1 ENHANCED UTILITY OF CARBON NANOTUBE FILM MORPHOLOGY ELECTROCATALYSIS AND ENERGY APPLICATIONS By RAJIB KUMAR DAS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011
2 2011 Rajib Kumar Das
3 To Maa and others
4 ACKNOWLEDGMENTS I am grateful to Professor Andrew Rinzler for his invaluable guidance, training, support and encour agem ent. I am very thankful to Professor John Reynolds and his group members for the ir wonderful collaboration I would like to thanks to Professor Arthur Hebard and his lab members for letting me use their thermal evaporation system and other instruments. I am also thankful to Marc, Bill, Mike and Ed for their help in the machine sho p. My thanks also go to the members of our group, who have provided me a joyful working environment and great help. My special thanks go to Evan for his help regarding this dis sertation. Finally, I would like to express my great gratitude to my mother for her unconditional love and support. I would also like to thank Shudhangshu Maharaja Swami Shibamayanandaji, Mr. B.K. Ghosh and Mrs. Ghosh for their great support. I would also like to thank Bapida, Mithuda, Late Gopalda a nd many others for their great help. My thanks also go to my friends for their company and friendship.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 1.1 Structure of Single Wall Carbon Nanotube ................................ ....................... 15 1.2 Energy Dispersion, Density of States and Energy Gaps of Single Wall Carbon Nanotubes ................................ ................................ ......................... 16 2 MORPHOLOGICAL MODIFICATION OF SINGLE WALL CARB ON NANOTUBE THIN FILMS FOR LARGE SURFACE AREA CONTACT APPLICATIONS ............ 23 2.1 Porosity of Conventionally Fabricated SWNT Films ................................ .......... 23 2. 2 Method of Fabrication of High Surface Area Porous SWNT Film ..................... 27 2.3 Optical Clarity and Electrical Properties of the Porous and Standard SWNT Film ................................ ................................ ................................ ................ 32 2.4 Capacitance of Standard and Porous Films ................................ ...................... 33 2.5 Permeation Rate of Fluid Through Standard and Porous SWNT Films ............ 35 3 HIGH SURFACE AREA, POROUS SWNT FILMS FOR SUPERCAPACITORS .... 40 3.1 Experimental Method and Results ................................ ................................ .... 42 3.2 Summary ................................ ................................ ................................ .......... 46 4 SWNT FILM ELECTRODE FOR HIGH PERFORMANCE ELECTROACTIVE POLYMER BASED SUPERCAPACITOR. ................................ .............................. 47 4.1 Symmetric PProDOPs Supercapacito rs ................................ ............................ 47 4.2 Interlude On Sticky Polymers ................................ ................................ ............ 49 4.3 Return To PProDOP Based Supercapcitors ................................ ..................... 52 5 REMARKABLY HIGH OXYGEN REDUCTION ACTIVITY OF PURE SINGLE WALL CARBON NANOTUBE FILM ................................ ................................ ........ 55 5.1 Background and Significance ................................ ................................ ........... 55 5.2 Experimental Set Up and Electrode Fabrication ................................ ............... 58 5.3 Experimental Results ................................ ................................ ........................ 60
6 5.3.1 Electrocatalytic Activ ity of SWNT Film ................................ ................... 60 5.3.2 Long Term Stability Test ................................ ................................ ....... 64 5.3.3 Carbon Monoxide (CO) Test and Methanol Test ................................ ... 66 5.4 Metal Air Battery ................................ ................................ ............................... 69 5.4.1 Mg Air battery ................................ ................................ ........................ 69 5.4.3 Zn Air Battery ................................ ................................ ........................ 72 5.5 Effect of Chemical Modification to the Electrocatalytic Activity of SWNT .......... 75 5.5.1 Effect of Oxygenated Species Created in Plasma Exposur e of the SWNT Film ................................ ................................ .......................... 75 5.5.2 Effect of Defect Sites Created by Controlled Functionalization of the SWNT ................................ ................................ ................................ .. 76 5.5.3 Effect of Edge S ites of the Nanotubes ................................ ................... 78 6 HIGH RATES OF HYDROGEN OXIDATION AND GENARATION ACTIVITY OVER SINGLE WALL CARBON NANOTUBE FILM ................................ .............. 81 6.1 Background and Significance ................................ ................................ ........... 81 6.2 Experimental Set Up and Electrode Fabrication ................................ ............... 84 6.3 Experimental Results ................................ ................................ ........................ 85 6.3.1 Activation of SWNT Film ................................ ................................ ....... 85 6.3.2 H 2 Oxidation Activity of SWNT Films ................................ ..................... 86 6 .3.3 Long Term Stability ................................ ................................ ............... 90 6.3.4 Lack of CO poisioning ................................ ................................ ........... 92 6.3.5 Demonstration of Fuel Cell With a SWNT Film Anode .......................... 93 7 SUMMA RY AND FUTURE WORK ................................ ................................ ......... 97 7.1 Summary ................................ ................................ ................................ .......... 97 7.2 Future Work ................................ ................................ ................................ ...... 98 APPENDIX A IN PRINCIPLE DENSITY OF HEXA GONAL CLOSED PACK ARRAY OF (10, 10) SWNT ................................ ................................ ................................ ......... 99 B ST ANDARD SWNT FILM DENSITY ................................ ................................ ..... 101 C E STIMATE OF POROUS FILM DENSITY ................................ ............................ 102 LIST OF REFERENCES ................................ ................................ ............................. 103 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 110
7 LIST OF TABLES Table page 2 1 Permeation time and permeation rate of 5 mL methanol through different samples ................................ ................................ ................................ .............. 37 2 2 Permeation time and permeation rate o f 5 mL methanol through different samples following 2 min under h igh pressure differential ................................ .. 38 3 1 The areal capacitances and t he specific capacitances of the three different RuO 2 /SWNT films ................................ ................................ ............................... 45
8 LIST OF FIGURES Figure page 1 1 The unrolled hexagonal lattice structure of a (5 5) nanotube. ............................ 15 1 2 Energy dispersion relation of the graphene shown through out the first Brillouin zone ................................ ................................ ................................ ...... 17 1 3 Energy disper sion relation of a n armchair (10, 10) tube and a zigzag (10, 0) nanotube ................................ ................................ ................................ ............ 18 1 4 Density of states of a (12, 8) semiconducting nanotube (solid red curve) and a (10, 10) metallic nanotube, ................................ ................................ .............. 20 1 5 Transmission spectra of a p type doped SWNT film and an undoped SWNT film (baked at 600c under argon). ................................ ................................ ....... 21 2 1 An AFM image of 25 nm thick SWNT film on silicon substrate ........................... 25 2 2 Schematic diagram of co filtration of SWNT and the poly styrene nanospheres and surface topography of such film ................................ ............. 29 2 3 SEM and AFM images of the composite film of SWNT and polystyrene nanosperes ................................ ................................ ................................ ......... 29 2 4 Comparison of AFM image of standard and porous film ................................ ..... 30 2 5 SEM images of a Standard and Porous SWNT fim ................................ ............ 3 1 2 6 Comparison optical clarity of two types of SWNT Films in air and liquid ............ 32 2 7 Comparison of discharge current curves and corresponding capacitances of the four different electrodes ................................ ................................ ................ 34 2 8 Schematic of the apparatu s for the permeation rate measurement .................... 36 3 1 Comparison of specific energy and specific power domains (Ragone plot) of various devices ................................ ................................ ................................ ... 40 3 2 AFM images of electrodeposited ruthenium oxide on SWNT ............................. 43 3 3 Cyclic voltammograms of the electrodeposited (300 cycles for each electrode) RuO x on three different electrode with equal geometric area ............ 43 3 4 Charging/Discharging curves of the electrodeposited RuO 2 deposited on three different electrodes ................................ ................................ .................... 44
9 4 1 Schematic diagram of supercapacitor device configuration and Faradic processes of PProDOPs. ................................ ................................ .................... 48 4 2 Chemical structure of sticky POF and schematic of a SWNT ............................. 49 4 3 Absorption spectra of sticky POF on SWNT and in solution ............................... 51 4 4 AFM image of dilute network of SWNT (a) before and (b) after application of sticky POF ................................ ................................ ................................ .......... 52 4 5 C yclic voltammograms of PProDOPs on sticky POF/SWNT Electrode .............. 53 4 6 C omparison of stability of SWNT device and gold ................................ .............. 54 5 1 Schematic of a H 2 O 2 fuel cell and a metal air battery ................................ ........ 56 5 2 Schematic of the Electrochemical cell ................................ ................................ 59 5 3 Chrono amperometry for a 120 nm thick SWNT film on nylon membrane in pH 7 phosphate buffer ................................ ................................ ........................ 61 5 4 Steady state conversion current of 120 nm SWNT fi lm electrode in air with p H 1, pH7 and pH 13 electrolytes. ................................ ................................ .... 62 5 5 Steady state current densities as a function of potential in pH 7 (phosphate buffer) for SWNT films of increasing thickness ................................ ................... 64 5 6 Long term testing at 0.4 volt versus Ag/AgCl with 120 nm SWNT electrode in 1M sulfuric acid. ................................ ................................ ................................ .. 65 5 7 Comparison of the ca talytic activity of the 120 nm SWNT film in pH 13 electrolyte for before and after the 60 day. ................................ ......................... 65 5 8 The carbon monoxide test of SWNT and commercially available Pt cathode. .. .. 66 5 9 CO test of Pt Anode ................................ ................................ ............................ 67 5 10 Effect of methanol exposure on the ORR activity of a 120 nm SWNT film electrode. ................................ ................................ ................................ ............ 68 5 11 Performance of different thickness SWNT film as cathode in Mg air battery. ..... 69 5 12 Short circuit current density of Mg air batteries for different f ilm thickness ......... 70 5 13 Comparison of the performance of a 1.5 m SWNT film and a commercial Pt loaded (0.5 mg/cm 2 ) electrode in Mg O 2 battery. ................................ ................ 71 5 14 Comparison of the performance of 120 nm SWNT film based Zinc air battery and commercial zinc air battery (Renate ZA5, Swatch, Switzerland) ................. 72
10 5 15 Stability of the out pu t voltage and current of a Zinc air battery with a 120 nm SWNT film on Nylon membrane ................................ ................................ ......... 73 5 16 Schematic of SWNT air electrode with a patterned hydrophobic membrane to achieve high surface area three phase interface ................................ ................ 74 5 17 Comparison of the electrocatalytic activity of an O 2 plasma exposed SWNT film and an as prepared SWNT film ................................ ................................ .... 76 5 18 Catalytic activity SWNT film in pH7 in air on PTFE membrane before and after functionalization in 1M sulfuric acid ................................ ............................ 77 5 19 Comparison of the catalytic activity of an uncut SW NT film (100%) and a film containing mixture of cut and uncut SWNT. ................................ ....................... 79 6 1 Diagrammatic representation of Sustainable Energy Economy .......................... 83 6 3 acid scanned (5 mV/s) ................................ ................................ ...................... 85 6 4 Chronoamperometry of a SWNT film exhibiting high hydrogen oxidation foll owing activation by exposure to the 1 M sulfuric acid electrolyte for 288 hr. .. 86 6 5 2 oxidation activity remains even though the acid was removed. .......................... 88 6 6 Comparison of the electrocatalytic activity of the SWNT film and the commercially available Pt loaded carbon (Pt C) electrode. ................................ 89 6 7 Longterm (10 hour) H2 oxidation and evolution carried out over SWNT film electrode at +0.3 volt (vs. NHE) and 0.3 volt (vs. NHE) ................................ ..... 91 6 8 The effect of CO exp osure to H 2 oxidation activity of the commercial Pt (0.5 mg/cm 2 ) loaded electrode (black curve) and a 1.5 m SWNT film ..................... 92 6 9 Polarization curve for SWNT film hydrogen oxidation anode operatin g in a H 2 air fuel cell ................................ ................................ ................................ ..... 94 6 10 Polarization curve of a H2 air fuel cell (completely Pt free) operating with SWNT film anode and SWNT film cathode ................................ ......................... 95 A 1 Hexagonal lattice structure of a (10, 10) nanotube ................................ ............. 99 A 2 Schematic diagram of a cross section of a nanotube bundle ........................... 100
11 LIST OF ABBR EVIATIONS AF M Atomic force microscopy CO Carbon monoxide DMFC Direct methanol fuel cell FWNT Few wall carbon nanotube HER Hydrogen evolution reaction HOR Hydrogen oxidation reaction MWNT Multiwall carbon nanotube ORR Oxygen reduction reaction SEM Scanning electron microscopy SWNT: Single wall carbon nanotube
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ENHANCED UT ILITY OF CARBON N ANOTUBE FILM MORPHOLOGY ELECTROCATALYSIS AND ENERGY APPLICATIONS By Rajib Kumar Das May 2011 Chair: Andrew Rinzler Major: Physics Single wall carbon nanotube films are promising as conducting electrode s in applications ranging from energy storage to catalysis due to their inherent high surface area. Here several such applications are pursued. First it is demonstrated h owever, that the flexibility of the individual nanotubes and their affinity for each other conspire to obstruct the porosity in such films limiting the accessible surface area W e demonstrate a simple, effective means to engineer controlled porosity into the nanotube films and apply such films to RuO 2 based supercapacitor. Given the expense of ruthenium a more cost effe ctive system: p oly (3, 4 propylenedioxypyrrole) based supercapacitors have also been explored Carbon supported platinum and its alloys with other precious metals are commonly used as the oxygen redu cing catalyst o n the air electrode side of various high v alue or experimental fuel cells and metal air batteries However, large scale application of these technologies is inhibited due to the scarcity of the precious metal s. Here we show that SWNTs alone, without any added metal catalyst, fabricated as thin film s by vacuum filtration (a readily scaled process) exhibit a remarkable O 2 conversion rate comparable
13 to a commercially available Pt base cathode with a mass basis activity of the SWNT s is nearly 5 times that of the commercially available Pt C cathode. Pt is also the catalyst of choice for the hydrogen oxidation reaction and the hydrogen evolution reaction requiring zero overpotential for each of these reactions (i.e. the reaction begins at the thermodynamic potential). Here, w e have report that upon activa tion of the SWNTs by acid exposure they too exhibit potentially useful hydrogen oxidation and hydrogen evolution reactions at zero overpotential These catalyti c activities are characterized.
14 CHAPTER 1 INTRODUCTION Carbon appears in nature in various form s (allotropes) diamond, graphite, carbon fibers, f ullerenes and carbon nanotubes. Single wall carbon nanotubes (SWNT) can be thought of as graphene sheet s rolled into a seamless, cylindrical shape with a typical diameter of about 1 2 nm and length s of a fe w hundred nanometers to m illimeters The nanotubes can be considered as quasi one dimensional object as they have very large aspect ratio (length to diameter ratio can be as large as 10 4 10 5 ). Multiwall carbon nanotubes (MWNT) are simply multiple cylinder s of single wall nanotubes arranged co axially. Since the pioneering identification of Multiwall carbon nanotube by Iijima in 1991 an enormous amount of theoretical and experimental studies have been performed on carbon nanotubes 1 2 The single wall carbon nanotubes showed very interesting electronic prope rties. They can be metallic or semiconducting depending upon the orientation of the hexagonal lattice structure relative to the axis of the nanotubes 2 The nanotubes are mechanically very strong. They have ex tremely high Young Modulus 3 (1 TPa) which is about 5 times higher than steel (200 GPa) and a very high tensile strength 4 5 (~ 45 GPa). The hexagonal sp 2 hybridized carbon lattice of the sidewall of the nanotube makes them highly stable materials in a variety of chemical en vironment. Due to their interesting geometrical structure, electronic and chemical properties carbon nanotubes remain a very active field of research with frequent surprises. In this C hapter I will briefly discuss the structure and electroni c properties of the nanotubes. The discussion follow s the excellent treatment given by Saito, Dresselhaus & Dresselh aus 2
15 1.1 Structure of Single Wall Carbon N anotube The structure of the single wall carbon nanotube is specified by a chiral vector C h = n a 1 + m a 2 (n, m) which corresponds to a vector perpendicular to the nanotube axis 2 the a 1 and a 2 are the unit vectors of the hexagonal graphene lattice as shown in Figure 1 1. This vector ( OA as shown in Figure1 1) also defines the circumference of the tube. An Armchair nanotube corresponds to the n=m with chiral vector C h = (n, n) and a zigzag nanotube corresponds to the case of m=0 with C h = (n, 0). The real space unit vectors a 1 and a 2 are not orthogonal rather they are aligned each other at angle 60. These correspond to a 1 a 1 = a 2 a 1 a 2 = (a 2 /2) where lattice constant a = (1.44 diameter of a (n, m) tube is given by h h h C h = a (n 2 + m 2 + nm) C h and the unit vector a 1 which yield s Cos = ( C h .a 1 h 1 = [ n+ ( m /2)] /[ (n 2 + m 2 + nm) ] Figure 1 1 The unrolled hexagonal lattice structure of a (5, 5) nanotube.
