WADC TECHNICAL REPORT 59-207 THE UNIVERS ITY OF MICHIGAN COLLEGE OF LITERATURE, SCIENCE, AND THE ARTS Department of Chemistry Final Report THE CHEMISTRY OF BORON HYDRIDES AND RELATED HYDRIDES R. W. Parry R. C. Taylor C. E. Nordman G. Kodama S. G. Shore E. R. Alton C. R. Peters J. C. Carter C. Reimann C. Cluff J. R. Weaver A. E. Emery M. Yamauchi C. W. Heitsch UMRI Project 2469 under contract with: WRIGHT AIR DEVELOPMENT CENTER AIR RESEARCH AND DEVELOPMENT COMMAND UNITED STATES AIR FORCE CONTRACT NO. AF 33(616)-3343 WRIGHT-PATTERSON AIR FORCE BASE, OHIO administered by: THE UNIVERSITY OF MICHIGAN RESEARCH INSTITUTE ANN ARBOR May 1959

FOREWORD The work reported herein was conducted under Contract No. AF 33(616)-3343 with the United States Air Force, the sponsoring agency being the Aeronautical Research Laboratory of the Wright Air Development Center, Air Research and Development Command, Direct technical responsibility for the investigation was vested in the Chemistry Research Branch of the Aeronautical Research Laboratory under the supervision of Dr. James Bierlein. Dr. William Ruigh served as task scientist, representing WADC in technical matters. The work was conducted in the Department of Chemistry of The University of Michigan through The University of Michigan Research Institute. WADC TR 59-207

TABLE OF CONTENTS Page LIST OF TABLES vi LIST OF FIGURES ix ABSTRACT xi I. REACTIONS AND STRUCTURES OF THE AMMONIA ADDITION COMPOUND OF DIBORANE AND OF ITS DERIVATIVES 1 A. Background 1 1. Early structural arguments 1 2. Experimental structural arguments for B2H6-2NH3 4 B. Further Studies on "The Diammoniate of Diborane," [H2B(NH3)21 [BH4],and Related Molecules 8 1. Synthesis of B2H6.2NH3 in one gram-lots 8 2. The preparation of microcrystalline B2H6.2NH3-the X-ray powder pattern of "the dianmmoniate of diborane 9 3. Additional chemistry of B2H6*2NH3. Evidence for and against the existence of "diammoniate of diborane II" with the formula [HB(NH3)31[BH4]2 10 4. Studies on the ammonotysis of the residues resulting from the reaction of an alkali metal with "the diammoniate of diborane" in liquid ammonia 18 C. Halide Salts of the General Form [H2B(NH3)2]X 23 1. Preparation of salts containing the complex cation [H2B(NH3)2]+ 23 2. Characterization of the halide salts of the [H2B(NH3)21+ cation 28 3. The crystal structure of [H2B(NH3)2 +salts 30 4. The nuclear magnetic resonance spectrum of the complex cation [H2B(NH3)2]+. The proton spectrum. 50 5, The reactions of the [H2B(NH3)2]X salts with sodium in liquid ammonia 52 Do The Synthesis of "The Diammoniate of Diborane" by Metathesis 52 1. The reaction between KBH4 and [H2B(NH3)2]Cl 53 2. The reaction between LiBH4 and [H2B(NH3)2]B3H8 in diethyl ether 53 E. A Summary of Developments on the Diammoniate of Diborane and Its Derivatives 55 WADC TR 59-207 iii

TABLE OF CONTENTS (Continued) Page II. REACTIONS, STRUCTURES, AND PROPERTIES OF OTHER DIBORANE ADDITION COMPOUNDS 57 A. Background 57 B. The Amine-Boranes 59 1o Preparation and stability of the mono-, di-, and trimethylamine-boranes 59 2. Vapor pressures and heats of vaporization for the amine-boranes 61 3. The dipole moments of the amine-boranes 64 4. The Raman spectra of the amine-boranes 78 5. The nuclear magnetic resonance spectrum of Me2HNBH3 83 C. Phosphorous Trifluoride-Borane 83 1. Background 83 2. The Raman spectrum of phosphorus trifluoride-borane and phosphorus trifluoride-borane-d3 84 3. The reaction between F3PBH3 and ammonia 86 D. Raman Spectra, Vibrational Assignments, and Force Constants for H3BCO and D3BCO 88 1. Experimental 88 2. Experimental results 89 3. Assignments 89 4. Normal coordinate treatment 93 E. The Structure of the Diammoniate of Borane-Carbonyl 98 1. Background 98 2. The model 99 III. THE REACTIONS AND STRUCTURES OF THE AMMONIA ADDITION COMPOUNDS DERIVED FROM TETRABORANE 100 A. Background 100 1. The problem 100 2. The preparation and properties of B4H1o.2NH3 101 3. Chemical evidence for the structure of B4Ho102NH3 102 4. The system ammonia-tetraborane in ether-excess ammonia present 107 5. The system ammonia-tetraborane without solvent 107 6. Discussion 112 7. Experimental 113 C. The Preparation and Properties of Ammonia-Triborane, H3NB3H7 120 1o Earlier studies 120 2. Methods for synthesizing H3NB3H7 121 WADC TR 59-207 iv

TABLE OF CONTENTS (Concluded) Page 3. Properties of H3NB3H7 123 4. Experimental 125 D. The Structure of H3NB3H7 from a Single-Crystal X-ray Study 132 1. Earlier structural information 132 2. High-temperature form 132 3. Low-temperature form 135 4. Discussion 140 E. Other Triborane Addition Compounds 145 1. Trimethylamine-triborane 145 2. Attempts to prepare F3PB3H7 148 IV. SOME ADDITION COMPOUNDS RELATED TO THE BORANE ADDUCTS 150 Ao Background 150 B. The Compound F3PA1C13 151 1. Earlier studies 151 2. The preparation of F3PA1C13 152 3. Experimental 153 C. The Compound H3A1[N(CH3) 3 12 158 V. PHYSICAL AND CHEMICAL STUDIES ON PURE BORON HYDRIDES AND BOROHYDRIDES 159 A. The Raman Spectra of Four Isotopic Varieties of Diborane in the Gas Phase 159 1. Experimental 159 2. Results 160 B. Raman Spectroscopy in Liquid Ammonia Solutions. Vibrational Frequencies and Force Constants for Isotopic Species of the Borohydride Ion Having Tetrahedral Symmetry 164 1. Introduction 164 2, Experimental 165 3. Experimental results and discussion 165 4. Normal coordinate treatment 168 C. The Dipole Moment of Tetraborane 172 1. The problem 172 2. Experimental 173 3. Results and discussion 173 D. Exchange Studies on Decaborane-14 175 1. Background 175 2a Experimental 176 3, Results 176 VI. CHEMICAL CORRELATTONS 178 WADC TR 59-207 v

LIST OF TABLES No. Page I. Inteplanar Spacings for the Crystalline Form of the "Diammoniate of Diborarne" 10 II. The Reaction of B2H6.2NH3 with HC1 in Dimethyl Ether. Evidence Against the Existence of Diammoniate (II) or [HB(NH3)3](BH4)2 18 III. The Preparation of [H2B(NH3)2]C1 from B2H6-2NH3 and HC1 in Various Ethers 25 IV. Summary of Analytical Data for the Characterization of the [H2B (NH3) a2X Salts 29 V. Observed and Calculated Structure Factors for [H2B(NH3)2]C1 38 VI. Comparison of Lattice Parameters for Chloride, Bromide and Iodide Salts of [H2B(NH3)2]+ 44 VIIo Observed and Calculated hkl Values for [H2B(NH3)2]C1 46 VIII. Observed and Calculated hkl Values for [H2B(INH3)2]Br 48 IXo Observed and Calculated hkl Values for [H2B(NH3)2]I 49 Xo Vapor-Pressure Data for Some Amine Boranes 63 XI. Summary of Dipole Moment Data 72 XII. The Dipole Moment of the Amine-Boranes in Dioxane as a Solvent 76 XIII. Assignment of Frequencies in the Spectra of the Ammonia-Boranes 80 XIV. Description of Fundamental Vibrations of H3N:BH3 81 XV. Observed Raman Frequencies and Assignments for Liquid PF3BH3 and PF3BD3 at -80~ C 85 XVI. Observed Raman Frequencies of Liquid H3BCO at -80~C 90 XVIIo Observed Raman Frequencies of Liquid D3BCO at -80~C 90 WADC TR 59-207 vi

LIST OF TABLES (Continued) No. Page XVIII. Force Constants for the H3BCO Molecule 95 XIX. Comparison of Observed and Calculated Values of the Fundamental Frequencies for the Various Isotopic Species of the H3BCO Molecule 96 XX. Comparison of Product Rule Ratios for Various Isotopic Combinations of the H3BCO Molecule 97 XXI. Interplanar Spacings and Relative Intensity Data for B4Hlo.2NH3 102 XXII. The Reaction of B4H1o and NH3 in Diethyl Ether 114 XXIII. Summary of Runs on B4H1oNH3 117 XXIV. The Addition of Excess B4H10 to Ammonia 118 XXV. Interplanar Spacings and Relative Intensity Data for H3NB3H7, High-Temperature Form (Tetragonal) 124 XXVI. Observed and Calculated Structure Amplitude (High-Temperature Form) 137 XXVII. Observed and Calculated Structure Factors (Low-Temperature Form) 141 XXVIII. Atomic Parameters and Their Standard Deviations 142 XXIX. Interatomic Distances and Bond Angles 143 XXX. Data for the Reaction of A12C16 and PF3 156 XXXI. Observed Raman Frequencies of Gaseous lB2H6 and BH 1061 XXXII. Observed Raman Frequencies of Gaseous llB2D6 and l~B2D6 162 XXXIII. Raman Active Fundamentals of Gaseous Diborane 163 XXXIV. Product and Sum Rules for the Ag Class Frequencies of Gaseous Diboranes 163 XXXV. Vibrational Frequencies and Assignments for the Borohydride Ion in Liquid Ammonia Solution 166 WADC TR 59-207 vii

LIST OF TABLES (Concluded) No. Page XXXVI. Vibrational Frequencies and Assignments for the Borodeuteride Ion in Liquid Ammonia and Deuteroammonia Solutions 166 XXXVII. Application of the Teller-Redlich Product Rule to the Borohydride Ion 170 XXXVIII. Comparison of Observed and Calculated Fundamental Vibrational Frequencies of the Various Isotopic Species of the Borohydride Ion 171 XXXIX. Force Constants for the Borohydride Ion 172 XL. The Dielectric Properties of Tetraborane 174 XLI. The Index of Refraction of Tetraborane 174 WADC TR 59-207 viii

LIST OF FIGURES No. Page 1. The evolution of H2 from B2H6'2NH3 in diethyl ether slurry. The role of NH4C1. 12 2. Phosphorus pentoxide drying tube for HC1. 16 35 Vacuum line filtration apparatus. 17 4. Rate of hydrogen evolution in the system KBH4-H2BNH2-NH3. 21 5. Apparatus for the preparation of hydrogen iodide. 27 6. Molecular weights of the halide salts of the cation [H2B(NH3)2] 351 7. Electron density projectior for [H2B(NH3)2]C1 on 010 plane showing arrangement of boron, nitrogen, and chlorine atoms. 40 8. Electron density projection on (001). Contours as in Fig. 7. 41 9. The structure of [H2B(NH3)2]C1. 42 10. Interatomic distances and bond angles in [H2B(NH3)2]C1. 43 11. X-ray powder patterns for [H2B(NH3)2] [BH4] prepared by metathesis reactions. 54 12. The structure of H3NBH3. 60 13. a) Solution tube for dielectric constant. b) Special vacuum flask. c) Solution density flask. 69 14. ae/6x versus mole percent solute. 77 15. The Raman spectrum of H3BCO at -80~C. 91 16. The Raman spectrum of D3BCO at -80~C. 91 17. The rate of H2 evolution in the reaction between sodium and B4Hzoo2NH3. 104 18. Evolution of H2 from the reaction of B4H10,2NH3 with excess of 105 HC1. WADC TR 59-207 ix

LIST OF FIGURES (Concluded) No, Page 19, Ammonia pressure above an ether solution of tetraborane. 108 20. Composition of system B4Hl, —excess NH3 as a function of time. 109 21. Pressure composition diagramfor system tetraborane ammonia. 111 22. Apparatus for recrystallization of H3NB3H7. 129 23. Trial models of the ammonia-triborane molecule. The z-direction is vertical. 134 24. Electron-density distribution at y = 0 in the unit cell of the disordered, high-temperature structure. 136 25. The structure of the H3NB3H7 molecule. 139 26. The packing of molecules in the low-temperature crystal of H3NB3H7- 146 27. Apparatus for the preparation of A12C16. 154 28. Raman spectra of 10B and "lB deuteroboranes in the gas phase. 164 29. Raman spectrum of Na~1BH4 dissolved in liquid ND3 (1) or liquid 5 N0 I (t). 169 30. The B-H stretching region in the Raman spectrum of "lBHD. 169 31, The B-D stretching region in the Raman spectrum of llBD4 dissolved in liquid ND3. 169 WADC TR 59-207 x

ABSTRACT Arguments to support the structure [H2B(NH3)2]BH4 for the "diammoniate of diborane" were presented in WADC Technical Report 56-318. The following additional information relative to the "diammoniate of diborane" and related structures has been obtained. (a) A microcrystalline form of B2H6*2NH3 has been obtained and the powder pattern is tabulated for ready compound identification. (b) Relatively strong evidence against the structure [HB(NH3)3](BH4)2 has been accumulated. The so-called "diammoniate of diborane II" which was formerly assigned this structure has been shown to consist chiefly of the regular "diammoniate" with the structure [H2B(NH3)2]BH4. (c) Improved methods for the synthesis of pure salts of the form [H2B (NH3)2]X have been developed. (d) A complete single-crystal X-ray crystallographic study which established the structure of [H2B(NH3)2]Cl has been completed and the data confirm the above structure assignments in every detail. The structures of the bromide and iodide salts have also been deduced. (e) A nuclear magnetic resonance study of [HB(NH3)2]I can best be interpreted in terms of the accepted structure of the cation [H2B(NH3)Jd. (f) The "diammoniate of diborane," [H2B(NH3)2]BH4, has been prepared by metathesis, a fact which offers final and complete chemical support for the above structure. The mono-, di-, and trimethylamines have all been prepared as stable compounds, and vapor-pressure measurements, dipole-moment measurements, and Raman spectra have been obtained for the series. The data represent part of a systematic examination of the amine-boranes, Phosphorus trifluoride-borane has been examined by means of Raman spectroscopy. Preliminary studies of the reaction between F3PBH3 and NH3 indicate the existence of the new compound, (NH2)3 PBH3. The Reman spectrum of carbon monoxide —borane is considered. A strong similarity has been found between the chemistry of diborane and that of tetraborane. The new compounds B4HIo-2NH3 and H3NB3H7 have been prepared. On the basis of chemical evidence the structure of B4Hlo 2NH3 is written as [H2B(NH3)2][B3H8]. A complete single-crystal X-ray study has established the detailed structure of H3NB3H7. All attempts to prepare F3PB3H7 have been unsuccessful. WADC TR 59-207 xi

The new compound C13AlPF3 has been prepared and characterized. The compound H3A1[N(CH3)21 described by Wiberg has been confirmed independently and structural studies on the molecule have been initiated. A detailed Raman spectral study of B2H6 has been completed and a detailed spectroscopic study of several borohydrides is described along with force constant calculations. The dipole moment of tetraborane has been measured as 0.6D. The bridge hydrogens of decaborane display acidic character on the basis of deuteron exchange studies. WADC TR 59-207 xii

I. REACTIONS AND STRUCTURES OF THE AMMONIA ADDITION COMPOUND OF DIBORANE AND OF ITS DERIVATIVES A. Background 1. EARLY STRUCTURAL ARGUMENTS In most of its reactions with electron donor molecules, diborane, B2H6, appears to cleave symmetrically to give characteristic base-borane adducts of the type: Base: BH3 For example, Burg and Schlesinger(l) found that trimethylamine reacts smoothly with B2H6 to give (CH3)3NBH3, and CO reacts smoothly under pressure to form the compound carbon monoxide-borane. The foregoing authors used processes of the above type as support for the postulate that the reactions of diborane proceed through a BH3 intermediate. Extensive experimental support for the postulate was provided by the researches of Schlesinger, Burg, their students, and other workers in the field. Indeed a large body of present-day evidence supports the postulated role of the BH3 broup,(2) and McCoy and Bauer(3) have estimated the energy relationships involved in the postulated borane-diborane equilibrium as indicated below: AH = 28.5 kcal/mole lo B2H6 U — 2BH3 AS = 34.2 e.u. at 3000 Although most reactions between diborane and electron donor molecules can be explained easily and logically in terms of the hypothetical borane intermediate, several boron hydride products have posed real interpretive problems. For example, one might expect that the reaction between diborane and ammonia would give the simple adduct ammonia-borane, H3NBH3, yet Stock and Pohland(4) (1) A. B. Burg and H. I. Schlesinger, JO Am. Chem. Soc. 59, 780-7 (1937). (2) R. Wo Parry and L. J, Edwards, Systematics in the Chemistry of the Boron Hydrides, Paper No. 54, the Division of Inorganic Chemistry Meeting of the American Chemical Society, San Francisco, April, 1958; J. Am. Chem. Soc., in press. (5) R. E. McCoy and S. H. Bauer, J. Am. Chem. Soc. 78, 2061 (1956). (4) A. Stock and E. Pohland, Ber. 58, 657 (1925)9 WADC TR 59-207 1

found that the reaction did not follow the expected pattern. Instead, the product was a salt-like solid whose molecular weight in liquid ammonia corresponded to that of a dimer. The dimeric solid has been called "the diammoniate of diborane." In order to rationalize this unexpected dimeric formula, Stock(5) assumed that two of the hydrogen atoms of diborane are acidic in character. Such an assumption then leads to the representation of the "diammoniate" as an ammonium salt, (NH4) 2 [B2H4 ] Since Stock's formulation of the "diammoniate of diborane" was in conflict with the borane reaction hypothesis of Burg and Schlesinger, these authors reinvestigated the ammonia-diborane addition product. In the resulting study(6) it was found that only one equivalent of hydrogen gas was liberated when sodium was allowed to react with B2H6*2NH3 in liquid ammonia at -77~C. It was justly held that such behavior was inconsistent with the model of Stock, since a molecule containing two ammonium ions would be expected to liberate two equivalents of hydrogen in the sodium reaction. The equation suggested by Stock's formula is: 2. (NH4)2[B2H4] + 2Na ~- H2 + Na2B2H4 Since only one equivalent of hydrogen or 1/2 mole was obtained, a new formulation, consistent with the sodium stoichiometry and the borane reaction hypothesis, was proposed; the structure was (NH4)(H3BNH2BH3). Additional evidence bearing on the question of acid hydrogens in the diborane molecule was presented by Burg. 7) Nyman(8) and his coworkers showed that when the BrbonstedLowry acid NH4C1 is dissolved in liquid ammonia, a rapid interchange of protons occurs between the acid cation and the solvent molecule. Burg conducted experiments with B2H6-2NH3 in liquid ND3 which showed that H-D interchange occurs only with the hydrogens attached to the nitrogen and not with those attached to boron in B2H6-2NH3. He interpreted his results as proof of the assumption that the hydrogens of diborane have no acidic character. If the hydrogens of diborane are not acidic in character, the model of Stock for the diammoniate becomes untenable. On the basis of the foregoing evidence, the model of Schlesinger and Burg (i.e., NH4[H3BNH2BH3]) was widely accepted for many years, particularly in the standard textbooks of the English-speaking countries. (9) (5) A. Stock, Hydrides of Boron and Silicon, Cornell Univ. Press, Ithaca, 1933. (6) H. I. Schlesinger and A. B. Burg, J. Am. Chem. Soc. 60, 290 (1938). (7) A. B. Burg, J. Am. Chem. Soc. 69, 747 (1947). (8) C. J. Nyman, S. C. Fung, and H. W. Dodger, J. Am. Chem. Soc. 72, 1033 (1950). (9) a) N. V8 Sidgwick, The Chemical Elements and Their Compounds, Oxford Univ. Press, London, 1950; b) To Moeller, Inorganic Chemistry, J. Wiley and Sons, N. Y., 1952; c) L. F. Audrieth and JL Kleinberg, Non-Aqueous Solvents, John Wiley and Sons, N. Y., 1953; d) D. R. Stranks and R. G. Wilkins, Chemical Reviews, 57, 743 (1957). WADC TR 59-207 2

An alternative proposal by Agromonov(10) received equally wide acceptance in continental Europe, particularly in Germany in the laboratories of Wiberg and his students. It was suggested that B2H6-2NH3 is indeed a monomer, H3NBH3, but that its salt-like character and dimeric nature in liquid ammonia are due to dipole-dipole interaction and hence to systematic errors in molecular weight measurements in liquid ammonia solution. The assumption that H3NBH3 would be a solid of salt-like appearance because of its dipole moment has indeed been verified by subsequent isolation of the solid monomer, and the assumption that dipole-dipole interaction could give rise to apparent dimerization in solvents of low dielectric constant found support in the observations of Bright and Fernelius(ll) on the molecular weights of boron trifluoride-amine addition compounds in benzene. It was observed that trimethylamine-boron trifluoride was 81% dimerized in an.08 molal solution in benzene, and phenyldimethylamine boron trifluoride was likewise 91% dimerized in a benzene solution which was.011 molal in solute. It has been shown more recently, however, that, because of the higher dielectric constant of liquid ammonia, H3NBH3 is not dimeric in that solvent, even at relatively high concentrations (1.0 molal).(12) Since the enigma of the B2H6-2NE{3 structure could only be resolved by additional investigation, intensive chemical work was carried on in a number of laboratories in the United States. Attempts to resolve the question by direct X-ray diffraction studies were unsuccessful because the material of the correct composition was always obtained in an amorphous condition. Although Stock reported that the "diammoniate" crystallizes out of a liquid ammonia solution in long thin needles, it was found in several American laboratories that the crystalline solid obtained from liquid ammonia appeared to be a solvate since removal of the ammonia to obtain the solid of formula 32H6*2NH3 invariably led to decrepitation of the crystals and formation of an amorphous powder. It is significant, however, that X-ray methods did provide the final indirect evidence for resolution of the structural question. George Schaeffer and his students(13) at the University of St. Louis examined the X-ray diffraction pattern of the residue produced by the reaction between B2H6-2NH3 and Na in liquid ammonia. They found that the pattern contained the characteristic lines of NaBH4. Schultz, Shore, and Parry, working in the Michigan laboratories, were told by Dr. Schaeffer of the NaBH4 identification and the Michigan group immediately proposed the structure [H2B(NH3)2][BH4] for B2H6'2NH3o Such a model had been mentioned brief(10) a) L. J. Agromonov, Jo Chim. gen. 9 (71), 1389 (1939); Chem. Zbl. 1941 I, 2362; b) E. Wiberg, A. Boltz, and P. Buckheit, Z. anorg. Chem. 256, 287, footnote p 307 (1948); c) Gmelin's Handbuch der anorganischen Chemie, 8 Auflage, System 13, pp. 100-101 (1954). (11) J. R. Bright and W. C. Fernelius, J. Am. Chem. Soc. 65, 735 (1943)(12) R. W. Parry, Go Kodama, and DI. R. Schultz, J. Am. Chem. Soc 80, 26 (1958). (13) G. W. Schaeffer, M. D. Adams, and F. J. Koenig, So Jo., J. Am. Chem. Soc. 78, 725 ( 1956). WADC TR 59-207 3

ly by Burg(7) in 1947 but had been rejected by him without serious examination. Schaeffer, Adams, and Koenig(13) were reluctant to accept the unprecedented cation (H2B(NH3)2]+, but proposed instead a structure containing both the ammonium ion and the borohydride group, NH+(H2BNH2)BH4. Since Schultz(14) had shown that NH4BH4 is unstable, it was suggested by the St. Louis group that H2BNH2 groups stabilized the ammonium borohydride structure by an unspecified mechanism. In summary, five widely recognized structural proposals for the compound B2H6'2NH3 have been made. The formulas suggested are: (1) (NH4)2(B2H4), (2) NH4[II3BNH2BH3], (3) H3NBH3, (4) [NH4][H2BNH2][BH4], and (5) [H2B(NH3)2] BH4. Evidence useful in making a choice between the above possibilities will now be reviewed. Since the detailed arguments have all been presented in earlier reports to WADC from this laboratory and in a recently published series of papers,(l5) only the major arguments and conclusions will be considered. 2. EXPERIMENTAL STRUCTURAL ARGUMENTS FOR B2H6_2NH3 (a) The Models [NH4]2[B2H4] and H3NBH3 Strong evidence against Stock's original model, [(NH4)]2B2H4, was provided by Burg's exchange experiments involving ND3 and B2H6'2NH3. The data indicated no acid hydrogens in diborane and thus were in direct conflict with Stock's model which assumed two acid hydrogens in the original B2H6 molecule. On the other hand, Burg's conditions were mild and his quantitative methods for determining the extent of exchange could be questioned since they were based on small differences in the vapor pressures of ND3 and NH3. A more rigorous and unequivocal argument against assuming acid character for the hydrogens of B2H6 was contained in the tracer study of Shore, Girardot, and Parry.(15f) If Stock's model were correct, the reaction of B2D6 and NH3 should produce;[NH3D]2 [B2D4] Subsequent reaction of this salt with sodium should then produce some HD or D2 in the evolved gas. Detailed tracer studies showed conclusively that all hydrogen liberated in the sodium reaction came from the rupture of hydrogennitrogen bonds, not from the breaking of hydrogen-boron bonds. In view of this (14) D. R. Schultz, Ph.D. dissertation, Univ. of Mich., 1954. (15) a) R. W. Parry, D. R. Schultz, and P. R. Girardot, Jo Am. Chem. Soc. 80, 1 (1958); b) D. R. Schultz and R. W. Parry, J. Am. Chem. Soc. 80, 4 (1958); c) So Go Shore and R. W. Parry, ibid. 80, 8 (1958); d) S. G. Shore and R. W. Parry, ibid. 80, 12 (1958); e) R. W. Parry and S. G. Shore, ibid. 80, 15 (1958); f) S. Go Shore, P. R. Girardot, and R. W. Parry, ibid. 80, 20 (1958); g) R. W. Parry, G. Kodama, and D. R. Schultz, ibid. 80, 24 (1958); h) R. C. Taylor, D. Ro Schultz, and A. R. Emery, ibid. 80, 27 ( 1958). WADC TR 59-207 4

evidence and Burg's earlier results, the model (NH4)2B2H4 became untenable. The ammonia-borane model, H3NBH3, for B2H6.2NH3 was eliminated unequivocally by Shore and Parry(15c) who synthesized an authentic sample of monomeric H3NBH3 by means of the reaction between, NH4Cl and B2H,6.2NH3 in ether slurry. It was clearly established that B2H6R2NH3 and H3NBH3 are different chemical species and that no labile equilibrium exists between them at the low temperatures found in the usual liquid ammonia solutions. (b) Evidence for The Borohydride Ion in B2H6.2NH3 Although the production of NaBH4 in the reaction between B2H6.2NH3 and Na originally suggested the existence of the borohydride ion in the diammoniate of diborane, it did not offer unequivocal proof of the B2H6.2NH3 structure since the possibility of a slight rearrangement of the solid product to give NaBH4 and H2NBH2 had not been eliminated. If such a rearrangement of solid products were postulated,(l3) the X-ray evidence could be regarded as consistent with any of the models for B2H6.2NH3 outlined earlier. Clearly, more definitive evidence was needed. Schultz and Parry(l5b) found that magnesium thiocyanate in liquid ammonia reacted with the diammoniate of diborane to give a precipitate of [Mg(NH3)6] (BH4)2. This test for the borohydride group is comparable to procedures normally used in qualitative analysis for anion identification; for example, S04= is identified by precipitation as BaS04. Further, it was found that solid NaBH4 in intimate mixture with authentic ammonium salts such as NH4C1 or NH4Br reacted to give hydrogen gas and a new microcrystalline solid phase containing boron, ammonia, hydridic hydrogen, and the halide anion. The equation indicated by the data is: MBH4 + NH4X - MX + [H2B(NH3)2]X + 2H2 solid phase It was found that B2H16*2NSE3 reacts with ammonium salts under comparable conditions to give comparable products thus suggesting the presence of the borohydride ion in the solid B2H16.2NH3. Furthermore, Shore and Parry(15c) showed that the reactions of ether slurries of B2H6.2NH3 with ammonium salts, lithium halides, and aluminum chloride paralleled completely the reactions under comparable conditions of authentic borohydrides such as NaBH4 or LiBH4. Additional chemical evidence supporting the existence of borohydride in the diammoniate was contained in an early observation of Stock(l6) who reported that solid B2H6.2N1L3 reacted slowly at -60'C with gaseous HC1 to give significant quantities of B2H6. Although B2H6 was produced in relatively good yield, Stock (16) Ao Stock, Hydrides of Boron and Silicon (Cornell Univ. Press, Ithaca, 1933), p9 134. WADC TR 59-207 5

considered that it was a product of a secondary reaction. He represented what he regarded as the main reaction by the equation: B2H6e2NH3 + 2HC1 -0 B2H4C12,2NH3 + 2H2. The nature of the secondary reaction which supposedly gave B2H6 was not clearly delineated in Stock's work. Modern information about borohydrides suggests representation of the process as MBH4 + HC1 —-- MC1 + 1/2 B2H6 + H2 where M is any cation. A complicating side reaction which would reduce the yield of B2H6 would be: 6B2H6 + 6HC1.6B2H5C1 L 2BC13 + 5B2H6 Finally, physical evidence supporting the existence of the borohydride ion in the original ammonia solution of B2H6'2NH3 was provided by Taylor, Schultz, and Emery. 15g) The Raman spectrum of B2H62NH3 in liquid ammonia solution at -781C revealed unequivocal evidence for the borohydride group. (c) Evidence Related to the Identity of the Cation in B2H6.2NH3 Three of the five formulas proposed for B2H6.2NHT3 contained one or more ammonium ions per mole of diammoniate; a fourth formula was the now discredited monomeric model [H3BNH3], and the fifth model contained the new and unrecognized cation [H2B(NH3)2]t In view of these facts, the evidence concerning the nature of the cation, and particularly the evidence supporting the existence of the widely accepted ammonium ion deserve careful examination. The original model containing an ammonium ion, (NH4)2B2H4, was written by Stock to rationalize his observations on the electrolysis of a liquid ammonia solution of B2H6-2NH3. It was found that a monomeric compound H3NB(CH3)3 which would be analogous to the monomeric H3NBH3, had a conductivity in liquid ammonia solution which was less than 1/100th of that of B2H6.2NH3. On the basis of this evidence an ionic structure was indicated, but any ionic model could be made to agree with the observations on conductivity. To support his position further, Stock attempted to identify the electrolysis products and account for them in terms of his molecular model, but his interpretations were by no means unique and unequivocal. The observations have been re-interpreted in a variety of ways. The principal evidence used by Schlesinger and Burg(6) to justify the use of an ammonium ion in their model, NH4(H313NH2BH3) was the reaction between WADC TR 59-207 6

B2H6'2NH3 and sodium in liquid ammonia. The equation assumed was: Na + NH4[H3BNH2BH3] liq J3. 1/2 H2 + NFI3 + Na[H3BNH2BH3] -78;0C Although the sodium reaction is indeed consistent with the presence of an ammonium ion in the compound, it is important to note that the reaction is difinitely not proof of the existence of an ammonium ion in the solid "diammoniate." Rather it is evidence for the presence of any Bronsted-Lowry acid which can oxidize the active sodium metal in liquid ammonia solution. For example, (CH3)3 BNH3(17), F3BNH3(18), and monomeric H3BNH3(19) all react slowly with sodium in liquid ammonia, yet contain no NH+ ions in the solid state. One might also compare the reactions of acids such as [Al(H20)+ ] in water systems. The acid character of A12(SO4)3-12H20 is not evidence for H30+ ions in solid aluminum sulfate hydrate, but is instead evidence for the existence of water molecules whose proton donor qualities have been enhanced by coordination to an ion of relatively high field strength. Similar arguments may be applied to ammonia molecules coordinated to the boron of [H2B(NH3)2]+. The reaction with sodium may be written: 1iq [H2B(NH]21X + Na i > NaX + H2BNH2 + 1/2 H2 + NH3 NH3 See pages 11 and 52 for further details. No direct evidence for an amnmonium ion in the "diammoniate of diborane" has ever been presented. On the other hand, strong evidence against the presence of this ion is now available. For example, it is known that NaBH4 in intimate mixture with authentic ammonium salts such as NH4C1 or (NH4)2S04 reacts rapidly to give off H2,(15b) yet Schultz and Parry(l5b) showed that the "diammoniate of diborane" does not react with excess NaBH4 under comparable conditions. On the other hand, B2H6*2NH3 will react with authentic ammonium salts because of the borohydride ion in the structure. If one were to assume a model for B2H6-2NH3 containing NH+ ions, the above observations could be rationalized only by assuming that something in the solid (H2NBH2 has been suggested) renders the NH+ ion in the solid "diammoniate" inactive toward BH- even though NH+ ion, when added as FI4Cln reacts vigorously with the BH' ion of the same structure. The mechanism of such stabilization without formation of the cation [H2B(NH3)+] remains obscure. (17) a) C.E. Coates, Organo Metallic Compounds (Methuen and Co. Ltd., London, 1956), p. 59; b) C. A. Krause, J. Chem. Ed. 29, 548 (1952); c) J. E. Smith and C. A. Krause, J. Am. Chem. Soc. 73, 2751 (1951). (18) W. J. McDowell and C. W. Keenan, J. Am. Chem. Soc. 78, 2065 (1956). (19) S. G. Shore and R. W. Parry, J. Am. Chem. Soc. 80, 8 (1958). WADC TR 59-207 7

In contrast, evidence for the complex cation [H2B(NH3)2+] is provided by isolation of its chloride and bromide salts. The two compounds are isostructural and are not a mixture of H2BNH2 and the corresponding NH4X salt. The bromide salt reacts with sodium in liquid ammonia to give hydrogen as does the "diammoniate of diborane," and the salts containing the complex cation should be electrolytes in liquid ammonia as Stock reported for B2H6"2NH3. Finally, Taylor, Schultz, and Emery(l5h) found that the Raman spectrum of the salt [H2B(NH3)2]Br plus the Raman spectrum of NaBH4 gives the Raman spectrum of B2H6 e2NH3o In summary, all evidence cited for the existence of the NH+ ion is equally good for the complex ion [H2B(NH3)+], and several discriminatory physical and chemical tests argue against the NH+ ion but are consistent with the ion [H2B(NH3)2.. Schultz and Parry(15bl reviewed all earlier arguments cited for the structure [NH+4[H3BNH2BH3] and found that every observation cited by earlier workers was consistent with the model [H2B(NH3)2](BH4). B. Further Studies on "The Diammoniate of Diborane," [H2B(NH3)21[BH4], and Related Molecules 1. SYNTHESIS OF B2H6-2NH3 IN ONE GRAM-LOTS In the past all attempts to synthesize as much as 1 g of B2H6*2NH3 in a single batch resulted in decomposition of the product. Schultz,(14) working in this laboratory, found that attempts to scale up the usual synthetic procedure resulted in the production of volatile products; B2HsNH2 was probably a major component of the reaction mixture. Temperature control was a major problem. A successful synthesis of 0.6 g of B2H6-2NH3 in one run has been effected using the procedure described below. A 10-millimole quantity of diborane was condensed at the bottom of a reactor tube, 30 mm in diameter and about 230 mm long. An excess quantity of ammonia was condensed along the walls of the tube by progressively raising a Dewar filled with liquid nitrogen from a height of about 50 mm above the condensed diborane to 50 mm from the top of the tube. A pentane bath at -145'C was placed around the tube. The thermostating pentane bath and the reaction tube warmed up gradually from -145~C to -80~C over a period of about 12 hours. The diborane vaporized and reacted with the ammonia~ The diammoniate was isolated by subliming away the excess ammonia at -78~C. The total yield of B2H6.2NH3 was about 0.6 g. The product was identified by its X-ray powder pattern. The first standard X-ray powder pattern of B2H6*2NH3 was obtained as described in the next section. In another attempt to prepare a relatively large quantity of "diammoniate" the reaction between diborane and ammonia was carried out in diethyl ether at -95~0 A mixture of the "diammoniate" and ammonia-borane resulted. IdentificaWADC TR 59-207 8

tion was by X-ray powder pattern. 2. THE PREPARATION OF MICROCRYSTALLINE B2H6.2NH3-THE X-RAY POWDER PATTERN OF "THE DIAMMONIATE OF DIBORANE" (a) Discussion As indicated in the literature review of Section A, all attempts to prepare crystalline B2H6-2NH3 for X-ray structural investigations had been unsuccessful in this and other laboratories. During the course of the study described in this report, a product with the empirical formula [H3NBH3]n was obtained which gave a definite and characteristic X-ray powder pattern. This product was originally prepared by a procedure which supposedly gave "diammoniate of diborane II," [HB(NH3)31(BH4)2.(l5e) However, repetition of the experiment by a procedure known to give an authentic sample of the "diammoniate" [H2B(NH3)2] [BH4] gave a product with an identical x-ray diffraction pattern. These facts cast serious doubt on the existence of diammoniate (II) [see page 10 for further information], but made possible identification of B2H6.2NH3 by X-ray methods. Why the crystalline product is easily obtained now, but was unobtainable earlier, remains in the realm of speculation. (b) Experimental (1) Diammoniate of Diborane (II).-Diborane was added to frozen, excess ammonia in the same manner employed to prepare the "Diammoniate I:'(15e) Then the system was aged for about 20 hours at -450 to cause conversion to "diammoniate (II)." Excess ammonia was then pumped from the system atabout -650C. Final traces of ammonia were pumped away at room temperature. The resulting solid product had the empirical formula H3NBH3. The following analytical ratios were obtained: B/N = 102; hydridic H/N = 3.02. X-ray powder data for the microcrystalline product are presented in Table I. (2) Diammoniate of Diborane (I). —The solid was prepared according to standard procedures(15e) and solvent ammonia was removed as completely as possible at -78~C. The product, presumably B2H6'2NH3 (I), was then warmed to room temperature, transferred rapidly in a dry box to an X-ray capillary tube, and a powder pattern was taken. The pattern obtained was identical to that in Table I except for four very weak lines at d = 6.05, 5.18, 4.09, and 3.63. WADC TR 59-207 9

TABLE I INTERPLANAR SPACINGS FOR THE CRYSTALLINE FORM OF THE "DIAMMONIATE OF DIBORANE" B2H6-2NH3 Intensity 0 d Intensity Q d W 5.78 7.62 MS- 19.40 2.322 W 8.18 5.42 W 21.03 2.149 MS 9.53 4.64 MW 21.55 2.099 S 10.35 4.28 W 22.75 1.979 S 12.65 3.52 MW 23.20 1.959 MS 15.18 2.94 W 25.40 1.798 MS 15.73 2.84 W 25.98 1.76z MS 16.63 2.69 W 26.35 1.737 MS- 17.23 2.60 W 26.98 1.701 MW 17.83 2.518 MS- 27.43 1.673 MW 18.65 2.413 Intensity scale: S>MS>MS->MW>W>VW CuK, radiation 3. ADDITIONAL CHEMISTRY OF B2H6-2NH3. EVIDENCE FOR AND AGAINST THE EXISTENCE OF "DIAMMONIATE OF DIBORANE II" WITH THE FORMULA [HB(NH3)31[BH4]2 (a) Decomposition of B2H6.2NH3 in Diethyl Ether Slurry In an earlier report(20) the preparation of monomeric H3NBH3 from [H2B(NH3)2] BH4 was described in terms of the following equation. NH4C + [H2B(NH3)2[BH4]Trace Diethy3 H2 + H3NBH3 + [H2B(NH3)2]C1 Diethyl ether The validity of this overall equation was questioned when it was found that the solid reaction residues, examined by X-ray diffraction, showed only small amounts of the salt [H2B(NH3)2]C1. The explanation for this discrepancy between theory and experiment arises from properties of [H2B(NH3)2]BH4. It is now known that the "diammoniate" undergoes a competing decomposition process in an ether slurry containing ammonia but no ammonium chloride: (20) R. W. Parry, R. C. Taylor, T. C. Bissot, D. H. Campbell, A. E. Emery, PO R. Girardot, G. Kodama, D. R. Schultz, S. G. Shore, J. T. Yoke II, WADC Tech. Report 56-318 (1956). WADC TR 59-207 10

[H2B(NH3)2] BH4 Diethylce 3 H + H3NBH3 + (H2BNH2)n ether The reaction is not unexpected in view of the known weak acidic character of the [H2B(NH3)2+] cation in liquid ammonia and the hydridic character of the hydrogens in the borohydride group. Data are summarized in Fig. 1. The two reactions written above proceeding simultaneously, would be consistent with all observed facts. In a typical run 1.73mmoles of [H2B(NH3)21[BH4] were prepared and isolated in one arm of a vacuum line filter system of the type described by Schultz(l4,l15a) [see Fig. 3o] About 10 ml of diethyl ether was then distilled into the reaction tube. About 0.15 mmole of rigorously dried ammonia was also distilled in; then the evacuated system was closed off and the reactor was warmed up to about 5 or 4 degrees below room temperature. A beaker of water just below room temperature served as a thermostating medium. This procedure prevented condensation of ether throughout the system and minimized attack on the stopcock grease by ether. The extent of reaction was determined at various intervals by quenching the reaction tube in liquid nitrogen and measuring volumetrically the quantity of hydrogen produced. Upon completion of the reaction, the contents of the reactor tube were filtered and extracted with ether. The filtration and extraction were carried out at about -75~ to take advantage of the negative temperature coefficient of solubility of ammonia-borane. Crystalline ammonia-borane was obtained from the filtrate by distilling away the ether as the receiver tube was warmed from -70 to -20~Co (b) The Reaction of "The Diammoniate of Diborane" with Sodium in Liquid Ammonia In all earlier work considerable structural and chemical significance has been attached to the stoichiometry and rate of the reaction between sodium and B2H6-2NH3 in liquid ammonia. For example, the original "diammoniate" formula of Schlesinger and Burg [NH4][H3BNH2BH3] was based primarily upon the stoichiometry of the sodium reaction, and Parry and Shore(l5e) used the stoichiometry of this same reaction as evidence for the suggestion that there exists a second form of B2H6-2NtR3 called the "diammoniate of diborane (II)" with the structure [HB(NH3)3][BH4]2o It was suggested that if the normal "diammoniate of diborane (I)," [H2B(NH3)21 [BH4], which was prepared and maintained below -78~C, is warmed to -45OC and stored at this temperature for about 20 hours, diammoniate (II) is produced. It is assumed that diammoniate (I) in liquid ammonia reacts with sodium to give one equivalent of H2 per mole of B2H6-2NH3 in accordance with the following equation: [H2B(Nf3)2] BH4 + Na.-.-. 1/2 H2 + NaBH4 + H2BNH2 + NH3 WADC TR 59-207 11

1.2 [H2B (NH3)2 (BH4)+ NH4C I [H2B( NH3)2](BH 4 0.8 1.. ZH 0.4 0.2 Small amount of NH3 odded 0 20 40 60 80 100 HOURS Fig. 1. The evolution of H2 from B2H6.2NH3 in diethyl ether slurry. The role of NH4C1. WADC TR 59-207 12

Diammoniate (II) was assumed to react to give the observed 1.33 equivalents of H2 per formula weight of B2H6-2NH3. The equation written was: [HB(NH3)3][BH4]2 + 2Na -- 2NaBH4 + H2 + HB(NH2)2 + NH3 Note: [HB(NH3)3][BH4]2 is equal to 1.5 formula weights of B2H6-2NH3. Thus' 2 equiv He = 1:33 equiv H2/B2H6.2NH3 1,5 formula wts Earlier in Section B it was reported that a sample of B2H6e2NH3 was prepared and stored under conditions assumed to give diammoniate (II) and a second sample of B2H6-2N1H3 was prepared carefully under conditions known to give diammoniate (I). Strangely enough, the X-ray powder patterns of both samples were identical except for four very weak lines in the diammoniate (I) pattern. To characterize further the presumed diammoniate (II), a portion of the sample from the batch used in X-ray analysis was dissolved' in liquid ammonia and allowed to react with excess sodium in liquid ammonia at -78~C. In about one hour 1.28 equivalents of H2 per formula weight of B2H6-2NH3 were obtained. The value is consistent with behavior attributed to diammoniate (II), but a molecular weight determination in liquid ammonia on this same material gave a value of about 55 which corresponds to diammoniate (I) [theoretical value = 611. A reasonable interpretation of the foregoing facts is that diammoniate (I) and diammoniate (II) are the same material, but that other factors such as traces of decomposition products or other as yet undefined variables in the higher-temperature product influence the rate and stoichiometry of hydrogen evolution in the sodium reaction to such an extent that the structural significance of the data is definitely limited. Such an interpretation is supported by data on the HC1 reaction described next, and by observations on the reaction of [H2B(NH3)2]C1 with sodium (p. 52) and of [H2B(NH3)2][B3H8] with sodium (po 114). The assumed importance of impurities in catalyzing hydrogen evolution in boron nitrogen systems has been demonstrated by Campbell, Bissot and Parry 21) and by Kodama and Parry,(22) and is amply supported in other sections of this report [see p. 591. One other alternate possibility cannot be eliminated on the basis of evidence now available. In view of the extreme sensitivity of the properties of the so-called "diammoniate I" to the methods used for its preparation, it is possible that this low-temperature material is in reality a mixture consisting of [H2B(NH3)2J]BH4 and a postulated intermediate(l5e) containing a single bridged (21) D. H. Campbell, T. C. Bissot and R. W. Parry, J. Am. Chem. Soc. 80, 1549 (1958); WADC Tech. Report 56-318 (1956). (22) G. Kodama and R. W. Parry, WADC Tech. Report 57-11 (January, 1957) p. 21; JO Am. Chemo Soc. 79, 1007 (1957); see also discussion in this report on B4Hlo 2NH30 WADC TR 59-207 13

bond, H H H H NH3ff On the basis of a detailed study of the sodium reaction, Shore, Girardot, and Parry suggested that the single-bridge structure would react slowly with sodium in liquid ammonia. Thus any appreciable quantity of the intermediate would reduce the rate of hydrogen evolution in the sodium reaction. A corollary to the above argument is that pure salts of the type [H2B(NH3)2]X would liberate one equivalent of hydrogen very rapidly in the sodium reaction and additional hydrogen more slowly in a secondary process. Studies reported on pages5 and 114 for the salts [H2B(NH3)2]C1 and [H2B(NH3)2][B3H8] support this argument. Additional information relative to the postulated forms I and II of B2H6.2NH3 is found in the next section. (c) The Reaction of "The Diammoniate of Diborane" with Hydrogen Chloride In his early studies of the "diammoniate of diborane" Stock(5) investigated the reaction between B2H6'2NH3 and HC1. He wrote the equation: B2H6*2NH3 + 2HC1 -- B2H4C122NH3 + 2H2 I but he reported the formation of B2H6 by an unspecified secondary reaction. Schultz and ParryL5b) reinterpreted Stock's observations and wrote the equation between the "dianmgoniate of diborane (t)" and HC1 as: [H2B(NH3)21[BH4] + HC1 >- [H2B(NHS)2]C1 + H2 + 1/2 B2H6 The analogous equation expected for the diammoniate of diborane (II) as suggested by Parry and Shore(15e) would be: [HB(NH3)31(BH4)2 + 2HC1- [HB(NH3)3]Cl2 + B2H6 + 2H2 A complicating side reaction between the B2H6 formed as a product and HC1 present in the reacting system would be: 6B2H6 + 6HC1 -B6B2H5C + 6H2 K 2BC13 + 5B2H6 To evaluate Stock's proposal more effectively and to differentiate more effectively between the two proposed forms of the dia-nmoniate Li.e., [H2B(NH3)21 (BH4) and [HB(NH3)3][BH4]2J, the reaction between B2H6.2NH3 and HC2 has been studied under rather well-controlled conditions. From the foregoing equations for the reaction between the diammoniate of diborane and HC1, it is apparent that the amount of hydrogen produced in the WADC TR 59-207 14

process will be complicated by the secondary reaction between B2H6 and HC1. Although the HCl-B2H6 reaction can be slowed up by holding the temperature at -780C, it is still difficult to attach great structural significance to the amount of hydrogen obtained in a given period of time. Of greater structural significance is the identity of the solid product which remains after gas evolution. Three possibilities must be recognized. Stock wrote B2H4C12'2NH3. Such a product would have a B/N/C1 ratio of 1/1/1. "Diammoniate of diborane (I)" as described by Schultz and Parry would produce [H2B(NH3)2]C1 which has a B/N/C1 ratio of 1/2/1, while "Diammoniate of diborane (II)" as postulated by Parry and Shore(l5e) would produce [HB(NH3)3]C12 which has a B/N/Cl ratio of 1/3/2. When a slurry of "diammoniate of diborane (I)" in dimethyl ether was allowed to react with an excess of HC1 at -78~C, the solid product was unequivocally identified as [H2B(NH3)2]C1 on the basis of its X-ray powder pattern. Some NH4C1 contaminated the solid. When a slurry of "the diammoniate of diborane (II)" in dimethyl ether was allowed to react with HC1 under conditions comparable to those used above, the solid product was again [H2B(NH3)2]C1 contaminated with small amounts of NH4C1; it was not [HB(NH3)3]C12 as expected. These data, coupled with the X-ray powder patterns, lead almost unequivocally to the conclusion that diammoniate of diborane (I) and diammoniate of diborane (II) have the same major phase. Differences between the two samples in the sodium reaction or molecular weight values are attributed to small amounts of impurities in one sample or the other. Two possibilities remain and neither can be eliminated conclusively on the basis of available evidence. The possibilities are: (a) the -780 product, B2H6'2NH3 (I), is contaminated by products representing incomplete reaction between B2H6 and NH3, or (b) the -45~ product, formerly B2H6.2NH3 (II), is contaminated by small quantities of decomposition products which give rise to instability of the compound. The problem now resolves itself into a question of compound purification. If [H2B(NH3)2][BH4] could be obtained in high purity, its utilization in the preparation of high purity [H2B(NH3)2]C1 becomes immediately apparent. Studies directed toward this objective are described under the synthesis of [H2B(NH3)2]C1 on page 23. The reactions referred to above were conducted as described below. Samples of "diammoniate of diborane (I)" and "diammoniate of diborane (II)" were prepared by previously described procedures.(l5e) Commercial ethers, used as solvents, were dried over LiAlH4 before use. Commercial anhydrous hydrogen chloride was purified by freezing it into one arm of the purification system shown in Fig. 2. By cooling the second arm of the purification system, the gas could be slowly passed through the P205 to dry it. Six passes through the dessicant were made; HC1 was then condensed into the reaction vessel for use. In a typical run a suitable sample of B2H6.2NH3 (about 1 to 2 millimoles) was placed in a 24/40 reaction tube on a vacuum filter apparatus (Fig. 3). For studies on diammoniate (I) it was desirable to make the B2H6o2NH3 right in the WADC TR 59-207 15

MEDIUM GLASS FRIT I f _ i 29/40 P2 05 - -GLASS WOOL PLUGS 20cm, Fig. 2. Phosphorus pentoxide drying tube for HC1.

