USF Libraries
USF Digital Collections

Endurance materials for hydrogen sulfide splitting in electrolytic cell

MISSING IMAGE

Material Information

Title:
Endurance materials for hydrogen sulfide splitting in electrolytic cell
Physical Description:
Book
Language:
English
Creator:
Mbah, Jonathan Chinwendu
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Solid Acid
Permeability
CsHSO4
Ionic Conductivity
RuO2/CoS2
Dissertations, Academic -- Chemical Engineering -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: This study describes the development of a novel thin membrane exchange assembly (MEA) from a solid acid material, cesium hydrogen sulfate (CsHSO₄), and from a composite anode electrocatalyst for electrolytic splitting of (100 %) H₂S feed content gas operating at 135 kPa and 150 °C. A new class of anode electrocatalyst with the general composition, RuO₂/CoS₂, and an improved proton conductor, CsHSO₄, have shown great stability and desired properties at typical operating conditions. This configuration demonstrated stable electrochemical operation for 24 h with a (100 %) H₂S fuel stream at 423 K. This same system showed a maximum current density of (19 mA/cm²) at 900 mV. The performance of this new anode electrocatalyst when compared to that of Pt black investigated in a previous study showed an overall superiority in application.We have achieved a 30 % reduction in the overall system performance by fabricating a thin (200 μm) CsHSO₄ electrolyte, which reduced the whole MEA thickness from 2.3 mm to 500 μm. The result of permeability measurements proved that this thin solid electrolyte is impermeable to H₂S gas and physical integrity was preserved throughout the experimental period. Further resistance losses were compensated by using a high energy planetary milling system to enhance the ionic conductivity of CsHSO₄. The difference in stability and electrochemical performance of these cells compared to that of Pt anode based systems is directly attributable to the anode materials developed in this project. Factorial experiments were used to characterize the effect of controllable process variables (electrolyte thickness, time, age of the electrolyte) on the cell current density and interfacial polarization resistances.As expected, cell current density and interfacial polarization resistances were a function of electrolyte thickness and age. Nevertheless, the effect of electrolyte thickness has a more prominent effect on the measured parameters. In addition, these experiments were used to identify regions of optimum system performance. Tafel plots were constructed to investigate the kinetic behavior of various anode based electrocatalysts. Exchange current densities, which are directly a measure of the electrochemical reaction, increased with RuO₂/CoS₂-based anodes. These experiments also suggested that high levels of feed utilization were possible using these materials. This was an impressive result considering the drastic improvement in electrochemical performance, current density, and sulfur tolerance compared to the other anode configurations.
Thesis:
Dissertation (Ph.D.)--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 Jonathan Chinwendu Mbah.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 123 pages.
General Note:
Includes vita.

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 - 002004905
oclc - 352926362
usfldc doi - E14-SFE0002693
usfldc handle - e14.2693
System ID:
SFS0027010:00001


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 2200409Ka 4500
controlfield tag 001 002004905
005 20101007125604.0
007 cr mnu|||uuuuu
008 090527s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002693
035
(OCoLC)352926362
040
FHM
c FHM
049
FHMM
066
(S
090
TP145 (Online)
1 100
Mbah, Jonathan Chinwendu.
0 245
Endurance materials for hydrogen sulfide splitting in electrolytic cell
h [electronic resource] /
by Jonathan Chinwendu Mbah.
260
[Tampa, Fla] :
b University of South Florida,
2008.
500
Title from PDF of title page.
Document formatted into pages; contains 123 pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2008.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: John T. Wolan, Ph.D.
520
ABSTRACT: This study describes the development of a novel thin membrane exchange assembly (MEA) from a solid acid material, cesium hydrogen sulfate (CsHSO), and from a composite anode electrocatalyst for electrolytic splitting of (100 %) HS feed content gas operating at 135 kPa and 150 C. A new class of anode electrocatalyst with the general composition, RuO/CoS, and an improved proton conductor, CsHSO, have shown great stability and desired properties at typical operating conditions. This configuration demonstrated stable electrochemical operation for 24 h with a (100 %) HS fuel stream at 423 K. This same system showed a maximum current density of (19 mA/cm) at 900 mV. The performance of this new anode electrocatalyst when compared to that of Pt black investigated in a previous study showed an overall superiority in application.We have achieved a 30 % reduction in the overall system performance by fabricating a thin (200 m) CsHSO electrolyte, which reduced the whole MEA thickness from 2.3 mm to 500 m. The result of permeability measurements proved that this thin solid electrolyte is impermeable to HS gas and physical integrity was preserved throughout the experimental period. Further resistance losses were compensated by using a high energy planetary milling system to enhance the ionic conductivity of CsHSO. The difference in stability and electrochemical performance of these cells compared to that of Pt anode based systems is directly attributable to the anode materials developed in this project. Factorial experiments were used to characterize the effect of controllable process variables (electrolyte thickness, time, age of the electrolyte) on the cell current density and interfacial polarization resistances.As expected, cell current density and interfacial polarization resistances were a function of electrolyte thickness and age. Nevertheless, the effect of electrolyte thickness has a more prominent effect on the measured parameters. In addition, these experiments were used to identify regions of optimum system performance. Tafel plots were constructed to investigate the kinetic behavior of various anode based electrocatalysts. Exchange current densities, which are directly a measure of the electrochemical reaction, increased with RuO/CoS-based anodes. These experiments also suggested that high levels of feed utilization were possible using these materials. This was an impressive result considering the drastic improvement in electrochemical performance, current density, and sulfur tolerance compared to the other anode configurations.
653
Solid Acid
Permeability
CsHSO4
Ionic Conductivity
RuO2/CoS2
690
Dissertations, Academic
z USF
x Chemical Engineering
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2693



PAGE 1

Endurance Materials for Hydrogen Sulf ide Splitting in Electrolytic Cell by Jonathan Chinwendu Mbah A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemical a nd Biomedical Engineering College of Engineering University of South Florida Major Professor: John T. Wolan, Ph.D. Yogi Goswami, Ph.D. Elias Stefanakos, Ph.D. Vinay Gupta, Ph.D. Matthias Batzill, Ph.D. Burton Krakow, Ph.D. Date of Approval: November 5, 2008 Keywords: Solid Acid, Permeability, CsHSO4, Ionic Conductivity, RuO2/CoS2 Copyright 2008, Jonathan Chinwendu Mbah

PAGE 2

ACKNOWLEDGEMENTS To my Creator from whom all good things co me, I glorify. Thanks to all my advisors, colleagues, friends, and family member s for making this journey possible.

PAGE 3

i TABLE OF CONTENTS LIST OF TABLES ............................................................................................................. ii i LIST OF FIGURES ........................................................................................................... iv ABSTRACT ...................................................................................................................... vii 1 INTRODUCTION ........................................................................................................... 1 2 BACKGROUNDS AND LI TERATURE REVIEW ..................................................... 10 2.1 Electrochemical Cell Description ............................................................................. 10 2.2 Essential Anode Properties ....................................................................................... 11 2.3 Phases and Ionic Conductivity of CsHSO4 ............................................................... 12 2.4 Previous Work on H2S Splitting ............................................................................... 18 2.5 Gas Permeability ....................................................................................................... 18 3 THEORY ...................................................................................................................... 21 3.1 Physical Properties and Sources of H2S .................................................................... 21 3.2 H2S Electrochemical Splitting .................................................................................. 21 3.3 Electrochemical Thermodynamics ............................................................................ 23 3.4 Nernst Equation ........................................................................................................ 24 3.5 Cell Performance ...................................................................................................... 2 5 3.5.1 Activation Polarization ( act) ......................................................................... 26 3.5.2 Potential and Rate: Butler-Volmer Equation ................................................. 26 3.5.3 Concentration Losses ( conc) .......................................................................... 27 3.5.4 Ohmic Losses ( Ohmic) .................................................................................... 28 3.6 Physical Meaning of Conductivity ............................................................................ 29 3.7 Ionic Conductivity in a Crystalline Solid Electrolyte ............................................... 30 3.8 Electrochemical Impedance Spec troscopy (EIS) Characterization .......................... 33 4 EXPERIMENTAL EQUIPMENT AND TECHNIQUES ............................................. 36 4.1 Introduction ............................................................................................................. ... 36 4.2 Experimental Apparatus............................................................................................ 36 4.3 Preparation of Anode Electrocatalyst ....................................................................... 39 4.4 Cell Fabrications ....................................................................................................... 40 4.5 Permeability of H2S on CsHSO4 Pellet ..................................................................... 41 4.6 Experimental Procedure ............................................................................................ 43 4.7 Synthesis and Ionic Conductivity of CsHSO4 .......................................................... 45 4.8 Design of Experiments .............................................................................................. 47

PAGE 4

ii 5 RESULTS AND DISCUSSION .................................................................................... 49 5.1 XRD Structural Characterization of Modified and Unmodified CsHSO4 ................ 49 5.2 Simultaneous Gravimetric and Calorimetric Analysis of CsHSO4........................... 51 5.3 Ionic Conductivity Modeling of CsHSO4 ................................................................. 53 5.4 Permeability of H2S on CsHSO4 Membrane ............................................................ 60 5.4.1 Morphological Characteri zation of the Surface ............................................. 60 5.4.2 Gas Chromatography and Mass Spectrometry Analysis ............................... 61 5.4.3 Micropore Analysis ........................................................................................ 63 5.4.4 XRD Analysis ................................................................................................ 64 5.4.5 Resistance Measurements of CsHSO4 Membranes ....................................... 65 5.4.6 Permeability ................................................................................................... 68 5.5 Electrochemical Synthesis ........................................................................................ 70 5.5.1 Current Density from H2S Splitting ............................................................... 70 5.5.2 Electrolysis of H2S ......................................................................................... 71 5.5.3 Material Balance ............................................................................................ 75 5.5.4 Stability of H2S Electrolytic Cell Anode Materials ....................................... 76 5.5.5 Electrochemical Performanc e of Anode Electrocatalysts .............................. 80 5.5.6 Resistivity of Anode Catalysts ....................................................................... 83 5.5.7 Fuel Utilization .............................................................................................. 84 5.5.8 Tafel Slope and Exchange Current Densities for Anode Configurations ...... 86 5.5.9 Factorial Experiment ...................................................................................... 87 5.5.9.1 Effect of Process Variab les on Cell Resistances for RuO2/CoS2 ............ 87 5.5.9.2 Effect of Process Variab les on Current Density for RuO2/CoS2 ............ 91 6 CONCLUSIONS AND RECOMMENDATIONS ........................................................ 93 6.1 Increase Reactant Concentration ............................................................................... 95 6.2 Electron Collection ................................................................................................... 9 5 REFERENCES ................................................................................................................. 98 APPENDICES .................................................................................................................104 Appendix A Cell Preparation and H2S Splitting ...........................................................105 A.1 CsHSO4 Synthesis ..........................................................................................105 A.2 Pellets Preparation..........................................................................................106 A.3 Computation of H2S Flowrate Required for both the 0.5’’ and 2’’ Cells ......108 A.4 Gasket and O ring for the ’’ Cell ................................................................111 A.5 System Pressure .............................................................................................111 A.6 Material Balance ............................................................................................114 A.7 Product Analysis ............................................................................................117 A.8 Graphical Comparison of the Cl aus and the Electrolytic Processes ..............121 A.9 LCR Calibration .............................................................................................122 A.10 Thermochemistry Data ...................................................................................123 ABOUT THE AUTHOR ....................................................................................... End Page

PAGE 5

iii LIST OF TABLES Table 4.1 Various catalyst s configurations of anode catalysts ..................................44 Table 4.2 Factors, treatment levels, and effects for metal sulfides. ...........................47 Table 4.3 Two-level 3-Factor full-factorial experime nt design pattern. ....................48 Table 5.1 Comparison of crysta l parameters for unmodified and modified CsHSO4 ......................................................................................................51 Table 5.2 Various anode configurat ions and their BET catal ytic active areas. .........79 Table 5.3 Cell area sp ecific ohmic resistance (RA) .................................................84 Table 5.4 Conversion effi ciency for severa l electrolytes. ..........................................86 Table A.10.1 Thermochemistry data (evaluated at T = 298 K) .....................................123

PAGE 6

iv LIST OF FIGURES Figure 2.1 Simplified schema tic of an electrochemical cell system ........................11 Figure 2.2 Proton conduc tion mechanism for solid acids (CsHSO4). ....................13 Figure 2.3 Viscosit y of sulfur, conductivity of CsHSO4 vs. Temperature. ..............17 Figure 3.1 Basic operating principle of H2S electrolysis ..........................................22 Figure 3.2 A sinusoidal voltage perturbation/ result ing sinusoidal current response...................................................................................................35 Figure 4.1 Schematic of electrolytic splitting cell. ................................................. 38 Figure 4.2 Schematic diagram of a fabricated cell (CsHSO4-based MEA). ........... 41 Figure 4.3 Illustration of the two-probe AC conductivity measurement setup. ...................................................................................................... 46 Figure 5.1 X-ray diffractogr ams of (a) unmodified and (b), modified CsHSO4. ................................................................................................. 50 Figure 5.2 Simultaneous differential scanning calorimetry and thermogravimetric analysis (SDT) measurement. ................................. 52 Figure 5.3 The elements of th e equivalent RC circuit and the fitting results of CsHSO4 sample. ............................................................................... 55 Figure 5.4 Fitting the c onductivity to Arrhenius law upon heating at 3 oC min-1. ...................................................................................................... 57 Figure 5.5 Nyquist plots at various temperatures upon heating (CsHSO4). ........ 59 Figure 5.6 Atomic force microscope images of CsHSO4 membranes used in the permeation study. ............................................................................. 61 Figure 5.7 Qualitative comp arison of permeability at 150 C by GC-MS instrument. ............................................................................................. 62 Figure 5.8 A typical Saito-Foley (SF) DV (d) method of pore size distribution isotherm. ............................................................................. 64

PAGE 7

v Figure 5.9 XR D diffractogram of CsHSO4 with different sample history. ............. 65 Figure 5.10 Nyquist plots at and above CsHSO4 superprotonic transition temperatures (Tsp) for two membrane thicknesses 1 mm and 0.2 mm. ...................................................................................................... 67 Figure 5.11 Permeability and pr essure drop as a function of time for CsHSO4 membrane. .............................................................................................. 69 Figure 5.12 Current density as a function of vo ltage generated with anode catalyst S3.. ............................................................................................ 71 Figure 5.13 Relationship of H2S conversion to time on stream with a voltage 900 mV................................................................................................... 72 Figure 5.14 Relationship of curre nt to time during a period of 12 h of operation with a voltage 900 mV.. ......................................................... 74 Figure 5.15 XRD pattern for (RuO2/CoS2 composite).............................................. 77 Figure 5.16 Multipoint B ET plot for anode metal sulfide (RuO2/CoS2) nanocomposite. ...................................................................................... 78 Figure 5.17 SEM images of surfaces of electrode. ................................................... 80 Figure 5.18 Testing for four cel ls with different anodes (Vcell = 0.9V, T = 150 oC, Fuel = 0.25 cm3/min) H2S was the fuel.. .................................. 81 Figure 5.19 Current density for different anode configurations with 100 % H2S feed gas content.. ............................................................................ 83 Figure 5.20 Tafel plots for anode configurations at operating temperature 150 oC and 100 % H2S feed content. ............................................................ 87 Figure 5.21 Measured effects of three proces s variables on cell polarization resistances.. ............................................................................................ 90 Figure 5.22 Factorial results showing the interaction effect between electrolyte age and electrolyte thickness on cell polarization resistance. ............................................................................................... 91 Figure 5.23 Measured effects of three process variab les on cell current density. ................................................................................................... 92 Figure A.5.1 A LabView block di agram for monitoring system pressure .................112 Figure A.5.2 A sample LabView repr esentation of an electrolytic system ...............113

PAGE 8

vi Figure A.7.1 X-ray diffraction comparison of sulfur .................................................118 Figure A.7.2 SEM and EDS images of sulfur ............................................................119 Figure A.7.3 Differential scanning calorimetry comparison of sulfur melting points .....................................................................................................120 Figure A.7.4 Gas chromatogr aph analysis of hydrogen produced .............................121 Figure A.8.1 Comparison of the Claus and the electrolytic processes ......................121 Figure A.9.1 LCR calibration wi th known resistors and capacitor.............................122

PAGE 9

vii ENDURANCE MATERIALS FOR HYDROGEN SULFIDE SPLITTING IN ELECTROLYTIC CELL Jonathan Chinwendu Mbah ABSTRACT This study describes the development of a novel thin membrane exchange assembly (MEA) from a solid acid materi al, cesium hydrogen sulfate (CsHSO4), and from a composite anode electrocatalyst for electrolytic splitting of (100 %) H2S feed content gas operating at 135 kPa and 150 oC. A new class of anode electrocatalyst with the general composition, RuO2/CoS2, and an improved proton conductor, CsHSO4, have shown great stability and desired properties at typica l operating conditions. This configuration demonstrated stable electrochemical operation for 24 h with a (100 %) H2S fuel stream at 423 K. This same system showed a ma ximum current density of (19 mA/cm2) at 900 mV. The performance of this new a node electrocatalyst when comp ared to that of Pt black investigated in a previous st udy showed an overall superiority in appl ication. We have achieved a 30 % reduction in the overall system performance by fabricating a thin (200 m) CsHSO4 electrolyte, which reduced the w hole MEA thickness from 2.3 mm to 500 m. The result of permeability measurements pr oved that this thin solid electrolyte is impermeable to H2S gas and physical integrity was preserved throughout the experimental period. Further resistance losse s were compensated by using a high energy planetary milling system to enhance the ionic conductivity of CsHSO4. The difference in

PAGE 10

viii stability and electrochemical performance of these cells compared to that of Pt anode based systems is directly attributable to th e anode materials developed in this project. Factorial experiments were used to charac terize the effect of controllable process variables (electrolyte thickness, time, age of the electrolyte) on the cell current density and interfacial polarization resistances. As e xpected, cell current density and interfacial polarization resistances were a function of electrolyte thic kness and age. Nevertheless, the effect of electrolyte thickness has a more prominent effect on the measured parameters. In addition, these experiments were used to identify regions of optimum system performance. Tafel plots were constructed to investigate the kinetic behavior of various anode based electrocatalysts. Exchange current densitie s, which are directly a measure of the electrochemical reaction, increased with RuO2/CoS2-based anodes. These experiments also suggested that high levels of feed ut ilization were possible using these materials. This was an impressive resu lt considering the drastic im provement in electrochemical performance, current density, and sulfur tolerance compared to the other anode configurations.

