USF Libraries
USF Digital Collections

Development of thin CsHSO₄ membrane electrode assemblies for electrolysis and fuel cell applications

MISSING IMAGE

Material Information

Title:
Development of thin CsHSO₄ membrane electrode assemblies for electrolysis and fuel cell applications
Physical Description:
Book
Language:
English
Creator:
Ecklund-Mitchell, Lars E
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Solid acid
Electrolyte
Methanol
Hydrogen
MEA
Dissertations, Academic -- Chemical Engineering -- Masters -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: In this work the use of the solid acid CsHSO₄ as an electrolyte in a hydrogen/oxygen fuel cell or the disassociation of water into hydrogen and oxygen has been investigated. Several issues have been cited in literature regarding the use of CsHSO₄ as a solid electrolyte; these include: difficulty interpreting proton conductivity profiles of real membranes, high permeability of the membrane to fuel and product gases, and low mechanical strength. In an attempt to improve our understanding and possibly eliminate these issues, performance characteristics of prepared CsHSO₄ membranes have been investigated utilizing various methods of synthesis and membrane fabrication. A consistent method of CsHSO₄ membrane construction was developed based on these investigations. In addition, a novel method of sintering to decrease the membrane's permeability to fuel gases was developed and evaluated. The effects of these measures were investigated and tested in a prototype cell for proof of concept of fuel cell and electrolysis applications.
Thesis:
Thesis (M.S.Ch.)--University of South Florida, 2008.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Lars E. Ecklund-Mitchell.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 106 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002001349
oclc - 319699930
usfldc doi - E14-SFE0002627
usfldc handle - e14.2627
System ID:
SFS0026944:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200385Ka 4500
controlfield tag 001 002001349
005 20090430122123.0
007 cr bnu|||uuuuu
008 090430s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002627
035
(OCoLC)319699930
040
FHM
c FHM
049
FHMM
090
TP145 (Online)
1 100
Ecklund-Mitchell, Lars E.
0 245
Development of thin CsHSO membrane electrode assemblies for electrolysis and fuel cell applications
h [electronic resource] /
by Lars E. Ecklund-Mitchell.
260
[Tampa, Fla] :
b University of South Florida,
2008.
500
Title from PDF of title page.
Document formatted into pages; contains 106 pages.
502
Thesis (M.S.Ch.)--University of South Florida, 2008.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
3 520
ABSTRACT: In this work the use of the solid acid CsHSO as an electrolyte in a hydrogen/oxygen fuel cell or the disassociation of water into hydrogen and oxygen has been investigated. Several issues have been cited in literature regarding the use of CsHSO as a solid electrolyte; these include: difficulty interpreting proton conductivity profiles of real membranes, high permeability of the membrane to fuel and product gases, and low mechanical strength. In an attempt to improve our understanding and possibly eliminate these issues, performance characteristics of prepared CsHSO membranes have been investigated utilizing various methods of synthesis and membrane fabrication. A consistent method of CsHSO membrane construction was developed based on these investigations. In addition, a novel method of sintering to decrease the membrane's permeability to fuel gases was developed and evaluated. The effects of these measures were investigated and tested in a prototype cell for proof of concept of fuel cell and electrolysis applications.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: John T. Wolan, Ph.D.
653
Solid acid
Electrolyte
Methanol
Hydrogen
MEA
690
Dissertations, Academic
z USF
x Chemical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2627



PAGE 1

Development of Thin CsHSO4 Membrane Electrode Assemblies for Electrolysis and Fuel Cell Applications by Lars E. Ecklund-Mitchell A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Chemical Engineering Department of Chemical Engineering College of Engineering University of South Florida Major Professor: John T. Wolan, Ph.D. Burton Krakow, Ph.D. Elias K. Stefanakos, Ph.D. Aydin Sunol, Ph.D. Defense Date: October 3, 2008 Keywords: solid acid, electro lyte, methanol, hydrogen, MEA Copyright 2008, Lars E. Ecklund-Mitchell

PAGE 2

Acknowledgments No one makes a journey entirely unassisted. I ow e any success that I meet from this work from the many people who gave me support, enc ouragement, and advice. First of all, my parents, without whom I could not have even be gun on this great adventur e. I would also like to thank Drs. Wolan and Krakow, whose advice and support provided motivation and direction for this project. Fi nally, all my colleagues in the Applied Surface Science Lab and Fuel Cell Research Lab, but specifically Benjamin Grayson, Eric Weaver, Don Payne, Timothy Fawcett, Ala’a Kababji, and Jonathan Mbah for their friendship and for providing technical assistance, editing, and sounding boards for my ideas. Thank you to all.

PAGE 3

i Table of Contents List of Tables ................................................................................................................ ........... iii List of Figures ............................................................................................................... ........... iv Abstract ...................................................................................................................... .............. vi 1Introduction .................................................................................................................. ........11.1Solid Acid Electrolytes and CsHSO4 .......................................................................21.1.1Structure .................................................................................................41.1.1.1Bonding ...................................................................................71.1.1.2Order-Disorder ......................................................................101.1.2Proton Conductivity .............................................................................121.1.3Phase Transitions .................................................................................151.2Applications ...........................................................................................................171.2.1Fuel Cells .............................................................................................181.2.2Electrolysis ...........................................................................................27 2Development of a Solid Acid Electrolyte ..........................................................................292.1Synthesis of the Electrolyte ...................................................................................292.1.1Chemical Synthesis ..............................................................................302.1.2Crystallization ......................................................................................312.1.3Pressing Methods .................................................................................322.2Impedance Spectroscopy .......................................................................................352.2.1Proton Conductivity .............................................................................352.2.2Experimental Measurements ................................................................382.3Thermal Analysis ...................................................................................................412.4X-ray Diffraction ...................................................................................................522.5X-ray Photoelectron Spectroscopy ........................................................................60 3Permeability Studies .......................................................................................................... 653.1Permeability ...........................................................................................................653.2Experimental Setup ................................................................................................683.3Gas Chromatography .............................................................................................703.4Effect of Methanol .................................................................................................71

PAGE 4

ii 4Fuel Cell and Electroly sis Applications .............................................................................774.1Experimental Setup ................................................................................................774.1.1Fuel Cell ...............................................................................................784.1.2Electrolysis ...........................................................................................794.2MEA Construction .................................................................................................794.2.1Electrolyte ............................................................................................794.2.2Electrodes .............................................................................................804.2.3Catalysts ...............................................................................................804.3Fuel Cell Performance ...........................................................................................814.4Electrolysis Performance .......................................................................................864.4.1Electrolysis of Methanol ......................................................................864.4.2Electrolysis of Steam ...........................................................................88 5Summary and Conclusions ................................................................................................915.1Materials Characterization .....................................................................................915.2Permeability Studies ..............................................................................................955.3Fuel Cell and Electrolyte Test Bed ........................................................................975.4Future Directions .................................................................................................100 References .................................................................................................................... ..........104

PAGE 5

iii List of Tables Table 1-1 ........Correlation Between Hydrogen Bond Types and Bond Character ............................................................................................................9 Table 1-2 ........Comparison of Elect ronic and Ionic Conductivity ..........................................12 Table 2-1 ........Impedance Spectroscopy Data for CsHSO4 .....................................................40 Table 2-2 ........DSC/TGA Data for CsHSO4 ............................................................................50 Table 2-3 ........Data for Thermal Stability of CsHSO4 ............................................................51 Table 2-4 ........Experimental PXRD versus CsHSO4-III Reference Pattern ............................54 Table 2-5 ........Experimental PXRD versus CsHSO4-III and CsHSO4-II Reference Patterns ...........................................................................................56 Table 2-6 ........Experimental PXRD versus CsHSO4-II Reference Pattern .............................57 Table 2-7 ........CsHSO4 Crystallite Size ..................................................................................60 Table 2-8 ........Binding Energies and Atom ic Sensitivity Factors for Species Present in CsHSO4 ...........................................................................................62 Table 3-1 ........Permeation Data for Methane through CsHSO4 ..............................................74 Table 4-1 ........Fuel Cell Data .............................................................................................. ....86

PAGE 6

iv List of Figures Figure 1.1 .......Structure of CsHSO4 at Room Temperature Phase ............................................6 Figure 1.2 .......Structure of CsHSO4 at High Temperature Phase .............................................7 Figure 1.3 .......The Mechanics of Proton Conduction in CsHSO4 ...........................................15 Figure 1.4 .......CsHSO4 Proton Conductivity from Boysen .....................................................17 Figure 1.5 .......Diagram of a Hydrogen Fuel Cell ....................................................................20 Figure 1.6 .......Polarization Losses in a Hydrogen Fuel Cell ...................................................26 Figure 2.1 .......CsHSO4 Membranes Pressed Under Ambient and Heightened Temperature x100 ............................................................................................34 Figure 2.2 .......Impedance Spectra of CsHSO4 ........................................................................39 Figure 2.3 .......High Temperature Impedance Spectra of CsHSO4 ..........................................40 Figure 2.4 .......CsHSO4 Crystallized by Acetone ....................................................................42 Figure 2.5 .......CsHSO4 Crystallized by Temperature Control ................................................43 Figure 2.6 .......CSHSO4 Crystallized by Temper ature Control, a Second Attempt ............................................................................................................44 Figure 2.7 .......DSC of CsHSO4 Using Closed Cup Q10 DSC ................................................46 Figure 2.8 .......DSC of CsHSO4 after Excessive Drying .........................................................47 Figure 2.9 .......DSC/TGA of CsHSO4 Rerun ...........................................................................48 Figure 2.10 .....Stability of CsHSO4 .........................................................................................51 Figure 2.11 .....PXRD Pattern of Monoclinic CsHSO4-III .......................................................54 Figure 2.12 .....PXRD of CsHSO4 Showing CsHSO4-II and CsHSO4-III Features ............................................................................................................55 Figure 2.13 .....PXRD Pattern of Monoclinic CsHSO4-II ........................................................57 Figure 2.14 .....X-ray Photoelectron Spec troscopy Survey Scan of CsHSO4 Powder Crystallized with Methanol .................................................................63 Figure 2.15 .....XPS High Resolution Scan of Sulfur 2p Peaks ...............................................64 Figure 3.1 .......Schematic of the Permeability Cell..................................................................70 Figure 3.2 .......Permeation Response of CsHSO4 Membrane to Methanol Vapor................................................................................................................72 Figure 3.3 .......Permeation of Methane through Fresh CsHSO4 Membrane ............................73 Figure 3.4 .......Permeation of Methane through CsHSO4 Membrane after Treatment with Methanol .................................................................................74 Figure 3.5 .......Effect of 50/50 Mix of Methanol/Steam on Permeability to Methane............................................................................................................76 Figure 4.1 .......Circuit Diagram for Polari zation Measurements of Fuel Cell .........................78 Figure 4.2 .......Catalyst/Elect rode Construction ......................................................................81 Figure 4.3 .......Polarization Curve of CsHSO4 Fuel Cell with H2 and Air ..............................82

PAGE 7

v Figure 4.4 .......Polarization Curve of CsHSO4 Fuel Cell with H2 and Air after Methanol Electrolysis ......................................................................................83 Figure 4.5 .......Polarization Curve of CsHSO4 Fuel Cell with H2 and Air after Methanol Sintering (Long-term) ......................................................................84 Figure 4.6 .......Polarization Curve of CsHSO4 Fuel Cell with H2 and Air after Methanol Sintering and Regeneration .............................................................85 Figure 4.7 .......Electrolysis of Methanol ..................................................................................87 Figure 4.8 .......Electrolysis of Methanol at Low Voltage ........................................................88 Figure 4.9 .......Electrolysis of Steam ....................................................................................... 89 Figure 4.10 .....Electrolysis of Steam Low Voltage .................................................................90

PAGE 8

vi Development of Thin CsHSO4 Membrane Electrode Assemblies for Electrolysis and Fuel Cell Applications Lars E. Ecklund-Mitchell ABSTRACT Abstract In this work the use of the solid acid CsHSO4 as an electrolyte in a hydrogen/oxygen fuel cell or the disassociation of water into hydrogen an d oxygen has been investig ated. Several issues have been cited in literatur e regarding the use of CsHSO4 as a solid electrolyte; these include: difficulty interpreting proton conductivity profiles of real membranes, high permeability of the membrane to fuel and product gases, and low mechanical strength. In an attempt to improve our understanding and possibly eliminat e these issues, performance characteristics of prepared CsHSO4 membranes have been investigat ed utilizing various methods of synthesis and membrane fabricati on. A consistent method of CsHSO4 membrane construction was developed based on these i nvestigations. In addition, a novel method of sintering to decrease the membrane’s perm eability to fuel gases was developed and evaluated. The effects of these measures were in vestigated and tested in a prototype cell for proof of concept of fuel cell and electrolysis applications.

PAGE 9

1 1 Introduction The main objective of this work is the investig ation of the viability of solid acid electrolytes, specifically cesium hydrogen sulfate, for use in thin membrane electr ode assemblies (MEAs) for improved performance in fuel cell or electr olysis applications. The formal introduction will describe solid acid electrolytes and how their structure, bonding types, and internal order-disorder determine their proton conductiv ity character. Moving from these general considerations, the structure a nd properties of cesium hydrogen su lfate in particular will be addressed, showing the reasoning behind its selection as the electrolyte to be studied. Considerations for the use of cesium hydrogen sulfate in memb rane electrode assemblies, and finally discussion of overall design cons iderations for electrolysis and fuel cell applications will be addressed.

