shift in the monomer oxidation potential was evident with the progression of each cycle in the case of Sticky-PF coated SWNTs. As explained in the prev ious sections, this shifting in the oxidation potential might be caused by a combination oxidation of monomers and coupling of soluble oligomers, which have more lowe r oxidation potential than monomers.142, 143 Moreover, the cyclic voltammetric system, after several repeated polymerizations occurr ing under identical conditions, the process of adding new polymer to the electrode surface chemically modifies the working electrode. These reasons combine to yi eld broader and shifted peaks appear with the progression of cycling in the cas e of Sticky-PF coated SWNTs. Not surprisingly, the oxidation potential of ProDOP (Ep,m) on Sticky-PF coated SWNTs was found to be 0.76V vs. Fc/Fc+, a slightly higher potential than Ep,m on the gold electrode 0.65V vs. Fc/Fc+. While the actual process of coating of the SWNTs surfaces with Sticky-PF produces a chemically modified current substrate electrode covere d with an essentially insulating polymer, the Helmholtz layer at the electrode surface was expected to be aff ected, thereby increasing the potential energy required for monomer oxidation due to th e addition of a t unneling barrier. On the other hand, there was no monomer oxida tion peak observed on a bare SWNTs film as it was on the other electrodes, yet polymer films were formed on both bare and Sticky-PF coated SWNTs film substrates. As explained earlier, delamination of PProDOP film, when potentiostatically deposited on a bare SWNTs film was observed during the redox cycling of the film in a monomer-free electrolyte solution. Th e polymer film was observed to fall from the electrode surface upon washing or after several sw itches and float in solution. As discussed above, this delamination was most likely due to th e inability of the PProDOP film to adhere to the SWNTs to their low energy graphene surfaces.157 The lack of both a well-defined monomer
97 oxidation peak and polymer waves in the CV of PProDOP deposition onto the bare SWNT films was probably due to this poor coordination of the molecules with nanotube surfaces. The increased currents due to the re peated scanning demonstrate the formation of a polymer film on both electrodes; however, the nucleation loop seen in the first scan of electropolymerization on Sticky-PF coated SWNTs, corresp onding to the presence of ad sorbed polymer increasing the working electrode surface area, and also the increased current densities, corresponding the polymer redox process in the case of Sticky-PF coated SWNTs, are indicative of a more efficient electropolymerization. These results were most likely due to the bett er interaction of the polymer molecules to the electrode surface through Sticky-PF as an interfacial compatiblizer. As explained in the previous section, the c onductivity of SWNTs film was not affected by Sticky-PF coverage. The voltammogram of the el ectropolymerization of ProDOP on Sticky-PF coated SWNTs film offer further proof that the non-covalent functionalization of carbon nanotube surfaces through pyrene units does not render the SWNTs electrochemically passivated. It is obvious that char ge carriers can tunnel th rough the thin layer of adsorbed StickyPF, as shown in Figure 3-18A. Moreover, no poly mer delamination was visually observed after subsequent electrodeposition of pol ymer film onto the Sticky-PF m odified SWNTs. These results effectively demonstrate that Sticky-PF as an in terfacial compatiblizer effectively enhances the interaction of PProDOP with the SWNT surface. Electrochemistry of PProDOP Film on Sticky-PF coated SWNTs Film In order to obtain uniform fil ms of PProDOP on the porous 3D-network structured SWNTs electrodes, the potentiostatic wa ve electropolymerization method wa s applied, in contrary to the constant potential electrochemical polymerization method that was used for the deposition of PProDOP on flat gold substrates. This method i nvolves step-wise constant potential application at determined time intervals. The potential st epped between the monome r peak potential of
98 0.76V vs. Fc/Fc+ plus 0.05V and a potential of 0.4V vs Fc/Fc+ where the resulting polymer is in oxidized and neutral state, resp ectively. The potential stepping wa s repeated for 10s intervals until the desired amount of material was depos ited. The solution was gently stirred during electropolymerization process to preserve the homogeneity of th e monomer solution especially in the vicinity of the electrode. This method is somewhat analogous to the cyclic voltammetry deposition method and projected to be the mo st suitable synthetic method for any porous structured substrates like SWNTs film. Application of the oxidative and reductive (neutralization) potentials alternatively with ti me intervals would allow the constant potential polymerization and also permit sufficient time fo r the monomer species to diffuse through the channels of porous network in order to mainta in the monomer-electrolyte concentration at equilibrium in the pores of th e nanotube film to prevent diffusion limited polymerization to occur. On the other hand, this method allows for the reorganization of the already formed polymers in between the pores once the neutraliza tion potential applied alternatively with the oxidative potential, to obtain smooth hom ogenous films of PProDOP. Once formed on the electrode, the polymer films were thoroughly soaked and rinsed with ACN and equilibrated with the monomer free electrolyte solution as explai ned in the previous sections and voltammograms of the switching of polymer films were then obtained. To determine the extent of capacitance c ontribution from Sticky-PF which acts as a compalitiblizer layer between the nanotubes a nd PProDOP, the voltammograms of PProDOP on a Sticky-PF coated SWNTs electrode were comp ared with the CV of Sticky-PF coated SWNTs electrode prior to PProDOP deposition as well as pristine SWNTs. As e xpected, no distinction was observed between the electrochemical respon se of pristine and Sticky-PF coated SWNT films, indicating that Sticky-PF did not appreciably perform as a charge storage material. The
99 capacitances obtained from those films were found to be 0.09mF/cm2 due to the double layer charging of the surface. In the Figure 3-19, CVs of Sticky-PF coated SWNTs electrodes before and after PProDOP deposition are shown at a scan rate of 50mV/s between -0.5and +0.5 V vs Fc/Fc+. It can be clearly seen that the capacitanc e contribution of the el ectrode before PProDOP deposition is negligible. -0.6-0.4-0.20.00.20.40.6 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -20 -16 -12 -8 -4 0 4 8 12 16 20 E (V) vs. Fc/Fc+BI (mA/cm2)C (mF/cm2) A Figure 3-19. Cyclic voltammograms of A) PProDOP on Sticky-PF|SWNTs and B) bare StickyPF|SWNTs film in LiBTI/ACN at 50 mV/s. As in the case for PProDOP on gold electrode s ubstrates, in order to relate the deposition charge and more importantly the capacitance values with the mass of PProDOP, calibration graphs were prepared for the films prepared on the Sticky-PF coated SWNTs film substrates. After the electrodes with varying amount of PPro DOP were prepared, the capacitance values of the films and the mass of the polymer material were measured. The ma ss of the 80nm thickness film of SWNTs was determined by using the de nsity of the SWNTs. The mass of the SWNTs were subtracted from the measured mass of th e PProDOP deposited substrates in order to accurately obtain the mass of PProDOP. As for PProDOP on gold electrode substrates, a linear relationship between the deposition charge fo r electropolymerization, shown in Fig3-20A, and
100 also the capacitances of the resulting films w ith the mass of the PProDOP, shown in Fig3-20B, were found. Unlike the PProDOP deposition on gold electrodes, or bare SWNTs films, after certain amount of PProDOP was deposited, oligomer formation, which was clearly observed by the coloration of the monomer so lution, started to occur during de position. After th is point, the mass of the polymers deposited was difficult to determine accurately. In Fig 3-20 the capacitances as a function of th e mass of PProDOP on Sticky-PF coated SWNTs film are shown. A very small intercept at the onset of polymer ization originates from the double-layer charging of the Sticky-PF coated SWNTs electrodes. From the slope of the curve, the average specific capacitances of the PProDOP films were calculated as 105F /g. Although, areal capacitances (mF/cm2) of PProDOP are found to increase with Stic ky-PF|SWNTs electrodes, compared to the Au electrodes (141F/g) the average mass capacitan ces of PProDOP were decreased by 26% with Sticky-PF|SWNTs. This is probably due to th e morphological differences of PProDOP films formed on gold and SWNTs electrodes. As previ ously mentioned, the capacitance of the polymer strongly depends on the morphol ogy of the polymer. The ion pe netration abil ity through the channels of the porous network of the polymer re stricts the capacitance of the material. As will be discussed below, PProDOP forms a more densely structured morphology when deposited on SWNTs than a flat metal surface. Therefore, a sma ller percentage of the total mass of the charge storage material contributes to the overall cap acitance, resulting decreased mass capacitances. The cyclic voltammograms of the films with highest deposited amount of PProDOP with the scan rate dependence for the redox switchi ng are shown in Figure 3-21. CVs of the films were taken at different scan rates from 50 to 250 mV/s between .5 and +0.5V vs Fc/Fc+. At slower scan rates of 50 and 100mV/s, the typical and previously obtained PProDOPs electrochemical response was observed. As the scan rate was increased, a deviation was
101 observed, which was attributed to slower ion di ffusion through the polymer matrix with respect to the fast potential change. This deviation wa s expected due to the relatively thick polymer layer. As the scan rates were in creased, the ability of ions to pe netrate through this thick polymer layer was lowered, and as a result the overall resistivity of the el ectrode material increased and the maximum capacitance became restricted by electrolyte diffusion kinetics. Figure 3-20. Calibration curves; A) Deposition char ges of ProDOP and B) capacitances of PProDOP as a function of the mass of PProDOP on Sticky-PF|SWNTs substrates. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 050100150200Qdep(C/cm2)mass (g/cm2)A 0 2 4 6 8 10 12 14 16 18 20 0 50 100 150 200C (mF/cm2)mass (g/cm2)B
102 -0.6-0.4-0.20.00.20.40.6 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 250mV/s 50mV/sI (mA/cm2)E ( V ) vs. Fc/Fc+A -0.6-0.4-0.20.00.20.40.6 -30 -20 -10 0 10 20 30 50mV/sC (mF/cm2)E(V) vs. Fc/Fc+B 250mV/s Figure 3-21. A) Cyclic voltammograms and B) capacitances as a function of applied potential of PProDOP films on Sticky-PF|SWNTs in Li BTI/ACN at 50 to 500mV/s with 50mV/s intervals. The highest capacitance achieved with PProDOP films formed on Sticky-PF coated SWNT electrode reached a stable value of 25mF/cm2 without degradation or delamination of the film. The cyclic voltammograms at different scan ra tes from 50 to 250mV/s with the capacitance graphs of this film are shown in Fig 3-21A and b respectively. The co rresponding mass of these PProDOP films was calculated as 239 g/cm2 from the calibration curve. Surface Analysis of PProDOP Film on Sticky-PF Coated SWNTs Film The structure of the PProDOP film s on S ticky-PF coated SWNTs electrodes were examined using a scanning electron microscope. A 45 tilted angle sample holder was used in order to obtain the highest quality resolution. In order to examine the growth morphology of PProDOP polymerization, films with varying am ount of PProDOP were prepared as in the calibration curve. PProDOP was deposited in various amounts using a wave potentiostatic electrodeposition method. Fig 3-22 demonstrates this deposition morphology of PProDOP on Sticky-PF coated SWNTs.
103 The deposition of the film begins with the form ation of an extremely thin uniform layer of PProDOP. This thin layer is a continuous layer of PProDOP dire ctly attached to the nanotubes and fill up the pores of the three dimensional nanotube network. In Fig 3-22A, the noodle-like structure of the nanotube bundles can be easily realized underneath the thin layer of the polymer. Some irregularities formed, as s een in the SEM image of this ear ly stage of film deposition. As the polymerization proceeded, the noodle-like stru cture of the SWNT film electrode could no longer be discerned since PProDOP layer becam e thicker. Further polymerization allowed thicker films of PProDOP to be formed. In Fig 3-22C, the SEM image of the cross section of a relatively thick film is shown and the nanotubes b undles that are broken off and sticking out of the polymer layer are clearly visible. Figure 3-22. Scanning electron micrographs of PProDOP films on Sticky-PF|SWNTs substrates; A) at the start of the deposition, B) during the progress of the de position and C) after formation of a thick film at varied magnification. The films demonstrate more compact and de nse morphologies when deposited onto the Sticky-PF coated SWNTs compared to deposition on gold. These structures show consistency with the calculated specific capacitance results. The polymers probably have a high affinity for the Sticky-PF coated porous nanotube substrate and interact well with (or diffuse into) the substrate, as shown in the Fig 3-22C. As pred icted, the three dimensi onal network of SWNTs serves as a structural support for PProDOP layer, much like a plant anchors itself via roots in the soil. This change allowed the delamination of the polymer from the electrode surface to be A 500nm 857nm 1 6 7 m B C
104 eliminated. Moreover, the conductivity of the SWNTs dispersed throughout the PProDOP structure must increase the electr ical conductivity of the polymer film, so the charge injection into the charge storage medium (PProDOP) is greatly enhanced by the accessible surface areas of the SWNTs film, compared to that of flat surface electrodes. Type I Supercapacitor Devices with Sticky-PF Coated SWNT Substrates The cyclic voltammograms and capacitance graphs of supercapacitors prepared using LiBTI gel electrolyte with PPro DOP deposited on Sticky-PF coat ed SWNTs film, at the sweep rates from 50 to 250mV/s with 50 mV increments are shown in Fig 3-23A&B respectively. As in supercapacitors with gold substrates, these su percapacitors also demonstrate shapes which are very close to the ideal parallel ogram plot of Type I devices. Although a slight deviation in the parallelogram shape occurs as the scan rates increases, the device still shows an effective switching curve at higher scan rate of 250mV/s. 0.00.20.40.60.81.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 250mV/sI (mA/cm2)E (V) 50mV/s A0.00.20.40.60.81.0 -10 -5 0 5 10 15 BC (mF/cm2)E (V) 250mV/s 50mV/s Figure 3-23. A) Cyclic voltammograms and B) capacitances as a function of applied potential of Type I PProDOP supercapacitor with Stic ky-PF|SWNTs substrates using LiBTI gel electrolyte at 50 to 250mV/s with 50mV/s intervals. The capacitance values achieved with PPro DOP on Sticky-PF coated SWNTs were found to be 9.7mF/cm2. As stated above, the corresponding ma ss of PProDOP of this film (25mF/cm2) was calculated as 239 g/cm2 from the calibration curve. So th e specific capacitance of the device
105 was calculated by dividing the measured capacitance (9.7mF/cm2) to the mass of two electrodes (2*239 g/cm2) and found as 20F/g. Since the capacitances of the device is expected to be the half of the capacitance of the film (25mF/cm2), as explained in the section 3.6. Th e capacitance of the device is 9.7mF/cm2 and 77% of the expected value of 12.5mF/cm2. This result demonstrates that the difference between the measured and the calculated capacitance of the device can be attributed to a substantial value of equivalent series resistance, especially switching from solution-based measurements for films to gel electrolyte systems in case of devices. The energy and power density of the device, shown in Fig 3-23, were calculated as 1.3Wh/kg and 466W/kg, respectively. Table 3-1 summarizes the areal and specific capacitances of PProDOP films, along with the energy and power densities determined for PProDOP Type I supercapacitors on gold and Sticky -PF coated SWNTs substrates discussed in this chapter in comparison to each other. Table 3-1. Capacitances, energy and power densities of PProDOP films and devices. Substrate Film Device C(mF/cm2) C (mF/g) C(mF/cm2) C (mF/g) E (Wh/kg) P(W/kg) Au|Kapton 8.1 141 3.2 30 2.6 584 Sticky-PF|SWNTs 25 105 9.7 20 1.3 466 Stability of Type I Supercapacitors wi th Sticky-PF coated SWNTs Substrate The cyclic stability of the Sticky-PF|SWNT based devices were examined by continuous CV switching for over 32700 non-stop cycles at a s can rate of 100mV/s, as shown in Fig 3-24. The corresponding voltammograms of films were shown in Fig 3-21, which are the highest capacitance values obtained with PProDOP on the Sticky-PF|SWNT s film substrates. Voltammograms of the device switching were reco rded in every 100 switc hes up to a total of 32767 switches. The average capacitance for each vo ltammogram was calculated and plotted as a
106 function of number of cycles. As mentioned in the previous sections, conducting polymers require a preconditioning period in order to ob tain reproducible CVs, permitting complete permeation of the electrolyte into the polymer. The fast decrease of capacitance (~1mF) seen in the first several hundred cycles ar e due to this break-in peri od. There after the capacitance of the device continues to decrease slowly for the duration of the experiment. In contrast to gold devices, the Sticky-PF|SWNT devices show a st eady loss of capacitance to about 35% of the original value. As mentioned in the previous sections, swe lling and shrinking of the polymer occur during doping and undoping processes. This phenomenon surm ised to cause the polymer to peel off from the gold substrate, especially as the adhe sion gets weaker with thicker polymer films. However using the porous structured SWNT film substrates prevented the poor adhesion issue particularly with thicker polymer layers. Neve rtheless, the intrinsic swelling and construction had a negative effect on the overall stability, theref ore performance constraint occured especially with increased film thicknesses. Since each com ponent of a device system from electrodes to separator paper contribute s to the overall stability, further op timization is necessary to increase the stability of these Sticky-PF| SWNT based devices. These devi ces were prepared on a bench top at ambient conditions. The aim of this study was to prepare an electrode substrate that would allow higher amounts of PProDOP material to be deposited fo r switching. However, while the switching of the devices was not as stable as gold devices, the capacitance values achieved with Sticky-PF|SWNT based devices were initially a lmost four times of the capacitance of gold devices, and remained as high as three times even after 15000 cycles switching. StickyPF|SWNT based devices maintain a greater capacitance than the gold device for overall 32000 cycles, which is very encouraging.
