USF Libraries
USF Digital Collections

Application and modeling of TiO2-supported gold nanoparticles for CO preferential oxidation in excess hydrogen

MISSING IMAGE

Material Information

Title:
Application and modeling of TiO2-supported gold nanoparticles for CO preferential oxidation in excess hydrogen
Physical Description:
Book
Language:
English
Creator:
Grayson, Benjamin Alan
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Supported Au/TiO2
Kinetic modeling
Catalysis
World Gold Council
FTIR
Dissertations, Academic -- Chemical Engineering -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: This work begins with a brief overview of heterogeneous, characterization techniques, and current hypotheses about gold mechanisms. This is followed by the initial characterization of custom two-phase-method gold nanoparticles provided by the Interfacial Phenomena and Polymeric Materials research group at USF, the anatase TiO2 support and reference Au/TiO2 catalyst provided by the World Gold Council. In order to verify the ability of the two-phase-method GNP catalyst provided to oxidize CO in excess hydrogen, it was necessary to develop an effluent testing protocol. The first experiments involved 24 hour runs to observe catalyst deactivation. Concerns over cycling effects observed in the absorbance integral calculations lead to the introduction of a reference gas. Corrections were made to the carbon monoxide absorbance integral calculations which allowed the direct comparison of results.These corrections included baseline adjustments for each species and N2 purging to eliminate background CO2 and H2O contamination. After these improvements, the two phase method GNP catalyst CO oxidation ability was investigated. Unfortunately, the supplied two phase method gold catalyst has been unresponsive for CO oxidation applications. One hypothesis for the problems is that the surfactants used to keep the gold nanoparticles from aggregating are preventing carbon monoxide transport to the surface of the particle. Another theory is that the gold may not be adhering to the surface of the TiO2 creating a cohesive metal/support interaction. The kinetics of CO preferential oxidation (PROX) catalyzed by the World Gold Council's nano-Au/TiO2 was studied to evaluate elementary and nonelementary empirical rate expressions. Information is readily available for CO fractional conversion for this catalyst below 0 degrees C.However, a comprehensive CO PROX kinetic model in which three reactions (CO oxidation, H2 oxidation and the water gas shift reaction) occur simultaneously is lacking. The reaction was carried out in a vertical packed bed micro-reactor testing unit; temperature was varied between 25 and 125 degrees C, and a range of feed rates were tested. In-situ Fourier transform infrared spectroscopy (FTIR) reaction data was analyzed; pre-exponential and activation energies are calculated for each kinetic model. Empirical rate expressions based on power law models were used to fit the experimental data. The reversible water gas shift reaction was found to play an important role when fitting the experimental data precisely and explained the selectivity decrease at higher reaction temperatures. The empirical kinetic model presented will be useful to simulate PROX operation parameters for many applications.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Benjamin Alan Grayson.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 118 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 - 001921017
oclc - 190844513
usfldc doi - E14-SFE0002131
usfldc handle - e14.2131
System ID:
SFS0026449:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Application and Modeling of TiO2-Supported Gold Nanoparticle s for the Preferential Oxidation of Carbon Monoxide in Excess Hydrogen by Benjamin Alan Grayson A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemical Engineering College of Engineering University of South Florida Major Professor: John T. Wolan, Ph.D. Stephen E. Saddow, Ph.D. Elias Stefanakos, Ph.D., P.E. Aydin K. Sunol, Ph.D. George S. Nolas, Ph.D. Date of Approval: June 4, 2007 Keywords: Supported Au/TiO2, Kinetic Modeling, Catalysi s, World Gold Council, FTIR Copyright 2007, Benjamin Alan Grayson

PAGE 2

ACKNOWLEDGEMENTS To all of my advisors, colleagues, friends and family, thank you for your support. I could not have done this without you.

PAGE 3

i TABLE OF CONTENTS LIST OF TABLES.............................................................................................................iv LIST OF FIGURES............................................................................................................v ABSTRACT.......................................................................................................................ix 1 INTRODUCTION AND THEORY...............................................................................1 1.1 Introduction to Catalysis.....................................................................................1 1.2 Catalytic Applications of Gold...........................................................................2 1.3 Synthesis Techniques of Au Nanoparticles and the Necessity of a Reference............................................................................................................6 1.4 Reactor Setup......................................................................................................7 1.5 Characterization Techniques...............................................................................8 1.5.1 X-ray Photoelectr on Spectroscopy (XPS).......................................................8 1.5.2 X-ray Diffraction (XRD)................................................................................9 1.5.3 Fourier Transform Infrared Spectroscopy (FTIR)........................................10 1.5.3.1 Aabspec In-situ FTIR Micro-reactor....................................................11 1.5.4 Gas Chromatography (GC)...........................................................................13 1.5.5 Thermogravimetric Analysis.........................................................................15 1.6 Discussion of Possible Reaction M echanisms and the Reactive Species of Gold..............................................................................................................15 1.6.1 Introduction...................................................................................................15 1.6.2 Theories for the Reaction M echanism of CO Oxidation on Supported Gold Nanoparticles......................................................................16 1.6.3 Catalytic Species of Gold..............................................................................19 2 INITIAL CHARACTERIZATION..............................................................................21 2.1 Introduction.......................................................................................................21 2.2 Two-Phase-Method Nano-Au/TiO2 Catalyst....................................................21 2.2.1 Fabrication and Pre-Treatment of the Two-Phase-Method GNPs................21 2.2.2 DLS Analysis................................................................................................24 2.2.3 X-ray Photoelectr on Spectroscopy (XPS).....................................................24 2.2.4 Scanning Electron Micrographs of the Two-Phase-Method GNPs on a Silicon Wafer.............................................................................................26 2.3 Titanium Dioxide (TiO2)...................................................................................27 2.3.1 XRD and SEM of the Experimental Support................................................27 2.3.2 Aabspec FTIR Micro-reactor........................................................................29 2.3.3 XPS Analysis................................................................................................31 2.4 World Gold Council Reference Catalyst A (nano-Au/TiO2)........................33 2.4.1 XPS Analysis................................................................................................34 2.5 Summary of Results..........................................................................................35

PAGE 4

ii 3 DEVELOPMENT OF TWO-PHASE-M ETHOD AU NANOPARTICLE TESTING PROCEDURES AND PROTOCOLS FOR THE OXIDATION OF CO.........................36 3.1 Introduction.......................................................................................................36 3.2 Twenty-four Hour Analysis of Two-Phase-Method GNP Catalyst FTIR Absorbance Integrals...............................................................................36 3.3 Data Analysis with Reference Methane in the Influent Stream........................43 3.4 Reevaluation of the FTIR Spectra to Account for Shifting Baselines..............44 3.5 Redesign of the Supported Gold Catalyst.........................................................47 3.6 Summary...........................................................................................................48 4 EMPIRICAL MODELS OF CARB ON MONOXIDE OXIDAT ION VIA WORLD GOLD COUNCIL Au/TiO2 IN EXCESS HYDROGEN.................................................51 4.1 Introduction.......................................................................................................51 4.2 Experimental.....................................................................................................52 4.3 Calculation of Thermodynamic Properties.......................................................52 4.4 Empirical Models..............................................................................................54 4.5 Elementary CO Oxidation Single Reaction Model...........................................54 4.5.1 Elementary CO Oxidation Single Reaction Model Results..........................55 4.6 Comprehensive PROX Models.........................................................................57 4.7 Comprehensive Elementary Reaction Model...................................................58 4.8 Comprehensive Non-Elementary Reaction Model...........................................59 4.9 Comparison of Comprehensive Models to Experimental Results....................59 4.9.1 Comprehensive Elementary Reaction Model Results...................................60 4.9.2 Comprehensive Non-Elementary Reaction Model Results..........................61 4.10 Linearly Independent Model Equations for CO Oxidation...............................64 4.10.1 Linearly Independent Elementary Reaction Model Results.........................64 4.10.2 Linearly Independent Non-elementary Reaction Model Results..................66 4.11 Verification of FTIR Efflue nt Concentrations via Gas Chromatography...............................................................................................68 4.12 Conclusions and Future Work..........................................................................74 5 FUTURE WORK..........................................................................................................77 5.1 Introduction.......................................................................................................77 5.2 Two-Phase-Method Gold..................................................................................77 5.3 Modification of the WGC Effluent Model........................................................79 5.4 Experimental Applications................................................................................80 REFERENCES.................................................................................................................81 APPENDICES..................................................................................................................85 Appendix A Matlab Code for FTIR Modeling............................................................86 A.1 Elementary Model without WGS.......................................................................86 A.2 Comprehensive Elementary Model Fit Routine................................................88 A.3 Comprehensive Non-Elementary Model Fit Routine.......................................90 A.4 General Comprehensive Model........................................................................94 A.5 Independent Elementary Model Fit Routine.....................................................96 A.6 Independent Non-Elementary Model Fit Routine.............................................98

PAGE 5

iii A.7 General Independent Model............................................................................102 A.8 Error Calculations for WGC Data...................................................................104 ABOUT THE AUTHOR.End Page

PAGE 6

iv LIST OF TABLES Table 1.1 FTIR gas cell specifications.........................................................................10 Table 1.2 GC operating parameters..............................................................................14 Table 4.1 Standard heats of reaction and Gibbs free energies of reaction...................53 Table 4.2 Gibbs free energies of reaction at each temperature.....................................54 Table 4.3 Calculated equilibrium constants at each temperature.................................54 Table 4.4 Experimental influent conditions to the tubular reactor...............................55 Table 4.5 Experimental reaction effl uent fractional conversions at each temperature and flow rate.............................................................................56 Table 4.6 Calculated error with each fractional conversion value...............................57 Table 4.7 Calculated pre-exponenti als and activation energies for the elementary model which includes CO oxidation, H2 oxidation and the WGS reaction...............................................................................................60 Table 4.8 Empirical mole balances for the elementary model which includes CO oxidation, H2 oxidation and the WGS reaction......................................60 Table 4.9 Calculated pre-exponential s and activation energies for the nonelementary model which includes CO oxidation, H2 oxidation and the WGS reaction...............................................................................................62 Table 4.10 Empirical mole balances for the non-elementary model which includes CO oxidation, H2 oxidation and the WGS reaction.......................62 Table 4.11 Calculated pre-exponentials a nd activation energies for the linearly independent elementary model reaction.......................................................65 Table 4.12 Empirical mole balances fo r the linearly indepe ndent elementary model reaction..............................................................................................65 Table 4.13 Calculated pre-exponentials a nd activation energies for the linearly independent non-elementary model.............................................................67 Table 4.14 Empirical mole balances fo r the linearly independent non-elementary model............................................................................................................67

PAGE 7

v LIST OF FIGURES Figure 1.1 Catalytic and noncatalytic potential energies versus reaction coordinates for an el ementary reaction...........................................................3 Figure 1.2 Carbon monoxide and Oxygen diffusing to the catalyst surface....................4 Figure 1.3 Both molecules adsorb to the surface.............................................................4 Figure 1.4 Oxygen radicals diffuse quickly along the surface.........................................4 Figure 1.5 Carbon monoxide and th e oxygen radical react to form carbon dioxide............................................................................................................4 Figure 1.6 After the reaction, the carbon dioxide molecule desorbs...............................4 Figure 1.7 The catalyst surface returns to its original active state enabling the process continue.............................................................................................4 Figure 1.8 Typical formulation of CO selective oxidation catalysts...............................6 Figure 1.9 FTIR/Mic roreactor setup................................................................................8 Figure 1.10 Microreactor setup..........................................................................................8 Figure 1.11 Aabspec microreactor for in si tu FTIR analysis of solid samples................12 Figure 1.12 In transmission mode, the cell l ooks like a standard unit modified for high temperature high pre ssure IR analysis..................................................12 Figure 1.13 For specular reflectance IR dete rminations, the angle of incidence is nearly normal................................................................................................12 Figure 1.14 For large angle re flectance IR, the incidence angle is near grazing.............12 Figure 1.15 Possible reaction mechanisms for the oxidation of carbon monoxide in a Au/TiO2 system.....................................................................................17 Figure 1.16 Representation of a possibl e mechanism for the oxidation of carbon monoxide using gold on an oxide support.33................................................19 Figure 1.17 Relationship between par ticle size and melting point of gold nanoparticles.................................................................................................20 Figure 2.1 Tetraoctylammonium bromide molecule (TOAB) [CH3(CH2)7]4NBr binds to the gold particles to prevent aggregation........................................22 Figure 2.2 Two-phase-method nano-Au catalyst preparation procedure.......................23

PAGE 8

vi Figure 2.3 4 mg GNPs/16 mL toluen e on a silicon wafer (single crystal 100 plane)............................................................................................................25 Figure 2.4 4 mg GNPs/16 mL toluene evaporated onto a Silicon Wafer after overnight 100oC evaporation........................................................................25 Figure 2.5 High resolution XPS Au 4f7/2 spectra of the two-phase-method GNP/silicon as-received after 100oC evaporation........................................25 Figure 2.6 High resolution XPS Au 4f7/2 spectra of the two-phase-method GNP/silicon after 3 hours at 500oC in air.....................................................26 Figure 2.7 SEM (400x) Micr ograph of 4 mg GNPs/16 mL toluene air dried onto a silicon wafer...............................................................................................27 Figure 2.8 SEM (5000x) micrograph of 4 mg GNPs/16 mL Toluene on a silicon wafer after exposure to 150oC......................................................................27 Figure 2.9 SEM (150000x) micrograph of 4 mg GNPs/16 mL toluene evaporated onto a silicon wafer after exposure to 150oC.............................27 Figure 2.10 XRD spectra of as-receiv ed titanium dioxide support.................................28 Figure 2.11 SEM micrograph (50000x) of the as-received titania support.....................28 Figure 2.12 XRD spectra of (a) titania after 500oC calcination (b) titania asreceived.........................................................................................................29 Figure 2.13 Representative FTIR respons e curve for the reaction effluent.....................30 Figure 2.14 FTIR absorbance spectra of the 30 mg 1:1 TiO2/KBr pellet........................31 Figure 2.15 FTIR transmission mode CO char acteristic peak absorbance integral of the 30 mg 1:1 TiO2/KBr pellet at 25-125oC.............................................31 Figure 2.16 Anatase TiO2 powder XPS spectra afte r air calcination at 500oC for 3 hours.............................................................................................................32 Figure 2.17 Anatase TiO2 powder XPS spectra afte r air calcination at 500oC for 3 hours and hydrogen reduction at 400oC.......................................................32 Figure 2.18 Anatase TiO2 powder XPS spectra afte r air calcination at 500oC for 3 hours and hydrogen reduction at 400oC and slight oxidation in air at 200oC............................................................................................................33 Figure 2.19 SEM micrographs of the two-phase-method gold reveal similar pictures to the As-received TiO2 used as a support......................................34 Figure 2.20 SEM micrographs of the WGC show much small particle sizes for the TiO2 support used in their samples.........................................................34 Figure 2.21 WGC nano-Au/TiO2 XPS spectrum.............................................................34 Figure 3.1 FTIR absorbance integral s of carbon monoxide, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 5 sccm air..........................................................................38

PAGE 9

vii Figure 3.2 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 5 sccm air..........................................................38 Figure 3.3 FTIR Absorbance integral s of carbon monoxide, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air........................................................................40 Figure 3.4 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air........................................................40 Figure 3.5 FTIR absorbance integral s of carbon monoxide, carbon dioxide and water at 300oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air........................................................................41 Figure 3.6 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 300oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air.....................................................41 Figure 3.7 FTIR absorbance integral s of carbon monoxide, carbon dioxide and water at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air........................................................................42 Figure 3.8 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air........................................................42 Figure 3.9 FTIR absorbance integr als of carbon monoxide carbon dioxide, water, and methane at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane.......................43 Figure 3.10 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide, water and methane at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane.........................................................................................................44 Figure 3.11 FTIR absorbance integral s of carbon monoxide carbon dioxide, water and methane at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane.......................45 Figure 3.12 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide, water and methane at 425oC after baseline subtraction. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane......................................................................46 Figure 3.13 Carbon monoxide absorbance integrals for all data collected reevaluated using baseline subtracti on to allow for direct comparison........47 Figure 3.14 FTIR results of redesigned two-phase-method GNP catalyst before and after exposing the sample to a 205oC calcination step..........................49

PAGE 10

viii Figure 3.15 TGA analysis of or iginal two-phase-method GNP/TiO2 catalyst formula.........................................................................................................50 Figure 4.1 Graph of W X f FoCO) ( ln vs. T-1 for the simple elementary model.............56 Figure 4.2 Kinetic model flow diagram.........................................................................58 Figure 4.3 Comparison of model results for the comprehensive elementary model............................................................................................................61 Figure 4.4 Comparison of model results for the comprehensive non-elementary model............................................................................................................63 Figure 4.5 Comparison of model results for the linearly independent elementary model............................................................................................................66 Figure 4.6 Comparison of model resu lts for the linearl y independent nonelementary model.........................................................................................68 Figure 4.7 Representative image of a GC response spectrum of the effluent gases which include hydrogen, oxygen, nitrogen, carbon monoxide, carbon dioxide, and water.............................................................................69 Figure 4.8 Integral of gas chromatogr aphy hydrogen spectrum data range versus influent flow rate at te mperatures ranging from 25oC-125oC......................70 Figure 4.9 Integral of gas chromatogra phy nitrogen spectrum data range versus influent flow rate at te mperatures ranging from 25oC-125oC......................71 Figure 4.10 Integral of gas chromatogr aphy oxygen spectrum data range versus influent flow rate at te mperatures ranging from 25oC-125oC......................73 Figure 4.11 Integral of gas chromat ography carbon monoxide spectrum data range versus influent flow rate at temperatures ranging from 25oC125oC............................................................................................................73 Figure 4.12 Conversion calculations of gas chromatography carbon monoxide spectrum data range versus influent flow rate at temperatures ranging from 25oC-125oC..........................................................................................74 Figure 5.1 Modified micr oreactor setup which incl udes bubbler and relative humidity gauge for moisture content calculations........................................80

PAGE 11

ix APPLICATION AND MODELING OF TiO2-SUPPORTED GOLD NANOPARTICLES FOR THE PREFERENTIAL OXIDATION OF CARBON MONOXIDE IN EXCESS HYDROGEN Benjamin Alan Grayson ABSTRACT This work begins with a brief overview of heterogeneous, characterization techniques, and current hypotheses about gold mechanisms. This is followed by the initial characterization of custom two-phase -method gold nanoparticles provided by the Interfacial Phenomena and Polymeric Material s research group at USF, the anatase TiO2 support and reference Au/TiO2 catalyst provided by the World Gold Council. In order to verify the ability of the tw o-phase-method GNP catalyst provided to oxidize CO in excess hydrogen, it wa s necessary to develop an effluent testing protocol. The first experiments involved 24 hour runs to observe catalyst de activation. Concerns over cycling effects observed in the absorbance integr al calculations lead to the introduction of a reference gas. Corrections were made to the carbon monoxide absorbance integral calculations which allowed the direct comparis on of results. These corrections included baseline adjustments for each species and N2 purging to eliminate background CO2 and H2O contamination. After these improvements, the two phase method GNP catalyst CO oxidation ability was investigated. Unfort unately, the supplied two phase method gold catalyst has been unresponsive for CO oxidati on applications. One hypothesis for the problems is that the surfactants used to keep the gold nanopa rticles from aggregating are preventing carbon monoxide transport to the su rface of the particle. Another theory is that the gold may not be adhering to the surface of the TiO2 creating a cohesive metal/support interaction.

PAGE 12

x The kinetics of CO preferential oxidati on (PROX) catalyzed by the World Gold Councils nano-Au/TiO2 was studied to evaluate elementary and nonelementary empirical rate expressions. Information is readily avai lable for CO fractional conversion for this catalyst below 0oC. However, a comprehensive CO PROX kinetic model in which three reactions (CO oxidation, H2 oxidation and the water gas shift reaction) occur simultaneously is lacking. The reaction was ca rried out in a vertical packed bed microreactor testing unit; temperature was varied between 25 and 125oC, and a range of feed rates were tested. In-situ F ourier transform infrared spectr oscopy (FTIR) reaction data was analyzed; pre-exponential and activation energies are calculated for each kinetic model. Empirical rate expres sions based on power law models were used to fit the experimental data. The reversible water gas shift reaction was found to play an important role when fitting the experimental data precis ely and explained the se lectivity decrease at higher reaction temperatures. Th e empirical kinetic model pres ented will be useful to simulate PROX operation parameters for many applications.

PAGE 13

1 1 INTRODUCTION AND THEORY 1.1 Introduction to Catalysis In a chemical reaction, there are five basic parameters one can control: temperature, pressure, concentration of sp ecies, contact time and pattern.1 While early attempts at improving reaction rates and conversions were successful by relying on high temperature and high pressure processes, these conditions are energy intensive, corrosive, and result in undesirable side products.1 These issue ushered in th e development of catalytic systems. To date, approximately 90% of all industrial processes are catalyzed in some fashion. The extensive use of selective catal ysts along with improvements in fluid flow characteristics have led to lower operati ng temperatures, lower operating pressures, higher efficiencies and cost reduction. The ma jority of these selective catalysts falling into two broad categories: heterogeneous and homogeneous.1 The difference being that heterogeneous reactions occur between at least two different phases while homogeneous reactions all occur in the same phase. This work will primarily focus on the heterogeneous reactions occurring between a gas phase influent and a solid phase nanocatalyst. One of the most basic definitions of a catalys t is a material that enhances the rate and selectivity of a chemical reaction and in the process is cyclically regenerated.1 This is a valid definition; however, it fails to describe the subtleties of the catalytic process. Figure 1.1 attempts to graphically explain some of the details of an elementary heterogeneous reaction. One of the processes by which a catalyst increases the reaction rate is through the lowering of the activati on energy (or energy barr ier) to the formation of products. The typical mechanism for hete rogeneous catalysis starts with incoming reacting compounds adsorbing onto th e surface of a solid catalyst (Figure 1.2 and Figure 1.3).1 This process must be energetically profitable for both species or it will not occur.

PAGE 14

2 Figure 1.1 demonstrates that the chemisorbed species do lower their energy states when adsorbing onto the surface. These adsorbed species are then rapidly and selectively transformed into ad sorbed products (Figure 1.4 and Figure 1.5).1 Although one can see that the energy barrier for this catalyzed re action is substantially lowered, one subtlety expressed in Figure 1.1 is that the apex of the catalyzed energy barrier is before the apex of the uncatalyzed reaction. If the reaction path can be assumed to be an approximation of the time needed for the reaction to occur, this difference in reaction time can result in a catalyzed reaction occurring se veral orders of magnitude quicker. The products then desorb from the surface returning the solid catalyst back to its original state to recycle the process (Figure 1.6 and Figure 1.7).1 These interactions provide a chemical shortcut in which reactants are converted to products mo re rapidly and at much milder conditions than if no surface interactions occurred.1 Additionally, catalysts can perform many other tasks. Just a few are listed below.2 Help initiate reactions Stabilize the intermediates of a reaction Hold reactants in close proximity or in the corre ct orientation Block side reactions Stretch bonds or make bonds easier to break Donate or accept electrons Act as an efficient means for energy transfer 1.2 Catalytic Applications of Gold The first indication that gold might be a usef ul catalyst came through the work of Haruta et al. when he discovered in the late 1980s that gold becomes considerably stickier when spread in tiny dots on certain metal and oxide compounds.3 Since then, his research has led to a renewed interest in gold applica tions previously unexpl ored. Many groups now report numerous applications for gold nanopartic les with the optimum gold particle sizes ranging from 2-50 nm depending on the application.4-6 Several examples for which gold nanoparticles demonstrate catalyt ic activity are listed below.7

PAGE 15

3 Oxidation of CO and hydrocarbons Water gas shift (WGS) Reduction of NO with propene, CO or H2 Reactions with halogenated compounds Removal of CO from hydrogen streams Hydrochlorination of ethyne Selective oxidation, e.g. epoxidation of olefins Selective hydrogenation Hydrogenation of CO and CO2 This work will be focusing on the extreme abili ty of gold nanoparticles to oxidize carbon monoxide at low temperatures and determin e any potential applications for proton exchange membrane fuel cells (PEMFCs). Figure 1.1 Catalytic and noncatalytic potential energies versus r eaction coordinates for an elementary reaction. Without catalyst; Ea is higher thus rate is lower Reaction Path Reactants Complex Products Hr Final State With catalyst; Ea is lower thus rate is faster Chemisorption Different Reaction Path Energy of Reacting Particles Ea Ea

PAGE 16

4 Figure 1.2 Carbon monoxide and Oxygen diffusing to the catalyst surface. Figure 1.3 Both molecules adsorb to the surface. Figure 1.4 Oxygen radicals diffuse quickly along the surface. Figure 1.5 Carbon monoxide and the oxygen radical react to form carbon dioxide. Figure 1.6 After the reaction, the carbon dioxide molecule desorbs. Figure 1.7 The catalyst surface returns to its original active state enabling the process continue. In many fuel cell systems, the influent hydrogen is produced through hydrocarbon reformation, and as a result of the reforming process, a small amount of carbon monoxide is typically present in the product stream.8 This is a major concer n in the direct hydrogen

PAGE 17

5 PEMFC industry due to the prevalence of plat inum based catalysts and the potential of poisoning due to CO. The Department of Energy is focusing on improvements regarding the tolerance of PEM membrane assemblies but reducing the CO infl uent concentration would be preferred.9 The Department of Energy 2011 t echnical target fo r CO tolerance of stationary PEM fuel stack systems (5 -250 kW) operating on reformate to be 500 ppm at steady state and 1000 ppm transient; the cu rrent status (c. 2005) is 50 ppm at steady state and 100 ppm transient.9 Several solutions to the problem include the development of alternative catalyst systems more tolera nt to CO poisoning, increasing the operating temperature to increase the catalyst kinetics and introducing catalyst additives to oxidize the CO.8 A breakdown of the typical formulatio n components of CO selective oxidation catalysts are shown in Figure 1.8. Increasing the operatin g temperature is beneficial because fuel cells which operate at higher temperatures, i.e. above 200oC, are better equipped to deal with the CO issue, because CO does not readily adsorb to the anode at these elevated temperatures; however, hydrog en based PEM fuel cells normally operate at 80oC though DOEs target for H2-PEMFCs by 2010 is to operate at 120oC.7 The temperature limitations of PEM fuels are due to the necessity of NafionTM to remain hydrated (~40% relative humidity) in or der to maintain proton conductivity.10 Within the NafionTM film exist nanometer size pores lined with sulfonic acid groups ( H SO3, and in the presence of water, these groups form hydronium ions that can only transport in the liquid phase.10 It is found that NafionTM exhibits proton conductivity similar to liquid electrolytes but only wh en hydrated with water.10 When the PEM operating temperature exceeds 100oC, the NafionTM begins to dehydrate and the proton conductivity drops.