16 Translational vector T = t 1 a 1 + t 2 a 2 is defined to be parallel to the nanotube axis and perpendicular to the chiral vector C h .The tra nslation al vector T (OB in Figure 1 1) and the chiral vector C h define the unit cell of the 1D carbon nanotube. As C h and T are orthogonal C h T = 0 (n a 1 + m a 2 ). (t 1 a 1 + t 2 a 2 ) = 0 (2n+m) t 1 + (2m+n) t 2 =0 t 1 = ( 2m+n )/d r and t 2 = (2n+m )/d r where d r is the gr eatest co mmon divisor of (2n+m ) and (2m+ n) The unit cell defined by OABB' for the (5, 5) nanotube is shown in Figure 1 1 where the translational vector defined by T 1). The ratio of the area of the C h T to the area of the unit cell of the hexagonal a 1 a 2 gives the number of hexagon (N) per unit cell as: C h T a 1 a 2 2 + m 2 + nm) / d r As each hex agon of the lattice is associated with two (whole) atoms there are 2N number o f atoms per unit cell of the carbon nanotube. For an armchair tube (n, n) there is N= 2n number of hexagons per unit cell of the nanotube. Thus the translational unit cell of a (10 10 ) nanotube contains 4 0 atoms. 1.2 Energy D ispersion, D ensit y of S tates a nd Energy Gaps of Single Wall Carbon N anotubes T he energy dispersion relation of the hexagonal graphen e lattice 6 is s hown in Figure 1 2. The energy dispersion relation can be obtained using the tight binding method and considering only the nearest neighbor interaction s of the carbon atoms as described by Saito et al 2
17 Figure 1 2 Energy dispersion relation of the gr aphene shown through out the first Brillouin zone [Reprinted with permission from Ph. Avouris, Z. Chen and V. Perebeinos, Nature Nanotechnology 2 605 (2007) Copyright 2007, Nature Publishing Group ] The e ne rgy dispersion relation of the s ingle wall nan otube can be obtain from the energy dispersion of the graphene by imposing periodic boundary condition along the circumference of the nanotube, C h K Where K is the reciprocal lattice vector and N is the number of hexagons in the uni t cell of the carbon nanotube. The reciprocal lattice vector k along the circumference of the nanotube (perpendicular to the tube axis) can be written as k = (2 C h [ 2 q ] / [ a (n 2 + m 2 + nm) ] The transformation of the wave vectors k x and k y wave vectors and application of the periodic boundary condition on graphene energy dispersion relation
18 leads to quantization of wave vector along circumferential direction of tube (k ) whereas b ecause of the long length of the nanotube (assumed infinite) take s continuous values. The energy dispersion relation of the single wall carbon nanotube accordingly becomes the number N of simple parallel line segments that cut the graphene energy dispersio n along the direction of periodic boundary condition of the specified tube This leads to 2N number of energy subbands (N valence bands and N conduction bands) in the first brillouin zone as a function of the 3 shows the zone folded energy di spersion relation of armchair (10, 10 ) nanotube and zigzag (10, 0) nanotube 2 Figure 1 3 Energy dis persion relation of (a) an armchair (10, 10 ) tube and (b) a zigzag (10, 0) nanotube [ Reprint ed with permission from Z.H. Chen, Thesis, University of Florida, 2003 ]. It is clear that the upper most valence band and lower most conduction band cross each other for the (10, 10 ) nanotube at the Fermi level within the 1 st Brill ouin z one ( a ) (b)
19 making it metallic. On the other hand for the (10, 0) there is clearly a gap between the valence band and conduction band making it semiconducting. In general for a n (n, m) nanotub e if 2n+m is multiple of 3 then the tube become metallic otherwise it is semiconducting. Thus (n, n) armchair nanotubes are always metallic. The d ensity of states of a SWNT is derived from the energy dispersion relation of the nanotube. The energy dispersi on relation of the SWNT can be divided into small energy intervals. S umming the states in each energy interval specifies the density of states of the nano tube. Figure 1 4 shows the density of states of a zigzag (12, 8 ) and a (10 1 0) nanotube s 7 The singularities in the density of states (called von Hov e singularities) of the SWNT are due to the extrema in the 1 D energy dispersion relations and are a characteristic feature of 1 D systems. The distance between the first valance and conduction band van Hove singularities (the band gap) of the semiconduct ing SWNTs is inversely proportional to the diameter of the tube 8 9 Scanning tunneling spectroscopy provides a powerful tool for probing the electronic structure of the nanotubes. Such measurements first confirmed that the band gap is inversely proportional to the diameter of the tubes 8,9 High yield methods for nanotube synthesis typically produce 2/3 semiconduct ing and 1/3 metallic nanotubes so that both types are present in films of SWNTs. The presence of these species can be detected by absorption spectroscopy where dips in the transmittance observed are due to photoinduced electronic transitions between valenc e and conduction band van Hove singularities, symmetry allowed for symmetric singularities about the intrinsic Fermi level. The individual electronic transitions are sharp (few meV) but the dips in the
20 transmittance spectra are typically broad because they are a composite of the absorption from nanotubes of varying diameter (~1.1 1.6 nm) in the sample. Figure 1 4 Density of states of a (12, 8) semiconducting nanotube (solid red curve) and a (10, 10) metallic nanotube dotted green curve, of approximately the same diameter. For nanotubes of the same diameter the spacing between the first pair of van hove singularities about the Fermi level for the metallic nanotube is ~3 times that for the semiconducting nanotube. The blue regi on corresponds to filled valence band states in these charge transfer doped nanotubes as discussed below. [Adapted from Z.Wu et.al Science 305 1273 1276 Copyright (2004) AAAS] The individual nanotube absorption lines are also broadened by perturbation s due to their bundling 2 As occurs for graphite (where the process is called interc a lation 10 ) the nanotubes are subject to charge transfer doping in which a donor or an acceptor atom or molecule will spontaneously exchange charge with a nanotube lowering the free energy of the system 11 Such doping is reversible simply by heating the nanotubes.
21 Figure 1 5 shows the transmission spectra of p type doped (black curve) 65 nm SWNT film (nitri c acid reflux doped) and an undoped (red curve) SWNT film of same thickness (baked at 600C for 30 min under argon). Figure 1 5 Transmission s pectra of a p type doped SWNT film and an undoped SWNT film (baked at 600c under ar gon). The S1 absorption peak corresponds to electronic transition s from the first Van Hove singularity of the valence band to the first Van Hove singularity of the conduction band of the semiconducting nanotube. Similarly S2 and M1 peaks represent the abso rption peak corresponding to the second sets of Van Hove singularities of the semiconducting tube and the first set of Von Hove singularities of the metallic tube respectively. Figure 1 4 shows such electronic transitions for (12, 8) semiconducting tube an d (10, 10) metallic tube. As tubes are acceptor doped the Fermi level shifts towards th rough the valence band singularities depleting the ground state of the
22 electronic transition Heavy acceptor doping of the tube can thus deplete the first and second Van Hove singularit ies significantly diminishing the S1 and S2 absorption peaks in the absorption spectra. As shown in Figure 1 5 the S1 and S2 transmittance dips are greatly diminished as the nanotube s were acceptor doped by the nitric acid reflux used in their purification. These fundamental properties of the nanotube electronic structure and optical characteristics will be referred to as needed in the pages to come.
23 CHAPTER 2 MORPHOLOGICAL MODIFICATION OF SING LE WALL CARBON NANOT UBE THIN FILMS FOR LARGE SURFACE AREA CONTACT APPLICATIONS Nanoscale materials afford new opportunities for their bulk assembly into forms that enhance their function. I t is common for electronic devices to have a two dimensional (planar) contact between the electrode and the mat erial to be contacted. However, a three dimensional distributed contact can be highly beneficial in a number of applications. For example, in the c ase of hydrogen production in an electrol yzer or hydrogen oxidation in fuel cells 12 a high surface area porous electrode is desirable to present larger areas over which the electrode reactions can occur and to avoid kinet ically limiting reaction rates by diffusion through narrow channels. As will be shown in Chapter 3 this is true for supercapacitors, where the capacitance can be greatly enhanced by a high surface area porous electrode 13 Such porous electrodes could also be advantageous in solar cell or photodetector applications 14 where the volume over which light absorption occur s is important to their operation. C onductive thin films of single wall carbon nanotubes (SWNTs) have recently been recognized as promising porous electrode s for diverse applications 15 However, as sho wn in this C hapter, SWNT films fabricated by known methods such as vacuum filtration 7 spray coating 16 or Langmuir Blodget deposition 17 possess relatively small pore volumes generated by the overlapping nanotube bundles W e first show this to be true for vacuum filtration fabricated films and then demonstrate a novel method developed to build in the desired porosity during the film assembly 2.1 Porosity of Conventionally F abricated SWNT Films Because t he previously reported vacuum filtration method for SWNT film fabrication is relevant to the present work I will briefly review the process. To fabricat e
24 SWNT film s by the vacuum filtration method we start with nanotubes in an aqueous surfactant suspension, of known concentration, and vacuum filter a known volume to the surface of a filtration membrane having pores too small for the nanotubes to permeate. The nanotubes form a homogeneous film (as a filtration cake) on the surfa ce of the membrane, which also permits the residual surfactant to be removed from the nanotube film by subsequent washing with pure water. The film is then dried to consolidate the nanotubes into a dense stable van der Waals bonded network of nanotubes. To transfer the film from the membrane to the substrate of choice the film is wetted again (with water or ethanol) and nanotube side of the membrane is pressed to the surface of the substrate while the wetting agent dries. Surface tension of the drying liqui d generates intimate contact between the nanotubes and the substrate providing sufficient adhesion that the filtration membrane can now be dissolved away with a solvent (acetone) in which the membrane polymer is soluble, leaving the film on the substrate. The SWNTs in the starting surfactant suspension are rarely individual nanotubes but rather bundles of parallel nanotubes held together by van der Waals forces. The van der Waals forces are weak on a per atom basis compared to chemical bonds but because t he flat nanotube surfaces permit many atoms to come into close registry along the lengths of the nanotubes the net binding energy becomes large. For the nitric acid purified nanotubes created via pulsed laser ablation 18 as used in this work the nanotube bundles have diameters ranging from 3 20 nm. The SWNT film s fabricated by th e filtration process have an open porosity between the nanotubes which is defined by the tortuous path between the self organized bundles of nanotubes. An AFM image of a
25 typical 25 n m thick film is shown in Figure 2 1 Inspect ion of such image suggests that pore dimensions between nanotube bundles are o f the order of tens to hundreds of na nometers; however, the material is not isotropic making the apparent pore dimensions from such images deceptive. Three straight, crossing nanotube bundles lying in a plane define a triangular area between them that can measure hundreds of nanometers across (bounded by bundles lying above and below) is effectively the diameter of the three crossed bundles (typically 5 10 nm). The flexibility of the nanotubes and surface tension forces of liquids from which the films are dried further conspire to collapse pores bringing the nanotubes into intimate van der Waals contact, reducing the open porosity, bundle diameter. Figure 2 1 An AFM image of 25 nm thick SWNT fi lm on silicon substrate To get a more quantitative idea of the porosity in the film produced by vacuum filtration method one can compare the theoretical density of a hexagonal closed packed
26 arrangement of (10, 10) nanotubes (with diameter~1.4nm) to the exp erimentally measured density of the films. It is found by calculation [ Appendix A] that a hexagonal closed packed array of (10, 10) SWNT have a density about 1.33 gm/cm 3 Experimentally [ Appendix B] it is found the nanotube films produced by vacuum filtrat ion method have a density about 0.71 gm/cm 3 So the SWNT films fabricated by filtration method have already achieved approximately 55% of the ir maximum density. Since the remaining 45% pore volumes are uniformly distributed through out the volume of the fi lm the average size of the pore volumes are of the order of size of the bundles which is ~10 nm in diameter. Means exist for measuring porosity in nanoporous media; however to be accurate, these require substantial quantities of the material, while the hi gh cost of the SWNTs and the labor intensive purification used to obtain the cleanest materials for film fabrication make dedicating such quantities a disagreeable proposition (porosimetry measurements are in principle nondestructive but redispersing a dri ed buckypaper is generally damaging to the nanotubes). Reliable(nonimaging derived) porosity data exist for multiwalled nanotube films 19 but the much greater intrinsic stiffness of multiwalled nanotubes makes this of little relevance for SWNTs. One study o f porosity in SWNT buckypapers 20 21 found pore dimension s in two broad ranges: 1 10 nm consistent with the micro mesoporosity determined by BET analysis of adsorption isotherms in bulk, solution processed, SWNT mat erial 22 23 and a larger range of 10 100 nm. The latter range is probably not intrinsic to pure SWNT films but likely rather due to contamination by micrometer scale particles clearly visible in the imaging shown there (perhaps graphitic particles, a common contaminant in laser ablation synthesized
27 n anotubes). Given the tortuous path nature of the connectivity, pore dimensions in the range of 1 10 nm are small for many applications, li miting the perfusion rate of liquids and gases through the films and decreasing the accessible surface area. 2.2 Method of Fabrication of High Surface Are a P orous SWNT Film To increase the utility of the SWNT films we have devised a simple, general mean s to enhance their porosity over a broad, tunable range of length scales. The strategy is to fabricate the films from a fluid suspension of intimately mixed nanotubes and sacrificial nanoparticles thus forming a composite nanotube/nanoparticle film. The sa crificial nanoparticles are subsequently eliminated by a process that does not disrupt the film framework thus retaining pores at the previous sites of the nanoparticles. The studies reported here used monodisperse sacrificial particles to increase the geo metric volume of the films (for the same quantity of nanotubes) by a factor of ~17, with a corresponding reduction in the film density [Appendix C] The use of polydisperse sacrificial particles of controlled size distribution could further provide control over the mesoporosity in the nanotube films. In our first implementation we used 200 nm diameter polystyrene nanospheres (Duke Scientific, Ma) as the sacrificial particles because they form stable aqueous surfactant suspensions and can be dissolved in or ganic solvents. The filtration method consists of vacuum filtering the nanotubes in a surfactant suspension to the surface of a filtration membrane possessing pores too small for the nanotubes to permeate; the surfactant is washed away and the film dried, which consolidates the nanotubes into intimate van der Waals contact forming the robust film. If needed, thick films can be peeled off of PTFE based membranes while thinner films can be transferred to alternative substrates by dissolving away a membrane se lected for its solubility In our
28 initial (unsuccessful) attempts to form the nanotube/nanosphere composite films the surfactant suspensions of the nanotubes and nanospheres were simply mixed by stirring and brief ultrasonication followed by the normal fil tration step. Atomic force microscopy (AFM) of the dried films on the membrane surface showed that rather than form the desired intimately mixed composite, the components segregated into distinct layers with the nanospheres arrayed on top of the nanotube l ayer. Such undesirable segregation was overcome by exploiting the hydrophobic character of polystyrene and nanotubes and a characteristic of the surfactant. The non ionic surfactant Triton X 100 TM possesses a limited ionic strength tolerance before it bec omes ineffective as a surfactant. After mixing the nanotube and nanosphere initiating flocculation. Since the rate of flocculation is pH dependent this permits a measure of control over the degree of flocculation at the time that the film is assembled. Intimately mixed nanotube/nanosphere composite films were readily formed for 1% (v/v) Triton concentrations acidified to a pH of 1.5 with the film formation (filtration) beg un 15 90 min after addition of the acid. The carbon nan otubes and nanospheres form a homogeneous composite film on the surface of the me mbrane. Figure 2 2a shows a schematic diagram of the co filtration of SWNT with nanospheres Figure 2 2b shows the surfa ce topography of a composite film of SWNT with polys tyrene nanospheres. In the case shown the quantity of nanospheres used, if stacked in hexagonal close pack arrangement (in the absence of the nanotubes), would have formed a deposit calculated to be 3 lay ers thick, having thickness ~530 nm. The quantity of nanotubes used, in the absence of the
29 nanospheres, would have formed a layer ~80 nm thick (determined by fabricating a pure nanotube film from the same concentration and volume of the SWNT suspension). Figure 2 2 (a) Schematic diagram of co filtration of SWNT and the polystyrene nanospheres (b) the surface topography of a composite film of SWN T Figure 2 3a and Figure 2 3b show scanning electro n microscopy (SEM) and AFM images, respectively, of such a composite film Figure 2 3. (a) A SEM i mage of the c omposite film of SWNT with the nanospheres (b) An AFM image of the Composite film of the SWNT with nanospheres a b a b
