THE VAPORIZATION THERMODYNAMECS 0F sous WWIDE OXIDE BROMIDES ' Thesis for the Degree of ma. MICHIGAN STATE UWSETY BALE E. WORK 1972 _—_ Hp .7. “ 1’ Michigan 3‘”. it? University This is to certify that the thesis entitled THE VAPORIZATION THEMDYNAMICS OF SOME LANTHANIDE OXIDE BROMIDES presented by Dale E. Work has been accepted towards fulfillment of the requirements for _2h..D..._ degree in W 34.” a £4 b‘jor professor Date—mm 0-7639 ABSTRACT THE VAPORIZATION THERMODYNAMICS OF SOME LANTHANIDE OXIDE BROMIDES By Dale E. Work The monoxide monobromides of Nd, Sm, Tm, and Yb have been prepared and their decomposition reactions igugaggg_have been examined. Pure samples of NdOBr and SmOBr were prepared by direct bromination of the respective sesquioxides; TmOBr and YbOBr were prepared by decomposition of the tribromide hydrates. Lattice parameters have been derived for each of these LnOBr phases. . i The high temperature decomposition modes of these phases in gaggg_ were examined by several techniques: (1) net weight-loss experiments in which the samples were converted to the sesquioxides; (2) solid state decomposition traces by X-ray diffraction in which the vaporization reactions were monitored by a series of X-ray diffraction patterns taken as the reaction progressed; (3) mass spectrometric analyses of the gaseous decomposition products; and (4) bulk effusate collection experiments in which the gaseous reaction products were condensed and analyzed by X—ray diffraction. The molecular beams which effused from Mo Knudsen cells were analyzed in the mass spectrometer; fragmentation patterns of each effusate were determined, as were appearance potentials for the gaseous ions NdBr2+, Nd+, SmBr+, TmBr +, TmBr + rs+, YbBr2+, 2 9 Yb+. Lattice parameters for the tribromides of Nd, Sm, and Yb were YbBr+, and obtained from the bulk effusate collection experiments. Each of the LnOBr phases was found to decompose to the sesquioxide by the two-step reaction sequence: Dale E. Work 4 LnOBr(s) - 3 Ln Br(s) + LnBr3(g) (l) 304 3 Ln304Br(s) - 4 Ln203(s) + LnBr3(g). (2) For the SmOBr and YbOBr decomposition processes, the mass spectrometric data indicated that the respective LnBr2(g) species were the primary gaseous bromides present; this discrepancy is discussed. For the TmOBr and YbOBr vaporization reactions in No Knudsen cells, three solid phases (LnOBr, LnBOaBr, Ln203) were observed to exist simulta- neously; this discrepancy is also treated. Pure samples of Nd304Br and SmBOQBr were prepared from equimolar mixtures of the respective Ln203 and LnOBr phases at 1000° in evacuated quartz ampoules. The TmBOABr and Yb3O4Br phases were identified but not isolated. Lattice parameters for each of these Ln3O4Br phases have been determined. Knudsen effusion target collection data were obtained for reactions (1) (the NdOBr and SmOBr systems) and (2) (the Sm3043r system only). Targets were analyzed for microgram quantities of lanthanide by an X-ray fluorescence technique. Molybdenum Knudsen cells were used in each study. The measured Knudsen pressures were found to depend on the orifice size and approximate vaporization coefficients for the NdOBr, SmOBr, and Sm3043r decomposition systems were determined to be 10-2, -3, and 10-3, respectively. 10 For each series of target collection experiments, pressure data obtained with the smallest orifice were taken as equilibrium pressure data, and the corresponding temperature ranges were 1683—1850 K (NdOBr system), 1547-1768 K (SmOBr system), and 1636-1775 K (Sm304Br system). The following second law enthalpy and entropy changes were obtained Dale E. Work . ° . t for the decomposition reactions (1) and (2). AH1766 109.17 0.76 eu (NdOBr system); AH° - 97.6 1657 3 * 0'3 kcal/gfw, A81766 = 48.22 t 0.43 O B . O 3 kcal/gfw, A8165? 46.63 i 0.19 eu (SmOBr system), AH1705 99.73 1 0.81 kcal/gfw, [88:705 = 46.21 i 0.48 eu (Sm304Br system). Estimative schemes were devised by which second law enthalpy and entropy changes 2 and third law enthalpy changes at 298 K were calculated. For the NdOBr decomposition system, AH§98 (second law) 8 115.3 3 3.6 kcal/gfw, o I o a A3298 55.1 * 4.1 eu, AH298 (third law) 108.9 0 . o "-" the SmOBr system, AH298 (second law} 103-4 * 3.2 kcal/gfw, A5298 1: ° = 1 . 0 4.0 eu, A3298 (third law) 100.4 7.5 kcal/gfw, for the o a i o I Sm304Br system, AH298 (second law) 104.9 3.5 kcal/gfw, A8298 52.0 t 4.3 an, AH§98 (third law) - 103.5 * 8.2 kcal/gfw. Standard enthalpies o o = of formation were also calculated for the oxide halides. AHf,298,Nd0Br * 8.2 kcal/gfw; for 53. -229 kcal/mole; AHf,298,SmOBr - -225 kcal/mole; AHf,298,Sm3OABr - -660 kcal/mole. The reasonableness of these data is discussed in terms of published thermochemical values for LnOCl species. Factors believed to contribute to the sub-unity vaporization coefficient are discussed with reSpect to the experimental data. THE VAPORIZATION THERMODYNAMICS OF SOME LANTHANIDE OXIDE BROMIDES By Dale E. Work A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1972 C1 ACKNOWLEDGEMENTS Many people have contributed to the completion of this thesis in a variety of ways, but the assistance of certain individuals is of sufficient magnitude to merit special recognition. Accordingly, I extend my sincere thanks to Dr. Harry A. Eick for the encouragement he has offered and the integrity he has exemplified; to my wife, Debbie, for her spirit of understanding and self-denial; and to my parents, whose steady encouragement and quiet expectations have, at times, provided my only incentive to complete this work. Additionally, the dialogue and debate with other members of the high temperature research group have been both valuable and vital. The financial support of the National Science Foundation, the Atomic Energy Commission, and the National Aeronautics and Space Administration is gratefully acknowledged. ii I. II. III. TABLE OF CONTENTS Introduction Previous Investigations of Lanthanide Oxide Bromides and Selected Lanthanide Bromide Systems A. Previous Investigations of Lanthanide Oxide Bromide Systems: A Detailed Survey 1. Preparative and Structural Studies 2. Thermochemical Studies 3. Other Studies Previous Investigations of the Bromides of Neodymium, Samarium, Thulium, and Ytterbium: A Cursory Review 1. Tribromides a. Preparative and Structural Studies b. Thermochemical Studies (1) Condensed Phase (2) Vapor Pressure Measurements (3) Other Studies 2. Dibromides a. Preparative and Structural Studies b. Thermochemical Studies Theoretical Considerations A. B. The Phase Rule Temperature Considerations 1. Temperature Measurement 2. Temperature Scale Thermodynamic Calculations 1. The Second Law Method 2. The Third Law Method 3. Estimation of Heat Capacities a. Lanthanide Tribromide Vapors b. Solid Lanthanide Oxide Bromides 4. Estimation of Entropies at 298 K Knudsen Effusion 1. The Method 2. Basic Equation for Knudsen Studies 3. The Target Collection Technique iii Page UIUIUJUJ CDNVVO‘O‘O‘MMU' 10 10 10 11 ll 12 14 14 14 15 15 15 16 16 TABLE OF CONTENTS (Cont.) IV. E. 4. Assumptions Common to Knudsen Effusion Target Collection Data Treatment 5. Qualitative Aspects of the Vaporization Coefficient 6. The Quantitative Determination of the Vaporization Coefficient 7. General Considerations Error Treatment Experimental Equipment and Procedures A. Equipment 1. Target Collection 2. X-Ray Fluorescence 3. Mass Spectrometric 4. X—Ray Powder Diffraction 5. Other Materials Preparative 1. Monoxide Monobromides of Neodymium, Samarium, and Europium . Monoxide Monobromides of Thulium and Ytterbium . Tetraoxide Monobromides of Neodymium and Samarium . Tetraoxide Monobromides of Thulium and Ytterbium . Other MkWN Analytical 1. Lanthanide Metal 2. Bromine X-Ray Diffraction X-Ray Fluorescence 1. Optimization of Spectrometer Sensitivity 2. Calibration of the Spectrometer 3. The Analysis Weight—Loss Experiments Decomposition Traces Mass Spectrometry Bulk Effusate Collections Vaporization (Target Collection) Experiments 1. Samples 2. Crucibles iv 17 22 25 27 28 29 29 29 3O 3O 30 3O 31 32 32 32 33 34 34 34 34 34 35 36 36 37 39 39 4O 41 42 42 45 45 TABLE OF CONTENTS (Cont.) 3. Targets 4. Temperature Measurement 5. Miscellaneous Details L. Sticking Coefficient Determination Results A.' Preparative l. Monoxide Monobromides of Neodymium, Samarium, and Europium 2. Monoxide Monobromides of Thulium and Ytterbium 3. Tetraoxide Monobromides of Neodymium and Samarium 4. Tetraoxide Monobromides of Thulium and Ytterbium B. Weight-Loss Experiments C. Decomposition Traces D. Mass Spectrometry E. Bulk Effusate Collections F. Target Collection Experiments 1. Crucible Interaction 2. Residue Analysis a. Neodymium Monoxide Mbnobromide Study b. Samarium Monoxide Mbnobromide and Samarium Tetraoxide Monobromide Studies 3. Vaporization Behavior of Neodymium Monoxide Monobromide a. Constant Orifice Size b. As a Function of Orifice Size 4. Vaporization Behavior of Samarium Monoxide Monobromide and Samarium Tetraoxide Monobromide a. Constant Orifice Size b. As a Function of Orifice Size 5. A/d Values 6. Vapor Pressure Equations a. Neodymium Monoxide Monobromide System b. Samarium Monoxide Monobromide System c. Samarium Tetraoxide Monobromide System 7. Data Reduction a. Neodymium Monoxide Monobromide System b. Samarium Monoxide Monobromide System c. Samarium Tetraoxide Monobromide System 8. Sticking Coefficient Determination 47 49 50 51 52 52 52 52 54 54 55 58 62 65 67 67 67 67 67 67 67 68 7O 70 71 71 75 75 75 76 77 77 79 81 81 TABLE OF CONTENTS (Cont.) VI. Discussion A. Evaluation of Experimental Work 1. Characterization of Decomposition Modes a. Weight-Loss Experiments b. Solid State Decomposition Traces (1) General Considerations (2) Decomposition Traces of the Monoxide Monobromides of Thulium and Ytterbium c. Mass Spectra (1) Fragmentation Patterns (2) Appearance Potentials d. Bulk Effusate Collections e. The High Temperature Decomposition Modes of Samarium Monoxide Monobromide and Ytterbium Monoxide Monobromide 2. Vapor Pressure Measurements a. Attainment of Equilibrium Conditions b. Target Collection Technique c. In Defense of the Knudsen Method B. Evaluation of Thermochemical Data 1. Second and Third Law Data 2. Enthalpies of Formation of Lanthanide Oxide Bromides C. General Comments and Observations 1. Experimental Routes to Definitive Answers a. Attainment of Equilibrium Pressures b. Vaporization Modes of Samarium Monoxide Monobromide and Samarium Tetraoxide Monobromide 2. Samarium as a Divalent Lanthanide 3. The Lanthanides: All Alike? 4. Future Investigations REFERENCES APPENDICES vi 82 82 82 82 83 83 84 85 85 87 88 89 95 95 99 99 100 100 100 104 104 104 104 106 106 107 108 116 12. LIST OF TABLES Spectrometer Settings Employed for X-Ray Fluorescence Analyses Orifice Areas and Associated Temperature Ranges Employed in the Target Collection Experiments Analytical Results Comparison of Derived and Published Lattice Parameters for Some LnO Br Phases x 3-2x Results of Weight-Loss Experiments Mass Spectrometric Results Published Fragmentation Patterns of Trichloride, Tribromide, and Triiodide Monomers Comparison of Derived and Published Lattice Parameters for the Tribromides of Nd, Sm, and Yb Comparison of Thermodynamic Values for BiClB(g) and AsI3(g) Fragmentation Patterns of SmBrx, YbBry, and EuBr LnBrZIP 2 ) Ratios As Functions 3 Estimated Knudsen Cell (P of T and PLnBr LnBr 2 Estimated Standard Enthalpies of Formation of LnOBr(s) Phases vii Page 37 48 53 55 59 63 64 66 77 86 91 103 FIGURE LIST OF FIGURES Apparatus Employed to Collect the Condensed Bromide Vapor from the Decomposition of SmOBr(s) Effusate Collection Apparatus Knudsen Cell/Oven Configuration Employed in Target Collection Experiments Measured Knudsen Pressures of NdBr3(g) over a NdOBr(s)-Nd3043r(s) Mixture Measured Knudsen Pressures of SmBr3(g) over a SmOBr(s)—Sm3O4Br(s) Mixture Measured Knudsen Pressures of SmBr (g) over a Sm304Br(s)-Sm203(monoclinic) Mixture Equilibrium Pressures of SmBri(g) over SmOBr(s)—Sm3OABr(s) and Sm O Br(s)-Sm O (monoclin c) Mixtures 3 4 2 3 , viii Page 43 44 46 69 72 73 74 LIST OF APPENDICES APPENDIX I. II. Comparison of Observed and Calculated Interplanar d-Values IA. Tetragonal NdOBr IB. Tetragonal SmOBr IC. Tetragonal TmOBr ID. Tetragonal YbOBr IE. Orthorhombic Nd304Br IF. Orthorhombic SmBOaBr IG. Orthorhombic Tm3O4Br IH. Orthorhombic Yb304Br II. Condensed Effusate from Vaporization of NdOBr (Orthorhombic NdBr3 plus extra lines) IJ. Condensed Effusate from Vaporization of SmOBr (Orthorhombic SmBr3 plus extra lines) IK. Condensed Effusate from Vaporization of YbOBr (Hexagonal YbBr3 plus extra lines) Maximum and Minimum Values of A/d for the Target Collection Experiments III. Values of Afef for Reactions (V-l), (V-2), and (V-3) IV. Equilibrium Pressure and Third Law Enthalpy Data IVA. The Vaporization of NdOBr; Reaction (V-l) IVB. The Vaporization of SmOBr; Reaction (V-2) IVC. The Vaporization of Sm304Br; Reaction (V—3) ix Page 116 116 116 117 117 118 118 119 119 119 120 120 121 122 123 123 123 124 CHAPTER I INTRODUCTION Today, more than at any time in recent years, external demands are being placed on the scientific community by the larger society, and the reality of these demands is evidenced by the hesitancy with which both government and industry invest their resources in chemical research. At no time since the launching of the first sputnik have scientists been called upon more forcefully to supply a sound rationale for their research efforts and to justify the tax and/or profit dollars which these efforts digest. These societal concerns are understandable when one considers the volume of manpower and money expended on projects which are designed to confirm an earlier report or to provide an nth independent determination of a physical property (albeit to one additional significant figure). In contrast to such studies, the thermochemistry of lanthanide compounds comprises a research area which has been neglected until recent years and in which, even now, only fragmentary data are available. Aside from the oxides, only a few prototype lanthanide binary phases have been well characterized; more pointedly, lanthanide oxide halides have been known for many years but the characterizations of their thermal decomposition modes have been reported only in the last decade. To date, no thermochemical data have been obtained for the thermal decomposition reaction of a LnOBr phase. In this context, an investigation of the SmOBr decomposition reaction was judged to be both needed and justifiable. The samarium 2 system was selected because of the anomalous valency characteristics of samarium relative to the other lanthanides; while most of the lanthanide bromides are thought to be characterized by a single dominant lanthanide oxidation state at elevated temperatures (~1000°), there is much doubt and little data on the relative stabilities of the di- and tribromides of samarium (and, to a comparable extent, ytterbium) at these tempera- tures. Consequently, the vaporization behavior of samarium oxide bromides is not so susceptible to accurate prediction as are those of the more characteristically divalent or trivalent lanthanide oxide bromides. In an effort to isolate some of the parameters involved in this decomposition process, the NdOBr, TmOBr, and YbOBr systems were examined concurrently. The purpose of this research was to describe as thoroughly as possible the decomposition reactions of these lanthanide oxide bromide systems, and considerable progress has been made toward that end. To be sure, many questions remain unanswered, but both the answers supplied and the questions raised support the rationale upon which the research was based, vindicate the associated expenditure of resources, and underscore the need for the determination of fundamental thermodynamic values for simple binary and ternary lanthanide systems. CHAPTER II PREVIOUS INVESTIGATIONS 0F LANTHANIDE OXIDE BROMIDE AND SELECTED LANTHANIDE BROMIDE SYSTEMS A. Previous Investigations of Lanthanide Oxide Bromide Systems: A Detailed Survey 1. Preparative and Structural Studies The earliest mention of a lanthanide oxide bromide was made in 1949 by Zachariasen who reported1 that NdOBr is isostructural with tetragonal PbFCl. Sixteen years elapsed before Mayer, 35H31.2 and Barnighausen,.g£_al.3 reported further data for the monoxide monobromides. In an investigation which spanned all of the lanthanides except promethium, Mayer, 35 al.2 reported the preparation of the respective LnOBr compounds by the controlled thermal decomposition of the tribromide hydrates. A temperature range of 650-700° was employed for the lighter lanthanides (except Ce), with a lower temperature used for the heavier lanthanides (and Ce). Heating was effected for 40-60 minutes. The resultant LnOBr phases were examined by Xeray powder diffraction and were found to possess the layered PbFCl structure, space group P4/nmm. About the same time, Barnighausen, Brauer, and Schultz3 published their findings on the LnOBr and Ln3O4Br compounds of Sm, Eu, and Yb. These LnOBr samples were prepared either by the thermal decomposition of the tribromide hydrates in air or, in the case of EuOBr, by direct bromination of the sesquioxide at 700° in a stream of nitrogen. In the former method, lower temperatures and longer ignition times were employed relative to those of Mayer,_g£_§l. Structural data were essentially in agreement with those reported by Mayer and co-workers. 3 4 The SmBOABr phase was prepared by heating an equimolar mixture of SmOBr and Sm203 at 900° for several hours under nitrogen or, alternatively, by carefully decomposing SmOBr under a stream of dried air. The Eu 0 Br 3 4 phase was prepared by the former method, Yb 04Br by the latter. All 3 three tetraoxide monobromide phases crystallized with orthorhombic symmetry, but the precise space group to which the structures belong was not definitively identifiable from the powder diffraction data. A, few years later (1971) a single crystal study“ of the Eu3048r phase clarified the uncertainties?”5 and specified the space group as Dii-Bbmm. In 1967, Schultz and Reiter6 prepared NdBOéBr and La3043r and determined the structure of the former compound. The two preparative techniques employed were analogous to those used for the corresponding Sm species, and Nd304Br was found to be isostructural with the three Ln3043r phases cited above. Thus, in a span of about two years, the preparations and structural determinations of all the LnOBr compounds (except PmOBr) and of several Ln3043r phases were reported. One year later, Scherer7 reported the tentative preparation of PmOBr. The only divalent lanthanide oxide bromide known was reported by Haschke and Eick5 in 1970. They prepared Eu3OBr4 from the monoxide and the dibromide at 650-700° in the absence of air. The hexagonal monoxide tetrabromide was found to be isostructural with the MAOBrG phases of Ba and Sr, and more recent evidence8 suggests that the proper stoichiometry of the Eu phase is Eu4OBr6. In each of the structural studies cited above, lattice parameters and, with one exception, error estimates were given. 2. Thermochemical Studies Mayer and Zolotov9 conducted a thermogravimetric study of the decomposition modes of the lanthanide tribromide hydrates in air and reported the occurrence of LnOBr phases as discrete steps in the decomposition processes for all Ln except Ce. They further stated that the thermal decomposition of the LnOBr phases directly yielded the respective sesquioxides, but this conclusion is clearly erroneous in light of the reported Ln3O4Br preparations (Ln - La, Nd, Sm, Eu§ and Yb). Haschke and Eicks have provided the only thermochemical description of a lanthanide oxide bromide decomposition process. They studied the vaporization thermodynamics of reaction (II-1) 3 EuBOaBr(s) = 4 Eu203(monoclinic) + EuBr2(g) + Br(g) (II-1) and reported second-law values of AH°1399 - 129.031.1 kcal and AS°1399 - 50.81.:_O.81 eu for the process. Thermodynamic estimates were also made for the EuOBr decomposition. 3. Other Studies The only other references to LnOxB systems are the IR and Raman ry spectral work done by Basile, st 31.10 and the electron-transfer spectra reported by Barnes and Pincott.11 Both studies examined LnOBr samples. B. Previous Investigations of the Bromides of Neodymium,g§amarium, Thulium,iand Ytterbium: A Cursory Review 1. Tribromides a. Preparative and Structural Studies Several preparative techniques have been employed in the syntheses of anhydrous lanthanide trihalides, and these have been reviewed 6 recently;12’13 accordingly, the details are not presented here. One preparative method not reviewed is that of Brown, ggflgl.1“ in which the controlled thermal decomposition of the tribromide hydrates yields the anhydrous tribromides. Both NdBr3 and SmBr3 crystallize with an orthorhombic, layered PuBr3 structure-type with four molecules per unit cell.15 Hexagonal TmBr3 and YbBr3, with larger Br-IM.+3 radius ratios, exhibit the FeC13 structure-type, with the metal atoms positioned in the center of a near perfect octahedron of bromine atoms.1” b. Thermochemical Studies (1) Condensed Phase Virtually no thermodynamic data have been determined for these compounds. Dworkin and Bredig16 have measured enthalpy increments for solid and liquid NdBr3 from 298-llOO°, including the enthalpy offusion, but no other data are available. Spedding and Daane17 report YbBr3 to decompose to the dibromide before melting, and, in what was surely a grossly contaminated system, Jantsch and Skalla18 report spurious decomposition behavior for molten SmBr3. These sketchy, descriptive data are all that are available and reflect the empirical void which exists. This lack of thermochemical measurements has both permitted and encouraged the development of estimative schemes for these and related systems.19"23 (2) Vapor Pressure Massurements Harrisonzu reported the first vapor pressure data for NdBr3 and SmBr3. By his own admission, however, the work was very approximate, with only three data points determined for the SmBr3 sublimation and fewer than six for NdBr3. The vapor pressures cited correspond to a temperature accuracy of‘: 25°. 7 Ten years later (1962) Shimazaki and Niwa25 published sublimation vapor pressure data for several lanthanide trihalides including NdBr3. Only six (in P, T) data points were obtained for the system and no error estimate was presented. Although the data must be viewed with some reservations, they are consistent among the several LnX3 systems studied. Vaporization data have not been reported for TmBr3 or YbBr3. (3) Other Studies Several spectral investigations have been made, but the only one pertinent to this document is an electronographic investigation of the structure of the Ndx3 vapor species. Akishin, 35.31.26 determined that gaseous NdBr3 has a flat configuration with the Nd atom in the center of a right triangle flggg] of bromine atoms (Nd-Br distance - 2.72 A). The same workers subsequently reported”:28 gaseous NdBr3 to have a planar symmetric structure. 