PAGE 18

6 Figure 1.8 Typical formulation of CO selective oxidation catalysts.8 1.3 Synthesis Techniques of Au Nanoparticle s and the Necessity of a Reference Various methods have been employed to s ynthesize supported gold nanoparticle (GNP) catalysts including deposition-precipitati on, impregnation, sol-gel techniques, coprecipitation, incipient wetness, metal organic-chemical vapor deposition, and dipcoating.11 Although there are literall y dozens of methods of prep aration, essentially all of these techniques can be reduced to three major procedures.12 In the first procedure, a gold metal precursor is mixed with a support to give a sy stem that, after controlled calcination, produces gold/support catalysts.12 The metal nanoparticles size is controlled by the calcination temperature and experime ntal results have shown that lower temperatures seem to favor the formation of smaller particles.12 Secondly, a gold precursor is deposited or grafted onto th e surface of the support, which has been preformed from the gas or from the liquid phase and subsequently ther mally decomposed to give the gold/support catalysts.12 Lastly, the gold nanopar ticles are pre-produced in a given solution where they are stabilized by so luble polymers or by kinetic conditions, and subsequently left to be adsorbed by the surface of the desired support particles.12 Preparation procedures devel oped by laboratories vary greatly and the effects of aging, stirring, washing, order in wh ich reactants are added, temp erature, concentration of reactants, and calcining conditions all appear to be important parameters to monitor.4

PAGE 19

7 With this in mind, the World Gold C ouncil (WGC) in 2002 commissioned four gold reference catalysts enabling researchers to benchmark their own catalyst systems against a common reference. These consist of three proprietary reference catalysts of nano-gold on metal oxide supports made by Sd Chemie, Japan under the supervision of Haruta et al., with characterization at AIST, Japan.7,13 A fourth reference catalyst of gold on carbon is produced by Rossi et al. of the University of Milan.7,13 Descriptions of the four catalysts are listed below. The choice of supports for each catalyst application is very important and will be discussed in Section 1.5.2. World Gold Council Reference Gold Catalysts7 A 1.5 wt% Au/TiO2 (P-25) as powder by deposition precipitation B 0.3 wt% Au/Fe2O3/Al2O3 as beads by deposition precipitation C 5 wt% Au/Fe2O3 as powder by co-precipitation D 10 wt% Au/C (Cabot XC72R) as powder using gold sol. 1.4 Reactor Setup The FTIR/reactor setup is shown in Figure 1.9 and Figure 1.10. All reactions were performed in a quartz tubular reactor, 24 inches in length, 4 mm ID which constricts to 2 mm ID at the midpoint, inside of a Lindberg /Blue split tubular furnace. One hundred milligrams of nano-Au/TiO2 was packed loos ely between high temperature quartz wool to prevent the powder from escaping. Two ba ll valves placed before and after the reactor tube allows for the influent stream to bypa ss the catalyst bed comple tely to comprise a reference unreacted sample. The input stre am is modeled after a typical single stack PEMFC inlet feed stream with CO contamin ation. Three mass flow controllers (MFCs) lead to a mixing tube to assure a consistent inlet gas concentration. The setup allows for injection points before and after the test bed in order to analyze the inlet and effluent of the process in the gas chromatograph. Also, a bypass is incorporated to send just the feed stream through to the effluent. After exiting the reactor system, the effluent is fed directly to the inline Fourier transform in frared spectrometer gas cell for real time compositional analysis.

PAGE 20

8 Figure 1.9 FTIR/Micror eactor setup. Figure 1.10 Microreactor setup. 1.5 Characterization Techniques The main characterization techniques for th ese experiments will be X-ray photoelectron spectroscopy, X-ray diffracti on, Fourier transform infrar ed spectroscopy, and gas chromatography. A brief overview of each tec hnique and its application to this work follows. 1.5.1 X-ray Photoelectron Spectroscopy (XPS) XPS spectra collected will attempt to determine the oxidation states of the Au nanoparticles and identify the catalysts elem ental surface atomic concentrations to examine changes during calcination and after reaction. All XPS anal yses were conducted in a Perkin-Elmer PHI 560 ESCA/SAM System (base pressure 5 x 10-10 Torr) equipped with a PHI 04-500 Dual Anode X-ray sour ce and a 25-270AR cylindrical mirror analyzer. All spectra were measur ed in normal emission using Mg K radiation. The spectrometer was calibrated to yield the standard Au 4f7/2 and Cu 2p3/2 lines at 84.00 eV and 932.66 eV, respectively. XPS will be used to determine elemental and chemical state information of near-surface species. X-ray Photoelectron Spectroscopy is a relativ ely non-destructive technique that exposes the sample to ultra high vacuum pre ssures with base pressures of 1 x 10-10 Torr, increases the temperature by approximately 10-20oC, and exposes the sample to soft X-rays between approximately 1000-1500 eV. These conditions are considered nondestructive Effluent PortH2, CO, Air H2, H2O, CO, CO2, N 2 FTI R Reacto r MFCs Bypass

PAGE 21

9 for most materials and systems with few excep tions. Due to the ultrahigh vacuum (UHV) requirements, samples that are run in the X PS system must have relatively low vapor pressures; hence, they must be solids. T ypical examples of solids analyzed are metals, glasses, semiconductors, and low vapor pressure ceramics. Ideally all samples should be free of fingerprints, oils, and other surfac e contamination. Furthermore, XPS is a beneficial technique wh en destructive techniques must be avoided 14. The sensitivity of the technique is about 0.1~1.0 atomic percen t. The approximate sampling depth is material dependent but averages about 3 na nometers, and the energy resolution is about 0.3 to 4 eV 15. One drawback of working with XPS is the time required to run a sample. A sample placed in the vacuum chamber requi res several hours in order to pump down before analysis. Qualitative analysis can be performed in 5 to 10 minutes. Quantitative analysis requires 1 hour to several ho urs depending on the information desired 16. 1.5.2 X-ray Diffraction (XRD) The XRD will be used to determine the crystalline state of the TiO2 dioxide before and after calcination. Thes e results will verify any changes in crystal structure due to typical calcination temperatures. X-ray diffraction is a versatile non-destructive analytical te chnique for identification and quantitative determination of the various cr ystalline forms of compounds present in powdered and solid samples.17 XRD starts with a gene rated X-ray focused onto a sample. When the X-ray beam hits an atom, the electrons around that atom oscillate with the same frequency as the incoming beam. X-ra ys are then generated; however, most of the waves destructively combine and result in no resultant energy leaving the sample. In a relatively few specific directions within a crystalline solid, the X-rays will combine constructively and produce a diffracted X-ray out of the sample. Us ing Braggs Law and the relationships for interplanar spacing, one can predict the diffraction angles in which X-rays should constructively combine.17,18 Identification is usually achieved by comparing the X-ray diffraction pattern, or diffractogram, obtai ned from an unknown sample with a database containing refere nce patterns. X-ray diffraction (XRD) was

PAGE 22

10 performed using a Panalytical XPert Diffr actometer, tension 45 kV, current 40 mA, and Cu K radiation. 1.5.3 Fourier Transform Infrared Spectroscopy (FTIR) All experiments were performed using a Bio-Rad Excalibur Series FTS 3000 system. From the FTIR analysis of the effluent da ta, individual component conversions can be calculated at each temperature and gas flow rate The gas cell specifications are given in Table 1.1. The detector used for all experime nts is mercury cadmium telluride (MCT). Table 1.1 FTIR gas cell specifications. Model Number: 2.4-PA (Ultra-mini Cell) Pathlength: 2.4 meters Body material: Borosilicate glass Mirror coating: Protected gold Body dimensions: Length, 11.5 cm.; I.D., 3.3 cm. Volume: 0.1 liter Transfer mirrors: Two plane mirrors on mounts for finger grip adjustment. Valves: Two stainless steel plug valves Fourier Transform Infrared spectroscopy examines the effects of electromagnetic radiation on solids, liquids, and gases.17,18 When the infrared ra diation passes through the samples, characteristic frequencies are absorbed that cause molecular vibrations to occur. The transmitted light which passes to the detector allows for the determination of composition through the calculation of transmittance and absorbance spectra.17,18 Transmittance is defined as the ratio of in tensity of transmitted radiation through the sample versus the intensity of incident ra diation, i.e. the background transmittance with no sample.17,18 Absorbance is defined as the logarithm, base 10, of the inverse transmittance and is proportional to molar concentration.17,18 Absorption band intensities, widths, and areas are all dependent on both temperature and pressure.19 Recommendations in the Digi lab manual suggest running ga s samples at atmospheric pressure to minimize pressure broadening and allowed to reach thermal equilibrium before collecting spectral data.19 Another problem with gas cells is selective adsorption of species onto the cell walls which can alter quantitative results.19 Flushing the system several times prior to running is necessary to clean the lines and erase cell memory.19

PAGE 23

11 The main component of the FTIR spectromete r is an interferometer which splits and recombines a beam of light such that th e recombined beam produces a wavelengthdependent interference patte rn or an interferogram.20 When radiation with more than a single wavelength hits the interf erometer, the output signal is the sum of all the cosine waves of the entering radiation.20 Even though the interferogram contains the basic information on frequencies and intensities characteristic of a spectrum in the time domain, the output is in a form that is not easily understood.20 Typically, a spectrum is converted from the time dependent interferog ram to the frequency domain by computing the cosine Fourier tr ansform of the signal.20 Calculations of absorbance use the Beer-Lam bert law which is shown in Eqns. 1.1-2, which states that there is a logarithmic dependence of the light transmitted through a substance and its concentration.21 Transmittance T = I/Io (1.1) Absorbance A = log(1/T) = log(Io/I) = ecL (1.2)Io = Intensity of incident radiation, I = Intensity of transmitted radiation, e = molar extinction coefficient, c = concentration (mole/ liter), L = sample pathlength (cm) 1.5.3.1 Aabspec In-situ FTIR Micro-reactor A specialized cell mounted within the FTIR a llows for additional IR analysis of solid samples and allows one to discriminate be tween surface gold sites according to their oxidation state.22 Since IR absorbance area is linearly proportional to the concentration of the species, one can calculate the integral of the absorbance and correlate that integral to the concentration of the species.18 The Aabspec cell structure ( Figure 1.11) consists of three part s: the body containing access ports for reactive gases, a sample probe containing a heating system, and a matched end plate designed sp ecifically for each probe.23 This allows for the three

PAGE 24

12 modes of operation, as shown in Figure 1.12, Figure 1.13, and Figure 1.14: transmission mode, specular reflectance mode, and large angle infrared reflectance (LARI) mode.23 Figure 1.11 Aabspec microreactor for in situ FTIR analysis of solid samples.23 Figure 1.12 In transmission mode, the cell looks like a standard unit modified for high temperature high pressure IR analysis. Figure 1.13 For specular reflectance IR determinations, the angle of incidence is nearly normal. Figure 1.14 For large angle reflectance IR, the incidence angle is near grazing. IRbeam Sample Sample Sample

PAGE 25

13 In transmission mode, the sample is mounted perpendicular to the optical axis of the infrared beam and the beam passes directly through the sample mount.23 The matched endplate is unnecessary and is used onl y to close the end of the cell body.23 However, some applications allow for liquid cooli ng through the endplate for low temperature studies.23 In specular reflectance mode, the IR beam strikes the sample nearly perpendicular to the surface of the sample w ith a typical angle of incidence near 10 degrees.23 A gold mirror mounted on the end plate reflects the entering beam so that it continues along the optical axis. In LARI mode the IR beam strikes at a glancing angle of approximately 20 degrees.23 Once again, gold mirrors on the end plates reflect the IR beam to continue its optical axis path.23 The LARI mode samples can be mounted vertically or horizontally.23 Horizontal mounting allows for samples which may melt during analysis at elevated temper atures, liquid samples, or powders.23 1.5.4 Gas Chromatography (GC) All experiments were performed with an Agilent Technologi es 6890N network GC system with Chemstations software. Sta ndard GC operating conditions are shown in Table 1.2. The GC will be used to verify th e effluent concentrations from the oxidation reaction along with the FTIR.24 Gas chromatography uses a thin capillary column to separate injected samples into its prim ary components through adsorption onto the column walls or onto the packing materials in the column.24 The velocity of the species progressing through the column is a function of th e strength of adsorption, which in turn depends on the type of molecule and on the colu mn materials which in turn forces each of the species to exit at different times.24 The outlet gases on then detected and the signal strength and time are noted to identify molecules and concentration.24

PAGE 26

14 Table 1.2 GC operating parameters. GC Operating Parameters Oven: 30oC for 10 minutes Front Inlet: Mode: Split Initial Temperature 200oC Pressure: 20 psi Split Ratio: 14.568:1 Split Flow: 67.1 mL/min Total Flow: 74.2 mL/min Gas Type: Helium Column: Capillary Column Model Number: Varian CP7534 Plot Fused Molsieve Column Nominal Length: 30 m Nominal Diameter: 320 m Nominal Film Thickness: 10 m Mode: Constant Pressure Pressure: 20 psi Nominal Initial Flow: 4.6 mL/min Average Velocity: 62 cm/sec Detector (TCD) Temperature: 250oC Reference Flow: 20 mL/min Mode: Constant column + makeup flow Combined flow: 7.0 mL/min Makeup flow: On Makeup Gas Type: Helium Filament: On Negative Polarity: Off Temperature must be closely monitored as both molecular adsorpti on and the rate of progression along the column are temperature dependent.24 Although lower temperatures produce the greatest level of se paration, the progression time to the outlet of the column is greatly increased and can be prohibitiv e. Temperature programs which modify the oven temperature try to compromise between various species to be analyzed between lower temperatures to separate and higher te mperatures that promote shorter analysis times.24 Ideally, hydrogen would be carrier gas used for most GC experiments because it provides the best component separation.24 However, helium is the most common choice due it being nonflammable, shares similar prope rties to hydrogen, and is compatible with most detectors.24 The two most common types of detectors for GC are thermal

PAGE 27

15 conductivity detectors (TCD) and fl ame ionization detectors (FID).24 Thermal conductivity detectors are useful in that they can detect any component other than the carrier gas injected.24 Flame ionization detectors are primary used to detect hydrocarbons.24 1.5.5 Thermogravimetric Analysis Thermogravimetric Analysis (TGA) is a ther mal analysis technique used to measure changes in the weight (mass) of a sample as a function of temperature and/or time. Examples of uses are the determination of polymer degradation temperatures, absorbed moisture content, and inorganic filler in composite materials.25 Analysis of the samples starts by placing a tared sample into a microbalance assembly located in a high temperature furnace. After determining the initial weight at room temperature, the sample is subsequently heated while the weight is constantly monito red. A weight profile is then generated for amount or percent weight loss at any given temperature.25 The TGA will be used to determine the temperatures at which each component of the Brust gold catalyst evaporates. This information is us eful when determining at which temperature the catalyst becomes stable and if any adju stments to the calcination temperature are necessary. 1.6 Discussion of Possible Reaction Mechanis ms and the Reactive Species of Gold 1.6.1 Introduction There are many factors to consider when one begins to work with catalytic materials. First of all, the process of picking a cataly st to promote a reaction is not well understood, therefore extensive trial and error is part of optimization.26 In addition, reproducing the chemical composition of a catalyst does not guarantee the reproduction of the catalytic activity.26 Also, the crystalline struct ure of the catalyst are just as if not more important than the composition26 As discussed in Chapter 1.1, cat alysts reduce the energy barriers to form products; however, they cannot shif t the equilibrium of a reaction which are governed by thermodynamics alone.26 All of these factors ar e necessary to note when beginning catalysis work.

PAGE 28

16 Accordingly, the difficulty of finding a mech anism which fully characterizes the catalyst system is that the favored mechanism must fit the data is such a way that all other mechanism possibilitie s can be rejected.26 For each rate controlling mechanism, there are usually three to seven parameters which must be fitted, verified, and precisely reproduced.26 Thus, it is difficult if not impossible to justify precisely a correct mechanism for a particular process.26 This means that the inhe rently that most proposed analytical mechanisms can only claim to be approximate calculations of each system under specific conditions.26 With the fundamental unde rstanding that a proposed mechanism can only be approximated, one shoul d try to find the simplest equation which adequately describes their system.26 When analyzing the PR OX reaction catalyzed by GNPs/TiO2, the three most important reactions are the oxidation of carbon monoxide, the oxidation of hydrogen, and the water gas shif t reaction (Eqns. 1.3, 1.4, 1.5) and attempts to model these reactions are discussed in Chapter 4. 2 22 1 CO O CO Ho 298 = -283 kJ/mol (1.3) O H O H2 2 22 1 Ho 298 = -242 kJ/mol (1.4)2 2 2H CO CO O H Ho 298 = -41.1 kJ/mol (1.5) 1.6.2 Theories for the Reaction Mechanism of CO Oxidation on Supported Gold Nanoparticles Some of the reasons put forth fo r the ability of nano-gold to be an active catalyst are the interactions between the metal and support, the coordinative unsaturation of the surface atoms, and quantum size effects.27 So, in order to fully co mprehend the entire system, it is necessary to begin with an examination of particle size effects and the interactions of molecules with the Au na noparticles and the support.28 Models for these GNP/support inte ractions can be categorized into three basic types: perimeter, surface, and intermediate as presented in Figure 1.15.29 Surface models

PAGE 29

17 propose that the oxygen molecules dissociate di rectly on the surface of the gold particles with the supporting material having little to no effect on th e reaction mechanism.29 Intermediate and perimeter models imply that the only active areas of the system are the metal-support interfaces.29 The main difference in these two models is whether or not the CO molecule interacts with a O2 molecule or a O radical.29 As shown in both of these models, it is important to note that the in terface density on the surface would drastically increase with decreasing particle size thus increasing the active sites for CO oxidation.30 The increase in catalytic activity with decreasing particle size is consistent with reported observations, and intermediate and perimeter models are the more likely mechanism for reducible metal oxide supported GNP catalysts.30 Figure 1.15 Possible reaction mechanisms for the oxidation of carbon monoxide in a Au/TiO2 system.29,31 Along with the possibility of providing a oxyge n source to the Au-CO oxidation reaction, the support introduces a defect structure which aids in th e formation and stabilization of small gold particles and promotes active edge sites.32 Mavrikakis et al. found that carbon monoxide (CO) was able to bind to the off-axis faces of gold where the presence of steps may be the most significant contri buting factor to the attraction.27 Density functional theory calculations indicate that CO ma y only chemisorb on the low-coordinated Au atoms such as those found in the steps and ki nks on the edges of the atom, and not on the regular (111) terraces.27 Therefore, smaller particle si zes may greatly increase the steps and kinks density within the catalyst.27

PAGE 30

18 The choices for nano-Au support materials ca n be divided into two broad categories: irreducible and reducible supports. A few catal yst supports that have been investigated include oxides of cobalt, magnesium, iron, alum inum, ceria, nickel, titanium, and tin with the most commonly used supports for gold catal ysts being titania, zirconia, silica, and alumina.11 It has been found that there is a sign ificant division between the activities of irreducible oxide supports, such as Al2O3, and reducible transition metals oxides, such as TiO2. When the reducible metals are coupled with gold, the catalytic activity of the system can be up to one order of magnitude higher than the same Au on irreducible supports.30 In addition, reducible supports are more tolerant to increased gold particle size while still maintaining a reasonable level of activity.30 A study comparing Au/TiO2, Au/Fe2O3, Au/Co3O4 found that Au/TiO2 was the most active catalyst system and will be the primary support for a ll of these experiments.7 One of the most current hypotheses for the gol d reaction mechanism (Eqns. 1.6-8) for CO oxidation is shown in Figure 1.16 proposed by Thompson et al.33 This is a representation of the early stages of the oxidation of carbon monoxide at the periph ery of an active gold particle. At the left, a carbon monoxide molecu le is chemisorbed onto a low coordination number gold atom, and an hydroxyl ion has moved from the support to an AuIII ion, creating an anion vacancy.33 At the right they have re acted to form a carboxylate group, and an oxygen molecule occupies the anion vacancy as O2 -.33 This then oxidizes the carboxylate group by abstracting a hydrogen atom, forming carbon dioxide, and the resulting hydroperoxide ion HO2 then oxidizes a further carboxylate species forming another carbon dioxide and restoring two hydr oxyl ions to the support surface. This completes the catalytic cycle. There was no attempt by Thompson to suggest the charges carried by the reacting species. Also, there is no experimental evidence as to whether the oxygen derives from the gas or the support.33 CO + OH COOH (1.6) COOH + O2 CO2 + HO2 (1.7) COOH + HO2 CO2 + 2OH (1.8)

PAGE 31

19 Figure 1.16 Representation of a possible mechanism for the oxidation of carbon monoxide us ing gold on an oxide support.33 1.6.3 Catalytic Species of Gold Another point of contention among researchers is whether the active species of nano-gold is metallic gold, an oxidized gold species, or a sub-oxide. Some researchers believe that oxidized gold species, stabilized by an interaction with the support, are more active than Au0 alone.34 Other researchers conclude that the active species of Au for CO oxidation is Au1+ with CO incorporating dual roles as bot h a reactant and a reducing agent for gold converting Au1+ to Au0.31,35 Haruta believes that it is unlikely that an oxidized gold species is a candidate due to th e fact that some of the most active catalysts are created by calcination in air at 573 K where the gold hydroxide a nd organic precursors are transformed into metallic species.31 Goodman et al. suggested that the primary source of the catalytic activity of gold was non-metallic nanoparticle clusters.31,36 It has been proposed that high temperature reduction/lo w temperature calcination results in higher catalytic activities for Au/TiO2 systems and implies that fully reduced metallic gold along with a small percentage of gold oxide to be catalytically the most active.37 However, an important fact to consider wh en approaching nanometer diameters of gold particles is that the melting point can d ecrease significantly compared to the bulk properties as shown in Figure 1.17. The particles will melt and agglomerate thus increasing the particle diameters and decreasing the catalyst activity.12

PAGE 32

20 Figure 1.17 Relationship between particle size and melting point of gold nanoparticles.12

PAGE 33

21 2 INITIAL CHARACTERIZATION 2.1 Introduction This chapter describes the initial charac terization of custom two-phase-method gold nano-particles fabricated by the Interfacial Phenomena and Polymeric Materials research group, the TiO2 support through all experiments, and the reference Au/TiO2 sample purchased from the World Gold Council. Af ter describing both the fabrication procedure of the two-phase-method gol d nanoparticles and the TiO2 supported Au catalyst, oxidation state is verified via XPS and th e morphology of each is examined with SEM micrographs. The crystallography of the su pport is confirmed via XRD. Initial examination of the adsorption of CO to the TiO2 support is presented. Lastly, the preliminary characterization of the GNPs pur chased from the World Gold Council via XPS and SEM which contrasts the similaritie s and differences between the purchased reference and the custom two-phase-method gold nanoparticles fabricated. Kinetic information will be described in detail for both catalysts in Chapters 3 and 4. 2.2 Two-Phase-Method Nano-Au/TiO2 Catalyst 2.2.1 Fabrication and Pre-Treatment of the Two-Phase-Method GNPs The experimental gold nanoparticles (GNPs) were synthesized in an organic medium following the method reported by Brust et al.38,39 The fabrication procedure of the twophase-method GNP solution, as described by Dayling Chaparro38-40, begins with a 38 mM aqueous solution of hydrogen tetrachloroaurate (HAuCl4) added slowly to a 13 mM solution of tetraoctylammonium bromide (TOAB) in toluene. The TOAB is used to stabilize the nanoparticles in so lution. The solution was then stirred for 1.5 hr to allow the metal salt to be transferred to the organic phase. To reduce the Au3+ ions to metallic gold (Au0), a freshly prepared aqueous soluti on of 0.13 M sodium borohydride (NaBH4) was added drop wise to the mixture while stir ring. The solution turned ruby red after the

PAGE 34

22 gold reduction which indicates the formation of nanoparticle s of gold. After stirring overnight, the organic layer was separated from the aqueous layer and washed with water, 0.1 M HCl, and 0.1 M NaOH respectiv ely. The rinsing procedure was repeated three times and a final wash with DI water wa s performed. The organic layer containing the GNPs/TOAB was allowed to dry w ith anhydrous sodium sulfate (NaSO4) for 1 hr. Finally, after recovery of the GNPs in toluen e, the solution was transferred to an amber bottle and refrigerated. As illustrated in Figure 2.1, after the formation of the GNPs, TOAB completely surrounds the Au particle clusters maintaining their diameters. While this is designed to prevent agglomeration, it may also be a hindran ce to gas transport to the particle surface if not fully removed prior to reaction. Figure 2.1 Tetraoctylammoni um bromide molecule (TOAB) [CH3(CH2)7]4NBr binds to the gold particles to prevent aggregation.41 The Au-nanoparticles, suspended in solution (4 mg GNPs/16 mL Toluene), were then combined with titanium dioxide (TiO2) via incipient wetness techniques. To begin the procedure, the entire Au solution was blended with anatase TiO2 ((4 mg GNPs/16 mL Toluene):100 mg TiO2) to form a 4% Au mixture. Th en, the solution was continuously stirred at room temperature for 1 hour prior to heating at 150oC. The Au-solution evaporated leaving a pinkish-red powder. Th e entire sample was then loaded into the tubular reactor as explained in Chapter 1.4.3. Beginning at room temperature, the sample was ramped 30oC/min to 500oC in air (100 sccm, 1 atm) and held at 500oC for 3 hours, as

PAGE 35

23 illustrated in Figure 2.2. Then it was allowed to cool back to room temperature in ambient atmosphere. The sample was then reduced in a hydrogen atmosphere (100 sccm, 1 atm) at 400oC for 3 hours, and then allowed to cool The final pre-treatment step was to slightly oxidize the sample for 2 hour s in air (100 sccm, 1 atm) at 200oC. This preparation procedure is based on pre-tr eatment techniques used by Haruta and Choudhary.31,37 The calcination step is used to remove any excess volatile compounds and promote Au/TiO2 contact; however, a side effect of the calcination process is gold agglomeration.31 Studies have shown that th e high temperature reduction (400oC in H2) induces some catalytic activity towards CO oxidation in the TiO2 and reduces the gold to its metallic state.37 The low temperature oxidation promotes the formation of a small amount of oxide and sub-oxide species in the gold.37 Several proposed mechanisms suggest that one of the active species of gold is the sub-oxide formed during this step as discussed in Chapter 1.5.2.33 Figure 2.2 Two-phase-method nano-Au catalyst preparation procedure.

PAGE 36

24 2.2.2 DLS Analysis Dynamic light scattering (DLS) verified th e gold nanoparticle diameters in solution before combination with the titania support. These experiments contain a range of particles with diameters of 1 0.25 nm. 2.2.3 X-ray Photoelectron Spectroscopy (XPS) In order to determine the effect of heat trea tments on the oxidation state of GNPs, a small amount (0.6 mL) of the 4 mg GNPs/16 mL tolu ene solution was applied to the surface of a silicon wafer (Figure 2.3) and evaporated at 150oC overnight in a furnace (Figure 2.4). A silicon wafer was chosen because the XPS binding energy reference peaks for silicon dioxide (Si 2p, 103.4 eV) would not in terfere with thos e of gold (Au 4f7/2, 84.00 eV). Additionally, the silicon substrate is able to withstand the calcination temperature (500oC) without a phase change or melting. After an initial XPS reference spectrum of the dried GNPs (Figure 2.5), the Au/silicon sample was then placed in a furnace for 3 hours at 500oC in air at ambient pressure. A sec ond XPS measurement was performed on the calcined sample to determine any changes in the oxidation state of the gold (Figure 2.6). Exposure to 500oC in air at atmospheric pressures resulted in a reduction in the full width at half maximum (FWHM) of the Au0 4f7/2 peak (84.00 eV) and Au0 4f5/2 peak (87.71 eV), which suggests that the GNPs become more single crystalline in nature and no oxidation did occur at this temperature.42 This is expected since gold oxide decomposes into its elements above 350oC.43

PAGE 37

25 Figure 2.3 4 mg GNPs/16 mL toluene on a silicon wafer (single crystal 100 plane). Figure 2.4 4 mg GNPs/16 mL toluene evaporated onto a Silicon Wafer after overnight 100oC evaporation. Figure 2.5 High resolution XPS Au 4f7/2 spectra of the two-phase-method GNP/silicon as-received after 100oC evaporation.

PAGE 38

26 Figure 2.6 High resolution XPS Au 4f7/2 spectra of the two-phase-method GNP/silicon after 3 hours at 500oC in air. 2.2.4 Scanning Electron Micrographs of the Two-Phase-Method GNPs on a Silicon Wafer SEM micrographs of the gold particles on the silicon surface after air drying were difficult to image (Figure 2.7). This would suggest that the particles are coated with a non conductive film. After the GNP soluti on was placed in a furnace overnight at 150oC in air at ambient pressure, the particles then appeared in the SEM micrograph (Figure 2.8). The 150000x micrograph, Figure 2.9, shows that the particle sizes are quite large.(~100 nm). This could be a result of the nanoparticle mobility on the surface of the wafer leading to agglomerat ion of the particles.

PAGE 39

27 Figure 2.7 SEM (400x) Micrograph of 4 mg GNPs/16 mL toluene air dried onto a silicon wafer. Figure 2.8 SEM (5000x) micrograph of 4 mg GNPs/16 mL Toluene on a silicon wafer after exposure to 150oC. Figure 2.9 SEM (150000x) micrograph of 4 mg GNPs/16 mL toluene evaporated onto a silicon wafer after exposure to 150oC. 2.3 Titanium Dioxide (TiO2) 2.3.1 XRD and SEM of the Experimental Support Titanium (IV) oxide powder (Sigma-Aldrich 99.8%) was used exclusively for the support material for all experiments. While not explicitly noted, XRD spectra of the support confirmed an anatase crystal structure (Figure 2.10). SEM micrographs revealed TiO2 particle diameters of ~100 nm (Figure 2.11). BET measurements (named for Brunauer, Emmett and Teller) of the titania su pport estimated the specific surface area to

PAGE 40

28 be ~7.16 m2/g with an external area of 4.51 m2/g and a micropore surface area of 2.65 m2/g. The average pore diameter is 84 angstroms. Figure 2.10 XRD spectra of as-received titanium dioxide support. Figure 2.11 SEM micrograph (50000x) of the as-received titania support. To determine the effect of calcination temperatures on the support, the titania was calcined in an oven for 8 hours at 500oC in air at atmospheric pressure. High temperature processes performed on TiO2 have been shown to cause a transformation from the anatase phase to rutile.37 XRD analysis verified that a small percentage of the titania recrystallized into the rutile phase ( Figure 2.12). It has been proposed that this partial reconstruction creates an optim ized interact ion between the GNPs and the titania support, leading to higher dispersion and increased CO oxidation activity.37 Ex p erimental Anatase TiO2Reference

PAGE 41

29 Figure 2.12 XRD spectra of (a) titania after 500oC calcination (b) titania as-received. 2.3.2 Aabspec FTIR Micro-reactor The next characterizatio n technique involved the use of th e Aabspec system described in Chapter 1.4.3.1. Influent flows to th e FTIR reaction chamber were 50 sccm N2, 99 sccm H2, 1 sccm CO, and 20 sccm of air. The reaction chamber was ma intained under steadystate operating conditions for 30 minutes before data collection with temperature ramping from 25-125oC in 25oC increments at atmospheric pressure. A 30 mg 1:1 titanium dioxide/potassium bromide (TiO2/KBr) pellet, 7 mm diameter 0.5-1.0 mm thick, was inserted into a variable temperature IR tr ansmission probe and positioned such that the pellet face intersected the in frared (IR) beam. KBr is a common inert binding agent which allows the formation of rigid pellet s and does not influence the measured IR spectrum wavenumber range (4000-400 cm-1). The FTIR is able to detect the three primary constituents of the gas stream other than hydrogen: CO, CO2, and H2O (Figure 2.13). One of the main limitations of the FT IR technique is its inability to analyze Rutile Phase (a) (b)

PAGE 42

30 diatomic species. This is the primary reas on that the GC was used in conjunction with the FTIR for effluent analysis. Figure 2.13 Representative FTIR respons e curve for the reaction effluent. FTIR transmission absorbance spectra of the 30 mg 1:1 TiO2/KBr pellet are shown in Figure 2.14. While the IR region of CO is located between 2000 and 2250 cm-1, a smaller representative R-branch spectral region (2145-2230 cm-1) for integration was chosen to eliminate interferences from both CO2 and H2O in the CO P-branch (21102175 cm-1).44 After baseline correction of the abso rbance spectral region, the integral of the defined R-branch in each data set was calculated. Since absorbance is linearly proportional to concentration, Figure 2.15 shows that the concentration of adsorbed carbon monoxide decreases with increasin g temperature, indicating a negative temperature dependence on the available reserve of CO on the surface of the support. Previous studies on both anatase and rutile TiO2 found that CO only weakly adsorbs to the support surface so it seems reasonable that the sticking coefficient would decrease with increasing temperature.45-48 Other reported results show that the sticking probability CO CO2 H2O H2O

PAGE 43

31 of CO on metal oxides such as ZnO decrease s with increased temperature which supports the results seen here on a similar oxide support, TiO2.49 All of these studies support the negative correlation reported in Figure 3 for the temperature dependence of CO adsorption on TiO2. Figure 2.14 FTIR absorbance spectra of the 30 mg 1:1 TiO2/KBr pellet. Figure 2.15 FTIR transmission mode CO characteristic peak absorbance integral of the 30 mg 1:1 TiO2/KBr pellet at 25-125oC 2.3.3 XPS Analysis The following three figures are high resolution XPS spectra for the titanium 2p1/2 and titanium 2p3/2 doublet peak after a three hour calcination at 500oC in air (Figure 2.16), a three hour reduction at 400oC in hydrogen (Figure 2.17), and a 2 hou r slight oxidation at 200oC in air (Figure 2.18). Gaussian curves labeled and are representative XPS photoelectron features for TiO2 (458.9 eV and 464.3 eV). Th roughout the entire process, these figures show that the TiO2 did not experience a change in oxidation state. The primary differences between the three figures are the measurements of the width of each Gaussian curve (FWHM full width at half maximum). The FWHM is an indication of the degree of crystallinity of the support with smaller values suggesting a more crystalline state. These result s would indicate a restructuring of the crystal lattice of the TiO2; yet, the oxidation stat e remains constant and does not oxidize to Ti2O3. The restructuring of TiO2 from anatase to rutile is observe d in the XRD analysis discussed in Section 2.3.1.

PAGE 44

32 Figure 2.16 Anatase TiO2 powder XPS spectra afte r air calcination at 500oC for 3 hours. Figure 2.17 Anatase TiO2 powder XPS spectra afte r air calcination at 500oC for 3 hours and hydrogen reduction at 400oC.

PAGE 45

33 Figure 2.18 Anatase TiO2 powder XPS spectra afte r air calcination at 500oC for 3 hours and hydrogen reduction at 400oC and slight oxidation in air at 200oC. 2.4 World Gold Council Reference Catalyst A (nano-Au/TiO2) According to the material safety data sheet (MSDS), the catalyst A contains 1.5 wt% Au with the balance TiO2, and it was prepared via deposition/precipitation. The average diameter of the gold particles, verified via TEM, is 3.3 nm with a standard deviation of 0.72 nm.13 Additionally, it has been shown, in a fixed bed flow reactor, the temperature at which there is a 50% conve rsion for CO oxidation is -46oC.13 Accordingly, the temperature at which there is a 50% conversion for H2 oxidation is 40oC.13 Therefore, the WGC reference is highly re active with both CO and H2 at the typical fuel cell operating temperature of 80oC. SEM micrographs of the WGC refe rence catalyst shows that the average particle size is a lot smaller than the Sigma-Aldrich support used in all experiments. This may be an advantage of the reference material that needs to be explored in my system. Ball-milling may be a requirement to lower the aver age diameter of the support material.

PAGE 46

34 Figure 2.19 SEM micrographs of the twophase-method gold reveal similar pictures to the As-received TiO2 used as a support. Figure 2.20 SEM micrographs of the WGC show much small particle sizes for the TiO2 support used in their samples. 2.4.1 XPS Analysis Analysis of the XPS spectrum (Figure 2.21) reveals that the World Gold Council supported nano-Au is primarily metallic gold, Au0 4f7/2 peak (84.00 eV) and Au0 4f5/2 peak (87.71 eV). This is comparable to the two-phase-method nano-Au provided by the Interfacial Phenomena and Polymeric Materials research group (Figure 2.6). Although there is the possibility of a suboxide species present, Au+1, the minute quantity of the sample and resolution make it difficult to distinguish. Figure 2.21 WGC nano-Au/TiO2 XPS spectrum.

PAGE 47

35 2.5 Summary of Results Initial characterizations of the two-phase-method gold nanoparticles, titanium dioxide, and the world gold council supported Au/TiO2 were performed. After the two-phasemethod gold nanoparticles were successfully fabricated and combined with TiO2 via incipient wetness techniques, a pre-treatm ent procedure based on techniques used by Haruta and Choudhary was developed. This method includes a calcination step at 500oC for solvent removal, a reduction step at 400oC to reduce the GNP to their metallic state, and a slight oxidation step at 200oC to promote the formation of sub-oxide gold species. XPS verified the metallic state of the GNPs after exposure to 500oC calcination temperature. Finally, SEM illustrated the high mobility of the GNPs under relatively mild conditions. This was a cause for concern because of the deactivat ion of the particles above ~100 nm diameters. Continuing with the characterization, the TiO2 was then examined. After calcination at 500oC in air at atmospheric pressure of the TiO2 support, XRD confirmed that the crystal structure changed from anatas e to rutile in a small percentage of the sample. The presence of the rutile phase is suggested to be advant ageous due to an optimized interaction between the GNP and the TiO2 support. Analysis of the TiO2 support via FTIR found that the availabl e CO concentration on the support surface to have a negative temperature dependence when varying the system temperature from 25oC-125oC. This result is proposed to be due to the decreasing sticking probability of CO on the TiO2 support with increasing temperature, the w eak adsorption of CO on the support surface, and the decreasing residence time of CO with increasing temperature. Lastly, the World Gold Council was characte rized via SEM and XPS. SEM micrographs of the World Gold Council Au/TiO2 catalyst revealed that th e support used for its catalyst have much smaller diameter particles than the anatase TiO2 used in the two-phasemethod gold experiments. XPS spectra showed that the WGC supported GNPs are metallic gold and very similar to the two-phase-method GNP provided by Dr. Gupta.

PAGE 48

36 3 DEVELOPMENT OF TWO-PHASE-M ETHOD AU NANOPARTICLE TESTING PROCEDURES AND PROTOCOLS FOR THE OXIDATION OF CO 3.1 Introduction In order to verify the ability of the tw o-phase-method GNP catalyst provided to oxidize CO in excess hydrogen, an effluent testing pr otocol was developed. To determine CO conversion, the wavenumber range for each sp ecies of the FTIR absorbance spectrum must be integrated and referenced to a background spectrum. Experimental artifacts corresponding to an eight hour cycle in the absorbance data were observed. These artifacts, identified as shifts in the calculated absorbance integrals, were first attributed to the FTIR heating/cooling cycle. This problem made conversion calcul ations impossible. After identifying the problem, attempts were made to remove the issue through the use of a non-reactive reference species such as methane. After noticing inconsistencies in the noise levels at different temperatures, baseline adjustments for each species and N2 purging to eliminate background CO2 and H2O contamination were found to be critical for direct comparison of results. After th ese calibration corrections were implemented, accurate assessments of the two-phase-method GNP catalyst CO oxidation activity was evaluated. 3.2 Twenty-four Hour Analysis of Tw o-Phase-Method GNP Catalyst FTIR Absorbance Integrals To determine the effluent characteristic s of the two-phase-method GNP catalyst, a twenty-four hour experiment was designed to record the absorbance integral of CO (2000-2250 cm-1), CO2 (2250-2400 cm-1), and H2O (1400-1900 cm-1) every hour (Figure 3.1). The absorbance integral is linearly proportional to the concen tration of the species and will be essential in determining the convers ion potential of these catalysts. As stated previously in Section 3.2, a 1% two-phase-method GNP/TiO2 catalyst was prepared by incipient wetness and dried, calcined, reduced and slightly oxidized for FTIR analysis.

PAGE 49

37 The influent stream of 99 sccm H2, 1 sccm CO, 50 sccm N2, and 5 sccm air was reacted isothermally at 200oC and 1 atm. Specie concentrations were then evaluated via ratios of the absorbance integrals. Initial observation showed a cycling pattern of the FTIR absorbance integrals approximately every 8 hours (Figure 3.1). The mercury cadmium telluride (MCT) detector reservoir is filled with liquid nitr ogen at the beginning of each experimental run and allowed to equilibrate for approximately 2 hours before data collection. After every 8 hours, the detector reservoir was refilled. It is hypothesized that the detector is warming over the period of 8 hours. Once the detector is cooled again with liquid nitrogen, the measured integral value readjusts to a new value. In the case of the CO integral, this can be up to a bout 10-20%. With this large m ovement in the integral value proposed to be due to the detector warming/ cooling over a period of time, it is impossible to determine if the concentration is changing because of a reaction. The second concern with the FTIR spectra is the negative inte grals for both the water and carbon dioxide. Since concentration is linearly proportional to the integral of the absorbance, negative values do not have any physical meaning. Th is can be explained through the calculation of FTIR absorbance (Eqn. 3.1). If the background response curve (Io) is approximately equal to the experimental response (I), then the calculation of absorbance can produce negative values caused by noise. The integral of negative values could lead to the values calculated. Absorbance = A = log(Io/I) = ecL (3.1) The FTIR experimental con centration ratios of CO/CO2, CO/H2O, and CO2/H2O established a stabilized system after ~2 hours (Figure 3.2) since the ratio values were found to be proportional to each other.

PAGE 50

38 Figure 3.1 FTIR absorbance integr als of carbon monoxi de, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 5 sccm air. Figure 3.2 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 5 sccm air. MCT detector filled

PAGE 51

39 In subsequent experiments, the air concentra tion was increased to 15 sccm to determine if the oxygen concentration may be limiting the reaction; all other influent components remained constant. To determine any temperature dependence on the absorbance integral, isothermal reactions were evaluated at 200oC (Figure 3.3 and Figure 3.4), 300oC(Figure 3.5 and Figure 3.6) and 425oC(Figure 3.7 and Figure 3.8). Initial results were inconclusive until the experimental artif acts were removed. The final results are discussed in Section 4.4. Once again, an 8 hour cyclic nature to the FT IR absorbance integrals was evident. This observation verified the result of the changing detector te mperature in conjunction with the possibility of a non consistent background atmosphere. One proposed solution was to introduce an inert gas into the influent whic h would remain at a constant concentration throughout the experiment. Adjusting the sign al strength such that the inert gas is a straight line will line up the other gas concentrations to their true values. Unfortunately, one of the major limitations of FTIR analysis is its inability to measure diatomic species concentrations. Therefore, the nitrogen inert added as air cannot be used as a reference. Another inert gas, una ffected by either the gold nanoparticles or the influent species must be chosen. The gas ultimately selected was methane. The results of those experiments are presented next in Section 4.3

PAGE 52

40 Figure 3.3 FTIR Absorbance integr als of carbon monoxide, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air. Figure 3.4 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 200oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air.

PAGE 53

41 Figure 3.5 FTIR absorbance integr als of carbon monoxi de, carbon dioxide and water at 300oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air. Figure 3.6 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 300oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air.

PAGE 54

42 Figure 3.7 FTIR absorbance integr als of carbon monoxi de, carbon dioxide and water at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air. Figure 3.8 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide and water at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air.

PAGE 55

43 3.3 Data Analysis with Reference Methane in the Influent Stream Although there have been some studies to oxi dize methane using gol d catalysts, there has been limited success in gas phase reactions.50 Significant reactions involving methane oxidation and gold particles have been made w ith liquid phase reacti ons utilizing selenic acid H2SeO4 as an oxidant.50 With no promoters available and at these temperatures, methane will pass unreacted thr ough the catalyst bed. This w ill be an excellent reference since the FTIR can easily de tect the presence of the me thane species. After the introduction of 10 sccm methane into the feed stream, it appears that there is an inconsistent methane signal which was unexpected (Figure 3.9). It is interesting to note that the methane signal follows the pattern of the carbon monoxide almost exactly. This would suggest that the carbon m onoxide is not reacting in this system since it is seems to be directly proportional to the unreactive methane. The ratios lead to very few conclusions other than the cyclic nature of the FTIR effluent absorbance integral of each species (Figure 3.10). Figure 3.9 FTIR absorbance integrals of carbon monoxide, carbon dioxide, water, and methane at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane.

PAGE 56

44 Figure 3.10 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide, water and methane at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane. 3.4 Reevaluation of the FTIR Spectra to Account for Shifting Baselines One observation that I have had w ith the last set of data runs is that the mean value of the noise level changes for each of the species as the runs proceeds th rough the 24 hours. In order to accurately determine the value of th e areas for each of th e data points, it was suggested that I reference the initial startin g point of each of the species ranges to the same point. Since the mean value of the noise level is supposed to be zero, I chose this value as a reference. My new evaluation procedure was a linear base line subtraction from the first point to the last point in the species evaluation range. Given a common starting and ending point the areas should line up. Evaluation of the integrals will then be on areas of the curves with the background subtracted. Each curve is evaluated on an individual basis and baseline corrected. After the baseline correction, one obvious change is noticed in Figure 3.12. The integrals for the components actively controlled and included in the influent flatten out and

PAGE 57

45 become relatively constant values with change s over the entire 24 hour time span approaching about 1% for both the carbon m onoxide and the methane. Before the reevaluation, the changes as observed in Figure 3.11 show fluctuat ions approaching 20%. While the signals for the gas stream appear to be constant, the values for CO2 and H2O are still cycling with the introduction of liqui d nitrogen. One possible explanation of this is that the sample chamber is not under su fficient vacuum thus water vapor and carbon dioxide are merely changing due to the una voidable experimental variations in the surrounding atmosphere around the gas cell. Figure 3.11 FTIR absorbance integrals of carbon monoxide, car bon dioxide, water and methane at 425oC. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane.

PAGE 58

46 Figure 3.12 Ratios of FTIR absorban ce integrals of carbon monoxide, carbon dioxide, water and methane at 425oC after baseline subtraction. Influent gas composition: 99 sccm H2, 1 sccm CO, 50 sccm N2, and 15 sccm air, 10 sccm methane. With this updated method of observation for the FTIR spectra, one can see if, in fact, any reactions occurring. The reevaluated CO abso rbance integral for each run is presented in Figure 3.13. While there is a slight decrease in the equilibrium concentration of carbon monoxide, there does not appear to be a signific ant change in the react ivity of the system. A couple things to note about the two runs are that at each temperature, the concentration of CO went down as the con centration of air went up. There was not a huge difference but the trend was as hypothesized that the system may be oxygen deficient.