30 A section of this film was transferred to a silicon substrate by dissolution of the mixed cellulose ester filtration membrane in acetone, which simultaneously dissolves the PS nanospheres A comparison of the morphology of AFM im ages for a standard (flat) nanotube film (Figure 2 4a) and a porous SWNT film, (Figure 2 4b) both using the same quantity of nanotubes is presented below. The vertical color scale is the same ( 400 nm ) in both images, making the flat film appear washed out b ecause it occupies such a small fraction of this scale. The porous film in contrast occupies the full color scale because of the large scale pores introduced by the dissolved nanospheres. Figure 2 4 (a) A n AFM image of standard (flat) SWNT film (b) an AFM image of porous SWNT film For the purpose of imaging a cross section of the nanotube film a drop of photoresist (saturating through to the Si substrates) protected a portion of the film from being consumed during a plas ma ash. Shipley 1813J photor esist soft baked for 1 hr at b a
31 80 C was used. Etching was done in an Anatech SCE600 plasma asher at 300 W power, 870 mTorr O2 pressure flowing at 450 sccm for 360 seconds. The etched film surface at the boundary between the protec ted and unprotected regions is not perpendicular to the plane of the film because the diluted photoresist (3:1) wicks in at an angle such that the footprint of the protected region is larger at the substrate than at the top surface of the film. Following t he plasma ash the photoresist was dissolved in successive acetone baths followed by methanol and the film dried. The SEM image was recorded along the etched edge, where the density of observable pores was a factor of ~2 times greater than those visible fro m the top surface. Figure 2 5b shows the SEM image of the resultin g porous SWNT film For comparison, Figure 2 5a shows the image of a standard nanotube film (no nanospheres). Figure 2 5 (a) SEM image of a standard (flat ) SWNT film (b) SEM image of plasma etched edge of a porous SWNT film Note that the film prepared for the imaging is dried after the removal of residual polymers in subsequent baths. Th is collapses some of the pore volumes as nanotube a b
32 bundles are brought i nto intimate contact by the surface tensions of the evaporating liquid during the drying. T he accessible pore volumes are thus anticipated to be even greater prior to such final drying. Whe n ever a second material must infiltrate such a porous film, if pos sible, it would be beneficial to effect such infiltration without a prior drying of the porous film. 2.3 Optical C larity and Electrica l Properties of the Porous and S tandard SWNT F ilm Figure 2 6a shows a photograph of a standard (left) and porous (right) SWNT films, made using the same quantity of nanotubes on transparency sheet, comparing their visual appearance for transmitted light. The porous film exhibits some haze due to random aggregates of the ~200 nm empty (air filled) pores. Figure 2 6 (a) Comparison optical clarity of two types of SWNT Films in air (b) Indistinguishable optical clarity of two types of SWNT Films in liquid When such porous films are bathed in methanol, having a refractive index more closely matched to that of the nanotubes, such haze vanishes. This has relevance in optical applications where the pores are filled with secondary media (e.g. electroluminescent or electrochromic polymers) likely to provide a closer refractive index a b Standard (Flat) Porous Standard (Flat) Porous
33 match to the nanotubes than air. Sheet resistances of the standard and porous films, measured by the 4 probe van der Pauw technique, were 75 respec tively, hence while the increased porosity necessarily implies a decreased tube tube overlap this results in litt le degradation in the film conductance. 2.4 Capacitance of Standard and Porous F ilms To provide a more quantitative measure of the changes induced in the films by the additional porosity we measured their electrolytic double layer capacitance. Double laye r capacitance is generally not a reliable indicator of surface area however in simple, single material systems lacking pseudocapacitive functional groups and for which the new porosity is not in the form of micro pores (that are too small for the electroly te ions to enter), double layer capacitance has been found to correlate well with BE T deduced surface areas 24 25 Double layer capacitance measurements were performed in a specially constructed, Teflon body, electrochemical cell possessing a circular opening through the cell sidewall connected to the electrolyte reservoir (Figure 2 7 a, inset). The opening in the cell sidewall was surrounded by a captured O ring. The film to be tested was deposited onto a glass plate as a rectangular area ~ 15 x 19 mm 2 overlapping a Pd contact pad by 3 4 mm along the 19 mm dimension (for making electrical contact to the film). Th e film side of the glass plate was sealed against the O ring with the film working electrode (film) area exposed to the electrolyte to be consistently defined (from s ample to sample) by the O ring diameter. Because the surface tension of drying liquids constitutes a dominant force at the nanoscale likely to cause partial collapse of the newly introduced porosity the films were not allowed to dry between nanosphere diss olution and introduction of the electrolyt e in the electrochemical cell
34 Figure 2 8 a shows the current discharge curves of 4 distinct films used as the working electrode in a step function change of the electrode potential from 0.5 to 0.0 volts versus a n Ag/AgCl reference. The 4 films were: a 100 nm thick sputtered Pd metal electrode (to provide comparison with a completely non porous material); an 80 nm thick standard SWNT film; a porous SWNT film as described above; and a porous SWNT film possessing a 2.5 times greater nanosphere concentration. Note that the quantity of nanotubes in each SWNT film was identical, the only difference being the film porosity. The discharge current was followed to 2 seconds but only the fir st half second is shown in Figure 2 7 a to highlight differences at early times Integrating these currents yields the accumulated charge discharged by the electrode. These are plotted in Figure 2 7 b normalized by the voltage and film area exposed to the electrolyte. Figure 2 7 (a) C omparison of discharge current curves of the four different electrodes with schematic of electrochemical cell in the inset. (b)The integrated curves of (a) normalized with applied voltage and area of electrode. The values at 2 seconds, when all the curves have leveled off, gives the total charge stored per unit area, which upon dividing by the voltage change gives the areal a b
35 capacitance of the electrode. These values label each curve in the plot. The measured density of the standard SWNT films is 0. 71 g/cm 3 ( Appendix B ) allowing calculation of the areal mass of the 80 nm thick film as 5.68 g/cm 2 Since all the SWNT films contained the same quantity of nanotubes their areal masses are the same. This permits calculation of the specific capacitance of the films (also given for each SWNT curve in the plot). The specific capacitances of the 1x p orous and 2.5x porous films exceed that of the standard film by a factor of 1.4 and 2.1, respectively. These are substantial enhancements in capacitance correlating nicely with the nanosphere template concentrations however the densities of the 1x and 2.5 x porous films are estimated to be approximately 1/7 and 1/17 of the standard film density, resp ectively (Appendix C ). The much smaller changes in the capacitances compared to the changes in the densities begs explanation. This is found in the dimensions o f the Helmholtz double layer, which in strong electrolytes is only of the order of 1 nanometer 26 At such length scale the capacitance measurements ar e simply not sensitive to the macro pores introduced into the films. This is supported by comparison of the standard film areal capacitance with that of the non porous Pd electrode. Although pore sizes in the standard film are only in the range of 1 10 nm, its areal capacitance already exceeds that of the completely non porous Pd electrode by a factor of 6.6. Thus, while the capacitance changes provide a measure of the newly exposed surface area in the porous films, they do not quantify the larger scale por osity introduced. 2.5 Permeation Rate of Fluid Through Standard and Porous SWNT Films To establish at least a comparative measure of the larger scale porosity we measured the permeation (flow) rate of a liquid passing through the films under identical
36 cond itions. For this purpose the SWNT films were transferred to a PTFE filtration membrane possessing much larger, 10 m pores (Millipore, LC filter disks). In this case the films were dried following their transfer to the PTFE membranes. Flow measurements were made using a two par t glass microanalysis filter holder (Millipore, XX1002500). In a typical experiment the PTFE membrane, SWNT film side up, was placed across the glass frit base with the funnel placed over the film clamping the fu nn el and base together (Figure 2 8 clamp not shown) Fig ure 2 8 Schematic of the apparatus for the permeation rate measurement Methanol was used because it wets both PTFE and SWNTs. The time taken for the liquid level to drop from 75 to 50 mm, corresponding to 5 mL of methanol permeated, without pumping on the outlet, was measured (average pressure differential across the film P = 613 Pa). Each permeation time measurement was repeated 4 times with
37 variations for any g iven sample being within 4%. Table 2 1 gives the averaged permeation time for each film and the corresponding flow rates As expected the permeation rates through the SWNT films increases with their increasing porosity. The rate through the 1x film was 1.42 times that of the standard film, while that through the 2.5x film was nearly 10 times that of the standard film. The much smaller enhancement in the 1x film agai n seems incommensurate relative to the estimated change in its density, however the permeation rate of a fluid Table 2 1. Permeation time and permeation rate of 5 mL methanol through different samples Sample Permeation time (sec ) /s) PTFE membrane (only) 90 55.6 standard SWNT film 2470 2.02 1x porous SWNT film 1740 2.87 2.5x porous SWNT film 255 19.6 depends not only on the porosity but also on the connectivity between pores. Consideration of Figure 2 7 a a nd these results suggest that the connectivity between the large pores in the 1x film is limited, while the connectivity is greatly increased in the 2.5x films. The 10 times enhancement of the flow rate in the 2.5x porous film is more in line with its esti mated 1/17 density. These permeation rate experiments can also provide information relating to the frailty of the pores. After the measurement set for each porous film sample, in which the pressure differential across the membrane was set merely by gravity and the height of the methanol column above the film, a vacuum of 12 cm Hg was applied to the base funnel for 2 minutes after which vacuum was released and the permeation time again measured as above. These results are listed in Table 2 2
38 Table 2 2 Per meation time and permeation rate of 5 mL methanol through different samples following 2 min under high pressure differential and subsequent releasing the pressure. Sample Permeation time (sec ) Permea /s) standard SWNT film 2470 2.02 (unchanged) 1x porous SWNT film 2340 2.14 2.5x porous SWNT film 3720 1.34 The flow rate through the standard film was unchanged while the brief (2 min.) pressure differential of P~16000 Pa across the poro us films was sufficient to greatly decrease the flow rate through the porous films, presumably by collapsing the pores. Somewhat surprising is that the flow rate through the 2.5x film has become substantially lower even than that of the standard film. The standard films, when they are made experience a much greater pressure differential (typically ~80,000 Pa), hence it is not possible for the comparatively weak vacuum applied here to further compact the 2.5x film explaining the decreased flow rate (recall t hat all the films have the same quantity of nanotubes). The effect can be explained however by a reorganization of the nanotubes in the 2.5x film during the pumping to assemble more material (compacted to a density approaching that of the standard films) o ver the pores of the PTFE support membrane. This additional SWNT material is coming from the non porous regions of the support membrane. Given the more tenuous tube tube contact that must be present in the 2.5x porous film framework compared to the standar d films such reorganization would seem plausible. In this C hapter, the limit ed accessible surface area of SWNT film s was discussed Flexibility of the individual tubes and their affinity for each other gives rise to a small porosity in the SWNT film s ; impo sing limitations on the perfusion rate of fluids as well as
39 on the accessibility of the electrochemical surface area. A novel method was demonstrated to engineer porosity into carbon nanotube films without sacrificing the electronic properties of the films The porosity introduced into the SWNT films was characterized by SEM imaging, AFM imaging and measurements of the perfusion of fluids. By electrochemical capacitance measurement it was shown that the accessible surface area of the nanotube films is enhan ced by introducing such porosity. Although both standard and porous films have the same mass of nanotube s the electrolytic double layer capacitance of porous SWNT films was increased by 110% compared to that of a standard (flat) nanotube film. These highl y accessible surface area SWNT films electrodes should be useful in a number of applications and are used in some of the studies discussed in the next Chapter
40 CHA PTER 3 HIGH SURFACE AREA, P OROUS SWNT FILMS FOR SUPERCAPACITORS In the area of mobile ene rg y storage and delivery electrolytic double layer capacitors (EDLC) and s upercapacitors are very promising for rapid storage and release of energy. 13 In Electrolytic double layer capa citors, the capacitance is based on the non Faradaic Helmholtz layer (double layer) formation due to the electrostatic attraction of opposite charges at the electrolyte /electrode interfaces. 13 A supercapacitor also uses an electrolyte however in the case of a supercapacitor the electrode s are coated with a material that undergoes a reversible electron transfer with the electrode s ( a Faradaic process ) while the electrolyte permeates the volume of the material compensating the charge exchanged with the electrodes The charge storage capability via this process is called pseudo capacitance 27 In case of supercapacitors this pseudo capacitive reaction process is highly reversible which distinguishes them from most batteries. Figure 3 1 shows a simplified R a g o ne plot for various energy storage devices which compares the ir energy and power storage domains 13 Figure 3 1 Comparison of specific energ y and specific power domains (Rago ne plot) of various devices [Reprinted with permission from Martin Winter et al Chem Rev 104 4245 4269 (2004) Copyright 2004 ACS ]
41 Batteries have higher energy storage capacities but can not deliver that energy quickly so they have relatively low power density Supercapacitors are useful because they have high power density with reasonable energy dens ity Transition metal oxides wi th reversible oxidation states have been found to act as good psuedocapacitive materials. Ruthenium oxide in particular has been shown to have a high specific capacitance (capacity/unit mass), with good reversibility over a wide potential window. The high specific capacitance in ruthenium oxide comes from following Faradiac reaction 28 29 RuO a (OH) b + H + + a (OH) Hence the reaction involves a reversible oxidation /reduction of the hydrated ruthenium complex while incorporating a proton from the electrolyte. Competing processes are involved in optimizing the capacitance of ruthenium oxide based supercapacitors. Granularity of the material is desired to admit the charge compensating protons from the electrolyte but such grains can also become electrically isolated (upon being surrounded by electrolyte) from the rest of the material thus reducing the overall capacitance (especially with repeated charging discharging cycles that involve s material swelling and contraction). The deposition of a thick continuous layer of ruthenium oxide (e.g. by electrodeposition) on a flat substrate suffers from the l imitation of getting protons into the matrix while the deposition of individual grains (e.g. via a sol gel process) suffers from grain isolation. These limitations can be overcome if the ruthenium oxide is deposited in thin layers on an electrically condu cting porous network like our carbon nanotube films. The open structure of the resulting network admits protons while the conducting nanotube
42 backbone provides electrical conductivity between grains that might otherwise become isolated. 3.1 Experimental Method and Results The RuO 2 was deposited via cyclic potential scan electrodeposition from an aqueous RuCl 3 bath 30 31 All measurements and cyclic depositions were performed in three terminal mode using a Perkin Elmer 283 Potentiostat/Galvanostat using a platinum counter electrode. versus Ag/AgCl at a potential scan rate of 50 mv/s for 300 cycles. Following the deposition cycles the RuCl 3 solution was dumped from the cell and replaced by DI water. The water was replaced 6 times wi th a minimum of 30 minutes between exchanges to ensure virtually complete removal of the RuCl 3 solution. Finally the cell was filled with1 M H 2 SO 4 Cyclic voltammetry was performed in the 1 M H 2 SO 4 electrolyte with potential scans from 0.1V to 1.1V at 50 mV/s, versus an Ag/AgCl reference electrode. Figure 3 2 shows the AFM image of a standard SWNT film following the RuO 2 deposition for 300 scan cycles showing that the material coats the nanotube surfaces effectively. Figure 3 3 shows the CV scans for the three distinct, RuO 2 coated films The trend is clearly for increasing current (and hence increasing areal capacitance) with increasing film porosity. Simultaneously recorded current time plots (Figure 3 4) allow extraction of the accumulated charge and th erefore the areal capacitance associated with each supporting electrode type. Following the measurements the electrolyte for each SWNT/RuO 2 electrode was exchanged for water, the film dried, removed from the EC cell and heated to 200 C to drive off the wat er. On removing the sample from the
43 cell, the film region in contact with the O ring generally st uck to it, hence the active film area exposed to the electrolyte was clearly demarcated. Figure 3 2 AFM images of electrodepos ited ruthenium oxide on SWNT Fig ure 3 3 Cyclic voltammograms of the electrodepo sited (300 cycles for each electrode ) RuO x on three different electrode with equal geometric area