107 0.00.20.40.60.81.0 -1.0 -0.5 0.0 0.5 1.0 I (mA/cm2)E(V)32767 A 1 0.00.20.40.60.81.0 -10 -5 0 5 10 E(V)C(mF/cm2)32767 B 1 05000100001500020000250003000035000 0 1 2 3 4 5 6 7 8 9 10 C(mF/cm2)number of cycles C Figure 3-24. Stability of Type I PProDOP supe rcapacitors with Sticky-PF|SWNTs substrates using LiBTI gel electrolyte by non-stop cycl ing at 100mV/s scan rate; A) cyclic voltammograms, B) capacitances as a function of applied potential and C) capacitances as a func tion of number of switching cycles (1st, 100th 500th, 1000th and every 5000th cycle).
108 Conclusions and Perspective In conclusion, this chapter presents the utilization of PProDOP as the charge storage material in Type I supercapacitors. While the ultimate goal of this chapter to demonstrate PProDOP as the charge storage material, differe nt electrode designs have been attained and many key concepts and methods were also demonstrated and established. The supercapacitor devices were constructe d using two different current collector substrates: Kapton-Au sheets and SWNT films. In the latter project, the goal was to utilize the charge storage material on a porous structured substrate to realize the penetration of the material through it, thus to led more loading of PProDOP per unit area. To the best of our knowledge, this is the first presentation of the use of SWNT s for such a purpose. Although a delamination problem was occurred between PProDOP and SWNT s layers, in order to overcome this problem, a non-covalent modification of the surface of carbon nanotubes have been conducted. On this purpose, a new polymer; Sticky-PF was synthe sized. It has been observed that PProDOP electrodes containing porous subs trates can provide a significant areal capacitance improvement over conventional non-porous flat metalized substrates. In summary, the non-covalently modified SWNT films with Sticky -PF are presented as a novel t ype of supercapacitor electrode substrate in this chapter. A ny supercapacitor utilizing varying charge storage materials can benefit from this method. All PProDOP films were deposited by electr ochemical polymerizati on techniques in the presence of different supporti ng electrolytes. The electrochemi cal and morphological properties of the resulting films were investigated by using cyclic voltammetry, galvanostatic chargingdischarging methods and also SEM and AFM techniques. PProDOP exhibit homogeneous morphology with porosities up to varying levels in the presence of different supporting electrolytes and current collector substrates. The results show that PProDOP exhibits almost
109 ideal rectangular CV behavior of a supercapacitor material with LiBTI as the suppoting electrolyte. Moving from electrode substrates with flat metalized surfaces to the ones with porous structures were led to more than 3 fo ld increase in capacitances of PProDOP per unit area. On the other hand the specific capacita nces (per total active material) show a 26% decrease. Correspondingly the specific capacitance values of the supercapacitors were found to be diminished with Sticky-PF/SWNTs substrates comp ared to Kapton-Au substrates, however, the areal capacitances were calculated as 3.2 mF/cm2 and 9.7mF/cm2 for the devices with KaptonAu and Sticky-PF/SWNTs substrates, respectively. The devices with both of the substrates, but especially the ones based on Kapton-Au, showed excellent stability during long cycles with a moderate loss of charge. Future directions for this wo rk would include optimization of the device performances for varying intentions; this would entail determinin g the optimum amount of the polymer material (or thickness) for the highest cap acitance performances. In furthe r studies, SWNT films with increased porosity or a SWNT film template with particular order (perpendicularly aligned bundles) can be utilized in conduc ting polymer based supercapacitors in order to increase the polymer loading per unit area. Moreover, ut ilization of solution pr ocessable conjugated polymers would be a major advantage for c onstruction of large area supercapacitors.
110 CHAPTER 4 TYPE IV SUPERCAPACITORS: DONOR-ACCEPTOR-DONOR SYSTEMS Introduction Pseudocapacitan ce arises from an electrochemical Faradaic charge transfer processes, where the extent of the reaction Q is a function of the cell potential. Duri ng discharging of an ideal battery cell exhibiting a Nernstian cell poten tial, the progressively increasing charge is being added at a relatively constant voltage that remains ideally constant until all the reactant materials have been electrochemically consumed. Un like this Nernstian behavior in batteries, in pseudocapacitance proce sses, the extent of Faradaically admitted charge depends almost linearly, on the applied voltage. In other words, the potential at which the charge is being passed is a continuous f unction of the charge accumulate d. Consequently, a capacitive phenomenon occurs with the arising derivative of dQ/dV having the properties of a capacitance. As introduced in Chapter I, the energy stored electrochemically in a battery is given as QV where V is the voltage of the cell and Q is the cell charge capacity ; the electrical charge transferred to the load during the chemical r eaction. On the contrary, the energy of a capacitor charged to a potential difference of V, scales with the square of the cell voltage and can be calculated due to Eq 1-2; E = CV2 = QV Hence, the energy stored in an electrochemical capacitor is half of the value for the equivalent charge stored in a battery14 Therefore, the key to achiev ing higher energy densities in a electrochemical supercapacitor is to expand the cell voltage. In conducting polymers, the redox pseudocapacitance arises from the doping and undoping processes of the material. As mentioned in the previous chapters, the versatility of the conjugated polymers enables different redox supercapacitor configurations. Among the different polymer
111 supercapacitor configurations proposed to date, the Type III and IV (n /p-type) devices are the most promising configurations in terms of ha ving a high operating voltage for delivering all the doping charge during discharge at high potential s. The operating voltage of the cell can be increased up to 3V in these kinds of devices. Hence, the highest ener gy densities are possible with these designs. Besides providing the br oadest cell operating vol tages, Type III&IV supercapacitors can deliver all the doping ch arge during discharge at high potentials; consequently, high power densities can be demonstrated. The Type III and IV supercapacitors are ba sed on conducting polymers that can be pand ndopable. Devices utilize the n-dopable polymer as the negative electrode and the p-dopable polymer as the positive electrode. Type III supercapacitors uses the same polymer on both the positive and the negative electrodes while Type IV utilizes two different polymers. The anode pdopable polymer is in the fully oxidized state an d the cathode n-dopable polymer is in the fully reduced form when the device is charged. Both polymers neutralize with the discharging of the supercapacitor. By this means, the total char ge capacity of both pand n-dopable polymers contribute to the overall cap acitance of the device. The p-&ntype conducting polymer based s upercapacitors are the most favorable configurations in term s of energy and power densities, but the difficulty of obtaining stable ndoping polymers is the largest obstacle for the de velopment of these supercapacitors for realworld applications. While there have been some n-dopable polymers synthesized, most of their coulombic efficiency is low upon repetitive cycli ng in the reduced state. The bulk of conjugated polymer systems possess positive ch arge carriers (hole transport) exhibiting p-type doping. To obtain a truly n-type doped system that exhibit stable coulombic efficiency is still a challenge in the field of conducting polymers since the envi ronmental stabili ties of n-type dopable systems
112 are poor due to the high reactivity of the radical anion charge carriers to oxygen and water. The concept of band-gap engineering method allows one to modulate the electronic properties of a polymer by controlling both the electron donating and electron accepting abilities, thus the valence and conduction band levels of the polym er. Conjugated polymer systems having stable oxidation and reduction redox states with lowe red band gaps are possible by incorporating donor-acceptor (D-A) units into a single polymer backbone. The D-A approach has been used in various combinations and has resulted in some of the most successful n-dopable conjugated systems to date.158-165 Several D-A type polymers were previous ly investigated in the Reynolds Group.158, 164, 166, 167 Recently a family of D-A-D monomers based on alternating alkylendioxythiophene donors and nitrogen based heterocycle acceptors, whos e structures are show n in Fig 4-1, were synthesized by Tim Steckler in the Reynolds Group.121 Figure 4-1. Structures of donor-acceptor-donor EDOT-benzobisthiadiazole and thiadiazolequinoxaline monomers: BEDOT-BBT, BEDOT-TQ-Me2, BEDOT-TQ-Hx2. The three D-A-D monomers based on alte rnating ethylenedioxyt hiophene (EDOT) as a donor along with benzobisthiadiazole and thiadiazol e-quinoxaline acceptors will be referred to as follows throughout this chapter: BEDOT-BBT, BEDOT-TQ-Me2 and BEDOT-TQ-Hx2. The Type IV supercapacitors utilizing PProDOP and th is family of D-A-D systems are discussed in this chapter. In the supercap acitor devices, PProDOP were used as the p-dopable polymer on the
113 positive electrode and the D-A-D systems (Fig 4-1) were used as the n-dopable polymers on the negative electrodes. Cyclic Voltammetric De position of DAD Systems Electrochemical po lymerization of the mono mers shown in Fig 4-1 were carried out via repeated scan cyclic voltammetry (CV) from a 5mM monomer, 0.1M electrolyte solution of TBAP or LiBTI in acetonitrile or propylene carbonate onto platinum button, gold button and Au/Kapton electrodes, respectively Several drops of methylene ch loride were added to achieve complete dissolution of the monomer due to thei r limited solubility in acetonitrile and propylene carbonate. All electrochemical experiments were pe rformed in a glovebox filled with argon having both H2O and O2 concentrations less than 1 ppm. The best quality homogeneous films with good el ectrochemical response of these polymers were found as the ones synthesized through CV on Au button electrodes from TBAP/ACN solution. In Fig 4-2A, the first 10 complete scan s of CV deposition of BEDOT-BBT at a scan rate of 50mV/s in TBAP/ACN is shown. Very similar deposition vo ltammograms of BEDOTBBT were obtained for different electrolyte so lution systems, but with very small current densities in LiBTI systems. As shown in th e voltammograms in Fig 4-2A, BEDOT-BBT gives a peak oxidation potential at 0.5V vs. Fc/Fc+. Upon continued cycling an increase in current densities of the monomer oxidation peaks is obser ved and the formation of the as made polymer film on the gold electrode is easily evidenced by the increase in current densities of the peaks appearing between 0 and +0.3 V vs. Fc/Fc+, as indicated by the arrows. In addition to gold button electrodes, Au/Kapton sheets were also used for the film deposition. The first 30 complete scans of CV deposition of BEDOT-BBT at a scan rate of 50mV/s in TBAP/PC on Au/Kapton (1cm2) sheets is shown in Fig 4-2B. Compared to the voltammogram of BEDOT-BBT deposition on Au butt on electrode that were shown in Fig 4-2A,
114 the results in Fig 4-2B on a larger electrode surf ace of Au/Kapton sheets are quite similar to each other in shape. However, in the case of Au/Kapton electrodes, the electrochemical polymerization leads to dramatical ly smaller current densities; almo st an order of magnitude less. Hence, the efficiency of the film formation was lo w as seen in Fig 4-2B. This is probably due to the limited solubility of the monomer in propyl ene carbonate since the mo re soluble monomers can polymerize more easily, forming homogeneous films that provide better electrochemical results. For that reason, the experiments were co nducted on Au button electrodes for the rest of the chapter. -0.8-0.6-0.4-0.20.00.20.40.60.8 -2.0 -1.0 0.0 1.0 2.0 3.0 I (mA/cm2)E(V) vs. Fc/Fc+ A-0.8-0.6-0.4-0.20.00.20.40.6 -0.10 -0.05 0.00 0.05 0.10 BI (mA/cm2)E(V) vs. Fc/Fc+ Figure 4-2. Electrochemical polymerizat ion (first 10 cycles) of BEDOT-BBT by potential scanning on A) Au button (0.02cm2) in TBAP/ACN and on B) Au/Kapton (1cm2) in TBAP/PC at 50 mV/s. (Several drops of met hylene chloride were added to assist in complete dissolution of the monomers.) The CV electropolymerization of BEDOT-TQ-Me2 and BEDOT-TQ-Hx2 were carried out using the same electrolyte solution; TBAP/A CN on Au button electrodes. The voltammograms are shown in Fig 4-3A and B respectively. Both of these monomers were polymerized forming electrochemical responses sligh tly different than the deposition of BEDOT-BBT but very similar to each other. As shown in Fig 4-3A and B, BEDOT-TQ-Me2 oxidizes at 0.57V vs Fc/Fc+and BEDOT-TQ-Hx2 oxidizes at 0.67V vs Fc/Fc+, respectively. The oxidation poten tial of BEDOT-
115 TQHx2 is found at a slightly hi gher potential than the Ep,m of either BEDOT-BBT and BEDOTTQMe2. Moreover, the second and third scans lead to slightly smaller current density of the monomer oxidation peak, which was previously s een for the electrochemical polymerization of similar types of monomer systems. 129, 168 The increase in the current densities of the peaks appearing between 0V and +0.5 V vs Fc/Fc+ with the progression of cy cling in both cases is due to the redox activity of the polymers being formed on the el ectrode surfaces. As a result of creating a chemically modified working electr ode and forming dimers, the monomers oxidation peaks become broader and negatively shift to 0.1V lower potential upon continued cycling. The overall current densities for the same number of cycles are higher in the case of BEDOT-TQHx2 than those of BEDOT-TQ-Me2. The results indicate that th e electropolymerization and switching of BEDOT-TQ-Hx2 is more efficient than BEDOT-TQMe2. One possible reason for this is the slightly more open morphol ogy caused by bulkier he xyl pendant groups. -0.8-0.6-0.4-0.20.00.20.40.60.8 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 AI (mA/cm2)E(V) vs. Fc/Fc+ -0.8-0.6-0.4-0.20.00.20.40.60.8 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 BE(V) vs. Fc/Fc+I (mA/cm2) Figure 4-3. Electrochemical polymerization (first 10 cycles) of A) BEDOT-TQ-Me2 and B) BEDOT-TQ-Hx2 by potential scanning on Au button (0.02cm2) in TBAP/ACN at 50 mV/s. (Several drops of methylene chlori de were added to assist in complete dissolution of the monomers in acetonitrile.) Cyclic voltammograms of all of the three DAD monomers exhibit the irreversible monomer oxidation peak with the absence of a reductive wave which is indicative of radical-
116 radical coupling showing ease of formation of po lymeric materials. The relatively low oxidation potentials of all three monomers can be attributed to the terminal 3,4-alkylenedioxythiophene (EDOT) units of the monomers, which are electron rich in character that allow easily accessible oxidation potentials for elect rochemical polymerization. Electrochemistry of P(DAD) Films The electrochem ically synthesi zed polymer films from all three monomers were examined for their redox processes in TB AP/ACN, after being washed with copious amounts of acetonitrile to remove any residual monomers. These polymer films possess unique properties when reductively (or n-type) doped. Introducing strong accepto r units of heterocycles contai ning nitrogens in between the terminal EDOTs within the polymer repeat unit lowers the LUMO level of the molecule since the nitrogen is more electrone gative compared to the carbon.121, 169 Subsequently, the reduction of the polymer occurs more easily. The benzobis( thiadiazole) and thiadiaz ole-quinoxaline units are strong acceptors. This help s to create relativel y positive reduction po tentials for the polymers, and enhance the reductive stability upon repetitive cycling due to the easy access to their reduced states. The donor-acceptor character of these monomers and polymers allows stabilizing separated charges. Anion radical structures demonstrat ing the hypothetical reductio n of PBEDOT-BBT are shown in Fig 4-4, as the representative of all three P(DAD) systems since PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 can be envisioned to behave in a similar way. In troducing an electron to the polymer backbone forms a radical anion, in whic h the electron density is stabilized on the nitrogen on the thiadiazole ring of the acceptor unit as shown in the scheme. The first electron would be stabilized in the same way upon reduction of P BEDOT-TQ-Me2 and PBEDOT-TQHx2 since the thiadiazole is a stronger acceptor unit than the quinoxaline unit The next