RECEIVER TUBE 25 mm PYREX GLASS FRIT, MEDIUM POROSITY \ 24/40 DRIP-TIP TO MERCURY BUBBLER TO J 14/35 " 24/40 DRIP TIP -i-SOL NOI D REACTOR TUBE, 25 mm PYREX Fig. 3. Vacuum line filtration apparatus. WADC TR 59-207 17

reaction tube so that it could be maintained at temperatures below -78~C until reaction was effected. For studies directed toward the synthesis of pure [H2B(NH3)2]C1, larger samples of diammoniate of diborane were prepared; then smaller aliquots were transferred to the reaction tube for use. Approximately 3 ml of anhydrous dimethyl ether was condensed onto the solid B2H62NH3, and this mixture was stirred with the conventional magnetically operated "hopper stirrer" until a uniform suspension was obtained. Then the desired quantity of previously dried HCl, ranging from an equivalent amount to a 5-fold excess, was dissolved in 1 ml of dimethyl ether and the solution was condensed onto the suspension. Hydrogen was evolved rapidly at first, then very slowly. In earlier runs where hydrogen evolution was measured, the reaction mixture was frozen down with liquid nitrogen and H2 was pumped from the reaction vessel for measurement. When hydrogen evolution amounted to less than 0.1 millimole/hour, the reaction mixture was filtered at -78~C. The washed precipitate was extracted with ammonia and the solvent NH3 was sublimed away. leaving a white residue. The residue was identified as [H2B(NH3)2]Cl which was contaminated by small amounts of NH4C1. X-ray powder patterns served as the basis for identification. The X-ray capillary was loaded in a dry box and sealed at both ends to prevent hydrolysis of the sample during a 10- to 12-hour exposure period. CuKa radiation was used. Representative data are contained in Table II. TABLE II THE REACTION OF B2H6-2NH3 WITH HC1 IN DIMETHYL ETHER EVIDENCE AGAINST THE EXISTENCE OF DIAMMONIATE (II) OR [HB(NH3)3 ](BH4)2 Vol H2 Evolved Sample mmoles mmoles (CH3)20 Time mmoles H2 Product Diammoniate II B2H6 HC1 Solvent B2H6 Prepared 2.16 9 9 3-5 ml 2 hr 1.18 IH2B(NH3)2C1 in situ + NH4C1 as 5 hr 1.33 Impurity Prepared 4.82 7.9 3-5 ml 1.5 hr 1.08 jH2B(NH3)2C1 in situ + NH4C1 as 5 hr 1 23 Impurity All reactions at -78~C 4. STUDIES ON THE AMMONOLYSIS OF THE RESIDUES RESULTING FROM THE REACTION OF AN ALKALI METAL WITH "THE DIAMMONIATE OF DIBORANE" IN LIQUID AMMONIA (a) Background Reports by Schaeffer, Adams, and Koenig(3) and independent observations in this laboratory(20) indicated that the solid residues from the alkali metalWADC TR 59-207 18

diammoniate reaction (these are MBH4 and H2NBH2) reacted with NH3 at room temperature and below to give off H2. A tracer study(15f) showed that such hydroger;n arises from the interaction of a hydridic hydrogen attached to boron and a protonic hydrogen attached to nitrogen. The equation assigned to the process in this and other laboratories was: H2BNH2 + NH3 - IHB(NH2)2 + H2 NH3 L > B(NH2)3 + H2 Although this equation appeared to be consistent with all earlier data on the ammonolysis reaction, it appears questionable when critically compared with analogous systems. In the first place, the ready reaction between the very slightly acidic ammonia and the residue (H2NBH2)n seems inconsistent with the properties expected for (H2NBH2)n, particularly when it is compared with the formally analogous compound (N2BNHCH3)3. The latter compound will float on a slightly acidic water solution at room temperature with only very slow hydrolysis. Complete hydrolysis is effected by a relatively long exposure to a boiling 20% hydrochloric acid solution. On the other hand, the equation written above postulates a reaction between H2NBH2 and the weakly acidic NH3 molecule at temperatures below 250C. The contrast in properties assumed for (H2NBH2)n and (CH3HNBH2)3 seems remarkable. Other observations are equally disturbing. Schaeffer, Adams, and Koenig(l3) reported that the presence of an alkali metal borohydride with [H2NBH2]n could promote hydrqgen evolution. LiBH4 was very effective while KBH4 was relatively ineffective; NaBH4 occupied an intermediate position. It was suggested that an explanation for the differences might be based upon the relative abilities of LiBH4 and KBH4 to form stable ammoniates and hold ammonia in the reaction zone as the temperature of the system was raised. The foregoing uncertainties in observation and interpretation suggested that a careful experimental reexamination of the system might be appropriate. (b) Experimental A mixture of H2NIBH2 and KBH4 was prepared in liquid ammonia by allowing H2~B(NH3)2][BH4] to react with a slight excess of potassium in about 2 cc of liquid ammonia. A sample containing 1,846 millimoles of B2H6.2NH3 was prepared; 0.922 millimole of H2 was evolved in 75 minutes; then gas evolution stopped. After reaction was complete, all excess potassium was removed quantitatively by amalgamating it with about 0.7 cc of liquid mercury at -350C. The solution was filtered to separate the alkali metal amalgam; then the solvent was removed from the filtrate at -78~Co The mixture of KBH4 and H2NBH2 was transferred under nitrogen to a dry box where it could be weighed. In a separate reaction vessel a sample (1e09 millimoles) of commercial LiBH4 was attached to the vacuum line and dissolved in 1 cc of liquid ammonia. WADC TR 59-207 19

Hydrogen, which evolved very slowly, was measured. After 24 hours a sample mixture consisting of.07 millimole of KBH4 and.07 millimole of H2NBH2 was weighed in the dry box and introduced into the system under dry nitrogen. The entire mass was dissolved in liquid ammonia; then the solvent was distilled away and the system was allowed to warm to room temperature in contact with gaseous ammonia (1 atm). Data in Fig. 4 show that hydrogen evolution was extremely slow under these conditions, less than.018 mmole of H2 being evolved from.07 mmole of H2NBH2 in 18 hours at room temperature. While dry nitrogen was allowed to flow through the system, a glass bulb containing a slight excess of sodium was broken into the reaction vessel. The reactor was then evacuated and the system frozen with liquid nitrogen. Ammonia was distilled into the reactor and the entire reaction mass was completely dissolved in liquid ammonia at -780C. The solvent was then removed at -78~C; then the solid residue was warmed up to room temperature in the presence of one atmosphere of ammonia vapor. As shown in Fig. 4, hydrogen evolution was rapid at room temperature as soon as sodium was added. Even in the presence of sodium, gas evolution was slow at -78~C. The data indicate unequivocally that sodium metal rather than MBH4 was effective in promoting gas evolution at room temperature. (c) Discussion If, as suggested by Schaeffer, Adams and Koenig, the reaction between H2BNH2 and ammonia were catalyzed by the presence of a metal borohydride which served to hold the ammonia in the reacting mass, gas evolution should have been rapid when H2BNH2, KBH4, NH3 and excess LiBH4 were present in the same vessel at room temperature. The fact that hydrogen evolution was slow under these conditions can be taken as strong evidence against the above explanation. On the other hand, the role of the alkali metal in hydrogen evolution seems well established by the foregoing data, particularly when one notes that in the work of Schaeffer, Adams, and Koenig excess alkali metal was not removed prior to ammonolysis reaction. With this point in mind, one notes that the earlier data are consistent with the somewhat more detailed studies described here. When excess alkali metal was rigorously removed from the system by mercury extraction, N2BNH2 and LiBH4 reacted very slowly with gaseous ammonia at room temperature. Addition of sodium metal accelerated the reaction in a dramatic fashion. On the basis of the above experimental evidence, the following speculative explanation may be advanced. It is postulated that the free alkali metal in the system reacts with H2NBH2 in accordance with the equation shown below to give a compound comparable to the "base" Na2[BH(CH3)2] described by Burg and Campbell: (23) 2Na + H2BNH2 - Na2(H2BNH2) (23) A. B. Burg and G. W. Campbell, Jr., J. Am. Chem. Soc. 74, 3744 (1952). WADC TR 59-207 20

, I 52 LiBH4%J NH3' BH2NH2 (AND KBH4).o{ z AND EXCESS No NH NH AND 12..,J ~ I I 0.103 |ROOM TEMP o.09w0.08 -LL -780C' BH2NH2 (AND KBH4) W I7C I IH/BH2NH2=2.05 J. 03 Na ADDED _.02._1- 350C~-7C-350C _78 o c 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 84 88 HR Fig. 4. Rate of hydrogen evolution in the system KBH4-H2BNH2-NH3.

Burg and Campbell found that the lone hydrogen bound to boron in their compound was extremely hydridic in character. Some evidence was accumulated which can be interpreted as indicating ammonolysis of Na2[HB(CH3)2] at temperatures as low as -500~C In view of these observations ammonolysis of the related compound Na2[H2BNH2] at room temperature would not be surprising. The following ammonolysis equations can be written: RT. Na2(H23NH2) + NH3 — Na2[HB(NH2)2] + H2 The above process would produce 1 molecule of H2 per molecule of BH2NH2 present (or 2H/BH2NH2); the formula suggests that a change in hydrogen evolution rate might be observed when this step is completed. Such a change in rate of hydrogen evolution is noted where the ratio of H2/H2BNH2 is equal to 1 (or H/H2BNH2 = 2; Fig. 4). The next step in the process would be: Na2[HB(NH2)2] + NH3 R.T Na2[B(NH2)3] + H2 On the basis of the above equation a maximum of 2H2 molecules per molecule of H2BNH2 should be obtained or 2-1/2 H2 molecules per original molecule of B2H6o2NH3. Schultz(14) observed as much as 2o15 H2 molecules per original molecule of B2H6.2NHs, and the data in Fig. 4 indicate that slow H2 evolution is continuing even after the total ratio of H2/B2H6-2NH3 exceeds 1.5 [or H/H2BNH2 = 3 (Fig. 4)]. If the compound Na2[B(NH2)3] were unstable at room temperature and dissociated in accordance with the equation Na2[B(NH2)31 > 2Na + B(NH2)3 a small amount of alkali metal could well be catalytic for hydrogen evolution. Alternatively, sodium could be consumed in the reaction. More data are needed on this point. If one recalls that LiBH4 is very proton-sensitive, NaBH4 less so, and KBH4 stable even in water solution, it is not difficult to rationalize Schaeffer's observation regarding the ammonolysis rates for the residues resulting from the reaction of Li, Na, or K with B2H6-2NH3. The relative proton sensitivity of Li2(H2BNH2), Na2(H2BNH2), and K2(H2BNH2) would decrease quite rapidly from Li to K; hence the rate of the ammonolysis reaction would be expected to fall sharply as was reported. (13) The above arguments are definitely speculative; isolation of pure compounds of the form M2BH3,, M2BH2TH22 M2BH(N1H2)2, and M2B(NH2)3 could provide a more definite test for the process proposed. WADC TR 59-207 22

C. Halide Salts of the General Form [H2B(NH3)2]X 1. PREPARATION OF SALTS CONTAINING THE COMPLEX CATION [H2B(NH3)2 ]+ (a) The Solid Phase Reaction of B2H6.2NH3 and NH+ Salts The first halide salt containing the complex cation [H2B(NH3)21+ was prepared by Schultz(14) using the solid phase reaction between ammonium bromide and the diammoniate of diborane, [H2B(NH3)2]BH4. The equation is [H2B(NH3)2] [BH4] + 2NH4Br -;- 2[H2B(NH3)2]Br + 2H2 A mixture of [H2B(NH3)2]BH4 and NH4Br was completely dissolved in about 5 cc of liquid ammonia and held for 2 hours with continuous stirring at low temperatures ranging from -78.5 to -35 (-63.5 was used in most runs). Under these conditions hydrogen evolution was negligible. Solvent ammonia was then sublimed from the solid mixture and the temperature was raised slowly to that of the room. Hydrogen was removed with a Toepler pump; the residue was redissolved in ammonia; after stirring ammonia was again removed; the system was warmed up and H2 was measured. This cycle was repeated 3 to 5 times. The product prepared by the above procedure was pure enough to permit original identification of the salt, but not as pure as the product obtained by the method of Kodama(24) described in the last procedure of this section. (b) The Reaction of B2H6-2NH3 with HX in Ethers In an earlier section the reaction occurring in dimethyl ether between B2H6*2NH3 and HC1 was described in some detail. The product always showed NH4C1 contamination. A number of other ether solvents have been investigated in an effort to prepare [H2B(NH3)2]C1, free of NH4C1. A suspension of B2H622NH3 in diethyl ether reacted too slowly with HC1 at -780~C to provide a basis for the synthesis of [H2B(NH3)2]C1. Use of Ansul 121 polyether (1, 2 dimethoxyethane), dried over LiAlH4 at room temperature, required a slight modification of the previously described procedure. Because of the higher melting point of the polyether, the reaction was carried out at -63OC (chloroform slush) until hydrogen evolution ceased (about 2 hours). Under these conditions diborane produced in the primary reaction was attacked readily by HC1, and BCl3tetherate contaminated the product. The polyether was removed by distillation at -45~C; then diethyl ether was condensed into the crude salt mixture and stirred at -63~C. The contaminating Et20BC13 was removed in the filtrate; the precipitate was washed with excess Et20 to remove last traces of Et20BC13; then [H2B(NH3)2]C1 was extracted from (24) G. Kodama, Ph.D. dissertation, Univ. of Mich. (1957). WADC TR 59-207 23

the washed precipitate using liquid ammonia in the usual manner. [H2B(NH3)2]C1 prepared in Ansul 121 showed no lines for NH4C1 in the X-ray powder pattern. The same procedure was used with Ansul 141 (diethyleneglycol dimethyl ether) except that ether had to be distilled from the reaction mixture in the vacuum system at room temperature instead of at -450C. Data are contained in Table III. (c) The Reaction of B4H1O.2NH3 and Acids in Diethyl Ether By methods to be described in a subsequent section, Kodama(24) in this laboratory synthesized the compound B4Hlo-2NH3. Chemical evidence cited in a later section led unequivocally to the structure [H2B(NH3)2](B3H8) for this compound. A reaction at -78oC between an acid HX and the ether soluble compound B4H1022NH3 in diethyl ether gives the ether insoluble salt [H2B(NH3)2]X. The equation is: Et2O [H2B(NH3)2] [B3H8] + HX - >f [H2B(NH3)2]X4 + H2 + Et2OB3H7 -78o Kodama(24) prepared pure [H2B(NH3)2]Cl and [H2B(NH3)2]Br by this procedure. Heitsch(25) working on this contract prepared and characterized a relatively pure iodide salt. Yamauchi(26), working in this laboratory, prepared a salt which has been tentatively identified as the fluoride [H2B(NH3)2]F, although its complete characterization has been complicated by difficult analytical problems. Procedures which have given satisfactory products for each halide salt are described below. (1) [H2B(NH3)2]C1. — In a typical run a 1.84-millimole sample of B4H10o2NH3 was placed in the reaction tube of a vacuum filtration assembly and dissolved in about 2 ml of diethyl ether. The solution was frozen with liquid nitrogen and a 1.78-millimole sample of HC1 was condensed into the tube. When the temperature was raised from -196 to -780~C, very rapid evolution of hydrogen gas and formation of a white precipitate were observed. In 15 minutes, 1.74 millimoles of H2 gas was liberated; then gas evolution almost stopped. Even if an excess of HC1 were used, H2 gas evolution was relatively slow (see Fig. 17) after 1 mole of H2 had been evolved from 1 mole of B4H1o.2NH3. The white precipitate was separated by filtration in the vacuum line filter system (Fig 3). The filter cake was washed with diethyl ether; then the receiver was changed and the precipitate was washed through the filter disc with liquid ammonia. Removal of the solvent ammonia at -450C left 119 mg of the compound [H2B(NH3)2]C1 or a recovered yield of 79%, based on the B4Ho102NH3 used. (25) C. Heitsch, doctoral dissertation, Univ. of Mich. To be published. (26) M. Yamauchi, doctoral dissertation, Univ. of Mich. To be published. WADC TR 59-207 24

TABLE III THE PREPARATION OF [H2B(NH 3)2jl FROM B2H6&2NH3 AND HC1 IN VARIOUS ETHERS Vol and H2 Evolved Sample mmoles mmoles Evolved Purity Solid B2H6s2NH3 B2H612NH3 HC1 Identity Time, mmoles H2 H2B(NH3)2C1 of Ether min. B2H6.2NH3 1. Prepared 1.97* 10.35 (CH3)20 30 0.86 Traces of NH4C1 in situ (5 ml) (X-ray) 60 0. 89 (99.5%) analysis 2. Taken from 1.9* 10.41 (CH3)20 40 1.15 Traces of NH4C1 stock (5 ml) 60 1.20 impurity from 135 1.36 X-ray 3. In situ 1.7 10.0 (CH3)20 50 1.0 Trace of NH4C1 (5 ml) 60 1.06 X-ray 90 1o 12 4. In situ 1.7 4 Ansul 121 45 0.64 No NH4C1l but 60 0.75 other impurities 90 1.10 present (X-ray) 11 hr 1.91 5. In situ 2.06 3.4 (CH3)20 15 0.87 Trace of NH4C1 (5 ml) 75 0.95 570 1.07 6. Stock** 1.29 2.1 Ansul 121 120 1.78 (4 ml) 7. Stock** 1.37 2.3 Ansul 141 120 1.83 Pure sample (4 ml) No NH4C1 8. Stock** 6.81 8.9 Ansul 121 135 1.13 (15 ml) *Small sample removed for X-ray powder pattern 0.1 mmole t*Reaction run at -63~C. Others at -78~ WADC TR 59-207 25

(2) [H2B(NH3)2]Br.-A 0.64-millimole sample of B4H10o2NH3 was allowed to react at -78~C with 3.83 millimoles of HBr in about 7 ml of diethyl ether. The procedure was identical to that given above. The hydrogen gas amounted to 0.64 millimole. After filtering the solid and washing it with 2 cc of diethyl ether, it was kept under high vacuum for one hour, then washed through the filter dise with liquid ammonia. Most solvent was removed at -78~C but the last traces were pulled off under vacuum at room temperature. The solid product was 0.50 millimole of [H2B(NH3)2]Br or a recovered yield of 79% based on the B4H10o2NH3 used. (3) [H2B(NH3)2]I.-Hydrogen iodide was prepared by the reaction: Pt H,? + I2 -60 > 2HI The procedure used was a modification of that described in Inorganic Synthesis, Vol. 1, p. 159. The apparatus is shown in Fig. 5. A is a catalytic deoxygenator attached to the pressure reduction valve on the hydrogen tank. B is a Pyrex tube filled with P205; D is a Vycore reaction tube. A layer of glass wool near the inlet of the reaction tube was covered with (a) a layer of iodine crystals (Mallinkrodt Analytical Reagent), (b) a second layer of glass wool, (c) a layer of platinized asbestos, and (d) a final layer of glass wool. Joints at a, b, and g were closed with T ygon tubing. Joints d and f were sealed with de Khotinsky cement. Joint e and stopcock g were lubricated with Dow-Corning High Vacuum Silicone grease. The furnace, E, was raised so that it surrounded the platinized asbestos but was above the iodine layer. The temperature of the furnace was then raised to 2000C and the platinized asbestos was dried in a hydrogen stream for one hour. During this period a stream of cold air was directed against the outside of that part of the tube which contained the iodine. This reduced sublimation of iodine during the drying of the asbestos. After the drying operation the furnace was dropped down around the iodine layer, its temperature was raised to 600'C, and the exit gas stream was diverted through the cold trap H (-78~) and into a trap on the vacuum system chilled with liquid N2 (-1960C). The product from the -196~ trap was sublimed through a trap cooled to -78~C and into a trap cooled to -196. The sublimation was repeated. The purified material showed no color due to iodine and had a vapor pressure of 170 mm at -63.50C [chloroform slush]. Stull(27) reports a value of 165 mm at this temperature. The product was stored at -1900C until used. The reaction between Hi and B4H0lo2NH3 was carried out in the same fashion as that described for the reactions of B4H10'2NH3 with HC1 and HBr. The stoichiometry of hydrogen evolution was not as clean-cut as that in the two preceding cases, apparently because of the instability of HI in high-vacuum manipulation. In a typical run about 1 gram of crude B4H1o'2NH3 (see p. 101 for its synthesis) was leached with dry diethyl ether (about 10 cc) through a medium porosity frit into a weighed tube on the vacuum line. The solvent ether was then removed and the tube was filled with dry nitrogen and stopperedo The tube and pure crystals (27) Do Ro Stull, ind. and Eng. Chem. 39, 517 (1947). WADC TR 59-207 26

TO VACUUM LINE E- HINGEDTUBULAR OVENH2 (FROM TANK) J - ABSOLUTE MERCURY MANOMETER G1m 1 B f A F A- CATALYTIC DEOXYGENATOR B - PHOROUSPENTOXIDE DRYING TUBE C - PINCH CLAMP d D - VYCOR REACTION VESSEL E - HINGED TUBULAR OVEN M F-'AIR CONDENSOR G-STOPCOCK O!Di * H- COLD TRAP' e J - ABSOLUTE MERCURY MANOMETER B M- METER FOR THERMOCOUPLE SCALE: / E 4" E PLATINIZED AS BESTOS IODINE WAJDC TR 59-207 27

were carefully weighed. The weight corresponded to 10.1 millimoles of B4Hio 2NH3. The tube was replaced on the vacuum filtration apparatus and about 10 cc of dry diethyl ether was distilled into the system; then a mixture of 11.5 millimoles of HI and 5 cc of ether was distilled into the system. The mixture was then warmed from -190~ to -78~C. Stirring was started as soon as the reaction mass became sufficiently fluid. Hydrogen evolution proceeded for about 2 hours. Then the rate of gas discharge dropped rapidly to almost zero. The mixture was filtered; the ether insoluble [H2B(NH3)2Iwas washed with more ether and finally leached into a clean tube with anhydrous liquid ammonia. A total sample consisting of 1.52 grams of salt was recovered from the tube representing a yield of 87% based on the B4H1o. 2NH3 used. The X-ray powder patterns showed weak lines for NH4Io (4) [H2B(NH3)2]F or Perhaps [H2B(NH3)2[BF4]. -A 2.12-millimole sample of [H2B(NH3)2](B3H8) was dissolved in 3 ml of diethyl ether in a translucent teflon tube which could be attached to the copper vacuum line. An excess of HF (about 3 millimoles), dried over commercial anhydrous CoF3, was condensed about an inch above the solution and this was allowed to warm up slowly to -780C with constant stirring. Rapid evolution of H2 (identified by molecular weight measurements)took place. The amount of H2 evolved could not be measured in this system. The ether solvent was removed at -780C and the white residue was transferred under dry nitrogen to the vacuum filter system. The residue was washed with toluene, and diethyl ether then extracted with dimethyl ether. The solid extracted by Me20 has been difficult to characterize by conventional techniques because of analytical difficulties. The best evidence to date suggests that it is [H2B(NH3)2]F or [H2B(NH3)2](BF4). (d) General Observations on the Synthesis of Salts of the Cation [H2B(NH3Y2]+ One of the most common contaminants of [H2B(NH3)2]X salts is the corresponding ammonium halide, NH4X. In general it seems that the more soluble the original diamnmoniate of diborane or tetraborane in the suspending solvent the less the contamination of the resulting product. For example, both bromide and chloride salts, prepared by the solid phase reaction between B2H6.2NH3 and NH4X, are contaminated with ammonium salts. B2H6~2NH3 has very low solubility in diethyl. ether; the reaction with HC1 in this solvent was too slow for practical use and the product obtained was of low purity. Dimethyl ether is a somewhat better solvent for B2H6-2NH3, and the resulting chloride salt, [H2B(NH3)2]C1, was fairly good but always contaminated with NH4C1. Finally, B2H6'2NH3 is fairly soluble in Ansul 121 and 141 and the resulting product was always free of ammonium salts. These generalizations seem to be applicable as well to the reactions using B4HIo-2NH3 as a starting material since B4HEo02NH3 is soluble in diethyl ether and the products obtained from this reaction system were always of relatively high purity. 2. CHARACTERIZATION OF THE HALIDE SALTS OF THE [H2B(NH3)2]+ CATION Analytical data and molecular weight values have been obtained for all the halide salts described in the preceding section. Analytical results are sumWADC TR 59-207 28

marized in Table IV and analytical procedures are described below. TABLE IV SUMMARY OF ANALYTICAL DATA FOR THE CHARACTERIZATION OF THE [H2B(NH3)2]X SALTS Salt %B ~ Hydridic H.oN o Halide Mol. Wt. Solvent Obso 6.15 1.09 19.50 170 Ammonia [H2B(NH3)2]I Theor. 6.22 1.16 19.59 173.8 Obs. 8.46 1.59 22.15 62.5 Br- 120 Ammonia Theor, 8053 1.59 22.10 63o0 Br- 127 Obs. 13.2 2)42 33-9 42.7 C1- 95 Ammonia [H2B(NH3)2]Cl Theor. 13.1 2.44 340o 430O C1- 82.3 Obs. * * * * 90-120 [H2B(NH3)2]F above 200 (Me)20 Theor. 16.4 3.05 42.5 28.9 F- 65.8 *Problems of obtaining reliable analysis for B, N, H in presence of F not yet resolved. Identity of sample still very uncertain. (a) Analysis for Hydridic Hydrogen The sample was sealed with acid (usually 6N HC1) into a tube equipped with a break-off tip. The tube and contents were then heated in a 1500~C oil bath for a period of about 3 days. The hydrogen gas was measured in the vacuum system by use of a Toepler pump and a gas burette. (b) Analysis for Boron A modification of the so-called identical pH method proposed by Foote(28) was used. The method permits the titration of boron in the presence of ammonium and substituted ammonium salts. The reliability of the procedure was considered by Bissot(29) and the magnitude of errors established. The method consists of adjusting the pH of the solution to a value of 6.7, then adding an excess of reagent such as mannitol to complex the boron, and titrating the liberated H+ with.05N NaOH. A Beckman, Model F pH meter with a glass electrode was used to follow the pH values. (28) F. J. Foote, Ind. Eng. Chem., Anal. Ed., 4, 39 (1932). (29) To Ho Bissot, Ph.D. dissertation, Univ. of Mich. (1955). WADC TR 59-207 29

(c) Analysis for Nitrogen Nitrogen was determined using standard micro-Kjeldahl methods.(30) The Beckman pH meter described above was used for titrations. Solutions of 0.01. N H2S04 and.05 N NaOH were used. (d) Analysis for Halogens Bromide and chloride were determined as silver halides by standard gravimetric procedures. No satisfactory procedure for the determination of fluoride in these compounds has yet been obtained. (e) Molecular Weight Measurements The procedure used has been described in detail elsewhere.(31) Occasional changes in thermostating liquids were demanded for some systems. The apparent molecular weight is shown as a function of concentration for each of the salts in Fig. 6, (f) Comments on the "Fluoride" Salt Since all results on the assumed fluoride salt are unsatisfactory at the present time, its characterization must be considered incomplete. The strongest argument supporting its identity is its use in a metathesis reaction to prepare [H2B(NH3)2]BH4, Work on its complete characterization is Still in progress. (g) X-ray Powder Patterns See next section. 3. THE CRYSTAL STRUCTURE OF [H2B(NH3)2]+SALTS Although a large amount of chemical evidence has been amassed to support the existence of the cation [H2B(NH3)2]+, no independent and unequivocal physical evidence has yet been presented in support of the identity of this ion. The following independent structural information, provided by single-crystal and powder X-ray methods, leaves no room for reasonable doubt regarding the identity of the [H2B(NH3)2] cation. (30) J. B. Niederl and V. Niederl, Micromethods of Quantitative Organic Analysis, John Wiley and Sons, N. Y., 1948. (31) R. W. Parry, et al., Chemistry of the Boron Hydrides and Related Hydrides, WADC Tech. Report 56-318 (1956), p. 85. See also JO Am. Chem. Soc. 80, 24 (1958). WADC TR 59-207 30

\J1.._q ||MOLECULAR WEIGHT OF [H2B(NH)9CI (M.W=82.4) BY VAPOR PRESSURE LOWERING He IN LIQUID AMMONIA 110 0_9 0 82.4 0.4 ____ o THEORY z 05 FORMULA WEIGHTS/ 1000 NH3 Fig. 6a. Molecular weights of the halide salts of the cation [H2B(NH3)21.

155.4 127.-..... MOLECULAR WEIGHT OF [(NH3)2 BH2] Br(M.W=126.8) 123 BY VAPOR PRESSURE LOWERING __o'T' 119|IN LIQUID AMMONIA 9119 010 91__ 0 0.2 0.4 0.6 0.8 1.0 1.2 FORMULA WEIGHTS/10009g NH3 SOLVENT Fig. 6b. Molecular weights of the halide salts of the cation [H2B(NH3)21

1-3 \i J I Molecular Weight H2 B(NH3)2 I in Liquid NHj \10 170, 10 |Observed /70.0 | Colc. for Monomer /73.8 160 150 w: 140 <\ w 130 _1\ z o w 120 aa110 100 90 0 1.0 2.0 FORMULA WTS/1000 GMS SOLVENT Fig. 6c. Molecular weights of the halide salts of the cation [H2B(NH3)2]+.

(a) The Crystal Structure of [H2B(NH3)2]Cl as Determined by Single-Crystal X-ray Studies Nordman and Peters,(32) working in the University of Michigan laboratories, succeeded in the difficult task of preparing a single crystal of the chloride salt and working out an unequivocal structure for it by single-crystal X-ray methods. (1) Experimental.-Samples of [H2B(NH3)21C wreprepared according to methods previously described from B4H10o2NH3. The preparation of a single crystal suitable for X-ray work proved difficult since the compound is decomposed in solution by traces of moisture and by solvents containing protonic hydrogen. Acetone, diethyl ether, and hydrocarbons do not dissolve [H2B(NH3)"Cl, and crystallization from liquid ammonia yields a microcrystalline powder which gives a characteristic powder pattern, but no crystals large enough for singlecrystal examination. A few imperfect single crystals were obtained, however, by dissolving the compound in Ansul 141 (diethylene glycol dimethyl ether) and layering the solution with diethyl ether. These crystals were sealed in thinwalled glass capillaries for X-ray examination. In spite of repeated attempts, only one crystal suitable for single-crystal work was obtained, Zero-, first-, and second-level precession or Weissenberg patterns of the three principal nets were recorded using MoKa or CuKa radiation. This yielded a total of 186 independent intensities. The X-ray work was discontinued at this point due to the deterioration of the crystal. (1.1) Unit Cell and Space Group.-The diffraction patterns showed nearly tetragonal symmetry and the axes were labeled accordingly. On the nets perpendicular to the c axis, however, systematic intensity variations were found which violated the fourfold symmetry, and led to the conclusion that the crystals were orthorhombic with two axes, a and b, of equal length. It was further observed that the only systematic absences in the hkO net occurred when h and k were both odd. Since this is not a space group extinction, it was concluded that the crystal was a twin consisting of two fragments turned with respect to each other by 90~ about the c axis. The observed absences in hkO could then be explained as occurring for odd h in each of the fragments. Since the a and b axes were equal in length, all hkO reflections with h and k even were overlaps of, in general, nonequivalent reflections from the two fragments. This was not the case for reflections with k odd, in view of the systematic absences for odd h. Thus, hkO reflections with h + k odd were "pure" and afforded a means of determining the relative size, or diffractive power, of the two fragments, The ratios of the intensity of hkl, divided by the Lorentz and polarization factors, to that of khl was found to be very nearly constant for all hkO with h + k odd, as well as for corresponding upper-level reflections. This constituted strong evidence that the crystal was indeed a twin and yielded a value for the (32) Co Nordman and C. Peters, Jo Am, Chem. Soc., in press; see earlier reports, this contract. WADC TR 59-207 34

ratio of the sizes of the fragments. O O The unit cell parameters are a = 10.20 ~ 0.02 A, b = 10.20 ~ 0.02 A, c = 8.71 ~ 0.02 A, where the limits of error are taken as three times the estimated standard deviation. The calculated density is 1o21 g/cm3 assuming eight formula units per cell. Systematic absences, referred to the untwinned crystal, were observed in hkl for h + 1 odd, in Okl for k odd and in hol for h odd. The possible space - - 18 - - groups are then D2h-Bbcm and C 7-Bba2. The first of these has sixteen general positions and consequently requires at least the chloride and boron atoms to lie in special positions. In addition to the reflections leading to the above unit cell assignment, a set of very weak and somewhat diffuse reflections was observed calling for a doubling of the c axis. The true unit cell parameters are then a = 10.20 + 0.02 A, b = 10.20 ~ 0.02 A, c = 17.42 ~ 0.04 A. In view of the relatively small number of clearly observable reflections in this category, no attempt was made to determine the space group applicable to the doubled cell. The reflections demanding the doubled cell were ignored in the structure analysis. (1.2) Structure Determination.-Intensity data were obtained from the diffraction photographs by visual comparison with a series of timed exposures of spots of similar shape. Application of the Lorentz and polarization factors yielded a set of raw values of the squares of the structure factors. For every spot observed on two films, the ratios of these raw F2 values were calculated and averaged to give the ratio of the scale factors to be applied to the two films. Through appropriate averaging of such ratios the scale factors themselves were found and the raw F2 values placed on a common scale. Due to the twinning of the crystal. most of the values obtained were derived from spots which were overlaps of contributions from the two fragments, whereas the rest of the values arose from one or the other of the two fragments. From the nature of the twinning and the symmetry of the reciprocal lattice, it follows that a reflection hkl from one of the fragments (A) will appear as "khl" when produced by the other fragment (B). If h + k is odd, either the reflection hkl or khl produced by A must be absent, since all reflections with h + 1 odd are absent. For example, if h is even and k odd, a reflection with even. 1 indexed as "khl" must in fact be the hkl reflection from fragment B. Furthermore, it must be "pure" inasmuch as it cannot contain a contribution from an overlapping reflection from A. For the same reason the reflection indexed as hkl must be a pure reflection from fragment A. Twenty-seven independent reflections with h + k odd were observed. The ratios of the raw F2 values of these reflections hkl to the corresponding values of the reflections "khl" were found to be constant withintbeerror of measurement. The average value of this ratio, l/r, expresses the size ratio, A to B, of the two crystal fragments. The value of r was found to be 0.26. WADC TR 59-207 35

All reflections for which h + k is even are overlaps of contributions from both fragments. If the raw F2 values of these reflections are denoted I(hkl), we have I(hkl) = F2(hkl) + rF2(kh1) and I(khl) = F2(khl) + rF2(hkl) Consequently, the expression F2(hkl) = [I(hkl) - rI(khl)]/(l - r2) yields the magnitudes of the individual structure factors expressed on the scale of the large fragment. The magnitudes of the structure factors with' all -indices even are, on the average, considerably larger than the rest. This observation led to the conclusion that the chlorine atoms, which dominate the scattering, must lie at or near the points of a simple tetragonal lattice having one half the axial lengths of the assumed cell. The hkO and hOl electron density projections were therefore computed using structure factors with even indices only, with signs based on the chlorine contributions. The hOl projection showed the nitrogen and boron atoms to lie in layers perpendicular to the c axis half-way between the assumed chlorine layers. This arrangement is consistent with Bbcm, the more symmetrical of the two possible space groups, if the boron and nitrogen atoms lie in mirror planes at z = 0 and 1/2 and the chlorine atoms on twofold axes parallel to the c axis at (x,y) = (0,0), (0,1/2), (1/2,0) and (1/2,1/2). Since any violation of these conditions was, at most, very slight, the space group was tentatively assumed to be Bbcm. The eightfold chlorine positions are then (0,0,0; 1/2,0,1/2) + 0,0,z; 0,O,z; 1/2, 1/2,z; 1/2,1/2,7 where the parameter z is approximately 0.25. The nitrogen and boron atoms are located at eightfold positions of the type (0,0,0; 1/2,0,1/2) + x,y,O; x,y,O; x,1/2-y,1/2; x,l/2+y,1/2. Approximate x and y coordinates of the two nitrogen atoms and one boron atom were derived from the hkO electron density projection. The suggested configuration was an angular N-B-N group with two approximately equal B-N distances. The seven atomic coordinates, the scale factor k, and the isotropic temperature factor B were now refined by least-squares methods. Specifically, the quantity Zw(hkl) kFo (hkl) - Fc(hkl)2// w(hkl)k2Fg(hkl) WADC TR 59-207 36

was minimized, where the weight w(hkl) of each reflection was taken as 3 min/F (hkl) if F(hkl) > 3Fmin and equal to unity otherwise The least-squares refinement proceeded to R1 = ZYklF| - JFJ/Z<F| = 0.17 0 c 0 at which point several Fourier sections and (Fo - Fc) syntheses of the electron density were computed. The chief reason for this was to check for indications of lower symmetry than that of Bbcm. Since no such indications were found, it must be concluded that the space group is Bbcm, at least to the accuracy of this determination. Anisotropy was observed in the thermal motion of the chlorine atoms, the peaks being somewhat elongated in the z direction. An anistropic temperature factor exp - [a(h2 + k2) + B12] was therefore applied to fCland by successive difference syntheses the best values a = -0.0030 and 5 = +0.0020 were found. To account for the somewhat stronger thermal motion in the nitrogen atoms than in the boron atom, an additional temperature factor exp (-0.75 sin2/2) was applied to fN in calculating the structure factors. It was not possible to locate the hydrogen atoms independently; however, their presence was assumed and an approximate hydrogen contribution included in the calculation of structure factors. Two hydrogen atoms were assumed to be attached to the boron atom, and six half-hydrogens to each of the nitrogen atoms. The hydrogen scattering factors mere multiplied by an additional temperature factor exp [-1.0 sin2@/%2]. With these refinements, including several more cycles of least-squares adjustment of the seven atomic coordinates, scale and overall temperature factors were carried out. The final agreement of observed and calculated structure factors is given by R1 = 0.137 and R2 = (kF - F )2/ k2F2= 0o036 The observed and final calculated structure factors are listed in Table V. An additional 96 structure factors which were part of the photographed nets were found to be unobservably small. The final calculated values of these structure factors showed no significant disagreement with their observed maximum limits. When h + k is odd, the geometrical structure factor has the form Ajhkl) = -16 sin 2thx sin 2tkyo cos 2ilz Consequently, the chlorine atoms do not contribute to these structure factors. As a further check on the nitrogen and boron coordinates, a separate leastWAlDC TR 59-207 37

T ABLE V OBSERVED AND CALCULATED STRUCTURE FACTORS FOREH2 NHa:,B(3) 21l td kO \20 o hkl kF o F c hkl kF o F c hkl k.Fo F c hkl kF o F c hkl kF o F c hkl k 000 - +352 0.12.2 15 - 17 246 20 -24 408 22 + 21 620 63 + 57 880 7+1 002 22 + 37 113 34 + 28 248 13 + 14 410 17 - 20 62238-4 82 004 119 +126 1i5 28 - 18 2. 4.10o 8 - 5 412 14 - 14 624 32 + 36 8.lo1.0 20+1 006 18 - 3 117 2 3 + 18 250 19 - 18 420 75 + 79 626 20 - 20 8.10.212-1 008 36 + 32 119 19 - 10 252 15 - 16 422 5 7681 i81. 0.0.10 8 + 1 121 47 - 48 260 75 +'61 424 47 + 50 630 14 + 11 10.0.0 9+2 020 89 + 72 123 25 - 31 262 44 - 42 426 27 - 24.640 41 + 39 10.0.2 31-2 022 176 -130 125 lo - 16 264 37 + 38 428 12 + 14 642 5 9l~. 024 35 + 42 131 20 18 266 16 - 19 430 26 + 24 650 15 - 14 10.0. 613-1 026 39 - 40 141 19 - 21 268 14 + 12 432 15 + 19 652 13 - 13 10.2.024+2 028 13 + 9 161 23 - 24 280 54 + 50 440 72 + 76 660 68 + 47 10.2.224-2 AN0. 2. 10 14 - 10 171 16 - 12 282 21 - 24 442 5 3622 41.. co o~ 182+147 181 9 + 8 284 36 + 34 460 41 + 36 680 36 + 29 10.4.0o6+2 042 ~35 - 33 200 67 + 75 286 8 -.11 462 58 - 52 682 23 - 23 10.4.222-3 044 58 + 69 202 110 -128 288 12 + 11 480 27 + 29 6.10.0 30 + 22 10.6.0o9+2 0461 18 - 16 204 49 + 43 2.10.0 28 + 26 482 38 - 36 6,10.2 13 - 16 10.6.215-6 048 21 + 21 206 46 - 40 2.10.2 24 - 25 4.10.0 26 + 22 6.12.2 11 - 11 10.8.0 9+1 0.4.10' 6 - 2 208 15 + 9 21o1.4 lo + 18 4.10.2 2 4711 2l.. 060 24 + 26 2.0.10 8 - 9 2.10.6 10 - 12 4.12.0 9 + 15 751 11 -91.. 062 93 - 77 210 46 - 51 2.12:0 20 + 20 4.12.2 13-571lo+6 202 064 17 + 18 212 31 - 40 2.12.2 10-13515-178o6 37 220 066 22 - 33 220 77 + 57 2.14.2 14- 1533-13824 -4o 222 068 10 + 5- 222 134 -122 311 21-19517-7842 23 24o 00 41 -39 224 33 + 34 321 26 + 25' 561 11 - 12 806 22-19 242 o8, — 3 38 226 33 - 38 323 13 + 17 571 11 - lo 820 57 + 51 12.6.213-1 o84 2 + 25228 8 + 1 3319 - 8 600 30 + 31 822 19 - 23 12.8.29-9 o8, 1 - 8 2..10 7 - 7 341 22 + 24 602 73 - 72 824 29 + 35 14.2.211-1 0.10.0 26 + 2220 2 7+731 0.102 28- 2 30 2 + 2 3519 - 8 604 21 + 21 826 11 - 10 9 232 19 + 2 400 137 +149 606 37 - 31 840 30 + 27 0.10.4 13 + 16 240 90 + 82 402 3 2681 4 3-3 o.1o6 12- 13242 69 - 65 404 73 + 70 610 30 - 28 860 36 + 29 0.120 22+ 1 244 43 52 4o6 23 - 16 612 24 - 24 862 23 - 23

squares refinement of these parameters was carried out using the 27 observed structure factors with h + k odd only. This yielded coordinates very nearly equal to those obtained with the complete set of data, and a value of R1 = 0.092 for the set of 27 structure factors. The final atomic coordinates (x,y,z) are (0, 0, 0.264) for C1, (0.236, o.o64, 0) and (o.o60, 0.237, 0) for N1 and N2, and (0.212, 0.219, 0) for B. Estimated standard deviations of these values are 0.0015, 0.002 and 0.003, respectively, for chlorine, nitrogen, and boron coordinates. The reliability of these estimates, which are based on the least-squares residuals, is doubtful since the effect of a possible error in r is unaccounted for. (2) Discussion.-Electron density projections on (010) and (001) are shown in Figs~ 7 and 8 and a perspective drawing of the structure in Fig. 9. Interatomic distances and angles are given in Fig. 10. Each NH3 group is surrounded by four chlorine atoms at the corners of a distorted square while each chlorine is surrounded by eight NH3 groups belonging to six different [H2B(NH3)2]1 at the corners of a strongly distorted cube. The arrangement of the chlorine atoms and the N... C1 distances indicate that the compound is ionic, namely, [H2B(NH3)21+C1-. The four symmetratelynonequivalent distances may be compared with the N... C1 distances in ammonium chloride (3.36 A) and particularly methylammonium chloride(33) (3.18 A) where the coordination is more comparable to that found in [H2(NH3)2B]C1. The lengthening of this distance in the present case is to be expected in view of the presumably smaller charge on each of the NH3 groups in the (NH3)2BH2 ion. The lengths of the two N-B bonds in the cation are equal within experimental error, and also equal to the N-B bond length found in ammonia-triborane, NH3B3H7.(34) The complete X-ray structural analysis of [H2B(NH3)2]C1 leaves no question regarding the identity of the cation [H2B(NH3)21+. Essentially unequivocal evidence now supports the arguments advanced earlier from the chemistry of B2H6" 2NH3 (b) Powder Patterns and Structures for Other Halide Salts (1) Cell Geometry. -X-ray powder patterns of L[2B(NH3)2]I and[H2B(NH3)2]Br, produced by the Debye-Scherrer method, show obvious similarities in low-angle reflections with the low angle reflections for [H2B(NH3)2]C1. As noted in the preceding section, the structure of the chloride salt is known. By assuming that the three structures are isomorphous, it has been possible to index the thirtyone observed lines for the bromide salt and the sixty-six lines for the iodide (33) E. W. Hughes and W. N. Lipscomb, J. Am. Chem. Soc. 68, 1970 (1946). (34) Nordman and Reimann, this report, p. 143. WADC TR 59-207 39

O x > a 0 0 C Fig. 7. Electron density projections for [H2B(NH3)2]C1 on 010 plane showing arrangement of boron, nitrogen, and chlorine atoms.