44 This permit ted the active region to be scraped from the glass into the p an of a microbalance for weighing. These masses, in turn, permit determination of the specific capacitance (capacitance per gram) of the films. Fig ure 3 4 Charging/Discharging curves of the electrodeposited RuO 2 deposited on three different electrodes Somewhat surprisingly the mass of RuO 2 deposited was found to change little with the underlying SWNT electrode being nearly independent of the film porosity so that the increasing areal capacitance reflects an increase in the specific capacitance. The res ulting masses, areal and specific capacitances are listed in Tabl e 3 1. The second to last column gives the specific capacitance for each RuO 2 /SWNT electrode. Since the mass of SWNTs in each electrode is known (4.92 g in the 0.866 cm 2 area), and the capac itance of each S WNT electrode was measured (Figure 2 8 b) these can be subtracted out to yie ld the specific capacitance of the RuO 2 alone. These are given for each electrode in the last column. The value of 1715 10 F/g obtained with
45 the 2.5x porous film app roaches the theoretical maximum capacitance of RuO 2 estimated as 2000 F/g 32 Notably, this is before any thermal treatments that typically increase the capacitance of RuO 2 electrodes by optimizing the trade off between their hydration (allowing for proton transport in to the structure) and th eir electronic conductivity 33 We can speculate that the enhanced specific capacitance with increasing film porosity is due to the morphology/hydration of the electrochemically deposited RuO 2 as it relates to the confinement of the local spaces within which the deposition occurs, but more work w ould be needed to establish the cause. Table 3 1 The areal capacitances and the specific capacitances of the three different RuO 2 /SWNT films SWNT film type Mass (g ) Areal capacitance (mF/cm 2 ) Specific capacitance (F/g) Specific capacitance mass RuO 2 only (F/g) Standard 12 8.1 0.1 587 10 979 10 1x porous 13 11.1 0.1 736 10 1165 10 2.5x porous 13 16.3 0.1 1084 10 1715 10 Besides supercapacitors the porous films should also find application as support scaffolds in catalysis, e.g. in fuel cells and in more general catalytic and electro catalytic reformation. To date the engineered porosity has been introduced using monodispersed sacrificial nanospheres. Applications would likely benefit from the greater meso porosity likely to result from the use of polydisperse sacrificial particles. Haddon and co workers, in their own recognition of the mass flow limitations of pure SWNT films have incorporated additional porosity by mixing single wall and multi walled nanotubes obtaining enhanced perf ormance from fuel cell cathodes 12 The addition of a small fraction of long MWNTs could also provide the advantage of additiona l strength. A
46 combination of these approaches could produce the ideal engineered porosity scaffold for a broad range of applications. 3.2 Summary In this C hapter it was demonstrated that the porosity engineered into the SWNT film s were beneficial for the i r application as supports for a pseudocapacitive material A high ruthenium oxide specific capacitance of 1715 10 F/gm was achieved approaching the theoretical maximum specific capacitance of RuO x
47 CHAPTER 4 SWNT FILM ELECTRODE FOR HIGH PERFORMANCE ELEC TROACTIVE POLYMER BASED SUPERCAPACITOR Although the ruthenium oxide base supercapacitors have high specific capacitance, the material is too expensive to be practical except in the highest value applications such as in space an d some critical military app lications 34 E lectrically conducting polymers are in principle less expensive and some of them have shown a useful reversible pseudocapacit ive b ehavior already exploited in demonstration supercapacitors 35 36 Prof. Reynolds group in the UF department of Chemistry has developed 37 38 the poly (3, 4 alkylenedioxypyr roles) (PXDOPs) family of polymer s 39 41 among which poly (3, 4 propyl enedioxypyr role) (PProDOP) has following structure. In collaboration with Merve Ertas of Prof. Reynolds group we have explore d this pol ymer as the pseudocapacitive material for in supercapacitor devices using the SWNT films as the electrode. 4.1 Symmetric PProDOPs Supercapacitors PProDOP was deposited on both gold substrates (as a control electrode) and on SWNT films The PProDOP was dep osited onto the SWNT film electrode by electrochemical polymerization in a 10 2 M ProDOP sol ution in 0.1 M lithium bis(trifluoro methanesulfonyl)imide (LiBTI) salt in acetonitrile (ACN) using platinum as counter electrode and Ag wire as pseudo reference el ectrode. 20 cycles of potential pulse were applied during the polymerization with each cycle consists of 10 sec on (1.1V vs Fc/Fc + )
48 state and 10 sec off (0V vs Fc/Fc + ). After the electropolymerization a symmetric device was fabricated using p olymeric gel e lectrolytes containing LiBTI salt used as the conductive media betwe en the electrodes as shown in Figure 4 1. Figure 4 1. Schematic diagram of supercapa citor device configuration and f aradic processes of PProDOPs. In t his de vice configuration the two electrodes are in the half oxidized state. During the charging cycle the positively biased electrode goes towards higher oxidation while the negatively biased electrode is reduced to its neutral state and vice versa during the di scharge cycle. In contrast to expectations, given the far greater surface area of the nanotube film electrodes over the gold electrodes, the PProDOP on standard SWNT films gave capacitances that were substantially lower than that of the PProDOP on gold. Co nsideration of the reasons for this led to the conclusion that the ProDOP was delaminating from the low energy nanotube surfaces. This was confirmed by building devices using ProDOP electropolymerized onto porous SWNT films as the underlying electrodes. In that case the performance exceeded that of the gold devices. The rational
49 explaining this behavior was that the open network of the porous SWNT films allowed for penetration of the PProDOP into the nanotube film providing a mechanical interlocking that av oided the delamination. Poor interfacial adhesion of various electro active polymers to low energy surface of SWNT is a separate severe problem for application of nanotubes in devices. 4.2 Interlude O n Sticky P olymers A general solution to this problem was developed by Prof. Reynolds, Dr. Ryan Walczak and Prof. Rinzler. The idea is illustrated in Figure 4 2 Shown there is one subunit of a poly (9, 9 dioctylfluorene) [POF] to which pyrene molecules have been tethered via flexible alkane linkers. Pyrenes are known to have a weak non covalent binding with the nanotubes through a stacking interaction 42 43 Figure 4 2 Chemical structure of sticky PO F and schematic of a SWNT By incorporating many such pyrenes along the polymer backbone the net binding energy is multiplied, making the binding affinity of the polymer to the SWNT large Important to note is that the non covalent interaction between pyren e and the nanotube designed t o couple w ell to PProDOP stack on graphitic surface of SWNT Poly (9, 9 dioctylflouorene) [POF] as back bone SWNT
50 does not degrade the conductivity of the nanotube. The polymer backbone provides an anchor for subsequently deposited polymers, such as PProDOP, in principle solving the delamination problem. Because the backbone can be any of a number of different polymers the family of polymers having the ability to stick to nanotubes by this mechanism has come to be called S ticky P olymers. The S ticky group. To verify the associat ion of t he sticky POF with the nanotubes a standard 45 nm thick nanotube film on glass w as placed in a c hloroform solution of 0.1 mg/mL concentration of the polymer for 12 hrs. To remove excess unassociated polymer the films were subsequently soaked in 6 successi ve chloroform baths ( 12 hours each) followed by a final bath of methanol. Figure 4 3 shows two absorption spectra of sticky POF. One spectrum (black curve) is S ticky POF dissolved in chloroform while the other is S ticky POF assembled on to the nanotube fil m as described above (red curve). The fact that the sticky POF remained associated with the nanotubes after the long solvent baths demonstrates the high binding affinity of the Sticky POF for the SWNT surface s To confirm that the conductivity of the SWNT films was unaffected by the coating, four probe sheet resistance measurements were performed before and after association of SWNT with the sticky POF. The sheet resistance increased negligibly from 90 / sq to 96 / sq To further confirm the association of the S ticky POF to the nanotube surface s we AFM imaged a dilute network of SWNT s onto which Sticky POF had been assembled following similar procedures to those described above. The dilute SWNT network
51 sample on a silicon substrate was placed into S ticky P OF dissolved in chloroform (0.1 mg/mL ) overnight. Fig ure 4 3 Absorption spectra of sticky POF on SWNT and in solution To remove the excess unassociated sticky POF the samples were given 3 consecutive baths of chloroform for 1 2 hour each. A final bath of methanol wa s given for 12 hour and then samples were blow dried with nitrogen. All the sample preparation and polymer association steps were performed in a class 100 clean room environment Figure 4 4 compares AFM images of a dilute layer of bare SWNTs (on silicon) to S ticky POF coated SWNT (deposited and washed as above). T he smooth surface of the nanotube bundles before the application of sticky POF [F igure 4 4a ] compared to the rougher surface after application of sticky POF [Figure 4 4 b ] is clear evidence of the association.
52 Fig ure 4 4 AFM image of dilute network of SWNT (a) before and (b) after application of sticky POF 4.3 Return T o PProDOP B ased Supercapcitors Sticky POF coated SWN T films (80 nm thick) were used as the working electrode for electropolymerization of the ProDOP. Following the electropolymerization the capacitance of this electrode was measured in a three terminal measurement by linear sweep voltammetry With S ticky PO F on the SWNT films coated with PProDOP we now obtained capacitance exceeding that of the PProDOP on gold electrodes. Figure 4 5 show s the charging and discharging current density of PProDOP on the StickyPOF/SWNT electrode for different sweep rate s in LiB TI /p ropylene carbonate (PC) a b a b
53 electrolyte. The semi rectangular shape of the curves demonstrates the highly reversible nature of the charging for this electrode. The deviation from a more rectangular shape at higher scan rates suggests a kinetic hindrance fo r the ion permeation through the PProDOP film. Figure 4 5 C yclic voltammograms of PProDOPs on sticky POF/SWNT Electrode We studied the charging and discharging properties in gel electrolyte devices for PProDOP deposited on gol d and Sticky POF/SWNT. The supercapacitors consisting of PProDOP on gold electrodes showed a capacitance of 3 2 mF/cm 2 Sticky P O F on SWN Ts film electrodes showed a capacitance of 8.8 mF/cm 2 The specific capacitance for PProDOP on sticky POF coated SWNT f ilm was 1 22 F/gm. The stability of the gel electrolyte based two terminal devices were studied for PProDOP deposited on the gold and the S ticky POF/SWNT by continuous CV cycling (0 volt to 1.0 volt) for over 32000 nonstop cycles at scan rate of 100 mv/sec onds. As
54 shown in F igure 4 6, although t he gold substrate de vices showed a better stability over the course of the cycling experiment because the Sticky POF/SWNT device started with a much higher initial areal capacitance it still had the higher capacitanc e at the end of the experiment. Figure 4 6 C omparison of stability of SWNT device and gold In this C hapter, it has been demonstrated that the SWNT film can be used as electrode for electropolymerization of a pseudocapacitive polymer for use in super capacitors. Because these polymers can in principle be made inexpensively ( compared to the precious metal oxide based psuedocapacitive materials ) such devices may find practical applications. It was found that the low energy surfac e presented by the native nanotubes makes materials deposited onto them prone to delamination. This study was the first to demonstrate a general solution to this problem in the form an effectively monolayer interfacial adhesion layer (Sticky POF) developed for this purpose
55 CHAPTER 5 REMARKABLY HIGH OXYG EN REDUCTION ACTIVIT Y OF PURE SINGLE WAL L CARBON NANOTUBE FILM 5 .1 Background and Significance Improved sources of mobile power are increasingly desired. Applications include: electric vehicles (automobiles delivery vans, buses), portable consumer electronics (cell phones, PDAs, GPS locators), portable medical electronics (hearing aids, diagnostic telemetry), and military needs (communications, telemetry, autonomous surveillance and weapons delivery). Fuel cells and metal air breathing batteries already serve or have been demonstrated in some of these applications 44 47 With further improvements f uel cells and metal air batteries could become the principle energy storage and energy conversion devices in the coming decades. A f uel cell is a device 13 consists of two electrodes in contact wi th an electrolyte T he fuel (h ydrogen, methanol, formic acid, glucose etc) is oxidized at one of the electrode s ( the anode) and oxidant (typically oxygen) is reduced at other electrode ( the cathode) provid ing electrical energy (Figure 5 1). In a hydrogen f uel cell, hydrogen (H 2 ) is oxidized to proton s (H + ) by giving electron s to the anode while oxygen is reduced to hydroxyl ion (OH ) by accepting those electrons which travelled from anode to cathode through the external circuit while providing power The by product formed is clean water (by reaction of H + and OH ). Similarly in a metal air battery 48 the m etal (M) anode is oxidized to metal ion s (M +n ) and oxygen is reduced at the cathode generat ing metal hydroxide (or oxide) as the end product while generating the power in the external load (Figure 5 1). E lectrocatalysts are necessary at cathode for oxygen reduction and also at anode for hydroge n oxidation for the power generation in these devices The best
56 catalyst for both of these reactions is finely divided (nanoscale Platinum (Pt) or its alloys with other precious metals. 13 49 Figure 5 1 Schematic of a H 2 O 2 fuel cell and a metal air battery Among the main barriers to the broad use of fuel cells and metal air batteries for emission free vehicles is t his need f or Pt as electrocatalysts for oxygen reduction. The activity of Pt towards hydrogen oxidation greatly exceeds its activity towards oxygen reduction so the major fraction of the Pt must be used in the oxygen reducing cathode. Reflecting its scarcity, Pt is prone price inflation under high demands; the resource is limited to a few countries (mainly in South Africa and Russia; 80% of world Pt resources held in South Africa alone) and is generally scarce in the earth s crust 50 It has been estimated that a fleet of 500 million f uel cell vehicles (the number of vehicles on the road world wide in year 2000) would consume the worlds estimated total Pt supply i n about 15 years even with recycling at a 50% recovery rate. 51 This estimate does not H 2 O 2 H + OH H + H + H + H + OH OH OH OH H 2 O H 2 O 2 fuel cell Metal air battery Metal (fuel) O 2 OH M +n OH OH OH OH M(OH) n M +n M +n M +n M +n
57 include the competition of other platinum uses such as stationary power fuel cells, industrial catalysts, or jewelery and it assumed no further growth in worldwide number of vehicles. Besides the finite resource the conventional Pt based oxygen reduction electrode suffers from other issues. The nanoscale metal catalyst particles are subject to migration and agglomeration by Os t wald ripening resulting in the loss of activity towards ORR 52 O xidation of the electroca talysts 53 and the corrosion of c arbon support 36 are the other problems faced by the conventional electrocatalysts Since the observation of oxygen reduction activ it y of metal Phthalocyanine first reported in 1964, 54 there has been much effort directed at the development of non precious metal catalysts by the pyrolysis of Fe N 4 or Co N 4 macrocycle s on a variety of carbon support s in inert atmosphere 55 or by pyrolysis of n itrogen co ntaining cobalt or iron precursors 56 57 T hese metall ic electrocatalysts suffer from some of the same problems as the conventional Pt based cathodes (oxidation, migration, deactivation and loss of the carbonaceous support) Recently n itrogen containing so called vertical forests of m ultiwall n anotube s prepa red by pyrolysis of i ron phthalocyanine h ave show n promising catalytic activity towards oxygen reduction even after removal of the Fe growth catalysts 58 However, retention of the vertical forest structure in devices requires an elaborate process for fixing to a conducting electrode support and removal from growth substrate. The growth of such forests is also subject to sever producti on rate limits, all of which present difficulties in scale up for commercial applications. We have found that pure single wall carbon nanotubes alone, without any added metal catalyst, fabricated as thin films by vacuum filtration (a readily scaled proces s) exhibit a high oxygen reduction activity. This activity is quantified below via polarization
58 access of the gas phase oxidant to the SWNT layer. 5.2 Experimental Set U p and Electrode Fabrication Purified SWNT material ( Laser vaporization synthe sized, nitric acid purified 18 ) was dispersed via brief, bath ultrasonication in 1 wt% Triton X 100 TM surfactant to form a dilute aqueous suspension. Thin films from 30 to 1500 nm were vacuum filtered to the surface of nanoporous ,hy drophobic membranes made of Teflon (PTFE, Sartorius, 0.2 m), Nylon (GE Osmonics,0.22 m) or P olyvinylden efluoride (PVDF, Millipore,0.2 m).In each case the membrane was pre wetted with methanol to facilitate initial passage of the aqueous solution through the membrane ma terial. To minimize the loss of nanotubes by permeation through pores of this size sulfu ric acid was stirred into surfactant suspension for 2 minutes before initiation of the vacuum filtration. The limited ionic strength tolerance of the non ionic Triton X 100 TM surfactant disrupts its stabilizing properties. The degree of flocculation in th at time forms aggregate too large to permeate the pores of the membrane while still forming homogenous film of the nanotubes across the membrane surface. DI water washing removed the residual surfactant and the films/membrane was dried under a heat lamp. T he resulting SWNT films were 15 mm diameter circles, as defined by the filtration funnel, offset toward one edge of the 47 mm diameter circular membranes. The offset was such that the longest distance from edge of the membrane to the edge of the film was ~ 27 mm. Electrical contact was made to the films by sputtering Pd metal as rectangular area across this longest distance from membrane edge, overlapping ~2 mm of one edge of the SWNT film with the metal.