2. Dibromides a. Preparative and Structural Studies Of the four dibromides of interest, only SmBrz and YbBrz have been characterized. Tetragonal SmBrz is isostructural with EuBrZ and SrBrzu’29’30 while YbBr2 exhibits the orthorhombic CaClz structure- type.” Both compounds have been prepared by hydrogen reduction of the tribromide at elevated temperatures.31’32 The dibromides of Nd and Tm have not been isolated and definitively identified. Dworkin, 55 31.33 have inferred the existence of NdBr2 from the conductivity and solubility behavior of a fused Nd-NdBr3 system; TmBrz has been prepared and partially characterized, but the results have not been published.31+ 8 b. Thermochemical Studies Thermochemical data are singularly lacking for the dibromides under consideration, although estimative techniques have been suggested.“"’23 CHAPTER III THEORETICAL CONSIDERATIONS A. The Phase Rule "Classical chemistry, by which is meant not something outmoded but what is still the bulk of chemical investigation, rests upon the identification and characterization of substances and upon the study of their interaction. Implicitly or explicitly we apply the principles of the Phase Rule in almost all such considerations."35 In this context, the gist of this investigation is classical in its reliance on phase rule principles, yet modern--even pioneering--in terms of the compounds studied: Concisely expressed, the familiar phase rule is f - c - p + 2 (111-1) where f is the number of degrees of freedom of the system, c is the number of components, and p the number of phases. Apart from the phase rule, there are no universally agreed upon definitions of the terms "phase" and "component"; to be properly understood, the phase rule must be treated as an integral expression.35 For systems in which one condensed phase thermally decomposes to a second condensed phase plus a vapor species, the phase rule dictates that at equilibrium, f - c - 1. Such systems are pseudobinary since the three pure phase components are subject to the restriction expressed by the net stoichiometric equation (and associated equilibrium constant) relating them (and their respective activities). Accordingly, 9 10 f - 1 and at a specified temperature the vapor pressure is fixed. (An important assumption underlying this discussion is the immiscibility of the condensed phases.) B. Temperature Considerations 1. Temperature Measurement Temperature data obtained by optical pyrometry must be corrected for intensity diminution between the hot body and the pyrometer due to absorption and reflection by windows, prisms, etc., in the optical path. Details of the correction are supplied by Margrave36 and have been discussed in recent theses.37’38 The net result is that the quantity (1/Tob8 - llTact) is a constant for the effective wavelength monitored by the pyrometer. The numerical value of the constant may be determined by independent temperature measurements of a constant temperature glowing object external to the vacuum chamber and free of its optical constraints; measurements are made both by sighting the object directly (Tact) and by viewing it via the optical path of the vacuum system. The theoretical formulation of the expression (l/Tob8 - l/Tact) a k stems from Wein's radiation law which, at the A and temperatures employed in this study, is an excellent approximation to Planck's exact radiation law. Application of Beer's law to Wein's general formulation directly yields the corrective expression, with k a function of the transmissivities of the optical elements, the emissivity of the radiating body, and the A monitored. 2. Temperature Scale Temperature measurements were made, and calculations effected, with respect to the International Practical Temperature Scale of 1948. A more recent scale, the ITPS-l968,39 went into effect on January 1, 1969, 11 but the difficulties involved in the conversion of numerous literature values, viewed in light of the improved accuracy, were judged to make such corrections impractical at this time. Appropriate conversion tables are available, however.“0 C. Thermodynamic Calculations Conversion of the equilibrium pressure data as a function of temperature to the more conveniently employed expressions AH° and AS° was effected in two ways commonly denoted as the second and third law methods. ”I l. The Second Law Method The second law method for determining AH°vap over the experimental temperature range is a direct application of the integrated form of the Clausius-Clapeyron equation, - - o — in P - ( AH vapIRT) + C. (III 2) For the specific case in which a condensed phase thermally decomposes to a second condensed phase plus a single vapor species, equation (III-2) can be written for the system at equilibrium .. _ ° ° _ 2n K — ( AH vapIRI) + AS vapIR. (III 3) A plot of in K.y§ l/T yields a straight line with slope -AH°/R and intercept AS°/R. There are at least two assumptions in the above treatment which merit elaboration. First, equation (III-2) implies that the fugacity of the gas is the same as its pressure. For the Knudsen effusion experiments conducted in this study, vapor pressures ranged from 10'3 to 10'7 atmospheres, and ideal gas behavior may be assumed. Second, 12 the linearity of the in K !g_l/T plot implies the constancy of AH° and 118° over the temperature range. Although this is not a perfectly valid assumption, it is a reasonable approximation in view of the relatively narrow temperature ranges involved (W 200°). Related to this latter assumption is the ambiguity involved in assigning the temperature to which these second law AH° and AS° correspond. In a statistical treatment of this problem, Horton“2 has shown that the resultant AH° and AS° values do not unambiguously correspond to any recognizable temperature in the experimental temperature range. He has further demonstrated that the temperature to which the second law value corresponds is not very different from the median temperature for a data set with a good distribution of experimental points over the temperature range. In this thesis, second law values are cited for the median temperature of the examined range, consistent with Horton's findings. The second law AH°T and AS°T values are easily corrected to the corresponding entities at 298 K by the straightforward equations: AH° = AH° — [T A0 dT (III-4) 298 T 293 p AS° = AS° - T (AC /T) dT. (III-5) 298 T I298 P 2. The Third Law Method Unlike the two parameter second law method, the third law technique is essentially a one parameter treatment71 which employs free energy functions: fefT . -(G°T - H°298)/T (III-6) AfefT I -(AG°T - AH°298)/T. (III-7) l3 Substitution of AG° = AH° - TAS° and AG° - -RT in K in successive steps yields: a O | O O I o and AH°298 = TAfef - RT in K. (III-9) For an incongruent vaporization reaction yielding a single vapor species and stoichiometrically expressed in terms of one mole of gaseous product, AH°298 3 TAfo - RT 2n Pe (III-10) q' From known and estimated heat capacities and entropies, equation (III—8) can be used to determine values of AfefT which, in conjunction with experimental T in P data, yield a value of AH°298 for each experimental point. (Contrast this treatment with the second law approach which treats the collective data set and yields a single AH°298 value.) Accordingly, the third law results are much better suited to an analysis of systematic errors in the experimental procedure as revealed by trends in the AH°298 values with increasing (or decreasing) temperatures and/or exposure times, or exposure sequence. When heat capacity and entropy data are not available estimated values may be employed; the resultant third law AH°298 values also provide a check on the suitability of the approximations used. Additionally, a comparison of second and third law AH°298 values provides an indication of the internal consistency of the estimations made since AC is used P twice in the third law calculation but only once in the second law treatment, whereas the entropy estimate enters into the third law method only. Large discrepancies between second and third law results 14 indicate grossly incorrect C or $0298 estimates, systematic errors P in the experimental procedures, non-equilibrium conditions, or an incorrectly formalized reaction. It should be noted, however, that for formalized reactions in which large stoichiometric coefficients are present, the effects of incorrect thermodynamic estimates are magnified. 3. Estimation of Heat Capacities a. Lanthanide Tribromide Vapors In the absence of measured values, heat capacities for vapor species can be estimated in several ways. Perhaps the best method involves the estimation of vibrational frequencies from measured data on chemically and structurally similar compounds. In the case of the gaseous lanthanide tribromides, and more generally for LnX3 species (X is a halogen other than fluorine), virtually no spectral data are available. Consequently, the needed heat capacities were obtained from correlation with heat capacity data for other non-lanthanide trihalides. b. Solid Lanthanide Oxide Bromides Several schemes have been suggested"‘3"'*5 for the estimation of solid phase heat capacities, virtually all of which stem directly or indirectly from Kopp's rule: the heat capacity of a solid compound is equal to the sum of the heat capacities of its constituent elements. Admittedly, this rule serves only as a rough guide at best, but in the absence of a more clearly definitive scheme, it provides the basis for the estimations used in this study. Within this estimative framework, errors can be minimized by a judicious selection of prototype compounds. In this study, the Dulong and Petit formulation of Kopp's rule is employed in conjunction with the use of a condensed LnBr3 prototype species for the estimation of the Ln and Br atomic contributions to the 15 overall heat capacity. By use of these data, the atomic Cp contribution of oxygen is derivable from a consideration of corresponding Ln203 species in light of Kopp's rule. This technique was used to estimate Cp,293 values for each of the LnOBr and Ln304Br compounds studied. This treatment is given added credence by the fact that the lanthanide oxides and oxide bromides possess similar OLna structural units."6 Higher temperature Cp estimates were based on Kubaschewski and Evan's"5 estimate of a 7.25 cal/g-atom contribution to the overall Cp at the first transition temperature. For lack of contrary evidence, this estimative scheme was used to obtain Cp values at the median temperature of the vaporization experiments. 4. Estimations of Entropies at 298 K The techniques employed for estimating entropies are neither new nor peculiar to this work, and do not merit extensive treatment. Entropies for the gaseous LnBr3 molecules are based on measured values and on the assumption that the Nd and Sm tribromides exhibit similar and comparable behavior. Solid phase entropies are estimated by combining Latimer'sl‘7 estimate with Westrum's"8 formulation for the magnetic contribution of the Ln metal to the entropy. This estimative approach is well documented in recent papers.5,“9-51 D. Knudsen Effusion 1. The Method The fundamentals of Knudsen effusion are conceptually simple: a Ianivariant system consisting of an inert, isothermal cavity (enclosed £33E¢ept for a small, thin orifice) which contains a volatile, condensed material, will eventually obtain equilibrium conditions with respect 16 to the phase distribution of the system. For equilibrium vapor pressures in the range 10'9 to 10'3 atmospheres, the Knudsen method is used to monitor effectively the gaseous molecular flux through the orifice. 2. Basic Equation for Knudsen Studies It is clear that the rate of mass transport through the orifice should be proportional to the vapor phase molecular density (P/RT for dilute gases), the velocity of the molecules (E, assuming a Maxwellian distribution of velocities), and the orifice area (a). Accordingly, the rate of mass transport (r, in units of mass/time) is given by equation (III-ll) r - (M/4)(a)(z)(P/RT). (III-11) where M.is the molecular weight of the effusing species. Details of the derivation are given by Present52 and Carlson.53 Rearrangement and substitution of E'- (8RT/nM)1/2, yield P = (r/a)(21rRT/M)1/2- (III-12) Since the rate of mass transport is m/t, p = (m/at)(21rRT/M)1/2. (III-13) The mathematical formulation for the monitoring of several vapor species, monomeric and polymeric, is given by Lyubitov.5“ It is interesting to note that equation (III-13) also describes the pressure of a vapor emanating from an ideally evaporating Langmuir Sample positioned at the orifice with evaporating surface area, a. 3. The Target Collection Technique The target collection technique is one of several55 which employ theanudsen effusion method. In this technique, the molecular flux 17 through the orifice is monitored by condensing a well-defined fraction of the effusing vapor on a chilled target positioned over the orifice. For a circular target, coaxial with the orifice (with the orifice area much smaller than the area of the collection surface), the solid angle of the effusing vapor intercepted by the target can be calculated precisely. Based on the cosine law of evaporation the following relationship between the vapor pressure inside the cell and the mass of effusate collected can be derived: P = (m/at)(2nRT/M)1/2(R2 + L2)/R2, (III—l4) where R is the radius of the collector and L is the perpendicular distance from the orifice to the target. 4. Assumptions Common to Knudsen Effusion Tapget Collection Data Treatment Inherent in the above paragraphs are several assumptions: a. The pressure inside the Knudsen cell is sufficiently low so that the vapor molecules experience free molecular flow; i.e., the number of gas phase collisions is negligibly small. Although there is general agreement that spurious vapor pressure data result when hydrodynamic flow conditions prevail, there is no agreement on the limiting experimental conditions at which gas phase collisions become measurably significant. Typically, these limiting conditions are expressed in terms of l/d, where A is the mean free path of the molecules and d is the orifice diameter. Estimates of the maximum A/d ratio permissible range from 10 to 0.05.53955‘61 b. All molecules which strike the orifice pseudosurface, and only those molecules, permanently exit from the cell; the orifice is infinitely thin. 18 In practice, infinitely thin orifices are approximated by knife-edged ones. Nevertheless, every orifice is in fact a channel and molecular collisions with the walls of this passageway yield vapor pressures lower than the equilibrium value. Accurate corrective factors have been 62'6“ and in practice it is calculated for these scattered molecules, often desirable to use decidedly channeled orifices. An opposite effect has been treated by Winterbottom and Birth,65 who provide an approximate mathematical formalism for treating the diffusional contribution to the net mass transport. Specifically, their calculations suggest that under some conditions the number of vapor molecules which adsorb on the cell lid and migrate to the orifice wall before re-evaporation contributes significantly to the total molecular flux, leading to anomalously high vapor pressure data. c. The Knudsen cell is not only chemically inert to the various condensed and vapor species present, but is also physically impervious to them. Ward, ggugl.56’57 have convincingly dramatized the drastic effect which an apparently slight loss rate per wall collision can have on the measured vapor pressure. Like the considerations in (III.D.4.b.), this effect applies to all Knudsen effusion measurements but is especially significant in Knudsen effusion weight-loss experiments. d. The system is at constant temperature. The establishment of a uniform temperature distribution on the cell is frequently very difficult. The most crucial consideration from a gross thermodynamic point of view is that any temperature gradient present be the same for each target collected. A small constant temperature gradient has little 19 effect on the resulting second law thermal values, while even 3 5° change in the gradient during a single experiment makes a marked difference in the resultant enthalpy and entropy values. Similar effects are noted by Richardson and Alcock..68 The effect of a temperature gradient (discontinuity) at the interface of the condensed and vapor phases will be noted later. e. It is evident from equation (III-l4) that independent identification of the vapor species must precede meaningful treatment of the data obtained from the target collection experiments. f. Equation (III-ll) was derived from the kinetic theory of dilute gases, and the validity of equations (III-13) and (III-14) rests on the assumption that such kinetic considerations are applicable to Knudsen systems. 5 Although definitive evidence will not be presented here either to verify or disprove the validity of this assumption, three published viewpoints, each by reputable researchers, are noteworthy: "The Knudsen effusion method is based on the kinetic theory of dilute gases . .";53 "In fact, the assumption that the evaporation rate of metals is given by the kinetic theory expression . . . is . . . more of a convenience introduced as a plausible postulate than a deduction";69 "It has been shown that the behavior of a Knudsen cell Operating in the molecular flow region is governed by the properties of the cell and sample surfaces rather than by kinetic theory."66 g. Motion of the molecules inside the Knudsen cell is equally probable in all directions, and the effusing beam (from an infinitely thin, small orifice) is characterized by a cosine distribution. (The cosine law states that the vaporizing flux at a specified distance from a differential source of vaporization is directly proportional to 20 the cosine of the angle between the direction of molecular flow and the normal to the vaporizing surface.) While the cosine law must describe the evaporation behavior of a system in which a gas is in equilibrium with the sample surface,70 there is not general agreement on the extent to which the law holds for Knudsen effusion experiments in which gas collisions are rare and the orifice introduces non-equilibrium conditions. WhereasPresent52 and Knacke and Stranski71 perceive the law to hold unequivocally, Loeb7o and Hirth and Pound72 express more guarded assessments. Ward, 35,31,65967:73 in an extensive combination of actual and computer simulated experiments, report that severe deviations from the cosine law are obtained in Knudsen experiments and they conclude that the effusate distribution reflects cell and sample geometry rather than kinetically diffuse evaporation and reflection (re-evaporation). h. Equilibrium conditions must exist in the cell if equation (III-l4) is to yield the equilibrium vapor pressure; at least, equilibrium conditions must be approximated within experimental error. This statement implies not only that the rates of evaporation and condensation are equal, but also that the pressure must be the equilibrium vapor pressure. Carlson, g5 £1.7u and Ward and Fraser75 have tabulated data which indicate that the measured vapor pressure approximates the equilibrium vapor pressure, within experimental error, for orifice areas (so) less than one per cent of the cell cross-sectional area (ac). Accordingly, one expects to find no discernible dependence of the measured vapor pressure (Pm) on the orifice area, and this observation has been substantiated by many vaporization studies, even where (ac/ac) > 0.01. 21 There are numerous systems, however, for which Pm has been found to vary drastically with the size of the orifice. This effect is typically traced to a non-unity vaporization (or Langmuir) coefficient, (av), 0.§_av.§.l, which mirrors the extent to which kinetic barriers inhibit free evaporation. The theoretical develOpment and treatment of vaporization coefficients may punfully be said to exist in a state of flux. While there have been numerous attempts to formalize this concept,72’7““76"87 neither theory nor experiment has yet provided a satisfactory description of the behavior of av. Much of the confusion evident in the literature stems from the improper, although not infrequent, use of the terms vaporization coefficient (av), condensation coefficient (ac), and accommodation coefficient (ca). These terms have distinct meanings, yet are often used interchangeably. The terms are defined as follows:69 experimentally observed evaporation rate ; rate at which the saturated vapor impinges on sample surface cc 8 the fraction of impinging vapor molecules which condense; and where E1 is the energy of the incident gas E1. is the energy of the reflected gas E8 is the energy with which fully accommodated molecules escape from the surface T1 is the r.m.s. temperature of incident gas T is the r.m.s. temperature of reflected gas is the substrate temperature. 22 5. Qualitative Aspects of the Vaporization Coefficient Very little data are available on the behavior of “v and ac. Accordingly, the customary practice is to assume, 3 priori, that “v - ac, presumably on the assumption that both of the coefficients are defined with respect to equilibrium conditions. From the definitions cited above, it is clear that “v - ac only at equilibrium; even if the less preferred but frequently used definition of ac is employed (i.e., “c is the ratio of the rate at which gas molecules condense to the rate at which the saturated vapor impinges on the sample surface), there is no reason to assume the equality of the two coefficients for the general case. Historically, a sub-unity vaporization coefficient is associated with an observed "orifice effect"; i.e., the smaller the orifice, the greater the measured vapor pressure. Based on the premise that steady state pressures result when av a dc, a common mental construct is that an observed "orifice effect" indicates that the cell is losing vapor molecules at a rate comparable to the evaporation rate. The term, av, need not equal ac, however, and two observations are noteworthy: an observed "orifice effect" is not a necessary manifestation of a sub-unity vaporization coefficient (see III. D.5..egueeg.); the magnitude of the observed "orifice effect" provides no information on the absolute value of av, but only on the relative magnitudes of av and ac. Nevertheless, the experimental difficulties encountered in attempts to isolate the two coefficients make other approaches impractical. Various factors which contribute to low values of av have been cited in the literature: 23 l) Grossly dissimilar structures of the condensed and vapor phase species; Brewer and Kane88 specifically cite the cases of phosphorus and arsenic, and they conclude that, as of 1955, the only clear demonstrations of sub-unity vaporization coefficients are for those systems in which the rigidity of the atoms or molecules of the condensed phase preclude free rearrangement to form the vapor species. For some systems, this effect may be largely negated by experimentally loosening up the lattice. 