::F C0 x, a a O \J1X I i b o o Fig. 8. Electron density projection on (001). Contours as in Fig. 7.

o 0 o~~~~~~~ o 0 00' OYQO a~~~~~F U).~ ~~ ~ ~~~ ~~~~~~~~~~~~~~~~~~~~~ \ X 0~~ 0 0' WADC TR 59-207 42

!x y 1 ( 0) R0) Io I X,'' (0 1060 Iv> / b Fig. 10. Interatomic distances and bond angles in HBNH C Fig. 10. Interatomic distances and bond angles in [H2B(NH3)2]C1.

salt. By calculating the powder pattern for the chloride salt from the singlecrystal data, it has been possible to determine hkl values for each reflection, and by analogy the hkl values for the corresponding bromide and iodide reflections.* Knowing hkl values, one can then calculate dimensions of the unit cell. It should be noted that while the bromide salt is apparently pseudotetragonal (actually orthorhombic with the a and b axes equal as in the chloride salt), the iodide salt is unambiguously orthorhombic with "b" larger than "a" using the same axis designation as used in the chloride salt. The patterns themselves do not provide an unequivocal choice between space groups Bbcm and Bba2; however, because of the striking similarities in the intensities of lines with the same index for all three salts, it is probably safe to assume that they are indeed isomorphous and all share the same space group Bbcm, determined by Nordman and Peters from their single-crystal study of the chloride salt. There are eight formula weights of each compound per unit cell. Unit-cell data are summarized in Table VI. TABLE VI COMPARISON OF LATTICE PARAMETERS FOR CHLORIDE,BROMIDE, AND IODIDE SALTS OF [H2B(NH3)21+ Density Apparent Vol.: ~a b c g/cc per Ion Pair; (A)3 [H2B(NH3)2]C1 10.20A 10.20A 8.71A 1o21 113o1 [H2B(NH3)2]Br lO.48~.02 10o48+.02 9.03~.02 1.70 12308 [H2B(NH3)2]I 10O75~0O2 11.01~.02 9.58~.02 2.04 141.9 (2) Procedure. -Data were fit by least-squares methods to minimize errors in cell parameters. Conventional data-reduction procedures were employed. One may define the following: d* = l/d, a* = 1/a, b* = l/b, c* = 1/c where d is the distance between reflecting planes of atoms and a, b, and c are the distances on the x, y, and z axes which define the unit cell. For orthorhombic cells (1) d = a*2h2 + b*2k2 + c*212 *In actual practice only the low-angle reflections of the chloride, bromide, and iodidce salts were sufficiently similar to provide unequivocal h, k, and 1 identification; however, identification of these lines provided tentative reciprocal lattice constants from which the complete indexing could be effected. WADC TR 59-207 44

where d* = the d value calculated from known h, k, 1 values and h, k, and 1 are defined as the Miller indices of the scattering plane. From the Bragg equation (2): d* 2 sin Q o where do = the d value observed from measurements on powder photographs. From the single-crystal data on the chloride salt, the values of a*, b*, and c* were known and d* could be calculated for each known hkl plane. The powder pattern for the chloride calculated by this procedure from the singlecrystal data checked the experimental powder pattern (Table VII), thus providing an unambiguous indexing for the chloride salt. It was then assumed that a line in the bromide pattern, corresponding to a given line in the chloride pattern would have the same h, k, 1 values as the chloride line. It was thus necessary to adjust values of a*, b*, and c* for the bromide and iodide salts until the calculated de values [Eq. (1)] showed the best agreement with the experimental bromide and iodide results obtained from the powder patterns [Eq. (2)]. If a*, b*, and c* are the approximate assumed reciprocal lattice constants from which the indices of each line are determined and Aa*, Ab*, and Ac* are the differences between the actual reciprocal lattice constant and the approximate a*, b*, or c* values then the accurate reciprocal cell parameters can be defined as a* + Aa*, b* + Ab*, and c* + Ac*. It is then desired to find Aa*, Ab*, and Ac*, all related to Ad* by the equation: Ad* = 2 (h2a*Aa* + h2b*Ab* + 12c*Ac*) (3) d c and chosen such that the function F is a minimum (based on least-squares procedures for fitting data). F = Z( d* -d* - Ad*)2 = min (4) By substituting relation (3) into (4), differentiating with respect to Aa*, Ab* and Ac* and equating to zero, one obtains: haF do-dc h2 - 2a*Aa*. h4 - 2b*Ab*. hk2 - 2c*Ac* h 6Aa* dC d 2 d2c __F d_-d h2k2 4k212 = d -d k2 - 2a*Aa*Z d - 2b*Ab* 2c*Ac* dAb* dc d2 d=2 dc 2 __ do-d_ h2 _ aa -2b*Ab* - 2c*Ac* dh2 0 WADC TR d- d7 42c WAIDC TR 59-207 45

TABLE VII OBSERVED AND CALCULATED hkl VALUES FOR [H2B(NH3)2C1 (Indexing of Powder Pattern) 4sin2% hkl __ d Intensity Obsetrved Calculated 200 0.0388 0.0384 5.08 ms 002 0.0524 0.0524 4.36 weak doublet 220 0.0772 0.0772 3.60 m 202 0.0912. 0908 3.31 vs 131 0.1104 0.1092 3.01 w 222 0.1304 0.1296 2.77 s 113 0.1380 0.0.384) 2.69 w 321 0.1380 0.1360j 2.69 w 400 0.1552 0.1536 2.54 s 420 0.1924 0.1920 2.28 ms 004 0.2104 0.2100 2.18 mw 422 0.2452 0.2448 2.02 ms 521 0.2920 0.2920 1.85 w 440 0.3088 0.3072 1.80 w 115 0.3500 0.3484 1.69 vw 404 0.3628 0.3548 1.66 m 620 0.3848 0.3848 1.61 w 602 0.4004 0.3988 1.58 m 622 o.4384 o0.4376 1.51 w WADC TR 59-207 46

These are three simultaneous equations in three unknowns, Aa*, Ab*, and Ac*. In the case of the bromide salt, since a* = b*, the system can be reduced to two equations in two unknowns. Approximate values of a*, b*, and c* were originally obtained in each case from the relation d* = a*2h2 + b*2k2 + c*212 By selecting three of the observed d* values and their accimpanying h, k, and 1 values, one could obtain three equations in three unknowns, the unknowns being the approximate a*, b*, and c* values for the lattice. These approximate methods gave very good results since the least-squares treatment indicated corrections of only.003A,.003A and.006A for a, b, and c of the iodide salt and a correction of only.002A for both a and c. A probable error of +0.02A in each of the cell parameters is suggested on the basis of the following analysis. The relation between a, b, and c and the observed d values can be expressed as: d2 = a2h2 + b2h2 + c212 Differentiating, we have: did = h2asa + k21bb + 12c6c The maximum value for the error in a, ba, can be found if one assumes bb and bc = 0. Then did H2a Since a = d and h2 > 1, we may say that ba' a/d 6d d d Md* The above assumes that bd/d 6 bd*/d*. The maximum error in a can then be computed as: 6a = a z|da dc ~~0~~~~n Thus a = +.012A for the iodide and +.014 for the bromide. The assigned error of ~.02A thus appears to be reasonably conservative. Tables VII, VIII, and IX give the observed and calculated hkl values for [H2B(NH3)2]C1, [H2B(NH3)21]Br, and [H2B(NH3)2II, respectively. WADC TR 59-207 47

TABLE VIII OBSERVED AND CALCULATED hkl VALUES FOR [HB(Pi43)2jBr 4si n2g hkl a_ k2 d Intens ity Observ4 Calcuatd 200 0.0365 0. 034 5.22 s 002. o486. 0490 4.51 w, 220 0.0727 0.0728 3.71 ms 202 0.0856 o. o8s5 3.42 vs 131 0.1036 0.1033 3.10 vw 222 0.1219 0.1219 2.86 s 113 0.1286 0.1286 2.79 vw 400 0.1457 0.1456 2.62 ms 420 0.1822 0.1820 2.34 ms 402 0.1943 0.1947 2.27 ms 422 0.2307 0.2311 2.08 s 143 0.2660 0.2651 1.94 w 440 0.2900 0.2912 1.85 w 343 503 0.3396 0.3379 1.72 m 620 0.3639 0.3640 1.66 w 602 0.3769 0.3767\* 1.63 m 424 0.3769 0.3782'/ 1.63 m 622 0.4117 o.4131 1.56 w 640 0.4728 0.4733 543 0.4834 o.4835 363 0.5215 0.5199 642 0.5215 0.5225 650 0.5558 0.5552 703 0.5558 0.5563 800 0.5807 0.5824 535 0.6159 0.6160 830 0.6653 o.6644 563 0.6653 0.6655 84o 0.7275 0.7281 383 0.7756 0.7747 850 0.8108 0.8100 547 0.7742 0.9741 *One line observed for both lines calculated. WADC TR 59-207 48

TABLE IX C) >-3 023~~OBSERVED AND CALCULATED hkl VALUES FOR H2B (NH3) 2]I o 4s in2 4sin2G 4s inO Intensity hkl k_ hkl A2 hkl k2 Observed Calculated Observed Calculated Observed Calculated vw 020 0.0330 0.0330 523 0.3475 0.3475 066 0.6883 0.6893 s 200 0o.o346.0346 602 0.3551 0. 3552 901 0.7134 0.7120 w 210 0.0429. 0429 135 0.3551 0.3554 616 0. 7134 0.7122 m 220 0.0678. 0676 054 0.3802 o.3806 266 0.7236 0.7239 vw 022 0.0761 0.0766 171 0.4227 0.4237 921 0.7449 0.7449 s 202 0.0784 0.0782 270 0.4389 0.4387 391 0.7581 0.7568 m 222 0.117 0.112 640 0.4428 O0.4435 672 0.7581 0.7593 vvw 232 0.1532 0.1524 711 0.4428 0.4432 517 0.7581 0.7587 m 420 0.1699 0.1715 036 0.4654 0.4666 636 0.7771 0.7782 Mo v oo004 0.1744 0.1744 462 0.4808 0.4790 0o10- 0.8255 0.8247 w 133 0.1810 0.1810 272 o.4808 0.4823 537 0.8255 0.8247 w 204 0.2086 0.2090 345 0.4808 0.4824 09og4 0.8421 0.8424 m 430 0.2124 0.2127 731 0.5089 0.5092 585 0.8421 0.8423 w 250 0.2408 0.2408 703 0.5229 0.5222 2-10-0 0.8583 o.8593 vw 440 0.2704 0.2704 525 0.5229 0.5219 682 0.8817 0.8830 vvw 115 0.2878 0.2894 416 0.5388 0.5391 547 0.8817 0.8824 w 343 0o3084 O.3080 181 0.5479 0.5474 816 0.9563 0.9546 w 442 0.3141 0o 3140 355 0.5562 0. 5566 286 0.9563 0.9548 125 0.3141 0.3141 082 0.5705 0.5714 961 1.0098 1.0088 503 0.3141 0.3145 13735 0.5812 0.5801 458 1.0427 1.0423 620 0.3432 0.3446 274 0.6131 0.6131 09og6 1.0602 1.0604 450 0o3432 0.3447 191 0.6883 0.6876 4-o106 1.3563 1.3556 w~~~~~~..

4. THE NUCLEAR MAGNETIC RESONANCE SPECTRUM OF THE COMPLEX CATION [H2B(NH3)2i. THE PROTON SPECTRUM (a) Background The technique of nuclear magnetic resonance absorption has been exploited in recent years as a tool for structural studies in both organic and inorganic chemistry. Although a number of erroneous structural proposals have resulted from improper interpretation of NMR data by overly enthusiastic workers, NMR data can, if properly interpreted and supported, give useful structural information. tr particular, it should be possible to differentiate between different types of hydrogens bound in the complex cation [H2B(NH3)2]+. Thus the resonance of two chemically identical hydrogens bound to the B -' nucleus will be split into four bands of equal intensity (I for B = 3/2 % 2I + 1 = 4), while the resonance of the six protons bound to chemically identical nitrogen atoms will be split into three bands of equal intensity (I for N' = 1.~ 2I + 1 = 3) if coupling with only nearest atoms is considered to be significant. For a proton of given chemical environment, we may define the chemical shift, a, as =HcHr x 105 Hr where Hc is the magnetic field strength necessary for nuclear magnetic resonance of the proton under consideration and Hr is the corresponding value for a reference compound, usually water. If both values are measured at constant frequency, the frequency dependence of the resonance cancels out and a may be interpreted as a measure of the diamagnetic shielding of the proton in question. It is of course obvious, then, that the diamagnetic shielding is an indicator of the chemical environment of the proton. Using the definitions above, one finds that more positive values of a should imply greater shielding and a more hydridic character for the hydrogens while less positive and more negative values of a imply more protonic character for the hydrogens, As yet, however, no complete correspondence between a values and chemical acidity can be claimed.-' Proton magnetic resonance studies have been very effective in structural studies on organic compounds; unfortunately, their application to boron-nitrogen systems has been much less effective, since the lines for protons bound to boron are split into smaller bands of reduced intensity by the magnetic moment of the boron nucleus: Since the major isotope of carbon has no nuclear moment, the proton magnetic resonance spectrum of an organic compound is usually much simpler than that of its boron counterpart. The limited solubility of the salts of H2B(NH3)2] in various solvents resulted in a weak signal at best and no suitable signal could be obtained from the solid salts of the cation. The fourfold (35) J D Roberts, Nuclear Magnetic Resonance (McGraw-Hill, N. Y;, 1959), p. 24. WADC TR 59-207 50

splitting by the boron nucleus of the weak signal obtained from solutions gave rise to such weak absorption that identification of peaks was difficult, and assignments as given below are based on the identification of very weak lines which appeared reproducibly in the spectrum. The observations and interpretations can be considered as consistent with the structural data already available on the cation, but one would need to be a real optimist to interpret the very weak lines as direct experimental evidence for the structure. In short, if one knows what pattern is anticipated, it is possible to find all the very weak absorptions expected; but if one were to look at the spectrum without initial prejudice, the separation of certain peaks from the background could well provide a problem. Conditions are still being studied in an attempt to obtain more unequivocal NMR data. (b) Experimental Results and Discussion Dihydridodiammion-boron (III) iodide was prepared from the diammoniate of tetraborane as described earlier (see page 26. The deuterated salt [H2B(ND3)2]I was prepared in an analogous fashion using B4Hloo2ND3 as the starting material. Diethylene dimethyl ether [CH30-C2H4-0-C2H4-0-CH3] (dried over CaH2) and liquid NH3 (dried over sodium) served as solvents. [H2B(NH3)2]I dissolved in the polyether (saturated solution) should show a triplet of moderate intensity for the six protons of the two ammonia molecules. The resonance of the six equivalent protons would be split into three peaks by the nitrogen nucleus (I for N = 1 and 2I + 1 = 3). Samples were examined using a Varian High Resolution spectrometer. Three such peaks were clearly recognized for [H2B(NH3)2]I in the polyether solvent with values of a, the chemical shift, at 0.048, -0.035, and -0.117. The average value for the six hydrogens of the coordinated ammonia molecules is thus -0.04 in the polyether solution with an average separation of peaks equal to 0.o8. Ammonium iodide in the same solvent has a 6 value of -0.20 and pure liquid ammonia has a b value of 0043 with a separation of peaks equal to 011. It is of interest that available chemical data indicate increasing acid character for protons in the series NH3 < [H2B(NH3)21+ < NH+; the increasing negative character for the 6 values of these similar compounds is consistent with the postulate made earlier, namely, that the smaller amount of diamagnetic shielding around the proton would show up as a more negative 6 value and a more acidic proton. The identity of the nitrogen-hydrogen peaks was confirmed by running the spectrum of [H2B(ND3)2]I in polyether. Since the precession frequency of the deuterium nucleus is 6.5 megacycles per 10,000 gauss and the oscillator used for the protium nucleus has a frequency of 40 megacycles, deuterium resonance does not appear with the oscillator used. The fact that the lines attributable to N-H absorption did not appear when [H2B(ND3)2]+ was used as solute supports the assignments given. As might be expected, identification of the N-H absorption was impossible when liquid ammonia was used as a solvent for [H2B(NH3)2]I. The B-H resonance was particularly difficult to detect in solutions using polyether as the solvent because of: (1) the low solubility of [H2B(NH3)2]I in the solvent; (2) the fact that only two B-H hydrogens are found per mole as WADC TR 59-207 51

compared to six N-H hydrogens, and (3) the fourfold splitting of the already weak signal by the boron nucleus. Only one of the four weak B-H absorptions could even be recognized tentatively in the polyether solution. In the liquid ammonia solution the dihydridodiammino-boron (III) iodide is somewhat more soluble than it is in the ether solvent. In this solvent four weak lines, identified as due to B-H absorption, were recognized at a values of 0.62, 0.37, [0.12 not visible] and -0.13. The average a value for the B-H proton of.[H2B(NH3)2]I is thus 0.24 with a separation of peaks corresponding to 0.25. Sodium borohydride in liquid ammonia showed the expected four peaks at a a value of 0.52 and a peak separation of 0.20. The fact that the average absorption of the B-H hydrogens in NaBH4 comes at a more positive value (0.52) than the average absorption of the B-H hydrogens in [H2B(NH3)2]I (0.24) implies greater electronic shielding and thus more hydridic character for the hydrogens of NaBH4 than for the B-H hydrogens of [H2B(NH3)2]I. This fact is in agreement with all available chemical information. The foregoing NMR study will be more conclusive if more definitive absorption can be obtained in subsequent work. 5. THE REACTIONS OF THE [H2B(NH3)2]X SALTS WITH SODIUM IN LIQUID AMMONIA The process expected was: Na + [H2B(NH3)2]X liq H2BNH2 + NH3 + 1/2 H2 + NaX NH3 Two samples, identified as [H2B(NH3)2]Cl by their X-ray powder pattern, were allowed to react with an excess of sodium in liquid ammonia at -78~ using conventional techniques. Values of H2 liberated prior to a sharp change in rate were: 0o9 and 1.4 equivalents of H/per mole of [H2B(NH3)2]C1. It is to be noted that the two values are in poor agreement even though conditions were formally the same. It is also significant that H2 evolution did not stop at the points indicated above but merely slowed up. On long standing a value as high as 2 equiv. H/[H2B(NH3)2]Cl was obtained. Since the solid reactant was identified as [H2B(NH3)2]Cl in each case by X-ray methods, it appears that the sodium reaction is not reliable as a means of compound characterization. It appears that these observations support earlier conclusions regarding the identity of B2H6.2NH3 (I) and B2H6o2NH3 (II). As noted on page 13, the question of compound purity seems to be a major factor in the rate and stoichiometry of the reaction of Na and the cation [H2B(NH3)2+]. Certainly the diagnostic value assigned to the sodium reaction in earlier studies is subject to serious question in view of current evidence. D. The Synthesis of "The Diammoniate of Diborane" by Metathesis Heretofore the diammoniate of diborane has been best prepared by the direct reaction between solid NH3 and gaseous B2H6 under very carefully controlled conWADC TR 59-207 52

ditions.(l5e The identification of both cation and anion in the solid suggested an alternative metathetical procedure for synthesis. If appropriate separative procedures were available to purify the products, one could effect synthesis by means of the general equation. [H2-(NI3)2]X + M[BH4] >- MX + [H2B(NH3)2][BHB4] The nature of the separation and purification procedure would of course be determined by the identity of M+ and X- and the properties of MX. 1. THE REACTION BETWEEN KBH4 AND [H2B(NH3)2]C1 If M+ in the foregoing type of reaction is K+ and X- is Cl-, one may take advantage of the low solubility of KC1 in liquid ammonia solution. The following equation then results: [H2B(NF3)2]C1 + KBH4 78C [H2B(NH3)2][BH4] +KC14 -780C The KC1 may be separated from the liquid ammonia solution by vacuum line filtration and the "diammoniate of diborane" should be recoverable from the filtrate. A sample of [H2B(NH3)2]Cl, free from NH4C1, was dissolved in liquid ammonia at -78~C along with a stoichiometric amount of pure KBH4. After 14 hours of stirring at -780C the solution was filtered (temperature held at -780C during filtration); the solid retained on the filter was identified as KC1 from its X-ray diffraction pattern. The solvent ammonia was removed from the filtrate by sublimation of the solvent, and the solid residue, presumably [H2B(NH3)2] [BHI4, was examined by means of X-ray diffraction. In the pattern lines of [H2B(NH3)21 [BH4] were found except that a slight displacement of all lines was noted (see Figs. Ila and b). It was assumed that the displacement resulted from isomorphous substitution of C1 for BH- in the B2H6e2NH3 lattice. The hypothesis was checked by dissolving a mixture of [H2B(NH3)2]BH4 and a small amount of [H2B(NH)2] C1 in liquid ammonia, and then removing the solvent ammonia by sublimation at -78~C. In the X-ray diffraction pattern of the resulting solid, no lines for [H2B(ITH3)2]C1 were found, and the lines for [H2B(NH3)21[BH4] were somewhat displaced, just as was found in the metathesis product. It is thus concluded that a sample of the "diammoniate of diborane," somewhat contaminated by chloride ion and K+ had indeed been produced in the metathesis reaction. 2. THE REACTION BETWEEN LiBH4 AND [H2B(NH3)2]B3H8 IN DIETHYL ETHER The synthesis, characterization, and structure proof for the compound [H2B(NHE3)2] [B3s] are described in a subsequent section. Assuming the above structure to be correct, the following reaction would be anticipated: [H2B(1N3)2J[b3H81 + LiBH4 Et2O [H2B(NH3)2]LBH4I+ + LiB3H8 WADC TR 59-207 53

'-......................... :'':''.:...:.:..::::::::::,::::X:::-::.::::::.:.: -,I....,,-: —:.:..:::::..:::::::!.::: -—.,., -W-:.:::::. —....1...1.... 1. ,III-. - -.I I... 11 I-1, 1-.1..11, —. II I 1-1.1I..11I- III- I ...,, ..: I..........''.1..., ---— ....... I......'..'.....'.,.,.,.... —- -....,.-.;.:.:,:!,-,:.:.:..,. I -. .. I 11 -.1- -.1..:.:.:..::.: 11 11 —., I I I.1. I. I.. - 11, I II::::::. - —::::::::..:..:..,:.,..":.::,:. - - -.",.-:,.:,.:-::..d: — -...... -.- - I 1. 111 I...............:::::.::::......,-..I..I.:...........:::::::::::.:::.,::....,. —..-,-'-....,:::::.-............ ....,-...''..........-........-.... I - -.. -- -- I.- -......,... - -- - ..11,.I'll. ....:-:..:.- -...,...,...I'::: -.-:!:::.:..-........'114 '.....11",! -, —-.......- -- I.... 11 -.. -.,-. ....................I.....:1, -,, "....,...,....... ::,:_:::::.,: -.,:,.!..:.. - - i o.......,";:.::.:::''::.:.'..,'.,:....::-::::::..-::,:.::-:-:,:,.:::::.":,:.:,..II.... - I... II i!!: iii`i!il'.-.......I... —...-...:-.-i.11,11.4-i", —..........-:::-..::::::.:,:.:;:............................11 -,.............. I.,.............,::, I...I....... I-. —-I..::..- - — I.-I.....-... -,...''.....- —::w:. -.,..::.,.I,,, :.,.:.; — -;;-;;;- ":. ..-'...:.:.,-,.-.-I.......- I.- ::::.:,:.:,:.. :.::...- I::,X::::-:::.X .:.. -..::..;_... q,:;::::::.-...,..-I...-.,... -... —:.:1.:.-.... I — -:..: -...::..:.,.. -:..I. -.,,,...,.-.......'...."......, — sq...::....,...... 1.... -.11- —.::: ]: -:''. - —.1..."'.........:.-m.,.s.,:,::.::::,. --... -..,. I...:-....I.I-.,:!:-.... - ........-......... ... -...:....11 -... -:I,..... -::..:::..-.....,I.-::.:... -—....'.. I 1.:.::.::::::::::::.::::: I:::.:::::::.:......,:::,::::.,........-........ ....;:.::..:!::::..w: I. I I....... I -:I::::;-::.;.::::. :::::......I........,,.....-'::::....,.%.....I . -I -,.,............... ":: ...X....... %.,,.-I- -......,...., -.. -.......-..1.............. -.......:::.:.:::.,. -..::::::::-v,...I................ :.:..%...- 1--... 4,.......,..,.. I .. ....,:I...-....... ........... .-.-I........I........-..... -:....:.....I....q..... .::!:::::::.::::::!:::.:... 1 I:1..::.........,...,............:::.:..::::::::::.. -. .,.,....,q............,,".."..-.,. I..... X.:....:.:::::.%.,,-... 1..:..::,::::::::::::::,:.-,.... .,..:- I......:. -:::.:. - -.::.......,.......,..... —.:::::;:::::...,.. I.:::,:::::::: ".......I...-............-.I........ : ::.,,,...... :,..v.I.I -...,I:...:.::.-.......-,.., ",.................I, -.:..... -:-:.:: X. I... .. , - I I I I.-:..:.:....:........I..........'...".1.1....j..."..,- ..."..'.:-:,1 ... -::.-.-11.. -—,.........,,..,...'.. II.11.II.....I-....-.......-.,-.I...I:::.:-.I....:,... 1,101.,,........I.............-.,I.,.-I...:::::::-:::..::::. - -.......I:,::::::.. 1%....:IX.:,-.:X.. —11.I-II, .. .::.::::.:.11'... I,.......,'..,.. ,,..... ...,::::::,;,:: w......,......,...,:.:.:.;. I.. 1.,.. q:::-:, I I, .., I.1,,.... - -. -.:.:"::-I. I .1, 1,.,;:!........:1.:..:.. .1 1,I...........-..... 1.11-....,..;,.I.,. —,.....- ....1.. I I :::..::.I.:::1. -1:... —.-:-:- -.':::-:.... I......I. ;: .. I: ...... I.. .............. I ..,.,..,.11... ... %. -"...!::: 11.-I. I....... -':....:,:.:.,:", . ", -.:.:.:.:.I.-....::. -. —..,-,. —,-.-,.:;.`:,.,.:-::-:-I.I I1.I......... -........ .. -,,.... :......,.1 11 I I., I 1, I I -...''... -.....................::.:.::::::.. ".............. I::-::::.:::: —:-:-:.I..'..::.:".III, 1:::.,:::::,..- I- 1.1. -........-...,:.. I I I, I. - I I,,... -V........, W.....,.I. —.I:::-.:-::-:;::.-::: .. .................:-.11 I ...., -...,.:::::....:::::..-......::-:-::::.:.::.............. ,:1:::.., _: - -.:-.: --:; ——....,:. --.:.:.....I. I........I.......-,..:. - I;..:...:-:., ,.....',.1:...:..:::::,;.:..::.::......:.:. w.:..... - . *i:::.::.:.:..::.:.... ......:-...:-..:..,.1.....-....... - -....... ....:::...,..::::: -- - I,.:::::: ::::,..:::.::::.. I.,......''... .....1:iiK::.:.::::::::...1.......... I...... I. 1..- :::.. 1. -:.... -.1.1.......... x -,.,.::::....:...-.-.-.-.1.1. V.-.........-.1.,........,-, -..........-...,.......................'. - .... I -..:iix'.11............-.........':.......,.,...,'....'.,..."..........''....:-I I.............''....-....:..:,.:::::,::::::::: F........14 —....",............-..... q........ -1-.....-I.-.....-.:::-::............''.., - -.:;.:..:::::::::.:..-::.::..:-....I.-....... :- - -.-.:.-.:.:.:..-.......... -....:..;.:::.......'.........."....................,...... ....... -.... I - -.:..:..:.:................".I ......... -:.::::::::::::..................1......"........;::..:::::::::::::::::::::::::::::,:::::::::::::..!:.:::.::::.I...: -:-:-:i:; -......,............,.....,..."..... - —,'.- -.... 11-1-......... -F........'....I.I..........,.........1.1 -I.:: -.::...........,.,...".,...''..,..,.,,,'..,ll.I...... ---—.I.. I......,....,.".......,.........",.,,.".,.,...,...,.,.,..". 1-1-1-1-1.- 11 - 11. III...'.......,.,.....''.,......'..:.... -............:..-.-: : ---- 1..11II..III.. I1,... I- I....1- 1- - - I.........,........,..."w...:!::,::-:::. ,.:.,,:.. ".::::.: - 11 I I .......... I.I.-I.,......................... ,:.:..;-.I'll. -- -.:::.::,.:-:::::-:.... -... - ...:::::.:. :. :..,:,:.:.:.,:.::.::,::...,.,,..,.....,,...I...... 1.11.1 I 11 I I . I ,;::::ii:,I................: "...::.::.::::. .... -- I.- II1. 11-11 I III I -11.. I ..:.....,.... .III.. I.1.I I........ . . - I I I I I. II.I...............1.... 1.11,I .,I -11II.I 1......... —...:-:::::::::-.::.:-.,-'....."...._ _ .,,......". 111-1 11 -I.I .11I.......:..:,.:.:.:... ........ -14.1 -....::I-I - III..:.I I.-I.,.11II.1.I:::::::!::.:::",.."..........., I. 1- 11-............I -.1... -.,::::::::::.::: -.... I -I.1 .I..I. -I............-,,-1141 — .I....':.:,:.::.:.,,,...,.-I.11II 11....-, -.11........-:. "....... ....................................-.,.- I II. 1,.1 . I..''-. I I................. -...,,.'':..::. -. - .,..1 . 11 - I I ". I - 1- 11 I .. 1........... —---..... --...'... ".,........ ..,:,:'' - I I. -.1 .-...1,:......,...,.....,..'........."..:....!:::.. - -.:::, -. I1,II I,.I::.,..:::..:: -:::_:.:.-::............,.. , - 1.. I- - - . I ...........I.1 I 11.. I — I-11'.. - -. — I...;!:;!;!!!! i.........., - I ..11.1 I I I 11 11 . -I,,I -.:::.::::.::::.:.:,::..,:.,...".,....,.,:.:,:.:......... -,,: - ...... I.-I-111'.- I. I.... -I I . I, -..- -.- -... ..., 11 1, . ".IIII ,, . I 1 .. 11. "..,". 1.11 —....... -—. . I......... . -1111. -,,,I, I — 11 - 11 1. II-....I. I11.1 I...:.."..,.F.-... I,...... -!., I1-111,-..111.1 —..1. - II -......... - I.- I.I.- .. - - -: -.:,: —::...... ''.1- 1 1, I I I , 11 I I -.I.-.. — I.,.. -X...:::.: - ".,...::.I:::-._.,.,.,,..I-I-. -..,11... 11 I.......... .... I,- I — -11. I -.... 1. .. III."Ill I I 1. I I 11'-, - -.1.. .I.1.........I...I1..... I...... ,... I...I........,III .:..::. I -.41-1...................... I... - .. I:I.1.. - I I 1- I - ......... —..1 —.I.II.1q:::::::.............I.1111.....I..,:::::::,,:,i,::::::::::!:::::. —-.. -.::;;:.:::::::.: .::.::.:::::..:I....,-.,;...-.........-,-....., .."... I.::.,::.::::.:.::.:.:.:::::::::,.:::.:,:1 %... -il!ig-.i, -1-:::!],M M..,.",..,,,:: -.1-1..:!i::, ".'..-..'......-..,...-.,.'.'...1-F., -- I -.iiiim —,. —. -.''....I. ....-I...... I....'',''.:.:,...::::::.:,:::-.......-.....,......-.,..... T.!."... .,:: 1. - ........ ::,:....:1;....... ...-... -11-1-1........ —. , .................''...::....... :!,.:!-.I 11I:.:,....,..,::-......I...'......."...,...............,: ::......I.....,......... I I. "....I.,....,....-......... . '...,.-...I-... . I I......- "',:: - -::: -::: -:.1-1. .... 1..... -:.::: A I I.- I .... I I...,......,.....I....,..,.......,."......"...;I:,::: ::::::; ii!;i...... - ..-....-...... - I...-I.,...... I......... :. -... I........... I —.... - I. ..-..::: -:-::::::- ff 1-1 -I.. -.1...14'......, 41,.,.... ...:. -:::::::.::..:::::::::.,:::::::::.:,:.,. - - -..-.........:.:: -...,-. —-,... ... —-.-....-:.:..:.".:.::;. - .............. —.....:1..:::, ,::s...:.-....,.:.::,::::;::...::...::-:,::,:,:,.,,.......,...........,.....:::.::: - 1.1 1............ -.....,. 1. - :.:::: II I:......:::...... -.. ......,".'..,-.,,.,.,...,.'. -...:..:::::: :.::.:.:::::.:..: -.."'... ..... ...... I. - -.., .1 ",..::: - I- -. . X.... --.. 1. I I,%......-'..,.-:.... ....,.-I....::::,.:;.:,.1. -.....,: . -.1,1 -I..... I"::..:.-;-.;... .. ", -F d I.....,...... ....I............ I,-,. -....::::::,-.... . -:.:.,..I.I. —l.....; .:.,::.,::. :::::::,", : ..,I.- I1..-..............,,,.,.-......... - ..-.-,,........ -.... -1.-.:....::.::;.. 1. I., -..''.. I I.1-1 I:,. - ............1.1...,.,......'......,......'...."........::..-!..., —-:I -::::-:::::::.-::. —,-:i:........ :.,.,...,: I..1. I.. —:...::,::.::::: -,..::: -;::: - -.. . ......':,,,....'.-................".., I I -: - - - -..,:.-......I.....-..,.::.: -......%................ 11.1.."'.....11 -, -.-....'...,.-........-...,.,.,-I'...:.I....."....q. I....,.....,%...........,.,.,................... 11 . -::::-.....,,.,...I..-.'.:: :::,:::::::::.:::::::.:::::::,:. I...,,.,.,."...:::::.: . .,,..%-. I...,.::..I.. I., I...,,::.!.,.1.II.I I. -........- -..;:::::.. -:-X :;:.::.:::::,....,... I - I -,., I. -,.I1.. I I ..:. X.:..: I.,. ....:..:: ...,....I.. I'll.I.11-......-..,.::-: ..1111,,..... - -. —:-....... -.1.11- ... I.I.- -.-.1...... -...I I —.......1...... -.-I.........w, - ..,.. ... ...-:........ -.,W.....1,'.,.,,'......I...:::::::::::::::::::,.:.:.-... .X. 41::.::.::::::::. -::::..,:::,,..4. ::: -:: -..,..I -11I:: 1: :,...,,:,......1 1, -::- I....11.... -....................................: :: ::........ -.... I....... -.. 1-........... - 11:.. 1,,.. -.,-,. :::::,:::::::;-.: :::::, — b :::: -:: -:...:.,......:.:...:,..I......".....:.;:::......-2'. ...I.. I...,.:.:.:::.... 1. ... -:::,::::::::::::.:::,:'::::::::: :,v.-,.........,.-.... —,..I -,..-I...: 1:1:...,............'.......I... .......:::...%,.,..-',.,.. -I...I............ -..-, -. -:;-:-.-....... —..,,,:::::::::........:',I. II.. 1-1-1:::::::::.;:..:...-.,'.,:::.:: ..:.,.,.. — -.-, ..I, I.I -,.,O...,,,....:::::::.:.:.:::,.. 11 I --... . -,- I II.. II..."......., ----.I - -::::::_ -.:.!...... 11 I...".... ..1. ....-........... 1.11 - 1-:- -..... ..,... ,,::...::::::..,.,.".,. - - -, 1, :.I.I .. 1.....................-...::::-:!::::::::::-::::::;:........-'.-........:..:!:::::.:-::-:.::.::::::........... I...,.,, - - ..-......11-1-I.I....-.1-..-...I. —.,-..,.. -, —.1.1............-,:.::::.:.:::::-::.:::.::: -::..'...... %.I--.....-....................,....."....,...',.....,... _ -... 11.111.. I... ....... - ...... —.. .......-...........I,.....-............-I ,....I.............. - -..I... —.I..'v.. .:: -:-, .....-........ -.- -I'l-,-.1- 11I. -....... 11.1 -.11.11 - - -:,,: I - -. -I ,::.:::!:::::::.::....- ".... I ... 11 .... -1-1.1-::,,, - -..I... 1. -....... -—.I-.,'... —....I.:::...''. -;.:,;...::................... ---........:, - 1 - —......I.... -—..F.., — - -,... -...1.11I-,- I,- I.:.'',.::,..'.......I..:.:.,.. I.- II -,......:.: -.-.......:::..:-:-::.:::- ,:::::,: - 11]:: .. I.- !:i..... - . -.1 1. I.,:''.1. 1. I-,..1... -—::::;.:::- -,,.... —......,-,.-...1. -.... I -- - .:I -....I.. -..............:1.11. -I:....:.:.I:I:.:: -...:.:-;.... -... - -... —... I. 111.I:,II. ---........... I —.,-`... I I.-...... I.-............... -.11 —::::: :.:.,:::::.::.::::::;::,.:,::,::::.:.......,..I.I...I..,.,.....,..."..,......,.':.:-:..I;- I —....I.....................:::::.....::::::: —..........I.......11-111-1-1 - I -,:-.... -..''...''- I 1..1 I. I. —...:::;::::::::. —-.... I...1- ,::::-::-::::::.::::::.:,:..: 1.1'. ''.. :,, -I;%......,......,.....,..".I....,.,.,......,..'...:::::,.::.:::.................. ..'' - - - .1-.... 11.1 .. II1..11 - -- -- -.-,,.- I...-..... -...... -.I... .11....... - I........ -II.- —..... -.....:. —-. -11.I.:II,. -.... -.......,..-..,.,-...'I...-..,..I...,,..I.,.....-.- - ::. - -..1... - -.1-,: 1-: , -.::...:::-::::::..... I-. I..''. I.. I - ,:: I.., —:::,: —-..:-....-:,...:. -......::.: -::::.:,:,.1.1 ,:. -:::.:::!::::::.:.:::.:.:...:,..:...,...:.:.:,...,:.:.:.,:.:...,:,.:.:,,:::,.. - I.v-........1 -......... -....... 1. - 1:. - I- -, I - 1. -. I... - -. I"....... ...........:.:.::,::::::.::::::::,:::::::::::;.,:..'..1.1'':,: ..... I.. -1 —- —.....",",.,....,'.....'.".:-, -.'.v-.X. -.-I..''.....I.I.-...''.1. -:111:':.1 —II -.-......:.-......-........ —11., -- I.... 11.2- --- I-.-I-,I.. -.I.1..-.. IIII I I.- ...IIII I. I... :..'...,.:...:...:,:.:.:.:.: .:,::.:::::::::::::::::::X: —-,.,.,:, -.::::,,.. 1.111.1. I- I I.-I-II —:: -::V::::.-. :;:.:.:.:.:.:-:.I...................... -- ........... -,,-.1.. I...II I.,............ .......... ....', q...: -::::: .: - .1 I II1 1., I 1- 1. . I I. I I.- - -...::::. 2 - 1- -....... -.......... -.,,-.1. 1 1, I - II. ................. 1. I —,-.......... - I11.I,.-.1.11I,...........,.",.....,.,.....".............I.......................-.-:-,.-..:: —-.::.....-.........-... 1.:::::;:.: —. —.... ''. 11.I 1.I. 1.-II- .. - II -. I..1. -. —.-....''!.,,::,,::::::::::,.:.::::..:::.-,-I -.1III.11,I-. I I...11,.11.,:,:,::,:.:,:,:,:.:.,...::,:.:,:.........'..,..,,:......:::::..::: - - -..............''.-....... 1. -.... - -,,.,I. -II I- I, I........ --........-.....:-:-::::,: I....,.,'..''.''.'..".,.111.lI I I... I...1 1. -1. -.1.1.1-.. -.: " -'::::,:,. -,.1. I.... I I.1 I- I I'l l.....................,...I.".....'....".'ll...:,..... -----....... ... —.,::: —.. I- 1. - II 11 I.. -.1. - -.:::...:.:::,.,, -. 11. . - -., 1, ..... - -... - .:.:::::::::::.::-.:::::::,:-::, I...'', - - - -......-., -, -I I.1 11...I.. - 1. I.111. I II .:::..::::.:"::.::::::::::::!::,:.:..::..::::.::::::....,........1,,.'..,....::::::.::..-..::.:!::::::::::.:...:::::::::::.....,.., —. —-... 11........1. -—.- . 11 ... II'.'-' —-..,. —-. —. .III11 — I.''...'''....., I... -_X.::.., -:.::,:.:::::,:::::::.::::::,... I:-. ——:::::::.:.::::.. -:, I. II....:.,.,,-.1I-.11.1 - I-...- -...................'..,,-.:::;::- :::::-::::-.... 11 I.. 11 -1- I - 11 II-..I - -......I...-..........e:...:.:... ,:.q.:,:,::.:,:::::-::.:::,:.::_:.:,:. % —,v-,-.,I1,., .:I.. ....I.I I-.I II..1,.. -I .:: - -::::,.:: -.: - ,.I. -. —...,I'',,.I..II. II..1,...,.. I ...... ........,.,..................,.,.....,.,..,....,... I -... I...:::::::,:::::.::::::.,:."::::::::::::.:.: - Xq,..::::.,:::::::.:::- -.1-1-........ 11 , - -...... -............... - - 11-.... I I..111..1. I I.......... -...... - -...... -.......I..... 2 - 1.'11,..,..,.,. ....... II I. ................' —v - -:::::-::::::.,;.:..:,..:..:::.::::,::. V.:..........,I.I.I.".."..,..,.............,..I... -I.-.....;.: 1. -.I..... I .. .....,..v,% I.......I.... 1:...: :::::-::.:::::-:-:-:.:::::.:::::::::::, -I.,I-... - - -.......-.1 -. - I.........,, ................ I I .... I I I. . . I.. -, I., I I . I.1... -1-1-1... -.-,...I —:.:::.,:::::::i::::.::,..::::::,::::::..,................ ...... .. 1. :.........:::::.::::,::::;:: .,. - I I -. 1. - -.- - --. — I .. .... 1. .,-... I.1111, .I -... I.".. -.......-.....,.:::::::::::;:,:,::::::., 11, -.,, -.-I- -..-..... ', ...,..." :..111, —..;... -:::::::.:.: :,::::::.::,. ............... ::I....,-,...,.:,:,:::.:.:::::::.,..... —... -...:',,:.X:.::::::.:::........::::-:::.:::::::::.::::,:. 11.. - .............. 1. -.1 I...II1- 1- 1-1- 1. .... -.:!:......... I..'..'....., ::::.:.::::::.:::.,.::.V:::- I.-::::.:::::::,::,.-.:.:.:.:...::,:,:.:-. -.-F.F.'... ...:::::::::.:::::::;::::::::::.:.:..,;..,_..'..'.,,......'...,.1..1-1... I. .-......-..... I -.....I-. -....... .,.,.......-..................-.... . ....II .:::i1-::::.::::.. ....:.::::::::::: 7:..........b.-,'. - X::...::..!:::"'I.1 III.:.: :,::.:::::::.:.,.'...'..,..1-11,::::,::::jjj;::::;i,',:. I.. - 1. -......., -... ...I....... -..,.............. I.....:....:.:.......1. II .:::: :::::::-:::: .:..,...-.,:.:::::-...... -.-...........,....,.,-., -I.-...::::::::,.:::::::..:.::::,: -:,:::::,...sji-........ 1. —-..........%.,,,,..-I I -- II 111.1 111-1-1 11 —-..... --.......I........... -... — ...11 1.'.......,...,.,,v,.,.,...."....,..,.,.........,::::::::,:!:::...,... -........,'...,. -1..::V.::: :::.:: -:::: :::::.:Xx....11,,.-:,:::': -..,",.". I I I I I.......I .,,-.,.,.,"...,.,,.,...-......,...,%........',".-.'-,.:..:,:.:.: 4..:..,W:::..::::: - -,. q.,... I1,.. I.1 X:,::::::.:.::: -, —.F I-., -.......... I...... .....'"',... -, ,.. -::.. -.............: - -.... -.1-.... —1............-...... -1-.... I-I..,'',,...''.,:..;.......: , -.............- -,, -......:,.:....::::!:.::::::: ::.:,.X ..... -:..::::,::::.-::!.:::::::,:::::.::::::::::::.::::::-,.. ::,:;::i]]i,v..,.,. I,....,.,.,.'-I ,.,...........",-: -..:,, ".:.:.,.... 11 II.I-,:-... I., V,'..., 11..... -.... —. I.. I I........,.,.., q....,. '....::.: -........................,,,..- 1 I::::::.::-:-::-:.,......... ............,:I:...:::: .:.,.....,...,.,...., I 1, I.-,-,-',. I-I..... - I -....... V. - ...... ..,.,.,...-::,:.I. 1-1- ,.. 1-1-...II...II.-..II..,I::: !:::'::::: q:::::.:::::.,.-.............I.,,,. -.. I.,...,::: :::::::::::::::::::::::::::.:::.::::::::::::,::.:.. I.. I,.:.:.'. I-, .:.:..,.'. . - ::..:'...:... -.-.-.,,II11. - -.11............ -...I..... ......I...-, -,...,::: -.-::::::::::.:,:::::::::.: -...... I......I....... - I, I.-,. i:ii:i.., -........I.........-11 -.........-.........-...I. I....I.I..-.."I..... --....- -,::,:..,::::-V..::,,, 1.1,:: -::: :::- -..... -............. ..::::::::::::::::..,-.:...,-..,..-I I.."..,.1I I I I....... — 11.111.,..... - _,....... —,,....: X.:::... 11-1 -.1-1.-.1...... i:;::::i:;:i;."....:.::::::::-:::::::::::-:-::: -............ -.-.- I,.. I I I... .-.......-.........-...I- 11-I,-., -..,.......'W-F". I..'......!:::.::-:::.q.;: V,::::,:.::..,-....:...:I11::.,I:.:.:.:,:-:.::..,:.:,.:,.:::::.::::::::::::::::::::I. I -.."..'. :.-... ,.,..,.,... -...... I........ —...I.IIii,,,,,',.......,,', .,.,.," "..,:: X.:,:::::::::.:...:.::.:::...I I I.....,;::::::::::.:::::::,:,.,:..:: V.::.1.. I —........... '. I..,. I.,,,'....".....,::.,-.I. .-..,,.-1:II.I:..1..''. --.... I I.....-: -,: -::: ---- -,-.,.. .-1.,,:::..,.:::.:::,.::::,:; -......-................iI.,..'' -.1.1- -.... I-.I -. -................. -.1-: —.::::::::-:-:.:::-:....,... -I 1.- ---—..-I-.-I....... ,.,.."...",.......,!.:.,::::::::::::::::::::::::.:.:::::..,:X:.:.:.: -.1-...... -....I11...- . ----..... ,: -...... 1. .1 -- -II'',, -— I......I.... —.-:-..::::::::- -., -::::..... :.:.:::.::: - -:- - -..- I. .1 -- I . --—.I-:...:.:.:.:.::.:.::....:::::::::::::::.:.:::: - -..:::...., .. I I: -,::::.1-....... - --- II II.... 11...1-1......-......-.................:... %::::::::::::::: .:........ ,.......... 1. 11''....-.......I.......''Ill.-..,,....,.,...,............'............,.,.....''.,...,.,.,.,, -::::. -......I. - 1. I.- I I..... - - - - - -.1..................... -.-:X:.::::::::.::::.;:.:::., - -..... - -...... - - - -.- -..... ,,- -.- I- 1- 1-.....,:::::::::.::::::::::::::.::..::::....'.::..:.:.:. I- I.- I....

When diethyl ether solutions of LiBH4 and [H2B(NH3)2]B3H8 were stirred at -78~C for a few minutes, no ether insoluble B2H6-2NH3 was formed. However, when this cold solution was allowed to warm up slowly, white crystals of pure B2H6*2NH3 (identified by X-ray pattern) started to form near -23~C. The white precipitate redissolved when the system was cooled to -780C and reappeared when the temperature was raised slowly. The filtrate obtained by removal of the precipitate gave a white, gluey material when the solvent was removed. The gluey residue contained a microcrystalline solid phase. An X-ray powder pattern of this gluey mass gave evidence for a new crystalline solid phase, presumably LiB3H8. The supposed impure LiB3H8 residue gave off H2 slowly at room temperature. The entire study is as yet incomplete. Details of other metathesis reactions for the synthesis of B2H6.2NH3 will be presented when more certain characterization of reactants has been effected. E. A Summary of Developments on the Diammoniate of Diborane and Its Derivatives Arguments to support the structure [H2B(NH3)2]BH4 for the "diammoniate of diborane" have been presented in earlier publications.(l5) The following additional information relative to the "diammoniate of diborane" and related structures has been reported in the foregoing pages. (a) A procedure for preparing B2H6-2NH3 in one-gram lots has been developed. (b) A microcrystalline form of B2H6.2NH3 has been obtained and the powder pattern is tabulated for ready compound identification. (c) Relatively strong evidence against the structure [HB(NH3)3](BH4)2 has been accumulated. The so-called "diammoniate of diborane II" which was formerly assigned this structure(15) has been shown to consist chiefly of the regular "diammoniate of diborane" with the structure [H2B(NH3)2]BH4. The stoichiometry of the sodium reaction (B2H6a2NH3 + Na —- NaX + 1/2 H2) which was formerly a major line of evidence supporting the structure [HB(NH3)3][BH4]2 appears to be very sensitive to compound purity and the structural significance of the earlier data on the sodium reaction is questionable. Molecular weight data which supported the dianmmoniate (II) data are also subject to re-evaluation. (d) Improved methods for the synthesis of pure salts of the form [H2B(NH-2]X have been developed. (e) A complete single-crystal X-ray crystallographic study which established the structure [H2B(NH3)2]C1 has been completed and the data confirm the above structural assignment in every detail. The structures of the bromide and iodide have also been deduced. WADC TR 59-207 55

(f) A nuclear magnetic resonance study of [H2B(NH3)2]I can be best interpreted in terms of the accepted structure of the cation [H2B(NH3)2]+. (g) The "diammoniate of diborane I," [H2B(NH3)2][BH4], has been prepared by metathesis, a fact which offers final and complete chemical support for the above structure. It is now possible to make the diammoniate of diborane without going through diborane, WADC TR 59-207 56

II. REACTIONS, STRUCTURES, AND PROPERTIES OF OTHER DIBORANE ADDITION COMPOUNDS A. Background The "normal base-borane adducts" have the form Base: BH3. It was a deviation from this expected pattern which made the previously discussed ammonia addition compounds of diborane anomolous. Trimethylamine-borane, first reported by Burg and Schlesinger,(36) can be prepared by a variety of reactions. These include: (a) H3BCO + (CH3)3N > (CH3)3NBH3 + CO (b) B2H6 + 2(CH3)3N - 2CH3NBH3 (c) Al(BH4)3 + (CH3)3N >-(CH3)3NBH3 + other products (d) NH4BH4 + (CH3)3N > (CH3)3NBH3 + H2 + other products The compound is a white, crystalline solid which melts at 94-94.50C, has a vapor pressure of 77.2 mm at 96.90C and an extrapolated boiling point of 171~C. It is monomeric in the vapor state and is relatively stable toward thermal decomposition; it is hydrolyzed slowly by water. Wiberg, Boltz, and Buckheit(37) reported that dimethylamine reacts with diborane to give an unstable product, (CH3)2HNBH3, which loses hydrogen at room temperature to give an amine borane, (CH3)2NBH2. On the other hand, G. Huff and co-workers at Callery Chemical Company reported that (CH3)2HNBH3 (or an isomeric product) could be prepared as a stable, nonhygroscopic solid which did not decompose in alkaline or negtral aqueous solution and which could be distilled without decomposition.(3 ) Huff and co-workers first prepared the stable product by electrolysis of a solution of borohydride salt dissolved in liquid dimethylamine. No further direct evidence regarding this apparent inconsistency in the properties of (CH3)2HNBH3 has been reported in the literature, although Bissot, Campbell, and Parry,(31) working under the sponsorship of WADC, found that the stability of a closely related molecule HO(CH3)2,BH3 was very much dependent upon the purity of the sample being studied. The pure compound HON(CH3)2BH3 was stable up to 550C while the impure product evolved (36) A. B. Burg and H. I. Schlesinger, J. Am. Chem. Soc. 59, 780 (1937). (37) E. Wiberg, A. Bolz, and P. Buckheit, Z. anorg. Chem. 256, 285 (1948)o (38) W. H. Schechter, C. B. Jackson and R. M. Adams, Boron Hydrides and Related Compounds, Second Edition, Callery Chemical Co., Callery, Pa., May 1954. WADC TR 59-207 57

hydrogen easily at room temperature. Such an observation suggested that the instability of Wiberg's compound might be related to impurities present in his sample, while Huff's stable adduct would be a pure form of the same compound. Monomethylamine-borane was first described by Wiberg(39) as an unstable solid which melted at 5-10~C with evolution of hydrogen. A sample of monomethylamine-borane prepared by Parry, Kodama, and Schultz(15g) using Wiberg's procedure melted just above 0~C with gas evolution and McCoy and Bauer(3) described both Me2HNBH3 and MeH2NBH3 as unstable solids at 0~C. No stable form of this complex has been reported in current literature. The properties of anmmonia-borane, first reported by Shore and Parry,(31ll5c) stand in sharp contrast to the properties reported by Wiberg for monomethyl and dimethylamine-boranes. The compound is a white crystalline solid with a negative temperature coefficient of solubility in diethyl ether from -78~ to 250C. Clear rigorously anhydrous ether solutions are stable at room temperature to the extent that only a small amount of precipitate appears after standing for several days. The precipitate has the empirical composition (H3NBH3)n. Trace quantities of hydrogen indicate that small amounts of (H2NBH2)n may contaminate the ether insoluble solid. In the: presence of trace quantities of moisture, the ether solution becomes very unstable and solid material is precipitated quite rapidly. Ammonia-borane is soluble and stable in anhydrous liquid ammonia. On the basis of X-ray powder pattern intensities, there was no detectable conversion to the diamnmoniate when a sample stood for 30 hours at -780C and an additional 18 hours at -450C in liquid ammonia. Reaction of ammonia-borane with sodium in liquid ammonia liberated one equivalent of hydrogen per mole of ammonia-borane. The stoichiometry of the process suggests the reaction: H3NBH3 + Na —--- 1/2 H2 + NaH2NBH3 The sodium amidotrihydridoborate (III) has not been isolated from this process in pure form; its pre aration by another procedure has been reported by Schlesinger and Burg(o) but has not yet been verified. Solid ammonia-borane appears to undergo slow conversion to the "diammoniate of diborane" at room temperature. A sample which stands at 250C in a dry atmosphere will not redissolve completely in ether. Very slow loss of hydrogen at room temperature occurs also. Its stability is orders of magnitude higher than that reported by Wiberg for CH3H2NBH3 and (CH3)2HNBH3. Solid ammonia-borane sublimes with difficulty under high vacuum. Its density is 0.74 g/cm3, The structure of solid H3BNH3 was determined independently from powder data by Lippert and Lipscomb(40a) and by Hughes.(40b) The unit cell is tetragonal (39) E. Wiberg, Naturwissenschaften 35, 182 (1948). (40) a) E. L. Lippert and W. N4 Lipscomb, J. Am. Chem. Soc. 78, 503 (1956); b) E. W. Hughes, ibid., 78, 502 (1956). WADC TR 59-207 58

with the C axis differing only slightly in length from the other two. The structure of H3NBH3 is shown in Fig. 12. The highly ordered boron-nitrogen arrangement implies a high dipole moment for the molecule. In addition to the nitrogen bases, many other electron-pair donor molecules can coordinate with the BH3 group. Even the weak secondary bases CO and PF3 can form borane adducts.(36,41) Adducts of these weak bases have not been easily justified by the usual theoretical prejudices used to interpret coordination phenomena and a more detailed study of the properties of these materials seemed to be of considerable interest. BE The Amine-Boranes 1. PREPARATION AND STABILITY OF THE MONO-, DI-, AND TRIMETHYLAMINE-BORANES (a) Background As noted in Section IIA, ammonia-borane and trimethylamine-borane have been described as stable, well-characterized solids while monomethylamine-borane was described as an unstable material which loses hydrogen below room temperature. To compound the confusion, stable and unstable forms of dimethylamine-borane have been described, and the possible existence of isomers has been suggested. During the systematic studies in this laboratory of the amine-boranes, monomethylamine-borane was prepared by the low-temperature interaction of methylamine and diborane in tetrahydrofuran. The resulting product from the reaction: 2MeNH2 + B2H6 Tetrahydrofuran> 2MeNHBH3 was surprisingly stable at room temperature and even melted sharply at 580C without observable decomposition. This behavior is reminiscent of an earlier observation made by Bissot(42) in this laboratory to the effect that HO(CH3)2 NBH3 is a low-melting unstable solid when impure, but is a stable solid melting sharply at 55~C when pure. Physical data obtained on the series of amine-boranes indicated rather clearly that the stable compounds under study are indeed legitimate members of a series and suggest strongly that the unstable samples are simply impure materials resulting from uncontrolled local conditions during compound formation. (41) RB W. Parry and T. C. Bissot, J. Am. Chiem. Soc. 78, 1524 (1956). (42) D. H. Campbell, T. Co. Bissot, and R. W. Parry, JO Am, Chem. Soc. 80, 1549 (1958).o WADC TR 59-207 59

I I 5.23402 AMMONIA-BORANE LIPPERT & LIPSCOMB. HUGHES Fig. 12. The structure of H3NJBH3. WADC TR 59-207 6o

(b) Experimental Procedure for the Preparation of Stable Methylamine-Borane Diborane was distilled twice at the temperature of methylcyclohexane slush to remove higher hydrides. A 50-millimole sample of B2H6 was added slowly at -78oC to 15 ml of liquid tetrahydrofuran (previously dried over LiAlH4). Commercial anhydrous methylamine (Matheson) was dried over sodium and about a 2-ml sample of the liquid was distilled into the reaction tube containing the tetrahydrofuran solution of B2H6. A slush bath at -112~C was placed around the reaction tube and the system was allowed to stand. The temperature rose gradually. The solvent and excess amine were removed when the temperature reached -450C; the system was then allowed to warm to room temperature for a few minutes to aid in solvent removal. Finally the reaction tube was surrounded with ice and the last traces of amine and solvent were removed by pumping on the system with the highvacuum pump for 10 hours. Trimethylamine-borane and dimethylamine-borane should sublime under these conditions. A cold finger was then inserted into the reaction tube, the sample was warmed to room temperature, and MeNH2BH3 was sublimed onto the cold finger as large rectangular crystals. After about half of the product had sublimed, the residue liquified. The sublimation was then discontinued and the residue discarded. The pure product melts sharply at 58~C. It is soluble in ethyl ether and dioxane, and dissolves with difficulty in benzene, Analysis of the sample showed: hydridic H = 6.64%, B = 23.8%; calculated values for CH3NH2BH3 are: hydridic H = 6.68%, B = 24.10.% 2. VAPOR PRESSURES AND HEATS OF VAPORIZATION FOR THE AMINE-BORANES As part of the program of study of the chemical and physical properties of the amine-boranes, it appeared desirable to obtain vapor-pressure data which could yield values for the heats of sublimation of the compounds. Since the sublimation pressures of the substances are all quite low, below one millimeter of mercury at room temperature, the usual manometric techniques are not suitable. The experimental procedure employed, therefore, was the Knudsen method which involves the measurement of the weight of material effusing through an orifice in a given time. This method, although fairly simple and straightforward experimentally, is subject to numerous sources of error which must be eliminated, or for which a correction must be applied. In particular, it is assumed that a saturated vapor exists in the vaporization chamber, that the vapor molecules effuse through the orifice into a region of zero pressure, that the effusion process is molecular (i.eo, the orifice diameter is smaller than the mean-free path), and that the vapor molecules are single molecular units. A correction customarily is applied for the back reflection of molecules from the sides of an orifice of firite thickness. Values of this correction due to Clausing are given by Dushman, who also discussed thoroughly the above and other sources (43) St Dushman, Scientific Foundations of Vacuum Technique, (John Wiley and Sons, N. Yo, 1949), p. 99WADC TR 59-207 61

of error in this experiment. The pressure is calculated from the relationship P(mm) = 17.4 w KAt M where W is the weight of the substance in grams effusing in time t seconds through an orifice of area A cm2. M is the molecular mass, T the absolute temperature and K is the correction factor for orifice thickness. The two orifices used had diameters of 0.0374 cm and 0.076 cm and ratios of thickness to radius of 0.163 and 0.314, respectively. The corresponding correction factors were 0.935 and 0.865. The compounds studied included mono-, di-, and trimethylamine-borane and ammonia triborane, NH3B3H7. The preparations of all but the trimethylamine compound are described elsewhere in this report;(44) the preparation of trimethylamine borane is available in the literature. All compounds were purified by vacuum sublimation immediately before use and were handled in a dry atmosphere. They appeared as white crystalline substances with no amine odor and sharply defined melting points. To eliminate the possibility of error from decomposition during the course of a determination, the stability of the compounds was checked under comparable conditions. This test consisted of a hydrogen evolution measurement in vacuum over a period of several days and indicated that no significant loss in weight from this cause occurred. The data obtained are given in Table X for the four compounds. It should be noted that part of the data listed were determined after the closing date for the termination of the contract covered by this report. They have been included here for the sake of completeness. To calculate the heats of sublimation of the compounds, the data were fitted to the integrated form of the Calusius-Clapeyron equation using least-squares methods. The values obtained from the slopes together with the vapor-pressure equations are given below for the various compounds. In the case of the trimethylamine compound, data determined by conventional manometric techniques are available in the literature(44) for temperatures between room temperature and the melting point. Since these data appeared consistent with those determined in the present investigation, all data were used in determining the constants of the equation. Methylamine-borane (CH3)NH2:BH3 Loglo P(mm) 4114 (1 ~ 0.053) + 11.411 (1 ~ 0.063) T range = 0~ to 450C AH = 18.8 ~ 1.0 kcal/mole (44) This report, Section IIB. A. B. Burg and H. I. Schlesinger, J. Am. Chem. Soc. 59, 780 (1937). WADC TR 59-207 62

TABLE X VAPOR-PRESSURE DATA FOR SOME AMINEL B0ORATNES R) I\) Methylamine Borane, Dimethylamine Borane, Trimethylamine Borane, Ammonia Triborane,* (CH3)NH2: BH, (CH3)2NH:BH3 (CH3)3N: BH3 NH3:B3H7 Temperature, Pressure, Temperature, Pressure, Temperature, Pressure, Temperature, Pressure, OK mm x 103xOK mm x 102 OK mm x 10 K mmx0 273.6 0.259 273.8 0.778 273.4 0.984 303.6 0.776 292.8 1.63 285.7 2.52 274.2 1.04 309.6 1.51 500.0 6.75 285.7 2.52 285.4 5.29 514.2 1.99 300.0 6.60 291.1 4.12 285.4 5.27 519.4 305.4 8.42 294.0 5.90 289.4 5.15 527.0 506.9 10.5 295.2 6.64 289.4 5.16 506.9 8.49 299.2 11.1 296.4 8.96 509.1 14.1 502.5 18.8 297.5 8.19 313.2 17.8 506.0 24.9 298.0 7.94 314.4 19.6 506.0 25.3 *Ammonia.triborane, although not a legitimate member of this series, was included here to conserve space since the procedure is the same as that already described.

Dimethylamine-borane (CH3)2NH:BH3 Loglo P(mm) 4034 (1 ~ 0.038) + 12.544 (1 ~ 0.042) range = 0~ to 35~C AH = 18.5 + 0.7 kcal/mole Trimethylamine-borane (CH3)3N:BH3 Loglo P(mm) = 2962 (1 ~ 014) + 9,894 (1 + 0,014) T range 0~ to 90~C AH = 13.6 ~ 0.2 kcal/mole Ammonia triborane H3N:B3H7 Loglo P(mm) 3739 (1 ~ 0.0075) + 9.200 (1 0.0oo096) T range = 30~ to 550C AH = 17.1 ~ 0.1 kcal/mole The uncertainties in the constants of the preceding equations represent standard deviations calculated from the residuals of the experimental data. Inspection of the heats of sublimation shows that the values for the monoand dimethylamine compounds are approximately the same and are some 5 kcal/mole higher than the value for the trimethylamine compound. This suggests that a single hydrogen bond involving a hydrogen attached to nitrogen is important in the crystal structure of the first two compounds. 3, THE DIPOLE MOMENTS OF THE AMINE-BORANES (a) The Implications of Dipole Moment Studies Many writers have made much of the fact that benzene and borazene are isosteric; in fact, borazene, B3N3H6, is frequently referred to as "inorganic benzene" The analogy is encouraged by a marked similarity in many of the physical properties of borazene and benzene, but the chemical implications of such an analogy are clearly eliminated by marked differences in the chemical behavior of the two reagents, After ammonia-borane was first prepared, many workers (i.e., E. Wiberg, E. Hughes} etc.) were quick to note that the molecule was isosteric with ethane and the trivial name "inorganic ethane" began to appear in discussions of H3NBH3. Although borazene and benzene have rather similar physical properties, ammoniaborane bears no obvious physical or chemical relationship to ethane, The former is a white crystalline solid which has a barely detectable vapor pressure at 400C while the latter is a gas with a boiling point of -92~C. The low volatility and highly ordered crystal structure of ammonia-borane suggests a high degree of electrical disymmetry in the molecule, a fact which should show up clearly in the WADC O'R 59-207 64

size of its dipole moment; indeed the measured and published(45) value of 5.9D confirms such expectations. The method of representing in conventional formulae the charge separations which give rise to such a dipole value is rather arbitrary. Two possibilities might be considered. In the first, the fact that the boron atom of the BH3 group accepts part interest in the electron pair of the nitrogen is frequently used to support the assumption that the boron becomes negative relative to its original state in the borane group, while the nitrogen of the ammonia becomes positive relative to its original state in the ammonia molecule. Charges are then assigned as follows: 6+ 6H3N. BH3 Alternatively, one can try to represent the fact that boron is more electropositive than nitrogen by the structure 6- 6+ H3NBH3 The latter representation is clearly suggested by the fact that reactions of compounds containing such boron-nitrogen structures usually occur to give products in which the more electropositive portion of the attacking reagent is bound to the nitrogen while the more electronegative portion is bound to the boron, H OH H C1 B / H i N H'OH N HO \ / H Cl\/Cl T H H B3N3H6" 320 B3N3H6 "3HC1 A comparison of the two formulae soon reveals that the difference between the two modes of representation is a result of the reference state chosen in each case. The first case implies that the boron of ammonia-borane is clearly negative when compared to the boron of an uncomplexed borane group, but the line of reasoning from which the formula was developed says nothing which demands that the boron bear either a relatively positive or negative charge in comparison to the nitrogen atom with which it is associated. In this sense the first representation can be misleading, particularly when applied to molecules containing atoms of distinctly different electronegativity values; it is much less misleading in carbon chemistry This situation can be clearly represented by reference to a molecule such as LiClo When the coordination compound Li(NH3)x C1 is (45) J. R. Weaver, SO G. Shore, and Ro W. Parry, Jo Chem. Phys. 29, 1 (1958). WADC TR 59-207 65

formed, the Li+ ion is in all probability less positive than it was in LiCl, but the donation of electronic charge by each nitrogen does not mean that Li is of necessity negative with reference to either C1 or the attached nitrogen atoms. Electronegativity arguments, on the other hand, clearly suggest in agreement with chemical information, that the boron is the relatively positive end of the ammonia-borane molecule. On the basis of the above argument, the representation preferred in this discussion will be 5- 6+ H3THB3 The usual inductive effect arguments of organic chemistry indicate that replacement of a hydrogen atom by a methyl radical should result in donation of relatively more electronic charge to the binding atom, Thus in CH3H2NBH3 the methyl group should donate relatively more negative charge to the nitrogen than did the hydrogen of H3NBH3. H H CH3 H H-N B —-H H N B -H H H H 6-_ _6> + _s_ e sultarnt Simple Dipole Vectors (a+6)- 6+ Simple Dipole Vectors The naive vector model shown above indicates rather clearly that the effect of the electronic shift on the dipole is a most sensitive function of the molecular angles; indeed, with very reasonable assumptions regarding simplified "bond moments" and angles, one can rationalize the experimental fact that MeH2NBH3 has about the same moment as H3NBH3. The best available values are ammonia-borane 5.01D and methylamine-borane 5.02D. (See experimental subsection of this section.) With two methyl groups replacing hydrogen as in dimethylamine, the inductive effect of the two methyl groups can be compensated in large degree by the angular orientation of the methyl groups. Thus the argument is not inconsistent with the observed value of 4.95D. Finally, in trimethylamine the fact that the three methyl groups are symmetrically placed about the nitrogen would demand a reduction in the dipole value. C 3+ \ -H hi_ Resultant of 2 methyl vectors WADC TR 59-207 66 +

The observed value of 4.69 is consistent with, but does not prove the simple model. The value is also consistent with the greater volatility of the compound Me3NBH3 v The above picture does predict, however, something yet to be verified as fact. The compound H3NBH2CH3 could vell have a moment comparable to that of ammonia borane, but the value for H3NB(Cfa) should indeed be larger than that of ammonia-borane. The measurement of moments for compounds of this type is of interest for the testing of both naive and sophisticated electronic models of the compounds involved. Indeed, dipole moment values for these materials offer a rather valuable testing ground for many theoretical concepts. Since the amine-boranes are not sufficiently volatile as a group to permit the highly desirable vapor-phase studies, it has been necessary to make measurements in solution and to try to estimate the magnitude of the solvent effect in each recorded value. Unfortunately, methods for establishing these corrections are not well defined, and it is necessary to examine the overall question of dipole moment measurement for this group of compounds to obtain defensible data for the series. Many theoretical approaches to the question are found in the literature; selection of the most appropriate model for the system is still under examination. (b) The Experimental Measurement of Dipole Moments for the Amine-Boranes (1) Materials and the Preparation of Solutions for Measurement. (1.1) Benzene —Merck Reagent grade benzene, dried by refluxing with LiAlH4, was used to make up the benzene solutions and to standardize the dielectric cell. It was used without drying for rinsing the dielectric cell and as a secondary standard. The dielectric constant of the undried material is only.003 to.004 higher than that of the dried substance. Moderate exposure to air does not affect the dielectric constant of benzene. It is a much more convenient standard than dioxane. (1.2) Dioxane.-Preliminary purification of best quality commercial dioxane involved refluxing it with dilute HC1 for 12 hours, preliminary drying involving storage of the liquid for several days in contact with successive batches of KOH, and finally distilling the liquid from LiAlH4 Even the product obtained from the last step was not completely dry as was shown by a high value for its dielectric constant (2,255 as compared to an accepted value of 2.209). When the final product was refluxed with LiAIH4 for several hours, and then distilled, the measured value of the dielectric constant was in good agreement with the accepted literature value (2.209 +.003). Dioxane picks up water rapidly from the air and great care is required in its handling. Vacuum techniques were employed in general manipulation of this solvent~ (1.3) Ethyl ether. —Commercial "anhydrous diethyl ether" was dried by refluxing with LiAlH4. Observed values for the dielectric constant were WADC TR 59-207 67

4.205, 4.209, 4.225, and 4.207 as compared with an accepted value at 25~C of 4.235. This may indicate the presence of a nonpolar impurity but was not regarded as serious, since it should not affect the increment in dielectric constant resulting from the addition of solute. (1.4) Ammonia-borane. -Samples from laboratory stock (see Ref. 15c for preparation) were placed on a filter in the vacuum line and leached into the solution tube (Fig. 13a) with the solvent. No weights were taken in the ammonia-borane determinations so the concentrations were available only from analysis. Samples for analysis were withdrawn from the solution at the time that it was transferred to the dielectric cello Although 1H1 B, and N values were obtained, the values for hydridic H were the most consistent with dielectric constant measurements. The boron values were the least reliable of the three. Typical hydridic H/N ratios were 2.93, 2.97, and 3.00. The final run on ammonia-borane was made by weighing out samples in the dry box as described below. The resulting solutions were completely clear, confirming the purity of the sample and the dryness of the solvents. (1.5) Monomethylamine-boraneo —The product resulting from the procedure described in Section IIB1 was used after sublimation. Solution preparation is described below. (1.6) Dimethylamine-borane.-Samples of dimethylamine-borane, prepared by the method of Huff at Callery Chemical Company, were generously donated by Callery. The authors wish to thank Dr. Huff and Callery Chemical Company for these samples. The product was always sublimed before using. In preparing the solution the product was either sublimed directly into the preparation tube (Fig. 12a) or the sample was weighed out in the dry box. Analysis of representative samples confirmed the identity and purity of the product used. A representative ratio of hydridic hydrogen to nitrogen was 2.96; hydridic hydrogen to boron values of 3.15 and 3.08 were obtained. (1.7) Trimethylamine-borane. —Trimethylami.ne was stored above LiAIH4 to remove water and lower amines. It was then used to prepare trimethylamineborane by conventional procedures. The compound was sublimed several times before use. (2) Techniques of Measurement. — A rigidly standardized procedure was followed in preparing solutions.* The solution tube (Fig. 13a) was cleaned with nitric and chromic acids, rinsed with tapwater and distilled water until the water film remained continuous, then dried in the oven at 110~C, The tube was then evacuated, filled with. dry nitrogen under a slight positive pressure, removed from the vacuum system and capped quickly with a rubber syringe stopper Stopcock grease was removed using a tissue soaked in CC14, static electricity was removed by wiping the entire assembly with a barely damp chamois, and the entire assembly was weighed, using as a tare a similarly shaped tube and stopper *Except the modification for ammonia-borane; see Section IIB3(b). WADC TR 59-207 68

b. 0. C. c) Solution density flask. WADC TR 59-207 69

assembly which had been treated in a similar fashion. If the solute was to be sublimed into the tube, the lower part of the ground glass joint was lightly greased, the stopper removed and the tube quickly evacuatedo After subliming in the solute, the tube was again filled with dry nitrogen, removed and weighed as before. For certain methylamine-borane and ammoniaborane runs, the tube was introduced into the drybox, opened, filled, capped, removed, and weighed as before. The solvent was distilled from LiAlH4 into the special flask shown in Fig. 13bo Dry nitrogen was admitted through the side arm and the flask was removed from the line and capped using a rubber syringe stopper. Whenever solvent was to be removed from this vessel by means of a hypodermic syringe, N2 was admitted through the side arm to maintain a positive pressure of dry N2 above the liquid. Using a 50-ml syringe, solvent could be transferred into the weighed solution tube (Fig. 13a) containing the solute. By reweighing, the quantity of a solvent added could be determined. Discrepancies in density determinations indicate that the rubber caps will absorb as much as 40 mg of benzene. (2.1) Dielectric constant measurements. -Dielectric constant values were obtained using a precision heterodyne heat apparatus generously donated to The University of Michigan by the Chrysler Corporation. Solutions were introduced into and removed from the cell of the instrument by a syringe. After removal of the solution, the cell was filled twice with benzene, a reading being taken after the second time to check the constancy of the cell and measuring system. After removal of the benzene, dry nitrogen was passed through for five minutes, then solution was again introduced into the cello The actual measurement required about 5 minutes but no change was detected as a result of longer standingo (2,2) Determination of the cell constant for the dielectric measurements,-The cell constant is the ratio between condenser reading and dielectric constant for a known standard, Benzene was used as the standard, using the National Bureau of Standards value of 2.274 for its dielectric constant at 250. The value of the constant was confirmed by reproduction of the accepted value for the dielectric constant of rigorously dried dioxane (2.209 ~.003) using the cell constant obtained from benzene, A new value for the cell constant was needed whenever the cell was resilveredo (2.3) Refractive index measurements — Solution remaining in the syringe after filling the dielectric cell was used for the refractive index measuremento Since the contribution of the refractive index term to the dipole moment value is almost negligible for these highly polar solutes, no attempt was made to obtain the dispersion; only the refractive index for the sodium D line, nD, was obtainedo A Bausch and Lomb precision Abbe refractometer was used and the solution was introduced directly between the plates and measurements were made rapidly. WADC TR 59-207 70

(2.4) Density measurements. —To obtain density measurements on certain solutions, the special solution tube in Fig. 13c was substituted for the tube in Fig. 13a. The solute was sublimed into the tube and glass caps were placed over the joints during weighing. After insertion of the weighed quantity of solute, the solvent was carefully added, by a syringe. The tube capped with a syringe was equilibrated in a thermostat; then the level of solvent in the tube was carefully adjusted to that of a fine scratch on each constriction. The upper parts of the tube were heated with a heat gun to volatilize any excess liquid in that region. The tube was then closed with glass caps and weighed. Because of the extra manipulation and difficulty in mixing, the dielectric constant measurements on these solutions were in general less reproducible than those made on solutions from the tube shown in Fig. 13a. (3) Results. -A large numbex (more than 60) of runs have been made. The technique is a difficult one but as procedures were refined, results became both more precise and more accurate (as based on reproduction of accepted values of dielectric constants of solvents, etc.). The values meriting the highest confidence are summarized in Table XI. The orientation polarization was obtained from the Cohen-Henriquez equation: -orientation[dx do(e +2)2 -dx where Porientation = the molar orientation polarization of the solute Mo = molecular weight of solvent do = density of solvent eo = dielectric constant of pure solvent. (noo) = the refractive index of benzene extrapolated to infinite wave length and x, e, and n~ are the mole fraction, dielectric constant, and index of refraction at infinite wave length for the solution. The molar orientation polarization of the solute is related to the dipole moment of the solute by the equation: u = [akT =?orientationj 1/2 WADC TR 59-207 71

TABLE XI SUMMARY OF DIPOLE MOMENT DATA A. DIOXANE AS THE SOLVENT 1. Ammonia-borane Mole % Dielectric+ radnDI Dipole H3INYBH3 Constan-t Lx Lj Moment in Solution of Solution Solute 0 2.209. _ _ _..392 2.346 35.5 -.15+o03.293 2.313 35.2...402 2.351 35.4+.8 D.840 2.504 35.1+.4. 1lo455 2.713 34.6+.2 o4+.oo005 1.99 2.89 34.2.03 Best value from region of 34.5+,5 o4+.oo5 4.91 D of highest precision (14) -1 - Value extrapolated to infinite solution (Fig. 13) -- -- 5001 D 2. Methylamine -borane Mole % Dielectric ranDl:d l Dipole CH3H2NBH3 C onstant I I Moment in Solution of Solution S olute 0 2,215 -.659 2.2448 35.4+1o0. 1.754 2.817 34.3+1.0.. 0 2.212 - 1.122 2.608 35.3+.4.02+.005 -.18+.o03 0 2.208 -. -..710 2.458 35.3+.6 1.654 2.787 35.oi,.3.02t3. 005 - 2l.O 02 Best value (measured) 35.2+.3.02+.005 -.20+.o03 4.99 D Best value (extrapolated) 35.5.02i.005 -.20+-.03 5 02 D +Values calculated using cell constant obtained with benzene as standard. *Rate of change dielectric constant of solution with concentration. **Rate of change refractive index (NaD line) of solution with concentration. ***Rate of change density of solution with concentration. WADC TR 59-207 72

TABLE XI (Continued) 3. Dimethylamine -borane Mole of Dielectric1 e nD] __ Dipole (CH3)2HNBH3 Constant Moment in Solution of Solution Solute 2..08 Value below Value below ~ 279 2. 304 34.5+. 8.9- 28 repeated on repeated on 757 2.470 345+ 4 solutons: 2 solutions 1.464 2.715 34.6+.2 -.005+.005 -.32+. 04 0 2.12...... 570 2.409 34.5+. 6... 1.097 2.594 34.8+.*3 -.005+.005 -- -- Best value (measured) 34.6+.3 -.005+.005.32+.04 4.96 Best value (extrapolated) 34.3 -. 00+.005.32+. 04 4.93 4. Trimethylamine-borane Mole * Dielectric e ] nD] rFd1 Dipole (CH3)3NBE3 Constant L i LDxI Moment in Solution of Solution Solute o 2.2135 --.115 2.249 31+4 --.315 2.311 31.1+1.5 -- -.65o 2.413 31.0+.6 -.02+.005.850 2.479 31.3+1.0 --.88- 2.490 31+2 -.03+.02 1.10 - -- -- -37 2.42 - - - -.43 -- Best value (measured) 31.2+.5 -.02+.005 -.43+.05 4.71 Best value (extrapolated) 31.0 -.02~.005 -.43+. 05 4.69 5o Ammonia triborane 0 2.202 -- -- -- --.262 2.357 59.2+4.541 2. 529 60.4+.2 -- -- 1.115 2.878 60.6+1 +.09+.02 - -- Best value (measured) 60.6+1 +,09+.02 -- 6.55 D Best value (extrapolated) 60 +.09+.02 - -- WADC TR 59-207 73

TABLE XI (Continued) B. BENZENE AS THE SOLVENT 1. Methylamine -borane o Dielectric nDl Dipole Solute Constant j L L Moment in Solution of Solution Solute 0 2.274 --.404 2. 375 25.1- 662 2.426 22.9 -. 045 -- 1.o 479 2o 562 19.4 -.039 -- 2. Dimethylamine-borane 0 2.274 (std) -- -- ~359 2.377 28.7 -.07.396 2.385 28.0 -.07...743 2.474 26.9 --.788 2.485 26.8 -.07 -.34 -- 1.066 2.550 25.9 -.07.. 1.594 2.668 24.7 -.073 3. 356 3.058 23.4 -.066 3o Trimethylamine-borane 0 2.274 (std) -- -- --.218 2.341 30.6 -.105 -.22.301 2,363 29.6 -.11 --.3435.375 29.5 -.10.665 2.470 29.4 -.10.. 1.275 2.652 29.7,098. 1.814 2.819 30.0 -.0 98 o 2. 245 2.950 30.1 -.097 -- Best value (measured) 29.9+3 -.098+.003 -.22+.05 4.64 Best value (extrapolated) 29.7 -.098+.003 -.22+.05 4.62 WADC TR 59-207 74

TABLE XI (Concluded) C. DIETHYL ETHER AS THE SOLVENT 1. Ammonia-borane Mole % DielectricnD Dipole of Solute Constant _e d DMoment in Solution of Soltuion ax x Solute 0 4.205 -- -- 1.390 4.696 35.2 -.14 2. Methylamine-borane 0 4.205 - -- --.709 4.507 42.6 +.07 -- 3. Dimethylamine-borane 0 4.205 -- --.339 4.347 41.9 +.o8.730 4.510 41.8 +.0o8.. 1.266 4.730 41.5 +.o8 0 4.225 -- —..990 4.645 42.4 +.09... 2.07 5.075 41,0 +.087.. 4. Trimethylamine -borane 0 4.209 - -- --.522 4.398 36.2 +.07 -_ 1.019 4.583 36.7 +.07 -- WADC TR 59-207 75

where T = the absolute temperature = 298"K k = the Boltzmann constant = 1<58 x 1016 NA = Avogadro's number = 6.02 x 1023 The most precise dielectric data for each solute-solvent pair are shown in Fig. 14. The rate of change of dielectric constant with concentration, (ae/ax), is plotted against the mole fraction of solute in the solution. The variation in the (Oe/6x) with mole fraction is a point of some importance and is discussed in the next subsection. In Table XI the best value (measured) of (ae/6x) for each solute-solvent pair is the measured value in the neighborhood of 1-2 mole percent where the precision of the data is highest. The entry labeled "best value (extrapolated)" is obtained by extrapolating the appropriate lines of Fig. 14 to infinite dilution o (4) Discussiono —All the amine-boranes show a high dielectric constant, as was anticipatedo As seen in Table XII, the values for ammonia-borane, methylamine-borane, and dimethylamine-borane in dioxane are comparable. The value for trimethylamine-borane in dioxane is significantly lower than the other three. TABLE XII THE DIPOLE MOMENT OF THE AMINE-BORANES IN DIOXANE AS A SOLVENT -Solute Extrapolated Dipole Moment Ammonia-borane [H3NBH3] 501 Debyes Methylamine-boranie [CH3H2NBH3] 5.02 Debyes Dimethylamine-borane [(CH3)2HNBH3] 4 93 Debyes Trimethylamine-borane [(CH3)3NBH3 ] 4o69 Debyes Several interesting and as yet unresolved points are evident from Fig. 14. These are listed below. The monomethyl and dimethyl compounds show an abnormal variation of dielectric constant with concentration in benzene. The nature of the variation is what would be expected if the solute exists in both a monomeric and polymeric form. WADC TR 59-207 76

4240in diethyl ether 3836 de 0 function of solute concentration — )- H3NBHj - (CH3HJ)NBH3.-..i CH3H2NBH3-e- H3N83H7 38 - - (CH3)2HNBH3 e 36 ax ~32t~ _e_ e @ ~in dioxane F _ 26 24 -' 22t. ~20 ~ ~`'QID~"...002.004.006.008.01.02.03 MOLE FRACTION SOLUTE Fig. 14. e/x versus mole percent solute. WJADC TR 59-207 77

The slope of the line would indicate that the association is greatest for the monomethyl compound. The low value for the single measurement of ammoniaborane in ether might possibly be due to the same phenomenon. It may be worth noting that the monomethylamine-borane in benzene and the ammonia-borane in ether dissolved with difficulty, while ammonia-borane is insoluble in benzene. It is interesting and probably significant that the slope of the trimethylamine-borane line appears to be positive in each solvent. A small positive slope is indeed expected from theory for a normal solute. On the other hand, ammonia-borane seems to give a negative slope in dioxane, while the monomethyl and dimethylamine-borane lines are nearly horizontal in the same solvent. Such deviation from theory (if real) might suggest a small amount of association or of solute-solute interaction. Such an interaction was suggested in the solid by the heats of sublimation. The negative slope of dimethylamine in diethyl ether could also suggest solute-solute interaction. As expected from theory, (Oe/6x) is higher in ether with a dielectric constant of 4.2 than in dioxane with a constant of 2.209. In benzene, however, the value is slightly lower than in dioxane, the reverse of what one would expect from the dielectric constants. Specific solvent-solute interaction is indicated. 4. THE RAMAN SPECTRA OF THE AMINE-BORANES (a) Background Despite the rather active chemical interest in boron hydrides and their derivatives, the amount of spectroscopic work which has been carried out on these compounds has been disproportionately small. This disparity is particularly marked if one compares the data available on vibrational frequencies, assignments, force constants, molecular parameters, and so on for the derivatives of simple hydrocarbons with the corresponding data for the various boron-hydride derivatives. Experimental difficulties caused by the high reactivities of the boron compounds are responsible to a large degree for this situation; however, if studies are carried out on condensed phases and at low temperatures, a great deal of spectroscopic information can be obtained which can yield significant values for molecular constants and provide a basis for comparisons of chemical properties. A systematic study of all the amine-boranes is in progress; the data on ammonia-borane are most complete. (b) Vibrational Frequencies and Assignments of Isotopic Varieties of AmmoniaBorane Having C3V symmetry (1) Background.- Ammonia-borane is of some interest spectroscopically since it provides a molecule of relative simplicity and high symmetry containing the B-N "dative bond." Since the hydrogen atoms on the boron and nitrogen WADC TR 59-207 78

atoms are sufficiently dissimilar chemically that exchange does not occur at a measurable rate in solution, it is possible to prepare a number of deuterated derivatives in which the symmetry of the parent molecule is preserved. The possibility of arriving at an unequivocal set of assignments is thus excellent and the results should be of considerable value in the interpretation of spectral results for more complicated boron compounds. In addition, the data provide a basis for the further comparison of the frequencies associated with the BH3 group as they are modified in complexes with different bases. (2) Experiment. — Ammonia-borane was prepared by standard methods,(5c) the appropriate deuterated materials being used where necessary to get isotopically substituted molecules. Deutero-ammonia was prepared from Mg3N2 and heavy water while B2D6 was prepared from BF3 and LiAlD4 according to standard procedures. All samples were purified before use by removing all material insoluble in diethyl ether. In view of the nature of the compound, Raman spectroscopy was most suitable for use in the investigation and samples were prepared in liquid ammonia and in dimethyl ether using essentially the same procedure as for the work on the borohydride ion (see page 165). The concentration range customarily employed was between 3 to 5 molar and the samples were maintained at temperatures of approximately -35~ during an exposure. Infrared spectra of some of the samples were obtained as Nujol mulls but the compounds proved too soft to mull readily and the spectra obtained were not entirely satisfactory. Details of the spectroscopic equipment are given in Ref. 15h. (3) Experimental Results. -The Raman frequencies observed for five isotopic varieties of H3N:BH3 dissolved in liquid ammonia or ammonia-d3 are listed in Table XIII. In some cases where solvent interference prevented observation of a band in the liquid ammonia solution, values obtained from dimethyl ether solutions or from the infrared spectra of Nujol mulls have been substituted. Less information was obtained from the dimethyl ether solutions than the ammonia due to the greater number of solvent bands and consequently greater interference. With one exception, the positions of the observed bands were not shifted significantly on going from one solvent to the other. The exception is the frequency assigned to the B-N stretching vibration which appears at 787 cm-1 in liquid ammonia solutions of N3H:BH3. This frequency was found at 755 cm-1 in dimethyl ether solution, indicating an appreciable solvent interaction. probably with the ammonia. The exact nature of the interaction is not certain but it seems likely that it involves a weak N-Ho..N hydrogen bond. Comparison of the solvent effect on the N-H bands would give most direct support to this hypothesis but unfortunately interference by solvent bands in the NH3 solution was too great to allow the solute bands to be distinguished. However, in molecules containing the ND3 group, it was possible to distinguish the N-D variations of the complex both in liquid ND3 solution as well as in dimethyl ether solution and positions of the bands in the two solvents agreed within a few cm-K. The failure of the N-D bands to shift is evidence against the hydrogen bond hypothesis and the exact nature of the interaction remains uncertain. For consistency in the WADC TR 59-207 79

TABLE XTII ASSIGNMENT OF FREQUENC IES IN THE SPECTRA OF THE AMMONIA-BORANES Solutions in Ammonia or Ammonia-d3 Observed Frequencies, cm-1 Symmetry Descriptiona H3N'iBH3 D3\NBH3 H3NED3 D31T1 D3 D D3 and No (185b)c b b (3185 ) 2322 -3b- 2323b -- a1 V1 v-TN 2263 (2260b) 1641 1643 1645 a1 V2 v-BEl 1175 1171 857 840 848 al v3 F)-BH3 1060od - - - - al V4 6-NH3 787 754 737 708 712 al v5 v-BN (3300b) (2465b) -- ( 460 ) -- e V7 v-NH 2316 2313 1743 1715 1741 e 3s -iBH 1600d- -- -- - e v9 5-NH3....... — -- -- e v1o 5-BH3 ~ — 1006 917 917 936 e z11 p-BH3.. __ 648 574 (570) e v12 p-NH3 1575? A1 v3+v5 1767 1775 E V3+v~1 1833 1834 1871 A1 + E H colta H taminat ion 2319 on BD group? 2372 2372b 1716 Al 2vs (2380b) 2378b A1 + E 2v9 2163 2162 (2161) Apparently due 3040 to to interaction 3100 2252 2262 (2260)j with solvent a Symbols sued: v = stretching, 6 = deformation, p = rocking~ b Value taken from the Raman spectrum of a dimethyl ether solution of the material. c Bracket indicates a rather large uncertainty in the value, 5 to 10 cm-. d Value taken from the infrared spectrum of a Nujol mull of the material. wADc TR 59-207 80

discussion of the frequency assignments, values from the liquid ammonia solutions will be used. (c) Discussion X-ray studies(40) have shown that H3N:BH3 has an ethane-like structure with C3v symmetry (see Fig. 12). Such a molecule should possess twelve vibrational frequencies classified into five Al, one A2 and six E frequencies, the latter being doubly degenerate. The A1 and E frequencies are allowed in both infrared absorption and the Raman effect but the A2 is forbidden in both. The notation used in describing the frequencies and an approximate description of the mode of motion is given in Table XIV. TABLE XIV DESCRIPTION OF FUNDAMENTAL VIBRATIONS OF H3N:BH3 Frequency Symmetry Description V1 Al Symmetrical N-H stretch v2 Al Symmetrical B-H stretch v3 Al Symmetrical H-N-H deformation V4 Al Symmetrical H-B-H deformation V5 Al B-N stretch v6 A2 Torsional motion around B-N axis V7 E Asymmetrical N-H stretch vw E Asymmetrical B-H stretch v9 E Asymmetrical H-N-H deformation V10 E Asymmetrical H-B-H deformation V1l E BH3 group rock V12 E NH3 group rock The isotopic substitution of deuterium for hydrogen should result in a marked lowering of the frequencies involving hydrogen motions, the ratio approaching 1.41. The effect on the other frequencies, while easily discernible, should be much less pronounced. Since a number of the fundamentals have not been observed in the spectra, application of the more rigorous product rule to the isotopic frequencies to confirm assignments does not appear feasible. The two N-H stretching modes of the normal compound were readily identified as the bands at 3185 and 3300 cm-1 in spectra of dimethyl ether solutions, the former being assigned to the symmetric mode on the basis of polarization measurements. These two frequencies shift, respectively, to 2323 and about 2460 cm-l with deuteration on the nitrogen. Deuteration of the boron had virtually no effect on the N-H frequencies, indicating little coupling between the two ends of the molecule. It is of interest to note that both N-H stretching frequencies WADC TR 59-207 81

are some 190 cm lower than the correspondirng frequencies of b504 a-rd 3594 observed:for free TlNiVT in a dinethyl ether soltion of the same o.cncentration.:~oron- h ydogen stretc!>hing frequlencies occur in the region amoiZd 270 cmn1F but the ident.ffitifl (_-ictiol of the f- udamerntals in the present- case was complic-ated scmewhat by the presernce of overtones of the bend-ing modes enhancned in intensity by FerinL rescnance with the f -,dameilntals- The two bands at 2263 and 2316 cmtin. the spectrum of -the olormal molecule have been assigned, respectively9 to the cTym:et r. al.nd -onsymmetrc(-al modes largely on the basis of their intensities ar,~,d positi on Tlhey shift to approximately 1640 and 1740 cm-l upon deuteration. As in. the case of the N-il bands, the B-H frequencies are relatively unaffected by deute-ratEio; n of the other end of the molecule. Both B-H stretching frequencJ.es are relatively low, the positioi of the symmetrical mode virtually coincidirlng with the.correspo;d_~ding freq'ency of the borohydride ion2 while the ansymmetrical1 is about 70 cm-m higher than v3 of the same ion. Previous work has shoa. tshJat; the position of the B-I stretchinrg frequencies of the BE3 gr-oup is a good i.ndex of the strength of the interaeticn with the Lewis base., In the prese:nt molecule, it appears that azmmonia is nearly the equal of the hydride ion, H> -.n the extent of'its interaction with BH3s Du-e to solvent interference in both solvents, it was not possible to identi fy e:l,ther of the N-H or N-D deformation modes in solution. Results from an infrrared rnu.ll of the normal material indicate that both occur at positions virtuall(3y -ich. ged from thei.r positions in the free ammonia molecule. The symmeAt: ral:-H ard B-D deformation fu-ndamentals were identified at about 1175 and -Mi8 -m2, theiLr expected positions, on the basis of polarization measurements. The c:-S:,mmetrical mode was not observed, either because its intensity was too low c hberlase it was acc.i"dentally degenerate with the symmetrical frequency, The remainlng syrnmetric al frequency is v5, the B-N stretching vibratioan and Jit-, plosSitirlon.i.s of parti ular interest, partly because of the in.rlormationr whFic'h r.r cc1 y-ield regardn.rng the dative bond and also bec.ause there has been some disagreement about, its position in related molecules, The situation is partic:;larly favorahble for its identification in the present case since the only c t;h en fun:damentals belonging to the A, symmetry class involve hydrogen motions almc:-s4 excl. qusively and consequently can be distinguished with little chance of T.,h:e cnly bar.d which appears as a possible choice for v5 occurs at 787 cm — in Yth e 1:i..qid ammoni1a spectrum of the normal compound. This band is moderately stcc:-,?ng naod'well polar7ized UJpon deuteration of both ends of the molecule, -the h'ain^d shiftiC to 708 cmZ9 gi. ving a frequency ratio of 0.90, If the NHT and 3Bi3L halves of the molecule are considered as approximate point masses,, the ratio of isotopic frequencies should be given by the square root of the inverse ratio of._ redoied masse. fu,*, This mass is calculated to be 0.91 for the comparrson of the normal and completely deuterated molecules, the differerce of only 1% pro;riifnrg exoellenit support for the cor~recetness of the assignment. Similar omparsceins cuing the frequenoies of thue partially deuterated speeies gives agree merts of the same order Cof magnlitude. The B-N force corlstant cal1culated using WADI I~B 595.207 82

the diatomic approximation is 2.8 millidynes/Angstrom. The value given for this force constant by Badger's rule using the B-N distance of 1.6 Angstroms(40) is 2.4 md/A. The agreement is satisfactory considering the empirical nature of the rule, the large (~0.2 A.U.) uncertainty in the B-N distance, and the other approximations involved. It is of interest to note that the B-N distance in this molecule is the same within limits of error as the B-N distance in the BF3 amine complexes(46) and that consequently the B-N frequency in these latter compounds should be found in the region near 800 cm-1 instead of the region near 1000 cm-1 where it has recently been assigned.(47) Of the three fundamentals remaining to be discussed, one, v6, is inactive and no information was obtained as to its position. The other two involve rocking motions of the BH3 and NT3 groups and may be expected to have low intensities in the Raman effect. The values assigned to the BH3 and BD3 rocking frequencies, 1006 and 917 cm-1, respectively, are somewhat higher than the values assigned to the corresponding motions in the BH3CO molecule and are not to be considered as final. In particular, the possibility exists that one or the other may be the asymmetric deformation mode which has not been assigned. The NH3 and ND3 rocking motions, on the other hand, appear fairly certainly identified as the bands at 648 and 574 cm-1 since they are the lowest in the spectra and nitrogen deuteration affects them appreciably while boron deuteration does not. The least certain of the above assignments are those for the deformation and rocking modes of the BH3 group. At the present time, the Raman data do not appear adequate to allow a clear decision to be made. Information from the infrared spectra of thin films at low temperatures, if the films can be prepared, may supply the necessary evidence, and it also appears that a normal coordinate treatment will be necessary for final confirmation. 5. THE NUCLEAR MAGNETIC RESONANCE SPECTRUM OF Me2HNBH3 The nuclear magnetic resonance spectrum of molten Me2HNBH3 was obtained. Three peaks attributable to B-H links and one to the N-H resonance were visible. The study is still incomplete. C. Phosphorous Trifluoride-Borane 1. BACKGROUND Phosphorus trifluoride-borane was first prepared in this laboratory under (46) J. L. Hoard, S. Geller, and W. M. Cashin, Acta. Cryst. 4, 396 (1951). (47) a) J. Goubeau and H. Mitschelen, Z. phys. Chem. 14, 61 (1958). b) H. Luther, D. Mootz, and R. Radwitz, J. Pract. Chem. 277, 242 (1958). WADC TR 59-207 83

sponsorship of WADC.(41) The molecule was of considerable theoretical interest since earlier arguments of Chatt had suggested that boron and aluminum would be incapable of forming coordination compounds with PF3 because they lacked electrons. Since publication of the original paper, it has been suggested by others, without any expe:rimental or real theoretical support, that the molecule is held together by "T-bonds" involving delocalization of the B-H electrons. Such speculation seems to do little beyond clutter the literature. Of the more immediate concern is an experimental study of the structure and chemistry of this molecule. Parry and Bissot assumed an ethane-like structure for the molecule. The reaction of the molecule with trimethylamine was that expected from conventional theory. If stoichiometric quantities of F3PBH3 and trimethylamine were mixed, the trimethylamine displaced the PF3 quantitatively to form (CH3)3NBH3 and free PF3. If an excess of trimethylamine were used, the liberated PF3 reacted in a one-to-one ratio to give gases and a solid which was never fully characterized. When F3PBH3 reacted with excess ammonia over the temperature range -178o to 8Oo, five molecules of ammonia seemed to be picked up by each molecule of the borane adduct. The reaction with stoichiometric quantities of ammonia was not carried out and the nature of the ammonia reaction product was never established. 2. THE RAMAN SPECTRUIM OF PHOSPHORUS TIRIFLUORIDE-:BORANE AND PHOSPHORUS TRIFLUORIDE -BORANE-d3 More definitive structural information of F3PBH3 and F3PBD3 has been obtained from a study of the Raman spectra of these molecules. All observed bands were assigned and the ethane-like model, assumed in the original work was supported. The molecule has C3V symmetry. The samples prepared as described after fractionation on the vacuum line were condensed into capillary Raman tubes which were then sealed off and maintained at liquid nitrogen temperatures until used. Details of the experimental equipment have been given previously.(48) The spectra of the liquid were obtained at -80~C, qualitative polarization measurements being made on the hydrogen compound only, using the two method and polaroid cylinders. The data reported in Table XV represent the averages from several spectra, the estimated probable error being approximately 1 cm-1 except where indicated. Reference has been made previously(47) to the chemical similarity between PF3 and carbon monoxide, and to the similarity in properties between phosphorous trifluoride-borane and carbon monoxide-borane. In view of the known structure of H3BC0,(49) it would therefore be expected that the phosphorous complex would (48) Go Lo Vidale and Ro C. Taylor, J. Am. Chem. Soc. 78, 294 (1956). (49) Gcrdy, Ring, and Bury, Phys. Rev, 78, 512 (1950)e WADC TR 59-207 84