59 Testing of the electrodes was performed in a special ly constructed Teflon electrochemical cell. The cell was designed to expose a known geometric working area of the SWNT film to the aqueous electrolyte, while simultaneously allowing gas access to the film through the supporting hydrophobi c membrane. As sh own in Figure 5 2 the cell body consisted of a rectangular Teflon block with a cylindrical well bored from the top to form the electrolyte reservoir. Fi gure 5 2 Schematic of the Electrochemical cell A hole through the sidewall of the cell surrounded b y a captured O ring defined the area of the nanotube film exposed to the electrolyte (0.5 cm 2 ). Screws in the 4 corners of the gas delivery cover secured the cover to the cell body sandwiching the SWNT film/electrode/membrane against the O ring, thereby se aling the electrolyte volume. The gas delivery cover incorporated a gas plenum lined up with the hole in cell sidewall. Gas inlet and outlet ports terminated with Luer Loc fittings fed the desired gas to and from the plenum. This flow through design allowe d for rapid switching out of the gas to which
60 the cathode was exposed. The outlet line was routed to a water bubbler to prevent back streaming of ambient air. When the SWNT film, offset on the oversized membrane, was aligned over the hole in the cell sidew all the membrane supported Pd electrode extended from the edge of the nanotube film to beyond the edge of the cell body on one side permitting its electrical contact by a copper clip. Care was taken to always ensure that the metal electrode lay well outsid e the electrolyte wetted region of the SWNT film defined by the O ring. Not shown in Figure 1 is a tight fitting cover that incorporated sealed feed throughs for the reference electrode, the counter electrode and inert gas purge lines (inlet & outlet). The latter continuously purged N 2 gas through the electrolyte to ensure that the principal source of oxygen to the nanotube film electrode arrived through the hydrophobic porous membrane via the gas entering the gas plenum. The counter electrode was a Pt flag electrode 5.3 Experimental Results 5.3.1 El e ctrocatalytic Activit y of SWNT F ilm Figure 5 3 shows the result of a three terminal chrono amperometry measurement on a SWNT film working electrode (120 nm thick film on 0.22 micron pore N ylon membrane) in 0.1 M phosphate buffer pH 7 electrolyte at 400 mV versus an Ag/AgCl reference electrode. At the times shown by the arrows the gas indicated was switched from pure nitrogen to air (not flowing, nitrogen line merely disconnected allowing air diffusion into the gas plenum) to oxygen (flowing oxygen line connected) and finally back to nitrogen again. A fraction of the current seen (~200 A seen before and after the air/oxygen exposure) involves side reactions unrelated to externally supplied oxygen but the jump in current on exposure to ambient air and the further increase on switching
61 to pure (99.99%) oxygen can only be due to oxygen red uction reaction activity which is remarkably high given that the SWNT film is only 120 nm thick Figure 5 3 Chrono amperometry for a 120 nm thick SWNT film on 0.22 N ylon membrane in pH 7 phosphate buffer at 0.4 volt (vs. Ag/AgCl reference) upon exposure to the indicate d gas The relative ORR activity in ambient air of such a 120 n m thick SWNT film on PTFE membrane (0.22 m pores) was tested in acidic neutral and basic electrolyte e nvironments. Figure 5 4 shows the st eady state current densities as a function of potential in pH 1 (0.1 M sulfuric acid), pH 7 (phosphate buffer) and pH 13 (0.1 M potassium hydroxide). Steady state values were recorded at each potential after holding the potential for 1 hour (solid lines ar e guides for the eye). In the acidic media, in addition to the oxygen reduction current, proton reduction (with hydrogen evolution) contributes increasingly to the current as the potential is made increasingly negative. At
62 0.9 V the current density actual ly measured was 27 mA/cm 2 To determine the current contributed by oxygen reduction alone (the pH 1 curve in the plot), inert N 2 gas was first fed to the cathode to determine the potential dependent proton reduction currents. These were subsequently subtra cted from the currents when air was admitted to the cathode yielding the pH 1 curve shown in Figure 5 4. Figure 5 4 Steady state conver sion current of 120 nm SWNT film electrode in air with pH 1(blue curve), pH7 (red curve) and pH 13 (green curve) elec trolytes. Each point is taken after continuous 1 hr of measurement. Solid lines are guide for the eye. The substantial activity at all pH is highly desirable and should permit application of the SWNT oxygen reduction cathodes to all types of fuel cells an d metal air batteries allowing electrolyte optimization for the other components of the system. The observed ORR activity of SWNT in buffered pH 7 electrolyte is very promising as ORR electrocatalys t for enzyme based biofuel cell
63 Recently poly(ethylenedio xythiophene) (PEDOT) films were demonstrated to have a metal catalyst free oxygen red uction activity 59 In pH 1 electrolyte the current density reported at 900 mV (vs. SCE) was 6 mA/cm 2 with an areal mass given of 50 g/cm 2 This gives a mass basis specific activity of 0.120 A/mg. From our Figure 5 4 the current density at 870 mV (vs. Ag/AgCl 900 mV Vs. SCE ) can be estimated as 9.6 mA/cm 2 while the areal mass of the 120 nm thick SWNT film is 8.5 g/cm 2 which yields a mass basis specific activity of 1.13 A/mg nearly 10 times that of the PEDOT films 59 Figure 5 5 compares the steady state current densities as a function of potential in pH 7 (phosphate buffer) for SWNT films of increasing thickness and against a commercially available Pt loaded (0.5 mg/cm 2 ) electrode (ELA T HT140EW The Fuel Cell Store). The oxygen in these measurements was derived from atmospheric air at ambient pressure (gas cover inlet and outlet open to the atmosphere) with the cell at room temperature. The symbols are data points and the lines are leas t squares fits of exponential growth curves. The oxygen reduction currents increase with increasing SWNT film thickness. The 1500 nm thick SWNT film compares well with the performance of the much thicker Pt loaded commercial GDE. The low overpotential requ ired by Pt to effect the reduction continues to make Pt the better catalyst at low cell potentials however at potentials in the operational range of cells designed to operate near their maximum power points the pure SWNT cathode performs comparably or bett er. These comparable current densities for the 0.5 mg/cm 2 Pt loaded electrode and the 1500 nm thick SWNT electrode (areal mass 0.106 mg/cm 2 ) implies a mass basis oxygen reduction activity for the SWNT films that exceeds that of Pt by a factor of ~5.
64 Figure 5 5 Steady state current densities as a function of potential in pH 7 (phosphate buffer) for SWNT films of increasing thickness and against a commercially available Pt loaded (0.5 mg/cm 2 ) electrode These values are obtained without any optimization and sufficient for enzyme based biofuel cells 60 61 Microbial fuel cells 62 64 Microfuel cells 65 66 small direct methanol fuel cells 45 and advanced metal air battery 67 68 technologies. 5.3.2 Long Term S tability Test To test the l ong term stabili ty of the SWNTs in a harsh environment a continuous measurement (Figure 5 6) was performed at 0.4 volt versus an Ag/AgCl in 1M sulfuric acid electrolyte for 60 days. The current plotted is after the subtracting the contribution of hydrogen evolution rea ction due to proton reduction (measured independently by flowing N 2 to the SWNT film).
65 Figure 5 6 Long term testing at 0.4 volt versus Ag/AgCl with 120 nm SWNT electrode in 1M sulfuric acid. Spikes correspond to periodic cha nges of the acidic electrolyte. No degradation of the performance was observed even after 2 months of continuous operation Figure 5 7 compare s the catalytic activity of the 120 nm film in pH13 (0.1M KOH) Figure 5 7. Comparis on of the catalytic activity of the 120 nm SWNT film in pH 13 electrolyte for before and after the 60 day measurement in acid
66 electrolyte before the long term test in 1 M H 2 SO 4 (blue curve) and after the long term test (red curve). The catalytic activity in the alkaline electrolyte actually improved after the long term test in the acid. 5.3.3 Carbon M onoxide (CO) Test and Methanol Test to poisoning by carbon monoxide. 58 59 Such poisoning is however not an issue on the oxygen reducing electrode side of fuel cells (Figure 5 8) Figure 5 8. The effect of 10% carbon monoxide (red curve) compared to 10% argon (black curve) in oxygen balance on the oxygen reduction activity, as indicated by t he oxygen reduction current density for 0.4 V vs. Ag/AgCl applied to a commercial Pt loaded (0.5 mg/cm 2 ) air cathode. Note the rapid recovery of the oxygen reduction activity on turn off of the CO feed. The green curve shows the complete lack of an effect of 10% CO on the SWNT film based air cathode indicating that the activity is not metal catalyst based. That is, while the metal based catalysts do have a substantially greater affinity for CO than for oxygen, thereby degrading their performance in the pre sence of relatively
67 high CO concentrations, the oxygen reduction activity quickly recovers when the CO concentration is reduced to low levels.This is demonstrated in Figure 5 8 which shows what happens when the commercial Pt loaded ELAT air cathode, initia lly exposed to 100% oxygen is exposed to 10% argon (black curve) or 10% CO (red curve) with a balance of oxygen. The much smaller current degradation upon equivalent concentration argon exposure shows that the current degradation due to the CO exposure is due to more than just the ~9% reduction in the oxygen concentration. Note how quickly, however, the full activity (and more) recovers when the CO stream is turned off (the activity does in time level off to the initial level). This must be contrasted with Pt catalyzed hydrogen upon 10% CO exposure (hydrogen balance) and remains zero after CO feed is turned off (returning to 100% hydrogen). oxidation (Figure 5 9) for which the activity drops to zero Figure 5 9 Effect of 10% CO exposure to the H 2 oxidation activity of a commercial Pt loaded (0.5 mg/cm 2 ) electrode. The electrode was held at 0.165 volt vs Ag/AgCl in 1M sulfuric acid electrolyte while the gas plenum was exposed to 100% H 2 10% CO and 90% H 2 and 100% H 2 consecutiv ely as indicated in the graph.
68 While notions of any permanent CO poisoning of the oxygen reduction catalyst are thus rendered moot, such exposure of the SWNTs to 10% CO in oxygen balance (green curve Figure 5 8) does have utility. Despite their purification 18 the SWNT material used still contains a small fraction of the nickel and cobalt metal catalysts used in their synthesis. The lack of any effect of the CO exp osure strongly suggests that the oxygen reduction activity in the nanotube films is not catalyzed by the residual metal particles in the material In direct methanol fuel cells m ethanol cross over from the anode to Pt loaded cathode is a problem that degr ades the oxygen reduction capability of the cathode. 69 We have measured the methanol sensitivity of the SWNT based air cathode Figure 5 10 shows the effect of 1% (vol/vol) methanol added to the pH 13 electrolyte at a potential of 0.4 volt versus Ag/AgCl. Figure 5 10 Effect of methanol exposure on the ORR activity of a 120 nm SWNT film electrode. The electrode was held at 0.4 volt vs. Ag/AgCl in 0.1M KOH electrolyte while the gas plenum was exposed to air.
69 The steady state current decreased about 10% a steady value. Upon exchanging the methanol contaminated electrolyte with fresh electrolyte the electrode recover ed the ini tial steady state current value 5.4 Metal Air B attery 5.4.1 Mg A ir battery The performance of the SWN T air cathode was tested in a two terminal metal air battery configuration with different metal anode s (e.g. z inc, magnesium aluminum ) in a variety of electrolytes. The terminal of the cells was connected to a variable resistor through an ammeter allowing the current measurement. A voltmeter across the resistor was used to measure the potential across the variable resistor load. The p erformance of a SWNT air cathode in a Mg air battery using a 1.50 cm 2 by 0.15 mm thick sheet of magnesium metal (99% Sigma Aldrich) as the anode in 2M NaCl electrolyte is shown in Figure 5 11 as a function of the SWNT film thickness. Figure 5 11 Performance of different thick ness SWNT film as cathode in Mg air battery.
70 The open circuit voltage in all cases was 2.1 V. Without further optimization (e.g. the distance between the anode and the cathode in the cell here was ~ 1 cm whereas in a true battery this would be minimized to below 200 m to minimize cell internal impedance) the 1.2 m SWNT prov ide s about 30 mA/cm 2 current density at 1 V output voltage, with air as the oxygen source The gas plenum in these experiments was merely exposed to the ambient air. The current density provided by the Mg air battery increases as the thickness of the SWNT film increases. The short circuit current density is plotted in Figure 5 12 as a function off the film thickness The current density follow s a saturating exponential growth with film thickness. A char act eristic thickness can be derived from such a plot wh ich can specify the thickness of the cathode for which the the electrocatalysts utilization is optimized. i.e. point of diminishing return in current density For our configuration the thickness constant was abou t 350 nm. Figure 5 12 Shor t circuit current density of Mg air batteries for different film thickness
71 Figure 5 13 compares the performance of a 1500 n m thick SWNT film versus the Pt loaded ELAT electrode used as the oxygen redu ction cathode in a magnesium air electrochemical cell. The electrolyte was again 2 M NaCl. To test the maximum currents that can be drawn the oxygen source was 99.99% pure oxygen flowing through the gas plenum. Decreasing the load presented by the variable resistor yielded the polarization curves shown in Figure 5 13. Fig ure 5 13. Comparison of the performance of a 1.5 m SWNT film and a commercial Pt loaded (0.5 mg/cm 2 ) electrode in Mg O 2 battery. The open circuit voltage for both electrodes was 2.1 V. Again, at low cell potentials Pt remains the better catalyst however by the maximum power point the SWNTs have caught up to yield equivalent performance. This maximum power density (55 mW/ cm 2 ) and high short circuit current densi ty (nearly 170 mA/cm 2 ) approaches the
72 performance of Mg H 2 O 2 semi fuel cell cathodes investigated by the Navy for its unmanned undersea vehicle s program 70 5.4.3 Zn Air B attery The performance of a Zn air battery in 1M KOH electrolyte was studied with 120 nm SWNT air cathodes on 0.22 m PTFE membranes using a zinc rod as anode (Figure 5 14). An open circuit voltage of 1.45 volt was measured in the cell. With the u se of a 120 nm SWNT film in 1M KOH without any optimization of the internal resistance the Zinc air battery showed current density of 14.5 mA/cm 2 at 0.27 volt. Figure 5 14 Comparison of the performance of 120 nm SWNT film based Zinc air batt ery and commercial zinc air battery (Renate ZA5, Swatch, Switzerland) Even better performance can be anticipated if the zinc electrode had a high surface area and a shorter distance between the anode and cathode as is done in commercial cells
73 The stability of a zinc air battery in 1M KOH electrolyte was also tested. A 120 nm SWNT film on a hydrophobic Nylon membrane was used as the oxygen reduction cathode. The battery was continuously discharged for 50 hour by connecting the two terminals to a 2K resistor. The voltage across the resistor and current through it are shown in Figure 5 1 5 The se current densities demonstrated without optimization are al ready sufficient for some metal air batter y, microfuel cell smal l direct methanol fuel cell, micro bial fuel cell and bio fuel cell technologies. Figure 5 15 Stability of the out put voltage and current of a Zinc air battery with a 120 nm SWNT film on Nylon membrane However, it is proposed that still higher current densit ie s c ould be achieved by engineering the three phase interface of the SWNT electrode for the applications
74 requiring higher power. In conventional Pt based cathode s such larger surface area of the three phase interface is achieved throughout a substantial fra ction of the body of the relatively thick meandering layers of hydrophobic Teflon, Pt and carbon particles. However, the construction of the conventional electrode suffers f rom two well known problems 71 : 1) the pores that should provide the pathways for gas oxygen penetration to the catalysts particles can flood with electrolyte (few of the pathways are purely hydrophobic, containing a mixture of hydrophilic cat alyst and carbon) slowing the diffusion of oxygen to the catalysts surfaces ; 2) there exists in this construction a kinetic barrier to the diffusion of hydroxyl ion s throughout the torturous pathways of the hydrophilic part of the membrane, resulting in an additional current limiting impedance. For our filtration fabricated SWNT based cathodes a high surface area three phase interface can be engineered in SWNT based air electrode by simply assembling (filtering) the SWNT film to the surface of a pre patter ned h ydrophobic membrane (Figure 5 16 ). Figure 5 16 Schematic of SWNT air electrode with a patterned hydrophobic membrane to achieve high surface area three phase interface