PAGE 59

47 Figure 3.13 Carbon monoxide absorbance inte grals for all data collected reevaluated using baseline subtraction to al low for direct comparison. 3.5 Redesign of the Supported Gold Catalyst Results from the FTIR analysis have been less than promising for the catalysts produced. After discussions with both Dr. Gupta and Da vid Walker from the Interfacial Phenomena and Polymeric Materials re search group at USF, an improved creation method was devised to increase the adherence of the GNP to the TiO2. This improved procedure is as follows: Walkers improved procedure begins with 0.25 mL 0.01 M aqueous solution of hydrogen tetrachloroaurate (HAuCl4) added slowly to a 0.50 mL 0.01 M sodium citrate solution to stabilize the nanoparticles, 0.1 M sodium bor ohydride as a reduction agent, and 9 mL of water. The solution was then stirred for 2 hr to allow the formation of GNPs. Approximately 100 mg of anatase TiO2 was then added, the pH of the entire system was adjusted to 4.45, and the catalyst was conti nuously stirred for a nother 30 minutes. An

PAGE 60

48 acidic solution induces a positive charge to the surface of the TiO2 which promotes the adhesion of the negatively charged GNP. After 30 minutes, the supported GNP/TiO2 catalyst was evaporated and dried in a vacuum oven. Initially, no attempt was made to expose the GNPs to excessive heat to determine if the calcination temperatures used may be result ing in deactivation. The results of the experimental run with the improved catalyst s till yielded very little difference in the values between the non-reacted bypass influe nt absorbance integrals and the reaction effluent concentrations. Figure 3.14 shows that the absorbance integrals at each temperature and flow to be approximately e qual. This correlates to a CO fractional conversion of approximately zero. Ne xt, the catalyst was heated to 205oC for 2 hours in hydrogen. This is based on TGA results which show a decrease in the catalyst weight of ~9% at 200.9oC for the previous catalyst formula (Figure 3.15). The evaporating species at this temperature has yet to be identified. Again, no sizable change in the absorbance integral for carbon monoxide is observed. Only the 25oC data is presented for the catalyst after exposure to the 205oC reduction step because that temperature has been shown to demonstrate highest conversion in the World Gold Council catalyst samples. 3.6 Summary The development of a FTIR effluent testing protocol which include s baseline subtraction and nitrogen purging is fully functional with both the catalyst provided by the Interfacial Phenomena and Polymeric Materials research group at USF as well as the World Gold Council. The baseline subtracti on of the absorbance data to allows direct comparison of data taken under dissimilar conditions and nitrogen purging is used to eliminate background CO2 and H2O contamination. Unfortunately, the supplied two-phase-method gol d catalyst has yet to show any promise for CO oxidation applications. One hypothesis for the problems is that the surfactants used to keep the gold nanoparticles from aggregating are preven ting carbon monoxide transport to the surface of the particle. A nother theory is that the gold may not be

PAGE 61

49 adhering to the surface of the TiO2 creating a cohesive metal/suppo rt interacti on. Future directions for this project are discussed in Chapter 6. Figure 3.14 FTIR results of redesigned two-phase-method GNP catalyst before and after exposing the sample to a 205oC calcination step.

PAGE 62

50 Figure 3.15 TGA analysis of original two-phase-method GNP/TiO2 catalyst formula.

PAGE 63

51 4 EMPIRICAL MODELS OF CARB ON MONOXIDE OXIDAT ION VIA WORLD GOLD COUNCIL Au/TiO2 IN EXCESS HYDROGEN 4.1 Introduction Numerous research groups ha ve reported CO oxidation frac tional conversion over metal oxide-supported nano-Au catalysts at temperatures below 0oC of 1% CO in Air.51,52 However, information regarding CO prefer ential oxidation (PROX) at temperatures above 0oC is lacking. Although conceptually simple, the oxi dation of CO in the presence of excess H2 and CO2, without oxidizing the hydrogen or remaking CO via the reverse WGS reaction, is a particularly difficult obj ective, which has so far only been achieved using a multistage reactor.53 The presence of excess CO2 presents an additional problem since if the catalyst is active for the reverse water gas shift reaction then CO will be remade; this is a key feature that has limited success in this field to date. Due to the recent growth in research on fuel cells and fuel processing, a large number of studies on CO selective or pref erential oxidation (P ROX) have been published. Indeed, a comprehensive kinetic model for this catalytic system which includes the contributions of CO oxidation, H2 oxidation and the WGS reaction above 0oC is clearly needed. Most recent papers are focused on catalyst form ulation, characterization, and basic performance such as activity and selectivity of CO. Few papers ha ve investigated the kinetics and rate expressions of the reactions involved. In this study, kinetic models are presente d which predict CO frac tional conversions at typical PEMFC operating conditi ons catalyzed by nano-Au/TiO2 provided by the World Gold Council (WGC). This WGC catalyst is pr omoted as a benchmar k for researchers to evaluate their own ca talyst systems against a common reference.7 Although many studies have examined the extreme activity of this pa rticular catalyst to oxidize CO at cryogenic temperatures, this study is one of the first attempts to provide a comprehensive kinetic

PAGE 64

52 model of the WGC nano-Au/TiO2 catalyzed PROX of CO in excess hydrogen at temperatures ranging from 25oC-125oC. Although these models do not take into consideration the effects of influent moisture content, this work provides some insight into the behavior of this reference catalyst and its use in possibl e hydrogen and direct alcohol fuel cell applications. 4.2 Experimental All reactions were performed in a 24 in. quartz tube, 4 mm ID, vertical packed bed micro-reactor inside of a Lindberg/Blue split tubular furnace operating at 1 atm. One hundred milligrams of nano-Au/TiO2 powder were packed loosely between high temperature quartz wool in order to stabil ize and ensure proper influent distribution through the bed. The reactor was operated unde r steady state conditions at temperatures ranging from 25-125oC in 25oC increments. The influent consists of a 1% CO/H2 mixture with a constant 10 s ccm of air. The 1% CO/H2 feed rate was varied from 20-100 sccm in 20 sccm increments. With the air included, this corresponds to a space velocity ranging from 18000 to 66000 mL/hr/gcat. No attempts were made to humidify or dehumidify the influent stream. The feed stream moisture cont ent was assumed to be ~0%. The reaction effluent is fed directly to an inline Bio-Rad FTS 3000 Excalibur Series Fourier transform infrared spectrometer gas ce ll for in-situ analysis. No attempt was made to quench any effluent reactions before entering the FTIR. Specifications for the gas cell are given in Section 1.5.3. 4.3 Calculation of Thermodynamic Properties Before examining the empirical kinetic calc ulations, one should determine if the given reactions are thermodynamically feasible through the analysis of Gibbs free energy. The thermodynamic properties which determine whet her or not a given reaction is favorable (spontaneous) are enthalpy and entropy.54 Some reactions are spontaneous because of the heat given off during the reaction ( Ho < 0). Other reactions are spontaneous because of an increase in the entropy or disorder of the system ( So > 0). However, sometimes, one property may be favorable while another is not Another term had to be defined which reflects the balance between the heat of reaction and the change in entropy. This value is

PAGE 65

53 known as the Gibbs free energy and is defined as the enthalpy of the system minus the product of the temperature times the entropy of the system.54 A negative Gibbs free energy indicates a favorable, spontaneous reaction and is most useful for thermochemical processes at constant temperatur e (isothermal) and pressure (isobaric).54 Calculations of both the heat of reaction a nd the Gibbs free energy at room temperature for each reaction are given in Table 4.1. All data for thes e calculations were obtained from Appendix C in Introduction to Chemical Engineering Thermodynamics .54 These calculations show that CO oxidation, H2 oxidation and the WGS are spontaneous while the RWGS is not based on the sign of the calculated Gibbs free energy of reaction. Table 4.1 Standard heats of reaction and Gibbs free energies of reaction. Ho 298 K (kJ/mol) Go 298 K (kJ/mol) CO + O2 CO2 -283 -257 H2 + O2 H2O(g) -242 -228 CO + H2O CO2 + H2 -41.2 -28.6 CO2 + H2 CO + H2O 41.2 28.6 There is very little change in the calculated values of the Gibbs energy of reaction over the temperature range evalua ted in these studies (25oC-125oC) (Table 4.2). Based on equations outlined in Smith and Van Ness54, the equilibrium constants are shown in Table 4.3. The equilibrium constant is another indication of the extent and direction of the given reactions. A large value indicates the fo rmation of the products in that reaction. Again, this validates the spontan eous reaction of CO oxidation, H2 oxidation, and the WGS.

PAGE 66

54 Table 4.2 Gibbs free energies of reaction at each temperature. Go 298 K (kJ/mol) Go 323 K (kJ/mol) Go 348 K (kJ/mol) Go 373 K (kJ/mol) Go 398 K (kJ/mol) CO + O2 CO2 -257 -255 -253 -251 -248 H2 + O2 H2O(g) -228 -227 -226 -225 -224 CO + H2O CO2 + H2 -28.6 -27.8 -26.5 -25.5 -24.4 CO2 + H2 CO + H2O 28.6 27.8 26.5 25.5 24.4 Table 4.3 Calculated equilibrium constants at each temperature. K298 K K323 K K348 K K373 K K398 K CO + O2 CO2 1.21x1045 1.76x1041 9.01x1037 1.27x1035 4.09x1032 H2 + O2 H2O(g) 1.17x1040 6.06x1036 9.17x1033 3.27x1031 2.33x1029 CO + H2O CO2 + H2 104000 28800 9620 3730 1640 CO2 + H2 CO + H2O 9.6x10-6 3.47x10-5 1.03x10-4 2.68x10-4 6.10x10-4 4.4 Empirical Models The next sections examine the potential of empirical models to predict the effluent concentrations of CO oxidation over the WGC nano-Au/TiO2 at various temperatures and space velocities. Matlab code for each model can be found in Appendix A. 4.5 Elementary CO Oxidation Single Reaction Model Based on previous studies at cryogenic temperatures55, a simple elementary reaction model was chosen first to predict the carbon monoxide oxidation reaction parameters (Eqn. 4.1): activation energy (E ) and pre-exponential factor ( ). This model assumes an elementary reaction order; a steady state, is othermal, isobaric reaction; Arrhenius reaction rate constants, and ignores the contribu tion from the water gas shift and hydrogen oxidation reactions.

PAGE 67

55 Table 4.4 Experimental influent conditions to the tubular reactor. Flowrate Air (sccm) 10 10 10 10 10 Flowrate 1% CO/H2 (sccm) 20 40 60 80 100 Flowrate O2 (sccm) ~2 ~2 ~2 ~2 ~2 Flowrate CO (sccm) 0.2 0.4 0.6 0.8 1 Space Velocity (mL/hr/g-cat) 18000300004200054000 66000 2O =FO2/FCO 10 5 3.33 2.5 2 *103 -3.33 -4 -4.28 -4.44 -4.54 2 22 1 CO O CO (4.1) Utilizing the design equation of a packed bed reactor, the following Eqns. 4.2-5 can be formulated to determine CO conversion. 2 12'O CO CO COC kC r W X Fo (4.2) RT Ek exp (4.3) X O COdX X X C X X fo0 2 1 2 3 2 3) 2 1 )( 1 ( 1 ) (2 (4.4) W X f F RT EoCO) ( ln ln (4.5) A graph of W X f FoCO) ( ln vs. T-1 should yield a slope of -E/R and a y-intercept of ln if the system can be modeled as a single elementary reaction. 4.5.1 Elementary CO Oxidation Single Reaction Model Results The FTIR CO absorbance integral (2145-2230 cm-1) of both an unreacted reference (AREF) and after reaction (ARXN) were used to calculate CO experimental conversion

PAGE 68

56 defined as (AREF ARXN)/AREF. These results are displayed in Table 4.5. The plot of W X f FoCO) ( ln vs. T-1 led to a completely non-linear plot which does not allow an accurate estimation of either the activation or of the pre-exponential factor (Figure 4.1). This plot shows that the carbon monoxide preferential oxida tion reaction is nonelementary in nature at thes e conditions and the effluent CO concentrations cannot be modeled by evaluating a single elementa ry CO oxidation reaction. A more comprehensive model which includes the contributions of the H2 oxidation and WGS reaction is necessary. Table 4.5 Experimental reacti on effluent fractional conversio ns at each temperature and flow rate. 25oC 50oC 75oC 100oC 125oC (20 sccm 1%CO/H2)/100 mg 0.3062 0.0371 0.0026 0.0057 0.0081 (40 sccm 1%CO/H2)/100 mg 0.1546 0.0702 0.0032 0.0898 0.0041 (60 sccm 1%CO/H2)/100 mg 0.0338 0.0218 0.0030 0.0851 0.0020 (80 sccm 1%CO/H2)/100 mg 0.0312 0.0177 0.0021 0.0017 0.0015 (100 sccm 1%CO/H2)/100 mg 0.0237 0.0153 0.0014 0.0023 0.0015 Figure 4.1 Graph of W X f FoCO) ( ln vs. T-1 for the simple elementary model.

PAGE 69

57 The calculated error for each fractional conversion is shown in Table 4.6. This calculation is based on the mean and standa rd deviation of the FTIR reference and reaction integrals. Each sample was collect ed four times at each temperature and space velocity. Propagation of error based on samp ling accuracy has not been included. The calculations of error are included in Appendix A. Table 4.6 Calculated error with each fractional conversion value. 25oC 50oC 75oC 100oC 125oC (20 sccm 1%CO/H2)/100 mg 0.55 0.40 0.32 0.33 0.28 (40 sccm 1%CO/H2)/100 mg 0.42 0.35 0.27 0.24 0.16 (60 sccm 1%CO/H2)/100 mg 0.37 0.34 0.24 0.20 0.14 (80 sccm 1%CO/H2)/100 mg 0.40 0.35 0.24 0.17 0.12 (100 sccm 1%CO/H2)/100 mg0.47 0.42 0.33 0.17 0.11 4.6 Comprehensive PROX Models After being unable to model the effluent conc entrations using a sing le reaction, a more comprehensive model set was examined. The following two kinetic models utilize all of the primary reactions in the PROX of CO (CO oxidation (Eqn. 4.6.a), H2 oxidation (Eqn. 4.6.b), and the WGS/RWGS reactions (Eqn. 4. 7.a-b)). These sets of experiments are based on similar work examining CO oxidation via Pt/TiO2.8 In both mode ls, the values of the kinetic models pre-e xponentials and activa tion energies were determined by minimizing the sum of the square of the di fference between predic ted and experimental CO conversion via nonlinear least squa res fitting (LSQCURVEFIT) in Matlab.56 All Matlab code for these models is provided w ith this text. The pr edicted CO conversion was calculated by solving the kinetic model s ordinary differential equations by the fourth order Runge-Kutta method and compared to experimental FTIR conversion data. A flow diagram of the kinetic model path is shown in Figure 4.2. 2 22 1 CO O CO O H O H2 2 22 1 (4.6.a-b)O H CO H CO2 2 2 2 2 2H CO O H CO (4.7.a-b)

PAGE 70

58 Figure 4.2 Kinetic model flow diagram. 4.7 Comprehensive Elementary Reaction Model After defining the total molar flow rate of th e gas stream as the sum of the components (Eqn. 4.8), the concentration of the individual species are ca lculated as the mole fraction times the total concentration (Eqn. 4.9). This model assumes elementary reaction orders of all components (Eqns. 4.10.a-b, 4.11.ab); isothermal, isoba ric reactions; and Arrhenius reaction rate consta nts, and steady state operation. After describing the mole balances for each reaction (Eqns. 4.12.a-b, 4. 13.a-b, and 4.14.a-b), fractional conversion can be predicted and compared to the experime ntal data. Fractional conversion of CO is defined as (FCO o FCO@W=100mg)/ FCO o. 2 2 2 2 2N O H H CO O CO TotalF F F F F F F (4.8) Total xx o Total xxF F C C (4.9) 2 1 1 12O COC C k r 2 1 2 22 2O HC C k r (4.10.a-b) Define Constants and Variables Flow rates, Temperature, Experimental FTIR CO Conversion Results Define the mole balance equations for each component Select starting values and any boundary conditions Compute models conversions Does the models predicted values far within tolerance? Output Empirical Values YES Adjust the variables to minimize the Least Squares Error NO Is this an elementary or non-elementary reaction? New Model Variables

PAGE 71

592 23 3H COC C k r O H COC C k r24 4 (4.11.a-b) 4 3 1r r r dW dFCO 2 12 1 2 12r r dW dFO (4.12.a-b) 4 3 12r r r dW dFCO 4 3 22r r r dW dFH (4.13.a-b) 4 3 22r r r dW dFO H 02 dW dFN (4.14.a-b) 4.8 Comprehensive Non-Elemen tary Reaction Model The non-elementary kinetic model also assumes steady state, isothermal, isobaric reactions, and Arrhenius reaction rate constants. However, this model removes the constraints of elementary reaction orders. A ll steps are repeated as in the elementary reaction except Eqns. 4.15.a-b and Eqns. 4.16.a-b replace E qns 4.10.a-b and Eqns 4.11.ab and the concentration dependence (reaction or der) for each is evaluated. Fractional conversions are then pred icted and compared to the experimental data. 2 2 11 1 Exp O Exp COC C k r 2 2 1 22 2 Exp O Exp HC C k r (4.15.a-b)2 2 1 23 3 Exp H Exp COC C k r 2 2 14 4 Exp O H Exp COC C k r (4.16.a-b) 4.9 Comparison of Comprehensive Mod els to Experimental Results At lower temperatures (25oC and 50oC), the CO fractional conversions did not follow as smooth a curve as expected between flow ra tes. One hypothesis is that the reacting system is in a transitional state betw een dominance by the carbon monoxide oxidation regime and one dominated by hydrogen oxidation and the WGS reaction. Another theory is that the production of liqui d water may be hindering th e catalyst at these lower temperatures.52,57 It is interesting to note that the highest CO c onversion occurred at the lowest temperature and flow rate which s eems counter intuitive to initial thoughts that CO conversion would increase w ith increasing temperature as seen in previous cryogenic studies.55

PAGE 72

60 4.9.1 Comprehensive Elementary Reaction Model Results As shown in Figure 4.3, the elementary model does not accurately depict the CO fractional conversions experime ntally determined. The empi rical parameters found are listed in Table 4.7. These calcula tions revealed that the CO concentrations cannot be modeled given the elementary reaction order restraints. Us ing equations 4.12.a-b, 4.13.ab, and 4.14.a-b along with Table 4.7 yields the empiri cal mole balances found in Table 4.8. Table 4.7 Calculated pre-exponentials and ac tivation energies for the elementary model which includes CO oxidation, H2 oxidation and the WGS reaction. Pre-exponential Activation Energy (kJ/mol) 2 1 1 12O COC C k r 683 ~0 2 1 2 22 2O HC C k r 1350 6.8 2 23 3 H COC C k r 1890 3.6 O H COC C k r24 4 0.126 43.0 Table 4.8 Empirical mole balances for the elementary model which includes CO oxidation, H2 oxidation and the WGS reaction. O H CO H CO O CO COC C RT C C RT C C dW dF2 2 2 2* ) / 0 43 exp( 126 0 ) / 6 3 exp( 1890 6832 1 O H CO H CO O CO COC C RT C C RT C C dW dF2 2 2 2 2* ) / 0 43 exp( 126 0 ) / 6 3 exp( 1890 6832 1 O H CO H CO O H O HC C RT C C RT C C RT dW dF2 2 2 2 2 2* ) / 0 43 exp( 126 0 ) / 6 3 exp( 1890 ) / 8 6 exp( 13502 1

PAGE 73

61 Table 4.7 (Continued) 2 1 2 12 2 2 2* ) / 8 6 exp( 1350 2 1 683 2 1O H O CO OC C RT C C dW dF O H CO H CO O H HC C RT C C RT C C RT dW dF2 2 2 2 2 2* ) / 0 43 exp( 126 0 ) / 6 3 exp( 1890 ) / 8 6 exp( 13502 1 Figure 4.3 Comparison of model results fo r the comprehensive elementary model. 4.9.2 Comprehensive Non-Elementary Reaction Model Results After removing the restrictions of elem entary reaction orders for each of the concentrations, Figure 4.4 shows that there is an accura te representation of the effluent concentrations predicted by the non-elementary kinetic model. The model parameters are displayed in Table 4.9. This kinetic model calcula ted positive values for all activation energies, and pre-exponentials. This allows comparisons to literature values for similar reaction mechanisms.

PAGE 74

62 After comparing these results to those found by Haruta for CO oxidation over the identical catalyst for 1% CO in air, ther e appears to be some correlation. Haruta discovered that the activation energy of the carbon monoxide reaction approached zero as one approaches standard temp eratures and pressures (25oC, 1 atm) from cryo temperatures.31 While the empirically derived valu e in this model is not zero, the activation energy of the CO oxidation reactio n is significantly lower than the other reactions, resulting in less temp erature dependence than all of the other reactions. He also found that the oxygen reaction order for CO oxidation to be between 0-0.25 which correlates well with the 0.15 reaction order calculated.31 However, he also found that the CO reaction order for CO oxidation to be ~0.31 This does not ma tch well with 0.91 I calculated. It is speculated that the di fference in the reaction order is due to the synergistic combination of the four reactions and the possibility of CO being reformed due to the RWGS reaction. Since in his experiments, Haruta did not have any H2 in the feed stream, there are no comparable numbers for the hydrogen oxidation or WGS/RWGS reactions. Table 4.9 Calculated pre-exponentials and activation energies fo r the non-elementary model which includes CO oxidation, H2 oxidation and the WGS reaction. Pre-exponential Activation Energy (kJ/mol) EXP1 EXP2 2 2 111 Exp O Exp COCCkr 99.0 0.4 0.91 0.15 2 2 1 222 Exp O Exp HCCkr 100.8 9.6 0.96 0.13 2 2 1 233 Exp H Exp COCCkr 117.4 8.3 0.38 0.80 2 2 144 Exp OH Exp COCCkr 109.0 8.7 0.75 0.51 Table 4.10 Empirical mole balances for the non-elementary model which includes CO oxidation, H2 oxidation and the WGS reaction. 510750 800380 1509102 2 2 2*)/7.8exp(*0.109 *)/3.8exp(*4.117 *)/4.0exp(*0.99OH CO H CO O CO COCCRT CCRT CCRT dW dF

PAGE 75

63 Table 4.9 (Continued) 510 750 800380 1509102 2 2 2 2*)/7.8exp(*0.109 *)/3.8exp(*4.117 *)/4.0exp(*0.99OH CO H CO O CO COCCRT CCRT CCRT dW dF 510750 800380 1309602 2 2 2 2 2*)/7.8exp(*0.109 *)/3.8exp(*4.117 *)/6.9exp(*8.100OH CO H CO O H OHCCRT CCRT CCRT dW dF 130960 1509102 2 2 2*)/8.6exp(*8.100* 2 1 *0.99* 2 1O H O CO OCCRT CC dW dF 510750 800380 1309602 2 2 2 2 2*)/7.8exp(*0.109 *)/3.8exp(*4.117 *)/6.9exp(*8.100OH CO H CO O H HCCRT CCRT CCRT dW dF Figure 4.4 Comparison of model results fo r the comprehensive non-elementary model.

PAGE 76

64 4.10 Linearly Independent Model Equations for CO Oxidation The preceding models in Sections 0-4.9 were based on similar research by Choi et al. for CO oxidation via Pt/TiO2.8 However, after examination of the four reactions in Eqns. 6.a-b and 7.a-b, one can se e that the CO oxidation, H2 oxidation and WGS/RWGS reactions are not linearly inde pendent. Equation 6.a minus equation 6.b equals the water gas shift reaction (Equation 7.b). Also, the WGS/RWGS reactions are not linearly independent of themselves. Removing the hydrogen oxidation reacti on (Eqn. 4.6.b) and restraining the reaction rate constants to be dependent on each based on the equilibrium constants calculated in Table 4.3 allows the system to modeled by a set of linearly independent reactions as shown in Equations 4.17-19. In the linearly independent elementary model, only four variables are undefined, 1, E1, 2 and E2. All previous assumptions (isothermal, isobaric reactions, A rrhenius reaction rate constants, and steady state operation) are still valid. For the nonelementary linearly independent kinetic model, the reaction orders will be undefined brings the total number of variables to ten: 1, E1, 2, E2, and six reaction orders. 2 1 1 12O COC C k r ( 4.17)2 22 2 H COC C k r ( 4.18)O H COC C K k r2* *(RWGS) K 298 2 3 ( 4.19) 4.10.1 Linearly Independent Elemen tary Reaction Model Results Results for linearly independent elementary mo del once again fail to accurately depict the effluent CO concentrations. The values of the pre-exponentials and activations energies for the closest fit are shown in Table 4.11-11 and the graphed in Figure 4.5.