PAGE 11

1 1 INTRODUCTION Hydrogen sulfide (H2S) is extremely toxic and undesirabl e; yet over 12 million tons of it is produced annually in the US alone as a result of many industrial activities1. The rotten egg like smell accompanying it is disagreeable and the ability to detect it via smell diminishes rapidly with exposure, hence one can dangerously be exposed to a high concentration and yet be oblivious. Highl y flammable under normal conditionshydrogen sulfide readily forms explosive mixtures w ith air (4 % lower explosion limit, LEL, H2S). Many processes are available to eliminate H2S but most of these processes are not efficient and generate other pollutants as well2. In addition, most processes are multistage which may require a huge capital inve stment. A common process used in H2S splitting is the Claus process. In this process H2S is extracted along with carbon dioxide in a stream called acid gas. Partial oxidation of the acid gas with air yields elemental sulfur and water with a waste stream of dilute carbon di oxide in nitrogen. Electrolytic splitting of the extracted hydrogen sulfide can yield sulf ur and hydrogen (instead of sulfur and water). The carbon dioxide is separated before electrolysis, which leaves a concentrated carbon dioxide waste stream which is easie r to sequester. The small Gibbs energy possessed by H2S enables it to cleanly split into el emental sulfur and diatomic hydrogen with a low energy and voltage requirement. The energy benefit of electrolytic splitting of

PAGE 12

2 the H2S is illustrated in Eqns. (1.1, 1.2). The value of the hydrogen makes the system more profitable. H2S (g) S (l) +H2 (g) – 2 Faradays @ 0.26V G o = 8.9 kcal/mole @ 400K (1.1) H2O H2 + O2 – 2 Faradays @ 1.2V G o = 57 kcal/mole (1.2) Therefore, removal of H2S from process gas streams electrochemically has many economic and environmental benefits. At present H2S is not employed as a hydrogen or high-grade energy resource such as found in methanol or ethanol. Therefore, except for a process such as oil refining where a use for the heat generated by the Claus process is present, the capital investment can not be economically justified. Ot her disadvantages of the Claus process are that its hydrogen sulfid e conversion efficiency is only about 92 percent and other pollutants such as CS2 and COS are normally also produced by the Claus process3. Thus, for many applications, H2S decomposition by the Claus method is not practical. An electrolysis process requires a membrane exchange assembly (MEA) to provide both the reaction interface and the ion migration route. In a ddition, it provides a good surface for electron dispersal away from the reac tion interface. Therefor e, a good electrolyte capable of splitting H2S gas to liquid sulfur and hydrogen is of utmost importance. The preference of solid electrolytes over liquid electrolytes has been proven in previous publications4. There is currently no published article on H2S splitting to produce liquid sulfur and hydrogen using a solid acid electr olyte such as the one we reported here, but many articles have been published on H2S splitting using differe nt solid electrolytes4. In

PAGE 13

3 most of these processes, H2S containing gas was utilized as a feed in fuel cells to generate power and produces sulfur and H2O. The need for a solid electrolyte that maintain s its integrity and prot on conductivity at high temperatures and also allows th e anode product, sulfur to exist in liquid form so that it can easily flow out of the electrolytic cell without hi ndrance cannot be overemphasized. A group of solid acid electrolytes (SAE) have e xhibited capability in there use as fuel cell membranes5. They belong to inorganic crystals, with intermediate properties between normal acids, such as H2SO4 or H3PO4, and normal salts, such as K2SO4 6. Solid acids still have the capacity to act as proton donors ev en with only some of their normal acid’s hydrogen atoms replaced7. To electrolyze H2S we exploited these recent developments in SAE whose proton conductivities rise rapidly with temperature. One other important components of the ME A are the electrocatalysts required to completely split H2S into sulfur and hydrogen prot on at the reaction interface. Electrocatalysis is the term us ed to indicate the catalysis of electrode reactions. This can be accomplished by the action of the electrode material8. In harmony with the field of heterogeneous electrocatalysis, heterogeneous el ectrocatalysis deals with the effect of the electrode material on the rate and the mechanism of electrode reactions9. In order to be a suitable candidate for the electrochemical splitting of H2S, an anode material must possess good electrical conductivity and sulfur to lerance at high temperatures, in addition to good catalytic activity. Pt or Pt based anode catalysts have good catalytic activity, but degrade over time in H2S streams. Moreover; most metals and metal oxides are severely

PAGE 14

4 corroded by H2S at elevated temperatures. A pyrochlore-based anode material, Gd2Ti1.4Mo0.6O7, showed remarkable tolerance to sulfur-containing fuels used in fuel cells application. The anode/electrolyte interfaci al resistance was only 0.2 cm2 at 950 C in a fuel gas mixture of 10 % H2S and 90 % H2, demonstrating a peak power density of 342 mW cm–2. The fuel cell operated under these conditions continuously for 6 days without any observable degradation, suggesting that Gd2Ti1.4Mo0.6O7 anode exhibits not only excellent stability but also good catalytic activity toward the splitting of H2S containing gas10. But a major draw back in this a pplication is the high temperature of operation. One attempt to overcome this problem in a fuel cell utilizing H2S is proposed in United States patent11. In this is disclosed a fuel cell, in which a redox couple is used as the negative electrolyte which is oxidized at the fuel cell anode and then subjected to reduction outside the electrolyt ic cell by reaction with H2S in a remotely located reaction column. The sulfur formed in the reaction colu mn is removed from the electrolyte before the reduced electrolyte is recirculated back in to the fuel cell. Consequently, the sulfur is prevented from inactivating the platinum or Ra ney nickel anode catalyst of the fuel cell. A major drawback of this patent anode catalys t is due to the fact that performance is generally based upon a crystalline structure. In a crystalline struct ure the catalytically active sites which provide the catalytic effect of such materials result primarily from accidently occurring, surface ir regularities which interrupt the periodicity of the crystalline lattice. A few examples of such surface irregularities are dislocation sites, crystal steps, surface impurities and foreign adsorbates.

PAGE 15

5 A major problem with a crystalline structure is that the number of such irregularities forming the catalytically active sites is relatively few and occurs only on the surface of the crystalline lattice12. This results in the catalytic material having a density of catalytically active sites which is relatively low. Thus, the catalytic efficiency of the material and the device in which it is utilized is substantially less than that which would be possible if a greater number of catalytic ally active sites were available for the hydrogen sulfide decomposition or other de sired reaction. For these reasons, high catalytic efficiency from a re latively low cost material whic h is resistant to poisoning and stable in the H2S cell environment, remain a desired re sults which must be attained before there will be commercial utilization of this devices. Catalytic materials must have a high density of active sites and have improved catalytic activity. It is therefore intuitive to consider me tal sulfides as alternative potentially useful catalysts. These materials having increased catalytic activity13 serve to increase operating efficiencies to thereby reduce operating cost s and result in more complete removal of hydrogen sulfide contaminants. Th e catalytic material s are also resistant to poisoning primarily due to their increased density of catalytically active si tes and can provide a stable performance over a long period of time14. The increased numbers of catalytically active sites also enables the material s to be more resistant to poisoning3. This is because with these materials a certain number of catalytically active si tes can be sacrificed to the effects of poisonous species while a large number of unpoisoned sites still remain to provide the desired reaction.

PAGE 16

6 Transition metals form good homogeneous or heterogeneous catalysts, for example iron is the catalyst for the Haber process. Vanadi um (V) oxide used for the contact process, nickel is used to make margarine and plat inum is used to speed up the manufacture of nitric acid15. This is because they are able to fo rm numerous oxidation states, and as such, are able to form new compounds during a reac tion providing an alternative route with a lower overall activation energy. The improved catalytic activity can be accomp lished by manipulating the local chemical order and hence the local structural order by the incorporation of selected modifier elements into a host matrix to create the desired disordered material16. The catalytic materials include at least one element forming a host matrix and at least one modifier element intimately incorporated into the hos t matrix. The element or elements forming the host matrix include at least one transiti on element. The host matrix or substrate can also be formed from or with carbon. The hos t matrix element is selected from the group consisting of C, Co, Mo, Fe Ru, Ti, W, Cu and Pb and modifying element is incorporated to provide a sulfide or oxide of said host matrix element. The modifier elements include sulfur or oxygen to form a sulfide, oxide of the transition metal or metals of the host matrix. The incor poration of the modifier element or elements acts to disorder the structure of the material and to create local structural chemical environments which are capable of acting as catalytically active sites for the hydrogen sulfide splitting reaction. The utilization of a disordered structure allows the creation of an increased density and a wide spectrum of catalytically active site s to yield materials

PAGE 17

7 which operate at high catalyt ic efficiency and are more resistant to poisoning and corrosion17. Metal sulfides ar e inexpensive when compared with precious metals like Pt. Among them, CoS2, a hexagonal layered n-type semiconducto r, is a widely used component in catalysts for a variety of hydrogenation/de hydrogenation and hydrodesulfurization processes in the petroleum i ndustry. Bimetallic sulfides of molybdenum with cobalt, iron, and nickel are each active catalysts for hydrodesulfurization processes18. The activity is attributed in part to the presence of CoS2-like aggregates. An impor tant attribute of these catalysts is sulfur tolerance. Although CoS2 has not been studied intensively for H2S decomposition, MoS2 has been shown to be an eff ective catalyst for reversible decomposition of H2S above 600 C, which indicates that MoS2 is chemically stable in a high temperature H2S stream and is not poisoned. The inte resting thing about Co is that it is one of the three noteworthy elements in the transition metals family. These elements are iron, cobalt, and nickel, and they are th e only elements known to produce a magnetic field. One major issue that affects the electrolyti c cell performance is the thickness of the membrane employed which tends to increase in ohmic resistance as the thickness increases. Therefore, gas permeability testing of electrolytes is critical to the development of ultrathin membranes, where gas leak s can prove catastrophic. Although reducing electrolyte thickness improves the cell performan ce, there are several practical issues that limit how thin the electrolyte can be made. As the electrolyte thickness is reduced, the

PAGE 18

8 crossover of reactants may increase. This l eads to an undesirable parasitic loss which can eventually become so large that furthe r thickness decreases are counter productive19. For solid electrolytes, the membrane cannot be made so thin that structural risks of breaking or developing pinholes become an issue. Membrane failure can result in catastrophic mixing of the fuel and cathodic product (h ydrogen). Even mechanically sound, pinhole-free electrolytes may fail if th e thickness varies considerably across the cell. Thin electrolyte areas may become hot s pots that are subject to rapid deterioration and failure. In addition extremely thin electr olytes (solid or liquid) risk electrical shorting, especially when the electrolyte thickness is on the same order of magnitude as the electrode roughness. Furthermore, part of the electr olyte resistance is associated with the inte rface between the electrolyte and the electrode. This contact resistance is independe nt of electrolyte thickness. Also, the ultimate phys ical limit to solid-electrolyte thickness is given by the electrolyte breakdown properties. This limit is reached when the electrolyte is made so thin that the electric field across the memb rane exceeds the dielectric breakdown field for the material20. The main research goal was to establish a novel environmentally benign one-step pathway for splitting H2S electrolytically by using e ffective MEA to produce hydrogen and liquid sulfur. In order to accomplish this task, we have investigated and developed: (1) a solid acid membrane, (2) fabrication patt ern for a MEA and electr olytic cell, and (3) implementation of exit gas control and mon itoring systems. In th is study, which we

PAGE 19

9 report here, an electrochemical process was applied in the splitting of (100 %) H2S feed content gas to yield liquid sulfur and hydroge n using a thin membrane MEA developed in our laboratory. This work was successful in achieving the electrochemical splitting of H2S in an electrochemical system. Furthermor e, it provides valuable insight into the nature of electrolyte material by incorpora ting a series of electrochemical analytical methods and techniques.

PAGE 20

10 2 BACKGROUNDS AND LITERATURE REVIEW 2.1 Electrochemical Cell Description The beauty of electrochemical devices is that the energy of chemical bonding is converted directly to electri cal energy. This is because el ectrochemical energy conversion is not based on the transfer of heat between a hot and cold reservoir, Carnot limitations are avoided and inherently more efficient processes are, in principle, possible. The basic concept of an electrochemical cell includes a test-bed with porous electrodes in which a solid proton conducting membrane separates the electrodes. A simplified schematic of an electrochemical cell system is shown in Figure 2.1 The process consists of passing H2S gas through the anode chamber to contact a catalytic anode, where electrochemical reaction takes place to produce elemental sulfur protons and electrons. The protons pass through the membrane from the anode chamber to the cathode chambe r, where they react with electrons from the catalytic cathode to produce hydrogen ga s. The electrolyte membrane separating both electrodes is impe rmeable to fuel and product flows, but allows the transport of ionic spec ies between both sides of the cell.

PAGE 21

11 Figure 2.1 Simplified schematic of an electrochemical cell system 2.2 Essential Anode Properties An effective anode material for the H2S electrolysis has to be electronically conductive, chemically and electrochemically stable, and catalytically active to split H2S at system operating conditions. It must pr ovide intimate gas contact at the triple phase boundary (TPB) of the fuel-electrolyte-e lectrode interface a nd a large interfacial area to maximize electrochemical oxidation rates. Thermal expans ivity properties of the anode have to be the same or close to those of other cell components, particul arly the cell electrolyte and current collector to avoid different layers fr om physical damage during system operation. In addition, it has to be chemically inert to ot her cell constituents as this may lead to drop in system performance.

PAGE 22

12 2.3 Phases and Ionic Conductivity of CsHSO4 Solid acids are chemical intermediates between normal salts (a salt that has neither hydrogen (H) nor hydroxyl (OH) in its formula) and normal acids. If a normal acid such as sulfuric acid reacts with a normal salt su ch as cesium sulfate, the product is a solid acid, cesium hydrogen sulfat e (cesium bisulfate): CsSO4 (salt) + H2SO4 (acid) CsHSO4 (solid acid) (2.1) this is the prototypical solid acid used in solid acid fuel cells (SAFCs). Their physical appearances are similar to salts, such as common table salt (NaCl) They show orderly structural arrangements at low temperatures. Nevertheless, at elevated temperatures some solid acids undergo phase transi tions to highly disordered structures which cause the conductivity to jump dramatically21. Cesium hydrogen sulfate (CsHSO4) belong to a group of crystal MeXAO4, family of compounds (where Me = Cs, Rb; X = H, D; A = S, Se) first observed to undergo a phase transition to a so-called supe rionic state. They show a high proton conductivity at elevated-temperature phase, usua lly tagged ‘‘a superprotonic phase’’22,23,24,25. In these compounds, tetrahedral XO4 anions form hydrogen bond networks. The assumption is that the proton transport takes place via the hydrogen bond, acting through by reorientation of the XO4 tetrahedron. Nevertheless, prot on dynamics at molecular is still being investigated. How fast is the reorientation of the HSO4 ions and how fast is the proton transfer between two neighboring SO4 tetrahedra? Are the two motions paired or not? In the first stage, hydrogen bond is broken, whereas the hydrogen bond and the chemical bond are exchanged in the second stage. CsHSO4 is one of the most basic

PAGE 23

13 systems among the solid acids and the phase di agrams and crystal structures have been studied extensively26,27,28,29,30. The bisulfate (HSO4 -) group of CsHSO4 forms a tetrahedron with an oxygen atom at each corner and a hydrogen atom sitting on one of the oxygen. At room temperature, all the sulfate groups are frozen into place. When th e temperature is raised, disorder takes over and the sulfate groups reorient, changing the positions of the hydrogen atoms as they do so (Figure 2.2). The reorientation takes about 1/6 of a minute to complete. Frequently, a proton from one sulfate oxyanion m oves over to the next, with a transfer rate close to 109 Hz31. Importantly, these oxyanions rotate almo st freely and approxima tely in every 101 reorientations or so, they are in exactly the right position for a proton transfer to occur32,33,34. Subsequently, as the material goes through this superprotonic transition, there is a sudden increase in conductivity of several or ders of magnitude. These conductivity values for the acid salts can be matched to those of Nafion and other polymer electrolytes, but at slightly higher temperatures. There are quite a number of such different solid acid compounds exhibiting this kind of behavior. Figure 2.2 Proton conduction mechanism for solid acids (CsHSO4). Protons (H+) attached to oxyanions are rapidly repositioned (101 Hz ) by rotations of the tetrahedra (1). Approximately once every one hundr ed rotations (109 Hz), the H+ finds itself in an ideal configuration to hop from one nei ghboring tetrahedra to another (2)35.

PAGE 24

14 CsHSO4 exists in 3 phases. Phase III has a hexa gonal array of cesium ions stacked along a bent direction at room temperature. Zigzag chains of hydrogen bonded HSO4 ions are perpendicular to the loose packed Cs+ ion plane and ties the Cs+ ions. The structure is an intermediate between the non-protonic glas rite structure and the ferroelectric CsH2PO4 structure with two different types of hydrogen-bonded chains surrounding the Cs+ cations35,36,37. X-ray diffraction shows the phase III structure, which is confirmed by infrared and Raman spectra, particularly the statistical distributi on of protons between two potential minima of the OH…O hydrogen bonded chains38,39 ,40. There are two first order phase transitions one obser ved at approximately 318K (III II) and the other at 417K (II I). As a result of weakening of the hy drogen bonds and a structural disorder of the anions, the III II transition has been e xplained on the basis of spectral vibration analysis where a convers ion of infinite chains into cy clic dimers occurs. In the low conductivity phases III and II the number of protons per unit cell is equal to the number of positions for them. Hydrogen bonds link SO4 tetrahedra so as to form zigzag chains. Hydrogen atoms are localized on the bonds and their mob ility is low. For the superionic phase I there is a further weakening of the hydrogen bonds and an increasing level of disorder. There is also evidence of rapid re-orientation of HSO4 species and translation disorder of Cs+ ions. The high proton mobility was observed by H and D NMR, as well as quasi-elastic neutron scattering41,42,43. When the temperature is lowere d gradually, metastable phases (I II III) occur, with long annealing below 280K being required to reach the phase III stage. Essentially, the kinetics is largely influenced by the presence of water traces. Samp les prepared by slowly

PAGE 25

15 evaporating solutions of Cs2SO4 and H2SO4 to produce small crystals (0.5 to 1 mm), do not undergo phase transitions when stored in a sealed container. However complex DSC traces are observed for the temperature ra nge 315-380K. According to the study, this is supported by the measured enthalpy values ( H) for the III II transition which varies reversibly between 4.8 and 14.4 kJ/mol being dependant on the concentration of the defect, while the II I (II I) phase transition H value yields approximately 14.4 or 29.2 kJ/mol. The stretching intensity of the SO and S-OH bands is directly affected by the grinding force during the sample prepara tion and is not influenced by the time factor. Weak and strongly ground samples show di fferent spectra. The Raman spectrum shows typical cyclic dimer occurrence as in phases II and III. Thus, CsHSO4 at room temperature can be structurally modified by mechanical treatment which converts the chains of HSO4 ions into cyclic dimers. The applied pressure also referred to as the mechanical treatment, is respons ible for the chain/dimer ratio44,45 ,46,47. CsHSO4 conductivity is influenced by two majo r factors: how many ca rriers are available to transport charge and the mobility of thos e carriers within the material. Therefore, conductivity is determined by carrier concentr ation and carrier mobility. Charge transport in the CsHSO4 crystalline ionic conductor is accomplished by mobile ions (H+) protons which hop from position to position within the lattice. The hopping process only occurs when lattice defects such as vacancies or in terstitials are present. Ion mobility for the material is dependent on the rate at whic h ions can hop from position to position within the lattice. The hopping rate is exponentially activated48,49. However, an activation barrier impedes the motion of the atoms as it hops between positions. We mi ght associate this

PAGE 26

16 energy barrier with the displacements that the atom causes as it squeezes through the crystal lattice between lattice sites50. Thus, carrier concentration in a crystalline electrolyte is controlled by the de nsity of the mobile defect species50. Most crystalline materials conduct via a vacancy mechanism. Th ese vacancies are intentionally introduced into the lattice by doping or by mechanic al pulverization such as ball milling51. Solid electrolytes are preferable to liquid due to a number of reasons. Most importantly, the temperature ranges of operation for such electrolytes are eith er very high (above 600 C) or rather low (below 90 C)52. The high-temperature elect rolytes are not currently practical in mobile applications such as au tomobiles-one of the major potential markets for fuel cells-due to start-up and other limitations53. Low-temperature fuel cells, on the other hand, are not as flexible in the fuel they can use, and much of the energy they release is needed just to maintain opera tion. State of the ar t polymer electrolyte membranes composed of a sulfonate d fluorocarbon, also known as Nafion have high current and power densities, but tend to lose integrity and proton conductivity as a result of dehydration of the membrane at temperatures above 90 oC. Another limitation of Nafion in the application under st udy is that the anode produc t is solid sulfur, which accumulates at the anode surface, and thereby reduces the activity of the system with time on stream. Although these classes of solid electrolytes are known to be soluble in water, they do not pose any problem to our application at 150 oC because steam instead of water exists at this temperature and does not affect the memb rane performance. At the transition phase