2) Surface depletion of the condensed phase; Surface depletion may result in a diffusion-controlled process for incongruent vaporizations. Peavler and Searcy82 cite convincing experimental evidence for this effect in a vaporization study of the M03Ge system. The surface evaporation rate is greater than the rate at which the "buried" vaporizing material can diffuse to the surface. The contention of Ackermann, e5 31.69 that most of the evaporating molecules escape from the outermost atomic layer of the condensed phase lends further credence to the argument. 3) Temperature gradient at the evaporation surface; Ackermann, e£.el.sg:73 focus their discussion of a on the interface v between the vapor and condensed phases, and they point out that a temperature discontinuity in this region is not only plausible but can also be expected to yield a sub-unity vaporization coefficient. The precise "surface" boundary at which the discontinuity occurs is related to the vaporization mechanism. This factor is further dramatized from an independent perspective. In target collection experiments, where the collection angle is frequently < 5°, the effusing beam comes entirely from the sample 24 surface.66 .It is this sample surface directly below the orifice, however, which suffers most directly from a lack of reflected radiation from the cell lid (i.e., the orifice pseudosurface). This argument, coupled with the observation that surface temperatures are exceedingly difficult to measure,69 gives added significance to the potential effects of such gradients. Littlewood and Rideal89 report that "self-cooling" as a result of a large net flux of vaporizing material also leads to a lower surface temperature and a corresponding value of av < l. 4) Surface contamination, surface strain; Any surface constraints which introduce entropies of activation can give rise to sub-unity vaporization coefficients. Such constraints may take any of several forms: a thin surface layer impervious to the buried vaporizing species, which yields a diffusion-controlled slow step; a condensed product which forms a solid solution with the vaporizing species and lowers its activity; a surface contaminant which exerts unusually strong binding forces on impinging vapor molecules, thus functioning as a sink for the vapor phase, increasing (1 c’ decreasing Pm, and giving rise to an apparently low value of av. 5) Supersaturation in the condensed state. Evaporation from one phase may occur more rapidly than nucleation of the succeeding phase in incongruently vaporizing systems. Super- saturation accompanied by a reduced vapor pressure occurs, followed by a rapid increase in volatility at the onset of nucleation. To be sure, several different kinetic barriers may contribute to the overall av; accordingly, the vaporization coefficient may properly be expressed as the product of several more specific coefficients: n “v - n av . Similar arguments would hold for ac. i=1 1 25 Finally, since the individual “V1 are indicative of kinetic barriers to free evaporation (with associated enthalpies and entropies of activation), it is reasonable to assume that the value of av will vary with temperature; further, Rosenblatt77 remarks that both theory and experiment indicate that neither av nor ac is independent of pressure. In practice, these coefficients are almost always considered invariant with respect to temperature and pressure. 6. The Quantitative Determination of the Veporization Coefficient By the mid-1950's equations had been proposed by Whitman,83 Motzfeldt,76 Rossman and Yarwood,85 Speiser and Johnston,8“ and Peavler and Searcy82 for the correction of measured vapor pressures (Pm) to equilibrium vapor pressures (Pe). Of these, the formulations of Whitman and Motzfeldt were the most rigorous and the resultant expressions were very similar. Because of its simpler form, Motzfeldt's formulation has been the more widely used. Motzfeldt derived his equation from a consideration of the differences between the "upward" pressure (toward the orifice) and "downward" pressure (away from the orifice) at each of two planes, one just above the sample surface and one just below the cell lid. The resultant expression is pm . pe - (i: + i: - 2)(me) , (III-15) where f is the ratio of the orifice area to the sample surface area, soles, and W8 is the transmission probability factor for the cell. For many Knudsen cells, including those used in this thesis research, Wa N 0.5; this number, plus the inclusion of the orifice transmission factor Wb, yields 26 Pm - Pe - (Wmef/ov) . (III-16) A plot of Pmyg'beP.m gives a straight line with slope —l/av and intercept Pe' Tacit assumptions included in this derivation are: (1) av - ac, and each is independent of temperature and pressure; (2) there is no radial pressure gradient in the respective theoretical planes; and (3) f is constant for a given experiment. As has already been noted, the first of these assumptions is not generally true; Paula and Margrave87 have separated av and “c and have reworked the derivation. Assumption (2) is definitely incorrect (the orifice introduces a marked radial pressure dependence) and Carlson, e5 21.7“ have derived a more exact formulation. Assumption (3) may_be satisfactory for vaporizing liquids and some solids; for porous solids, however, the evaporating surface cannot be determined precisely. Further, for those solids which experience a significant volume (surface) change (i.e., sintering) during an experiment, f is not constant. Rosenblatt77 treats this topic and derives an expression similar to Mbtzfeldt's but which incorporates an "effective surface area". Noteworthy in his treatment is the marked dependence of av on a Additionally, Balson86 derived c' a more sophisticated expression than did Motzfeldt, but his treatment is conceptually more difficult and practically more cumbersome than Motzfeldt's; like Motzfeldt, he did not consider the radial pressure dependence in the cell. As has been evidenced for several of the theoretical facets of Knudsen effusion, there is not uniform agreement on the utility of (III-16). Gross evaluations of the expression range from Miuradov's80 "clearly erroneous" to Hollahan and Gregory's90 "the validity of such 27 an extrapolative procedure is questionable, but a better model has not been found as yet", to Ward and Fraser's75 contention that, despite the inconsistencies, equation (III-l6) gives Pm values correct to within 12 for av's on the order of 0.01, and Knudsen cells with ilD - 2 (l is the cell height, D the cell diameter). To be sure, (III—l6) incorporates simplifying assumptions and some inconsistencies; but whatever the accuracy, the Motzfeldt equation is the least cumbersome of the more rigorous formulations and it has, accordingly, enjoyed the most widespread use.9°‘9” It should be noted that values of “v so obtained have been employed to calculate values of AH* and AS* (the enthalpy and entropy of activation),95a96 as well as values of the root mean square of the surface diffusion distance.97 7. General Considerations To date, no procedure exists for the precise determination of “v by a series of vaporization experiments with various orifice sizes. Paule and M'argrave87 suggest a Knudsen cell to be employed in such studies, but ultimately their procedure is subject to the fundamental limitation: the vaporization coefficient, like the condensation coefficient, is an exclusive prOperty of non-equilibrium systems, and ". . . it is fruitless to attempt to understand the vaporization coefficient in terms of an equilibrium theory."69 Indeed, under equilibrium conditions, where “v equals ac, both coefficients lose their intrinsic meaning. In this context, the most enlightened approach to date is that of Ackermann, Thorn, and Winslow“,78 who conclude from non-equilibrium thermodynamic considerations that the most important factors contributing to a are a temperature discontinuity V at the sample surface and/or surface strain, depending on the rate determining step of the evaporation mechanism. 28 E. Error Treatment The uncertainties associated with various uncertainty expressions have been detailed by Eisenhart.98 The procedure employed in this thesis parallels that used by'llaschke:"3 if the sum of X1 (1.: 1.5 n) variables is Y, then u 2 1/2 ‘ - Cy (2 01 ) ’ (III 17) where 01 is the standard deviation (or, when the standard deviation is not available, a subjective error estimate) for the corresponding variable Xi. CHAPTER IV EXPERIMENTAL EQUIPMENT AND PROCEDURES A. Eguipment 1. Target Collection The target collection apparatus used in the vaporization experiments is essentially the same as that described by Kent99 and Ackermann, e£_el,5° as modified by Haschke."3 There are three differences: (1) a second ground glass joint has been introduced in the vertical effusion chamber about 4 cm below the bottom of the Cu target holder, the joint itself being cooled by an extention of the water jacket which surrounds the lower section; (2) the middle section of the 3-part effusion chamber has been made from Pyrex; and (3) two water-cooled bottom sections were employed, one fashioned from quartz and the other from Pyrex. Symmetrical cylindrical effusion cells, with a knife-edged orifice at each end opening into identical cavities, were heated by a push-pull type 20-kva Thermonic high frequency induction generator. Temperature measurements were made with Leeds and Northrup disappearing filament optical pyrometers (serial numbers 1572579 and 1524388) which had been calibrated previously by the National Bureau of Standards. Sample cavity-to-target distances were measured with a precision cathetometer (Gaertner Scientific Co.) and orifice areas were determined by circumscribing greatly enlarged photomicrographs (6400K, lOOOOX or 40000X in square units) of the orifice with a compensating polar planimeter. The photomicrographs were taken with a Bausch and Lomb DynaZoom Metallograph. 29 30 2. X-Rey Fluorescence A four-position Norelco Universal Spectrograph was used to analyze the condensed effusate from the target collection experiments. The analyzing crystal was graphite and the scintillation detector employed a thallium-activated NaI crystal. Incident radiation was supplied by a broad focus tungsten X-ray tube in conjunction with a Norelco XRG-SOOO generator. 3. Mass Spectrometric A Bendix Model 12-107 time-of-flight mass spectrometer was used to effect the several mass spectra. The instrument was equipped with a high temperature source region and the high temperature molecular beams effused from non-symmetric Knudsen cells with channeled orifices. Details of both the spectrometer and the 2«piece, single chamber cell design are provided elsewhere.37 4. X-Ray Powder Diffraction The hundreds of powder diffraction patterns were taken by a Haegg type Guinier camera, Cu Kal radiation, Isl - 1.54051 A, T - 24 1 1°. The fine focus X-ray tube was powered by a Picker 809B generator. 5.221195. In addition to the induction heater cited above, a single tube ZO-kva Thermonic high frequency induction generator was used as a general purpose heater for temperatures above 750°. Other furnaces used include a Marshall Products Co. Pt-4OZ Rh wound tube furnace, routine horizontal tube furnaces, and a make-shift two-zone furnace effected by connecting two of the horizontal tube furnaces end-to-end. High temperature heating in air was routinely carried out in a muffle furnace (Thermolyne Corp., Dubuque, Iowa). 31 In addition to the disappearing filament optical pyrometers cited above, temperature measurements were made by use of a chromel- alumel thermocouple and assorted general purpose low temperature thermometers. ‘ An Ionalyzer Specific Ion Meter (Orion Research, Inc., Cambridge, Mass.) was employed for potentiometric titrations. B. Materials Sesquioxide powders used included those of samarium, neodymium, ytterbium, and europium, each obtained 99.9% pure from Michigan Chemical Co., St. Louis, Mich. Other preparative reagents were: thulium sesquioxide (99.9%, Research Chemical Co., Phoenix, Ariz.), elemental bromine (99.82, J. T. Baker Chemical Co., Phillipsburg, N.J.), hydrobromic acid (47.92, Fisher Scientific Co., Fair Lawn, N.J.), and oxalic acid dihydrate (99.8%, J. T. Baker Chemical Co., Phillipsburg, N.J.). Sample containers were fashioned from molybdenum (Kulite Tungsten Corp., Ridgefield, N.J.), graphite (Becker Brothers Carbon Co., Cicero, IlL), platinum (J. Bishop and Co. Platinum Works, Malvern, Pa.), thermal recrystallized alumina (Thermal Syndicate LTD, England), quartz (boats from Thermal American Fused Quartz Co., Montville, N.J.; tubing from Engelhard Industries, Inc., Hillside, N.J.), and thoria (Zirconium Corporation of America, Solon, Ohio). The copper targets used in the neodymium oxide halide studies were fashioned from stock copper obtained from McMaster-Carr Supply Co., Chicago, Ill. The platinum powder employed as an internal standard in the powder diffraction samples was purchased from J. Bishop and Co. Platinum Works, Malvern, Pa. 32 C. Preparative 1. Monoxide Monobromides of Needymium, Samarium, and Europium Neodymium monoxide monobromide was prepared most conveniently by placing a preheated quartz or alumina boat containing annealed sesquioxide in a stream of bromine at 800-lOOO° for 3-5 hours. Helium, dried over liquid nitrogen and bubbled through the bromine reservoir, was used as a carrier gas and also as a purging agent before and after the bromine flow. Samples of SmOBr and EuOBr were prepared analogously; in no case did the temperature nor the time seem particularly critical. Poorly annealed NdOBr was also prepared from the tribrmmide hydrate by ignition at 650° (in air) for 50 minutes. Similarly, SmOBr was prepared from the hexahydrate by heating in air at 375° for 5 hours. 2. Monoxide Monobromides of Thulium and Ytterbium The most convenient preparation of YbOBr involved dissolution of the sesquioxide in concentrated hydrobromic acid, evaporation to dryness, and further drying by sequentially heating in air to 100°, crushing, and drying over P205 in a vacuum dessicator. After one or two repetitions of the drying cycle, the pink-orange hydrate was situated in a preheated tube furnace at 450° under a stream of HZO-saturated He and heated for 40 minutes. The residue was quickly transferred to a quartz ampoule which was evacuated, sealed, and heated to 700° for a few hours. Well annealed YbOBr was obtained. Similarly, TmOBr was prepared by heating the dried, crushed hydrate under a stream of HZO-saturated He for 45 minutes at 450°, followed by annealing in a flamed, evacuated quartz ampoule at 650-750°. Tetragonal YbOBr was prepared alternatively by heating the dried bromide hydrate under a Brz-saturated He stream for 12 hours at 220°. 33 The poorly annealed, very lumpy residue was crushed and mixed with 1 or 2 ml of elemental bromine in a flamed quartz test tube. The tube was evacuated, sealed, and placed in the two-zone furnace with the solid residue decidedly elevated in the hot zone (900°). CAUTION. Heating was continued for 2-3 hours before removal to the dry box. This latter preparative procedure is tedious, but is believed to yield a YbOBr sample of higher purity. Direct bromination of the oxalate at 450° for 1 hour under dried He yielded a chocolate brown residue which, when annealed in an evacuated quartz enclosure at 775°, was found to be the monoxide monobromide. None of the LnOBr species was vigorously hygroscopic. Nevertheless, they were exposed to air as little as possible, with sample storage and handling effected in a dry box. 3. Tetraoxide Monobromides of Neodymium and Samarium Equimolar amounts of sesquioxide and monoxide monobromide were heated in an evacuated quartz ampoule for several hours at lOOO°. In one experiment, the SmOBr-Sm203 mixture was situated in a platinum crucible inside the quartz ampoule. Both Nd3043r and Sm3O4Br were also prepared by carefully vaporizing the LnOBr species ipuyegug in a molybdenum or graphite effusion cell at 1200-1250°. With experience the reaction can be reasonably controlled, but special care must be taken to ensure that: (l) the heating be sufficiently prolonged to decompose all of the LnOBr species; and (2) the heating not be so prolonged that the Ln3043r species begins to decompose. The difficulty rests with the fact that at l300° there is less than an order of magnitude difference between the equilibrium vapor pressures over the two decomposition processes; at higher temperatures, the difference is smaller. 34 All samples used in the Sm3O4Br vaporization studies were prepared by the former method. 4. Tetraoxide Monobromides of Thulium and Ytterbium Attempts were made to prepare each of these compounds as a pure phase by the careful vaporization of the respective LnOBr phases. In addition, prolonged heating of an equimolar mixture of TmOBr and Tm203 at 900-1100° in an evacuated quartz ampoule was effected in an effort to prepare pure Tm3O4Br. $191.19.]: Although no specific attempt was made to prepare additional YbOxBr3_2x phases, the results of several attempts to prepare YbOBr suggested that such phases, more bromine rich than YbOBr, do exist. D. Analytical 1. Lanthanide Metal All of the bulk compounds employed in these investigations were analyzed for metal content by conversion to the sesquioxide. Samples, typically weighed in the dry box, were ignited to constant weight at W 900° in quartz, alumina, or platinum boats which had been preheated to constant weight. In some instances, the samples were treated with a few drops of concentrated nitric acid just prior to ignition. Special attention was taken to weigh the resultant sesquioxide within 5-7 minutes after removal from the muffle furnace in order to minimize hydrolysis (and associated) problems. 2. Bromine Four methods were employed to effect bromine analyses. Bulk preparations of NdOBr and SmOBr were analyzed most conveniently by noting the weight gain upon bromination. The favorable gravimetric 35 factors involved (1.427 for NdOBr, 1.412 for SmOBr) yielded accurate determinations. One analysis was made by the standard gravimetric procedure for bromide determination, with a standardized AgNO3 solution used as the precipitating agent. In a few cases, bromide samples were dissolved in dilute nitric acid and analyzed by the Volhard method.100 Standard NaCl samples were titrated as a procedural check. When only very small (< 5 mg) samples of bromide were available, the analyses were effected by a potentiometric titration technique which employed a standard calomel electrode (saturated KCl solution) as the anode and a silver cathode. The half cells were connected by an NH4NO3 (agar-agar matrix) salt bridge. Each sample was dissolved in a stock solution of NH4NO3 and dilute HNO3 and titrated with a standard AgNO3 solution (0.00735 M). The apparatus, including solvent, was standardized with bromide solutions of known concentration by another worker101 who reported the endpoint to occur at +191 my (the reported value is 200 my = 10-12.l).102 for KSP ,AgBr E. X-Ray Diffraction The most frequently used analytical tool in this study was that of X-ray powder diffraction. All sample preparations were analyzed by X-ray diffraction; the solid residues from vaporization experiments were examined by powder diffraction to ascertain the phases present and to check for interfacial contamination; solid phase decomposition modes were monitored by a series of X-ray films. Among more isolated uses, the patterns were employed to identify a compound and to determine the purity of a phase, the number of phases present, the extent of crucible 36 interaction, the physical distribution of a phase or phases, and for the determination of lattice parameters. Sample preparation for diffraction patterns has been reported 5’103 for those substances reasonably inert to air, which . elsewhere includes all of the oxide bromide phases examined. Particularly hygroscopic samples, most notably the lanthanide bromides, were sealed in small plastic envelopes in the dry box and mounted as usual. In every case, platinum powder (a - 3.92373: 0.0003 A)1°“ was included as an internal standard; patterns were obtained while the sample chamber was under a rough vacuum (lO'l-lO"2 torr). F. X-Ray Fluorescence Condensed effusates from the target collection experiments, typically containing 4-10 ug of lanthanide metal, were analyzed by an X—ray fluorescence procedure. For each element analyzed, three steps were required: (1) optimization of the spectrometer sensitivity for that element; (2) calibration of the spectrometer; and (3) the actual analysis. 1. Optimization of Spectrometer Sensitivity The optimization procedure paralleled that outlined by Neff105 with approximately 8 ug of the element (M) deposited on a target (T). The ratio (IMHT)1/2/(rM+T - rT) was minimized as a function of the pulse height discriminator settings for a selected spectral peak with a specified target material (r has the units of counts/time). For each optimization procedure, the process was executed for Al and Cu targets and for the several spectral peaks of the element. In order to identify troublesome peaks in the background spectrum, three important considerations were: (1) use of a sample size comparable to that of 37 the samples to be analyzed; (2) deposition of the element from a solution which does not contribute to the background spectrum; and (3) examination of the metal (bromine) spectral peaks for possible interference from the bromine (metal) which co-deposits on the targets in the Knudsen effusion experiments. One further consideration was to avoid those spectral peaks which shouldered even larger background peaks; although these spectral peaks may give rise to suitable counting rates, they do not permit satisfactory reproducibility due to slight 26 instrument drifts. A summary of target materials, spectral peaks and discriminator settings used are given in Table 1. For the analysis of Br on Al targets, suitable settings were employed but the outlined optimization procedure was not followed. Table l. Spectrometer Settings Employed for X-Ray Fluorescence Analyses Discriminator Target Settings: Spectral Element Material .E E“ Peak Nd Cu 1.5 3.5 Lal Sm A1 2.0 4.0 Lol Br Cu 4.5 3.9 Kal, Ka2 Br Al . 