PAGE 77

65 Table 4.11 Calculated preexponentials and activation en ergies for the linearly independent elementary model reaction. Pre-exponential Activation Energy (kJ/mol) 2 1 1 12O COC C k r 231 5.8 x 10-5 2 22 2 H COC C k r 957 2.97 x 10-3 Table 4.12 Empirical mole balances for the linearly independent elementary model reaction. O H CO H CO O CO COC C K RT x C C RT x C C RT x dW dF2 2 2 2* ) / 10 97 2 exp( 957 ) / 10 97 2 exp( 957 ) / 10 8 5 exp( 231(RWGS) K 298 3 3 2 1 5 O H CO H CO O CO COC C K RT x C C RT x C C RT x dW dF2 2 2 2 2* ) / 10 97 2 exp( 957 ) / 10 97 2 exp( 957 ) / 10 8 5 exp( 231(RWGS) K 298 3 3 2 1 5 O H CO RWGS K H CO O HC C K RT x C C RT x dW dF2 2 2 2* ) / 10 97 2 exp( 957 ) / 10 97 2 exp( 957) ( 298 3 3 2 1 52 2* ) / 10 8 5 exp( 231 2 1O CO OC C RT x dW dF O H CO H CO HC C K RT x C C RT x dW dF2 2 2 2* ) / 10 97 2 exp( 957 ) / 10 97 2 exp( 957(RWGS) K 298 3 3

PAGE 78

66 Figure 4.5 Comparison of model results for the linearly independent elementary model. 4.10.2 Linearly Independent Non-elem entary Reaction Model Results The last model in this series is the linearly independent non-elementary kinetic model. This model does represent the experimental data (Figure 4.6). The concentration dependence of the CO oxidati on is not significantly differe nt than what was calculated via the comprehensive non-linearl y independent model found in Table 4.9. The calculated activation energy and pre-exponential are much higher in this model than in the previous incarnations. Once again the CO reaction order in CO oxidation is ~1.0 which is contrary to the expected value of ~0.31 This last model proves that the PROX of CO in excess H2 can be modeled within the temper ature and space velocity constraints using a linearly independent set of reactions.

PAGE 79

67 Table 4.13 Calculated preexponentials and activation en ergies for the linearly independent non-elementary model. Preexponential Activation Energy (kJ/mol) EXP1EXP2 2 2 111 Exp O Exp COCCkr 10400 3.7 1.02 0.26 2 2 1 222 Exp H Exp COCCkr 71.9 6.9 0.34 0.45 2 2 1* *(RWGS)K 29823 Exp OH Exp COCC Kkr NA NA 0.32 2.89 Table 4.14 Empirical mole balances for th e linearly independent non-elementary model. 892320 (RWGS)K 298 450340 2600212 2 2 2* *)/9.6exp(*9.71 *)/9.6exp(*9.71 *)/7.3exp(*10400OH CO H CO O CO COCC KRT CCRT CCRT dW dF 892320 (RWGS)K 298 450340 2600212 2 2 2 2* *)/9.6exp(*9.71 *)/9.6exp(*9.71 *)/7.3exp(*10400OH CO H CO O CO COCC KRT CCRT CCRT dW dF 892320 (RWGS)K 298 4503402 2 2 2* *)/9.6exp(*9.71 *)/9.6exp(*9.71OH CO H CO OHCC KRT CCRT dW dF 2600212 2*)/7.3exp(*10400* 2 1O CO OCCRT dW dF 892320 (RWGS)K 298 4503402 2 2 2* *)/9.6exp(*9.71 *)/9.6exp(*9.71OH CO H CO HCC KRT CCRT dW dF

PAGE 80

68 Figure 4.6 Comparison of model results fo r the linearly independe nt non-elementary model. 4.11 Verification of FTIR E ffluent Concentrations via Gas Chromatography Although the FTIR is highly se nsitive to the concentratio ns of carbon monoxide, carbon dioxide, and water, there is an extreme limita tion to the technique. It is unable to measure diatomic molecules effl uent concentrations (i.e. H2, O2, and N2). So, concurrently with each of the FTIR efflue nt runs, a 1 mL effluent gas sample was extracted and compared to the FTIR results. The GC column used in all experiments was a Varian CP7534 plot fused silica 30 m x 0.32 mm ID coating Molsieve 5A (DF = 10 m). All GC experimental parameters are listed in Chapter 1.4.4. A representative image of a GC effluent signal is shown in Figure 4.7. Four peaks located at approximately 0.8, 1, 1.8, and 8 minutes correspond to the effluent concentrations of hydrogen, oxygen, nitr ogen and carbon monoxide respectively.

PAGE 81

69 Unfortunately, this column/detector combination was unable to detect either water or carbon dioxide for comparison. Similar to the FTIR, area integrals of each curve ar e linearly proportional the concentration of the species a nd allow for the calculation of conversion. All peak signals should be positive except for hydrogen. The reason that the hydrogen signal is pointing in the opposite direction than the other peaks is becaus e of the thermal conductivity detectors carrier gas being he lium. The GC mV response is directly proportional to the molecular species entering the detector relativ e to a carrier gas re ference. Oxygen, nitrogen, and carbon monoxide have a lower th ermal conductivity than the carrier gas and the detector signal represents this as a positive peak. Oppositely, hydrogen has a higher thermal conductivity than helium and graphs as a negative peak. Figure 4.7 Representative image of a GC response spectrum of the effluent gases which include hydrogen, oxygen, nitrogen, carbon monoxide, carbon dioxide, and water.

PAGE 82

70 I will begin the discussion of the GC results by examining the experimental integral data for hydrogen. Figure 4.8 shows the integral of the hydrogen signal becoming more negative as the flow rate of the 1%CO/H2 mixture is increased. This is reasonable since an increasingly negative value of the inte gral indicates an increasing amount of H2 in the effluent stream. Figure 4.9 graphs the integral of nitrogen in the effluent versus temperature and 1%CO/H2 flow rate. The nitrogen in the feed stream to this process is treated as an inert, and therefore it does not react with any species at these operating temperatures. The decreasing integral is indicat ive of the fact that the 10 sccm air flow rate was kept constant. With increasing CO/H2 flow rates, the fractional concentration of nitrogen decreased. Another observation is that as flow ra te increases, the standard deviation between integrals at each temperat ure decreases. This discrepancy between samples is mirrored in other calculated inte grals and is one rationale to install an autosampler to obtain more consistent results. Figure 4.8 Integral of gas chromatography hydr ogen spectrum data range versus influent flow rate at temper atures ranging from 25oC-125oC.

PAGE 83

71 Figure 4.9 Integral of gas chromatography ni trogen spectrum data range versus influent flow rate at temper atures ranging from 25oC-125oC. Figure 4.10 is one of the most interesti ng of the GC spectra. As shown, the oxygen concentrations plummet to almost zero in each of the sample runs. This indicates that the amount of oxygen in the influent is a limiting factor in the reaction kinetics and is completely depleted due to the carbon monoxide oxid ation and hydrogen oxidation reactions. Future work will include the eff ects of increasing the oxygen content. Care must be taken to avoid the potential fire and explosion hazard of having a hydrogen/oxygen mixture entering an exothermic reaction. Figure 4.11 represents the concentration of carbon monoxide in the effluent stream. The general trend of the data fo llows that found by the FTIR (Figure 4.4) with the largest conversions of carbon monoxide occurri ng at the lowest temperature, 25oC (Figure 4.12). Once again, the GC experimental integral results (Figure 4.11) are inc onclusive at the lowest flow rate. This is most likely due to the extremely low flow rate of the influent

PAGE 84

72 gases. Even though the influents were allowed to equilibrate over ~1 hour, this may not have been enough time. Another possibility is that since these inaccuracies did not appear as significantly in the FTIR spectra, the flow regime does not allow for adequate mixing and, thus, syringe extrac tion is not extremely accurate at low flow rate regimes. The conversion for 20 sccm was excluded due to experimental inaccuracies. The calculations for CO conversion are identical to th e FTIR model with the conversion being defined as: Conversion = (Bypass integral Experime ntal Integral)/Bypass Integral (4.20) One would expect to have a nice correlation between the resu lts of the FTIR and the GC; however, the GC conversion results (Figure 4.12) are much higher than the FTIR calculations although the trend is the same. However, with the sensitivity of the FTIR being much greater than the GC, the FTIR being plumbed directly, and the syringe transportation method to the GC introducing error into the pro cedure, the FTIR data is a more accurate representation of the effluent concentration.

PAGE 85

73 Figure 4.10 Integral of gas chromatography oxygen spectrum data range versus influent flow rate at temper atures ranging from 25oC-125oC. Figure 4.11 Integral of gas chromatography carbon monoxide spectrum data range versus influent flow rate at te mperatures ranging from 25oC-125oC.

PAGE 86

74 Figure 4.12 Conversion calcula tions of gas chromatogra phy carbon monoxide spectrum data range versus influent flow ra te at temperatures ranging from 25oC-125oC. Legend: 25oC(x), 50oC(o), 75oC( ), 100oC( ), 125oC( ) 4.12 Conclusions and Future Work Several elementary and non-elementary models were presented which describe the effluent CO fractional conversions of a PR OX reaction catalyzed by the World Gold Council nano-Au/TiO2 catalyst. Each model attempts to represent the effect of CO oxidation, hydrogen oxidation, a nd the water gas shift reacti on on the reaction effluent concentrations of a 1% CO mixture in ex cess hydrogen with a small amount of air. These models included a single CO reaction elementary model, a set of comprehensive elementary/non-elementary, and linearly inde pendent elementary/non-elementary kinetic models. Beginning with the calculations of Gibbs free energy for each reaction at each temperature, it was verified that the CO oxidation, H2 oxidation and WGS reactions are spontaneous at these conditions. The reverse-water-gas-shift reaction is not favorable at the conditions examined. The calculation of equilibrium cons tants verified the results.

PAGE 87

75 After examining the results of the predicte d models, none of the elementary models accurately depicted the CO oxidation experiment al results. The conclusions gained from these tests find that the PROX of CO is unlikely to be an elementary reaction and should not be modeled as such. The non-linear mode ls were better suite d to calculating the effluent CO fractional convers ions. Both the comprehensive and the linearly independent models accurately depicted the CO fractional conversion. Since the non-elementary model accurately approximated the CO fractional conversion, correlation with previo us studies are possible. Haruta discovered that the apparent activation energy of the carbon monoxide react ion approached zero as one approaches standard temperatures and pressures (25oC, 1 atm). While the empirically derived value in this model is not zero, the activatio n energy of the CO oxidation reaction is significantly lower than the other reactions, re sulting in less temperature dependence than all of the other reactions. Also, the oxyge n reaction order for CO oxidation in the comprehensive kinetic model falls within the range sited in other papers for CO oxidation via Au/TiO2. After the analysis of the fits to the FTIR data, comparisons were made to the simultaneous GC measurements. The genera l trends of the GC matched the data collected from the FTIR with the highest CO conversion occurring at the lowest temperature. One interesting observati on was the limiting amount of oxygen in the system. Another observ ation is that the GC conversion results are much higher than the FTIR calculations altho ugh the trend is the same Inconsistencies are most likely due to the syringe transportation me thod introducing erro r into the procedure and will be addressed in future experiments. Although, the experimental resu lts show that the most eff ective CO oxidation for this catalyst system occurs much lower temperatures than PEMFCs operate which would suggest that incorporation of this catalyst into a PEMFC anode for CO reduction would

PAGE 88

76 not be advantageous. A lower temperature pre-filter not inco rporated into the fuel cell anode would be a better solution to remove CO contamination from the influent stream. List of Variables for Chapter 4 = pre-exponential factor AREF = FTIR CO absorban ce integral (2145-2230 cm-1) of an unreacted feed stream ARXN = FTIR CO absorban ce integral (2145-2230 cm-1) after reaction CCO = concentration of carbon monoxide (mol/L) CCO2 = concentration of carbon dioxide (mol/L) CH2 = concentration of hydrogen (mol/L) CH2O = concentration of water (mol/L) CN2 = concentration of nitrogen (mol/L) CO2 = concentration of oxygen (mol/L) CO = carbon monoxide CO2 = carbon dioxide CTotal o = total concentration Cxx = concentration of a single species E = activation energy Exp = exponent FCO = flow rate of carbon monoxide (sccm) FCO2 = flow rate of carbon dioxide (sccm) FCO o = initial flow rate of carbon monoxide (sccm) FH2 = flow rate of hydrogen (sccm) FH2O = flow rate of water (sccm) FN2 = flow rate of nitrogen (sccm) FO2 = flow rate of oxygen (sccm) Fxx = flow rate of a single species (sccm) GNP = gold nanoparticle H2 = hydrogen H2O = water k = reaction rate constant N2 = nitrogen PEMFC = proton exchange membrane fuel cell PROX = preferential oxidation r = reaction rate W = weight of catalyst (gram) WGC = World Gold Council WGS = water-gas shift reaction which incl udes the reverse wate r gas shift reaction

PAGE 89

77 5 FUTURE WORK 5.1 Introduction In this study, two catalyst systems were ev aluated: the two-phas e-method GNPs provided by Dr. Gupta (Interfacial Phenomena and Poly meric Materials research group at USF) and the reference GNP/TiO2 purchased from the World Gold Council. Future work with the two-phase-method gold nanoparticles incl udes optimization of pre-treatment and fabrication to prevent nanoparticle agglomeratio n and to activate gold sites. The goals in this future work are to redu ce carbonization, improve CO tr ansport to the catalyst surface, and to increase the surface-gold concentra tion. Future work with the WGC Au/TiO2 catalyst includes the determination of moistu re effects on catalyst activity. This will necessitate the introduction of more complex models and analysis. 5.2 Two-Phase-Method Gold The proposed difficulties with the cataly tic activity of the two-phase-method GNP catalysts are the surfactant layer used to pr event agglomeration of the particles and the inability of the gold particles to form a cohe sive bond to the substrate. While these two problems are related, I will begin with the hypoth esized solutions to the surfactant layer. The initial solution to this problem was the evaporation of the solvent through the calcination step. This has been proven to be unsuccessful. The carbonization of the solvents used may be coating the gold partic les and preventing transport to the surface. While this may not be the ultim ate source of the inactivity of the catalyst, this is one aspect to consider. Another possible cause is that the concentration of gold nanopart icles is too low even in a 4% mixture. Although the initia l loading for the 100 mg of TiO2 is 4 mg which would lead to an ideal 4% mixture, there are inherent losses with the incipient wetness procedure. Difficulties in quantifying the conc entration after creation may have led to a

PAGE 90

78 misrepresentation final catalysts gold loading percentage. It has b een speculated by both David Walker and Dr. Gupta that the gold na noparticles may be attracted to the quartz containers and not sticking to the TiO2 support. However, the redesign of the experiment to include in situ mixing of TiO2 while the nanoparticles are being formed still did not lead to improved results. Incr easing the initial concentration of gold to an 8% initial may not be the most elegant of solutions, but it may overcom e the inherent losses of fabrication. The next iteration of tests wi ll include analysis of the pr e-treatment procedure. Though this process has been developed for similar GNP pre-treatment in other research, this catalyst may require a more specialized process. The high mobility of the GNP on silicon shown in Section 3.2.4 may be responsible for the low activity due to reduction of low coordinated gold active sites. The protocol for determini ng the activation of the TiO2 support Brust GNPs will be as follows: 1. Calcination/Oxidation experiments a. Create 100 mg of 8% Au/TiO2 catalyst i. Redesigned method to improve Au-TiO2 adhesion (Section 4.5) b. Load 100 mg sample into FTIR micro-reactor i. No pretreatment c. Start at 100oC in an oxygen environment i. Can vary time in oven from 30 minutes-2 hours d. FTIR effluent analysis to determine activity i. 20 sccm 1% CO/H2 ii. 10 sccm oxygen e. Repeat procedure for 150oC-600oC in 50oC increments to determine if calcination is advantageous i. This should verify the necessity of a slight oxidation to end the pre-treatment procedure

PAGE 91

79 2. Reduction Experiments a. Repeat 1.a-e but in a hydrogen rich environment to promote GNP and TiO2 reduction. 5.3 Modification of the WGC Effluent Model The most significant modification to the WGC model of the PROX reaction kinetics would be the inclusion of moisture effect s on the reaction kinetics. For all models discussed in Chapter 5, dry gases were used for the experiments and no attempt was made to humidify the influent. Therefore, the in fluent gas moisture c ontent was approximated to be 0%. This allowed the removal of th at parameter from the equation. This next set of experiments would necessita te several additions to the current setup, primarily a set of bubblers and relative humidit y sensors. A diagra m of the modified setup is shown in Figure 5.1 which shows a bubbler inline with the microreactor. A relative humidity gauge after th e bubbler will indicate the moisture content of the influent stream. The models will remain exactly th e same except for a slightly modified Arrhenius relationship for each of the speci es. Equation (3) in Chapter 5 will now become RT ET M f k exp ) ( (5.1) where f(M,T) is a function of moisture content and temperature.

PAGE 92

80 Figure 5.1 Modified micr oreactor setup which incl udes bubbler and relative humidity gauge for moisture content calculations. 5.4 Experimental Applications Experimental results from the World Gold C ouncil supported gold cata lyst indicate that their use in fuel cel l anodes may have limited results due to the pred ominance of the water gas shift reaction and hydrogen oxida tion at normal operating conditions (80oC, 1 atm). The GNPs will mostly be useful for a room temperature pre-filter for the influent gases. Having the temperature near room temp erature or lower shifts the selectivity of the catalyst towards the production of CO2 over H2O. H2, CO, Humid Air H2, H2O, CO, CO2, N 2 H2, H2O, CO, CO2, N 2 H2, CO, Dry Air Bubbler Bypass

PAGE 93

81 REFERENCES 1 C.H. Bartholomew and R.J. Farrauto, Fundamentals of Indu strial Catalytic Processes Second ed. (John Wiley & Sons, Inc., 2006). 2 R.I. Masel, Chemical Kinetics and Catlaysis (John Wiley & Sons, Inc., 2001). 3 A. Cho, Science 299(14 March 2003), 1684-1685 (2003). 4 M.B. Cortie and E.v.d. Lingen, Materials Forum 26, 1-14 (2002). 5 M. Haruta, Journal of New Materi als for Electrochemical Systems 7, 163-172 (2004). 6 P. Landon, J. Ferguson, B.E. Solsona, T. Ga rcia, A.F. Carley, A.A. Herzing, C.J. Kiely, S.E. Golunskic, and G.J. Hutchings, Chemical Communications, 33853387 (2005). 7 D. Cameron, R. Holliday, and D. Thompson, Journal of Power Sources 118, 298303 (2003). 8 Y. Choi and H.G. Stenger, Journal of Power Sources 129, 246-254 (2004). 9 http://www1.eere.energy. gov/hydrogenandfuelcells/mypp/ "3.4 Fuel Cells", 2006. 10 R. O'Hayre, S.-W. Cha, W. Colella, and F.B. Prinz, Fuel Cell Fundamentals (John Wiley & Sons, Inc. 2006). 11 S. Schimpf, M. Lucas, C. Mohra, U. R odemerck, A. Bruckner, J. Radnik, H. Hofmeister, and P. Claus, Catalysis Today 2592, 1-16 (2002). 12 G. Schmid and B. Corain, European Journal of Inorganic Chemistry, 3081-3098 (2003). 13 http://www.gold.org/discover/sci_indu/gold_catalysts/refcat.html 14 Metals Handbook Vol. 10, 9th ed. (1986).

PAGE 94

82 15 N.A. Hodgea, C.J. Kiely, R. Whyman, M.R.H. Siddiqui, G.J. Hutchings, Q.A. Pankhurst, F.E. Wagner, R.R. Rajaram, and S.E. Golunski, Catalysis Today 72, 133-144 (2002). 16 M. Okumura, S. Nakamura, S. Tsubota, T. Nakamura, M. Azuma, and M. Haruta, Catalysis Letters 51, 53-58 (1998). 17 B.D. Cullity, Elements of X-ray Diffraction, 2nd edition (Addison-Wesley Publishing Company, Inc., 1978). 18 B.C. Smith, Fundamentals of Fourier transform infrared spectroscopy (CRC Press, Boca Raton, 1996). 19 Win-IR Pro online help, "Q uant Analysis error s ources"(Version 3.4.2.025). 20 H. Phan, in Fundamental Infrared Spectroscopy (http://www.midac.com/apnotes/Tn-100.PDF) 21 B. Stuart, Infrared Spectroscopy: Fu ndamentals and Applications (John Wiley & Sons, Ltd, 2004). 22 T. Venkov, K. Fajerwerg, L. Delannoy, H. Klimev, K. Hadjiivanov, and C. Louis, Applied Catalysis A: General 301, 106-114 (2006). 23 V. Rossiter, Research & Development(February 1988), 94-97 (February 1988). 24 Agilent Technologies 6890N Gas Chro matograph User Information Manual May (2001). 25 http://www.impactanalytical.com/tga.html 26 O. Levenspiel, Chemical Reaction Engineering Third ed. (John Wiley & Sons, Inc., 1999). 27 M. Mavrikakis, P. Stoltze, and J.K. Nrskov, Catalysis Letters 64, 101-106 (2000). 28 Z.-P. Liu, S.J. Jenkins, and D.A. King, Physical Review Letters 93(No. 15), 156102(4) (2004). 29 M. Haruta, The Chemical Record 3, 75-87 (2003). 30 M.M. Schubert, S. Hackenberg, A.C.v. Veen, M. Muhler, V. Plzak, and R.J. Behm, Journal of Catalysis 197, 113-122 (2001). 31 M. Haruta, Gold Bulletin 37(1-2), 27-36 (2004).

PAGE 95

8332 S.K. Shaikhutdinov, R. Meye r, M. Naschitzki, M. Baumer, and H.-J. Freund, Catalysis Letters 86(No. 4, March), 211-219 (2003). 33 D.T. Thompson, Topics in Catalysis 38(4), 231-240 (2006). 34 N.M. Gupta and A.K. Tr ipathi, Gold Bulletin 34(4), 120-128 (2001). 35 J. Guzman and B.C. Gates, Journal of the American Chemical Society 126, 26722673 (2004). 36 M. Valden, X. Lai, and D.W. Goodman, Science 281, 1647-1650 (11 September 1998). 37 T.V. Choudhary, C. Sivadina rayana, C.C. Chusuei, A.K. Datye, J. J.P.Fackler, and D.W.Goodman, Journal of Catalysis 207, 247-255 (2002). 38 M. Brust, D. Bethell, D.J. Schiffrin and C.J. Kiely, Advanced Materials 7, 795 (1995). 39 C. Demaille, M. Brust, M. Tsionsky, and A.J. Bard, Analytical Chemistry 69, 2323-2328 (1997). 40 http://www.eng.usf.edu/~vkgupta/ 41 I. Khramtsov, G.B. MacDonald, Z. Fakhraa i, J.H. Teichroeb, and J.A. Forrest, (Department of Physics and Guelph-Waterl oo Physics Institute, University of Waterloo, Canada, ON). 42 B.C. Beard, H.W. Sandusky, B.C. Glan cy, and W.L. Elban, Surface and Interface Analysis 20(2), 140-148 (1993). 43 L. Maya, M. Paranthaman, T. Thundat, and M.L. Bauer, Journal of Vacuum Science and Technology B 14(1), 15-21 (1996). 44 M. Grutter, Atmosfera 16, 1-13 (2003). 45 Z. Yang, R. Wu, Q. Zhang, and D. W. Goodman, Physical Review B 63, 045419(6) (2001). 46 D.C. Sorescu and J.T. Yates, Jr., Journal of Physical Chemistry B 102, 4556-4565 (1998). 47 Z. Dohnalek, J. Kim, O. Bondarchuk, J. M. White, and B.D. Kay, Journal of Physical Chemistry B 110, 6229-6235 (2006). 48 K. Hadjiivanov, J. Lamotte, and J.-C. Lavalley, Langmuir 13, 3374-3381 (1997).

