LEGANE) EXCHANGE REACTEONS 0F SGME GRQUP EHB B-DEKETONATE COMPLEXES Thesis for the Degree of M. S. MICHEGAN STATE UNIVERSITY 3.. OGUZIE NWEKE 1968 TH ESIS . Univ=3z‘*2..'" 4 fl J 9}? Iva-nan 9'." ‘r' WV .35 BINDING BY HMS] SflflS’ MK llflDEPV NC. I lnnnnv n“, D ABSTRACT LIGAND EXCHANGE REACTIONS OF SOME GROUP IIIB fi-DIKETONATE COMPLEXES BY S. Oguzie Nweke Nuclear magnetic resonance Spectroscopy has been used to study ligand exchange reactions between M(acac)3 and M(bzbz)3 complexes, where M = A1, Ga, or In, and acac = acetylacetonate and bzbz = dibenzoylmethanate. The ex- change of ligands between Ga(acac)3 and Ga(hfac)3 (hfac = hexafluoroacetylacetonate) has also been investigated. Equilibrium constants for formation of each mixed ligand complex from the corresponding parent complexes have been determined in benzene solution. The equilibrium constants at room temperature for forma- tion of M(acac)(bzbz)2 and M(acac)2(bzbz) complexes are only '93. 15 to 25% smaller than the values expected for a random statistical distribution of ligands. Enthalpy changes for the exchange reactions when M = A1 or Ga are zero or nearly zero; within experimental error, the entropy changes are equal to the values expected for a statistical distribution of ligands. In the Ga(acac)3-Ga(hfac)3 system, the mixed ligand complexes are very strongly favored at the expense of the S. Oguzie Nweke parent complexes, the equilibrium constants (extrapolated to 25°) being 23- 400 times larger than statistical values. The deviations from statistical behavior are due almost entirely to enthalpy effects. LIGAND EXCHANGE REACTIONS OF SOME GROUP IIIB B-DIKETONATE COMPLEXES BY -".r>k SE Oguzie Nweke A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1968 4 £352 Q/M/H To my family ii ACKNOWLEDGMENTS I am very grateful to Dr. T. J. Pinnavaia for his good understanding, guidance, encouragement,and help throughout the period of this research. Financial support from the Agency for International Development is very highly appreciated. iii TABLE OF CONTENTS Page I. INTRODUCTION . . . . . . . . . . . . . . . . . . 1 II. EXPERIMENTAL . . . . . . . . . . . . . . . . . . 8 A. Preparation of Compounds . . . . . . . . . . 8 1. Reagents . . . . . . 8 2. Tris 2 ,4-pentanedionato gallium(III) . . 9 3. Tris 2, 4—pentanedionato aluminum(III) . . 10 4. Tris 2, 4-pentanedionato indium(III) . . . 10 5. Tris 1, 3- -diphenyl-1, 3- propanedionato)— gallium(III). . . . . . . . . . . 10 6. Tris(1, 3- -diphenyl—1, 3-propanedionato)— aluminum(III). . . . . . 11 7. Tris(1, 3- -diphenyl-1, 3-propanedionato)- indium (III) . . . . . . . . . . 12 8. Tris(1, 1, 1, 5, 5, 5-hexafluoro-2, 4-pentane- dionato)gallium(III) . . . . . . . . . . 12 B. Preparation of Solutions for Nmr Studies . . 13 C. Proton Magnetic Resonance Spectra . . . . . . 13 D. Signal Intensity Measurement . . . . . . . . 14 III. RESULTS AND DISCUSSION . . . . . . . . . . . . . 15 IV. REFERENCES . . . . . . . . . . . . . . . . . . . 50 iv TABLE II. III. IV. VI. VII. VIII. LIST OF TABLES Nomenclature for diketonate ligands . . . Dependence of K1 and K2 for the Ga(acac)3- Ga(bzbz)3 system on solute composition . . Dependence of K1 and K2 for the Ga(acac)3- Ga(bzbz)3 system on total solute molarity. Temperature dependence of K1 and K2 for the Ga(acac)3-Ga(bzbz)3 system in benzene . . Temperature dependence of K1 and K2 for the Al(acac)3-Al(bzbz)3 system in benzene . . Temperature dependence of K1 and K2 for the Ga(acac)3-Ga(hfac)3 system in benzene . . Thermodynamic data for ligand exchange in benzene at 25° . . . . . . . . . . . . . . Thermodynamic data for formation of mixed ligand complexes in benzene at 25° . . . . Page 37 38 41 42 43 44 45 Figure 1. LIST OF FIGURES Page Representation of the nonequivalent methyl proton enviornments, a and b, in Ga(acac)2(bzbz) 17 Methyl proton nmr spectra for equilibrium mix- tures of Ga(acac)3 and Ga(bzbz)3 in benzene at 36° . . . . . . . . . . . . . . . . . . . . 19 Methyl proton nmr spectra for equilibrium mix- tures of Ga(acac)3 and Ga(hfac)3 in benzene at 36° . . . . . . . . . . . . . . . . . . . . 22 Methyl proton resonance spectrum of an equi— librium mixture of Ga(acac)3 and Ga(hfac)3 in di— chloroethane at -46° . . . . . . . . . . . . . 24 Methyl proton resonance spectra for an equi— librium mixture of Ga(acac)3 and Ga(bzbz)3 showing temperature dependence of the two Ga(acac)2(bzbz) methyl lines . . . . . . . . . 26 Methyl proton nmr specta for equilibrium mixtures of (a) In(acac)3-In(bzbz)3 and (b) Al(acac)3- Al(bzbz)3 in benzene at 36° . . . . . . . . . . 28 Ligand exchange equilibria for the Ga(acac)3- Ga(bzbz)3 system in benzene at 36° . . . . . . 33 Ligand exchange equilibria for the Ga(acac)3- Ga(hfac)3 system in benzene at 36° . . . . . . 35 Vi INTRODUCTION Reactions in which distinguishable ligands on two cen— tral metal moieties undergo exchange to give mixed ligand products have been discussed in the literature under several synonymous titles ranging from diSproportionation reactions to scrambling, redistribution, or ligand exchange reactions. These reactions, which will be referred to as ligand exchange reactions throughout this thesis, may be represented in general form by the equation: K. 1 up I BI > 2 I I 11"1-1 1+1 <_ n-1B1 MA + MA n—i+1Bi—1 where A and B are the exchangeable ligands, n is the num- ber of ligands, i is 1,2,3,...(n-1), and Ki is the equi- librium constant. The recognition of these reactions as a class of chemical reactions was due to the work of Calingaert and Beatty (1), who studied the exchange of alkyl radicals and halogen atoms in metal alkyls and organometallic halides. As a result of the intitial studies of Calingaert and Beatty, ligand exchange reactions have been widely used for the preparation of many new compounds which are not readily obtainable by other types of reactions. Ligand exchange reactions are employed even on a commercial scale. For example, phosphorous oxychloride is prepared by ligand 1 2 exchange between phosphorus(v) oxide and phosphorus(v) chloride, and organotin(IV) chlorides are obtained by reac- tion of the corresponding tetra-alkyltin(IV) with tin(IV) chloride. A variety of experimental methods have been utilized for the quantitative characaterization of ligand exchange equilibria. ~For systems that are kinetically inert at ordinary temperatures, standard separation techniques, such as chromatography (2-6), solvent extraction (7), distilla- tion (1,8), and vapor density measurements (9), have been exploited very successfully. On the other hand, analysis of labile systems is best achieved using instrumental techniques which do not require the separation of any of the components of the equilibrium mixture. The following instrumental methods have been successfully employed in the study of exchange reactions: polarography (2) and infrared (10), Raman (11), mass (12), and nuclear magnetic resonance (nmr) (13,14,15), Spectroscopy. Nuclear magnetic resonance spectroscopy has proved 13mg most rewarding in terms of the accuracy, the wealth of information available, and the ease with which such information are obtained. Using nmr spectrosopic methods one may obtain,in addition to equilibrium constants, rates of intra- and intermolecu- lar rearrangements and the stereochemistry of mixed ligand molecules. Recently, five review articles (16-20) on ligand ex- change reactions have appeared in the literature. Earlier 3 studies of ligand exchange reactions dealt almost exclusive— ly with exchange ofnmmodentateligands in non—labile sys- tems. Numerous examples can be found in the review articles cited above. Although a considerable amount of quantitative information is available for ligand exchange of monodentate ligands, relatively few studies of exchange of bidentate ligands have been reported. The relative lack of informa— tion for bidentate ligands can be traced to the difficulties that would often be encountered in applying spectrOSOpic methods for determination of equilibrium concentrations. One group of the bidentate ligands that lends itself to exchange studies by nmr spectroscopy is the 5-diketonates. This fact probably explains why most of the published re- sults on ligand exchange reactions of bidentate ligands is for metal 5—diketonate complexes. Since the researches described in this thesis are concerned with a study of ligand exchange reactions for some gallium(III) B-diketonates by nmr spectroscopy, a brief review of diketonate exchange will be given here. For convenience, structural formulae and abbreviations for the dieketonate ligands discussed are pre- sented in Table I. 4 Table I. Nomenclature for diketonate ligands. ._I_ V ~vw— ‘_‘___ ([Rr- COCHCO -R' ] — Abbreviation Name R R' acac acetylacetonate CH3 CH3 bzbz dibenzoylmethanate C6H5 C6H5 thd 2,2,6,6-tetramethyl-3,5- E7C4H9 EfC4H9 heptanedionate hfac hexafluoroacetylacetonate CF3 CF3 tfac trifluoroacetylacetonate CH3 CF3 Lowry and Fay (14) studied exchange of B-diketonate ligands between Ti(acac)2F2 and Ti(tfac)2F2. The equilib- rium quotient was found to be about four times larger than the value expected for a random statistical distribution of ligands, the mixed ligand complexes being favored. The enthalpy change for the reaction was zero within the limits of experimental error. Equilibrium constants for the system Hf(acac)4— Hf(tfac)4 (21) were found to be about 2.5 times the statis— tical value. Similar results were obtained for the analogous zirconium system (13,21) and for the [Y(tfac)4]- -[Y(hfac)4]— system (22). Diketonate exchange in the Al(thd)3—Al(acac)3, Al(thd)3— Al(hfac)3, and Al(acac)3-Al(hfac)3 systems in chlorobenzene have been studied by Fortman and Sievers (15). For the Al(thd)3-Al(acac)3 system, the equilibrium concentrations 5 of mixed ligand complexes were found to be slightly less than the amount expected on the basis of a statistical distribution of ligands. Concentrations very much larger than statistical values were observed for the other two sys— tems. Using solvent extraction methods, Newman (23) found that the system In(acac)3-In(bzbz)3 behaves almost statistically. In most of the above investigations only equilibrium constants were determined, and the results were compared with those calculated on the basis of a random distribution of ligands. Attempts to explain the deviations of some of the systems from randomness had to be Speculative. In the study of Zr(acac)4-Zr(tfac)4 system, however, Pinnavaia and Fay (13) used fluorine nmr spectroscopy to determine the equilibrium constants, enthalpy and entropy changes for the diketone exchange. They found that the equilibrium constants were about 2.5 times greater than statistical, the mixed complexes being favored. The deviations from randomness were attributed to entropy effects; entropy changes were either zero or nearly zero. They noted that small devia- tions from randomness could be attributed to entropy changes, but very large deviations from randomness almost certainly involve an enthalpy effect. This modifies the suggestions by Van Wazer (17a) that deviations from randomness are at— tributable to changes in enthalpy since enthalpy changes are never extreme. 6 From the information presently available, it appears that for practical purposes the behavior of any system with reference to randomness is independent of the central metal atom but is dependent to a large extent on the nature of the ligands that are undergoing exchange. For exchange be— tween non-fluorinated ligands and fluorinated ligands, the mixed complexes are favored at the expense of the parent complexes, whereas for the exchange of similar or closely related ligands (eg., both fluorinated or both non-fluor- inated) the equilibrium constants are very close to statis- tical values. Therefore, as noted by Fay (20), the equilib- rium constants are appreciably larger than statistical when the ligands differ in charge or electronic structure but close to statistical when the ligands have the same charge and electronic structure. This generalization is in accord with the theoretical considerations of Kida (24) and Marcus and.EUezer (25), who have shown that the mixed ligand com- plexes are favored at the expense of the parent complexes when the ligands are electronically dissimilar, because of a decrease in the average ligand-ligand repulsive energies. In the present work, equilibrium constants in benzene solution for the following systems have been determined by nmr spectroscopy: Ga(acac)3-Ga(bzbz)3, Ga(acac)3-Ga(hfac)3, In(acac)3—In(bzbz)3, and Al(acac)3-Al(bzbz)3. Enthalpy and entropy changes for the systems Ga(acac)3-Ga(bzbz)3, Ga(acac)3—Ga(hfac)3, and Al(acac)3-Al(bzbz)3 have also been determined. These investigations have stimulated the study 7 of kinetics of the intramolecular exchange in the molecules Ga(acac)2(bzbz) and Ga(acac)2(hfac). II . EXPERIMENTAL A. Syntheses 1. Reagents Gallium metal, indium(III) chloride, dibenzoylmethane and aluminum isopropoxide purchased at about 99.5% purity were used without further purification. 2,4-Pen- tanedione and 1,1,1,5,5,5-hexafluoro-2,4-pentanedione were freshly distilled before use. Gallium(III) chloride was prepared by reaction of the elements at 22 200°. The compound was purified by subli— mation ig_y§gug and subsequently handled under anhydrous conditions. Hydrous gallium(III) nitrate was prepared using the following procedure. Gallium metal was dissolved in the minimum amount of hot EHEE.£SS$§' and the solution was evaporated almost to dryness on a steam bath. The resulting paste was dissolved in distilled water, and the solution was evaporated to dryness. This process of dissolution and evaporation was repeated until the paste no longer had the odor characteristic of nitric acid. The paste was then dis- solved in water, and the gallium was precipitated as the hydrous hydroxide by addition of dilute ammonia solution. 8 9 The hydroxide was repeatedly redissolved in 3ggnitric acid and reprecipitated with dilute ammonia until the mother liquor gave no test for chloride ion with acidic silver nitrate solution. Finally, the gelatinous hydroxide was dissolved in 3M_nitric acid and the solution evaporated to dryness. A similar procedure was used for the preparation of hydrous indium(III) nitrate. All solvents used in the synthesis were reagent grade. Benzene and hexane were dried by refluxing over sodium, lithium aluminum hydride,or calcium hydride. Carbon tetra- chloride was distilled from calcium hydride. Dioxane was dried by distilling from freshly cut sodium. 2. Tris(2,4-pentanedionato)gallium(I;;). The method of Morgan and Drew (26) was adopted for the preparation of this compound. Acetylacetone (6.0 ml, 59 mmoles) was added to a solution of hydrous gallium(III) nitrate (2.15 g) in distilled water. While the mixture was refluxing, 20 ml of approximately 1.5M;ammonia solution was added dropwise with constant stirring. The gallium acetyl- acetonate formed was filtered, dried i§_yggug at room tem- perature, and recrystallized from benzene-hexane. The com- pound melted at 196-1970 d. (sealed capillary); lit. (26) 194-195°. 10 3. Tris(2,4-pentanedionato)aluminum(III). The compound was prepared following the method of Piper and Fay (27) for synthesis of other aluminum 5- diketonates. A solution of acetylacetone (6.09 ml, 59.4 mmoles) in 80 ml of benzene was added to a solution of aluminum isopropoxide (4.04 g, 19.8 mmoles) in 160 ml of benzene at room temperature. The solvent and the isopropyl alcohol produced were removed by vacuum distillation at room temperature. The white, crystalline product was re- crystallized from benzene—hexane, and dried in vacuo at 80° for 0.5 hour. The compound melts at 194—1950; lit (26,28,29) 192-1940, 196-1970, 194.60. 4. Tris(2,4-pentanedionato)indium(III). The method of synthesis was analogous to that used for the preparation of gallium acetylactonate (26). The very pale yellow crystalline product was recrystallized from benzene-hexane and dried in vacuo at 80° for about 0.5 hour. Melting point is 186—1880; lit (26) 186—1870. 5. Tris(1,3-diphenyl-1,3-propanedionato)gallium(III). The procedure used for the preparation of this compound is similar to that described by Funk and Paul (30). Di- benzoylmethane (9.1 g,,40 mmoles) in 240 ml of dioxane was heated with freshly cut sodium (0.92 g, 40 mmoles) at reflux temperature to form a solution of the sodium salt. A 11 solution of gallium(III) chloride (0.14 g, 13 mmoles) in 95 ml of dioxane was added to the hot solution of sodium dibenzoylmethanate, and heating was continued for one hour. The hot solution was filtered through a sintered glass fun- nel and the filtrate was concentrated to about half the original volume by boiling off some solvent. The concen- trated solution was cooled to zero degree and the yellow crystals that separated were filtered off, washed with ether and recrystallized from benzene—hexane. The crystals were dried in vacuo at 100° for about 0.5 hour. The melting point is 294-295°; lit. (30) 284°. The discrepancy in the melting points necessitated elemental analysis. 53:31. Calcd. for Ga(C15H1102)3: c, 73.09; H, 4.5; Ga, 9.43. Found: C, 73.19: H, 4.62; Ga, 9.61. 6. Tris(113-diphenyl-1,31propanedionato)aluminum(III). The compound was obtained in 59% yield using the same method that was described for the preparation of the gal- lium analog. The yellow crystalline product was recrystal- lized from benzene—hexane and dried in vacuo at 100°. The melting point is 305-307°. This compound has been reported (32,33) previously, but no synthetic details, melting point. or analytical data were given. Therefore, the compound was analyzed. 5:131. Calcd. for Al(C15H1102)3: c, 77.58,- H, 4.77; Al, 3.87. Found: C, 77.80; H, 4.90; Al, 4.05. 12 7. Tris(1,3-diphenyl-1,3-prgpanedionate)indium(III). The compound was obtained in 56% yield using the same method that was described for the preparation of a similar compound of gallium(III). The yellow crystalline product was recrystallized from benzene—hexane and dried in vacuo at 100°. The melting point is 254—256°. This is the first reported synthesis of the compound. Anal. Calcd. for In(C15H1102)3: C, 68.89; H, 4.24; In, 14.63. Found: C, 69.01; H, 4.24; In, 14.39. 8. Tris(1,1,1,5,5,5-hexafluoro—2,4-pentanedionato)— gallium(IIrIT This compound has not been reported previously but was readily obtained using the method described by Morris, Moshier and Sievers (31) for the preparation of aluminum analog. Hexafluoroacetylacetone (9.00 g, 43.5 mmoles) was added to gallium(III) chloride (2.5 g, 14 mmoles) in carbon tetrachloride at room temperature. The mixture was shaken until evolution of hydrogen chloride ceased and the solution was refluxed for 0.5 hour. The solution was concentrated by boiling off some solvent under a stream of dry nitrogen. The resulting needle-shaped crystals were recrystalized from carbon tetrachloride, washed with hexane and dried in vacuo at room temperature. Melting point is 70-72°. Anal. Calcd. for Ga(C5HF602)3: c, 26.08; H, 0.44; F, 49.50; Ga, 10.09. Found: C, 26.03; H, 0.50; F, 49.70; Ga, 9.94. 13 B. Preparation of Solutions for Nmr Studies All complexes were found not to be sensitive to atmos- pheric moisture, but prolonged exposure to the atmOSphere was avoided. The samples were weighed into vials and the desired volume of solvent added with a one milliliter syringe. Aluminum(III), gallium(III), and indium(III) di- benzoylmethanate complexes are only slightly soluble in benzene. Therefore, mixtures containing these compounds were heated gently to effect dissolution. However, all Ga(acac)3-Ga(hfac)3 mixtures were dissolved at room tempera- ture despite the inconvenient time period required. When Ga(acac)3-Ga(hfac)3 mixtures were warmed, the undissolved crystals readily formed an immiscible oil which then dis- solved, and under these conditions unidentified decomposi- tion products were observed in the nmr spectrum of the solu- tion mixture. If oil formation was avoided in the dissolu- tion step, then equilibrium mixtures stable at temperatures even as high as 120° were obtained. The rather long time required for dissolution at room temperature is attributed to the low rate at which the less soluble component, Ga(bzbz)3 or Ga(hfac)3, is consumed in the ligand exchange reaction. C. Proton Magnetic Resonance Spectra Proton magnetic resonance spectra were obtained using a Varian A-60 analytical spectrometer at 60.000 Hz. The 14 instrument was equipped with a Varian variable temperature controller, Model V-6040. Temperatures were determined by measuring the chemical shift differences for methanol (low temperatures)and ethylene glycol (elevated temperatures). The magnetic sweep width was checked using the audiofre- quency side—band technique or a standard sample containing seven non-interacting compounds with chemical shifts uni- formly distributed over a magnetic sweep width of 500 Hz. D. Signal Intensitpreasurement Signal areas used in the determination of equilibrium constants for aluminum(III) and gallium(III) were determined by electronic integration. The integrations were checked by planimetry, and the equilibrium constants obtained by both methods agreed within experimental error. (In general, each Spectrum was integrated 12 to 25 times and the results were averaged in order to reduce errors caused by variations in magnetic field sweep. Also, errors due to saturation effects were avoided by employing small radiofrequency fields. The chemical Shift differences between the methyl reso- nance lines in In(acac)3-In(bzbz)3 mixtures were too small for accurate integration. An estimate of the areas was ob- tained by preparing solutions with compositions such that the two closest lines had equal heights. The two overlap- ping lineS were therefore assumed to have equal intensities. Areas were measured with a compensating polar planimeter. III. RESULTS AND DISCUSSION Prior to determining the ligand exchange equilibria which occur in Ga(acac)3-Ga(bzbz)3, Ga(acac)3-Ga(hfac)3, Al(acac)3-Al(bzbz)3 and In (acac)3—In(bzbz)3 mixtures, the time required to achieve equilibrium in each system had to be investigated. Equal molar amounts of Ga(acac)3 and Ga(bzbz)3 at a total molarity of 0.050Mgin benzene required 23, 0.5 hour to dissolve at room temperature. A study of the relative intensities of nmr lines as a function of time showed that about an additional 40 minutes was required for the system to attain equilibrium after complete dissolu— tion of the parent complexes. On the other bani, less thau 15 min. was required for the Ga(acac)3-Ga(hfac)3 System. Unlike the mixture of Ga(acac)3 and Ga(bzbz)3, the indium system equilibrated almost instantaneously. Investigations Similar to those carried out for Ga(acac)3-Ga(bzbz)3 system were adopted for Al(acac)3 and Al(bzbz)3 mixture using a solution containing equimolar amounts of the two parent complexes and of total concentration 0.214M, Dichloroethane was used as solvent instead of benzene, because of the low solubility of Al(bzbz)3 in the latter solvent. A plot of the relative concentration of Al(acac)2(bzbz) in the solution Kg. time indicated that the system was still not in equilibrium after 15 16 162 hours at room temperature; however, at 80° the system attained equilibrium within 8 hours. For each of the systems, there are four possible Species in solution at equilibrium, yi§., the two parent complexes and two mixed complexes, M(acac)2(dik') and M(acac)(dik')2 where dik' is bzbz or hfac. The methyl protons on the acetylacetonate ligands of the M(acac)(dik')2 and M(acac)3 complexes are all equivalent, and each is expected to give one proton resonanCe line. The M(acac)2(dik') species, however, should give rise to two non—equivalent groups of methyl protons. The sketch of Ga(acac)2(bzbz) in Figure 1 serves as an example. Therefore, the M(acac)2(dik') complexes Should give two equally intense nmr absorption signals. No methyl resonance is possible, of course, for either M(bzbz)3 or M(hfac)3. Therefore, a total of four methyl proton resonance lines Should be observed for equilibrium mixtures of each ligand exchange system. The resonance lines were assigned by studying the changes in relative Signal inten- sities as a function of ligand composition. The ligand composition is specified by the quantity facac’ which is defined as fraction of total ligand present as acetylaceton- ate. Proton nmr spectra for equilibrium mixtures of the sys- tem Ga(acac)3-Ga(bzbz)3 in benzene solution at 36° are shown in Figure 2. The lines at T8.13 and 18.32 are assigned to Ga(acac)2(bzbz), and the lines at 18.19 and 18.25 are assigned to Ga(acac)(bzbz)2 and Ga(acac)3, respectively. 17 Figure 1. Representation of the nonequivalent methyl proton environments, a and b, in Ga(acac)2(bzbz). 18 CH | 3 H‘C/C\O (a) I r HaC-C\ _C/C6H5 \ r ------- Ga JU/C-HH-J ‘°’ ’ % -C‘C H H3C-C\ $ 6 5 ,CI H C1:03) CH 19 Figure 2. Methyl proton nmr spectra for equilibrium mix- tures of Ga(acac)3 and Ga(bzbz)3 in benzene at 36°; total solute molarity = 0.105, 20 fococ =33 fococ = 89 M )\_JL___ I 8.08t 8.23t 8.38: Figure 2. 21 Unlike the Ga(acac)3-Ga(bzbz)3 system, a total of only three lines were observed for Ga(acac)3-Ga(hfac)3 mixtures as Shown in Figure 3. The resonance lines at 18.51, 18.38 and 18.25, respectively, are assigned to Ga(acac)(hfac)2, Ga(acac)2(hfac), and Ga(acac)3. The appearance of one resonance Signal for methyl protons of Ga(acac)2(hfac) is attributed to rapid intramolecular rearrangement processes which average the expected non-equivalent methyl group en— vironments. This was confirmed by the methyl proton nmr spectrum of an equilibrium mixture of Ga(acac)3-Ga(hfac)3 at facac value of 0.667 in methylene chloride at -46°. At this temperature, the rate of:hu1amolecular exchange for Ga(acac)2(hfac) is sufficiently Slow to observe two lines. The Spectrum at —46° is shown in Figure 4. A time averaged methyl resonance may also be observed for Ga(acac)2(bzbz), as shwon in Figure 5. Above 60° the lines broaden, collapse completely at about 80°, and then at about 109° they are coalesced to a single, rather sharp line which appears be- tween the Ga(acac)(bzbz)2 and Ga(acac)3 resonances. Figure 6 Shows methyl proton Spectra for an In(acac)3- In(bzbz)3 mixture and an Al(acac)3—Al(bzbz)3 mixture in benzene at 360. In the spectrum of the In(acac)3— In(bzbz)3 mixture, the line at lowest field is assigned to In(acac)(bzbz)2, the line at next highest field to In(acac)2(bzbz), and the line at highest field to In(acac)3. The assignments for the Al(acac)3—Al(bzbz)3 mixture are analogous to the assignment made for the corresponding gal- lium(III) system. 22 Figure 3. Methyl proton nmr spectra for equilibrium mix- tures of Ga(acac)3 and Ga(hfac)3 in benzene at 36°; total solute molarity = 0.10M, 23 . . {000: =79 24 Figure 4. Methyl proton resonance spectnunof an equi— librium mixture of Ga(acac)3 and Ga(hfac)3 in dichloromethane at -46°; facac = 0.67, total solute molarity = 0.20M, ‘5 egg Figure 5. 26 Methyl proton resonance spectra for an equie librium mixture of Ga(acac)3 and Ga(bzbz)3 showing the temperature dependence of the two Ga(acac)2(bzbz) methyl lines; facac = 0.67; total solute molarity = 0.20M, Figure 6. 28 Methyl proton nmr spectra for equilibrium mix- tures of (a) In(acac)3-In(bzbz)3 in benzene at 36°; facac = 0.29, total solute molarity = 0.20M; and (b) Al(acac)3-Al(bzbz)3 in benzene at 36°; facac = 0.62, total solute molarity = 0.205. 29 famC -.—. .62 M (a) M W 12mC . .67 (b) 5 cps. 30 It may be noted that In(acac)2(bzbz) is Similar to Ga(acac)2(hfac) in that both compounds exhibit a single Sharp methyl line at room temperature. On the other hand, Al(acac)2(bzbz) is similar to Ga(acac)2(bzbz) in that two methyl lines are observed. The simplified Gutowsky--Holm equation (34) states that the first order rate constant, k, at the coalescence temperature for exchange of protons between two equally populated, nonequivalent sites is given by k: —7T—-0v JE' where 0v is the frequency separation between the resonance components in absence of exchange. Since values of 0v for all four MCacac)2(dik') complexes are of the same order of magnitude, and since the coalescence temperature for Ga(acac)2(bzbz) and Al(acac)2(bzbz) are much larger than the coalescence temperatures of Ga(acac)2(hfac) and In(acac)2(bzbz) the rate of methyl group exchange for the latter two compounds must be much larger than the former compounds. For each of the equilibrium mixtures, the concentrations of complexes containing acetylacetonate were determined by integration of their methyl resonance lines. Concentrations of complexes not containing acetylacetonate were determined by difference. In the discussion of the equilibria, the quantity f represents the molar fraction of total M(acac)n(dik')3_n 31 complex present as M(acac)n(dik')3_n, where n = 0, 1, 2, or [M(acac)n(dik')3_n] 3, and is given by §?‘ . For a ran- 2 [M(acac) (dik') ] n=0 n 3-n dom statistical distribution of ligands, it may be shown _ 11 3-n 31 that fM(acac)n(dik')3_n - facac fdik' nt(3-n)i where facac and fdik" respectively, are the molar fractions of total ligand present as acetylacetonate and dibenzoylmeth— anate or hexafluoroacetylacetonate. The equilibria may be described by Specifying two arbitrary, independent equi— librium constants. The constants that were determined ex- perimentally are: K1 <— M(dik')3 + M(acac)2(dik') > 2 M(acac)(dik')2 (1) K2 <___> 2 M(acac)2(dik') (2) M(acac)3 + M(acac)(dik')2 However, it will be convenient to discuss the equilibrium in terms of the formation of one mole of the mixed ligand complexes from the parent complexes: K 2/3 M(dik')3 + 1/3 M(acac)3{:§é> M(acac)(dik')2 (3) 1/3 M(dik')3 + 2/3 M(acac)3<_£:? M(acac)2(dik') (4) 1 where K = Ki/a K2/3 and K 2 = Ki/3K3/3. Based on the £1 E ‘Statistical expression for fM(acac)n(dik')3_ at facac = fdik" it may be Shown that the statistical value of both K1 and K2 is 3.00. It follows that Kf1 = Kf2 = 3.00. 32 vs. f for the A plot Of fGa(acac)n(bzbz)3_n -—— acac system Ga(acac)3-Ga(bzbz)3 is shown in Figure 7. The Curves in broken lines were calculated assuming a random statistical distribution of ligands. A comparison of the theoretical curves with the experimental curves shows that the concentrations of the complexes formed in an equilibrium mixture of Ga(acac)3 and Ga(bzbz)3 are very close to those computed for a random scrambling of ligands. Similar plots for the Ga(acac)3-Ga(hfac)3 system are Shown in Figure 8. Unlike the Ga(acac)3-Ga(bzbz)3 system, the concentrations of the different Species in an equilibrium mixture of Ga(acac)3 and Ga(hfac)3 are very much different from those based on a random distribution of ligands. Figure 8 clearly Shows that the mixed ligand species are strongly favored at the expense of the parent complexes. Equilibrium constants defined by reactions 1 and 2 were studied as functions of ligand composition and of total solute molarity at 36° in benzene solution for the system Ga(acac)3-Ga(bzbz)3. The results are shown in Tables II and III. Within the limits of experimental error, the equilibrium quotients are independent of both ligand com- position and total solute molarity. Therefore, the solu— tions are probably fairly ideal; at least the activity quotient is constant over the ranges studied. On the basis of the results of these investigations, concentration and composition studies on the other systems were considered Figure 7. 33 Ligand exchange equilibria for the Ga(acac)3- Ga(bzbz)3 system in benzene at 36°. Experimental curves,._____; calculated curves assuming a ran- dom statistical distribution of ligands, Total solute molarity is 0.10M, 34 [q q _ _ q 4 S q _ I I I I \\ II 0” \o I \ I \ / I I 3 II \ \ 3 IA \ n e \ I. \ 2 . x. 1 = . \ n. \ [I ’. \ '1 / I I x \ / x \ / \ I 0‘ o 4 \ / 1 I / I = h z n z 21 O x , I. : . n \\\ / \\ / \\\\ I II \ I.’ \\\ I” \\\\ I’ll . . . _ _ h . _ _ II 8 6 4 2 LO FnEBVcAoooovooop LO .8 Figure 7. Figure 8. 35 Ligand exchange equilibria for the Ga(acac)3- Ga(hfac)3 system in benzene at 36°. Experi- mental curves,.______; calculated curves assuming a random statistical distribution of ligands, ——————— . Total solute molarity is 0.1034. 36 ID I l l l l l l t n=l ":2 08 h ‘ I I r \ 7 \\ I \ c \ A \ - O n: O \ O \ a. .c _ .’ A ( I O \ I O \ 9.-..“ ‘ a"- I 0 I’ ‘8 ‘~ ‘l\ 3 04 '" I \\ ‘ \ / \ O I \ ‘ ’ 0 / \ I .‘ I q.— I . \ .l \ I \ , ’ h- , \\ .I I I I, \II \/ o /\ \ \. 2 I / \ \ \ o " I 0’ \ \O \\ I \ \ I \ / \ n I. I, \ \. I’ ,’ K ’I ’ \\ ’ ”. § \ I ._.L--' I I I ~ '54 L .2 .4 .6 .8 f0 COC 37 Table II. Dependence of K1 and K2 for the Ga(acac)3- Ga(bzbz)3 system on solute composition /--——-Average Valuesb acac K1 K2 0.291 2.22 i .25 . . 0.387 2.40 i .16 2.48 i .19 0.420 2.60 i .17 2.52 i .17 0.490 2.38 i .23 2.58 i .18 0.598 2.63 i .15 0.813 . . . 2.54 i .22 aIn benzene at 36°; total solute molarity is 0.0503, b Average of at least 12 spectral measurements. CAll errors are estimaed at the 95% confidence level. 38 Table III. Dependence of K1 and K2 for the Ga(acac)3— Ga(bzbz)3 system on total solute molaritya . w b Total Molarity ,—————e-Average Values K1 K2 0.05 2.38 i .23C 2.58 i .18 0.13 2.51 i .10 2.62 i .11 0.18 . 2.54 i .20 2.57 i .08 0.20 2.50 i .37 2.62 i .25 aIn benzene at 36°; f = 0.490. acac b Average of at least 12 spectral measurements. gAll errors are estimated at the 95% confidence level. 39 not necessary. It may be noted that the values of K1 and K2 in Tables II and III are only 15—25% smaller than the statistical value of 3.00. On the other hand, the analogous constants for the Ga(acac)3-Ga(hfac)3 system are at least 100 times larger than the statistical value, as will be shown later. Equilibrium constants for the In(acac)3-In(bzbz)§ and Al(acac)3--Al(bzbz)3 systems were also determined in benzene solution. In both systems the constants K1 and K2, re- spectively, were determined at facac values near 0.3 and 0.6. For the In(acac)3-In(bzbz)3 system at 36°, K1 = 2.6 i .2 and K2 = 2.4 i .2. It was not possible to obtain equi— librium constants for the aluminum system at 36°, because of the low rate at which the system attains equilibrium. However, at 790 the constants were found to be 2.37 i .07 (K1) and 2.56 i .09 (K2). Thus the equilibrium constants indicate that the redistribution of acac and bzbz on indium(III) and aluminum(III) is very similar to that ob— served for the analogous gallium(III) system. The dependence of the equilibrium constants on tempera- ture for the Ga(acac)3-Ga(bzbz)3, Al(acac)3-A1(bzbz)3, and Ga(acac)3-Ga(hfac)3 systems was studied in order to deter— mine enthalpy and entropy changes for the exchange reactions. For the Ga(acac)3«Ga(hfac)3 system, the temperature range was limited by the freezing point of the solvent (5.20) and the broadening of Ga(acac)2(bzbz) lines at temperatures above 92, 60°. For the Ga(acac)3-Ga(hfac)3 system, however, 40 equilibrium constants were determined at temperatures as high as 98° using thick—walled nmr tubes. Equilibrium constants for the Al(acac)3-Al(bzbz)3 system at elevated temperatures were determined by heating the solutions for at least 8 hr, quenching with cold tap water, and then re— cording the spectra within 1 hr at 36°. The temperature dependence for the three systems are shown in Tables IV to ’1 VI. Enthalpy and entropy changes for reactions 1 and 2 were determined from the slope and intercept of log K gs. 1/T plots. The data were treated by least squares analysis » including 25 data points at each temperature. The thermo- v dynamic data, along with extrapolated values of K1 and K2 at 25°, are summarized in Table VII. The experimentally observed thermodynamic parameters in Table VII were used to calculate enthalpies and entropies for formation of the mixed ligand complexes from the cor- responding parent complexes, as defined by reactions 3 and 4. These latter thermodynamic data, along with calculated values of K and K at 25°, are presented in Table VIII. f1 f2 Within experimental error, enthalpy changes for formation of Ga(acac)2(bzbz) and Ga(acac)(bzbz)2 are zero or nearly zero and the entropy changes are equal to the values ex- pected for a random statistical scrambling of ligands. The results obtained for Al(acac)(bzbz)2 and Al(acac)2(bzbz) are similar to those obtained for the analogous mixed ligand complexes of gallium(III). For both systems, the 95% con— fidence level estimates of error indicate that either small 41 Table IV. Temperature dependence of K1 and K2 for the Ga(acac)3-Ga(bzbz)3 system in benzene. f.“- Tem .. OC Avera e Valuesa * 3 P /—_—__—_jb g C T\ t 5.7 2.60di .19e 2.49f: .13 '4 16.0 2.77 i .19 2.24 :t .19 J 30.3 2.50 i .27 2.42 i .19 43.0 2.90 i .16 2.27 i .22 aAverage of 25 spectral measurements. b facac 0.340; total molarity is 0.10g_unless otherwise noted. C facac 0.677; total molarity is 0.20g_unless otherwise noted. dTotal molarity is 0.076g, eAll errors are estimated at the 95% confidence level. fTotal molarity is 0.10g, 42 Table V. Temperature dependence of K1 and K2 for the Al(acac)3-Al(bzbz)3 system in benzene. a Tem ., oC /-—-—-—- Avera e Values p b g c \ K1 K2 d 79 2.37 i .07 2.56 i .09 99 2.51 i .10 2.69 i .06 115 2.48 i .08 2.67 i .07 139 2.54 i .07 2.57 i .05 aAverage of 25 spectral measurements. bf = 0.337; total molarity is 0.11M. acac _. C = 0.668; total molarity is 0.20M. acac _. d All errors are estimated at the 95% confidence level. 43 Table VI. Temperature dependence of K1 and K2 fgr the Ga(acac)3-Ga(hfac)3 system in benzene Temp., 0C /——————:;Avgrage Valuesb _2 d 10 K1 10 K2 e 36.0 8.88 i 1.10 . . . 58.5 4.81 i 0.81 . 63.3 . . . 5.54 i .21 78.9 3.98 i .53 3.77 i .51 98.0 3.22 i .33 2.89 i .33 aTotal molarity is 0.25fl, bAverage of 25 spectral measurements. C acac d acac 0.335. 0.671. eAll errors are estimated at the 95% confidence level. 44 .Hm>mH mocmoncoo Rmm mnu um omumEHumm mum muouum HH wH N oo o co m wH N oo o co m HMUHumHumum . I . . I . .I I . I . .I . «momma mo vv m+hw HI N H+Hm VI monHm +Nm.H mm.H+NH H N®.+wh MI mOmeH +No H ImAUmom m0 m N N ®¢.Hmm.H bH.Hmo.o ®O.Hom.m mu.Hbm.N wN.va.o mo.HNH.N In UMUMWMM .I . .I . .I . .I . .I . .I . a Nana mu mu +hm 0 mm +mm OI mo +vm N mm +mo v om +Hm o mmH +mb N In umom mo mHOE\Hmox . mHoE\Hmox 5m .m< .m< «x 5m .m< .mq . HM Emummm /K N coHuommm \//I H GOHuummm \ 6mm um mamucmn CH mmcmnoxm UcmmHH How mumo UHEmcmoofiumLB .HH> mHQmB 45 m: I .mfilr .Hm>mH mochHmcoo Ram may um UmumEHumm mum mnouum HHmm . . . . . . mmsHm> wH N oo o co m wH N oo o co m HmoHumHumum I I I a 0mm; mu m.NHm.OI o.H+m.vI nonbN.+NN.H v.NH H.o w.+ H.¢I «OHXHN.HNH.H In omom we .I . .I . .I . .I . .I . .I . «Sun 2 mp +bm N HN +mH 0 mo +mv N vm +m® N vN +om o no +>N N In omum Hm .I . . . .I . .I . .I . .I . «SB 8 Hm +Hm H vN five O OH +>v N vm +mm N mN +Nm o mHH +mm N In omom m0 mHoE\Hmo# am mHOE\Hmox Hm 5m .2 .3 M 5m .m4 .3 z Emummm //WI v coHuomwm \ //w m coHuommm omN um mcmncmfl CH mmmemEoo ocmmHH UmxHE mo coHumEuow How mumo UHEmcmonHmQB .HHH> mHQNB 46 enthalpy or entropy effects could account for the slight deviations from statistical scrambling. Enthalpy changes for formation of Ga(acac)(hfac)2 and Ga(acac)2(hfac), how— ever, are exothermic by at least 3.3 kcal/mole, and the entropy changes are probably somewhat smaller than the statistical values. Thus the deviations from a random dis- tribution of ligands are due almost entirely to enthalpy effects. Previous studies of ligand exchange reactions between zirconium(IV) (13,21), hafnium(IV) (21), titanium(IV) (14). yttrium(III) (22), aluminum(III) (15), and copper(II) (23) B-diketonates indicate that the mixed ligand complexes are appreciably favored at the expense of the parent complexes whencme ligand contains highly electronegative trifluoro- methyl groups and the other ligand is non-fluorinated. The equilibria lie closer to a random statistical distribution of ligands when the two ligands are both fluorinated or both non-fluorinated. The only thermodynamic data reported is for exchange of acetylacetonate and trifluoroacetylacetonate (tfac) between Zr(acac)4 and Zr(tfac)4 (13) and between Ti(acac)2F2 and Ti(tfac)2F2 (14). In these acac—tfac.ligand exchange systems the equilibrium constants for formation of the mixed diketonate complexes are three to seven times larger than the statistical values, and the deviations from random scrambling are due to entropy effects; enthalpy changes are zero within experimental error. The results reported here are in agreement with the general behavior exhibited by other metal B-diketonate systems. The Ga(acac)3—Ga(hfac)3 47 system, however, is the first system reported in which devia— tions from statistical scrambling are due mainly to enthalpy effects. Favorable enthalpy changes can probably be ex- pected for similar systems in which the equilibrium quoti- ents are two or more orders of magnitude larger than sta— tistical values (15). Using simple electrostatic arguments, Kida (24) has shown that a decrease in average ligand-ligand repulsive energies will always favor the mixed ligand species at the expense of the parent complexes whenever the effective charge on the exchangeable ligands differ. Similar electro- static effects have also been considered by Marcus and Eliezer (25). Pay (20) has pointed out that deviations from statistical behavior for exchange of non-fluorinated and fluorinated diketonates, such as acac and hfac, can be expected to receive some contribution from the difference in effective charge on the donor oxygen atoms of the two ligands, caused by the inductive effect of the fluorine atoms. When the inductive effects of terminal groups on the two ligands are similar, which should be the case for acac and bzbz, little stabilization energy is gained by the mixed complexes, and the equilibrium constants are close to the statistical values. It is of interest to estimate the charge differential on the donor oxygen atoms of acac and hfac that would be necessary to account for the observed stabilization energy for Ga(acac(hfac)2 and Ga(acac)2(hfac). Based on a point 48 charge model, the stabilization energy relative to the parent complexes is given by = ‘13E .' ’23E . AEe1 EGa(dik)2(dik') / Ga(d1k )3 / Ga(d1k)3 where EGa(dik)2(dik‘) is the repulsive energy between oxygen atoms in Ga(dik)2(dik'), and E and E Ga(dik')3 Ga(dik)5 are the oxygen-oxygen repulsive energies in the parent com- plexes. For a regular octahedron: _ 4L5f2 + 1) 16J2 + 2) , {2_ ,2 EGa(dik)2(dik') ' Zr 92 + 2r ee + 2r 9 - ’Ii2f2 + 3; e,2 EGa(dik')3 - 2r E _ (12J2 +Agl e2 Ga(dik)3 _ 2r where r is the mean gallium-oxygen bond distance, and e and e', respectively are the effective charges on the donor oxygen atoms of dik and dik'. It follows that AB = (3J2 + 1) (e _ e.)2 e1 2r This result is identical to that given by Kida for the sta— bilization energy of gingX4Y2 complexes. For an estimated (35) gallium-oxygen distance of 2.03, a difference of 22, 0.1 electron charge units between the oxygen atoms of acac and hfac would be sufficient to attain a stabilization energy of 4 kcal/mole for Ga(acac)(hfac)2 and Ga(acac)2(hfac). Although the value of 0.1 electron charge units is plausible, 49 it probably would not be reasonable to attribute the observed enthalpies for formation of Ga(acac)(hfac)2 and Ga(acac)2(hfac) entirely to lower ligand—ligand repulsive energies in the mixed ligand complexes. Ga(acac)(hfac)2 and Ga(acac)2(hfac) are expected to have larger net dipole moments than the corresponding Ga(acac)n(bzbz)3_n complexes; hence, contributions to the enthalpies of reaction 3 and 4 due to differences in solvation energies of reactants and products should be more important for the former complexes. There is some experimental data available which suggest that differences in solvation energies may be important in stabilizing M(acac)n(hfac)3_n complexes. Based on the data reported by Fortman and Sievers (15), equilibrium constants for formation of Al(acac)(hfac)2 and Al(acac)2(hfac) in chlorobenzene are 40 to 100 times larger than statistical values. The deviations from randomness are appreciably smaller than the deviations reported here for formation of the analogous gallium(III) complexes in benzene solution. If electrostatic effects were primarily responsible for the deviations observed for the gallium(III) complexes, then it would be difficult to explain the smaller deviations for the aluminum(III) complexes in absence of solvation effects. 5 10. 11. 12. 13. 14. 15. REFERENCES G. Calingaert and H. A. Beatty, J. Am. Chem. Soc., 61” 2748 (1939). A. Davison, J. A. McCleverty, E. T. Shawl,and E. J. Wharton, J. Am. Chem. Soc., 82” 830 (1967). R. G. Linck and R. E. Sievers, Inorg. Chem., 5, 806 (1966). K. Moedritzer and J. R. Van Wazer, Inorg. Chem., 3, 268 (1964). J. B. Carmichael and J. Heffel, J. Phys. Chem., 62, 2218 (1965). J. A. Semlyen, C. R. Walker, R. E. Blofeld, and C. S. G. Philips, J. Chem. Soc., 4948 (1964). T. Sekine and D. Dyrssen, J. Inorg. Chem., 26, 2013 (1964). D. w. Scott, J. Am. Chem. Soc., _6_8_, 2294 (1946). G. E. Coates, F. Glocking, and N. D. Huck, J. Chem. Soc.I 4495 (1952). M. L. Delwaulle, M. G. Buisset, M. J. Delhaye, J. Am. Chem. Soc., 14, 5768 (1952). S. R. Gunn and R. H. Sanborn, J. Chem. Phys., 33, 955 1 60). P. D. Zemany and F. P. Price, J. Am. Chem. Soc., 19, 4222 (1948). T. J. Pinnavaia and R. C. Fay, Inorg. Chem., 5” 233 (1966). R. N. Lowry and R. C. Fay, Abstracts, 154th National Meeting of the American Chemical Society, Chicago, 111., Sept. 1967, No. 0, 50. J. J. Fortman and R. E. Sievers, Inorg. Chem., 6, 2022 (1967). 50 16. 17. 17a. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 51 J. C. Lockhart, Chem. Rev., 65, 131 (1965). R. E. Dessy, T. Psarras, and S. Green, Ann. N.Y. Acad. Sci., 125, Art. 1, 43 (1965). J. R. Van Wazer, Am. Scientist, 59, 450 (1962). K. Moedritzer, Adv. Organometal. Chem., 6, 171 (1968). K. Moedritzer, Organometal. Chem. Rev., 2, 179 (1966). R. C. Fay, Ann. N.Y. Acad. Sci., (in press). A. C. Adams and E. M. Larsen, Inorg. Chem., 5, 228 (1966). F. A. Cotton, P. Legzdins, and S. J. Lippard, J. Chem. Phys.,;4_§_I 3461 (1966). ' L. Newman and P. Klotz, Abstracts, 153rd National Meeting of the American Chemical Society, Miami Beach, Fla., April 1967, No. L119. S. Kida, Bull. Chem. Soc. Japan, 34, 962 (1961). Y. Marcus, and I. Eliezer, J. Phys. Chem., 66, 1661 (1962). G. T. Morgan and H. D. K. Drew, J. Chem. Soc., 119, 1058 (1921). R. C. Fay and T. S. Piper, J. Am. Chem. Soc., _5, 500 (1963). J. A. S. Smith and E. J. Wilkins, J. Chem. Soc., 1749 (1966). R. C. Young, Inorg. Syntheses, g, 25 (1946). H. Funk and A. Paul, Z. Anorg. Allg. Chem., 330, 70 (1964). M. L. Morris, R. W. Moshier, and R. E. Sievers, Inorg. