THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING THE EFFECT OF CHANGES IN CATALYST COMPOSITION ON THE HYDROGEN-DEUTERIUM EXCHANGE REACTION ON COBALT FERRITE Robert Go Squires A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the University of Michigan Department of Chemical and Metallurgical Engineering 1962 October, 1962 IP-587

ACKNOWLEDGEMENTS The author wishes to express his appreciation to the members of his committee for their guidance during the course of this work. Special thanks are due to Professor Go Parravano, who suggested the topic for this research and served as chairman, for his numerous helpful suggestions and criticisms. The author would also like to give special thanks to Professor JO J, Martin for his encouragement given throughout the author's studies, and to Professor D, Ro Mason for arousing the author's interest in the application of semiconductor theory to catalysis. The author is indebted to the Standard Oil Company of California and to the National Science Foundation for their financial aid through the award of fellowships for four years. To the laboratory and shop personnel of the Department of Chemical and Metallurgical Engineering-the author expresses his thanks. In particular, Mr. F. Bo Drogosz deserves the author's thanks for his assistance in solving some of the analytical problems presented by this work, Finally, the author would like to thank his wife for her constant encouragement and for the typing of both the rough and final drafts of this manuscripto The drafting of the figures by Mr, To Fo Beals and the printing of this manuscript by the Industry Program of the College of Engineering are also very much appreciated, ii

TABLE OF CONTENTS Page ACKNOWLEDGEME NTSS................................................... i i ACKNOWLEDGEMENTS o o o o o o o o D o o o o o o o o o o o o o. o o o o o o o o o o o o o o o o o o o o o o o o o o o LIST OF...................., oaeDOOQ~0000006000 vi LIST OF TABES ooo oooooooo ooo oooooooooooooooooo o ooooo oooo oo Vix LIST OF FIGURES0. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 vi ABSTRACT........................ o o o o oo ooooooooooooooooooooooooooooooo o I. INTRODUCTION,,,oo,,,,,o 0 0 0.. 0 0 0. 0 0 0 0 0 0 0 O ooooo o o o 1 Ao Background and Scope of Research... oooo...oo.oo.ooo 1 B, Literature Survey.oo o o o o o o lo The Electron Theory of Catalysiso o o o o o o.. o o 4 a. Development of Theory,00.........000 0000000 4 bo Chemisorption Studies...oo............... 6 Co Reaction Studies... o...o o o o.. o o..o 0 6 do Hydrogen-Deuterium Exchange Studieso o..... 8 e. Reactions on Ferrite Catalystso o oo.o o o 9 2o Cobalt Ferriteo o o o o o o 0 0 o................. 0 0 0 10 a. Preparationo... o 0o o o 0... o o.o 10 bo Phase Behavioro o,,........ o o o o o o o o o o o o o o 10 co Semiconducting Properties,,. o...oo..o o.. 11 IIe THEORY.. O.......ooo oo o.o.o o oo.. o o o 0...... o.. o 12 A. Semiconducting Properties of Cobalt Ferrite..oo.... 12 B. The Effect of Chemisorption on Thermoelectric Properties, o o o... o o........... o.oooo..... o o.o o 24 Co The Effect of Changes in Fermi Level on Adsorption and Catalytic Activityo 0 oo0o00000...00000..0000....0 28 D. Determination of Activation Energy and Preexponential Factor oo....o.........oo.................. 34 lo Activation Energyo,,o OOO. o0 o o oo. o 0 0 0 0 0 35 2o Pre-exponential Factor.,00..0.....00.0000..0..0. 37 Eo Isotopic Analysis by the Mass Spectrometero o...... 38 III, EXPERIMENTAL APPARATUS AND TECHNIQUES. o 00...00.......... 42 Ao Kinetic Experiments o oo o o. o o o o o o o. o o o o... o o o o o o o 42 iii

TABLE OF CONTENTS CONT'D Page 1o Catalyst Preparation and Analysis o...o,00 0.... 42 2. Apparatus 0.000.0000.0000.0 00 0 o o D.0 OO.... 43 35 Experimental Procedures, o o... o.......o......... 50 4o Analytical Methodso o..ooo......oooooooooooooo.. 51 Bo Thermoelectric Power Studieso ooo....o....o.... ooooo 53 1. Catalyst Preparation and Analysiso.... o.. 00.. 53 2. Apparatus.0 0 0 0 0 0 0 0 o.............. 0. 0. o 54 3. Experimental Procedureso.000ooooo.....0oo o oo o 61 C. Experimental Program...0000..00o00.0o oo.oo0ooo ooooo 62 IV. EXPERIMENTAL RESULTSo o 0................ oo............... 64 Ao Hydrogen-Deuterium Exchange Studies..o o0oo..o..... 64 1. Ferrite Catalyst Characterization 00.....ooo0000 64 2. Exchange Runs...OOOO0OOOOOOOO0ooo00000000000000. 69 B. The Effect of Chemisorption on Thermoelectric Powero 77 1, Cobalt Ferrite Characterizationo..o,,.....oo.o. 77 2o Thermoelectric Power Measurements During Chemisorption of Hydrogen and Oxygen Gases ooO,,,o.o, 81 V. DISCUSSION OF RESULTS.o.oooooooooooooooooooooo.., oo.ooo0 88 Ao Catalyst.D00000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 0O 88 Bo Thermoelectric Power Changes During Chemisorptiono.. 91 1. High Temperature Runs, oo..........oooo....... oo 91 2o Intermediate Temperature Runs. oo...o..oo...oooo 95 30 Low Temperature Runs..oooo........ooo..oo 0000.. 95 Co Kinetic Studies o o........ 00...........,,.0o o. 96 D. Proposed Reaction Mechanism.......0000000000000000. 0 96 1o Thermoelectric Power Studies During Chemisorption 98 20 Hydrogen-Denterium Exchange Datao o.. o........ 99 VIo CONCLUSIONS..... ooooooo.0o o o o o o o o oooo o o o o o 102 APPENDICESeo o o o o o o o o.o o o o o o o o o o o o o o oo o o o o o o o o o o o o o o o o o o o o o.104 NOMENCLATUE..0000000000000000i000000000000000000 12 iJv

LIST OF TABLES Table Page I Constants Needed to Describe the Semiconducting Properties of Co3X Fe 04o,,o..00. a...Oo,...oooo..o.o.o 23 II Ions From a Mixture of the Hydrogens in Which H is More Abundant Than Do, o. IO......o o ao.o.o o o o oa oo o.o.o..o 39 III Comparison of Standard Deviations of X-Ray Analyses of the Four Catalyst Sampleso ooOOo.oo.oooooooooooooo.oo.. 66 IV Summary of Activation Energy and Pre-exponential Factors For Hydrogen-Deuterium Exchange Experiments..........ooo 76 V Activation Energies and Pre-exponential Factors For Runs Using Mixed Sintered Catalysts, oo o, o o o.o o o, o.o oo 77 VI Variation of EMF at AT 0 For Different Gas AtmospheresO, ooooooooo....O..oO.^Oooooo oooooooooo.v. 82 VII Thermoelectric Power Change During the Adsorption of Hydrogen and Oxygeno,.,,o,,o..o.oo,, o.aooooo..o...... 87 VIII Preliminary Runs at Constant Temperature (Run 72) and Constant Flow Rate (Run 74),.o. o O.o o. a o o...o. o. o o. 105 IX Mixed P- and NWType Catalyst Runs o,,o,,,ooo,,,,,,.o..o 105 X High Temperature Runso o o o o o,,,,, o o o.......... o 106 XI Low Temperature Runs.,.. ooo oo o....ooooo..oooooooooo..o 107 XII Adsorption Runs,,,o ooooooooooooooooooooooooo 108 XIII Co0 96Fe2 0404 X-Ray Diffraction Pattern Calculation..... 118 XIV X-Ray Flourescent Data for Catalyst Samples.............. 119 XV X-Ray Flourescent Data for Mixtures With Known Fe/Co Ratios o..O*........aOO..o. oOOOooooa..a...o o O 0.......OB. 119 XVI Analyses of Fe203 and CoCo 03, o o o o o o o DO. oo o o 120 v

LIST OF FIGURES Figure Page 1 Energy Level Scheme for Cobalt Ferrite........... 1,, 15 2 Schematic Survey of the Seebeck Effect of a Number of Compounds as a Function of Hole Concentration, n, Extrapolated to the Intersection Points with the Abscissa 22 3 Cross Section of Compressed Powder Sample.o............. 27 4 Temperature Distribution in Idealized Spherical Particles. o o. o o. o, o o o........,,, O *. * *, 27 5 Energy Picture for Adsorption,,,,.,,. *,. *....,.* 29 6 Kinetic Study Apparatus,, o,......4.. *,.,,.,,.*, **4* 30 7 Typical Data Plot,...,....o.... *......., o ~ ~ 35 8 Kinetic Study Apparatus..............,,,,,,A... 44 9 Mixing Tank Detailso........o.......4........,. o. 47 10 Reaction Vessel Details., o,,,,,..,,,,, o..,.,*,. 49 11 Thermoelectric Power Apparatus,,o,,,fi** 4.... l.,.o....,, 55 12 Thermoelectric Power Cell Details,... o.,.,..44.,... 57 13 Thermoelectric Power Apparatus: Sample Holder and Internal Wiring Details.,,... 6 O, *,*..*, 4.. 4. 6 4,..4 58 14 Thermoelectric Power Apparatusl Wiring Diagram,,. o..,.. 60 15 Catalyst Composition Calibration Curve for X-Ray Flourescent Spectrometer Data,.,,...., o,,,,,,.,* 65 16 Thermoelectric Power of Cobalt Ferrite as a Function of Composition,, o.oo4 4 b4 *4 a6 4 o * O o. 4...* S 6 67 17 Resistivity of Cobalt Ferrite as a Function of Composition4,,, 4 * a * a O * * * * 4 4s 4 4 4 O 67 18 Photomicrograph (11,5X ) of the Crushed Pellet Catalyst Particles. *. o * 6 4 O O O O 4* 44 O 4*. 69 vi

LIST OF FIGURES CONT'D Figure Page 19 The Effect of Changes in Flow Rate in Percent Conversion at Constant Temperature,,oo. oo,. o....,,.*.... 71 20 Percent Conversion as a Function of Temperature at Constant Flow Rate,, oo o o o o o o o o.o. o,..... o o. o... 72 21 A Typical Hydrogens-Deuterium Exchange Run..o...o,..... 73 22 Activation Energy as a Function of Catalyst Composition.. 74 25 Pre-exponential Factor as a Function of Catalyst Compositiono..O.Oo....... o o oo o... o.. o. n........ 74 24 Compensation Effect Between Activation Energy, E, and Pre-exponential Factor, J2ko..........o..o5.5..e 75 25 Surface Area of Ferrite Powder as a Function of Firing Time....... o o on o o o oo o 4 5 a o o o. a. A. ~. o 78 o *... 78 26 Photomicrograph (60x) of Sintered Ferrite Powder Agglomerates,,........ o o o o a a o 4.. o o o o o... o, o. o.,...... 79 27 Electron Photomicrograph (llOOOx) of Separated Particles. 80 28 Electron Photomicrograph (1600x) of Separated Ferrite Particleso a o o o * o o o 5 o* o e o o o d o * 81 29 Particle Size Distribution for Separated Ferrite Particles 5. oa... S o o oa oo 5 a o o. t o o * o o o.* 81 30 Variation of EMF of Ferrite Pellet with A T Across the Pellet, n. o-o * a a o e o... *., o.,.. *,,,.. *, 83 31 Variation of Thermoelectric Power with Time in Helium, Hydrogen, and Oxygen Atmospheres; Temperature = 2500C.... 84 32 Variation of Thermoelectric Power with Time in Helium and Hydrogen Atmosphereso Temperature = 150~C.. oo........ 85 33 Variation of Thermoelectric Power with Time in Helium, Hydrogen and Oxygen Atmospheresj Temperature = 88~ C...,. 86 vii

LIST OF FIGURES CONT'D Figure Page 34 Run Number 155a, Variation of Thermoelectric Power with Time in Oxygen and Hydrogen Atmospheres.............. 94 35 Comparison of X-Ray Diffraction Patterns and True Bulk Densities of Ferrite Materials Used in Exchange Studies and Thermoelectric Power Studies........... o...... 117 viii

LIST OF APPENDICES Appendix Page I Experimental Data,,,. o 0., o o o 4 o *..i..*.*,A., 1 04 A. Raw Data and Calculated Per Cent Conversions for Hydrogen-Deuterium Exchange Runs,,,,. o,.......,. 105 B. Raw Data and Calculated Thermoelectric Power for Adsorption Runs, 0 o **,0 0, o,, * oo,,,, 0,,, 108 II Sample Calculations. o o,,,,, o,.....o..... 112 III Comparison of Ferrite Materials Used in Exchange Studies and Thermoelectric Power Studies,,,, o.^,*.. o *.., 116 IV X-Ray Flourescent Spectrometer Analyses Data,,....,.... 119 V Chemical Analyses of Raw Materials,.,, ^,.... <o *,,..,.,. 120 VI Energy of Formation of CoFe204, *.,*. a is 0 o.,.. 0,,..*, * 121 ix

ABSTRACT The catalytic activity of the hydrogen-deuterium exchange reaction on cobalt ferrite catalysts and the change in thermoelectric power of the ferrite during the adsorption of hydrogen and oxygen were investigated as functions of the catalyst composition. The relationships between the semiconducting and catalytic properties of cobalt ferrite were analyzed in the light of recent developments in the electron theory of catalysis. The hydrogen-deuterium exchange reaction was investigated on ferrite catalysts, Co Fex04, with four different compositions: two n-type catalysts with x ) 2, Co0 93Fe2 0704 and CO~ 98Fe2.0204; and two p-type catalysts with x<2, Co1 03Fe 19704 and Co. 07Fe. 9304. The study was made using a flow reactor at approximately 750C, 1 atmosphere pressure. Analyses of the gas samples were made by means of a mass spectrometer. The activation energy increased from 19 Kcal/mole to 24Kcal/mole and the pre-exponential factor increased from 1030 to 1037 as the composition progressed from x< 20 to x>2.0. The change of thermoelectric power of compressed powder ferrite pellets during the adsorption of hydrogen and oxygen was investigated in the temperature range 88~-250~C on two samples of cobalt ferrite, one n-type (Co0 96Fe2. 404) and one p-type (Co 1. 0Fel 910 At temperatures above 1200C hydrogen was adsorbed on the cobalt ferrite as an electron donor, and oxygen was adsorbed x

as an acceptor, No change in thermoelectric power was observed during the adsorption of hydrogen or oxygen in the lower temperature (88~C) runs, indicating that little electron transfer occurs between the adsorbed molecules and the catalyst surface in this temperature range. It is concluded that the hydrogen-deuterium exchange reaction on cobalt ferrite occurs in two stages: (1) The first stage is an activation stage, in which the catalytic activity increases with time, This step might be associated with the reduction of oxygen on the surface of the ferrite, and the corresponding formation of OH and OD groups. (2) In the second stage the exchange reaction occurs with constant activity, In this step, it seems likely that exchange occurs between hydrogen and deuterium, and the OD and OH groups formed in stage (1)o This study has furnished data which indicates that the change in thermoelectric power of particulate systems due to gas chemisorption may be used to gain an insight into the electron exchange at the catalyst surface. xi

I. INTRODUCTION A. Background and Scope of Research The main objective of this research was to investigate and analyze in the light of recent developments in the electron theory of catalysis, thewrelationship between the semiconducting and catalytic properties of a semiconducting oxide catalyst, More specifically, the catalytic activity of a cobalt ferrite catalyst for the hydrogendeuterium exchange reaction and the change in thermoelectric power of the ferrite during adsorption of hydrogen gas were investigated as a function of catalyst composition. Cobalt ferrite, Co._Fexe 0 was chosen for the following reasons - 1) In the composition range lo 9 x <2,1, single phase spinels are formed with a metal to oxygen ratio of 3:4* The con. centration of lattice vacancies is small enough to be neglected in the theoretical considerations, 2) Numerous previous workers in the literature have.-imestigated oxide catalysts which are single carrier semiconductors, The catalyst, in these cases, was always p-type or n-type and only the carrier concentrations could be changed by doping with small concentrations of foreign elements, When Fe is added to CoFe204, it enters the crystal structure as Fe, which acts as an electron donor and causes the ferrite to become an n-type semiconductors When Co is added to CoFe204 it enters the crystal structure as Co II which acts as an acceptor, causing the ferrite to become a,1

-2p-type semiconductor, The opportunity is therefore afforded to study a reaction on an oxide catalyst having a single chemical substrate with either electrons (n-type catalyst) or holes (ptype catalyst) in excess, Furthermore, no foreign impurities need be added to the ferrite, 3) The bulk thermoelectric properties of polycrystaline sintered cobalt ferrite have been reported by Jonker(45). As the catalyst composition increases from 1,9 to 2,1 Jonker observed a 7 2 resistivity change from 10 to 102 ohm-cm and a thermoelectric power change from +800 to -600ov/~C. If a relationship does exist between the electrical properties and the catalytic properties of the ferrite, one might expect a corresponding change in catalytic activity in this composition range, The hydrogen deuterium exchange reaction was chosen since it was relatively simple and preliminary studies indicated that it proceeded at a conveniently low temperature. Due to the experimental difficulties involved in obtaining meaningful conductivity data in high impedance particulate systems, the thermoelectric power was chosen as the semiconducting property to be measured so that an insight could be gained into the nature of the electron transfer process taking place between the adsorbed gas molecules and the catalyst surface. Since the cobalt ferrite is a narrow band, or localized level semiconductor, no appreciable Hall effect, photoconductivity, or carrier injection effect would be expected, These effects, which are often used in studying the

53electrical properties of semiconducting materials, were, therefore, not investigated. The simultaneous measurement of thermoelectric power and catalytic activity was originally proposed. Experimental difficulties soon demonstrated that this approach was impractical. Therefore, the kinetic and thermoelectric properties of cobalt ferrite were investigated independently, Some effort was devoted to considering possible reaction mechanisms, A possible reaction mechanism is proposed which is consistent with the data in this study and with other work in the literature, Several runs were also made on mechanically mixed: no and p-type catalysts and on sintered n- and p-type catalysts to qualitatively determine if the formation of p-n junctions would affect the catalytic activity of the ferriteo B. Literature Survey The pertinent literature can be divided into two main sections, the first of which is concerned with the theoretical development and experimental verification of the role which elec. tron transfer takes in adsorption and in heterogeneous catalysis, The investigations of the hydrogen-deuterium exchange reaction on various catalysts and of kinetic studies using ferrite catalysts for various reactions pertain more directly to this thesis and are therefore reviewed in more detail in separate sections. Although much work has also been done on metals, this survey will emphasise

-4. the adsorption and heterogeneous catalysis on oxide semiconductors. The second category includes the methods of preparation, phase behavior, and semiconducting properties of cobalt ferrite. These sections of the literature are reviewed below and those articles which pertain directly to this work are discussed in more detail in other chapters of this thesis. 1. The Electron Theory of Catalysis a, Development of Theory A theoretical approach to surface catalysis was (60) considered in 1916 by Langmiur who suggested that chemical forces hold adsorbed particles to the surface. Roginskii and Schultz(80) in 1928 emphasized the electronic considerations and Rideal and Wansbrough-Jones(76) proposed a relationship between the work function of metals and the speed of catalytic reaction, De Boerl6) studied the relationship between the work function and the ionization potential of the adsorbed gas during ionic adsorption, (13) (83) (71) Brewer13) 1928, Schmidt 3, 1933^ and Nyrop (71 1935 suggested that during some catalytic reactions the adsorbed species must be present on the surface in an ionized form, Lennard-Jones(62) in his electron theory of chemisorption on metal surfaces, formulated the problem of the electron transfer process in chemisorptiono With the development of quantum mechanical treatments of solids and the application of Fermi-Dirac statistics to electrons, detailed studies of the electronic factor in catalysis on metals

-5and semiconductors became possible, Reviews of the theoretical and experimental development of the electron transfer process in metals have been given by Garner0 and Culver and Tompkins( The electron theory of catalysis relating chemisorption, catalytic activity and semiconducting properties for semiconducting oxides cannot yet be regarded as complete. At least three approaches to the problem have been proposed., The boundary-layer theory, independently developed by Aigrain and Dugas( ) Hauffe and Engell(38) and Weisz(15) emphasizes the electron transfer between the semiconductor and the chemisorbed layer, The density and energy level of the surface electrons are changed by the space charge which builds up in the boundary layer between the interior of the semiconductors and the adsorbed species on the surface; this change in surface electron charge density causes corresponding changes in the heat of adsorption and the reactivity of the chemisorbed gas, Wolkensteiin(lO9) (110) (111) in another approach, emphasizes the covalent and ionic bond formation between adsorbate and semiconductor using conduction electrons or electron holes of the semiconductor, Wolkenstein(109) (110 (11 differentiates between "weak" and "strong" chemisorption, Dowden, Mackenzie, and Trapnell (19) emphasize covalent bonding by means of atomic orbitals or electrons of the metal ions of the oxide, Recent articles by Wolkenstein (112) (113) indicate his views on the present state of the electron theory of catalysts. Rognskii (79) has attempted to formulate rules for the selection Roginskii' has attempted. to formulate rules for the selection

-6of catalysts based on the electron theory of catalysis. An attempt (52) has been made by Garrett (3 to indicate how the electron theory may be applied in a quantitative manner. Garrett's paper follows (56) the approach of Krusemeyer and Thomas to the adsorption and charge transfer on semiconducting surfaces, b, Chemisorption Studies The electron theory of chemisorption predicts a variation in the semiconducting properties of the solid surface, such as the electrical conductivity, Hall effect, and thermoelectric power, with gas adsorption. Experimental verifications of this effect have been given, for example, for the adsorption of various (5, 14, 24, 28, 40o 47, 54, 58, 69, 91, 1 101, 103) gases on ZnO (7, 8, 24, 48, 101) and u2(31 34, 47, 74) Other examples may be found in review articles by Wolkenstein(112) Parravano and Boudart(73), Winter(100), Hauffe(37) and Morrison 68) c, Reaction Studies Many experimental studies have been reported which attempt to relate the catalytic activity to the semiconducting properties of the catalyst surface, The decomposition of nitrous oxide, oxidation of carbon monoxide, and hydrogen-deuterium exchange, in particular, have been frequently used to study this effect. These investigations fall into two categories, The first group of experiments measured changes in reaction rate or activation energy as a function of the hole and electron concentration cf the catalyst, which

-7may be controlled by bulk doping, Wagner(l02) first used this technique in trying (unsuccessfully) to improve the rate of the nitrous oxide decomposition on ZnO by doping the latter with Ga203 Later experiments, however, were successful, For example, Schwab and Block(86) have investigated the carbon monoxide oxidation on Li- and Cr-doped NiO and Ga- and Li-doped ZnO, In both cases a change in activation energy with doping was observed. Similar correlations between catalytic activity and doping have been reported by Schwab et al (87) for the carbon monoxide oxidation on various mixed oxides; Molinari and Parravano(66) for the hydrogen-deuterium exchange on ZnO; Hauffe et al.(39) for the nitrous oxide decomposition on NiO; Block and Chon(lO) for the carbon monoxide oxidation on CoOs Keier et al'49) for the carbon monoxide oxidation on NiO; Otwinowska et al(72) for the dehydration of isopropanol on ZnOJ and Dogramadzi and Maticl(8) for the hydrogen-deuterium exchange on ZnO and Ni0. All of the experiments mentioned above were made on single carrier oxide semiconductors A few experiments have been conducted on two carrier semiconductors in which not only the carrier concentration but also the type of carrier may be changed by doping. Watson(l ) for instance, has investigated the catalytic activity of the Friedel-Crafts reaction on p- and. n-type germaniumn Penzkofer(75) reported that the activation energy for the dehydro,genation of formic acid on germanium was less for p-type (32 Kcal/mole) than for n-type (40 Kcal/mole), The hydrogenation

=8of ethylene on p- and n-type germanium and silicon has been reported by (54i) Krawczynski5 The activation energy was found to be 22 Kcal/mole for n-type and 3 to 6 Kcal/mole for p-type germaniumo The similar study on silicon gave activation energies of 11 Kcal/mole and 5 Kcal/mole, respectivelyo Kuchaev and Boreskov(59) reported that, at 150~C, intrinsic germanium was an order of magnitude more active for the hydrogen-deuterium exchange than were either n-type or p-type germaniumo The second category of experiments measures both the electrical conductivity and catalytic activity during the course of a reaction. This type of investigation has been made by Matveev and Boreskov (65 for the dissociation of methyl alcohol on ZnoJ by Bielanski et alo 6) for the dehydration of ethyl alcohol on various oxides3 Weller and Voltz (6)for the hydrogen-deuterium exchange on Cr203 Otwinowska(72) et alo for the dehydration of isoproponal on ZnOo Schwab 85) reported a variation of electrical conductivity with catalytic activity for various mixed catalysts. A similar investigation was also made by Alkhazov and Bielanski ) on Fe0OA12l03o Further discussion may be found in review articles by Wolkenstein 112) Law (61 Parravano and Boudart(73) ( 108) (37) (68) Winter 108) Hauffe ) and Morrison (68 do'Hydrogen-Deuterium Exchange Studies Hydrogen-deuterium exchange and reversible chemisorption have been observed for ZnO in the temperature range -190~ to 200~C by Taylor et al.(93)(9) and Harrison and McDowell(36) Heckelsberg et alo measured the simultaneous variation of con

-9ductivity and catalytic activity of the hydrogen-deuterium exchange on Li-, Al-, and Cr-doped ZnO, Molinari and Parravano(66) measured the catalytic activity of the hydrogen-deuterium exchange on ZnO as a function of Li-, Al-, and Ga-doping, The catalytic activity of chromic oxide for hydrogen-deuterium exchange after pretreatment in either hydrogen or oxygen was measured by Weller and Voltz(l). (19) Dowden et al measured the rate of hydrogen-deuterium exchange on the oxides of most of the transition metals, Dogramadzi and Matic(18) studied the effect of Li- and Ga-doping of ZnO and NiO on the catalytic activity of the hydrogen-deuterium exchange. The hydrogen-deuterium exchange on p-type, n-type and intrinsic germanium was studied by Kuchaev and Boreskov(59) Other investigations using the hydrogen-deuterium exchange on oxides are discussed in review articles by Parravano and Boudart(73) and Halpern(35) eo Reactions on Ferrite Catalysts Schwab et al. (87) reported that the carbon monoxide oxidation proceeds faster and with lower activation energy on zinc ferrite than on magnesium ferrite. Svaalenak and Scott(92) investigated the ortho-para hydrogen conversion on iron-zinc oxide catalysts as a function of increasing ferrite content. In his review on the properties of oxide catalysts based on semiconducting properties, Solymosi() included a detailed discussion of the relationship between the catalytic properties of oxide mixtures and spinels and their electrical properties, Linde et al(63) studied the catalytic

-10activity of Co-Mn spinels for the oxidation of propane. The catalytic activity for the conversion of water gas on ferrites of NiOj ZnO, CoO, MnO, and MgO was studied by Fukutome and Kusanr (29) Most exhibited weak activity, NiO-Fe203 and CoO-Fe203, however, were found to be very active after a short initiation period in which the ferrite is reduced by the water gas4 2, Cobalt Ferrite a. Preparation The general methods of ferrite preparation have (20) (21) been outlined by Economos 20 Economos2) also compared these different methods, giving the percent conversion to ferrite as a function of firing time and temperature for Ni ferrite, The effect of iron oxide particle size on Ni ferrite formation was f2pp reported by Economos and Clevenger 22 ) Methods of preparation of cobalt ferrite are outlined by Robin and Benard (78) Smiltens(89) and Jonker (5) b, Phase Behavior Robin and Benard(78) constructed a phase diagram for the Fe-CO-O systems based on x-ray diffraction data of a series of mixed iron and cobalt samplesheated at temperatures up to 1000~C. Smiltens(89) studied the 1200~C, 1400~C and 1626~C isotherms of the Fe-Co-0 systems. Smiltens(89) used the triangular diagram method. based on chemical analysis of the quenched samples and x-ray diffraction data. The data of Smiltens(89) was substantiated by Jonker(5) at a temperature of 1350~C0 Both Smiltens(89) and

^11l Jonker reported that in the vicinity of CoFe204 the ferrite has a spinel structure with a ratio of metal to oxygen ions of 5:4. c. Semiconducting Properties The semiconducting properties of transition metal oxides, including cobalt ferrite, were reported bly Jonker and (46) Van Houten A detailed analysis of the semiconducting properties of cobalt ferrite was made by Jonker(45) From measurements of resistivity. activation energy and Seebeck effect, (45) Jonker derived an energy level scheme by which the semiconducting properties of CoFe204 can be described,

IIo THEORY A. Semiconducting Properties of Cobalt Ferrite Cobalt ferrite belongs to the group of transition metal oxides whose semiconducting properties depend on a partially filled 3d band, The importance of understanding the fundamental properties of the transition metal oxides has been emphasized in recent years, (25 ) due to their increased use in heterogeneous catalysis, (4) (25) electric and magnetic circuits, and corrosion applications A review of the properties of oxides of the 3d transition metals has been given by Morin 7) In the case of a relatively simple oxide such as zinc oxide, the energy bands may arise from the filled 2p levels of the O and the empty 4s levels of the Zn++ which are broadened when the ions form the solid. The energy gap, or forbidden region, which determines the semiconducting properties of the oxides is the energy separation of the 2p and 4s bands. However, the existance of a partially filled 3d band in the transition metal oxides makes possible other energy levels in the 3d band, The semiconducting properties of 3d oxides will be controlled by the relative energy of the 3d and 2p levels of the anions and the 4s levels of the cation, Morin7) points out that calculations of the energy levels of the 3d metals(88) indicate that the 4s band is at least 10 ev wide and overlaps the 3d band which, in Ni, for example, is approximately 2~8 ev wide. Since the 3d band can hold 10 electrons as -12

-513 compared to two electrons for the 4s band, the 3d band in Ni, per electron, is only 1/15 as wide as the 4s band Morin(67) postulates that the overlap of the 3d wave function in the already narrow 3d band present in the metals is reduced further due to the increased distance between the ions in the oxide. When the 3d band becomes extremely narrows it can no longer be considered a continuous band, and the 3d charge carrier can be treated as occupying energy levels localized on the cationso The conduction in such a narrow band or "localized level" semiconductor is due to the lattice vibrations which cause the electron wave functions on adjacent cations to overlap with a higher degree of probability. Transport of this type, resulting in an extremely low mobility (10'4 to 10-8 cm2/o-sec, for CoFe204) which increases exponentially with temperature, is given by d e = d-de) e (1) kT where / = mobility d = jump length =) lattice frequency Q= activation energy for lattice deformation Due to the very low mobility, no Hall effect, photoconductivity, or carrier injection would be expected in a localized level semiconductor, For conduction to take place in the 3d levels, there must be present in pure CoFe204, some Co and FeII(hole electron pair)

14which can exchange electrons with CoI and Fe I These states, however, can exist only at high temperature. However, the room temperature semiconducting properties of CoFe2O0 can be varied by introducing excess CoIII (holes) or FeII (electrons) into the lattice, Cobalt ferrite has an inverse spinel structure, with the cobalt ions and half of the iron ions on the octahedral sites, and the remaining iron on the tetrahedral siteso Since excess Fe and Co also occupy octahedral sites, cobalt ferrite, Co 3_xex04, may be considered as the following series of mixed crystals. For x>-2 the mixed crystal consists of 3-x molecules of FeII(ColIeIII)04 and x-2 molecules of FeIII(Fe FeI )04 thus x-2 represents the FeII or electron content per molecule of the n-type ferrite, For the p-type ferrite (x<2) the crystal can be considered x-l molecules ITT ITIF IeI1 of Fe (Co Fe )04 and 2-x molecules of Fe (CoIICoII) thus 2-x represents the Co or hole concentration of the ferrite, Since the n- and p-type conductivities of cobalt ferrite can be related to the concentration of excess Fe and Co in the lattice, cobalt ferrite may be classified as a controlled-valency semiconductor [Verwey et alo (99)(10)] Note, however, that no foreign doping elements need be added to the latticeo Jonker(45) proposed the following localized band model to describe the semiconducting properties of cobalt ferrite (Figure 1 )o The following discussion (through equation 18) outlines Jonkers (45 6) treatment of the semiconducting properties of cobalt ferrite,

-15"CONDUCTION Ei -- d..t.. Fe Ef "VALENCE -> 0 E -~Co LEVEL" d q2 Figure 1, Energy Level Scheme for Cobalt Ferrite. Ef is the Fermi level (for a p-type ferrite in this case). q_ and q2 are activation energies for jumps of electrons and holes. a and P are the contribution to transport levels of electrons and holes. Energy of CoII is taken as zero. Eg is energy of FeII level. Ea and Ed are the energies of acceptor and donor levels, Since the distribution of charge carriers over the different energy levels affects both the conductivity and Seebeck effect, a brief description follows concerning the application of Fermi statistics for interpreting the measurements. The method of calculation is analogous to that for the energy band

-16scheme in normal semiconductors [i*e. Kittel (51), Mason (64) with the bands replaced by localized energy levels, For the case of p-type ferrite.s, with excess Co, the application of Fermi statistics gives Nv _ H~ I - exp ( 4)kT I+ exEa-E+)I I + exp[ (-Ef)/kT (2) (holes in valence level) - (electrons in + (electrons in acceptor level) conduction level) Since the Co concentration is small in the series of mixed crystals in this study, the number of available valence levels and conduction levels, designated by NfV and NC, will be approximately equal to the number of cobalt or iron ions, respectively, which occupy octahedral sites. Since there are eight Co ions and eight Fe ions occupying octahedral sites in the cobalt ferrite unit cell, and since the cell constant is 8o39A, it follows that NV = NC'= 1.35 x 10 cm - Since NV = NiC in the following discussion the subscript will be dropped (ioeo., NV NC N)o Equation (2), relating the hole concentration, the Fermi level, temperature and acceptor level concentration, can be simplified by making various assumptions which are valid over limited temperature ranges.. At low temperatures the last term of Equation (2) may be neglected, and the resulting equation which is quadratic in the

-17factor, exp(-Ef/kt), and may be solved for Efo When kT <( Ea, the resulting equation simplifies to F N (3) In this temperature region where the donor levels are completely ionized, but the intrinsic conduction is still negligible, the concentration of holes in the valence level, n2, will equal the concentration of acceptor levels, Na, Therefore, Equation (3) may be written Es = kTfaor - ~ NeXp ( / E kT) For an n-type material, with excess Fe' an analogous development gives E,-E = kTa, (6) or N1 = (Nexp[-(E)iE)/ (7)

18The difference between n1 and n2, fixed by the chemical composition of the ferrite, is independent of temperatures The equilibrium between holes and electrons is given by the law of mass action: 2= NZ E/kexp) E(T (8) Therefore, a cobalt ferrite with a majority carrier concentration appreciably greater than Jn-n2 will have a negligible concentration of minority carriers and the majority carrier concentration will be practically independent of temperatureO The general expression for the electrical conductivity,, for a two carrier semiconductor is Cr = n^ e/ P + t e/, (9) The variation of d- with temperature for nl >) | nln2 and n2 ) J nln2 may therefore be attributed to the variation of 1 and.A2 with temperature, Insight into the energy level scheme of cobalt ferrite may also be gained by measurement of the Seebeck effect, or thermoelectric power, A temperature difference A T between the ends of the sample causes an emf of /\AT microvolts, where ), the thermoelectric power, is given in microvolts/degree,

~19The significance of the Seebeck effect is often explained in terms of its inverse, the Peltier effect, which is a measure of the heat absorbed at one semiconductor-metal junction, and liberated at the other, when current is passed through the semiconductor, The Peltier coefficient is expressed in units of joules/coulomb, or volts. The thermoelectric power and Pelti coefficient, II, are related by the Kelvin relationship T = I (10) These effects arise from the fact that the energy level of the current carrying electron in the semiconductor differs from that of the metal, In normal band type semiconductors an additional term must be added to account for the transfer of kinetic energy of the carriers in moving from a hot to a cold region. In a localized level semiconducor oner indicates that a similar additional level semiconductor, Jonker indicates that a similar additional term must be added to account for the transport of the energy in excess of that indicated by the electron levels, He further points out that not only is it difficult to predict the magnitude of the contribution of the activation energy, q, to the transport energy, but even the sign of the contribution may be in doubt since the activation energy may be associated with either the emitting or receiving ion;. If the energies, o( and /, are associated with

-20- (45) the transport of electrons and holes respectively, Jonker5) points out that the Peltier effect for cobalt ferrite is given by -nYi (E.-Ef + + a 2 ( E4 3 P) TT - Mlu e h wz/^2e (11) In the concentration ranges where one type of carrier predominates Equation (1) reduces to ell = e T =-(E- + ) >> 2 (12) or e ell e(T= (u+/S) 1 (13) Combining Equations (5), (13) and (7), (12) gives eCT = - o- kTA Y>i 2 (14) for an n-type ferrite, and eT - + I<T..P_ N -^ - — +k2NY2 1 27 (15) for a p-type ferrite.

-21N (45) Jonker reports that both o and. / are linear functions of temperature, Therefore, Equations (14) and (15) can be solved for S to give.k f N t 2 (16) or 8=,kr1\ ~ L1 /a37 > 1 (17) In the temperature range for which Equations (5) and (7) apply, Equations (16) and (17) predict that 0 will be independent of temperature, Jonker(45 ) found the Seebeck effect for cobalt ferrite to be temperature independent between room temperature and 160~C. The difference between the band type and localized level -type semiconductors is emphasized by a plotting ~vs n, for various materials (Figure 2, from Jonker( 46) The theoretical slope, of 198 /Av/*CC per decade of concentration,is observed for all cases. -At the intersection of the extrapolated lines with the abscissa, as can be seen from Equations (16) and (17), n is approximately equal to N, the number of sites available for electrons or holes, Band theory predicts a value for N given by

-22(18) If m* = 1, at room temperature, N = 25 x 10 cm-3 The CdS(Kroger et al(55)), PbS(Bloem ) and CdTc(de Nobel(17)) curves, Group I in Figure 2, intersect the abscissa at approximately this value. The localized level semiconductor, Group II in Figure 2, including data on NiO(van Houten(97)), COFe204(Jonker 4), (44) F'e2OTiO2 and LaFeO3(Jonker ), intersect the abscissa at extra22 polated values of N 10 which is approximately equal to the number of metal ions/cc, present in the oxide, This emphasizes the different character of the two groups of semiconductors, J: I 7T i 00 Boo e GATe 2oo 400- PdS o do o6 1o, 0o14 /6 20 22 -3 Figure 2 o Schematic Survey of the Seebeck Effect of a Number of Compounds as a Function of Hole Concentration, n, Extrapolated to the Intersection Points With the Abscissa,

f235 From measurementsof both electrical conductivity and thermoelectric power, Jonker(45) calculated a set of values of the pertinent parameters which characterize the room temperature semiconducting properties of Co3. xex04O (See Table I), TABLE I, Constants Needed to Describe the Semiconducting Properties of Co3, ex045 Eg 0.55 ev o t 0.025 ev 1 kT -3 0.15 ev = 6 kT * 0*205 0.175 ev 2 e 0.51 - 0475 ev 0,7 1,3 x 104cm2/v.sec o2 = 06 - 21 10-8cm2/vsec, _D W 60.5 x 1012sec. - > 1,0 x 1014sec.1?,//~Z 10o The energy gap, Eg, of cobalt ferrite may be estimated by calculating the energy change in the valency reaction III III CoTI + Fe - CoI + FeI, E3 = I3 (co)- 3 (Fe) + c.f (19)

-24where 13 refers to the third ionization potential and. c.f refers to the crystal field stabilization energies. If 13 values from Finkelnburg and Humback (27) and cf from Jonker and von Houten(46) are used. the calculations give Eg 335549 - 30.64 - 1.5 - 1.35 ev. This value compares reasonably well with Jonker's experimental value of Eg =.55 ev, since elastic energy contributions have been neglected. Jonker and von Houten(46) also point out that, since the acceptor levels are completely ionized at room temperature in samples with an excess carrier concentration as high as 10%l Ea and Ed must be smaller than 0o06 ev, B, The Effect of Chemisorption on Thermoelectric Properties The preceeding development assumes that only one type of donor or acceptor level exists, However, the electron transfer associated with the adsorption of either donor or acceptor atoms would introduce other levels on the surface of the ferrite. These surface states, due to the nature of the bond formed between the adsorbed ions and the surface (which might be characterized by the degree of electron sharing, or degree of covalent bonding), might be at energy levels which are different from the donor and acceptor levels formed in the semiconductor due to additions of excess Fe or Co to the crystal structure. Assuming that these surface acceptor states, NB, are all at the energy level, Eb, the application of Fermi statistics in the surface region gives

-25" electrons in electrons in electrons in holes in = acceptor + acceptor + conductor valence level A level B band N - NA N NB N J Jex +exf;E _ _ L t —- _F _+ ex +e-4p [f T' pfkT' 1+ (20) At low temperatures, the last term may be neglected and the resulting equation is cubic in exp(-Ef/kT), For the temperature range in which Ea<< kT, Eb << kT, this cubic equation can be solved for Ef to give N = T- NAN (21) Equation (21) is analogous to the previous case, but the denominator now contains the total number of acceptor states, due to excess III CoI and due to adsorption, A similar solution may be derived for the n-type material. Therefore, in the temperature range where both Ea and Eb are much smaller than kT, an increase in the number of acceptor atoms adsorbed on the surface is analogous to an increase in the ZIlt number of holes (Co ), and an increase in the donor atoms on the surface corresponds to an increase in the number of electrons (FeII), Thus the Fermi level at the surface is related to both the stoichiometry and the degree of adsorption. (43) Huston points out that, although the thermoelectric power is sometimes thought of as characteristic of the bulk material, in the case of a compressed powder or sintered sample a large fraction

-26of the total temperature drop through the sample would probably be localized at the sintered "necks" or regions of particle to particle contact, Consider, for example, a compressed powder, such as illustrated in Figure 3 5 The space charge layer, formed as a result of the electron transfer between the adsorbed gas and the surface, is indicated by the shaded borders of the particles, One would not expect the temperature gradient within such a compressed powder to be uniform, If an idealized spherical particle is con(74) sidered, Parravano and Domenicali(7) point out that the temperature distribution in a collinear row of particles will resemble tie distribution given in Figure 4 o The solid lines inside the spheres (Figure 4a ) represent isothermal surfaces, and the dashed curves represent lines of heat flow. Figure 4b shows qualitatively the temperature variation along a line connecting the centers of a collinear row of particles, This model illustrates that the temperature gradient in the region of particle to particle contact may differ appreciably from the gradient farther inside the particle} causing the measured thermoelectric power to be "weighed" heavily in favor of the contact region, If this were the cases this surface layer contribution to the thermoelectric power (74) would be large, Parravano and Domenicali measured the thermoelectric power of powdered NiO under different gas atmospheres and presented a theoretical analysis showing how the ratio between the thickness of the space charge layer at the semiconductor surface and the "thickness" of the thermal gradient affects the change in

-27Figure 3. Cross Section of Compressed Powder Sample. (Shading Indicates Space Charge Region) z 2.0 IJ 2wS I I/ -.. A ~ ~ ~ ~ ~ ~~rI I W -> X, DISTANCE a. Figure 4. Temperature Distribution in Idealized Spherical Particles (74).:4~~~~~~~~~~~~~~~~~~ LLI'~::n X DSTNC Fib~~~~~~~~~~~~~~~ re.TmeaueDsrbtini daie peia Patcls(7)

.28thermoelectric power resulting from chemisorption. A similar study on the thermoelectric behavior of pellitized 60 x 104 a germanium particles exposed to oxygen and. water vapor has been made by Kmetko(52). Thus, at the surface of the cobalt ferrite the distance between the conduction band (or valence band) and the Fermi level, indicated by measurements of the thermoelectric power of a compressed powdered sample, is related not only to the composition (Fe/Co ratio) of the ferrite but also to the amount and type (donor or acceptor) of ions adsorbed on the surface. C. The Effect of Changes in Fermi Level on Adsorption and Catalytic Activity Boreskov 12) presented the following derivation of the relationship between the Fermi level and the rate of adsorption and desorption. Consider the adsorption of an atom. A, such as hydrogen, on the surface of an oxide catalyst. The heat to adsorption, Q, is Q = ( T - Ia + w), where A is the electron work function, Ia is the ionization energy, and w is a term referring to the interaction of the adsorbed atom with the surface. Let us consider the effect on chemisorption of a small change in d associated with a change in the Fe/Co ratio of our ferrite catalystO We will neglect any change in w and will, at present, not consider any changes of ( with surface coverage, The simple band picture of the p-type material is shown in Figure 5,

-29- -. Refere vce E erg T 0^^ I L E to!e I Q _ E, FAdsorbed f f Figure 5 Energy Picture For Adsorption The change in heat of adsorption will then be equal to the change in Fermi level, Efs caused by the new Fe/Co ratio, qtc2 - Qo + ^ - Qo {(22) The degree of surface coverage, assuming Langmuir adsorption, is b Pl I - bPP^ (23) where wh~ere (o/kT) (4+/kT) b - boe e (24)

-30is the adsorption coefficient and PA is the pressure of gas A. The rates of adsorption and desorption are determined by the activation energy of the absorbed complex. This energy also depends on the level of chemical potential of the electrons. Temkin( 95) has shown that a change in activation energy of adsorption forms some part of the change in heat of adsorption. j I \E Iw - ------ 0 a.'^~i Eo+ (1-Y) 4 QtQ X, DISTANCE --— > FROM SURFACE TO ADSORBED SPECIE Figure 6. Change in Activation Energy of Adsoption With Changes in Chemical Potential of the Electrons (2 In Figure 6, curve I represents the Van der Waals interaction energy of the molecule A. As a molecule of A approaches the surface it is acted upon by a repulsive force and thereby increases in potential energy. Curves II and III represent potential energy of the chemically absorbed specie as a function of the distance from the

531surface of catalysts with different compositions, ioe,, different Fe/Co ratios, in the case of cobalt ferrite, At the intersection of curves I and II (or I and III) the particle will follow the chemically absorbed curve (II or III) if x decreases, and will follow the Van der Waals curve as x increases, This intersection, then, represents the activated complex, The activation energy of adsorption is E - E -" t < EI, 1~~01'~E0~ ~(25) and the activation energy for desorption is E~ 7 E02 + (1- ) (26) where ~ is between 0 and 1. The rate of adsorption is6&0/kT, =k,P,(l-e)-=,l e AdI-e) (27) The rate of desorption is - (1- 0) ~Q/kT uz k2e ko2e (28)

-32At equilibrium Qo/kT ^r/kr POE e es,v p i ^,o kTe (29) Equation (29) indicates, that at equilibrium, as the wo-rk function increases, the rates of adsorption and desorption increase for low degrees of surface coverage, pass through a maximum at e =, and decrease at high degrees of surface coverage. (12) Boreskov applies these results to clarify the effects of displacement of the Fermi level of oxide catalyst on the activity with respect to the hydrogen-deuterium exchange reaction. If the hydrogen-deuterium exchange can be represented by S(H) + 2D -e S + HD (50) s 2 e S (H) (51) and if Equation (30) is rate controlling, the rate will be given by r- k PH^6e (32) where 0 is the fraction of the surface covered with hydrogen. Since Equation (30) is the rate controlling step, e will be given by the equilibrium adsorption isotherm corresponding to Equation (31):

-33(33) Therefore, *y 1.0^ (Q/kzT ^0)/kT Ik G k0 OP e 1 ( - [A+ k Pa, [ X + O/icr heave't's'(34) Analogous equations may be written if the role of H2 and D2 in Equations (30) and (31) are interchanged. If the hyd-rogen-deuterium exchange is represented by the reactions S()+S oD) - 25 + HD (35) 5 2- 1 9 5 (36) S + { e4 - SCD) (37) when reaction (35) is rate controlling, the rate will be given by r= k e2 (38) where e is the fraction of the surface covered. In this case,

-34Boreskov(12) points out that 9 _ v - J — (39) and r'; ~' ko6 QLe e p r cw w h is In either case, lowering the Fermi levels which is equivalent to increasing the work function, decreases the activation energy of adsorption and increases the activation of desorption, see Equations (25) and (26), Therefore, in regions of low surface coverage, ($ < $) lowering the Fermi level should increase the catalytic activity (see Equations (34) and (40), D. Determination of Activation Energy and Pre.exponential Factor The kinetic data for the reaction Hp + D2 -* 2HD were computed from the percentage conversion p, defined as /CHD3\ P. vwt,,X^-x CD 00 CH 1)] Dz ( HJ ~2; (LH9 (41)

-35Values of p were determined at three flow rates as a function of temperature, Figure 7 shows a typical data plots The method of calculating the activation energy and pre-exponential factor is outlined below, 1 Activ ation Energy (See solid lines in Figure 7)o If the rate of change in deuterium concentration is assumed to be equal to the first power of the deuterium concentration dt (42)

-36Therefore _ diDl 2], kt (43) At constant conversion the integral on the left side of Equation (43) is a constant, A. Writing k in the Arrhenius form, k = k e E/T 0 and substituting / for t, gives E/kT t A -V k e (44) where V/ = void volume of reactor in cc, _ = flow rate through reactor, cc,/min. The flow rate va, measured at ambient temperature and pressure is related to v by:,p = T (45) Therefore E/kT /q oV To P A = T /P (46) or r=Tj. t/ky +AUVTcPkoJ (1 Llk T + TC4, (47) -A'P.C

;37Since the quantity, j -,- P- ] is constant during the run, the activation energy, E, may be determined from the slope, -E/k, of a plot of.dTujjvv T ~ The procedure was repeated for three values of conversion. Note that this method of determining the activation energy is independent of the kinetic order of the reaction (ihe., the assumption of first order kinetics is unnecessary). 2o Preexponential Factor The preexponential factor was calculated from value of the flow rate and corresponding percent conversion at constant temperature (see dotted lines, Figure 7 ), If the reaction is first order, -= D K D (48) at ^~~ ~ ]o = i t (49) D2]t However, since the HD concentration at time m 0 is negligible, r D2o 2 Z Dz Et J rH t t C o,2t 2 DZ ] - pl- (50)

538Equation (49) then becomes J [ -- J~ - ft (51) Combining Equations (44), (45) and (51) gives A r___ fv ^ P e E/kT %v..L- i- PJ - ^ T e (52) Since the temperature is a constant, a plot of O -pn E<T, results in a straight line of slope m, where.oV T> P /kT eT~ PGIÆ t (53) T PII The preoexponential factor, k0, can then be calculated, since 2vi? = E/kT +T [v T -P] (54) This procedure was repeated for three values of temperature, E, Isotopic Analysis By the Mass Spectrometer The analysis of hydrogen isotopes on the mass spectrometer was developed by Bleakney( 9 ) Since this technique is frequently used in tracer studies, it has been developed to a high accuracy, and, in some laboratories, is a routine method. A detailed descrip

539tion of isotopic analysis by a mass spectrometer has been given by Kirshenbaum( 50) A brief outline of the factors which are involved is presented below, When the mixture of H2, HD and D2 enters the mass spectrometer, monotomic, diatomic, and triatomic hydrogen ions are produced (see Table II), TABLE II, Ions From a Mixture of the Hydrogens in Which H is More Abundant Than D(50) Ion Mass Intensity Dependence of Peak Height on Pressure H 1 weak aP+, blP2: H 2 very strong a2P D 2 very weak a3P+b3P2 H3 3 weak b4 2 -D 3 weak a5P D2 4 very very weak a6P HED 4 very very weak. b52 HD2 5 very very weak b6P D3 6 very very weak b7P The height of the peak at a particular mass number is proportional to both the number of molecules having that mass and the ionization efficiencies of the molecules, In the case of the hydrogen isotopes, the peak height is a function only of the number of molecules, since the ionization efficiencies of hydrogen and deuterium do not

-40omeasurably differ(50 ) (77), (96) The H2i, ED D2 mixture used in this study was approximately 98% H2. Table II shows concentration of the various ions present in the mass spectrometer as a function of pressure. The peak height I, for masses 2, 35 and 4, are, therefore given by *TZ^,~~ =^~~ taP ~(55) J = 5 P + b P2 (56) J = P (5) Since the gas is 98% H2, P I2/a or I3 - [ osQl z+[ (C+6 I:(58) or L 7L l2 (59) Therefore 3 -EHD] + rH37 I Iz EH2] c4N21 2 (60) If the results of analyses made at different pressures ame plotted, the intercept of an 3 v graph will equal D rS. rH2

.41. For a 1-2% D mixture, IflD+J << D2] and may be neglected,. Therefore l4 - i__ _D__ (61) IZ ^ ^2 r ^z] (70) NjLer, Stevens, and Rustad. indicate that it is desirable to operate the mass spectrometer at sufficient high pressures to minimize dilution or memory effects, which become pronounced at low pressures,

III. EXPERIMENTAL APPARATUS AND TECHNIQUES A, Kinetic Experiments 1. Catalyst Preparation and Analysis Reagent grade cobalt carbonate (J. T0 Baker Co) and iron oxide (General Chemical Division, Allied Chemical and Dye Corporation) were weighed out, mixed, and wet ball milled in acetone (in a Szegvari Attritor, Type SV, Size 01, Union Process Company) for four hours. An analysis of the cobalt carbonate and iron oxide is given in Table XVI, Appendix V. The sample was then dried overnight. The resulting cake was crushed into a powder, loaded in platinum boats, and fired at 10500~C for four hours in air in a Harper Model HL 7618 furnace (Harper Electric Furnace Corporation) controlled by a Brown Electronik Controlled (Minneapolis Honeywell Model 152C15PS-226-91Q2). The material was crushed with a motar and pestle and put through a No. 40 mesh screen. The resulting powder was poured into a 3/8"! diameter die, and compressed at a pressure of 20,000 psi (Buehler Ltd., Type 1315AB Hydralic Specimen Mount Press) to form pellets, 3/8" diameter, and approximately 1/2" higho These pellets were then placed on top of a layer of the No, 40 mesh powder in platinum boats and fired at 13500C. for ten hours in air in a Burrell high temperature furnace Type B-7 (Burrell Technical Supply Co, )o The samples were then quenched in air. The resulting sintered pellets were crushed and screened, and -42

-43the fractions that would pass a No. 40 mesh screen but would not pass a No, 100 mesh screen were used as the catalyst in the hydrogen-deuterium exchange studies, The surface area of the catalyst powder was measured using nitrogen adsorption at liquid nitrogen temperatures, The results were plotted according to Bo E, T, theory. The density of the catalyst was measured in a 10 cc, picnometer, Catalyst samples were made having four different Fe/Co ratios, The Fe/Co ratios were determined by comparing the catalyst pellets to three pellets with known Fe/Co ratios by means of a Norelco X-ray Flourescent Spectrometer (Phillips Electronics, Inc. Type 42202), The three pellets with known Fe/Co ratios were made from the same starting materials as the catalysts. However, they were wet ball milled in acetone for ten hours, and thenpressed at 4,000 psi into 1" pellets. The catalyst pellets were mounted in bakelite and polished with a fine diamond polishing wheel. The samples were then analyzed twelve times on each side. In order to check the homogeniety of the pellets, two samples were rough surface ground (600 A paper) and then polished with the fine diamond wheel, exposing a new layer of material, between each analysis, X-ray diffraction patterns were taken to determine the crystal structure of the powder samples and microscopic examination of the polished surfaces of the pellets were made, 2, Apparatus A diagram of the kinetic study apparatus is shown in Figure 8, The cylinder of prepurified hydrogen (Matheson Company,

TO THERMOELECTRIC MASS SPECTROMETER VENT 7 DDV POWER APPARATUS SAMPLE TUBES -mm PYREX A TUBING j| ^ HIGH VACUUM STOPCOCKS TO VACUUM < — PUMP J~ II IMERCURY MANOMETER /< -ROTAMETER L I FLOW MCLEOD GAUGE *-SAMPLE SWITCHING ECOPPER E STOPCOcK Z4TUBING DRYING COLUMN RERPRFE DEOXO 5 ~^~ ___ UNIT VENT Fi" BRASS K tCONSTANT TEMPERATURE BELLOWS SEAL /BATH NEEDLE VALVES I-RP E ^DEUTERIUM MIXING TANK HYDROGEN Figure 8. Kinetic Study Apparatus.

-45Inc ) was specified by the Matheson Company to be 999%. pure and to contain less than 20 ppm oxygen, The purity of deuterium (General Dynamics Corporation) was specified by General Dynamics Corporation to be 99,5%. The pressure of the gases leaving these cylinders was controlled by Matheson No, 1 single stage pressure regulators. The hydrogen then passed through a Deoxo Type 5-50 catalytic hydrogen purifier (Baker and Company, Inc,) which was designed to reduce the oxygen content of the hydrogen stream to less than one ppm, The oxygen content of the hydrogen at this point was below the level which could be detected on the mass spectrometer, 10 ppm. The water formed in the Deoxo unit was removed from the hydrogen stream in a phosphorus pentoxide drying column~ The gas flowed through 1/4" copper tubing with brass compression fittings and 1/4" helium leak tested brass bellows seal needle valves (Hoke, Type A433) into a glass lined mixing tank, The pressures of hydrogen and deuterium were regulated so that a 2% deuterium 98% hydrogen mixture was formed in the mixing tank at a total pressure of 60 psi, Flow of the hydrogen-deuterium mixture from the mixing tank was manually regulated by a Matheson No, 1 pressure regulator and 1/4" needle valve and was measured by a rotameter (Fischer and Porter, Precision Bore Flowrator), In the low pressure side of the flow system, downstream from the rotameter, the gas flowed through 7 or 8 mm Pyrex tubing fitted with high vacuum oblique bore stopcocks (Pyrex brand No, 7544)~ The gases then passed into the preheating coil and

a46reactor, which were contained in a constant temperature bath. A sample of the gas stream was taken as it left the reactor. The gas then passed through a soap bubble flowmeter in which the volumetric flow rate was measured. and was then vented. A vacuum pump (Cenco Hyvac-7) was connected to the system to facilitate removing air from the lines and sample tube before each run. The copper high pressure side of the system, including the mixing tank, was statically pressure tested at 80 psi and the entire system was tested with a vacuum of 104 mm Hg., measured by a McLeod fitting type vacuum gage (Scientific Glass Apparatus Company, Inc. ) The system was also checked with a helium leak detector (Consolidated Electrodynamics Corporation Type 24-210). Details of the mixing tank are shown in Figure 9, A Pyrex glass mixing vessel was chosen since the hydrogen-deuterium exchange reaction is not appreciably catalyzed by Pyrex glass at room temperature. A double walled vessel, able to withstand a pressure as high as 80 psi, was designed, The pressure in the annulus between the glass inner vessel and the steel outer shell was manually controlled so that the pressure differential across the glass did not exceed 5 psi. The outer shell was fabricated from 5" 0, D, heavy wall steel pipe, A 1/2" thick steel bottom plate was welded to the pipe. The 1/2" thick top plate was attached with eight 1 1/2" 8-32 steel bolts, The rubber gasket between the top plate and the steel shell was coated with SEAL-ALL (Allen Products Corporation) to prevent leakage, The glass inner

-471! 8-32 TO PRESSURE BOLTS REGULATOR h / r COUPLING IIO"RI -- 3AVE$/4"VECO VACUUM RUBBER GASKET PYREX INNER VESSEL PIPE NIPPLE "~^ "-~~~~',"4_: -; /STEEL | \ OUTER VESSEL 5" Figure 9. Mixing Tank Details.

vessel was held in place by means of a 3/4" Veeco vacuum coupling (Vacuum Electronics Corporation) which was silver soldered to the top plate, Gas entered the annuilus through a standard 1/4" pipe fitting, The annulus gas pressure was measured by a 0-100 psi pressure gauge (U, S. Gauge Co. ), The pressure of the hydrogendeuterium gas mixture in the inner glass vessel was measured by the pressure gauge in the Matheson No, 1 pressure regulator which was attached to the top plate by means of steel pipe fittings, Details of the constant temperature bath and reactor vessel are shown in Figure 10. A 2000 ml. stainless steel beaker was clamped to an aluminum support rod, An asbestos shell, with walls 1 1/2" thick, was fabricated to fit snuggly around the stainless steel beaker0 This insulating shell, attached to the aluminum support rod with an adjustable slide, could be removed to facilitate rapid cooling of the bath, A 1 1/2" thick asbestos lid was constructed. The stainless steel beaker was filled to approximately 1" of the top with Dow Corning 550 heat stable silicon fluid, The current input to the 250 watt knife type immersion heater (Cenco Cat, No, 16551) was regulated by a gas thermometer type thermoregulator and relay (Supersensitive Relay No, 4-5400, American Inst. Company, Inc.)o Agitation was provided by a variable speed stirrer (Eastern Industried Inc, Model 3), The bath temperature was measured with a Chromel P-Alumel couple in conjunction with a Leeds and Northrop type 8662 portable potentiometer, The hydrogen-deuterium gas mixture was preheated to the

-49TO SAMPLE TUBES STIRRER f KNIFE HEATER GAS INLET PREHEATER I -I / y/ rT TEMPERATURE ASBESTOS —-- - - CONTROL BULB LID - REACTOR ALUMINUM SUPPORT ROD / \ 1\ \{ ~- GLASS WOOL REMOVABLE --— C C ATALYST ASBESTOS INSULATION 1I2 THICK A / 20 mSTAINLESS STEEL \ ADJUSTABLE- 2000 ml BEAKER L SILICONE BATH FLUID SLIDE Figure 10. Reaction Vessel Details.

a50bath temperature in a coil of 7 mm Pyrex tubing and then passed through the reactor, The reactor was a section of Pyrex tubing, 25 mm long, with the diameter enlarged to 10 mm, A slight constriction in the line below the reactor supported a small wad of glass wool which, in turn, supported the catalyst, A small wad of glass wool was also inserted on top of the catalyst to prevent the gas from blowing the catalyst out of the reactor, The temperature controller was essentially a gas thermometer with two electrical contacts in the mercury column, which activated a supersensitive relay, regulating the current to the bath heater. The temperature setting could be controlled by varying the height Of mercury in the pressure leg. In order to avoid oxidation of the mercury at the upper electrode, hydrogen was used to fill the bulb, 35 Experimental Procedures Hydrogen and deuterium were introduced into the mixing tank, The pressures were regulated manually during the filling operation in order to achieve a final pressure of 60 psi, with 2% deuterium, 98% hydrogen in the gas mixture, A weighed amount of catalyst was placed in the reactor and a gas mixture flow of 30 cc/minm was started through the reactor, After flowing at this rate for 1/2 hour to flush air out of the system, the flow was reduced to 10 ce/min, and the reactor temperature was raised to 200~Co The catalyst was activated for twelve hours at 200~C, The

-51bath temperature was then lowered to 70-75~C, and held at this temperature for at least four hours. The catalyst had now been activated and the run was started, The gas flow rate was measured on the soap bubble flowmeter, and the reactor temperature was noted. The stopcocks on the mass spectrometer sample tube were closed, and the sample switching stopcock was turned, diverting the gas flow through the second sample tube. The gas sample was removed and an empty sample tube was put in its place for the next measurements The experimental conditions (temperature and/or flow rate) were then changed to the next set of desired values. While the system was coming to a new steady state (15-20 minutes), the gas sample was analyzed on the mass spectrometer, As the run progressed and the pressure in the mixing tank dropped, gas was periodically bled from the annulus so that the pressure differential across the glass inner vessel would not exceed 5 psi, In order to avoid air contamination of the premixed gases, the pressure in the mixing tank was not allowed to fall below 15 psig, 4J Analytical Methods The analysis of the gas samples were made on the mass spectrometer (Consolidated Engineering Corporation Type 21-0135B Modified to Type 21-103C specifications)0 A quantitative analysis of a gas mixture is, in general, obtained by comparing the sample cracking pattern against standard samples of the pure components,

-52However, since the sensitivity of the mass spectrometer for hydrogen and deuterium are approximately equal (1356 div// for hydrogen, 1355 div/ for deuterium)( 3 ) in this special case the ratio of peak heights will equal the ratio of partial pressures, If the sensitivity of the spectrometer for hydrogen deuteride is assumed to be also 135.5 div> 4 then the ratio of the peak heights for all three gases in the mixture will equal the ratios of their partial pressures and no standard samples need be run. The validity of this assumption was verified by comparing the mass balances of samples taken before and after reaction0 More specifically, the %D in the unreacted H2 + D2 mixture was calculated from the mass spectrometer analysis. The sample was then passed over the catalyst and the resulting H2 + HD + D2 mixture was analyzed on the mass spectrometer. The calculated %D, based on an HD sensitivity of 13.5 div//, was then compared to the %D calculated for the unreacted gas. The most accurate method for determining the composition of H, + HD + D2 mixtures, described in the theory chapter, requires that each sample be analyzed four or five times, at different pressures, Since, in following the course of the exchange reaction, a new sample was taken each 15-20 minutes, it was impossible to make multiple analyses of each sample. Instead, all of the samples of any one run were analyzed only once, and each analysis was made using the same gas pressure in the mass spectrometer, This technique would introduce a consistent error (approximately +3%) in the HD

-53and D analyses which could be eliminated during the calculation of 2 the activation energy andpre-exponential factor, since these calculations involved only differences in percent conversions. (See sample calculations, Appendix II ), Bo Thermoelectric Power Studies 1. Catalyst Preparation and Analysis In order to produce a high surface area ferrite powder for the thermoelectric power studies the following preparation procedure was used, The reagent grade iron oxide and cobalt carbonate were weighed. mixed and then wet ball milled in acetone for four hours, The sample was dried overnight at 120~C, and fired for four hours at 10355~C in a Harper Model HL7618 furnaceo The material was crushed with a mortar and pestle and put through a INo. 40 mesh screen, It was then mixed on a rolling mill for seventeen hours. The powder was placed in mullite combustion boats, McDanielj high temperature mullite, fired in air for eight hours at 1100 C, (Harper Type HL7618 furnace), air quenched, crushed in a mortar and pestle and passed through a No,, 40 mesh screen, The powder was then poured into a 3/8" diameter die and compressed at 2,000 psig, into pellets 3/8" in diameter and approximately 1/2" long which were used in the thermoelectric power studies, The Fe/Co ratio of the powder was determined by chemical analysis, The ferric ion was precipitated from 10% H2SO4 solution by cupferron, the ammonium salt of nitrosophenylhydroxyamine, separating the iron from the cobalt, The iron was

-54then reduced with stannous chloride and titrated with standard potassium dichromate, The cobalt was weighed as cobalt sulfate, obtained by evaporating the oxide with H2SO4 and igniting at 5500C. X-ray diffraction patterns were taken in order to determine the crystal structure of the material, The surface area of the powder was measured using nitrogen adsorption at liquid nitrogen temperatures. The results were plotted according to B. Ed To theory. The partial size distribution of the ferrite powder was determined from electron photomicrographs (RCA EML Electron Microscope) at 110OX magnification, The density of the powder was measured by means of a 10 cco picnometer, 2. Apparatus A diagram of the thermoelectric power apparatus is given in Figure 11, The gas inlet manifold was constructed with 7 mm Pyrex tubing, and high vacuum oblique bore Pyrex stopcocks (Pyrex Cat, No. 7544). The prepurified hydrogen used in the hydrogen-deuterium exchange study was also used in this work. The helium (General Dynamics Corporation, 99.5% pure) pressure was regulated by a Matheson No 1 single stage regulator. The helium was purified by passing it through hot copper oxide 350~C, and activated charcoal cold trap, cooled with dry ice in isopropanol, Pressure measurements were made with a McLeod tilting type vacuum gauge (Scientific Glass Apparatus Company, Inc. ) and a

HIGH VACUUM OBLIQUE- BORE 7mm PYREX TUBING STOPCOCKS TO VACUUM PUMP AND HYDROGEN CYLINDER - THERMOELECTRIC POWER CELL TO MCLEOD FROM OXYGEN GAUGE CYLINDER MERCURY MANOMETER I FROM HELIUMCuO CYLINDER / \350~C i -— / ACTIVATED CHARCOAL / lX --- ISOPROPANOL-DRY ICE COLD TRAP ASBESTOS INSULATION Figure 11. Thermoelectric Power Apparatus.

-56mercury manometer, The details of the thermoelectric power cell are shown in Figures 12 and 135 A modified Dresser coupling (Dresser Industries, Inc., Style 65) was used to support the sample holder and Pyrex cover~ Tightening the ends of the Dresser coupling forced a copper sleeve to expand the "0" ring, forming a vacuum tight seal between the coupling and the sample holder and Pyrex cover. The vacuum system and purification train were attached to the cell by an "0" ring seal in a 5/16" Veeco vacuum coupling which was silver soldered over a 5/16" hole drilled in the coupling, The electrical leads passed through Kovar tubes The primary coil, made from Noo 26 gauge Chromel A heating wire, was wound on the Pyrex envelope and covered by approximately 1/2" of asbestos insulation, The current to this coil was controlled by a variable transformer (Variac Type 200-CM)o Details of the sample holder and electrical wiring are shown in Figure 13 The sample holder was mounted on a stainless steel support which was bolted to the brass sample holder mount, The sample was mounted between two 20 mil, platinum discs (Baker Platinum Division of Englehard Industries), supported by 2 mm, bore capillary Pyrex tubing. These capillary tubes were mounted inside hollow stainless steel cylinders which were welded to the sample holder support0 When the sample had been placed between the discs, a stainless steel knob, threaded into the outer cylinder,

-KOVAR GLASS TO BRASS SLEEVE METAL SEALS -0 RINGS PYREX ENVELOPE -HIGH VACUUM,\\~~~ ~~ ~ /~ ~~ \ / / \STOPCOCK STAINLESS STEEL _-SAMPLE SUPPORT ELECTRICAL — _ LEADS \J FACE PLATE BRASS SAMPLE \2 ASBESTOS INSULATION HOLDER MOUNT \ MODIFIED-^^ i ^ ** \PLATINUM CONTACTS DRESSER COUPLING, VEECO VACUUM TYPE 65 COUPLING PRIMARY HEATER COILS TO VACUUM SYSTEM AND GAS PURIFICATION TRAIN Figure 12. Thermoelectric Power Cell Details.

PLATINUM SEEBECK EMF LEADS SECONDARY HEATER LEADS CHROMEL P-ALUMEL 20ml PLATINUM DISCS THERMOCOUPLE LEADS 2mm BORE PYREX CAPILLARY TUBING 2 HOLE CERAMIC _._ t +-^ -- 7 15 / —- - -.Z - /1 / INSULATOR TUBE INSULATOR__-______ _ A // _ / CERAMIC - - - --- ^|' __i== --- U TUBES 1 -STAIN-LESS STAINLESS SAMPLE SPRING STAINLESS CHROMEL P-ALUMEL STEEL STEEL PELLET STEEL THERMOCOUPLE SUPPORT CYLINDER TIGHTENING LEADS KNOB Figure 13. Thermoelectric Power Apparatus; Sample Holder and Internal Wiring Details.

-59was tightened, firmly clamping the pellet in place. The secondary Chromel A heating coil, controlled by a variable transformer (Variac, Type 200-CM), was wound on the inner glass capillary tube. A No. 40 Chromel-P-Alumel thermocouple (Hoskins Manufacturing Co.) and a No. 30 gauge platinum lead (for EMF measurement) were spot welded on each platinum electrode on the side opposite the sample. The Chromel-P and Alumel leads passed through a 2-holed ceramic insulating tube mounted in the bore of the capillary tubing. The platinum leads were threaded through a similar ceramic insulator which was mounted outside of the stainless steel cylinders. The eight leads (4 thermocouple wires, 2 platinum EMF wires, and 2 secondary heater wires) passed through the brass sample holder mount and through the Kovar glass seals in the face plate. Details of the external electrical wiring are shown in Figure 12, The Chromel-P-Alumel thermocouple wires lead to the 0~C. cold junctiono Copper wires then connected the thermocouple wires to a thermocouple switch (Leeds and Northrup Ten Point rotary switch Cat, No, 8240). The output terminals of this switch were connected to the "B" contacts of a double pole, double throw thermal free switch (Leeds and Northrup Cat. No. 3294)o The inner contacts of the DPDT switch were connected to a potentiometer (Leeds and Northrup Type K-3) and electronic DC null detector (Leeds and Northrup Cat. No, 3834). The platinum EMF wire lead to the inner terminals of a similar double pole, double throw switch, The "A" terminals of

HEWLETT- LEEDS a PACKARD NORTHRUP STANDARD MICROVOLT- D.C. NULL CELL AMMETER DETECTOR BATTERY I9 9 L 9 I0 Q., I SHORTING SWITCH DPDT THERMAL FREE R kA C A o ALUMELSWITCHES LEEDS a NORTHRUP L ~~~TYPCOLD JUNCTIOETER Figure 14. Thermoelectric Power Apparatus, Wiring Diagram. \^^y CHROM E L-P — LEEDS 6 NORTHRUP TEN POINT ALUMEL [ THERMOCOUPLE SWITCH SAMPLE -COLD JUNCTION Figure 14. Thermoelectric Power Apparatus, Wiring Diagram.

m61l the two switches were then connected through a DC micro voltammeter (Hewlett Packard Model 425A). A switch was provided to short out the micro volt-ammeter for zeroing purposes. As Figure 14 indicates, when both double pole, double throw switches were connected to terminals tA"t and the null detector was switched off, the EMF between platinum discs was measured through the platinum leads by means of the micro voltammeter, (-used as a null instrument) in conjunction with the K-3 potentiometer, When both double pole, double throw switches were connected to terminals "B", the temperatures on the platinum discs were measured through the Chromel-P-Alumel couples and thermocouple switch, by means of the DC null detector and K-3 potentiometer. 3., Experimental Procedures To minimize the electrical contact resistance, the ends of the ferrite pellet were first coated with graphite by rubbing with a pencil~ This technique was used by Jonker ) ( 98) Van Uitert reported that indium-mercury and graphite form good electrical contacts on ferrite, The ferrite pellet was clamped between the platinum discs and the Pyrex envelopeput in place, The Dress:er -fitting was then tightened so that the "0" ring made a vacuum tight seal on the Pyrex envelope, The cell was evacuated to 10 4 mm Hg, It was then filled with purified helium and evacuated to 10-1 mm Hg. This cycle was repeated five times in order to flush any residual air

-62w from the cell. The cell was then filled with helium and the primary heater coil voltage was set at the desired value, The secondary heater voltage was adjusted to impose a temperature difference of 1-3~ C across the pellet. The EMFs were then measured with a Leeds and Northrup Type K-3 potentiometer, coupled with the DC electronic null detector (Leeds and Northrup Cat, Noo 3834) for temperature measurements (Chromel-P-Alumel couples), and coupled with the Hewlett Packard Model 425A micro volt-ammeter (used as a high impedence null detector) for the thermoelectric EMF measurement (platinum-ferrite couples.) The resistance of the sample and platinum leads was measured with anohm meter (Triplett, Model 630, The helium pressure was then reduced to 30 cm, Hg, and 2-6 cm, of hydrogen gas (or oxygen) was introduced, The pressure was increased to 76 cm, Ho with prepurified helium, The temperatures, thermal EMF and resistance were recorded as function of time. The measurements were continued until steady state was reached. The cell was then evacuated (10-4 cm), flushed five times with helium, and a helium-oxygen (or hydrogen) mixture was introduced and similar measurements were taken, Catalyst composition, gas composition and temperature were the independent variables of this study, C. Experimental Program The hydrogen-deuterium exchange reaction was initially studied in the 110-130~Co temperature range, using an eight hour

!63activation time at 200~C, Four different catalyst compositions were used, two p-type and two n-type, Per cent conversion data were taken as a function of temperature at two different flow rates, The activity of the catalyst did not remain constant when data for a third flow rate was taken. In order to avoid this effect, all subsequent exchange data were taken in the 55-75~C, temperature range, using a twelve hour activation period at 200~C, Using this technique, reproducible kinetics data were taken, as function of temperature, using four catalyst compositions (two p-type, two n-type), three flow rates (10, 15, 20) cc/min The thermoelectric power studies, investigating the direction of the change in Seebeck coefficient during gas (hydrogen' and oxygen) adsorption, were made in order to gain some insight into the electron transfer process during adsorption. These studies were made as a function of temperature and gas composition, using two catalyst compositions, one n-type and one p-type. Four hydrogen-deuterium exchange runs were also made on a series of mixed, sintered, p- and n-type ferrite catalysts, to see what effect the formation of p-n junctions in the catalyst would have on the kinetics of the reaction.

IV. EXPERIMENTAL RESULTS Ao Hydrogen-Deuterium Exchange Studies 1. Ferrite Catalyst Characterization Cobalt ferrite catalyst samples, Co 3xFe 04, having four different compositions were prepared in pellet form, two with x > 2.0 and two with x < 2.0o The Fe/Co ratios of these samples were determined by comparing the catalyst pellets to samples with known Fe/Co ratios by means of an x-ray flourescent spectrometero Tables XIV and XV, Appendix IV, shows the Fe/Co ratios and standard deviations which resulted from the twenty-four measurements made on each known sample and each catalyst pellet. Figure 15 is the catalyst composition calibration curve determined by the three samples with known Fe/Co ratios of 0o9/2.1, 1.0/2.0, and o1/1. 9o The resulting catalyst compositions, COo Fel 0Q4 Co Fe 704, Co Fe 04, and Co Fe 04 are lo~07 l0 loo03 197T4 o0.98 2.02 4 0.95 2.07 also indicated on Figure 15 o In order to check on the homogeniety of the catalyst pellets, catalyst sample Co0 98Fe2 0204Y was surface ground between each of the twenty-four x-ray analyses in order to expose a new layer of material to the spectrometer~ As Table I1 indicates, no significant difference was observed in the standard deviation of this sample as compared to the standard deviation of the three other catalyst samples, in which the twenty-four measurements were repeated on the same surface. The thermoelectric power and resistivity of the four catalyst pellet samples were measured. In Figures 16 and 17 the results of these measurements are compared to similar data reported by Jonker~(45) these measurements are compared to similar data reported by Jonker() -64~

-65-'.3 tlo Ie,1' CO00.93 Fe2.07 04 1.20o 0.98 Fe2.02 04 o0 LL. w w Co Fe 001.03 1.97 04 i w 0 I 9 Flourescent Spectrometer Data.: Co1.07 Fel.93 04 O CATALYST SAMPLES 0.9 ---— I -- l - 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 MOLES FMOLEs Co Figure 15. Catalyst Composition Calibration Curve for X-Ray Flourescent Spectrometer Data.

-66Note that in both Jonkerts data and the data of this study an extreme change occurs in both the electrical resistivity and thermoelectric power as the catalyst composition passes through x = 2.0. For x ( 2.0 a high resistivity, p-type ferrite resulted, while for x > 2,0 a lower resistivity, n-type ferrite was observed, No variation in thermoelectric power or resistivity during the adsorption of hydrogen or oxygen was observed for these pelletized, sintered samples after 48 hours at 250~C, TABLE III. Comparison of Standard Deviations of X-Ray Analyses of the Four Catalyst Samples Catalyst Side Peak Height Ratio Standard Deviation Sample Fe/Co Co1 07Fel. 90 4 1 1,001 0 0128 2 0o986 0.0118 Co1o 02Fel 9804 1 1.055 0.0114 2 10043 0,0022 Co0 93Fe2.0204* 1 1137 0 0128* 2 1,140 0o0141* Co 093Fe2 0704 1 lo232 0.0139 2 1.222 0.0125 *Surface ground between measurements A microscopic examination of the polished surfaces of the catalyst pellets, and x-ray diffraction patterns of the samples of

- JONKER'S DATA- JONKER'S DATA (ROOM TEMPERATURE) (ROOM TEMPERATURE) 0- DATA OF THIS WORK 0- DATA OF THIS WORK (600~C) (600C) 800=1 400- 7- W 200- 6 o X w I -200 - --------- Co —-s —i —- 4 — w -400 H 3o /w /U -- ---- ---- — I- _ —---- -0 —- - -800 -_ 01.90 1.95 2.00 2.05 2.10 1.90 1.95 2.00 2.05 2.10 X X Co3-x Fex 04 C03.x Fex 04 Figure 16. Thermoelectric Power of Cobalt Figure 17. Resistivity of Cobalt Ferrite Ferrite as a Function of as a Function of Composition. Composition.

-68crushed pellets indicated that single phase spinel compounds were formed for all four compositions. These results are in good agreement (45) with reports of the phase behavior of cobalt ferrite by Jonker Smiltens(89), Robin and Benard(78) and Roiter and Paladino(81 ). A typical x-ray diffraction pattern is given in Figure 35a, Appendix III. All of the lines, except two, in the x-ray diffraction pattern shown in Figure 35a may be attributed to the diffraction of Co radiation on a spinel structure, The two extra lines are caused by iron contamination in the cobalt target in the x-ray diffraction tube, This error may be checked by applying a correction factor equal to the ratio of the wavelengths of the iron and cobalt radiationo (See footnote, Table XIII, Appendix III. ) Sample calculations of "d" values for the x-ray diffraction pattern are given in Table III, Appendix XIII. The catalyst pellets were crushed and screened. The fraction which would pass through a No5 40 mesh Tyler screen but would not pass through a No, 100 mesh Tyler screen was used as the catalyst for the hydrogen-deuterium exchange studies, The surface area of this crushed powder, measured in a B. E, T. apparatus, was found to be 0,08 m2/gm. No variation in surface area, within the experimental accuracy of the B. E, T, measurements, was observed between the different catalyst samples. Figure 18 is a typical photomicrograph of the crushed catalyst particles, As would be expected due to the screening procedure employed, the particle sizes lie in the range 0.15 to 0.5 mm.

.69Figure 18, Photo!micx-norajh (l, 5x) of the Crushed Pellet Catalyst Particles The density of the ferrite catalyst, measured In a 10 cc. picnometer~ was 5,2 gl/cc, 2, Exctnge Ruins Preliminary runs indicated that for given flow rate and temperature, the per cent conversion would remtain constant as a function of time (the constant flow was continued for as long as four hours), after the catalyst was activated for 12 hours at 200~C

-70in a hydrogen atmosphere, It was further found that activated samples would be rapidly deactivated if exposed to air or oxygeno Preliminary runs were also made to test the variation in per cent conversion with flow rate and temperatureo Run 72, shown in Figure 19, illustrates that consistent values of the per cent conversion were observed as the flow rate was alternated from 20 cco/min. to 37 cc,/mino at a constant temperature of 126~C. In Figure 20, Run 74 indicates that at a constant flow rate (20 cc./min ) the per cent conversion observed at decreasing temperatures was the same as that observed when the temperature was increasing0 A plot of the data from a typical hydrogen-deuterium exchange experiment is shown in Figure 21 o A summary of the activation energies and pre-exponential factors for the hydrogen-deuterium exchange experiments are given in Table IVo The experimental data used in calculating Table IV is given in Tables X and XI, Appendix IA In Figures 22 and 23 the activation energy and preexponential factor, respectively, are plotted as a function of the Co3_xFexO4 composition variable, xo The activation energy increases 4 to 5 Kcal/mole and the pre-exponential factor increases 6 to 7 orders of magnitude as the composition progresses from x < 2.0 to x > 2.0o.This compensation effect is emphasized in Figure 24, which is a plot of activation energy vso X (ko)o

-710.5...l 0.4 -E - 00 0 0 ~ 0 E 000 0o3 - z A AA AA'A 0 - 0.2 o.-1 RUN: 72) 0.1 - RUN: 72 0 20 cc/min A 37 cc/min TEMPERATURE 126~ 0 IiI0 100 200 300 400 500 TIME, MINUTES Figure 19. The Effect of Changes in Flow Rate on Percent Conversion at Constant Temperature.

-720.5 4*. %o 0.4e 0.3 z 0 0.2 0 0 RUN: 74 0.1 * TEMP DECREASING o TEMP INCREASING FLOW RATE 20 cc/min 0 120 130 140 150 TEMPERATURE, ~C Figure 20. Percent Conversion as a Function of Temperature at Constant Flow Rate.

0.6 0.5 ~7/ / 0.4 X C, w 0.3 0.2 0.1 - RUN: III CATALYST- Col3 FeL97 0 I 50 60 70 80 TEMPERATURE ~C Figure 21. A Typical Hydrogen-Deuterium Exchange Run.

-742524 o w 0 2322 0 21 X 20 z 11W 0 19. Z S * 750C RUNS 18 5>~~~~~~ ~~~o IIO~C RUNS 17,,, 1.90 1.92 1.94 1.96 1.98 2.00 2.02 2.04 2.06 2.08 2.10 2.12 C03-x Fex 04 Figure 22. Activation Energy as a Function of Catalyst Composition. 40 39, 38- 5 3736 35 0 34 33 - 32 — 31 - z w 30- - 0 29 n 26 o 110OC RUNS a- 26 25- - -, ------ 1.90 1.92 94 1.96 1.98 2.00 2.02 2.04 2.06 2.08 2.10 2.12 X Co3-x Fex 04 Figure 23. Pre-Exponential Factor as a Function of Catalyst Composition.

38 36 -'34 __ oc 0^ ^^^ 0r~~~~~~~~~~~~~~~~0 - 32 - (3 IL~ LJ z o a. tU 28 - ~ 75~ RUNS cr a. o / y^0~ I Il0o 1|10~ RUNS 26 -- 17 18 19 20 21 22 23 24 25 ACTIVATION ENERGY K CAL/MOL Figure 24. Compensation Effect Between Activation Energy, E, and Preexponential Factor, Ln ko.

m76. TABLE IV d Summary of Activation Energy and Pre-exponential Factors for Hydrogen-Deuterium Exchange Experiments 1100C Runs 75~C Runs Act, Pre-exp. Act. Pre-exp. Run Energy Factor Run Energy Factor Sample _ Type No. Kcal/mole ln(kN) No, Kcal/mole ln(k^) Col 07Fel 9304 p 76 19 4 29.2 108 18.3 30.1 77 19o 9 30,2 110 18.5 30 2 78 18,4 27.6 Col 03Fe 19704 p 80 18.7 27.5 111 19.0 30.8 81 18 5 26.9 112 18.6 30.1 82 18.8 27,2 Co0 98Fe2 020 4 n 83 23.2 33. 4 120 23.9 37.8 84 23.0 5330 121 2351 36.1 85 22.9 33o0 86 22.3 32,0 Co0 93Fe 20704 n 89 255 335 8 123 25.4 37 The results of four runs made on mixed, sintered, p- and n-type catalysts are given in Table V o The activation energies and pre-exponential factors calculated for these runs, which were made to see what effect the formation of p-n junctions in the catalyst might have on the exchange kinetics, do not differ appreciably from the results for the unmixed catalyst listed in Table 1V V

,77TABLE V o Activation Energies and Pre-exponential Factors for Runs Using Mixed Sintered Catalysts Pre-exp. Sintering Run Act. Energy Factor Sample Time (hr) Noo Kcal/mole ln(ko) 50% Col 07Fel 9304 0 125 20.7 33.5 50% Co.93Fe2 074 1 hour 126 216 34.4 850oC 1 hour 127 21,1 34.2 990 C 1 hour 128 22.1 34.3 11200C Bo The Effect of Chemisorption on Thermoelectric Power 1, Cobalt Ferrite Characterization Cobalt ferrite powder having two different compositions, one with x ( 2,0 and one with x > 2,0, were prepared. The compositions of these samples, determined by means of wet chemical analysis, were ol.09 Fe1.910 4 and Co o96Fe2o0404 X-ray diffraction patterns of the samples indicated that a single phase spinel compound was formed in each case, A typical x-ray diffraction pattern is presented in Figure 35B, Appendix III, All lines in Figure 35B are characteristic of the spinel structure except for the two extra lines caused by the iron contamination in the cobalt target of the x-ray tube (see footnote, Table XIII Appendix III )o

-78The surface area of the ferrite powder was measured by B. E. To techniques. Figure 25 shows the surface area of the ferrite powder as a function of the firing time. 2 \ <rf %3 k3 0 6 /2 18 24 F-IRtM5G 7i bE H OL ORs Figure 25. Surface Area of Ferrite Powder as a Function of Firing Time (1035~C for 4 hours followed by 1100~C for 20 hours) The samples used in the chemisorption studies were fired for four hours at 1055~C followed by eight hours at 1100~C, which corresponds to a surface area of 2m /gmo Figure 26 is a photomicrograph of the resulting sintered agglomerates of small particles. These agglomerates were separated by means of a mortar and pestle in accordance with the techniques outlined by Schuster and Fullam () Figures 27 and 28 are electron micrographs of the resulting separated particles. From eight

.f79.. electron microgx apls at 3l.OO0x magni:fl catison, i3..nc.Idxmlg th.e micro. grzaph.n F:itu.re 2'y te part t;-ic2X scze diz distr.tt')btioA0 o:.7 the ~er I:te wits obt3tained.o t.h1isx pta tL'cVle size d.istr bjlp)tirb.oxn'e.J.s n at n ox "tmal.., ctistrnibtxioxlot when t.. p.jx)tlted, on a logariXtmitct' scaele' (see Figurte 29 ) birit gi-tes a sRkewed. ctxive wthen tpl.ot'ted on a.lin ear sf cale, tPhis "isk;rewed" di.:ltrtlluti.on, found. l in previ.ous woxrk (7(, i p troatJbly dUBle to Jtte.X \.g of tfhe ferArite peoder.5 vigxre 26, PhotoMicrograph ( 6x) of Siyitered Fevr.ite Po~wer Agglmofferat e s. 4 ~ > A~gl]oine.va'tes$

0~ 0 p~ p 4, Ai,111 ii-0 rS~~~~~ % f >14.0i~~~~~~~ C A~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Al ~~~~~~~"t;~ ~ ~ ~~~~~~~~~~~~V --- -- -- --- - -- - - -- -- ~~~fJ~~~~~~~SA~ ~ ~ ~ X 3~~~~~~~~~~~~~~~~~~~~cv 0 C''-5

G0 81-? 20 tO 0 — Figure 29. Particle Size Distribution For Separated Ferrite Particles The density of the ferrite powder, measured by means of a 10 cco picnometer, was 5.3 gm/ccs If a spherical particle shape is assumed, the average particle size can be estimated by where A is the surface area and.o the solid density. This value agrees quite well with the value determined electronmicroscopically, 2, Thermoelectric Power Measurements During Chemisorption of Hy rogen an. Oxygen Gases -- Preliminary runs were made to test the variation of emf with / T and the emf at zero A T under varying gas atmospheres.

-82Figure 30 in which the emf is plotted for Co0 93Fe2 070 4 for eight different A TIs, illustrates that no variation in thermoelectric power was observed for different A T values, Table VI illustrates that the emf's produced by temperature differences near A T = 0 is constantly small in helium, hydrogen, and oxygen atmospheres, The A T values in this table, given in millivolts (Chromel-Alumel couples) are all less than 0002~C (ie,,.01~C =,0004 millivolts)o This table indicates that the intercept of the emf vs0, T curve, the slope of which determines the thermoelectric power, does not appreciably vary in helium, hydrogen, or oxygen atmospheres TABLE VIo Variation of EMF at A T = 0 for Different Gas Atmospheres Helium Atmos, Hydrogen Atmos, Oxygen Atmos, Sample AT emf X T emf AT emf Co, oFel 9104 0003 0,0021 0000 0,0310 0,0007 Oo0112 Coo 96Fe2 o04 Oo 00001 03,002 0,0005 0.0092 0.0003 0o0412 Typical plots showing the change in thermoelectric power during gas adsorption are given in Figure 30 (high temperature 250~C), Figure 32 (intermediate temperature 150~C), and Figure 33 (low temperature 88~C)o A summary of the results of all of the thermoelectric power runs is given in Table VII o Raw data for these runs are presented in Tables XII, Appendix IB.

-83+ 5.0 I EMF, 4.03.0 2.0 ____ _________ __ __ -31.0o-AREA _ __PA _ __ _ - -4 - 2.0 — -0.7 -0.6 -0.5 -0.4 o-0.3 -0.2 -.1 +.i 0.2 0.3 0.4 0.5 0.6 0.7 -ot-ohe - o,,,/_____ AT, MILLIVOLTSP ---- --- --- --— 1.0 — - Cr-AI COUPLES _-3__.0. AREA EXPANDED IN GRAPH BELOW -4.0-5.0 EMF,_ - +0.5- MILLIVOLTS 0'3 0.4 ----- -- -0.3 —- 0 -- -0.08 -0.07-0.066-005 -004 -03-0X32 01 + 0.01 0.02 0.013 0.04 0.05 0.06 0.07 0.08 |. _ A/T, MILLIVOLTS Cr-AI COUPLES'B r -- - 0.2 -0.3 - -0.4- SAMPLE: Co0.93 Fe2.07 04 J- -- -- -- -- -- -0.5 — ~=-276jV/~C Figure 30. Variation of EMF of Ferrite Pellet with AT Across the Pellet.

oO0 0 -5005 -450 — ---- a- -350 LC) 0 INTRODUCED:S H RUN NUMBER 155 H2 G. -300 - CATALYST - Co093 Fe27 04 TEMPERATURE 250~C -250 - - - t INTRODUCED 02, _ 0 I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 25 TIME, HOURS Figure 51. Variation of Thermoelectric Power with Time in Helium, Hydrogen, and Oxygen Atmospheres; Temperature = 250~C.

-520 j I II III -500 INTRODUCED H2 -4801-460 He- -- --- --- H -420 -- 0 Temp-40ratu - - - - - - W -360 -340-____ RUN NUMBER 157 CATALYST- Co093 Fe207 04 -320 TEMPERATURE 1500C -300 - -,. -,-. -.-, 0 I 2 3 4 5 6 7 8 9 10 II TIME, HOURS Figure 32. Variation of Thermoelectric Power with Time in Helium and Hydrogen Atmospheres; Temperature = 1500C.

-350 —-i-o INTRODUCED He _ _ ~I-3 00 kI -300 -- -- -- -- -- -- -- -- -- -- -- -- -- - / — INTRODUCED 02 w INTRODUCED H2 0 t -2 -200 C 0 (K \3 o ZD 0 W 150 > w- 0 W I- RUN NUMBER 150 t )H-100 — - --- W CATALYST- Co93 Fe207 04 TEMPERATURE 880C 0 2 3 4 5 6 7 8 9 0 11 12 13 02 3 4 5 6 7 8 TIME, HOURS Figure 33. Variation of Thermoelectric Power with Time in Helium, Hydrogen, and Oxygen Atmospheres; Temperature = 88~C.

-87TABLE VII *tThermoelectric Power Change During the Adsorption of Hydrogen and Oxygen Catalyst Gas Adsorbed Run Semiconductivity Temperature No. Type _ Oxygen* Hydrogen*C 112 p (1) t 250~ 113 P (1) 250 114 n (2)4 (1) 250 119 n (2)4 (l)+ 250 120 p (2) () + 250 126 p (1) (2) 4 250 138 n (2) (1)C 250 139 n (1) 250 141 p (l) t (2) 4 250 142 n (2) (1)i 250 143 p (2)t (1) 250 147 P (2)t (1), 250 149 p (2)* (1)-+ 140 150 n (2)-+ (l)- 88 152 p ()t (2) + 250 153 P (1) (2) 4 250 155 n (2)t (1)'3 250 155a n **(1) *+ *+(2)1~\ 250 157 n (1) 4 160 158 n (1)- 88 158a n (1) 120 159 n (1) 4 180 x The arrows show the direction of change of the absolute value of thermoelectric power upon admittance of the gas. The numbers (1), (2), refer to the order which the gases were admitted,, i.e,, compare Runs No~ 155, 157 150, and 155awith Figures 31 32, 33, and 54. ~** In Run 155a the thermoelectric power changes from - to + as oxygen was first introduced and from + to - when hydrogen was introduced. The + and - signs on the arrows indicate this change.

V. DISCUSSION OF RESULTS Ao Catalyst (45) The methods outlined by Jonker were used in preparing the cobalt ferrite catalyst, A hard, sintered, non-porous ferrite pellet resulted, Crushed ferrite pellets having four different compositions were used as catalysts for the hydrogen-deuterium exchange studies. If the hydrogen-deuterium exchange reaction occurs by means of an adsorption-desorption mechanism on the cobalt catalyst, the investigation of the electron transfer process by means of thermoelectric power measurements during the chemisorption of hydrogen and oxygen would be valuable in gaining an insight into the mechanism of the reaction, However, no change was observed in the thermoelectric power of these sintered ferrite catalyst pellets during the adsorption of helium, hydrogen, or oxygen. This result would be expected if, as discussed in Chapter II, the ratio of the surface layer to particle diameter was so small that the contribution of the surface layer to the thermoelectric power, which would be influenced by the chemisorbed gas, was negligible. In this case, the measured thermoelectric power would be characteristic of the bulk thermoelectric power of the pellet and would be independent of the degree of adsorption on the surface, In order that the surface contribution to the thermoelectric power would be an appreciable part of the total measured thermoelectric -88

-89= power, a ferrite pellet with a much larger surface to volume ratio was required. To achieve this end, the technique used by Parravano and ( 74) Domenicali(74) in studying the thermoelectric behavior of solid particulate nickel oxide was applied to the ferriteo In this technique high surface area powder is compressed to form mechanically strong pellets without the need for additional heating. In order to make the high surface area powder needed for this technique, a ferrite preparation procedure was used in which final firing of the ferrite material occurred at a lower temperature (1100~C as compared to 1350DC) and in a powder form (as compared to a compressed pellet form). The observed variation of thermoelectric power under different gas atmospheres for the: mechanically compressed pellets made from the powdered ferrite indicates that the surface contribution to the thermoelectric power is appreciable. The thermoelectric power measurements made on these powdered samples shows the same trend (io.e, swing from + to - at x = 2.0) with composition as was observed for the sintered pellets. However, due to the contribution of the surface layer, the magnitude of the thermoelectric power of the powdered samples would not be expected to agree with the value reported by Jonker(45) which correspond to the bulk thermoelectric power of the ferrite, By observing the direction of the change in thermoelectric power of the compressed powder pellets, an insight can be gained into the electron transfer mechanism during adsorption and desorption. This insight may then be applied to better understand the reaction mechanism of the heterogeneous catalysis of hydrogen-deuterium on the cobalt surface,

90go If the results of the thermoelectric power studies made on this higher surface area ferrite are going to apply to the kinetic study results made on the pelletized ferrite, it must be demonstrated that the two preparation techniques result in the same ferrite material, with different surface areaso This assumption is supported by Figure 35, Appendix III, in which the x-ray diffraction patterns and true solid densities of the two materials are compared. The x-ray diffraction patterns agree both in the line position and relative intensity, indicating that the two methods of preparation result in the same single phase spinel structure. The true solid density compares well also. A surface area of 2 m2/gm. (Figure 25) and average particle size of.5 microns (Figure 29) resulted from the preparation of cobalt ferrite in the powder form, This particle size is small enough that the space charge layer formed by the transfer of electrons between the adsorbed ions and the catalyst is of the same order of magnitude as the particle size, Parravano and Domenicali ) reported changes in the thermoelectric power during the adsorption of various gases on nickel oxide, a localized level semiconductor, which was prepared in a manner similar to the ferrite and had approximately the same particle size (ioe,o fired at 11000C in air), with an average particle size of.57 microns. They reported a lower limit of 600 for the thickness of the space charge region. Due to thesimilarity in type of material (localized level semiconductor), firing temperature, and particle size, this value, 600X, is probably a good estimate of the space charge layer thickness at the surface of the cobalt ferrite.

,91B. Thermoelectric Power Changes During Chemisorption The measured thermoelectric power, as pointed out in Chapter II, is a function of the hole and electron carrier concentration. Consequently, factors such as catalyst composition, temperature, and chemisorbed ions which change the carrier concentration will also alter the thermoelectric power, The temperature and catalyst composition, however, were held constant during each run, so that the observed change in thermoelectric power may be attributed to changes in carrier concentration caused by the electron exchange between the adsorbed ions and the catalyst surface. The change in thermoelectric power during chemisorption was observed over the temperature ranges 80~ - 250~C. The results of these runs fall into three categories which will be designated as high temperature (approximately 2500C), intermediate temperature (14P - 1800C), and low temperature (800 - 100~C), 1. High Temperature Runs At high temperatures, the acceptor levels and donor levels, which are at most.06 ev above the valence band or below the conduction band respectively, have become ionized, and some electrons are thermally excited across the.55 ev gap from the valence band, into the conduction band. When this occurs, the ferrite becomes a two carrier semiconductor, with the electrons in the conduction band contributing to the n-type conductivity and the holes in the valence band contributing to the ptype conductivity. In this temperature range the thermoelectric power,, I is given by -n, u, E5- Ef + ] z/ E (11) 17,,we -+ Hz e ( 11)

-92where the Fermi level, E and (Eg-Ef) are logarithmic functions of carrier concentration, Equation (11) shows that when nl = n2 the thermoelectric power will be directly proportional to the carrier concentration, If nl>> n2 (or n2<< nl) however, Equation (11) reduces to eo T = - d - kT^^ ( l7 "2) (14) e@T ~ +/? -+ (eTA N (a,>2I) (15) and S becomes inversely proportional to the logarithm of nl or n2. Jonker(45) reports the cobalt ferrite becomes a two carrier semiconductor at temperatures above approximately 160o~C Experimental evidence of two carrier ferrite at high temperatures (250~C) was also observed in this work (see next paragraph), For compositions in the neighborhood of CoFe204, this intrinsic dissociation occurs at lower temperatures 45). At 2500C, hydrogen ions, adsorbed on a two carrier n-type ferrite, would give up electrons to the catalyst surface, thereby increasing the absolute value of the n-type thermoelectric power as predicted by Equation (11). When the electron concentration increased to the point where n1)> n2, Equation (11) reduces to Equation (14) and the magnitude of n-type thermoelectric power should begin to decrease. If oxygen is now admitted to the surface, electrons will be transferred to the adsorbed oxygen atoms, increasing the absolute value of the thermoelectric power. This effect, is clearly demonstrated in Runs No.

-93138, 139, 142, 155 and 155a. See Table VII and Figure 34. Run 155a, Figure 34, is particularly interesting. In this run, oxygen was adsorbed at 2500C on an n-type ferrite, As electrons were transferred, n1 became smaller than n2 and the thermoelectric power switched from n-type to p-type, changing from an initial value of -170/ v/0C to a value of +600 / v/~C where it began to level off as n2 )> nl. At this point hydrogen was introduced and the transfer of electrons from the hydrogen to the surface (and probably the reduction of the adsorbed oxygen to water) caused the thermoelectric power to swing from +600/ v/~C to a value of -420 q v/~C. At this point n > > n2 and, as predicted by Equation (14), the absolute value of thermoelectric power began to decrease as nl increased further, The amounts of hydrogen and oxygen needed for a transition from the two carrier case, Equation (11), to the single carrier case, Equation (15) may be estimated by using the data given by Jonker, Table Io This data predicts, at room temperature for instance, at a maximum thermoelectric power, 2% of the acceptor levels are ionized and a minimum thermoelectric power at 0,. (This minimum is 0% excess Fe, rather than some larger value, due to the large mobility ratio,/( / 2 = 104), If the acceptor levels are furnished by the adsorbed gas, and if the surface contribution to the thermoelectric power controls the measured value, a 2% coverage would be necessary to swing the thermoelectric power from the negative minimum to the positive maximum (see Figure 34), The magnitude of this calculated per cent coverage is certainly reasonable. These runs demonstrate that at 250~C the ferrite behaves as a two carrier semiconductor, with hydrogen being adsorbed as an electron

INTRODUCED HYDROGEN eo — 600 | — --- -j, l l l l ] ] CATALYST - Co09 Fe 04 500 - 7- - - l - - i I TEMPERATURE 250~C 400 0 300 T 200 +100 0 0 a -100Hydrogen —-mospher — - - -2 INTRODUCED w OXYGEN 0 -300 I-400 -500 0 1 2 3 4 5 6 7 8 9 10 11 12 13 TIME, HOURS Figure 34. Run Number 155a; Variation of Thermoelectric Power with Time in Oxygen and Hydrogen Atmospheres.

-95donor and oxygen being absorbed as an electron acceptor. 2, Intermediate Temperature Runs In the intermediate temperature range, the acceptor levels (or donor levels) are thermally ionized, but the intrinsic conduction is negligible. Therefore, either n1 or n2 predominates and Equations (14) and (15) apply, with the absolute value of the thermoelectric power being inversely proportional to the logarithm of the concentration of the majority carriers. In this case the absolute value of the thermoelectric power of an nRetype ferrite should be decreased by hydrogen adsorption and increased by oxygen adsorption. In a similar manner the magnitude of the thermoelectric power of a pR-type ferrite should be increased by oxygen adsorption and be decreased by hydrogen adsorption. Runs No. 157 and 159 (TABLE VII) demonstrate this effect, 53 Low Temperature Runs At still lower temperatures (800-120~C) the ferrite is a single carrier semiconductor with completely ionized donor and acceptor levels, Therefore, the effect of the transfer of electrons from the adsorbed hydrogen or oxygen would be expected to be the same as the effect observed in the intermediate temperature rangeo However, in this low temperature range, which corresponds to the temperature of the kinetic exchange experiments, no change in thermoelectric power was observed during the adsorption of hydrogen and oxygen. In this temperature range, the adsorption on the powdered ferrite, therefore, takes place with littleelectron exchange between the adsorbed molecules and the catalyst surface,

~96~ C, Kinetic Studies The present data show that the hydrogen-deuterium exchange on cobalt ferrite proceeds in two stages: (a) an activation stage, in which the catalyst's activity increases with time, followed by (b) a second stage, in which, the catalyst activity is constant. During the activation stage the 98% hydrogen, 20 deuterium gas mixture flows over the catalyst for twelve hours at 200~C. The activated catalyst could be de-activated by exposing it to air or oxygen. The data also indicate that the activation energy of the hydrogen-deuterium exchange reaction on cobalt ferrite, (Co xFex04), is a function of the catalyst composition (see Figure 22 ) For the p-type catalysts, with x< 2,0, the activation energy was 18 to 19 Kcal/mole, while for the n-type catalysts, with x > 2, the activation energies increased to 23 to 24 Kcal/mole. A similar change was observed in the sko factor, which increased from 30 to 31 (for x<2.0) to 56 to 38 (for x >2,O0)o Thus a compensation effect was observed between the activation energy E, and the iko 0 (see Figure 24 ), D. Proposed Reaction Mechanism The exchange reaction occurs on cobalt ferrite in two stages: (a) an activation stage in which the catalytic activity increases with time and (b) a second stage in which the exchange reaction occurs with constant catalytic activity, The proposed mechanism for the hydrogen pretreatment during the activation stage is given by the following reactions:

-97S(O-) + 2H — S(OH') (62a) S(O0) + -D2 - S(OD-) (62b) S(O0) + ~ H2- S(OH') + e- (63a) S(O -) + D2 - S(OD-) + e- (63b) 2S(OH ) -- H20 + S(O') + e (64 ) where S refers to a surface site. In this proposed mechanism the hydrogen pretreatment corresponds to the reduction of adsorbed oxygen. The oxygen is present on the surface of the ferrite as 0" and 0"", The reaction of hydrogen with O (reaction 62a) will be referred to as Type A adsorption, reaction with 0O' (reaction 63), will be called Type B adsorption. Morrison(68) has proposed this scheme to describe the adsorption of hydrogen on ZnO. He points out that the singly ionized oxygen atom should provide an attractive adsorption site for hydrogen since the heat involved for the reaction IH2 + 0 - - OH" is approximately 2,6 ev, calculated from known heats of reaction for various similar reactions. Morrison also reports, that for ZnO, the activation energy is greater for Type B adsorption than for Type A adsorption. Other similar activation pretreatments have been reported in the literatures Molinari and Parravano(66) proposed reaction 62a and reaction 64 to describe the activation of ZnO. Holmn and Blue(42) and unpublished data from the Frick Chemical Lab(66) indicate that several molybdenum, tungsten, and uranium oxides are not

.98active for the hydrogen-deuterium exchange until they are subjected to similar pretreatments. After the catalyst has been activated, the proposed exchange reaction mechanism is given by the reactions: S(OHO) + -D2? -* S(O) + HD (65a) S(ODi) + H2 - S(0) + HD (65b) S(OH") + S(OD-)2S(O0) + HD (66) The catalytic reaction occurs by exchange between the deuterium or hydrogen and the surface OH or OD groups, respectively, reactions 65a and 65b, or between the adsorbed OH and OD" group on the surface, reaction 66. Wicke(l07) has shown that reactions 65a and 65b describe the mechanism of the hydrogen-deuterium exchange reaction on alumina catalysts. This mechanism has also been proposed by Molinari and Parravano66) to explain the hydrogen-deuterium exchange in ZnO. In the present study the following observations substantiate the plausibility of this mechanism: 1l Thermoelectric Power Studies During Chemisorption At high temperatures, during hydrogen adsorption, a shift in thermoelectric power corresponding to an electron transfer to the catalyst was observed. This electron transfer could be explained by reactions 63a, 63b and 64, which occur during the high temperature pretreatment of the catalyst. If the activation energy for Type B ad, sorption is greater than for Type A adsorption, which isthe case for ZnO.Morrison(68)], then reactions 63a and 63b would occur at higher

-99temperatures, As the temperature is lowered, the semiconductor becomes a single carrier semiconductor, and the observed change in thermoelectric power with hydrogen adsorption decreasesuntil, in the temperature range 800 to 1000, no change in thermoelectric power is observed during the adsorption of hydrogen, This effect may be caused by the increased predominance of Type A adsorption over Type B adsorption as the temperature is lowered. In Type A adsorption, reactions 62a and b, there is no electron transfer to the catalyst surface and consequently, no change in thermoelectric power would be expected. Since the amount of electron transfer is small at temperatures corresponding to the exchange reaction temperature, the bending of the bands at the catalyst surface due to the build up of surface charge will be small. This means that the distance between the conduction band (and valence band) and the Fermi level will be approximately the same at the surface as it is in the bulk of the catalyst. Consequently, it would be expected that changes in the Fermi level caused by varying the r'e/Co ratio of the ferrite would influence both the catalytic properties and the surface contribution to the catalyst's thermoelectric power. 2. Hydrogen-Deuterium Exchange Data The reaction proceeds in two stages: phase one. in which the catalytic activity increases with time, followed by phase two of constant activity. Phase one could then correspond to a reduction of the surface oxygen and an increase in the surface OH and OD concentration. The second stage is a truly catalytic stage in which the hydrogen-deuterium is

100 exchanged with the surface OH and OD groupso Since the OD and OH groups are formed in phase one, the measured catalytic activity of phase two should increase with time during phase one, These effects have been observed, Air or oxygen treatment of the activated surface would reoxidize the hydrogenated surface, reducing the OH and OD concentrations and lowering the catalytic activity,. It was observed that the introduction of air or oxygen during an exchange run would immediately deactivate the catalyst. The catalyst could then be reactivated by reducing the catalyst with hydrogen (or deuterium) at high temperature, If the surface oxygen, S(O~) or S(O"), is considered as the active site for adsorption of hydrogen and deuterium on the cobalt surface, Boreskov's relationship (see Theory section) between catalytic activity and Fermi level, described in the Theory Chapter, would apply, Reaction 63a and 63b would be analogous to Equation (30), resulting in an exchange rate of the form of Equation (54). Similarly, reaction 66 would be analogous to Equation (55), resulting in a rate of exchange of the form of Equation (4o), This mechanism, then, would predict that lowering the Fermi level (becoming more p-type) should increase the catalytic activity. The activation energy for hydrogen-deuterium exchange reaction was observed to be lower for the p-type ferrite than for the n-typeo Although the pre-exponential factor varied in such a manner as to compensate for the change in activation energy, causing only small variation in the overall reaction rate with Fermi level, the direction of the observed change in catalytic activity is

-101consistent with the direction of the change in Fermi level of the catalyst, as predicted by Boreskov's relationships.

VI, CONCLUSIONS The hydrogen-deuterium exchange reaction on cobalt ferrite occurs in two stages4 (1) The first stage is an activation stage, in which the catalytic activity increases with timeo This step might be associated with the reduction of oxygen on the surface of the ferrite, and the corresponding formation of OH and OD groups. (2) In the second stage, the catalytic exchange reaction occurs with constant activity. In this step, it seems likely that exchange occurs between hydrogen and deuterium and the OD andCH groups formed in stage one. This reaction mechanism is consistent with the data of this study and has been used in previous investigations to explain the hydrogendeuterium reactions on other catalysts. The activation energy for the hydrogen-deuterium exchange reaction on the cobalt ferrite is lower on p-type than on n-type ferrite. A compensation effect was observed between the activation energy and pre-exponential factor. In four runs made to determine the effect of p-n junctions on catalytic activity, no increase in catalytic activity was observed on mixed, sintered p- and nRtype catalysts. Since the relative surface area of pan junction, produced by sintering the mixed ferrite, could not be determined, it may only be concluded that either (1) the junction is not appreciably more active for the exchange reaction, or (2) the relative area of p-n junction produced by the sintering process was so small that any increase (or decrease) in catalytic activity at the junction would not affect the overall activity of the catalyst. -102

-103This study has also furnished data which, along with previous investigations reported in the literature, indicate that the changes in thermoelectric power of particulate systems due to gas chemisorption may be used to gain an insight into electron exchange occuring at the solid surface. Future work of interest would be to study the oxygen and hydrogen adsorption isotherms on cobalt ferrite and the electron transfer during adsorption. These investigations would shed further light into the mechanism of the exchange reaction.

APPENDIX I EXPERIMENTAL DATA,104

-105A. Raw Data and Calculated Per Cent Conversions for Hydrogen-Deuterium Exchange Runs TABLE VIII. Preliminary Runs at Constant Temperature (Run 72) and Constant Flow Rate (Run 74) Sample Time Temp. Flow Mass Spec. Peak Height Per Cent Sample Time Temp. Flow Mass Spec. Peak Height Per Cent (min) (C) (cc. min.) H2 D D2 Conversion (min) (C) (cc. mi ) 2 D D2 Conversion Run 72, Catalyst Col.07Fel.93q4 Run 74, Catalyst Co1 07Fe1 9^~4 Run 72, Catalyst Col 07^1.970.4 9 55.6 19.6 F~ C oo100 124.5 20 2418 36.3 30.4.374 A 58 144.25 20 1779 33.6 19.6.462 O 123 i26.25 20 2190 33.5 27.2.381 B 56 144.0 20 2064 39.4 22.3.469 E 140 126.25 20 2352 36.3 29.2.383 C 120 143.75 20 2160 41.1 23.4.468 F 160 126.25 20 2274 35.5 28.4.384 D 140 137. 2 20 1773 9.4 21.4.407 G 180 126.25 37 2589 28.3 38.1.271 E 160 137.0 20 2181 34.4 27.0.389 H 200 126.25 37 2385 25.7 25.1.268 F 180 136.9 20 2376 37.3 29.6.387 I 220 126.25 37 2124 26.2 35.8.268 G 200 131.0 20 2172 29.2 29.14.332 I 247 126.25 37 2298 25.2 33.8.272 H 224 11. 20 2412 52.0 23.2.325 K 275 126.25 20 2492 38.7 30.2.390 I 240 130.8 20 231 30.1 32.84 319 L 305 126.25 20 21490 40.8 32.0.390 J 260 122.75 20 1950 19.2 29.4.246 N 126.5 37 2481 27.6 36.4.275 L 300 143.2 20 2328 23.0 35.9.213 T 140 26.25 20 2304 35.3 29.1.378 R 42o 136.2 2 0 2115 36.4 30.7.372 U 475 126.25 20 2499 38.1 31.3.378 S 44o 143.6 20 21460 45.1 27.1.1456 T 1460 1143.8 20 2397 44.1 26.6.9452 ABLE IX. Mixed P- and H-Type Catalyst Runs Sample Time Temp. F low Mass Spec. Peak Height Per Cent Sample Time Temp. Flow Mass Spec. Peak Height Per Cent Ru 125 0C (noatsintered) Ru n781, Catalyst; o^ 4 50Co.93e2 04 (sintered Run 127, Catalyst 50CO1* 0 7Fel 9304, 500Coo.93F2 4 nter Run 126 Ctalyst 07 F193O 850 1 hour) O 60 71-4.0 20.0 4350 96.6 58.1.454 A 90 77.8 0.0 3940 81.7 49.6.454 80 714.1 20.0 4530 103.5 9.7.8461 B 110 74.7 20.0 3940 70.1 54.7.391 100 72.3 19.9 14350 86.2 63.1.408 C 10 72.6 20.0 4030 62.3 60.3.31 O 115 70.0 20.5 14380 75.7 63.4.353 150 70.2 20.0 4070 53.3 66.2.287 H 130 67.0 19.9 4400 62.0 77.1.287 E 175 72.14 114.9+ 080 78.5 514..1420 I 150 71.0 15.0 14370 97.8 56.8.463 F 205 70.5 14.9+ 14ooo 58.1 57.9.370 E 170 69.2 15.0 43800 89.2 63.2.414 0 230 68.25 15.0 4020 59.1 63.8.318 F 190 67.0 15.0 1430 75.9 68.1.358 F 255 65.8 15.0 4110 149.1 62.14.261 L 210 64.5 15.0 4330 614.3 7.8.301 I 325 68.2 9.8 060 78.1 53.7.421 M 2140 68.6 9.8 4380 115.8 50.2.3355566. 99 68.7 58.9.365 N 270 66.2 9.9 4530 100.2 8.4.461 K 1425 64.0 99 14o6o 60.1 63.2.322 0 300 61.5 9.85 4390 89.2 63.3.1413 L 1455 61.8 9~9 4o6o 50.9 67.9.273 P 330 62.0 9~9+ 4350 74.9 69.4.350 Q. 360 59.5 9.9 4560 65.8 79.0.294 Hun 127 Catalyst 501C- aO-Fee. 9304, 507Co,a.372. 04 (sintered un 128, Catalyst R 0uCol o7Fe 1.93014 50CooCo 93Fe2. 0704 (sintered Run 127, _07 - u —'' L~V I 990'0, 1 hr.) 1100'C, 1 hr.) A 85 71.2 20.0 3720 75.4 48.8.436 A 140 74.2 20.0 1+00 77.9 54.2.418 B 105 69.5 19.9 3950 72.5 56.1.393 B 160 72.5 19.9 3920 68.1 57.1.374 C 15 67.0 20.0 4140 64.8 64.4.335 C 180 70.2 20.0 4100 61.2 54.5.322 0 115 65.1 20.0 14190 6.6 69.0 291 200 68.0 20.0 1+150 52.14 70.2.272 H 170 67.7 15.5 1409 81.5 53.9.131 H 225 70.3 114.0 3880 71.8 53.6.401 F 195 65.9 15.0 4130 74.7 58.8.388 F 250 68.5 15.0 14150 68.2 61.9.355 O 225 63.7 15.0 14200 614.6 65.9.329 0 275 66.3 15.0 1410 57.7 66.4.303 H 250 61.14 15.0 4140 54.6 69.4.282 H 00 614.0 15.0 4144 148.8 71.8.2514 I 275 62.8 10.0 1+180 82.5 56.1.142 I 330 65.0 9.8 4090 71.9 58.8.379 J 305 61.0 9.9 4270 75.3 61.8.379 i 360 63.3 9.8+ 14220 67.2 64.8.341 K 335 59.0 9.9 4180 614.14 65.5.330 H 390 61.0 9.85 3970 53.14 65.9.288 L 370 57.0 9.9 1414 514.7 69.4.283 L 420 59,0 9.9 4090 46.8 71.7.2146

-106TABLE X. High Temperature Runs Sample Time Temp. Flow Mass Spec. Peak Height Per Cent Sample Time Temp. Flow Mass Spec. Peak Height Per Cent (min) (~C) (cc. min.) H2 HD D2 Conversion (min) (~C) (cc. min.) H2 HD D2 Conversion Run 76, Catalyst Co1.07Fel. 9304 Run 83, Catalyst Co0 98Fe2. 0204 B 115 110.25 20.2 2575 48.2 23.8.503 K 520 120.2 19.5 2430 46.1 26.8.462 C 170 105.5 19.8 2514 45.1 28.9.427 L 540 120.2 19.2 2540 45.7 25.0 477 D 215 100.4 19.7 2520 54.9 3355.5.344 M 56o 120.2 19.2 2591 46.6 25.9.474 E 260 95.6 19.8 2487 26.1 57.1.260 N 590 116.0 19.2 2459 4o.6 50.1.399 F 515 104.8 56.8 2405 25.8 55.9.264 0 430 109.75 19.0 2456 29.1 55.5.292 G 560 110.4 36.4 2469 3355.6 533.6.33555 P 460 105.1 19.5 2430 21.4 59.0.215 H 429 114.5 57.1 2428 51.5 25.6.381 Q 500 115.4 55.7 2469 25.4 58.1.250 I 470 119.75 56.0 2465 45.0 27.1.448 R 525 119.5 54.9 2415 30.7 24.2.310 S 548 124.5 55.5 2274 56.2 28.9.385 T 575 150.0 55.5 2576 46.6 25.6.477 Run 77, Catalyst Co1.07Fel. 904 A 90 115.4 19.5 2601 51.0 26.2.495 Run 84, Catalyst Co0. 98Fe2.0204 B 155 110.0 19.4 2571 42.1 50.0.412 C 180 104.8 19.0 2589 34.9 34.4.33552 A 20 125.0 19.4 2504 47.1 22.9 507 D 225 100.7 19.3 B 4o 125.0 19.4 2292 49.1 22.4.525 E 270 109.8 36.4 2759 29.1 59.9.267 C 60 125.25 19.8 2515 50.5 22.6.527 F 515 115.1 56.0 2808 58.7 56.8.545 D 90 119.8 19.1 2528 41.5 27.1.434 G 560 120.2 56.0 2505 42.2 29.1.420 E 120 114.9 19.4 2564 3355.5 51.4.347 H 405 124.75 56.0 2865 56.0 29.7.484 F 150 110.1 19.5 2564 26.0 35.4.269 G 175 120.0 55.7 2564 29.4 34.4.299 H 190 120.1 55.7 2564 29.5 34.3.299 I 220 124.5 55.7 2585 35.5 51.5.366 Run 78, Catalyst Co10e. 9504 J 250 150.1 55.5 2576 45.0 26.7.457 Run — 78 — Ca —aly- 1.0 —'el.-93 K 280 154.8 55.5 2588 53.6 22.7.541 A 80 120.5 19.5 2406 55.7 21.1.560 B 150 115.0 19.5 2570 46.5 24.8.483 C 185 109.5 19.5 2567 58.5 28.7.4o01 D 250 105.0 19.2 2430 3355.4 52.5.541 Run 85, Catalyst Co. 98Fe2 020 E 575 114.0 55.2 243355 52.7 3355.2.33550 F 520 120.5 55.0 2469 40.7 29.9.405 D l1o 124.9 19.2 2406 53.1 25.6.55 G 565 125.1 55.1 2412 46.1 25.8.472 E 140 124.75 19.5 2424 55.4 25.6.531 H 410 150.1 55.2 2259 48.4 21.5.550 F 170 120.2 19.6 2591 44.4 27.4.448 G 200 115.0 19.7 2588 56.0 51.9.561 H 250 110.2 19.7 2594 28.1 55.9.281 I 265 119.8 48.8 2588 24.8 58.0.246 Run 80, Catalyst Co. 03Fe1 9704 J 290 119.75 48.6 2564 24.6 58.0.245 -----------— 1-~-27 ~K 520 124.8 48.5 2412 51.5 55.2.508 E 95 119.5 19.1 2519 41.7 27.0.436 L 550 130.2 48.4 2400 38.8 51.7.583 F 115 119.5 19.1 2507 42.8 26.5.447 M 585 155.1 48.5 2400 46.9 27.4.461 G 155 119.5 19.1 2516 43.5 26.7.449 H 166 115.5 19.2 233554 57.7 29.9.387 I 180 115.6 19.4 2522 57.2 29.8.584 J 215 109.8 19.5 2564 29.7 534.3.502 Run 86, Catalyst Co0 98Fe2 0204 K 250 109.8 19.1 2575 29.7 34.6.300 L 250 105.4 19.1 2561 25.2 57.7.256 D 80 125.15 19.5 2406 45.9 24.9.479 M 275 115.0 54.9 2406 25.8 38.4.257 E 100 125.2 19.5 2364 46.8 24.8.485 N 290 114.8 54.7 2406 24.2 38.5.259 F 120 125.3 19.6 2412 47.9 24.8.491 0 515 109.8 54.9 2415 50.5 55.5.501 G 150 119.8 19.6 lost sample P 54o0 125.0 34.7 2418 28.0 51.4.3577 H 180 114.8 19.6 2421 51.4 33.5.519 Q 565 150.2 34.8 2412 45.4 27.7.450 I 210 110.1 19.6 2424 24.0 37.0.245 R 4oo 117.1 19.7 2412 41.4 29.5.412 J 240 120.0 55.5 2424 26.6 36.4.267 K 270 124.8 55.1 2427 54.1 52.5.5345 L. 500 129.8 55.5 2415 42.0 28.7.421 M 330 155.25 35.5 2421 29.6 24.5.505 Run 81, Catalyst Co1 05Fe1 9704 E 170 125.0 19.5 2442 41.0 29.2.412 F 190 120.1 19.7 2541 35.7 33.8.346 Run 89, Catalyst Co0 93Fe. 04 G 210 120.1 19.7 2442 34.1 52.9.329 H 245 115.25 19.7 2709 31.6 39.5.282 A 70 115.2 15.6 2322 38.9 24.9.439 I 290 109.9 19.4 2430 21.6 38.2.220 B 85 115.4 19.6 2346 41.2 25.7.450 J 520 119.8 54.9 2520 25.2 40.2.224 C 100 115.4 19.6 233554 41.4 25.4 450 K 550 124.5 54.8 2436 27.5 55.9.275 D 145 115.2 19.6 2364 41.6 25.6.445 L 580 129.7 54.8 2436 33.0 3355.5.33554 E 175 109.8 19.6 2549 52.1 50.2.547 M 420 156.25 54.7 2481 42.5 29.8.415 F 205 104.8 19.6 2364 25.5 33.7 274 G 235 100.0 19.6 2376 20.7 56.2.222 H 265 110.0 35.0 2373 22.8 35.6.245 I 295 114.9 55.0 2576 29.0 32.8.307 J 325 120.6 34.9 2364 36.4 28.7.388 K 355 125.0 35.2 2564 45.2 25.2.467 Run 82, Catalyst Col. 03Fel. 9704 A 75 125.6 19.6 2546 534.4 29.2.371 Run 90, Catalyst Co 95Fe2 0704 B 100 125.6 19.6 2540 55.2 29.1.377 C 150 120.6 19.4 2274 28.2 51.1.312 A 80 120.7 19.6 2519 46.6 25.8.495 D 160 115.5 19.4 2358 25.7 55.5.251 B 100 120.7 19.6 2528 48.6 23.9.5o4 E 195 110.0 19.8 2545 18.2 57.9.194 C 120 120.9 19.6 2313 48.9 23.5.509 F 220 120.5 54.7 2575 19.1 58.2.200 D 150 114.7 19.6 2525 39.2 28.6 407 G 250 120.5 35.0 2448 19.4 58.5.201 E 180 110.0 19.7 2540 32.7 32.1 537 H 280 125.7 54.8 2555 25.4 35.4.249 F 210 105.8 19.6 2549 26.6 55.1 274 I 516 129.5 54.5 2450 29.5 54.1.295 G 240 114.8 56.6 2359 28.0 35.2.285 J 550 154.2 54.5 2579 55.7 51.4.549 H 265 120.1 56.6 2358 56.1 51.2.566 I 295 126.5 35.9 2358 46.2 26.2.468 J 525 154.0 56.1 2570 52.1 20.7.587

-107TABLE XI. Low Temperature Runs Sample Time Temp. Flow Mass Spec. Peak Height Per Cent Sample Time Temp. Flow Mass Spec. Peak Height Per Cent (min) (~C) (cc. min.) H2 HD D2 Conversion (min) (~C) (cc. min.) H2 Hi D2 Conversion Run 108, Catalyst CO.07Fel. 9304 Run. 120, Catalyst Co. 98Fe2.0204 H 240 69.41 15.1 2880 5.6 38.0.427 B 25 71.0 20.0 2427 37.8 28.1.402 I 255 70.5 20.0 2910 7.7 42.6.359 C 45 71.2 20.0 2433 43.5 31.5.408 J 270 70.5 20.0 2900 47.5 42.5.358 D 60 71.3 19.9 2484 46.8 32.9.416 K 285 68.5 20.0 2920 42.3 45.5.317 E 75 71.5 20.0 24.66 47.8 33.0.420 L 300 68.5 20.0 2930 42.2 45.7.316 F 95 71.5 19.9 2514 49.7 33.3.427 M 315 67.25 15.0 2950 50.2 47.1.374 G 125 69.5 20.1 2526 43.7 36.8.373 N 325 67.25 15.0 2950 50.0 42.7.369 H 140 67.0 20.0 Lost sample 0 345 65.75 10.0 2950 62.1 36.1.403 I 155 64.25 20.0 2562 29.7 44.8.249 P 355 65.75 9.9 2960 62.5 36.4.462 J 170 64.25 19.9 2559 29.4 44.7.247 Q 370 63.4 9.9 2970 55.4 40.4.407 K 185 65.8 20.0 2568 33.2 42.9.279 R 380 63.4 9.9 2980 54.6 40.8.401 L 200 68.1 20.0 2580 40.2 39.5.337 S 395 64.25 15.0 3000 41.9 47.4.307 M 225 68.1 15.0 2568 50.4 350.419 T 405 64.25 15.1 3000 41.8 47.6.305 N 240 65.75 15.0 2559 42.o 38.4.354 U 420 64.8 20.0 3020 33.9 51.5.248 0 255 63.4 15.1 2607 35.8 42.8 V 430 64.8 20.0 3010 33.9 51.3.248 P 270 61.0 15.0 2592 29.2 4,-2 2 W 445 62.6 20.0 3020 28.9 54.3.210 Q 290 64.7 15.0 2598 4o.o 40.0.333 X 46o0 61..67 14.9 3040 35.5 51.2.257 R 310 65.0 9.9 2607 54.5 33.3.450 Y 475 60.3 9.9 3030 45.9 45.9.333 S 330 63.2 9.85 2622 48.3 36.6.398 Z 490 58.5 9.9 2850 40.9 48.4.297 T 350 60.75 9.85 2622 41.o 40.7.335 a 505 59.7 14.9 3030 30.4 53.9.220 U 370 58.3 9.85 2646 33.4 44.3.274 b 520 60.25 19.9 3060 42.2 57.3.174 Run 121, Catalyst Co0 98Fe2.0204 Run 110, Catalyst Co1.07Fe. 9-04 ----— o 110, Catalyst C01~ 1.934 4A 60 70.0 19.9 2452 40.1 37.3.350 M 100 75.0 25.0 2670 39.1 4o.6.325 B 95 71.8 20.0 2427 48.1 32.4.426 N 115 74.8 25.0 2700 40.3 41.0.329 C 115 72.0 20.0 2175 43.9 28.6.434 0 130 74.8 25.0 2710 41.3 40.7.337 D 130 72.3 20.0 1914 42.0 26.2.445 P 145 74.5 25.0 2730 41.9 40.7.340 E 150 70.3 20.0 1989 35.1 28.1.384 Q 175 74.0 20.0 2730 51.2 36.5.417 F 165 68.3 19.9 1986 29.6 31.1.322 R 190 74.0 20.0 2760 52.7 36.5.419 G 180 65.7 20.0 1971 24.4 33.4.268 S 205 73.75 15.0 2730 63.7 30.5.511 H 195 62.3 20.0 1965 32.8 29.2.360 T 225 73.75 15.0 2750 63.6 30.7.506 I 210 69.1 15.0 2010 41.6 26.2.443 U 240 73.0 9.9 2750 79.2 22.8.635 J 225 67.3 15.0 1986 36.1 268.4.389 V 255 73.0 10.0 2700 79.4 23.1.635 K 240 65.1 15.0 1974 30.2 30.7.330 W 285 70.8 9.9 2770 72.5 27.3.570 L 255 62.75 14.9 1977 25.0 33.4.272 X 310 71.6 9.9 2780 55.5 35.6.438 M 270 60.25 15.0 1998 20.6 36.1.222 Y 325 70.9 20.0 2800 43.7 42.0.342 N 285 65.9 9.9 1971 43.7 24.3.473 z 340 68.6 20.0 2810 38.2 44.1.300 0 305 63.5 9.8 1986 37.2 28.2.397 a 355 68.6 15.0 2780 47.8 39.2.379 P 325 60.0 9.9 1992 28.4 32.4.305 b 4oo 68.0 9.9 2800 63.9 32.1.499 Q 345 56.8 9.9+ 1955 21.1 34.6.234 c 425 65.75 9.9 2830 56.4 36.2.438 a 445 65.5 15.0 2880 40.7 45.3,310 e 600 65.25 20.0 2850 29.8 49.3.231 Run 123, Catalyst Coo, 9Fe2.0704 A 25 72.5 20.0 4540 67.0 65.1.340 Run 111, Catalyst Col 03Fel. 9704 B 40 72.5 20.0 4570 71.4 67.8.345 C 55 72.5 19.9 456o 71.6 67.8.344 E 160 76.0 20.0 2550 46.6 31.5.425 D 70 70.0 20.0+ 4590 59.0 75.2.282 F 175 76.0 20.0 2571 49.6 33.3.427 E 85 66.6 20.0 4630 44.6 82.9.212 G 190 76.0 20.0 2571 51.0 34.0.429 F 100 63.6 20.0 4660 34.6 88.8.163 H 205 73.25 20.0 2568 43.8 38.0.366 G 130 68.4 15.0 466o 66.5 73.5.311 I 220 73.25 15.0 2598 55.7 33.1.457 H 145 65.7 15.0 4700 54.1 80.7.250 3 240 62.5 9.9 2592 60.4 26.2.491 I 160 62.5 15.0- 4710 41.2 86.5.152 K 255 70.20 15.0 2725 47.0 38.2.381 J 175 60.0 15.0- 4740 32.8 91.1.153 L 270 29.75 20.0 2673 36.6 44.1.293 K 195 64.8 9.8 4730 71.2 72.3.330 M 285 67.25 20.0 2661 30.8 46.7.248 L 215 62.6 9.85 4730 60.1 78.2.278 N 300 67.25 15.0 2643 4o.0 42.1.322 M 235 60.0 9.85 4770 48.9 84.8.224 0 320 67.0 9.9 2661 52.8 36.0.423 N 255 57.0 9.8+ 4800 37.1 20.4.170 P 340 64.4 9.9 2655 45.6 39.2.368 Q 353 64.6 15.0 2706 32.6 47.1.257 R 375 62.0 9.9 2664 38.2 43.6.307 Run 124, Catalyst Co0 93Fe2.0704 A 0 75.1 19.9 4810 98.1 6o.4.448 Run 112, Catalyst Co1. 0Fe1.070 4 B 15 75.1 20.0 4890 99.9 61.9.447 C 30 72.7 20.0 4910 83.8 70.3.373 G 25 77.4 20.1 2592 53.4 32.0.455 D 45 69.4 20.0 4810 63.6 78.4.289 H 45 77.4 20.1 2664 55.7 43.3.448 E 60 66.1 20.0 4910 48.8 88.0.217 I 60 77.3 20.0 2667 55.8 34.9.444 F 80 70.9 14.9 4930 90.9 67.6.402 J 75 77.15 20.0 2709 55.8 35.9.437 G 100 68.7 15.0- 4890 75.1 74.1.336 K 105 74.2 20.0 2715 47.4 39.9.373 H 115 65.8 15.0 4960 61.6 82.3.272 L 110 74.1 20.0 2667 57.7 33.8.46o I 130 62.5 15.0 4950 46.3 89.8.205 M 140 70.8 9.9 2718 65.0 31.3.510 J 155 66.2 9.8 4860 91.2 66.6.406 N 155 71.25 14.9 2739 49.1 35.4.356 K 175 64.75 9.9 4970 77.7 74.1.344 0 170 71.5 20.0 2745 40.2 44.3.312 L 195 62.0 9.9 5000 63.1 83.3.275 P 185 68.2 20.0 2784 33.3 43.6.255 M 215 58.5 9.9+ 4930 46.8 90.8.205 Q 200 68.3 15.0 2787 43.0 43.8.329 R 220 68.3 9.9 2766 57.2 36.6.439 S 240 66.0 9.85 2784 50.9 46.0.389 T 255 65.2 15.0 2778 34.3 48.1.263 U 275 63.0 9.9 2781 47.0 44.3.322

-108B. Raw Data and Calculated Thermoelectric Power for Adsorption Runs TABLE XII. Adsorption Runs Time Atmosphere 6T (millivolts EMF Thermoelectric Time Atmosphere AT (millivolts EMF Thermoelectric Cr-Al couple) (millivolts) Power,,A4V/~C Cr-Al couple) (millivolts) Power,AeV/~C Run No. 112, Catalyst- Col.09Fel. 9104 Temperature- 250~C Run No. 119, Catalyst- Co0 96Fe2. 0404, Temperature- 250~C 1:43 He+2cm 02 0.8576 10.6270 +532 4:18 Helium 0.8215 -6 419 -302 1:48 0.3370 3.7800 4:22 0.0970 -Q.4059 1:55 " 0.2955 3.5269 +538 4:29 He+2cm H2 0.8612 -6.4121 -305 2:00 0.8518 10.8557 4:30 -7.100 -338 4:31 * -7.325 -348 2:15 " 0.8704 11.1547 +545 4:32 * -7.519 -358 2:22 0.3161 3.7912 4:33 * -7.711 -367 4:34 -7.883 -375 2145 0.3035 3.7266 +551 4:35 * -8.024 -383 2:50 0.8557 11.1630 4:36 "-8.116 -386 4:37 -8.225 -391 3:30 " 0.8759 11.5384 +561 4:38 * -8301 -396 3:35 0.3230 3.9865 4:39 -8.305 -398 4:4o0 * -8.387 -399 4:38 " 0.2928 3.6725 +561 4:45. 8561 11. 3563 4:49 He+2cm 02 0.6591 -8.471 -408 4:50 " -7.700 -366 7:26 " 0.9306 12.5239 +574 4:51 -7.525 -358 7:35 0.2876 3.5255 4:52 -7 354 -350 4:53 7.091 -338 10:40 " 0.2700 3.4520 +571 4:54 * -6.811 -324 11:00 0.9016 12.2210 4:55 * -6.46 -311 4:56 * -6.351 -302 Run No. 113, Catalyst- Col. 09el.9104, Temperature- 250~C 4:57 -*6.125 -291 8:15 Helium 0.9023 9.8052 +470 4158 -5864 -279 8:30 " 0.2811 2.6827 5 358 -28 5:00 -5.358 -255 8;43 0.2620 1.850 +220 8:49 0. 8308* 4.905 Run No. 120, Catalyst- Co,96Fe2,o044 7:15 Helium 0.8240 6. 767 +300 8:50 He+lcm H2 4.7116 +212 7:20 0.0890 0.4585 8:51 " 4.3450 +195 8:52 " 4.0490 +182 7:29 He+2cm H2 * 6.8960 +296 8:53 " 3.8520 +176 7:30 6.742 +290 8:54 " 3.6840 +165 7:31 * 6.505 +279 8:55 " 3.5490 +159 7:32 * 6.441 +276 8:56 " 3.4570 +155 7:33 * 6.277 +270 8:57 " 3.3790 +152 7:34 * 6.105 +263 8:58 " 3.317 +148 7:35 * 5.929 +255 8:59 " 3.261 +146 7136 * 5.740 +247 9i00 " 3.211 +144 7:37 5.530 +237 9:01 " 3.167 +142 7:38 * 5.267 +227 9:02 " 3.125 +14o 7:39 5.071 +218 9:03 " 3.085 +139 7:40 * 4.860 +209 9:04 " 3.047 +137 9:05 " 3.015 +135 8:00 He+2cm 02 4.770 +205 9:d6 " 2.984 +134 8:01 " * 10.650 +461 8:02 10.875 +470 Run No. 114, Catalyst- Coo 96Fe2.0404, Temperature- 250~C 8:03 " 10.972 +474 8:04 "11.025 +477 10:30 Helium 0.8507 -7.2632 -356 8:05 * 11.045 +477 10:4o " 0.2037 -1.6271 8:06 * 11.053 +477 8:07 * 11.052 +477 11:35 He+lcm Ho 0.8527 -7.545 -370 8:08 * 11. 046 +477 11:36 " * -7.839 -384 8:09o * 11.022 +477 11:37 " -7.970 -391 11:38 " -8.148 -400 Run no. 126, Catalyst- Col.09Fel. 9104, Temperature- 250~C 1139 * -8.267 -406 1:30 Helium 0.6025 6.629 +455 11:40 * -8.4o4 -413 1:40 6.890 +465 11:41 * -8.46 -419 1:45 o. 6050 6.825 +463 11J:42 * -8.631 -423 1:50 6.760 +46 11:43 * -8.782 -431 1:55 0.5964 6.606 +456 11:44 * -8.850 -435 2:00.5872 6.520 +456 1145 " -8.901 -437 2:05 0.5809 6.456 +456 11:46 -8.939 -439 2:10 0.5804 6.493 +457 11:47 * -8.966 -440 11:48 -8.981 -441 212 He+2cm 02 0. 5856 7.707 +541 11:49 * -8.990 -442 2:15.5806 7.817 +553 11:50 " -9.000 -443 2:20 "0.5807 7.939 +56 2:30 0.5752 8.041 +572 3120 Helium 0.8108 -7.4514 -386 2:45 0.5808 8.200 +578 3:25 " 0.2162 -1.8501 3:00 " 0.5805 8.242 +582 3:040 " 0.5876 8.290 +582 5130 He+lcm 0, 0.2072 -1.6901 -387 3:40 0.5876 8.290 +582 3:35 " 0.7873 -7.1873 4:13 Helium 0.5913 7.905 +548 4:00 0. 8364 -7. 7946 -387 4:25 He+2cm H2 0.5732 4.63 +321 4:05 0.2225 -1.9769 4:26 4.27 +305 4:30 " 0.1438 -1.3420 -358 4:27 3.98 +269 4:35 "0.7863 -6.9451 4:29 365 +269 4:30 3.45 +246 5:00 ". 8276 -6.8115 -337 4:3 * 3.30 +236 5:10 0.1899 -1.4865 4:34 3 +222 4:35 * 2.98 +213 6:05 " o.1404 -1.0660 -209 4:40 2.70 +193 6:15 ".8106 -4.4575 4:45 2.46 +176 4:50 * 2.23 +162 7:25 0.6319 1.16 + 75 8:00 0.6018 1.03 + 69 9:00p.m. " 0.6403.95 + 61 * T assumed unchanged from previous reading. * 1T assumed unchanged from previous reading.

-109B.. Raw Data and Calculated Thermoelectric Power for Adsorption Runs TABLE XII. (continued) Adsorption Runs Time Atmosphere AT (millivolts EMF Thermoelectric Time Atmosphere AT (millivolts EMF Thermoelectric Cr-Al couple) (millivolts) Power, AV/~C Cr-Al couple) (millivolts) Power,.MV/~C Run No. 149, Catalyst- Co1. 0Fe1 9104, Temperature 140~C Run No. 1535, continued 9:05 Helium 0.3469 5.071 +598 3:00 He+2cm H2 0.5421 5.0212 +380 9:08 0.3157 4.941 +640 3:30 0.5401 3.9485 +300 9:26 0.3341 5.211 +64o 4:10 0.5399 3.4877 +265 9137 " 0.3414 5.203 +630 4:30 "0.5378 2.9453 +225 5:00 " 0.5365 2.6848 +205 9i40 He+6cm H2 0.3410 5.226 +633 6:00 " 0.5378 2.4307 +185 9:55 0.3441 5.206 +627 6:30 "0.5377 2.2351 +170 10:17 " 0.3408 5.266 +633 7:00 "0.5384 2.2362 +170 10:44 0/3369 5.200 +634 7:30 "0.5212 2.0995 +165 11:25 "0.3362 5.132 +628 8:00 "0.5271 20.310 +158 11 55 ". 3349 5.135 +699 10:00 "0.5317 1.8794 +145 1:10 " 0.3373 5.139 +626 12:00 "0.5301 1.7442 +135 1:17 " 0.3368 5.102 +626 1:36 " 0.3362 5.128 +626 Run No. 155, Catalyst- Co0. 6Fe2.0404, Temperature 250'C 2:17 0.3338 5.668 +625 8:00 Helium 0.7939 -7.1658 -369 2:55 0.3348 5.561 +619 8:05 "0.7907 -7.2012 -372 3:20 03332 5.034 +620 9:25 ". 8021 -7.1765 -367 4:06 0.3316 4.936 +613 4:15 " 0.3223 4.984 +672 9:30 He+2cm H2 0.7990 -7.2013 -371 10:06 0.7900 -9.3950 -486 8:25 Helium 0.3430 5-068 +607 10:12 0.7917 -9.6226 -499 8:45 0.3334 5.020 +618 10:24 0.7929 -9.3721 -483 9:00 " 0.3350 4.9885 +616 10:35 0.7939 -8.7067 -450 10:45 " 0.7965 -7.9370 -409 9:01 He+6cm. 02 0.3377 5.431 +659 10:58 " 0.7895 -7.3278 -384 9:11 " 0.3400 5.645 +680 11:15 0.7917 -6.9735 -361 9:25 " 0.3355 5.415 +664 11:35 0.7947 -6.6719 -344 9:47 0.3338 5.536 +674 12:05 0.7956 -6.4026 -332 10:21 " 0.3331 5.555 +684 1:10 0.7903 -5.9917 -311 11:15 " 0.3366 5.336 +676 2:00 0.7838 -5.7496 -301 5:45 0.8000 -5.3270 -275 Run No. 150, Catalyst- Co 96gFe2. 0404, Temperature 88'C 7i20 " 0.7707 -5.0251 -269 10135 Helium 0.2485 -1. 4825 -244 10:40 " 0.3025 -1.7264 -235 7:35 He-+2cm 02 0.8034 -7.4235 -380 10:45 "0.2621 -1.5405 -241 7:50 "0.7925 -7.7265 -4oo 1l:00 0.2561 -1.5015 -241 8:05 0.7969 -8;1215 -420 11:15 " 0.2525 -1.4785 -241 8:30 ". 8037 -8.4259 -432 9:00 0.8128 -8.8361 -446 11:20 He+6cm Ho 0. 2497 -1.4762 -243 9:30 "0.8075 -9.0535 -459 11:26 " 0.2461 -1. 492 -239 1000 o.8027 -91718 -467 11:44 " 0.2445 -1.4322 -240 8:00a.m. 0.8097 -10.112 -512 12:05 0.2456 -1.4535 -243 12:40 " 0.2479 -1.4575 -242 Run No. -155a, Catalyst- Co0.96Fe20404, Temperature 2500C 12:59 "0.3483 -1.4605 -242 10:35 Helium 0.7700 -3.002 -160 1:15 0.2486 -1.4667 -242 3:50 " 0.2489 -1.4615 -241 10:50 He+2cm 02 0.7705 4.844 +258 1200 " 0.2600 -1. 5115 -235 10:55 "0.7737 7.1925 +382 10:56 * 7.7920 +412 11s55 0.2617 -1.4722 -232 10:57 " * 8.160 +432 12:36 0.2482 -1.429 -246.10:59 " * 8.700 +46o 12:47 " 0.2422 -1.448 -245 ll:00 * 9.000 +476 12:55 0.2436 -1.454 -245 11:01 9.181 +486 1:12 " 0.2422 -1.458 -246 11:04 " 9.412 +498 11:06 9.576 +507 1:15 He+6cm 0 0.2749 -1.489 -246 11:12 " * 9.926 +526 1:01 2 0.2499 -1.408 -246 ll:16 * 10.183 +538 2:02 0.2494 -1.503 -249 ll25 " * 10.590 +560 2:50 " 0.2531 -1.536 -249 4:00 " 0.2489 -1.562 -247 11,35 He+2cm H * 10. 955 +580 4:530 " 0.2470 -1.483 -246 1145 " 11.273 +595 4:55 " 0.2440 -1.471 -247 11:50 * 9.0665 +480 8:05 0.2484 -1.490 -246 11:54 " 8.1665 +426 11:55 " 7.400 +391 Run No. 152, Catalyst- Co. 09e1. 9104, Temperature 250'C 11:58 " * 6.300 +333 11:00 Helium 0.6843 6.4105 +385 12:10 2.464 +131 8:00 " 0.6522 6.2801 +395 12:22 " * -0.549 - 28 12:35 " * -1.934 -102 8:35 He+2cm02 0.6219 6.9265 +460 12:44 * -2.558 -155 9:00 0.6321 8.0105 +519 12:55 -3.33 -176 9:30 " 0.6575 8.9979 +56o 9:30 0. 6575 8. 9979 +560 1:01 " * -3.648 -193 10:10 0.6596 9.4334 +590 1:37 * -5.129 -271 10:55 0.6604 9.5825 +595 1:50 -5.504 -291 11:55 " 0.6672 9.9505 +603 2:10 0.7829 -6.2412 -326 2:25 0.7814 -6.649 -348 1:10 He+2cm H2 0.6465 8.1565 +520 2:45 0.7738 -7.0435 -373 1:30 0.6277 5.7830 +378 3:00 0.7775 -7.360 -388 2:00? 0.6329 4.4892 +290 3:45 0.7810 -7.9317 -416 2:30 ".6199 3.6301 +240 4:15 0.7786 -8.o0414 -424 3:00 ".6319 3.1598 +205 4:45 0.7781 -8.2676 -43 3:30 " 0.6355 2.9401 +190 6:000.0770 -8.1015 -431 4:00 0.6216 2.7356 +186 7000.8067 -7.9513 -4o4 4:30 " 0.6242 2.6045 +172 11:00 o0.80o -6.4329 -328 5:00 " 0.6314 2.5911 +168 Run No. 157, Catalyst- Co0.96Fe2.0404, Temperature 160~C Run No. 153, Catalyst- Col. oel, 9104q Temperature 250'C 12:50 Helium 0.5036 -5.8138 -572 12:20 Helium 0.5408 4.3696 +332 1:02 "0.5034 -5.7602 -470 7:30 0". 5330 4.8789 +335 1:10 0.5038 -5.7654 -470 7:45 He+2cm 02 0.5212 4.9102 +385 8:00 " 0.5117 5.2398 +420 1:20 He+6cm H2 0.5067 -5.7917 -470 8:30 "0.5218 5.7422 +42 1:25 0.5003 -5.7457 -471 9:00 " 0.5099 5.8382 +470 140 0.5028 -5.6978 -463 9:30 " 0.5100 5.9810 +481 2:00 0.5000 -5.5129 -455 10:00 0.5212 6.1585 +485 225 o.4971 -5.3922 -445 10:30 "0.5222 6.1207 +482 2:45 o. 4908 -5.2933 -441 11:00 0.5177 6.1900 +488 3:30 0.4911 -5.2250 -436 11:30 " 05206 6.1524 +425 4:25 "0.4933 -5.1359 -429 12:00 " 0.5311 6.3385 +488 5:55 0. 4947 -5.0171 -419 11:15 " 0.4946 -4. 7628 -395 8:00a.m. " 0.4960 -4. 6227 -382 *~T assumed unchanged from previous reading.

-110B. Raw Data and Calculated Thermoelectric Power for Adsorption Runs TABLE XII. (continued) Adsorption Runs Time Atmosphere AT (millivolts EMF Thermoelectric Time Atmosphere &.T (millivolts EMF Thermoelectric Cr-Al couple) (millivolts) Power,jMV/~C Cr-Al couple) (millivolts) Power,.uV/~C Run No. 126, continued Run No. 142, Catalyst- Co0 96Fe2 0404, Temperature 2500C 8:35a.m. He+2cm H2 0.5985 0.035 + 2.3 9:30 -0.279 -19.0 7:35 Helium 0.6590 -4.605 -287 9:45 0.6029 -0.33 -22.0 7:40 0.6532 -4.690 -290 10:05 0.5863 -0.295 -20.0 7:50 " 0.6581 -4.782 -298 10:30 " 0.5592 -0.286 -21.0 11100 " 0.5708 -0.752 -18.0 7:52 He+lOcm H2 * -5.090 -311 11:50 " 05680 -0O247 -18.0 755 " -5.235 -325 1100p.m. 0.5655 -0.245 -18.0 8:02 " * -.5321 -331 2:00 " 0.5545 -0.915 -16.0 8106 " -5.370 -333 8:27 " * -5.370 -333 Run No. 138, Catalyst- Coo. 6Fe2,0404 Temperature 250~C 8:37 " * -5.327 -331 10:45 Helium 0.1148 -8.386 -300 845 * -5.308 -330 8:48 " * -5.295 -328 10: 52 He+6cm,,H 0.ll42 -8.366 -299 9:31 " -5.067 -314 10153 * -8.415 -301 9:53 -4.995 -310 10:56 -8.310 -298 1105 * -4.795 -295 1l:00 * -8.138 -291 11:55 * -4 473 -277 ll:02 " * -7.976 -285 1:10 * -4.429 -274 l: 06 "-7.700 -275 1110 * -7.514 -269 3:55 Helium * -4.538 -290 11 16 "-7.214 -258 1122 "-6.976 -249 3:57 He+lOcm 02 * -4.821 -308 11:26 -6.858 -245 3:59 * -4.210 -314 11:34 -6.654 -238 400 -5.004 -320 11:55 -6.294 -225 4:02 "0.6410 -5.281 -338 12:12 -6.067 -216 4:05 0.6431 -5.475 -350 12:7 -5.897 -211 4:10 " * -5.656 -362 12:42 "-5.860 -209 4:17 * -5.855 -374 11:0 * -5.552 -198 4:23 * -5.967 -386 617 * -4.718 -169 4:30 "0.6615 -6.073 -389 1O "* -4.260 -155 4:40 * -6.182 -395 010 4.26155:45 * -6.231 -399 11:17 He+6cm 02 0.1103 -3.95 -1430 " 7.9 469 ll:19 " * -4.920 -178 12:00 * -7.468 -478 11:20 * -5.100 -185 1:23 * -5.640 -204 Run No. 143, Catalyst- Col. 0el. 9104 Temperature 250~C 1126 * -5.935 -215 9:10 Helium 0.6374 7.309 +468 11:31 * -6.206 -296 10:10 " o.6792 7 082 +462 11:38 * -6.359 -230 11:50 " -6.420 -233 10:25 He+6cm H2 0.6327 6.938 +447 1200 * -6.695 -242 10:30 "0.6287 6.820 +446 12:07 * -6.916 -251 10:35 0.6298 6.743 +439 12:19 "-7.005 -254 10:50 0.6257 6.359 +417 12:50 * -7.233 -262 1137 " 5.795 +379 12O " * 5.443 +356 Run No. 139, Catalyst- Co0.96Fe20404, Temperature 250'C 1257 " 0.6275 5.114 +334 3:33 Helium 0.0732 -4. 724 -265 1155 " 0.6230 4.835 +317 2:42 4.690 +306 3:38 He+2cm H2 0.0728 -4.825 -271 3:40 * -5.19 -290 4:02 Helium 4.931 +324 3:44 "-5.255 -294 3:153 -5.30 -297 4:05 He+6cm 02 0.6073 8.156 +544 3; 56 -5.574 -12 4:lo " 0.6023 8.297 +563 4:00 "-5.297 -297 4:19 * 8.555 4:o4 * -4.860 -272 4i26 * 8.761 4:18 * -4.358 -244 4:40 * 8.879 4:23 "-4.323 -241 5:00oo " 0.6326 8.951 +579 4:40 " -4.208 -236 6150 0.6473 9.135 +579 4:52 -4.190 -234 7i40 9.141 8:4o * -3.845 -215 8:40 8.944 9:50 " 8.963 Run No. 141, Catalyst- Col. 09Fel. 9104 Temperature 250"C 10:35 " 8.908 9:10 Helium * 7.216 +384 11:20* 9.287 9:30 * 7.160 +381 12:20 * 9.290 9:45 * 7.227 +385 9:55 7.777 +385 Run No. 147, Catalyst- Col. 0Fel.2104, Temperature 250"C 10:20 Helium 0. 5370 5.5662 +445 9:56 He+6cm 02* 10.010 +532 1027 " 0.5220 5.774 +452 9:58 lo10.184 +540 10:41 0.54o4 5.913 +450 10:02 10.265 +545 10:55 0.5497 5.946 +445 10:07 * 10.339 +550 " ll07 0.5514 5.955 +444 10:15 0.7702 10.386 +552 10:20 0.7690 10.436 +554 ll e+cm 0.5482 5.652 +422 10:25 0.7683 10. 462 +556 11114 0.5463 5.534 +435 10:30 0.7722 10.504 +557 11126 0.5416 5.528 +399 10:45 0.7744 10.589 +558 11:8 0.5471 5.057 +379 11:00 " 0.7608 10.577 +53 5439 +75 12:10 0.5469 4.879 +367 1:00 Helium 10.119 1:10 ". 5528 4.870 +362 1:04 He+6cmH 0.7605 9.773 +527 2:49 Helium 0.5727 4.559 +336 1:06 0 9.503 +513 3:06 0.5377 4.579 +349 1:09 9. 284 +501 3:40 " 0.5457 4.625 +347 1:10 9.109 +496 4:47 0.5578 4.656 +347 1:20 * 8.654 +466 8:20 0.5873 4.970 +346 l: 25 " * 8.435 +454 1:25 * 8007 +432 8:29 He+2cm02 0.5776 7.491 +533 1:50 7.567 +406 8:31 "0.5730 7.540 +538 21053 7.526 +394 8:46 0.5764 7.618 +542 2:21 * 7.023 +378 8:56 0.5776 7.590 +543 2:55 "* 6.510 +351 9:00 " 0.5724 7.593 +543 4:05 " * 6.120 +330 9:15 0.5710 7.836 +546 6:30 "* 5.601 +302 7:50 "* 5.462 +295 ~* T assumed unchanged from previous reading *AT assumed unchanged from previous reading

-111B, Raw Data and Calculated Thermoelectric Power for Adsorption Runs TABLE XII (continued) Adsorption Runs Time Atmosphere AT (millivolts EMF Thermoelectric..Cr-Al couple) (millivolts) Power AV/~C Run Noo 158, Catalyst- CoO 96Fe2 o4~4 Temperature 88~C 11:20 Helium 0 1130 _-0o8968 -500 11:50 Tt 0 1061 -07759 -299 12 05 " 0 1055 -0 7689 -300 12:15 He+6cm H2 0o1051 -0.7689 -500 12: 25 oo 1041 -0o7650 -501 12: 40 r 01036 -0o7605 -501 1:12 r O 1042 -0o 7656 -500 1:55 T 0.1050 -0.7650 -299 5:25 rr 0 1095 -0 7940 -298 Run No, 158a, Catalyst- Co0. 96Fe2 O404 Temperature 120~C 5^40 He+6cm H, 0o2045 -1.5015 -505 5:45 0r O 2840 -2.0390 -294 5 47 t 0,3184 -2 3477 -303 5 55 0,3721 -2 6989 -297 6 05 oT 0.4056 -29261 -296 6: 3o 0,4276 -350847 -296 8:00 0rr 4349 -351286 -296 Run No, 159, Catalyst- Co0 96Fe2 o004^ Temperature 180~C 8530 Helium Oo5795 -4,5194 3-06 8:45 Tr 0 5776 -4 3094 -306 9t00 He+lOcm H2 0o5668 -4, 2454 -206 9:10 r 05664 -4,2436 -306 10:00 O0 5719 -4,2201 -301 11:45 tr 0.5738 -4 1733 -298 2:45 tr 0o5699 -4,o800 -293 6:25 tr 0 5706 -4 0277 -289 9:30 t 0,5762 -4,o454 -287 8:OOam. " 0,5653 -359265 -285

APPENDIX II SAMPLE CALCULATION Sample calculations of activation energy and pre-exponential factor for a typical hydrogen-deuterium exchange run are given belowo Run No, 111, using Co1 03Fel 9704 catalyst, is selected for this example solution. The data plot for Run No, 111 is given in Figure 21. The calculation techniques for determining the activation energy and pre-exponential factor are outlined in Chapter II, section D. A. Activation Energy Calculation As indicated by Equation (41), the activation energy may be determined from the slope of the RvaTijvs. 1/Ti plot, where Ti is the temperature required at each flow rate vi to reach the constant per cent conversion, po The procedure was repeated for three values of per cent conversion, (ioe., i = 1, 2, 3)o Per Cent Conversion Va Temperature fvaT\ 1 p cc,/min. OK vaT /rv 1 T.275 9 95 333 5 3318 5,804.0029985 150 0 338,2 5073 6 229.0029568 20 00 341, 9 6838 6o 527 o0029248 350 9 95 337.1 3354 5o815.0029664 15,00 341o9 5129 6,240.0029248. 425 20,00 345o7 6914 6.538.0028927 9.95 340,1 3384 5 284.0029403 15.00 345o 5176 6,249,0028977 20,00 348,8 6977 6,548.0028666 112

-1134,4:,3. \! \ \ —'.3 L..Lj (,.a 280 285 2so z S'5 o!/T w lo5Per Cent Activation Energy Calculation Conversion E (10 987)(slope) 6 r 5 - 5.8 (103)o,275 E (1.987) [299 8 292 6 (10-5 ) = 19.0 Kcal/mole.550 E - (1,987) [2968 - 289 6 (105) = 19,2 Kcal/mole r 65 5 8 1 (10),425 E (1,987 t294,3.- 286.8 J (10o5) = 186 Kcal/mole B, Pre:-exponential Factor Calculation As indicated by Equation (54), ~ -do- E/T + T. P _ (54)

-l14where mn is the slope of a plot of L\ tL]| vs-. L, The values of p ~a and va for this plot are determined from Figure 21, at constant temperature. This procedure is repeated for three values of temperature. The smaple calculation below is made at a temperature of 67.5~Co 1 [11 1 Va p 1=p miPL va 20.0 250 1.333.288.0500 15.0.324 1.480.392.0667 9.9.437 1.778.576.1005 -.3 - --...........i..,a.o;,, 0 iI1/ 0,,04,oG,1/

-115~ko [VTa + -- ~0 LaVT P kT ok = At (5a6)(3l405)(l146) + 19.0 6( 43 )(296.o)(14 6)j J (1o 987)(540.5) Ako' 2.7 + 28.1 = 3508.

AP!PENDIX III COMPARISON OF FERRITE MATERIALS USED IN EXCHANGE STUDIES AID THERMOELECTRIC POWER STUDIES -116

A. Hydrogen-Deuterium B. The1rmoeletric Exchtange Studies Power Stu;dies C0. 98iFe2. 02041 Co. le. 040 Crushed Pellet Powder I. X4-Ray Diffracti. on Pattern 2, Bulk Densi)ty 5.2 gm./cc. 5.5 gm./cc. Figure 55. Comiparis on of X-Ray Diffraction Patterns and True Bu:lk DensiLies of Ferrite Materials Used in Exchange Studies and Theimtoe-lectric Power Studies,

-118TABLE XI.; C0o. 96Fe2. 0404 X-Ray Diffraction Pattern Calculation Line Line 2 2 2 Number Intensity** mm 2G dh +k +1 1 30 29,880 21 32 4.84 3 2 50 28.495 35.17 2.91 8 3 100 27.865 41.47 2,256 12 4 4 27.665 43.47 2.415 12 5 2 27.490 45 22 2.326* 6 50 26.950 50.62 2.092 16 7 2 26 475 55 37 1.925 19 8 30 250700 63512 1.718 24 9 70 25.265 67.47 1,610 27 10 80 24 575 74 37 1,480 32 11 3 24,175 78 37 1,415 35 12 2 253850 81.62 1.368 * 15 5 235515 84.92 1.374 40 14 20 23 120 88 92 1 277 43 15 5 22,990 90.22 1.263 44 16 5 22.465 95 47 1 209 48 17 2 22 070 99.42 1 172 51 18 10 21,415 105.97 1,120 56 19 50 21.010 110,02 1 091 59 20.20 20 290 117 22 1 098 64 21 10 19 035 129, 77.988 72 22 45 18,510 135502.968 75 23 16 18 315 136.97.961 76 24 30 17.485 145.27.937 80 25 5 16 720 150 92.920 83 *These two lines are caused by iron contamination in the cobalt target used in the x-ray diffraction apparatus, These lines are apparent for the two strong intensities, 100 and 80, The ratio of the d values of the strong intensity lines to the d values of the extra lines should equal the ratio of the iron and cobalt wave lengths, 1.098o This calculation is given below: 45,22/41.97 = 1,098 and 81,62/743.7 1o098 **Center = 32.012

-119APPENDIX IV X-RAY FLOURESCENT SPECTROMETER ANALYSES DATA TABLE XIV. X-Ray Flourescent Data for Catalyst Samples Sample Side Peak Height Fe/Co Average Std. Dev. Sample Side Peak Height Fe/Co Average Std. Dev. Fe Co Fe Co Col. 07Fe. 9304 1 75.2 76.2.987 Co0.98Fe2.0204* 1 83.2 74.8 1.112 77.6 77.6 1.000oo 85.2 74.1 1.150 76.3 78.3.974 84.7 74.2 1.142 78.0 76.8 1.016 85.3 74.0 1.153 76.8 77.8.987 85.5 75.6 1.131 77.2 77.2.1.000 84.8 74.6 1.137 78.1 77.1 1.013 84.9 95.0 1.132 78.1 77.7 1.005 85.3 74.5 1.145 78.1 78.2.999 85.7 74.3 1.153 78.3 78.4.999 84.9 75.3 1.127 78.7 77.0 1.022 83.9 75.1 1.117 79.4 79.0 1.005 85.6 75.0 1.141 1.001.0128 1.1367.0141 2 82.9 84.8.978 2 76.1 66.2 1.150 81.9 84.8.966 76.0 66.8 1.138 82.4 84.2.979 75.6 66.0 1.145 83.4 83.0 1.005 75.5 65.6 1.151 82.4 84.3.977 73.9 65.0 1.137 84.0 83.6 1.005 74.8 66.5 1.125 81.8 83.8.976 74.8 66.4 1.127 83.0 84.5.982 76.8 65.4 1.174 82.5 83.9.983 75.1 67.0 1.121 84.5 84.5 1.000 74.4 66.1 1.126 84.0 84.8.991 75.5 66.2 1.140 83.3 84.2.989 74.7 64.9 1.151 0.9859.0118 1.140.0141 *Sample reground between each analysis Co 03Fel. 0704 1 74.5 71.2 1.046 Co 93Fe2 0704 1 78.6 64.3 1.222 74.4 71.9 1.035 77.6 63.9 1.214 76.3 71.6 1.066 78.9 63.4 1.244 76.0 71. 3 1.066 78.4 64.2 1.221 75.7 71.2 1.063 78.9 64.5 1.223 76.7 71.8 1.068 77.9 63.1 1.235 75.8 72.7 1.043 79.3 64.2 1.235 75.6 72.6.o041 78.2 64.4 1.214 75.5 72.7 1.039 78.3 63.2 1.239 75.6 72.2 1.047 78.8 64.o 1.231 75.4 71.4 1.056 79.0 62.4 1.266 75.9 71.6 1.060 78.7 63.7 1.235 1.0525.0114 1.2316.0139 2 83.5 79.5 1.050 2 86.1 71.7 1.201 84.5 81.0 1.043 86.4 70.7 1.222 84.0 81.2 1.034 85.4 70.4 1.213 84.2 80.2 1.050 86.7 70.1 1.237 84.7 81.3 1.042 86.5 70.0 1.236 84.3 81.5 1.034 86.5 71.5 1.210 83.6 79.7 1.049 85.7 70.1 1.222 83.4 80.2 i.o4o 86.5 71.4 1.211 84.5 80.8 1. 46 87.5 70.7 1.238 83.4 80.8 1.032 87.9 70.9 1.240 84. 3 81.2 1.038 86.6 71.4 1.213 85.4 81.0 1.054 86.4 70.5 1.226 1. 0427.0022 1.222.0125 TABLE XV. X-Ray Flourescent Data for Mixtures with Known Fe/Co Ratios Sample Side Peak Height Fe/Co Average Std. Dev. Sample Side Peak Height Fe/Co Average Std. Dev. Fe Co Fe Co Co0.90/Fe.0 1 83.4 64.3 1.297 Col1/Fel.90 1 75.2 79.6 0.945 83.8 64.5 1.299 75.9 79.2 0.958 83.2 64.2 1.296 75.6 79.7 0.949 83.5 64.7 1.291 76.0 80.o 0.950 83.5 64.6 1.293 75.3 79.3 0.950 83.2 64.4 1.292 75.5 81.2 0.930 1.295.0091.947.0085 2 85.5 67.0 1.276 2 76.8 80.3 0.956 85.9 66.7 1.289 76.7 80.8.o040 85.2 66.6 1.279 76.3 80.2 0.951 85.4 67.0 1.275 76.0 80.o 0.950 86.0 67.3 1.278 75.9 80.0 0.949 85.1 67.4 1.263 76.4 80.7 0.947 1.277.0077.951.0029 Co. 0/Fe2.0 1 80.8 72.9 1.108 81.4 73.3 1.111 81.1 72.9 1.112 81.2 72.7 1.117 80.9 72.7 1.113 81.3 73.4 1.108 1.112.0098 2 83.0 75.4 1.101 83.7 75.6 1.107 83.4 75.9 1.099 83.2 76.5 1.089 83.5 75.5 1.106 83.1 76.5 1.086 1.098.0091

APPENDIX V CHEMICAL ANALYSES OF RAW MATERIALS TABLE XVI. Analyses of Fe203 and CoC03 Fe203 (Baker and Adamson Quality, Reagent grade, General Chemical Division, Allied Chemical and Dye Corp.) Assay (Fe20O) min. 99.0 % Maximum Limits of Impurities Insoluable in HC1 0.20 Sulfates (So4) 0,20 Copper (Cu) 0.005 Zinc (Zn) 0.005 Substances not precipitated by NH40H (as sulfates) 0.10 CoC0 (Reagent grade cobalt carbonate, Jo To Baker Chemical 3 Company) Assay as Co 47.8 % Insoluable in HC1 0.005 Chloride (C1) 0.001 Nitrogen Compounds (as N) 0 005 Sulfates (So4) 0.003 Lead (Pb) 0.003 Copper (Cu) 0o002 Iron (Fe) 0.001 Nickel (Ni) 0.05 Alkalides and Earths (SO4) 0o24 L-120

-121APPEMX VI ENERGY OF FORMATION OF CoFe204 The standard free energy of formation of CoFe204 may be calculated from the following data: Standard Free Energy of Temperature Reaction* Reaction AGT, cal. Accuracy Kcal. Range ~K Reference 2<C 0> 2<Co> + (02) 111,800 - 33.8T 2 298 - 1400 57 <FeO) <Fe> + 1(02) 55,620 - 10.83T 3 298 - 1642 57 <Fe304) =5 FeO> + (02) 74,620 - 29.9T 3 298 - 1642 57 3<Fe20> 2(Fe304> + 1(02) 59,620 - 33.62T 8 298 - 1460 57 (Co> + <Fe20> = (CoFe2O4> -5,000** 1275 * (H2) + (02) = (H20) -58,900 + 13.1T 1 298 - 2500 57 Calculated Standard Free Energy of CoFe2r04 <Co> + 2(Fe> + 2(02) )(CoFe204)> GT -241,490 + 69.67T Equilibrium Oxygen Pressure,:~ o S - 241,490 + 69.67T 7. - 26 420 10 l 2O2 = T 7.64 - 0 ^ t- q9.14T T PH20 If P(02) is replaced by the equivalent -, the following equations apply: <Co> + 2<'e) + 4(<H20> --- oFe204> + 4(<B. GT = -5,890 + 17.27T PH20 -5,890 + 17.27T 522'lo 1. 18_28T 7.6411 T * ( >indicates solid, ( ) indicates gas **Estimate from data by Schmalzried( 82) for similar spinel reactions.

BIBLIOGRAPHY lo Aigrain, Po and Dugas, C, Z,, Z, Elektrochem,, 56, 3635 (1952). 2, Alkhazov, To G,, Belenskii, Mo S,, Izvestia Vysshikh Ucheb, Zavendenii, Neft i Gas, 3, 735 (1960), 35 A, P, I, Project 44, "Mass Spectral Data", Serial No, 452,453, (1950), 4. Becker, J, A., Green, Co B,, Pearson, Go L,, Trans. Am. Inst. Elect, Engrso 65 711, (1946), 5, Bevan, D. Jo M, and Anderson, Jo So, Disc, Faraday Society, 8, 235, (1950). 6, Bielanski, A., Deren J,, and Haber, Jo, Nature, 179, 668, (1957). 7o Bielanski, A., Deren J,, Haber, J,, and Sloczynski, J,, Proc. International Congress on Catalysis,2nd, Paris, 1960, 2, 16535 (1961)f 8 Bielanski, A,, Deren J., Haber J,, Sloczynski, J., and Wilkowa, J,, Bull, acada polon, sci,, Sera Sci,, Chim geol, et geograph, 1^ 3335 (1959), 9, Bleakney, W,, Phys, Rev. 40, 496, (1932) J 41, 32, (1932). 10, Block, JO and Chon, H, Z,, Elektrochem,, 60o 912, (1956). 11. Bloem, J., Philips Research Repts,, 13, 167, (1958), 12. Boreskov, G. K,, Doklady Akad, Nauk SSSR, 127, 591, (1959). 13. Brewer, A, K,, J, Phys. Chem,, 32 1006, (1928)o 14o Cimino, A,, Molinari, E,, Cipollini, E,, Gass. Chim, ital, 90, 79, 91, 120, (1960). 15 Culver, R, V, and Tompkins, F, C,, Advances in Catalysis, 11, 68, (1959), 16, de Boer, Jo H,, "Electron Emission and Adsorption Processes", Cambridge University Press, London, 19355 17. de Nobel, D,, Philips Research Repts,, 14, 361, (1959)o "122"

-12318, Dogramadzi, N, N,. Matic, Z. Bo. Bull. Inst, Nuclear Sci, "Boris Kidrich", 11L 155, (1961)4 19o Dowden, D. A,, Mackenzie, N,, and Trapnell, B, M, W., Proco Roy. Soc. A237, 245, (1956), 20, Economos, G,,' JO Am. Ceram. Soc., 38, 241, (1955 ) 21, Economos, G, JS Am, Ceram,, Soc., 38, 628, (1959)a 22, Economos, G, and Clevenger, T, R., Jro J, Am. Ceram, Soc,, 435 48, (1960). 235 Emmett, P. H., "Catalysts", Volz, Reinhold Publishing Corp., New York, 1955. 24. Enikeyev, E, H,, Margolis, L, I,, and Roginskii, S, Z., Doklady Akad. Nauk SSSR, 124, 606, (1959), 25, Evans, U, R,, "Metallic Corrosion, Passivity, and Protection", Edward Arnold and Co, y London, 1937, 26, Farrar, R, L,, and Smith, Ho A, J, Phys, Chemo 59, 7635 (1955). 27, Finkelnburg, WV and Humbachg W,, Naturwiss, 422, 355 (1955), 28. Frilzche, H,, Z, Physik, 133, 422, (1952)o 29,1 Fukutome, M. and Kusano, K,, Kogyo Kagaku Zasshi, 6, 1186, (1960), 30. Garner, W, E,, Advances in Catalysis, 9, 169, (1957). 31. Garner, W. E.. Gray, T. J,, and Stone, F, S,, Proc. Roy. Soc. A197, 296, (1949), 32. Garrett, C. G. B., Jo Chem, Phys., 335 966, (1960)o 33. Gorter, E- W,, Proc, I, R, E, 43 1945 (1955). 34, Gray, T, J,, Disc, Faraday Society, 8, 331, (1950), 35. Halpern, J,, Advances in Catalysis, 11, 301, (1959). 36, Harrison, La G. and McDowell, C, A,, Proco Roy, Soc,, A228, 66, (1955). 37, Hauffe, K., Advances in Catalysis, 7. 2135 (1955), 38, Hauffe, K, and Engell, H. LS,, Z Elektrochem,, 56_ 366, (1952)o

-12439. Hauffe, K, Glang, R,, and Engell, H. J., Z, physik Chem., 201_ 223, (1952), 40~ Hauffe, K, and Vierk, A, Lo, Z, physik Chem., 196, 160, (1950). 41. Heckelsberg, Lo F., Clark, A., and Bailey, G, C, J. Phys. Chem. 60, 559, (1956), 42, Holm, Vo C. F. and Blue. Ro W,, Ind. Eng. Chem,, 44, 107, (1952). 43O Huston, A, R,, in "Semiconductors", (Hannay, N, B., editor), p. 541, Reinhold Publishing Corporation, New York, 1959. 44. Jonker, Go Ho, unpublished measurements, (mentioned in Ref. 46). 45, Jonker, G, H,o JO Phys. Chemo Solids, 9, 165 (1959)o 46, Jonker, G, HR and van Houten, S,, in "Halbleiterprobleme Band VI", p. 118, Verlag Friedro Vieweg and Sohn, Braunschweig, 1961. 47, Keier, N. P. and Chizhikova, Go I,, Doklady Akad, Nauk SSSR, 120, 8301 (1955)0 48, Keier, N. P. and Kutseva, L. N,, Doklady Akado Nauk SSSR, 117, 259, (1957). 49. Keier, N. P,, Roginskii, So Z., and Sazonovo, Io S., Izvest, Akad. Nauk SSSR, Ser Fiz, 21, 183, (1957). 50. Kirshenbaumn I., "Physical Properties and Analysis of Heavy Water", (Urey, Ho C. and Murphy, Go M,, editors), McGraw-Hill Book Company, Inc,, New York, 1951, 51. Kittel, C', "Introduction to Solid State Physics", John Wiley and Sons, New York, 1956. 52, Kmetko, E. A,, Phys, Rev, 99 1642A, (1955)o 535 Korsunovskii, G, A,, Doklady Akad. Nauk SSSR, 134, 1394, (1960). 54, Krawczynski, Dissertation, Munich, (1956). 55. Kroger, F. A,, Vink, H, J., and Volger, J., Physica, 20, 1095, (1954)> Philips Research Repts. 10, 39, (1955). 56~ Krusemeyer, H, JO and Thomas, D, G., Jo Phys, Chem, Solids, 1, 78, (1958), 570 Kubaschewski, 00o and Evans, E, Lo,, "Metallurgical Thermodynamics", Pergamon Press, New York, 1958o

-12558, Kubokawa, Y, and Toyama, 0,, J. Phys, Chem, 60, 833, (1956)o 59. Kuchaev, J, L. and Boreskov, GC K,, Problemy Kinetiki i Kataliza, 10, 108, (1960)4 60, Langmiur, I,, J, Am. Chem. Soc. 38, 2221, (1916) 40, 1361, (1918); Trans, Faraday Society, 17, 607, 9(T922). 61. Law, J. T, in "Semiconductors", (Hannay, N, B,, editor), po 676, Reinhold Publishing Corporation, New York, 195 9 62, Leonard-Jones, J, G,, Trans, Faraday Society, 28 5333, (1932). 635 Linde, VT R, Margolis L. Y, and Roginskii, S, Z,, Doklady Akad. Nauk SSSR, 136, 86o, (1961), 64, Mason, D, R,, "Semiconductor Theory and Technology", Engineering Summer Conference, The University of Michigan, (1962). (To be published by McGraw-Hill Book Company, Inc., New York). 654 Matveev, K. and Boreskov, GC K,, Problemy Kinetiki i Kataliza, 8, 165, (1955). 66, Molinari, E. and Parravano, G,, J. AmChem, Soc., 75, 5233, (1955). 67, Morin, F, J, in "Semiconductors", (Hannay, N, B., editor), p. 600o Reinhold Publishing Corporation, New York- (1959). 68, Morrison, S, R., Advances in Catalysis, 7, 259, (1955)o 69, Myosnikov, I. A. and Pshezhetsky, S, Y,, Problemy Kinetiki i Kataliza, 8, 175, (1955), 70o Nier, A. 0 C,, Stevens, C, M,, and Rustad, B,, S. A, Mo Report A-5735 March 17, 1943, (71, Nyrop, J, E,, "The Catalytic Action of Surfaces", Williams and Norgate, London,. 1937 72, Otwinowska, H., Treszczanowixz, E,, and Ciboroski, S,, Actes intern. Congr. Catalyse, 2nda Paris 1960, 2, 1733, (1960), 735 Parravano, G, and Boudart, M,, Advances in Catalysis 7, 47, (1953)0 74, Parravano, G, and Domenicali, C, A,, J, Chem, Physo, 26, 359, (1957)o 75, Penzkofer, Dissertation, Munich, (1956),

-126 760 Rideal, E. Ko and Wansbrough-Jones, O0 H,, Proc, Roy, Soc., A123, 202, (1929)o 77O Rittenberg, D., Bleakney, W,, and Urey, Ho Co, J, Chem. Phys,, 2, 48, (1934). 78. Robin, Jo andBenard, Ja,g Compt. rend. 232, 1830, (1951); Compt. rendo 23, 734, (1952). 79. Roginskii, So Z., Problemy Kinetiki i Kataliza, Akado Nauk SSSR, Trudy Konf,, 1958, 10, 5, (1960), 80, Roginskii, S, Z, and Schultz, E., Z. physik0 Chem, A138, 21, (1928)0 81o Roiter, B. D, and Paladino, Ao E., J, Am, Ceram, Soc, 45, 128, (1962)o 82. Schmalzried, H,, Z. physik Chem, Neue Folge 25, 178, (1960), 83. Schmidt, 0,, Chem. Revso, 12, 363, (1933). 84. Schuster, M. Co and Fullamn Ed F,, Ind, Eng. Chemon, 18 6535 (1946), 85. Schwab, G, -M,, Angew, Chem,, 7 399, (1961), 860 Schwab, Go -M., and Block, J,, Z. Elektrochem, 58 756, (1954); Z. physik Chem, N. F,, lo 42, (1954), 87, Schwab, G. -M., Roth, E., Grintzos, CO,, and Mavrakis, N,, in'Structure and Properties of Solid Surfaces", p. 464, University of Chicago Press, Chicago, 19553 88. Slater, J, C., "Handbuch der Physik"'t Volo 19, J, Springer, Berlin, 1956, 89. Smiltens, R,, J. AmO Chemo Soc., 79, 4881, (1957)o 90, Solymosi, Fo, Magyar Tudomanyos Akad. Kens. Tudomanyak Ostalyonak Kozlemenyei, 13, 97, (1960), 91, Stockmann, F,, Z. Physik, 127, 5635 (1950)o 92. Svadlenak, R. Eo and Scott, A, B,, J, Am. Chem, Soc., 79, 5385, (1957). 935 Taylor, Ho S, and Liang, So CO, J. Am. Chem. Soc., 69, 1306, (1947).

-12794o Taylor, H. S. and Strother, C. 0O. J. Am. Chem. Soc9, 56, 586, (19534) 95. Temkin, M, Io, J. Phys. Chem, (UoS. SR.) 15, 296, (1941). 96, Urey, Ho C. and Teal, G, K,, Revs, Modern Phys. 7, 3 4, (1935), 97, van Houten, S., J, Physo Chem, Solids, 1 7, (1960), 98, van Uitert, L. G,, J, Chem, Phys,, 23, 1883, (1955)3 24, 306, (1956), 99. Verwey, E, Jo W., in "Semiconducting Materials", p, 151, Butterworth Scientific Publications, Lnondon 1951o 1000 Verwey, E, J. W., Haaijman, P. W,, Romeijn, Fo C,, and- van Oosterhout, G, W,, Philips Research Reptso, 5 173, (1950). 101, Wagner, C., Z, physik, Chem, B22, 181, (1933). 102. Wagner, C., J, Chem, Phys,, 18, 69, (1950), 103. Wagner, C. and Hauffe, Ko, Z, Elektrochemo, 44_4 172, (1938)o 104. Watson, H., Jr., J. Applo Physo, 32 120, (1961). 105a Weisz, P. B., J. Chem. Phys. 20, 1483, (1952)J ibid 21, 1551 (1955)o 106o Weller, SO W, and Voltz, S. E,, J. Am. Chem< Soc, 75, 5227, (1953)3 Z physik, Chenm Frankfort, N, S., 5, 100 (1955)o 107. Wicke, E,, Z, Elektrochem., 53, 279, (1949). 108. Winter, E. Ro S., Advances in Catalysis, 10, 196, (1958). 109, Wolkenstein, Th,, Uspekhi Fiz, Nauk, 60, 249, (1956). 110o Wolkenstein, Th, J, chim. phys., 54, 175, (1957), 111, Wolkenstein, Th,, Advances in Catalysis, 9, 807, 818, (1957). 112. Wolkenstein, Th,, Advances in Catalysis, 12, 189, (1960). 1130 Wolkenstein, Th,, "Theorie Electronique de la Catalyse sur les Semi-conducteurs", Mason and Cie, Paris, 1961,

NOMENCLATURE A Surface area, m2/gm. al*, a6 Proportionality constant between pressure runs and mass spectrometer peak height (see Table II) bl...b7 Proportionality constant between (pressure)2 and mass spectrometer peak height (see Table II) b Adsorption coefficient bo Defined by Equation (24) d Electron jump length, cm, see Equation (1) d Average particle diameter, microns d Average interplanar spacing (see Table XIII) E Activation energy in Arrheius equation, Kcal/gm. mole E1 Energy of adsorption E2 Energy of desorption E01 Defined by Equation (25) E02 Defined by Equation (26) Ea Energy of acceptor level Eb Energy of surface acceptor level Ed Energy of donor level Ef Fermi level E Energy of conduction level g e Charge on electron e Base of natural logarithms h Planck's constant h,k,l Miller indices I Ionization potential -128

-129I. Mass spectrometer peak height (i=2,3,4.-refers to mass of ~1 molecule) k Boltzmann s constant k Rate Constant ko Pre-exponential factor in Arrhenius equation kl Rate constant for adsorption k2 Rate constant for desorption k01 Defined by Equation (27) k02 Defined by Equation (28) A' ^Natural logarithm (to the base e) m* Effective mass m Defined by Equation (53) N Concentration of sitesavailable for electrons and holes, cm'NA Total concentration of acceptors (ionized and unionized) NB Total concentration of surface sites NC Concentration of states in conduction level ND Total concentration of donors (ionized and unionized) N Concentration of states in valence level V nl Concentration of electrons in conduction level n2 Concentration of holes in valence levels P Pressure p Per cent conversion of deuterium to hydrogen-deuteride Q Heat of adsorption q Activation energy for lattice deformation

-130r Rate of reaction S Surface site for adsorption T Temperature, ~K t Time V Void volume of reaction, cc, v Flow rate through reactor at S.T.P. cc./min. Va Flow rate through reactor at actual temperatures and pressure, cco/min, w Energy of interaction between adsorbed atom and surface wl Rate of adsorption w2 Rate of desorption o <sExtra contribution to transport energy levels of electrons I< (kT) Extra contribution to transport energy levels of holes a^ Se (kT) f' Defined by Figure 6 and Equations (25) and (26) Indicates a difference (i.e., AT = T2 - T1) Thermoelectric power, ccv/~C Og Per cent surface coverage +/ Mobility, cm2/v sec. /7 Abbreviation for micron 0 Frequency of lattice vibrations, sec" aI Peltier coefficient /O Catalyst density rz LSummation Electrical conductivity

-131l Work function Subscripts 1 Electrons, or adsorption 2 Holes, or desorption A Acceptor level, or molecule A a Actual conditions B Surface acceptor level C Conductor level 0 Initial state t Time V Valence level

UNIVERSITY OF MICHIGAN 3 901115 03526 93211II 3 9015 03526 7932