7.0 2.0 Kol, Kaz 2. Calibration of the Spectrometer The determination of calibration curves for the elements analyzed was a painstaking endeavor and not a particularly reproducible one. Once the spectral peak to be monitored was selected, the precise 26 value corresponding to this peak was re-determined at each counting session by scanning this part of the spectrum by using a bulk lanthanide 38 or bromide sample. With the scanning arm of the spectrograph locked on this peak, cleaned, dried blank targets were analyzed (counted) to determine the background count rates. In all cases, both blank and sample targets were counted in target holder position No. l, with each target rotated slowly during the analysis. The aluminum targets were cleaned by an abrasive treatment with fine emery paper followed by a thorough rinsing with distilled water; copper targets were either similarly polished or cleaned with steel wool, bathed in concentrated HCl, and rinsed thoroughly. Standard solutions of Nd and Sm were prepared by the dissolution of precisely weighed samples of the sesquioxide in 25 ml of distilled water to which 3-5 drops of concentrated HCl were added. The solutions were gently heated in 250 ml volumetric flasks, cooled, and diluted to the mark. For a typical determination, triplicate standard solutions were prepared with 600, 1200, and 1800 ug Ln(Br)/ml. In the case of the Sm standard solutions, it was critical to vaporize the excess HCl in order to prevent subsequent reaction with the Al targets. A similar procedure was followed for the preparation of standard Br solutions, with reagent grade NaBr03 used as the solute. Known amounts of the standard solutions were placed on the precounted targets by a weighing technique; typically, 10-12 targets were employed for a single calibration with each standard solution placed on 3 or 4 targets. After considerable experimentation, it was determined that the calibration results were a function of the size, number, and distribution of the drops placed on the several targets. (Ideally there would be a uniform distribution of sample on the target surface.) Accordingly, the standard solutions were transferred to the pre-weighed (to 0.00001 g) targets by a glass capillary which, with practice, 39 allowed the routine placement of 20-30 droplets on the 2 cm: target surface in less than 10 seconds. Care was taken to effect a symmetrical distribution. The targets were immediately weighed and subsequently reweighed at 20-25 second intervals. The weight 32 time data were extrapolated to t - O to determine the weight of standard solution used, where t0 is the mid-time between the placement of the first and last droplets on the target. The weighing process itself, with the slight drift of the balance zero, appeared to be the greatest single source of error in the analyses. Typical weights of standard solutions deposited per target were 5-9 mg. The dried sample targets were recounted in the fluorescence spectrometer. For every calibration, standard blank and/or sample targets were counted in order to permit use of the calibration data over the period of time in which the X-ray generator experienced power drifts and in which the detector crystal hydrolyzed. Typical counts/ug results for Nd, Sm, Br (on Cu), and Br (on A1) calibrations were 1900, 2650, 1500, and 1050, respectively; the results were considered accurate to i 102. 3. The Analysis Targets from the target collection experiments were removed from the effusion chamber and generally analyzed within 2-5 hours to minimize the effects of sample decay. In all target analyses (blank, calibration, and effusate collection), the countable target area was defined precisely by a circular insert, fashioned from the same metal as the target. G. Weight-Loss Experiments In order to confirm a vaporization mode and/or to determine the extent of sample-crucible interaction, net weight-loss experiments were 4O executed in which the LnOBr species was heated lpngeeug (10-5-10-6 torr) until decomposition to the oxide was complete (constant weight). The resultant weight-loss was compared to that calculated on the assumption that 3 moles of LnOBr yield 1 mole of sesquioxide. Sample/cell combinations used were NdOBr/Mo; SmOBr/Mo, ThOZ, 8102, graphite; YbOBr/Mo; and TmOBr/Mo. The cells employed were of the same design and fabrication as those used in the target collection experiments. H. Decomposition Traces The solid phase high temperature decomposition processes were monitored by a series of powder diffraction patterns taken as the decomposition progressed. Typically, 6-8 patterns were taken after consecutive beatings of the sample at successively higher temperatures (increments of 50-100°) in a Knudsen cell. Careful readings of these patterns, with attention given both to absolute sinze values and to sequential relative intensities, provided detailed information on the solid state decomposition mode. Traces were made for NdOBr (in a graphite cell) and for NdOBr, SmOBr, TmOBr, and YbOBr (in Nb cells). In addition, the decomposition of YbOBr (in Mo) was traced over a much narrower temperature range (100° total), as was that of SmOBr (in platinum). In all of the above-mentioned traces, heating was effected in a dynamic vacuum (10"5-10'6 torr). The thermal decomposition of SmOBr in air was similarly monitored as a check on Zolotov and Mayer's earlier report.9 It should be noted that, for most of the above traces, temperature measurements were effected with an optical pyrometer which focussed into the cell through a window and glass prism for which no absorption corrections were made; accordingly, the recorded temperatures were approximately 35-40° too low. 41 1. Mass Spectromeepy The gaseous products from the several oxide bromide decompositions were examined by a mass spectrometric technique in which the single cavity effusion cells were heated by radiation and electron bombardment. Ionizing energies of 35 eV were employed to study the fragmentation patterns of the effusates from LnOBr (Ln - Nd, Sm, Tm, Yb) samples confined in Mo crucibles. Additionally, fragmentation patterns of the effusate over NdOBr were obtained with 20 and 50 eV ionizing energies. The effusate over SmOBr in a graphite cell was also examined. In each case, the Br peaks observed were largely residual and were not meaningful for comparison purposes. Appearance potentials were determined by a linear extrapolation technique with Hg and H20 vapors used as references. Potentials were determined for the species: NdBr2+, Nd+, SmBr+, TmBr3+, TmBr I, rm+, YbBr2+, YbBr+, and Yb+. Background spectra were taken at several temperatures as each sample was heated, and for the NdOBr system (and for the YbBr3+ peaks of the YbOBr system), spectra were also recorded after decomposition to the oxide was complete. The shutterability of the several m/e peaks was generally not determined, and peak height assignments were made by a comparison of the particular spectral peaks with the background spectra. Temperature measurements were effected with an optical pyrometer which sighted the hot cavity through a window and prism for which no absorption corrections were made. 42 J. Bulk Effusate Collections Several attempts were made to collect bulk samples of the anhydrous condensed effusates over the LnOBr species. The use of an inverted quartz cup positioned over an effusion cell ipuyeeug_was only marginally successful. In the case of SmOBr, the effusate was collected in a 10"6 torr vacuum by use of the apparatus sketched in Figure 1. After prolonged sample heating at 1200°, the furnace was lowered and the narrow quartz tubing which contained the condensed effusate was sealed at both ends and removed to the dry box. The secondary trap was checked for elemental bromine. The most efficient apparatus used for collecting the condensed effusates as the anhydrous bromides is outlined in Figure 2. Only the lower 1/3-1/2 of the Mo cell was positioned in the hot zone. Condensed samples for powder diffraction analyses were scraped from the underside of the Mo lid which was several hundreds of degrees cooler than the cell base. This type of apparatus was required since, at a given temperature, the equilibrium bromide pressure above the condensed bromide is considerably greater than that above the corresponding LnOBr-Ln304Br phases. Effusates over NdOBr, SmOBr, EuOBr, and YbOBr were collected and analyzed by X-ray diffraction. The effusates over SmOBr and YbOBr were also analyzed for Br content. K. Vaporization (Tagget Collection) Experiments The salient features of the vaporization technique employed in this laboratory have been treated amply by others,“3999’106 but several details merit special attention here. 43 H. sfl- DIFFUSION <—_—_—_, i PUMP V COLD TRAPS 3 mm 0.0. QUARTZ ' SmOBr POWDER FURNACE Figure 1. Apparatus Employed to Collect the Condensed Bromide Vapor from the Decomposition of SmOBr(s) 44 ORIFICES Mo 4313/ AI203 .J .. 1/ Mo H. L. / C , SAMPLE I I I CAVITY (actual size) Figure 2. Effusate Collection Apparatus 45 1. Samples Typically, samples used in the LnOBr vaporization studies consisted of well mixed charges of 300-400 mg LnOBr and 25-50 mg Ln3043r. In a few instances, the residue from a previous vaporization experiment, consisting of LnOBr and Ln304Br, was mixed with a fresh charge of LnOBr. For Sm304Br vaporization experiments the sample consisted of 300-400 mg of Sm3O4Br and 25-50 mg of the annealed sesquioxide. For those experiments in which the cell orifice area was < 10'3 cm2, sample sizes I were increased to approximately 500 mg of the primary vaporizing phase. In all cases, the crucibles were loaded in the dry box. Following each experiment the solid residue was examined critically by X-ray diffraction in order to ascertain the phases present; frequently powder diffraction samples were extracted from two or more local sites of the highly sintered residue. 2. Crucibles In the earlier stages of these investigations, thoria-lined graphite crucibles were used for the NdOBr study and graphite crucibles were employed for the Sm304Br work, and both the above noted crucible types, in addition to a quartz-lined graphite crucible, were used for the SmOBr system. For those experiments which generated the actual thermo- dynamic data used as a basis for further calculations, Mo crucibles were used exclusively for all three decomposition systems. The 3-piece crucible assemblies have been schematically represented by Haschke“3 (see Figure 3) and were described earlier in this document. Only crucibles and liners which had been preheated $3 33222 were employed. The single greatest experimental difficulty was in the maintenance of a uniform temperature distribution over the entire length of the “P-I- Q \\,\‘\x Q/ CHAMBER p R I, /< R / S SIDE 3&3? 23,6 I _ T‘— .. % VIEW b i \ ‘ * \ .J" \ ' ‘ \ OPTICAL ‘ I CHAMBER L l'62cm I I' 'I I 2°OOcm 1 ”0"” OVEN TOP OF Mo /KNUDSEN CELL . T OP 6/ VI E w 2-54cm I H O-47cm Figure 3. Knudsen Cell/Oven Configuration Employed in Target Collection Experiments 47 crucible. This consideration was important not only because of the thermodynamic implications of a temperature gradient in the sample chamber, but also because the recorded temperature was read as black body radiation from an orifice in the cell base whereas the effusate escaped from the cell lid. Accordingly, after singularly unsuccessful attempts to remove the gradient by adjusting the induction Coil, a cylindrical 2-piece oven, fabricated from a W—Mo alloy, was used. The sample cell was positioned inside the oven as depicted in Figure 3, and crucible heating was effected by radiative emission from the inductively- heated oven. The oven top had a centered circular hole, d - 0.47 cm, and thus provided no resistance to the effusion beam from a reasonably well aligned crucible. The suitability of the crucible-oven configuration for vaporization studies was confirmed by another researcher107 who used the apparatus to determine the vapor pressure of silver as a function of temperature. Cell orifice areas were measured both before and after each target collection experiment. No difficulty was experienced from orifice clogging due to condensed effusate, but in a few cases (particularly when extremely small orifices were employed) the crucibles had to be outgassed thoroughly in order to prevent serious orifice area diminution due to the migration of a slight surface oxide coating. The temperature range corresponding to each orifice area is listed in Table 2 for those experiments in which the oven/Mo cell combination was used. 3. Targets The majority of the experiments were conducted with 12-15 cleaned, precounted targets stacked in the target magazine. Since inverted dummy targets were used only as the bottommost and topmost ones in the stack 48 Table 2. Orifice Areas and Associated Temperature Ranges Employed in the Target Collection Experiments Orifice Area Vaporizing 4 2 Temperature Phase Run X 10 cm Range (°K) NdOBr 8 10.2 1598 - 1725 NdOBr 9 69.7 1467 - 1592 NdOBr 11 43.6 1544 — 1678 NdOBr 12 7.8 1635 - 1709 NdOBr 14 104.3 1472 - 1606 NdOBr 15 11.3 1609 - 1731 NdOBr ‘ 16 70.6 1532 - 1617 NdOBr 21 2.45 1683 - 1850 NdOBr 22 2.45 1683 - 1793 SmOBr 1 69.2 1416 - 1542 SmOBr 2 10.8 1481 - 1628 SmOBr 3 40.5 1428 - 1568 SmOBr 4A 40.5 1445 - 1561 SmOBr 4B 43.1 1438 - 1481 SmOBr 5 0.95 1611 - 1768 SmOBr 6 0.95 1547 - 1582 SmOBr 7 0.95 1603 - 1726 Sm304Br 1 10.0 1557 - 1672 SmSOABr 2 10.0 1525 - 1642 SmBOaBr 3 0.8 1636 - 1775 Sm 0 Br 4 0.8 1659 - 1739 3 4 49 (i.e., no spacers were used between exposed targets), careful alignment of the magnetically-operated shutter was required. Target-to-crucible distances were measured with the precision cathetometer. The lowest surface of the targets was measured after the system had been subjected to high vacuum (10'5-10"6 torr) for several minutes with liquid nitrogen in the target magazine dewar; an allowance of 0.193 cm was made for the elevated collection surface of the targets. Cathetometer readings of the crucible top were made after the collection experiment upon disassembly of the effusion chamber. As mentioned previously, analyzed exposure areas were precisely defined by a circular metal insert fitted to the targets. A single Al insert was employed for all SmOBr and Sm304Br experiments; a Cu insert was used for the NdOBr work. In each case the exposed area was taken as 1.998 cmz. Collection times ranged from 6 to 120 minutes for the NdOBr system, 5 to 112 minutes for the SmOBr system, and 5 to 94 minutes in the case of Sm3043r. 4. Temperature Measurement Temperatures were determined with NBS-calibrated optical pyrometers which monitored the black-body radiation emanating from the optical cavity. Correction was also made for intensity absorption due to the optical elements used. The latter correction was made by a comparison of two readings of a lamp, one with the lamp sighted directly and one with it viewed via the optical elements used in the Knudsen experiments. In addition to the use of the oven (see XI.B.), attempts were made to control temperature gradients by use of a second Optical pyrometer. Although the measurements from this second pyrometer were made by sighting between turns of the induction coil and through the water 50 jacket (rendering absolute readings meaningless), relative temperatures of the upper and lower oven walls were measured accurately and gradients too small to be definitively identified visually were detected. This technique was used to indicate needed adjustments in the coil position; a gradient of < 5° over the length of the oven wall was considered insignificant. It was determined that a coil position which yielded no detectable gradient at one power setting could produce a substantial gradient at another. Accordingly, monitoring of the temperature was required for each target collected, and this was done for the last several Knudsen runs. It was determined empirically that a 10° gradient on the oven wall produced a 1 or 2° gradient on the Mo cell, with a 15° oven gradient resulting in a 2 or 3° cell gradient. When applicable, corrections were made. 5. Miscellaneous Details To make a specific check on the relationship between the orifice area and the measured vapor pressure over a NdOBr-Nd3O4Br mixture, two vaporization experiments, in which particular care was given to reproducing experimental conditions except for the orifice area, were run within a 4 day period. The two orifice areas used were 11.3 x 10"4 and 70.6 x 1o“I cm2. For those vaporization experiments in which temperatures above 1600 K were attained, it was necessary to seal the ground glass joint closest to the hot zone with Apiezon Wax W rather than the Apiezon Grease T employed at lower temperatures. All target collection experiments were performed in a dynamic vacuum (10"5-10'6 torr). 51 L. Sticking Coefficient Determination Haschkem’108 reported the sticking coefficient of gaseous EuBrZ on chilled Cu targets to be nominally unity. An experiment patterned closely after that of Haschke, but which employed an A1 disc, was performed to determine the sticking coefficient of the bromide over SmOBr on chilled aluminum targets. No analogous determination was made for the sticking coefficient of the gaseous bromide over NdOBr on Cu targets. CHAPTER V RESULTS A. Preparative l. Monoxide Monobromides of Neodymium, Samarium, and Eur0pium The light blue NdOBr, cream-colored SmOBr, and cream-white EuOBr were prepared in essentially pure form. When prepared by direct bromination of the sesquioxide at high temperatures, the LnOBr products were very fine powders which gave sharp diffraction patterns. Lower temperature preparations from the thermal decomposition of the bromide hydrates (for Nd and Sm) yielded poorly annealed powders of a coarser texture. Results of the Ln and Br analyses are given in Table 3. 2. Monoxide Monobromides of Thulium and Ytterbium All attempts to prepare YbOBr by direct bromination of the sesquioxide with elemental bromine failed. This result reflects the relative instability of the heavier LnOBr species with respect to the oxides. Reasonably pure samples of YbOBr and TmOBr were prepared by the careful decomposition of the bromide hydrates, but even by this method it was difficult to prepare sesquioxide-free samples. Only one of the YbOBr samples did not contain a slight oxide contaminant as shown by the very weak occurrence of the most intense line of Yb203 in the powder patterns; this lone sample yielded a pattern which included a doublet of very weak intensity due to Yb3O4Br. Analytical data are cited in Table 3. The resultant YbOBr and TmOBr samples were very coarse powders comprised of isoluble granules, unlike the analogous Nd, Sm, and Eu powders. The YbOBr varied in color from grey-black to white; TmOBr was white with a yellowish tinge. 52 53 ~.qu.a.me Hw.mn Assemsm wm.en eo.mm «smsomsz s.qusm.ss em.eo anon» om.mo an.mo HmoaH Ha.o+ m.am m~.Nm mm. Hm.He pmosm so.ounm.~m mo.~m m.o+ o.Ho oo.Ho Amoam s~.omx~.mm nm.mm H.o+ w.mn 00.09 Hmovz muasmom assessaas< umovz mo unaunaaa some» n cfimueoo ou sebum“ meowummflauoaoa mo .oz sssom .um N Hmowumuoonh .Am N maoauoawauouon mo .02 sass .sa N Hmoauouoona afi‘H N ommnm .m mHan 54 3. Tetraoxide Monobromides of Neodymium and Samarium Pure samples of Nd3O4Br and Sm304Br were prepared as outlined in Chapter IV. Since these preparations were effected in closed systems, chemical analyses generally were not made; rather, powder diffraction data were relied upon. Analytical results are given in Table 3. Both the dark blue Nd3O4Br and the cream-colored Sm3O4Br were very fine powders, much like the corresponding LnOBr species. Samples prepared at high temperatures in quartz ampoules interacted with the quartz, and an extremely thin, brittle sheet surrounded the sample at the sample-ampoule interface. This easily isolable contaminant was present only in minute amounts (<< 1% of sample weight) and was not considered to render the sample unsuitable for vaporization studies. At 1400° Iipnyeegg, a mixture of the contaminant, Sm304Br, and szo3 did not react when confined in a Mo cell, but the contaminant annealed as a very hard, thin red blade (see V.B.). Corresponding Ln3O4Br species, free of this slight impurity, were prepared by lining the quartz ampoule with platinum foil. 4. Tetraoxide Monobromides of Thulium and Ytterbium Neither Yb3O4Br nor Tm3O4Br was prepared in reasonably pure form in this investigation, although controlled decomposition of the respective LnOBr species in air was not attempted. Equimolar mixtures of TmOBr and Tm203, heated in an evacuated quartz ampoule at T > 900°, yielded a mixture of the reactants and TmBOaBr, even after several hours of heating. The resultant Guinier patterns showed intense lines for the oxide and the monoxide monobromide, but only weak lines for Tm3O4Br. Lattice parameters have been determined for the LnOxBr3_2x compounds prepared (including Yb304Br and Tm304Br) and these are compared in Table 4. Phase NdOBr SmOBr TmOBr YbOBr Nd3O4Br Sm304Br Tm3048r Yb304Br 55 Comparison of Derived and Published Lattice Parameters for Some LnOx Br 3- 2x Crystal System Tetragonal Tetragonal Tetragonal Tetragonal Orthorhombic Orthorhombic Orthorhombic Orthorhombic Phases Derived Lattice . Parameters (A) 12.049 0.003 11. 929 0.003 4.141 f 0.002 1. i 11.5 Published Lattice . Parameters (A) Ref. a = 4.024 2 c - 7.597 a-3..952:0001 3 c-7..9l4_-l_-_0002 a a 3. 806 2 c - 8. 288 a = 3.7847 2‘. 0.0005 3 c - 8.309 f 0.001 a a 12.223 0.003 6 b - 12.079 0.003 c - 4.179 i 0.002 1'. 1'. a .. 12.049 1 0.003 3 b - 11.928 1 0.003 c - 4.141 it 0.001 c - 4.010 2: 0.001 56 Table 4 with the corresponding literature values. Of particular note are the parameters cited for Tm304Br, since this compound has not been reported previously. B. WeiggEfLoss Experiments After prolonged heating ipnyeegg in a Mo cell, NdOBr decomposed to the sesquioxide; the weight lost in the process was 100.0% (one determination) of that calculated for the solid state process in which 3 moles of NdOBr give 1 mole of Nd 03. The Nd203 reacted with the Mo 2 crucible at temperatures higher than those employed in the target collection and weight-loss experiments. Similar experiments with SmOBr, TmOBr, and YbOBr in Me cells gave comparable results: 100.1%, 99.12, and 98.62, respectively; it should be noted that the YbOBr sample was known to contain a trace Yb203 impurity. In the case of SmOBr, an additional check was made to determine if the Mo crucible had changed weight during the decomposition process; it had not. i That these LnOBr phases did not react with Mo at their respective experimental decomposition temperatures was surprising in view of Haschke's report"3 that EuOBr reacted appreciably. Accordingly, a weight—loss experiment with EuOBr in a Mo cell was conducted, and the results confirmed Haschke's findings: the actual weight lost was .2 114.5% of theoretical. Diffraction lines were observed in the X-ray pattern of this residue which were not due to B- or C-Eu203. Elemental analysis of the residue on the fluorescence spectrometer showed that Eu and Mo were present, but not Br. In the earlier stages of the SmOBr investigation, metal crucibles were avoided because of the presumed interaction of the metal with 57 vaporous bromine and/or bromides. Consequently, a graphite crucible was employed for much of the early work. Later, however, a weight-loss experiment in graphite revealed that the weight lost during the decomposition to the oxide was 1132 of the calculated value. That this discovery was made only after extensive use of graphite cells is attributed to two factors: (1) there was no visible or X-ray evidence for interaction; and (2) the graphite cell itself did not lose appreciable weight (< 1 mg) during the experiment. Many months later, the weight-loss experiment was repeated with a different graphite cell and good agreement was found between the actual and theoretical weight—loss data. This discrepancy is not understood. In the quest for a suitable cell material, Th02 and $102 liners were tested and found to be unsatisfactory, but the weight-loss experiments of SmOBr in Th02 and SiO yielded additional interesting data. In the 2 case of the Th02 liner, the residue after the decomposition was white except for the top surface, which was pink. X-Ray analysis of the bottom surface showed pure B-Sm203, but the diffraction pattern of the top surface revealed a different phase. That this contaminant should appear on the tap surface rather than at the sample-cell interface suggests a vapor transport mechanism. Elemental analysis of this phase showed Sm to be definitely present and Br to be definitely absent; the presence of Th was questionable. No Th-Sm-O ternary phase has been reported. Similarly, interaction between the sample and the quartz liner appeared to occur by a vapor transport mechanism. After completion of the weight-loss experiment of SmOBr in this liner the outer surface of the initially clear cup was coated with a white powdery substance. The diffraction pattern of this coating closely resembled that of 58 the spurious phase found in the analogous experiment in Th02. Additionally, this phase appeared to be the same as the isolated contaminant in the Sm3OABr sample prepared in quartz (see V.A.3). The powder diffraction data approximate closely those expected for hexagonal Sm4(8104)3 (a = 9.497, c = 6.949 A).109 Lattice parameters derived from this study for the SmSixOy and SmTthy phases are a = 9.476 i 0.008, c = 6.936 3: 0.009 3., and a - 9.493 _4; 0.006, c = 6.968 t 0.007 A, respectively. The fair agreement in lattice parameters, coupled with the striking similarity in appearance, suggest that the SmSixOy phase isolated in this study is the reported Sm4(3104)3. The analogous phases of europium were studied concurrently and the resultant implications for an unusual Ln ionic environment have been reported.110 Weight-loss data are summarized in Table 5. C. Decomposition Traces The determination of the solid state decomposition modes of the several LnOBr species revealed that, in Mo crucibles, NdOBr and SmOBr decompose to the Ln3O4Br phases which further decompose to the sesquioxides (hexagonal Nd203 and monoclinic Sm203). The same results were obtained when the decomposition of NdOBr in a graphite cell was monitored. In every instance, the LnOBr phase disappeared completely from the powder pattern before any sesquioxide appeared. (Experimental checks on these results were effected each time a vaporization experiment was conducted and the residue analyzed by powder diffraction.) All of these traces were made ip.yeegg over approximately the same temperature ranges as those used in the vaporization experiments in which the larger orifices were employed. 59 meme» «0 unseen oomuu samucoo ou azoaxe mo» may 02 oz oz mo» 02 oz woman mmqu meowunam _ coauunuouau mm» mow oz 02 02 wow 02 oz HHoOIOHaamm sanamfi> w.moa m.oo~ ooH x Aoamov anon .us o.eHH m.moa mafia o.mm H.mm m.eHHA H.0oa o.OOH AmAOV smog .u3 Non\umoam Nosy\umoam o\umoam oz\sumonw oz\umose oz\umo:m oz\nmoam oz\umomz Hamo\umoaq mucoafiumaxm mmounanoz mo muasmom .m manna 60 The decomposition mode of SmOBr in air was determined as a check on an earlier report9 that SmOBr decomposes directly to the sesquioxide. A Guinier trace of the process showed that decomposition to the oxide proceeded via the Sm304Br species. However, another trace of the SmOBr decomposition, effected lgflyeggg (10-5 torr) at 675-725° (W 600° below the temperatures employed in the target collection experiments) over a period of weeks, showed the monoxide monobromide to decompose directly to the cubic sesquioxide. This was the only instance in the entire study that the cubic form of Sm203 was evidenced. The TmOBr and YbOBr decomposition processes proved to be more complex. In the case of TmOBr, 7 diffraction patterns were taken as the compound was heated from 775 to 1323°. The patterns show three solid phases to exist at 1035-1150° under apparently equilibrium conditions, and present serious problems from a thermodynamic point of view. These problems, and their implications for the LnOBr decomposition systems, are discussed in the next chapter. It is significant that the diffraction lines due to Tm304Br, at their maximum intensity, are only of weak to medium strength. The YbOBr decomposition followed a similar course. Three different X-ray traces were made, each from a Mo cell; two different preparations of the oxide bromide were employed, one of which contained a slight impurity of Yb203, the other a slight impurity of Yb3oaBr. Two of the traces examined the decomposition over the temperature range 800-1150° and in each case, three solid phases were observed to exist simultaneously in the temperature range 950-1050°. The same result was obtained when the decomposition was monitored at T a 950° i;50°. As in the TmOBr system, the Ln203 and Ln3O4Br phases were first 61 observed on the same Guinier film, with the oxide diffraction lines the more intense. As the decomposition progressed, the Ln203 pattern became successively more intense until it was the only phase observed, while the Ln3O4Br pattern was evident only as a few lines of weak or medium intensity. For the TmOBr decomposition process, the TmOBr and Tm304Br phases disappeared concurrently, and the same was true for two of the examined YbOBr decompositions. In one of the YbOBr traces, however, it was clear that the YbOBr phase disappeared just before the Yb304Br phase. At this stage of the process, with only Yb3O4Br and Yb203 contributing to the overall diffraction pattern, the residue was analyzed for metal content and was found to contain 39.8 mole per cent of the sesquioxide and 60.2 mole per cent of the tetraoxide monobromide. In view of the very weak diffraction lines observed for Yb3043r relative to those of Yb203, it is clear that either the sample used for the x-ray analysis was not representative of the bulk material or the Yb304Br structure inherently gives only weak reflections. The observed decomposition processes for TmOBr and YbOBr are particularly puzzling in terms of the thermodynamics of the systems. If a specified LnBr3 gas, or a mixture of LnBr2 and Br gases, is the vapor product(s) for both the LnOBr and Ln304Br decompositions, one would not expect to find the three solid phases present under equilibrium conditions, as was apparently observed. Some insight into this puzzle was provided by an X—ray investigation of such a three phase residue in a YbOBr decomposition trace. This residue was critically examined and found to consist of particles of two general types: a large number of small but discrete granules, very similar in appearance to the product of the YbOBr preparations; and a much finer 62 powder which could not be separated readily into discrete particles. A sample of each type from the same residue was examined; the intensities of the YbOBr and Yb203 lines were found to be comparable in the two films whereas the lines due to Yb3O4Br were much more intense in the sample prepared from a single granule. Implications of this finding are discussed in the next chapter. Finally, it was desirable to determine the thermal decomposition mode of a LnOBr hydrate in order to assess the effect of possible hydrolysis products on the target collection studies. A sample of SmOBr was exposed to air for one week, at which time its diffraction pattern revealed spurious lines. This sample was heated to 1120° for 6 hours (10'5-10'"6 torr), and the resultant powder pattern of the residue showed only lines assignable to SmOBr and Sm3O4Br. It was inferred that the monoxide monobromide hydrate did not decompose to a third oxide bromide phase and, accordingly, no difficulty was anticipated for the target collection experiments, even if the samples were hydrolyzed slightly. D. Mass Spectrometry Results of the mass spectral investigations of the effusates over the several LnOBr species are presented in Table 6. The estimated error in the appearance potentials is 1_O.7 eV. Considered in light of published results for several trichlorides, tribromides, and triiodidesln"116 (Table 7), and in light of the known mass spectra of EuBr2,1°8 EuC12,116 Eu12,117 and YbClZ,118 the data indicate the principal lanthanide effusates to be NdBr3, SmBrz, YbBrZ, and TmBr3. A particularly noteworthy observation is that for every MX3 (X - Cl, Br, 1) species for which mass spectral data for the monomer 63 n.ea 0.0H 0.0H O: CH 8.... .2: 93 TB 03 N m A+z +§ + xx + N5 A>OV mamauemuom.ouamumomm< me “ OOH u no u e mm 0: HOOD» we a ma u OOH u Hm mm o: HOOBH mm H OOH “ mm H II mm o: umoam me “ OOH " NH " II oe O umOam mm A an " OOH " II Om 0: umOOz mm x mm " OOH " II mm o: umomz ma " me " OOH " II ON oz umOOz A z u N: " Nxz u mxzv A>Ovewuonm HHOO maaawm + + + + we finanOH cuouumm magnumsoawmum moanmom oauuoaouuoomm mmmz .O manna 64 Table 7. Published Fragmentation Patterns of Trichloride, Tribromide, and Triiodide Monomers Fragmentation Pattern Vapor + + + + Ionizing Species (MX3 : sz : MX : M ) Energy (eV) Ref. 3013 35 : 100 : 6 : 2 75 111 BBr3 37 : 100 : 49 : 1 -- 112 313 O : 100 : 22 : 9 -- 112 AlCl3 43 : 100 : 26 : 16 100 113 AlBr3 71 : 100 : 18 : 40 100 113 PCl3 37 : 100 : 19 : 10 70 114 PBr3 6O : 100 : 31 : 19 70 114 YCl3 25 : 100 : 5 5 70 115 LaCl3 1.5 : 100 : 0 : 0 50 116 L001 26 : 100 : 3S : 70 50 116 65 are available, the MKZ+ peak is the most intense in the fragmentation pattern. While this fact may not preclude the MBr3 parent species for the effusates over SmOBr and YbOBr, it suggests that one consider other possibilities. The small yet distinct presence of the YbBr3+ ion was surprising and suggests that both YbBrz and YbBr3 may be present in the effusing beam. Implications of this observation will be discussed later. E. Bulk Effusate Collections X-Ray analysis of the light-blue condensed effusate over NdOBr showed NdBr315 and some tribromide hydrate to be present. A Many attempts were made to collect the effusate from SmOBr, and the resultant powder patterns typically revealed the yellow-green tribromide.15 These patterns were rendered complex and unsatisfactory, however, by extensive contamination with the tribromide hydrate, identified by comparison with the pattern of the residue obtained from a solution of Sm203 in hydrobromic acid. Use of the apparatus sketched in Figure 2 at 1400-1500° yielded a much purer product, with diffraction lines due to the hydrate only of very weak intensity. This condensed effusate was analyzed for Br content by potentiometric titration and found to contain 62.92 Br; theoretically, SmBr3 contains 61.452 Br. It should be noted that a 0.1 mg uncertainty in the mass of the 2.7 mg sample gives a 3.72 uncertainty in the Br analysis. No trace of the dibromide was observed in any of the several X-ray investigations; in one experiment, however, in which the apparatus of Figure 1 was used, in addition to SmBr3 a phase (or phases) was observed which was not the familiar tribromide hydrate. In this experiment, three effusate samples were examined, each taken from 66 different sites in the narrow quartz tubing. In each sample the SmBr3 pattern was strong, but at successively higher positions in the tubing, the diffraction lines due to the spurious phase became more intense. The bottommost of the three samples (nearest the furnace) was essentially pure SmBr3; the topmost sample, in addition to SmBr3 and the unknown phase, contained a small amount of the familiar tribromide hydrate. It is hypothesized that the spurious phase is an intermediate hydrate. After one target collection experiment over an SmOBr-SmBOgBr mixture, the effusates were analyzed for both Sm and Br on the fluorescence spectrometer; a Br:Sm ratio of 3°15.i.0-09 was determined. The pale green effusate from YbOBr yielded the diffraction pattern of the tribromide.11+ No dibromide lines were observed, although the pattern was contaminated by spurious diffuse lines, presumably an indication of hydrolysis. The Br content of this condensate was 52.72 (for YbBr3, Z Br - 58.1; for YbBrz, 2 Br - 48.0). SmBr Derived lattice parameters for the NdBr and YbBr3 3’ 3’ phases are compared with published values in Table 8. Table 8. Comparison of Derived and Published Lattice Parameters for the Tribromides of Nd, Sm, and Yb Derived Published Crystal Lattice . Lattice . Phase System Parameters (A) Parameters (A) Ref. NdBr3 Orthorhombic a - 4.109 1: 0.006 a - 4.11 i 0.03 15 b - 12.66 i 0.02 b - 12.65 1; 0.05 c - 9.165 _+_- 0.009 c - 9.17 i 0.04 SmBr3 Orthorhombic a - 4.045 1: 0.004 a = 4.04 i 0.03 15 b = 12.68 _+_- 0.02 b - 12.64 i 0.05 c - 9.116 1; 0.009 c - 9.08 i 0.04 YbBr3 Hexagonal a - 6.980 1’0.001 a - 6.981 t 0.002 14 c - 19.106 1 0.002 b = 19.115 :1; 0.006 67 The effusate from EuOBr was collected in the same apparatus as that employed for the NdOBr, SmOBr, and YbOBr samples cited above. The resultant diffraction pattern matched that of EuBr23° and confirmed Haschke's earlier finding.5 F. Terget Collection Egperiments l. Crucible Interaction No evidence was found for interaction between the Mo cells and the several samples used. 2. Residue Analysis a. Neodymium Monoxide Mbnobromide Study The compacted solid material in the crucible after the vaporization experiments unfailingly consisted of two layers with a visibly distinct interface: a dark blue upper layer of Nd3oaBr and a light blue lower layer of NdOBr. The solid state transition from an extremely fine powder to a highly sintered mass was sufficiently complete to permit dissection of the residue with a knife. No spurious diffraction lines were observed. b. Samarium Mbnoxide Monobromide and Samarium Tetraoxide Monobromide Studies Residues from the SmOBr and Sm304Br vaporization experiments were highly sintered masses which could be removed easily from the cell as single bricks. X-Ray analyses revealed only the expected lines. 3. Veporization Behavior of Neodymium Monoxide Monobromide a. Constant Orifice Size For a given orifice the vaporization of NdOBr proceeded as expected: the same pressure values were obtained at a given temperature when this temperature was approached from both the high and low 68 directions. This observation, in conjunction with the mass spectrometric, weight-loss, decomposition trace, and effusate collection data, confirms the decomposition reaction as: 4 NdOBr(s) - Nd3oaBr(s) + NdBr3(g). (V-l) b. As A Function of Orifice Size The measured vapor pressures of NdBr3(g) over the NdOBr-Nd3O4Br mixture showed a definite trend with orifice area. The smaller the orifice, the higher the measured vapor pressure. Although this trend is not evident from the data obtained from experiments in which relatively large orifices were employed and in which possible temperature gradients on the oven were not monitored, later experiments pointedly confirm this dependence. The target collection data are presented in Figure 4. This trend was confirmed by a pair of consecutive vaporization experiments in which the only parameter changed was orifice size; orifice areas of 11.3 X 10-4 and 70.6 X 10-4 cm2 were used, and the measured pressures differed by > 502 at T - 1617 K. By use of Motzfeldt's formula, a vaporization coefficient of W 10'2 was calculated (at 1700 K). This number is only an order of magnitude value, however, since Pm is not a linear function of me and av is very sensitive to cell temperature gradients. Behavior such as this indicates that the vapor pressures measured are not the equilibrium values,at least for the data obtained with the larger orifices. These data must be interpreted in terms of a steady state rather than an equilibrium pressure; the effect of a kinetic constraint on the observed pressures is evident. 69 l 1 T j T I I 1 RUN,0RIFICE AREAXIOIcmz . _ U 8;IO'2 60 0 9;69'7 V ||;43'5 ‘ h l2;7°8 7-0 - o l45|04°3 <1 l5;||'3 A I6:70-6 ’5‘ 8-0 F ._ x 2|;2-46 .- 9 22,245 3 . ’6 90 ~ v '9 1. B I In .. "’ 1‘74) r . ‘s 2 3 a O z I .o _ —I' I ‘c' ' . I2-o - \ I3-O - \ J J J l J l l L 5-4 5-6 5-8 6-0 6-2 6-4 6-6 6-8 Io‘xr Figure 4. Measured Knudsen Pressures of.NdBr3(g) over a NdOBr(s)-Nd304Br(s) Mixture 7O 4. Vaporization Behavior of Samarium Mbnoxide Monobromide and Samarium Tetraoxide Monobromide These two sets of experiments are considered together because of the similarity of their vaporization behavior. a. Constant Orifice Size Unlike the NdOBr system, neither of these decomposition processes yielded the same vapor pressure for a specified temperature as this temperature was approached from opposite directions. The only exceptions were that in each system normal vaporization behavior was observed with the smallest orifice used and, for the SmOBr system, this effect was not examined for the next to smallest orifice size. Reasons for this anomalous behavior will be discussed later. At least for the smallest orifice used, the vaporization data were consistent with a decomposition which involved two solid phases. Considered in view of the weight-loss, decomposition trace, and effusate collection data, the decomposition reactions are thought to be: 4 SmOBr(s) = Sm3O4Br(s) + SmBr3(g) (V-2) and 3 Sm304Br(s) = 4 Sm203(s) + SmBr3(g) (V-3) over the respective temperature ranges examined. It cannot be denied that some of the previously cited results, specifically the mass spectral data, suggest the vapor decomposition products to be the dibromide and atomic Br; this discrepancy will be treated later. 71 b. As a Function of Orifice Size Vapor pressures over both of the samarium oxide bromide systems varied with the size of the orifice used, and for the smaller orifices, the dependence was seen to be the same for both decomposition processes. Results of the target collection experiments for both of these processes are presented in Figures 5 and 6. Figure 7 shows the data for the SmOBr and Sm3048r decompositions on the same set of axes (for the smallest orifice). At 1650 K the use of Motzfeldt's equation yields for the SmOBr system “v = 1.3 X 10'3 and for the Sm304Br system av - 1.8 X 10'3. (These calculations were based on experimental data collected with the smallest two orifices.) Again, these values should be regarded only as approximate. For example, use of data from several crucibles with orifice areas ranging from 9.5 X 10'5 to 6.9 X 10'3 cm2 did not yield a linear plot for the SmOBr system, and the vaporization coefficient estimated from these data was N 4 X 10'3 at 1575°K. It is instructive to compare the results of the Sm3043r decomposition in Mo cells with 4 vaporization experiments conducted in graphite cells; in each case, the cells were situated inside the W-Mo oven. For comparable orifice areas (1.0 X 10‘3 cm2 in Mo and 1.1 X 10.3 cm2 in graphite), the slopes of the in P‘ye l/T plots agree to within 2.5% and the absolute pressures (at mid-T) to within 6%. By use of data from only the graphite cells, an av of 10"2-10-3 is calculated at 1575 K, depending on the weight assigned to each experimental point. 5. A/d Values Values of A/d for the three vaporization processes are given in Appendix II. The calculated diameters of the LnBr3 gaseous molecules r I I r T T I I T I I I I —I I 4.6 ‘_ Emmet AREA giO‘cmz- .~ 0 IIINCR. T);69’2 r a I(DECR. T) " _ s 2 (lNCR.T);lO-8 5’6 r- K. V 3(‘NCR.T,;40‘5 - v 3(DECR.T) , - s. o 4A(lNCR.T);40-5 '* E 5'5 ' a 9 4B (lNCR.Tl;43-l - 8 . J ,5 . V . ,_ '. "' n 7 6 L. In . .. Us) . o. 8-6 - , ‘. .. Z r ‘ _I , . I as ~ '~ 4 MmRIFICE AREAon‘ cm2 :_ h x 5; 0-95 . d 2:22:22 9 J h 0’-P\: a; lie; 1 I .I L .1. 1 I J 1 1 J, I l 1, 57: 5-7 5-9 GI 83 6-5 6-7 6-9 7-I 4 l0 /T Figure 5. Measured Knudsen Pressures of SmBr3(g) over a SmOBr(s)-Sm3043r(s) Mixture 73 I“ ‘r 1 I I I I’ I 1 r MIORIFICE AREA Inc‘s“2 5-25 I: l (INCR. T);lO-0 -I n I (DECRT) 5-75 - g o 2 (INCR. T); IO-O 'j 0 3;O°8 6'25 *- . x 4;O°8 -‘ A g \. 0 6-75 . . '1 v A 3 7-25 h- \ ‘ _+ '0 . _ . 5 E 775 I- » 4 (D g 825 I ‘ 4 J 1. 19125II- g d 9.75 — LI VI . J_ L l L l L l L L L 5-6 5-8 6-0 6-2 6-4 6-6 4 I0 /T Figure 6. Measured Knudsen Pressures of SmBr (g) over a Sm304Br(s)-Sm203(monoclinic) Mixture 74 r I I 1 I I I I I 4.3 MNAPORIZATION X 5;SmOBr 6 6;SmOBr D 7;SmOBr A E I- c.- 3 A 5-9 - 3 IV) _ L m ‘5; GVI' P CL. 2 h .J I '7'5 r- 5343 - _L _L i LIL l l _L L 5'5 5'7 5'9 6" 6'3 6'5 I04/T Figure 7. Equilibrium Pressures of SmBr (g) over SmOBr(s)—Sm304Br(S) and Sm3048r (s) ~Sm203 (monoclinic) Mixtures 75 were based on a trigonal planar model in each case and employed a Ln-Br distance of 2.72 A27’28 and ionic radii of 0.995 and 0.97 A for Nd+3 and Sm+3, respectively.13 6. Vapor Pressure quations a. Neodymium Monoxide Monobromide System The linear least-squares fit of the 28 in PNdBr3-33 l/T data points (R - 1.987 eu, Mo cell, orifice area - 2.45 X 10'4 cmz), is described with standard deviation as: 3 _ [(54.94 i 0.38) x 10 ] 2n PNdBr3 -I- T + (24.2 7;: 0.22) (1683-1850 K). (V-4) From this equation the thermodynamic data for reaction (V-l) at the median temperature (1766 K), together with their standard deviations, are: AH°1766 = 109.17 i 0'76 kcal/gfw NdBr3 (V-5) O 3 .- AS 1766 48.22 i 0.43 eu. (V 6) b. Samarium Monoxide Monobromide System By use of vapor pressure data obtained for reaction (V-2) from the Mo cell with the smallest orifice area, the linear least-squares fit of the 28 data points is: [(49.1 + 0.1 x 103] 6-— 6 in P = — SmBr3 T + (23.47 i 0.10) (1547-1768 K). (V-7) 76 Associated thermodynamic values at the median temperature (1657 K) are: - 97.6 $10.32 kcal/gfw SmBr3 (V-8) O A“ 1657 8 AS°1657 3 46.63 i 0.19 eu. (II-9) In light of the mass spectral evidence for an SmBrz gaseous decomposition product, the results of the target collection experiments of the SmOBr decomposition system can be expressed most generally in terms of an SmBrx effusing species: PSmBrx [(49.16 i 0.16) x 103] _________ = - + (20°49.i.°°10) (MWSmBrx)1/2 T in (1547-1768 K). (V-10) c. Samarium Tetraoxide Menobromide System Again, the best data are considered to be those obtained with the smallest orifice. Expressed in terms of an SmBr3 effusate, the 24 in PSmBr3 XE l/T data points are represented, with standard deviation, by the linear least-squares equation: [(50.19.1'0.41) x 103] 2n PSmBr3 = - T + (23.25 i 0.24) (1636-1775 K). (V-ll) Conversion to standard thermodynamic functions at the median temperature (1705 K) yields: Asol7os ‘3 46.21 _‘t 0.48 eu. (II-13) 77 Again, the more general expression of the target collection results in terms of an SmBrx effusate yields: PSmBrx [(50.19.:,0.41) x 103] in = - +-(20.27 i 0.24) (MW )1/2 T SmBrx (1636-1775 K). (v-14) 7. Data Reduction a. Neodymium Monoxide Monobromide System Second law data reduction to 298 K required extensive use of estimates in view of the absence of experimental values. The estimative technique employed for C arallels that used for the estimation p.NdBr3(8) p of C 0 Kelley119 reports heat capacity data for 16 gaseous p.8mF3(g)‘5 MX3 species (M not necessarily a metal) with masses ranging from 71 to 456 amu. About 1000° these data show the heat capacity of MX3(g) to be virtually mass independent, but at lower temperatures the magnitude of CI) becomes decidedly mass dependent, especially for the lighter masses. From this pool of data the Cp values for BiCl3(g) and AsI3(g) were selected as prototypes since the average molecular weight of these two compounds is 100.4% of the molecular weight of NdBr3 and 98.8% of the molecular weight of SmBr3. The heat capacities of BiCl3(g) and AsI3(g) are very nearly equal from 298 to 2000 K (Table 9); for this study, the Cp values of NdBr3(s) and SmBr3(s) were estimated to be that of BiC13(g). Table 9. Comparison of Thermodynamic Values for BiClB(g) and AsI3(g) o _ 0 ° - ° VaP°r Cp298 Cp1000 Cp1500 Cp2000 H1657 H298 S1657 S298 BiCl3(g) 19.02 19.78 19.82 19.83 26.85 kcal/gfw 33.70 eu AsIB(g) 19.28 19.82 19.85 19.86 26.94 kcal/gfw 33.85 eu 78 The heat capacity of NdOBr was assumed to vary linearly over the temperature range 298-1766 K. This assumption is not based on experimental data; no C data have been measured for LnOX (X = Br, P Cl) compounds, but the assumption of linear C 'y§_T behavior has been P employed by other researchers for LnOCl systems.120 The heat capacity of NdOBr at 1766 K was estimated as 21.75 en.“5 The heat capacity at 298 K was determined by the following two-step estimative scheme: 1. From the measured heat capacity of NdBr3(s)16 at 298 K (24.01 eu), the atomic heat capacity contributions of Nd and Br were each assigned a value of 6 eu on the basis of Kopp's Rule. 2. The oxygen contribution to the total Cp,298,NdOBr value was determined by subtracting 12 eu (the Nd contribution) from the measured heat capacity of Nd203121 at 298 K and dividing this difference by 3. The resultant value for Cp,298,NdOBr is 16.9 eu and the overall heta capacity expression for NdOBr from 298-1766 K is: -3 Cp,NdOBr(s) 3 3.30 X 10 T + 15.92. (V‘lS) Estimated error limits for this scheme are 0.5 eu for Cp,298’ 1.0 eu for Cp,l766’ and 0.75 eu for the uncertainty in the shape of the C vs T plot. p— The heat capacity of Nd304Br was taken to be the sum of the heat capacities of NdOBr and Nd203 over the entire temperature range, with uncertainties of 0.3 eu in the summation at a given tempearature and 0.3 eu in the temperature dependence of Cp. Use of the second law data at 1766 K, together with the above estimates, yields: 79 AH°298 = 115'3-i-3'6 kcal/gfw NdBr3 (V-l6) O - — as 298 55.1 i 4.1 eu. (v 17) Third law values of AH°298 were obtained by use of the above estimations in conjunction with $0298 estimates for NdOBr(s), Nd3048r(s), and NdBr3(g). Values of 80298 for NdOBr(s) and Nd3O4Br(s) were estimated by Latimer's“7 technique with inclusion of a magnetic contribution for Nd as outlined by Westrumf‘8 the estimated error is .1 2 eu for the difference: 4 The entropy S° -S° . 298,NdOBr 298,Nd304Br of NdBr3(g) at 298 K was estimated from Shimazaki and Niwa's sublimation data25 at the melting point (with an estimated uncertainty of 1 eu), from the measured heat capacity of NdBr3(s),16 and from the estimated heat capacity of NdBr3(g). A value of 96.9 eu was obtained, with an associated uncertainty of 1.3 en. The resultant free energy functions yielded an average AH°298 of 108.9 kcal/gfw NdBr3 for reaction (V-l). The range of third law AH°298 values was 108.5-109.2 kcal, with an associated 0.5 kcal temperature trend, and the uncertainty in the average third law enthalpy value 18;: 8.2 kcal. The magnitude of the discrepancy between the second and third law AH°298 values is put into perspective by noting that it corresponds to an error in SoZ98,NdOBr of 1.0 eu, an error in Cp,NdOBr(s) of 0.7 eu, or an error in 80298,NdBr3(g) of 3.8 en (42). For example. use of a Cp,NdOBr(s) value 0.7 eu lower than that employed above yields second and third law AH°298 values of 112.2 and 112.4 kcal/gfw NdBr3, respectively, with no discernible temperature trend in the third law values. b. Samarium Monoxide Mbnobromideg§ystem The estimations used in reducing second law AH°1657 and AS°1657 80 data to the corresponding values at 298 K parallel those used for the NdOBr decomposition. The heat capacity of SmBr3(g) was estimated to be that of BiC13(g). The assumption that Cp,SmBr3(s) - Cp,NdBr3(s) was employed to get the Sm and Br atomic contributions to Cp,Sm0Br at 298 K and the atomic contribution of oxygen was derived from data for Sm203.121 For SmOBr, Cp,1657 was taken as 21.75 eul‘3 and an assumed linear variation of Cp with T over the temperature range 298-1657 yields 0 = 3.42 x 10'3 r + 16.08. (V-18) p,SmOBr(s) The heat capacity of Sm3O4Br was taken as the sum of Cp,SmOBr and Cp,$111203 over the temperature range of interest. The resultant reduced data for the SmOBr decomposition system (V-2) are: AH°298 = 103.4 : 3.2 kcal/gfw SmBr3 (V-19) AS°298 = 53.0 1 4.0 eu. (v-20) The error limits reflect the individual error estimates cited for the NdOBr data reduction plus an uncertainty of 0.2 an in the assumption that SmBr3(s) and NdBr3(s) have equal heat capacities at 298 K. An average third law AH°298 value was calculated with the above estimates plus an estimated value of AS°298 with respect to reaction (V-2). The 80298 values for the oxide bromides were obtained from Latimer"7 and Westrum,"8 and 80298,SmBr3(g) was taken as $0298,NdBr3(g)’ except for an adjustment for the different magnetic contributions of the metal.“8 Assigned uncertainties parallel those for the NdOBr system and the resultant average AH°298 value is 100.4 1 7.5 kcal/gfw SmBr . Individual third law AH°298 values varied from 100.1 to 3 100.6 kcal and displayed a 0.3 kcal temperature trend. 81 c. Samarium Tetraoxide Monobromide System The estimates employed here are the same as those cited for the SmOBr system. Equation (V—18) was employed in the calculation of cp,Sm3O4Br» even though the temperatures attained were higher than those of the SmOBr vaporization experiments. Additionally, SoZ98,Sm203 has been estimated from Latimer and Westrum for consistency. The resultant second and third law data for reaction (V-3) are: AH°298 = 104.9 1 3.5 kcal/gfw SmBr3 (Second law) (V-21) AS°298 - 52.0':_4.3 eu (Second law) (V-22) AH°298 = 103.5 1 8.2 kcal/gfw SmBr3 (Third law). (V-23) The individual third law AH°298 values ranged from 103.4 to 103.7 kcal, with no discernible temperature trend. 8. Sticking Coefficient Determination the bouncing experiment effected to determine the sticking coefficient of the samarium bromide effusate on chilled A1 targets yielded the expected result: no measurable quantity of Sm was found on the side of the A1 disc which faced the target. In retrospect, this result is not considered significant, however, due to (1) the oversized hole in the A1 disc, and (2) the close proximity of the target to the disc. CHAPTER VI DISCUSSION It is no secret to those familiar with scientific literature that the naked presentation of data only imperfectly reflects an experimenter's work, with the associated subtleties. Data, presented without comment, may serve to misrepresent or at least to mislead. Accordingly, this chapter attempts to place the foregoing results in their proper context, in the hope that the data will be accorded their full value, but only that value, by the scientific community. A. Evaluation of Experimental Work 1. Characterization of the Decomposition Modes a. WeightrLoss Experiments For incongruently vaporizing systems, weight-loss data play an integral part in the determination of high temperature decomposition modes. The data are both precise and generally conclusive, although no definitive information is revealed about the discrete vapor species. In the course of this investigation, two unanticipated subtleties were found. First, the decomposition mode of SmOBr in_vgggg depended on the temperature to which the sample was heated. At 1350°, decomposition progressed to Sm304Br; at 675°, SmOBr decomposed to the cubic sesquioxide. Weight-loss data indicate the latter process to result from the reaction of SmOBr with the residual 02 and H20 in the evacuated chamber (10.5-10'6 torr), and the advisability of using a better vacuum system is obvious. Second, the initial weight-loss data from the decomposition of SmOBr in graphite (1350°) revealed considerable sample-cell interaction, 82 83 yet the cell did not lose appreciable weight (0.6 mg) and an X-ray analysis of the residue revealed no sample contamination. This apparent discrepancy can be explained from a consideration of reaction (VI—1) Sm203(s) + 3 C(s) = 3 C0(g) + 2 Sm(g). (VI-l) Although this reaction may not accurately describe the anomalous weight-loss data, it has been reported to occur at 11350°,389"’9 and it is important to note that sample-cell interaction can occur without sample contamination or a marked change in cell weight. Clearly, the sample and cell should be examined independently. b. Solid State Decomposition Traces (1) General Considerations The use of sequential X-ray diffraction patterns to monitor a solid state decomposition mode is a sensitive technique for the study of incongruently vaporizing substances. For high temperature vacuum systems the procedure can be tedious but the resultant data make the effort worthwhile. The variation of sin26 values provides information on subtle lattice shifts (e.g. solid solution formation or a gradual change in the stoichiometry of a phase) and indicates discrete changes in the number of phases present; sequential intensity data reveal trends in the relative abundances of the various phases. Like most experimental tools, however, such a trace yields data which must be interpreted carefully. Theoretically, only nucleated, annealed phases will give sharp powder patterns. Experimentally, care must be taken that no phases are overlooked due to an insufficient number of patterns poorly distributed over the decomposition time and 84 temperature intervals. Further, the assumption that the phases identified at room temperature are the same as those existent at the experimental decomposition temperature may not be valid. (2) Decomposition Traces of the Monoxide Monobromides of Thulium and Ytterbium The utility of this research technique is evidenced in the TmOBr and YbOBr decomposition systems. In each case, three phases (LnOBr, Ln304Br, and Ln203) were observed to exist in apparent equilibrium at elevated temperatures. Thermodynamically, it is not possible for these three phases to exist at equilibrium under a specified vapor pressure of the corresponding lanthanide bromide. It is hypothesized that the surface Ln304Br phase which results from the initial decomposition of the monoxide monobromide forms a coating which effectively shields the buried LnOBr from the vapor phase. Because the LnOBr cannot diffuse to the surface rapidly enough, the vapor pressure drops sufficiently to provide a driving force for the decomposition of Ln304Br to the sesquioxide. Similarly, the resultant oxide coating inhibits diffusion of the Ln304Br and LnOBr phases to the surface, and non-unity vaporization coefficients are observed. While this model may describe satisfactorily the decomposition process in finely divided particles, it does not obviously explain the relatively greater occurrence of the Ln304Br phase in the granules. However, if one considers that the LnOBr phase is able to diffuse through the Ln304Br layer in these granules, but is unable to diffuse through the relatively thick Ln203 shell there exists an environment very similar to that used for the preparation of pure 85 Sm304Br and Nd304Br samples. The oxide and monoxide monobromide react in the solid state to form the tetraoxide monobromide; this is a much more efficient conversion than that resulting from the thermal decomposition of the LnOBr phase. In the former method, one mole of LnOBr yields one mole of Ln304Br; in the latter, four molecules of LnOBr yield one mole of Ln304Br. .Accordingly, the granules, with their thicker, less penetrable oxide layers, become relatively rich in the Ln304Br phase. As the tetraoxide monobromide layer in turn becomes thicker, less LnOBr can diffuse to the oxide layer and under an undersaturated bromide pressure, the system may become essentially static: LnOBr cannot diffuse to the oxide layer (to form Ln304Br), and Ln304Br cannot diffuse to the sample surface (to vaporize). Such a static environment was apparently attained in the thulium system when a mixture of TmOBr, Tm304Br and Tm203 appeared to be stable after several hours of heating in an evacuated ampoule. c. Mass Spectra The mass spectrometric experiments represent the only direct measurements in which the properties of the discrete gaseous species were examined in the vapor phase. The resultant data are the most disturbing of any encountered in this study, and have been the focus of months of concerted contemplation. (l) Fragmentation Patterns Unfortunately, the theoretical foundations of fragmentation processes are not formulated sufficiently well to permit the prediction of fragmentation patterns for most inorganic compounds. Accordingly, the interpretation of a particular spectrum rests largely on a comparison with spectra of chemically and/or structurally similar compounds. Kiser,_e__t_.§_l__.111+ suggest a guideline for the interpretation 86 of the mass spectra of inorganic halides: the (parent minus one halogen)+ ion is typically the most intense one in the spectrum. This guideline seems to describe trihalides very well (see Table 7), dihalides much less accurately. For the gaseous lanthanide bromides examined in this thesis research, the guideline suggests the effusing species to be NdBr3, SmBrz, TmBr3, and YbBrz. For the NdOBr and TmOBr systems, these data are in agreement with the other data obtained in this research. For the SmOBr and YbOBr systems, however, these data are in striking disagreement with the results of the bulk effusate and (for SmOBr) target collection experiments. The small but distinct YbBr3+ peaks indicate the presence of a tribromide parent species. In Table 10, the fragmentation patterns of SmBrx, YbBry, and EuBr2 are compared. The SmBrx and EuBr2 molecules yield very similar patterns; in contrast to the EuBrz pattern, the YbBry vapor yields MBr3+ peaks and also gives a relatively more intense MBr2+ signal, characteristic of a tribromide vapor. Thus, the fragmentation patterns indicate that SmBr2 is the lanthanide vapor product from the decomposition of SmOBr, while a mixture of YbBrz (and Br) and YbBr3 comprises the vapor phase over the decomposing YbOBr. Table 10. Fragmentation Patterns of SmBrx, YbBry, and EuBr2 Fragmentation Pattern Vapor + + + + Ionizing Species (LnBr3 : LnBr2 : LnBr : Ln ) Energy (eV) SmBrx —- : 27 : 100 : 57 35 YbBry 4 : 65 : 100 : 48 35 *EuBr —- : 15 : 100 : 50 2 3! Ref. 108 87 The different fragmentation patterns obtained for NdBr3 and TmBr3 are interesting, but it is difficult to assess accurately the significance of this difference. (The most abundant isotope of TmBr3 comprises 37.9% of the species while, for NdBr3, only 21.3% of the total signal is due to the most abundant isotope). Only two other lanthanide trihalides (other than fluorides) have been examined,116 and it is noteworthy that, like TmBr3, the LuC13 vapor species gives an appreciable parent ion intensity, while LaCl3, like NdBr3, gives virtually no parent ion signal. Whether or not this correlation reflects a real trend from the lighter lanthanides to the heavier ones is not clear, but it is apparent that different fragmentation pathways are accessible for various lanthanide trihalides. (2) Appearance Potentials One is tempted to treat appearance potential data as precise thermochemical measurements from which dissociation energies and enthalpies of formation may be calculated. Such treatments are of questionable value, however, when the ionization reactions are not well defined. Accordingly, the appearance potentials measured in this study are not used for the calculation of derived thermodynamic data; rather, they are simply presented for what they, in fact, are: the first appearance potential data for the lanthanide tribromides (or for any MBr3 monomeric vapor where M is a metal heavier than A1), and the only appearance potential data available for lanthanide dibromides, aside from EuBrz.108 An empirical check on the consistency of the TmBr spectral 3 data is suggested by Kiser 25.51.;11” they state that for inorganic halides, an increase of approximately 1.5-2.0 eV occurs for the lowest appearance potential above the ionization potential, if a significant 88 amount of the parent ion is observed. This generalization is in good agreement with the spectrum of TmBr3, where the TmBr3+ intensity is relatively strong and the appearance potentials of TmBrz+ and TmBr3+ are, respectively, 11.1 and 9.6 eV. d. Bulk Effusate Collections For air-insensitive effusates which are thermodynamically stable from room temperature to the experimental decomposition temperature of the substrate, the bulk effusate collection technique is straightforward and the resultant data are subject to unambiguous interpretation. Problems arise, however, when a compound decomposes to yield an extremely hygroscopic vapor. A collection experiment in which the hydrolyzed form of the effusate is the predominate condensate does not constitute a definitive identification of the vapor species. (For example, Haschke and Eick3o have reported that anhydrous EuBrz hydrolyzes to the tribromide hexahydrate.) Fortunately, the lanthanide dibromides are more stable to air (water vapor) than are the corresponding tribromides. Consequently, an effusate which, when condensed and examined, consists primarily of the tribromide with slight contamination with tribromide hydrates, may be assumed to contain little or no dibromide. Another complication which can arise in effusate collection experiments is illustrated by a LnOBr phase which decomposes thermally to yield gaseous LnBr2 and Br. To condense the dibromide, a relatively cold collection surface must be employed, but as the temperature is lowered below the empirical decomposition temperature, the condensed tribromide becomes more stable with respect to the dibromide and bromine. Thus, once the dibromide has condensed on the collector, it 89 may react with the impinging Br to form the tribromide. Since the collisions of the Br atoms with the condensed LnBrz are not believed to be 100% efficient, some control may be exercised over the extent of the conversion. Consider, for example, the effect of a variable collector temperature. The colder the collection surface, the more stable the tribromide with respect to the dibromide, and the greater the driving force favoring a recombination reaction. Accordingly, the apparatus sketched in Figure 1 might well yield different condensates depending on the distance of the condensation site from the furnace. To dramatize this effect, consider the target collection apparatus where the target surface is sufficiently cold to condense the bromine as well as the dibromide. In this case, recombination would be favored not only by thermodynamic considerations, but also by the greatly enhanced contact between the reactants. A consequence of this reasoning is that the measured 3:1 Br:Sm ratio determined by an analysis of the target surfaces does not definitively specify a tribromide effusate. The fact that no trace of a dibromide was observed for any condensed effusate other than EuBrz, while the tribromide was clearly evident in the vapor decomposition products from NdOBr, SmOBr, and YbOBr, suggests strongly that the effusing vapors were, respectively, NdBr3, SmBr3, and YbBr3. e. The high-Temperature Decomposition Modes of Samarium Monoxide Mbnobromide and Ytterbium Monoxide Monobromide It is believed that YbOBr decomposes igLygggg at 1000° according to reaction (VI-2): 4 YbOBr = Yb304Br + YbBr3(8). (VI-2) 90 This reaction is consistent with the results from the effusate collection, weight-loss, and X-ray decomposition trace experiments. It is not inconsistent with the mass Spectral data since the spectra were taken at 1100° and exhibited appreciable tribromide character. It is instructive to calculate the PYbBr2/PYbBr3 ratio as a function of PYbBrz at 1100° by a consideration of the isomolecular exchange reaction YbBr3(g) + Br(g) - YbBr2(g) + Br2(g). (VI-3) No measured thermochemical data are available for the ytterbium 23 o bromides, but Feber provides an estimate of AH f,298,YbBr2(g) - , and 8° values may be calculated from the methods 0 AH f,298,YbBr3(g) 298 47 48 ~ o . of Latimer and Westrum, coupled with AS sub,298,YbBr2 46.7 eu (average of 7 lanthanide and alkaline earth dichlorides and di-iodides) and Asosub,298,YbBr3(g) - 52.0 en.25 The heat capacity of YbBr2(g) is taken as that of HgBr2(g)1°8’119 and C is estimated to p,YbBr3(g) be Cp,BiC13(g)°119 Thermochemical data for Br2(g) and Br(g) are taken from JANAF Interim Thermochemical Tables.122 The calculated ratios are given in Table 11. If one assumes a YbBrz pressure of 10"5 atm a PYbBrz/PYbBrB ratio of N 200 is calculated at 1400 K. In light of the approximations involved, this estimation is in good agreement with the observed fragmentation pattern; the fact that there is agreement within two orders of magnitude lends credence to the conclusion that both YbBrZ and YbBr3 are present in the effusate at 1100°. Nevertheless, based on the results of the bulk effusate collection experiment, YbBr3 is thought to be the primary gaseous apecies at lOOO°. 