75 As the nanotubes are drawn to the entire accessible surface of the membrane vi a hydrodynamic drag forces of the permeating solution the film will assembl on all regions of the membrane greatly increasing the size of the three phase interface The pre pattering of the membrane c ould be accomp lished by casting the co solvent polymer d opes (e.g. Nylon, Polysulfone, PVDF) used to fabricate flat membranes onto a substrate template that consists of protrusions or pits across its surface fabricated by conventional micro machining. Advantages of such SWNT based electrode s over the convention al electrode is that the membrane on to which SWNT film is assembled is sufficiently hydrophobic to completely prevent flooding of pores thus avoiding the diffusion barrier for oxygen and as the nanotubes films are comparatively thin the hydroxyl ion dif fusion barrier is minimized 5.5 Effect of Chemical M o dification to the Electrocatalytic A ctivity of SWNT 5.5.1 Effect of Oxygenated S pecies Created in Plasma Exposure of the SWNT F ilm In attempts t o obtain information regarding the active ORR sites of the SWNT film cathodes some experiments were performed to probe the effect of the defect introduction into the SWNTs. As the SWNTs are purified by refluxing in nitric acid 18 one may consider that the fuctionalization ( COOH OH, C=O or other oxygenated species) of the nanotubes might play a role in the oxygen reduction. To see if introducing further oxygenated sites on the SWNTs improved the activity such sites wer e introduced by brief exposure to reactive oxygen in a plasma asher. A 120 nm thick film was formed on a mixed cellulose ester (MCE) membrane followed by transfer of the film onto a 10 m pore PTFE membrane ( by dissolving away the MCE membran e) 7 The SWNT film on PTFE membrane was exposed to the oxygen plasma in an Anatech SCE600 p lasma a sher at 300 W power, 870 mTorr O2 pressure flowing at 450 sccm for
76 5 seconds The oxygen reduction activity of the plasma exposed SWNT film electrode was tested in the el ectrochemical cell. Figure 5 17 shows chronoampero metry of an as prepared SWNT film (blue curve) and an oxygen plasma exposed SWNT films (red curve) at 0.6 volt vs Ag/AgCl in air with pH7 electrolyte. Both film s were prepared on an MCE membrane the catalytic activity of the SWNT film was drastically degraded by the brief reactive oxygen exposure Figure 5 17 Comparison of the electrocatalytic activity of an O 2 plasma exposed SWNT film and an as prepared SWNT film Th is is likely due to a degradation of the conductivity of the quasi 1D SWNTs in which only a few defects can disrupt the conducting channels 5.5.2 Effect of Defect Sites Created by Controlled F unctionalization of the SWNT Recently, Go l dsmith et al. have demonstrated 72 the point controlled functionalization of SWNTs by oxidation and reduction of surface attached species in
77 acidic solution. We also probed the eff ect of such more controlled chemical functionalization of SWNTs on the oxygen reduction activity of the film s. A 120 nm film was fabricated on a 0.2 m pore PTFE membrane with two Pd contact electrodes at the two opposite edges of the 15 mm dia. circular f ilm. The two contacts permitted prob ing the two terminal resistance of the film before and after electro oxidation of the SWNT film. The electro chemical modification was performed in 1M sulfuric acid electrolyte. The defect sites were introduced by applyi ng a potential square wave with each cycle consist ing of V oxidation (1.45 volt versus Ag/AgCl for 10 seconds) and V reduction ( 0.1 volt versus Ag/AgCl for 10 seconds). The oxygen reduction activity was then probed in pH 7 buffer in air at 0.4 volt versus Ag/AgCl after the acid was replaced by the buffer (with multiple washing steps to ensure removal of the acid) Figure 5 18 shows that the oxygen reduction activity was increased (blue curve) by about 32 % after the treatment with four cycles of square pluse s. Figure 5 18. Catalytic activity SWNT film in pH7 in air on PTFE membrane before and after functionalization in 1M sulfuric acid
78 The two probe r esistance measured in pH7 buffer was increased by about 20% after this electroch emical treatment. The nanotube film was also treated with 8 cycles of square pluses instead of 4 cycles The catalytic activity was degraded (green curve) compared to the film modified by 4 cycles of potential pulses. The two probe resistance of the 8 cycl e treated film increased by 46% compared to the resistance of an as prepared film. The initial enhancement (4 cycles) followed by the degradation (8 cycles) of the catalytic activity along with the general conductivity degradation suggests that defect site s may play an important role in the ORR reaction but that this method of introducing the defects (whatever they may be) involves two c ompeting processes: one which puts in the right kind of defects but another which degrades the film conductivity and is pe rhaps the cause of the reduction in the ORR activity. 5.5.3 Effect of E dg e Sites of the N anotubes It is known that the graphite edge sites are comparatively much more active towards chemical reaction than the basal p l ane of the graphite due to the presence of unsaturated carbon bonds at the edge sites. 73 As nanotube ends are also opened by their nitric acid based purification the possibility exists that the nanotube ends or functional groups at the ends are responsible for the ORR activity. To test this hypothesis a film made with nanotubes purposely cu t to make more end sites. T he SWNTs were cut by ultra sonication of the nanotubes in a mixture of concentrated sulfuric and nitric acid (3:1, 98% and 7 0% respectively) at 40 C 74 A 120 nm thick SWNT film w as prepared on a 0.2 m PTFE membrane from an aqueous surfactant suspension of nanotubes containing 25% the acid cut tubes and 75% uncut tubes (without the uncut tubes the cut nanotubes would have passed through the membrane without forming a film) The r esulting film was tested in the electrochemical cell in pH 7
79 buffer electrolyte using air as the oxygen source Figure 5 19 compares the ORR activity of the composite film (25% cut, 75% uncut nanotubes ) with a film of 100% uncut tubes. Although the composi te film contains a much greater number of tube ends no significant change in the catalytic activity was observed. The experiment indicates that the nanotube end sites do not play a major role in the observed ORR activity of the SWNT film s Figu re 5 19. Comparison of the c atalytic activity of an uncut SWNT film (100%) and a film containing mixture of cut and uncut SWNT. It is unclear at present if the active sites in the SWNT film s involve site specific defects in their side wall s (e.g. chemical defects, topological defects), or simply the tight curvature of their sidewalls. Elucidating the mechanism for this high oxygen reduction activity remains for future work to addr ess. In this C hapter it was demonstrated that pure single wal l carbon nanotubes (SWNTs) alone, without any added metal catalyst, fabricated as thin films by vacuum filtration (a readily scaled process) exhibit a remarkably high oxygen reduction activity.
80 Th e activity wa s demonstrated via three terminal steady state electrochemical provides for access of the gas phase oxidant to the SWNT layer. The activity measured wa s found to be nearly 10 times that of recently published 59 results on polyethylenedioxythiophene (PEDOT) films, also shown to exhibit a Pt fre e oxygen reduction activity The activity in the pure SWNT films is shown to persist independent of the pH of the environment allowing for their use in highly corrosive basic or acidic device environments, without evidence of degradation o ver weeks of continuous operation. For simple initial device demonstrations the performance of the films as the cathode in metal air batteries w as characterized. These compare favorably with published results on metal air batteries. 67 68 70 To more directly compare the SWNT films with a Pt catalyst based cat hode we measured the steady state electrochemical performance of a commercially available Pt based cathode under the same experimental conditions in our electrochemical cell and then compare d their performance as cathodes in metal air batteries. A 1.5 m thick pure SWNT film achieve d current densities comparable with the much thicker commercial Pt based cathode. The conversion current densities demonstrate d are already sufficient for a number of technologies includin g for enzyme based biofuel cells 60 61 Microbial fuel cells 62 64 Microfuel cells 65 66 small direct methanol fuel cells 45 and advanced metal air battery 67 68 technologies
81 CHAPTER 6 HIGH RATES OF HYDROG EN OXIDATION AND GEN ARATION ACTIVITY OVE R SINGLE WALL CARBON N ANOTUBE FILM 6.1 Background and Signific ance disease, water, terrorism and war, Richard Smalley ranked energy as number one. With impeccable logic he noted that once the problem of a clean, carbon free, inexpensive sourc e of energy was developed the other problems became far less severe (e.g. poverty, disease and terrorism can all be linked to a low standard of living which cheap energy could resolve) 75 The supp ly of secure, carbon neutral energy is the most important challenge facing the scientific community 76 W ind and solar energy are the most widely available re newable energy sources available to us h owever, because of the sporadic nature of these energy sources and the asynchronicity between their availability and the demand a major challenge is the need for an efficient means to convert and store the energy an d then resupply when it is needed 77 80 One of the most efficient means to achieve such energy storage is the generation of a high energy density fuel: hydrogen via the electrol y sis of water (powered by solar or wind energy), followed by its reconversion into energy and water in a hydrogen air fuel cell. 81 82 Electrol y sis requires the following reactions: At anode 2H 2 2 + 4H + + 4e At Cathode 4H + + 4e 2 Overall reaction 2H 2 O 2H 2 + O 2 At the anode two molecules of water break up to produce one oxygen molecule and four protons. The protons are reduced at the cathode to produce H 2 gas (fuel)
82 accepting four e lectrons from the power source (wind or solar) The gases are stored for the future use. When the energy is needed the electrolyzer runs in reverse as a fuel cell supplying the energy. Stored H 2 and O 2 is fe d to the fuel cell at the anode and cathode respe ctively where the following reactions occur At anode 2H 2 + + 4e At cathode O 2 + 4H + + 4e 2 O Overall reaction 2H 2 + O 2 2H 2 O At the anode of the fuel cell two hydrogen molecules are oxidized to four protons leaving four electrons at the anode electrode. The electrons at the anode perform work in the exte rnal circuit in their journey to the cathode. At the cathode one oxygen molecule reduced by accepting those four electrons react s with four protons recreating water. Figure 6 1 is a diagrammatic representation of this sustainable energy economy scheme. 77 80 Efficiently perform ing these reactions (i.e. without the dissipating too much of the solar or wind derived energy) requires electrocatalysts at the electrodes. However, the conventional state of art electrodes in fuel cells and electrolyzers are based on the precious metal platinum (Pt) the scarcity of which has already been discussed in C hapter 5. Most non precious metals that have been tried as hydrogen evolution catalysts are subjected to dissolution in the typically strong acidic environment of the electrolyte and the ir activity is very low compare to Pt 81 83 Recently Nickel bisdiphosphine electrocatalyst supported on multiwall carbon nanotube (MWNT) ha s show n some very promising catalytic activity towa rds the hydrogen (H 2 ) oxidation and h ydrogen generation reactions 84 In control experiment s associated with those studies it
83 was found that the MWNT support was not an effective electrocatalysts for either hydrogen oxidation or generation. The observed catalytic H 2 oxidation and generation were thus s olely attributed to the Nickel b isdiphosphine electrocatalyst. Figure 6 1 Diagramm a tic representation of sustainable e nergy e conomy 77,80 Given the high oxygen reduction activity of the SWNT film s discussed in C hapter 5 it was natural for us us to explore the catalytic activity of SWNT towards H 2 oxidation. Initial experiment with the SW NT films did not show catalytic hydrogen oxidation or evolution without the application of high over potential s ( consistent w ith the previous observation 84 ). However, during our studies of oxygen reduction in sulfuric acid it was noticed that proton reduct ion currents (with hydrogen evolution) increased with time in the acid. In another, unrelated recent study a moderate hydrogen generation activity 85 of a conducting polymer PEDOT PEG, composed of poly (3,4 ethylenedioxythiophene) H 2 O O 2 Conversion H 2 H + H + H + H + H + H 2 O H 2 O 2 H + H + H + H + H + Storage EMF e Direct use FUEL CELL Use on demand ELECTROLYZER
84 (PEDOT) and the polyethylene glycol (PEG) was observed In that report it was noted that t he hyd rogen generation activity of PEDOT PEG improved after 48 hour immersion in 1M sulfuric acid The overpotential there remained too large to be truly useful but the improvement over the initial activity after the 48 hr of immersion in the acid was appreciab le. Coupled with our own observations that the H 2 generation activity of SWNT films increased with exposure time in sulfuric acid we decided to explore the H 2 generation activity and changes in the required overpotential for the SWNT film s with the time o f exposure to the acid electrolyte. Remarkably we found that after sufficient exposure of the SWNT film to concentrated sulfuric acid (activation) the H 2 evolution reaction (HER) starts at zero overpotential. This observation also led us to explore the per formance of acid exposed SWNT film towards the Hydrogen oxidation reaction (HOR). As shown below the hydrogen oxidation activity of the SWNT film s exceeds that reported by Le Goff et al for the Nickel Bisdiphosp hine MWNT device s 84 In fact on a per unit mass basis it is shown below that the H 2 evolution reaction activity of the SWNT films exceeds that of a commercially available Pt loaded electrode. 6.2 Experim ental Set U p and Electrode Fabrication F abrication of the SWNT film electrode on the hydrophobic Teflon (PTFE) membrane for hydrogen oxidation and generation is accomplished by same method as described in C hapter 5 T he hydrogen oxidation reaction (HOR) a nd hydrogen evolution reaction (HER) experiments were performed in the cell already described in C hapter 5. For the H 2 oxidation reaction (HOR) experiments the H 2 gas wa s fed at low flow rate to the gas plenum, through one port with excess gas exit ing thro ugh the other port
85 into a water bubbler that prevents back streaming of the ambient atmosphere. During the HOR and HER experiments the electrolyte was continuously purged with argon. 6.3 Experimental Results 6.3.1 Activation of SWNT F ilm The Hydrogen evol ution reaction (HER) activity of the SWNT films wa s studied in 1M sulfuric acid at successively increased exposure time to the acid electrolyte. During these measurements argon wa s fed to the gas plenum, through one port exit ing with the evolved hydrogen g as through the other port into a water bubbler that prevents back streaming of the ambient atmosphere. Argon was also purged into the electrolyte. Figure 6 3 shows the result of three terminals measurement s of the HER current density Figure 6 3 SWNT film (1.5 sulfuric acid scanned (5 mV/s) at 0, 1, 2, 4, 24, 48, 72, 96 and 120 hr of acid exposure.
86 versus potential for a 1.5 m thick film on 0.22 m pore Teflon membrane, scanned (5 mv/s) a fter 0, 1, 2, 4, 24 48, 72, 96 and 120 hr of acid exposure. N ote that the HER shifts to lower overpotentials as the HER activity increases with increasing time in the acid. W ith further time of exposure to the acid ( 288 hr) the overpotential for HER gradually decreased to z ero (relative to NHE) 6.3.2 H 2 Oxidation Activity of SWNT F ilm s Such activation is also necessary to observe hydrogen oxidation from the SWNT films at low overpotential. Figure 6 4 shows the activity of the SWNTs for hydrogen oxidation with and without t he H 2 SO 4 exposure. Figure 6 4 Chronoamperometry of a SWNT film exhibiting high hydrogen oxidation following activation by exposure to the 1 M sulfuric acid electrolyte for 288 hr.