PAGE 27

17 temperature of CsHSO4, sulfur is a low viscous liquid and can flow out of the electrolytic cell quickly and easily (Figure 2.3). 120 7 150160140 130 -1 -2 -3 -4 -5 -6 -7 -8 6 12 11 10 9 8Log [ (1/ cm)] of CsHSO4Viscosity of Sulfur (Centipoise)Temperature (Degree Celsius) Viscosity of Sulfur Log of conductivity ( ) of CsHSO4 Figure 2.3 Viscosity of sulfur, conductivity of CsHSO4 vs. Temperature54. Thermomechanical and electrochemical studi es conducted previously show that this material is stable for more than 85h of c ontinuous voltage measurements in an oxygen atmosphere and that the ionic transference num ber is~1 Based on extensive studies of the material, it is clear that th e mobile species are protons53. But there are still some issues to be resolved. A major concern with this SAE is the possibility that performance will be lost gradually through the deple tion of the sulfur in the CsHSO4 electrolyte since the compound can react with molecular hydrogen to form hydrogen sulfide. A thorough investigation of the decomposition and behavi ors at transition temperatures of CsHSO4 could provide valuable insight into the ther mochemical requirements necessary to create a stable solid acid electrolyte for electrolytic cell applications. For complete splitting of H2S, the electrolyte must demons trate high ionic conductivity: th at is, the entirety of the

PAGE 28

18 current should be carried by i ons rather than electronic species and the membrane must be impermeable to H2S. 2.4 Previous Work on H2S Splitting The ability to electrochemically split H2S in a fuel cell was carried out in the late 1980s55. Subsequently, this led to the study of altern ative electrolytes and a node electrocatalysts for H2S splitting in electrolytic cell12. O2conducting yttria-sta bilized zirconia (YSZ) electrolytes were the most comm only used electrolytes in H2S splitting cell. Recently, ceria-based O2conducting electrolytes56,57 and H+ conducting electrolyte s have proven to have some new advantages for H2S electrolysis operated at temperatures below 700 C. Active anode materials (Li2S/CoS1.035 and WS2) were found to enhance H2S electrolytic cell performance tremendously. In order to be a suitable candidate fo r the electrochemical splitting of H2S, an anode material must possess good electrical conductivity and sulfur tolerance at high temperatures, in addition to good catalytic activit y. Pt anode catalysts have good catalytic activity, but degrade over time in H2S streams. Moreover; most metals and metal oxides are severely corroded by H2S at elevated temperatures. 2.5 Gas Permeability High proton migration which subsequen tly leads to high c onductivity of CsHSO4 at elevated temperature is eliminat ed if the cell electrolyte exhibits a ny type of permeability. Permeability measures the ease with which gases move through a material. Even low porous materials can experience permeability if most of their pores are open or connected. In this work it was found, via el ectrochemical impedance spectroscopy that by

PAGE 29

19 reducing the electrolyte membrane thickness to one-fifth improved the cell performance by over 30 %. Gas permeability testing of these ultrathin membranes whose thicknesses range from (0.1 -0.02 cm) is critical to the development of high perf ormance electrolytes, where gas leaks can prove catastrophic Although reducing el ectrolyte thickness improves the cell performance, there are severa l practical issues that limit how thin the electrolyte can be made. As the electrol yte thickness is reduced, the crossover of reactants may increase. This leads to an unde sirable parasitic loss which can eventually become so large that further thickness decreases are counterproductive19. For solid electrolytes, the membrane cannot be made so thin that physical breaking or developing pinholes become an issue. Membrane failure can result in sudden and violent mixing of the fuel and cathodic product (hydrogen). Even mechanically sound, pinhole-free electrolytes may fail if the thickness varies considerably across the cell. Thin electrolyte areas may be subjected to rapid deterioratio n and failure. In addition, extremely thin electrolytes (solid or liquid) risk electrical shorting, espe cially when the electrolyte thickness is on the same order of magnitude as the electrode roughness. Furthermore, part of the electrolyte resistance is associated w ith the interface between the electrolyte and the electrode. This contact resistance is independent of electrolyte thickness. Also, the ultimate physical limit to solid-electrolyte thickness is determined by the electrolyte breakdown properties. This limit is reached when the electrolyte is made so thin that the electric field across the membrane exceed s the dielectric breakdown field for the material. Our system can be described by E qn. 2.2. Permeability (K) is determined by measuring the volume of gas ( V) that passes through a sample in a given period of time ( t) when driven by a given pressure drop ( p = p1-p2):

PAGE 30

20 p p p p t V p I K 2 1 22 (2.2) Where I = D l is a constant with a diffusion coefficient, D, l is the thickness per area of the membrane, and p1 and p2 are the pressures in the delivery and receiving compartments of permeation cell respectively51. It was found, via electrochemical impeda nce spectroscopy, that by reducing the electrolyte membrane thickne ss by one-fifth improved the cell performance by over 30 %. In this investigation, possible H2S gas permeation through various thicknesses of CsHSO4 ulta-thin membranes at 150 oC is reported for the first time. Herein, we will show that CsHSO4 membrane is impermeable to H2S gas, with innocuous effect, over prolonged periods of operation. Previous study has shown th at at the superprotonic transition phase of CsHSO4 (140-160 oC) permeance to hydrogen gas decreased58. However, analyses conducted by a material science group at Cali fornia Institute of Technology report that the low open circ uit voltage (OCV) observed using CsHSO4 electrolytes could be attr ibuted to hydrogen permeability53. Herein we will show that CsHSO4 membrane is impermeable to H2S gas, with innocuous effect, over prolonged periods of operation.

PAGE 31

21 3 THEORY 3.1 Physical Properties and Sources of H2S H2S information was obtained from Specialty Gases of America, purity = 99.5 %, Mwt: 34.08, specific volume: 1.23 ft3/lb, labels: *P, *F: *P = poison-inhalation hazard, *F = flammable gas, fH0gas, -20.5 kJ/mol (exothermic), autoignition temperature, 500 oF or 260 oC, density, 1.363 g/L. Hydrogen sulfide, H2S, at room temperature is a colorless, extremely poisonous vapor (MSDS data: IHL-HMN LC50 800 pp m 5 min). Highly flammable under normal conditi ons hydrogen sulfide readil y forms explosive mixtures with air (4 % lower explosion limit, LEL, H2S). It occurs naturally in crude petroleum, natural gas, volcanic gases and hot springs at concentrations ranging from a few ppm to 50 % or higher. It also can result from b acterial breakdown of organic matter including human and animal wastes. Other source incl udes, industrial activ ities, such as food processing, coal processing in the integrat ed gasification combined cycle (IGCC) power plants, desulfurization of hydrocarbon resour ces, coke ovens, Kraft paper mills, tanneries, petroleum refineries and ma ny other such processes. 3.2 H2S Electrochemical Splitting The fundamental principle underlying the splitting of H2S electrochemically consist of having an anode and cathode compartments in which a solid proton conducting membrane separates an anode chamber from a cathode chamber (Figure 3.1). The process

PAGE 32

22 consists of passing H2S containing gas through the anode chamber to contact a catalytic anode, where reaction Eqn. 3.1 takes pla ce to produce elemental sulfur, protons and electrons. H2S S2 + 2H+ + 2eEo (423 K) = 0.19 (3.1) The protons pass through the membrane from the anode chamber to the cathode chamber, where reaction (Eqn. 3.2) occurs with elec trons from the catalytic cathode to produce hydrogen gas. 2H+ + 2eH2 (3.2) Electrolyte H+Electrolyte H2 DCe Cathode Anode H2S S H+H+H+H+O O O O Electrolyte H+Electrochemical Oxidation H2S S2+ 2H+ + 2e-Electrochemical Reduction 2H+ + 2eH2Chemical Reaction H2S S2+ H2 Figure 3.1 Basic opera ting principle of H2S electrolysis The chemical reaction taken place is given by reaction (Eqn. 3.3) H2S S2 + H2 (3.3)

PAGE 33

23 3.3 Electrochemical Thermodynamics Any electrochemical reaction involves the tr ansfer of electrons between an electrode surface and a chemical species adjacent to the electrode surface. Thermodynamically favorable electron transfer processes is harnes sed in order to extract the electrical energy requirement for the oxidation process. To carry out a thermodynamic analysis, it is pertinent to relate thermodynamic (reversib le) potential to state variables. For a reasonable analysis to be performed, it is common to assume electrochemical operation occurs at constant temperature a nd pressure. The choice of fuel (H2S) makes it easy to define the electrochemical processes occurring w ithin the cell as well as the reaction free energies associated with the components. Furt hermore, if a closed system operating under these conditions is considered, then a th ermodynamic analysis can be carried. This implies that the electrical work output by a syst em can be related to its total change in Gibbs free energy. G = electrical (3.4) A well known expression for maximum amount of electrical energy which can be deliverd by a system is (Eqn. 3.5)48: electrical = nFE (3.5) Where electrical is the maximum electrical work, n is the number of electrons (per ionic species) taking part in the reaction, F is Faraday’s consta nt (96,487 A sec/equiv), and E is the reversible cell potential. The latter is the necessary driving force for any electrochemical process to occur. Combini ng these two expressions (Eqns. 3.4, 3.5),

PAGE 34

24 yields the key equation relating electrical work to traditional thermodynamics (Eqn. 3.6). Eqn. 3.6 statement represents the possible maximum predicted elec trical work output. G = nFE (3.6) However, since this analysis is for a reversible (ideal) system, the elec trical work in a real electrochemical cell device will be smaller. Eqn. 3.6 is the key e quation relating electrical potential to the traditional thermodynamic fr amework. The free energy of reaction can be calculated from free energy of formation f iG, data by i f iG si G, (3.7) where is is the stoichiometric coefficient (positive for products and negative for reactants48). 3.4 Nernst Equation The ideal standard potential for an electroche mical reaction may be obtained either from tabulated data or by using standard Gibbs-f ree energy data in Eqn. 3.6. The Nernst equation is given by Eqn. 3.8 which relate s concentration changes to changes in reversible potential59. vi i Oa n nF RT E E 1 (3.8) Here R is the gas constant (8.314 J/mol K), T is the absolute temperature in K, ia is the activity and vi is the stoichiometric coeffi cient of electro -active speciesi. Ideal gas behavior is assumed, where the activity of gaseous species equa ls their partial pressure as in (Eqn. 3.9):

PAGE 35

25 vi i Op n nF RT E E 1 (3.9) where Pi is the partial pressure of species iin the gas phase. Since sulfur is the product in the electrochemical splitting of H2S by (Eqn. 3.3) the Nernst equation takes the form: S H H S Op p p n F RT E E22 2 / 11 2 O E (423 K) = 0.19 (3.10) It is clear from the Nernst equation that a set of concentrations exists that will result in a reversible potential of zero. This state corresponds to the thermodynamic equilibrium and the equilibrium constant K can be obtained from (Eqn. 3.11). nK nF RT E EO1 = 0 Therefore, nK nF RT EO1 (3.11) 3.5 Cell Performance There several factors that contribute to irreve rsible losses associated with electrochemical performance commonly referred to as polariza tion losses. Such factors include: activation losses ( act) due to kinetics, ohmic losses ( ohmic) from ionic and electronic resistance, and concentration losses ( conc) due to mass transport60.

PAGE 36

26 3.5.1 Activation Polarization ( act) Activation polarization represents the amount of voltage sacrif iced (lost) to overcome the activation barrier associated with the electroch emical reaction. This usually occurs as a result of slow kinetics which impedes the rate of electrochemical reactions at the electrodes. Consequently, activ ation polarization is a strong function of reaction rates, i.e., oxidations or reductions at e ither the cell anode or cathode. 3.5.2 Potential and Rate: Butler-Volmer Equation The Butler-Volmer equation basically states that the current produced by an electrochemical reaction increa ses exponentially with activati on overvoltage (Eqn. 3.12) RT nF RT nFe e j j/ 1 / 0 (3.12) where j is the current density, 0j is the exchange current density for the reaction and represents the state of dynamic equilibrium at which the net reaction rate is zero for both the forward and reverse reactions, is the transfer coefficient expresses how the change in the electrical potential ac ross the reaction interface changes the size of the forward versus reverse activat ion barrier (values lie s between 0 and 1), is the activation polarization, and n is the number of electrons transferred per reaction. Kinetics can be simplified with two useful approximations to the Butler-Volmer equation when the activation overvoltage ( act) is either very small or very large: a. Whenact is very small For small act (less than about 15 mV at room temperature), a taylor series expansion of the exponential te rms gives a linear relationship of current and overvoltage (Eqn. 3.13).

PAGE 37

27 RT nF j jact 0(3.13)bWhen act is large (greater than 50-100 mV at room temperature), the second exponential in the Butler-Volmer equation be comes negligible. In other words, the forward–reaction direction dominates, correspon ding to a completely irreversible reaction process. Simplifying and solving the equation for act, yields the well known Tafel equation59 (Eqn. 3.14). nj nF RT nj nF RTact1 10 (3.14)In generalized form, this equation becomes j b aactlog (3.15) where b is the Tafel slope. A plot of act versus 1n j is a straight line fr om which the values of 0j and are obtained. This proves more us eful in most discussions. 3.5.3 Concentration Losses ( conc) Overpotential associated with mass transport is called concentration overpotential. This form of overpotential ( conc) is caused by mass transport lim itations occurring as a result of reactant depletion (or product accumulation) within the catalyst layer which adversely affect performance. It is also caused by inad equate flow of reactants to or removal of products from the cell electrode. Gas transpor t within the electrodes is dominated by diffusion because of the tortuous, sheltering ge ometry of the electr odes which insulates

PAGE 38

28 gas molecules from the convective forces present in the flow channels50. The 3 major transport processes are: a. diffusion of fuel and product flows in the bulk, b. diffusion of fuel and pr oducts in the electrode, and c. proton diffusion in the electrolyte. The first two types of diffusion, which are not covered by this study, can be modeled by Fick’s law of diffusion; listed below are the tw o effects these transport processes have on the performance of the system: a. Nernstian losses: The reversible cell volt age will increase as predicted by the Nernst equation since the reactant concentration at the catalyst layer is decreas ed relative to the bulk concentration and the produc t concentration at the catalys t layer is increased relative to the bulk concentration. b. Reaction losses: The reacti on rate (activation) losses wi ll be increased because the reactant concentration at the catalyst layer is decreased relative to the bulk concentration and the product concentration at the catalys t layer is increased relative to the bulk concentration. 3.5.4 Ohmic Losses ( Ohmic) These are caused by resistance to flow of ch arged species (electr ons and ions) at cell electrodes and electrolytes respectively. It re sults primarily from the concentration cell that is established between the bulk electr olyte and the electrode surface. At the cell anode, sulfur poisoning tends to increase th e polarization resistan ce which physically affects the anode structure bri nging about de-lamination of th e cell electrode. As a result,

PAGE 39

29 an insulation layer is formed which inhibits the transport of ions and electrons through the electrolyte and the electrode respectiv ely causing the cell pe rformance to drop. Ohmic losses obey Ohms law: Total OhmicIR (3.16) I is the current drawn from the cell and RTotal is the total cell resistance; this includes contributions from the electr olyte and the whole of the electrochemical system. The intrinsic resistance from the electrolyte is given by Eqn. 3.17. A x Re electrolyt e electrolyt (3.17) where e electrolyt is the ionic resistance of the electrolyte, x is the electrolyte thickness, and A is the electrolyte area. Impe dance spectroscopy or other av ailable methods are usually used to determine cell resistance due to interface contacts and lead wires. 3.6 Physical Meaning of Conductivity Conductivity quantifies the ability of a material to permit the flow of charge when driven by an electric field. In other words, conductivity ( ) is a measure of how well a material accommodates charge transport. A material conductivity is influenced by two major factors: how many carriers are available to transport charge and the mobility of those carriers within the material. The following equation defines in those terms: i i i iu c F z | | (3.18) where ic represents the molar concentration of charge carriers (how many moles of carrier is available per unit volume) andiu is the mobility of charge carriers within the material. The quantity F zi| |is necessary to convert charge carrier concentration from

PAGE 40

30 units of moles to units of coulombs. Here, iz is the charge number for the carrier, the absolute-value function ensures that conductiv ity is always a posit ive number, and F is Faraday’s constant. A material’s conductivity is therefore determined by carrier concentration ci and carrier mobility ui. These properties are in turn set by the structure and conduction mechanisms within the material. Up to this point, the charge transport equations apply equally well to both electronic and ionic conduction. However, their path will dive rge because electronic and ionic conduction mechanisms are vastly di fferent hence; the el ectronic and ionic conductivities are also quite different. 3.7 Ionic Conductivity in a Crystalline Solid Electrolyte In contrast to electron transport in metals where valence electrons detach from immobile metal atom cores and move freel y in response to an applied field, charge transport in crystalline ionic conductors is accomplished by mobile anions which hop from position to position within the lattice. The hopping process only occurs when lattice defects such as vacancies or interstitia ls are present. As a result, conduction hoping process leads to a very different expression for mobility as co mpared to a metallic electron conductor. Ion mobility for the material is dependent on the rate at which ions can hop from position to position within the lattice. The hopping rate is exponentia lly activated. A calculation of the net flux (net movement) of atoms in an imaginary plane A which lies between two real atomic planes in the materi al is shown below. The assumption made is that the flux of grey atoms hopping in the forward direction (and therefore through plane

PAGE 41

31 A) is simply determined by the number (con centration) of grey at oms available to hop times the hopping rate49,59: x vc JA 12 1 (3.19) where AJis the forward flux through plane A,v is the hopping rate, 1cis the volume concentration (mol/cm3) of grey atoms in plane 1, x is the atomic spacing required to convert volume concentration to planar concentration (mol/cm2), and the 2 1 accounts for the fact that on average only half of the jump s will be forward jumps. Similarly, the flux of grey atoms hopping from plane 2 back ward through plane A will be given by x vc JA 22 1 (3.20) where JA is the backward fl ux through plane A and 2c is the volume concentration (mol/cm3) of grey atoms in plane 2. The flux of grey atoms across plane A is therefore given by the difference between the forwar d and backward fluxes through plane A: 1 22 1c c x v Jnet (3.21) A familiar expression for diffusion is given by Fick’s first law, dx dc D J / Eqn. 3.21 can be expressed in terms of a concentration gradient as x c c x Jnet 1 222 1 (3.22) x c x v 22 1 dx dc x v22 1 (for small x ) (3.23)

PAGE 42

32 Comparison with the Fick’s fi rst law of diffusion equation dx dc D J/ allows us to identify what we call the diffusivity as 22 1x v D (3.24) The above expression illustrate s that diffusivity is a functi on of the intrin sic hopping rate for atoms in the material and the atomic length scale associated with the material. However, an activation barrier impedes th e motion of the atom s as it hops between positions. We might associate this energy barri er with the displacem ents that the atom causes as it squeezes through the crystal lattice between latti ce sites. We can write the hopping rate as RT Gacte v v/ 0 (3.25) Where actG the activation barrier for the hopping process and 0v is the jump attempt frequency. Based on this activation model fo r diffusion, we can then write a complex expression for the diffusivity as RT Gacte v x D/ 0 22 1 (3.26) Or, lumping all the preexpone ntial constants into a 0D term, RT Gacte D D/ 0 (3.27) Where 0Dis a constant reflecting the attempt fr equency of the hopping process, R is the gas constant, and T is the temp erature (K). The overall mob ility of ions in the solid electrolyte is then given by RT FD zi u| | (3.28)

PAGE 43

33 Where | | zi is the charge number on the ion, F is the Faraday’s constant, R is the gas constant, and T is the temperature (K). Inserting the expression for ion mobility given by (Eqn. 3.28) into equation for conductivity (Eqn. 3.18) yields RT D ziF c2) ( (3.29) Combining Eqns. 3.27, 3.28 and 3.29 in what is known as the Einstein-Smoluchowski equation we obtain an expres sion for ionic conductivity of SAE in the form of an Arrhenius relationship (Eqn. 3.30) KpT Ha Oe A u c e T/. ) (3.30) where RT D ziF c AO O2) ( is the preexponential or the frequency factor, K is the Boltzmann constant, Ha is the migration enthalpy. Thus, carrier concentration in a crystalline el ectrolyte is controlled by the density of the mobile defect species. Most crystalline electrolytes c onduct via a vacancy mechanism. These vacancies are intentionally introduced into the lattice by dopping or by mechanical pulverization. 3.8 Electrochemical Impedance Spectroscopy (EIS) Characterization Impedance measurements are usually made by applying a small sinusoidal voltage perturbation,wt V t Vcos0, and monitoring the system’s resultant current response,

PAGE 44

34 wt i t icos0. In these expressions, t V and t i are the potential and current at timet, 0V and 0i are the amplitudes of the voltage and current signals, andwis the radial frequency. The relationship between radial frequency w(expressed in radians per second) and frequency f(expressed in hertz) is61 f w 2 (3.31) In general, the current response of a system may be shifted in phase compared to the voltage perturbation. This phase shift effect is described by A graphical representation of the relationship between a sinusoidal voltage perturbati on and phase-shifted current response is shown in Figure 3.2 (for a linear system). Impedance is given by the ratio between a time-dependent voltage and a time-dependent current: t i t V Z (3.32) Following Eqn. 3.32 we can write the sinusoidal impedance response of a system as wt wt Z wt i wt V Z cos cos cos cos0 0 0 (3.33) Alternatively, complex notation can be used to write the im pedance response of a system in terms of a real a nd imaginary component: sin cos cos0 0 0 0j Z e Z e i wt V Zj j jwt (3.34) The impedance of a system can therefore be expressed in terms of an impedance magnitude 0Z and a phase shift or in terms of a real component ( cos0Z Zreal) and

PAGE 45

35 an imaginary number (j Z Zimag sin0 ). The expression j represents the imaginary number ( 1 j), not the current density used later in the text. Figure 3.2 A sinusoidal voltage perturbation/ resulting sinus oidal current response. The current response will possess the same period (frequency) as the voltage perturbation but will generally be phase shifted by an amount Normally, impedance data are plotted in term s of the real and imaginary components of impedance (realZ on the x axis and imagZ on the y axis). Such graphical representations of impedance data are known as Nyquist plots. Because impedance measurements are made at dozens or even hundreds of differe nt frequencies, Nyqui st plots generally summarize the impedance behavior of a sy stem over many orders of magnitude in frequency.