91 Table 11. Estimated Knudsen Cell (PLnBrZIPLnBr3) Ratios As Functions of T and PLnBr2 YbBrZ-YbBr3—Br—Br2 System PYbBr2(atm) TK 10’2 10’“ 10'6 10‘8 500 9 x 10"11 x 10‘10 x 10‘9 x 10'8 800 2 x 10"5 x 10" x 10‘“3 x 10’2 1100 5 x 10’3 0.1 9 x 102 1400 0.3 21 x 103 x 105 1700 7 x 102 x 104 x 106 2000 85 x 103 x 105 x 107 SmBrz—SmBr3—Br-Br2 System P mBr2(atm) TR 10'2 10'“ 10'6 10‘8 500 3 x 10'12 x 10‘11 x 10'10 x 10'9 800 3 x 10'6 x 10"5 x 10"4 x 10'2 1100 1 x 10’3 x 10'2 3 x 102 1400 0.1 9 x 102 x 104 1700 4 x 102 x 104 x 106 2000 60 x 103 x 105 x 107 92 What is the decomposition mode of SmOBr (or Sm304Br)? The answer to this question has successfully eluded this researcher for several months; a clear-cut answer supported by evidence from each experimental tool employed has not yet been deduced, but an analysis of the evidence is instructive. Concisely expressed, the dilemma is this: data from the bulk effusate collection and Knudsen effusion experiments point to a tribromide effusate, while mass spectrometric data suggest a dibromide effusate. Additionally, consideration of the isomolecular exchange reaction SmBr3(g) + Br(g) - SmBr2(g) + Br2(g) , (VI-4) where thermodynamic estimates are made analogously to those of the corresponding Yb system (VI-3), suggests that the dibromide should predominate at 1400 K (Table 11). By far the single most definitive indicator of the vaporization mode is the series of Knudsen effusion target collection experiments. Since Sm analyses of the targets were used to calculate the number of moles of bromide which effused, the data can be treated either on the assumption of a tribromide decomposition product (as was done in Chapter V), or on the assumption that the effusate is the dibromide. As noted previously, the assumption of a tribromide product leads to a consistent set of thermodynamic values; especially noteworthy is the good agreement between second and third law AH°293 values for reactions (V-2) and (V-3). The assumption of a dibromide (and Br) effusate leads to very different results. By use of the estimative scheme cited in the last 93 . 108,119 chapter, and taking Cp,SmBr2(g) Cp,HgBr2(g)’ second and third law values of AS°298 for reaction (VI-5) are, respectively, 97.45 and 4 SmOBr(s) = Sm3048r(s) + SmBr2(g) + Br(g) (VI-5) 79.75 eu. This 18 eu difference contrasts with the 2 en difference if the calculations are based on a tribromide product. Clearly the target collection data do not describe reaction (VI—5) and in this context there can be little doubt that the vapor pressure data correspond to a tribromide effusate. How does one explain the mass spectrometric data? Several possibilities may be formulated: (1) The observed fragmentation pattern is that of SmBr3. It can be argued that the fundamentals of the fragmentation process are poorly understood, and thus it is possible that, surprisingly enough, SmBr3 fragments to the pattern shown in Table 10. This argument, however, is more of a postulate than a deduction and reflects an arbitrariness which this researcher is not eager to display. (2) The effusing vapor is the tribromide, which is efficiently reduced to the dibromide in the source region of the spectrometer before intercepting the ionizing beam. While this is a plausible explanation for the observed spectra, it is thought to be an unlikely occurrence since the molecular beam which enters the ionization chamber is not reflected from any identifiable surface. (3) The effusing vapor is a mixture of SmBr2 (and Br) and SmBr3 analogous to the mixture in the YbOBr system. That no tribromide+ peaks were observed in the spectra is not surprising since the NdBr3 vapor also did not give a parent ion signal. 94 This argument suggests that at the temperature at which the mass spectra were obtained (1275°), the dibromide and the tribromide are comparably stable, while at the temperatures of the target collection experiments (1274-149S°) the tribromide is decidedly more stable. This reasoning is difficult to accept for at least two reasons. First, this approach does not explain the close similarity of the observed fragmentation pattern with that of EuBrz.108 Second, the fact that a bulk effusate collection experiment at 1275° yielded the tribromide, with no trace of the dibromide, discredits this explanation. @) The molecular beam effusing from the cell is comprised of SmBrz and Br. This explanation readily accommodates the mass spectral data, and it suggests that either (1) the other experimental data need to be re-evaluated; or (2) the bromide vapor effusing from the spectrometer cell is not the thermochemically stable one. Although this hypothesis may appear strained initially, it is believed to describe best the vaporization process in the mass spectro- meter. To feign a thorough understanding of this process would be naive, but it is known that severe kinetic constraints are exhibited in the vaporization mechanism, as reflected by the marked pressure- orifice area dependence. Further, the X-ray analyses of the TmOBr and YbOBr decomposition processes indicate that the kinetics of vaporization are complex, and the drastic sintering of the decomposition residues (with the accompanying lattice shrinkage and reduction in porosity) suggests that the surface vaporization process may be inhibited severely. This subject will be discussed further in a later section, but it is introduced here as a credible rationalization by which the various experimental observations can be correlated. 95 Two other isolated pieces of information deserve mention with respect to the vaporization mode. First, no trace of the blood-red color characteristic of the Sm+2 ion was observed in the course of the bulk effusate collection experiments. Second, the fact that the gaseous bromide of Sm did not attack the Mo cells, while the effusate over EuOBr (EuBrz + Br) strongly reacted with Mo and apparently formed a volatile Mo bromide, suggests that atomic bromine is not present in significant quantities in the effusate over SmOBr. The discrepancy between the experimental results and the data of Table 11 is large, but it should be noted that no data are available for SmBr2(g) and the tabulated values simply reflect one researcher's estimate of the enthalpy of formation of SmBr2(g). 2. Vapor Pressure Measurements a. Attainment of Equilibrium Conditions There is no doubt that equilibrium conditions were not attained for the Knudsen experiments in which larger orifices were employed. It is believed, however, that for the NdOBr system the pressure data obtained with the smallest orifice closely approximate equilibrium data. This conclusion is substantiated by the fact that, for a specified temperature, the measured pressure is nearly constant for orifice areas_f 10'3 cmz. For the SmOBr and Sm304Br decompositions, one cannot state that the systems were at equilibrium with the same degree of confidence, even with the smallest orifice used. Nevertheless, it was not feasible to extend the experiments to yet smaller orifices, and in light of the good agreement of the data from the SmOBr and NdOBr studies, the deduction that equilibrium conditions were closely realized with the smallest orifice is a reasonable one. 96 On the other hand, one cannot deny that kinetics played an important role in the vaporization measurements when larger orifices were employed. The precise kinetic barrier to the decomposition process has not been identified and, indeed, the experiments were not designed to provide detailed kinetic data. (Not only is this study the first to examine the details of the LnOBr decomposition processes, but it is also the first in this laboratory to isolate definitively the orifice effect as a measurable parameter.) The systems clearly exhibit a non-unity av and/or ac, and it may well be suspected that the anomalous mass spectral (for the SmOBr system) and X-ray decomposition trace data (for the TmOBr and YbOBr systems) are related to the observed vaporization (and/or condensation) coefficient. It is tempting to suggest that the measured vaporization rates are not the equilibrium decomposition rates of the oxide bromides, but rather those of a slow, rate determining step in the overall reaction mechanism. For example, consider the set of reactions: LnOBr = LnOBr* (VI-6) LnBr3(g) (VI-7a) 4 LnOBr* = Ln304Br +' or LnBr2(g) + Br(g). (VI-7b) Suppose reaction (VI-6) represents a solid state kinetic constraint; perhaps the Ln304Br phase is not kinetically susceptible to nucleation, and the lattice rearrangements required for nucleation proceed much more slowly than the equilibrium rate at which surface molecules vaporize. In this situation, target collection data would reflect the kinetics of (VI-6) rather than the thermodynamics of (VI-7), and the data would not depend on the nature of the effusing 97 vapor. The strength of this argument is that it correlates the mass spectral data for the SmOBr and Sm304Br decomposition vapors with the other experimental evidence. The attractiveness of such an argument is rendered artificial, however, by the good agreement between the second and third law thermodynamic values calculated for reactions (V-l), (V-2), and (V-3), and by the temperature independent third law AH°298 data obtained for these processes. These thermodynamic consistencies are convincing evidences that the vapor pressures observed with very small orifices approximate equilibrium pressures. It is believed that the pressure data collected with the larger orifices do reflect a kinetic barrier to the vaporization process. The X-ray data for the TmOBr and YbOBr systems suggest that for LnOBr decomposition reactions at high temperatures, the vaporization rate may be controlled by the rate at which the vaporizing phase diffuses to the sample surface. With larger orifices the rate of effusion is comparable to the rate of condensation and--since the rate of mass transport through the orifice for a steady state system is a function of the relative rates of vaporization and condensation--anoma1ously low pressure measurements result. With smaller orifices, a greater fraction of the vapor molecules are reflected to the sample surface where condensation (and re-evaporation) occurs. This reasoning is borne out by experimental data. For the NdOBr system, the orifice effect is less pronounced than for the SmOBr system, but nevertheless marked. The change in slope of the in PNdBr3 '35 l/T plots with change in orifice area, coupled with the vaporization behavior as T was both increased and decreased, indicates that either or both 01V and ac vary with T and/or P in the NdOBr system. 98 For the SmOBr decomposition, the effect of orifice area on vaporization behavior is more striking. For the larger orifices, in PSmBr3 !§_1/T data were not reproducible as the temperature was successively increased and decreased. (It should be noted that in every experiment the first target was collected at a relatively low temperature; the temperature was increased incrementally for several targets then decreased.) It is presumed that the kinetic inhibition to the vaporization process is related to the thickness of the SmOBr- depleted surface layer. As successive targets are collected there is a very gradual decrease in the rate of the diffusion process at a given temperature, but a marked drop in the diffusion rate as T is increased (at higher temperatures sintering, surface area shrinkage, and surface shell thickening occur more rapidly). The further the reaction proceeds, the greater the deviation of the measured pressure from the equilibrium value. As a result, different pressures are measured at a given temperature if several targets have been collected at higher temperatures in the interim period, while the sequential collection of targets at a specified temperature may yield essentially constant pressure data. To be sure, much of the foregoing discussion is little more than speculation, but it is speculation founded on experimental evidence. The hypothesis of a solid state diffusion controlled vaporization process suggests that an investigation of the crystal structures of the solid phases may reveal some details of the diffusion mechanism. Further, the fact that the Ln3O4Br phase is the only one common to each system studied hints that the key to the mechanism may lie in the intricacies of the tetraoxide monobromide structure. The most striking 99 feature of this structure is the position of the Br atoms. Each is sandwiched symmetrically between two mirror-image octagons with vertices of alternating Ln and 0 atoms. The Br atoms appear to be trapped in these 16-member cages, and it is suggested that the principal kinetic barrier to vaporization involves the migration of the Br atoms to the sample surface. b. Target Collection Technique As detailed above, the Knudsen effusion method is not ideally suited to the study of the LnOBr and LnBOABr decomposition reactions. In this context, the target collection technique is not to be faulted. Aside from temperature measurements, which have been discussed earlier, the greatest source of error in the pressure determinations lay in the target analysis procedure.) The difficulty does not rest with the fluorescence method itself, but with the weighing technique employed for spectrometer calibration. The weighing process used is satisfactory for the determination of thermodynamic data, but is less preferred than well-conceived volumetric and coulometric calibration techniques for the measurements of absolute pressures and the stoichiometric composition of the condensed vapor(s). c. In Defense of the Knudsen Method It would be easy for the casual reader to assign the multitudinous experimental uncertainties to the Knudsen method itself; admittedly, the method is imperfectly suited to the examination of systems which exhibit appreciable orifice effects. This is not to say, however, that other methods commonly used to effect pressure measurements would be more advantageous. In fact, no satisfactory technique has been developed for the definitive measurement of equilibrium vapor pressures .III «III J; \Illli 100 for systems with av << 1; of the several less-than-satisfactory techniques employed, the Knudsen method has the fewest disadvantages. The resultant uncertainties are traced more truthfully to the inherent non-equilibrium properties of the system than to the imprudent selection of an experimental tool. B. Evaluation of Thermochemical Data 1. Second and Third Law Data After the experimental difficulties were eliminated or minimized, vapor pressure data which yielded reasonable enthalpy and entropy values were obtained. The second and third law agreement was satisfactory for the SmOBr and Sm304Br decomposition reactions, and was reasonable for the NdOBr experiments when one considers the magnitude of the approximations used. Specifically, the fact that the difference between the second and third law AH°298 values for reaction (V-l) reflects only a 0.7 eu error in the estimated heat capacity of NdOBr testifies to the overall consistency of the estimative scheme. Further, the temperature-independent third law values of AH°298 obtained for reactions (V-2) and (V-3) and for the adjusted data (0.7 eu lowering of cp,NdOBr) describing (V-l) indicate both the correctness of the assumed decomposition modes and the compatibility of these modes with the pressure data and thermochemical approximations. The suitability of the approximations employed is evidenced further by the successful application of an analogous estimative shceme to the data reduction of the EuOCl vaporization system.101 2. Enthalpies of Formation of Lanthanide Oxide Bromides A lack of experimental data for gaseous lanthanide trihalides limits severely the accuracy to which many thermochemical properties I Llf IA II, will)" 101 can be estimated. Nevertheless, attempts to make such estimations can provide rough guidelines and, accordingly, such an attempt is made here. While AHOf,298,SmBr3(8) has not been measured, it has been estimated to have values from -209 to -216 kcal/mole,2°-23 with an average value of -213 kcal/mole. (The standard state of bromine is taken as the perfect diatomic gas at 1 atm pressure.) When this average is used in conjunction with an extrapolation of Shimazaki and Niwa's sublimation data for the lighter lanthanide tribromides25 to SmBr3 (which is converted from melting point data to 298 K values by use of ACp’sub - -7 eu), AHOf,298,SmBr3(g) - -l4l kcal/mmle is calculated. The standard enthalpy of formation of B-Sm203 has been measured as -433.9 kcal/mole.123 Use of these data, together with the average of the second and third law values of AH°293 for reaction (V-3) yields AH f,298,Sm3O4Br(s) a -660 kcal/mole. (VI-8) By use of (VI-8), the estimate of AHOf,298,SmBr3(g)’ and an average of the second and third law values of AH°293 for reaction (V-2), one obtains AH f,298,Sm0Br(s) = -225 kcal/mole. (VI-9) The sensitivity of (VI-8) and (VI-9) to the estimated AHOf,298,SmBr3(g) is indicated by the fact that a 3 kcal/mole error in the estimated standard enthalpy of formation of the tribromide leads to a l kcal/mole error in the derived standard enthalpies of formation of the oxide bromides. 102 O . O employs the estimative technique outlined above (which gives AH.f,298,NdBr3(g) . -l45 kcal/mole), A check on the reasonableness of these data is obtained from an examination of the thermochemical data for (l) the hydrolysis reactions of NdC1312‘“125 and SmC13126 and (2) the dissolution reaction of Nd203 in hydrochloric acid.127 When these data are combined with the published enthalpies of formation for H20(g),122 HCl(g),128 and the respective trichlorides,129'131 standard enthalpies of formation for NdOCl and SmOCl are calculated to be -239 and -237 kcal/mole, respectively. These numbers indicate that the standard enthalpies of formation of the LnOBr phases are about 11 kcal/mole more positive than those of the corresponding LnOCl phases. This measured difference seems reasonable in light of published estimateszo’m’23 for the standard enthalpies of formation of the lanthanide trichlorides and tribromides. Based on the gaseous bromine standard state, estimates of AHOf,298,SmC13(8) - AHOf,298,SmBr3(8) range from -32 to -34 kcal/mole, or -11 kcal per Sm-X bond. Similarly, estimates of AHOf,298,NdCl3(S) - AH ) range from -31 to -33 kcal/mole, or N -11 kcal per of,298,NdBr3(s Nd-X bond. In light of the several approximations employed for each of the LnOBr and LnOCl data reductions, this close agreement may be more apparent than real. Nevertheless, the agreement does suggest an overall consistency in the experimental and estimated data. The estimated standard enthalpy of formation for EuOBr(s) (-206 kcal/mole) is in excellent agreement with the estimation {-203 kcal/mole)S based on thermodynamic data for the EuBr2108 and Eu304Br5 vaporization processes. 103 By use of analogous data,12°’122’12"”125-129’132-136 the standard enthalpies of formation of several other LnOBr species can be estimated (Table 12). (Drobot and Korshunov137’138 have also suggested values for the enthalpies of formation of the oxide chlorides, but their data are in marked disagreement with those of the other researchers cited above; accordingly, their data have not been employed in these calculations.) Table 12. Estimated Standard Enthalpies of Formation of LnOBr(s) Phases Reported Estimated -AHf,298,LnOCl -AHf,298,Ln0Br Ln (kcal/gfw) Ref. (kcal/gfw) La 243 120,125, 232 127,129 Ce 241 125 230 Pr 243 124,125, 232 (229) 135 Nd 239 124,125, 228 129 Pm 245 132,136 234 (227) Sm 237 126,131 226 Eu 217 101 206 Gd 235 126,127, 224 135 Tb (223) Dy 234 132 223 Ho 232 133 221 Er 231 132,134 220 Tm 229 132,134 218 Finally, it is instructive to compare the standard enthalpies of formation of condensed SmOBr and SmOCl with the value reported previously for SmOF:so AH - -275 kcal/mole. Estimateslg’zz’lzz O f,298,Sm0F(s) of AHOf,298,SmF3(8) - AHOf,298,SmBr3(B) range from -189 to -200 kcal/mole, with an average estimated value of -193 kcal/mole. This enthalpy 104 difference corresponds to a value of -64 kcal per Sm-X bond, compared to the value of -50 kcal per SmrX bond derived from the vaporization studies of SmOBr and SmOF. The agreement is not strikingly good, but is realistic in the context of the large number of approximations involved. C. General Comments and Observations 1. Experimental Routes to Definitive Answers Evidence has been presented by which the high temperature decomposition modes of LnOBr and Ln3O4Br (Ln - Nd, Sm, Tm, Yb) have been characterized. The thermodynamic description of the NdOBr decomposition process has been detailed by a Knudsen effusion technique which gives data compatible with that obtained from each experimental tool used. The decomposition processes of SmOBr and Sm304Br have been described thermochemically but the descriptions are not in agreement with the mass spectrometric data. One might well ask how the experimental uncertainties may be clarified. a. Attainment of Equilibrium Pressures No method has been devised whereby equilibrium pressure measure- ments can be made unequivocally over condensed phases which exhibit vaporization coefficients << 1. b. Vaporization Modes of Samarium Monoxide Monobromide and Samarium Tetraoxide Monobromide Definitive identification of the stable vapor product from the thermal decomposition reactions of SmOBr and Sm3O4Br may be effected in at least three ways. 105 1) Combination of torsion effusion data with the target collection data The torsion effusion technique is used to measure the momentum recoil of a suspended Knudsen cell with orifice pseudosurfaces in vertical planes, parallel to the suspending fiber. The recoil is calculated from the extent to which the supporting fiber is twisted as the vapor effusates. The torque is related to the vapor pressure inside the Knudsen cell by an expression independent of the molecular weight of the effusing gas. Once the pressure at a specified temperature and orifice size has been determined by this method, target collection data can be utilized to calculate the molecular weight of the effusate by (III-l4). For systems exhibiting significant orifice effects, extensive calibration efforts would be required. 2) Examination of the effusate by an IR matrix isolation technique Vapor species effusing from Knudsen cells have been isolated successfully in inert matrices and their IR spectra have been obtained.139-1“1 Such a measurement should distinguish unambiguously between a dibromide and a tribromide, although it is noted that experiments of this type have not always produced definitive structural determinations.139'1“° 3) Elucidation of the vaporization thermodynamics of lanthanide bromides If accurate thermodynamic data were available for the lanthanide di- and tribromides, the relative stability of the species (for a given lanthanide) could be determined from a consideration of the respective free energies and heat capacities. These data would be 106 definitive for the SmOBr and Sm3O4Br decomposition systems only if equilibrium conditions were approximated closely in the Knudsen cells. 2. Samarium as a Divalent Lanthanide While thermodynamic estimates for SmBr2(g) have been made, one should note that there is no direct experimental evidence for the existence of the species. The results of this study indicate that the divalent character of Sm is markedly less than is often assumed. This point will not be belabored here but is simply mentioned as a word of caution to those who would blindly accept extrapolations based on experimental data for a few compounds. 3. The Lanthanides: All Alike? The similarities among the several lanthanides, both as metals and in corresponding compounds, have been trumpeted for many years; the notion that the lanthanides exhibit strikingly similar behavior is widespread among chemists and is fostered by esteemed texts. That this notion is less than satisfactory, however, has been dramatized in this and recent related studies in this laboratory. For example, why is no orifice effect observed for the vaporization study of Eu304Br? Why does thoria react with EuOF110 and EuOBr“3 at their respective experimental decomposition temperatures, but not with EuOCl?107 How does one account for the dramatically different reactivities of EuOBr and SmOBr (or EuOF)110 with quartz under Knudsen conditions? These and other questions suggest some fundamental differences among the lanthanides, differences which too often go unreported in cursory treatments of the elements as a group. This particular research effort provides no answers to these questions but it does emphasize their reality, and it suggests that the 107 unifying key to a proper understanding of general lanthanide chemistry may as yet be undiscovered. 4. Future Investigations In retrospect it seems clear that the undertaking of this research may have been more ambitious than warranted by the available thermochemical data. The lack of data for the lanthanide halides is especially apparent. Accordingly, it is deemed wisest to postpone further vaporization studies of ternary lanthanide oxidehalide systems until the respective halides have been characterized more fully. Specifically, investigations of the SmBrz-SmBr3 and YbBrz-YbBr3 systems would be especially welcome, and a study of the former system has been initiated.”2 Additionally, vaporization studies of lanthanide thiohalide phases would be valuable in light of the oxidehalide data. The advisability of such investigations at this time is questionable, however, pending a more detailed elucidation of the thermochemistry of the lanthanide halides and the several lanthanide sulfide phases. REFERENCES 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) ll) 12) 13) 14) 15) l6) 17) 18) REFERENCES W. H. Zachariasen, Acta Crystallogr., 2, 388 (1949). I. Mayer, S. Zolotov, and F. Kassierer, Inorg; Chem., 4, 1637 (1965). H. Barnighausen, G. Brauer, and N. Schultz, Z. Anorg. A113. Chem., 328, 250 (1965). H. Barnighausen, H. P. Beck, and H. W. Grueninger, "Proceedings of the Ninth Rare Earth Research Conference," P. E. Field, Ed., Blacksburg, Va., 1971, p 74. J. M. Haschke and H. A. Eick, J. Amer. Chem. Soc., 22, 4550 (1970). N. Schultz and G. Reiter, Naturwissenschaften, 54, 469 (1967). V. Scherer, U.S. At. Eneggy Comm., BMXFKXRKKRQZRR, 1968; Chem. Abstr., 69, 102592m (1968). H. Bfirnighausen, personal communication, 1971. I. Mayer and S. Zolotov, J. Inpgg. Nucl. Chem., 21, 1905 (1965). L. Basile, J. Ferraro, and D. Gronert, J. Inorg;_Nucl. Chem., 33, 1047 (1971). J. Barnes and H. Pincott, U.S. Dept. Com., Clearinghouse Sci. Tech. Inform., Agmgglggk, 111 (1965); Chem. Abstr., 65, 9947b (1971). M. D. Taylor, Chem. Rev., 62, 503 (1962). D. Brown, "Halides of the Lanthanides and Actinides," John Wiley and Sons Ltd., New York, N.Y., 1968. D. Brown, S. Fletcher, and D. G. Holah, J. Chem. Soc. (A), 1889 (1968). W. H. Zachariasen, Acta Crystallogr., k, 265 (1948). A. S. Dworkin and M. A. Bredig, High Temp. Sci., 3) 81 (1971). F. H. Spedding and A. H. Daane, Met. Rev., 2’ 297 (1960). V. Jantsch and N. Skalla, Z. Anorg. Allg. Chem., 13%, 391 (1930). 108 l'l'l’l‘c'll.‘ I'll-II!" 19) 20) 21) 22) 23) 24) 25) 26) 27) 28) 29) 30) 31) 32) 33) 34) 35) 36) 109 L. Brewer, "Chemistry and Metallurgy of Miscellaneous Materials: Thermodynamics," L. L. Quill, Ed., McGraw-Hill Book Co., Inc., New York, N.Y., 1950, paper 7. ' L. Brewer, L. A. Bromley, P. W. Gilles, and N. L. Lofgren, ibid, paper 6. P. G. Maslov and U. P. Maslov, Izvest. Vysshikh. Ucheb. Zavedeniip Khim. i Khim. Tekhnol., g, 516 (1959). D. E. Wilcox and L. A. Brmmley, Ind. Egg; Chem., 55, 32 (1963). R. C. Feber, AEC Report LA-3l64, Los Alamos, N.M., 1965. E. R. Harrison, J. Appl. Chem., 2, 601 (1952). E. Shimazaki and K. Niwa, Z. Anggg. Allg. Chem., 314, 21 (1962). P. A. Akishin, V. A. Naumov, and V. M. Tatevskii, Nauchn. Dokl. Vysshei Shkoly Khim. i Khim. Tekhnol., 229 (1959); Chem. Abstr., 52, 19493e (1959). P. A. Akishin, V. A. Naumov, and V. M. Tatevskii, Vestnik Moskov. Univ., Ser. Mat., Mekh.,rAstron,,.Fiz., Khim., 14, 229 (1959); Chem. Abstr., 24, 31f (1960). P. A. Akishin, N. G. Rambidi, and V. P. Spiridonov, "The Characterization of High Temperature Vapors," J. L. Margrave, Ed., John Wiley and Sons, Inc., New York, N.Y., 1967, chapter 12. w. Doell and w. Klemm, z. Anorg. Allg. Chem., gap, 239 (1939). J. M. Haschke and H. A. Eick, J. Inorg. Nucl. Chem., 32, 2153 (1970). H. Czeskleba, Michigan State University, personal communication, 1970. H. H. Thomas, Jr., Ph. D. Thesis, Syracuse University, Syracuse, N.Y., 1968; University Microfilm, Inc., Ann Arbor, Mich. A. S. Dworkin, H. R. Bronstein, and M. A. Bredig, J. Phys. Chem., 61, 2715 (1963). A. H. Daane, Kansas State University, personal communication, 1971. J. E. Ricci, "The Phase Rule and Heterogeneous Equilibrium," Dover Publications, Inc., New York, N.Y., 1966, p vii. J. L. Margrave, "Physicochemical Measurements at High Temperatures," J. Bockris, J. White, and J. Mackenzie, Eds., Academic Press, Inc., New York, N.Y., 1959, chapter 2. 110 37) P. A. Pilato, Ph. D. Thesis, Michigan State University, East Lansing, Mich., 1968. 38) R. L. Seiver, Ph. D. Thesis, Michigan State University, East Lansing, Mich., 1971. 39) F. D. Rossini, J. Chem. Thermodynamics, 447 (1970). 40) T. B. Douglas, J. Res. NBS, 73%, 451 (1969). 41) J. R. McCreary and R. J. Thorn, High Temp. Sci., 2’ 330 (1971). 42) W. S. Horton, J. Res. NBS, 18A, 533 (1966). 43) J. M. Haschke, Ph. D. Thesis, Michigan State University, East Lansing, Mich., 1969. 44) G. N. Lewis and M. Randall, "Thermodynamics," revised by K. Pitzer and L. Brewer, Second Edition, McGraw-Hill Book Co., New York, N.Y., 1961. 45) 0. Kubaschewski and E. Evans, "Metallurgical Thermochemistry," Pergamon Press, New York, N.Y., 1958, pp 180—187. 46) P. E. Caro, J. Less-Common Metals, 16, 367 (1968). 47) W. M. Latimer, "Oxidation Potentials," Second Edition, Prentice— Hall, Inc., Englewood Cliffs, N.J., 1952, Appendix III. 48) E. F. Westrum, Jr., "Advances in Chemistry Series 71. Lanthanide/ Actinide Chemistry," R. F. Gould, Ed., American Chemical Society, Washington, D.C., 1967, pp 25-50. 49) R. L. Seiver and H. A. Eick, High Temp. Sci., 3, 292 (1971). 50) D. E. Work and H. A. Eick, J. Phys. Chem., 74, 3130 (1970). 51) A. V. Hariharan and H. A. Eick, High Temp. Sci., 3, 123 (1971). 52) R. D. Present, "Kinetic Theory of Gases," McGraw-Hill Book Co., Inc., New York, N.Y., 1958, pp 55-61. 53) K. D. Carlson, "The Characterization of High Temperature Vapors," J. L. Margrave, Ed., John Wiley and Sons, Inc., New York, N.Y., 1967, chapter 5. 54) Y. N. Lyubitov, "Molecular Flow in Vessels," trans. by W. Furry and J. Wood, Consultants Bureau, New York, N.Y., 1967, pp 69-75. 55) J. L. Margrave, "Physicochemical Measurements at High Temperatures," J. Bockris, J. White, and J. Mackenzie, Eds., Academic Press, Inc., New York, N.Y., 1959, chapter 10. 56) 57) 58) 59) 60) 61) 62) 63) 64) 65) 66) 67) 68) 69) 70) 71) 72) 73) 74) 75) 76) 111 R. J. Ackermann, P. W. Gilles, and R. J. Thorn, J. Chem.‘Phys., 25, 1089 (1956). D. A. Schultz and A. W. Searcy, J--PEX§° 6999;) 62, 103 (1963). T. H. Johnson, Phys. Rev., 31, 103 (1928). F. Knauer and 0. Stern, Z. Physik, 32, 765 (1926). J. W. Ward and M. V. Fraser, J. Chem. Phys., 42, 5129 (1968). S. Dushman, "Scientific Foundations of Vacuum Technique," John Wiley and Sons, Inc., New York, N.Y., 1949, chapter 2. P. Clausing, Ann. Physik, 961 (1932). W. C. DeMarcus, Oak Ridge Gaseous Diffusion Plant Technical Report K-1302, 1957, part 3. R. D. Freeman and J. G. Edwards, "The Characterization of High Temperature Vapors," J. L. Margrave, Ed., John Wiley and Sons, Inc., New York, N.Y., 1967, Appendix C. W. L. Winterbottom and J. P. Hirth, J. Chem. Phys., 31, 784 (1962). J. ward and R. N. R. Mulford, J. Chem. Phys., 41, 1710 (1967). J. W. Ward, R. N. R. Mulford, and R. L. Bivins, J. Chem. Phys., 41, 1718 (1967). F. D. Richardson and C. B. Alcock, "Physicochemical Measurements at High Temperatures," J. Bockris, J. White, and J. Mackenzie, Eds., Academic Press, Inc., New York, N.Y., 1959, chapter 6. R. J. Ackermann, R. J. Thorn, and G. H. Winslow, presented at 135th National Meeting of the American Chemical Society, Boston, Mass., 1959. L. B. Loeb, "The Kinetic Theory of Gases," Third Edition, Dover Publications, Inc., New York, N.Y., 1961, pp 331-335. 0. Knacke and I. Stranski, Progr. Metal. Phys., 6, 181 (1956). J. P. Hirth and G. M. Pound, Progr. Mater. Sci., 11, 1 (1963). J. W. Ward, J. Chem. Phys., 47, 4030 (1967). K. D. Carlson, P. W. Gilles, and R. J. Thorn, J. GREEtiEPYS" 38, 2064 (1963). J. W. Ward and M. V. Fraser, J. Chem. Phys., 42, 3743 (1968). K. Motzfeldt, J. Phys. Chem., 52, 139 (1955). 112 77) G. M. Rosenblatt, J. Electrochem. Soc., 110, 563 (1963). 78) R. J. Ackermann, R. J. Thorn, and G. H. Winslow, "The Characterization of High Temperature Vapors," J. L. Margrave, Ed., John Wiley and Sons, Inc., New York, N.Y., 1967, chapter 14. 79) G. A. Komlev, Russ. J. Phys. Chem., 38, 1498 (1964). 80) V. G. Muradov, Russ. J. Phys. Chem., 41, 1266 (1967). 81) G. M. Rosenblatt, J. Phys. Chem., 11, 1327 (1967). 82) R. J. Peavler and A. W. Searcy, J. Amer. Chem. Soc., 78, 2076 (1956). 83) C. I. Whitman, J. Chem. Phys., 28, 161 (1952). 84) R. Speiser and H. Johnston, Trans. Amer. Soc. Metals, 42, 283 (1950). 85) M. Rossman and J. Yarwood, J. Chem. Phys., 21, 1406 (1953). 86) E. W. Balson, J. Phys. Chem., 65, 1151 (1961). 87) R. C. Paule and J. L. Margrave, "The Characterization of High Temperature Vapors," J. L. Margrave, Ed., John Wiley and Sons, Inc., New York, N.Y., 1967, chapter 6. 88) L. Brewer and J. S. Kane, J. Phys. Chem., 22, 105 (1955). 89) R. Littlewood and E. Rideal, Trans. Faraday Soc., 52, 1528 (1956). 90) J. R. Hollahan and N. W. Gregory, J. Phys. Chem., 68, 2346 (1964). 91) D. Bonderman, E. D. Cater, and W. E. Bennett, J. Chem. Eng. Data, 15, 396 (1970). 92) D. L. Hildenbrand and W. F. Hall, J. Phys. Chem., 61, 888 (1963). 93) B. D. Pollock, J. Phys. Chem., 65, 731 (1961). 94) M. Farber and A. J. Darnell, J. Chem. Phys., 25, 526 (1956). 95) N. W. Gregory, J. Phys. Chem., 61, 618 (1963). 96) R. R. Hammer and N. W. Gregory, J. Phys. Chem., 68, 314 (1964). 97) N. A. Gokcen, J. Phys. Chem., 62, 3538 (1965). 98) C. Eisenhart, Science, 160, 120 (1968). 99) R. A. Kent, Ph. D. Thesis, Michigan State University, East Lansing, Mich., 1963. 100) 101) 102) 103) 104) 105) 106) 107) 108) 109) 110) 111) 112) 113) 114) 115) 116) 117) 118) 119) 113 G. H. Ayres, "Quantitative Chemical Analysis," Harper and Row, Publishers, New York, N.Y., 1958, p 388. A. V. Hariharan, Michigan State University, personal communication, 1972. N. J. Bjerrum, G. Schwarzenbach, and L. G. Sillen, "Stability Constants," The Chemical Society, London, 1958, Part II. J. J. Stezowski, Ph. D. Thesis, Michigan State University, East Lansing, Mich., 1968. J. D. H. Donnay, Ed., "Crystal Data Determinative Tables," Second Edition, American Crystallographic Association, Washington, D.C., 1963, p 841. H. J. Neff, Arch. Eisenhuettenw., 34, 203 (1963). J. M. Haschke and H. A. Eick, J. Phys. Chem., 12, 4235 (1968). A. V. Hariharan, Michigan State University, personal communication, 1971. J. M. Haschke and H. A. Eick, J. Phys. Chem., 74, 1806 (1970). G. J. McCarthy, W. B. White, and R. Roy, J. Inorg. Nucl. Chem., 42, 253 (1967). S. L. Bacon, M.S. Thesis, Michigan State University, East Lansing, Mich., 1971. J. Marriott and J. D. Craggs, Int. J. Elect., 4, 194 (1957). W. S. Koski, J. J. Kaufman, and C. F. Pachucki, J. Amer. Chem. Soc., 44, 1326 (1959). P. F. Porter and E. E. Zeller, J. Chem. Phys., 44, 858 (1960). R. W. Kiser, J. G. Dillard, and D. L. Dugger, "Advances in Chemistry Series 72. Mass Spectrometry in Inorganic Chemistry," R. F. Gould, Ed., American Chemical Society, Washington, D.C., 1968, chapter 12. J. D. McKinley, J. Chem. Phys., 44, 2245 (1965). J. W. Hastie, P. Ficalora, and J. L. Margrave, J. Lessfgommon Metals, 44, 83 (1968). A. V. Hariharan and H. A. Eick, to be published. A. V. Hariharan, N. A. Fishel, and H. A. Eick, to be published. K. K. Kelley, U.S. Bur. Mines, Bull. 584, 1960. 120) 121) 122) 123) 124) 125) 126) 127) 128) 129) 130) 131) 132) 133) 134) 135) 136) 137) 138) 114 C. W. Koch, A. Broido, and B. B. Cunningham, J. Amer.‘Chem. Soc., 14, 2349 (1952). L. B. Pankratz, E. C. King, and K. K. Kelley, U.S. Bur. Mines, Rep. Invest. 6033, 1962. "JANAF Interim Thermochemical Tables," D. R. Stull, Project Director, Dow Chemical Co., Midland, Mich., 1960 and Supplements. C. E. Holley, Jr., E. J. Huber, Jr., and F. B. Baker, "Progress in the Science and Technology of the Rare Earths," Vol. 3, L. Eyring, Ed., Pergamon Press, New York, N.Y., 1968, pp 379-384. C. W. Koch and B. B. Cunningham, J. Amer. Chem. Soc., 14, 1471 (1954). A. K. Baev and G. I. Novikov, Russ. J. Inorg. Chem., 19, 1337 (1965). C. W. Koch and B. B. Cunningham, J. Amer. Chem. Soc., 15, 796 (1953). V. P. Glybin and R. B. Dobrotin, Zh. Fiz. Khim., 45, 1582 (1971); Chem. Abstr., lg, 81067a (1971). C. E. Wicks and F. E. Block, U.S. Bur. Mines,ABu11. 605, 1963. D. A. Johnson, J. Chem. Soc.y(A), 2578 (1969). F. H. Spedding and C. F. Miller, J. Amer. Chem. Soc., 44, 4195 (1952). R. L. Montgomery and T. D. Hubert, U.S. Bur. Mines, Rep. Invest. 5525, 1959; Chem. Abstr., g4, 4134f (1960). V. F. Vishnevskii, Munich Univ., Rep., 1968; Nucl. Sci. Abstr., £2, 21678 (1969). F. Weigel and H. Haug, Chem. Ber., 24, 1548 (1961); Chem. Abstr., £2, 19446 (1961). F. Weigel and V. Wishnevsky, Chem. Ber., 49%, 193 (1970); Chem. Abstr., 1%, 48300m (1970). F. H. Spedding and J. P. Flynn, J. Amer. Chem. Soc., 14, 1474 (1954). F. Weigel, V. Wishnevsky, V. Scherer, and M. van Ghemen, Radiochim. Acta, 44, 38 (1969); Nucl. Sci. Abstr., 44, 24164 (1969). D. V. Drobot and B. G. Korshunov, Inorg. Mater., 4, 1978 (1965). D. V. Drobot and B. G. Korshunov, Inorg. Mater., 4, 931 (1966). 139) 140) 141) 142) 115 R. H. Hauge, J. W. Hastie, and J. L. Margrave, J. Les§:Comm9n Metals, 44, 359 (1971). R. D. Wesley and C. W. DeKock, J. Chem. Phys;, 44, 3866 (1971). C. W. DeKock, R. D. Wesley, and D. D. Radtke, High Temp. Sci., 4, 41 (1972). J. M. Haschke, University of Michigan, personal communication, 1971. APPENDICES 116 APPENDIX 1: Comparison of Observed and Calculated Interplanar d-Values Appendix IA: Tetragonal NdOBr (ref. 2) Iobs Cl(Mobs hkl d(Mealc Iobs d(A)obs hkl d(mcalc w 7.634 001 7.597 w 1.752 121 1.751 vw 3.802 002 3.799 w 1.717 104 1.718 m 3.563 101 3.556 ms 1.627 122 1.626 3 2.850 110 2.845 w 1.579 114 1.580 3 2.764 102 2.762 w 1.423 220 1.423 w 2.277 112 2.277 w 1.329 222 1.332 ms 2.014 200 2.012 w 1.273 310 1.273 w 1.894 113 1.892 w 1.265 006 1.266 w 1.779 202 1.778 032 1.264 Appendix IB: Tetragonal SmOBr (ref. 3) m 7.936 001 7.914 m 1.725 121 1.725 w 3.963 002 3.957 8 1.614 122 1.614 m 3.534 101 3.536 114 1.615 5 2.794 110 2.794 ms 1.397 204 1.398 102 2.796 m 1.318 214 1.318 w 2.633 003 2.638 222 1.318 m 2.283 112 2.283 006 1.319 8 1.975 200 1.976 8 1.250 032 1.250 m 1.917 021 1.917 m 1.134 007 1.131 m 1.769 202 1.768 117 Appendix IC: Tetragonal TmOBr (ref. 2) O O O O Iobs d(Mobs hkl d(“calc Iobs d(Mobs hkl d(Mcalc 3 8.387 001 8.288 w 1.728 202 1.729 m 4.162 002 4.144 wm 1.666 211 1.667 m 3.465 101 3.459 ms 1.642 114 1.642 w 2.988 3 1.574 212 1.574 vw 2.930 w 1.520 105 1.520 vs 2.805 102 2.803 m 1.448 213 1.449 vs 2.689 110 2.691 m 1.411 115 1.411 ms 2.561 111 2.560 m 1.401 204 1.402 m 2.256 112 2.257 w 1.345 220 1.346 m 2.237 103 2.236 w 1.327 221 1.328 wm 2.074 004 2.072 w 1.315 124 1.315 w 1.928 113 1.928 w 1.299 106 1.298 vs 1.901 200 1.903 ms 1.213 302 1.213 m 1.853 201 1.855 ms 1.202 310 1.204 m 1.820 Appendix ID: Tetragonal YbOBr (ref. 2) 3 8.343 001 8.362 3 1.641 114 1.647 w 4.162 002 4.181 3 1.568 212 1.567 3 3.450 101 3.444 w 1.523 105 1.529 3 2.802 102 2.804 ms 1.445 213 1.445 3 2.677 110 2.673 m 1.412 115 1.418 3 2.549 111 2.546 ms 1.399 204 1.402 ms 2.251 112 2.252 3 1.338 220 1.336 ms 2.237 103 2.243 m 1.321 221 1.320 m 2.078 004 2.090 m 1.312 214 1.314 m 1.925 113 1.929 w 1.301 106 1.308 3 1.893 200 1.890 w 1.274 222 1.273 3 1.846 201 1.844 m 1.249 205 1.252 m 1.824 104 1.829 w 1.231 116 1.236 w 1.722 202 1.722 3 1.208 302 1.206 ms 1.659 211 1.657 5 1.197 310 1.195 l 118 Appendix IE: Orthorhombic Nd3O4Br (ref. 6) O O O O Iobs d(Mobs hkl d(Mcalc Iobs d(mobs hkl d(mealc m 6.154 200 6.111 vw 2.402 141 2.400 ms 3.971 101 3.954 w 2.249 250 2.247 m 3.774 111 3.758 3 2.093 341 2.098 3 3.371 320 3.378 w 2.063 151 2.062 ms 3.317 211 3.317 w 2.040 600 2.037 vw 3.111 w 1.994 521 1.992 vw 3.061 400 3.056 w 1.932 620 1.930 w 2.970 410 2.962 wm 1.913 260 1.912 3 2.904 031 2.900 m 1.871 531 1.869 w 2.841 311 2.834 vw 1.817 630 1.817 5 2.827 131 2.821 ms 1.794 161 1.794 m 2.732 420 2.727 ms 1.775 232 1.775 m 2.712 240 2.707 3 1.730 541 1.729 mw 2.631 321 2.627 w 1.708 412 1.707 ms 2.439 430 2.434 Appendix IF: Orhtorhombic SmaoaBr (ref. 3) m 6.037 200 6.024 ms 2.070 002 2.070 ms 3.919 101 3.916 wm 2.038 151 2.037 m 3.721 111 3.721 wm 2.008 600 2.008 3 3.320 230 3.318 wm 1.987 060 1.988 m 3.275 121 3.274 wm 1.979 610 1.980 m 2.920 410 2.921 w 1.904 620 1.903 3 2.884 301 2.883 m 1.889 260 1.888 m 2.803 311 2.802 m 1.844 531 1.845 3 2.790 131 2.790 m 1.773 161 1.773 m 2.691 420 2.689 m 1.756 232 1.757 m 2.673 240 2.673 m 1.708 541 1.708 w 2.596 321 2.596 ms 1.637 242 1.637 m 2.401 430 2.401 m 1.576 711 1.576 w 2.218 250 2.218 Appendix IC: 119 Orthorhombic Tm 0 Br 3 4 O O O 0 Iobs d(Mobs hkl d(Mealc Iobs d(A)obs hkl d(Mealc w 5.797 200 m 2.786 301 vw 3.807 101 w 2.583 420 w 3.614 111 w 2.572 240 m 3.193 230 w 2.013 002 m 3.173 121 Appendix IH: Orthorhombic Yb3048r (ref. 3) w 5.789 200 5.754 w 2.558 240 2.555 m 3.798 101 3.787 wm 2.294 430 2.294 vw 3.603 111 3.594 vw 2.124 250 2.121 m 3.179 230 3.172 m 2.005 002 2.005 m 3.163 121 3.155 vw 1.802 260 1.805 m 2.775 301 2.772 vw 1.768 531 1.767 ms 2.686 131 2.683 vw 1.698 161 1.699 w 2.569 420 2.569 w 1.633 451 1.633 Appendix II: Condensed Effusate from Vaporization of NdOBr (Orthorhombic NdBr3 plus extra lines) (ref. 15) B m £§gm£883§€§€8§£€3 6.344 4.585 3.965 3.710 3.595 3.526 3.251 3.203 3.148 3.045 2.975 2.942 2.866 2.838 2.802 2.749 2.602 2.475 020 002 022 111 112 130 131 023 042 132 6.325 4.585 3.712 3.596 2.975 2.943 2.802 2.752 2.603 2.477 sszzzzzszgzg £3525 2.405 2.290 2.119 2.099 2.052 1.980 1.927 1.916 1.878 1.810 1.788 1.763 1.645 1.612 1.555 1.526 113 004 133 151 061 114 062 202 134 153 025 223 242 064 006 2.408 2.292 2.120 2.097 2.055 1.977 1.916 1.875 1.810 1.761 1.761 1.647 1.613 1.552 1.528 120 Appendix IJ: Condensed Effusate from Vaporization of SmOBr (Orthorhombic SmBr3 plus extra lines) (ref. 15) O O 0 O Iobs d(Mobs hkl d(Mealc Iobs d(Mobs hkl d(A)ca1c vw 6.354 020 6.320 5 2.389 113 2.379 vw 6.096 wm 2.280 004 2.270 vw 4.551 002 4.540 w 2.108 133 2.100 vvw 3.955 vw 2.093 151 2.086 w 3.857 110 3.848 m 2.022 200 2.020 m 3.552 111 3.543 vvw 1.925 220 1.924 3 2.946 112 2.935 w 1.848 202 1.846 vvw 2.921 130 2.916 w 1.628 171 1.622 wm 2.784 131 2.776 vw 1.546 064 1.544 wm 2.744 023 2.730 wm 1.520 006 1.513 vw 2.610 042 2.594 wm 1.513 204 1.509 Appendix IK: Condensed Effusate from Vaporization of YbOBr (Hexagonal YbBr3 plus extra lines) (ref. 14) tmégzagém 6.351 6.018 5.950 3.486 3.184 3.057 2.675 2.353 2.123 003 110 006 113 116 009 6.372 3.490 3.186 3.061 2.353 2.124 2.014 1.921 1.814 1.683 1.593 1.530 1.292 1.249 1.219 300 303 119 223 00 12 226 413 30 12 416 .015 .921 .814 .683 .593 1.531 1.292 1.250 1.219 P'F‘P‘h‘h) 121 APPENDIX 11: Maximum and Minimum Values of l/d for the Target Collection Experiments Vaporizing Phase Run (A/d)max (A/d)min NdOBr 8 5.89 0.56 NdOBr 9 60.4 3.67 NdOBr 11 13.3 0.97 NdOBr 12 3.28 0.83 NdOBr 14 50.5 2.81 NdOBr 15 5.46 0.60 NdOBr 16 15.9 3.34 NdOBr 21 2.15 0.13 NdOBr 22 2.22 0.35 SmOBr l 8.50 1.17 SmOBr 2 4.68 0.34 SmOBr 3 7.43 0.67 SmOBr 4 10.3 0.84 SmOBr 5 1.03 0.07 SmOBr 6 3.09 1.56 SmOBr 7 0.97 0.13 Sm304Br 1 2 83 0.38 Sm304Br 2 5.20 0.64 Sm3048r 3 1.54 0.14 Sm 0 Br 4 0.97 0.24 3 4 122 * APPENDIX III: Values of Afef for Reactions (V-l) , (V—2), and (V-3) i ‘v v v v ~—r i w Afef, vaporization Afef, vaporization Afef, vaporization TK of NdOBr of SmOBr of Sm304Br 1525 48.57 1550 48.52 1575 48.47 1600 48.42 1625 48.36 48.59 1650 48.31 48.54 1675 50.14 48.26 48.50 1700 50.12 48.20 48.46 1725 50.10 48.15 48.41 1750 50.07 48.10 48.37 1775 50.05 48.04 43.32 1800 50.02 43.27 1825 50.00 1850 49.97 * Includes the 0.7 cu adjustment of C p,NdOBr noted in text 123 APPENDIX IV: Equilibrium Pressure and Third Law Enthalpy Data * Appendix IVA: The Vaporization of NdOBr ; Reaction (V-l) AH° AH° v,298 v,298 TK — In P (atm) (kcal/gfw) TK - 1n P (atm) (kcal/gfw) 1683 8.339 112.27 1714 7.785 112.40 1704 7.950 112.32 1712 7.805 112.34 1724 7.630 112.51 1735 7.439 112.55 1746 7.185 112.36 1733 7.442 112.44 1770 6.837 112.64 1753 7.078 112.46 1770 6.803 112.52 1754 7.103 112.58 1791 6.458 112.59 1793 6.357 112.36 1790 6.449 112.50 1793 6.364 112.38 1820 5.879 112.27 1770 6.761 112.38 1821 5.889 112.37 1770 6.771 112.41 1850 5.426 112.37 1731 7.446 112.32 1850 5.417 112.34 1731 7.473 112.42 1794 6.338 112.35 1703 7.995 112.41 1794 6.341 112.36 1683 8.377 112.40 *Includes the 0.7 eu adjustment of Cp,NdOBr noted in text Appendix IVB: The Vaporization of SmOBr; Reaction (V-2) 1611 7.039 100.49 1660 6.143 100.42 1641 6.494 100.48 1615 6.953 100.45 1675 5.886 100.42 1636 6.597 100.53 1684 5.738 100.43 1663 6.110 100.48 1710 5.296 100.36 1688 5.661 100.40 1737 4.813 100.21 1725 5.018 100.26 1738 4.785 100.16 1699 5.482 100.38 1768 4.316 100.14 1682 5.785 100.48 1768 4.313 100.12 1655 6.252 100.49 1747 4.696 100.34 1628 6.746 100.54 1747 4.683 100.30 1606 7.110 100.45 1721 5.113 100.40 1547 8.304 100.59 1720 5.115 100.35 1565 7.935 100.55 1681 5.813 100.52 1582 7.598 100.53 124 Appendix IVC: The Vaporization of Sm Br; Reaction (V-3) 304 V‘ V ‘ V— ‘—w W ‘ — v—wv—v T 1 p AH;,298 T 1 AH;,298 K ’ “ (atm) (kcal/gfw) K ‘ n P (atm) (kcal/gfw) 1636 7.466 103.73 1685 6.499 103.45 1647 7.197 103.52 1710 6.070 103.45 1673 6.762 103.63 1739 5.598 103.49 1680 6.628 103.59 1739 5.569 103.39 1681 6.571 103.46 1728 5.817 103.62 1702 6.266 103.66 1728 5.812 103.60 1729 5.803 103.63 1714 6.018 103.51 1734 5.721 103.63 1692 6.383 103.47 1757 5.289 103.42 1659 6.987 103.54 1757 5.321 103.53 1694 6.380 103.58 1775 4.994 103.39 1732 5.753 103.63 1682 6.573 103.53 1729 5.753 103.46 1771!}!!! I jilll‘l‘l