117 introduced electron upon the second reduction of PBEDOT-BBT could occur in different ways. Only one of the hypothetical structures of the second reduction is shown in the Fig 4-4, where the electron density is stabilized on the other thiadiazole ring, indicating the ability of this acceptor unit to accommodate more than one nega tive charge density. Moreover, the resonance states of this structure would e xhibit varying delo calization points for the two electron densities, distributed between the two thiadiazole rings. In the same way, the second electron density upon the second reduction of P BEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 could be delocalized either on the thiadiazole or quinoxaline rings. Cyclic voltammograms of a ll three compounds showed two reductive doping peaks, which is consistent with the proposed stru ctures in Fig 4-4. Figure 4-4. Anion radical structur es demonstrating the hypothetical mechanism of the reductive doping (first and second reductions) of PB EDOT-BBT with the proposed resonance structures of the reduced forms of the polymer. All P(DAD) systems could possess dual p-type and n-type doping with in a single polymer. In order to examine the n-type doping prope rties, only reductive doping experiments were performed separately from oxidative doping of the same polymer. Experiments conducted by repeatedly cycling between oxidized (p-doped) and reduced (n-doped) states can often lead to the
118 formation of peaks that may be misinterpreted as redox doping peaks which result from trapped charge carriers. Accordingly, the polymers were switched by repetitive potential scanning from neutral to reduced state for ten times to condi tion (break-in) the films until a reproducible CVs were obtained. The reductive cyclic voltammetry behaviors of the all three polymers, as a function of the scan rate between 50 and 500mV/ s with 50mV increments are shown in Fig 4-5. As can be seen in the figure, the reproducible cyclic voltammetric be havior upon reduction of each polymer is similar to each other. Two reduction processes occur for each polymer due to the strong acceptor unit that allows for the insertion of two electrons per polymer repeat unit. The ntype doping electrochemistry show s sharp and discernible peaks. In Fig 4-5A, PBEDOT-BBT exhibi ts two redox processes with E1/2 values centered at -0.98 and -1.78 V vs. Fc/Fc+. This demonstrates the high electron accepting ability of the polymer having two thiadiazole rings on the acceptor that it can stabilize two negative charges. Both of the reductive doping peaks exhibit comparable current densities, and the anodic to the cathodic peak charge ratios exhibits values close to the id eal ratio of 1, even at a scan rate as high as 500mV/s. This demonstrates the quasi-rever sibility of both the first and second n-doping processes of PBEDOT-BBT. Moreover, the peak curre nt densities in both peaks increase linearly as a function of scan rate up to 500mV/s, indicating that the electroac tive sites are surface bound to the electrode and that the reductio n processes are not diffusion limited. PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 systems contain thiadiaz ole-quinoxaline units as acceptor different than the benzobis(thiadiazole) unit in PBEDOT-BBT. Since quinoxaline based unit is a weaker acceptor than thiadiazole based un it, it is expected to observe noticeable changes in the voltammograms of these pol ymers in moving from the stronger to the weaker acceptor.
119 -2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 A500mV/sI (mA/cm2)E(V) vs. Fc/Fc+50mV/s -2.4-2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -10.0 -8.0 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 BE(V) vs. Fc/Fc+I (mA/cm2)500mV/s 50mV/s -2.6-2.4-2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0 CE(V) vs. Fc/Fc+I (mA/cm2)500mV/s 50mV/s Figure 4-5. Reductive cyclic volta mmograms of A) PBEDOT-BBT, B) PBEDOT-TQ-Me2 and C) PBEDOT-TQ-Hx2 at 50-500 mV/s on Au button electrode (0.02cm2) in 0.1 M TBAP/ACN solution.
120 In Fig 4-5B&C, PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 exhibit an E1/2 value at -1.2 V and -1.25 V for the first reduction, and an E1/2 value at -1.9 V and -1.95V for the second reduction, respectively. The first reduction peak s of polymers are quite stable and quasireversible with adequate charge compensation of the anodic to the cathodic peak. On the other hand, the second reduction does not e xhibit a similar behavior. The cu rrent densities of the first reduction peaks of these polymers are on the same order of magnitude with the first reduction of PBEDOT-BBT. However, the current densities observed for the second reduction processes of both PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 are about 1/5 of the first peaks. These results confirm the strong ability of the thiadiazole ring to delocalize electron densities since the high coulombic efficiency of the second reduction of PBEDOT-BBT, compared to that for PBEDOTTQ-Me2 or PBEDOT-TQ-Hx2 is evident. Most likely the second reduction of PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 is stabilized on the quinoxalin e ring, which is a weaker acceptor. Furthermore, the peak current densities of the first reductive doping peaks for both polymers increase linearly up to 500mV/s scan rate. Conve rsely, the current densities for the second peaks do not show a linear dependence on the scan rate. Although both the anodic and cathodic peak currents of the second reductive doping peaks increas e with the increasing s can rate, the ratio of them to each other do not exhibit a stable number with the increasing scan rate. This illustrates the relatively low reversibility of the second reductive doping processes of these polymers. Furthermore, the peak to peak separation in both the firs t and the second reductions of PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 is much larger than in PBEDOT-BBT. This result is indicative of the the structural differences in the morphologies of these polymers. Incorporating different (weaker) acceptor units in between EDOT units on the polymer backbond and also coupling the alkyl pendant groups on the polym er backbonds may result in a more ordered
121 configuration of the polymer chains resulting in a more compact film, and making it harder for the supporting electrolyte to m ove in and out of the film. Although the relative shape of the electrochemical re sponses of PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 upon reduction is very similar; th e difference of 0.05V between the E1/2 values of the first reductions between two pol ymers exactly matches the difference between the E1/2 values of the second reductions of these two polymers. These slightly different reduction potentials are likely due to the structural differences in the morphology of these polymers caused by the different methyl and hexyl pendant groups. Conversely, the more positive potentials observed for both first and second reductions of PBEDOT-BBT compared to PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 is likely a consequence of the st ronger benzobisthiadiazole acceptor lowering the valance conduction band level of the DAD polymer system. Capacitances of P(DAD) Films In order to d etermine the achievable maximum capacitance from these polymers, the fresh films containing varying amounts of polymers were synthesized from the exact same electrochemical systems while va rying the number of deposition cycles. Electropolymerizations were terminated once a certain amount of redox cy cles had been reached. In this manner, the mass of the polymers is conveniently been contro lled. Then the capacitance values obtained for these films containing varying masses of polymers were determined for all three of the DAD systems. The cyclic voltammograms at a scan rate of 50mV/s, and the capacitances (mF/cm2) as a function of applied potentials are shown in Figs 4-6, 4-7, 4-8 A&B, respectively for the varying amount of P(DAD) systems. The charge examined by integrating the voltammograms of anodic and cathodic scans over the potential rang e during reductive switching of the polymers are also shown in Figs 4-6C, 4-7C, 4-8C as a function of time.
122 -2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 A 10cycles 20cycles 30cycles 40cycles 50cycles 60cycles 70cyclesI (mA/cm2)E(V) vs. Fc/Fc+ -2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 BE(V) vs. Fc/Fc+ 10cycles 20cycles 30cycles 40cycles 50cycles 60cycles 70cyclesC (mF/cm2) 010203040506070 -25 -20 -15 -10 -5 0 5 CQ (mC/cm2)time(s) 10cycles 20cycles 30cycles 40cycles 50cycles 60cycles 70cyclescathodic scan anodic scan Figure 4-6. A) Cyclic voltammograms, B) capacitances as a function of applie d potential and C) charge densities as a function of tim e of varying amounts of PBEDOT-BBT (depsoition cycles) at 50mV/s.
123 -2.4-2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 A 20cycles 30cycles 40cycles 50cycles 60cyclesE(V) vs. Fc/Fc+I (mA/cm2) -2.4-2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -40 -30 -20 -10 0 10 20 30 40 50 60 70 B 20cycles 30cycles 40cycles 50cycles 60cyclesE(V) vs. Fc/Fc+C (mF/cm2) 01020304050607080 -20 -16 -12 -8 -4 0 C anodic scan 20cycles 30cycles 40cycles 50cycles 60cyclesQ (mC/cm2)time (s) cathodic scan Figure 4-7. A) Cyclic voltammograms, B) capacitances as a function of applie d potential and C) charge densities as a function of tim e of varying amounts of PBEDOT-TQ-Me2 (depsoition cycles) at 50mV/s.
124 -2.6-2.4-2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 A E(V) vs. Fc/Fc+I (mA/cm2) 5cycles 7cycles 10cycles 15cycles 20cycles 30cycles -2.6-2.4-2.2-2.0-1.8-1.6-1.4-1.2-1.0-0.8-0.6-0.4 -20 -10 0 10 20 30 40 BE(V) vs. Fc/Fc+ 5cycles 7cycles 10cycles 15cycles 20cycles 30cyclesC (mF/cm2) 0102030405060708090 -14 -12 -10 -8 -6 -4 -2 0 2 C 5cycles 7cycles 10cycles 15cycles 20cycles 30cyclesQ (mC/cm2)time (s) cathodic scan anodic scan Figure 4-8. A) Cyclic voltammograms, B) capacitances as a function of applie d potential and C) charge densities as a function of tim e of varying amounts of PBEDOT-TQ-Hx2 (depsoition cycles) at 50mV/s.
125 Table 4-1. Charge densities and capac itances of PBEDOT-BBT, PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 with different deposition cycles. # of dep. cycles PBEDOT-BBT (1.5V) PBEDOT-TQ-Me2 (1.75V) PBEDOT-TQ-Hx2 (2.0V) Q (mC/cm2) C (mF/cm2) Q (mC/cm2) C (mF/cm2) Q (mC/cm2) C (mF/cm2) 5cycles 11.3 5.7 7cycles 10.5 5.2 10cycles 5.7 3.7 11.6 5.8 15cycles 8.1 4.1 20cycles 6.7 4.3 8.7 4.9 9.1 4.5 30cycles 10.8 6.9 9.9 5.7 9.3 4.7 40cycles 11.8 7.6 12.1 6.9 50cycles 13.1 8.5 13.5 7.7 60cycles 15.5 10.0 21.4 11.6 70cycles 21.9 14.2 80cycles In Table 4-1, the charge and the capacitance values with different amount of deposition cycles were summarized for all three DAD polym er systems. The average capacitance values were calculated from the voltammetric charges as follows; Cp = (Qa + Qc)/2.E(V) (4-1) The capacitance values were expected to increas e linearly with increasing number of deposition cycles, since the amount of charge consumed fo r each deposition is expected to be equal. However, this was not observed for any of the polymers, likely due to the unavoidable experimental errors. In the case of PBEDOT-TQ-Me2, the highest amount of capacitance of 11.6mF/cm2 was obtained with the 60 cycles of depos ition. For the higher number of deposition cycles, the response of the reductive switching of the polymer resulted in voltammograms with the distorted shape. This was not surprising since the conducting polymer films lose their electrical response upon charging an d discharging as the films get thicker, probably due to the kinetic limits of the ion interc alation and depletion processes. The same behavior was observed for PBEDOT-BBT with the maximum capacitance obtained 14.2mF/cm2. On the other hand, the
126 electrochemical responses of the PBEDOT-TQ-Hx2 films were observed to be more sensitive to the loading of the polymer ma terial. The highest capacitances obtained with PBEDOT-TQ-Hx2 were calculated as 5.8mF/cm2 for 30cycles deposition and the redox response of the polymers were decreased for the further deposition cycles. Th e slight inconsistency of the dependence of redox responses of all three polymers to the deposition cycles was probably caused by the difficulty to control all the electroche mical parameters for each experiment. The environmental stabilities of n-type dopable systems are very poor due to the high reactivity of the radical anion charge carriers to oxygen and water. As such, obtaining a truly ntype doped system that exhibits reproducible coul ombic efficiency with adequate stability is still a challenge in the field of conducting polymers. The stability of n-type doping upon long-term charging and discharging is important for the use of these materials in de vice applications. As mentioned in the previous chapters, CPs undergo volume expansion and contraction when the oxidation level is changed due to the insertion and expulsion of ions during the redox doping and dedopi ng processes. Moreover, the stability of n-type doping is poor due to the high reactivity of the radical anion charge carriers to oxygen and water. Therefore, the stability of a ll three polymers were examined by exposing the films to nonstop continuous dopi ng and dedoping processes by cyclic voltammetry in their potential range including both the first and second reduction processe s at a scan rate of 50mV/s. The voltammograms of the films were recorded every 10 switch es up to 1400 switches. The average charge consumed for anodic and cathodic scans in each voltammogram were calculated and the normalized charge values were plotted as a function of number of switching cycles for each P(DAD)s in Fig 4-9. Although the experiment s were conducted in an argon-filled glove box (an oxygen and moisture free environment), a fast decrease in anodic and cathodic current
127 responses was observed. The capacitan ces of PBEDOT-BBT and PBEDOT-TQ-Hx2 were dropped to 20% of the original values dur ing the few hundreds of cycles. PBEDOT-TQ-Me2 showed better stability than th e other two DAD systems, however its capacitance also dropped to 20% of the original valu e after 1400 cycles. The instability of th ese polymers can be attributed to the less reversible second reductive doping proce ss. Another reason might be the unavoidable trapped charges on the polymer system upon re ductive doping causing the irreversible redox reactions, which are typically observed for conducting polymer systems. Consequently, the stabilities of the DAD polymers need to be improved. 02004006008001000120014001600 0.0 0.2 0.4 0.6 0.8 1.0 B CNormalized QNumber of cycles A Figure 4-9. Stability of n-doping of A) PBEDOT-BBT, B) PBEDOT-TQ-Me2 and C) PBEDOTTQ-Hx2 in TBAP/ACN by non-stop cycling at 50mV/s shown as normalized redox charges as a function of nu mber of switching cycles. Type IV Supercapacitors As was explained in the Introduction, the Type IV supercapacitors utilize conducting polym ers that can be pand ndopable. They us e an n-dopable polymer as the negative electrode and a p-dopable polymer as the positive electro de. During charging of the device, charges are being removed from the p-dopable polymer on positiv e electrode and are being transferred to the
128 n-dopable polymer on negative electrode. The creat ed electron deficiency and the excess of electrons on the polymers were comp ensated by the appropriate counter ions; anions and cations fro m the electrolyte, respectively. Accordingly, when the anode p-dopable polym er is in the fully oxidized state and the cathode n-dopable polymer is in the fully reduced form, the device is in the fully charged state. Upon discharging of the device, both polymers neut ralize and the ions are released back to the electrolyte solution. Cons equently, it is very important to match redox (switching) charges of the two complementary polymers, prior to device construction for a balanced switching. Type IV devices have an asymmetric conf iguration since they utilize two different polymers on the positive and the negative elec trodes. In this study, the three DAD systems discussed so far, were combined with PProDOP as the negative and the positive electrode materials, respectively, in a Type IV supercap acitor device configura tion. As was mentioned earlier, the aim of this study was to expand the ce ll voltage to achieve high er energy density in a polymer based electrochemical supercapacitor sin ce the energy of a capacitor scales with the square of the cell voltage. Util izing the DAD systems with PProDOP in a Type IV configuration expands the operating voltage range of the devices up to 3V. In Fig 4-10 the cyclic voltammograms of positive doping of PProDOP and the negative doping of PBEDOT-TQ-Me2 as the representative of DAD systems, in TB AP/ACN at 50mV/s, are shown in the same potential scale. This figure demonstrates the extended operati ng voltage of the cell up to 2.75V. The voltage ranges for PBEDOT-BBT, PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2 were approximately found as 1.6V, 1.75V and 2V, resp ectively. Hence, the combination of these polymers with a 1V range of PProDOP in a Type IV configurati on result in a total of 2.6V, 2.75 and 3V device working potential ranges.