PAGE 10

2 The overall structure of this work will then present the three major project areas explored. Firstly, the analysis of different synthesi s methods for cesium hydrogen sulfate with the verification and characterization of the resulting materials by a variety of solid state methods will be presented. Second, the effect of pressing and sintering methods on the gas permeability of cesium hydrogen sulfate will be shown and discussed. Finally, the knowledge gained from the above noted inve stigations will be used to fa bricate, characterize, and test prototype membrane electrode a ssemblies for performance metrics. A detailed discussion of results, conclusions and suggestions for future work will be given. 1.1 Solid Acid Electrolytes and CsHSO4 An electrolyte is a medium in which charge transfer is produced by the mobility of ions, charged particles, through that medium. Electrolytes, which may exist as solutions, gases, molten liquids, or solids, do not allow direct electrical conduction. When an electrical potential is applied to the elec trolyte a chemical reaction near the sources of the potential strips or adds electrons to mobile species wi thin the electrolyte. These charged species may then travel through the electrolyte to neut ralize charge build up near the electrodes. Solid electrolytes are electrolytes that allow the transfer of charged ions through their solid structure. Often these materials are polymers or ceramics. Larger ion species are typically constrained within the solid structure by direct bonding, as well as small pore size. Conduction of charge is therefor e accomplished by the transfer of smaller ionic species, often single protons. Solid acids have advantages of being self-supporting, not requiring the greater

PAGE 11

3 support as molten or liquid electrolytes. Also unlike molten or liquid electrolytes, solid electrolytes are not subject to mass transfer of the electrolyte medium to reactant or product gases by evaporation. Finally, with proper pro cess design, a solid electrolyte can form a barrier between reactant and product gases, ei ther creating a pure product gas or a mixture that can be easily separated. Th is can drastically reduce costs associated with separation and purification of process gases1. A solid acid electrolyte ideally displays th e high ionic conductivity of an acid with the advantage of a solid matrix at standard operating temperatures and pressures. Cesium hydrogen sulfate (CsHSO4) has been investigated as a solid acid electrolyte by several authors recently2-6. CsHSO4 belongs to a category of solid acid electrolytes known as MHXO4 type solid acids. The M represents an atom in the alkali group, such as potassium or rubidium, while the X is atom of the chalcoge n group such as sulfur or selenium. Thus, in addition to CsHSO4, potassium hydrogen selenate (KHSeO4) and sodium hydrogen sulfate (NaHSO4) are representatives of this classification. These solid acids are characterized by structures composed of alternat ing cations and oxy-anions, wher e the ratio of the two is 1:1, with the oxy-anions bounded by hydrogen bonding th roughout the structure. The ratio of hydrogen atoms to oxy-anions within the struct ure (1:1), directly impacts the electrical character of these materials depending on the number of n earest neighbor oxy-anions around a given oxy-anion, as shown in the subsections below.

PAGE 12

4 Previous work with CsHSO4 includes study of proton conduc tivity in the superprotonic phase3, 5-9, study of the phase transitions and phase diagrams of CsHSO4 10-16, modeling of the molecular structure of room-tempe rature and superprotonic CsHSO4 17-19, developments in the high-temperature stability of CsHSO4 under a variety of atmospheres20, and the construction of H2/O2 fuel cells2, 4. The result is a body of knowledge that describes the electrical properties, phase transition temperatures and en ergies, and expected structural measurements of CsHSO4. The construction of a working H2/O2 fuel cell using CsHSO4 has been successfully reported2, 4. In addition, certain difficulties, including the formation of consistent membranes, adhesion of electrodes and catalyst layers, cross-barrier diffusion of fuel ga ses, the relative fragility of manufactures membranes, and the stability of membranes under different atmospheres at high temperature, invol ving the use of CsHSO4 have hampered attempts, requiring novel solutions to preparati on and operating conditions. 1.1.1 Structure The structure of solid acids can be descri bed using the principles of atomic bonding, coordination, and order-disorder. These three fact ors describe how the structure of solid acids arises, and also how the properties of solid ac ids, particularly thei r changes in phase and resulting changes on electr olytic behavior, arise.

PAGE 13

5 Solid acids adhere to the general chemical formula: MaHb(XO4)c, where M is a monovalent or divalent cation, XO4 -2 is a tetrahedral oxy-an ion, and a, b, and c are integers. Structurally, solid acids are typically described as two overlapping lattices, the hydrogen-bonded tetrahedral oxy-anions (XO4 -2) making up one lattice, interspersed with a second lattice of the cations (M+) resulting in a charge balance. At low temperatures, CsHSO4 has a monoclinic structure of stacked chains of SO4 -2 and Cs+ ions. Two such phases have been observed18, 19, both monoclinic but w ith different unit cell parameters. These phases are labeled CsHSO4-III for the lower temperature phase, and CsHSO4-II for the higher temperature phase. The tr ansition temperature for these two phases ranges from 69 to 73 C. A m onoclinic structure of CsHSO4 is shown in Figure 1.1.

PAGE 14

6 Figure 1.1 Structure of CsHSO4 at Room Temperature Phase The monoclinic structure of CsHSO4-III or CsHSO4-II, present from room temperature to 140-145 C. At higher temperatures (i n excess of 145 C) CsHSO4 exists as a solid phase tetrahedral structure. The Cs+ and SO4 -2 ions form two interlocking three dimensional tetrahedral matrices with neutral charge distri bution. This phase is labeled CsHSO4-I. A tetrahedral CsHSO4 structure is shown laterally to il lustrate nearest neighbors of the SO4 -2 matrix in Figure 1.2.

PAGE 15

7 Figure 1.2 Structure of CsHSO4 at High Temperature Phase The tetrahedral structure of CsHSO4-I, present at temperatures above 145 C. 1.1.1.1 Bonding21 The structure of solids is dominated by the interactions resulti ng from atomic bonding between constituent atoms in the solid. These bonds are a consequence of electronic forces between negatively charged outer shell electr ons and the positively charged nuclei of the atoms. These bonds are classified into categor ies of electrostatic bonding, between positively and negatively charged ions due to electron transfer, covalent bondi ng, a result of electron

PAGE 16

8 sharing between two atoms, a nd metallic bonding, at traction due to delocalized electrons. These categories do not descri be all possible bonding as observed bonding may exhibit character of more than one type. Of these types of bonds, the electrostatic bond, also known as ionic bonding, best describes the bonding in CsHSO4, like solid acids in general. Elect rons are transferred locally from electropositive cesium to the electronegative species, SO4 -2, creating a charge imbalance resulting in an electros tatic force. This creates a strong ove rall pattern of alternating anion and cation within the structur e of the solid, a three-dimens ional pattern dictated by the electrostatic attraction between positively a nd negatively charged elements resisting the repulsion of like-charged elements to set the si ze and distance between elements within the overall matrix. The other important type of bonding in so lid acids is hydroge n bonding. In hydrogen bonding, the attraction of an atom of hydrogen to two atoms acts as a bond between them. The hydrogen atom acts as a sort of “bridge.” This type of bonding is much weaker than ionic bonding, but still has a la rge effect on the shape and properties of the solid. Hydrogen bonding can be described by its intra-hydrogen b onding, the local geometries of the hydrogen about the two atoms being bonded, and the inter-hydrogen bonding, th e overall spatial geometry of the solid lattice22. In CsHSO4, hydrogen bonding occurs between two oxygen atoms on adjacent SO4 -2 groups. The intra-hydrogen bond geometry between the two oxygen atoms participating in the

PAGE 17

9 bonding, and the hydrogen atom, vary in strength with distance and the covalency of the bonding. At closer distances, the hydrogen bond is st ronger, with a more covalent character. The single valance electron of the hydrogen is sh ared between the two oxygen atoms. As the inter-oxygen atom distance increases, the bond strength decreases and the hydrogen atom becomes more firmly attached to one of the oxygen atoms, creating a charge imbalance across the bond, resulting in a more ionic character. At certain bond lengths, the hydrogen atom can switch positions between two energy wells along the bond, creating hydrogen bond disorder. This is only seen at a discrete range of bond lengths—at longer bond lengths the hydrogen closely attaches to one of the oxygen atom s, and the energy required to jump to the other potential well is greater than the actual bond st rength, while at shorter bond lengths the hydrogen atom is equally shared between the two bonding oxygen atoms. These bonds strengths and corresponding characters are summarized in Table 1-1. Table 1-1 Correlation Between Hydr ogen Bond Types and Bond Character22 Bond Strength Interoxygen distance () Char acter Hydrogen Bond Disorder Strong 2.4 covalent Not observed—hydrogen is shared Medium 2.4 – 2.9 polar covalent Can be observed at higher temperatures Weak 2.9 ionic Not observed—bond is too weak Because the ratio of hydrogen bonds to SO4 -2 groups is one to one, traditional predictions would describe the structure of CsHSO4 as cyclic dimmers, rings, or chains. This is because of the generally accepted rule relating the ratio of hydrogen to tetrahedral oxy-anions. In this regime, the dimensionality of the network is related to the density of hydrogen bonds by:

PAGE 18

10 1 22 4 XO H D Where D is the dimensionality of the lattice, (H+/XO4 -2) is the ratio of hydrogen atoms (or hydrogen bonds) to oxy-anion tetr ahedrals, in this case SO4 -2 groups. As we will see, this accurately describes CsHSO4 in the lower temperature phase but other factors lead to a different structure in th e higher temperature phase. 1.1.1.2 Order-Disorder21 The discussion of order-disorder concerns th e effects on the structure and properties of a solid when a species partially occupies more th an one position within the solid structure. The given species will be energeti cally stable in more than one position, and will oscillate between the two positions. A hierar chal system exists for the clas sification of disorder within a solid lattice, with th e two main types being structural disorder, which looks at the multiple positions that can be occupied by a single spec ies, and chemical disorder, involving a single position but multiple species. Structural disorder is further classified as static disorder, in which the basic structure of the solid include s multiple positions that a given species may occupy, with a randomized distribution based on local factors, and dynamic disorder, in which the disorder of the species arises as it is energetically excited into a state in which freely can move between multiple positions. The superprotonic conduction ability of CsHSO4 at higher temperatures is due to the dynamic disorder created as the la ttice is thermally excited. The disorder created is expressed

PAGE 19

11 throughout the crystal lattice, and is described using two cl assifications based on the two species that are involved. Inter-hydrogen bond disorder is disorder created as the structure of CsHSO4 changes upon heating. Multiple positions for hydrogen bonding between SO4 -2 anions become available as the crystal lattice changes from a monoclinic st ructure at low temperature to a tetrahedral structure at high temperature. This phase transition increases the number of nearest neighbor SO4 -2 anions around any given SO4 -2 anion from two to four, and the ratio of possible interhydrogen bonds to SO4 -2 anions from one to two. This al lows for the transition of hydrogen bonds between the available cr ystallographic positi ons, dramatically changing the proton conduction character of the solid. Strongly related to the change in inter-hydrog en bond disorder is an increase in oxy-anion disorder. The orientation of the SO4 -2 anion within its location in the crystal lattice is subject to constraints based on hydrogen bonding and va n der Waals forces from neighboring ions. After the transition to the higher temperature tetrahedral phase, the constraints of hydrogen bonding are relaxed due to the increase in inte r-hydrogen bond disorder. At the same time, van der Waals forces from neighboring SO4 -2 anions and cesium cations is normalized from a linear mode in the monoclinic, to a more diffuse three-dimensional mode in the tetrahedral. Certain orientations of the SO4 -2 anion are still preferred, but th e energy barriers that separate them are much lower, allowing easy, thermally ex cited transitions between them. As a result, the SO4 -2 anion may freely rotate about its loca tion in the crystal lattice after CsHSO4 takes a tetrahedral structure.

PAGE 20

12 1.1.2 Proton Conductivity Proton conduction through solid ac ids is several orders of magn itude higher than that of electron conduction. This allows solid acids to be used in ap plications as proton transport membranes while also acting as electrical insulato rs. This feature is a defining behavior for a successful electrolyte in a me mbrane electrode assembly. Table 1-2 shows typical valu es for solid acid electri cal and proton conductivities. Table 1-2 Comparison of El ectronic and Ionic Conductivity23 Electronic Conductivity Ionic Conductivity Conductivity range 10 S/cm < < 105 S/cm Conductivity range 10-3 S/cm < < 10 S/cm Electrons carry current Ions carry current Conductivity increases as temperature decreases Conductivity decreases as temperature decreases Solid acids display five general methods fo r proton transfer: atom ic diffusion, protondisplacement, molecular reorientation, vehi cle mechanism, and the Grotthus mechanism4. The mechanism which dominates in that solid acid is determined by the acid’s structure, electron density, the diffusivity of larger atom ic species through the solid lattice, and the rotational frequency of cation groups. Atomic diffusion is only possible in those materials where the proton can share electron density with the host material. The proton then diffuses through the lattice coupled with the

PAGE 21

13 diffusion of electrons. These materials have high electron mobilit y, and are therefore electrically conductive. Thus they make poor electrolytes. Proton displacement is triggered when protons become thermally excited to energy levels above the height of the surrounding potential we ll they typically occ upy within the lattice. This creates ferroelectric behavior in solid acids as electric forces cha nge faster then elastic forces can compensate, creating a shift in ion position and a dipole moment within the structure. This effect shows hysterisis behavior when the applied heat is removed. Only a few solid acids show this behavior, and it is genera lly at too low of a le vel for an electrolyte application. Molecular reorientation, the ve hicle mechanism, and the Grott hus mechanism all involve the proton attaching to a larger molecular or ioni c species. The motion of this larger species within the solid lattice pr ovides the primary motive transport for the proton. These mechanisms have the advantages of being inde pendent of the electric al conductivity of the material while still allowing for fast proton transport. In molecular reorientation the proton is tr ansported by a molecular species undergoing a rotation within the solid lattice. This allows the proton to “jump” between multiple locations within the crystal structure. The rate of proton transfer is then limited by the ability of protons to attach to whole mo lecular species, as well as the dynamics of the molecular species.

PAGE 22

14 The vehicle mechanism involves the proton bonding with a molecu lar or ionic species with the capability to diffuse throughout the solid. Thes e larger species act as carriers, while their diffusivity within the lattice determines the rate of proton transfer. To maintain equilibrium within the solid, empty carriers must move c ounter to the direction of proton transfer, forming a kind of “conveyor belt” of molecular or ionic species carry ing the protons through the solid. The Grotthus mechanism combines features of proton displacement and molecular reorientation. Unlike molecular reorientati on, the proton bonds with an oxy-anion whose rotation within the matrix transports the prot on between two crystallig raphically equivalent positions. From here the proton undergoes a displacement along the hydrogen bond with the former oxy-anion to a closer position to a s econd oxy-anion. This second oxy-anion then undergoes a similar rotation and the proton moves to another position within the crystal. In this way the proton “hops” from rotating oxy-an ion to oxy-anion. This requires that the oxyanions possess a high degree of dynamic mobility within the lattice, but it has been determined that the proton displacement be tween oxy-anion occurs at a frequency three orders of magnitude less then the rotational rearra ngement of oxy-anions, making it the rate limiting step. The Grotthus mechanism is illustrated in Figure 1.3 below.