TABLE XV OBSERVED RAMAN FREQUENCIES AND ASSIGNMENTS FOR LIQUID PF3BH3 AND PF3BD3 AT -80~0C F re quency Intensity in cm-1 and Poin cm and Symmetry Assignment PolarizaP'F3BH3 PF 3BD 3 tion 197 169 m, dp e'V12 PF3 rock 370 362 vw e Vll F-P-F deformation 441 421 m, p al V5 F-P-F deformation 607 572 s, p al V4 P-B stretch 697+2 -- vw e vio BH3 rock 799 vw Diborane? 886+5 -- vw, p? A1 2v5 2x441=882 920 -- w, p A1+A2+E v8-v12 1117-197=920 944 944+5 m, p? a3 P-F stretch 957+3 958+2 m, dp e P-F stretch 1040+3 -- W, p A1 441+607=1048 1077 -- w, p al H-B-H deformation 1117 807 s, dp e Vs H-B-H deformation -- 1756 w A1+A2+E V8+v9 807+958=1765 _- 1797 vvw E v2+vg? 842+958=1800 -- 1980 vvw Diborane-d6? 2112 -- vvw Diborane? 2140+4 1672+2 vw A1 2v2 2x1077=2154 2247+2 1602 vw A1 2v8 2x1117=2234 2328+4 -- vvw 2v3+v5? Calc 2329 2385 1717 vs, p al vl B-H stretch 2455 1845 vs, dp e 7 B-H stretch -- 2431 w B-H stretch (H impurity) 2530 -- vvw Diborane? 2655+5 -- vvw A1+A2+E v7+vl2 197+2455=2652 WADC TR 59-207 85

have an ethane type configuration with C3v symmetry. The vibrational frequencies of such a structure are twelve in number and can be thought of in terms of the four vibrational frequencies of each of the two halves, considered as free molecules with C3vr symmetry, plus four- 77'jr p. rising as a consequence of the bond between the apex atoms of the two pyramids. Since free BH3 is not known, reference can be made to the Raman frequencies of H3BCO which have been determined recently.(50) In the B-H stretching region, two strong bands appear in the spectrum of PF3BH3 at 2385 and 2455 cm-4 which appear to be al and e type modes, respectively. Corresponding frequencies are found at 2380 and 2434 cm-1 in the spectrum of liquid H3BC0. Boron-hydrogen deformation bands are observed at 1077 and 1117 cm-1 for the al and e modes which do not differ greatly from the corresponding bands at 1073 and 1101 cm-L in the carbon monoxide complex. The frequencies associated with the PF3 group likewise show a close similarity in pattern to those found in the Raman spectrum of liquid PF3, which was also obtained in the present work. In this case, the al and e P-F stretching frequencies at 832 and 874 cm-1 shift to 944 and 958 cm1l respectively, in the complex while the a1 and e bending frequencies shift from 484 and 351 to 441 and 370 cm-', respectively. The remaining modes may be described as a P-B stretch, BH3 and PP3 rocking motions, and the inactive torsional mode. The first was easily identified as the strong polarized band at 607 cm-1 while the low depolarized band at 197 cm-1 is certainly the PF3 rock. The BH3 rock was assigned to the rather weak band at 697 cm-l. No information as to the torsional frequency was obtained. Two fundamentals of PF3BH3 were not observed directly. The position of the first, the symmetrical B-D band, is estimated at 842 cm-1 from its first overtone, and the position of the second, the BD3 rock, is estimated to be close to 600 cm-1 from the comparison with the frequencies of D3BCO. The calculated product rule ratios using these estimated values are 1.98 and 2.53 for the a1 and e classes, respectively, which may be compared with theoretical values of 1.97 and 2.61. The spectra thus can be interpreted satisfactorily in term of the C3v or ethane-like model. 35 THE REACTION BETWEEN F3PBH3 AND AMMONIA As noted in an earlier section, the reaction between F3PBH3 and trimethylamine can be summarized by the equations: Me3N + F3PBH3 -. Me3NBH3 + PF3 PF3 + ME3N(excess) > New Products (50) R. C. Taylor, J. Chem. Phys. 26, 1131 (1957); this report, p. WADC TR 59-207 86

It is tempting to extrapolate the behavior of NMe3 and to predict that a similar reaction might take place if NH3 were used in place of NMe3, but a wellsupported axiom of coordination chemistry indicates that the chemistry of ammonia cannot be predicted from observations on other amines. The following studies were undertaken to establish the nature of the F3PBH3 - NH3 reaction. Addition of one mole of NH3 to F3PBH3 without solvent over a temperature range of -780C to -1700C did not result in liberation of PF3 or the isolation of H3NBH3. Only NH4F was formed and 1/4 of the original F3PBH3 was recovered from the system. It was also found that PF3 reacted with ammonia at -111~C to produce NH4F as one of the products. The NH3 + F3PBH3 reaction was then tried under somewhat milder conditions using as excess of NH3. Diethyl ether was used as a solvent and a controlled low temperature, ranging from -111~C to -35~C, was employed. Under these conditions the reaction is best represented as: 6NIH-3 + F3PBH3 Et. t 4 3NH4F + (H2N)3PBH3 -111 to -350C The latter compound, triamido-phosphorus-borane, is a:new ether-soluble crystalline solid with a characteristic X-ray diffraction pattern. It is stable when pure but undergoes decomposition when contaminated. Analytical data for characterization show: H2 = 32.0 mmoles/g; N = 45.5 ~.5 mmoles/g; B = 12.1 mmoles/g. Theoretical values for (H2N)3PBH3 are H2 = 32.3 mmoles/g;N = 45.3 mmoles/g; B = 11.55 mmoles/g. It has been pointed out by others that the BH3 group isisoelectronic with the oxygen atom and indeed certain early molecular orbital arguments compared B2H6 and the oxygen molecule. While the original arguments were rendered inapplicable.by geometric differences, one might extrapolate the above unestablished argument to note a similarity between F3PO and F3PBH3 with ammonia: 6H3N + F3PBH3 --- 3NH4F + (NH2)3PBH3 6H3N + F3PO —-- 3NH4F + (NH2)3 PO The above formalistic relationsip does not prove the validity of the analogy. A combination of more conventional reaction processes might well account for the observation. For example, one might write NH3 NH3 + F3PBH3 -- H3NBH3 + PF~-' NH4F + F2PNH2BH3 Such a process could well account for the observations of the one to one reaction. Alternatively, one might write NH3 + F3PBH3 - H3NBH3 + PF3 WADC TR 59-207 87

PF3 + 6NH3 —-> 3NH4F + P(NH2)3 H3BNH3 + P(NH2) 3- (H2N)3PBH3 + NH3 More detailed mechanistic studies are underway, but the formalistic analogy which serves at least as a reaction guide is still of interest and is worthy of more explicit testing. The work is continuing. D. Raman Spectra, Vibrational Assignments, and Force Constants for H3BCO and D3BCO Spectroscopic studies of compounds such as H3BCO are of interest since theycan provide information not only on the BH3 group but also on the nature of the acidbase "interaction between the BH3 and the CO. In the present work, the Raman spectrum of H3BCO, has been obtained, a complete assignment of fundamental frequencies has been made, and a set of valence force constants has been determined which agrees with experimental data for four isotopic combinations. Onl one previous spectroscopic paper on H3BCO has appeared, a paper by Cowan(51) reporting the infrared spectrum of the vapor. Five fundamentals reported by him agree with the values found in the present work, two he did not observe, and his assignment of the last appears incorrect. No data for the D3BCO molecule have been found. 1. EXPERIMENTAL Both H3BCO and D3BCO were prepared by reaction of B2H6 or B2D6, respectively, with CO in a sealed tube at several atmospheres pressure. After several days, the tubes were opened and the contents carefully fractionated at low temperatures on the vacuum line. After fractionation, the sample was distilled into the Raman cell which was then sealed off. To reduce thermal decomposition, the vapors were never allowed to come in contact with surfaces warmer than about -500C during all transfer operations. The spectra obtained showed no bands attributable to diborane or CO which would be produced as decomposition products. The sample of HSIBCO examined was about 1 ml in volume while the D3BCO was about 0.2 ml. During the exposures, the samples were maintained at approximately -80~C, at which temperature the decomposition occurring in the liquid is negligible. A general description of the apparatus and spectrograph has been given previously.(48) Exposure times varied from ten minutes to three hours using Eastman 103a-J plates. Measurements were made with a comparator directly on the plates and on enlarged tracings made with a Leeds and Northrup microphotometer. The estimated probable (51) R. D. Cowan, J. Chem. Phys. 18, 1101 (1950). WADC TR 59-207 88

error for most lines reported is approximately 1 cm-'. 2. EXPERIMENTAL RESULTS The experimentally observed frequencies for H3BCO are listed in Table XVI and those for D3BCO are listed in Table XVII. Tracings of spectra of the two substances selected to show the fundamentals most clearly are shown in Figs. 15 and 16. The agreement between the frequencies here reported and those found previously in the infrared of the vapor is very satisfactory, the differences at most amounting to a few cm-1 and being well within the normal shifts in frequency observed in the transition from vapor to liquid. Several overtones and combinations were observed on some of the longer exposures on H3BC0 which are not shown in the figure. No bands attributable to diborane or CO were observed in any of the spectra indicating a fairly high purity for the compounds. However, a weak band was observed at 2411 cm-1 in the spectrum of the deuterated compound which indicates a small amount of hydrogen to be present. 3. ASSIGNMENTS The H3BC0 molecule has C3V symmetry which predicts eight fundamentals, all active in the Raman effect, which are either totally symmetric (Al) or doubly degenerate (E). Previous work on boron compounds has shown that B-H stretching frequencies fall in the range between 2000 and 2600 cm-1. Three frequencies appear in this range in the H3BCO spectrum. Deuterium substitution affects only two, however; and, on the basis of their polarization characteristics, the band at 2380 cm-1 is assigned as vl and the band at 2434 cm-1 as v5s In the D3BCO spectrum, the asymmetric frequency vs occurs at 1825 cm-1 but the position of v1 cannot be determined exactly because of Fermi resonance with the overtone of the fundamental at 860 cm-1. The two members of the Fermi doublet occur at 1678 and 1749 cm-1. The latter is assigned as the overtone and the former to the fundamental on the basis of the BO1 satellite appearing on the high-frequency side of 1749 at 1777 cm-'. Comparison of the intensities of the two bands indicates that the coincidence between the overtone and the fundamental is very close and consequently the unperturbed value of vl probably is not far from 1700 cm-, on the high-frequency side. vl was not observed in the infrared spectrum of the hydrogen compound(51) but its predicted value agrees with that given above. The third band in the 2000 cml region is immediately identified as the C-0 stretch, both from its nearness to the carbon monoxide frequency at 2143 cmrl and from the fact that deuterium substitution does not shift its position. The assignment of the band is further confirmed by its polarization characteristics. The situation with regard to the B-H bending modes is somewhat more complicatedo In the hydrogen compound, a triplet is observed in the 1100 cml1 region with maxima at 1073, 1101, and 1133 cm-l, the first being the most intense and probably polarized~ In the ~3BCO spectrum, two bands of approximately equal intensity appear at 808 and 860 cm-1, the latter having a weak satellite on its high-frequency side at 881 cm-l1 Since polarization measurements were not made WADC TR 59-207 89

TABLE XVI OBSERVED RAMAN FREQUENCIES OF LIQUID H3BCO AT -80~C Band Position (in cm'1) Intensity Assignment 317 m vS-e fjndamental 632 vvw 2 v8 692 w (pol.) v4-a fundamental 705+2 vw v4'-B10 isotopic species 816 w v7-e fundamental 1073 s (pol.?) v3-al fundamental 1101 m v6-e fundamental 1133 w v7+vg 1626 vvw 2V7 1761 vvw V3+V4 1887 vvw v6 +V7 2129 vw 2V3 2169 s (pol.) v2-al fundamental 2380 s (pol.) vl-al fundamental 2434 s v5-e fundamental 2703 vvw v1+v8 TABLE XVII OBSERVED RAMAN FREQUENCIES OF LIQUID D3BCO AT -80~C Band Position Intensity Ass ignment (in cm-1) 264 w v8-e fundamental 619 m v4-al fundamental 706 w v7-e fundamental 808 m v6-e fundamental 860 m v3-al fundamental 881 w v3'-B10 isotopic species 991 w? 1678 s vl-al fundamental 1749 s 2v3 1777 w 2v3'-B1O isotopic species 1825 s v5-e fundamental 2169 s v2-al fundamental 2411 w B-H stretch WADC TR 59-207 90

BH$CO Fig. 15. The Raman spectrum of H3BCO at -80~C. BD3CO I I I I i 0 1000 2000 CM"' Fig. 16. The Raman spectrum of D3BCO at -80~C. TWADC TR 59-207 91

on the deuterated spectrum, the product rule plus the results of the normal coordinate treatment were necessary to arrive at a satisfactory assignment. Fortunately, the dimensions and moments of inertia of the four possible isotopic molecules of C3v symmetry, H3BliCO, D3Bll-C0 H3BLOC0O D3B:OCO0 have been determined from microwave results(49) so that the theoretical product ratios can be calculated with no assumptions. The closest agreement with the theoretical values is obtained by assigning V3 to the 1073 cm-1 and v6 to the 1101 cm-1 band in the hydrogen compounds and v3 to the 860 cm-1 and v6 to the 808 cm-1 bands of the deuterium species. Confirmation for the interchange in the relative position of the two bands in the deuterium case comes from the normal coordinate calculation of the B10 shift. For the Al frequency of the deuterated molecule, the calculated shift is 22 cm-1, while for the E frequency it is only 5 cm-1, a separation that would not be resolved by the equipment used. The presence of a weak satellite 21 cm-1 higher than 860 cm-1 is therefore accepted as additional evidence that the higher band is the Al frequency, the satellite being assigned as V3 of the BO0 isotopic species. In the hydrogen compound, the B10 isotope shifts are calculated to be +12 and +4 Acm- respectively for v3 and v6. Since the observed spacings between the members of the observed triplet are 28 and 32 cm-1, it appears that neither can easily be assigned to the B10 species. However, the combination of the two E modes at 816 and 317 cm1l has a calculated value of 1133 cm-1 and the correct symmetry to resonate with an E fundamental and borrow sufficient intensity to appear as a weak band. The band at 1101 cm-1 accordingly is assigned as v6. Cowan, in his infrared work.51)assigned v6 to a band at 1392 cm-!. No band at this position was observed in the Raman spectrum, and it appears that the infrared band most likely is V3 + 98 which, from the Raman data, is calculated at 1390 cm-1. He assigned V3 to a band at 1105 cmil'hut comments that this band had some peculiar features which now appear explained. The remaining fundamentals may all be classed as skeletal modes, The only polarized, fairly intense band left occurs at 692 cm-1 in the H3BCO spectrum and shifts to 619 cm-1 upon deuteration, This is assigned to v4 in agreement with the infrared results. A satellite was observed at 705 cm-! in the more intense exposures on the hydrogen compound and, on the basis of a calculated shift of +16 cm-1 from the force constant treatment, is assigned to v4 of H3BloCO. The corresponding shift in the deuterated molecule is calculated to be only 5 cm1 and accounts for the failure to observe a satellite to the 619 cm"l band. The two fundamentals v7 and v8 can be considered as bending motions of the axial chain of atoms. The second, v89 the B-C-0 deformation, is to be expected at a rather low frequency inr view of the masses of the atoms involved. It consequently is assigned to the moderately intense depolarized band at 317 cm-1 in the H3B T0 spectrum. This fundamental was not observed in the infrared work, but its position was predicted quite accurately. In the deuterated spectrum it appears at 264 cm-o. The last fundamental, v7, which is most simply described as a- BH3 rock, is assigned to 816 cm1l partly by a process of elimination and partly from the infrared evidence. The corresponding band at 706 cm-1 in the deuterated compound is rather weak, but the correctness of the assignment is substantiated by the WADC TR 59-207 92

product rule calculations (Table XX). 4. NORMAL COORDINATE TREATMENT Cowan(51) carried out a normal coordinate treatment of the H3BCO molecule based on the results of his infrared study and obtained a set of force constants which produced reasonably satisfactory agreement with his assignments. However, in view of the incorrect assignment for v6 and of the fact that data on the deuterated molecule were not available, it would appear that a better approximation can now be obtained. Since his equations did not include interaction force constants, the molecule was reanalyzed using the FG method of Wilson and the following symmetry coordinates: A1 species: SO = (3 + 3a2)-1/2 x[Acl12 + A9c23 + Acz31 + a(AP1 + AP2 + AP3)] - 0 S1 = AT S2 = AR S3 = 3-y/2 (Arl + Ar2 + Ar3) S4 = (3 + 3a2) 1/2 [-a(AcE12 + AEc23 + AQ31) + AS1 + A~2 + AP3] where c cos - COS 2 E species: S5 2-1/2 (Ar2-Ar3) S6 2-1/2 (AP2-AP3) S7 = 2-1/2 (Ac31-Aa12) S8 = Abx In terms of the molecule parameters, T refers to the C-O bond, R to the B-C bond, ri to the ith B-H bond, Cq'j to the H-B-H angle between ri and rj, Pi to the ith H-B-C angle and 6x to the B-C-O angle. The equilibrium values for these parameters taken from the microwave work(49) are as follows: T = 1.131A, R = 1.540A, r = 1.194A, a = 113~52', P = 104037', b = 180~. The elements of the inverse kinetic energy (G) matrix were evaluated from the tables of Decius(52) and the note by Ferigle and Meister.(53) As a check on (52) J. Co Decius, J. Chem. Phys. 16, 1025 (1948). (53) S. Mo Ferigle and A. G. Meister, J. Chem. Phys. 19, 982 (1951). WADC TR 59-207 93

the correctness of the equations, the force constants of Cowan were substituted into the secular equation and the roots were found to agree with his calculated values. The present calculations were carried out in terms of the symmetry force constants, Fi and Fij, where the single index indicates principal force constants and the double, interaction constants. It was found that the calculated frequencies were rather insensitive to most of the possible interaction force constants but were quite sensitive to two. It was not necessary to give values to any of the insensitive constants to obtain a satisfactory fit, and the potential energy, therefore, is given adequately by the following expression containing only ten constants: 2V = F1S12 + F2S22 + F3S32 + r2F4S42 + 2RF24S2S4 + F5S52 + r2F6S62 + r2F7S72 + RTF8S82 + 2R2F68SeS8 This reproduces the sixteen frequencies of the H3B11CO and D3B"1CO molecules with a standard deviation from the observed values of 0.3% which undoubtably is less than the error introduced by the harmonic approximation. Since there are four product rule relations among the sixteen fundamental frequencies, twelve independent data exist, and the problem is slightly overdetermined. By inspection of the potential function, it will be seen that the F matrices are nearly diagonal, only one off-diagonal element appearing in each symmetry block. The off-diagonal element in the Al block is F24 linking the symmetrical bending coordinate and the B-C stretching coordinate, while the off-diagonal element of the E block, F68, links the BH3 rocking coordinate with the B-C-0 deformation. In both cases, the two lowest frequencies in the respective symmetry classes are involved, and considerable mixing of the two motions is indicated. The valence force potential constants are related to the symmetry force constants through the elements of the F matrix and in principle may be obtained by solving a set of simultaneous equations. Frequently, however, the number of valence force constants exceeds the number of equations available so that a number of constants must remain undetermined. In such circumstances, the practice has been either to tabulate appropriate unresolved combinations of the valence force constants or to make arbitrary but reasonable assumptions regarding the magnitude of certain of the constants which permits the assignment of numerical values for the rest. The usual assumptions involve setting some of the interaction constants equal to zero; the particular ones selected being chosen on the basis of previous experience or other physical arguments. An alternative and perhaps more objective way of introducing enough arbitrary assumptions to allow giving values to all valence force coordinates is to retain the redundant symmetry coordinate, So, in the transformation from the f matrix to the F matrix and then to equate the elements of the F matrix involving the redundant coordinate to zero. In the present case, this procedure provides enough equations so that values for all constants can be given. These are tabulated in Table XVIII. It must be emphasized that this method is just as arbitrary as the older procedure; the only justification is that it is objective and can be applied in a consistent manner to molecules of different structures. WADC TR 59-207 94

TABLE XVIII FORCE CONSTANTS FOR THE H3BCO MOLECULE Symmetry Valence Value Value Force o Force o Constant (md/A) Constant (md/A) F1 17.9800 kT 17.9800 F2 2.7900 kR 2.7900 F3 3.2920 kr 3.1587 F4 0.3486 ka 0.2768 F24 0.1753 k5 0.2801 k5 0.2744 F5 35.0920 krr 0.0667 F6 0.2203 Ka5 -0.0703 F7 0.2434 keg -0.0339 F8 0.2744 ka-=k' -0,0567 F68 0.0793 kRa -o00633 kR5 0.0790 ks o0.0647'k k k k -kt k= k -0 kRT = krR = kr = k r = = k k = kr = The values calculated for the fundamental frequencies of carbon monoxideborane by the normal coordinate treatment are compared with those observed in Table XIX. Since two frequencies attributed to B10 molecules were observed, the calculations were extended to include both boron isotopic species. The calculated values for the B10 molecular frequencies are given in Table XIX in the form of the shifts in frequency from the calculated values of the corresponding B11 frequencies. As an additional check on the calculations, the Bll/Bl~ product rule ratios were calculated using a set of B10 frequencies obtained by adding the calculated isotope shifts to the experimentally observed B11 frequencies. The agreement as shown by the numerical values in Table XX is quite satisfactory. Comparisons of force constants from different molecules are not always as meaningful as one would like because published values inevitably depend on such factors as the type of potential function used and the number of interaction constants retained, the closeness of fit secured, the magnitude of the anharmonicity errors, and so on. Bond stretching force constants are the least affected by such variations and when derived by the usual valency force field can be used qualitatively in the same fashion as bond energies and bond lengths to give an indication of the electron density concentrated in a given bond. In the carbon monoxide borane molecule one of the interesting observations is the small effect which the presence of the borane group has on the C-O bond. The force constant of the C-O bond in carbon monoxide gas calculated from the observed infrared frequency of 2143 cm-l is 18.5 md/A. This is decreased only to 17.98 md/A WADC TR 59-207 95

TABLE XIX \- COMPARISON OF? OBSERVED AND CALCULATED VALUES OF THE FUNDAMENTAL FREQUENCIES \O FOR THE VARIOUS ISOTOPIC SPECIES OF THE H.BCO MOLECULEa 0 HRB -1CO D, B 10C H3B 10CO D3B 10CO Fundamental Obs Calc Diff. % Obs Calc Diff. Ca Ob Calc Ob Shift Shift A1 vl 2380 2378 -2 0.1 (1700)b 1701 -- + 2.5 + 4.3 v2 2169 2169 0 0 2169 2169 0 0 0.4 -- 0.4 -- V3 1073 1070 -3 0.3 860 863 +3 0.3 12.1 -- 22.5 21+2 \o1 V4 692 691 -1 0.1 619 621 +2 0.3 16.0 13+2 4.8 E v5 2434 2431 -3 0.1 1825 1827 +2 0.1 15.0 -- 21.3 v6 ( 101)b 1101 -- 808 805 -3 0.4 3.6 5.4 -- v7 816 816 0 0 706 704 -2 0.3 6.3 -- 3.2 -- v8 317 315 -2 0.6 264 265 +1 0.4 0.0 -- 0.1 -- a Standard deviation for 14 frequencies = 0.35%. b The data enclosed in parentheses are for fundamentals disturbed by Fermi resonances and are estimated values.

TABLE XX COMPARISON OF PRODUCT RULE RATIOS FOR VARIOUS ISOTOPIC COMBINATIONS OF THE H3BCO MOLECULE Isotope IspHeld Isotopes Symmetry Frequency Product Ratio Held Constant Compared Class Theoret. a Calcb Dev. B"-1 /D Al 1.931 1.953 1.1% B1" H/D E 2.513 2.522 o.4 B10 H/D Al 1.929 1.951 1.1 B'l H/D E 2.498 2.507 0.4 H B" /B o Al 1.036 1.036. 0 H B"l/B'l E 1,017 1.017 0.0 D Bll/Blo Al 1.037 1.037 0.0 D Bll/Blo E 1.024 1.023 -0.1 a Calculated from moments of inertia of reference 5 and the masses involved. b The frequencies for the Bll molecules were those observed; the frequencies for the B'0 molecules were obtained by adding the shifts determined in the force constant treatment to the experimental values of the B11 frequencies. in H3BCO, whereas in nickel carbonyl the C-O force constant is 15.9 md/A and in carbon dioxide it is 15.5 md/A. This small change is consistent with the small (ca 0.001 A) increase in CO bond length and with the chemical instability of the molecule. However, the adjacent B-C bond is not as weak a bond as one might expect from this, its force constant of 2.79 md/A being only slightly less than the value 2.9 md/h predicted by Badger's ruleo Unfortunately, there appear to be no other data available on B-C bond constants for further comparisons. However, the calculated value for H3BCO does not seem entirely consistent with the rather large amount of no-bond character which has been proposed(49) for this bond. The two molecules, perhaps, with which it would be of most interest to compare B-H stretching force constants are diborane and the borohydride ion, A value of 3.42 md/A has been used in a normal coordinate treatment of diborane;(55) however, this figure was transferred unaltered from an earlier treatment of borazine,(56) and it is of doubtful significance to do more than note that it is of the same magnitude as the value found in the present work. A more reliable (54) B. L. Crawford and P. C. Cross, J. Chem. Phys. 6, 525 (1938). (55) Ro P. Bell and H. C. Longuet-Higgins, Proc. Roy. Sc. (London) A183, 357 (1945). (56) B.o L.o Crawford and J. To Edsall, J. Chem. Phys. 7, 223 (1939). WADC TR 59-207 97

value has been calculated recently by Scherer(57) in an investigation of the infrared intensities of the vibrational bands of diborane but this datum has not yet been published. The value of 3.16 md/A found for the B-H constant in H3BCO appears to be significantly greater than the value 2.77 md/i calculated for the borohydride ion (see page 172). This difference can be interpreted as indicating that the electronic configuration around the boron atom is somewhere intermediate between the trigonal sp2 hybridization expected for the hypothetical free BH3 molecule and the tetrahedral sp3 hybridization in the borohydride ion and as indicating that the lone pair of electrons on the carbon atom have not been utilized fully in forming the boron-carbon bond. The same conclusion, of course, can be drawn from the larger than tetrahedral angle (113~0) in H3BCO. Finally, it may be pointed out that the experimental product rule ratios for the hydrogen-deuterium substitution listed in Table XX are greater than the theoretical values, even though the differences are not large. Exact agreement, of course, is not to be expected since the observed rather than harmonic values have been used in calculating the ratios. However, the normal effect of anharmonicity is to cause the experimental ratios calculated in this way to be less than the theoreticall58) In view of the weight of other evidence and of the fact that all experimental ratios are greater than the theoretical, it does not seem probable that there is an error in the assignments. The difference, therefore, may also reflect the specific electronic structure of the molecule or the BH3 group and, if so, should be found in other molecules containing the borane group, E. The Structure of the Diammoniate of Borane-Carbonyl 1. BACKGROUND When borane-carbonyl was first prepared by Burg and Schlesinger,(59) it was noted that the complex reacted with trimethylamine to give trimethylamine-borane and free CO, but the compound appears to react in a different fashion with ammonia. A solid product containing 2 moles of ammonia per mole of borane-carbo.yl was formed. The structure of this so-called "diammoniate of borane-carboryl" remains a mystery, and the mystery is being examined in this laboratory. (57) J. Scherer, thesis, Univ. of Minn., April, 1958. (58) F. Halvorsen, Revs. Modern Phys. 19, 87 (1947). (59) A. B. Burg and H. I. Schlesinger, J. Am. Chem. Soc. 59, 780 (1937). WADC TR 59-207 98

2. THE MODEL In rationalizing the behavior of borane-carbonyl, a formalistic analogy between F3PO and F3PBH3 was noted. The BH3 group seemed to be serving in a formal sense as a coordinated oxygen. An extrapolation of the above argument would suggest that H3BCO might be analogous under some circumstances to CO2 and the diammoniate of borane-carbonyl might well be considered analogous to ammonium carbonate. 2H3N + C02- 4 C-O LNH2J 2H3N + H3BCO -*NH+ H3B-C ]2 The resulting solid H3BC0a2NH3 does not easily form a crystalline solid which will yield either a powder pattern or single-crystal photograph. The problem is similar to that encountered in the early studies of B2H6'2NH3. As in the former case, chemical evidence may well provide the first structural information. Samples of H3BCO.2NH3 have been prepared for study. The diammoniate is much more stable than the parent H3BCO molecule. WADC TR 59-207 99

III. THE REACTIONS AND STRUCTURES OF THE AMMONIA ADDITION COMPOUNDS DERIVED FROM TETRABORANE A. Background Although the reaction between diborane and ammonia has been studied rather intensively, little work other than the pioneer study of Stock and his colleagues(60) has been carried out on the tetraborane-ammonia reaction. The solution of the structural problem associated with B2H6.2NH3 increased interest in the ammonia addition compounds of tetraborane, since the possibility of a systematic extension of concepts seemed most challenging. Indeed, as will be shown in the following sections, the chemistry of tetraborane parallels that of diborane rather closely. B. The Preparation and Structure of the Diammoniate of Tetraborane 1. THE PROBLEM In their initial study of the reaction between tetraborane and ammonia Stock, Wiberg, Martini, and Nicklas(60) froze excess ammonia and tetraborane together using liquid air; then they held the system at -780~C for one hour. When excess ammonia was removed, the composition of the residue corresponded to the formula B4H10o4NH3o The solid decomposed when warmed to room temperature; gaseous boron compounds, hydrogen, and ammonia gas were liberated, and the "salt-like solid turned to a liquid." When excess tetraborane was used in the reaction instead of an excess of ammonia, an unstable solid residue of composition B4Hlo' 2.5NH3 resulted. In a careful study of the same system made in this laboratory, neither the formula B4H1o.4NH3 nor B4H10*2.5NH3 could be verified despite all efforts to reproduce the experimental conditions described in the original study. A modification of the experimental procedures resulted, however, in the isolation of a pure diammoniate of tetraborane B4Hlo02NH3, which is stable at room (60) A. Stock, E. Wiberg, and H. Martini, Ber. 63, 2927 (1930); A. Stock, E. Wiberg, H. Martini, and A. Nicklas, ibid., 65B, 1711 (1932). WADC TR 59-207 100

temperature. A preliminary description of this solid has been given.(61) In this report a more complete account of the investigation is recorded, and evidence is presented in support of the belief that the original "compounds" described by Stock and his co-workers were indeed complex and unstable mixtures. 2. THE PREPARATION AND PROPERTIES OF B4Hlo02NH3 When ammonia was added slowly to a cold (-780C) ether solution of tetraborane, the ammonia was absorbed completely. An excess of tetraborane (in relation to ammonia) was always used. If the system was aged at low temperature (4 days at -78, 12 hours at -45), and then the solution filtered and the solvent removed from the filtrate at -450C, a white, dry, microcrystalline solid identified by analysis as B4H1o.2NH3 was obtained. The molecular weight of the product as determined by vapor pressure depression in liquid ammonia was 81 (61c). In diethyl ether values obtained by vapor pressure depression ranged from 150 to 250. The theoretical value for B4Hlo.2NH3 is 87. The marked discrepancy between results in liquid ammonia and in ether would suggest that B4Ho0-2NH3 is an ionic type of solid. Ionic solids such as NaBH4, NH4Br, and [H2B(NH3)2]BH4 exist in liquid ammonia as ion pairs, thus giving a molecular weight in this solvent equal to the normal formula weight.[62) On the other hand, certain ionic or semi-ionic solids which dissolve in ether [e.g., FeC13, Na2HB(CH3)2](63) give ion clusters and high values for the apparent molecular weight. The compound B4H0o*2NH3 undergoes no visible decomposition in air at room temperature. It dissolves in cold water with only very slow hydrogen evolution. Acid or platinized platinum accelerates hydrogen evolution markedly in water. The compound is soluble in liquid ammonia and, judging from its X-ray powder pattern, it is unchanged upon recovery from the liquid ammonia solution. If perfectly dry, the compound will dissolve in anhydrous diethyl ether to give a stable solution at room temperature; the compound can be recovered unchanged by evaporation of the solvent. If, however, the solid has been handled in moist air prior to its solution in ether, a precipitate slowly forms. The diammoniate of tetraborane is insoluble in aromatic as well as in aliphatic hydrocarbons. The crystalline nature of the material is indicated by the X-ray powder data. The interplanar spacings and relative intensities of the reflections are given in Table XXI. A single-crystal X-ray study is being conducted in the laboratory of Dr. C. Nordman of this department. (61) a) G. Kodama and R. W. Parry, J. Am. Chem. Soc. 79, 1007 (1957); b) G. Kodama and R. W. Parry, paper presented before the Inorganic Section of the Sixteenth International Congress of Pure and Applied Chemistry, (Paris, 1957), p. 483; c) G. Kodama, Ph.D. dissertation, Univo of Mich., 1957. (62) R. W. Parry, G. Kodama, and D. R. Schultz, J. Am, Chem. Soc. 80, 24 (1958). (63) A. B. Burg and G. W. Campbell, Jr., J. Am. Chem. Soc. 74, 3744 (1952). WADC TR 59-207 101

TABLE XXI INTERPLANAR SPACINGS AND RELATIVE INTENSITY DATA FOR B4H10' 2NH3 Ihtensity* d(A) Intensity d(A) w 6.50 w 2.36 w 6.10 mw 2.29 m 4.63 w 2.18 s 4.10 w 2.03 m 3.69 vw 1.97 w 3.48 vw 1.89 m 3.29 vw 1.84 w 3006 vw 1.81 ms 2.88 vw 1.78 m 2.78 vw 1.74 w 2.62 vw 1.60 m 2.57 vw 1.54 m 2.52 vvw 1.50 w 2.46 *s = strong, m = medium w = weak, v = very. 3. CHEMICAL EVIDENCE FOR THE STRUCTURE OF B4o10'2NH3 Although no unequivocal physical evidence on the structure of B4Hlo-2NH3 is yet available, it is possible to assign a structure with a high degree of certaintyon the basis of chemical evidence. The arguments listed earlier supporting an ionic or semi-ionic model are reminiscent of those used in the case of the diammoniate of diborane which has the structure [H2B(NH3)21[BH4]. The structures of diborane and tetraborane can be represented in planar projection as: ",iiH H H and H\ B H XPH/ 1-B -4? 9\H It is significant that the end borons of tetraborane are bound to the rest of the molecule only by double bridge bonds. This feature resembles rather closely, but is not identical to, the double bridge bond structure of diborane. Nonsymmetrical cleavage of the double bridge bond of diborane by ammonia (see WADC TR 59-207 102

dotted line on formulas) results in the structure [H2B(NH3)2][BH4]. An application of the same argument to tetraborane suggests for the diammoniate of tetraborane the structure [H2B(NH3) +]1[B3H8]. Such a model immediately suggests a number of discriminating chemical reactions which should be of structural significance. These are outlined below. (a) The Reaction of B4H10.2NH3 with Sodium The reaction between sodium and the diammoniate of diborane was one of the basic experimental points upon which proposals for the structure of B2H6*2NH3 were based. The data may be interpreted(64) in terms of the structure [H2B(NH3)21 [BH4] by using the following equation: [H2B(NH3)2][BH4] + Na liq NH3 _ 1/2 H2 + NaBH4 + H2BNH2 -780C Since the reaction is fundamentally that of the complex cation It2B(NH3)2]+ and sodium, one would expect the following reaction if the proposed structure for B4Hlo'2NH3 were accepted: [H2B(NH3)2][B3H8] + Na —-.-liq NH3 > 1/2 H2 + NaB3H8 + H2BNH2 -780c The prior isolation and unequivocal characterization of NaB3H8 in an excellent independent study by Hough, Edwards, and McElroy(65) provided an easy and meaningful framework for the study of the above reaction. A solution of B4H10-2NH3 in liquid ammonia will react with dissolved metallic sodium at -780C to liberate one half mole of hydrogen per mole of B4Ho102NH3 (see Fig. 17 for related data). This reaction is complete in 20 minutes; further hydrogen evolution proceeds more slowly (Fig. 17). If the reaction is stopped after the initial rapid hydrogen evolution process has been completed NaB3H8 can be extracted in 60 to 70% yield from the solid residues by leaching them with diethyl ether. The NaB3H8 was identified by comparing its X-ray powder pattern with that of an authentic pattern for NaB3H8 which was kindly furnished by Dr. L. J. Edwards of Callery Chemical Company. (b) The Reactions of B4H.o-2NH3 with HC1 and HBr Schlesinger and Brown(66) found that under proper conditions metal borohydrides will react with HC1l as indicated by the following equation: (64) Do R. Schultz and R. W. Parry, J. Am. Chem. Soc. 80, 4 (1958). (65) Wo V@ Hough, L. J. Edwards, and A. D. McElroy, J. Am. Chem. Soc. 78, 689 (1956). (66) H. I. Schlesinger and H. C. Brown, J. Am. Chem. Soc. 62, 3429 (1940). WADC TR 59-207 103

~j (-2 1.1 \J1 Run 2 z.8 0 0 2>~~~ I ~~Run# I Table I <.7 0 I m.6 w H~~~J o 0.5 0~~~ 4='4.4 or) LL! J 0.3.2.1 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 TIME (HOURS) Fig. 17. The rate of H2 evolution in the reaction between sodium and B4H02NH3.

MBH4 + HC1 -.MC1 + H2 + 1/2 B2H6 In his early studies on the diammoniate of diborane Stock(67) noted that HC1 acts on B2H6.2NH3 to give significant quantities of B2H6. The observation was interpreted by Schultz(64) in this laboratory as evidence for a borohydride ion in B2H6'2NH3. [H2B(NH3)2][B1H4] + HX —- [H2B(NH3)2]x + +2 + 1/2 B2H6 An extrapolation of such an argument to the structure proposed for B4H1o'2NH3 gives the equation: [H2B(NH3)]2[B3H8] + HX-. —-> [H2B(NH3)2]X + H2 + B3H7 (X = C1- or Br-) The B3H7 group would immediately undergo further reaction unless stabilized by coordination with a suitable base used as solvent. The interaction of HC1 with a stoichiometric quantity of B4H10*2NH3 in diethyl ether solution at -780C gave rapid liberation of the expected mole of hydrogen and immediate precipitation of the ether insoluble solid [H2B(NH3)2]X where X = C1- or Br- (Fig. 18). The complex chloride and bromide salts were shown to be purer forms of the salts first prepared earlier from the diammoniate of diborane.(64) Indeed, it is F.. I.0.2 I 1 O 1 2 3 HOURS Fig. 18. Evolution of H2 from the reaction of B4H102NH3 with excess of HC1. (67) A. Stock, Hydrides of Boron and Silicon, Cornell Univ. Press, Ithaca, 1933. WADC TR 59-207 105

proper to note that the reaction between B4H,0-2NH3 and HC1 or HBr in diethyl ether at -78oC is the best method currently available for preparing pure samples of chloride and bromide salts of the dihydride-diammineboron (EIi) cation. Analytical values aria molecular weight data based on vapor-pressure depression of a liquid ammonia solution are contained in the experimental section, and a single-crystal X-ray study of the chloride salt by Nordman and Peters(68) removes all doubt about the structure of this solid, (c) The Preparation of Ammonia-Triborane from B4Hjoo2NH3 When diethyl ether was used as the solvent at -780C, the Bl3H7 fragment, resulting from the action of HC1 on B4Hjoo2NH3, was stabilized at -78~C by formation of the etherate, (C2H5)20B3H7o If such a solution is warmed, the etherate dissociates and the B3H7 group decomposes. If, however, ammonia is added to the cold (-78~C) ether solution, ammonia, being a stronger base than ether, will displace it to give the stable solid. NH3 + (C2H)20B3H7 -78C (2H5)20 + H3NB3H7 Et20 The ammonia adduct produced was identical in all respects to the product prepared by other methods and described elsewhere~(69) Yields were about 45%. In summary, strong evidence has been presented for the cation [H2B(NH3)21 and the anion B3H8 in B4H:,0oo2NH3o It is suggested that the anion has the general form: H H —- B H H I H H —]~ H resulting from the original B4H-o geometry. Such a structure is suggested by the known geometries of the parent B4H10 and the daughter compound H3NB3H7,(ThY]-) H3N —- -(i H H (68) C, Nordman and Co, Peters, J. Am~ Chem, Soc, in press; this report, P.34. (69) G.o Kodama and Ro Wo Parry, JO Am. Chemo SOCo, in press; this report, p. 120. (70) CO Nordman and C. Reimann, Jo Amo Chem, Soc., in press. (71) This report, p. 132o WADC TR 59-207 106

4. THE SYSTEM AMMONIA-TETRABORANE IN ETHER-EXCESS AMMONIA PRESENT Using excess ammonia in the direct reaction of ammonia and tetraborane instead of excess tetraborane, Stock(60) found the unstable compound B4H1o*4NH3. The following study in ether was undertaken in an effort to isolate or obtain evidence for a stable form of Stock's tetraammoniate of tetraborane. A given quantity of tetraborane was dissolved in diethyl ether. Ammonia was then introduced above the liquid at -78~C. The apparent equilibrium pressure (ammonia plus ether) above the solution at -780~C was measured as the ammonia content of the system was increased. Results are summarized in Fig. 19. Although a choice between the formula B4Hj0E2NH3 and the formula B4Hjo.2.5NH3 is arbitrary from the data of Fig. 19, the former composition is preferred in view of previously described observations. It is particularly significant that no evidence suggesting the expected composition B4Hio-4NH3 is contained in Fig. 19. On the other hand, certain, as yet incomplete, observations suggest the possibility of higher ammoniates. The ether solution remained clear as long as the ratio NH3/B4H1o did not exceed 2, but a noticeable precipitate formed when the ratio exceeded 2.5. Data on the preparation of B4Ho102NH3 indicate that compound formation proceeds very slowly at -78~C. A long aging period was required. Since the entire experiment of Fig. 19 was completed in 5 hours, it is doubtful that the formation of the final compound, B4H1o-2NH3, was ever completed. Intermediate ammoniates are a more reasonable possibility.(72) The formation of the trace of precipitate suggests the possibility of an ether-insoluble higher ammoniate, but the amount of solid available up to the present time has not permitted characterization. 5o THE SYSTEM AMMONIA-TETRABORANE WITHOUT SOLVENT The earlier study(60) on the system B4HjoENH3 involved mixing quantities of B4Hio and NH3 together and then pumping off the excess ammonia or tetraborane after equilibration for 1 hour at -78~C. Similar experiments conducted in this laboratory indicated that many hours of aging were required for apparent equilibration. This fact alone indicates that the earlier products were in all probability mixtures rather than single solid phases. In a typical experiment, patterned after that of Stock, an excess of ammonia was frozen with a measured quantity of tetraborane; then the system was equilibrated at -78~C for 5 hours. After 5 hours the reaction vessel was opened to a trap cooled with liquid nitrogen and excess ammonia was removed. The overall composition of the B4Hio-NH3 adduct was followed as a function of time. Data are shown in Fig. 20. Comparable data of Stock for a sample without the preliminary 5-hour aging period are shown for comparison. Under the conditions (72) This study was carried out before the preparation of B4Hzo'2NH3, and the observations made herein suggested the procedure which led to the isolation of B4Ho102NH3o WADC TR 59-207 107

U ~~~Q ~24 F-3 \Jl~~ ~22 20 I E 18 E 0 16 w4. I (/) o 14 LL z: I.-.12 - 0 0 -c0 or,~10 0 0- 10 o W C D~ U 8 a:,~~~~~~~~~~~~~ w a.6 4 2 s 4 5 6 7? NH3 MOLECULES PER B4HIO Fig. 19. Ammonia pressure above an ether solution of tetraborane.