PAGE 96

8449 T. Becker, C. Boas, U. Burghaus, and C. Woll, Physical Review B 61(7), 45384541 (2000). 50 D.E. De Vos and B.F. Sels, Angewandte Chemie International Edition 44, 30-32 (2005). 51 S. Tsubota, A. Yamaguchi, M. Date, and M. Haruta, http://www.gold.org/discover/ sci_indu/gold2003/pdf/s36a1383p1030.pdf 52 M. Dat and M. Haruta, Journal of Catalysis 201, 221-224 (2001). 53 N. Edwards, S.R. Ellis, J.C. Frost, S.E. Golunski, A.N.J. van Keulan, N.G. Lindewald, and J.G. Reinkingh, Journal of Power Sources 71, 123-128 (1998). 54 J.M. Smith, H.C. Van Ness, and M.M. Abbott, Introduction to Chemical Engineering Thermodynamics Sixth ed. (McGraw-Hill Companies, Inc., New York, NY, 2001). 55 http://www.gold.org/discover/ sciindu/gold2003/pdf/s36a1383p1030.pdf 56 H.S. Fogler, Elements of Chemical Reaction Engineering Fourth ed. (Prentice Hall Professional Technical Reference, 2006), p. 327. 57 M. Dat, M. Okumura, S. Tsubota, and M. Haruta, Angewandte Chemie International Edition 43, 2129-2132 (2004).

PAGE 97

85 APPENDICES

PAGE 98

86 Appendix A Matlab Code for FTIR Modeling A.1 Elementary Model without WGS function ElementaryModel_without_WGS clc; close all; clear all % Model of the reaction of the WGC and the Brust Gold % CO + 1/2O2 = CO2 (Equation 1) % A + 1/2B = C (Equation 1.a) % Temperature T = [25*ones(1,5) 50*ones(1,5) 75*ones( 1,5) 100*ones(1,5) 125*ones(1,5)]; R = 0.0821; % Gas Constant (L*atm/mole*K) P = 1; % Pressure W = 0.1; % grams of catalyst % Flow rates of the 1% mixtur e of CO in balance H2(sccm) COH2 = [20 40 60 80 100 20 40 60 80 100 20 40 60 80 100 20 40 60 80 100 20 40 60 80 100]; % Amount of CO in the mixture(sccm) CO = 0.01*COH2; % Flow rates of Air(sccm) Air = 10*ones(1,25); % Amount of O2 in the Air(sccm) O2 = 0.21*Air; %Conversion conversionCO = [0.5583 0.4195 0.3694 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.3533 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.2837 0.1576 0.1439 0.1229 0.1165]; yao = (CO./(COH2+Air)); % mole fraction of carbon monoxide ybo = (O2./(COH2+Air)); % mole fraction of oxygen Cao = yao*P./R.*T; % Concentration A initially = moles/liter Cbo = ybo*P./R.*T; % Concentration B initially moles/liter Fao = Cao.*CO*1/1000*1/60; % Initial molar flowrate of A (moles/sec) Fbo = Cbo.*O2*1/1000*1/60; % In itial molar flowrate of B (moles/sec) % theta_b = Fbo/Fao theta_b = Fbo./Fao;

PAGE 99

Appendix A (Continued) 87 % epsilon = (c/a b/a a/a)*Va/Vt epsilon = (1/1 (1/2)/1 -1/1)*(CO./(COH2+Air)); % ratio b to a; coefficients of CO and O2 in Equation 1.a ratio_ba = (1/2)/1; % % PBR Design Equation % % Fao*dX/dW = -ra % -ra = k*Ca*Cb^1/2 % Ca = Cao(1-X)/(1 + epsilon*X) % Cb = Cbo(theta_b rati o_ba*X)/(1 + epsilon*X) % -ra = k*Ca*Cb^1/2 % k = alpha*exp(-E/RT) % Fao*dX/dW = alpha*exp (-E/RT))*Ca*Cb^1/2 % (Fao/alpha*exp(-E/RT))*Ca*Cb^1/2)dX = dW % (Fao/alpha*exp(-E/RT))*Integral of (Ca*Cb^1/2) = W % (Fao/W)*Integral = alpha*exp(-E/RT) % ln((Fao/W)*Integral = ln alpha E/R*1/T % Therefore a plot of % ln((Fao/(W*Integral of (Ca*Cb^1/2)) ) vs (1/T) should give me activation % energy. From there I can get the preexponential factor. Then, calculate % k and have a model of the ac tivity of the oxidation reaction. % Experimentally determine conversion for n = 1:length(T) whos X = linspace(0, conversionCO(n)); Ca(n,:) = Cao(n).* (1-X)./(1+epsilon(n).*X); Cb(n,:) = Cao(n).*(theta_b(n) ratio_ba*X)./(1+epsilon(n).*X); Y = 1./(Ca(n,:).*(Cb(n,:).^(1/2))); Integral(n) = Simpson(X,Y) Answer(n) = log((Fao(n)/W)*Integral(n)); end hold on

PAGE 100

Appendix A (Continued) 88 plot(1./T(1:5), Answer(1:5), 'ko') plot(1./T(6:10), Answer(6:10), 'ko') plot(1./T(11:15), An swer(11:15), 'ko') plot(1./T(16:20), An swer(16:20), 'ko') plot(1./T(21:25), An swer(21:25), 'ko') % % poly = polyfit(1./T, Answer, 1) % hold on % plot(linspace(1./T(1), 1./T(e nd)), poly(1).*linspace(1./T(1), 1./T(end))+ poly(2), 'k:') xlabel('1/Temperature (K^{-1})') ylabel('ln((Fao/W)*Integr al of (Ca*Cb^1/2)))') A.2 Comprehensive Elementary Model Fit Routine function Final = Nonlinear_ leastsquares_fit_WGC_vers ion3_Ele(Initialguesses) % Edit % I removed some of the erra nt points to get a better fit close all; clc % conversionCO = [0.5583 0.4195 0.3694 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.3533 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.2837 0.1576 0.1439 0.1229 0.1165]; conversionCO = [0.5583 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.1576 0.1439 0.1229 0.1165]; Temps = [25 50 75 100 125]; Air = [10 10 10 10 10]; COH2 = [20 40 60 80 100]; data = [Temps; Air; COH2]; [Final,RESNORM,RESIDUAL,EXITFLAG,OUTPUT,LAMBDA,JACOBIAN] = lsqcurvefit(@myfun, Initialguesses, data, convers ionCO, [0 0 0 0 0 0 0 0], [+inf +inf +inf +inf 50000 50000 50000 50000]) function Conversionresults = myfun(VAR, data); Temps = data(1,:); Air = data(2,:); COH2 = data(3,:);

PAGE 101

Appendix A (Continued) 89 % Conversionresults = NewmodelwithWGS_ve rsion2(Temperature, alpha, E, [CO O2 CO2 H2 H2O N2]) Conversionresults(1) = NewmodelwithWG S_version2(Temps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*Air(1)], [1 0.5 1 0.5 1 1 1 1]); % 10 sccm Air 20 sccm COH2 % Conversionresults(2) = NewmodelwithWGS_version2(T emps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VA R(8)], [0.01*COH2(2) 0.21*Air(2) 0 0.99*COH2(2) 0 0.79*Air(2)]); % 10 sccm Air 40 sccm COH2 % Conversionresults(3) = NewmodelwithWGS_version2(T emps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VA R(8)], [0.01*COH2(3) 0.21*Air(3) 0 0.99*COH2(3) 0 0.79*Air(3)]); % 10 sccm Air 60 sccm COH2 Conversionresults(2) = NewmodelwithWG S_version2(Temps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*Air(4)], [1 0.5 1 0.5 1 1 1 1]); % 10 sccm Air 80 sccm COH2 Conversionresults(3) = NewmodelwithWG S_version2(Temps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*Air(5)], [1 0.5 1 0.5 1 1 1 1]); % 10 sccm Air 100 sccm COH2 Conversionresults(4) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 20 sccm COH2 Conversionresults(5) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 40 sccm COH2 Conversionresults(6) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 60 sccm COH2 Conversionresults(7) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 80 sccm COH2 % Conversionresults(10) = Newmodelwith WGS_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10]); % 10 sccm Air 100 sccm COH2 Conversionresults(8) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 20 sccm COH2 Conversionresults(9) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 40 sccm COH2 Conversionresults(10) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 60 sccm COH2

PAGE 102

Appendix A (Continued) 90 Conversionresults(11) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 80 sccm COH2 Conversionresults(12) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 100 sccm COH2 Conversionresults(13) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 20 sccm COH2 Conversionresults(14) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 40 sccm COH2 Conversionresults(15) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 60 sccm COH2 Conversionresults(16) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 80 sccm COH2 Conversionresults(17) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 100 sccm COH2 % Conversionresults(21) = Newmodelwith WGS_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10]); % 10 sccm Air 20 sccm COH2 Conversionresults(18) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 40 sccm COH2 Conversionresults(19) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 60 sccm COH2 Conversionresults(20) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 80 sccm COH2 Conversionresults(21) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [1 0.5 1 0.5 1 1 1 1 ]); % 10 sccm Air 100 sccm COH2 A.3 Comprehensive Non-Elementary Model Fit Routine function Final = Nonlinear_ leastsquares_fit_WGC_vers ion3_Non(Initialguesses)

PAGE 103

Appendix A (Continued) 91 % Edit % I removed some of the erra nt points to get a better fit close all; clc % conversionCO = [0.5583 0.4195 0.3694 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.3533 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.2837 0.1576 0.1439 0.1229 0.1165]; conversionCO = [0.5583 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.1576 0.1439 0.1229 0.1165]; Temps = [25 50 75 100 125]; Air = [10 10 10 10 10]; COH2 = [20 40 60 80 100]; data = [Temps; Air; COH2]; [Final,RESNORM,RESIDUAL,EXITFLAG,OUTPUT,LAMBDA,JACOBIAN] = lsqcurvefit(@myfun, Initialguesses, data, conversi onCO, [0 0 0 0 0 0 0 0 -2 -2 -2 -2 -2 -2 -2 -2], [+inf +inf +inf +inf 50000 50000 50000 50000 2 2 2 2 2 2 2 2]) function Conversionresults = myfun(VAR, data); Temps = data(1,:); Air = data(2,:); COH2 = data(3,:); % Conversionresults = NewmodelwithWGS_ve rsion2(Temperature, alpha, E, [CO O2 CO2 H2 H2O N2]) Conversionresults(1) = NewmodelwithWG S_version2(Temps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*Air(1)], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 20 sccm COH2 % Conversionresults(2) = NewmodelwithWGS_version2(T emps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VA R(8)], [0.01*COH2(2) 0.21*Air(2) 0 0.99*COH2(2) 0 0.79*Air(2)]); % 10 sccm Air 40 sccm COH2 % Conversionresults(3) = NewmodelwithWGS_version2(T emps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VA R(8)], [0.01*COH2(3) 0.21*Air(3) 0 0.99*COH2(3) 0 0.79*Air(3)]); % 10 sccm Air 60 sccm COH2 Conversionresults(2) = NewmodelwithWG S_version2(Temps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*Air(4)], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 80 sccm COH2

PAGE 104

Appendix A (Continued) 92 Conversionresults(3) = NewmodelwithWG S_version2(Temps(1), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*Air(5)], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 100 sccm COH2 Conversionresults(4) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 20 sccm COH2 Conversionresults(5) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 40 sccm COH2 Conversionresults(6) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 60 sccm COH2 Conversionresults(7) = NewmodelwithWG S_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 80 sccm COH2 % Conversionresults(10) = Newmodelwith WGS_version2(Temps(2), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10]); % 10 sccm Air 100 sccm COH2 Conversionresults(8) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 20 sccm COH2 Conversionresults(9) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 40 sccm COH2 Conversionresults(10) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 60 sccm COH2 Conversionresults(11) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 80 sccm COH2 Conversionresults(12) = NewmodelwithWG S_version2(Temps(3), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0

PAGE 105

Appendix A (Continued) 93 0.99*COH2(5) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 100 sccm COH2 Conversionresults(13) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 20 sccm COH2 Conversionresults(14) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 40 sccm COH2 Conversionresults(15) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 60 sccm COH2 Conversionresults(16) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 80 sccm COH2 Conversionresults(17) = NewmodelwithWG S_version2(Temps(4), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 100 sccm COH2 % Conversionresults(21) = Newmodelwith WGS_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10]); % 10 sccm Air 20 sccm COH2 Conversionresults(18) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 40 sccm COH2 Conversionresults(19) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 60 sccm COH2 Conversionresults(20) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 80 sccm COH2 Conversionresults(21) = NewmodelwithWG S_version2(Temps(5), [VAR(1) VAR(2) VAR(3) VAR(4)], [VAR(5) VAR(6) VAR(7) VAR(8)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [VAR(9) VAR(10) VAR(11) VAR(12) VAR(13) VAR(14) VAR(15) VAR(16)]); % 10 sccm Air 100 sccm COH2

PAGE 106

Appendix A (Continued) 94 A.4 General Comprehensive Model function ConversionresultsCO = Newmodelwith WGS_version2(T, alpha, E, Flowrates, Exponents) % NewmodelwithWGS(T, alpha, E, Flowrates, Exponents) % Temperature(C), alpha, E, Flowrates(sccm), Exponents( CO, O2, CO2, H2, CO, H2O) % Information needed, % Temperature, reaction rate constants( alphas and E's), Flow rates of gases % % Information attained % Approximate conversions for all species clc; % New thought process % CO + 1/2O2 = CO2 (1) % H2 + 1/2O2 = H2O (2) % CO2 + H2 = CO + H2O (3) % CO + H2O = CO2 + H2 (4) T = T + 273.15; % Temperature (K) R = 8.314; % R = J/mol*K k(1) = alpha(1)*exp(-E(1)/(R*T)); k(2) = alpha(2)*exp(-E(2)/(R*T)); k(3) = alpha(3)*exp(-E(3)/(R*T)); k(4) = alpha(4)*exp(-E(4)/(R*T)); % Fo = [1 2 0 99 0 8]'; % Initial Flowrates (sccm) Fo = Flowrates./600; % In itial Flowrates (L/s) CTo = 1/(0.0821*T); % CTo = Po/(R*To) % mol/L where R in this case is (L*atm/mol K) weight_int = 0; % Weight (g) weight_final = 0.1; % Weight (g) [W, C_rk1] = RK(@funct, wei ght_int, weight_final weight_final/1000, Fo, 4, k, CTo, Exponents); % Conversionresults = (C_rk1(:,1) C_rk1(:,end ))./C_rk1(:,1); Conversionresults = (C_rk1(1,1) C_rk1(1,end))./C_rk1(1,1) ; % Changed "all rows" to one to speed up calculations C onversionresults = (C_rk1(:,1) C_rk1(:,end)). /C_rk1(:,1); ConversionresultsCO = Conversionresults(1);

PAGE 107

Appendix A (Continued) 95 function dCs = funct(W, C, k, CTo, Exponents) FCO = C(1); FO2 = C(2); FCO2 = C(3); FH2 = C(4); FH2O = C(5); FN2 = C(6); FT = FCO + FO2 + FCO2 + FH2 + FH2O + FN2; CCO = CTo*FCO/FT; CO2 = CTo*FO2/FT; CCO2 = CTo*FCO2/FT; CH2 = CTo*FH2/FT; CH2O = CTo*FH2O/FT; r1CO = -k(1)*CCO^Exponent s(1)*CO2^(Exponents(2)); r1O2 = 1/2*r1CO; r1CO2 = -r1CO; r2H2 = -k(2)*CH2^Exponent s(3)*CO2^(Exponents(4)); r2O2 = 1/2*r2H2; r2H2O = -r2H2; r3H2 = -k(3)*CCO2^Exponent s(5)*CH2^Exponents(6); r3CO2 = r3H2; r3CO = -r3H2; r3H2O = -r3H2; r4CO = -k(4)*CCO^Exponent s(7)*CH2O^Exponents(8); r4H2O = r4CO; r4CO2 = -r4CO; r4H2 = -r4CO; rCO = r1CO + r3CO + r4CO; rO2 = r1O2 + r2O2; rCO2 = r1CO2 + r3CO2 + r4CO2; rH2 = r2H2 + r3H2 + r4H2; rH2O = r2H2O + r3H2O + r4H2O; dFCO = rCO;

PAGE 108

Appendix A (Continued) 96 dFO2 = rO2; dFCO2 = rCO2; dFH2 = rH2; dFH2O = rH2O; dFN2 = 0; dCs = [dFCO dFO2 dFCO2 dFH2 dFH2O dFN2]'; A.5 Independent Elementary Model Fit Routine function Final = IndependentReact ionModelFit_Non(Initialguesses) % Edit % I removed some of the erra nt points to get a better fit close all; clc; % conversionCO = [0.5583 0.4195 0.3694 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.3533 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.2837 0.1576 0.1439 0.1229 0.1165]; conversionCO = [0.5583 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.1576 0.1439 0.1229 0.1165]; Temps = [25 50 75 100 125]; Keq = [104000 28800 9620 3730 1640]; Air = [10 10 10 10 10]; COH2 = [20 40 60 80 100]; data = [Temps; Keq; Air; COH2]; [Final,RESNORM,RESIDUAL,EXITFLAG,OUTPUT,LAMBDA,JACOBIAN] = lsqcurvefit(@myfun, Initialguesses, data, conversionCO, [0 0 0 0], [+inf +inf 50000 50000]) function Conversionresults = myfun(VAR, data); Temps = data(1,:); Keq = data(2,:); Air = data(3,:); COH2 = data(4,:); % Conversionresults = IndependentReactionM odel_3reaction(Temperature, alpha, E, [CO O2 CO2 H2 H2O N2])

PAGE 109

Appendix A (Continued) 97 Conversionresults(1) = I ndependentReactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 20 sccm COH2 % Conversionresults(2) = IndependentR eactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10]); % 10 sccm Air 40 sccm COH2 % Conversionresults(3) = IndependentR eactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10]); % 10 sccm Air 60 sccm COH2 Conversionresults(2) = I ndependentReactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 80 sccm COH2 Conversionresults(3) = I ndependentReactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 100 sccm COH2 Conversionresults(4) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 20 sccm COH2 Conversionresults(5) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 40 sccm COH2 Conversionresults(6) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 60 sccm COH2 Conversionresults(7) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 80 sccm COH2 % Conversionresults(10) = IndependentR eactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10]); % 10 sccm Air 100 sccm COH2 Conversionresults(8) = I ndependentReactionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 20 sccm COH2 Conversionresults(9) = I ndependentReactionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 40 sccm COH2 Conversionresults(10) = IndependentReac tionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 60 sccm COH2

PAGE 110

Appendix A (Continued) 98 Conversionresults(11) = IndependentReac tionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 80 sccm COH2 Conversionresults(12) = IndependentReac tionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 100 sccm COH2 Conversionresults(13) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 20 sccm COH2 Conversionresults(14) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 40 sccm COH2 Conversionresults(15) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 60 sccm COH2 Conversionresults(16) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 80 sccm COH2 Conversionresults(17) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 100 sccm COH2 % Conversionresults(21) = IndependentR eactionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10]); % 10 sccm Air 20 sccm COH2 Conversionresults(18) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 40 sccm COH2 Conversionresults(19) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 60 sccm COH2 Conversionresults(20) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 80 sccm COH2 Conversionresults(21) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [1 0.5 1 1 1 1]); % 10 sccm Air 100 sccm COH2 A.6 Independent Non-Elementary Model Fit Routine function Final = IndependentReact ionModelFit_Non(Initialguesses)

PAGE 111

Appendix A (Continued) 99 % Edit % I removed some of the erra nt points to get a better fit close all; clc; % conversionCO = [0.5583 0.4195 0.3694 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.3533 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.2837 0.1576 0.1439 0.1229 0.1165]; conversionCO = [0.5583 0.3932 0.3696 0.4674 0.3999 0.3545 0.3361 0.4173 0.3181 0.2714 0.2414 0.2388 0.3318 0.2410 0.1982 0.1697 0.1667 0.1576 0.1439 0.1229 0.1165]; Temps = [25 50 75 100 125]; Keq = [104000 28800 9620 3730 1640]; Air = [10 10 10 10 10]; COH2 = [20 40 60 80 100]; data = [Temps; Keq; Air; COH2]; % [Final,RESNORM,RESIDUAL,EXITFLAG,OUTPUT,LAMBDA,JACOBIAN] = lsqcurvefit(@myfun, Initialguesses, data, conversi onCO, [0 0 0 0 -2 -2 -2 -2 -2 -2], [+inf +inf 50000 50000 2 2 2 2 2 2]) [Final,RESNORM,RESIDUAL,EXITFLAG,OUTPUT,LAMBDA,JACOBIAN] = lsqcurvefit(@myfun, Initialguesses, data, convers ionCO, [0 0 0 0 0 0 0 0 0 0], [+inf +inf +inf +inf 3 3 3 3 3 3]) function Conversionresults = myfun(VAR, data); Temps = data(1,:); Keq = data(2,:); Air = data(3,:); COH2 = data(4,:); % Conversionresults = IndependentReactionM odel_3reaction(Temperature, alpha, E, [CO O2 CO2 H2 H2O N2]) Conversionresults(1) = I ndependentReactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 20 sccm COH2 % Conversionresults(2) = IndependentR eactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10]); % 10 sccm Air 40 sccm COH2 % Conversionresults(3) = IndependentR eactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10]); % 10 sccm Air 60 sccm COH2

PAGE 112

Appendix A (Continued) 100 Conversionresults(2) = I ndependentReactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 80 sccm COH2 Conversionresults(3) = I ndependentReactionModel_3reaction(Temps(1), Keq(1), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 100 sccm COH2 Conversionresults(4) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 20 sccm COH2 Conversionresults(5) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 40 sccm COH2 Conversionresults(6) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 60 sccm COH2 Conversionresults(7) = I ndependentReactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 80 sccm COH2 % Conversionresults(10) = IndependentR eactionModel_3reaction(Temps(2), Keq(2), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0.01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10]); % 10 sccm Air 100 sccm COH2 Conversionresults(8) = I ndependentReactionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 20 sccm COH2 Conversionresults(9) = I ndependentReactionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 40 sccm COH2 Conversionresults(10) = IndependentReac tionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 60 sccm COH2 Conversionresults(11) = IndependentReac tionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0

PAGE 113

Appendix A (Continued) 101 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 80 sccm COH2 Conversionresults(12) = IndependentReac tionModel_3reaction(Temps(3), Keq(3), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 100 sccm COH2 Conversionresults(13) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 20 sccm COH2 Conversionresults(14) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 40 sccm COH2 Conversionresults(15) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 60 sccm COH2 Conversionresults(16) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 80 sccm COH2 Conversionresults(17) = IndependentReac tionModel_3reaction(Temps(4), Keq(4), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 100 sccm COH2 % Conversionresults(21) = IndependentR eactionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(1) 0.21*10 0 0.99*COH2(1) 0 0.79*10]); % 10 sccm Air 20 sccm COH2 Conversionresults(18) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(2) 0.21*10 0 0.99*COH2(2) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 40 sccm COH2 Conversionresults(19) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(3) 0.21*10 0 0.99*COH2(3) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 60 sccm COH2 Conversionresults(20) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(4) 0.21*10 0 0.99*COH2(4) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 80 sccm COH2

PAGE 114

Appendix A (Continued) 102 Conversionresults(21) = IndependentReac tionModel_3reaction(Temps(5), Keq(5), [VAR(1) VAR(2)], [VAR(3) VAR(4)], [0 .01*COH2(5) 0.21*10 0 0.99*COH2(5) 0 0.79*10], [VAR(5) VAR(6) VAR(7) VAR(8) VAR(9 ) VAR(10)]); % 10 sccm Air 100 sccm COH2 A.7 General Independent Model function ConversionresultsCO = Independent ReactionModel_3reaction(T, Keq, alpha, E, Flowrates, Exponents) clc; % CO + 1/2O2 = CO2 (1) % CO2 + H2 = CO + H2O (2) % CO + H2O = CO2 + H2 (3) T = T + 273.15; % Temperature (K) R = 8.314; % R = J/mol*K k(1) = alpha(1)*exp(-E(1)/(R*T)); k(2) = alpha(2)*exp(-E(2)/(R*T)); k(3) = k(2)*Keq; % Fo = [CO, O2, CO2, H2, CO, H2O] '; % Initial Flowrates(sccm) Fo = Flowrates./600; % In itial Flowrates (L/s) CTo = 1/(0.0821*T); % CTo = Po/(R*To) % mol/L where R in this case is (L*atm/mol K) weight_int = 0; % Weight (g) weight_final = 0.1; % Weight (g) [W, C_rk1] = RK(@funct, wei ght_int, weight_final weight_final/1000, Fo, 4, k, CTo, Exponents); % Conversionresults = (C_rk1(:,1) C_rk1(:,end ))./C_rk1(:,1); Conversionresults = (C_rk1(1,1) C_rk1(1,end))./C_rk1(1,1) ; % Changed "all rows" to one to speed up calculations C onversionresults = (C_rk1(:,1) C_rk1(:,end)). /C_rk1(:,1); ConversionresultsCO = Conversionresults(1); function dCs = funct(W, C, k, CTo, Exponents) FCO = C(1); FO2 = C(2);