129 -2.0-1.5-1.0-0.50.00.5 -2 -1 0 1 2 3 PProDOPu 1VI (mA/cm2)E(V) vs. Fc/Fc+ PBEDOT-TQ-Me2 1.75V Figure 4-10. Expanding cell voltage of Type IV supercapacito r by combining PProDOP with PBEDOT-TQ-Me2. Positive doping of PProDOP a nd negative doping of PBEDOTTQ-Me2 in TBAP/ACN at 50mV/s. The devices were fabricated using the films of each of three DAD systems that possessed the maximum capacitance values. To the best of our knowledge, the charging and discharging of conducting polymer based Type IV supercapacito rs with two redox peaks, shown by cyclic voltammograms were not reported in the literature before. In order to examine the switching behavior of the two terminal device systems, the electrochemically deposited DAD films were combined with PProDOP films containing differe nt amount of material into a device using TBAP/ACN as the conductive media between the electrodes. The PProDOP films were chosen as to display the redox charges upon switching in an excess, equal and limited amounts compared to the redox charges exhibited by P(DAD) of interest. In Figs 4-11, 4-12 and 4-13, the related experimental results were shown for each of the DAD syst ems. First, the redox switching charges of the DAD systems and of the varyi ng PProDOPs were monitored during cathodic and anodic scans in a three-electrode cell containing 0.1M TBAP/ACN and were shown in the same
130 graph. In order to plot the switching charges of the polymers in the same scale, the normalized times were used. Additionally, the cyclic voltammograms at a scan rate of 50mV/s and the capacitances (mF/cm2) as a function of applied potentials of the TypeIV supercapacitors were also shown in the figures. Furthermore, th e redox charges examined by integrating the voltammograms of devices over the entire potential range were also monitored and plotted as a function of time as shown in the following figures. 0.00.20.40.60.81.0 -20 -15 -10 -5 0 5 10 15 20 25 PProDOP excess equal limited_1 limited_2 AQ (mC/cm2)normalized time PBEDOTBTD 0.00.51.01.52.02.5 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 BI (mA/cm2)E(V)Device excess equal limited_1 limited_2 0.00.51.01.52.02.5 -20.0 -16.0 -12.0 -8.0 -4.0 0.0 4.0 8.0 12.0 16.0 Device excess equal limited_1 limited_2E(V)C(mF/cm2)C020406080100 0 5 10 15 20 Device excess equal limited_1 limited_2Q (mC/cm2)time (s)D Figure 4-11. A) Charge densitie s as a function of time of PBE DOT-BBT and varying amount of PProDOPs in TBAP/ACN at 50mV/s. B) Cyclic voltammograms, C) capacitances as a function of applied voltage and D) charge densities as a func tion of time of the corresponding devices in TBAP/ACN at 50mV/s.
131 0.00.20.40.60.81.0 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 PProDOP limited equal excessQ (mC/cm2)normalized time PBEDOT-TQ-Me2 A 0.00.51.01.52.02.53.0 -1.0 -0.5 0.0 0.5 1.0 I (mA/cm2)E(V)Device excess equal limitedB 0.00.51.01.52.02.53.0 -20 -10 0 10 20 Device excess equal limitedC (mF/cm2)E(V)C 020406080100120 0 5 10 15 20 25 Device limited equal excessQ (mC/cm2)time(s)D Figure 4-12. A) Charge densities as a function of time of PBEDOT-TQ-Me2 and varying amount of PProDOPs in TBAP/ACN at 50mV/s. B) Cyclic voltammograms, C) capacitances as a function of applied potential and D) ch arge densities as a function of time of the corresponding devices in TBAP/ACN at 50mV/s. The average capacitance values for supercapacito rs containing different (excess, equal and limited) amounts of all P(DAD) systems were calcu lated from the redox charges obtained from the integrated voltammograms using Equation 4-1. The anticipated capacitances of devices were also calculated using the following equation with the capacita nces of PProDOP and P(DAD) systems that were previously measured experimentally; (4-2)
132 0.00.20.40.60.81.0 -12 -8 -4 0 4 8 12 16 20 PProDOP equal excess limitedQ (mC/cm2)normalized time PBEDOT-TQ-Hx2A 0.00.51.01.52.02.53.0 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 I (mA/cm2)E(V)Device excess equal limitedB 0.00.51.01.52.02.53.0 -12 -8 -4 0 4 8 12 Device excess equal limitedC (mF/cm2)E (V) C020406080100120 0 2 4 6 8 10 Device excess equal limitedQ (mC/cm2)time (s) D Figure 4-13. A) Charge densities as a function of time of PBEDOT-TQ-Hx2 and varying amount of PProDOPs in TBAP/ACN at 50mV/s. B) Cyclic voltammograms, C) capacitances as a function of applied potential and D) ch arge densities as a function of time of the corresponding devices in TBAP/ACN at 50mV/s. In Table 4-2, the results of Figs 4-11, 412 and 4-13 were summarized. The overall redox charges involved in the electrochemical processes of the DAD films of interest were reported in the first column with the name of P(DAD). In the second column, the redox charges exhibited by the chosen PProDOP films were presented. Fina lly, the redox charges displayed by the devices were shown along with the calculated capacitances of the devices.
133 During the charging and discharging process of the device, the charges are being released and loaded from one electrode to the other. As a result, the devices cont aining the equal and the excess amount of PProDOP were expected to show a similar cyclic voltammogram behavior to P(DAD) systems with two distinct redox processe s. This is expected because the amount of charge needed by P(DAD) of interest will be compensated by the complementary PProDOP. In Figs 4-11B, 4-12B and 4-13B, the cyclic volta mmograms of the supercapacitors exhibit two distinct redox processes, sim ilar to the voltammograms of P(DAD) films measured in a threeelectrode cell. These redox charge s calculated for the devices were expected to be equal to the redox switching charges calculated for P(DAD)s. For example, the redox charges calculated for PBEDOT-TQ-Me2 based devices cont aining equal (21.9mC/cm2) or excess (27.4mC/cm2) amount of PProDOP compared to P(DAD), demonstrated 21.7 and 22.3mC/cm2 redox charges, as shown in Fig 4-12. These values are found to be quite close to 21.4mC/cm2 which is observed originally for the switching of PBEDOT-TQ-Me2. Alternatively, the devices c ontaining the limited amount of PProDOP compared to the P(DAD) of interest were exhi bited cyclic voltammogram sim ilar to PProDOP for the first reduction for each system. The second reduction wa s observed to exhibit a distroded shape like an over oxidation peak generally observed for the conducting polymer systems. This was not surprising at all since there were not sufficient charges exhibited by PProDOP to compensate the charges needed for the full switching of P(DAD) of interest. As a result, the voltammograms show first reduction similar to switching of PPr oDOP itself with probably a combination of a compressed peak of the second reduction of P(DAD) and extended over oxidation peaks for PProDOP. However, it is very hard to predict the exact electroche mical processes occuring in the devices containing the unbalanced positive and negative charges.
134 Table 4-2. Charge densities and capacitances of devices prepared by PBEDOT-BBT, PBEDOTTQ-Me2 and PBEDOT-TQ-Hx2 with different amount of PProDOP films. PProDOP Q (mC/cm2) Device Q (mC/cm2) Device C=Q/V (mF/cm2) Anticipated Capacitance 1/C = 1/CPProDOP + 1/CP(DAD) p(BEDOTBBT) Q=16.1mC/cm2 25 17.1 6.8 7.1 16.0 12.8 5.1 6.2 10.0 10.2 4.1 5.6 4.6 5.4 2.2 3.1 p(BEDOT-TQ-Me2) Q=21.4mC/cm2 13.0 18.9 6.5 6.1 21.9 21.7 7.6 7.5 27.4 22.3 8.1 8.0 p(BEDOT-TQ-Hx2) Q=10.5mC/cm2 4.5 5.2 1.9 2.4 10.1 7.7 2.8 3.5 17.6 8.9 3.2 4.0 The experimentally observed average capacitance values were calculated for each device from the voltammetric charges over the operating volta ge ranges as in the case of films. As seen in Table 4-2, the capacitances for the devices that were experimentally observed and the ones that were calculated from the separate film va lues are all in agreemen t with each other. The differences were most likely due to the unavoidable experimental errors. The highest amounts of capacitances, were observed as 6.8, 8.1 and 3.2mF/cm2, for PBEDOT-BBT, PBEDOT-TQ-Me2 and PBEDOT-TQ-Hx2, respectively. The difference between the cyclic voltamm ograms of the Type IV supercapacitors compared to Type I devices discussed in Chapter 3 is apparent. No current flow was observed until the potential of the device reached a certain point (which is about 0.7V) during switching of
135 Type IV supercapacitors, alternatively to the immediate increase observed for Type I supercapacitors. This was caused by the design of the Type IV configuration. All of the charge is delivered at higher voltages for Type IV devi ces upon discharging, while the voltage decays to 0V in the case of Type I supercapacitors. Ther efore, higher power densities were predicted for Type IV device configurations. 020040060080010001200140016001800 0.0 0.2 0.4 0.6 0.8 1.0 C ANormalized QNumber of cycles B Figure 4-14. Stabilities of Type IV supercap acitors of A) PBEDOT-BBT, B) PBEDOT-TQ -Me2 and C) PBEDOT-TQ-Hx2 in TBAP/ACN at 50mV/s. In regards to the films, the stability of devices upon long-te rm charging and discharging is very important. The cyclic stability of of all three supercapacitor systems containing the materials possessing equal amounts of charge upon switching, that were summarized in Table 42, were examined by continuous cyclic volta mmetry at a scan rate of 50mV/s. The voltammograms of the films were recorded in every 10 switches until the films lost 90% of original capacitance. Here again the experiments were conducted in an oxygen and moisture free environment. The average redox charges invol ved in anodic and cathodic scans in each voltammogram were calculated and the normalized charge density values were plotted as a function of number of switching cycles for each supercapacitor sy stem in Fig 4-11. The stability
136 of the devices are similar to the stabilities of th e of P(DAD) films. Unfortunately, a fast decrease in anodic and cathodic current resp onses was observed. In almost 1000 cycles, devices lost 10% of the original capacitances. Am ong three P(DAD) based supercapac itors, the best stability was obtained for the PBEDOT-TQ-Me2 based device. Conclusions and Perspective In conclusion, this chapter presents the utilization of a new family of n-dopable electroactive polymers synthesi zed in the Reynolds group, along with PProDOP in Type IV supercapacitors. This family includes BEDOT-BBT, BEDOT-TQ-Me2 and BEDOT-TQ-Hx2 which were designed based on an alternating donor and acceptor units. The energy stored in an electrochemical capacitor scales with the square of the cell voltage. Therefore, the key to achieving higher energy densities in an electroc hemical supercapacitor is to expand the cell voltage. In Type IV supercap acitors, PProDOP were used as the p-dopable polymer on the positive electrode and DAD systems were used as the n-dopable polymers on the negative electrodes, which lead the opera ting voltage ranges of the devi ces to be enhanced up to 3V. In this work, all the electroactive films have been electrochemically deposited on gold button electrodes. Further work to be conducted af ter this dissertation is anticipated to focus on optimizing the proper combination of these materi als on larger scale elec trodes. In order to achieve higher voltage ranges, ionic liquids can be a good candidate as the conducting media in between the electrodes.
137 CHAPTER 5 HYBRID SUPERCAPACITORS: RUTHENIUM OXIDES| PPRODOP Introduction The m aterials studied for supercapacitor applic ations have focused on carbon, metal oxides and conducting polymers. As mentioned in Chapte r 1, the electrochemical capacitors based on pseudocapacitance (or re dox-capacitors) involving metal oxi des and conducting polymers can provide higher energy density than conven tional double layer capac itors based on carbon.1, 6, 14, 15, 24, 170, 171 It has been demonstrated that PProDOP is a highly electron rich system having a quite low oxidation potential, easy to charge/disch arge and show superior stability, especially when prepared as thin films. PProDOP is a prom ising energy storage material with a potential to be utilized in commercial supercapacitor devices. Although conducting polymers offers significant advantages as supercapacitor electr ode materials, includi ng fast and efficient switching between redox states high electrical co nductivity, suitable morphology, especially lightweight nature; their intrin sic degree of doping is limited to capacitance values of 100250F/g. On the other hand, metal oxides are attractive electrode materials for supercapacitors due to their high specific capacitance and low resistance. Among the transiti on metal oxides, hydrous ruthenium oxide (RuOxnH2O) with a microporous structure has been recognized as the most promising supercapacitor material due to its intrinsic high electroc hemical reversibility1, fast charge/discharge ability, ultrahigh specific pseudo capacitance, and excellent cycle life in acidic electrolyte solutions of H2SO4.55, 172, 173 Depending on preparation procedures and measurement conditions, the specific capacitance ranges report ed to date for hydrous ruthenium oxide have been between 600 to 860 F/g.54, 55, 172-183 Recently a specific cap acitance of 1340 F/g at a potential scan rate of 25 mV/s was reported for sol-gel derived rutheniu m oxide nanodots loaded
138 on activated carbon.184 Although the best charge storage capabi lities of any capacitor reported to date have been achieved with hydrous ruthenium oxides, the high co st of ruthenium prevents its large-scale use. The high capacitance density of RuOxnH2O has been attributed to the mixed protonic electronic conduction. The charge storage process in ruthenium oxide electrodes is surface limited and progresses in several one electron steps, resulting in a broad range of capacitance values. It has been proposed that the hydrous regi ons within the nanoparticles allow facile proton permeation into the bulk material for efficient charge storage while the interconnected ruthenium oxide region accounts for the electronic conduction.55, 185 Thus, a constant pseudocapacitance occurs over the full ope rating voltage range. Combinations of conducting polymer materials with metal oxides have been prepared for different purposes in the literature. Composite s of polyaniline doped with layered transition metal halides (IrCl3nH2O)186 have been synthesized to increase the stability of polymer material and improve its electrochemical properties to a level suitable for practical applications. Moreover, polypyrrole|RuOxnH2O composite material have been synthesized by in situ chemical polymerization.187 In one study, electrodes made by electr ochemically deposited ruthenium oxide on conducting PEDOT were studied fo r supercapacitor applications and it has been observed that PEDOT|RuOxnH2O composite electrodes exhibited large capacitance in acidic electrolytes. There have been some studies attempting to improve the capacitance of carbon nanotube materials by introducing metal oxides into their surfaces. The nanotubes have high conductivity, large surface areas and a 3D por ous structure, and serve as a substrate for ruthenium oxide deposition.21, 188 This work focuses on combining PProDOP with RuOxnH2O in order to improve energy storage capabilities of each component in the composite.