\-3 Ld \10 1.04 Millimoles 84H10 O r0 12.58Millimoles NH3 Originally 0 5-Hr Reaction Time a I CID 8~,Temperature Raised From -780 To -63.5OC LU -i 780C 0 2 w z >-Data Reported By o \..' Stock, Wiberg, and Martini IW 1' B4 af.,6, 2927 (1930)] < ~~~~~~~~~~~~~~~~~~1.12 Millimoles B4 H0o- Second Run: INH LIJ ~~~~~~~~~~~~~~~~8.34Millimoles NH,3DitleOu-35C 0 Ratio= 3.25/I: 0 2 ~~~~~~~~~~~~~~~~~Reaction Time=12Hr+ H2 Evolved Ratio H2 /B4 HF0 3/ 0 I 2 3 4 5 6 7 8 9 10 I1 12 13 14 15 TIME (HOURS) Fig. 20. Composition of system B4Hio-excess NH3 as a function of time.

described above, the composition of the residue at -780C corresponds to the formula B4H1oo77TMH3 When no change in composition was detected at -78~C, the temperature was raised to -63.50C. At this higher temperature, ammonia was lost without significant hydrogen evolution. The composition of the residue appeared to be approaching B4H10o3NH3, although the stoichiometry was rather poor (Fig. 20). At temperatures above -63.50C additional ammonia was evolved and the ratio N/B4 slowly approached 2. It is significant that in systems of this type H2 evolution became quite vigorous above 00C, and the ratio H2/B4H1o became as large as 0.82 on some samples after several days. A pressure composition diagram was then constructed for the system B4H1oNH3. A sample of ammonia was placed in a tube and a measured excess of ammonia was added. The apparent equilibrium pressure above the solid adduct was measured at -780C, -635C, -450~C and -35~C. Part of the ammonia was removed and the pressure measurements were repeated. Many days were required to reach- a steady pressure at each temperature-composition point. Data are presented in Fig. 21o At temperatures above -78~C the highest definite ammoniate appears to be B4Hloo6NH3o At -78~0 the data do not exclude the composition B4H1oo7NH3 shown in Fig. 20. It is also of interest to note that the data of Fig. 20 are not in conflict with the existence of B4Hlo06NH3 at -65350C, but simply imply that such a hexammoniate has a measurable dissociation pressure at that temperature; thus ammonia can be removed to a trap cooled with liquid nitrogen. Stock's observations on the instability of the above product at temperatures above -23~0C were confirmed. The adduct of B4H1o-NH3 prepared without solvent evolved ammonia and hydrogen gas and became liquid when warmed to room temperatureo In fact, even "stable B4Hlo-2NH3" lost hydrogen and ammonia and turned to a liquid when added to the viscous liquid residues from the above system. Thus even if B4Ho.2NHI3 were formed by the direct reaction of B4H1o and ammonia, the unstable and catalytically active byproducts formed in the same reaction would complicate the isolation of pure B4Hloo2NH3 at room temperature. Some evidence for [H2B(NH3)2] [B3H] was indeed found in the product resulting from the reaction described by Fig. 20 and in related processes. If B4H1o and ammonia were frozen together with liquid nitrogen and then warmed to -780C, aged for 3 days at this temperature, dissolved in liquid ammonia, and treated with metallic sodium, hydrogen gas was evolved (Fig. 17) and NaB3H8 could be extracted from the solid product in yields up to 60%. Such an observation suggests that the solid product resulting from the prolonged interaction of tetraborane and excess ammor;iaat -780C is largely an ammonia addition compound or ammonia solvate of [H2B(NH3)2L[B3H8s; the data would argue against the hypothesis that any major portion of the BH8 ion has been destroyed through attack of the ammonia on its double bridge bonds. The problem of higher ammoniates is still being studied. The utilization of a solvent such as ether to moderate the direct reaction between a boron hydride and an electron donor molecule was initiated by Steindler WADC TR 59-207 110

Ln,Xj -350C 400 -d E E -l/ E _-450C z 300 0 Cf) w / 0- 200 wp a do wow | 18 - 780 C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 NH3/ B4HiO - Fig. 21. Pressure composition diagram for system tetraborane ammonia.

and Schlesinger(73) in a study of the reaction between diborane and hydrazine. They obtained excellent stoichiometry in diethyl ether, while Emeleus and Ston174) failed to isolate well-defined compounds in an earlier investigation involving direct combination of reagents. Similarly, Campbell, Bissot, and Parry(75) were able to moderate the reaction of diborane and hydroxylamine by using ether as a solvent; yet without the solvent the reaction either would not go or resulted in a violent explosion. All available data thus suggest that Stock's earlier direct reaction products (i.e., B4Hlo.2o5NH3 and B4Hlo.4NH3) were in all probability complex and unstable mixtures, which contained some undefined precursors of the structure [H2B(NH3)+]~J3Hs]o 6. DISCUSSION The foregoing observations suggest that the use of a solvent is important in the isolation of a stable tetraborane-ammonia adduct. It is equally clear that the choice of the solvent ether is of great importance in determining the nature of the product formed. If the ether is a relatively strong electron donor, it will cleave the double bridge bond of the tetraborane molecule symmetrically to give one half mole of B32H6 and the etherate of B3H70 Addition of ammonia to this system then results in the displacement of the ether by the stronger base to give the compound H3NB3H7.(69) On the other hand, if the ether is a weak electron donor and it does not attack the diborane molecule under the conditions used, it will simply serve as a diluent and heat-transfer medium to minimize localized overheating. Under such conditions the ammonia attacks the tetraborane molecule directly in the ether solution and effects nonsymmetrical cleavage of the double bridge bond without further decomposition. It is clear that B4H1o is not attacked by the diethyl ether at low temperature since it was found in this and other laboratories(76) that B2H6, the normal decomposition product, is not evolved and B4Hlo can be separated unchanged by fractionation. The resemblance between the chemistry of diborane and the tetraborane is indeed striking. Nearly every diborane reaction finds a predictable and verifiable counterpart in tetraborane chemistry. Such analogies are apparent in this and another section of this report (69) (73) Steindler and Ho. I. Schlesinger, J. Amo Chemo Soc. 75, 756 (1953). (74) Ho Emeleus and F.G.Ao Stone, JO Chemo Soco 840 (1951), (75) D. Ho Campbell, To C. Bissot, and R. W. Parry, JO Am. Chem. Soc. 80, 1949 (1958). (76).Ho Io Schlesinger, Univo of Chicago, Navy Contract No. N173 S-9820, Final Report (1946-47)o WADC TR 5 9-207 112

7. EXPERIMENTAL (a) Reagents Used (1) B4Hio.-The majority of the tetraborane used in this investigation was donated by Professor Thomas Wartik of Pennsylvania State University and by Drs. George Huff and L. J. Edwards of Callery Chemical Company. Some of the product was prepared in this laboratory. The high-pressure storage of B2H6 in a stainless-steel cylinder at room temperature was the best preparative method. This method was described to R. W. Parry by Dr. Wartik and Dr. Edwards and will be described in more detail by them. The product was fractionated at -85~C or -95"C and stored at liquid nitrogen temperature. It was again fractionated at -1000C before its use. The vapor pressure of the purified compound was 386:+ 1 mm at O0C. (2) Other Reagents.-Commercial NH3, (C2H5)20, LiAlH4, BF3, NaH, Na, etc., were used in top quality grade. Customary methods of purification and handling were used and have been described in earlier reports.(77) (b) The Preparation of B4Hlo'2NH3 The slow addition of ammonia to a cold (-78~) ether solution of B4H1o gave complete ammonia absorption. An excess of B4H1o was always used. After ammonia addition the system was aged for four days at -780 and for twelve hours at -45~. The solution was filtered on a vacuum line filter at -45~ to remove a slight turbidity; then the ether was removed at -45~ from the clear filtrate. The white solid which was left after removal of the ether was washed with cold dry benzene and transferred to a filter disc. It was then washed through the filter disc with dry ether. Removal of the solvent left a white, dry microcrystalline solid which was characterized as follows. Anal. N, 32.3; B, 49.4; H2 124.3 mmoles/g. Theory for B4H1o'2NH3: N, 32.1; B, 49.5; H2, 125.8 mmoles/g based on the hydrolysis equation B4H1o'2NH3 + 8H20 + 2H+ —--- 4HB02 + 2NH+ + 11H2 The molecular weight as determined by vapor pressure depression in liquid ammonia was 81. Theoretical for B4H10-2NH3 is 87. Yields ran as high as 86% based on the NH3 used or 78% based on the B4H1o (slight excess used). In Table XXII a number of reaction conditions and their results are listed. (77) R. W. Parry, et al, The Chemistry of Boron Hydrides and Related Hydrides WADC Tech. Report 56-318 (1956). WADC TR 59-207 113

TABLE XXII THE REACTION OF B4H1o AND NH3 IN DIETHYL ETHER Yield of B4H1o- 2NH3 B4H1o NH3 Ether NH3 Temp. Based Based (mmoles) (mmoles) (ml) B4H1o (C) Time on on Remarks B4HIO (oC) on on NH3 B4H1o (pct.) (pet.) 4.18 7.56 3 1.81 -78 to 30 min 0 0 (a) -70 4.20 1.91 2 0.455 o 1 hr 54 12.3 (b) 2.18 3.35 3 1.54 -95 1 hr low low 3.0o8 5.55 5 1.8 -78 4 hr 40.5 37.4 (e) 2.05 3.84 10 1.88 -78 3 days 54 51 (c) -78 4 days 2.37 4.25 10 1.8 86.5 77.7 (c) -45 12 hr a Product became liquid at room temperature. b Purity of product in doubt because methods of purification were not well developed at the time of the run. c Solvent ether removed at room temperature. (c) The Reaction of B4Hlo-2NH3 with Sodium in Liquid Ammonia A sample of B4H1,02NH3 weighing 74.8 mg (0.856 mmole) was placed in a dried nitrogen-filled tube and an excess of sodium in a bulb was crushed into the tube. The tube was attached to the vacuum line and 2 ml of liquid ammonia was condensed onto the sample. The temperature was rapidly raised to -78~Co In 5 minutes 0.358 mmole of H2 was liberated. On standing for an additional 15 min at -75~C, 0.101 mmole of H2 gas was evolved. (Total H2 = 0.459 mmole = 1.07 equiv H2/B4Hlo' 2NH3.) The excess sodium was eliminated by shaking with mercury and filtering at -350~C H2 gas evolved during this operation was 0.027 mmole. Most of the solvent ammonia could be removed at -35~C by distillation. Warming the residue to room temperature resulted in rapid evolution of H2 gas. (Total H2 then equaled 1.35 mmoles.) The solid residue was leached with dry diethyl ether at room temperature in the vacuum line extraction system. From the filtrate NaB3H8 was obtained by evaporation of the solvent at room temperature. The yield of NaB3H8 was 69% based on the equation: [H2B(NH3)2][B3H8] + Na —.. H2NBH2 + NH3 + NaB3H8 + O05 H2 WADC TR 59-207 114

(d) The Reactions of B4Hlo-2NH3 with HC1 and HBr. The Formation of [H2B(NH3)2]C1 and [H2B(NH3)2]Br. In a typical run 1.84 mmoles of B4Ho102NH3 was placed in a reaction tube and dissolved in about 2 ml of diethyl ether. The solution was frozen with liquid nitrogen and 1.78 mmoles of HC1 was condensed into the tube. When the temperature was raised to -78"C, hydrogen gas (1.74 mmoles) was evolved in 15 minutes and a white precipitate formed. The precipitate was separated in the vacuum line filtration assembly; the solid was washed through the filter disc with liquid ammonia, and the solvent ammonia was distilled from the filtrate at -45~C. A sample of H2B(NH3)2]C1 weighing 119 mg remained (yield = 79%). Analytical data showed B = 13.2%; N = 33.9%; hydridic H = 2.41%; C1- = 42.7%. Theoretical values for [H2B(NH3)2]Cl: B = 13.1%; N = 34.0%; hydridic H = 2.43%; C1- = 43.05. The molecular weight was determined by vapor pressure depression in liquid ammonia using standard methods(62) (see Fig. 6). The value obtained by extrapolation to infinite dilution was 95. The result is in only fair agreement with the theoretical value of 82.3, but the deviation is in the range of uncertainty of the method. The X-ray powder pattern was the same as that for [H2B(NH3)2]C1 made in an earlier study(64) and is identical to the powder pattern calculated from the single-crystal data of Nordman and Peters.(68) The compound [H2B(NH3)2]Br could be obtained by a similar procedure if HBr were substituted for HC1. Analysis of the bormide salt gave: B = 8.46%; N = 22.15%; Br- = 62.5%; hydridic H = 1.56%. Values calculated for [H2B(NH3)2]Br are: B = 8.53%; N = 22.10%; Br- = 63.02%; hydridic H = 1.58%. The molecular weight in liquid ammonia was about 120 as compared to a theoretical value of 127 (Fig. 6). Detailed powder pattern is given elsewhere.(64) (e) The Preparation of H3NB3H7 from B4Hloo2NH3 The filtrate obtained from the reaction between B4Hlo02NH3 and HC1 (see previous section) was agitated constantly at -78~C and 1.44 mmoles of ammonia gas was slowly introduced into the system. It was absorbed completely. The solvent ether was removed at room temperature and from the viscous liquid left in the tube H31NB3H7 was sublimed in the vacuum sublimation system at 40 to 50~C. The yield was 36.5 mg or 45% on the basis of ammonia added. (f) The System Ammonia-Tetraborane in Ether-Excess Ammonia Present A sample of tetraborane (1.02 mmoles) was dissolved in 4 ml of diethyl ether and maintained at -780C under constant agitation by means of a pulsating magnetic stirrer. Measured amounts of ammonia were introduced into the system. A constant pressure was reached in less than 5 minutes after each addition of NH3. The pressure above the system was measured using an oil manometer. The blank run was made without B4Hlo in the ether. Data are given in Fig. 19. WADC TR 59-207 115

(g) The System Ammonia-Tetraborane Without Solvent The stoichiometry obtained for the direct reaction of tetraborane and ammonia without solvent was strongly dependent upon experimental conditions. The two cases differentiated by Stock were reaction systems containing an excess of ammonia and those containing an excess of tetraborane. (1) Excess Ammonia Present. -A measured amount of B4Ho10 was condensed by liquid nitrogen at the bottom of a reaction tube and a measured amount of ammonia condensed in a ring above it. The system was allowed to warm slowly to the reaction temperature. Data for a typical run are plotted in Fig. 20 and several runs under different conditions are summarized in Table XXIII. A rapid examination of runs 1 and 3 indicates that the formula approached by the resulting residue is strongly dependent upon the aging period provided at -78~C and upon the temperature of the sample during removal of excess ammonia. It will be noticed that Stock, Wiberg, and Martini used well-specified time and temperature for their initial reaction and well-specified time and temperature for the removal of ammonia to obtain the results reported. The marked time dependence strongly suggests a reaction mixture which had not reached its equilibrium condition. (2) Excess Tetraborane Present. -The reactants were frozen together as indicated previously. Excess B4Hjo was distilled from the mass. Typical data are shown in Table XXIV. In run 2 a small amount of solid, stable at room temperature, appeared on the wall of the tube above the decomposing residue. The X-ray powder pattern of the solid was that of B4Hlo02NH3. (h) The System B4H1o-NH3-B4Hlo 2NH3 About 30 mg (0.3 mmole) of freshly prepared B4Hlo'2NH3 was placed in a reaction tube, and 1.25 mmoles of B4H1o and 7.30 mmoles of NH3 were condensed above the solid. The system was allowed to warm slowly to -78~C and maintained for 12 hours at this temperature. After the removable ammonia was distilled out at -780C, the system was allowed to warm to room temperature, The B4H1o-NH3 adduct formed by the direct interaction became liquid, evolving gas. The crystalline B4Ho102NH3 gave off gas as it came into contact with the decomposing liquid. In 4 hours 1.02 mmoles of H2 was collected and a viscous liquid phase removed in the reaction tube. At this stage the H2 gas evolution from the system was very slow. At this point more B4Hjo-2NH3 (0.16 mmole) was added to the system. Slow evolution of gas was observed at the solid-liquid interface. In about an hour close to 0.2 mmole of H2 gas was collected. A small amount of solid phase remaining in the viscous mass was identified as B4Hl0 2NH3 by its X-ray powder pattern. WADC TR 59-207 116

TABLE XXIII SUMMARY OF RUNS ON B4H IoNH3 Millimoles Reaction NH3 Removal NH3/B4H1o Run Rat io H2/B4H1o, R NH3 at Temp. Time Temp. Time No. B4H NH3 Approached mmoles Start (0C) (hr) (~C) (hr) by System Data from This Study 1 1.04 12.58 -196 5 -78 8.5 7.05 nil to -78 -63.5 6.00 3 to 3,2 nil -165 2 1.12 8.34 -65 12+ -63.5 -- 3.2 0.053 to -65 3 0.325 4.66 - 78 1 -78 24 3.2 0.15 -63 6 2.91 0.15 -45 3 2.83 0.15 -23 8 2.34 0.17 0 25 2.16 0.70 25 6 2.16 0.76 4 0.337 5.17 - 78 24 -23 3.3 3.05 0.016 25 18 2.14 0.72 70-80 5 2.14 0.81 25 24 2.14 0.82 Comparable Data from Stock, Wiberg, and Martini iS 0.445 4.7 - 75 1 -75 Ab.3a 4.0 N.R.b 2S 0.455 4.8 - 75 1 -75 Ab.3 4.0 N.R. 3s 0.442 4.6 - 75 1 -75 Ab.3 4.0 N.R. aAbout 3. bNot reported. WADC TR 59-207 117

TABLE XXIV THE ADDITION OF EXCESS B4Hlo TO AMMONIA Time at mmoles mmoles Temp Ratio Run BHTep Listd NH/BHo Remarks (min) 1 1.62 5.53 -196 to -126 20 Unstable at room -95~ C 30 2.25 temperature. -780~C 90 2 2.46 8.37 -196 to -95 20 Mostly unstable at -780C 30 1.8 room temperature but some B4HloE 2NH3 iso-350C 150 Llated. (i) The Reaction Between Sodium and the Solid Product Resulting from the Direct Interaction of B4Hjo and Excess NH3 To the product from run number 1 of Table XXIII 3 ml of liquid ammonia was added; then a bulb containing an excess of sodium metal (0ol g) was crushed and added to the frozen system. The temperature was permitted to rise to -780C and then held at this value. H2 evolution as a function of time is shown in Fig. 17. Similar data for run number 2 are shown on the same figure. In both cases one half mole of H2 per mole of B4H1o was liberated rapidly, but the rate of the secondary H2 evolution process varied widely depending apparently on sample purity. In a second run intended to isolate solid sodium salts before they decomposed in the system, the following procedure was followed. About 3 ml of ammonia was condensed above 1.86 mmoles of B4H1o in a reaction tube. The system was allowed to warm slowly to dry-ice temperature and then was held at -780C overnight. An excess of sodium was added to the system as described before; then the temperature was raised rapidly to -78~C. After one-half hour, ammcnia and evolved hydrogen were removed, and 1.17 equivalents of H2 per mole of B4Hlo were found. The last traces of ammonia were pumped off at room temperature, The solid residue was leached with dry diethyl ether in the vacuum line extraction system. A crystalline compound was isolated by evaporating the ether from the filtrate. The solid was identified as NaB3H8 by its powder pattern. The foregoing experiment was confirmed by a second experiment involving only one significant change in experimental detail. Ammonia was removed from WADC TR 69-207 118

the system at -780C but the solid residue was aged for 3 days at -780C before addition of solvent ammonia and of sodium. After 15 minutes the ratio of hydrogen evolved to B4Hlo was 0.51/1.73. A total of 0.725 mmole of H2 gas was collected in all. The ratio H/B4Hlo was 1.06. NaB3H8 was again leached from the solid residue and identified by its X-ray powder pattern and by analysis for hydridic hydrogen, and boron. The ratio H/B was 2.96; the theoretical value is 3o00. Although the yield of NaB3H8 was not determined in the above run, yields were good; a third run with a 2-day aging period at -780C and -50'C gave a 60% yield of recovered NaB3Hs. (j) The Reactions of Mg(AsF6)2 with B4Hlo-NH3 Adducts - A Precipitation Test for the Borohydride Ion In an earlier study from this laboratory,(6) it was noted that the addition of Mg(SCN)2 to a liquid ammonia solution of a borohydride resulted in the precipitation of [Mg(NH3)61[BH412. The process served as a qualitative means of identifying the BH4 ion in the liquid ammonia solution. Subsequent study in this laboratory has indicated that Mg(AsF6)2 is a better reagent than the thiocyanate since it is not as subject to attack by the borohydride ion. The reagent was prepared as follows. KAsF6 was prepared from KH2As04 using the method of Dess and Parry.(78) The K(AsF6) was converted to Mg(AsF6)2 by shaking a water solution with successive portions of a magnesium ion exchange resin. The dried compound dissolved readily in liquid ammonia and was recrystallized once from this solvent before use. The recrystallized solid was stable in open air and had the formula [Mg(NH3)6][AsF6 ]2. [Mg(NH3)6](AsF6)2 serves as an excellent precipitant for BH- ion in liquid ammonia and as such serves as a discriminating test for this presence of the borohydride ions in the liquid ammonia aolution. Mg(NH3)6 does not precipitate the B3H8 ion from liquid ammonia, as was shown by the following experimentso About 20 mg of freshly prepared B4Hlo02NH3 were mixed with [Mg(NH3)e] (AsF6)2 in a reaction tube, and about 3ml of liquid ammonia were condensed into the system. A temperature of -78~C was maintained for over 12 hours. No precipitate was formed in the solution. Similarly, a mixture of NaB3H8 and Mg(AsF6)2 gave no precipitate when a liquid ammonia solution was formed by condensing NH3 on to the solids. In contrast, B2H6'2NH3 and NaBH4 gave precipitates of [Mg(NH3)6](BH4)2 when [Mg(NH3)6] (AsF6)2 was added to their liquid ammonia solutions. Although properly prepared B4H1o02NH3 gave no evidence for the presence of the BH4 ion, it was found that the product formed by the direct interaction of B4Hlo and excess ammonia without solvent gave some evidence for the BH4 ion. It is assumed that this contaminant resulted from the more vigorous reaction conditions in the absence of solvent. A large excess of NH3 and 0,974 mmole of B4H1o were condensed together in a reaction tube. The tube was then surrounded by dry ice and allowed to react at -78~0C for 12 hours. The system was frozen (78) Ho M. Dess and R. W. Parry, J. Am. Chem. Soc. 79, 1589 (1957). WADC TR 59-207 119

and [Mg(NH3)6][AsF6]2 was introduced onto the frozen product under a dry nitrogen atmosphere. About 3 ml of solution was prepared by condensing more ammonia onto the system. The white precipitate formed was filtered in the vacuum line filtration system. This solid appeared to be reactive in moist air. In its Xray powder pattern every line of [Mg(NH3)6][BH4]2 was found. No attempt was made to identify or separate the other components. C. The Preparation and Properties of Ammonia-Triborane, H3NB3H7 1. EARLIER STUDIES The initial study of the reaction between tetraborane and trimethylamine by Burg and Stone(79) indicated that the tetraboraie molecule can give 3BH3 groups which appear as 3 molecules of H3BN(CH3)3. A'polymeric residue remained in the reaction vessel. The careful and definitive study by Edwards, Hough, and Ford(80) on the system (CH3)2N-B4Hlo indicated that under properly controlled reaction conditions N(CH3)3 and B4H1o will combine to give the new compound (CH3)3NB3H7. The data suggest immediately the possible existence of the compound H3NB3H7, which would bear thesame relationship to (CH3)3NB3H7 that H3NBH3 bears to (CH3)3NBH3. Indeed, one may note a striking parallel between the reactions of tetraborane and diborane with N(CH3)3 and the reactions of tetraborane and diborane with NH3. It is known that trimethylamine brings about symmetrical cleavage of the double bridge bonds in diborane to give (CH3)3NBH3: H/ H B B + 2N(CH3)3 >->2(CH3)3NBH3 H/ A/ whereas ammonia brings about a nonsymmetrical cleavage of the double bridge bonds to give the diammoniate of diborane: (81) HH H N1+ H H/B,/ + 2NH3 -b F B E,I8f~ H HH NH3 H Similarly, the direct reaction of tetraborane and trimethylamine to produce (CH3)NB3H7 can be interpreted as a symmetrical cleavage of the double bridge (79) A. B. Burg and F.G.A. Stone, J. Am. Chem. Soc. 75, 228 (1953). (80) L. J. Edwards, W. V. Hough, and M. Do Ford, to be published. (81) D. R. Schultz, So G. Shore, R. W. Parry, G. Kodama, P. R. Girardot, R. C. Taylor, and A. R. Emery, J. Am. Chem. Soc. 80, 1-30 (1958). WADC TR 59-207 120

bonds in tetraborane (82)' H HB + 2N(CH3)3 -*(CH3)3NBH3 + (CH3)3NB3H7 H and the direct reaction of ammonia with tetraborane can be interpreted as a nonsymmetrical cleavage of the double bridge bonds to give the diammoniate of tetraborane(61) [this report, p. ]. H,4-H&H +- -- LF — AH /B B\ + 2NH3 > \ H the dic o of aH Because the direct reaction of ammonia with diborane gives nonsymmetrical was delayed for a number of years and was achieved only by indirect methods. Similarly, the triborane analog of H3NBH3, namely, H3NB3H7, cannot be synthesized by direct reaction of NH3 and B4Hjo but can be prepared by indirect routes. Fortunately, the analogy with diborane chemistry is close and comparable reactions are applicable; however, such an analogy was not established at the beginning of this research 2. METHODS FOR SYNTHESIZING H3NB3H7 Shore and Parry(83) showed that the interaction of LiBH4 with an ammonium halide salt proceeded as indicated below: LiBH4 + NH3C Et H2 + H3NBH3 + LiCl~ 250C An extension of Shore's reaction to the new compound, NaB3H8, described by Hough, Edwards, and McElroy(84) gave a process which can be represented by the equation: (82) L. J. Edwards, W. Vo Hough, and M. D. Ford, XVIth International Congress of Pure and Applied Chemistry, Section of Inorg. Chem., p. 475, Paris, July, 1957; L. J. Edwards, et al., to be published. (83) S. G. Shore and R. W. Parry, J. Am. Chem. Soc. 80, 8 (1958). (84) W. V. Hough, L.o J. Edwards, and D. McElroy, Jo Am. Chem. Soc. 78, 689 (1956). WADC TR 59-207 121

NaB3H8 + NH4C1 Et —2- H2 + H3NB3H7 + NaCl4 250C The new compound ammonia-triborane, H3NB3H7, was sublimed in vacuum at 40 to 500C from the solid residues which remained after removal of the solvent ether. The reaction was not as clean-cut as that involving LiBH4. Rather complex side reactions appeared to be taking place. Yields of recovered H3NB3H7 ranged from a trace to 38% on the basis of the NaB3H8 used. Most yields were in the range of from 20 to 30%. A second extrapolation of diborane chemistry provided a direct synthesis of H3NB3H7 from the diammoniate of tetraborane, B4H1o.2NH3; this process provided important information in establishing the structure of B4H10o2NH3(7) and has been described in Section IIIB of this report. The most effective method for the synthesis of 10-gram quantities of H3NB3H7 is somewhat analogous to the procedure recommended by Shore and Parry(83) for the synthesis of H3NBH3 from the dimethyl etherate of borane: B2H6 + 2(CH3)20 > 2(CH3) 20BH3 (CH3)20BH3 + NH3 ---- H3NBH3 Edwards, Hough, and Ford(82) reported that a number of ethers will react directly with tetraborane at room temperature to give diborane and the etherates of the B3H7 group. Yields were reported as quantitative when tetrahydrofuran was used. The reaction of such etherates with ammonia might then give H3NB3H7. The overall process can be represented as: B4H1o + R20 25~C > 1/2 B2H6 + R20B3H7 R20 R20B3H7 + NH3 Diethyl ether H3NB3H7 + R20 -780C By proper selection of the ether, good yields of H3NIB3H7 have been obtained. Certain characteristics of the ether appear to be of major importance. First it should be a strong enough base to react smoothly with B4H1o to give a relatively stable B3H7 etherate at room temperature. Diethyl and dimethyl ethers were not suitable because they are too weak as a base and the B3H7 group decomposes so rapidly at room temperature that the subsequent low-temperature ammonia reaction cannot be effected. On the other hand, the triborane-etherate, R20B3H7, should not be extremely stable or difficulty might be encountered in WADC TR 59-207 122

displacing the ether by ammonia at -78~C. It was found that the tetrahydrofuran adduct of B31H7, C4HsB03H7, reacted with ammonia in a solution of diethyl ether at -780C, but the yields of recovered H3NB3H7 were only about 70% and a fair amount of unidentified liquid always contaminated the sublimed product. It was assumed, although not established with certainty, that the high stability of the tetrahydrofuran adduct contributed to the low yield. According to McLaughlin, Tamres, and Searles,(85) tetrahydropyran is intermediate in base strength between diethyl ether and tetrahydrofuran toward BF3 as a reference acid; it should thus be suitable for H3NB3H7 synthesis. When tetrahydropyran was used as the original ether in small-scale runs, yields of H3NB3H7 were as high as 94% on the basis of the original B4Hlo used. Yields of 70 to 80% were obtained in larger-scale (20-gram lots) operation. The product obtained directly from the ether displacement procedure appeared to contain small amounts of impurities which rendered the compound somewhat less stable than the H3NB3H7 obtained directly from the reaction between INH4X and NaB3H8. When the ether displacement procedure was utilized in the preparation of 20-gram lots of H3NfB3H7, the product was crystallized from benzene and then from toluene to obtain a solid of high purity. 3. PROPERTIES OF H3NB3H7 Pure ammonia triborane is a white crystalline solid which can be sublimed very slowly under high vacuum. It melts at 73-75~C with slow evolution of hydrogen. It can be handled in air without obvious decomposition. It is surprisingly resistant to complete hydrolysis; periods of several days at 1200C in 6N HC1 were used to hydrolyze the product for analysis. It is extremely soluble in diethyl ether; in fact, it picks up ether vapor avidly to give an ether solution. The original compound can be recovered unchanged on vaporization of the ether. It can be dissolved in and recovered unchanged from liquid ammonia at -750C and from benzene. It is soluble in acetone and alcohol, and very slightly soluble in petroleum ether. X-ray powder data for identification of the solid are presented in Table XXV. Westrum and Levitin(86) have shown that it undergoes an apparent firstorder transition involving an enthalpy increment of 1233 cal/mole and an entropy increment of 4.15 cal/mole x deg at 297.100K. The crystal structure of the lowtemperature form has been worked out by Nordman and Reimann(87) (see page 132 of this report). Details of compound characterization are in the experimental section. (85) M. Tamres and D. Eo McLaughlin, private communication (1957); D. E. McLaughlin, M. Tamres, and S. Searles, presented at the 133rd meeting of the American Chemical Society, April, 1958. (86) E. F. Westrum, Jr., and N. E. Levitin, J. Am. Chem. Soc., in press. (87) C. E. Nordman and C. Reimann,J. Am. Chem. Soc., in press; page 132 of this report. WADC TR 59-207 123

TABLE XXV INTERPLANAR SPACINGS AND RELATIVE INTENSITY DATA FOR H3NB3H7, HIGH-TEMPERATURE FORM (TETRAGONAL) Intensitya d(A) hkjb s 4.48 101 s 4.33 110 m 3.29 002 m 3.06 200 m 2.62 112 m 2.52 211 w 2.24 202 w 2.07 103 w 1.95 301 a s = strong, m = medium, w = weak. b These indices were assigned and verified on the basis of intensity data from the singlecrystal work by Nordman and Reimann. At -780C, H3NB3H7 was not attacked by trimethylamine, but at room temperature it reacted with an equimolar quantity of the base to give trimethylamineborane [(CH3)3NBH3], and an unidentified solid which decomposed slowly through evolution of hydrogen gas. It was not possible to recover any trimethylaminetriborane, (CH3)3NB3H7, from the products. Ammonia-triborane in liquid ammonia solution at -780C reacted with dissolved sodium metal (present in excess) to give one equivalent of hydrogen per mole of H3NB3H7 in 20 minutes. Slow hydrogen evolution continued until a total of 1 mole of H2 per mole of H3NB3H7 was obtained, The residue left by removal of the solvent contained NaBH4 on the basis of the X-ray powder pattern. No ether soluble component was detectable in the residue, and no other components were recognized in the mixture. Additional work on the chemistry of H3NB3H7 is in progress. As noted on page 73 of this report, H3NB3H7 has a dipole moment of 6.5 Debyes, and its vapor pressure (see pages 63 and 64) is given by the equation logjo (mm) = -3739 (1.0075) x 9.200 (1 ~.0096) T over the range of 30 to 55~C. AHvap = 17.1 + 0.1 kcal/mol WADC TR 59-207 124

4. EXPERIMENTAL Conventional vacuum lines were used for handling volatile and/or air-sensitive reactants and products. Moisture-sensitive solids and nonvolatile liquids were handled in a dry box. (a) Materials Used (1) B2H6. —Diborane was prepared from BF3 and LiAlH4 by conventional techniques.(88) (2) B4Hbo. -Most of the tetraborane used in this investigation was donated by Professor Thomas Wartik of the Pennsylvania State University and by the Research Department of Callery Chemical Coo The authors wish to express their sincere appreciation. Some additional supplies of B4Hlo were prepared by converting B2H6 to B5Hll by the method of Burg and Stone.(79) B5Hll was then converted to B4Hlo by heating it in the presence of H2(89). Tetraborane was also prepared by high pressure storage of B2H6 at room temperature.(90) B4H10 used in small-scale runs was fractionated at -85~C or -950C and was stored at -1960C until used. It was fractionated at -1000C just before use. The vapor pressure of the purified compound was 386 ~ 1 mm at 0'C. Tetraborane used in the larger-scale preparations was distilled from a -950C bath after initial removal of B2H6 at -1260~C (3) LiAlH4,CaH2, and NaHI-Commercial products from Metal Hydrides, Inc. (4) BF3. -Commercial diethyl etherate from Baker and Adamson. Solution was distilled immediately before use. (5) NH3. —Commercial, dried in vacuum line over sodium before use, or stored with dissolved sodium in steel tank and distilled directly into vacuum system. (6) Ethers. - (6o1) (C2H5)20. -Reagent grade, stored in vacuum line over LiAlH4 for several days before use. For larger-scale operation anhydrous grade ether was used directly from a fresh can. (88) I. Shapiro, H. G. Weiss, M. Schnick, S. Skolnik, and G.B.L. Smith, J. Am. Chem. Soc. 74, 901 (1952-). (89) A. Bo Burg and He I. Schlesinger, J. Amo Chem. Soc., 55, 4009 (1933). (90) W. H. Schechter, C. Bo Jackson, and R. Mo Adams, Boron Hydrides and Related Compounds, (Second Edition, Callery Chemical Co., Callery, Pa., 1)54, p. 17. WADC TR 59-207 125

(6.2) Tetrahydrofuran.-Product of E. I. duPont de Nemours Co., distilled once and stored in vacuum line over LiAlH4. (6.3) Tetrahydropyran.-Product of E. I. duPont de Nemours Co., distilled once and dried over LiA1H4. For larger-scale operation tetrahydropyran was refluxed for 2 hours over CaH2 and stored above Cal2 until used. (7) Benzene. —Reagent grade benzene was dried and stored over CaHll2. (8) Toluene. —Reagent grade toluene was dried and stored over CaH2. (9) Methylcyclohexane.-Technical grade methylcyclohexane was refluxed over Call2, distilled, and stored over CaHll2. (10) NaB3He. —Sodium triborohydride, first described by Hough, Edwards, and McElroy, was prepared in this laboratory by a new and easier procedure suggested from the work of Hough, Edwards, and McElroy.(91) (11) All Other Reagents. —These were always the best available commercial product and were always dried by appropriate methods prior to use. (b) The Preparation of H3NB3H7 by the Reaction Between NaB3H8 and Ammonium Halides In a typical run, 1 to 3 millimoles of NaB3H8 and an amount of ammonium chloride or bromide salt about 10% larger than the NaB3H8 sample (on an equivalent basis) were loaded in a dry box into a special reactor tube. The tube was equipped with an electromagnetically activated hopper type stirrer and consisted of a 25-mm pyrex tube, about 200 mm long and fitted on the top end with a 24/40 inner joint. This joint was sealed into an outer joint with De Khotinskey Cement. The outer joint was fastened through a U-bend to a stopcock which was equipped with a standard type of mercury protection plug to eliminate attack of ether on the stopcock grease.(92) The stopcock assembly was then fastened through a second U-bend to the vacuum system by a greased 24/40 joint. Dry diethyl ether (about 5 ml) was distilled into the reactor. The contents of the system were frozen with liquid nitrogen and the reactor was removed from the line at the greased joint. The mercury plug was put into place by inverting the unit. Then the system was warmed to room temperature. The ether slurry was magnetically stirred at room temperature for periods ranging from 20 to 40 hours. At intervals the mercury plug was removed from the stopcock by a reversal of the above procedure, the system was replaced on the vacuum systems and hydrogen was pumped into a gas burette with a Toepler pump. The reaction was (91) W. V. Hough, Lo J. Edwards, and Ao D. McElroy, to be published. (92) G. Kodama, Ph.D. dissertation Univ. of Mich., 1957, po 82; see also for related structure R. T. Sanderson, Vacuum Manipulation of Volatile Compounds, (J. Wiley and Sons, No Y., 1948), p. 69, Type II, Fig. 19o WADC TR 59-207 126

continued until the ratio between evolved H2 and NaB3H8 used approached one. In general NH4C1 seemed to react more rapidly than NH4Br. The white precipitate which formed in the ether was removed by filtration in the vacuum line filter assembly and was identified by its X-ray powder pattern as a sodium halide and a small amount of unreacted ammonium halide. When the solvent was distilled under vacuum from the filtrate, either a white solid or a viscous liquid was left in the tube. The tube was then transferred to a cold-finger type of vacuum sublimation apparatus. When the outer tube was immersed in a 40 to 500C water bath, H3NB3H7 was sublimed from the residue to the cold-finger condenser wall (100~C). H3NB3H7 could be scraped off the cold finger in open air. Yields ranged from a trace to 38%. The microcrystals of H3NB3H7 were first characterized on the basis of an elemental analysis and by a molecular weight determination. The observed values were: B = 57.28, N = 24.4%, H2 on acid hydrolysis = 141 mmoles/g. Values calculated for H3NB3H7 are: B = 57.4%, N = 24.8%, H2 = 141.5 mmoles/g based on the hydrolysis equation: H3NB3H7 + 9H20 + 9H0 +- H - NH+ + 3B(OH)3 + 8H2 The molecular weight as determined by vapor pressure depression in diethyl ether at 210C was 55 in the concentration range of 0.35 to 0.65 molal; the theoretical value for H3NB3H7 is 56.6. The molecular weight procedure is the same as that described earlier(93) except for changes necessitated by the change in solvent. Analytical procedures have been described elsewhere.(94) (c) The Preparation of H3NB3H7 by Displacement of an Ether by Ammonia in An Etherate of Triborane (1) Large-Scale Laboratory Synthesis Involving Displacement of Tetrahydropyran. -The process described below is that which was used in preparing approximately 25-gram samples of ammonia-triborane for calorimetric studies (see Ref. 106 on p. 145) and other investigations. At the present time, it represents a relatively large-scale synthesis for compounds of this type. It is proper to note that the rapid evolution of diborane during the formation of the etherate may be dangerous if uncontrolled, and that localized heating, which may arise in large-scale operation, is a major factor in reducing yields of H3NB3H7 o The reactor consisted of a 500-ml three-necked flask. One side neck was fitted with a glass plug and the other neck with a low-temperature reflux condenser, the outlet of which was connected to a manometer and then through a stopcock to the vacuum system. The center neck was equipped with a commercial "Lew" magnetic stirrer (Sci. Glass Apparatus Co., Bloomfield, N. J.). The (93) R. W. Parry, G. Kodama, and D. R. Schultz, J. Am. Chem. Soc., 80, 24 (1958) (94) T. C. Bissot, D. H. Campbell, and R. Wo Parry, J. Am. Chem. Soco 80, 1868 (1958). WADC TR 59-207 127

reactor was evacuated to remove moisture; dry nitrogen was then admitted to the system and the glass plug was removed. A steady stream of dry nitrogen through the system minimized moisture entry while the plug was out, A 100-ml quantity of dry tetrahydropyran was added through the opening, the plug was replaced; the tetrahydropyran was cooled to -1960C and the system was evacuated. A 55-ml sample of liquid tetraborane, measured at 00C, was condensed into the reactor. The temperature was allowed to rise slowly from -1960C until evolution of gaseous diborane was noted, At this point it was necessary to immerse the flask again in liquid N2 until the reaction was moderated. CAUTION. B2H6 evolution may be sudden and vigorous. As diborane evolution proceeded, B2H6 was removed to the vacuum line where it was purified and stored. After most of the diborane had been evolved, the reaction flask was cooled to -780C and allowed to stand overnight. The remainder of the diborane, the excess tetrahydropyran, and traces of hydrogen gas were then removed from the reactor over the temperature range of from -78 to 250C. The tetrahydropyran-triborane which remained in the reactor melted just below room temperature and was unstable in contact with air. The reactor flask was flooded with dry nitrogen under slight pressure, the glass plug was removed from the side neck, and 200 ml of dry diethyl ether were added, The ethyl ether solution was cooled to -78~C, and the system was evacuated. A 15-ml quantity of anhydrous liquidammonia (measured at -450C) was then cooled to -780C and opened to the reactor. Ammonia, maintained at a constant pressure near 46 mm in the reactor, was thus absorbed slowly by the constantly stirred ether solution. Approximately 12 hours elapsed between the beginning of NH3 absorption and the beginning of the subsequent step. The flask was next warmed to room temperature and the excess ammonia, displaced tetrahydropyran, and ethyl ether solvent were distilled from the system and discarded, The white solid remaining in the flask was primarily ammonia-triborane. It was extracted with three 50-ml portions of benzene. Each portion of solution was removed through a fritted filter stick inserted through the neck of the reactor flask, An excess of methylcyclohexane (about 300 ml) was added to the benzene solution, and ammonia triborane separated as a precipitate. The product was filtered out under a stream of dry nitrogen (not in vacuum system) and was dried under vacuum, Further purification was effected using the special apparatus shown in Fig. 22. After drying the apparatus shown in Fig. 22 by passing dry nitrogen through it, a 6- to 8-gram aliquot of the crude i3NB3H7 from the process above was dissolved in about 60 ml of dry benzene and placed in funnel I. Methylcyclohexane from funnel II was introduced into the precipitation vessel and the benzene solution of H3NB3H7 was filtered through the frit in the stem of funnel I and allowed to drip into the methylcyclohexane in the precipitation vessel. By adjusting flow rates, methylcyclohexane was maintained in a six to one excess during precipitate formation, After precipitation and filtration the solid was washed with a small aliquot of methylcyclohexane and dried with dry nitrogen. WADC TR 59-207 128

METHYLCYCLOHEXANE - DRY N2 H3NB3H7 -C6H6' e k,'SOLUTION 80 mm DROPPING \\65mm DROPPING FUNNEL IS E FUNNEL I 250 ml 0 125 ml E 2 19/38 o is~L 24,4 FRITTED FILTER E DISK (C) 14/3.5 _ PRECIPITATION CHAMBER 35 mm FRITTED FILTER ~~~ I DISC (M) m:. it 24/40 Hg BUBBLER VACUUM RECRYSTALLIZATION E TUBE O 30 mm Fig. 22. Apparatus for recrystallization of H3NB3H7. WADC TR 59-207 129

While the dry nitrogen stream passed through the system, the flask used to collect the liquids from the filter was replaced by the recrystallization tube. Funnel I was removed and a 35-ml portion of toluene was added to the precipitate in the precipitation vessel. About 30 minutes were allowed for the toluene to become saturated with H3NB3H7; then the toluene solution was pulled through the frit into the recrystallization tube. The tube was removed under an N2 stream and capped. This clear solution was cooled to -95~C in a toluene slush and the H3NB3H7 which crystallized was filtered at -78oC under a dry N2 atmosphere. All retained toluene was removed from the product under vacuum at 250C. The toluene solution which was recovered contained a considerable amount of ammonia-triborane and was used further to extract the precipitate, The yield of crude ammonia triborane was 70 to 80% based on the tetraborane used. Final purification reduced the overall yield to about 50%. The high purity of the product obtained from the above process was indicated by its stability and by the analysis. Values found were: N = 24.9%; B = 57.2%; H2 on acid hydrolysis = 141 mmoles/g. Theory: N = 24.8%, B = 57.4%, H2 = 141.5 mmoles/g. (2) Small-Scale Laboratory Synthesis Involving Displacement of Tetrahydrofuran. -About 5 ml of tetrahydrofuran and 0.97 mmole of B4Ho were condensed together into a reaction tube; then the temperature was allowed to rise slowly. No reaction was observed up to -235C but at room temperature o.46 mmole of B2H6 was evolved in half an hour. A trace of H2 gas (0.027 mmole) was collected. Crystalline tetrahydrofuran-triborane remained in the tube when excess tetrahydrofuran was removed. The solid was dissolved in excess tetrahydrofuran again and 0.96 mmole NH3 was frozen above the etherate (-196~ bath). The system was allowed to warm to room temperature and stirred for half an hour, The solvent was removed at 00~C From the white solid residue a 42.5-mg sample of H3NB3H7 was sublimed (X-ray identification). The yield of still impure product was 79%. Diethyl ether was also used as a solvent for the displacement reaction but neither the yield nor product quality was improved (69%). (3) Attempted Small-Scale Laboratory Synthesis Using Diethyl Ether.-A sample of B4Hjo (1.40 mmole) was dissolved in about 3 ml of diethyl ether and the solution was allowed to warm to ice temperature. A very slow and small vapor pressure rise due to B2H6 evolution was observed. The system was maintained at ice temperature for 4 hours. Then to permit continuous B2H6 removal, the stopcock was opened very slightly to the fractionation train while the reaction mixture was maintained at room temperature. B2H6 which entered the train was carefully fractionated by vacuum distillation (-78 and -126~C traps) until 0.632 mmole of B2H6 was obtained. BH3/B4H1o ratio = 0.975. During the foregoing process less than 0.04 mmole H2 was liberated, Then 1.25 mmoles of NH3 was introduced slowly to the stirred solution at -78~Co A small amount of solid B4H1o'2NH3 remained in the tube (identified by its X-ray powder pattern) along with a viscous liquid~ Only a trace of H3NB3H7 could be sublimed from the reaction mixture, WADC TR 59-207 130