PAGE 46

36 4 EXPERIMENTAL EQUIPM ENT AND TECHNIQUES 4.1 Introduction Selecting the most efficient approach from various different routes to make the same product is not always simple. Inthis study, our most successful anode material, pdichlorobenzene/ CsHSO4/ ruthenium (IV) oxide/ cobalt sulfide and 200 m modified electrolyte disc is used as an illustration. Our approach is to develop a stable solid electrolyte first, and then identify which anode materials improve ionic and electronic properties. 4.2 Experimental Apparatus The experimental apparatus us ed in splitting (100 %) H2S content gas is illustrated in Figure 4.1. The electrol ytic splitting of H2S in a solid acid electrochemical cell was performed in laboratory constr ucted equipment consisting of : (a) a gas handling system; (b) an electrolysis cell; and (c) an exit gas control and mo nitoring system, as described here. Electrolytic cell, with a 1. 25 cm electrolyte disc is opera tional. It operates at 150 oC with gaseous H2S and liquid sulfur in the anode co mpartment and hydrogen gas in the cathode compartment. The cell has thick walls to accommodate pipe fittings. The housing is chrome plated (316) stainless steel. Aflas o-rings and gaskets are employed to resist the corrosive anode contents. The total exposed surface on each side of the electrolyte pellet is 0.4 cm2. The active area is less because of electrode coverage. The

PAGE 47

37 pellet thickness is 200 m. The high area to thickness ratio implies a small ohmic resistance. The sulfur compartments are modeled after the sulfur compartments in sodium sulfur batteries which are filled with carbon felt. This carbon felt, which is in direct contact with the electr olyte, conducts electrons from the electrolyte in the anode compartment. The same material is being used in the cathode compartment. Apparatus for monitoring and control of the hydrogen su lfide sulfur and hydroge n flows have been designed and tested. A liquid sulfur collec tion system was fabricated and a water displacement apparatus for th e hydrogen produced was utili zed. Hydrogen sulfide is delivered to the anode compartm ent of the electrolytic cell vi a (316) stainless steel tubing regulated by a pressure regul ator on the tank, monitored by a flow meter and controlled by a micrometering valve in the line. The pressure is continuously monitored.

PAGE 48

38 Figure 4.1 Schematic of electrolytic split ting cell. (a) electrol yzer; (b) cathode purge line; (c) H2S line; (d) flowmeter 1; (e) flowmeter 2; (f) sulfur line; (g) sulfur collection apparatus; (h) NaOH solution; (i) nitr ogen tank; (j) anode purge line; (k) H2S tank; (l) to vent; (m) infrared cell; (n) pressure transducer ; (o) ejector for GC analysis; (p) pressure differential; (q) cathode produc t outlet; (r) water displaceme nt apparatus; (s) hydrogen gas; (t) vacuum pump; (u) to vent; (v) H2S inlet; (w) cathode product; (y) H2S line; (z) computer interfaced LabView. The differential pressure of the hydrogen exiting the ce ll and the hydrogen sulfide entering is measured directly for greatest sens itivity. Our intentions are to make the electrolyte disc as thin as possible to maximize conductance. This requires minimizing the pressure difference they must withstand. Hence this difference is carefully measured and kept small. There are nitrogen purge lin es for purging both compartments before and after experiments. The cell was sealed using hermetic sealing at all joints and insulation of electrical components was ensured by using an aflas gasket. The velocity of the H2S

PAGE 49

39 gas stream was controlled using a microm etering valve (Upchurch). Two pressure transducers are used to mon itor the pressures in the anode and cathode compartments for greatest sensitivity and the pressures in each half cell were maintained using two back pressure micrometering valves (Upchurch). The pressure was kept constant throughout the duration of the experiment. Before each ex periment the hermetic sealing of the cell was confirmed at room temperature and then at the operating temperature, using nitrogen as the test gas. 4.3 Preparation of Anode Electrocatalyst A mixture of CsHSO4, p-dichlorobenzene, CoS2, and RuO2 was used to prepare the anode electrocatalyst. Throughout the experiment, Pt black served as the cathode electrocatalyst. The ratio of the resp ective components was 1:0.5:3:3 by mass. Compressed carbon felt was used for both anode and cathode electrod e with a specific mass of 0.00136 g/m2 ; this also served as current collectors for both electrodes and a source of mechanical strength for the thin electrolyte disc. The anode catalyst materials, with the ratios specified, are mixed for several hours in a high energy planetary ball mill. To get a un iform mixture, ethanol is added to the combination to make a suspension that is mi xed thoroughly for 3 h. Then ethanol will be allowed to evaporate slowly to leave the mixture as powders. The powdery mixture are heat-treated in N2 atmosphere at 150 oC for 2 h and allowed to cool under N2 to room temperature in an autosorb apparatus in the outgassing station. The resulting powders were then used to prepare the anode catalys ts. The specific surface area determinations of

PAGE 50

40 the catalysts were carried out with Auto sorb-1 software using a BET Multipoint approach. 4.4 Cell Fabrications CsHSO4-based MEAs were fabricated by pre ssing in an evacuated die set. An electrocatalyst layer, an electrolyte laye r, and a second gas diffusion layer with electrocatalyst layer were se quentially added to a carbon felt gas diffusion electrode. After addition of each layer, the structure was uniaxially pressed to promote adhesion between layers, forming the final layer of th e sandwiched structure, shown in Figure 4.2. The electrolyte thickness was calcu lated from the weight of CsHSO4 incorporated into the electrolyte. Prior to the electrolysis run, the p-dichlo robenzene was removed under gentle heating (12 h) at 70 oC in inert atmosphere, followed by a 2 oC/min ramp to 174 oC, the boiling point of p-dichlorobenzen e, leaving behind open porosity in the electrocatalyst layers. Porous carbon felt, with average pore size and porosity of 35 m and 38-40 %, respectively, served as a curren t collector/gas-diffusi on electrode and also served to provide mechanical s upport for the thin membrane disc.

PAGE 51

41 0.5 mm Figure 4.2 Schematic diagram of a fabricated cell (CsHSO4-based MEA). 4.5 Permeability of H2S on CsHSO4 Pellet A CsHSO4 membrane was fabricated by pressing in an evacuated die set at (55 MPa). The membrane thicknesses whose permeations were examined are 0.1 cm, 0.05 cm and 0.02 cm. The test bed consists of a 1.25 cm diameter permeation cell made of chrome plated stainless steel with both the delivery and receiving compartments separated by the membrane under study. The membrane as used he re consists solely of the disc material investigated. The permeation cel l was placed in a gravity convection electrical oven (Yamato DX 300) with programmable te mperature controlled within 1 oC. H2S is delivered to the delivery compartment of the permeation cell to contact the solid membrane via 316 stainless stee l tubing regulated by a pressure regulator on the tank,

PAGE 52

42 monitored by a flow meter and controlled by an in-line micrometering valve and system pressure is continuously monitored. Before each experiment the hermetic sealing of the cell was confirmed at room temperature a nd then at the operating temperature using nitrogen as the test gas. N itrogen continued to flow into the two compartments at a regulated flowrate to maintain a zero pressu re differential in the system. The cell is allowed to attain a steady-state temperature of 150 oC before replacing nitrogen with H2S at a flow rate of 0.25 cm3 min-1. Again, the flows in the tw o compartments are regulated to maintain a zero pressure differential after which, the flow of nitrogen is cut off from the receiving compartment and the system is allowed to attain steady-state at which the differential pressure remained constant with time. Once steady-state was achieved, the pressures in the two compartments and the differential pressure between the two compartments were noted. If the membrane material is permeable, H2S will pass through the membrane to the receiving compartment driven by a pressure drop in the delivery compartment. This pressure drop would lead to a decrease on th e existing pressure differential as more H2S permeates through the membrane. Gas samples of 22 nmol were taken from the receiving compartment every 10 h for GC-MS analysis using a gas-tight syringe for an extended period of 96 h. Sin ce the volume of the delivery compartment is 127 L, 0.39 % cell volume is utilized for each sample analysis. The permeability measurements were carried out separately fo r each membrane thickne ss (0.1 cm, 0.05 cm and 0.02 cm) investigated.

PAGE 53

43 4.6 Experimental Procedure In this work, experiments were conducted us ing the constructed cel l and fabricated cell (CsHSO4-based MEA) prepared earlier to electrolyze H2S gas at different voltage applications in order to study the e ffect of variation of voltage on H2S splitting. Subsequently, this was followed by a series of experiments to elucidate the performance of the MEA materials by applyi ng different anode electrocat alysts configurations as shown in Table 4.1 at constant applied voltage 900 mV to split H2S. Although this voltage is high compared to the therm odynamic decomposition potential, commercial application requires keeping en ergy consumption as low as po ssible. This entails working at high current and low voltage. Because of poisoning of the anode catalyst by sulfur in the presence of Pt, RuO2 was used which acts as the host matrix while CoS2 is used as the modifier as outlined in the factorial experi ment. The electrochemical cell was placed in a gravity convection electrical oven (Yamat o DX 300) with programmable temperature controlled within 1 oC. H2S introduced into the anode co mpartment of an electrolytic cell contacts the solid electrolyte and catalytic anode, and splits electrolytically to form liquid sulfur at the anode and gaseous hydrogen at the cathode at a temperature at which liquid sulfur has a low viscos ity and can flow out of the electrolytic cell quickly and easily. The solid curve in Figure 2.3 gives the viscosity of liquid sulfur and shows that it is minimized near 150 oC. The cell operating pressure is 138 kPa. The proof of this principle is now accomplished.

PAGE 54

44 Table 4.1 Various catalysts conf igurations of anode catalysts. No Anode catalyst Cathode catalyst S1 RuO2/p-dichlorobenzene/CsHSO4/Pt black Pt black S2 RuO2/p-dichlorobenzene/CsHSO4 Pt black S3 RuO2/CsHSO4 Pt black S4 CoS2/p-dichlorobenzene/CsHSO4 Pt black S5 CoS2/RuO2/p-dichlorobenzene/CsHSO4 Pt black Electrons are removed in the anode compartmen t to yield hydrogen ions and liquid sulfur. The liquid sulfur pools at the bottom of the compartment and runs out of a drain tube there for collection. The liquid pool seal s the drain tube agai nst outflow of hydrogen sulfide from the compartment. Hydrogen ions pass through the solid electrolyte to the cathode compartment where they gain electron s to form hydrogen gas that flows out at the top. The hydrogen produced flows into a ma nifold from which samples are taken for analyses. The gas is collected for analysis by displacement of water in a graduated container that immediately shows its volume. For safety, the water includes dissolved alkali to absorb any acid gas contaminants The hydrogen so collected is analyzed by using an Agilent 6890N GC gas chromatogra phy. The current ge nerated and the cell resistances were measured using a computer interfaced LabView program.

PAGE 55

45 4.7 Synthesis and Ionic Conductivity of CsHSO4 CsHSO4, used both in the electrocatalyst and electrolyte layers, was synthesized by chilled acetone-induced precipitation from aqueous solution s of stoichiometric quantities of CsSO4 and H2SO4 (see Eqn. (4.1)) followed by vacuum filtration. The resulting CsHSO4 powder was vacuum dried in a Schlenk f iltration manifold for a period of 24 h at 80 oC under nitrogen atmosphere to remove any acid gas and other contaminants. Cs2SO4 + H2SO4 2CsHSO4 (4.1) The mechano-chemical process employing hi gh energy planetary milling was initiated by loading pure CsHSO4 powders into a ball mill referred to as Fritsch pulversette planetary mono mill, P6 in an inert atmo sphere. 10 grams of powder and 15 stainless steel balls (1.5 cm in diameter, 14 g in mass) were sealed under inert, into the stainless steel vials (80 cm3 volume). The rotation speed of the milling vials ( anticlockwise rotation) which are fixed onto a rotating disk and the rotation speed of disk ( clockwise rotation) can be set independently. Milling condition was associated by 3 essential parameters ( (rpm)/ (rpm)/ t (h)) where t was the duration of the milling process. The milling parameters such as ball to powder mass ratio and milling speed were optimized to 20/1 and 300 rpm, respectively. The milling duration was 3 h after which the fine crystal powder was stored in a dessi cator prior to characterization. The CsHSO4 that did not pass through the process of ba ll milling was also stored in a desicator. Proton conductivity measurements of unmodified and modified CsHSO4 were carried out by electrochemical impedance spectroscopy (EIS) in the temperature range of 60–280 oC.

PAGE 56

46 The unmodified and modified, as used in this context, represent CsHSO4 that did not pass through the mechanical ball milling and that which was subjected to mechanical ball milling process respectively. The ionic conductivity of CsHSO4 disc was measured by two-probe AC measurements techniques. In the 2-probe setup only two leads wires are located at the endpoints of th e copper strips as shown in Figure 4.3. A disc of CsHSO4 was made by pressing a measured amount of modified CsHSO4 in a vacuum die set. This disc was then sandwiched between two copper gol d plated strips serving as the electrode material and placed in a programmabl e furnace with a heating rate of 3 oC min-1. The impedance measurements was carried out with an Agilent HP 4284A LCR impedance analyzer interfaced with a computer in th e frequency range of 20 Hz–1 MHz with an applied voltage of 1.0 V. The whole proc ess starting from disc preparation to conductivity measurements were rep eated for the unmodified CsHSO4. Figure 4.3 Illustration of the two-probe AC conductivity measurement setup. Square Jig Teflon material Gold plated Copper strip CsHSO4 membrane

PAGE 57

47 4.8 Design of Experiments A full factorial experiment was utilized to i nvestigate different combinations of process variables factors at different treatment levels in the electrolytic system. This technique was employed to identify po ssible interactions between va riables and also directs the modeling and characteriza tion of the system. Table 4.2 shows the tested variable factors, the levels of testing (h igh and low), and the measured effects for metal sulfides based a nodes. The number of experimental runs is limited by using a 2k-1 design. Fractional designs are expressed as lk 1, where l is the number of levels of each factor investigated, k is the number of factors investigated. Table 4.3 gives the experimental combinations of chosen variables. Other factors are kept constant except for the chosen variables in order to maintain consistency and minimize error in the experiment. These other factors includes H2S flow rate (0.25 cm3/min), current collector wires, cathodes, H2S delivery tubes, etc. in order to reduce any outside interference to the measurements. Table 4.2 Factors, treatment leve ls, and effects for metal sulfides. Variable Factor Variable Name Treatment Level Measured Effects High Low Electrolyte thickness a 0.2 mm 1 mm lA Rcell Time on stream b 12 hr 6 hr lA, Rcell Electrolyte age c 1 day 7 days lA, Rcell

PAGE 58

48 Table 4.3 Two-level 3-factor full-fact orial experiment design pattern. + and –s refer to the treatment levels (high-low) of each factor. Run Order Comb. Factors a b c 1 (1) 2 a + 3 b + 4 ab + + 5 c + 6 ac + + 7 bc + + 8 = 23 abc + + +

PAGE 59

49 5 RESULTS AND DISCUSSION 5.1 XRD Structural Characterization of Modified and Unmodified CsHSO4 The powder X-ray diffraction (XRD ) of the synthesized CsHSO4 was carried out by a Philips X’pert diffractometer with Cu K radiation of known wavelength, = 1.54060 . The incident and the diffraction slit wi dth employed for the measurements are 1o and 2 o respectively and mask size is 0.10 cm. Prior to this measurement the diffraction from the paraffilm tape was calibrated and the angle of occurrence 13 o was obtained. Phase identification and particle size calculations have been carried out by PANanalytical X’pert Highscore so ftware version 1.0f62. XRD characterization was employed for struct ural characterization of solid acid for purpose of phase identification, structure determination, a nd crystal orientation. Figure 5.1 (a) and (b) represent the XRD diffractogr ams of both the unmodified and modified CsHSO4 respectively. In Figure 5.1 (b), the spec trum obtained has a reduction in size at maximum intensity as a result of decrease in particle si ze resulting from the mechanochemical process of the 3 h ball mill under a N2 ambient environment.

PAGE 60

50 Figure 5.1 X-ray diffractograms of (a) unmodi fied and (b), modified CsHSO4. Reflections at high intensities in all and the miller indices assigned to the reflection peaks correspond to CsHSO4 (phase III). The XRD pattern of powders pr oduced under these conditions, ( (rpm)/ (rpm)/ t (h)), 300/50/3, of power values of shocks was se lected to test the performance of our conductivity model. It is expected that ball milling process will improve the concentration of defects resulting from disorder of this material and facilitate the rate of proton diffusion and hence increas e its ionic conductivity as a result of increase in the number of vacant sites for proton migration51,51 It has been known from previous studies that the behavior of this material is pr eceded by order-disorder phase transition. Also from Scherrer’s calculation shown in Table 5.1, it was observed that the particle size decreased with mechanical pulverization of the ball milling process as compared to a non milling condition.

PAGE 61

51 Table 5.1 Comparison of crystal paramete rs for unmodified and modified CsHSO4. Preparation method Unmodified Modified Crystal size calculation from Scherrer’s equation 33.4 nm – 75 nm 18.6 nm – 28.2 nm FWHM 0.0984 0.1574 The full width at half maximum (FWHM) intensity of the modified CsHSO4 increased (at least two times) with ball milling as depicted in Table 5.1 reflecting a reduction in intensity of the reflected spectra in Figure 5.1a. 5.2 Simultaneous Gravimetric and Calorimetric Analysis of CsHSO4 Analysis using Simultaneous differential s canning calorimetry and thermo gravimetric analysis (SDT) for the weight loss and the he at flow for the reaction enthalpy during the superprotonic phase transition and the ther mal decomposition temperatures of CsHSO4 were estimated by the TA SDT-Q600 analytic al tool. The calibrati on of SDT was carried out in four steps with an empty pan and st andard sapphire disc. The four calibration subroutines such as thermo gravimetric analysis (TGA) weight, differential thermo analysis (DTA) baseline, and temperature and differential scanni ng calorimetry (DSC) heat flow were performed before sample measurements were conducted. A pre-weighed sample was loaded into the ceramic pan and c overed with the ceramic lid to prevent air or moisture from getting into the sample. The ramp rate for the SDT measurements was 2 oC/min. TA’s universal analysis 2000 softwa re program was used to analyze the SDT profiles.

PAGE 62

52 A typical thermal analysis result for a CsHSO4 sample at fixed heating rate (3 oC min-1) is presented in Figure 5.2. The TGA and DSC results are presented. The TGA analysis shows the temperature at which CsHSO4 is decomposed and that is in consonance with the literature reported value for this material which place the decomposition temperature of this material between 200 oC to 220 oC20. For the decomposition of CsHSO4 by the following reaction: 2CsHSO4 Cs2S2O7 + H2O (5.1) a weight loss of 0.8 % was observed at 197.91 oC while an expected weight loss at 220 oC was ~1 %. Figure 5.2 Simultaneous differential scanning calorim etry and thermogravimetric analysis (SDT) measurement. (a) TGA spect ra showing weight loss, CsHSO4 decomposes into: 2CsHSO4 Cs2S2O7 + H2O; (b) DSC spectra showing the phase transition and decomposition temperatures of CsHSO4 52.

PAGE 63

53 It is evident from these thermal measuremen ts that the decomposition of this material progressed with increasing temp erature and rapidly degrades at higher temperatures. Quite significantly, regardless of sample surface area, in all cases (modified or unmodified) a phase transition, independent of decomposition, is clearly evident at ~ 142 oC. Although the impact of this transition on conductiv ity cannot be assessed from thermal analysis alone, we assign this transformation as the superprotonic transition described earlier32,63. 5.3 Ionic Conductivity Modeling of CsHSO4 Protonic conductivity was the prin cipal property of interest in the material examined here. For this, extensive use of alternating current (AC) im pedance spectroscopy has been made to characterize the protonic conductiv ity of solid acid compound (modified and unmodified) as a function of temperature at ambient pressure. This is necessary since no information on the ionic c onductivity of mechano-chem ically processed CsHSO4 exist so, this will provide a datum for comparison w ith known data from previous works on unmodified method. As men tioned earlier, superprotonic behavior of solid acids compound is (to date) always associated with an order-disorder phase transition and they conduct via a vacancy mechanism. Therefore, ef fort has been made to characterize these phase transition. Thermal analysis to measure th e temperature of this transition, as well as to evaluate decomposition behavior, has been an essential tool in these investigations. Conventionally, impedance data is analyzed us ing an equivalent ci rcuit consisting of a parallel resistor/ capacitor (RC) elements connected in series64. As shown in Figure

PAGE 64

54 5.3(a), R and C are the resistance and cap acitance of the circuit components. The complex impedance of the resistor and capa citor of RC element in parallel can be described as: jw R jwCR R Z 1 1 (5.2) where RC and is the time constant of the RC element. Each parallel RC element of the equivalent circuit used to represent an electro-active region within the sample will result in a semicircular arc in the complex impedance plane, Z plot as shown in Figure 5.3 (c). The frequency at wh ich the semicircular arc maximum occurs is determined by the time constant of the parallel RC element as described by: 1 1 max RC w (5.3) where maxw is the angular frequency at the top of the semicircular arc. According to Eqns. (40) and (41), the elements of the e quivalent RC circuit can be extracted from Z plots, and the fitting results of CsHSO4 sample was shown in Figure 5.3 (d). The curve n<1in the Figure 5.3(d) is draw according to the fitting results. Based on the simple RC equivalent circuit, the semicircle in Z plane should be a full one with its center on the real axis of the complex plane.

PAGE 65

55 Figure 5.3 The elements of the equivalent RC circuit and the fitting results of CsHSO4 sample. (a) The RC effective circuit; (b) the modified model: the resistor and the CPE element connected in parallel; (c) Nyquist plot of (a); (d) experiment al data (square dots) and fitting results (solid lines) of the CsHSO4 sample. Z' and Z" are the real part and image part of the complex impedance Z, respectively. Curve n=1 represents the RC model (a) and Curve n<1 represents the CPE model (b). However, as shown above, experimental da ta are only rarely found to yield a full semicircle, instead depressed one with its cen ter below the real axis. The observed plot was indeed the arc of a circle, but with the ce nter some distance below the x-axis. This is mainly due to the distribu tion of the relaxation time continuous or discrete, around a mean1 m mw. The observed plot is related to ch arge storage phenomena. The interface impedance behaves as if ther e are two components in para llel. The fictitious element responsible for this behavior is known as the constant phase angle element (CPE) in place

PAGE 66

56 of the capacitor. In order to circumvent that and obtain a good fitting, the CPE was introduced to replace the capacitor in the RC circuit, as shown in Figure 5.3 (b). And the impedance of CPE can be described as65: jw A ZCPE1 (5.4) where A and are frequency-independent para meters which usually depend on temperature, and 1 0 The CPE describes an ideal capacitor with A C for 1 and an ideal resistor with A R1 for0 and the experimental data is corresponding to 1 0 generally. The complex impedance of th e equivalent circuit of the resistor and CPE in parallel can be described as: jw R RA jw R Z 1 1 (5.5) where RA and Eq. (5.3) may be expressed as: 1 1 max RA w (5.6) According to the Eqns. (5.5, 5.6), equivalent circuit elements can be fitted from the Z plane. Curve n<1 in Figure 5.3(d) is fitting result by using CPE model, and fitting parameters (e.g. A R, and ). Curve n<1 is closer to the experimental data than Curve n=1, so the CPE element is an effective tool to fit the experimental data. The refined material resistance R is then normalized by using area specific resistance (ASR), then: ASR L (5.7)

PAGE 67

57 L is the length of the conductor, the lower the resistance and ASR = RA (representing the pellet resistance and area respectively). It is intuitive that a shorter path results in less resistance. The results of AC impedance measurements for polycrystalline samples under ambient pressure conditions are presented in Arrhenius plot in Figure 5.4(a, b) for modified and unmodified CsHSO4 respectively. 1.00E-05 1.00E-04 1.00E-03 1.00E-02 1.00E-01 1.00E+00 1.00E+01 1.00E+02 0.00150.0020.00250.0030.0035 T 1/K 1 Log conductivity,S/m Tsp a b 50 100 150 200 T/ oC Figure 5.4 Fitting the conductivity to Arrhenius law upon heating at 3 oC min-1. (a) modified; (b) unmodified (CsHSO4) respectively. TSP is the superprotonic phase transition temperature (142 oC). The activation enthalpy or migration enthalpy (Ha) of H+, carrying one elementary charge Ha 0.59 eV and the frequency factor, Ao obtained for modified and

PAGE 68

58 unmodified samples were 2.0 105 -1 cm-1 K and 0.5 105 -1 cm-1 K respectively. The activation enthalpy value obtained agrees well and falls within the reported literature value of Ha 0.3-0.66 eV reported for this material66. Representative Nyquist plots obtained at temperatures a bove the superprotonic phase transition temperature TSP are presented in Figure 5.5. The conductivity of both sample conditions increased sharply at the transition temperature. Mor eover, the jump in conductiv ity was significantly lower for the unmodified pellet than for the modified crystal sample. An explanation for this behavior is that more vacancies are intent ionally introduced into the lattice by ball milling thereby providing a faster route for hydrogen proton mobility. Above the TSP, neither sample type exhibited a linear, A rrhenius region, as is typical for other superprotonic conductors67. Below the superprotonic phase transition (T < TSP), Figure 5.5(a), well-resolved semi-circles were observed. The real resistance of the sample at low temperatures was identified by the intercept of the semicircle with the real axis (Z ). The apex of the semicircle, o, is the characteristic frequency defined in terms of the resistive and capacitive response of the materi al under test. In contrast, at T TSP, the resistance approaches that of a single crystal, Fi gure 5.5 (b), and the intercept with Z was used as the estimated resistance of the sample. In both cases (modified and unmodified), the conductivity increased by 3 to 4 orders of magnitude at the transition temperatures.

PAGE 69

59 Figure 5.5 Nyquist plots at various te mperatures upon heating (CsHSO4). Below and above its superprotonic transi tion temperature (Tsp) for modi fied polycrystalline pellet; (a) at T< Tsp; (b) at T> Tsp.

PAGE 70

60 In conclusion, high energy planetary milling process was applied successfully to improve the ionic conductivity of CsHSO4. The electrochemical impedance spectroscopy measurements indicate that the three hour mechano-chemical pulverization process by ball milling quadrupled the ionic conductivity of CsHSO4. However, simultaneous differential calorimetry and thermogravimetri c analysis indicate that the decomposition temperature of 202oC for both the modified and unmodifi ed remained constant. It is hoped that the application of ball milled CsHSO4-based membrane to electrolytic splitting of hydrogen sulfide gas will show a great promise in reducing the activation barriers associated with this type of process. 5.4 Permeability of H2S on CsHSO4 Membrane 5.4.1 Morphological Characterization of the Surface Any surface morphology changes of the CsHSO4 membranes due to H2S exposure were investigated by atomic force microscope (AFM). The surface structure was analyzed before and after the permeability study at several different locations in order to obtain an average roughness value. The AFM was utili zed in contact mode making use of a sample-stationary scanning device by means of a Digital Instrument piezoscanner with x, y, z cantilever displacements. The 1D AFM images with area of 1 m x 1 m of the CsHSO4 membranes before and after permeability testing are shown in Figur e 5.6. The surface analysis performed show that both samples in Figure 5. 6 (a, b) have a smooth textur ed surface constituted of rounded grains with a RMS roughness of 5 nm. Furthermore, the structures are continuous, with a well-defined geometry and a rounded profile. This observation

PAGE 71

61 indicates that prolonge d exposure to (100 %) H2S does not significantly affect surface morphology. A possible reason for the surface mo rphology retention may be attributed to the compact layered structure of the memb rane which resists diffusion of any kind (molecular, Knudsen and surface) on the surf ace of the membrane in contact with H2S even after prolonged hours (100 h) of operation. b a b b a a Figure 5.6 Atomic force microscope images of CsHSO4 membranes used in the permeation study. (a) image before permeability measurement; (b) image after permeation process. 5.4.2 Gas Chromatography and Mass Spectrometry Analysis Samples drawn from the receiving compartmen t were analyzed by gas chromatography and mass spectrometry (Shimadzu GC-MS-QP 5000) utilizing an Agilent technology column DB-5 with column length, thickness and diameter (30 m, 0.25 m, and 0.25 mm) respectively with helium be ing used as carrier gas. A GC analysis is shown in Figure 5.7. In Fi gure 5.7(b-f)), the spectra obtained indicated that no H2S gas was detected af ter the passage of H2S through CsHSO4 membrane for the three membrane thicknesses investigated. None of the spectra observed has a peak that matched the retention time of H2S peak. MS analyses identifi ed the remaining peaks as a

PAGE 72

62 composition of diatomic nitrogen and water, which infiltrated the system from the nitrogen purge gas and water displacement apparatus. The GC-MS results are further supported by the absence of pressure drop across the membranes tested as well the information obtained from the graphical repr esentation of Eqn. (2 .2). This is also consistent with an earlier study68 which involved splitting of H2S electrolytically using a CsHSO4-based membrane in which no H2S gas was detected in the product stream. Based on findings in this study and in co rroboration with earlier analyses, CsHSO4 is impermeable to H2S gas. Though a theoretical result vi a mathematical modeling capable of simulating quantitative H2S-transport has not been explored and is beyond the scope of this work. Figure 5.7 Qualitative comparison of perm eability at 150 C by GC-MS instrument. (a) Pure H2S gas; (b) measurement of H2S permeability through 1 mm thick CsHSO4 membrane; (c) permeability test with a 0.5 mm thick CsHSO4 membrane (d-e) effect of further reduction in CsHSO4 membrane thickness on H2S permeability at 150 oC during a period of 24 h and 100 h respectively.

PAGE 73

63 5.4.3 Micropore Analysis To compare pore size distribution in the ma terial before and after the permeability measurements, the distribution of pore volume with respect to pore size was determined with Autosorb-1 software, ASI WIN from Quan tachrome instruments before and after the permeation tests. The Autosorb-1 operates by measuring the quantity of gas adsorbed onto or desorbed from a solid surface at some equilibrium vapor pressure by the static volumetric method. This volume-pressure data can be reduced by the Autosorb-1 software into pore size and surface area di stributions and, micropore volume using an extensive set of built-in data reduction procedures. The analysis most suitable for this study used a type of ASI WIN soft ware based on the Saito-Foley (SF)69 method which assumes cylindrical pore geometry to cal culate pore size dist ributions within a microporous range. The apparent pore dist ribution isotherms us ing a computational approach based on the Saito-Foley (SF) met hod for the membrane before and after the permeation study are shown in Figu re 5.8. It is clea rly seen that both isotherms (a, b) remain congruent during 100 hours of data acquisition, which implies that the passage of H2S has no significant influence on the a pparent pore diameter distribution. Differentiation of weight (volume) of gas adsorbed relative to the total uptake, W/W0, with respect to the effective pore width yi elds a pore size distribut ion in the micropore range expressed as a Gaussian isotherm70. W is the weight adsorbed at P/P0 and T, and W0 is the total weight adsorbed. P/P0 and T are the relative pressure and adsorption temperature (K) respectively. The mean pore diameter values are randomly distributed around the average dSF = 10.5 . The two isotherms bot h exhibited a narrow range of pore size distribution with no evidence of any mesopores or macropores present as shown

PAGE 74

64 in Figure 5.8. One conclusion that can be inferr ed from this observation is that the pores in CsHSO4 are not connected. Furthermore, due to the range of pore diameter sizes and the narrow pore walls of the membranes, when compared to the mean free path of H2S molecules, could have contributed to the non porosity observed. No previous work on the pore size distribution of CsHSO4 could be found to compare our results; theref ore we rely on the accuracy of the measurements obtaine d before and after the permeability study in addition to our earlier observations to arrive at this conclusion. Figure 5.8 A typical Saito-Foley (SF) DV (d) method of pore si ze distribution isotherm. Measured on, (a) CsHSO4 membrane befo re permeation investigation; (b) CsHSO4 membrane after permeation investigation. 5.4.4 XRD Analysis The phases of CsHSO4 are sensitive to external c onditions, so the aim of the XRD analysis is for structural characterization of CsHSO4 membrane in order to identify the phases present, structure determination, a nd crystal orientation. The XRD pattern of CsHSO4 membrane before and after permeation te sting is presented in Figure 5.9 (a, b).

PAGE 75

65 In agreement with Belushkin et al.71 the XRD pattern for CsHSO4 membrane before and after permeation testing correspond to a mixtur e of monoclinic structur al phases III/II. It is clear from the diffractogram in Figure 5.9 (b) that there is no mixture of different phases from passage of H2S gas through the membrane. As a result, the structure and phases of the membrane are preserved. Th e diffractogram depict s a typical sample comprising one crystalline phase with the same average crystalline sizes. All XRD measurements were performed on samples at room temperature, but the different sample history exhibits the same XRD patterns. Figure 5.9 XRD diff ractogram of CsHSO4 with different sample history. Here, (a) XRD pattern of CsHSO4 membrane before permeability measurements; (b) XRD pattern of CsHSO4 membrane after permeability study. 5.4.5 Resistance Measurements of CsHSO4 Membranes Resistance measurements as a f unction of temperature of CsHSO4 were examined by EIS in the temperature range of 142–170 oC. This was carried out by two-probe AC

PAGE 76

66 measurements techniques. In the 2-probe se tup only two leads wire s are located at the endpoints of the copper strips. The apparatus is similar to the ionic conductivity measurements (see Figure 4.3). The use of alternating current (AC) im pedance spectroscopy has been made to characterize the resistances at two differe nt thicknesses (1 mm and 0.2 mm) of CsHSO4 membranes investigated for permeability as a function of temperature at ambient pressure. Since we have confirmed the impermeability of 0.2 mm thickness to H2S gas, it will be worthwhile to also study the effect such reduction will have on the membrane resistance. The temperatures selected fo r this examination were 142, 150, and 170 oC. These are the temperatures at which CsHSO4 experience a jump in its protonic conductivity the so-called the “superprotonic phase”72,73. Also, in the electrolytic splitting of H2S the viscosity of sulfur product is minimized at 150 oC such that it can flow out of the electrolytic cell without hindrance54. The Nyquist plot at th ree different temperatur es is shown in Figure 5.10 for the two cases of 1 mm and 0.2 mm CsHSO4 membranes. As was in the ionic conductivity measurements, real resistance of the samples at these high temperatures was identified by the intercept of the spectr a with the real axis (Z ). At the superprotonic transition phases (T TSP), only an electrode response was observed, so the intercept with Z was used as the estimated resistance of the sample.

PAGE 77

67 Figure 5.10 Nyquist spectra at and above CsHSO4 superprotonic transition temperatures (Tsp) for the two membrane thicknesses 1 mm and 0.2 mm. Reduction in resistance with membrane thickness is shown by the inter cept of the spectra with the real (Z ) axis. The decrease in resistance as the thickness is reduced from 1 mm to 0.2 mm is confirmed in the spectra for the three te mperatures. Although, it was e xpected that the reduction in thickness by 80 % will have equal effect on the resistance but, only about 30 % reduction in the resistance was achieved as shown by the Nyquist spec tra observed. Nevertheless, the correlation between the membrane thickn ess and its resistance as observed in the Nyquist spectra shows that thinner membra nes will be more effective in minimizing ohmic polarization experienced when this membrane is used for electrochemical synthesis59.

PAGE 78

68 5.4.6 Permeability The permeability of the membrane was tested by measuring the pres sure differences in the permeation compartments which could only loose pressure by the passage of material through the membrane74,75. Figure 5.11 shows permeability of H2S gas through CsHSO4 membrane and the pressure dr op observed as a function of time. The pressure drop observed was not sufficient to sustain any driv ing force necessary for material crossover through the membrane. Therefore, pressure di fferential in the cell remained absolutely constant even with continuous influx of H2S gas through out the experimental period. Also, the duration of the experimental peri od has no effect on the permeability of H2S on CsHSO4 membrane since this value remained fairly constant. The membrane maximum permeance is 0.09 Barrer (6.75 x 10-19 m2.s-1.Pa-1) which is insignificant to support any material crossover. The permeation profile of H2S for CsHSO4 membrane is a function of pressure difference across the membrane73.