139 PProDOP, with its 3D porous structure, good conductivity, excel lent reversible charge/discharge characteristics and superior stab ility, was expected to provide an ideal substrate for the deposition of ruthenium oxide. In a composite material, it is essential for th e working potentials of each material to match in order to take the advantage of the capacita nce contribution of both materials. This was another reason to combine PProDOP which has a potential range from -0.4 to +0.6V vs. Ag/Ag+ and RuOxnH2O which exhibits redox processes between -0.2 and 1V vs. Ag/AgCl. It has been previously calculated that; E1/2(Ag/Ag+) = E1/2(Ag/AgCl)-0.30V (5-1) Therefore, both these materials exhibits redox pr ocesses, resulting high pseudocapacitance, in a similar potential range. As a result, the high ch arge storage ability of PProDOP, unlike bare nanotube substrates, would contri bute to the overall capacitance of the composite material resulting in enhanced charge stor age capabilities. Therefore, both the utilization of the ruthenium oxide on a specific area and the capacitance of the substrate itself would be improved by introducing a composite of PProDOP and RuOxnH2O into an electroche mical capacitor of high quality. Cyclic Voltammetric Deposition o f Hydrous Ruthenium Oxide Hydrous ruthenium oxide was electrochemically prepared using cyclic voltammetry on the three different substrates: Au plate, Pt plate and SWNTs film. Electrode position was carried out on different substrates by potential cycling between -0.2V and 1.1V vs. Ag/AgCl at a scan rate of 50 mV/s, from an aqueous solution 5 mM RuCl3 x H2O, 0.1 M NH4Cl, and 0.01M HCl, with a pH of 2. The cyclic voltammogram s of the deposition of RuOxnH2O of 300 cycles on bare SWNTs are shown in Fig 5-1A.
140 -0.4-0.20.00.20.40.60.81.01.2 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 AI (mA/cm2)E(V) vs. Ag/AgCl -0.4-0.20.00.20.40.60.81.01.2 -4.0 -2.0 0.0 2.0 4.0 6.0 B I (mA/cm2)E(V) vs. Ag/AgCl Figure 5-1. Electrochemical de position (300 cycles) of RuOxnH2O from 5 mM RuCl3xH2O, 0.01 M HCl, and 0.1 M NH4Cl (pH=2) on A) bare SWNTs film and B) PProDOP|Sticky-PF|SWNTs substrate at 50mV/s. From the voltammogram of ruthenium oxide deposition on bare SWNTs in Fig 5-1A, the presence of at least two redox couples at 0.3 an d 0.9 V are apparent. The cathodic peaks at 0.05V on the negative sweeps is due to the reducti on of Ru(III) species as it is deposited on the substrate. The peak at approximately 0.4V indi cate the oxidation of already deposited ruthenium species (Ru metal and/or oxy-chloro-rut henium species) to hydrous oxides (RuOxnH2O) and then further oxidation to a highe r oxidation state (i.e., hydroxyl Ru(VI) species) at 1V on the positive sweeps. Finally these further oxidized hydrous ruthenium oxides are reduced back to hydroxyl Ru(IV) species on the negative sweeps, then gradually reduced to lower oxidation states at more negative potentials. With th e progression of the elecrodeposition more RuOxnH2O is deposited onto the electrode which is evident by the increasing current densities as the CV cycles continue. However, a gradual distorti on of the voltammograms of both redox processes occurs with increasing CV cycles. This shows that the reversibility of both redox processes decreases gradually since the p eak potential differences increas es by the progression of the CV deposition. This is attributed to the IR drop th at occurs on more thickly deposited ruthenium
141 oxide layers due to an increased proton diffu sion path/barrier. The redox behavior of RuOxnH2O also exhibits hysteresis w ith increasing CV cycles. Fig 5-1B shows the voltammograms of ruth enium oxide deposition on PProDOP which had been previously deposited on Sticky-PF coated SWNTs films. The initial capacitance of this PProDOP layer was measured as 8mF/cm2. Compared to Fig 5-1A, the two redox peaks at 0.3 and 0.9 V are not as distinct as they appear on Fig 5-1A. The general appearance of the voltammogram in Fig 5-1B resembles the voltamm ograms on Fig 5-1A with positively shifted anodic and negatively shifted cathodic peaks. Th ese peaks are broader probably, due to the PProDOP layer undergoing capacitive charging/discharging as RuOxnH2O is deposited. Therefore the peaks appear as a combination of the redox p eaks due to the deposition of RuOxnH2O with the ox/red peaks of the PProDOP layer. The two distinct cathodic peaks can still be observed as in Fig 5-1A with a negative shift of approximately -0.1V. The cathodic peak at -0.05V is the indication of the deposition of RuOxnH2O as the Ru(III) species reduced and the cathodic peak at +0.5V is indicative of the re duction of the already oxidized hydrous ruthenium oxides to hydroxyl Ru(IV) species as it does in Fi g 5-1A. The positive and ne gative shifts seen in anodic and cathodic scans respectively are due to the more resistive modified electrode surface with PProDOP. Besides being shifted and br oader, the voltammogram peaks in Fig 5-1B show higher current densities compared to those in Fig 51A by almost 3 times. The progression of the anodic peak current densities of the rutheniu m oxide deposition at 0.6V on bare SWNTs and at 0.7V vs. Ag/AgCl on PProDOP|Sticky-PF|SWNTs with the deposition cycle numbers up to 300 cycles are shown in Fig 5-2A and B, respect ively. The peak current density for the oxide deposition process on both substrates shows l ogarithmic increase with increasing deposition
142 cycle number. The amount and rate of rutheniu m oxide deposition on PPro DOP is clearly faster than its deposition on bare SWNTs, resulting high er current densities for the same value of the deposition cycle numbers and higher slope in th e graphs. The higher in itial current density observed in Fig 5-2B is due to the capacitive response of the already existing PProDOP layer. These results likely indicate the deposition of more ruthenium oxide due to the increased surface area of the electrode. The rough and porous surf ace of PProDOP possibly allows ruthenium oxide deposition throughout the vo lume of the polymer layer. 050100150200250300 0.0 1.0 2.0 3.0 4.0 5.0 B AI (mA/cm2)# of cycles Figure 5-2. Peak current densities of RuOxnH2O deposition on A) bare SWNTs film at 0.6VAg/AgCl and on B) PProDOP (8mF/cm2) on Sticky-PF|SWNTs substrate at 0.7VAg/AgCl as a function of number of deposition cycles. Characterization of RuOxnH2O and Composite Films of RuOxnH2O with PProDOP Electrochemical Characterization Cyclic voltammetry was used to determ ine the electro chemical properties of electrochemically prepared ruthenium oxide films and composites of ruthenium oxide and PProDOP. In order to determine the capacitance of PProDOP before the deposition of ruthenium oxide, it is essential to look at the CV behavior of PProDOP in an acidic medium compared to its
143 voltammograms in acetonitrile. In Fig 5-3A and B, the voltammograms a nd capacitances as a function of applied voltage of PProDOP deposited on Sticky-PF coated SWNTs film, respectively, first in LiBTI/ACN, then in 0.5M H2SO4 solution and finally again in LiBTI/ACN media are shown to confirm the reproducibility of the CV behavior of PProDOP in an acidic medium. The differences between the CV behavior and the capacitance values of PProDOP in these two different media were found to be insignificant as seen in the Fig 5-3. -0.20.00.20.18.104.22.168 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 A (ii) (i)I (mA/cm2)E(V) vs. Ag/AgCl (iii)-0.20.00.20.22.214.171.124 -15 -10 -5 0 5 10 15 20 B E(V) vs. Ag/AgClC (mF/cm2)(ii) (i) (iii) Figure 5-3. A) Cyclic voltammograms and B) capacita nces as a function of applied voltage of PProDOP on Sticky-PF|SWNTs first in (i)LiBTI/ACN, then in (ii) 0.5M H2SO4 and finally again in (iii) LiBTI/ACN. (Not e that LiBTI/ACN experiments vs Ag/Ag+ reference was adjusted to Ag/AgCl refe rence for comparison of the voltammograms in the same sale.) Typical CV diagrams measured in a 0.5 M H2SO4 solution at potential scan rates from 10 to 50mV/s of electrochemically prepared RuOxnH2O/SWNTs film and RuO2nH2O |PProDOP|Sticky-PF|SWNTs film electrodes are shown in Fig 5-4A and C, respectively. The respective deposition of RuOxnH2O was halted at the equivalent of 300cycles on both substrates. In Fig 5-4A, voltammograms exhi bit typical pseudocapacitive behavi or of ruthenium oxide. The voltammograms maintained their shape at all potentia l scan rates from 10 to 50mV/s as shown in Fig 5-4A, revealing the reversib le high-power characteristics of ruthenium oxide. The current densities of the redox peaks in the potential re gion of 100 to 800 mV of each voltammograms at
144 different potential scan rates are li nearly proportional to the scan rate. This result implies that the redox transitions between the oxyruthenium species at various oxidation states are not diffusion limited at these scan rates. Slight gradua l distortions of the CV curves with the increase in peak potential differences of the redox pairs we re observed which can be attributed to more resistive switching as the potential scan rate increases. However, the pseudocapacitive electrochemical response of rutheniu m oxide was still quite reversible. -0.20.00.20.126.96.36.199 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 A 50mV/sI (mA/cm2)E(V) vs. Ag/AgCl 10mV/s -0.20.00.20.188.8.131.52 -30 -20 -10 0 10 20 30 B E(V) vs. Ag/AgClC (mF/cm2)50mV/s 10mV/s -0.20.00.20.184.108.40.206 -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 C E(V) vs. Ag/AgClI (mA/cm2)50mV/s 10mV/s -0.20.00.20.220.127.116.11 -100 -80 -60 -40 -20 0 20 40 60 80 100 D E(V) vs. Ag/AgClC (mF/cm 2 )50mV/s 10mV/s Figure 5-4. A&C) Cyclic voltammograms and B& D) capacitances as a function of applied potential of RuOxnH2O (300 deposition cycles) on A&B) bare SWNTs film and B&D) PProDOP|Sticky-PF|SWNT s substrates in 0.5M H2SO4 at 10 to 50mV/s with 10mV/s intervals. The high capacitive characteristics of the hydrous ruthenium oxides are due to the redox transitions of the surface sites on the very large surface area of the material. Although the charge
145 storage mechanism of ruthenium oxide is not ye t fully understood, the inte rfacial electroactive oxyruthenium species at various oxidation states pres ent redox transitions by exchanging protons with the electrolyte H2SO4 solution corresponding to th e following nonstoichiometric equation:173, 189 The broad redox peaks seen in the potential region of 100 to 800 mV in Fig 5.3A can be attributed to the combination of redox processe s due to the transitions of Ru(IV)/Ru(III)/Ru(II) and Ru(VI)/Ru(IV) redox couples in H2SO4 media.189 In Fig 5-4C, voltammograms of the composites of RuOxnH2O with PProDOP exhibits mostly featureless very broad redox peaks in both anodic and cathodic scans resembling a combination of a typical CV beha vior of PProDOP itself and RuOxnH2O itself in Fig 5-4A. The peak potentials were found to be shifted and the differences between the redox peak potentials are found to be greater than RuOxnH2O itself as was explained for the electrochemical deposition process of the ruthenium oxide on PPro DOP. In addition, the gradual distortion of the CV curves with increasing scan rate resembles the voltammograms of RuOxnH2O itself as seen in Fig 5-4A. Since all voltammograms in Fig. 5-4C show electrochemically reversible behavior in the H2SO4 solution, the conductivity of these composite films must be signi ficant. The current densities of the redox peaks in each voltammogram s at different potential scan rates are again linearly proportional to the scan rates, reveali ng that the redox transi tions of the composite material are not diffusion limited even at these s can rates. Accordingly, there is no significant proton diffusion barrier on composite films. Over all, the composites exhi bit voltammograms that are more rectangular in shape than those of RuOxnH2O itself, which is indicative of ideal pseudocapacitive behavior.
146 Fig 5-4B and d show the capacitance as a function of applied potential of RuO2nH2O|SWNTs film and RuOxnH2O|PProDOP|Sticky-PF/SWNTs film electrodes at different potential scan rates from 10 to 50mV/s in 0.5M H2SO4, respectively. The average capacitance values are calculated from the vo ltammetric charges examined by integrating the anodic and cathodic scans ove r the potential ranges; Cp = ( Qa + Qc) / 2 E The effect of the substrates on the utilization of the rutheniu m oxide is obvious from the current densities and capacitance values s hown in Fig 5-4C and D, respectively. The capacitances of RuO2nH2O calculated at 50mV/s for the film s deposited on bare SWNTs film (Fig 5-4B) and on porous PProDOP (F ig 5-4D) were found to be 18mF/cm2 and 45mF/cm2, respectively. A comparison of these capacitan ces of the analogous 300 cycle-deposition of ruthenium oxide on these two substrates shows that the capacitance increased by a factor of 2.5 with RuO2nH2O on PProDOP. This indicates the high charge storage capability of the RuO2nH2O|PProDOP composite film electrodes. An increase in the capacitance value of the composite material of RuO2nH2O|PProDOP compared to the RuOxnH2O itself is expected assuming that there would be a contribution from the PProDOP to the total capacitance. However, the contribution ratio of the capacitance of PProDOP to the composite is important. In Fig 5-5A and B, the voltammogram and the charges as a function of the applied potential taken fr om the same voltammograms of PProDOP before ruthenium oxide deposition and the voltammogram of the composite material (Fig 5-4C) at 50mV/s in 0.5M H2SO4 are shown. As mentioned before, the capacitance of PProDOP was determined to be 8mF/cm2 before ruthenium oxide deposition and 45mF/cm2 after ruthenium oxide deposition. The capacitance of PProDOP in creased by 5.6 times with the introduction of RuO2nH2O to the matrix.