(4) Small-Scale Laboratory Synthesis Using Tetrah)ydropyran.-A sample of tetraborane (1.69 mmoles) was condensed into 2 ml of tetrahydropyran and the system was warmed to room temperature. At room temperature the reaction was much slower than that with tetrahydrofuran.'To minimize attack of the tetrahydropyran on the stopcock grease in the system, the temperature of the reactor was reduced to 0~C and maintained for 15 hours. The B2H6 evolved (0.817 mmole) was separated by fractionation. The excess tetrahydropyran was removed at 0~C; 3 ml of diethyl ether was condensed into the system; then a sample of ammonia (3.4 mmoles) was added to the reactor and a temperature of -780C was maintained for 10 hours. On removal of the ethers a dry solid was left in the system. H3NB3H7 was sublimed at 52~C to give an 81% yield (77.8 mg). In a second run using 4.75 mmoles of B4H1lo and 1.5 ml of tetrahydropyran conditions were the same as above except the system was allowed to stand for 24 hours at -780C after addition of 9.80 mmoles (twofold excess) of ammonia. A 94% yield (.253 g) of H3NB3H7 could be sublimed from the solid residue. (d) The Reactions of H3NB3H7 (1) The Reaction of H3NB3H7 with Sodium Dissolved in Liquid Ammonia. -A sample of H3NB3H7 (.189 mmole) was dissolved in about 1.5 ml of liquid ammonia; sodium metal (0.1 g) in a small glass tube was added in two portions by breaking the tube immediately before addition. After introducing the first portion the solution was warmed to -78~C. The blue color of Na faded rapidly and 0.10 mmole of H2 was evolved. No more H2 was evolved at -78~C over 12 hours. The second portion of Na was then added as above. In 20 minutes 0.012 mmole of H2 was given off. After 5 hours an additional 0.057 mmole appeared. Total H2 = 0.169 mmole H2/H3NB3H7 = 0.895. Further gas evolution was very slow and the solution remained blue. The excess of Na was removed from solution by amalgamation at -35~C. Additional H2 (.007 mmole) was given off during' the amalgamation. The sodium-free clear solution was filtered from the amalgam, and the solvent ammonia was distilled from the filtrate. In the white solid residue NaBH4 was detected by its X-ray powder pattern, but no component soluble in diethyl ether could be extracted from the residue. (2) The Reaction of H3NB3H7 with Trimethylamine. —A sample of H3NB3H7 (.55 mmole) was placed in a tube and trimethylamine (0.565 mmole) was condensed above it. The system was allowed to warm slowly. At -78~C no visible reaction was observed, but as the temperature rose and solid H3NB3H7 began to dissolve in liquid trimethylamine, fairly rapid reaction was noted. Without waiting for the completion of the reaction, the noncondensable gas and the volatile components were removed from the system. They were.: H2 =.102 mmole, (CH3)3N =.24 mmole, and (CH3)3NBH3 about 0.35 mmole. From the solid polymeric residue 0.20 mmole of H3NB3H7 was sublimed on warming to 40 to 50~C. No (CH3)3NB3H7 could be detected as a product. WADO TR 59-207 131

D. The Structure of H3]NiB3H7 from a Single-Crystal X-ray Study 1. EARLIER STRUCTURAL INFORMATION The cleavage hypothesis used as a basis for correlating the chemistry of diborane and tetraborane [see earlier sections] suggests that the B3H7 group arises from symmetrical cleavage of the double bridge bond of tetraborane. H B H3B B3H7 Since the structure of B4Hlo is accurately known,(95) an accurate determination of the structure of H3NB3H7, including the possible identification of B3H7 as a fragment of B4Hlo, would be a significant contribution to our understanding of the chemistry of these compounds. 2. HIGH-TEMPERATURE FORM (a) Experimental Samples of pure NH3B3H7 were kindly furnished by Dro Kodamao The compound is a crystalline solid at room temperature, stable enough to allow recrystallization from a number of common solvents such as acetone, alcohol, and ether. Powder and single-crystal diffraction patterns showed the crystals to be tetragonal with a = 6.11 and c = 6.57A, giving a calculated density of 0.765 g cm-3, assuming two formula units per cell. Systematic absences observed for h+k+~ odd call for a body-centered lattice. These facts require each molecule in the crystal to have fourfold symmetry. This is unreasonable in view of the chemical composition of the molecule. We were therefore led to conclude that the structure is a disordered one with the molecules either axially rotating or having random orientation so as to conform to the fourfold symmetry in a statistical sense. Diffracted intensities were recorded and visually estimated using powder, Weissenberg and precession techniques. Due to the rapid falloff of the intensities with increasing angle of diffraction, no more than 25 different intensities (95) M. E. Jones, K. Hedberg, and V. Schomaker, J. Am. Chem. Soc. 75, 4116 (1953); C. E. Nordman and W. N. Lipscomb, ibid., 75, 4116 (1953); J. Chem. Phys. 21, 1856 (1953); E. Bo More, R. E. Dickerson and W. N. Lipscomb, Jo Chem. Phys. 27, 209 (1957). WADC TR 59-207 132

(hki) could be measured, The presence of disorder led us to investigate the possible existence of a phase transition at lower temperature. A transition was indeed found, but the equilibrium temperature of the transition could not be accurately determined due to considerable supercooling. It was, however, established that the transition point is no lower than -160C. Single crystals cooled through the phase transition shattered completely. (b) Structure Determination Since disorder about the c axis direction was known to exist, the space groups having a complete set of mirror planes parallel to that direction appeared to be the most likely choices. These are C9v - I4mm and D4h - I4/mmm. The latter of the two requires an additional disorder with the molecules pointing up or down the c axis at random, barring the chemically very unlikely possibility that the molecules possess a mirror plane perpendicular to their axis of disorder. The high dipole moment of 645 Debye units found for ammonia-triborane (see page 73 ) speaks in favor of the polar space group I4mm, which permits an energetically favorable arrangement of dipoles pointing in the direction of the c axis. This kind of arrangement has previously been found in structures composed of sterically simple or axially disordered molecules with high dipole moments such as hydrogen cyanide(96) and ammonia-borane.(97) The structure was determined by trial and error. Since only 25 independent intensities have been observed it was essential to keep the number of parameters describing the assumed models as low as possible. The axially disordered molecules were therefore assumed to have cylindrical symmetry about the c axis. The structure factors are then given by: N/2 F(hki) = 2 f Jo(ir h2 + (cos 2+J Lz + i sin 2r z )z j=l_ -j — j where j ranges over the atoms in one molecule, r. is the radius, and z. is the J.J z coordinate of the "doughnut" formed by the jth-atom. The model shown in Fig. 25(a) was first tried. This model has two adjustable parameters, z and r; the origin is taken as the center of the boron ring. These two parameters, the temperature factor parameter B, and the scale factor were refined by several least-squares cycles, yielding a final value of R = YZ I Fol - IFc j /ZIF0I of 0.16. A p(xOz) Fourier section computed at this stage revealed an appreciable electron density in the region between the boron ring and the nitrogen atom. This fact and the abnormally long B-N distance in the (96) W. Jg Dulmage and Wo N. Lipscomb, Act Cryst. 4, 55330 (1951). (97) Eo W. Hughes, J. Am. Chem. Soc. 78, 502 (1956); E. L. Lippert and W. N. Lipscomb, ibid., 73, 503 (1956). WADC TR 59-207 133

Ld N N 0 I Fig / 23.Tral odls f'he mmB Bm B~~~~ B'B 03v 02v C Fig. 25. Trial models of the arnmonia-triborane molecule. The z-direotion is vertical.

least-squares result led us to try the three-parameter model in Fig. 23(b). Several cycles of least-squares refinement now produced an R value of O.18, A p(xOz) electron-density section and the corresponding difference synthesis now showed the distribution of the central boron atom BT to be distinctly diskshaped, requiring a nonzero rB.o To a lesser degree the same was found to be true of the nitrogen atom. The two additional parameters -IB' and _N were therefore introduced [Fig. 23(c)] and a series of least-squares refinements of all seven parameters (including B and the scale factor K) was carried out. The agreement now improved considerably, giving an R value of 0.096. An approximate hydrogen contribution with a strong additional temperature factor was now calculated assuming a reasonable distribution of the hydrogen atoms about the nitrogen and boron atoms. Following a final, slight refinement of the seven aforementioned parameters the final R factor was 0.074. The p(xOz) electron-density section calculated at this point is shown in Fig. 24. The corresponding (Fo-Fc) synthesis showed no significant region of disagreement. The observed and final calculated structure factor magnitudes are listed in Table XXVI. To examine the validity of the assumption that the disordered molecules have cylindrical symmetry, the p(xyO) and p(xyz, ) sections were computed. Neither one revealed any noticeable departure from cylindrically distributed molecules. This may be due to strong angular motion of the molecules or to poor resolution afforded by the available data, or both. The X-ray data do not warrant any detailed conclusion regarding the angular distribution or the amount of rotational freedom of the disordered molecules. The final least-squares parameters are rB = 0.151, rB' = 0.159, _B' = 0.190, rN = 0.040 and =- o.408. The parameter B in the temperature factor exp(-B sin2Q/X2) is 8,4A2, The configuration of the molecule, shown in Fig. 23(c), can be described as a triangle of boron atoms with an out of plane NH3 group attached to one corner. No bond distances can be deduced unambiguously, but certain limits can be placed on them. The diameter 2rB or the two atom boron ring is 1.85 A; this is the maximum value of the B-B distance. The B-B' distances are approximately 1l8A, assuming that they are equal and that B-B equals 2rB. The B'-N distance must be in the range 1.7 ~ Oo.2A. These bond lengths agree satisfactorily with values previously found in molecules containing comparable bonds. 3. LOW-TEMPERATURE FORM (a) Experimental To prepare crystals of the low-temperature form, a diethyl ether solution of the compound was evaporated while maintained under a slow stream of dry nitrogen in a vial immersed in a chlorobenzene slush bath at -45~Co In this way WADC TR 59-207 15

\J1 0 / \ ~ 1 —'' R) - I\ 0 / \, / I// "' x I / o/ -J~x-' \,'', I I / X /~~~~~~~~~~~~ Fig. 24. Electron-density distribution at y = O in the unit cell of the disordered, high-temperature structure. Contours are at intervals of 0.4 e A-3; zeroe,'dontour broke

TABLE XXVI OBSERVED AND CALCULATED STRUCTURE AMPLITUDES (HIGH-TEMPERATIURE FORM) h k X IFOI PFc h k IFoI IFcI h k o IFCF 0 0 0 -- 64.0o 2 1 1 6.53 5.6 2 1.4 1.5 0 0 2 14.9 159 2 1 3 3.1 5.1 3 3 0 <o0.8 0.0 0 0 4 4.1 4.4 2 1 5 1.5 1.2 3 3 2 <1.9 1.53 1 0 1 135.1 14,2 2 2 0 35.1 35. 4 0 0 <0.9 0.1 1 0 3 6.2 5~7 2 22 35.2 35.53 4 0 2 <1.8 1.5 1 0 5 37 3.5 2 2 4 <1.3 103 4 1 1 2.2 2.4 1 1 0 25.5 25.6 3 0 1 4.5 4.2 4 1 3 <1.6 1.0 1 1 2 9.4 8.4 3 0 3 1.9 2.0 4 2 0 <1.2 0.0 1 1 4 2.7 2.6 3 1 0 1.5 1.7 4 2 2 <1.9 1.0 2 00 0 126 11.5 3 1 2 2.0 2,7 4 3 1 <1.2 1.1 2 0 2 61 56 3 1 4 1. 1.5 501 145 1.1 2 0 4 2.0 17 5 2 1 2.7 35. 5 1 0 <1.3 0.2 satisfactory crystals could be obtained in a few hours. The crystals were mounted and sealed in thin-walled glass capillaries on a simple microscope cold stage. While on the X-ray camera the specimen was maintained at -80 + 100C by means of a stream of cold nitrogen gas, The crystals are monoclinic and belong to the space group P21/n. The lattice parameters and their estimated standard deviations are a = 10.40 ~ 0.015A b = 4.824 + o.oo6 c = 9,997 + 0.012 5 = 115.2 ~ 0.15~ These values were derived from measurements of high-angle reflections on a number of precession patterns of zero-level principal and diagonal nets. With four molecules of H3NB3H-7 per unit cell the calculated density is 0.827 g cm-3, an increase of 8% over the density of the high-temperature form, No molecular symmetry is demanded, Data were collected as zero- and upper-level patterns on the Buerger precession camera using MoKC radiation. Reciprocal space was covered systematically to sin Q/k = 0.572, corresponding to a zero-level precession angle of 24~. Within this range 69% of the diffraction maxima was of measureable magnitude, 25% was unobservably weak and 6% was not evaluated. Including 29 spots observed outside the range, a total of 502 reflections were observed. WADO TR 59-207 137

Intensities were measured visually by comparison with a scale of timed exposures, and reduced to structure amplitudes in the usual way. The calculation of the Lorentz-polarization factor(98) and all subsequent least-squares refinements and Fourier syntheses were performed on an IBM 650 computer. (b) Structure Determination With the knowledge of the structure of the boron-nitrogen skeleton gained from the high-temperature form, approximate boron and nitrogen coordinates were easily found. The hOL Patterson projection was readily interpreted; x and z coordinates for the boron and nitrogen atoms were obtained from the corresponding Fourier projection. Approximate y coordinates were found by trial and error. A three-dimensional least-squares refinement of these coordinates and the scale and temperature factors, K and B, was then carried out. The quantity R' = _w(KFo - Fc)2/F w Ka2Fo2 was minimized taking w(hki) as 3 1Fmin l/Fo(hki)| or unity, whichever was smaller. McWeeny(99) atomic scattering factors were employed; in the case of boron the "average" scattering factor f was used. Toward the end of the refinement a contribution was included for six half-hydrogens forming a ring about the nitrogen atom. This was done to account in an approximate way for the three ammonia hydrogen atoms, whose approximate locations could be inferred. No other assumptions were made regarding hydrogen atoms. An R factor of 0.19 was reached. The seven hydrogen atoms on the boron skeleton were now located by means of three-dimensional (KFo - EN,B) difference Fourier syntheses. This was done in two steps; only the strongest of the presumptive hydrogen peaks in the first difference synthesis were included in an intermediate model, which was then subjected to further refinement of N and B coordinates. A second (KFo - FN B) synthesis yielded the rest of the hydrogens. This procedure was probably unnecessarily conservative, since in essence all peaks found in the first difference Fourier were ultimately confirmed as hydrogen atoms. Anisotropy was observed in the thermal motion of the nitrogen and, to a small extent. boron atoms. This was accounted for in an approximate way by applying two anisotropic temperature corrections, one to fN and another to fB, common to all boron atoms. With the hydrogen atoms and the partial anisotropy correction included least-squares refinement of K, overall B, and nitrogen and boron coordinates improved the agreement to R = 0.117. A set of sections of the three-dimensional (KFo - FN,B) synthesis computed at this stage are shown in Fig. 25. All spurious peaks are less than one half the height of the lowest hydrogen peak. This does not include the region of moderately high electron density found in the neighborhood of the nitrogen atom and presumably due to the three unresolved ammonia hydrogens. (98) J. Waser, Rev. Scio Instr. 22, 563 (1951); R. D. Burbank, ibid.,23, 321(1952). (99) R. McWeeny, Acta Cryst. 4, 513 (1951). WADC TR 59-207 138

I 2 0 C Fig. 25. The structure of the H3NB3H7 molecule. Left: The electron density distribution in the H3NB3H7 molecule represented by sections near the centers of the hydrogen atoms. The nitrogen and boron atoms have been subtracted out and are indicated schematically. Contours are at -0.25 (broken), +0.25 and +0.50 e A-3. Right: Schematic drawing of the same molecule and key to the numbering of atoms.

The method employed in refining the assumed model structure was now revised to allow all atoms to assume individual, but isotropic, thermal parameters, and several cycles of least-squares refinement of all coordinates and thermal parameters were carried out, The shifts in coordinates were slight and led to final values of R = 0.107 and R' = 0.019. An attempt was now made to find the locations of the three ammonia hydrogens, which had not been found as resolved peaks in any of the difference syntheses. Vaguely suggested maxima in the neighborhood of the nitrogen atom were taken as a starting point for a series of least-squares refinements of all parameters except those relating to the seven hydrogens already found. This refinement failed to improve the agreement previously obtained; in seven cycles the values R = 0.113 and RI = 0.021 were reached. Perhaps more significantly, the thermal parameters of the three ammonia hydrogens refined to very high values, 5.3, 111ol and 11.8, suggesting a high degree of rotational freedom. Since, in addition, the H-N and H-H distances in the I\3 group refined to unconvincing values, we conclude that the three hydrogens cannot be located unambiguously, presumably due to strong angular motion about the B-N bondo The observed and final calculated structure factors are given in Table XXVII. The values and standard deviations of the atomic coordinates and the thermal parameters are listed in Table XXVIIIL The standard deviations were deduced in the following way: The observed structure factors were divided into six equal sets with the same distribution in sin Q/k, but otherwise selected at random. Starting with the parameters given in Table XXVIII, each set was separately subjected to least-squares refinement. Usually the refinement was virtually completed in four or five cycles. The resulting six sets of coordinates and thermal parameters constitute, in effect, six independent structure determinations. The standard deviations of the final parameters can be taken as the standard deviations of the means of these quantities as given by (E r2) /2/ln(n-l) ]/2 ii where ri's are the deviations from the means of the individual values, and n = 6 in this case. The difference between the mean of a parameter as obtained from the six sets and the final (Table XXVIII) value was in all cases less than the standard deviation. 4. DISCUSSION Int.eratomic distances and bond angles in the ammonia-triborane structure are given in Table XXIX, The standard deviations were obtained from an analysis of the interatomic distances calculated from the six separate sets of coordinates. They nclude contributions from the standard deviations in the cell parameters, which in the case of the B-N and B-B distances are comparable to WAD3 TR 59-207 140

TABLE XXVII OBSERVED AND CALCULATED STRUCTURE FACTORS (LOW-TEMPERATURE FORM) h k I KFo F0 h k I KFO F0 h k I KFO F0 h k I KFO Fe h k I KFO F0 h k I KFO F0 o 0 0 -- 128.0+ 1 5 4- 7.0 6.6- 3 0 7- 6.7 5.2+ 4 2 4- 8.9 8.5+ 6 1 10- 1.5 2.1- 8 1 1- 1.1 1.40 0 2 2.9 1.4- 1 3 5- 10.0 10.5+ 3 0 5- 8.9 9.3- 4 2 5- 1.9 2.4+ 6 1 8- 5.6, 5.2- 8 1 0 2.7 2.1004 12.4 1.8+ 1 3 2- 5.0 2.1- 3 0 3- 17.3 19.7+ 4 2 2- 3.7 3.1+ 6 1 7- 8.6 8.9- 8 1 1 3.4 3.3o o 8 6.0 5.9+ 1 3 1- 6.7 6.4+ 3 0 1- 22.3 28.2+ 4 2 1- 5.9 4.8+, 6 1 6- 6.6 6.4+ 8 1 2 2.4 2.30 0 10 2.8 2.6+ 13 4.8 3.9+ 3 0 1 16.5 l6.3+ 4 2 0 6.8 6.0+ 615- 2.5 2.8+ 8 1 5.6 5.50 1 1 24.8 24.7+ 1 5 1 10.5 10.1+ 3 0 3 3.1 2.7- 4 2 1 11.5 11.0+ 6 1 4- 4.5 5.4- 8 1 4 3.3 4.o+ 0 1 2 0.4 0.1+ 1 5 3 1.9 1. 5 0 5 15.5 12.6+ 4 2 2 2.7 2.5+ 6 1 5- 2.5 1.5- 8 1 5 1.7 0.90 1 3 21.3 20.0+ 1 3 4 3.3 3.9- 3 50 9 7.2 6.4+ 4 2 3 9.2 7.1+ 6 1 2- 7.4 6.9- 8 2 8- 3.5 3.50 1 4 8.7 8.9- 1 3 5 4.3 4.5+ 3 1 9- 3.8 4.1- 4 2 4 8.1 6.9- 6 1 1- 4.7 4.7+ 8 2 6- 5.3 5.9+ 0 1 5 1.6 1.6+ 1 3 8 3.3 2.9- 3 1 8- 2.8 2.7- 4 5 6- 3.6 3.7+ 6 1 o 2.6 5.2- 8 2 5- 3.4 4.2+ 0 1 6 11.5 0o.8+ 1 4 4- 5.8 6.9+ 3 1 7- 8.1 7.5- 4 3 3- 1.9 2.4- 6 1 1 1.9 1.4- 8 2 4- 4.0 4.00 1 7 5.8 5.8+ 1 4 2- 4.9 4.5+ 3 1 6- 1.7 1.7- 4 3 2- 9.5 10.7+ 6 1 2 2.5 2.4- 8 2 5- 7.9 8.3+ 0 1 8 2.4 1.6+ 1 4 6 4.4 4.0+ 3 1 5-5.5 4.8+ 4 3 1- 14.1 13.5+ 6 1 4.5 3.6+ 8 2 2- 2.7 3.10 19g 1.7 1.8t 1 5 41.8+ 4-1.81.9+ 14-5.6 5.7+ 451 4.6 5.2- 4.2 3.6+ 82 1- 3.8 3.00 1 10 1.1 1.2- 1 5 5- 5.0 5.4+ 3 1 3- 6.6 6.1- 4 3 2 7.6 7.9+ 6 1 5 5.2 5.1+ 8 2 1 2.9 2.0+ 0 1 11 2.6 0.6+ 1 5 2- 3.0 2.8+ 3 1 2- 16.8 18.1+ 4 3 4 2.81.+ 6 1 6 1.2 1.8- 8 3 8- 2.9 3.2+ 0 2 0 5.5 5.8+ 1 5 1- 4.0 5.2+ 3 1 1- 10.9 11.0- 4 4 7- 3.4 3.5- 6 2 9- 2.7 2.6- 8 3 7- 5.5 3.2+ 0 2 1 12.9 15.0- 1 5 0 3.8 4.2- 3 1 0 12.4 1o.g+ 4 4 5- 5.9 6.5+ 6 2 8- 2.8 2.5- 8 3 6- 2.5 2.5+ 0 2 2 6.5 6.4+ 1 5 1 3.8 4.0+ 3 1 1 9.9 8.57 - 44 - 4.1 3.9+ 6 2 5- 3.1 3.2- 8 3 4- 3.2 3.4+ 0 2 3 7.0 6.1+ 2 o 10- 6.8 6.2- 3 1 2 11.4 10.1+ 4 4 1- 5.0 5.7+ 6 2 4- 3.02.7- 8 3 1- 4.8 4.90 2 4 12.4 12.7+ 2 o 8- 1.4 2.0- 3 1 3 4.6 45- 4 5 2- 5.0 6.5+ 62 3- 3.3.2- 8 3 0 3.5 2.9+ 0 2 5 5.2 2.7+ 2 0 6- 3.3 3.2- 3 1 4 5.9 3.3+ 4 5 i- 3.4 4.0- 6 2 2- 5.0 2.4- 8 4 5- 1.5 1.80 2 7 5.5 5.4+ 2 0 4- 22.8 24.7- 3 1 5 7.6 7.5+ 5 0 11- 2.9 2.4+ 6 2 1- 12.0 12.7- 9 0 9- 5.6 4.70 2 8 4.6 4.0+ 2 0 2- 40.1 52.6- 3 51 6 7.6 7.7+ 5 0 9- 5.2 4.0- 6 2 5 3.2 2.9+ 9 0 7- 2.3 2.60 2 5 3.2 2.5+ 2 0 0 6.2 6.4+ 3 1 8 4.3 3.5+ 5 0 7- 7.1 5.8+ 6 2 2 4.7 5.9+ 9 0 5- 2.1 2.40 3 2 3.8 5.5+ 2 0 2 19.1 18.0- 5 2 10- 7.5 5.7- 5 0 5- 18.9 20.2- 6 2 5 1.7 1.5- 9 0 1- 5.5 5.90 3 3 3.5 2.6+ 2 0 4 6.2 5.2+ 3 2 8- 2.2 2.5- 5 0 3- 7.4 8.9- 6 3 8- 2.3 2.9- 9 0 1 5.1 2.7o 3 4 3.8 4.0+ 2 o 6 10.7 9.2- 3 2 7- 3.8 3.6+ 5 0 1- 9.3 10.5- 6 3 7- 3.3 2.6+ 90 3.14..9+ 0 3 5 8.6 8.4+ 2 0 10 5.9 3.7+ 3 2 6- 6.5 6.3- 5 0 1 5.9 5.4- 6 3 6- 4.0 5.9- 9 1 lo- 4.1 4.70 3 8 2.8 1.9+ 2 1 11- 5.9 4.1- 3 2 3- 8.9 9.0+ 5 0 5 9.9 9.2- 6 3 5- 3.7 5.7- 9 i 8- 1.9 2.2o 4 1 5.1 2.8+ 2110- 6.0 5.2+ 3 2 2- 6.3 5.3- 505 5.5.- 6 3 4- 5.060- 9 1 7- 4.2 4.4+ 0 4 3 6.1 5.8+ 2 1 7- 7.6 7.2- 3 2 1- 8.7 8.4- 5 0 7 5.1 4.1- 6 3 5- 5.5 5.5- 9 1 6- 4.0 4.50 4 4 2.7 2.2+ 2 1 6- 1.2 1.6- 5 2 0 11.5 11.5- 5 1 11- 5.4 5.1+ 6 3 1- 4.4 4.4+ 9 1 5- 2.5 2.60 4 5 2.5 1.9- 2 1 5- 6.8 7.4- 3 2 1 16.1 15.5+ 5 1 7- 5.2 2.6- 6 3 0 5.9 6.5- 9 1 4- 1.1 0.5+ 0 4 7 1.7 3.0+ 2 1 4- 2.4 1.4- 5 2 2 5.0 4.1- 5 1 6- 8.6 9.1- 6 3 2 2.2 1.5- 9 1 5- 5.5 5.70 5 2 2.4 2.2+ 2 1 3- 23.6 23.7- 3 2 3 5.1 5.2+ 5 1 5- 3.8 3.9+ 6 3 3.1 4.5+ 9 1 2- 7.3 7.30 5 4 2.0 2.2+ 2 1 2- 7.8 8.7- 3 2 4 4.2 3.9+ 5 1 4- 8.3 8.7- 6 4 5- 5.5 5.9- 911- 1. 5+ 1 0 11- 5.8 5.5+ 2 1 1- 16.4 16.2- 3 2 5 5.5 4.9+ 5 1 5- 1.6 i.o- 6 4 5- 5.2 5.5- 9 1 0 2.0 1.53+ 1 0 5- 9.6 9.4+ 2 1 0 14.3 15.7+ 3 2 8 2.9 2.2+ 5 1 2- 12.0 12.8- 6 4 1 2.4 2.2- 9 1 2 2.5 2.51 0 3- 2.6 2.8+ 2 1 1 21.2 19.5- 3 3 9- 4.5 5.6- 5 1 1- 8.9 7.9+ 7 0 7- 12.1 14.3+ 9 1 3 1.2 1.6+ 1 0 1 20.1 22.5- 2 1 2 s.6 4.8- 3 3 8- 2.7 2.1- 5 1 0 5.4 2.4- 7 0 3- 8.1 8.7+ 9 2 7- 5.5 4.01 0 3 28.1 28.6- 2 1 3 0.7 0.9- 3 3 6-11.1 9.4+ 5 11 5. 4.9- 70 1- 2.7 3.4+ 9 2 3- 4.3 4.61 0 5 7.2 7.6+ 2 1 4 17.4 15.5- 3 3 5- 8.9 8.3- 5 1 2 10.4 9.7- 7 0 1 7.1 6.2+ 9 2 2- 4.3 5.71 0 7 11.6 9.5- 2 1 5 9.3 8.4- 3 3 4- 3.5 3.9- 5 1 3 7.2 6.9- 7 0 3 6.8 6.7+ 9 2 1- 3.4 2.9+ 1 0 9 4.3 3.1- 2 1 7 3.4 3.0+ 3 3 1- 5.6 5.4- 5 1 4 5.3 5.1- 7 0 5 1.5 2.0- 9 2 1 2.4 1.61111- 2.8 1.2- 2 1 8 2.2 1.7- 3 3 1 1.7 2.1- 51 5 1.7 2.4+ 7 1 11- 1.7 1.6+ 9 3 7- 3.0 3.81 1 10- 3.8 3.7+ 2 1 9 1.7 1.7- 3 3 2 5.0 2.7+ 5 1 6 5.7 5.2- 7 1 8- 5.9 7.2+ 9 3 6- 6.6 6.51 1 7- 2.0 1.8+ 2 2 10- 5.5 5.1- 3 3 4 5.5 5.9+ 5 2 8- 5.6 5.0+ 717- 1.1 o.6+ 9 3 3- 3.5 3.21 1 6- 5.8 5.9+ 2 2 9- 4.1 3.8+ 3 36 2.5 1.8- 5 2 7- 2.62.5- 7 1 6- 2.9 3.3+ 1000-.4 3.5+ 1 1 5- 10.8 10.7+ 2 2 6- 12.5 12.9- 3 4 6- 5.1 5.7- 5 2 6- 6.1 4.6- 7 1 5- 2.2 2.7- 10 0 8- 4.3 4.4+ 1 1 4- 0.9 0.9- 2 2 5- 5.8 5.5- 3 4 4- 5.9 5.7- 5 2 5- 4.4 5.8- 7 1 4- 9.5 9.6+ 10 0 2- 6.4 6.8+ 1 1 3- 6.9 6.6+ 2 2 3- 4.7 4.6- 5 4 2- 10.2 10.8- 5 2 4- 6.4 6.6+ 7 1 3- 4.4 3.8- 10 0 0 2.4 2.1+ 1 1 2- 15.8 16.0+ 22 2- 9.2 8.7- 3 4 o 6.0 6.7+ 5 2 2- 6.3 5.9+ 7 1 2- 1.0 1.2- 10 0 2 5.7 5.7+ 1 1 1- 20.1 19.5- 2 2 1- 4.2 4.5+ 3 4 2 3.8 4.5- 5 2 1- 16.9 16.4- 7 1 1- 9.7 9.5+ 10 1 7- 3.3 3.9+ 1 1 0 30.6 52.5- 2 2 o 1o.9 9.6- 3 5 3- 2.8 2.0- 5 2 2 5.2 2.8- 7 1 0 11.6 10.6+ 10 1 6- 2.9 5.61 1 1 27.5 28.0+ 2 2 1 12.1 10.2- 4 0 12- 5.1 4.5+ 5 2 5 7.0 6.6- 7 2 6- 6.2 6.2- 10 1 4- 1.2 1.81 1 2 6.5 5.1+ 2 2 2 11.7 9.8- 4 o lo- 1.6 1.2- 5 2 4 2.4 2.5- 7 2 5- 4.6 5.1+ 10 1 5- 5.0 5.1+ 1 1 3 10.0 8.2+ 2 2 3 4.9 5.9- 4 0 8- 3.3 3.7+ 5 2 5 5.2 2.0+ 7 2 3- 6.4 5.9+ 10 1 2- 5.6 4.5+ 1 1 4 14.1 15.6- 2 2 4 4.1 3.3- 4 0 6- 15.8 16.4+ 5 s 8- 5.2 2.8- 7 2 1- 5.7 3.0+ 10 1 1- 2.5 5.1+ 1 1 5 5.7 5.0+ 2 2 5 10.0 9.0- 4 o 4- 3.7 4.5+ 5 3 7- 3.6 2.7+ 7 2 0 5.4 5.4+ 1o 1 1 4.6 4.8+ 1 1 7 2.5 1.8- 2 2 8 2.6 2.5+ 4 0 2- 11.6 12.9+ 5 3 6- 0. 5 2.2+ 7 2 1 3.8 2.6+ 10 2 8- 2.2 5.21 1 8 5.8 5.4- 2 3 9- 3.8 2.6- 4 0 0 5.5 4.3- 5 3 5- 5.0 2.8+ 7 2 2 5.5 3.5+ 10 2 7- 4.2 4.11 19 5.2 2.7- 2537- 6.16.1- 4 02 1.8 1.9+ 555- 2.2 1.9+ 7537-5 3.05.5- 10 2 2- 5.2 3.3+ 1 1 10 3.4 2.8- 2 3 6- 4.1 3.5+ 4 0 4 7.8 7.4+ 5 3 0 6.3 6.7- 7 3 6- 2.7 2.2- 10 5 5- 4.0 4.7+ 1 210- 1.6 2.0- 2534-53.553.- 40o6 4.8 4.4- 5531 2.4 0.8- 7535- 2.82.5+ 11 0 9- 3.553.5+ 1 2 9- 4.1 3.6+ 2 3 3- 5.0 5.3- 4 0 8 4.9 4.1- 5 5 2 5.9 4.4- 7 3 2- 5.7 6.2+ 11 0 7- 4.0 5.91 28-53.4 2.7+ 2531-5 3.25 3.9+ 4 112- 4.1 5.7- 5534 4.3 4.7+ 7531- 2.52.1+ 11 0 5- 3.9 4.6+ 1 27- 5.8 5.4+ 2530 9.2 7.4- 4 110o-2.5 2.5+ 5355 4.9 5.5- 7532 2.6 2.5- 11053-2.0 2.9+ 1 26- 9.7 9.5+ 2531 11.79.8- 4 19-7.7.4 71+ 5 48-2.15 3.5+ 755 3 5.6 4.0+ 11015 3.75 3.41 25- 4.0 5.7- 2532 2.2 1.8- 41i6- 8.48.1+ 5 46-5.5 5.0+ 7 46- 2.42.8+ 11 1 7- 3.4 3.8+ 1 2 4- 4.4 4.4+ 2 5 3 5.1 2.6- 4 1 5- 1-5.7 17.0+ 5 4 4- 2.9 2.6+ 8 0 6- 3.9 3.4- 11 1 5- 1.1 1.6+ 1 2 5- 2.2 1.1- 2 3 4 4.9 4.5- 4 i 4- 1.4 1.5- 5 4 2- 5.9 6.8- 8 0 4- 7.4 7.2- 11 1 4- 1.7 1.8+ 1 2 2- 10.6 10.5+ 2 5 5 5.5 2.8+ 4 1 5- 1.9 1.5+ 5 4 l- 2.9 2.0- 8 0 2- 8.2 7.6+ 11 1 5- 1.2 1.5+ 1 2 1- 5.0 5.2- 2 3 6 3.4 2.5- 4 1 2- 6.2 4.8- 5 4 0 5.35.6+ 8 0 0 9.8 11.4- 11 2 5- 5.1 5.71 2 0 10.5 9.4+ 2 4 7- 5.2 2.9- 4 1 1- 12.7 11.6+ 6 0 12- 1.8 1. o+ 8 0 2 1.2 1.6- 12 0 6- 2.1 2.51 21 8.7 7.4+ 2 45-7.7.8 77+ 4102 20.4 18.7+ 6010l- 5.4 4.5- 80o4 8.1 6.6- 120o 4- 1.4 1.71 2 2 9.5 8.,5+ 2 4 5-. 2 5.3- 4 1 1 2.4 2.7- 6 0 8- 7.2 8.0- 8 1 11- 2.1 1.8+. 12 0 2)- 4 ".5, 5.

TABLE XXVIII Q Fa3~~~~~ ~ATOMIC PARAMETERS AND THE IR STANDARD DEVIATIONS 0 Atom x y z B N 0.2007 + 0000012 o.4383 + 0.0003 o.4799 + 0.00013 2.91 + 0.15 Bs -0.0641 + o.ooO5 o.5437 + 0.0009 0.2441 + o.ooo4 3.18 + 0.09 B2 0.0629 + 0.0003 o.6860 + o.oolo 0.1935 + o.ooo4 3.14 + o 14 B3 0.1000 + 0.0002 o.6805 + o.ooo5 0.3868 + 0.0003 2.62 + 0.08 H1 -0.153 + o.0o6 o.678 + o.olo 0.232 + 0.005 2.8 + 1.1 H2 -o.o67 + o.oo4 0.313 + o.0o6 0.283 + 0.0053 3.0 + 07 H3 -o.o63 + 0.007 0.538 + o.0o8 0.122 + 0.003 2.3 + 1.4 H4 0.042 + 0.009 o.893 + 0.016 0.138 + 0.007 4.4 + 0.8 H5 0.148 + 0.003 o.549 + o.oll 0.190 + 00005 1.8 + 0.8 H6 0.169 + 0.005 o.834 + 0.007 0.365 + o.0o4 2.0 + 0.7 H7 0.052 + 0.005 0.802 + 0.011 0.452 + o.oo4 1.6 + 0.7 k~~.........

TABLE XXIX INTERATOMIC DISTANCES AND BOND ANGLES A. Bond lengths B. Nonbonded intramolecular distances 0 N -B3 1.581 + 0.003A N... B1 2.806 + 0.0 BB2 1.744 + 0.005 N... B2 2.861 + 0.005 B s:3 1.820 + 0.006 N... H6 2.18 + 0.4 B2-B3 1.803 + 0.006 N... H7 2.28 + 0.07 B-11- 1.09 + 0.03 B1... H7 2.28 + 0. 4 3H2 1.18 + 0.04 B-H3 1. o23 + 0.03 C. Bond angles 2-3 1.39 + 0. 05 NB3B 1 111.0 + 50 B2-H4 1.12 + 0.05 NB3B2 115.3 + (5 B2-H5 1.11 + 0.04 NB3(B plane): 117.2 + 50 B2-H6 1.75 + 0.03 133-H6 1.12 + 0.03 133-17 1.14 + 0.07 D. Short intermolecular distances Atom of Atom of Molecules Molecule s reference neighbor related by Distance moleculea molecule H7 H2 Center at (0 1/2,1/2) 2.65A H7 H7 Center at (~0,1, 1/2) 2.o 57 H6 H5 2, axis at x=z=l/4 2.43 H4 H4 Center at (0,1,0) 2.72 aAtomic coordinates of the reference molecule are given in W TR 59 - 20 TT7 T T

those resulting from the standard deviations in the coordinates. It is clear that standard deviations estimated in this manner fail to reflect any systematic errors in the bond lengths conceivably introduced by incorrectly assumed scattering factor curves. (lOO) Topologically, the B3H7 group is readily identified with the fragment of tetraborane produced by symmetrical cleavage of the double bridge inferred on chemical grounds (see sections II B and C). A comparison of the detailed geometry with that of tetraborane(l) nevertheless reveals several significant differences. The nonbridged boron-boron distance B1-B3 is considerably longer than the value.712A found for its presumptive counterpart in tetraborane. This may be related to the change of the neighboring Bj-H2 from a bridge to a regular boron-hydrogen bond. On the other hand, the bridged B1-B2 distance is much shorter than any of the four bridge B-B distances in tetraborane, all of which lie in the range 1.842.007A. Perhaps the most striking difference is the strong asymmetry of the2 —.hydrogen bridge. While the B3-H6 distance is indistinguishable from a single B-H bond, the B2-H6 distance of 1.75A is almost O.4A longer than the corresponding B-H bridge distances in B4Hlo or any other boron hydride. Boron-hydrogen distances approximating this value are found in BH1( ) but they are not partof B4H-B bridges of the usual kind. The overall effect of these distortions from the B4Hio geometry may be interpreted as a tendency toward the structure of a br dge substituted diborane, (H3NBH2)B2H5. The nearly planar configuration of H1H2BiB2H4H5 and the close ness of the B-B2 distance to the value 1.770 ~ 0.013O found in diborane(102 would seem to support this interpretation, which presumably would require the free molecule to have a plane of symmetry perpendicular to Bl-B2. Since B2-H6 B.11.7 are clearly different, as are N0...Bl and N...B2, a molecular symmetry plane certainly is not present in the crystal. The H3NBH2 group., however, has a plane of symmetry, within experimental error, as shown by its B-H and N....H distances. The violations of the molecular symmetry p lane can then largely be accounted for in terms of a tilt of the H3R`BH2 group with respect to B2H5, such as to "crowd" H6 into B2. Support for the hypothesis that this might be due to intermolecular repulsions could be sought in the geometry of the close intermolecular contacts involving 116 and 117 (Table XXIX). This approach is inconclusive, however, since the H3NTBH2 group is about equally closely surrounded by neighbors on'both sides. In spite of the above considerations, we feel that the asymmetry of the moLecule., which also includes the -presumable significant asymmetry of the B. —H3-B2bridge,, is probably too large to be accounted for by mere intermolecular (lao) F. L. Hirshfeld., K. Eriks, R. E. Dickerson, E. L. Lippert and W. N. Lipscomb, J. C_(he-m. Phys. 28, 56 (19)58).

packing forces. Using the formulation of Eberhardt, Crawford, and Lipscomb(l03) our results suggest that the boron triangle in NH3B3H7 is held together by two -H-Bbridgebonds and one (BB3) electron pair bond, but the alternative descripBtion in terms of one hydrogen bridge 1-H3-B2 and a central three center bond (B-B23 not be entirely ruled ut. The N-B distance is in satisfactory agreement with other, less accurately determined N-B distances in boron hydride derivatives(104); it shows particularly close agreement with the more accurately determined boron-nitrogen distances in the addition compounds of boron trifluoride with ammonia and methyl amines.(lO5) The packing of molecules in the low-temperature structure viewed down the b axis is shown in the hO Fourier projection of Fig. 26. As in the high-temperature modification, the end-to-end arrangement is apparent, in this case in the direction of the diagonal glide translation. Molecules in neighboring chains lie next to each other rather than staggered as in the high-temperature form; this arrangement amounts to a favorable stacking of dipoles pointing in two opposite directions. A low-temperature heat-capacity study by Westrum and Levitin,(l06) carried out after the completion of the experimental part of the X-ray study, has established the transition temperature as 297.10~K. As pointed out by these authors, the entropy of transition, 15 cal. deg. - mole-1, is approximately equal to Rn8. It is interesting to note that the number of equivalent general positions in the space group I~-mm is 8, that is, if the only effect of the phase transition were a randomization over these positions of the angular orientation of each molecule, independent of its neighbors, the entropy change would very nearly equal the experimental value. E. Other Triborane Addition Compounds 1. TRIMETHYLAMINE-ThIBORANE (CH3)3NB3H7 was first prepared by Edwards and his collaborators at Callery Chemical Company(lO7) using the reaction between B4Hio and N(CH3)3. It Was (105) Wb H. Eberhardt, B. Crawford and W. N. Lipscomb, J. Chem. Phys. 22, 989 (1954-); R. E. Dickerson and W. N. Lipscomb, ibid.,, 27, 212 (1957). (104) E. W. Hugh-es., J. Am. Chem. Soc. 78, 502 (1956); E. L. Lippert and W. N. Lipscomb,, ibid., 78~, 505 (1956); C. E. Nordman and C. R. Peters, ibid., in press.

0 ci ~~~~~~~/ km / / \ P0~~~~~~~~~~~~~~~~ -.., f'C —.. I % N./~'~~~~~~~~~~~~~~~~~~~~~~~~~~~~'- % /- -\.' N~f ~~~~~~~~~~~~~- > - \ N._..., ~~~~~~~~\ "~~ -- NN /~~~ / NN~~\% % N~~~~~~~~~~~~l'N \~~~~~~~~~~ %~~* \-, - ~~~~~~~~~/ \'I I. ~-,~~~~ I. —— " \~~~~/ // " " packing~~~I of moeuls Cotusaeat1(rkn,2,4/ n

later found that a process involving an initial preparation of a B3H7 etherate followed by replacement of the ether by N(CH3)3 was superior for preparative purposes. Kodama had found that tetrahydropyran is a particularly suitable ether for such a purpose (see page of this report). The following process, modeled after that used by Kodama for the preparation of H3NB3H7, was developed in this laboratory by J. C. Carter and G. Kodama under the joint sponsorship of WADC (Kodama) and Callery Chemical Company (Carter). The sample thus prepared was used by Westrum and his collaborators to run the low-temperature heat capacity of (CH3)3NB3H7. The reactions involved are: (1) 2B4H~o + 2CH2(CH2)40 - 2(CH2)(CH2)40B3H7 + B2H6 (2) CH2(CH2)40B3H7 + (CH3)3N (CH3)3NB3H7 + CH2(CH2)40 The reaction vessel, a 500-cc three-necked flask, was fitted with an efficient magnetically driven paddle stirrer and a low-temperature reflux condenser attached at the top to a vacuum line and manometer. The third neck of the flask was fitted with a standard taper stopper and used to introduce solvents. One hundred milliliters of tetrahydropyran and one hundred milliliters of toluene were placed in the flask, cooled to -196~C, and the reactor evacuated. The toluene was used to dilute the tetrahydropyran. Fifty milliliters of tetraborane (liquid measured at 0~C) were then distilled into the reactor and the temperature raised slowly, when the first evidence of a reaction occurred, as noted by evolution of diborane., the reactor was immersed in liquid nitrogen. After the reaction had subsided and the pressure had dropped to zero., the reactor was again warmed slowly and the diborane removed. The initial reaction was quite exothermic and very difficult to control, the diborane pressure building up extremely rapidly. After cooling and rewarming., the reaction was generally quite moderate and the diborane could be removed without difficulty. When a 100-cc sample of liquid tetraborane was used, heat-transfer problems and localized heating in the solution resulted in the reaction going out of control for a short period of time. Unless more efficient methods of heat transfer and temperature control are available, the reaction cannot be recommended for greater than 50 cc of tetraborane and even then only if all precautions of safety are observed. After the major portion of diborane had been removed, the reaction was cooled to -780C and allowed to stand overnight. The remainder of the diborane was them removed and the excess tetrahydr~bpyran and toluene -were condensed into a tube for removal from the system. The tetrahyd-ropyran-tri'borane remaining in the reactor melts just below room temperature; it ignites spontaneously when ex

until the tetrahydropyran triborane had completely dissolved. The solution was cooled to -780Cand 7 mlof liquid trimethylamine (measured at 0~C) was allowed to distill very slowly into the system. A period of 24 hours was required for the amine to be completely absorbed. It was necessary to use a stoichiometric amount of trimethylamine because of the possibility of side reactions involving the desired product and excess amine.. At this time the reactor was allowed to warm to room temperature and the ethyl ether and liberated tetrahydropyran were distilled from the system and discarded. The method of recovery and purification of the trimethylamine triborane differed from that used with the ammonia compound. The trimethylamine-tri borane could not be precipitated from a benzene solution upon addition of methylcyclohexane. It was recovered and purified by recrystallizing twice from anhydrous toluene at -95C and filtering through a glass frit cooled to -80C under a stream of dry nitrogen. The yield of pure product was 65-70. The sample was hydrolyzed by heating 100 hours at 1500C with 10% HC1 in a sealed tube. Active hydrogen evolved was measured with a Toepler pump and calibrated gas burette. Boron was titrated as boric acid in the presence of mannitol. Trimethylamine was determined by the Kjeldahl procedure. It was necessary to bubble the liberated amine very slowly through a considerable depth of standard acid to obtain efficient absorption. Results of analysis were: Found Calculated * of mmoles mmoles Calculated Active Hydrogen 7.28 %_ 7.5]. 99.6 Boron 0.546 0.5)48 99.6 Trimethylamine 0.181 0.1827 99.0 Ratio: (CH3)3N B H 0.99)4 3.000 7.000 Reagents were the same as those described for H3]NB3HT or were the best commercial product. 2, ATTEMPTS TO PREPARE F3)PB3H7 The isolation in this laboratory of the compound F3PBH3, formally analogous to the compounds (NHm3)3NBH3 and H3N\BH3, and the isolation of the compounds H3IANB3:,H7 and (CH3)3]NB3H7 all suggested that it should be possible to prepare the

to stabilize the B3H7 fragment long enough to permit compound identification. Systematic chemical arguments given earlier suggest the following equation to represent the interaction of B4Hzo and PF3 B4Hlo + PF3(excess) ) F3PBH3 + F3PB3H7 The system was studied by sealing B4Hjo and PF3 in glass tubes equipped with break-off tips. Procedures were analogous to those used for the synthesis of H3BPF3.(108) When the PF3 pressure was 2 atmospheres or less and the reaction time at room temperature was half a day or less, the original PF3 and B4Hjo could be recovered unchanged. If the pressure was above 5 atmospheres and the reaction time at room temperature was 3 days, B2H6, F3PBH3, and a yellow viscous residue along with very small amounts of fairly volatile, unidentified components were isolated from the reaction mass. Some of the unidentified components had a volatility about the same as or a little less than that of B4Hio; this material decomposed rapidly to PF3, B2H6, and other residues when warmed to room temperature. The very small amount of substance available permitted only crude qualitative identification of products. The apparent low stability of any F3PB3H7 compound at room temperature suggested a low-temperature synthesis in liquid PF3 which can be represented by the following equation: NaB3H7 +HC1 + P3 liquid NaCl + F3-PB3H7 + H2 +H~l PIP3 PF3 solvent at -78o a Because of the very low solubility of NaB3H8, the evolution of H2 was very slow; several days were required to give off one mole of H2. With a large excess of HCl the reaction proceeded somewhat faster. The resulting reaction mixture was complex and difficult to separate. The material left after the removal of excess PIP3 and HCl decomposed easily when warmed to room temperature. PIP3, B2H6,~ B4H~o and a solid yellow residue were recognized as products. Current data suggest that if FP3PB3H7 is to be identified., low-temperature techniques will be required for the complete characterization. Further work is in progress.