PAGE 23

15 Figure 1.3 The Mechanics of Proton Conduction in CsHSO4 The Grotthus mechanism of proton conduction in CsHSO4. A proton transfers along hydrogen bonding between two adjacent SO4 -2 groups (B). The SO4 -2 group rotates bringing the attached proton into pr oximity with a different SO4 -2 group (C). The proton then transfers along the newly formed hydrogen bond to the next SO4 -2 group (D). 1.1.3 Phase Transitions Because CsHSO4 undergoes such a remarkable change in proton conductivity when transitioning from the lower temperature to th e higher temperature phase, the mechanics of this phase change are of interest and will be briefly discussed here.

PAGE 24

16 The phase transition of CsHSO4 from monoclinic CsHSO4-II to tetrahedral CsHSO4-I is a first order solid-solid transition. Since this is a first order ch ange, the changes in extensive thermodynamic variables are negligible compar ed to the discontinuous change in entropy. This allows the use of Differential Sca nning Calorimetry (DSC) to characterize the thermodynamics of the transition. Although this transition is a solid-solid transi tion, the change in physic al properties between the two phases is similar to that of a solidliquid transition. Specifica lly, the increased orderdisorder in the tetrahedral SO4 -2 ions leads to a malleable superprotonic phase, variously described as being similar to “clay or plastic ine.” The room temperature phase, on the other hand, is a solid crystalline stru cture, brittle and prone to fr acture in the aggregate. Figure 1.4 below, shows the typical pr oton conductivity behavior of CsHSO4 with increasing temperature and phase change.

PAGE 25

17 Figure 1.4 CsHSO4 Proton Conductivity from Boysen4 Typical proton conductivity behavior of CsHSO4. Note linear (near-linear on this plot) behavior within a given phase, contrasted with the sharp rise in proton conductivity associated with the superprotonic transition. 1.2 Applications Solid acid electrolytes have been investigat ed for their use in traditional electrolyte applications, namely, as fuel cell electrolytes and for active electrolysis. These two possible applications for solid acids are discussed in mo re detail below, considering currently used electrolyte materials.

PAGE 26

18 1.2.1 Fuel Cells Fuel cells are devices that conve rt a chemical change into an electric current by harnessing the transfer of ionic motion and the resulting el ectric potential. Fuel streams react to produce ions, which must move through an electrolyte to complete the chemical reaction. This motion through the electrolyte creates a charge imbalance between the two sides of the electrolyte barrier, which creates an electric current. In this way, a fuel cell resembles an intermediary device between a battery, which produces electri c potential due to inte rnal chemical reaction, and a combustion engine, which uses the energy released by the chemical reaction of fuel streams24. Many types of fuel cells have been put to use in a variety of applications. Solid acids in general, and CsHSO4 in particular, are used in a type known as proton exchange membrane (PEM) fuel cells. The acronym PEM is occasi onally interpreted as “Polymer Electrolyte Membrane” given the widespread use of commer cial polymer electrolytes such as Nafion in commercial fuel cells. Fuel cells of this type are also often grouped as part of Ion Exchange Membrane (IEM) fuel cells, a lthough that group also includes those whose electrolyte barriers can tr ansfer larger ion complexes than just protons. A vital component in a fuel cell is the electrolyte barrier that separates th e initial reactants. In a fuel setup, the electrolyte fo rms part of a combined electr ode/catalyst/elect rolyte complex commonly called a membrane electrode assembly (MEA). The construction of MEAs results in a complex interdependent series of considerations including the chemical compatibility

PAGE 27

19 and kinetics of fuel streams, electrode materials, catalyst ma terials and deposition techniques, and the electrolyte itself25, 26. The internal mechanism of a PEM fuel cell is best described by using an example. Consider a cell for the reaction of hydrogen and oxygen to produce water. The overall reaction for this fuel cell is: O H O H2 2 22 2 The reaction at the anode is the dissociation of hydrogen to adsorbed hydrogen atoms and electrons. e H H 4 4 22 The electrons follow an external path to the cathode, creating an elec tric current through a load. The hydrogen atoms migrate through th e membrane to the cathode, where they recombine with the electrons and adsorbed oxygen to produce water. O H O e H2 22 4 4 A schematic of a hydrogen-oxygen fuel cell is shown in Figure 1.5.

PAGE 28

20 H2H2H-HH-H H+H+H+H+O2O-O H2O H2O Anode Cathode Electrolyte 4e 4e Catalyst 4e Catalyst Fuel Oxidant Load 2H2 4H++ 4eO2+ 4H++ 4e2H2O Figure 1.5 Diagram of a Hydrogen Fuel Cell Hydrogen fuel enters and reacts in the anode chamber, resulting in H+ ions that migrate through the electrolyte, where they react wi th oxygen in the cathode chamber to produce water. A current of electrons is produced through a load connecti ng the two electrodes. The total Gibbs free energy of th e reaction in the cell is the sum of these two half reactions: e H H 4 4 22 mol kJ Gf0 0 O H O e H2 22 4 4 mol kJ Gf6 482 O H O H2 2 22 2 mol kJ Grxn6 482 The total Gibbs free energy of the fuel cell reaction is related to the theoretical open cell voltage by the Nernst equation:

PAGE 29

21 0nF Grxn where Grxn is the change in standard Gibbs free en ergy for the reaction, n is the number of charge carriers, F is Faraday’s constant, and 0 is the theoretical open cell voltage. Open cell voltage at operating conditions, howev er, will be dictated by the temperature of reaction and the chemical activities of the r eactants and products. The Nernst equation is expanded, to show these terms, thus: q s revP R nF RT10 0log 303 2 Where rev is the theoretical reversible cell potenti al at the operating conditions, T is the absolute temperature, [R] and [P] are the ch emical activities of th e reactant and product, respectively, and q and s are the coefficients of reaction for the reactant and product, respectively. The theoretical reve rsible cell potential accounts fo r the loss of electric potential due to the thermodynamic losses associated with reaction. A maximum efficiency (Emax) of an electrochemical cell can th en be defined as the ratio of the change in Gibb’s free ener gy at the reaction temperature ( GT), and the ideal change in enthalpy for the reaction ( H0). Since the change in Gibb’s free energy for an electrochemical cell nearly re duces to the electrical work, this efficiency can be

PAGE 30

22 conceptualized as the maximum amount of elect rical work that can be produced out of the total chemical energy rel eased in the reaction. 0 maxH G ET An additional definition of efficiency is the voltage efficiency, the ratio of the observed voltage with the theoretical reversible cell poten tial. This efficiency s hows losses that are not due to thermodynamic constraints because of operating conditions, but rather due to losses from the fuel cell and MEA design, or from polarization effects. rev cell vE Finally, an overall efficiency can be define d as the product of the maximum and voltage efficiencies, showing the ratio of observed cell potential to the theoretical open cell voltage, combining losses due to thermodyna mic and polarization effects. v overallE E E *max

PAGE 31

23 Polarization effects in a fuel ce ll are dependent on several fact ors relating to the construction techniques of the fuel cell apparatus itself, the membrane electrode assembly, and the choice of electrolyte and operating c onditions. The primary losses of power from a fuel cell are electrolyte conductivity, catalytic activity, gas transport to and from the surface of the electrolyte, and gas tran sport across the electrolyte barrier without reaction27. Loss of power due to electrol yte conductivity is termed Ohmic polarization and is due to the resistance to proton transport of the electrolyte. At operating conditions, the electrolyte will behave as a proton resistor of a given resistance. At low curre nts, the amount of loss from Ohmic polarization is zero, but as current increases, measured voltage between electrodes decreases linearly, as transported protons th rough the electrolyte begin to equalize the electric potential. The slope of this relationshi p is equal to the measured resistance of an equivalent resistor from pr oton impedance measurements of the electrolyte. As proton conductivity of a given el ectrolyte is increased, the slope of this line decreases and higher measured potentials between the electrodes continue to persist at higher currents, resulting in higher power (P = x I). Loss of power due to catalytic activity is termed activation polarization and is a result of the surface kinetics at the anode or cathode. A poor choice in catalyst, or a poisoned catalyst, at either electrode, can result in severe activati on polarization. Loss of cell voltage decreases quickly with increasing current at low current values, as reduced reaction sites or poor electrolyte/catalyst compatibility increases th e initial energy required for intermediate surface reactions. The losses due to activation polarizati on cause an initial drop in cell potential as

PAGE 32

24 current equilibriums within the immediate area of the electrodes are disrupted. As current increases, the relative change in current with voltage decreases, making these deviations from equilibrium less severe. However, initial loss of cell potential due to activation polarization can significantly reduce the overall potentia l of the cell as curr ent is increased. Gas transport to and from the el ectrode reaction areas can cause concentration polarization at higher currents. This can either be an issue of the fuel gas being consum ed too quickly at the surface of the anode, setting up a concentration gradient between the bulk fuel gas in the anode compartment through the porous electrode an issue of oxidant undergoing a similar concentration gradient in the cathode compartment due to th e speed of reaction, or the product gas not diffusing quickly enough from th e cathode surface to the bulk gas, setting up a concentration gradient in that direction. Concentration losses are primarily due to the latter situation, and are the result of slow gas diffusion of product away from the cathode. Since concentration polarization is due to quickly proceeding reactions, it is not seen until higher current.

PAGE 33

25 Diffusion of fuel or oxidant gases across th e electrolyte barrier can cause loss of cell potential or measured current. If the gases are diffusing wit hout reaction, then less current will be observed simply because of lower convers ion rates of fuel (imagining the fuel cell as a chemical reactor). If fuel or oxidant gase s are penetrating a sufficient distance into the electrolyte and then undergoing reac tion, this creates a larger zone of reaction with its local potential gradient than expected. This lowers m easured cell potential at the electrodes, as the local potential gradient about the electrode actually extends into the electrolyte beyond the point at which it is obs erved by the electrode. Figure 1.6 illustrates a typical current versus electric po tential curve of a fuel cell (polarization curve), along with the effects of losses and peak power.

PAGE 34

26 CurrentElectric PotentialPower A ctivation Polarization Ohmic Polarization Concentration Polarization Figure 1.6 Polarization Losses in a Hydrogen Fuel Cell A typical polarization curve for a fuel cell show ing electric potential (solid line) and power (dotted line) versus current. Ac tivation, Ohmic, and concentration polarization lowers the electric potential output by the cell with increasi ng current. All axis are in arbitrary units.

PAGE 35

27 1.2.2 Electrolysis Electrolysis, especially in the case of electrolysis performed w ith a solid electrolyte, can be thought of as a fuel cell operation in reverse. An electric current is used to create a chemical reaction. In a solid elec trolyte setup, the reaction occurs at electrodes just as in a fuel cell, and results in ionic transpor t between the electro des through the solid electrolyte. Thus, instead of inserting fuel and oxi dant into the unit process and producing a current as output, a current is added to produce the reverse reaction. Water can be electrolyzed as the opposite reaction to that in a H2/O2 fuel cell: 2 2 22 1 O H O H This reaction is usually carried out with water in the gas phase. Many of the same considerations with regards to power loss in a fuel cell have an analogue to loss of efficiency in an electrolysis operation. As voltage between the electrodes increases, a minimum electric potential is reached at which the electrolysis proce ss begins. In an ideal case, this voltage is equal to the ideal open cell voltage of a fuel cell operating with the reverse reaction. As the voltage increases, current passes through the circuit connecting the electrodes as ionic species move through the elec trolyte, completing the reaction. As voltage is increased, activation losses due to ionic species in the electrolyte resisting the formation of further species at the electrode similar to that seen in a fuel cell are observed. For similar

PAGE 36

28 reasons as explained in the fuel cell secti on above, Ohmic losses due to the conductivity of the electrolyte to ionic species, and concentration losses due to mass transfer to and from the electrodes from the bulk of the in let gases are also observed. These losses combine to shape the current versus voltage response of the el ectrolysis process. A maximum current is eventually reached at total conversion.

PAGE 37

29 2 Development of a Solid Acid Electrolyte The design and construction of the membrane el ectrode assembly is the primary objective of this work. To optimize the different steps in th e construction process, a variety of techniques were employed. The individual steps taken and the methods employed to achieve those steps, are detailed in this section. CsHSO4 powders created by different variations of the synthesis methods were analyzed using thermal analysis (inc luding differential scanning cal orimetry, DSC, and thermogravimetric analysis, TGA), impedance spec troscopy, X-ray photoe lectron spectroscopy (XPS), and powder X-ray diffraction (PXRD). Those results appear in the following subsections. 2.1 Synthesis of the Electrolyte A standardized method for producing CsHSO4 membranes was one of the primary goals of this work. To this end, various methods of chemically synthesizing CsHSO4, crystallizing it out of solution, and pressing the resulting crystals into a membrane were explored.

PAGE 38

30 2.1.1 Chemical Synthesis Several methods of synthesis we re employed for producing CsHSO4. The first method reacted cesium sulfate, Cs2SO4, with sulfuric acid as shown below: 4 4 2 4 22CsHSO SO H SO Cs An excess of sulfuric acid was used to insure the cesium sulfate progr essed all the way to CsHSO4. A second method reacted cesium carbonate, Cs2CO3, with sulfuric acid by the following reaction: 2 2 4 4 2 3 22 2 CO O H CsHSO SO H CO Cs Once again, an excess of sulfuric acid was used to insure complete reaction. The second reaction was eventually abandoned in favor of the first. The evolution of CO2 created splatters of heated sulfuric acid solution. In addition to this safety concern, as well as questions as to the extent of the reaction, lead to this decision.