PAGE 115

Appendix A (Continued) 103 FCO2 = C(3); FH2 = C(4); FH2O = C(5); FN2 = C(6); FT = FCO + FO2 + FCO2 + FH2 + FH2O + FN2; CCO = CTo*FCO/FT; CO2 = CTo*FO2/FT; CCO2 = CTo*FCO2/FT; CH2 = CTo*FH2/FT; CH2O = CTo*FH2O/FT; r1CO = -k(1)*CCO^Exponent s(1)*CO2^(Exponents(2)); r1O2 = 1/2*r1CO; r1CO2 = -r1CO; r2H2 = -k(2)*CCO2^Exponent s(3)*CH2^Exponents(4); r2CO2 = r2H2; r2CO = -r2H2; r2H2O = -r2H2; r3CO = -k(3)*CCO^Exponent s(5)*CH2O^Exponents(6); r3H2O = r3CO; r3CO2 = -r3CO; r3H2 = -r3CO; rCO = r1CO + r2CO + r3CO; rO2 = r1O2; rCO2 = r1CO2 + r2CO2 + r3CO2; rH2 = r2H2 + r3H2; rH2O = r2H2O + r3H2O; dFCO = rCO; dFO2 = rO2; dFCO2 = rCO2; dFH2 = rH2; dFH2O = rH2O; dFN2 = 0; dCs = [dFCO dFO2 dFCO2 dFH2 dFH2O dFN2]';

PAGE 116

Appendix A (Continued) 104 A.8 Error Calculations for WGC Data % clc; close al l; clear all; load FTIRdata_vars clc; close all; [Xrange, COrange, CO2range, H2Orange] = define ranges(background_3.data(:,1)); % Extra function defined elsewhere % Full FTIR signal of flow rates bypass_20rat1 = -(log10(bypass20_1.data(:,2) ./background_1.data(:,2))); % Determine absorbance bypass_20rat2 = -(log10(bypass20_2.data(:,2) ./background_1.data(:,2))); % Determine absorbance bypass_40rat1 = -(log10(bypass40_1.dat a(:,2)./background_1.data(:,2))); bypass_40rat2 = -(log10(bypass40_2.dat a(:,2)./background_1.data(:,2))); bypass_40rat3 = -(log10(bypass40_3.dat a(:,2)./background_1.data(:,2))); bypass_40rat4 = -(log10(bypass40_4.dat a(:,2)./background_1.data(:,2))); bypass_60rat1 = -(log10(bypass60_1.dat a(:,2)./background_1.data(:,2))); bypass_60rat2 = -(log10(bypass60_2.dat a(:,2)./background_1.data(:,2))); bypass_60rat3 = -(log10(bypass60_3.dat a(:,2)./background_1.data(:,2))); bypass_60rat4 = -(log10(bypass60_4.dat a(:,2)./background_1.data(:,2))); bypass_80rat1 = -(log10(bypass80_1.dat a(:,2)./background_1.data(:,2))); bypass_80rat2 = -(log10(bypass80_2.dat a(:,2)./background_1.data(:,2))); bypass_80rat3 = -(log10(bypass80_3.dat a(:,2)./background_1.data(:,2))); bypass_80rat4 = -(log10(bypass80_4.dat a(:,2)./background_1.data(:,2))); bypass_100rat1 = -(log10(bypass100_1.da ta(:,2)./background_1.data(:,2))); bypass_100rat2 = -(log10(bypass100_2.data (:,2)./background_1.data(:,2))); bypass_100rat3 = -(log10(bypass100_3.da ta(:,2)./background_1.data(:,2))); bypass_100rat4 = -(log10(bypass100_4.da ta(:,2)./background_1.data(:,2))); Integralbypass_201 = integratedata(bypass20_1.data(:,1), bypass_20rat1, COrange); % Extra function defined elsewhere Integralbypass_202 = integratedata(bypass20_2.data(:,1), bypass_20rat2, COrange); % Extra function defined elsewhere Integralbypass_401 = integratedata(bypass40_1.data(:,1), bypass_40rat1, COrange); % Extra function defined elsewhere Integralbypass_402 = integratedata(bypass40_2.data(:,1), bypass_40rat2, COrange); % Extra function defined elsewhere Integralbypass_403 = integratedata(bypass40_3.data(:,1), bypass_40rat3, COrange); % Extra function defined elsewhere

PAGE 117

Appendix A (Continued) 105 Integralbypass_404 = integratedata(bypass40_4.data(:,1), bypass_40rat4, COrange); % Extra function defined elsewhere Integralbypass_601 = integratedata(bypass60_1.data(:,1), bypass_60rat1, COrange); % Extra function defined elsewhere Integralbypass_602 = integratedata(bypass60_2.data(:,1), bypass_60rat2, COrange); % Extra function defined elsewhere Integralbypass_603 = integratedata(bypass60_3.data(:,1), bypass_60rat3, COrange); % Extra function defined elsewhere Integralbypass_604 = integratedata(bypass60_4.data(:,1), bypass_60rat4, COrange); % Extra function defined elsewhere Integralbypass_801 = integratedata(bypass80_1.data(:,1), bypass_80rat1, COrange); % Extra function defined elsewhere Integralbypass_802 = integratedata(bypass80_2.data(:,1), bypass_80rat2, COrange); % Extra function defined elsewhere Integralbypass_803 = integratedata(bypass80_3.data(:,1), bypass_80rat3, COrange); % Extra function defined elsewhere Integralbypass_804 = integratedata(bypass80_4.data(:,1), bypass_80rat4, COrange); % Extra function defined elsewhere Integralbypass_1001 = integratedata(bypa ss100_1.data(:,1), bypass_100rat1, COrange); % Extra function defined elsewhere Integralbypass_1002 = integratedata(bypa ss100_2.data(:,1), bypass_100rat2, COrange); % Extra function defined elsewhere Integralbypass_1003 = integratedata(bypa ss100_3.data(:,1), bypass_100rat3, COrange); % Extra function defined elsewhere Integralbypass_1004 = integratedata(bypa ss100_4.data(:,1), bypass_100rat4, COrange); % Extra function defined elsewhere bypass_20 = [Integralbypass_201 Integralbypass_202] bypass_40 = [Integralbypass_401 Inte gralbypass_402 Integralbypass_403 Integralbypass_404] bypass_60 = [Integralbypass_601 Inte gralbypass_602 Integralbypass_603 Integralbypass_604] bypass_80 = [Integralbypass_801 Inte gralbypass_802 Integralbypass_803 Integralbypass_804] bypass_100 = [Integralbypass_1001 Inte gralbypass_1002 Integralbypass_1003 Integralbypass_1004] mean20 = mean(bypass_20) mean40 = mean(bypass_40) mean60 = mean(bypass_60) mean80 = mean(bypass_80) mean100 = mean(bypass_100)

PAGE 118

Appendix A (Continued) 106 std20 = std(bypass_20) std40 = std(bypass_40) std60 = std(bypass_60) std80 = std(bypass_80) std100 = std(bypass_100) % 25oC FTIR data pr25oC_20rat1 = -(log10(pr25_20sccm1 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr25oC_20rat2 = -(log10(pr25_20sccm2 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr25oC_20rat3 = -(log10(pr25_20sccm3 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr25oC_40rat1 = -(log10(pr25_40sccm1.dat a(:,2)./background_1.data(:,2))); pr25oC_40rat2 = -(log10(pr25_40sccm2.dat a(:,2)./background_1.data(:,2))); pr25oC_40rat3 = -(log10(pr25_40sccm3.dat a(:,2)./background_1.data(:,2))); pr25oC_60rat1 = -(log10(pr25_60sccm1.dat a(:,2)./background_1.data(:,2))); pr25oC_60rat2 = -(log10(pr25_60sccm2.dat a(:,2)./background_1.data(:,2))); pr25oC_60rat3 = -(log10(pr25_60sccm3.dat a(:,2)./background_1.data(:,2))); pr25oC_80rat1 = -(log10(pr25_80sccm1.dat a(:,2)./background_1.data(:,2))); pr25oC_80rat2 = -(log10(pr25_80sccm2.dat a(:,2)./background_1.data(:,2))); pr25oC_80rat3 = -(log10(pr25_80sccm3.dat a(:,2)./background_1.data(:,2))); pr25oC_100rat1 = -(log10(pr25_100sccm1.da ta(:,2)./background_1.data(:,2))); pr25oC_100rat2 = -(log10(pr25_100sccm2.da ta(:,2)./background_1.data(:,2))); pr25oC_100rat3 = -(log10(pr25_100sccm3.da ta(:,2)./background_1.data(:,2))); pr25oC_100rat4 = -(log10(pr25_100sccm4.da ta(:,2)./background_1.data(:,2))); Integral25oc_201 = integratedata(pr2 5_20sccm1.data(:,1), pr25oC_20rat1, COrange); % Extra function defined elsewhere Integral25oc_202 = integratedata(pr2 5_20sccm2.data(:,1), pr25oC_20rat2, COrange); % Extra function defined elsewhere Integral25oc_203 = integratedata(pr2 5_20sccm3.data(:,1), pr25oC_20rat3, COrange); % Extra function defined elsewhere Integral25oc_401 = integratedata(pr2 5_40sccm1.data(:,1), pr25oC_40rat1, COrange); % Extra function defined elsewhere Integral25oc_402 = integratedata(pr2 5_40sccm2.data(:,1), pr25oC_40rat2, COrange); % Extra function defined elsewhere Integral25oc_403 = integratedata(pr2 5_40sccm3.data(:,1), pr25oC_40rat3, COrange); % Extra function defined elsewhere Integral25oc_601 = integratedata(pr2 5_60sccm1.data(:,1), pr25oC_60rat1, COrange); % Extra function defined elsewhere Integral25oc_602 = integratedata(pr2 5_60sccm2.data(:,1), pr25oC_60rat2, COrange); % Extra function defined elsewhere

PAGE 119

Appendix A (Continued) 107 Integral25oc_603 = integratedata(pr2 5_60sccm3.data(:,1), pr25oC_60rat3, COrange); % Extra function defined elsewhere Integral25oc_801 = integratedata(pr 25_80sccm1.data(:,1), pr25oC_80rat1, COrange); % Extra function defined elsewhere Integral25oc_802 = integratedata(pr 25_80sccm2.data(:,1), pr25oC_80rat2, COrange); % Extra function defined elsewhere Integral25oc_803 = integratedata(pr 25_80sccm3.data(:,1), pr25oC_80rat3, COrange); % Extra function defined elsewhere Integral25oc_1001 = integratedata(pr 25_100sccm1.data(:,1), pr25oC_100rat1, COrange); % Extra function defined elsewhere Integral25oc_1002 = integratedata(pr 25_100sccm2.data(:,1), pr25oC_100rat2, COrange); % Extra function defined elsewhere Integral25oc_1003 = integratedata(pr 25_100sccm3.data(:,1), pr25oC_100rat3, COrange); % Extra function defined elsewhere Integral25oc_1004 = integratedata(pr 25_100sccm4.data(:,1), pr25oC_100rat4, COrange); % Extra function defined elsewhere Data25_20 = [Integral25oc_201 Integr al25oc_202 Integral25oc_203] Data25_40 = [Integral25oc_401 Integr al25oc_402 Integral25oc_403] Data25_60 = [Integral25oc_601 Integr al25oc_602 Integral25oc_603] Data25_80 = [Integral25oc_801 Integr al25oc_802 Integral25oc_803] Data25_100 = [Integral25oc_1001 In tegral25oc_1002 Integral25oc_1003 Integral25oc_1004] mean25_20 = mean(Data25_20) mean25_40 = mean(Data25_40) mean25_60 = mean(Data25_60) mean25_80 = mean(Data25_80) mean25_100 = mean(Data25_100) std25_20 = std(Data25_20) std25_40 = std(Data25_40) std25_60 = std(Data25_60) std25_80 = std(Data25_80) std25_100 = std(Data25_100) % 50oC FTIR data pr50oC_20rat1 = -(log10(pr50_20sccm1 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr50oC_20rat2 = -(log10(pr50_20sccm2 .data(:,2)./background_1.data(:,2))); % Determine absorbance

PAGE 120

Appendix A (Continued) 108 pr50oC_20rat3 = -(log10(pr50_20sccm3 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr50oC_20rat4 = -(log10(pr50_20sccm4 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr50oC_40rat1 = -(log10(pr50_40sccm1.dat a(:,2)./background_1.data(:,2))); pr50oC_40rat2 = -(log10(pr50_40sccm2.dat a(:,2)./background_1.data(:,2))); pr50oC_40rat3 = -(log10(pr50_40sccm3.dat a(:,2)./background_1.data(:,2))); pr50oC_40rat4 = -(log10(pr50_40sccm4.dat a(:,2)./background_1.data(:,2))); pr50oC_60rat1 = -(log10(pr50_60sccm1.dat a(:,2)./background_1.data(:,2))); pr50oC_60rat2 = -(log10(pr50_60sccm2.dat a(:,2)./background_1.data(:,2))); pr50oC_60rat3 = -(log10(pr50_60sccm3.dat a(:,2)./background_1.data(:,2))); pr50oC_60rat4 = -(log10(pr50_60sccm4.dat a(:,2)./background_1.data(:,2))); pr50oC_80rat1 = -(log10(pr50_80sccm1.dat a(:,2)./background_1.data(:,2))); pr50oC_80rat2 = -(log10(pr50_80sccm2.dat a(:,2)./background_1.data(:,2))); pr50oC_80rat3 = -(log10(pr50_80sccm3.dat a(:,2)./background_1.data(:,2))); pr50oC_80rat4 = -(log10(pr50_80sccm4.dat a(:,2)./background_1.data(:,2))); pr50oC_80rat5 = -(log10(pr50_80sccm5.dat a(:,2)./background_1.data(:,2))); pr50oC_100rat1 = -(log10(pr50_100sccm1.da ta(:,2)./background_1.data(:,2))); pr50oC_100rat2 = -(log10(pr50_100sccm2.da ta(:,2)./background_1.data(:,2))); pr50oC_100rat3 = -(log10(pr50_100sccm3.da ta(:,2)./background_1.data(:,2))); pr50oC_100rat4 = -(log10(pr50_100sccm4.da ta(:,2)./background_1.data(:,2))); Integral50oc_201 = integratedata(pr5 0_20sccm1.data(:,1), pr50oC_20rat1, COrange); % Extra function defined elsewhere Integral50oc_202 = integratedata(pr5 0_20sccm2.data(:,1), pr50oC_20rat2, COrange); % Extra function defined elsewhere Integral50oc_203 = integratedata(pr5 0_20sccm3.data(:,1), pr50oC_20rat3, COrange); % Extra function defined elsewhere Integral50oc_204 = integratedata(pr5 0_20sccm4.data(:,1), pr50oC_20rat4, COrange); % Extra function defined elsewhere Integral50oc_401 = integratedata(pr5 0_40sccm1.data(:,1), pr50oC_40rat1, COrange); % Extra function defined elsewhere Integral50oc_402 = integratedata(pr5 0_40sccm2.data(:,1), pr50oC_40rat2, COrange); % Extra function defined elsewhere Integral50oc_403 = integratedata(pr5 0_40sccm3.data(:,1), pr50oC_40rat3, COrange); % Extra function defined elsewhere Integral50oc_404 = integratedata(pr5 0_40sccm4.data(:,1), pr50oC_40rat4, COrange); % Extra function defined elsewhere Integral50oc_601 = integratedata(pr5 0_60sccm1.data(:,1), pr50oC_60rat1, COrange); % Extra function defined elsewhere Integral50oc_602 = integratedata(pr5 0_60sccm2.data(:,1), pr50oC_60rat2, COrange); % Extra function defined elsewhere

PAGE 121

Appendix A (Continued) 109 Integral50oc_603 = integratedata(pr5 0_60sccm3.data(:,1), pr50oC_60rat3, COrange); % Extra function defined elsewhere Integral50oc_604 = integratedata(pr5 0_60sccm4.data(:,1), pr50oC_60rat4, COrange); % Extra function defined elsewhere Integral50oc_801 = integratedata(pr 50_80sccm1.data(:,1), pr50oC_80rat1, COrange); % Extra function defined elsewhere Integral50oc_802 = integratedata(pr 50_80sccm2.data(:,1), pr50oC_80rat2, COrange); % Extra function defined elsewhere Integral50oc_803 = integratedata(pr 50_80sccm3.data(:,1), pr50oC_80rat3, COrange); % Extra function defined elsewhere Integral50oc_804 = integratedata(pr 50_80sccm4.data(:,1), pr50oC_80rat4, COrange); % Extra function defined elsewhere Integral50oc_805 = integratedata(pr 50_80sccm5.data(:,1), pr50oC_80rat5, COrange); % Extra function defined elsewhere Integral50oc_1001 = integratedata(pr 50_100sccm1.data(:,1), pr50oC_100rat1, COrange); % Extra function defined elsewhere Integral50oc_1002 = integratedata(pr 50_100sccm2.data(:,1), pr50oC_100rat2, COrange); % Extra function defined elsewhere Integral50oc_1003 = integratedata(pr 50_100sccm3.data(:,1), pr50oC_100rat3, COrange); % Extra function defined elsewhere Integral50oc_1004 = integratedata(pr 50_100sccm4.data(:,1), pr50oC_100rat4, COrange); % Extra function defined elsewhere Data50_20 = [Integral50oc_201 Integral50oc_2 02 Integral50oc_203 In tegral50oc_204] Data50_40 = [Integral50oc_401 Integral50oc_4 02 Integral50oc_403 In tegral50oc_404] Data50_60 = [Integral50oc_601 Integral50oc_6 02 Integral50oc_603 In tegral50oc_604] Data50_80 = [Integral50oc_802 Integr al50oc_804 Integral50oc_805] Data50_100 = [Integral50oc_1001 In tegral50oc_1002 Integral50oc_1003 Integral50oc_1004] mean50_20 = mean(Data50_20) mean50_40 = mean(Data50_40) mean50_60 = mean(Data50_60) mean50_80 = mean(Data50_80) mean50_100 = mean(Data50_100) std50_20 = std(Data50_20) std50_40 = std(Data50_40) std50_60 = std(Data50_60) std50_80 = std(Data50_80) std50_100 = std(Data50_100)

PAGE 122

Appendix A (Continued) 110 % 75oC FTIR data pr75oC_20rat1 = -(log10(pr75_20sccm1 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr75oC_20rat2 = -(log10(pr75_20sccm2 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr75oC_20rat3 = -(log10(pr75_20sccm3 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr75oC_20rat4 = -(log10(pr75_20sccm4 .data(:,2)./background_1.data(:,2))); % Determine absorbance pr75oC_40rat1 = -(log10(pr75_40sccm1.dat a(:,2)./background_1.data(:,2))); pr75oC_40rat2 = -(log10(pr75_40sccm2.dat a(:,2)./background_1.data(:,2))); pr75oC_40rat3 = -(log10(pr75_40sccm3.dat a(:,2)./background_1.data(:,2))); pr75oC_40rat4 = -(log10(pr75_40sccm4.dat a(:,2)./background_1.data(:,2))); pr75oC_40rat5 = -(log10(pr75_40sccm5.dat a(:,2)./background_1.data(:,2))); pr75oC_60rat1 = -(log10(pr75_60sccm1.dat a(:,2)./background_1.data(:,2))); pr75oC_60rat2 = -(log10(pr75_60sccm2.dat a(:,2)./background_1.data(:,2))); pr75oC_60rat3 = -(log10(pr75_60sccm3.dat a(:,2)./background_1.data(:,2))); pr75oC_60rat4 = -(log10(pr75_60sccm4.dat a(:,2)./background_1.data(:,2))); pr75oC_80rat1 = -(log10(pr75_80sccm1.dat a(:,2)./background_1.data(:,2))); pr75oC_80rat2 = -(log10(pr75_80sccm2.dat a(:,2)./background_1.data(:,2))); pr75oC_80rat3 = -(log10(pr75_80sccm3.dat a(:,2)./background_1.data(:,2))); pr75oC_80rat4 = -(log10(pr75_80sccm4.dat a(:,2)./background_1.data(:,2))); pr75oC_80rat5 = -(log10(pr75_80sccm5.dat a(:,2)./background_1.data(:,2))); pr75oC_100rat1 = -(log10(pr75_100sccm1.da ta(:,2)./background_1.data(:,2))); pr75oC_100rat2 = -(log10(pr75_100sccm2.da ta(:,2)./background_1.data(:,2))); pr75oC_100rat3 = -(log10(pr75_100sccm3.da ta(:,2)./background_1.data(:,2))); pr75oC_100rat4 = -(log10(pr75_100sccm4.da ta(:,2)./background_1.data(:,2))); pr75oC_100rat5 = -(log10(pr75_100sccm5.da ta(:,2)./background_1.data(:,2))); pr75oC_100rat6 = -(log10(pr75_100sccm6.da ta(:,2)./background_1.data(:,2))); Integral75oc_201 = integratedata(pr7 5_20sccm1.data(:,1), pr75oC_20rat1, COrange); % Extra function defined elsewhere Integral75oc_202 = integratedata(pr7 5_20sccm2.data(:,1), pr75oC_20rat2, COrange); % Extra function defined elsewhere Integral75oc_203 = integratedata(pr7 5_20sccm3.data(:,1), pr75oC_20rat3, COrange); % Extra function defined elsewhere Integral75oc_204 = integratedata(pr7 5_20sccm4.data(:,1), pr75oC_20rat4, COrange); % Extra function defined elsewhere Integral75oc_401 = integratedata(pr7 5_40sccm1.data(:,1), pr75oC_40rat1, COrange); % Extra function defined elsewhere Integral75oc_402 = integratedata(pr7 5_40sccm2.data(:,1), pr75oC_40rat2, COrange); % Extra function defined elsewhere

PAGE 123

Appendix A (Continued) 111 Integral75oc_403 = integratedata(pr7 5_40sccm3.data(:,1), pr75oC_40rat3, COrange); % Extra function defined elsewhere Integral75oc_404 = integratedata(pr7 5_40sccm4.data(:,1), pr75oC_40rat4, COrange); % Extra function defined elsewhere Integral75oc_405 = integratedata(pr7 5_40sccm5.data(:,1), pr75oC_40rat5, COrange); % Extra function defined elsewhere Integral75oc_601 = integratedata(pr7 5_60sccm1.data(:,1), pr75oC_60rat1, COrange); % Extra function defined elsewhere Integral75oc_602 = integratedata(pr7 5_60sccm2.data(:,1), pr75oC_60rat2, COrange); % Extra function defined elsewhere Integral75oc_603 = integratedata(pr7 5_60sccm3.data(:,1), pr75oC_60rat3, COrange); % Extra function defined elsewhere Integral75oc_604 = integratedata(pr7 5_60sccm4.data(:,1), pr75oC_60rat4, COrange); % Extra function defined elsewhere Integral75oc_801 = integratedata(pr 75_80sccm1.data(:,1), pr75oC_80rat1, COrange); % Extra function defined elsewhere Integral75oc_802 = integratedata(pr 75_80sccm2.data(:,1), pr75oC_80rat2, COrange); % Extra function defined elsewhere Integral75oc_803 = integratedata(pr 75_80sccm3.data(:,1), pr75oC_80rat3, COrange); % Extra function defined elsewhere Integral75oc_804 = integratedata(pr 75_80sccm4.data(:,1), pr75oC_80rat4, COrange); % Extra function defined elsewhere Integral75oc_805 = integratedata(pr 75_80sccm5.data(:,1), pr75oC_80rat5, COrange); % Extra function defined elsewhere Integral75oc_1001 = integratedata(pr 75_100sccm1.data(:,1), pr75oC_100rat1, COrange); % Extra function defined elsewhere Integral75oc_1002 = integratedata(pr 75_100sccm2.data(:,1), pr75oC_100rat2, COrange); % Extra function defined elsewhere Integral75oc_1003 = integratedata(pr 75_100sccm3.data(:,1), pr75oC_100rat3, COrange); % Extra function defined elsewhere Integral75oc_1004 = integratedata(pr 75_100sccm4.data(:,1), pr75oC_100rat4, COrange); % Extra function defined elsewhere Integral75oc_1005 = integratedata(pr 75_100sccm5.data(:,1), pr75oC_100rat5, COrange); % Extra function defined elsewhere Integral75oc_1006 = integratedata(pr 75_100sccm6.data(:,1), pr75oC_100rat6, COrange); % Extra function defined elsewhere Data75_20 = [Integral75oc_201 Integral75oc_2 02 Integral75oc_203 In tegral75oc_204] Data75_40 = [Integral75oc_401 Integral75oc_ 402 Integral75oc_403 Integral75oc_404 Integral75oc_405] Data75_60 = [Integral75oc_601 Integral75oc_6 02 Integral75oc_603 In tegral75oc_604] Data75_80 = [Integral75oc_801 Integral75oc_ 802 Integral75oc_803 Integral75oc_804 Integral75oc_805]

PAGE 124

Appendix A (Continued) 112 Data75_100 = [Integral75oc_1001 In tegral75oc_1002 Integral75oc_1003 Integral75oc_1004 Integral75oc _1005 Integral75oc_1006] mean75_20 = mean(Data75_20) mean75_40 = mean(Data75_40) mean75_60 = mean(Data75_60) mean75_80 = mean(Data75_80) mean75_100 = mean(Data75_100) std75_20 = std(Data75_20) std75_40 = std(Data75_40) std75_60 = std(Data75_60) std75_80 = std(Data75_80) std75_100 = std(Data75_100) % 100oC FTIR data pr100oC_20rat1 = -(log10(pr100_20sccm1. data(:,2)./background_1.data(:,2))); % Determine absorbance pr100oC_20rat2 = -(log10(pr100_20sccm2. data(:,2)./background_1.data(:,2))); % Determine absorbance pr100oC_20rat3 = -(log10(pr100_20sccm3. data(:,2)./background_1.data(:,2))); % Determine absorbance pr100oC_20rat4 = -(log10(pr100_20sccm4. data(:,2)./background_1.data(:,2))); % Determine absorbance pr100oC_20rat5 = -(log10(pr100_20sccm5. data(:,2)./background_1.data(:,2))); % Determine absorbance pr100oC_20rat6 = -(log10(pr100_20sccm6. data(:,2)./background_1.data(:,2))); % Determine absorbance pr100oC_40rat1 = -(log10(pr100_40sccm1.da ta(:,2)./background_1.data(:,2))); pr100oC_40rat2 = -(log10(pr100_40sccm2.da ta(:,2)./background_1.data(:,2))); pr100oC_40rat3 = -(log10(pr100_40sccm3.da ta(:,2)./background_1.data(:,2))); pr100oC_40rat4 = -(log10(pr100_40sccm4.da ta(:,2)./background_1.data(:,2))); pr100oC_40rat5 = -(log10(pr100_40sccm5.da ta(:,2)./background_1.data(:,2))); pr100oC_60rat1 = -(log10(pr100_60sccm1.da ta(:,2)./background_1.data(:,2))); pr100oC_60rat2 = -(log10(pr100_60sccm2.da ta(:,2)./background_1.data(:,2))); pr100oC_60rat3 = -(log10(pr100_60sccm3.da ta(:,2)./background_1.data(:,2))); pr100oC_60rat4 = -(log10(pr100_60sccm4.da ta(:,2)./background_1.data(:,2))); pr100oC_60rat5 = -(log10(pr100_60sccm5.da ta(:,2)./background_1.data(:,2))); pr100oC_60rat6 = -(log10(pr100_60sccm6.da ta(:,2)./background_1.data(:,2))); pr100oC_60rat7 = -(log10(pr100_60sccm7.da ta(:,2)./background_1.data(:,2))); pr100oC_80rat1 = -(log10(pr100_80sccm1.da ta(:,2)./background_1.data(:,2))); pr100oC_80rat2 = -(log10(pr100_80sccm2.da ta(:,2)./background_1.data(:,2))); pr100oC_80rat3 = -(log10(pr100_80sccm3.da ta(:,2)./background_1.data(:,2)));

PAGE 125

Appendix A (Continued) 113 pr100oC_80rat4 = -(log10(pr100_80sccm4.da ta(:,2)./background_1.data(:,2))); pr100oC_80rat5 = -(log10(pr100_80sccm5.da ta(:,2)./background_1.data(:,2))); pr100oC_100rat1 = -(log10(pr100_100sccm1.da ta(:,2)./background_1.data(:,2))); pr100oC_100rat2 = -(log10(pr100_100sccm2.da ta(:,2)./background_1.data(:,2))); pr100oC_100rat3 = -(log10(pr100_100sccm3.da ta(:,2)./background_1.data(:,2))); pr100oC_100rat4 = -(log10(pr100_100sccm4.da ta(:,2)./background_1.data(:,2))); pr100oC_100rat5 = -(log10(pr100_100sccm5.da ta(:,2)./background_1.data(:,2))); Integral100oc_201 = integratedata(pr 100_20sccm1.data(:,1), pr100oC_20rat1, COrange); % Extra function defined elsewhere Integral100oc_202 = integratedata(pr 100_20sccm2.data(:,1), pr100oC_20rat2, COrange); % Extra function defined elsewhere Integral100oc_203 = integratedata(pr 100_20sccm3.data(:,1), pr100oC_20rat3, COrange); % Extra function defined elsewhere Integral100oc_204 = integratedata(pr 100_20sccm4.data(:,1), pr100oC_20rat4, COrange); % Extra function defined elsewhere Integral100oc_205 = integratedata(pr 100_20sccm5.data(:,1), pr100oC_20rat5, COrange); % Extra function defined elsewhere Integral100oc_206 = integratedata(pr 100_20sccm6.data(:,1), pr100oC_20rat6, COrange); % Extra function defined elsewhere Integral100oc_401 = integratedata(pr 100_40sccm1.data(:,1), pr100oC_40rat1, COrange); % Extra function defined elsewhere Integral100oc_402 = integratedata(pr 100_40sccm2.data(:,1), pr100oC_40rat2, COrange); % Extra function defined elsewhere Integral100oc_403 = integratedata(pr 100_40sccm3.data(:,1), pr100oC_40rat3, COrange); % Extra function defined elsewhere Integral100oc_404 = integratedata(pr 100_40sccm4.data(:,1), pr100oC_40rat4, COrange); % Extra function defined elsewhere Integral100oc_405 = integratedata(pr 100_40sccm5.data(:,1), pr100oC_40rat5, COrange); % Extra function defined elsewhere Integral100oc_601 = integratedata(pr 100_60sccm1.data(:,1), pr100oC_60rat1, COrange); % Extra function defined elsewhere Integral100oc_602 = integratedata(pr 100_60sccm2.data(:,1), pr100oC_60rat2, COrange); % Extra function defined elsewhere Integral100oc_603 = integratedata(pr 100_60sccm3.data(:,1), pr100oC_60rat3, COrange); % Extra function defined elsewhere Integral100oc_604 = integratedata(pr 100_60sccm4.data(:,1), pr100oC_60rat4, COrange); % Extra function defined elsewhere Integral100oc_605 = integratedata(pr 100_60sccm5.data(:,1), pr100oC_60rat5, COrange); % Extra function defined elsewhere Integral100oc_606 = integratedata(pr 100_60sccm6.data(:,1), pr100oC_60rat6, COrange); % Extra function defined elsewhere

PAGE 126

Appendix A (Continued) 114 Integral100oc_607 = integratedata(pr 100_60sccm7.data(:,1), pr100oC_60rat7, COrange); % Extra function defined elsewhere Integral100oc_801 = integratedata( pr100_80sccm1.data(:,1), pr100oC_80rat1, COrange); % Extra function defined elsewhere Integral100oc_802 = integratedata( pr100_80sccm2.data(:,1), pr100oC_80rat2, COrange); % Extra function defined elsewhere Integral100oc_803 = integratedata( pr100_80sccm3.data(:,1), pr100oC_80rat3, COrange); % Extra function defined elsewhere Integral100oc_804 = integratedata( pr100_80sccm4.data(:,1), pr100oC_80rat4, COrange); % Extra function defined elsewhere Integral100oc_805 = integratedata( pr100_80sccm5.data(:,1), pr100oC_80rat5, COrange); % Extra function defined elsewhere Integral100oc_1001 = integratedata(pr 100_100sccm1.data(:,1), pr100oC_100rat1, COrange); % Ex tra function defined elsewhere Integral100oc_1002 = integratedata(pr 100_100sccm2.data(:,1), pr100oC_100rat2, COrange); % Ex tra function defined elsewhere Integral100oc_1003 = integratedata(pr 100_100sccm3.data(:,1), pr100oC_100rat3, COrange); % Ex tra function defined elsewhere Integral100oc_1004 = integratedata(pr 100_100sccm4.data(:,1), pr100oC_100rat4, COrange); % Ex tra function defined elsewhere Integral100oc_1005 = integratedata(pr 100_100sccm5.data(:,1), pr100oC_100rat5, COrange); % Ex tra function defined elsewhere Data100_20 = [Integral100oc_201 In tegral100oc_202 Integral100oc_203 Integral100oc_204 Integral 100oc_205 Integral100oc_206] Data100_40 = [Integral100oc_401 In tegral100oc_402 Integral100oc_403 Integral100oc_404 Integral100oc_405] Data100_60 = [Integral100oc_601 In tegral100oc_602 Integral100oc_603 Integral100oc_604 Integral100oc_605 In tegral100oc_606 Integral100oc_607] Data100_80 = [Integral100oc_801 In tegral100oc_802 Integral100oc_803 Integral100oc_804 Integral100oc_805] Data100_100 = [Integral100oc_1001 Inte gral100oc_1002 Integral100oc_1003 Integral100oc_1004 In tegral100oc_1005] mean100_20 = mean(Data100_20) mean100_40 = mean(Data100_40) mean100_60 = mean(Data100_60) mean100_80 = mean(Data100_80) mean100_100 = mean(Data100_100) std100_20 = std(Data100_20) std100_40 = std(Data100_40) std100_60 = std(Data100_60)

PAGE 127

Appendix A (Continued) 115 std100_80 = std(Data100_80) std100_100 = std(Data100_100) % 125oC FTIR data pr125oC_20rat1 = -(log10(pr125_20sccm1. data(:,2)./background_1.data(:,2))); % Determine absorbance pr125oC_20rat2 = -(log10(pr125_20sccm2. data(:,2)./background_1.data(:,2))); % Determine absorbance pr125oC_20rat3 = -(log10(pr125_20sccm3. data(:,2)./background_1.data(:,2))); % Determine absorbance pr125oC_20rat4 = -(log10(pr125_20sccm4. data(:,2)./background_1.data(:,2))); % Determine absorbance pr125oC_20rat5 = -(log10(pr125_20sccm5. data(:,2)./background_1.data(:,2))); % Determine absorbance pr125oC_20rat6 = -(log10(pr125_20sccm6.da ta(:,2)./background_1.data(:,2))); pr125oC_40rat1 = -(log10(pr125_40sccm1.da ta(:,2)./background_1.data(:,2))); pr125oC_40rat2 = -(log10(pr125_40sccm2.da ta(:,2)./background_1.data(:,2))); pr125oC_40rat3 = -(log10(pr125_40sccm3.da ta(:,2)./background_1.data(:,2))); pr125oC_40rat4 = -(log10(pr125_40sccm4.da ta(:,2)./background_1.data(:,2))); pr125oC_40rat5 = -(log10(pr125_40sccm5.da ta(:,2)./background_1.data(:,2))); pr125oC_60rat1 = -(log10(pr125_60sccm1.da ta(:,2)./background_1.data(:,2))); pr125oC_60rat2 = -(log10(pr125_60sccm2.da ta(:,2)./background_1.data(:,2))); pr125oC_60rat3 = -(log10(pr125_60sccm3.da ta(:,2)./background_1.data(:,2))); pr125oC_60rat4 = -(log10(pr125_60sccm4.da ta(:,2)./background_1.data(:,2))); pr125oC_60rat5 = -(log10(pr125_60sccm5.da ta(:,2)./background_1.data(:,2))); pr125oC_80rat1 = -(log10(pr125_80sccm1.da ta(:,2)./background_1.data(:,2))); pr125oC_80rat2 = -(log10(pr125_80sccm2.da ta(:,2)./background_1.data(:,2))); pr125oC_80rat3 = -(log10(pr125_80sccm3.da ta(:,2)./background_1.data(:,2))); pr125oC_80rat4 = -(log10(pr125_80sccm4.da ta(:,2)./background_1.data(:,2))); pr125oC_80rat5 = -(log10(pr125_80sccm5.da ta(:,2)./background_1.data(:,2))); pr125oC_100rat1 = -(log10(pr125_100sccm1.da ta(:,2)./background_1.data(:,2))); pr125oC_100rat2 = -(log10(pr125_100sccm2.da ta(:,2)./background_1.data(:,2))); pr125oC_100rat3 = -(log10(pr125_100sccm3.da ta(:,2)./background_1.data(:,2))); pr125oC_100rat4 = -(log10(pr125_100sccm4.da ta(:,2)./background_1.data(:,2))); pr125oC_100rat5 = -(log10(pr125_100sccm5.da ta(:,2)./background_1.data(:,2))); pr125oC_100rat6 = -(log10(pr125_100sccm6.da ta(:,2)./background_1.data(:,2))); Integral125oc_201 = integratedata(pr 125_20sccm1.data(:,1), pr125oC_20rat1, COrange); % Extra function defined elsewhere Integral125oc_202 = integratedata(pr 125_20sccm2.data(:,1), pr125oC_20rat2, COrange); % Extra function defined elsewhere Integral125oc_203 = integratedata(pr 125_20sccm3.data(:,1), pr125oC_20rat3, COrange); % Extra function defined elsewhere

PAGE 128

Appendix A (Continued) 116 Integral125oc_204 = integratedata(pr 125_20sccm4.data(:,1), pr125oC_20rat4, COrange); % Extra function defined elsewhere Integral125oc_205 = integratedata(pr 125_20sccm5.data(:,1), pr125oC_20rat5, COrange); % Extra function defined elsewhere Integral125oc_206 = integratedata(pr 125_20sccm6.data(:,1), pr125oC_20rat6, COrange); % Extra function defined elsewhere Integral125oc_401 = integratedata(pr 125_40sccm1.data(:,1), pr125oC_40rat1, COrange); % Extra function defined elsewhere Integral125oc_402 = integratedata(pr 125_40sccm2.data(:,1), pr125oC_40rat2, COrange); % Extra function defined elsewhere Integral125oc_403 = integratedata(pr 125_40sccm3.data(:,1), pr125oC_40rat3, COrange); % Extra function defined elsewhere Integral125oc_404 = integratedata(pr 125_40sccm4.data(:,1), pr125oC_40rat4, COrange); % Extra function defined elsewhere Integral125oc_405 = integratedata(pr 125_40sccm5.data(:,1), pr125oC_40rat5, COrange); % Extra function defined elsewhere Integral125oc_601 = integratedata(pr 125_60sccm1.data(:,1), pr125oC_60rat1, COrange); % Extra function defined elsewhere Integral125oc_602 = integratedata(pr 125_60sccm2.data(:,1), pr125oC_60rat2, COrange); % Extra function defined elsewhere Integral125oc_603 = integratedata(pr 125_60sccm3.data(:,1), pr125oC_60rat3, COrange); % Extra function defined elsewhere Integral125oc_604 = integratedata(pr 125_60sccm4.data(:,1), pr125oC_60rat4, COrange); % Extra function defined elsewhere Integral125oc_605 = integratedata(pr 125_60sccm5.data(:,1), pr125oC_60rat5, COrange); % Extra function defined elsewhere Integral125oc_801 = integratedata( pr125_80sccm1.data(:,1), pr125oC_80rat1, COrange); % Extra function defined elsewhere Integral125oc_802 = integratedata( pr125_80sccm2.data(:,1), pr125oC_80rat2, COrange); % Extra function defined elsewhere Integral125oc_803 = integratedata( pr125_80sccm3.data(:,1), pr125oC_80rat3, COrange); % Extra function defined elsewhere Integral125oc_804 = integratedata( pr125_80sccm4.data(:,1), pr125oC_80rat4, COrange); % Extra function defined elsewhere Integral125oc_805 = integratedata( pr125_80sccm5.data(:,1), pr125oC_80rat5, COrange); % Extra function defined elsewhere Integral125oc_1001 = integratedata(pr 125_100sccm1.data(:,1), pr125oC_100rat1, COrange); % Ex tra function defined elsewhere Integral125oc_1002 = integratedata(pr 125_100sccm2.data(:,1), pr125oC_100rat2, COrange); % Ex tra function defined elsewhere Integral125oc_1003 = integratedata(pr 125_100sccm3.data(:,1), pr125oC_100rat3, COrange); % Ex tra function defined elsewhere

PAGE 129

Appendix A (Continued) 117 Integral125oc_1004 = integratedata(pr 125_100sccm4.data(:,1), pr125oC_100rat4, COrange); % Ex tra function defined elsewhere Integral125oc_1005 = integratedata(pr 125_100sccm5.data(:,1), pr125oC_100rat5, COrange); % Ex tra function defined elsewhere Integral125oc_1006 = integratedata(pr 125_100sccm6.data(:,1), pr125oC_100rat6, COrange); % Ex tra function defined elsewhere Data125_20 = [Integral125oc_201 In tegral125oc_202 Integral125oc_203 Integral125oc_204 Integral 125oc_205 Integral125oc_206] Data125_40 = [Integral125oc_401 In tegral125oc_402 Integral125oc_403 Integral125oc_404 Integral125oc_405] Data125_60 = [Integral125oc_601 In tegral125oc_602 Integral125oc_603 Integral125oc_604 Integral125oc_605] Data125_80 = [Integral125oc_801 In tegral125oc_802 Integral125oc_803 Integral125oc_804 Integral125oc_805] Data125_100 = [Integral125oc_1001 Inte gral125oc_1002 Integral125oc_1003 Integral125oc_1004 Integral 125oc_1005 Integral125oc_1006] mean125_20 = mean(Data125_20) mean125_40 = mean(Data125_40) mean125_60 = mean(Data125_60) mean125_80 = mean(Data125_80) mean125_100 = mean(Data125_100) std125_20 = std(Data125_20) std125_40 = std(Data125_40) std125_60 = std(Data125_60) std125_80 = std(Data125_80) std125_100 = std(Data125_100) % Error Error25_20 = sqrt((mean20/mean25_20^2)^ 2*std20^2 + (1/mean20)^2*std25_20^2) Error25_40 = sqrt((mean40/mean25_40^2)^ 2*std40^2 + (1/mean40)^2*std25_40^2) Error25_60 = sqrt((mean60/mean25_60^2)^ 2*std60^2 + (1/mean60)^2*std25_60^2) Error25_80 = sqrt((mean80/mean25_80^2)^ 2*std80^2 + (1/mean80)^2*std25_80^2) Error25_100 = sqrt((mean100/mean25_100^2)^2*std100^2 + (1/mean100)^2*std25_100^2) Error50_20 = sqrt((mean20/mean50_20^2)^ 2*std20^2 + (1/mean20)^2*std50_20^2) Error50_40 = sqrt((mean40/mean50_40^2)^ 2*std40^2 + (1/mean40)^2*std50_40^2) Error50_60 = sqrt((mean60/mean50_60^2)^ 2*std60^2 + (1/mean60)^2*std50_60^2) Error50_80 = sqrt((mean80/mean50_80^2)^ 2*std80^2 + (1/mean80)^2*std50_80^2)

PAGE 130

Appendix A (Continued) 118 Error50_100 = sqrt((mean100/mean50_100^2)^2*std100^2 + (1/mean100)^2*std50_100^2) Error75_20 = sqrt((mean20/mean75_20^2)^ 2*std20^2 + (1/mean20)^2*std75_20^2) Error75_40 = sqrt((mean40/mean75_40^2)^ 2*std40^2 + (1/mean40)^2*std75_40^2) Error75_60 = sqrt((mean60/mean75_60^2)^ 2*std60^2 + (1/mean60)^2*std75_60^2) Error75_80 = sqrt((mean80/mean75_80^2)^ 2*std80^2 + (1/mean80)^2*std75_80^2) Error75_100 = sqrt((mean100/mean75_100^2)^2*std100^2 + (1/mean100)^2*std75_100^2) Error100_20 = sqrt((mean20/mean100_20^2) ^2*std20^2 + (1/mean20)^2*std100_20^2) Error100_40 = sqrt((mean40/mean100_40^2) ^2*std40^2 + (1/mean40)^2*std100_40^2) Error100_60 = sqrt((mean60/mean100_60^2) ^2*std60^2 + (1/mean60)^2*std100_60^2) Error100_80 = sqrt((mean80/mean100_80^2) ^2*std80^2 + (1/mean80)^2*std100_80^2) Error100_100 = sqrt((mean100/mean100_100^2)^2*std100^2 + (1/mean100)^2*std100_100^2) Error125_20 = sqrt((mean20/mean125_20^2) ^2*std20^2 + (1/mean20)^2*std125_20^2) Error125_40 = sqrt((mean40/mean125_40^2) ^2*std40^2 + (1/mean40)^2*std125_40^2) Error125_60 = sqrt((mean60/mean125_60^2) ^2*std60^2 + (1/mean60)^2*std125_60^2) Error125_80 = sqrt((mean80/mean125_80^2) ^2*std80^2 + (1/mean80)^2*std125_80^2) Error125_100 = sqrt((mean100/mean125_100^2)^2*std100^2 + (1/mean100)^2*std125_100^2)

PAGE 131

ABOUT THE AUTHOR Benjamin Alan Grayson, a native of Mize, MS, was valedictorian of his 1995 high school graduating class. He began his undergraduate education at Jones County Junior College in Ellisville, MS before transferring to Missis sippi State University to finish his B.S. in chemical engineering in May, 2000. Be njamin received his M.S.Ch.E. from the University of South Florida in December, 2002. He continued at USF to finish his Ph.D. in August, 2007.


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 Ka
controlfield tag 001 001921017
003 fts
005 20080116110715.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 080116s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002131
035
(OCoLC)190844513
040
FHM
c FHM
049
FHMM
090
TP145 (ONLINE)
1 100
Grayson, Benjamin Alan.
0 245
Application and modeling of TiO2-supported gold nanoparticles for CO preferential oxidation in excess hydrogen
h [electronic resource] /
by Benjamin Alan Grayson.
260
[Tampa, Fla.] :
b University of South Florida,
2007.
520
ABSTRACT: This work begins with a brief overview of heterogeneous, characterization techniques, and current hypotheses about gold mechanisms. This is followed by the initial characterization of custom two-phase-method gold nanoparticles provided by the Interfacial Phenomena and Polymeric Materials research group at USF, the anatase TiO2 support and reference Au/TiO2 catalyst provided by the World Gold Council. In order to verify the ability of the two-phase-method GNP catalyst provided to oxidize CO in excess hydrogen, it was necessary to develop an effluent testing protocol. The first experiments involved 24 hour runs to observe catalyst deactivation. Concerns over cycling effects observed in the absorbance integral calculations lead to the introduction of a reference gas. Corrections were made to the carbon monoxide absorbance integral calculations which allowed the direct comparison of results.These corrections included baseline adjustments for each species and N2 purging to eliminate background CO2 and H2O contamination. After these improvements, the two phase method GNP catalyst CO oxidation ability was investigated. Unfortunately, the supplied two phase method gold catalyst has been unresponsive for CO oxidation applications. One hypothesis for the problems is that the surfactants used to keep the gold nanoparticles from aggregating are preventing carbon monoxide transport to the surface of the particle. Another theory is that the gold may not be adhering to the surface of the TiO2 creating a cohesive metal/support interaction. The kinetics of CO preferential oxidation (PROX) catalyzed by the World Gold Council's nano-Au/TiO2 was studied to evaluate elementary and nonelementary empirical rate expressions. Information is readily available for CO fractional conversion for this catalyst below 0 degrees C.However, a comprehensive CO PROX kinetic model in which three reactions (CO oxidation, H2 oxidation and the water gas shift reaction) occur simultaneously is lacking. The reaction was carried out in a vertical packed bed micro-reactor testing unit; temperature was varied between 25 and 125 degrees C, and a range of feed rates were tested. In-situ Fourier transform infrared spectroscopy (FTIR) reaction data was analyzed; pre-exponential and activation energies are calculated for each kinetic model. Empirical rate expressions based on power law models were used to fit the experimental data. The reversible water gas shift reaction was found to play an important role when fitting the experimental data precisely and explained the selectivity decrease at higher reaction temperatures. The empirical kinetic model presented will be useful to simulate PROX operation parameters for many applications.
502
Dissertation (Ph.D.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 118 pages.
Includes vita.
590
Advisor: John T. Wolan, Ph.D.
653
Supported Au/TiO2.
Kinetic modeling.
Catalysis.
World Gold Council.
FTIR.
690
Dissertations, Academic
z USF
x Chemical Engineering
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2131