IV. SOME ADDITION COMPOUNDS RELATED TO THE BORArE ADDUCTS A. Background Data outlined in earlier sections on the reactions of B2H6 and B4H10 have been interpreted in terms of a symmetrical and nonsynmetrical cleavage of the double bridge bonds of these two molecules. H\ /A /H H / H H Ho, H H HH H /~~~~~~~~~~~~~~~~~ I H H H\ B \/H B\ / i I B i GAB\ H/'HL- B3 H H-B- H H H Symmetrical Cleavage Nonsymmetrical Cleavage of the Double Bridge Bonds of the Double Bridge Bonds It has been found that certain reagents such as ammonia promote nonsymmetrical cleavage while others such as trimethylamine and various ethers promote symmetrical cleavage. Since the double bridge bond is found in many structures other than the boron hydrides, it is of interest to explore the cleavage reactions as they may apply to compounds such as A12C16, A12(CH3)5, and even [AlH31n* Both symmetrical and nonsymmetrical cleavages were reported by Gibson and later by Chatt for certain coordination compounds such as CH3ClPtCl2PtClCH3 and Et2AuBr2AmuEt2 (109

CH3N > ClN /Cl [CH3 1 [Cl CH3, Pt K,PtX + dipyridyl — Pt,) Pt t Cl /Cl CH3 [C1 DpyridylJ LC1 Ci I / Nonsymmetrical Cleavage of Pt Complex CH3N /Cli /C1 CH3s / Cl / Pt N /Pt + N(C3H7)3 - 2 Pt Cl Cl CH3 Tri-n-propyl c N(3H7)3 amine Symmetrical Cleavage of Pt Complex The current sequence of experiments was undertaken in an effort to gain information on the reactions of the bridge compounds of aluminum. Intelligent correlations between the reactions of such coordination compounds and the electronic configuration of the central metal atom demand more extensive data on the nature of the reaction between bridge molecules and various reagents. Many of the explanations and hypotheses based on t bonding have little basis in experimental fact and the current studies were initiated to establish a more reliable experimental basis for current speculation. B. The Compound F3PA1C13 1. EARLIER STIIUDIES The ability of PF3 and PC13 to form adducts with PtCl2 was first described by SchultzenbergerWlO) and by MoissorPlD) but it remained for Chat44li to recognize that the resulting compounds were coordination compounds similar to Pt(CO)2Cl2 and [Pt(CO)C12]2 described extensively in earlier literature on complex compounds. Chatt pointed out that F3BPF3 does not form and he contrasted this with the fact that [PtCl2(PF3)]2 can be formed with ease. He attempted to rationalize the experimental difference in terms of the electronic structure of the boron and platinum atoms. He noted that platinum with the electronic configuration ls2, 2s2, 2p6, 3s2 3p6 3dl~, 4s2, 4p6 4d'~, 4f14, 5s2, 5p6, 5d8 6s2 has 5d electrons which might be donated to the d orbitals of the phosphorus atom and pointed out that the attachment of F to P would lower the d levels of P to a bonding region; thus a "A or dative double bond" (110) P. Schultzenberger, Bull. Soc. Chim. 17, 482 (1872). (111) H. Moisson, Bull. Soc. Chim. ~, 454 (1891)o (112) J. Chatt, Nature 16, 637 (1950). WADC TR 59-207 151

was assumed to be important in the linking of PF3 to PtCl2. In such a bond, electrons from the metal are donated to the phosphorus atom. Aluminum, with the electronic configuration s2, 2s2, 2p6, 3s2, 3pl, has no such d electrons and would be unable to form a complex with PF3 by such a mechanism. In support of such ani argument, Chatt and Williamsll3) noted that no addition compound resulted when PF3 was passed over sublimed A12C16 or Al2Br6 at 2500C. More recently Holmes and BrownU 14)reported that AlC13 would not combine with Pl~3 in a low temperature study. Earlier work in this laboratoryl5) indicated that the bridged compound B2H6 reacts with PF3 by symetrical bridge cleavage to give H3BPF3, a molecule which in the earlier arguments of Chatt, cannot be stabilized by ET bonding of the type used for platinum complexes. It has since been suggested by Graham and Stonell) and others that such a complex owes its stability to a delocalization of the B-H electrons to give a "it bond" in a manner comparable to that invoked in the hyperconjugation arguments of organic chemistry. The prime argument for such it bonding was the difference in stability between complexes of borine and BF3 with bases such as PF3 and NH3. In this laboratory it was suggested that those Lewis acids which are dimeric in the free state would combine with very weak bases such as PF3 whereas thosewhich are monomeric would not. Thus B2H6, A12C16, and A12(CH3)6 should combine with PF3 while BF3 B(CH3)3, BBr3, etc., should not. In an effort to test this prediction, the system A12C16-PF3 has been restudied using the methods employed in the earlier 2H6-PF3 study. The empirical generalization suggested above can be rationalized by an appropriate combination of steric and electronic arguments which will be presented with a more detailed account of the experimental data in a subsequent publication. The existence of F3PAlC13 and the nonexistence of F3PBC13 are rationalized with more difficulty by " it bonders." 2. THE PREPARATION OF F3PAlC13 Very pure resublimed A12C13 (99-9+ * A12C16) reacts with PF3 in a bomb tube under 8 atmospheres pressure in accordance with the following equation: A12C16 + 2PF3 ->2Cl3AlPF3 About 4 hours at room temperature were required for reaction. Shorter periods of time or lower pressures gave no reaction. Longer periods of time permitted group exchange on the complex

nCl3AlPF3 - nPCl3 + [AlF3]n PC3 and AlF3 were recovered as products. The identity of the complex Cl3AlPF3 was established by determining the amount of PF3 combining with a given amount of A12C16. The molecular weight of the solid product was determined by vapor pressure depression in liquid PF3. An experimental value of 202 as compared to a theoretical value of 221.7 for AC3PF3 supports the proposed monomeric formula. The deviation of almost 10 between the experimental and theoretical values is understandable in terms of experimental difficulties of the method and the slow conversion of C13AlPF3 to A1F3 and PC3 in liquid PF3 even at -112~C. It is also appropriate to note that F3PAlC13 is soluble in liquid PF3 whereas A12C16 and PC13 are not. Further evidence to support the proposed formula is found in a base displacement reaction. When diethyl ether was added to the solid complex at -112~C, the solid dissolved but no P3 was liberated; addition of an equimolar quantity of NMe3 to the cold solution (-78~C) resulted in liberation of an amount of PF3 equal to the amount of Me3 used. Me3NAlCl3 was obtained as a product of the reaction. The process may be summarized by the equation: F3PAlCl3 + Me3N > Me3NAlCl3 + PF3. The shift of halide groups in the complex is an interesting process. It goes slowly in liquid PF3 at temperatures as low as -1120C and is rather rapid for the pure solid at temperatures above 0~C. Earlier studies in this laboratory indicated that CH30NHCH3BH3 undergoes a shift of groups to give B(OCH3)3 and NH2R as initial products but it was noticed that such a process was always preceded by a loss of H2. Bissot., Campbell., and Parry&17) interpreted this fact as evidence for the loss of 112 from N and B., thus opening up a coordination site for group transfer. Since-both Al and P can have coordination numbers above 4i rather commonly., it would be expected that the shift of groups would occur even at low temperatures and without initial loss of any coordinated groups. Indeed., such was observed. 15. EXPERIMENTAL (a) Compound Preparation IMetallic aluminum was burned in a dry chlorine stream in an all-glass syst'Uem using conventional good. techniques (see Fig. 27 ). The resulting A12C16 was sublimed into a small bulb which was then sealed off. Approximately 1 mM of A12C16 was trapped. in each bulb by means of the glass wool f ilter A in Fig. 27. (17M.ClsoD.Capel n R Pry J. Am Chem Soc. 80,187

\J1 0 ~~~~~~ARSTREAM FOR COO LI NG SEAL H20 GLASS WOOL ~ 4/35 Al PORCLAN SIIAGLaPLUGS Fig 2. Apaatu fr te repraionof EAL16

thus the weight Of A12C16 could be accurately determined. The bulb was placed in a heavy-walled glass bomb tube along with a magnet sealed in a glass case (a magnetic hammer). The tube was evacuated and opened to the diffusion pump for several hours prior to use. The tube was also warmed with a flame to aid in drying. The bulb containing the A12C16 was then crushed with the magnetic hammer and the dry nitrogen it contained was pumped out (i.e., the magnet inside was pidked up by means of an external magnet and dropped on the glass bulb). A measured quantity of PF3 was condensed into the tube (enough to give a desired pressure of about 8 atmospheres); then the tube was sealed off and transferred to the room-temperature storage area. After standing for the desired length of time, the contents of the tube were frozen with liquid nitrogen, the tube was opened to the vacuum line, and excess PF3 was removed at -112~C. The amount of PF3 recovered was measured and the stoichiometry of the reaction between PF3 and A2C16 was determined. Typical data for a number of runs are summarized in Table XXX. The course of the run after product formation was determined by product treatment. A number of separate observations are summarized below. (b) Decomposition of F3PCl3 upon Standing If in the preparation of F3PAlCl3 an unusually long time was allowed for the interaction of F3P and A12C16, no evidence for complex formation could be detected, but PC13 and AlF3 could be recovered in stoichiometric quantities. (See runs 1,2.,3.) The PC13 was identified by its molecular weight which was measured by vapor density. An observed value of 152 compared favorably with the theoretical value of 157. An analysis of the PC13 for chloride gave 80.0* C1- as compared to a theoretical value of 80.5* C1- for PC13. The AlF3 in run 3 was identified after removal Of P013., by analyzing for Al by precipitation as 8-hydroxyquinoline. The fluoride was determined on the same sample by precipitation as Pb~lF. The observed ratio of F_ to Al+3 was 2.8 to 1. In a number of other cases the complex formed and no evidence for PC13 was obtainable, but as the solid was allowed to warm up over a relatively long period., PC13 appeared (run 6). (c) The Reaction of the Complex with Trimethylamine,, with Dimethyl Ether, and with Diethyl Ether Trimethylamine (about 0.5 mmole'in run 2) was frozen onto the solid complex F3:PAl.-1- lC3 (about 5.20 mmoles) and the system was warmed slowly to -78oC., at which temperature it was allowed to stand for about 1/2 hour. No PF3 was liberated from this system. As the temperature was allowed to rise, P013 could

TABLE XXX u oQ~~~~~~ ~~DATA FOR THE REACTION OF A12C16 AND PF3 \31 kO ~~~~~~~~~~~~~PF3, Run A12C16, PF3 Pressure ) un ACl, Initial Recovered Used AC Atmosphere 1 0.735 20.46 19.01 1.44 1.96/1 8.1 2 wks PF3 was removed at -112~, fractionation at higher temperatures produced PC13 (mol. wt. = 132; theo. = 137). Analysis for C1- gave 80% vs. 80.5% calc. (see Item 3(b), p. 155). 2 1.60 19.55 16.98 2.67 1.67/1 7.7 4 days Addition of 0.5 mM (CH3)3N gave no PF3; some PC13 was recovered. 3 0.855 20.61 18.91 1.70 1.99/1 8.1 4 days Recovered 1.68 mM PC13 and analyzed for AlF3 (see Item 3(b) p. 155). 4 1.16 20.05 17.57 2.42 2.08/1 8.0 4.5 hr No PC13 was present originally but some was recovered upon warming to room temperature. 5 0.60 3.30 3.30 -- N.R. 400 mm 7 hr No reaction with liquid PF3 in slight excess at -112~. Ohen~ ~ 6 1.03 20.15 18.06 2.09 2.01/1 8.0 3 hr Addition of (CH3)N produced no PF3 but PC3 was obtained when tube was warmed to room temperature. 7 1.17 20.33 20.13 0.20 N.R. 8.1 1 hr Reaction does not go in 1 hr. 10 1.45 20.52 17.40 3.12 2.12/1 8.2 4 hr Addition of about 2 cc ethyl ether gave no PF3 at -78~; subsequent addition of about 0.8 mM of trimethylamine gave off about 1 mM material at -780 which had a mol. wt. = 83 (PF3 mol. wt. = 88). 11 0.91 20.99 19.05 1.94 2.14/1 8.3 4 hr Addition of 2 cc ethyl ether and 0.603 mM trimethylamine gave o.606 mM PF3 (mol. wt. = 87) at -78~. 14 0.855 25.02 22.98 2.04 2.3/1 8.1 4 hr Prepared for mol. wt. determination but methyl ether solvent displaced PF3 from the addition compound. 16 1.02 19.49 17.44 2.05 2.00/1 8 4 hr Molecular weight determination (see Item 3(d) p. 157). 17 0.86 19.71 17.99 1.72 2.00/1 8 4 hr Molecular weight determination (see Item 3(d) p. 157).

Addition of trimethylamine to the ether solution of the complex resulted in displacement of the PF3 at -78~C. The most convincing data were obtained in run 11. About 2 ml of diethyl ether was frozen onto the solid complex (about 1.9 millimoles) with liquid nitrogen; then a sample of trimethylamine amounting to 0.603 millimole was frozen on the ether. The system was allowed to warm u to -78~C and 0.606 millimole of PF3 was liberated. The PF3 was identified by its molecular weight. The value obtained by vapor density was 87. The theoretical value for PF3 is 88. (d) The Determination of the Molecular Weight of F3PAlC13 Molecular weight studies on this solid posed very difficult experimental problems. Conventional solvents such as benzene froze at such a high temperature that halide interchange occurred. Diethyl ether dissolved the solid complex, but because the solvent has two solid phases it is not particularly suitable for freezing point studies and its vapor pressure at suitable temperatures was too low for vapor-pressure depression studies. Dimethyl ether was not suit able as a solvent because it displaced PF3. Finally, liquid PF3 was used as a solvent at -112~C. A carbon disulfide slush was used as a thermostat at -112C. The sample bulbs were immersed directly in the slush; otherwise the procedure was the same as that previously describedfl5g) In a typical run the sample weighing some.454 gm was prepared directly in the molecular weight tube. The empty tube was weighed first and the tube was weighed again after the completion of the molecular weight measurement since exposure of the sample to room temperatures during weighing would result in halogen shift. About 23.0 millimoles of PP3 was used as a solvent. Typical data at about -112 0C showed: PMm AP APmm MOl1. Wt. 446 43. 0.0975 202 540 52.8 0.0977 201 Theory for F3PAlC13 =221.4 When more PF3 was added to get another point., a precipitate formed and AP dropped to zero. It is signif icant that the decomposition of FSPAlCl3 in liquid PP3 at -1120C was confirmed in subsequent studies. It is also noteworthy that AlP3,, PCl3., and A12C16 are all insoluble in liquid PP3. A duplicate run on a second sample (0.581 g) also gave a value of 202 for the molecular weight. The entire area of aluminum complexes is still under active investigation.

C. The Compound H3Al[N(CH3)3]2 One of the procedures used to dry trimethylamine in this laboratory involved the storage of the liquid above lithium aluminum hydride. After the trimethylamine was removed from above the lithium aluminum hydride, a white volatile solid with a sharp melting point of 940C could be sublimed from the reaction residues. The solid decomposed slowly at room temperature, evolving trimethylamine and other residues. When heated in glass, the solid left an aluminum mirror. Analysis of the solid showed the following: Al = 18.1, N = 18.9, H = 2.07%. These values support the formula H3Al[N(CH3)32 for which analytical values are: Al = 18.1*, N = 18.8*: H- = 2.03*. The formula immediately suggested that yields could be improved if AlH3 were made first and treated with trimethylamine. Two new preparative processes were then tried. They are described by the equations listed below: Process 1 ether AlC3 + 3LiAlH4 - ) 4H3A1-etherate + 3LiCl H3Al-etherate + 2N(CH3)3 -) H3A1-2N(CH3)3 + ether Process 2 ether LiAlH4 + HCl - H3Al-etherate + H2 + LiCl H3Al-etherate + 2N(CH3)3 - H3Al-2N(CH3)3 + ether The second process gave recoverable yields of more than 65*o based on the LiAlH4 used. It also offered an explanation for the original preparative process in which HOH replaced the HCl. Yields were poor when the water content of the amine wras low. The molecular weight of the solid was measured by vapor-pressure depression in dimethyl ether solution over a temperature range from -358 to -280C. The result of 14]3 ~ 8 clearly indicates a monomer. Shortly after complete characterization of the solid it was found that Wiberg and his collaborators~ll8) had prepared and characterized. the sane compound. His melting point of 950C and monomeric nature of the complex in ether were thus verified independently. The coordination number of 5 for Al+3 in this complex suggested interesting

V. PHYSICAL AND CHEMICAL STUDIES ON PURE BORON HYDRIDES AND BOROHYDRIDES A. The Raman Spectra of Four Isotopic Varieties of Diborane in the Gas Phase(l) The compounds llB2H6, llB2D6 l~B2H6 and B2D6 were prepared during the course of a spectroscopic investigation of the borohydride ion and aluminum borohydride. Since the Raman spectra of only the first and last of the above isotopic varieties have been reported in detail, and those only for the liquid state, it was deemed of interest to obtain the gaseous Raman spectra for the complete series. 1. EXPERIMENTAL The starting material in the preparation of the compounds was the diethyl ether complex of BF3. This was converted to the appropriate diborane by means of lithium aluminum hydride or deuteride essentially according to standard methods. The boron-ll compounds contained approximately 81% of the B isotope, the normal isotopic abundance, while the boron-lO compounds* contained 96% of the desired isotope. The hydrogen content of the deuterated compounds was estimated at less than 2% from the spectral evidence. Purification was accomplished by repeated fractionation on the vacuum line until a constant and reproducible vapor pressure was reached. The values observed for ordinary diborane agreed with those in the literature. The use of BF3 etherate as a starting material eliminates possible contamination with SiF4, which is difficult to remove by fractionation. The spectra were recorded photographically, using the spectrograph and light source described previously.(l2O) The Reman tubes had an outside diemeter of 25mm with an illuminated length of 150 mm and contained the gases at approximately )4 atmospheres pressure. Collimating baffles largely eliminated light scattered from the walls and window. Saturated sodium nitrite solution served as a filter for all exposures; an isopropyl alcohol solution of rhodemine 5 GDN Extra and paranitrotoluene was utilized as an additional filter for a few of the spectra. Exposure times varied from 12 to 72 hr with Eastman 103a-J plates. The resolution of the spectrograph was quite adequate to resolve clearly the isotopic triplet of ordinary diborane at about 800 cm. Wavelength measurements were

made both on the plates and on enlarged tracings made by a Leeds and Northrup microphotometer. Estimated probable errors accompany the averaged values listed for the frequencies. 2. RESULTS Observed frequencies, estimated intensities, and assignments for the hydrogen and deuterium compounds, respectively, are shown in Tables XXXI and XXXII. Tracings of typical spectra of the deuterated compounds are shown in Fig. 28. No difficulty was met in identifying the Ag fundamentals, since the characteristic sharp appearance of totally symmetric bands in the Raman spectrum of gaseous compounds is at least as conalusive as depolarization measurements. The other Raman-active fundamentals were typically broad and diffuse and lower in intensity so that frequently they could not be distinguished from the background. In some cases, they were obscured by Ag fundamentals or overtones. This appears to be the reason for the great breadth of the band near 700 cm-1 in the deuterated compounds. Assignment of the overtones and combination bands was made with the help of the infrared active fundamentals of the gas listed by Lord and Nielsen;(121) where Raman-active fundamentals could not be identified from the present work, liquid values from the same paper were used. Frequently alternate possibilities existed for the assignment of weak bands. In such cases, the assignment with Ag symmetry if one existed, was preferred on the grounds that bands of this class would likely be more intense in the Raman effect. Several cases in which the intensity of the overtone or combination was enhanced by Fermi resonance were noted~, the most pronounced case occurring in the deuterated compounds and involving 2v3 and vj. The doublet appearing at the position of 2V3 in the spectrum of 11B2D6 is attributed to the presence of roughly 30% of 1JiB-10B molecules. The unexpectedly high intensity, of the isotopic satellite at 1824 cm-i above that expected from the ratio of isotopic molecules arises from a closer resonance with the v1 fundamental in the lighter molecules~. The fundamentals derived from the present work are collected in Table XXXIII. Since they agree quite well with those given by Lord and Nielsen(121),, who have considered all previous work, they will not be discussed in detail. It may be pointed out that the general agreement found here supports the essential correctness of the assignments made by these authors. The application of the Teller-Redlich product rule and also one of the isotopic sum rules to the frequencies of the Ag class is shown in Table XXXIV. The data are satisfactory considering the lack of knowledge of the anharmonicities and the number of fundamentals involved in Fermi resonances.

TABLE XXXI OBSERVED IRAMAN FIREQUENCIES OF GASEOUS 11-B2H6 AND 1c>B2H6 11B2H6 10B2H6 -Frequency Frequency Intensity Shape Symmnetry Assi-gnmen -(Cm') I 6 (cm-') +6 -- - ~~~685 5 VWA sh Ag V9-Vi -- -786 2 VWbr V4(2 2B11B) 788 1 -- - s sh AgV41B 802 1 -- - Mn sh Ag V 0 818 2 816 1 s sh Ag -- - 852 5 vvw sh V7?,Yv15? 1026 6 1025 5 w br V3 F-J ~~~1184 1 1186 1 Mn sh Ag v+1 H ~~~~~1510 5 1518 2 w-vw sh Ag V 1755 5 1768 8 w br Bg 2v,5 2011 2 - - - vw sh Ag V2 2109 1 2110 1 s sh Ag 2v3,2 v18 -- - 2552 2 w sh Ag V4+2v5? 2485 7 -- - VVXw Ag V j 2552 1 2557 1 Vs sh Ag vil (2600) 10 2640 10 w brB2 6 = probable error; s= strong, m= medium, w = weak, sh = sharp, br= broad; frequencies enclosed in parentheses are estimated; assignments followed by a question mark are based on fundamental frequencies obtained indirectly.(l21)

TABLE XXXII Q 1-3>~~~~ ~OBSERVED RAMAN FREQUENCIES OF GASEOUS 11B2D6 AND'~B2D6 \J1 R) 0 < 11B2>~ ~~~~~~~~~~~~ lo2 Frequency Frequency Intensity Shape Symmetry Assignment F re que ncy F reque ncy (cM-1) 1+ +5 712 1 721 1 s sh Ag V4 -- - 748 5 w br v7v2v5? 912 1 928 1 m sh Ag V 971 3 962 3 vw sh Ag V 1287 4 1285 5 w br B V6 1414 3 1446 3 vw sh Ag 2v4 1458 3 -- - vw sh A 2v4 Ag 2 V14 aH~ ~1459 3 -- - vw sh A 2V7,2v12? 1488 4 -- - vw sh A 2v15? 1515 1 1515 1 vs sh Ag v2 1748 2 1758 5 vw sh Ag 2vl1 1809 1 1836 1 s sh Ag 2v3 1824 1 -- - m sh Ag 2v3(B-JB) 1867 1 1883 1 vs sh Ag v 1912 2 1924 2 w sh Ag v4+2V4 1968 2 1990 2 m br B2g vl = estimated probable error; s = strong, m = medium, w = weak, sh = sharp, br = broad; assignments accompanied by a question mark are based on fundamental frequencies obtained indirectlyi.J2l

TABLE XXXIII RAMAN ACTIVE FUNDAMENTALS OF GASEOUS DIBORANE l1B2H6 1OB2H6 11B2D6 1OB2D6 Ag ~1 2532 2537 (1860) (1870) v2 2109 2110 (1512) (1514) V3 1184 1186 (912) (928) v4 788 816 712 721 Big V6 1755 1768 1287 1285 V7 -- B2g Vll 2600 2640 1968 1990 Vl2 B3g V15 -- Frequencies enclosed in parentheses are involved in Fermi resonances and have been corrected by an estimated shift. TABLE XXXIV PRODUCT AND SUM RULES FOR THE Ag CLASS FREQUENCIES OF GASEOUS DIBORANES Product Ratio Calculated Theoretical ( lB2H6/llB2D6) 2.73 2.825 ( ioB2H6/1oB2D6) 2.74 2.825 ( 1~B2H6/"lB2H6) 1.04 1.049 (B0B2D6/1"B2D6) 1.o04 1. 049 Sum difference Z [vi2(l~B2H6)-v22(llB2He)1] = 7.9 x 104 Z1 [v 2(.O2D )v22( 11B2D)1 = 8.6 X 104

B2D6 B2 D6 I I I I l 0 500 1000 1500 2000 CM Fig. 28. Raman spectra of 10B and l1B deuteroboranes in the gas phase. B. Raman Spectroscopy in Liquid Ammonia Solutions. Vibrational Frequencies and Force Constants for Isotopic Species of the Borohydride Ion Having Tetrahedral Symmetry 1. INTRODUCTION In a previous article(l22) the Reman spectra and assignments of the 1"BH-4 and 1JBD- ions dissolved in liquid ammonia were reported. However, solvent interference prevented the observation of one fundamental of the borohydride ion and confused the interpretation of the B-P stretching region-of the borodeuteride spectrum. Since this ion is one of the simplest boron hydride structures., it is desirable to have a complete and unequivocal assignment of fundamentals, and consequently the work has been extended making further use of isotopic substitution for both hydrogen and boron. In particular, employment of the isotopic solvent,, ND3, has successfully eliminated interference by solvent bands and has made possible the observation of all the fundamentals of the various isotopically substituted ions. The use of ammonia as a solvent was dictated by

hydrogen between the borohydride ion and ammonia occurs extremely slowly, if at all, the two isotopic solvents could be used interchangeably depending on the spectral region of interest. The final assignment which has been made agrees with all available data and confirms the tetrahedral structure of the ion. The values of the fundamentals have been used to evaluate the force constants of a complete valence force potential function. 2. EXPERIMENTAL Sodium and lithium borohydrides containing boron-ll in its natural abundance of 81% were obtained from Metal Hydrides, Inc., and purified before use by removing all ammonia insoluble material. Potassium borohydride was prepared by metathesis between NaBH4 and KOH according to a recently described procedure.(123) Borodeuterides and all compounds containing boron-10 were prepared by one of two procedures depending on the alkali metal. Lithium compounds were obtained by reacting an ether slurry of lithium hydride or deuteride with the appropriate diborane in a sealed reaction tube. Sodium and potassium salts were prepared by the reaction of the appropriate diborane with sodium or potassium tetramethoxyborate. The reactions involved have been reported by Schlesinger et al.(124) All compounds were purified to remove ammonia insoluble material before use. Boron-10 was obtained in 96% enrichment as the calcium fluoride-boron trifluoride complex from Union Carbide and Carbon Corporation, Oak Ridge, Tennessee. The BF3 was converted to diborane by reaction of its etherate with lithium aluminum hydride or deuteride. Although a small amount of hydrogen was introduced in the synthesis of the deuterated compounds, its amount was estimated at not more than 2-3% of the total deuterium present and no difficulty was met in the interpretation of the spectra because of the presence of a small percentage of isotopically mixed ions. Deuterated ammonia was prepared in a small steel cylinder by the reaction of D20 with freshly prepared Mg3N2. Its spectrum showed insignificant amounts of hydrogen to be present. The preparation of the samples has been described previouslyJ (122) In most cases they consisted of 1-2 ml of solution whose concentration lay in the range'between 2-4. molar. No significant changes of the spectra with concentration were noted. Reference to the spectroscopic equipment has been given in the earlier paper. (122) 5. EXPERIMVENITAL RESTULTS AND DISCU~SSION The observed frequencies, assignments., and polarization characteristics of the bands of the alkali metal borohydrides in ammonia solution are listed in (124)- SchlesI Inger eta. JIIAm. ICem Soc. 75, 18-22 (1955).

TABLE XXXV VIBRATIONAL FREQUENCIES AND ASSIGNMENTS FOR THE BOROHYDRIDE ION IN LIQUID AMMONIA SOLUTION (in cm-1) LillBH4 NallBH4 KllBH4 LilOBH4 NalOBH4 Assignments 1082+2a 1080+2 1084+2 1085+2 1093+2 V4 1202+3 1210+3 12133 1206+5 1208+3 v2 2146+5 2146+3 2154+3 2150+3 2161+3 2v4 2244+5 2250+15 v3 2265+2 2264+2 2266+3 2268+2 2270+2 vl 2402+ 3 2398+3 2404+3 2409+5 2405+5 2v2 aEstimated standard deviations. TABLE XXXVI VIBRATIONAL FREQUENCIES AND ASSIGNMENTS FOR THE BORODEUTERIDE ION IN LIQUID AMMONIA AND DEUTEROAMMONIA SOLUTIONS (in cm-l) LillBD4 NallBD4 KllBD4 LilOBD4 NalOBD4 Assignments 822+2a 823+2 827+2 833+3 834+2 V4 854+2 855+2 864+3 858+4 v2 1569+3 1570+3 1572+3 1570+2 1571+2 V1 1636+4 1636+4 1636+4 1645+ 161650+5 2v4(E or F2) 1668+3 1668+3 1670+5 1683+3 1680+2 2v4(A1) 1698+3 1696 1696+5 1694+3 1704+3 1706+4 v3 2240+3 2238+4 2238+5 2243+3 2245+2 BD3H- ion aEstimated standard deviations.

plates, no distinction having been made between NH3 and ND3 solutions.* A small but real shift in band position with cation was apparent and data accordingly have been given for each salt. Data obtained from the sodium salt have been used for purposes of the discussion and in the force constant calculations, since they form the most complete set and are roughly intermediate between the values for the other two alkali metals. Some of the previously published frequencies of the boron- compounds(122) have been revised on the basis of a larger number of better quality spectra; these changes are largely within the experimental uncertainty of the previous values. Satisfactory assignments have been made assuming a tetrahedral configuration for the ion belonging to the Td point group. This model predicts one Al, one R, and two F2 fundamentals, all Reman active. Isotopic substitution of deuterium for hydrogen should result in a marked shift in all frequencies while isotopic substitution of' the central boron atom should not affect the Al and E frequencies but shift the F2 fundamentals by a small amount. The Al stretching frequency is easily identified as the 2264 cm-' band in the hydrogen compounds and the 1570 cm-1 band in the deuterium from intensity and polarization characteristics. A shift in the apparent position of this fundamental in the borohydride ion on substitution of'0B for "1B probably is due to the fact that v, is superimposed on v3 making the observed band maximum sensitiveto shifts in the latter. In the borodeuteride ion, a Fermi resonance occurs between vt anbd the overtone of the band at 823 cm-1 which shifts the fundamental some 50 cm-' from its unperturbed value. The assignment of V2., the doubly degenerate bending vibration, to the band at 1210 cm-' in the hydrogen ion and 855 cm-1 in the deuterium is confi~rmed in the present work by the absence of a 10B isotope shift. The other bending fundamental., the triply degenerate V4,1 was not observed previously because of interference by a solvent band; its position was inferred to be near 1080 cm-1 from the assignment of a band near 2150 cm-1 as its first overtone. When ND3 is used in place of NHl3, no solvent bands occur from about 900 to 1175 cm-'. The spectrum of BH- in this solvent showed the expected fundamental as a band of moderately low intensity with a maximum at 1080 cm-1. The corresponding band of BD_ apaeat 823 cm-1. Confirmation of the assignment was obtained from the 1OB isotope shifts of 15 and 11 cm-li, respectively. *Evidence suggesting small systematic differences in band -position between the two -solvents was obtained in the case of one or two of the fundamentals. The specific nature of solite-solvent interaction which this suggests is supported by the observation that the maximum of the symmetrical stretching frequency of BHZ differs by about 25 cm-'1 between Nil3 and H20 as solvents, and by the fact that the alkali metal borohydrides form stable arnmoniates. For most of the bands,,

The remaining triply degenerate fundamental, v3, was the most difficult of the four to identify because of near superposition by v or overtones in both the hydrogen and deuterium cases. In Fig. 29, v3,appears as a shoulder on the low-frequency side of the 2264 cm-1 band of the BHZ ion. Confirmation of its presence and a better measurement of its position was obtained by polarizing the incident light so that the intensity of vj was greatly reduced. In theory, the Al mode of a tetrahedral XY4 molecule is completely polarized in the Raman effect but because of convergence error in the incident light, this is difficult to demonstrate. In Fig. 30 a spectrum taken with the electric vector of the incident light polarized parallel to the observation direction shows v3 as a weak ibut distinct maximum. Solvent interference necessitated the use of ND3 for the BD- ion, and again a satisfactory assignment of the four bands observed in the B-D stretching region could not be attained without the help of polarized spectra. The two polarized bands shown in Fig. 31 at 1570 and 1668 cm-1 are asigned to the v1 and the A1 component of 2v4, respectively. They both have been shifted from their expected positions by Fermi resonance. Of the other two bands at 1636 and 1696 cm-1, the latter is assigned to v3 since it gives better agreement with the product rule and is more intense. It shifts 10 cm-1 on 10Bsubstitution. The apparent position and breadth of this band may have been influenced by some contribution from 2v2 which falls at approximately this position but does not have the correct symmetry to resonate. The remaining band at 16536 cm-1 is assigned to the E and/or F2 components of 2v4, the F2 component perhaps being enhanced in intensity by resonance with V3o It is also possible that this band may involve a B-D stretching motion of the BD3H- species which was present in small amounts. Comparison of the observed frequency product ratios of the four isotopic ions with those predicted by the product rule is shown in Table XXXVII. The agreement in general is satisfactory with the exception of the A1 class where some of the fundamentals are shifted by resonance. The normal effect of anharmonicity is to cause the observed product ratios as written to be somewhat less than the theoretical. Also included in Table XXXVII are data for the isotopic sum rule applied to the F2 class frequencies. 4. NORMAL COORDINATE TREATIVENT The vibrational analysis was carried out by means of symmetry factored F and G matrices as described by Wilson~125) Specific details of the application of this method to a five atom tetrahedral molecule have been given(l25,126) previously and will not be repeated here. The B-H bond length in the borohydride ion is not known with a high degree of certainty but it is certainly longer than (125) Wilson., Decius., and Cross, Molecular Vibrations,, McGraw-Hill Book Company,

xx x Hg 2 1/4 1/4 x x ~~~0 1000 2000 3000 cm-1' Fig. 29. Raman spectrum of Nal,~BH4 dissolved in liquid ND3 (1) or liquid NH3 (2). Solvent bands indicated by (x). 2 2 2000 2200 2400 cm-' 1500 1700 190.? cm Fig. 30. The B-H stretching region in Fig. 51. The B-ID stretching region in the Raman spectrum of "1BHZ. 1., per- the Rarnan spectrum of 1"BDZ dissolved pendicularly polarized incident light; in liquid ND3. 1, perpendicularly po2, parallel polarized incident light. larized incident light; 2., parallel polarized incident light. WADC TB 59-207 169A

TABLE XXXVII APPLICATION OF THE TELLER-REDLICH PRODUCT RULE TO THE BOROHYDRIDE ION Frequency Product Ratio Ratio C~assDeito Theoretical Observed ~11BD Al 1.o414 1.443 +2.1% 1 4E 1.414 1o414 0 F2 1.776 1o736 -2. BD- A1l 1.414 1.445 +2.2 J0-E 1.414 1.o408 -.4 2 1. 761 1.727 -1.9 rV i2(loBH - Vo 2(11 4H4) ] 5.52 x 104 (~2 class) D[vi2(1OBD) - Vi-2( =-D) 5o23 x 104 the 1.19 A value founLd(127) in BH30 since the bond in the latter molecule retas an appreciable amouat of sp2 character. Price(128) has suggested 1.26 A for NaBH9 from empirical relationships between vibrational frequency and internuclear distance. while Ford and Richards(l29) have calculated a distance of 1.255 A from measurements involving line breadths in nuclear magnetic resonance spectra of the same compound. The value of 1.25 A used in the calculations thus probably is not seriously in error, and irn any case the force constants are not particularly sensitive functions of this parameter. The potential function in te-nnis of the symmetry coordinates employed(126) can be written as 2V F= S7 + r2F2S2+ F3532 + rF-34S3S4+rFS4 where SI is the nonred-undant sy etry coordlinate belonging to the Al class., S2 the symmetry coordinate of the E class,, and 53 and 54 the stretching and bending coordinates,, respectively2, of the F2 class. Principal force constants are indicated by a single subslcript and interaction by a double; r represents the assumed equilibrium value of the inte.rnuclear distance. Comparison of the calculated wit~h experimental val'ues of the fundamentals of the four isotopic molecules is made in Table XXXVIIIT The standard deviation between the two sets (excllrdding the two Al frequencies involved in Fermi resonances) is 0.5% wituh no de~vi'aticr,,n exceeding 0.9%, Since this agreement probably is better than the uncertainty in the da-ta caused'by experimental error and the neglect of anharmonicity (12;7) GordyIT-,P I Ring2 andn -Bu1"rg D~- Ptiys Rev. 782 51 ( 1950)

-3 TABLE XXXVIII I'0 0 ~~~~~COMPARISON OF OBSERVED AND CALCULATED FU]N1DAM4ENTAL VIBRATIONAL FREQUENCIES OF TBE VARIOUS ISOTOPIC SPECIES OF THE BOROHYDRIDE ION 11BH4 1OBH-4 1 1BD4 10BDD Fundamental B 4 4 Obs Calc Dev Obs Calc Dev Obs Calc Dev Obs Calc Dev A1 vl 2264 2267 +0.1 2270 2267 -0.1 (1570)a 1604 -- (1571) 1604 E v2 1210 1210 0 1208 1210 +0.2 855 856 +0.1 858 856 -0.2 F2 v3 2244 2252 +o. 4 2250 2265 +0 7 1696 1688 -0. 5 1706 1707 +0. 1 v4 1080 1089 +0.9 1093 1097 +0.3 823 823 0 834 827 -O.9 aThose values enclosed in parentheses are fundamentals shifted by Fermi resonance.

other approaches using more refined potential functions such as the orbital valency type were considered unnecessary. The symmetry force constants yielding the best fit are given in Table XXXIX. By employing the defining equations for the symmetry coordinates, these data may be used to calculate values for the more common valence force constants. However, since seven valence force constants can be defined in the complete quadratic potential function whereas only five symmetry force constants are available, it is not possible to give a unique set of values for all of the valence force constants. One can, of course, introduce two arbitrary but reasonable assumptions and so obtain numbers for all the constants, but rather than introduce this artificiality, the unresolved combinations of valence force constants have been listed as such in Table XXXIX. The notation employed is largely self-explanatory, the primed constants being used for interactions between coordinates having only one atom in common. TABLE XXXIX FORCE CONSTANTS FOR THE BOROHYDRIDE ION Symmetry Force Value Valence Force Value Constant (105 dynes/cm) Constant (105 dynes/cm) F1 3.0511 kr 2.768 F2 0. 2897 krr O.o094 F3 2.6737 k-ka- 0.290 F4 0.2905 k —k' o0.291 F34 0.0150 kra-kral 0.011 C. The Dipole Moment of Tetraborane 1. THE PROBLEM By a molecular orbital treatment of boron hydride structures using a threecenter bond approximation., Eberhardt., Crawford., and Lipscomb(l5O) were able to rationa ize the known dipole moment values of 2.13D for B5H9(151) and 3.52D for BjoH146'132) Their treatment also predicted a zero dipole moment for B4Hjo if the boron skeleton, without moments due to attached hydrogens,, were considered. We have measured the dipole moment of tetraborane in benzene solution using a heterodyne beat method and have obtained the relatively small value of 0.56~.10 IDebye units. The small value could easily imply slight protonic character (152)\ — III A. WILabenaye'an R. otei J.Am Chem. N1 — - TY Soc. 7 161 (1952).

in the bridge hydrogens and slight negative character in the hydrogens attached to the single bonded borons; these low moments were neglected in the earlier theoretical treatment, and hence the theoretical arguments are not in serious disagreement with the dipole information. 2, EXPERIMENTAL Tetraborane from laboratory stock was purified by low-temperature fractionation on the vacuum line. The purified product had a vapor pressure of 388 + 1 mm at 0~0. When the sample was allowed to stand at 0~C for more than a few minutes, decomposition was indicated by a slow buildup in pressure. Reageit-grade benzene was dried by refluxing with calcium hydride and distilling on the vacuum lineo Solutions were prepared by condensing benzene and a measured volume of tetraborane vapor into a special weighing cell. Transfer to the dielectric cell was made under an atmosphere of dry nitrogen, since it was considered undesirable to evacuate the cell because of damage to the silvered surfaces. The cell was of a conventional design(133) with a capacitance of about 30 micromicrofarads. The measuring circuit has been previously described.(l34) It was necessary to make the capacitance measurements rapidly because decomposition of the solute in the dielectric cell produced a slow rise in the apparent dielectric constant value. Measurements of the refractive index of the solutions were made with a Bausch and Lomb, precision Abbe Refractometer. No difficulty was encountered in obtaining reproducible readings before appreciable decomposition took place. The refractive index of each solution was measured using light of three wavelengths (43581.9 541P60AY 5893A). A plot of the refractive index for each solution against l/ %2 permitted evaluation of the index of refraction for light of infinite wavelength. A plot of n0* against mole fraction permitted evaluation of Oi/O)0 5., RES'U TS MriD DISCUSSION Results of the dielectric constant and index of refraction measurements are show nabeXadXL, respectively. The value of the dielectric constan:t of pure benzene was obtained from a Bureau of Standards publication(135) andswaused. -to calibrate -the dielectric cell. (155) LeFevre, Dipole Moments (Methuen and Co., Ltd., London)-, p.36

TABLE XL THE DIELECTRIC PROPERTIES OF TETRABORANE Mole Fraction Dielectric Constant at 25~C (589) 0.000 2.274 (155) 1.4971 ~~~~~0.024 2.277 1.494 o0~~~.o0~44 2.278 -1.4918 0o~.0o~88 ~2.286 1=.16 + TABLE XLI THE INDEX OF REFRACTION OF TETRABORANE Refractive Index 5893A 5461A 4358U. 0.000 1.4975 1.5015 1.5191 1.4716 O,~036 1.4955 1.4975 1.514.6 1. 4678 0.057 1.4907 1.4949 1.5118 1.,4655 T 2-520C MO =78.1.107 +.01 d =.879 e = 2,274 Pd 6.5 + 7 cm3 no=1.472

It is of interest to note that the addition of tetraborane to benzene produces very little change in the dielectric constant, but a considerable decrease in the refractive index, indicating that the contribution of the electronic polarization of tetraborane to the total polarization of the solution is somewhat lss than that of the benzene that it displaces, but that the orientation polarization effectively makes up this difference. Because of the instability of the solutions, no attempt was made to obtain the time-consuming volume measurements necessary for determining the density of the solution. To calculate the apparent dipole moment, use was made of the extrapolation method of Cohen-Henriquez as given by Bottcher.(136) In this method the orientation polarization" is obtained directly from the dielectric constant and index of refraction measurements without use of the density. The equation used for the calculation was: [1]O = 5Mo 0e _ 2(n*)o [od 2-n~ do(eo+2)2 0 where [Po]d is the molar orientation polarization of the solute, Mo, do, and eo are the molecular weight, density, and dielectric constant of benzene, no is the refractive index of benzene extrapolated to infinite wavelength, x, e, and n are the mole fraction, dielectric constant, and index of refraction of the solution. This equation differs somewhat from the one given by Bottcher, having been obtained, not by neglecting the effect of the atomic polarization of the solvent, but by using Guggenheim's approximation(l37) that the atomic polarization of the solute and solvent are proportional to their partial molar volumes. Since there is no reliable way of estimating the atomic polarization of the solute, this is probably as good an approximation as any. D., Exchange Studies on Decaborane-l14 1. BACKGRWiND Chemical-experiments at Callery Chemical CoO(138) have demonstrated that four of the hydrogen atoms in decaborane, BjoHI4,, are much more acidic in character than the rest. Inasmuch as structure studies have shown the presence of four three-center or bridge-type bonds in the molecule,, identification of the acid-ic: hydrogens with the bridge locations appears logical. To obtain evidence on the validity of such a hypothesis, advantage may be taken of the fact that

boron-hydrogen vibrational stretching frequencies fall into two well-defined and well-separated regions of the spectrum according to the nature of the bond. In general, bands occurring in the region from about 2300 to 2600 cm- are associated with the presence of terminal hydrogen atoms while bridge hydrogens are indicated by bands in the region of 1400 to 2000 cm-'. Chemical exchange of the active hydrogens with deuterium from heavy water followed by spectroscopic examination should indicate the location of the substituted atoms from the shift of the appropriate bands. 2. EXPERIMENITAL The exchange of BloH14 with D20 was carried out in the presence of small amounts of CH3COOD and ethyl acetate since the pure decarborane did not exchange readily. To insure that the compound had not been altered by its treatment, the deuterated sample was re-exchanged with ordinary water after its spectrum had been obtained and the spectrum of the resultant material compared to that of the starting material. The identity of the two spectra showed that the procedure was suitable. The infrared spectra of the compounds were obtained as nearly saturated solutions in C4 using a Perkin-Elmer Model 21 spectrometer equipped with a rock salt prism. The concentration of these solutions is estimated to be in the neighborhood of 0. molar. 5. RESULTS The bands observed in the spectrum of the original BloH14 agreed quite well with those in the infrared spectrum of a CS2 solution reported by Keller and Johnston.(139) Those differences which were noted were minor and were confined to a few very weak bands in the skeletal region. In the hydrogen stretching region., three bands were observed., each undoub~tably containing several unresolved or partially resolved molecular frequencies. The most intense band showed no structure and ex-tended from about 2500 to 2640 cm-1; to it are assigned the several terminal hydrogen stretching frequencies. The next most intense band exhibited a principal maximum at 1515 cm-1 with secondary maxima of moderate intensity at 1475 and 1580 cm-1. The third band consisted chiefly of two rather weak maxima at 1890 and 1942 cm-1 together with two or three very weak frequencies on either side. The latter two bands are assigned to motions of the bridge hydrogen atoms. The shape and position of band I in the 2500 cm-1 region was not altered significantly after exchange with deuterium. Both band II and band III, however, were shifted to lower frequencies but without change of shape or-relativ

at 1398 and 1425 cm-1. The ratios of the hydrogen frequencies to those of the deuterium compound thus fall between 1.30 and 1.35 and are quite comparable to the ratios observed for the analogous bands in, for example, diborane and deuterodiborane. In addition to these frequencies, a fourth band was present in the spectrum of the deuterated sample. This fourth band showed a remarkable similarity in appearance to band II but had only about half the intensity or less. It principal maximum was at 1724 cm1 with the secondary maxima at 1710 and 1712 cm-1. The origin of this band is not clear. Since it appears some 200 cm- higher than band II in the original (undeuterated) sample, it cannot be attributed to incompletely exchanged material, and similarly it does not seem likely that it arises from "nonexchangeable" bridge hydrogens. It also does not appear to be due to the replacement of terminal hydrogens by deuterium since the ratio of any of the terminal hydrogen frequencies to any of those in band IV is larger than the limiting square root of two. It is possible that it represents a combination of the frequencies in band II with a frequency in the region of 550 cm:which was beyond the range of the spectrometer in the present investigation. This suggestion is not entirely satisfactory but appears to be the only reasonable one. Despite the uncertainty in the assignment of this last band, the spectroscopic evidence seems to indicate quite conclusively that it ise theo bridge ihydrogen atoms which can be replaced by deuterium exchange in heavy water and not the terminal, and that these, therefore, are acidic in nature. Several subsequent studies(l140) which have been published since the termination'of this work confirm that the bridge hydrogens exchange most rapidly but indicate also that the terminal hydrogens will exchange slowly in a secondary step. The foregoing preliminary study was completed prior to the other more detailed investigations.(1111)

VI. CHEMICAL CORRELATIONS The foregoing data together with data available in the Research Division of Callery Chemical Company and in the literature of boron suggest a number of systematic generalizations covering the reactions of the boron hydrides. Such generalizations were made by one of the senior authors (WP) in collaboration with Dr. L. J. Edwards of Callery Chemical Company. The results are summarized in a paper entitled "Systematics of the Boron Hydrides."( ) While the preparation of this paper was not done directly under this project, part of the data used was a result of these investigations and the assistance of the Wright Air Development Center in providing data for the summary paper is hereby gratefully acknowledged.

UNIVERSITY OF MICHIGAN 3 9015 03695 5618