PAGE 39

31 2.1.2 Crystallization CsHSO4 was then crystallized out of solution us ing one of two methods. The first method used the slow evaporation of solution in an oven at 60 C to produce a highly concentrated solution. The solution container was then rapidl y cooled to 10 C in a cold water bath, causing large crystals of CsHSO4 to appear. The crystals were then quickly placed in a vacuum filter. The second crystallization method took adva ntage of the insolubility of CsHSO4 in most organic solvents. A large amount of meth anol or acetone was added to the CsHSO4 solution, typically resulting in 70-90 % methanol or ac etone by volume. Small crystals of CsHSO4 quickly formed as the concentration of metha nol or acetone increased. These crystals were vacuum filtered, washing with methanol acet one. Crystals formed by acetone crystallization appeared quickly, as finely structured, fluffy crystals, floating at the interface between the water and acetone phases. When methanol was used, crystals formed in clumps at the bottom of the container, as the methanol di ssolved into water creating a solution. The second method of crystallization proved faster and easier to perform, and resulted in a greater percentage of the CsHSO4 crystallizing out of solution. Also, the CsHSO4 did not dissolve back into solution during the filtering pr ocess, which occurred in the first method as the samples warmed back up to room temperature. Finally, the crystals produced in the second method formed faster and were smaller th en the crystals formed by the first method. This reduced entrapment of solution within th e crystal structure, ma king drying easier. Later

PAGE 40

32 DSC/TGA analysis of the crystals showed irregu larities for the crystals that were grown with the first method. Those crystals formed with methanol instead of acetone were more difficult to dry completely, often requiring several cycl es of drying in an oven and grinding with mortar and pestle before satis factory material was obtained. For these reasons, the use of acetone to crystallize CsHSO4 out of solution was adopted as standard practice in this work. After crystals were produced, they were drie d in an oven at 60 C for 10 to 12 hours. The powder was taken out and ground with a mortar a nd pestle at several points during the drying process. This was continued until the powder presented a uniform, dry appearance with minimal clumping and discoloration. The powder was then stored in a vacuum desiccator until use. 2.1.3 Pressing Methods Powder was uniaxially compressed in a Intern ational Crystals Labor atory E-Z Press 12-ton hydraulic press, to achieve 50 MPa of pr essure its surface as described by Boysen4. The pellet sizes were ” in diameter, so the press was operated with 7.5 tons of force on the ram to obtain the desired pressure. The force was ap plied for ten to twenty minutes before the resulting pellet was removed. When 0.41 g of dry material was pressed, the resulting pellet had a density of 3.26 g/cm3, and a thickness of 1 mm. Since CsHSO4 becomes a more malleable solid above the superprotonic transition temperature, these pressing conditions were duplicated on a Carver 4386 heated hydraulic

PAGE 41

33 press. It was discovered that th e electrolyte stuck to stainless steel anvils after the heated pressing, making the removal of the pellet without damage extremely difficult. Using tungsten carbide anvils from International Crystal Laboratories w ith optically polished surfaces alleviated these problems. Membranes produced with these methods were ex amined using optical microscopy. A typical micrograph of membranes pressed at room temp erature and at heightened temperature is shown in Figure 2.1.

PAGE 42

34 Figure 2.1 CsHSO4 Membranes Pressed Under Ambien t and Heightened Temperature x100 CsHSO4 membranes pressed under ambient temper ature conditions (top) versus those pressed under heightened temperature (bottom) Grain and flake size are larger, and better defined in the sample pressed at ambient temperature.

PAGE 43

35 2.2 Impedance Spectroscopy Impedance spectroscopy measures the proton con ductivity of a sample against an intrinsic variable such as temperature. This gives a picture of the change in proton conductivity associated with a phase change in CsHSO4. This section will contain a di scussion of the importance of impedance spectroscopy, and what we hope to accomplish by using it. 2.2.1 Proton Conductivity Proton conductivity of the electr olyte is the primary indicator of the performance of the electrolyte in a fuel cell or elec trolysis process. This is because the transfer of proton across the width of the electrolyte is the pr imary limiting step of the reaction. The proton conductivity in solids can be descri bed with the same models used for ionic conductivity. What follows is a brief description of ionic co nductivity, and its dependence on temperature28. The conductivity ( ) of charge carriers in an isotropic solid can be described as the product of the concentration of charged ca rriers (n), the charge per carrier (q), and the charge carrier mobility (u) within the solid. nqu

PAGE 44

36 A form of the Nernst-Einstein equation relate s charge carrier mobility to the charge per carrier, the diffusion coefficient from Fick’s first law describi ng the flux of particles through the solid (D) and the absolute temperature (T). T k qD uB The diffusion coefficient can be described usi ng a random walk model as the product of the frequency of charge carrier transfer ( ), the square of the dist ance between crystallographic sites (a0) and a geometric factor depending on the structure of the solid ( ). 2 0a D As the frequency of charge carrier transfer is a thermally activated process, it is best described as an Arrhenius-t ype temperature dependence. T k GB aexp0 Where 0 is the attempt frequency of charge carrier transfer and Ga is the Gibbs free energy for activation of charge carrier transfer. Simplifying these equations we can arrive at an expression for the temperature dependence of conductivity within the solid:

PAGE 45

37 T k V P E A T k H A TB a a B aexp exp0 0 The pre-exponential factor, A0, can be expressed in terms of constants associated with the charge carrier transfer. B a Bk S k nq a A exp2 0 2 0 0 with Sa equal to the change in en tropy for activation of charge carrier transfer. From these expressions, we can see that a plot of the log T versus 1/T should produce a straight line, with the slope equal to the negative of the enthalpy for activation of the charge carrier transfer divided by the Boltzmann constant. This is valid as long as th e crystal structure of the solid and its appropriate a0, and 0 values remain constant. The conductivity of the sample is defined as the real admittance (reci procal of the real resistance) divided by a geometric factor. In this case, the area of the sample exposed to the probe divided by the thic kness of the membrane. l A R / / 1 It should be noted that with th e conductivity of the material and cross sectional area of the pellet fixed, the real resist ance is directly proportional to the thickness of the pellet.

PAGE 46

38 2.2.2 Experimental Measurements Impedance spectroscopy allows for the analysis of the real resistance and capacitance of a solid which might otherwise resist direct curr ent measurements. The primary charge mobility in CsHSO4 is carried by proton transfer, not elec tron mobility as in a typical conductor. Under direct current with out a proton source, CsHSO4 acts as an insulator, not yielding useful data. Impedance spectroscopy, however, allows the use of alternating current to determine the proton conductivity of the sa mple. Because there is no overall flow of electrons through the sample, th e frequency and amplitude of the applied electric field become translated into the alternate movements of protons throughout the crystal lattice. This creates a measurable alternating current. Typical impedance measurements employ a four probe technique applied to one surface of the sample. Either a linear or square configur ation of probes is typically used, with a high current pass and a low current pass, allowing el ectrode potentials to be canceled out in the final data analysis. Samples of CsHSO4 were analyzed with impedance sp ectroscopy using an Agilent 4284A Precision LCR Meter in a twopoint configuration. Half-i nch diameter membranes of CsHSO4 were manufactured, and then placed be tween two copper strip electrodes. An alternating electric potential at 1 MHz and one volt was imposed acr oss the width of the membrane. The temperature of the sample was regulated using Linberg/BlueM MO1420SA1 Mechanical Oven from 70 to 180 C. Proton conductivity was calculated from the resulting resistance measurements.

PAGE 47

39 Figure 2.2 shows proton conductivity versus temp erature for three preparation methods used. Sample A is CsHSO4 crystallized by acetone, sample B was crystallized by temperature variation, and sample C was cr ystallized by acetone, but presse d at heightened temperature, at 150 C. -7 -6 -5 -4 -3 -2 -1 0 70 80 90 100 110 120 130 140 150 160 170 A B C Temperature (C) Figure 2.2 Impedance Spectra of CsHSO4 Impedance spectra of CsHSO4 under different preparation conditions from ferroelectric CsHSO4-II phase (low temperature) to superprotonic CsHSO4-I phase (high temperature). Samples shown are CsHSO4 crystallized from solution by a cetone (A), crystallized from solution by temperature variation (B), and cr ystallized by acetone, then pressed at high temperature (150 C) (C). Figure 2.3 shows just the high temperature, superprotonic area of the above figure.

PAGE 48

40 -3 -2.5 -2 -1.5 -1 140 145 150 155 160 165 170 A B C Temperature (C) Figure 2.3 High Temperature Impedance Spectra of CsHSO4 Impedance spectra of CsHSO4 showing superprotonic CsHSO4-I phase (high temperature) only. Samples shown are CsHSO4 crystallized from solution by acetone (A), crystallized from solution by temperature variation (B), a nd crystallized by acetone, then pressed at high temperature (150 C) (C). Proton conductivities at high te mperature, activation enthal pies for conduction, and preexponential factors all showed little variation based on synthe sis and pressing method. These values are summarized in Table 2-1 below. Table 2-1 Impedance Spectroscopy Data for CsHSO4 Sample Superprotonic Transition Temperature Range (C) Proton Conductivity at 150 C (S/cm) Activation Enthalpy for Conduction (eV) Pre-exponential Factor (SK/cm) A B C 138 to 145 139 to 146 137 to 143 0.0106 0.0094 0.0126 0.227 0.239 0.229 2350 2900 2840

PAGE 49

41 2.3 Thermal Analysis To further test their high temper ature properties and verify thei r identity, electrolyte samples were examined using differential scanning calorimetry (DSC) and thermal gravimetric analysis (TGA). This data was used to determine the temperature at which the CsHSO4 underwent transition into transition into the superprotonic phase, as well as the melting temperature. DSC/TGA data also identified de hydration activity at hi gher temperatures, and was used to test dehydration/deliques cence activity under a humid gas stream. DSC uses measurements of the heat loss or ga in of the sample chamber as the sample as heated. The heat into the chamber is contro lled so that a constant ramp rate of the temperature is maintained. If the sample unde rgoes an endothermic transition, such as a phase change, this appears as a peak on the DS C spectrum. The location of that peak serves as a marker for the temperature at which that phase change occurs. In addition, by integrating beneath the peak the standard enthalpy of that transition can be determined. Ramp rates change the accuracy of these measurements, w ith slower ramp rates yielding more accurate results. TGA measures the mass of the sample as the temperature is increased. A phase change from solid to superprotonic solid, or from solid to liquid, should show no change in a TGA spectrum. A change from liquid to gas, or deliquescence of the sample, should appear as decreases or increases in the TGA plot respectively.

PAGE 50

42 For this work, a TA Instruments SDT Q-600 simultaneous DSC/TGA was used to analyze the samples. Powder samples of 15 to 25 mg were placed in ceramic pans and ramped from 50 to 250 C at a ramp rate of 5 C/min. For studies of dehydration and deliquescence at high temperature, samples were held for longer pe riods at temperatures ranging from 170 to 200 C, and their changes in mass were recorded under dry air streams. A DSC/TGA of a sample of CsHSO4, crystallized by acetone, is shown in Figure 2.4 below. Figure 2.4 CsHSO4 Crystallized by Acetone DSC/TGA of CsHSO4 crystallized by exposure to acetone. Solid line equals Weight % versus Temperature, dashed line equals Heat Fl ow versus Temperature. First-order solid to solid phase change at 140.66 C, evolution of gas (H2O) from 162.12 C to 195.90 C. This figure shows a sharp phase endothermic transition at 140.66 C, corresponding with the transition from the room temperature to the superprotonic phase. After transition to the

PAGE 51

43 superprotonic phase, there is a reduction in mass, also an endothermic process. This is likely the evolution of water, as the e xposed sample dehydrates from CsHSO4 to Cs2S2O7. This dehydration continues until a second transition occurs at 195.90 C. This same experiment was also run with CsHSO4 that was crystallized by evaporation of the solvent, followed by slow chilling. The DS C/TGA profile is show in Figure 2.5. Figure 2.5 CsHSO4 Crystallized by Temperature Control DSC/TGA of CsHSO4 crystallized by temperature cont rol. Solid line indicates Weight % versus Temperature, dashed line indicat es Heat Flow versus Temperature.

PAGE 52

44 A sharp transition, possibly corresponding to the phase ch ange from monoclinic to superprotonic phases, is stil l visible at 137.88 C. However, the DSC profile (heat flow versus temperature) becomes hi ghly erratic after the transition. There is still a reduction in mass, but it is a less smooth profile. The low te mperature solid-solid transition is visible on this profile at 70.18 C. To verify the result show in Fi gure 2.5, other samples of CsHSO4 that were crystallized by temperature control were examined using DSC/T GA. The results were all similar, with an erratic DSC profile after the superprotonic tran sition. Another example of this profile is shown in Figure 2.6. Figure 2.6 CSHSO4 Crystallized by Temperature Control, a Second Attempt DSC/TGA of CsHSO4 crystallized by temperature contro l. The solid line indicates Weight % versus Temperature, dashed line Heat Flow versus Temperature.

PAGE 53

45 The reduction in mass is still evident, but is follows a rougher pr ofile. The DSC profile becomes erratic after the superprotonic transi tion. The temperature at which the transition occurs is slightly higher in this run, at 138. 60 C, but still lower than what was observed with samples that had been crystallized using me thanol or acetone. The exact temperature at which the transition occurred varied in samples that were crystallized by temperature control by 2 C. To mitigate the effects of dehydration of the sample, CsHSO4 was also analyzed using a TA Instruments Q-10 DSC, with a sealable cup a nd controlled atmosphere sample area. Only a limited number of runs could be performed with this instrument, because of concerns of H2SO4 in the sample damaging the equipment. This tool is not equipped with TGA capabilities, further limiting the direct useful ness in tracking moisture losses. A sample DSC profile taken with CsHSO4 crystallized by acetone is shown in Figure 2.7.

PAGE 54

46 Figure 2.7 DSC of CsHSO4 Using Closed Cup Q10 DSC DSC of CsHSO4 using TA Instruments Q-10 DSC with a closed cup to prevent dehydration effects. The low temperature solid-solid transition is visible from th e two monoclinic phases at 71.37 C, as well as the superprotonic transiti on at 144.55 C. The peak observed for the dehydration of water is not visible, most likely an indicator that water is not evolved due to the closed cup. An erratic peak is shown at 208.98 C. This could be due either to the rapid formation of a liquid phase, or, more likely, th e pressure buildup within the closed cup of evolved water finding releas e, thus allowing the rapi d evolution of water. To verify that the evolution of water from the sample was producing the effects observed in the superprotonic phase, seve ral attempts were made to produce DSC profiles of Cs2S2O7. The first of these was excessive drying of CsHSO4 in an oven at 170 C, followed immediately by a DSC run. The resulti ng profile is shown in Figure 2.8.

PAGE 55

47 Figure 2.8 DSC of CsHSO4 after Excessive Drying DSC of CsHSO4 after excessively dried at 170 C to determine the formation of Cs2S2O7. This profile still shows elements of the CsHSO4 profile, leading to suspicions that the Cs2S2O7 may have begun rehydrating back to CsHSO4 while being transferred from the oven to the closed cup, reabsorbing moisture in the lab. The superprotonic tr ansition temperature is shifted higher to 155.89 C, and is smaller relative to the sample size, with 13.40 J/g versus 14.89 J/g in the non-dried sample. This shift coul d be due to a proportion of the sample still being Cs2S2O7 when the DSC was run. The higher temp erature peak is also visible on this profile, but is shifted higher a nd is less erratic. If the sample contains a greater proportion of Cs2S2O7, this can be explained as less available water to evolve from the sample, delaying the breaching of the closed cup a nd the rapid evolution of moisture.

PAGE 56

48 Another attempt to observe the DSC profile for Cs2S2O7 was made using the Q-600 DSC/TGA. A sample of CsHSO4 was subjected to the standard ramp of temperature to determine its DSC/TGA profile. It was then qu ickly cooled and subjec ted to the rise in temperature without removing it fr om the tool. The resulting prof ile is shown in Figure 2.9. Figure 2.9 DSC/TGA of CsHSO4 Rerun DSC/TGA of CsHSO4 run immediately after previous run, but without removing sample from instrument.

PAGE 57

49 There is no serious weight loss associated with the evolution of water in this sample. The weight of the sample varies w ith 0.02 % over the length of the run, likely due to vibrations in the sample arms because of the rapid cooling and heating. Two peaks are still visible, as in the sample of the vigorously dried sample (F igure 2.8), but are at lower temperatures. A steady, though slight, drop in the weight of the sample associated with the higher temperature suggests that it is still the evolution of water. Transition temperatures, heats of transitions and mass losses associated with the high temperature phase are summarized in Table 2-2 below.

PAGE 58

50 Table 2-2 DSC/TGA Data for CsHSO4 High Temperature Transition Heat (J/g) 102.3 Not Observed Not Observed 6.914 Temperature (C) 192.14 226.28 184.02 Mass Loss After SP Transition (%) 2.7 4.1 4.5 Not Measured Not Measured Superprotonic Transition Heat (J/g) 22.66 Unable to Determine 14.89 13.40 16.71 Temperature (C) 140.66 137.88 140.27 144.55 155.89 137.13 Low Temperature Transition Heat (J/g) Not Observed 2.998 1.564 1.635 Not Observed Not Observed Temper ature (C) 70.18 73.42 71.37 Sample Type Crystallized by Acetone Crystallized by Temperature Control Crystallized by Acetone, Closed Cup Excessive Drying Run Immediately After Previous Run

PAGE 59

51 Several long-term runs were made with CsHSO4 in the TGA to determine the average amount of weight loss and the rate at which it is lost. Th ese are shown in Figure 2.10. Figure 2.10 Stability of CsHSO4 Weight losses of CsHSO4 kept at constant temperatures to show relative degree of dehydration. The weight loss, time to steady state, and maximu m rate of weight loss associated with these four temperatures are summarized in Table 2-3 below. Table 2-3 Data for Thermal Stability of CsHSO4 Temperature (C) Weight Loss (%) Time at Weight Loss Stability (minutes) Maximum Rate of Weight Loss (%/minute) 150 2.81 450 0.013 170 2.88 420 0.011 180 3.96 410 0.013 190 4.31 350 0.027

PAGE 60

52 The weight loss of all the samples was 3 to 4 %. This corresponds to the complete dissociation of CsHSO4 into Cs2S2O7 through the evolut ion of water. mol gCsHSO 2304 mol gO S Cs4427 2 2 961 0 230 2 442 mol g mol g Total dissociation of water from CsHSO4 should cause a 3.9% reduction in mass. 2.4 X-ray Diffraction The atomic structure of electrolyte produced fo r this work was verified using powder X-ray diffraction (XRD). By observing the diffraction pattern created from an X-ray beam striking the sample, information about the relative position of atoms in the lattice can be determined. This information allows for the identificati on of phases present from different synthesis methods. In addition, the width of peaks observed in an XRD pattern allows for a determination of crystallite size. Powder X-ray diffraction techniques allow for th e analysis of structural formation and phase identification of samples which do not exist as single crystals29, 30. A polycrystalline sample, such as the CsHSO4 samples synthesized in this work, are formed of small granules or flakes, further composed of randomly oriented crystall ites. An incident X-ray beam would therefore

PAGE 61

53 encounter crystallites in all possible three-dimensional rotatio ns, and would diffract across all possible hkl planes simultaneously. This produces mu ltiple cone shaped diffractions from a given incident angle. As the sample rotates, these cone-shaped diffractions do not change, as at any given angle from the s ource, the sample appears the same, as a collection of randomly oriented planes. However, the detector st ill passes through these diffracted cones in sequence, producing a diffraction pattern identical to that observed with single-crystal diffraction methods. For this work, a Philips X’Pert PRO diffractometer with Cu K radiation (1.5418 ) was used to probe the samples. Applied voltage and current were 45 kV and 40 mA. ” pellets created were mounted unto the sample stage, or powder samples were placed in sample holders using a “backloading” technique. Early samples were scanned over range of 2 = 0 to 90 degrees, with a scan rate of 2.4 degrees /minute. Later samples were probed over a narrower range of 2 = 10 to 50 degrees, with a scan rate of 1.2 degrees/minute. The data was analyzed using Phillips X’Pert Highscore and the Inorganic Crystal Structure Database and the International Centre for Diffraction Data. Figure 2.11 shows an experime ntal pattern from CsHSO4 powder crystallized by acetone. Also shown is a reference pattern31 for CsHSO4-III showing a strong match to the experimental pattern.

PAGE 62

54 CsHSO4-III Experimental Scan 203040Position (2theta) -111 002 012 -202 210 020 -121 Figure 2.11 PXRD Pattern of Monoclinic CsHSO4-III Experimental PXRD pattern of CsHSO4 (top) and reference pattern of CsHSO4-III (P21/n) (bottom) showing solid correlati on between peaks in PXRD versus peaks from single crystal. Dominant peaks in the experime ntal pattern and their most likely matches from the reference pattern are shown in Table 2-4 below. Table 2-4 Experimental PXRD versus CsHSO4-III Reference Pattern Comparison of peaks observed in experi mental pattern to those from CsHSO4-III reference pattern31, showing strong correlation. Experimental Pattern (2 ) CsHSO4-III Reference Pattern Position (2 ) h k l 18.509 18.496 0 0 2 19.650 19.637 -1 1 1 24.071 24.039 0 1 2 24.811 24.813 -2 0 2 27.312 27.290 2 1 0 30.670 30.719 0 2 0 33.071 33.220 -1 2 1

PAGE 63

55 An experimental pattern of CsHSO4 crystallized by methanol is shown in Figure 2.12 below. This pattern showed features of CsHSO4-III and CsHSO4-II reference patterns, as shown in the figure. 2030 40Position (2theta)CsHSO4-III CsHSO4-II Experimental Scan101 -111 012 -202 210 110 -111 200 002 120 -202 -221 Figure 2.12 PXRD of CsHSO4 Showing CsHSO4-II and CsHSO4-III Features Experimental PXRD pattern of CsHSO4 (top), reference pattern of CsHSO4-III (P21/n) (middle), and reference pattern of CsHSO4-II (P21/c) (bottom) showing a mix of features from reference scans found in experimental pattern. Strong peaks from the experimental pattern al ong with likely candidates from both reference patterns are shown in Table 2-5 below.

PAGE 64

56 Table 2-5 Experimental PXRD versus CsHSO4-III and CsHSO4-II Reference Patterns Comparison of peaks observed in experi mental pattern to those from CsHSO4-III and CsHSO4-II reference patterns31, showing elements of both reference materials in experimental sample. Experimental Pattern (2 ) CsHSO4-III Reference Pattern CsHSO4-II Reference Pattern Position (2 ) h k l Position (2 ) h kl 16.390 16.435 1 0 1 16.321 1 1 0 17.634 -1 1 1 19.691 19.637 -1 1 1 24.039 0 1 2 24.445 2 0 0 24.680 0 0 2 24.855 24.813 -2 0 2 25.094 1 2 0 27.290 2 1 0 28.031 27.920 -2 0 2 32.064 31.977 -2 2 1 33.211 A sample of CsHSO4 was excessively dried in an oven ove rnight, to determine the effect of the drying step on electrolyte structure. The PXRD pattern is shown in Figure 2.13, along with the reference pattern for CsHSO4-II.

PAGE 65

57 203040Position (2theta)CsHSO4-II Experimental Scan110 -111 200 002 120 -202 -221 Figure 2.13 PXRD Pattern of Monoclinic CsHSO4-II Experimental PXRD pattern of CsHSO4 (top) and reference pattern of CsHSO4-II (P21/c) (bottom) showing solid correlati on between peaks in PXRD versus peaks from single crystal. The strong peaks from this experimental scan are matched with the most likely candidates from the CsHSO4-II reference pattern in Table 2-6 below. Table 2-6 Experimental PXRD versus CsHSO4-II Reference Pattern Comparison of peaks observed in experi mental pattern to those from CsHSO4-II reference pattern31, showing strong correlation. Experimental Pattern (2 ) CsHSO4-II Reference Pattern Position (2 ) h k l 16.412 16.321 1 1 0 17.672 17.634 -1 1 1 24.711 24.445 2 0 0 24.790 24.680 0 0 2 25.230 25.094 1 2 0 32.132 31.977 -2 2 1

PAGE 66

58 Crystallite size can be estimated from the wi dth of the peaks observed in the diffraction pattern30. For a given peak, Bragg’s Law must be satisfied for the crystal spacing, wavelength of incident light, and angle at which the beam strikes the crystal plane: Bd sin 2 where is the wavelength of the incident X-ray beam, d is the spacing between crystal planes, and B is angle which satisfies Bragg’s Law for the given d and (note that typically XRD patterns are shown as intensity versus 2 B). For a given peak, the full width at half maximum can be approximated by assuming the peak’s shape as a triangle. 2 1 2 12 2 2 1 B where B is the full width at half max, and 2 1 and 2 2 are the endpoints of the peak observed on the pattern. Bragg’s Law can be applied to the two angles at the endpoints of the peak, assuming a number of crystal layers m: 1 sin 2 1 sin 22 1 m t m t

PAGE 67

59 Where t is the total thickness of the crystallite, t = md. Approximating: B 22 1 and 2 2 sin2 1 2 1 Then solving for t results in: BB K t cos where K is a correction factor based on the assu med shapes of the crystallites. Values of K vary from 0.7 to 1.05, and are unitless quanti ties. For this work, a shape factor of 0.9 associated with parallel planes was used. Table 2-7 shows calculated crystal sizes from the three samples shown in Figures 2.11 to 2.13, as well as crystallite sizes calculated for CsHSO4 crystallized by temp erature control. Samples crystallized by temperature control prod uced peaks, but did not match any reference pattern for CsHSO4.

PAGE 68

60 Table 2-7 CsHSO4 Crystallite Size CsHSO4 crystallite sizes calculated from PX RD data using Sherrer’s Formula. Phase Average Crystallite Size () Standard Deviation () Sample A (Acetone) CsHSO4-III 3308 621 Sample B (Methanol) CsHSO4-III and CsHSO4-II 3748 316 Sample C (Excessively Dried) CsHSO4-II 2997 838 Sample D (temp control) Unknown 4228 692 Crystallite sizes ranged from approximate ly 3000 to 4200 , with larger crystallites associated with crystallization using temperat ure control. Crystallization using methanol yielded crystallites larg er than those yielded by acetone crystallization. 2.5 X-ray Photoelectron Spectroscopy X-ray photoelectron spectroscopy (XPS) was us ed to probe the surface atoms of CsHSO4 powders and pellets, to determine the presence or absence of absorbed species, particularly alcohols left over from the crysta llization or sintering steps. When an atom is struck by an X-ray, a photoel ectron may be ejected. Th is electron will have a kinetic energy equal to the difference between the energy of the X-ray photon that struck it, and the binding energy required to excite the electron from its bound orbital state to a free electron32, 33.

PAGE 69

61 In addition, electrons in higher orbitals may relax down to orb itals vacated by electrons. The difference in energy is transferred to a neighbor ing electron, which is emitted with a kinetic energy equal to the difference between the ch ange in binding energy experienced by the relaxing electron and the binding energy of the electron emitted. XPS can also determine the chemical state of an atom. Forming molecular bonds changes the binding energy of electrons in an atom, resulti ng in a shift in the kinetic energy observed for emitted photoelectrons. XPS data for this work was obtained usi ng a Perkin-Elmer PHI 560 ESCA/SAM system, operated at approximately 5 x 10-10 Torr. The X-ray source was a magnesium filament in a PHI 04-500 dual X-ray source, with a voltage of 15 kV and a current of 20 mA. For survey scans, a pass energy of 100 eV was used, while a 50 eV pass energy was used for high resolution scans. The kinetic energy of emitte d electrons was measured using a 25-270AR cylindrical mirror analyzer, while the data was analyzed using AugerScan. From a CsHSO4 sample we would expect to see peak s associated with the binding energies of Cs, S, and O, as well as carbon that has form ed on the surface of the sample from the lab. Expected binding energies for these species, as we ll as atomic sensitivity factors for selected peaks, are shown in Table 2-8.

PAGE 70

62 Table 2-8 Binding Energies and Atomic Sens itivity Factors for Species Present in CsHSO4 34 Element Atomic Shell Binding Energy (eV) Atomic Sensitivity Factor Cs Cs Cs Cs 3s 3p1/2 3p3/2 3d3/2 1219 1069 1002 740 Cs 3d5/2 726 7.041 O 1s 531 0.711 C 1s 285 0.296 Cs S Cs S 4s 2s 4p1/2 2p1/2 234 228 173 165 S 2p3/2 164 0.666 Cs Cs Cs Cs O S 4p3/2 4d3/2 4d5/2 5s 2s 3s 161 80 77 25 23 18 A typical XPS survey scan for CsHSO4 is shown in Figure 2.14. Th is sample was crystallized by added methanol to solution.

PAGE 71

63 0 5000 10000 15000 20000 25000 30000 35000 0 100 200 300 400 500 600 700 800 900 1000 1100Binding Energy (eV)CountsCs 3d3/2Cs 3d5/2O 1s C 1s Cs 4s & S 2s Cs 4p & S 2p Cs 4d Figure 2.14 X-ray Photoelectron Spect roscopy Survey Scan of CsHSO4 Powder Crystallized with Methanol This is a typical survey scan of CsHSO4 showing peaks associated with the binding energies of different atomic shells within the sample. Determining atomic ratios for CsHSO4 is hampered by the low atomic sensitivity of sulfur compared to that of cesium, and that characteri stic sulfur peaks lie very close to cesium peaks (sulfur 2s is very close to cesium 4s, and sulfur 2p1/2 and 2p3/2 are very close to cesium 4p1/2 and 4p3/2) meaning that the cesium peaks overwhelm the sulfur peaks. Figure 2.15 shows a high resolution scan about the cesium 4p and sulfur 2p peaks. The sulfur peaks are all but indecipherable against the larger cesium peaks.

PAGE 72

64 180177.5175172.5170167.5165162.5160157.5155 Binding Energy (eV) Cs 4p1/2Cs 4p3/2S 2p1/2and S 2p3/2 Figure 2.15 XPS High Resolution Scan of Sulfur 2p Peaks High resolution XPS scan (background subtracted) of sulfur 2p1/2 and 2p3/2 peaks, showing their low sensitivity factor versus energetically similar cesium 4p1/2 and 4p3/2 peaks.

PAGE 73

65 3 Permeability Studies Since the electric potential and current density of a fuel cell can both be lowered by fuel or oxidant gases permeating through the electrolyte, a study of ga s permeability characteristics of CsHSO4 was undertaken. 3.1 Permeability Experimentally, the flux through a membrane is typically measured directly by monitoring the concentration of the de sired solute on the acceptor side of the membrane35. This concentration function can be derived from a basic mass balance performed on the acceptor side. Starting with a mass balance expression: Fluxin – Fluxout = Accumulation The flux in the above expression includes both the flux due to diffusion across the membrane as well as flux due to the convection of so lvent carrying solute through the membrane. However, the proposed experimental setup has no pressure difference across the membrane,

PAGE 74

66 and the volume of material on both sides of the membrane is approximately equal. Thus the flux due to convection is considered negligible. From sources discussing the behavior of membra nes, flux due to diffusion alone is shown to be35, 36: c L K D A Flux Where A is the cross-sectional area of the membrane, L is the membrane thickness, D is the diffusion coefficient for the solute between the membrane and the adjacent solution, K is the partition coefficient for the solute between the membrane and the adjacent solution, and c is the concentration providing the driving force. Combining terms back into the balance expression: B A B Bc c L K D A dt dc V Where cB is the concentration on the accepto r side of the membrane and cA is the concentration on the donor side of the membrane. At the operating conditions of the experiment the concentration on the acceptor side of the membrane is always negligible relative to th e concentration on the donor side, so we assume cB goes to zero on the RHS of the above equation. That is, there is negligible diffusive flux

PAGE 75

67 backwards across the membrane due to con centration on the acceptor side during the experiment. A B Bc L K D A dt dc V Assuming the concentration on the donor side does not change during the experiment and that the initial condition for th e concentration on the acceptor is zero, this equation can be solved for cB. The following expression is the result: ) ( ) (0t t c L P V A t cA B B where P is the membrane permeability defined as the product D x K. This equation includes a time delay, t0, the result of the constant of integration. This time delay is due to the initial diffusion through the membrane immediately after th e addition of solute to the donor side, as well as the brief time before linearity is esta blished on the acceptor side. This time lag is explicitly related to the diffusivity by36: D L t62 0 It is desirable to determine how the permeation through a membrane varies with temperature.

PAGE 76

68 The permeability at a given temperature can be obtained from the slope of the curve of the concentration in the acceptor side versus time. In general, the relation between permeability of a membrane with temperature is expressed as a form of the Arrhenius Law, as shown37: RT E P Paexp Where P is the permeability at an infinitely high temperature, Ea is the activation energy for the permeation of the given solute across the membrane, R is the gas constant, and T is absolute temperature. Taking th e log of both sides results in: RT E P Pa ln ln Thus the slope of plot of the ln P ve rsus 1/T will have a slope equal to Ea/R, allowing one to determine the activation energy. 3.2 Experimental Setup An experimental cell was constructed to test permeability of manufactured membranes. This cell had two chambers, separated by the membrane to be tested. Gas inlets could be regulated into and out of either chamber by means of a manifold and rotometers. The cell was kept in a convection oven to ensure even heating, while all inlet lines passed through coils within the

PAGE 77

69 oven to insure that inlet gases were not lo wering the temperature at the surface of the membrane. Samples were removed from one of th e chambers (the acceptor side) by means of an Ultra-Torr diaphragm with a 10 l syringe. This sample size was small enough to not cause a significant loss of gas in the acceptor side compared to the size of the chamber. Samples were then analyzed using gas chro matography (see below) to determine the concentration of the species permeating through the membrane. Gases to be tested were fed into the donor si de of the cell at 100% concentration, while nitrogen was fed into the acceptor side. This was done for thirty minutes to insure a steadystate starting condition of 100% of the test gas on the donor side and 0% on the acceptor side. Flow of nitrogen to the acceptor side was halted at time zero, when measurements were taken from the acceptor side. All valves on the accep tor side were closed, while those on the donor side remained open, so that the concentrati on on the donor side remained nearly constant. The flow of test gas through th e donor side of the cell was ke pt at low values of 10 to 15 l per minute, so that pressure differentials across the membrane could be neglected. Figure 3.1 shows a cross-section of the permeability cell used for experimental measurements.

PAGE 78

70 Acceptor SideVariable conc. Membrane Donator Side100% conc.Acceptor Inlet (A)Acceptor Outlet (B) Donator Inlet (C) Donator Outlet (D) Sampling Port Figure 3.1 Schematic of the Permeability Cell Cross-sectional schematic of e xperimental setup for permeability experiments. Gas to be tested is fed through the donator side from C to D. Nitrogen is fed through the acceptor side from A to B, until a steady state has been achieve d. Then the nitrogen flow is closed off at both A and B. Samples are taken over time from the sampling port. 3.3 Gas Chromatography Gas chromatography allowed quantitative evalua tion of gases produced during electrolysis or for permeation testing of membranes38, 39. A gas separation was performed within the chromatograph column, and as the different species left the column, they were detected producing a signal proportional to the amount of that species. Using a calibration curve of signal strength to concentration, the makeup of the gas could be determined.

PAGE 79

71 Quantitative analysis of gas species was obtained using an Agilent Technologies 6890N Network Gas Chromatograph. Gas samples were injected into one end of an HP-5 Phenyl Methyl Siloxane column (Agilent 19091J-413) 30 m long with an i nner diameter of 320 m and an average film thickness of 0.25 m. Injection temperature was 250 C. Helium at 229 ml/min was used as a pass-through gas, though an oven at 40 C. A flame ionization detector operating at 300 C with a H2 / air mix of 1:10 was used to detect gas species leaving the column. The data was analyzed using Chemstation Rev. A.10.02. 3.4 Effect of Methanol Methanol was chosen to be the first gas inves tigated. Initial runs with methanol showed a steadily increasing concentrati on on the acceptor side with ti me. This concentration curve reached an upper limit, after which it did not in crease. One would expect the concentration of methanol to continue increasing to 100% at some long time, but this was not the case. Subsequent runs were attempted without ch anging the membrane in the permeability cell. Both sides were flushed thoroughl y with nitrogen, and the same experimental procedure was followed. When this was done methanol perm eation was drastically reduced versus that through a fresh pellet. Permeation profiles versus time for fresh CsHSO4 and those already subj ected to the runs are shown in Figure 3.2.

PAGE 80

72 0 5 10 15 20 25 30 0153045607590105120Time (min)Methanol Concentration (%) Fresh Pellet Treated Pellet Figure 3.2 Permeation Response of CsHSO4 Membrane to Methanol Vapor Methanol concentration in the acceptor side of the permeation cell versus time for a fresh membrane of CsHSO4 (black circles, dashed line) and a membrane of CsHSO4 already treated with methanol permeation (white circles, solid line). The same permeation experiment was also atte mpted with methane as the permeation gas. Like methanol, methane permeated freely th rough freshly made membranes of CsHSO4. Unlike methanol, methane did not level off at as low of a value, and continued to rise throughout the length of the expe riment. If the experiment was run long enough, eventually 100% methane was all that remained in the acceptor side of the cell. Figure 3.3 shows the results of several perm eation runs of methan e through fresh CsHSO4 membranes. Also shown is a weight ed average of experimental data.

PAGE 81

73 0 10 20 30 40 50 60 70 80 90 0102030405060Time (minutes)Methane Concentration (mole %) Experimental Values Average Value Figure 3.3 Permeation of Methane through Fresh CsHSO4 Membrane Methane concentrations in the acceptor side of the permeability cell versus time for several membranes of CsHSO4 (dashed lines). The solid line is a weighted average of these profiles. Experiments were also conducted with meth ane permeating through membranes that had already been treated with methanol. These pr ofiles showed a marked decrease in the permeability of CsHSO4 to methane after it had been treated with methanol. Molar flux across the membrane was decreased by nearly th ree full orders of magn itude after treatment with methanol. Figure 3.4 shows the results of several pe rmeation runs of methane across a CsHSO4 membrane that had been treated with methanol.

PAGE 82

74 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 01002003004005006007008009001000Time (minutes)Methane Concentration (mole %) Figure 3.4 Permeation of Methane through CsHSO4 Membrane after Treatment with Methanol Methane concentrations in the acceptor side of the permeability cell versus time for several methanol-treated membranes of CsHSO4. Data for methane permeability through fresh and treated membranes of CsHSO4 is shown in Table 3-1. Table 3-1 Permeation Data for Methane through CsHSO4 Fresh Membrane Membrane Treated with Methanol Flux of methane across membrane for short times (gmol/(m2s)) 3.74 x 10-5 4.22 x 10-8 Permeability of membrane to methane (s-1) 1.03 x 10-5 1.17 x 10-8 Steady state concentration of methane 45 to 100% 0.1 to 0.5% Time to achieve steady state value 30 to 50 minutes 100 to 300 minutes

PAGE 83

75 An investigation of the interaction between a methanol-treated CsHSO4 pellet and steam was also undertaken in this study. Earlier tests has shown that introducing steam into a cell containing CsHSO4, even at 150 C, often produced cat astrophic effects. It is well known that CsHSO4 is soluble in liquid wate r, but it is thought that water vapor produces little effect. However, membranes of CsHSO4 that were exposed to water vapor lost structural integrity, even in water to air concentrations as low as 20%. In fact, to prevent interaction between CsHSO4 membranes and any ambient moisture that might have migrated from the lab atmosphere into the experimental tubing, all tubes were flushed with dry nitrogen at 150 C for several hours before membranes were added to the permeation cell and experiments were run. Since permeation to neutral species (such as me thane) change after exposure to methanol, it was investigated whether methanol could al so improve the survivability of a CsHSO4 membrane in the presence of steam. A feed st ream of 50/50 methanol/water was fed to the permeability cell, containing a CsHSO4 membrane that had been treated with methanol, for periods of time, between which methane permeabili ty runs were performed just as before. The CsHSO4 membrane resisted the methanol/water stream for two hours. The third hour of exposure, however, showed a marked increase in methane permeability. By the fourth hour, the permeation profile was closer to that of an untreated membrane. Additional exposure beyond four hours did not dr astically degrade the me mbrane any further.

PAGE 84

76 The resulting methane permeati on profiles for steadily increasing exposure times for steam are shown in Figure 3.5 below. 0 5 10 15 20 25 30 35 40 45 0102030405060708090100Time (minutes)Methane Concentration (%) 1 hour 2 hours 3 hours 4 hours 8 hours Figure 3.5 Effect of 50/50 Mix of Methanol/Steam on Permeability to Methane Methane concentrations in the acceptor side of the permeability cell versus time for a methanol-treated membrane of CsHSO4 after increasing exposures to 50/50 methanol/water.

PAGE 85

77 4 Fuel Cell and Electrolysis Applications The previous two chapters have described me thods for the evaluation of the synthesis and optimization of CsHSO4 as an electrolyte in membrane electrode assemblies. The following chapter will explore construction techniques of MEAs and will evaluate the performance of several fully constructed MEAs in a variety of applications. Specifically, the use of CsHSO4 in an MEA for a H2/air fuel cell and the electrolysis of methanol and steam will be explored. 4.1 Experimental Setup A fuel cell test bed from the Fuel Cell Stor e (product FC25-01DM) w ith direct methanol capability was used for all fuel cell and elec trolysis testing. Temperature was regulating using thermal pads included in the kit, as well as a Yamato DX300 Gravity Convection Oven, which also served to keep the inlet ga s streams by heating long coils present in the oven. All experiments were conducted at 150 C.

PAGE 86

78 4.1.1 Fuel Cell The performance of the fuel cell experiment s was determined by measuring the electric potential and current across a variable re sistor using a Yokogawa WT230 Digital Power Meter. Open cell voltage was measured by removing the resistor from the current. A schematic of the experimental setup is shown in Figure 4.1. V A Fuel Cell AmmeterVoltmeterVariable Resistor Figure 4.1 Circuit Diagram for Polariza tion Measurements of Fuel Cell Experimental setup for measuring electric potenti al and current across load from a given fuel cell configuration.

PAGE 87

79 4.1.2 Electrolysis Direct current for electrolys is reactions was provided by an Agilent E3640A DC Power Supply, which also provided curr ent flow through the electrolyte. 4.2 MEA Construction MEAs for electrolysis and fuel cell experime nts were constructed in the same way. Electrolyte membranes and electrode/catalyst as semblies were fabricated separately and then placed together in the fuel cell test bed. When the test bed was closed and tightened, this produced a single membrane electrode assembly. 4.2.1 Electrolyte All electrolyte used in the following experiments was CsHSO4 synthesized as described in Appendix A. Based on the ease of crystallization, and a desire to avoi d any negative changes associated with the superprotonic phase cha nge as shown in Chapter two with DSC, all electrolyte was crystallized out of solution by adding aceto ne. Two inch diameter, one millimeter thick CsHSO4 membranes were uniaxially pressed with 100 tons of force, to create a densely-packed solid acid disc.

PAGE 88

80 4.2.2 Electrodes Carbon fiber paper was used as electrode mate rial. Two inch diameter discs were cut from larger sheets of density 12 g/m2 provided by Technical Fibre Products. These electrodes were then impregnated with catalyst material before being integrated into the completed MEA. 4.2.3 Catalysts A mixture of platinum black, car bon black, naphthalene, and CsHSO4 in a mass ratio of 10:1:1:6. This mixture was added to small amou nt of water to create a thick slurry. This slurry was spread onto the surface of the car bon fiber electrode. The electrode was then placed onto a crystal filter over a vacuum fl ask. The imposed vacuum pulled water through the fiber electrode, leaving the catalyst mixture impre gnated throughout the mesh. The impregnated electrode was then dried in a c onvection oven at 70 C for several hours. This left carbon fibers coated with electrolyte, catalyst, carbon black, and naphthalene. The electrolyte improved electrical connection between the bulk el ectrolyte membrane and the catalyst, while the carbon black improved electr ical connection between the catalyst and the electrode. When the MEA is heated to ope rating temperature (150 C) the naphthalene evaporates out of the catalyst layer, leaving a porous structure to increase available active sites. A schematic of the catalyst impregnation proces s into a carbon fiber electrode is shown in Figure 4.2 below.

PAGE 89

81 Vacuum Crystal Filter Carbon Fiber Electrode Catalyst/Electrolyte Slurry Figure 4.2 Catalyst/Electrode Construction A catalyst slurry of platinum black, carbon blac k, electrolyte, and naphthalene is pulled by vacuum into a carbon fiber electrode, creating a porous electrode/catalyst assembly with superior electrolyte to catalyst and catalyst to electrode connection. 4.3 Fuel Cell Performance The following is experimental polarization curves of H2/O2 fuel cells with CsHSO4-based membrane electrode assemblies with a variety of preparation methods. Figure 4.3 below shows a polarizatio n curve of a freshly-made MEA.

PAGE 90

82 Figure 4.3 Polarization Curve of CsHSO4 Fuel Cell with H2 and Air Electric potential and power density with increasing current density of a freshly made CsHSO4 MEA. This MEA shows an open cell voltage that appr oaches ideal (1.17 V), but current density, and by extension, power density is low. A H2/O2 polarization curve was generated with an ME A that had been used in the electrolysis of methanol (see next section). It is shown in Figure 4.4 below.

PAGE 91

83 Figure 4.4 Polarization Curve of CsHSO4 Fuel Cell with H2 and Air after Methanol Electrolysis Electric potential and power density with increasing current density of a CsHSO4 MEA after use in methanol electrolysis. This is compared to a simila r polarization curve of a CsHSO4 MEA that had been simply treated with a long exposure to methanol at high temperature in a manner similar to the permeability tests (Chapter 3). This cu rve is shown in Figure 4.5 below.

PAGE 92

84 Figure 4.5 Polarization Curve of CsHSO4 Fuel Cell with H2 and Air after Methanol Sintering (Long-term) Electric potential and power density with increasing current density of a CsHSO4 MEA after methanol treatment. These two polarization curves show a marked reduction in open-cell voltage, being a little over half that observed for a fresh MEA. However, current density is increased. The MEA in the previous experiment was subj ected to a longer st ream of hydrogen while under load for several hours. With time, open ce ll voltage increased, although not back to the pre-methanol treatment levels. Current density remained similar to that of a treated MEA, resulting in a higher maximum pow er density. This polarization curve is shown in Figure 4.6.

PAGE 93

85 Figure 4.6 Polarization Curve of CsHSO4 Fuel Cell with H2 and Air after Methanol Sintering and Regeneration Electric potential and power density with increasing current density of a CsHSO4 MEA after methanol treatment Data from these polarization curves is summarized in Table 4-1 below.

PAGE 94

86 Table 4-1 Fuel Cell Data Cell Run Open Cell Voltage (mV) Efficiency of Open Cell (EV) Maximum Current Density (mA/cm2) Maximum Power Density Cell Potential (mV) Efficiency (EV) Current Density (mA/cm2) Power Density (mW/cm2) Fresh CsHSO4 MEA 936 80.0% 7.76 485 41.5% 4.48 2.17 MEA post methanol electrolysis 633 54.1% 18.4 260 22.2% 11.7 3.05 MEA after methanol treatment 514 43.9% 17.5 144 12.3% 9.67 1.39 Treated MEA after extended H2 exposure 902 77.1% 23.3 395 33.8% 11.8 4.64 4.4 Electrolysis Performance MEAs made with the constructi on methods described earlier were also used in electrolysis application using much of the same experime ntal setup. Methanol and steam electrolysis were investigated. 4.4.1 Electrolysis of Methanol Methanol electrolysis was performed using a 50/50 mix of water vapor and methanol. The following figures show how measured current density increased after minimum voltage was applied.

PAGE 95

87 0 1 2 3 4 5 6 7 0123456789Electric Potential (V)Current Density (mA/cm2) Figure 4.7 Electrolysis of Methanol Current density and power density versus increasi ng voltage for the electrolysis of methanol. Figure 4.8 shows the behavior of current dens ity versus increasing vol tage just above the minimum voltage for methanol electrolysis.

PAGE 96

88 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 00.10.20.30.40.50.60.70.80.911.11.2Electric Potential (V)Current Density (mA/cm2) Figure 4.8 Electrolysis of Methanol at Low Voltage Electrolysis of methanol at lower volt ages showing minimum voltage required. 4.4.2 Electrolysis of Steam The following figures show increasing current de nsity with increasing applied voltage for the electrolysis of steam. Feed flows were 10/90 me thanol/water vapor in an effort to prevent structural degradation of CsHSO4 upon contact with water vapor.

PAGE 97

89 0 1 2 3 4 5 6 7 8 9 0123456789Electric Potential (V)Current Density (mA/cm2) Figure 4.9 Electrolysis of Steam Current density and power density versus increas ing voltage for the electrolysis of steam. Figure 4.10 shows the behavior of current dens ity versus increasing vo ltage just above the minimum voltage for electrolysis of steam.

PAGE 98

90 0 0.5 1 1.5 2 2.5 3 00.511.522.5Electric Potential (V)Current Density (mA/cm2) Figure 4.10 Electrolysis of Steam Low Voltage Electrolysis of methanol at lower volt ages showing minimum voltage required. Notable on the plot of current ve rsus electric potential for the electrolysis of steam is the low minimum voltage for current flow, at approximat ely 0.6 volt. The linear relation of current to voltage takes a quick upturn, and begins ri sing at a steeper slope, after 1.3 volts.

PAGE 99

91 5 Summary and Conclusions The following chapter will present the conclusi ons derived from the data presented in Chapters 2, 3, and 4. These conclusions are separated by chapter subject, so will proceed from materials characterization, to permeability studies, to conclusions about the fuel cell and electrolysis applications. Finally some suggestions for future directions for further research based on the findings of this work will be presented. 5.1 Materials Characterization The primary conclusions of the materi al characterization methods of CsHSO4 are presented below. CsHSO4 was investigated with impedance spectroscopy, differential scanning calorimetry (DSC), thermal gr avimetric analysis (TGA), powder X-ray diffraction (PXRD), and X-ray photoelectron spectrosc opy (XPS) in order to determine the relative benefits and problems associated with the different synt hesis and pressing me thods investigated. A secondary goal with materials characterizati on was the determinati on of feasibility and applicability of the different ch aracterization methods to CsHSO4 as part of an effort to design a consistent, efficient characterization procedure for future samples.

PAGE 100

92 Crystallization by introduction of organic produces smaller cr ystallite sizes than those produced by slow evaporation, while crystals produced using acetone were easier to dry producing acceptable material. Crystals produced with methanol or with water tended to clump together, and had to be aggressively ground and dried. Slight variations were observed between di fferent crystallization and pressing conditions with impedance spectroscopy. Membranes that had been crystallized by acetone showed improved proton conductivity against those that were crystallized by temperature control, while membranes that had been pressed under heated conditions were still better. Since proton conductivity can be dete rmined from thermodynamic and structural constants, one would expect all samples of CsHSO4 to display identical prot on conductivity at a given temperature. The primary difference is, however, th at samples investigated in this work were all pressed powders and non single crystals. Inte rface gaps between crysta llites, variation in the orientation of crystallites, and gaps between larger flakes of clumped crystallites would all result in different measured pr operties from impedance spectroscopy. It is also notable that despite the differences that these different cons tants would impose, that the ultimate proton conductivities at 150 C for all samples were similar values. This result allows whichever production method that is the eas iest, most efficient, or most economical to be chosen without concern of undue effect on proton conductivity. Loss of mass in the superprotonic phase was observed with TGA consistent with the degradation of CsHSO4 to Cs2S2O7 and water. This loss was also associated with an

PAGE 101

93 endothermic peak on DSC. When the sample was tested in DSC with a closed cup to prevent evolution of water, or when the sample wa s retried in an open cup immediately after a previous experiment, this loss of mass was not observed. This leads to the conclusion that CsHSO4 in the superprotonic phase will evolve water, degrading to Cs2S2O7, particularly in the presence of dry fuel or oxidant streams. Sp ecial consideration should be paid to the effect of the evolution of water in operating conditions, as Cs2S2O7 lacks the proton conductivity of CsHSO4. The jagged peaks observed on DSC with samples that had been crystallized by temperature variation suggest a structural effect on the e volution of water in the superprotonic phase. Given the greater tendency for CsHSO4 that has been crystallized by temperature variation to clump into larger pieces, this jagged profile may be due to pockets of moisture forming within the powder clumps as with other samp les. With the greater clumping, however, these moisture pockets build up pressure within th e clump, and are then released in bunches, resulting in many small endothermic peaks obs erved by DSC. So, in addition to worrying about the effect of the ev olution of water has on the proton conductivity of CsHSO4, an additional problem with samples that have been crystallized by temperature control is the effect on structural integrity of the membrane fr om this physical bursting as water is released. Phases were identified using PXRD. Solid matc hing of experimental patterns with reference patterns for CsHSO4-III and CsHSO4-II confirmed the reproducibility of structural features from synthesis methods. Also present were samp les that showed peaks associated with both CsHSO4-III and CsHSO4-II, meaning that samples of mixed phases were also produced.

PAGE 102

94 CsHSO4-II was found in those samples that had been aggressively dried at temperatures in excess of 70 C, while those that were not as ag gressively dried (typical ly those samples that were crystallized using acet one) showed just the CsHSO4-III phase. The time frame between when aggressively dried samples were remove d from the drying oven and when they were investigated with PXRD occasionally a period of days, suggests that the phase change from CsHSO4-III to CsHSO4-II is essentially irreversible. Although XPS survey scans produced characteristic peaks that would be expected for cesium and oxygen, the sulfur peaks were obscured by much larger cesium peaks that have similar binding energy. This made using XPS to determ ine atomic ratios and chemical shifts, and therefore the presence of excess acid, crystallizing agents, or surface reactants with different crystallization methods problematic. A standardized production proce dure was established for CsHSO4 membranes for use in all future work. Synthesis of CsHSO4 from Cs2SO4 was chosen as this was an easier and safer reaction than that of synthesis from Cs2CO3. From the results of materials characterization tests it was decided that all standard synt hesis would use acetone-crystallized CsHSO4, instead of those crystallized by methanol or temperature variation. The primary factors for this choice were the ease of use of acetone crys tallization, and concerns about the effects of clumping flakes, particularly after the superp rotonic transition as observed by DSC and TGA. Pressing conditions can be chosen based on the ease of manufacture, as significant changes in membrane conductivity were seen to be independent of the pressing method.

PAGE 103

95 For the evaluation of completed CsHSO4 samples, use of PXRD and DSC/TGA produced the most useful results. Conductivity test s confirmed the presence of CsHSO4 and its superprotonic character with the higher temper ature phase, but did not show much variation between different sample hist ories or easily repeatable re sults. XPS was hindered by the overlapping of photoelectron peaks from cesium over those from sulfer, limiting its use as a diagnostic tool. PXRD and DS C/TGA, however, allowed for the evaluation of phase presence, crystallite size, and the dynamics of dehydration at highe r temperatures. These methods show a good promise for the robust ev aluation of the quality of future CsHSO4 samples. 5.2 Permeability Studies The results and conclusions of methanol a nd methane permeability studies are presented below. Permeability to fuel and oxidant gases goes directly to the pe rformance of a given electrolyte within an MEA. Excessive permeab ility in an MEA can lower the current density and the output electric potential for a fuel cell. CsHSO4 membranes created using the standard pr ocedure developed from Chapter 2 proved to be highly permeable to gases. Flux of gases across the surface of the membrane was sufficient to cause a large change in concentrat ion on the other side of the membrane in the space of minutes. Membranes with this high degree of permeability would not be suitable as MEAs in fuel cells.

PAGE 104

96 Those membranes that were exposed to meth anol at high temperat ures, later showed a remarkable reduction in their permeability. Flux of fuel gas across their surface was reduced by nearly three orders of magnitude, and the membranes could be run in the cell for hours with only a slight change in th e concentration of gases on the acceptor side of the cell. This change seems irreversible, as membranes so treated kept these permeability characteristics long after the initia l runs were completed, and later runs up to days later, showed similar results. Membranes treated with methanol also s howed increased resistance to water vapor. Untreated membranes, when exposed to water va por, would lose their st ructural integrity and collapse into a loose deposit on the bottom of the cell within minutes of exposure. Treated membranes could withstand a 50/50 water vapor/ methanol mix for up to three hours before showing any effect. Even this situation, membranes showed an increased permeability to neutral gases (methane), and not a tota l collapse like untreated membranes. These results lead to the conclusion that a fuel cell of electrolysis process that would use CsHSO4 as an electrolyte should in clude treatment with metha nol, or a similar treatment process to prevent a highly permeable MEA. The result of increased resistance to water vapor is also promising, as it allows for the use of CsHSO4 in a direct methanol fuel cell, which requires water vapor in the fuel stream to allow complete reaction on the anode.

PAGE 105

97 5.3 Fuel Cell and Electrolyte Test Bed H2/O2 fuel cells were successfully run with CsHSO4 MEAs. These produced fuel cells with open cell voltages that were close to the exp ected values (1.17 V), but with much lower current densities than that of commercial fuel cells using Nafion or phosphoric acid. These lower current densities may be due to gas permeability of the untreated CsHSO4. Flow of fuel gas to the cell reaches a rate at which the fl uid flow dynamics limit the amount of gas that can reach or escape the surface of the membrane but a significant portion passes through the electrolyte without reaction, resu lting in a lower observed current. Membranes that had been exposed to metha nol at high temperatures, either through the treatment method developed in Chapter 3, or as a result of the methanol electrolysis experiment, showed an increase in current density, but a decrease in open cell voltage. The increase in current could be a re sult of the decrease in permeability to fuel gases leading to a greater proportion of the gases reacting at the surface versus passing through the membrane without reacted.

PAGE 106

98 The decrease in observed open cell voltage is more problematic, as it implies that the methanol treatment may cause a change in th e proton transport mechanism of membranes so treated. A possible explanation is that meth anol absorbed during the treatment phase is reacting to produce formaldehyde within the electrolyte. Since methanol is permeable through a fresh membrane, it may be present in th e electrolyte to some depth. Methanol can react in an anode reaction to produce hydroge n and formaldehyde, followed by the hydrogen being transported through the elect rolyte as in the simple diss ociation of molecular hydrogen: e H HCHO OH CH2 23 If this reaction is not in equilibrium, which gi ven the finite amount of absorbed methanol that would remain following the treatment, a reductio n in observed voltage can occur, as protons evolved by this reaction produce a potential b ack through the electrolyte, similar to an activation polarization. It is difficult to estimate the exact amount that this may actually reduce the observed potential, howev er, because of lack of data of the amount of methanol so entrained. The ideal potential of a half-cell with this reaction is only 0.189 volts, but if the amount of methanol is greater than the available r eaction sites for hydrogen as a fuel gas, there may be a stoichiometrically greater amount of methanol reacting than hydrogen, causing a larger voltage. Other possibilities include the creation of an oxide or other chemical reaction between methanol and CsHSO4, or poisoning of the pl atinum catalyst. The crea tion of an oxide layer should not change the open cell voltage of the cell, but may produce a profound lowering of

PAGE 107

99 voltage with current due to Ohmic polarizati on or activation polarization due to changing surface chemistry. Catalyst poisoning essentially competes with hydrogen for reaction sites, which also is seen primarily at higher currents but since the catalyst is part of the electrode assembly, poisoning of the catalyst might al so reduce observed voltage by creating an electric potential gradient within the electrode. When runs were continued with hydrogen and oxygen with treated membranes for longer periods, observed potentials increased. If the lo wered potential was due to the formation of formaldehyde, than the gradually lowering metha nol concentration coul d explain this rise. Cell potentials never returned to those of fresh membranes, s uggesting that additional losses due to chemical reaction of methanol and CsHSO4 or catalyst poisoning might be present. Electrolysis reactions for methanol and steam were also successful. Methanol electrolysis began at 0.4 volts, but consiste nt current was not seen until th e potential was raised to over 0.6 volts, after which current increased linearl y with potential. Maximum current density was achieved at approximately 6 volts, caused by the li mits of gas transport either to or from the surface of the membrane. Electrolysis of steam first showed current as low as 0.6 volts, significantly lower than the expected 1.2 or greater. Notably, the linea r relation of current to voltage increases significantly at voltage higher than 1.3 volts. The lower current fr om 0.6 to 1.3 volts is most likely the result of methanol in the feed (10% methanol was used in the feed to prevent CsHSO4 from losing structural integrity). Afte r the potential was raised to allow the

PAGE 108

100 dissociation of water, higher currents were observed. Like with methanol electrolysis, a maximum current was eventually achieved due to limits of gas transport to and from the surface of the membrane. The successful use of CsHSO4 in fuel cell and electrolysis applications in this work represents a confirmation of the feasibility of CsHSO4 as a proton conducting electrolyte in real applications. As of this work, the successful use of CsHSO4 has been reported at the California Institute of Technology2, 4. Data presented in this work affirms this finding. 5.4 Future Directions The ability to construct workable applications for CsHSO4 in a laboratory setting shows the potential of CsHSO4 as a proton conducting electrolyt e has value for future research. Problems with fuel gas permeability, chemical stability in the superprotonic phase, and current density have to be overcome, but there is ample evidence that a CsHSO4 based MEA could provide significant advant age over current commercial ME As in specialty applications.

PAGE 109

101 Although its use in competitive fuel cells is in question due to low current densities and losses due to polarization, CsHSO4 may still prove useful in specialty applications where its operational temperature range ( 144 to 180 C) allows for a special advantage that cannot be realized by existing electrolyt es like Nafion or phosphoric aci d. One example would be the electrolysis of hydrogen sulfide, a toxic byproduct of certain el ectricity generation processes that must be aggressively treated with stack scrubbers to conform to EPA regulations. H2S can be electrolyzed in a reaction sim ilar to the electrolysis of steam: S H S H 2 2 Sulfur is a liquid at temperatur es in excess of 115 C, allowing product sulfur to pour out of an electrolysis cell, where it might be collected as a commer cial product. With sulfur and hydrogen as reaction products, this process might be considered as both a pollution control step and a commercial chemi cal production step. As an a dded advantage, reaction and separation is combined into a single pro cess negating the need for an additional gas separation step. Other specialized reactions requiring ion mobility from heterogeneous reaction sites within the operational temperature range may be possible. To determine any future feasibility of CsHSO4 in any type of elec trolysis process, the problems identified within this work would need to be addressed. The possibility of using a gaseous organic to improve permeability charac teristics was explored. Unfortunately, this procedure produced unforeseen eff ects once performed on an MEA in a real application. It is currently unclear whether the loss of permeability and the loss of performance (open cell

PAGE 110

102 voltage) are related, or are due to separate effects. It is also unclear whether the change to CsHSO4 as a result of exposure to methanol is phys ical, such as crystal growth, dehydration, or the absorption of methanol, or if the change is chemical, such as the production of cesium methyl sulfate: O H SO CH Cs CsHSO OH CH2 4 3 4 3 This reaction is similar to that already reported in literature40 as possible between methanol and sulfuric acid, producing me thyl hydrogen sulfate (this reac tion may also be occurring with unreacted sulfuric acid st ill present in the electrolyte): O H HSO CH SO H OH CH2 4 3 4 2 3 Another possibility is poisoning of the platinum catalyst when exposed to methanol in the fuel cell configuration, which would not be a consideration with earlier permeability tests which did not include electrode or catalyst layers. Whatever the exact effect that methanol has on CsHSO4 membranes, there is no denying that the procedure improves impermeability. It is unclear if this effect is due to a chemical change in the electrolyte upon exposure to methanol, or a physical change. It is known that CsHSO4 crystallizes out of aqueous solution when an organic such as methanol is introduced. A similar physical reaction of crystal growth promotion may be occurring when the polycrystalline sample is exposed to gaseous methanol. The ability of CsHSO4 to absorb

PAGE 111

103 moisture from the laboratory environment was a constant concern during experiments. If water is absorbed into the membrane duri ng normal handling procedures (deliquescence), this may partially dissolve CsHSO4 within the membrane. When exposed to methanol, this partially dissolved CsHSO4 would then precipitate back out of solution, its crystal growth sealing gaps between grains within the polyc rystalline sample. Although the interaction with methanol seems to produce unintended losses of pe rformance, the possibility exists that other gaseous organic solvents may also have a beneficial effect on CsHSO4 or other solid acid electrolytes without th ese losses. The interaction of vari ous organic vapors on different solid acid electrolytes and catalysts for permeability and performance characteristics is an area of future investigation.

PAGE 112

104 References [1] T. Norby, The Promise of Protonics, Nature, 410, 877-878, 2001. [2] S. M. Haile, D. A. Boysen, C. R. I. Chishol m, and R. B. Merle, Solid Acids as Fuel Cell Electrolytes, Nature, 410(6831), 910–913, 2001. [3] X. Ke and I. Tanaka, Proton Tr ansfer Mechanism in Solid CsHSO4 by First-Principles Study, Solid State Ionics, 172(1-4), 145-148, 2004. [4] D. A. Boysen, Superprotonic Solid Acids: Structur e, Properties, and Applications, California Institute of Technology, 2004. [5] S. Hayashi and M. Mizuno, Proton Diffusion in the Superprotonic Phase of CsHSO4 Studied by 1H NMR Relaxation, Solid State Ionics, 171( 3-4), 289-293, 2004. [6] M. Mizuno and S. Hayashi, Proton Dynamics in Phase II of CsHSO4 Studied by 1H NMR, Solid State Ionics, 167(3-4), 317-323, 2004. [7] A. V. Belushkin, R. L. McGreevy, P. Ze tterstrom and L. A. Shivalov, Mechanism of Superprotonic Conductivity in CsHSO4, Physica B: Condensed Matter, 241-243, 323325, 1997. [8] W. Mnch, K. D. Kreuer, U. Traub a nd J. Maier, Proton Transfer in the ThreeDimensional Hydrogen Bond Network of the High Temperature Phase of CsHSO4: A Molecular Dynamics Study, Journal of Molecular Structure, 381(1-3), 1-8, 1996. [9] T. Norby, M. Friesel and B. E. Malland er, Proton and Deuteron Conductivity in CsHSO4 and CsDSO4 by In Situ Isotopic Exchange, Solid State Ionics, 77, 105-110, 1995. [10] L. Kirpichnikova, M. Polomska, J. Wolak a nd B. Hilczer, Polarized Light Study of the CsHSO4 and CsDSO4 Superprotonic Crystals, Solid State Ionics, 97(1-4), 135-139, 1997. [11] A. V. Belushkin, M. A. Adams, S. Hu ll and L. A. Shuvalov, P-T Phase Diagram of CsHSO4. Neutron Scattering Study of Structure and Dynamics, Solid State Ionics, 77, 9196, 1995. [12] J. Baran and M. K. Marchewka, Vibrat ional Investigation of Phase Transitions in CsHSO4 Crystal, Journal of Molecular Structure, 614(1-3), 133-149, 2002.

PAGE 113

105 [13] A. R. Lim, J. H. Chang, H. J. Kim a nd H. M. Park, Phase Transition and Ferroelastic Property Studied by Using the 133Cs Nuclear Magnetic Resonance in a CsHSO4 Single Crystal, Solid State Communications, 129(2), 123-127, 2004. [14] V. Varma, N. Rangavittal and N. R. Rao, A Study of Superionic CsHSO4 and Cs1xLixHSO4 by Vibrational Spectrosc opy and X-ray Diffraction, Journal of Solid State Chemistry, 106, 164-173, 1993. [15] N. Rangavittal, T. N. Guru Row, a nd C. N. R. Rao, Thermally Induced Phase Transitions of CsHSO4: A Reexamination, Journal of Solid State Chemistry, 117, 414415, 1995. [16] B. Baranowski and J. Lipkowski, On the Phase Transitions of Cesium Hydrogen Sulfate (CsHSO4), Journal of Solid State Chemistry, 117, 412-413, 1995. [17] C. R. I. Chisholm, Superprotonic Phase Transitions in Solid Acids, California Institute of Technology, 2002. [18] C. R. I. Chisholm and S. M. Hail e, X-ray Structure Re finement of CsHSO4 in Phase II, Materials Research Bulletin, 35(7), 999-1005, 2000. [19] A. V. Belushkin, M.A. Adams, S. Hull, A.I. Kolesnikov, L.A. Shuvalov, Structure and Dynamics of Different Phases of the Superprotonic Conductor CsHSO4, Physica B, 213&214, 1034-1036, 1995. [20] B. Yang, A. M. Kannan and A. Manthi ram, Stability of the Dry Proton Conductor CsHSO4 in Hydrogen Atmosphere, Materials Research Bulletin, 38(4), 691-698, 2003. [21] I. N. Levine, Physical Chemistry, McGraw-Hill Book Company, New York, 1988. [22] P. Schuster, G. Zundel, C. Sandorfy, The Hydrogen Bond, North-Holland Publishing Company, Amsterdam, 1976. [23] R. E. Hummel, Electronic Properties of Materials, Springer-Verlag, Berlin Heidelberg, 1993. [24] U.S. Department of Energy, Fuel Cell Handbook, National Energy Technology Laboratory, Strategic Center for Natural Ga s, Morgantown WV, Pittsburgh, PA, Tulsa, OK, 2002. [25] A. J. Appleby and F. R. Foulkes, Fuel Cell Handbook, Van Nostrand Reinhold, New York, 1989. [26] J. Larminie and A. Dicks, Fuel Cell Systems Explained, 2nd Ed., Wiley, West Sussex (2003). [27] C. Berger, Handbook of Fuel Cell Technology, Prentice-Hall, Englewood Cliffs, NJ, 1968.

PAGE 114

106 [28] J. R. MacDonald, Impedance Spectroscopy, Wiley, New York (1987). [29] A. Guinier, X-ray Diffraction in Crystals, Impe rfect Crystals, and Amorphous Bodies, Dover, New York, 1994. [30] B. D. Cullity and S. R. Stock, Elements of X-ray Diffraction, Prentice-Hall, Upper Saddle River, NJ, 2001. [31] J. Lipkowski, B. Baranowski, and A. Lunden, Pol. J. Chem., 67, 1867, 1993. [32] V.I. Nefedov, X-ray Photoelectron Spectroscopy of Solid Surfaces, Utrecht, Netherlands, 1988. [33] J. G. Ferreira and M. T. Ramos, X-ray Spectroscopy in Atomic and Solid State Physics, Plenum Press, New York, 1988. [34] J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Bomben, Handbook of X-ray Photoelectron Spectroscopy, Physical Electronics, Inc., USA (1995). [35] V. Tricoli, Proton a nd Methanol Transport in Poly (perfluorosulfonate) Membranes Containing Cs+ and H+ Cations, J. Electrochem. Soc., 145, 3798 (1998). [36] J. Crank and G.S. Park, Diffusion in Polymers, Academic Press, London (1968). [37] M. Mulder, Basic Principles of Membrane Technology, Kluwer Academic Publishers, Dordrecht, Netherlands (1991). [38] R. L. Grob, Modern Practice of Gas Chromatography, Wiley-Interscience, New York, 1995. [39] D. Ambrose, Gas Chromatography, Butterworths, London, 1971. [40] L. L. Van Loon and H. C. Allen, Methan ol Reaction with Sulfuric Acid: A Vibrational Spectroscopic Study, J. Phys. Chem. B, 108(45), 17666-17674, 2004.