87 Shown is the results of a three terminal chr onoampermetry measurement with a SWNT film working electrode (1.5 m thick film on PTFE membrane) in 1M sulfuric acid held at +0.3 Volt versus Normal Hydrogen electrode (NHE). The gas fed to the gas plenum and thus to the SWNT film through the PTFE mem brane is switched repeatedly from pure hydrogen to pure argon as indicated in the plot. The black curve is the initial response prior to H 2 SO 4 exposure, while the red curve shows the behavior after the SWNTs have been exposed to 1M sulfuric acid for 288 ho ur. Clearly the acid exposure activates the response and the activity is substantial. To determine if the long activation time (288 hrs) was reversible for a substantial fraction of such time out of acid the film was washed in 15 water baths at 2 min inter val s followed by 12 subsequent water baths at 12 hr intervals over 6 days. Following these exhaustive water baths the SWNT film was again tested for electocatalytic H 2 oxidation activity in acid electrolyte immediately upon exposure to the 1M sulfuric acid electrolyte. Figure 6 5 shows results of this three terminal chronoampermetry measurement under the same conditions as above. Interestingly the response is even greater than before so the activation is robust against water washing. Figure 6 6 shows the st eady state e lectrocatalytic activity towards H 2 oxidation and H 2 evolution of the activated (288 hr in 1M H 2 SO 4 ), 1.5 m thick SWNT film (red curve) scanned at 2 mv/s as a function of applied potential (referenced to NHE) using a Pt counter electrode. Th e currents in this voltammogram at this slow scan rate are representative of the steady state values at each potential as confirmed by testing discrete points along the curve and seeing that
88 the current holds steady for 10 minutes. Figure 6 5 Chronoamperome s electrocatalytic H 2 oxidation activity remains even though the acid was removed given exhaustive water baths for 6 days. Once the SWNTs are activated it remains activated even after complete removal of the acid. In the Figure t h e H 2 oxidation and H 2 evolution for the SWNT film is compared to the same measurements made on a commercially available, finely divide Pt loaded (0.5 mg/cm 2 ) carbon electrode (ELAT HT 140 EW, The fuel cell store) .For the purpose of the wetting the commercial Pt electrode and maximizing its performance the electrode was exposed to 1M sulfuric acid for 2 days and also cycled between +0.2 V to 0.5 V for 10 cycles over a few minutes ( while the argon gas was fed to the gas plenum ) The improvement seen from these treatments had fully saturated before the measurement s were made. Remarkably the H 2 oxidation and H 2 evolution reactions for the SWNT
89 electrode occur at zero overpotential just like for the Pt electrode No other material that that does not contain a conventional metal based catalyst has ever exhibited such behavior. Figure 6 6 Comparison of the electrocatalytic activity of the SWNT film and the commercially available Pt loaded carbon (Pt C) electrode. The 1.5 m thick SWNT film has a areal density of 0.106 mg/cm 2 This permits a mass basis activity comparison between the SWNTs and the Ni bisd iphosphine MWNT electrocatalyst as well as the Pt in the ELAT commercial electrode. Compared to the Ni bisd iphosphine MWN T electrocatalyst the mass basis activity for both H 2 evolution and H 2 oxidation (at 0.3 and +0.3 V respectively) is 5 times greater for the SWNT film. The Ni bisdiphosphine MWNT electrocatalyst thickness used had saturated in terms of the catalytic curre nt density for each reaction. T his is likely also true for the 1.5 m thick
90 SWNT film (assuming a thickness dependence like that shown for oxygen reduction in Figure 5 12) Comparing the catalytic current densities at at 0.3 and +0.3 V for the H 2 evolution and H 2 oxidation, respectively the SWNT film current densities a re about 10 times greater. While the thick, highly refined, Pt loaded commercial electrode outperforms the SWNT film in the catalytic current densities for the H 2 evolution and H 2 oxidation reactions, the large density advantage of the SWNT films means tha t on a mass basis the SWNT film actually outperforms the Pt based electrode. For example, at +0.2 V the hydrogen oxidation reaction current density for the Pt loaded electrode at 17.3 mA/cm 2 is 2.4 greater than that for the SWNT film of 7.2 mA/cm 2 Howeve r, as the mass of the Pt loading in the Pt catalyzed electrode exceeds the mass of the SWNT electrode by a factor activity for hydrogen oxidation exceeds that of the Pt electrode by a factor of 2. The mass normalized activity of th e SWN T film electrode towards the hydrogen evolution reaction is 4 fold greater than that of the Pt electrode where the currents are 30.7 mA/cm 2 and 36.1 mA/cm 2 for the SWNT and Pt electrodes at a cell potential of at 0.2V, respectively. 6.3.3 Long Term Stability M ost metal based electrocatalysts are prone to corrosion and dissolution in acidic environment s (9 11). Multiple hour H 2 oxidation reaction (at +0.3 V) and evolution reaction (at 0.3 V) measurements were performed for a 1.5 m thick SWNT fi lm in 1M sulfuric acid el e ctrolyte The results are shown in Figure 6 7 where the hydrogen oxidation (at +0.3 V) is shown in red and the hydrogen evolution (at 0.3 V) in black. The drop seen in the hydrogen oxidation current is not due to a reduction in t he electrode activity but rather some other long term drift as confirmed by the fact that when the
91 experiment was stopped and restarted the original value of 7.5 mA/cm 2 was recovered. Figure 6 7 Longterm (10 hour) H2 oxidation and evolution carried out over SWNT film electrode at +0.3 volt ( vs. NHE) and 0.3 volt ( vs. NHE) respectively in 1M sulfuric acid electrolyte. The noisiness observed in the (black) hydrogen evolution curve is because the evolved gas collects as bubbles on the electrode occluding parts of its surface until they grow large enough to float to the surface. The long term stability of the SWNT film electrode during these measurements omit s the remote possibility that the catalytic activity in the SWNT film i s due to residual nickel and cobalt nanotube growth catalyst in the SWNT material F rom the Pourbaix diagram, the oxide hydroxide or elemental form of these metals (Ni, Co) would be unstable under these assay conditions 86 87
92 6.3.4 Lack of CO poisioning One of the major problems in fuel cell technology is that platinum and all other previously known metal electrocatalysts for the H 2 oxidation reactions are readily poisoned by carbon monoxide (CO) even at concentration s as low as 10 ppm 13 88 89 Figure 6 8 shows the effect of CO exposure on the H 2 oxidation activity of the Figure 6 8 The effect of CO exposure to H 2 oxidatio n activity of the commercial Pt (0.5 mg/cm 2 ) loaded electrode (black curve) and a 1.5 m SWNT film (exposed to the acid for 120 hr).Note that H 2 oxidation activity of SWNTs recovers as CO turned off while activity of Pt remains zero as the CO turned off. c ommercial Pt loaded electrode and a 1.5 m SWNT film (acid exposed for 120 hour). Initially both electrodes were fed with 100% H 2 gas They were subsequently exposed to 500 ppm CO with the balance H 2 for about 400 seconds followed by a switch back to the p ure H 2 gas As can be seen the activity of both electrodes went to zero during the relatively high concentration CO exposure however in contrast to the activity of the Pt
93 electrode and all other metal based catalysts, the SWNT film electrodes activity rec overs once the CO stream is turned off. Unlike the metal catalysts the SWNT films are not poisioned by such CO exposure. This also demonstrates that residual metal SWNT growth catalysts in the films are not responsible for the activity observed. 6.3. 5 Demo ns tration of Fuel Cell W ith a SWNT F ilm A node To demonstrate the SWNT film as the H 2 oxidizing anode in a hydrogen air fuel cell configuration a two terminal measurement was performed in our electrochemical cell making use of a second wall opening (surroun ded by a captured o ring) and a second gas plenum cover. The SWNT film used was a film activated for 144 hr on a PTFE membrane ( as above ) and was situated over one sidewall opening. For the air cathode situated over the second wall opening a h alf membrane electrode assembly (MEA) was used. The half MEA consisted of a Pt (1.0 mg/cm 2 ) loaded gas diffusion electrode painted with a Nafion solution and hot pressed to one side of Nafion 212 m embrane The Nafion membrane faced the 1M sulfuric acid electrolyte in the cell. The SWNT film anode was fed H 2 gas flowing through its associated plenum cover at a low flow rate while the MEA was simply exposed to the ambient air by leaving the fittings to its plenum open to ambient. The anode was connected t o the cathode through a variable resistor. Voltage across the resistor and the current through it were monitored as the resistance was decreased. Figure 6 9 shows the resulting polarization curve.
94 Figure 6 9 Polarization curv e for SWNT film hydrogen oxidation anode operating in a H 2 air fuel cell A s imilar test was performed with SWNT air cathode instead of the Pt loaded air electrode. As described above a 1.5 m thick SWNT film on PTFE membrane was used as anode. Another 1.5 m thick SWNT film on PTFE membrane was used as cathode which was also exposed to the acid electrolyte. A Nafion 112 membrane was used as physical separator between two electrolyte filled chambers to prevent the diffusion of unreacted H 2 to the cathode. On the cathode side a Nafion 112 membrane was placed across the o ring with hemispherical dimple pressed into the membrane across the wall opening to form a reservoir for the electrolyte. U pon filling this reservoir with the acid electrolyte the SWNT air cathode was pressed against the reservoir by the gas delivery
95 cover sealing the fluid volume The SWNT film anode was fed H 2 gas flowing through its associated plenum cover at a low flow rat e while the SWNT film cathode was simply exposed to the ambient air. The anode was again connected to the cathode through a variable resistor and the v oltage across the resistor and the current through it monitored as the resistance was decreased. Figure 6 10 shows t he resulting polarization curve for the first completely Pt free (SWNT only) fuel cell. Figure 6 10 Polarization curve of a H2 air fuel cell (completely Pt free) operating with SWNT film anode and SWNT film cathode To become competitive with existing fuel cells such cells would have to improve by at least 2 orders of magnitude However, g iven the 150 year head start of Pt catalyzed fuel cells and the early stage of this research this is a good beginning This C hap ter discussed the discovery that upon activation by long term exposure to sulfuric acid SWNT films can act as electrocatalysts for hydrogen oxidation and
96 hydrogen evolution at zero potential (with respect to NHE), and thus with zero overpotential. Experime nts demonstrating that residual metal growth catalyst for the SWNTs possibly present in the films were not responsible for this activity. This is the first catalyst to this so that does not derive this activity from conventional metals. The activity of the SWNT films was compared to that of Nicke l Bisdiphosphine on MWNTs and on a commercially available, finely divided Pt loaded carbon electrode. The SWNT film activity towards both hydrogen evolution and oxidation compared favorably to each of these electrod es. The first ever conventional metal free fuel cell was demonstrated.
97 CHAPTER 7 SUMMARY AND FUTURE W ORK 7 .1 Summary This dissertation reports a simple, effective means to engineer controlled p orosity into nanotube films which promises to be beneficial i n applications ranging from energy storage to catalysis due to the enhanced accessible surface area of the film s The newly incorporated porosity was shown to modif y the film electrolytic capacitance and comparative pe rfusion rates. Application of the se high porosity SWNT films was demonstrated in RuO 2 based supercapacitor devices with record RuO 2 specific capacitance. Supercapacitor devices with PProDop as a polyme ric electroactive material on the SWNT film electrodes were also demonstrated for a more ec onomically viable alternative to expensive ruthenium, P ure single wall carbon nanotubes (SWNTs) alone, without any added metal catalyst, fabricated as thin films by vacuum filtration (a readily scaled process) were shown to exhibit a remarkable O 2 conversi on rate comparable to a much thicker commercially available Pt base cathode with a mass basis catalytic activity for oxygen reduction nearly 5 times that of the commercial Pt C c athode. The demonstrate d performance of such SWNT electrode s is already suffi cient for various technologies such as microbial fuel cells, bio fuel cells, micro fuel cell s direct methanol fuel cells and metal air batteries. Performance of the SWNT based air electrode was demonstrated in metal air batteries. It was shown that acid e xposed SWNTs become activated as an electrocatalyst for hydrogen generation and hydrogen oxidation at zero overpotential The catalytic activity of the activated SWNT films towards H 2 generation and oxidation was shown to exceed
98 that the activity, on a mas s basis, of a commercially available Pt loaded electrode. These findings suggest the possible replacement of Pt in a unitized regenerative fuel cell technolog y exploit ing a major renewable energy sources (e.g. solar, wind) for energy storage in the form o f the evolved hydrogen followed by its consumption when needed running in reverse as a fuel cell A Pt free f uel cell device based on a SWNT anode and a SWNT cathode was demonstrated. 7 .2 Future Work The most exciting and potentially important new findings to come out of this work are the previously unknown catalytic activity of the SWNTs for reactions important in energy storage and conversion. While such activity was demonstrated and characterized there remains much work to do in optimization and more imp ortantly in developing an understanding of the catalytic sites in this new class of catalytic material. It is unknown at present if these sites are specific regions of the SWNTs: sidewalls, ends, specific defects or functional groups.
99 APPENDIX A IN PRINCIPLE DENSITY OF HEXAGONAL CLOSED PACK ARRAY OF (10, 10) SWNT The average nanotube diameter in our dual pulsed laser vaporization grown material is about that of the (10, 10) nanotube having the armchair edge highlighted in Figure A 1. T he length of th e C C bond in the graphene lattice is 1.42 A the distance between two atoms ( D and E in Figure A 1) is 2.46 Angstrom 2 This is also the length of the unit cell which contains 40 carbon atoms. Figure A 1 Hexagonal lattice struc ture of a (10, 10) nanotube So a (10, 10) SWNT has 1 63 10 9 carbon atoms per cm. From X Ray diffraction it was reported 90 that the distance between adjacent nanotubes in a bundle (of ~(10, 10) diameter i s 16.95 (Figure A2 ). For a hexagonal close packed array of (10, 10)
100 nanotubes the area PQR= 1 244 10 1 4 cm 2 which then contains 8.15 10 8 carbon atoms for 1 cm of nanotube length This area fits into 1 cm 2 8.039 10 13 times giving 6.55 10 22 carbon atoms per cm 3 The mass of a carbon atom is 12 1.66 10 24 g. Hence th e density of this hypothetical hexagonal close packed array is 1.31 g/cm 3 Figure A 2 Schematic diagram of a cross section of a nanotube bundle
101 APPENDIX B STANDARD SWNT FILM DENSITY To determine the density of the standard nanotube films two films w ere made from the same thoroughly dispersed nanotube suspension (in 1% v/v Triton X 100 surfactant) by the filtration method. One film was fabricated on a PTFE membrane (200 nm pores, Sartorious) and thoroughly washed with methanol to remove the surfactant This film was thick enough to permit its clean separation from the membrane as an intact buckypaper. This was oven dried at 200C and weighed in a microbalance. The mass of this film and the volume of the original suspension that was used to make it gave the nanotube mass concentration in the suspension. The second thinner film was made on a cellulose ester membrane (Millipore, VC membrane) using a smaller known volume of the now known nanotube concentration. This gives the mass of nanotubes in the thinne r film. The inner diameter of the filter funnel where it contacts the membrane surface lated from sharp step edge was defined by photolithography and reactive oxygen etchi ng. Atomic force microscopy across the step edge gave the average film thickness. This procedure yields the measured density for such films of 0.71 g/cm 3
1 02 APPENDIX C ESTIMATE OF POROUS F ILM DENSITY The 1x porous film was made using a quantity of 200 nm diameter PS nanospheres that would form a hcp stack 3 layers high across the filter area. Simple geometric considerations lead to a thickness for this of 530 nm for a geometric volume per cm2 of 530x10 7 cm 3 Perfect sphere packing would yield a 74% volum e filling, leaving 26% of the volume empty. Hence per cm 2 of this stack there is (0.26 ) (530x10 7 cm 3 ) = 138x10 7 cm 3 of empty volume. A standard SWNT film of 80 nm thickness has volume per cm2 of 80x10 7 cm 3 Hence per cm 2 of hcp nanosphere stack area all the nanotubes used could fit within the intersticies between the nanosperes. As a crude estimate of the porous film density we take the areal mass of the 80 nm film (5.68 g/cm 2 ) and divide by the volume of the stack (per cm 2 ) of 530x10 7 cm 3 yielding a 1 x porous film density of 0.107 g/cm 3 Any pore collapse would make this an underestimate but compensating substantially in the other directio n is the lack of order in the films. Hence we accept this as a not unreasonable estimate of the 1x porous film dens ity. The 2.5x films scale accordingly. These are then ~1/7 and ~1/17 of the measured standard film density (0.71 g/cm 3 ).
103 LIST OF REFERENCES 1 Iijima, S. Helical Microtubules of Graphitic Carbon. Nature 354 56 58 (1991). 2 Saito, R., Dresselhaus, G. & Dresselhaus, M. S. Physical properties of carbon nanotubes (Imperial College Press, 1998). 3 Treacy, M. M. J., Ebbesen, T. W. & Gibson, J. M. Exceptionally high Young's modulus observed for individual carbon nanotubes. Nature 381 67 8 680 (1996). 4 Lourie, O. & Wagner, H. D. Transmission electron microscopy observations of fracture of single wall carbon nanotubes under axial tension. Appl Phys Lett 73 3527 3529 (1998). 5 Walters, D. A. Ericson L. M. Casavant, M. J. Liu, J., Col bert, D. T. et al. Elastic strain of freely suspended single wall carbon nanotube ropes. Appl Phys Lett 74 3803 3805 (1999). 6 Avouris, P., Chen, Z. H. & Perebeinos, V. Carbon based electronics. Nat Nanotechnol 2 605 615 (2007). 7 Wu, Z. C. Chen, Z. Du X.,Logan J. M. et al. Transparent, conductive carbon nanotube films. Science 305 1273 1276 (2004). 8 Wildoer, J. W. G., Venema, L. C., Rinzler, A. G., Smalley, R. E. & Dekker, C. Electronic structure of atomically resolved carbon nanotubes. Nature 391 59 62 (1998). 9 Olk, C. H. & Heremans, J. P. Scanning Tunneling Spectroscopy of Carbon Nanotubes. J Mater Res 9 259 262 (1994). 10 Enoki, T., Suzuki, M. & Endo, M. Graphite intercalation compounds and applications (Oxford University Press, 2003 ). 11 Bernier, P. & North Atlantic Treaty Organization. Scientific Affairs Division. Chemical physics of intercalation II (Plenum, 1993). 12 Ramesh, P., Itkis, M. E., Tang, J. M. & Haddon, R. C. SWNT MWNT hybrid architecture for proton exchange membr ane fuel cell cathodes. J Phys Chem C 112 9089 9094 (2008). 13 Winter, M. & Brodd, R. J. What are batteries, fuel cells, and supercapacitors? Chem Rev 104 4245 4269 (2004). 14 Han, L. T. Fukui, A., Chiba, Y., Islam, A. Komiya, R. et al. Integrated d ye sensitized solar cell module with conversion efficiency of 8.2%. Appl Phys Lett 94 013305 (2009).