PAGE 79

69 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 020406080100 Time, hPermeability ( K), Barrer0 5 10 15 20 25 30 35Pressure drop ( p), kPa Permeability Pressure drop Figure 5.11 Permeability and pressure drop as a function of time for CsHSO4 membrane. Both diffusive and convective gas transpor t contributed to the total flux over the membrane. We assumed that the mean free path of the gas molecule is greater than the pore diameter due to the microporous nature of our membranes and the low pressure regiment of operation. Therefore, it will be a reasonable assumption to state that Knudsen diffusion dominates over molecular diffusion th rough out the experimental time. Also the difractogram obtained from XRD analysis show only one crystalline phase present which is an indication that H2S did not adsorb on the surface of the membrane thereby ruling out any contribution from surface diffus ion. Diffusion coefficient of H2S through CsHSO4 (1.38 x 10-8 cm2 s-1) were based on estimation by Gilliland et al.76. The impact of molecular diffusion on the diffusivities is beyon d the scope of this investigation. But the resistance to diffusion along the pore could be attributed primar ily to molecular collisions

PAGE 80

70 with the wall rather than with each other, as in ordinary diffusion. Nevertheless, in the intermediate regime both wall collisions and intermolecular collisions contributed to the diffusional resistance and the effective diffusivity depends on both the Knudson and molecular diffusivities. 5.5 Electrochemical Synthesis 5.5.1 Current Density from H2S Splitting The current density generated at a constant cell pressure of 138 kPa and H2S flowrate 0.25 cm3/min is shown as function of voltage applied for modified and unmodified CsHSO4-based MEAs. Figure 5.12 (a, b) show current density as function of voltage based on the total exposed surface for modi fied and unmodified cases respectively. The results of Figure 5.12 show that higher cu rrent density was attainable by applying modified (Figure 5.12 (a) than the unmodifi ed (Figure 5.12 (b) CsHSO4-based MEAs. One explanation proposed is that the reductio n in particle size, which results from mechano-chemical pulverization of the ball m illing process, creates more defect structure in the material. This in turn facilitates the rate at which hydrogen protons are being transported from the reaction sites he nce, favoring the reaction kinetics.

PAGE 81

71 Figure 5.12 Current density as a function of voltage generated with anod e catalyst S3. (a) using modified (CsHSO4-based MEA); (b) applying unmodified CsHSO4-based MEA). For both scenarios in (Figure 5.12), the curv e is linear as Ohms law at lower current density. At higher currents diffusion limitations seem to be coming into play. However, other factors such as, mass transport lim itations (concentration losses), electronic conduction and activation losses (losses due to electrochemical reaction) will affect the cell performance. In considera tion that this is a novel invest igation, our preliminary data was on the proof of principle, therefore, sy stem optimization will take into account the effects of these factors on the system perfor mance which is the topic of the factorial experiments covered later in the document. 5.5.2 Electrolysis of H2S Examination of the performance of the MEAs at a constant applied voltage of 900 m V was carried out on three differe nt anode catalysts configurat ions as shown in Table 4.1 for both cases of modified and unmodified CsHSO4. Our intention is to minimize the

PAGE 82

72 overvoltage required to split H2S electrolytically, hence 900 m V was selected for this study. The sustainable application of anode catalyst (S1) for 8 h of operation has now been demonstrated. The conversion of H2S (feed: 100 %), during the first 8 hours of operation, is shown in Figure 5.13. In Figure 5.13 (a, b), curves (S1) (representing the anode catalysts used) shows that conversion increased steadily until it attained a constant value of 100 % and 92 % respectively after 5 hours of operation. Th e conversion rate is limited by the membrane current density attain able. Figure 5.13 (a, b), curves (S2) and (S3), which represent the anode catalyst used, did not show appreciable H2S conversion under the same conditions. Figure 5.13 Relationship of H2S conversion to time on stream with a voltage 900 mV. (a) modified (CsHSO4-based MEA) with anode catalysts configurations: S1, S2, S3; (b) unmodified (CsHSO4-based MEA) at different anode catalysts combinations: S1, S2, S3. The exclusion of Pt black in curves (S2 and S3) decreased the rate at which electrons are being conducted away from the electrolyte surf ace after reaction. Pt black as an electron conductor facilitated the rate at which these electrons are being tr ansported away from the reaction surface as H2S conversion proceeds as eviden t in curves (S1) which have 0 20 40 60 80 100 0246810 Time, hH2S Conversion, % a b c 0 20 40 60 80 100 120 0246810H2S Conversion, % a b c Time, ha b

PAGE 83

73 high rates of H2S conversion. Curve (S3) shows the lo west t performance because of the absence of a good porous media for gas diffusion which decreases the rate at which feed gas is transported to the reaction surface. The inclusion of p-dichlorobenzene in catalysts (S1) and (S2) provides a good porous media needed for gas to contact the reaction surface easily and rapidly. In the final analysis, modified CsHSO4-based MEA (Figure 5.13 (a)) has a higher rate of H2S conversion when compared to the unmodified CsHSO4based MEA (Figure 5.13 (b)) which exhibi t a lower conversion rate, although, this increase is more prominent in the presence of Pt black catalyst and p-dichlorobenzene as evident in Figure 5.13 (a) curves (S1 and S2). A proposed explanation of this is the action of the mechano-chemical pulverization of CsHSO4-based MEAs which enhanced the rate at which hydrogen protons hop from position to position within the lattice and ensures that reaction proceeds smoothly by transporti ng hydrogen proton away from the reaction surface as soon as they are formed. This conclusion is supported by conductivity measurements which indicate that modified CsHSO4 conductivity is improved by the action of the ball milling process. The relationship of current generated verses time is shown in Figure 5.14 for the two cases of modified and unmodified CsHSO4-based MEAs under study. As expected, anode catalyst (S1) proved to be a s uperior catalyst since it prov ides favorable kinetics over catalyst configurations (S2 and S3). The current generated was sustainable over an extended operating period (5 h), as evident fr om the values between the third and eight hours of operation.

PAGE 84

74 Figure 5.14 Relationship of current to time dur ing a period of 12 h of operation with a voltage 900 mV. (a ) modified (CsHSO4-based MEA) at diffe rent anode catalysts configurations: S1, S2, S3; (b) unmodified (CsHSO4-based MEA) with different anode catalysts combinations: S1, S2, S3. As seen in Figure 5.14, the performance of catalyst (S1) degrades after 8 hours of operation. Such degradation was not seen for ca talysts S2 and S3. It is proposed that this behavior is due to concentra tion polarization resulting from sulfur poisoning of Pt black catalyst. For sustainability over a long operational period, additional investigation is warranted as future work in order to minimi ze this effect. The durability of the catalysts and membranes, and sensitivity to pore bl ocking or catalyst deactivation by deposited sulfur, will be presented in a subsequent section. Again, the current generated was less for the unmodified than for the modified CsHSO4 as expected given that current generated is a function of ionic conductivity of the electrolyte. Nevertheless, the correlation of current generated and H2S conversion proves that the process is indeed the due to the electrolytic splitting of H2S, and not as a result of interaction between MEA materials. 0 1 2 3 4 5 6 051015 Time, hCurrent, mA S1 S2 S3 a 0 1 2 3 4 5 051015 Time, hCurrent, mA S1 S2 S3 b

PAGE 85

75 5.5.3 Material Balance H2S was electrolyzed for 5 hours at a steady voltage of 900 mV with anode catalyst S1 (see Figure 5.13 (a)) which produced a sustained cu rrent of 0.00531 A. At the anode, the half redox equation for electrolysis is: H2S 2H+ + 2e+ (1/8) S8 (5.8) 2 moles of electrons (2 Faradays) ar e required to produce 1/8 mole of S8 and 2 moles of hydrogen ions. A current of 0.00531 A for 5 hours represents 95.58 coulombs or (95.58/96,500) = 0.001 Faraday. This is sufficient to pr oduce 0.001 moles (0.001 g) of hydrogen ions and 6.25E-5 moles (0.016 g) of S8. At the cathode, the half redox equation for electrolysis is: 2H+ + 2eH2 (5.9) The 2 moles of hydrogen ions produced in Eqn. 5.8 are converted to 1 mole of hydrogen gas in Eqn. 5.9. The 0.001 moles (0.001 g) of hydrogen ions produced at the anode by 0.001 Faraday are converted to 0.0005 moles (0.0005 g) of hydrogen gas at the cathode. For the gas measurement, the gas issuing from the cathode compartment was collected and measured by displacement of water in a 50 cm3 burette at 25 oC and 102.4 kPa. The observed volume was 10 cm3. Gas collected by displacement of water is saturated with water vapor whose vapor pressure at 25 oC is 3.17 kPa. The partial pressure of hydrogen in the burette was then (102.4 -3.17) = 99.2 kPa. By the ideal gas law, 0.0004 moles of gas has a volume of 10 cm3 at a partial pressure of 99.2 kPa and temperature of 25 oC. Thus the measured volume of hydrogen gas is about 80 % of the theoretical limit for the electrolysis as calculated in the previous section. We are unable to account for the

PAGE 86

76 remaining 20 % which may have resulted due to some current leakage in the system; this is currently under investigation. 5.5.4 Stability of H2S Electrolytic Cell Anode Materials Electrocatalysis offers a convenien t route to the synthesis of H2S. Requisites of the anode material are a low overpotential for product evolution and high stability. We have tested a group of anode electrocatalysts for their suitability as endurance materials for H2S splitting. A particular one of interest is the metal sulfide group comprising RuO2/CoS2 composite which has shown in this application to be stable and efficient in splitting H2S gas. Figure 5.15 illustrates the XRD pattern for RuO2/CoS2 as prepared, after 12 h exposure to H2S, and after 24 h exposure to 100 % H2S content feed gas at 150 oC. This composite shows no phase or compositional changes after exposure to (100 %) H2S gas at the cell operating temperature. The chemical stability of this material in H2S was expected, as the preparation of RuO2/CoS2 anode composite was done with the intention to increase its catalytic activity via the mechano-chemical pulverization process in inert (99.5 % nitrogen) atmosphere at elevated temperature 150 oC. The stability of RuO2/CoS2 when exposed to H2S (for prolonged periods of time) made it an attractive an ode candidate for H2S splitting.

PAGE 87

77 253545556575Intensity, a.u.ab c Cu K 2 degree Figure 5.15 XRD pattern for (RuO2/CoS2 composite). (a) as pr epared; (b) after 12 h exposure to 100 % H2S feed content at 150 oC; (c) after 24 h exposure to 100 % H2S feed content at 150 oC. The catalytic active surface area of the composite was obtained by using a Multipoint BET method which requires a minimum of th ree points in the appropriate relative pressure range as shown in Figure 5.16. At hi gh relative pressures, the rate of adsorption equals the rate of desorption which is very essential for material catalytic performance. The specific surface area calculated from the total surface area for the composite is (10.96 m2/g). This is sufficient to sustain the reaction without interruption by providing sufficient compensation for sulfur poisoning as well as providing enough area for the reaction for the react ion to proceed.

PAGE 88

78 Table 5.2 shows a list of different anode configurations and their corresponding BET surface areas. The small catalytic active areas exhibited by both RuO2 and CoS2 when compared to RuO2/CoS2 composite could be responsible for the blockage of the anode surface by sulfur as observed in the SE M image (see Figure 5.17). As mentioned previously, this occurrence decreases the ra te of chemical reaction which adversely affects the overall performance of the system. Figure 5.16 Multipoint BET plot for anode metal sulfide (RuO2/CoS2) nanocomposite.

PAGE 89

79 Table 5.2 Various anode configurations and their BET catalytic active areas. Catalysts BET Specific Surface Areas (m2/g) CoS2 6.77 RuO2 6.41 RuO2/CoS2 10.96 The composite RuO2/CoS2 has a relatively high surface and expanded area metallic nanoparticles facing away from the electrolyt e. Figure 5.17 shows a scanning electron micrograph of a porous electrode (composite nanoparticles on carbon mesh support for strength). These electrodes have tortuous pa thways within them to expose orders of magnitude larger surfac e areas to reacting H2S and to also allow the escape of the liquid sulfur product. The increased numbers of catal ytically active sites as mentioned earlier enables the materials to be more resistant to poisoning. This is because these materials exhibit a certain number of catalyt ically active sites that can be sacrificed to the effects of poisonous sulfur production as shown in Figure 5.17 (b) while a large number of unpoisoned sites are preserved. In Figure 5.17 (c, d) whic h consists of non composite materials, the catalytic efficiency of the material is substantially less than that possible if a greater number of catalytically active sites were available. This was due to concentration polarization caused by sulfur bl ockage of the diffusion layers. In all the runs involving RuO2/CoS2 composite, there was only a tr ace amount of sulfur deposition on the anode surface.

PAGE 90

80 Figure 5.17 SEM images of surfaces of electrode. (a) RuO2/CoS2 electrocatalyst nanocomposite before electrolysis; (b) RuO2/CoS2 electrocatalyst nanocomposite after electrolysis; (c) RuO2 electrocatalyst after electrolysis; (d) CoS2 electrocatalyst after electrolysis. 5.5.5 Electrochemical Performance of Anode Electrocatalysts The first sets of experimental runs, as show n in the previous sections, were aimed at examining the stability and performance contri bution of the Pt electrocatalyst. It was shown that Pt degrades with time even though it gave the best performance but at a short period of time. Figure 5.18 compares the perfor mance of four cells using: Figure 5.18 (a) modified with RuO2/CoS2 and Figure 5.18 (b-d) modified with CoS2, modified with RuO2 and unmodified with RuO2 respectively as anode materials at 150 oC. Unmodified and modified as used in this context is the same as in pr evious sections The conversion process did not affect any of the anode materi als used and no de-lamination of the anode from the CsHSO4 electrolyte was apparent. The performance of the RuO2/CoS2

PAGE 91

81 composite anode was far superior to any of the anode materials. All the configurations initially increased, and subse quently maintained their pe rformances for over 12 hours. Consequently, prior to measuring performance in all subsequent experimental runs, cells were operated at an input voltage of 900 mV and allowed to stabilize. If no noticeable decline in output after this time period was observed, the cell was deemed adequate for performance testing. Both systems had simila r electrochemical behavior and deviations were probably caused by external factors (current collection, lead wires, etc.). Factorial experiments supported this observation. 0 2 4 6 8 10 12 14 16 18 20 0246810121 4 Time hCurrent density, m A/cm 2a b c d Figure 5.18 Testing for four cells with different anodes (Vcell = 0.9V, T = 150 oC, Fuel = 0.25 cm3/min) H2S was the fuel. (a) modified with RuO2/CoS2; (b) modified with CoS2; (c) modified with RuO2; (d) unmodified with RuO2.

PAGE 92

82 The main goal of this study was to iden tify and develop endurance materials for H2S splitting electrolytically, thus full system optimization was considered a secondary goal. The large number of anode materials and experi mental cells used in this project required the use of low cost materials with no system optimization configurations. For commercial applicability of this work, full system optimi zation will need to be carried out shown in Figure 5.19, are the effects of various anode configurations on current densities as a function of external ap plied voltages. Curve (a) represen ts the performance of the cell based on modified electrolyte with composite the RuO2/CoS2 anode. Curves (b-d) represent the performances of cel ls based on modified with CoS2, modified with RuO2 and unmodified with RuO2 anodes respectively. At the low region where there are uniform potential gradients, cu rrent densities are proportiona l to potential gradients and are uniform; hence Ohm’s law is appropriate. At high voltages, concentration gradient comes into play. By utilizing the anode composite RuO2/CoS we were able to maintain a high current density at low potential. This is because the catalytic surface area provided by the composite highly minimi zed the activation barrier and also reduced most of the effects associated with concentr ation polarization; more gases were able to diffuse to the reaction interface and the reaction equilibrium is shifted toward product.

PAGE 93

83 Figure 5.19 Current density for differen t anode configuratio ns with 100 % H2S feed gas content. (a) modi fied with RuO2/CoS2; (b) modified with CoS2; (c) modified with RuO2; (d) unmodified with RuO2. 5.5.6 Resistivity of Anode Catalysts The electrical resis tivity of anode catalyst sample s with the structure Au/anode catalyst/Au are listed in Table 5.3. The resistivity of composite catalyst pellet RuO2/CoS2 at 150 oC is less compared to the values obtained for CoS2 and RuO2 electrocatalysts. In the light of this observation, the pr esent anode catalyst with the RuO2/CoS2 configuration was utilized due to superior electrical conductivity for use as anode catalysts under operating conditions. The contribu tions from these resistances are small compared to the total cell resistances. This sugge sts that the large different in current densities observed may not be significantly due to the resistances of these materials, but primarily due to the availability of cataly tic active sites provided by the composite anode materials. However, it was necessary to reduce the internal ohmic resistance of the cell slightly using

PAGE 94

84 RuO2/CoS2 anode composite. Note, as mentioned previously, the total resistances experienced by the electrolytic cell result s from contributions due to, activation overvoltage loss ( act) that from ohmic loss ( ohmic) and from concentration loss ( con). Furthermore, the contribution from ohmic comes from both the ionic and electrical resistances i.e. j Rohmic = j(Relec + Rionic). Maximum current density as shown in Table 5.4 at 900 mV measured for the RuO2/CoS2 composite was 19 mA/m2, which was higher than that for CoS2 (9 mA/cm2 ) using the same electrolyte thickness. To this end, we have achieved minimizing the c ontributions from both act and con by using the anode composite RuO2/CoS2 and also reducing the effect of Relec. The ohmic resistance was determined from impedance measurements. Table 5.3 Cell area specific ohmic resist ance (RA). This shows different anode electrocatalysts at oper ating temperature of 150 oC. Anode materials RuO2/CoS2 CoS2 RuO2 RA ( cm2) 3.12 7.30 9.23 5.5.7 Fuel Utilization Table 5.4 summarizes the measured data in tabular form at specif ic operating voltage, comparing typical conditions to tested para meters. The data shows high reproducibility. The conversion efficiency depends on many factors which include the flowrate of H2S to the reaction sites and the trans portation of product away from th ese sites, the availability of active sites, and the thickness and nature of the electrolytes As shown in Table 5.4, for

PAGE 95

85 the same electrolyte thickness, H2S consumption was twice the value for modified with RuO2 as compared to unmodified with RuO2. This suggests that the difference could be attributed to faster electrode ki netics observed with modified RuO2 as explained in the previous sections. The fuel consumption rate s had equal effects on the system efficiency for these two systems. By using a different ca talyst configuration co mprised of modified electrolyte with CoS2 and reducing the electrolyte thic kness by 80 % of its value boosted the system efficiency by 30 %. This led to a high er rate of fuel utilization almost tripling that observed in the case of m odified electrolyte with the RuO2 configuration. The contribution from the low resistance value of CoS2 when compared to RuO2 and the effect of reduction in ohmic losses as a resu lt of decrease in the electrolyte thickness contributed to the increase observed. As pointed out prev iously, the reduction in the thickness of the electrolyte enhanced the fuel consumption. The introduction of the modified electrolyte with anode the composite RuO2/CoS2 using the same electrolyte thickness used in the modified with CoS2 doubled the system effi ciency and the fuel utilization. The increased electr ode kinetics is attributed to the large active sites available for the reaction to progress, some of which were sacrificed to th e effect of sulfur poisoning. This was also supported by the SEM image (Figure 5.17) after 12 hours of electrolysis.

PAGE 96

86 Table 5.4 Conversion efficiency for several el ectrolytes. Electrolytes prepared from high surface composite nanometals (RuO2/CoS2) demonstrate the best performance. 0.9 V Fuel utilization j, A/cm2 Cell area, cm2 Efficiency H2S flowrate, cm3/min Electrolyte thickness, cm 100 % utilization 0.03000 0.4 1 0.16 Unmodified with RuO2 0.002475 0.4 0.06 0.01 0.1 Modified with RuO2 0.003075 0.4 0.13 0.02 0.1 Modified with CoS2 0.009025 0.4 0.31 0.05 0.02 Modified with RuO2/CoS2 0.019125 0.4 0.69 0.11 0.02 5.5.8 Tafel Slope and Exchange Current Densities for Anode Configurations Improving kinetic performance stems from increasing j0. To increase the value of j0, we have increased the number of possible reaction sites (i.e., increase the reaction interface roughness) by incorporating a novel composite metal sulfide, RuO2/CoS2 electrocatalyst, in the electrode. The exchange current density is determined experimentally by extrapolating plots of n j vs. to = 0. This is a direct m easure of the reaction rate constant at the electrolytic cell electrode. Figure 5.20 show s various plots for the anode configurations. One can use the slope t of the straight line to estimate the number of electrons involved in the el ectrochemical process. In all four cases of anode configurations n = 2 indi cating the oxidation of H2S to S2. The anode composite consisting of RuO2/CoS2 shows great potential in reducin g the activation barrier as a result of the increased catalytic active sites available for the reaction to proceed without

PAGE 97

87 sulfur poisoning. This is reflected by th e high exchange current density obtained compared to the configurations RuO2 and CoS2 which have lower values of j0. For the composite configuration, the transfer coefficient is 0.68 and j0 is 2.53 x 10-2 A/cm2 signifying a rapid elec trochemical reaction. Figure 5.20 Tafel plots for anode confi gurations at operating temperature 150 oC and 100 % H2S feed content. 5.5.9 Factorial Experiment 5.5.9.1 Effect of Process Variables on Cell Resistances for RuO2/CoS2 Cell polarization resistances are functions of charge transfer (electrochemical reaction) and mass transport (reactant and products) activities in the cell. A Labview program was

PAGE 98

88 used to measure the effect of the three pr ocess variables on cell polarization resistances. A simplified graphical representa tion of results is shown in Figure 5.21. In this figure, straight lines indicate a zero e ffect for that particular process variable on cell polarization resistance. As expected, electrolyte thickness and age were the dominating factors on cell polarization resistance. It is not surprising that resistance decreased with age as this is in agreement with a previous study39 which showed that the presence of H2O molecules on the surface of CsHSO4 crystal increases io nic conductivity. Since th is material was not stored in a vacuum, it was prone to water va por interaction attack on the surface of the material. However, examination of these variables shows that the most prominent dominating factor in performa nce is the electrolyte thickn ess. The interaction between these two variables is shown in Figure 5.22. As mentioned previously resistance scales with electrolyte thickness and when normalized with area-specific resistance, a shorter path results in less resistance. Thus, improvement of electrochemical cell pe rformance can be achieved by utilizing thin and hydrated electrolyte membranes even though the latter contribution is not prominent. This new technique would either be more electronically conductive than previous method, or catalytically activ e towards the splitting of H2S. XRD analysis showed all of MEA materials were stable when pellets of this material were exposed to H2S-rich environments. XRD analysis performed after th e electrochemical testing of cells was also unable to detect changes in e ither the crystal structure or composition of MEA materials after extended ope ration with 100 % H2S feed gas. A second e xplanation was obtained

PAGE 99

89 from kinetic experiments performed on the RuO2/CoS2 composite anode (see the exchange current density section of this document).

PAGE 100

90 0 25 50 5791113 Time, hCell resistance, ohms cm2 modified with RuO2/CoS2 30 55 80 00.511.5 Thickness, mmCell resistance, ohms cm2 modified with RuO2/CoS2 47.35 47.4 47.45 02468 Age, daysCell resistance, ohms cm2 modified with RuO2/CoS2 Figure 5.21 Measured effects of three process va riables on cell polarization resistances. Time effect (top), electrolyte thickne ss (center), electrolyte age (bottom).

PAGE 101

91 30 50 70 00.30.60.91.2 Electrol y te thickness m m Cell resistance, ohms cm2 Thickness-modified with RuO2/CoS2 Age-modified with RuO2/CoS2 Figure 5.22 Factorial results showing the inte raction effect between electrolyte age and electrolyte thickness on cell polarization resistance. 5.5.9.2 Effect of Process Variables on Current Density for RuO2/CoS2 Cell current density is invers ely proportional to cell polari zation resistances. Since all other experimental variables were left unalt ered (i.e. H2S flow, lead wires, etc.), cells with the lowest polarization resistance had the maximum current density. The effect of process variables on cell power is inversely proportional to that of cell resistance as shown in Figure 5.23. The benefit from factor ial experiments is the ability to predict regions of optimum system performance (for future experiments) based on past results. With everything else equal, making the membrane thinner reduces the ohmic loss. However, note that the payoff does not scale directly with membra ne thickness. Although the membrane thickness was cut by 80 %, th e ohmic loss was only reduced by one-third.

PAGE 102

92 This is because not all of the electrolytic cell’s resistance contributions come from the electrolyte. 0.0189 0.01915 0.0194 5791113 Time, hCurrent density, A/cm 2 modified with RuO2/CoS2 0.01 0.03 0.05 00.511.5 Thickness, mmCurrent density, A/cm2 modified with RuO2/CoS2 0.019 0.0195 0.02 048 Age, daysCurrent density, A/cm 2 modified with RuO2/CoS2 Figure 5.23 Measured effects of three pro cess variables on cell current density. Time effect (top), electrolyte thickness (c enter), electrolyte age (bottom).

PAGE 103

93 6 CONCLUSIONS AND RECOMMENDATIONS In this investigation, a novel thin MEA from a solid acid material (CsHSO4) and innovative composite anode electrocatalysts have been developed for the electrolytic splitting of (100 %) H2S feed content gas operating at 135 kPa and 150 oC. A new class of anode electrocatalyst with general composition, RuO2/CoS2, and an improved proton conductor, CsHSO4, have shown great stability and desi red properties at typical operating conditions. This configuration showed stable electrochemical operation for 24 hours with a (100 %) H2S fuel stream at 423 K. This same system exhibited a maximum current density of (19 mA/cm2) at 900 mV. The performance of this new anode electrocatalyst when compared to that of Pt black investig ated in a previous study showed an overall superiority in its applicati on with a performance of over 40 % improvement in current density. This was an unexpected turnout considering that Pt has always been regarded as the state-of-the-art electrocatalyst used both in catalyzing the anode and the cathode in many applications. In the case of Pt black, su lfur poisoning of the anode material was a major issue after 8 h of exposure to H2S, which degraded the sy stem performance. This poisoning effect is irreversible which led to de-lamination of the MEA materials and blockage of the diffusion layers making it difficu lt for feed to be transported effectively. This detrimental effect is due to the fact that the state-of-the art Pt catalytic effect is based on crystalline catalytic active sites, which depend on the surface irre gularities and may not be sufficient to accommodate the effect of sulfur poisoning. Th is was circumvented

PAGE 104

94 by introducing a composite anode material whic h provided superior ca talytic properties compared to Pt as well as reduced cost. Howe ver, of all the resist ances encountered, ionic (electrolyte) resistance domina ted over others. We have ach ieved a 30 % increase in the overall system performance by fabricating a thin (200 m) CsHSO4 electrolyte, which reduced the whole MEA thic kness from 2.3 mm to 500 m. The result of permeability measurements proved that this thin solid electrolyte is impermeable to H2S gas and physical integrity was preserved throughout the experiment al period. Further resistance losses were compensated by using high energy planetary milling system to enhance the ionic conductivity of CsHSO4. The difference in stability and electrochemical performance of our cells compared to that of Pt anode based systems is directly attributable to the anode materi als developed in this project. Factorial experiments were used to charac terize the effect of controllable process variables (electrolyte thickness, time, age of the electrolyte) on the cell current density and interfacial polarization resistances. As e xpected, cell current density and interfacial polarization resistances were a function of electrolyte thic kness and age. Nevertheless, the effect of electrolyte thickness has a more prominent effect on the measured parameters. In addition, these experiments were used to identify regions of optimum system performance. Tafel plots were constructed to investigate the kinetic behavior of various anode based electrocatalysts. Exchange current densitie s, which are directly a measure of the electrochemical reaction kine tics, increased with RuO2/CoS2-based anodes. These

PAGE 105

95 experiments also suggested that high levels of fuel utiliz ation were possible using these materials. This was an impressive resu lt considering the drastic improvement in electrochemical performance, current density, and sulfur to lerance compared to e other anode configurations. 6.1 Increase Reactant Concentration The thermodynamic benefit to increasing reac tant concentration is minor due to the logarithmic form of the Nernst equation. In contrast, the kinetic benefit to increasing reactant concentration is significant, with linear rather than logarithmic impact. By operating the electrolytic cell at higher pr essure, we can increase the reactant gas (H2S), improving the kinetics commensurately. Un fortunately, the kinetic penalty due to decreasing reactant concentrati on is likewise significant. Again, reaction concentration tends to decrease at cell electrodes during high current density operation due to mass transport limitations. Essentially, the reactant s are being consumed at the TPB faster than they can be replenished, causing the local r eactant concentrations to diminish. This interaction between kinetics and mass transport was not part of this investigation. In this study, we have improved jo by decreasing the activation barriers and increasing the number of possible reaction sites using the novel anode composite RuO2/CoS2. 6.2 Electron Collection The bulk of the resistance in CsHSO4/RuO2/CoS2-based cells can be directly attributed to the current collector and l ead wires. They accounted for over 70 % of the bulk cell resistance. The system was setup such that leads and current collector (carbon mesh) are

PAGE 106

96 attached to the electrolytic anode. The same type of setup was carri ed out in the cathode section. Differences in the ther mal expansion properties of RuO2/CoS2, the mesh current collector, and the CsHSO4 electrolyte lead to a partial de-lamination of the anode from the electrolyte pellet after pr olonged operation. Despite that the anode was still in direct contact with the electrolyte the contact region between both components was reduced, causing an increase in the bulk cell resistance. One way of minimizing the cell resistance will be to replace the carbon mesh current collector. The MEA was expected to have lower electric conductivity compared to the carbon mesh, and this will result in low electro chemical activities and increase in ohmic polarization of the anodic cell. Another possi bility would be the layering of conductive material over the MEA anode pr ior to cell operation. Thin me tal plates with a corrosion resistant surface can significantly reduce the vol ume and weight of an electrolytic system, although the long term stability of such coatings needs improvement77. This will ensure smooth electron and gas transport and conse quently, minimize ohmic polarization of the cell. The most commonly used material for low temp erature fuel cell flow plates is graphite. This can be applied to electrolytic cells as well because it satisfies most of the criteria discussed except: ease of manufacturability, cost, and high mechanical strength. These criteria are not fulfilled because of costly m achining requirements and intrinsic brittleness of the material. New materials have to be sought or promising technology capable of

PAGE 107

97 making graphite cost competitive and ease of manufacture are in place. This material is stable and electronically conductivity over a wide temperature range in H2S atmosphere.

PAGE 108

98 REFERENCES 1 J. K. Kroschwitz, M. Howe-Gra nt (eds.), Encyclopedia of Chemical Technology, Vol. 23, Wiley Press, pg. 276, New York (1997). 2 N. Keller, C. Pham-Huu, C. Cr ouzet M. J. Ledoux S. Savin-Poncet J.B. Nougayrede, J. Bousquet Catalysis Today 53, 535–542, (1999). 3 D.A. Boysen, C.R.I. Chisholm, S. M. Haile, S.R. Narayanan, J. Electrochem. Soc. 47 3610 (2000). 4 B. Zhu, Fuel Cell Bull., 4, 9, (1999). 5 H. Iwahara, in Proc 17th Ris Int. Symp. Mat. Sci. (eds Poulsen, F. W. et al.) 13– 28 (Ris National Laboratories, Roskilde, Denmark, (1996). 6 S. M. Haile, D. A. Boysen, C. R. I. Chisolm, R. B. Merle, Nature 410, 910 (2001). 7 K. D. Kreuer, J. Membr. Sci. 185, 29 (2001); Solid State Ionics 125, 285 (1999). 8 R. Parsons, Trans. Faraday Soc. 54, 1053 (1958). 9 S. Trasatti, J. Electroanal.Ch em, 111, 125 (1980). 10 J. A. Rodriguez and M. Kuhn, J. Vac. Sci. Technol. A, 14, 1609 (1996). 11 J. Donini, K.T. Chuang, S.V. Sl avov, A. R. Sanger and V. Stanic, US 6,241,871 B1, June 5 (2001). 12 T. J. Kirk and J. Winnick, J. Electrochem. Soc., 140, 3494 (1993).

PAGE 109

99 13 Tetsuya Uda and Sossina M. Haile ,z, Electrochemical and Solid-State Letters, 8, 5, A245-A246 (2005). 14 T. Chivers, J. B. Hyne, and C. Lau, Int. J. Hydrogen Energy, 5, 499 (1980). 15 http://www.wpbschoolhous e.btinternet.co.uk/pg. 04/4_75trans.htm. 16 Jurado J F, Ortiz E and Vargas R A Meas. Sci. Technol. 8 1151–5, (1997). 17 R. Fitzgerald, Solid State Ionics 125, 285 (1999). 18 J. A. Rodriguez an d M. Kuhn, J. Vac. Sci. Technol. A, 14, 1609 (1996). 19 T. Uda, S. M. Haile, Electrochemical and Solid-State Letters, 8, 5, A245-A246 (2005). 20 T. Uda, D. A. Boysen, S. M. Haile, Solid State Ionics 176, 127–133 (2005). 21 B. Yang, A.M. Kannan, A. Ma nthiram, Materials Research Bulletin 38 691–698 (2003). 22 A.I. Baranov, L.A. Shuvalov, N.M. Shchagina, JETP Lett. 36 459 (1982). 23 A. Pawlowski, Cz. Pawlaczyk, B. Hilzcer, Solid Stat e Ionics 44 17 (1990). 24 T. Norby, M. Friesel, B.E. Mellander, Solid Stat e Ionics 77 105 (1995). 25 V.V. Sinitsyn, A.I. Baranov, E.G. Ponyatovsky, L.A. Shuval ov, Solid State Ionics 77 118 (1995). 26 A.V. Belushkin, M.A. Adams, S. Hull, L.A. Shuvalov, Solid State Ionics 77, 91 (1995). 27 K. Itoh, T. Ukeda, T. Ozaki, E. Nakamura, Acta Crystallogr., C Cryst. Struct. Commun. 46, 358 (1990).

PAGE 110

100 28 A.V. Belushkin, W.I.F. David, R.M. Ibberson, L.A. Shuva lov, Acta Crystallogr., B47 161 (1991). 29 J. Lipkowski, B. Baranowski, A. Lunden, Pol. J. Chem. 67 1867 (1993). 30 C.R.I. Chisholm, S.M. Haile, Mater. Res. Bull. 35 999 (2000). 31 J. Baran, M. K. Marc hewka, Journal of molecular structure, pg. 133149 (2002). 32 C. R. I. Chisholm and S. M. Haile, Materials Research Bulletin 35 999–1005 (2000). 33 A.M. Balagurov, A.I. Beskrovnyi B.N. Savenko and B.V. Merinov. Phys Status Solidi A 100, pg. K3–K7 (1987). 34 P.L. Antonucci, A.S. Arico, P. Creti, E. Ramunni, V. Antonucci, Solid State Ion 125 431 (1999). 35 T. Ishii, J. Kawamura, J. Phys. Soc. Jpn., 67 (10), pg. 3517 (1998). 36 M.G. Lazarraga, J. Ibanez, M. Tabellout, J.M. Rojo, Composite Science and Technology, 64, pg. 759-765 (2004). 37 S.M. Haile, Solid State Ionics and Electroceramics Research Group, Improved solid acid electrolyte at Caltech, Materials Science Department, Pasedena, California, USA, (2004). 38 J. Otomo, N. Minaga, Solid State Ionics, 156, pg.357-369 (2003). 39 E. Ortiz, R.A. Vargas, B.E. Me llander, J. Phys.: Condens. Matter 18 pg. 9561– 9573 (2006). 40 J. Maier, J. Phys. Chem. Solids, 46 (3), pg.309 (1985). 41 Ph. Colomban, M. Pham-Thi, A. Novak, Solid Sate Ionics, 24, pg. 193203 (1987).

PAGE 111

101 42 K. Itoh, T. Ozaki, E. Nakamu ra, International Union of Chrystallography, 37B, pg.1908-1909 (1981). 43 H. Matsunaga, K. Itoh, E. Nakamura Journal of the Physics Society in Japan, 48, pg. 2011-2014 (1980). 44 M. Pham-Thi, Ph. Colomban, A. Novak and R. Blink, Journal of Raman Spectroscopy, 18, pg. 185 (1987). 45 A. Goypiron, J. de Villepin and A. Novak, Journal of Ra man Spectroscopy, 9, pg. 297 (1980). 46 L. Borjesson and L.M. Torell, Physics Review, 32, pg. 2471 (1985). 47 Ph. Colomban, M. Pham-Thi and A. Novak, Solid State Ionics, 21, 1986, pg.125. 48 C. Berger (Ed.). Handbook of Fuel Cell Technology. Prentice-Hall, Englewood Cliffs, NJ, (1968). 49 R. Bird, W. Stewart, and E. Lightfoot. Transport Phenomena, 2nd ed. Wiley, New York, (2002). 50 E. L. Cussler. Diffusion: Mass Tran sfer in Fluid Systems. Cambridge University Press, Cambridge, (1995). 51 F. Delogu, G. Cocco, Journa l of Alloys and Compoun ds 436, pg. 233-240 (2007). 52 Sossina M. Haile, Dane A. Bo ysen, Calum R. I. Chisholm & Ryan B. Merle Nature |Vol 410 | 19 April 2001 |www.nature.com. 53 Sossina Haile, Fuel Cells : Powering Progress in the 21st Century 54 B. Krakow, E. Weaver, L. Ecklund-Mitchell, E. Stefanakos, G. Moore, and M.Smith, in The 22nd Annual International Coal Conf erence, University of Pittsburgh, pg. 191/1-191/8 (2005).

PAGE 112

102 55 T. J. Kirk and J. Winnick, J. Electrochem. Soc., 140, 3494 (1993). 56 Zha, Shaowu Electrochem. Solid -State Lett., Volume 8, Issue 8, pg. A406-A408 (2005). 57 D.A. Agievskii, M.V. Landau, a nd A. A. Slinkin, kine t. Katal., 29,792 (1989). 58 B.J. Bladergroen, F. Meyer, S. Pasupathi, V. Linkov, International Journal of Hydrogen Energy 33 3031-3035 (2008). 59 C. Berger (Ed.). Handbook of Fuel Cell Technology. Prentice-Hall, Englewood Cliffs, NJ, (1968). 60 L.I. Antropov, Theoretical Electroc hemistry (Moscow: Mir Publishers), pg. 418 (1981). 61 http://www.gamry.com. 62 PANalytical X’pert Highsco re (PW 3209) Software, Version 1.0F, (2004). 63 C. R. I. Chisholm, Y. H. Jang, S. M. Haile, and W. A. Goddard III, Phys. Rev. B 72, 134103 (2005). 64 F.D. Morrison, D.C. Sinclair A.R. West, J. Appl. Phys. 86 (11) 6355/ 6365 (1999). 65 J.R. Macdonald, Impendance Spectroscopy, Wiley, pg. 191/ 205 (1987). 66 T. Ito, H. Kamimura, J. Phys. Soc. Jpn, Vol. 67, No. 6, pg. 1999-2007, June, (1998). 67 E. Ortiz, R. A. Vargas and B. E. Mellander, J. Phys.: Condens. Matter 18, 9561– 9573, (2006). 68 J. Mbah, B. Krakow, E. Stef anakos, and J. Wolan, J. Electrochem. Soc. 155 (11) (2008) E166-E170.

PAGE 113

103 69 A. Saito and H.C. Fo ley, AIChE Journal, 37, (1991) 429. 70 G. Horvath and K. Kawazoe, Met hod for the Calculation of Effective Pore Size Distribution in Molecular Siev e Carbon, J. Chem. Eng. Japan, Vol 16, No. 5, pp (1983) 470-475. 71 A.V. Belushkin, M. A. Adams, S. Hull, A. L. Kolesnikov, L. A. Shuvalov, Phys. B Condens Matter 213-214, 1034-6 (1995). 72 A. I. Baranov, L. A. Shuvalov and N. M. Shchagina JETP Lett. 36 459–62 (1982). 73 T. Uda, D. A. Boysen, S. M. Haile, Solid State Ionics 176, pg.127–133 (2005). 74 M. H. Chahbani and D. Tondeur, F undamentals of Adsorption, Vol. 6, F. Meunier (Ed.). Elsevier, New York, (1998). 75 G. Dagan, Flow and Transport in Porous Formations, Springer-Verlag, Berlin, (1989). 76 E. R. Gilliland, R.F. Baddour, G. P. Perkinson, and K. J. Sladek, Ind. Eng. Chem. Fundam., 13, 95, 100 (1974). 77 R.C. Makkus, A. H. H. Janssen, F. A. de Bruijn, and R. K. A. M. Mallant. Use of stainless steel for cost competitive bipolar plates in the spfc. Journal of Power Sources, 86 (1): 274-282, (2000).