147 -0.20.00.20.18.104.22.168 -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 A (i)I (mA/cm2)E(V) vs. Ag/AgCl (ii) 01 02 03 04 0 0 10 20 30 40 50 B (ii)Q (mC/cm2)time (s) (i) Figure 5-5. A) Cyclic voltammograms and B) charge densities as a f unction of time of (i) PProDOP before the RuOxnH2O deposition and of (ii) composite films in 0.5M H2SO4 at 50mV/s. In the assumption of the total capacitance of the composite mate rial to be the sum of the individual capacitances exhibited by each material; (CComposite=CPProDOP+CRuOx), it is essential to calculate the contributions of each components. In order to determine the capacitance of ruthenium oxide, the capacitance of PProDOP itself (8mF/cm2) measured before the ruthenium oxide deposition was subtracted from the fi nal capacitance of the composite material (45mF/cm2). The contribution of only rutheniu m oxide is found to be 37mF/cm2. Compared to the capacitance values of the ruthenium oxide for same 300 cycles of deposition on bare SWNTs electrode (18mF/cm2), the capacitance of the rutheniu m oxide increased by 100% on PProDOP substrates, which indicates extr emely high electrochemical utiliz ation of ruthenium oxide on a polymer substrate, even with the assu mption that the that the whole 8mF/cm2 capacitance of PProDOP, exhibited before the ruthenium oxide de position, contributed to the final capacitance of the composite. The results in Fig 5-2 and 5-5 indicate that the introduction of ruthenium oxide into the PProDOP matrix greatly enhan ce the capacitive respons e of composite material. Moreover, the utilization of ruthenium oxide is greatly improved through the introduction of a polymer
148 substrate with a porous structure compared to relatively thin layer of SWNTs film. As determined for the PProDOP supercapacitors in Chapter 3, here the capacitance of ruthenium oxide is also found to depend st rongly on the substrate employed. To follow the pathway of ruthenium oxide deposition on PProDOP, with each sample having near the same amount of ruthenium oxide ; a 300cycles worth was de posited on different amount of PProDOP electrodes. In Fig 5-6, the re sulting capacitances of the composite material were plotted as a function of the initial capacit ance of the PProDOP layer before the ruthenium oxide deposition. The different capacitance values of PProDOP ar e reasonably indicative of the different film thicknesses. In Fig 5-6, with increasing PProDOP mass, the capacitance of the composite film increases linearly, stab ilizes at a level of about ~45mF/cm2 and then drops slowly to a value almost equal to the initial PProDOP capacitance. As mentioned above, the increase in the capacitance of the composite material is expected with the increase in the amount of PProDOP, considering the contribution of PProDOP to the total composite capacitance. However, it is obvious from the Fig 5-6 that th e ruthenium oxide deposition becomes ineffective with increased PProDOP thickness. Figure 5-6. Capacitances of the composite (RuOxnH2O and PProDOP) as a function of the initial capacitance of PProDOP before th e ruthenium oxide deposition. 0 10 20 30 40 50 60 05101520253035CComposite(mF/cm2)CPProDOP(mF/cm2)
149 Moreover, the capacitance of RuOxnH2O (deposited for 300cycles) calculated by subtraction of the capacitance of the initial PProDOP (before the ruthenium oxide deposition) from the total capacitance of the composite material is plotted as a function of the capacitance of PProDOP and is shown in Fig 5-7. The negative values found for RuOxnH2O (for the values beyond 15mF/cm2 in Fig 5-6) are not included in Fig 57. An increase in the capacitance of RuOxnH2O with the increasing PProDOP thickness, up to a certain point, is observed, as seen in the total capacitance of composite in Fig 5-6. Th is result indicates that the deposition of the RuOxnH2O increases with the increasing PProDOP amount. Therefore it is obvious that the increase in the total capacitance of the composite material is not solely due to the contribution of the increased amount of PProDOP. Figure 5-7. Capacitances of the RuOxnH2O (in the composite) as a function of the initial capacitance of PProDOP before th e ruthenium oxide deposition. The results in Fig 5-6 and 5-7 show that ru thenium chloride species diffuse through the polymer matrix to the current collect or substrate and formation of RuOxnH2O occurs throughout the polymer matrix, probably covering the polymer surface. Therefor e, the initial linear dependence of the capacitance of th e composite to the relatively thinner PProDOP layers (up to 12mF/cm2) is observed in Fig 5-6. If RuOxnH2O were to be forming only as a top layer on the 0 5 10 15 20 25 30 35 40 0 5 10 15CRuO2(mF/cm2)CPProDOP(mF/cm2)
150 surface of PProDOP but now diffusing through the matrix, the capacitance of ruthenium oxide would be expected to show a stable dependence instead of linear increa se with the increasing PProDOP thickness since there would be no signif icant difference between the surfaces of the polymer layer with different thicknesses. Accord ingly, a barrier occurs in the diffusion of ruthenium chloride species through the polymer matr ix to the current collector substrate as the PProDOP layer gets thicker. Therefore, there is a low or almost i nhibited deposition of ruthenium oxide occuring on thicker layer of PPr oDOPs, resulting in the lower capacitances of the final composite. In conclusion, optimization of the thickness of the conducting polymer film layer to a certain amount of ruthenium utilizat ion is necessary to re alize large specific capacitance from the composite material. Furthermore, either the ruthenium deposition rate changes or some portion of deposited ruthenium oxide must be not pa rticipating in the el ectrochemical response of the composite material, in the range where the capacitance st abilizes with increasing PProDOP thickness, which would eventually reduces the specific capacitance. Consequently, the electrochemical utilization of ruthenium oxide is conf ined to the thin films of PProDOP. The mass of the composite material with the capacitance (45mF/cm2) was measured by directly weighing the material and found as 68 g/cm2. Consequently, the maximum specific capacitance of the composite film was calculated as 660F/g. Howeve r, an inconsistency arises compared to the mass of the PProDOP before the RuO2nH2O deposition. These results reasonably direct to the conclusi on that the most of PProDOP is being dissolved into the solution during the deposition of RuOxnH2O, yet PProDOP still increase s the deposition amount of RuO2nH2O per unit area, compared to the deposition of RuOxnH2O on bare SWNTs.
151 SEM Analysis The surface morphologies of the composites of RuOxnH2O and PProDOP were examined by SEM. The SEM images of RuOxnH2O deposited for 300cycles on bare SWNTs film and on PProDOP/Sticky-PF/SWNTs film electrodes are shown in Fig 5-8A and B respectively. Figure 5-8. Scanning electron micrographs of RuOxnH2O (300 deposition cycles) on A) bare SWNTs film and B) PProDOP|Sticky-PF|SWNTs substrates. The SEM images of both films have demonstrated homogeneous morphologies. Compared to the AFM image of SWNT film in Fig 3-15, it is clear in Fig 5-8 that RuOxnH2O layer forms a sufficiently thick layer to cover the surface of SWNTs since the noodle-like structure of the nanotube bundles cannot be realized underneath this layer. In gene ral the surface images of both films; RuOxnH2O only and the composite of RuOxnH2O with PProDOP shown in Fig 5-8A&B, respectively, reveal similar nodular structures sh owing spherical grains in various sizes with a compact morphology. There have been several cracks observed under SEM throughout the both films. Since the pseudocapacitance of rutheniu m oxide is directly proportional to its active surface area and the electrochemical response of thes e films were observed to be very reversible, the micropores must be distributed in between the grains in order to supp ort the efficient proton diffusion. This morphology might be due to the colla psing of the porous structured films forming B A 3 00nm 2 00nm
152 a compacted structure during the drying processes used to evaporate the residual solvent trapped inside the matrix of the films. There also no dist inctive trends that were observed from the cross section of the films. The surface morphologies of these films also show similar structures compared to that of PProDOP alone (Fig 3-22) which probably have a less rough surface. These images also support the hypothesis that RuOxnH2O deposits are distributed throughout the matrix of the PProDOP film a nd also covering the surface. XPS Analysis For the e lemental inspection of the prepared composite film, the XPS spectra of films both with and without RuOxnH2O were taken as shown in Fig 5-9A and B, respectively. In both spectra the peaks due to C, N, and S along with the advantageous oxygen peak observed. Since the binding energies of Ru 3d and C 1s, wh ich is the dominant element in the composite material, interfere with each other, it was necessary to exam ine Ru 3p1 and 3p3 peaks as the evidence of the RuOxnH2O existence in the spectrum of the composite. The peaks at binding energies of both Ru 3p1 and 3p3 are broad which i ndicates that the oxide film are consists of mixed hydroxyl ruthenium species in various oxid ation states. From the comparison of the Ru 3p3 with N 1s peaks, the percentages of th ese species were found to be 52.3% and 47.5, approximately in 1:1 ratio. The absence of any F 1s peak in the spectra of the composite as in the spectra of PProDOP|Sticky-PF|SWNT s electrode and also the increas ed percentage of S 2s and 2 p3 peaks in the spectrum of the composite co mpared to that of PProDOP|Sticky-PF|SWNTs electrode indicate that the PPr oDOP is probably doped with SO4 or HSO4 instead of (N2(SO2CF3)dopant anion. However, it is difficult to ma ke an exact quantitative analysis since there were small impurities found on both sample s. Accordingly, XPS results only show the existence of the mixed oxidation states of ruthenium oxide species in the composite.
153 Figure 5-9. XPS survey scans of A) RuOxnH2O|PProDOP|Sticky-PF|SWNTs film and B) PProDOP|Sticky-PF|SWNTs films.
154 Characterization of the RuOxnH2O|PProDOP Supercapacitor The supercapacitor devices were constructe d utilizing the same composite of RuOxnH2O and PProDOP materials in a symmetric confi guration on both positive and negative electrodes. Sticky-PF coated SWNTs films (1cm2) were used as current colle ctor substrates and 0.5M H2SO4 was used as the conductive electrolyte media. As was in the polymer-only Type I devices, one of the electrodes was initially set in the fully oxidized state and the other in the fully neutralized form to establish the initial charge state of the device. Subsequent discharging of the device causes neutralization of the p-doped PProDOP and reduction of oxyruthenium species proceeding with concurrent oxidation of the cat hode PProDOP films with the oxidation of the ruthenium oxide film in the device. Since the quantity of charge that leaves one electrode must be equal to that injected onto th e other electrode in two-electrode systems, the charge capacity of each electrode prepared by electrodepos ition of first PProDOP and then RuOxnH2O (so that the amount of materials), was matched carefully be fore switching to the two-electrode system. Accordingly, no over accumulation of charges could have occurred when the device was switched between its charged and uncharged states, achieving optim al and balanced switching. 0.00.20.40.60.81.0 -1.0 -0.5 0.0 0.5 1.0 I (mA/cm2)E(V) A0.00.20.40.60.81.0 -30 -20 -10 0 10 20 30 C (mF/cm2)E (V) B Figure 5-10. A) Cyclic voltammograms and B) capaci tances as a function of applied potential of TYPE I Hybrid RuOxnH2O|PProDOP supercapacitor with Sticky-PF|SWNTs substrates using 0.5M H2SO4 at 10 to 50mV/s with 10mV/s intervals.
155 Fig 5-10A and B shows the cyclic voltammogr ams and the capacitance as a function of applied potential of the supercapacitor device at different s can rates from 10 to 50mV/s with 10 mV intervals in a 1V potential range, respectivel y. As in PProDOP-only Type I devices, here again there is a complete overlap between the electrochemical responses of anode and cathode electrodes since same material is utilized on both electrodes. Thus, the shape of the cyclic voltammogram of the supercapacitors would typica lly be a rectangular or parallelogram plot, showing a constant current with in the operating voltage window of the device. As mentioned earlier, the matching working-potential windows of RuOxnH2O and PProDOP materials become critical here. As seen in Fig 5-10A, as expect ed, the voltammograms present the near-rectangular shape. The voltammograms indicate almost mirro r images in between the anodic or cathodic scanning. A slight deviation in the voltammograms taken at relativ ely fast scan rates of 40 and 50mV/s, compared to slower s can rates of 10mV/s was observed. As seen in the voltammograms of the electrodes in a three-electrode system, here again, the linear dependence of the current density on the scan rate was observed, indicative of the nondiffusion limited redox switching. The vo ltammograms demonstrate that RuOxnH2O |PProDOP composite devices show reproducible, re latively fast and efficient switching. The capacitance graphs of the devices show n in Fig 5-10B were calculated using the relationship: C=i/s (Equation 2-2). The capacitances show a constant value that is independent of the applied potential for the larger portion of potential ranges as dictated by the equation. Similar deviation were observed in the voltammograms taken at the relatively faster scan rates of 40 and 50mV/s, compared to slower scan rates of 10mV/s were also seen in the capacitance graphs.
156 The average capacitance values were cal culated from the anodic and cathodic voltammetric charges, which were examined by integrating the voltammograms in the working potential range as had been performed for each of the film electrodes. Each electrode separately presented a capacitance of 34mF/cm2. The capacitance values obtai ned from the supercapacitor device was calculated as 16.9mF/cm2 as shown in Fig 5-10B. Since the configuration of the device consisting of the composite material was very similar to polymer-only Type I devices, the relationship between the capacitanc e values of the device was expected to be one half the capacitances of the two individua l electrodes connected in seri es in a two electrode device system, as explained in Chapter 3. Therefore, th e capacitance of the device was expected to be 17mF/cm2, which exactly matches to the observed value of 16.9mF/cm2.The corresponding mass value was calculated as 103 g/cm2 for both positive and negative electrodes and the specific capacitance of the device was determined to be 165F/g. Although there was not any stability experiments were conducted, the fi lms and devices were switched at least 50 cycles to collect the presented data and exhibit stable and reproducible performance. The energy density of the device, shown in Fig 5-10, was calcu lated using the relationship; E = QV (Equation 1-2). The averag e charge density was determined from the cyclic voltammograms and the average voltage of the de vice was taken as the half of the potential difference between the electrodes at the initial charged state of the device. The energy density was found as 11Wh/kg. The average power densit y was determined using the relationship: P= iV (Equation 2-9). The average current density wa s determined form the cyclic voltammograms taken at different scan rates and the aver age power density was calculated as 2289W/kg. Compared to PProDOP devices in Chapter 3, the energy density was increased 4.3 folds and power density was increased 4 folds. These energy and power densities are comparable with the
157 battery and electrochemical capacito rs values estimated in the Ra gone Plot (Fig 1-1). Table 5-1 summarize the areal and specific capacita nces of the composite film of RuOxnH2O|PProDOP and device shown in Fig 5-10, along with the energy and power values. Table 5-1. Capacitances, energy and power densities of the film and devices of RuOxnH2O|PProDOP composite. Film Device C(mF/cm2) C (F/g) C(mF/cm2) C (F/g) E (Wh/kg) P(W/kg) 45 660 17 165 11.3 2289 Conclusions and Perspective In summ ary, this chapter demonstrates a novel device design util izing a composite of conducting polymer and hydrous metal oxide as the charge storage material in Type I supercapacitors. A two step a ll electrochemical method for pr eparation of interpenetrating PProDOP and hydrous ruthenium oxide composite was introduced. In this method, PProDOP films have been electrochemically deposited on SWNT films covered by Sticky-PF, and then these films have been used as the scaffolds on which amorphous ruthenium oxide could be electrodeposited. The control of the electrochemical parameters allow some control of the thickness of the component and hence the poros ity. The electrochemical properties of the corresponding films were investigated by using cyclic volta mmetry and galvanostatic chargingdischarging methods. The results show that ru thenium oxide deposition was increased 100% in capacitance per unit area on PProDOP compared to bare SWNT films. The method introduced in this chapter effectively demonstrated the contri bution of each composite materials to the final capacitance outcome of the film The morphological properties of the resulting films were investigated by using SEM and the images of th e composite film were compared with the SEM images of SWNTs substrate itself and also th e PProDOP itself on SWNTs substrate. Moreover, the elemental inspection of the composite film was investigated by XPS and the spectra of
158 PProDOP films before and after ruthenium oxide deposition were taken. In all electrochemical, morphological and elemental analysis indicate the synergy effect between the components of the composite film with incr eased capacitance values. It is expected that further work to be conducted after this di ssertation will focus on optimizing the proper combination of materials a nd electrode/device designs to achieve the best performance for varying intentions. In further morphological studies of these sponge-like electrode materials, ESEM (Environmental Scanning Electron Microscope) th at operates in a gaseous environment in the specimen chamber, rather than in vacuum like all other electron microscopes, could provide enorm ous advantages to investigate the porosity of the com posite film.
159 LIST OF REFERENCES 1. Conway, B. E. Electrochemical Capac itors: Scientific Funda mentals and Technological Applications ; Kluwer Academ ic/Plenum: New York, NY 1999. 2. Becker, H. E. Low voltage electrolytic capacitor ; General Electric Co, U. S. Patent 2,800,616, 1957. 3. Rightmire, R. A., Electrical energy storage apparatus U. S. Patent 3,288,641, 1966. 4. Endo, M.; Takeda, T.; Kim, Y.J.; Koshiba, K.; Ishii, K. Carbon Science 2001, 1, 117. 5. High Capacitance Capacitors [Internet]. NEC-Tokin Coorporatio n; [cited 2009 Feb 02]. Available from: http://www.nectokin.net/product/ hypercapacitor/outline02.htm l 6. Conway, B. E. Journal of the Electrochemical Society 1991, 138, 6, 1539. 7. Conway, B. E.; Kozlowska, H. A. Accounts of Chemical Research 1981, 14, 2, 49. 8. Jordanov, S. H.; Kozlowska, H. A.; Vukovic, M.; Conway, B. E. Journal of the Electrochemical Society 1978 125, 1471. 9. Galizzioli, R.; Tantardini, F.; Trasatti, S. Journal of Applied Electrochemistry 1974, 4, 57. 10. International Energy Agency, Report on the first phase of the IEA implementing agreement for hybrid electric vehicle technology and programmes 1993-1999 ; International Energy Agency, Paris 1999. 11. The charge of ultracapacitors. [Internet]. Sh indall, Joel; Institute of Electrical and Electronics Engineers Spectrum ; [cited 2009 Jan 15]. Available from: http://www.spectrum.ieee.org/nov07/5636 12. Kotz, R.; Carlen, M. Electrochimica Acta 2000, 45, 15-16, 2483. 13. Winter, M.; Brodd, R. J. Chemical Reviews 2004, 104, 10, 4245. 14. Conway, B. E.; Birss, V.; Wojtowicz, J. Journal of Power Sources 1997, 66, 1-2, 1. 15. Burke, A. Journal of Power Sources 2000, 91, 1, 37. 16. Hughes, M.; Chen, G. Z.; Shaffer, M. S. P.; Fray, D. J.; Windle, A. H. Chemistry of Materials 2002, 14, 4, 1610. 17. Niu, C. M.; Sichel, E. K.; Hoch, R.; Moy, D.; Tennent, H. Applied Physics Letters 1997, 70, 11, 1480. 18. Mund, K.; Richter, G.; We idlich, E.; Fahlstrom, U Pacing and Clinical Electrophysiology 1986, 9, 1225.
160 19. Sawai, K.; Ohzuku, T. Journal of the Elect rochemical Society 1997, 144, 3, 988. 20. Rossier, J. S.; Girault, H. H. Physical Chemistry Chemical Physics 1999, 1, 15, 3647. 21. An, K. H.; Kim, W. S.; Park, Y. S.; Choi, Y. C.; Lee, S. M.; Chung, D. C.; Bae, D. J.; Lim, S. C.; Lee, Y. H. Advanced Materials 2001, 13, 497. 22. Conway, B. E. Progress in Surface Science 1984, 16, 1, 1. 23. Conway, B. E.; Pell, W. G. Journal of Solid State Electrochemistry 2003, 7, 9, 637. 24. Sarangapani, S.; Tilak, B. V.; Chen, C. P. Journal of the Electrochemical Society 1996, 143, 11, 3791. 25. Burke, A.; Arulepp, M. Electrochemical Society Proceedings 2001, 21, 576. 26. Barbieri, O.; Hahn, M.; Herzog, A.; Kotz, R. Carbon 2005, 43, 6, 1303. 27. Zhou, H.; Zhu, S.; Hibino, M.; Honma, I. Journal of Power Sources 2006 122, 219. 28. Sevilla, M.; Alvarez, S.; Centeno, T. A.; Fuertes, A. B.; Stoeckli, F. Electrochimica Acta 2007, 52, 9, 3207. 29. Li, L.; Song, H.; Chen, X. Electrochimica Acta 2006, 51, 5715. 30. Fernandez, J. A.; Morishita, T.; Toyoda, M.; Inagaki, M.; Stoeckli, F.; Centeno, T. A. Journal of Power Sources 2008, 175, 1, 675. 31. Salitra, G.; Soffer, A.; Eliad, L.; Cohen, Y.; Aurbach, D. Journal of the Electrochemical Society 2000, 147, 7, 2486. 32. Vix-Guterl, C.; Frackowiak, E.; Jurewicz, K. ; Friebe, M.; Parmentier, J.; Beguin, F. Carbon 2005, 43, 6, 1293. 33. Janes, A.; Permann, L.; Arulepp, M.; Lust, E. Electrochemistry Communications 2004, 6, 3, 313. 34. Chmiola, J.; Yushin, G.; Gogotsi, Y.; Portet, C.; Simon, P.; Taberna, P. L. Science 2006, 313, 5794, 1760. 35. Frackowiak, E.; Beguin, F. Carbon 2001, 39, 6, 937. 36. Pandolfo, A. G.; Hollenkamp, A. F. Journal of Power Sources 2006, 157, 1, 11. 37. Shukla, A. K.; Sampath, S.; Vijayamohanan, K. Current Science 2000, 79, 12. 38. Wang, J.; Zhang, S. Q.; Guo, Y. Z.; Shen, J.; Attia, S. M.; Zhou, B.; Zheng, G. Z.; Gui, Y. S. Journal of the Electrochemical Society. 2001, 148, 6, D75.
161 39. Nelson, P. A.; Owen, J. R. Journal of the Electrochemical Society 2003, 150, 10, A1313. 40. Moore, V. C.; Strano, M. S.; Haroz, E. H.; Hauge, R. H.; Smalley, R. E.; Schmidt, J.; Talmon, Y. Nano Letters 2003, 3, 10, 1379. 41. Wu, Z. C.; Chen, Z. H.; Du, X.; Logan, J. M.; Sippel, J.; Nikolou, M.; Kamaras, K.; Reynolds, J. R.; Tanner, D. B.; Hebard, A. F.; Rinzler, A. G. Science 2004, 305, 5688, 1273. 42. Lee, K.; Wu, Z.; Chen, Z.; Ren, F.; Pearton, S. J.; Rinzler, A. G. Nano Letters 2004, 4, 5, 911. 43. Frackowiak, E.; Jurewicz, K.; Delpeux, S.; Beguin, F. Journal of Power Sources 2001, 97, 822. 44. Frackowiak, E.; Beguin, F. Carbon 2002, 40, 10, 1775. 45. Barisci, J. N.; Wallace, G. G.; Chatt opadhyay, D.; Papadimitrakopoulos, F.; Baughman, R. H. Journal of the Electrochemical Society 2003, 150, E409. 46. Shiraishi, S.; Kurihara, H.; Okab e, K.; Hulicova, D.; Oya, A. Electrochemistry Communications 2002, 4, 7, 593. 47. Talapatra, S.; Kar, S.; Pal, S. K.; Vajtai, R.; Ci, L.; Victor, P.; Shaijumon, M. M,; Kaur, S.; Nalamasu, O.; Ajayan, P. M. Nature Nanotechnology 2006 1, 112. 48. Pushparaj, V. L.; Shaijumon, M. M.; Kumar, A.; Murugesan, S.; Ci, L.; Vajtai, R.; Linhardt, R. J.; Nalamasu, O.; Ajayan, P. M. Proceedings of the National Academy of Sciences of the United States of America 2007, 104, 34, 13574. 49. Schindall, Joel Institute of Electrical and Electro nics Engineers Power Electronics Society Newsletter 2008, First Quarter, 20, 1, 32. 50. Wohlfahrt-Mehrens, M.; Schenk, J.; Wilde, P. M.; Abdelmula, E.; Axmann, P.; Garche, J. Journal of Power Sources 2002, 105, 182. 51. Huggins, R. A. Solid State Ionics 2000, 134, 1-2, 179. 52. Roberson, S. L.; Finello, D.; Davis, R. F.; Lui, W.; Pell, W .; Conway, B. E. Proceedings of the 7th International Se minar on Double-layer Capacitors and Similar Energy Storage Devices Deerfield Beach, December 1997. 53. Kurzweil, P.; Schmid, O.; Loffler, A. Proceedings of the 7th International Seminar on Double-layer Capacitors and Similar Energy Storage Devices Deerfield Beach, December 1997. 54. Long, J. W.; Swider, K. E.; Merzbacher, C. I.; Rolison, D. R. Langmuir 1999, 15, 3, 780. 55. Zheng, J. P.; Jow, T. R. Journal of the Electrochemical Society 1995, 142, 1, L6.
162 56. Cimino, A.; Carra, S. Electrodes of Conductive Metallic Oxides Part A .; Trasatti, S., Ed.; Elsevier: New York, NY 1980; 97. 57. Miousse, D.; Lasia, A. Journal of New Materials fo r Electrochemical Systems 1999, 2, 71. 58. Jeong, Y. U.; Manthiram, A. Electrochemical and Solid State Letters 2000, 3, 5, 205. 59. Lee, H. Y.; Manivannan, V.; Goodenough, J. B., Comptes rendus de l'Academie des sciences Series Iic Chemistry 1999 2, 11-13, 565. 60. Lee, H. Y.; Goodenough, J. B. Journal of Solid State Chemistry 1999, 144, 1, 220. 61. Drezen, T.; Kwon, N.-H.; Bowenb, P.; T eerlinck, I.; Isono, M.; Exnar, I. Journal of Power Sources 2007, 174, 949. 62. Wang, Y.; Takahashi, K.; Lee, K.; Cao, G. Z. Advanced Functional Materials 2006, 16, 9, 1133. 63. Rudge, A.; Raistrick, I.; Gotte sfeld, S.; Ferraris, J. P. Electrochimica Acta 1994, 39, 2, 273. 64. Arbizzani, C.; Mastragostino, M.; Scrosati, B. Handbook of Organic Conductive Molecules and Polymers ; Nalwa, H. S., Ed.; John Wiley & Sons, New York, NY, 1997; 11-1. 65. Novak, P.; Muller, K.; Santhanam, K. S. V.; Haas, O. Chemical Reviews 1997, 97, 1, 207. 66. Mastragostino, M.; Arbizzani, C.; Soavi, F. Journal of Power Sources 2001, 97-8, 812. 67. Ghosh, S.; Inganas, O. Advanced Materials 1999, 11, 14, 1214. 68. Mastragostino, M.; Arbizzani, C.; Paraventi, R.; Zanelli, A. Journal of the Electrochemical Society 2000, 147, 2, 407. 69. Fusalba, F.; El Mehdi, N .; Breau, L.; Belanger, D. Chemistry of Materials 1999, 11, 10, 2743. 70. Gurunathan, K.; Murugan, A. V.; Marimuthu, R.; Mulik, U. P.; Amalnerkar, D. P. Materials Chemistry and Physics 1999, 61, 3, 173. 71. Fusalba, F.; Belanger, D. Journal of Physical Chemistry B 1999, 103, 42, 9044. 72. Carlberg, J. C.; Inganas, O. Journal of the Electrochemical Society 1997, 144, 4, L61. 73. Rudge, A.; Davey, J.; Raistrick, I.; Gottesfeld, S.; Ferraris, J. P. Journal of Power Sources 1994, 47, 1-2, 89.
163 74. Irvin; J. A.; Irvin D. J.; Stenger-Smith, J. D. Handbook of Conducting Polymers Electroactive Polymers for Batteries and Supercapacitors .; Skotheim, T. A.; Reynolds, J. R., Ed.; Taylor and Francis Gr oup, LLC: Boca Raton, FL, 2007; 9-1. 75. Leroux, F.; Koene, B. E.; Nazar, L. F. Journal of the Electrochemical Society 1996, 143, 9, L181. 76. Mu, S.; Ye, J.; Wang, Y. Journal of Power Sources 1993, 45, 153. 77. Mohammadi, A.; Inganas, O.; Lundstrom, I. Journal of the Electrochemical Society 1986, 133, 5, 947. 78. Moon, D. K.; Padias, A. B.; Hall, H. K.; Huntoon, T.; Calvert, P. D. Macromolecules 1995, 28, 18, 6205. 79. Meyer, W. H. Advanced Materials 1998, 10, 6, 439. 80. Ratner, M. A.; Johansson, P.; Shriver, D. F. Materials Research Society Bulletin 2000, 25, 3, 31. 81. Gray, F.; Armand, M. Handbook of Battery Material s: Polymer Electrolytes ; Besenhard, J. O., Ed.; Wiley-VCH, New York, NY, 1999; 8-1. 82. Novak, P.; Vielstich, W. Journal of the Electrochemical Society 1990, 137, 6, 1681. 83. Bittihn, R.; Ely, G.; Woeffler, F.; Muenstedt, H.; Naarmann, H.; Naegele, D. Makromolekulare Chemie Macromolecular Symposium 1987, 8, 51. 84. Naegele, D.; Bittihn R. Solid State Ionics 1988, 28-30, 2, 983. 85. Panero, S.; Prosperi, P.; Scrosati, B. Electrochimica Acta 1987, 32, 10, 1465. 86. Trinidad, F.; Alonsolopez, J.; Nebot, M. Journal of Applied Electrochemistry 1987, 17, 1, 215. 87. Naoi, K.; Sakai, H.; Ogano, S.; Osaka, T. Journal of Power Sources 1987, 20, 3-4, 237. 88. Prasad, K. R.; Munichandraiah, N. Electrochemical and Solid State Letters 2002, 5, 12, A271. 89. Ofer, D.; Crooks, R. M.; W righton, M. S. Journal of the American Chemical Society 1990, 112, 22, 7869. 90. Kinlen, P. J., Mbugua, J.; Kim, Y.-G.; Jung, J. -H., Viswanathan, S.; Liu, J., Abstract of papers, 236th ACS National Meeting, Philadelphia, PA, 2008. 91. Ibrisagic, Z. Chemical and Biochemical Engineering 1993, 7, 4, 191.