104 15 Hu, L. B., Hecht, D. S. & Gruner, G. Carbon Nanotube Thin Films: Fabrication, Properties, and Applications. Chem Rev 110 5790 5844 (2010). 16 Song Y. I., Yang, C. M., Kim, D. Y., Kanoh, H. & Kaneko, K. Flexible transparent conducting single wall carbon nanotube film with network bridging method. J Colloid Interf Sci 318 365 371 (2008). 17 Krstic, V., Duesberg, G. S., Muster, J., Burghard, M. & R oth, S. Langmuir Blodgett films of matrix diluted single walled carbon nanotubes. Chem Mater 10 2338 (1998). 18 Rinzler, A. G. Liu J., Dai H., Nikolaev P. et al. Large scale purification of single wall carbon nanotubes: process, product, and charact erization. Appl Phys A Mater 67 29 37 (1998). 19 Whitby, R. L. D., Fukuda, T., Maekawa, T., James, S. L. & Mikhalovsky, S. V. Geometric control and tuneable pore size distribution of buckypaper and buckydiscs. Carbon 46 949 956 (2008). 20 Eswaramoort hy, M., Sen, R. & Rao, C. N. R. A study of micropores in single walled carbon nanotubes by the adsorption of gases and vapors. Chem Phys Lett 304 207 210 (1999). 21 Cooper, S. M., Chuang, H. F., Cinke, M., Cruden, B. A. & Meyyappan, M. Gas permeability of a buckypaper membrane. Nano Lett 3 189 192 (2003). 22 Du, W. F. Wilson, L., Ripmeester,J. Dutrisac R. et al. Investigation of the pore structure of as prepared and purified HiPco sing le walled carbon nanotubes by N 2 /Ar adsorption Implication for H 2 storage. Nano Lett 2 343 346 (2002). 23 Yang, C. M., Kaneko, K., Yudasaka, M. & Iijima, S. Effect of purification on pore structure of HiPco single walled carbon nanotube aggregates. Nano Lett 2 385 388 (2002). 24 Lin, C., Ritter, J. A. & Popov, B N. Correlation of double layer capacitance with the pore structure of sol gel derived carbon xerogels. J Electrochem Soc 146 3639 3643 (1999). 25 Tabata, S., Isshiki, Y. & Watanabe, M. Inverse opal carbons derived from a polymer precursor as electrode materials for electric double layer capacitors. J Electrochem Soc 155 K42 K49 (2008). 26 Hamann, C. H., Vielstich, W. & Hamnett, A. Electrochemistry 2nd, completely revised and updated edn, (Wiley VCH, 2007).
105 27 Conway, B. E. Transition from Superc apacitor to Battery Behavior in Electrochemical Energy Storage. J Electrochem Soc 138 1539 1548 (1991). 28 Conway, B. E. Electrochemical supercapacitors : scientific fundamentals and technological applications (Kluwer Academic/Plenum, 1999). 29 Dmows ki, W., Egami, T., Swider Lyons, K. E., Love, C. T. & Rolison, D. R. Local atomic structure and conduction mechanism of nanocrystalline hydrous RuO 2 from X ray scattering. J Phys Chem B 106 12677 1 2683 (2002). 30 Kim, I. H., Kim, J. H., Lee, Y. H. & Kim K. B. Synthesis and characterization of electrochemically prepared ruthenium oxide on carbon nanotube film substrate for supercapacitor applications. J Electrochem Soc 152 A2170 A2178 (2005). 31 Hu, C. C., Chang, K. H., Lin, M. C. & Wu, Y. T. Design a nd tailoring of the nanotubular arrayed architecture of hydrous RuO 2 for next generation supercapacitors. Nano Lett 6 2690 2695 (2006). 32 Hu, C. C. & Chen, W. C. Effects of substrates on the capacitive performance of RuOx. nH 2 O and activated carbon RuO x electrodes for supercapacitors. Electrochim Acta 49 3469 3477 (2004). 33 Barbieri, O., Hahn, M., Foelske, A. & Kotz, R. Effect of electronic resistance and water content on the performance of RuO 2 for supercapacitors. J Electrochem Soc 153 A2049 A205 4 (2006). 34 Liu, X. R. & Pickup, P. G. Ru oxide supercapacitors with high loadings and high power and energy densities. J Power Sources 176 410 416 (2008). 35 Rudge, A., Davey, J., Raistrick, I., Gottesfeld, S. & Ferraris, J. P. Conducting Polymers a s Active Materials in Electrochemical Capacitors. J Power Sources 47 89 107 (1994). 36 Tu, L. L. & Jia, C. Y. Conducting Polymers as Electrode Materials for Supercapacitors. Prog Chem 22 1610 1618 (2010). 37 Zong, K., Groenendaal, L. & Reynolds, J. R A new and efficient synthetic route toward 3,4 alkylene dioxypyrrole (XDOP) derivatives via Mitsunobu chemistry. Tetrahedron Lett 47 3521 3523 (2006). 38 Walczak, R. M. & Reynolds, J. R. Poly (3,4 alkylenedioxypyrroles): The PXDOPS as versatile yet un derutilized electroactive and conducting polymers. Adv Mater 18 1121 1131 (2006). 39 Cirpan, A., Argun, A. A., Grenier, C. R. G., Reeves, B. D. & Reynolds, J. R. Electrochromic devices based on soluble and processable dioxythiophene polymers. J Mater Ch em 13 2422 2428 (2003).
106 40 Reeves, B. D. Christophe Grenier, R. G. Argun, A. A., Cirpan, A. et al. Spray coatable electrochromic dioxythiophene polymers with high coloration efficiencies. Macromolecules 37 7559 7569 (2004). 41 Sonmez, G., Schottland, P., Zong, K. K. & Reynolds, J. R. Highly transmissive and conductive poly[(3,4 alkylenedioxy)pyrrole 2,5 diyl] (PXDOP) films prepared by air or transition metal catalyzed chemical oxidation. J Mater Chem 11 289 294 (2001). 42 Chen, R. J., Zhang, Y. G., Wang, D. W. & Dai, H. J. Noncovalent sidewall functionalization of single walled carbon nanotubes for protein immobilization. J Am Chem Soc 123 3838 3839 (2001). 43 Nakashima, N., Tomonari, Y. & Murakami, H. Water soluble single walled carbon nanotubes via noncovalent sidewall functionalization with a pyrene carrying ammonium ion. Chem Lett 638 639 (2002). 44 Srinivasan, S., Mosdale, R., Stevens, P. & Yang, C. Fuel cells: Reaching the era of clean and efficient power generation in the twenty first ce ntury. Annu Rev Energ Env 24 281 328 (1999). 45 Han, J. & Park, E. S. Direct methanol fuel cell combined with a small back up battery. J Power Sources 112 477 483 (2002). 46 Goldstein, J., Brown, I. & Koretz, B. New developments in the Electric Fuel Ltd zinc air system. J Power Sources 80 171 179 (1999). 47 Hasvold, O. et al. CLIPPER: a long range, autonomous underwater vehicle using magnesium fuel and oxygen from the sea. J Power Sources 136 232 239 (2004). 48 Linden, D. Handbook of batteries 2nd ed edn, (McGraw Hill, 1995). 49 Hamlen, R. P., Jerabek, E. C., Ruzzo, J. C. & Siwek, E. G. Anodes for Refuelable Magnesium Air Batteries. J Electrochem Soc 116 1588 & (1969). 50 Sealy, C. The problem with platinum. Mater Today 11 65 68 (2008). 51 Gordon, R. B., Bertram, M. & Graedel, T. E. Metal stocks and sustainability. P Natl Acad Sci USA 103 1209 1214 (2006). 52 Wilson, M. S., Garzon, F. H., Sickafus, K. E. & Gottesfeld, S. Surface Area Loss of Supported Platinum in Polymer Electrolyte F uel Cells. J Electrochem Soc 140 2872 2877 (1993).
107 53 Yu, X. W. & Ye, S. Y. Recent advances in activity and durability enhancement of Pt/C catalytic cathode in PEMFC Part II: Degradation mechanism and durability enhancement of carbon supported platinu m catalyst. J Power Sources 172 145 154 (2007). 54 Jasinski, R. New Fuel Cell Cathode Catalyst. Nature 201 1212 (1964). 55 Faubert, G. Lalande, G., Ct, R., Guay, D. et al. Heat treated iron and cobalt tetraphenylporphyrins adsorbed on carbon black : Physical characterization and catalytic properties of these materials for the reduction of oxygen in polymer electrolyte fuel cells. Electrochim Acta 41 1689 1701 (1996). 56 Gupta, S., Tryk, D., Bae, I., Aldred, W. & Yeager, E. Heat Treated Polyacrylo nitrile Based Catalysts for Oxygen Electroreduction. J Appl Electrochem 19 19 27 (1989). 57 Bashyam, R. & Zelenay, P. A class of non precious metal composite catalysts for fuel cells. Nature 443 63 66 (2006). 58 Gong, K. P., Du, F., Xia, Z. H., Durst ock, M. & Dai, L. M. Nitrogen Doped Carbon Nanotube Arrays with High Electrocatalytic Activity for Oxygen Reduction. Science 323 760 764 (2009). 59 Winther Jensen, B., Winther Jensen, O., Forsyth, M. & MacFarlane, D. R. High rates of oxygen reduction ov er a vapor phase polymerized PEDOT electrode. Science 321 671 674 (2008). 60 Gao, F., Viry, L., Maugey, M., Poulin, P. & Mano, N. Engineering hybrid nanotube wires for high power biofuel cells. Nat Commun DOI: 10.1038/ncomms 1000 (2010) (2010). 61 Dav is, F. & Higson, S. P. J. Biofuel cells Recent advances and applications. Biosens Bioelectron 22 1224 1235 (2007). 62 Logan, B. E. Exoelectrogenic bacteria that power microbial fuel cells. Nat Rev Microbiol 7 375 381 (2009). 63 Logan, B., Cheng, S. Watson, V. & Estadt, G. Graphite fiber brush anodes for increased power production in air cathode microbial fuel cells. Environ Sci Technol 41 3341 3346 (2007). 64 Chaudhuri, S. K. & Lovley, D. R. Electricity generation by direct oxidation of glucose in mediatorless microbial fuel cells. Nat Biotechnol 21 1229 1232 (2003).
108 65 Tominaka, S., Ohta, S., Obata, H., Momma, T. & Osaka, T. On chip fuel cell: Micro direct methanol fuel cell of an air breathing, membraneless, and monolithic design. J Am Chem Soc 130 10456 (2008). 66 Morales Acosta, D., Rodriguez, H., Godinez, L. A. & Arriaga, L. G. Performance increase of microfluidic formic acid fuel cell using Pd/MWCNTs as catalyst. J Power Sources 195 1862 1865 (2010). 67 Malone, E., Berry, M. & Lipso n, H. Freeform fabrication and characterization of Zn air batteries. Rapid Prototyping J 14 128 140 (2008). 68 Chao, W. K., Lee, C. M., Shieu, S. Y., Chou, C. C. & Shieu, F. S. Clay as a dispersant in the catalyst layer for zinc air fuel cells. J Power Sources 177 637 642 (2008). 69 Choi, W. C., Kim, J. D. & Woo, S. I. Modification of proton conducting membrane for reducing methanol crossover in a direct methanol fuel cell. J Power Sources 96 411 414 (2001). 70 Patrissi, C. J., Bessette, R. R., Kim Y. K. & Schumacher, C. R. Fabrication and rate performance of a microfiber cathode in a Mg H 2 O 2 flowing electrolyte semi fuel cell. J Electrochem Soc 155 B558 B562 (2008). 71 Bidault, F., Brett, D. J. L., Middleton, P. H. & Brandon, N. P. Review of ga s diffusion cathodes for alkaline fuel cells. J Power Sources 187 39 48 (2009). 72 Goldsmith, B. R. Coroneus J. G., Khalap V. R.,Kane, A. A. et al. Conductance controlled point functionalization of single walled carbon nanotubes. Science 315 77 81 (2 007). 73 Zaghib, K., Song, X. & Kinoshita, K. Thermal analysis of the oxidation of natural graphite: isothermal kinetic studies. Thermochim Acta 371 57 64 (2001). 74 Liu, J. Rinzler, A. G., Dai, H., Hafner, J. H. et al. Fullerene pipes. Science 280 1253 1256 (1998). 75 Smalley, R. E. Future global energy prosperity: The terawatt challenge. Mrs Bull 30 412 417 (2005). 76 Rockstrom, J. ,Steffen, W., Noone, K., Persson A. et al. A safe operating space for humanity. Nature 461 4 72 475 (2009). 77 Turner, J. A. A realizable renewable energy future. Science 285 687 689 (1999). 78 Lewis, N. S. & Nocera, D. G. Powering the planet: Chemical challenges in solar energy utilization. P Natl Acad Sci USA 103 15729 15735 (2006).
109 79 Turner, J. A. A reali zable renewable energ y future Science 285 1493 1493 (1999). 80 Hambourger, M. & Moore, T. A. Nailing Down Nickel for Electrocatalysis. Science 326 1355 1356 (2009). 81 Cook, T. R. Dogutan, D. K., Reece, S. Y., Surendranath, Y. et al. Solar Energy Su pply and Storage for the Legacy and Non legacy Worlds. Chem Rev 110 6474 6502 (2010). 82 Hoffert, M. I. Caldeira, K., Benford, G., Criswel D. R. et al. Advanced technology paths to global climate stability: Energy for a greenhouse planet. Science 298 981 987 (2002). 83 Greeley, J., Jaramillo, T. F., Bonde, J., Chorkendorff, I. B. & Norskov, J. K. Computational high throughput screening of electrocatalytic materials for hydrogen evolution. Nat Mater 5 909 913 (2006). 84 Le Goff, A. Artero,V., Jou sselme, B., Tran P. D. et al. From Hydrogenases to Noble Metal Fre e Catalytic Nanomaterials for H 2 Production and Uptake. Science 326 1384 1387 (2009). 85 Winther Jensen, B., Fraser, K., Ong, C., Forsyth, M. & MacFarlane, D. R. Conducting Polymer Compos ite Materials for Hydrogen Generation. Adv Mater 22 1727 1732 (2010). 86 Beverskog, B. & Puigdomenech, I. Revised Pourbaix diagrams for nickel at 25 300 degrees C. Corros Sci 39 969 980 (1997). 87 Powell, D., Cortez, J. & Mellon, E. K. A Laboratory E xercise Introducing Students to the Pourbaix Diagram for Cobalt. J Chem Educ 64 165 167 (1987). 88 Markovic, N. M., Grgur, B. N., Lucas, C. A. & Ross, P. N. Electrooxidation of CO and H 2 /CO mixtures on Pt(111) in acid solutions. J Phys Chem B 103 487 4 95 (1999). 89 Garcia, A. C., Paganin, V. A. & Ticianelli, E. A. CO tolerance of PdPt/C and PdPtRu/C anodes for PEMFC. Electrochim Acta 53 4309 4315 (2008). 90 Thess, A. Lee, R., Nikolaev P. Da i, H. et al. Crystalline ropes of metallic carbon nanotub es. Science 273 483 487 (1996).
110 BIOGRAPHICAL SKETCH Rajib Das was born in Kapista, West Bengal, India. He had very fascinating childhood as he grew up close to the nature while his father and some ancient philosophical books unknowingly influenced th e his thought process reasoning and questioning about the life and various facts of nature. He holds good academic records since fifth grade He struggled through the severe financial crisis which always threatened his academic life He grew his immense interest in mathematics and physics during his high school educations and was surprised by the advancement of human understanding through the research. Although he scored top in his school in 12 th grade final exam and also topped in the county he could not had enough financial support to pursue his dream of higher education in physics. He has a vivid memory of an incident which changed the track of his academic life. One after noon while he was wondering in the woods he met a monk named S h udhangshu Maharaja Su rprisingly enough, after knowing his exam scores the monk decided to write a letter for him to Ramakrishna Mission for s cholarships Finally, with the great help of Shudangshu Maharaja and Swami Shibamay a nandaji Rajib received a scholarship from Ramakr ishna mission for pursuing Bachelor of Science in physics at R.K. Mission Vidyamandira in Calcutta University Rajib received his Bachelor of Science ( B. Sc. ) (Honors) in p hysics from Calcutta University in 2003. He was invited experimental physics worksho p in summer 2003 as he was topped in the country in National Gradu ate Physics Examination (NGPE) which was held 96 centers across India This experimental workshop was exposed him various fascinating experiments in the field. With his good academic record s and interview he was accepted for M.Sc in Physics in Indian Institute of Technology (IIT) Kanpur, one of the best universities in
111 India. During summer 2004, he worked on theoretical project on magnetic nanoparticles with Professor Sushanta Dattagupta in S.N. Bose National center for Basic Sciences, Kolkata. During 2004 he did experimental work on low temperature measurements on magnetic materials and the phase transition behaviors of these materials with Professor A.K. Majumdar in Low temperature physics laboratory of IIT Kanpur. He received Master of Science (M.Sc.) in physics from IIT Kanpur in 2005. In fall 2005, he came to USA to pursue Ph.D. program in physics in University of Florida. He decided to do experimental work on carbon nanotube and joined Professor He received his Ph.D. in physics from University of Florida in the spring of 2011.