PAGE 114

104 APPENDICES

PAGE 115

105 Apendix A Cell Preparation and H2S Splitting A.1 CsHSO4 Synthesis Solid Acid (CsHSO4) can be synthesized by the met hod of the two reactions below: a. Cs2SO4 + H2SO4 2CsHSO4 (A.1.a) b. Cs2CO3 + 2H2SO4 2CsHSO4 + CO2 + H2O (A.1.b) Sulfuric acid has a molecular weight of 98.08, specific gravity of 1.84gm/ml and a concentration of 95.7 %. The molecular weig hts of cesium sulfate and cesium carbonate are 361.88 and 325.82 respectively. Using pure H2SO4 with 95.8 % purity gives, 0.957 x1.84gm/ml = 1.76088 gm/ml (specific gravity of H2SO4). From equation (A.1a), the amount of Cs2SO4 required for every 1 mil of H2SO4 = (1.76272 x 361.88)/98.08 = 6.497 gm/ml Similarly, from equation (A.1b) the amount of CsCO3 required for every mil of H2SO4 = (1.76272 x 325.82)/2 x 98.08 = 2.9278 gm/ml Preparation of CsHSO4 using CsSO4 a. A measured quantity of CsSO4 (6.504 g) was dissolved in 4 ml of DI water at 34 oC. b. A (1.0 ml) H2SO4 was added to the solution, which raised the temperature to 64 oC (exothermic reaction).

PAGE 116

106 Appendix A (Continued) c. At the completion of reaction, the produc t is left to cool down in the hood until it starts to precipitate at 24 oC. At this point, the precipitation process was accelerated by the addition of 5 ml of acetone which was previously cooled with liquid nitrogen. d. The crystalline precipitate which was at 8 oC is separated using a vacuum pump. At this temperature the filtrate is clear wit hout any colloidal suspension. Also the liquid nitrogen helped to provide chilling effect to the result ing mixture on precipitation. e. Then the resulting crystalline powder is vacuum dried usi ng Schlenk filteration manifold at 80 oC overnight. f. The crystalline powder was then ball m illed in a high energy planetary mill for 3 h. g. The next stage involved baking of the pow dered sample in an Autosorb apparatus in outgassing station in nitr ogen atmosphere at 150 oC to produce a pure homogeneous crystal sample. h. Using an evacuated die set, the crystallin e sample was pressed to make pellets of 0.5’’ and 2’’ diameter depending on the application desired. A.2 Pellets Preparation Pellet sizes of CsHSO4 were prepared using pellet diam eters 0.5’’ and 2’’ and thickness range (1-0.2 mm). The calculations outlined below show the amount of CSHSO4 required for each diameter using 1 mm thick pellet for illustration. a. 0.5’’ diameter pellet and 1 mm thickness: Volume =Area x 1 x 252 = ( D2 x 252 x 1)/4

PAGE 117

107 Appendix A (Continued) = ( x 0.52 x 252 x 0.001)/4 = 0.1227 ml Density = mass/volume Mass = Density x volume Density of CsHSO4 = 3.352 gm/ml Mass of CsHSO4 required = 3.352 x 0.1227 = 0.4113 gm For 0.5 mm thickness we need Volume =Area x 0.5 x 252 = ( D2 x 252 x 0.0005)/4 = ( x 0.52 x 252 x 0.0005)/4 = 0.061328 ml Density = mass/volume Mass = density x volume Density of CsHSO4 = 3.352 gm/ml Mass of CsHSO4 required = 3.352 x 0.0613 = 0.2056 gm For 0.2 mm thickness Volume =Area x 0.2 x 252 = ( D2 x 252 x 0.0002)/4

PAGE 118

108 Appendix A (Continued) = ( x 0.52 x 252 x 0.0002)/4 = 0.02453 ml Density = mass/volume Mass = density x volume Density of CsHSO4 = 3.352 gm/ml Mass of CsHSO4 required = 3.352 x 0.02453 = 0.0822 gm b. 2’’ diameter pellet Volume = ( x 22 x 252 x 0.001)/4 =1.9635 ml Mass of CsHSO4 required = 1.9635 x 3.352 = 6.5816 gm A.3 Computation of H2S Flowrate Required for both the 0.5’’ and 2’’ Cells Hydrogen sulfide will split according to the following equation: H2S 2H+ + S2 + 2e(A.3.1) The current density target obtained from literature, for this process is 0.030 A/cm2 Current, A = C/s For 0.5’’ diameter cell: For an active area of pe llet of 0.281 diameter, A = D2/4

PAGE 119

109 Appendix A (Continued) = (0.281)2/4 = 0.062in2 = 0.4 cm2 Faraday requirement: 96400 Coulombs = 1 Faraday 1 Faraday = 1 mole of electron From equation (A.3.1), (96400 C/mole electron) x (mole 2 e-/moles H2S) = 192800 C/moles H2S The amount of current required = current density x active area of electrolyte (Amp/cm2) x cm2= C/s=Amp Amp = (0.030A/cm2) x 0.4cm2 = 12 mA Therefore, flowrate of H2S per electron is given by, H2S flowrate = 12 x10-3 C/s x [1/ (96400 C/moles H2S)] = 1.24 x 10-7 moles/s 1 mole of gas contains 22400 cm3 (1.24 x 10-7 moles/sec) x 22400 cm3/mole = 0.16 cm3/min Time Required to Electrolyze 1 mole of H2S Amperes x time = C Time (sec) = 192970 C/moles H 2 S =1.61 x 107 s/mole H2S 12 x 10-3 C/s And 1.61 x 107 s/mole or 1 mole of sulfur in 186 days

PAGE 120

110 Appendix A (Continued) 1 mole S = 32g of sulfur Density of sulfur = 2.07g/cm3 Atomic volume = 15.46 cm3 1 mole 15.46 cm3 sulfur 186 days 1 ml sulfur 186/15.46 = 12 days 1 day 1/12 = 0.08 ml/day of sulfur For 2’’ diameter Cell: Active area of pellet = 1.4760 in 1.5’’ Area in cm2 = x (381000)2/4 = 11.395 cm2 Current density, j = 0.030A/cm2 Amp = 0.030A/cm2 x 11.4 cm2 = 0.342 C/s H2S Flowrate = (0.342 C/s) x [1/ (192970 C/moles H2S)] = 1.772 x 10-6 moles/s = 2.38 ml/min of sulfur Time Required to electrolyze 1 mole of H2S Time (s) = (192970 C/moles H2S)/ (0.342 C/s) = 81012.907 s/mole H2S = 1 day 1 mole sulfur 32g sulfur Density of sulfur = 2.07 g/cm3 Atomic volume of sulfur = 15.46 cm3

PAGE 121

111 Appendix A (Continued) 1 mole sulfur 15.46 cm3 sulfur 1 day 1 ml sulfur = 1/15.46= 0.06 ~1.5 h A.4 Gasket and O ring for the ’’ Cell Both the O-rings and the gasket were from (C TG). These are all made from aflas material which can withstand the effect of H2S on them. The inner O-ring has the dimension: OD (0.441’’) and the ID (0.301’’) while the outer O-ring has the dimension: OD (2.270’’) and the ID (1.850’’). O-ring thickness = (OD-ID)/2 Gasket dimension is calculated from the formula below. Gasket thickness = Pellet thickness + 2 [O -ring (inner) thickness] – 2 (grove depth) =1 mm + 2(0.07’’) 2(0.05’’) = 0.04’’ + 0.14 – 0.1’’ = 0.08’’ Gasket size = OD (1.8’’) x ID (0.64’’) x Thickness (0.0625’’) ID of gasket is a little bigger than ID of pellet to allow for pe llet fitting, and the OD is a little smaller than OD of cell to be able to accommodate the outer O-ring. The thickness is made to be a little sma ller than the calculated value. A.5 System Pressure A PX481A pressure transducer from Omeg a engineering with 9-30 VDC/output: 1-5 VDC and range: 0-30 psig and accuracy of 0.5 % FS maximum was used with the

PAGE 122

112 Appendix A (Continued) data acquisition (DAQ) instrument. This wa s linked to LabView DAQ and the conversion for the transducer is shown in Figure A.5.1: V = (1-5) V P = (0-30) psig P = 30/4 (V-1) P = 7.5 (V-1) Figure A.5.1 A LabView block diagra m for monitoring system pressure.

PAGE 123

113 Appendix A (Continued) 0 2 4 6 8 10 12 9:36:00 AM12:00:00 PM2:24:00 PM4:48:00 PM7:12:00 PM9:36:00 PM TimeCurrent, mA0.9 V 1.2V 1.4V 1.6V 1.8V2.0V Figure A.5.2 A sample LabView representation of an electrolytic system. Comprising of anode based RuO2/CoS2 and cathode based Pt black at different voltage input. The sulfur collection vessel is filled with mineral oil to the same height where the liquid sulfur exits the cell to avoid oil being sucked into the cell as a result of this vacuum created by a higher oil level than the sulfur exit point in th e cell before the start of the experiment. Patm oil = Patm cell (before experiment) Pabs cell = Patm cell + gauge pressure (~1-2 psig ) (experiment start up) The flow of sulfur will then push oil in the lin e back since pressure in the cell is higher than the oil pressure. It is assumed that the rise in oil level due to sulfur production will not be enough to off set this balan ce and reverse the flow of sulfur.

PAGE 124

114 Appendix A (Continued) A.6 Material Balance The electrolysis of H2S produces hydrogen at the cathode and sulfur at the anode. A steady voltage of 0.9V is a llowed to flow through hydrogen sulfide for 5 hours, splitting it at a sustainable current of 5.31 mA. Mass of sulfur deposited at the anode. Method: Sulfur analysis a. The half equation for the electrolysis. H2S 2H+ + 2e+ (1/8) S8 (A.6.1) 2 moles of electrons (2 Faradays) are required to deposit 1/8 mole of sulfur. b. Number of Faradays that have pass ed through the hydrogen su lfide in 5.31 hours. Q = I x t (A.6.2) 5.31 hours contains (5.31 x 60 x 60) seconds = 18,000 seconds. Q = (5.31 x 18,000)/1000 = 95.58 Coulombs. 1 Faraday = 96,400 coulombs. 95.58 coulombs = 95.58 96,400 Faradays = 0.001 Faradays. c. From the proportion in (A.6.1), 2 Farada ys (2 moles of elec tron) are required to produce (1/8) mole of sulfur a nd 2 moles of hydrogen ions. This is sufficient to produce

PAGE 125

115 Appendix A (Continued) 0.001 moles (0.001g) of hydrogen ions and 6.25E-5 moles (0.016g) of S8. Hydrogen analysis: The half e quation for the electrolysis 2H+ + 2eH2 (A.6.3) The 2 moles of hydrogen ions produced in Eqn. A.6.1 are converted to 1 mole of hydrogen gas in Eqn. A.6.3. The 0.001 moles (0.001g) of hydrogen i ons produced at the anode by 0.001 Faraday are converted to 0. 0005 moles (0.0005g) of hydrogen gas at the cathode. For the gas measurement, the gas issuing from the cathode compartment was collected and measured by displacement of water in a 50 ml burette at 25 oC and 102.4 kPa. The observed volume was 10 ml. Gas collected by displacement of water is saturated with water vapor whose vapor pressure at 25 oC is 3.17 kPa. The partial pressure of hydrogen in the burette was then (102.4 -3.17) 99.2kPa. By the ideal gas law, 0.0004 moles of gas has a volume of 10 ml at a partial pre ssure of 99.2 kPa and temperature of 25 oC. Thus the measured volume of hydrogen gas is abou t 80 % of the theoretical limit for the electrolysis as calculate d in the previous section. When the water level is equal in the test tube and the trough, the pressure inside the test tube wi ll be equal to the atmospheric pressure. Ideal gas law is used to determin e the number of hydrogen moles in the test tube. The water displaced by hydrogen has vapor pressure that will distort the equation if not accounted for because of the Dalton's Law of partial pressure: the pressure in the test

PAGE 126

116 Appendix A (Continued) tube is from both the hydrogen and the wate r. To find just the hydrogen, the vapor pressure of the water has to be subtracted. Vapor pressure of water is published in most chemistry books, and varies by the temperature of the water. Calculations using Dalton's law: In our lab, the atmospheric pressure is 102.4 kPa (768 torr or 1 atm) The temperature of our water is 25 C. We used a 10 mL test tube to collect th e hydrogen. The pressure of the hydrogen will be calculated, and then mo les of hydrogen are obtained using the ideal gas law. Step 1: We need to know the vapor pressu re of the water. A common table lists the pressure at 25 C as 23.76 torr. A torr is 1 mm of mercury at sta ndard temperature. In kilopascals, that would be 3.17 (1 mm mercur y = 7.5 kPa). We should also convert the 10 mL to 0.010 L and 25 C to 298 C. Step 2: We can use Dalton's Law to fi nd the hydrogen pressu re. It would be: PTotal = PWater + PHydrogen 102.4 kPa = 3.17 kPa + PHydrogen So the pressure of Hydrogen would be: 99.23 kPa or 99.2 kPa. Step 3: We use the Ideal Gas Law to get th e moles. Recall that the Ideal Gas Law is: PV = nRT where P is pressure, V is volume, n is moles, R is the Ideal Gas Constant (0.010 Latm/mol-K or 8.31 L-kPa/mol-K) and T is temperature. Therefore, our equation would be:

PAGE 127

117 Appendix A (Continued) 99.2 kPa x 0.010 L = n x 8.31 L-kPa/mol-K x 298 K This can be re-arranged so: n = 99.2 kPa x 0.010 L / 8.31 L-kPa/mol-K / 298 K n = .00046 mol or 4.6 x 10-4 mol Hydrogen 1 coulomb is the amount of electrical charge in 6.2415061018 electrons or other elementary charged particles. The charge of one electron equal to -1.602210-19 C 1 C = 6.2415061018 = Amp x s For a theoretical 12 m A, the rate of prot on (and electron) genera tion is approximately [(6.2415061018 x 12 A)/1000] = 5.47 x 1016 electrons/s for 100 % conversion of H2S at 48.50 Psi or 334.39 kPa,150 oC, and a flow rate of 0.25 cm3/min. The sustainable current under these conditions is 5.31 m A corresponding to 3.12 x 1016 electrons/s for a cell that uses RuO2, Pt.black, p-Dichlorobenzene and CsHSO4 as the anode catalyst. This value agrees well with the previous calculation. Gi ven the severe limitations on precision of the above measurements, it is seen that the values for the H2S conversion and current generated are consistent with each other. A.7 Product Analysis Sulfur produced in the anode compartment was characterized usi ng XRD and SEM. XRD analysis indicates the molten sulfur produced fr eezes to monoclinic crys tals on cooling.

PAGE 128

118 Appendix A (Continued) Figure A.7.1 is an XRD pattern observed on the sulfur produced and that from pure sulfur as received. Sulfur from elec trolysis product matches that of pure sulfur as shown in spectra (a) and spectra (b) respectively. 102030405060Intensity, arb. units sulfur from electrolysis product Pure sulfur (2 2 2) (2 0 6) (2 0 6) (0 2 6) (0 2 6) Cu k 2 (degree), degree ( 2 2 2 ) Parafilm Tape a b Figure A.7.1 X-ray diffraction comparison of sulfur. (a) pure sulfur ; (b) sulfur from electrolysis product. Figure A.7.2 (a) and (b) show the SEM-EDS of the electrolyzed pellet. Figure A.7.2 (a) shows SEM image of layers of yellowish sulf ur deposit on the surface of the electrolyzed pellet (the area is depicted by the white background). Figure A.7.2 (b) is the EDS showing sulfur deposited on top of the electrolyzed pellet.

PAGE 129

119 Appendix A (Continued) Figure A.7.2 SEM and EDS images of sulfur Product resulting from the electrochemical splitting of H2S: (a) SEM image showing sulfur de posit on the pellet, (b) EDS spectra showing sulfur formed from H2S electrolysis within an ar ea of (216) microns Sulfur formed in this region is 65 wt %. DSC analysis indicates the molten sulfur produced is the monoclinic crystal when compared to the commercial supplied sulfur. Figure A.7.3 compares the melting point of sulfur from the electrolysis product to that of pur e sulfur as received. The two spectra shown both matched each other with the melting point of pure sulfur at 119.87 oC and that from electrolysis product at 119.21 oC but in the electrolysis other allotropes of sulfur compound are formed such as orthorhom bic sulfur which has a lower melting point than the monoclinic sulfur. It is not clear at this time whether the low rate of production led to formation of sulfur molecules smaller than S8.

PAGE 130

120 Appendix A (Continued) -6 -5 -4 -3 -2 -1 0 1 2 020406080100120140160 Temperature, o CHeat Flow, W/g Sulfur from electrolysis product Pure sulfur, 99.9 %119.87 o C 119.21 o C 110 o C Figure A.7.3 Differential s canning calorimetry comparison of sulfur melting points. Hydrogen gas produced was collected and meas ured by displacement of aqueous solution and identified by gas chromatography. The chromatogram indicated the gas contained about 15 % air. This probabl y infiltrated the sample even with a “gas tight syringe” while it was being transferred to gas ch romatograph. Figure A.7.4 shows the gas chromatograph of hydrogen sample from the el ectrolysis product. The hydrogen peak produced by the sample containing about 65 % hydrogen was small because the thermal conductivity of hydrogen is almost the same as the helium carrier gas. It was too small for accurate integration to get a direct measure of the hydrogen concentration.

PAGE 131

121 Appendix A (Continued) Figure A.7.4 Gas chromatograph analys is of hydrogen produced. This was by displacement of water from the electrolytical splitting of H2S gas using helium as a carrier gas. A.8 Graphical Comparison of the Claus and the Electrolytic Processes Figure A.8.1 Comparison of the Claus and the electrolytic processes. (a) Claus process: the reactions are highly exothermic and this heat is wasted; (b) one step electrolytic process: energy can be r ecovered as electricity.

PAGE 132

122 Appendix A (Continued) A.9 LCR Calibration Figure A.9.1 LCR calibration with known resistor s and capacitor.

PAGE 133

123 Appendix A (Continued) A.10 Thermochemistry Data Table A.10.1 Thermochemistry data (evaluated at T = 298 K). Material H(k J/ mol) G(k J/ mol) S (J/K mol) H2S (g) -20.1 -33.9 43.1 H2 0 -38.96 130.68 S 0 31.8 O2 0 -61.12 205 Reactions: ) ( 2 2 ) ( 22 / 1g gH S S H H = HH2S =-20.1 k J/mol S = [0.5(31.8) + 130.68] – 43.1 = 103.48 J/Kmol G = H T S G = -20.1(298) (0.10348) = 50.937 k J/mol = -12.17 kcal/mole G = -n F E E0 = g0/n F mol C molreac mole mol J E400 96 2 / 509370 = 0.26 V

PAGE 134

ABOUT THE AUTHOR Jonathan Chinwendu Mbah comes from the Ibo tr ibe in the eastern regi on of Nigeria. He received his B.S.Ch.E and M.S.Ch.E from University of Lagos, Nigeria, and North Carolina Agricultural and Technical State Univ ersity, North Carolina respectively. In the Spring of 2006, he enrolled into doctoral pr ogram in chemical engineering at the University of South Florida where he received his Ph.D. in the Fall of 2008.