164 92. Lee, J. Y.; Ong, L. H.; Chuah, G. K. Journal of Applied Electrochemistry 1992, 22, 8, 738. 93. Echigo, Y.; Asami, K.; Takahashi, H.; Inoue K.; Kabata, T.; Kimura, O.; Ohsawa, T. Synthetic Metals 1993, 57, 1, 3611. 94. Arbizzani, C.; Mastragostino, M.; Meneghello, L. Electrochimica Acta 1996, 41, 1, 21. 95. Kitani, A.; Iwai, K.; Ito, S. Electrochemistry 1999, 67, 1262. 96. Nishio, K.; Fujimoto, M.; Yoshinaga, N.; Fu rukawa, N.; Ando, O.; Ono, H.; Suzuki, T. Journal of Power Sources 1991, 34, 2, 153. 97. Hashmi, S. A.; Kumar, A.; Tripathi, S. K. European Polymer Journal 2005, 41, 6, 1373. 98. Wu, H.-M.; Shy, H.-J.; Ko, H.-W. Journal of Power Sources 1989, 27, 59. 99. Shacklette, L. W.; Jow, T. R.; Maxfield, M.; Hatami, R. Synthetic Metals 1989, 28, 1-2, C655. 100. Yamamoto, T.; Zama, M.; Hishinuma, M.; Yamamoto, A. Journal of Applied Electrochemistry 1987, 17, 3, 607. 101. Stenger-Smith, J. D.; Webber, C. K.; Anderson, N.; Chafin, A. P.; Zong, K.; Reynolds, J. R. Journal of the Electrochemical Society 2002, 149, 8, A973. 102. Haringer, D.; Novak, P.; Haas, O.; Piro, B.; Pham, M. C. Journal of the Electrochemical Society 1999, 146, 7, 2393. 103. Levine, K.; Iroh, J. O. Journal of Materials Chemistry 2001, 11, 9, 2248. 104. Arbizzani, C.; Mastragostino, M. Electrochimica Acta 1990, 35, 1, 251. 105. Rasch, B.; Novak, P.; Vielstich, W. Synthetic Metals 1991, 43, 1-2, 2963. 106. Ramasamy, R. P.; Veeraraghavan, B.; Haran, B.; Popov, B. N. J. Power Sources 2003, 124, 1, 197. 107. Mastragostino, M.; Scrosati, B. Journal of the Elect rochemical Society 1985, 132, 5, 1259. 108. Shacklette, L. W.; Toth, J. E.; Murthy, N. S.; Baughman, R. H. Journal of the Electrochemical Society 1985, 132, 7, 1529. 109. Tourillon, G.; Garnier, F Journal of the Electrochemical Society 1983, 130, 10, 2042. 110. Sato, M.; Tanaka, S.; Kaeriyama, K. Synthetic Metals 1986, 14, 4, 279. 111. Walker, C. W. Journal of the Electrochemical Society 1991, 138, 6, 1559.
165 112. Nagatomo, T.; Omoto, O. Journal of the Electrochemical Society 1988, 135, 9, 2124. 113. Mastragostino, M.; Marinangeli, A. M. ; Corradini, A.; Arbizzani, C. Electrochimica Acta 1987, 32, 11, 1589. 114. Jow, T. R.; Jen, K. Y.; Elsenbaumer, R. L.; Shacklette, L. W.; Angelopoulos, M.; Cava, M. P. Synthetic Metals 1986, 14, 1-2, 53. 115. Bartley, T. Procedings of Advanced Capacitor World Summit Washington, DC, 2004, paper 13. 116. Burke, A. F.; Hardin, J. E.; Dowgiallo, E. J. Proceedings of the 34th Power Sources Conference Cherry Hill, NJ, 1990. 117. Miller, J. R. Electrochemical Capacitors ; Delnick F. M.; Tomkiewicz, M., Ed.; The Electrochemical Society Proceed ings Series, Pennington, NJ, 1996. 118. Miller, J. R. Proceeding of 56th Annual Meeting of the International Society of Electrochemistry Busan, Korea, 2005, paper 3C-001-IL. 119. Schottland, P.; Zong, K.; Gaupp, C. L.; Thomps on, B. C.; Thomas, C. A.; Giurgiu, I.; Hickman, R.; Abboud, K. A.; Reynolds, J. R. Macromolecules 2000, 33, 19, 7051. 120. Walczak, R. M.; Reynolds, J. R. Advanced Materials 2006, 18, 9, 1121. 121. Steckler, T. T.; Abboud, K. A.; Craps, M. ; Rinzler, A. G.; Reynolds, J. R. Chemical Communications 2007, 46, 4904. 122. Reynolds, J. R.; Walczak, R. M.; Van der Aa, E. M.; Rinzler, A. Supramolecular polymers associative to carbon nanotubes U.S. Patent Application 20080287638, 2008. 123. Conway, B. E.; Pell, W. G Journal of Power Sources 2002, 105, 2, 169. 124. Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications 2nd ed.; Wiley: New York, NY 2001. 125. Arbizzani, C.; Mastragostino, M.; Soavi, F. Journal of Power Sources 2001, 100, 1-2, 164. 126. Ghosh, S.; Inganas, O. Journal of the Electrochemical Society 2000, 147, 5, 1872. 127. Matsum oto, M. Electrical Phenomena at Interfaces : Fundamentals, Measurements, and Applications 2nd ed.Vol. 76; Ohshima, H.; Furusawa, K., Ed.; Marcel Dekker: New York, NY, 1998. 128. Dietrich, M.; Heinze, J.; Heywang, G.; Jonas, F. Journal of Electroanalytical Chemistry 1994, 369, 1-2, 87.
166 129. Groenendaal, B. L.; Jonas, F.; Freitag, D.; Pielartzik, H.; Reynolds, J. R. Advanced Materials 2000, 12, 7, 481. 130. Heywang, G.; Jonas, F. Advanced Materials 1992, 4, 2, 116. 131. Kumar, A.; Welsh, D. M.; Morvant, M. C.; Piroux, F.; Abboud, K. A.; Reynolds, J. R. Chemistry of Materials 1998, 10, 3, 896. 132. Welsh, D. M.; Kloeppner, L. J.; Madrigal, L. ; Pinto, M. R.; Thompson, B. C.; Schanze, K. S.; Abboud, K. A.; Powell, D.; Reynolds, J. R. Macromolecules 2002, 35, 17, 6517. 133. Cirpan, A.; Argun, A. A.; Grenier, C. R. G.; Reeves, B. D.; Reynolds, J. R. Journal of Materials Chemistry 2003, 13, 10, 2422. 134. Mishra, S. P.; Krishnamoorthy, K.; Sahoo, R.; Kumar, A. Journal of Polymer Science Part a-Polymer Chemistry 2005, 43, 2, 419. 135. Reeves, B. D.; Grenier, C. R. G.; Argun, A. A.; Cirpan, A.; McCarley, T. D.; Reynolds, J. R Macromolecules 2004, 37, 20, 7559. 136. Walczak, R. M., Synthetic methodology as a basis for conducting polymer design. Dissertation, University of Florida, Gainesville, FL, 2006. 137. Reynolds, J. R.; Ruiz, J. P.; Child, A. D.; Nayak, K.; Marynick, D. S. Macromolecules 1991, 24, 3, 678. 138. Ruiz, J. P.; Dharia, J. R.; Reynolds, J. R.; Buckley, L. J. Macromolecules 1992, 25, 2, 849. 139. Child, A. D.; Sankaran, B.; Larmat, F.; Reynolds, J. R. Macromolecules 1995, 28, 19, 6571. 140. Pfluger, P.; Street, G. B. Journal of Chemical Physics 1984, 80, 1, 544. 141. Thomas, C. A.; Zong, K.; Schottland, P.; Reynolds, J. R. Advanced Materials 2000, 12, 3, 222. 142. Zotti, G.; Schiavon, G.; Berlin, A.; Pagani, G. Chemistry of Materials 1993, 5, 4, 430. 143. Sonm ez, G.; Schwendeman, I.; Schottla nd, P.; Zong, K. W.; Reynolds, J. R. Macromolecules 2003, 36, 3, 639. 144. Roncali, J. Chemical Reviews 1992, 92, 4, 711. 145. Schwendeman, I., Optical and transport properties of conjugated polym ers and their application to electrochromic devices. Dissertation, University of Florida, Gainesville, FL, 2002.
167 146. Snook, G. A.; Peng, C.; Fray, D. J.; Chen, G. Z. Electrochemistry Communications 2007, 9, 1, 83. 147. Wu, M. Q.; Snook, G. A.; Gupta, V.; Shaffer, M.; Fray, D. J.; Chen, G. Z. Journal of Materials Chemistry 2005, 15, 23, 2297. 148. Pyo, M.; Bohn, C. C.; Smela, E.; Reynolds, J. R.; Brennan, A. B. Chemistry of Materials 2003, 15, 4, 916. 149. Frackowiak, E.; Jurewicz, K.; Szostak, K.; Delpeux, S.; Beguin, F. Fuel Processing Technology 2002, 77, 213. 150. Shinomiya, T.; Gupta, V.; Miura, N. Electrochimica Acta 2005, 51, 21, 4412. 151. Zhihong Chen, Z., Electric field induced transparency m odulation in single wall carbon nanotube ultra-thin films and a method to separate metallic and semiconducting nanotubes, Dissertation, University of Florida, Gainesville, FL, 2003. 152. Sinani, V. A.; Gheith, M. K.; Yaroslavov, A. A.; Rakhnyanskaya, A. A.; Sun, K.; Mamedov, A. A.; Wicksted, J. P.; Kotov, N. A. Journal of the American Chemical Society 2005, 127, 10, 3463. 153. Bohn, C.; Sadki, S.; Brennan, A. B.; Reynolds, J. R. Journal of the Electrochemical Society 2002, 149, 8, E281. 154. Ehli, C.; Rahman, G. M. A.; Jux, N.; Balbinot D.; Guldi, D. M.; Paolucci, F.; Marcaccio, M.; Paolucci, D.; Melle-Franco, M.; Ze rbetto, F.; Campidelli, S.; Prato, M. Journal of the American Chemical Society 2006, 128, 34, 11222. 155. Britz, D. A.; Khlobystov, A. N. Chemical Society Reviews 2006, 35, 7, 637. 156. Yuan, W. Z.; Sun, J. Z.; Dong, Y. Q.; Haussler M.; Yang, F.; Xu, H. P.; Qin, A. J.; Lam, J. W. Y.; Zheng, Q.; Tang, B. Z. Macromolecules 2006, 39, 23, 8011. 157. Chen, R. J.; Zhang, Y. G.; Wang, D. W.; Dai, H. J. Journal of the American Chemical Society 2001, 123, 16, 3838. 158. DuBois, C. J.; Abboud, K. A.; Reynolds, J. R. Journal of Physical Chemistry B 2004, 108, 25, 8550. 159. Thompson, B. C.; Kim, Y. G.; McCarley, T. D.; Reynolds, J. R. Journal of the American Chemical Society 2006, 128, 39, 12714. 160. Colladet, K.; Fourier, S.; Cle ij, T. J.; Lutsen, L.; Gelan, J.; Vanderzande, D.; Nguyen, L. H.; Neugebauer, H.; Sariciftci, S.; Ag uirre, A.; Janssen, G.; Goovaerts, E. Macromolecules 2007, 40, 1, 65. 161. Kulkarni, A. P.; Kong, X. X.; Jenekhe, S. A. Macromolecules 2006, 39, 25, 8699.
168 162. Thompson, B. C.; Madrigal, L. G.; Pinto, M. R.; Kang, T. S.; Schanze, K. S.; Reynolds, J. R. Journal of Polymer Science Part A: Polymer Chemistry 2005, 43, 7, 1417. 163. Sonmez, G.; Sonmez, H. B. Journal of Materials Chemistry 2006, 16, 25, 2473. 164. Galand, E. M.; Mwaura, J. K.; Argun, A. A. ; Abboud, K. A.; McCarley, T. D.; Reynolds, J. R. Macromolecules 2006, 39, 21, 7286. 165. Sonmez, G. Chemical Communications 2005, 42, 5251. 166. Irvin, D. J.; DuBois, C. J.; Reynolds, J. R. Chemical Communications 1999 20, 2121. 167. Thomas, C. A.; Zong, K. W.; Abboud, K. A.; Steel, P. J.; Reynolds, J. R. Journal of the American Chemical Society 2004, 126, 50, 16440. 168. Dubois, C. J., Donor-acceptor methods for band gap control in conjugated polymers. Dissertation, University of Florida, Gainesville, FL, 2003. 169. Van Mullekom, H. A. M.; Vekemans, J. A. J. M.; Havinga, E. E.; Meijer, E. W. Materials Science and Engineering: R: Reports 2001, 32, 1, 1. 170. Trasatti, S. Electrochimica Acta 1991, 36, 2, 225. 171. Takasu, Y.; Murakami, Y. Electrochimica Acta 2000, 45, 25-26, 4135. 172. Zheng, J. P.; Cygan, P. J.; Jow, T. R. Journal of the Elect rochemical Society 1995, 142, 8, 2699. 173. Hu, C. C.; Huang, Y. H. Journal of the Elect rochemical Society 1999, 146, 7, 2465. 174. Fang, Q. L.; Evans, D. A.; Roberson, S. L.; Zheng, J. P. Journal of the Electrochemical Society 2001, 148, 8, A833. 175. Zheng, J. P.; Huang, C. K. Journal of New Materials fo r Electrochemical Systems 2002, 5, 1, 41. 176. Zheng, J. P. Electrochemical and Solid State Letters 1999, 2, 8, 359. 177. Jow, T. R.; Zheng, J. P. Journal of the Elect rochemical Society 1998, 145, 1, 49. 178. Zheng, J. P.; Xin, Y. Journal of Pow er Sources 2002, 110, 1, 86. 179. Long, J. W.; Swider, K. E.; Merzbacher, C. I.; Rolison, D. R. Langmuir 2003, 19, 6, 2532. 180. Long, J. W.; Ayers, K. E.; Rolison, D. R. Journal of Electroanalytical Chemistry 2002, 522, 1, 58.
169 181. McKeown, D. A.; Hagans, P. L.; Carette, L. P. L.; Russell, A. E.; Swider, K. E.; Rolison, D. R. Journal of Physical Chemistry B 1999, 103, 23, 4825. 182. Vukovic, M.; Cukman, D. Journal of Electroanalytical Chemistry 1999, 474, 2, 167. 183. Kim, I. H.; Kim, K. B. Electrochemical and Solid State Letters 2001, 4, 5, A62. 184. Hu, C. C.; Chen, W. C.; Chang, K. H. Journal of the Electrochemical Society 2004, 151, 2, A281. 185. Sugimoto, W.; Iwata, H.; Yokoshima K.; Murakami, Y.; Takasu, Y. Journal of Physical Chemistry B 2005, 109, 15, 7330. 186. Hu, C. C.; Chu, C. H. Materials Chemistry and Physics 2000, 65, 3, 329. 187. He, B. L.; Zhou, Y. K.; Zhou, W. J.; Dong, B.; Li, H. L. Materials Science and Engineering A Structural Materials: Properties Microstructure and Processing 2004, 374, 1-2, 322. 188. Park, J. H.; Ko, J. M.; Park, O. O. Journal of the Electrochemical Society 2003, 150, 7, A864. 189. Wen, T. C.; Hu, C. C. Journal of the Electrochemical Society 1992, 139, 8, 2158.
170 BIOGRAPHICAL SKETCH Merve Ertas was born an d raised in Turkey. She received her B.S. in 2002 followed by her M.S. in 2003 in chemistry at METU, Ankara, wh ere she worked in the research group of Professor Levent Toppare. Soon after, she moved to the University of Florida, Department of Chemistry to begin her doctoral studies under th e supervision of Professor John R. Reynolds. During her time at the University of Florida, sh e also conducted research in several projects in collaboration with Professor Andrew G. Ri nzler in the Department of Physics.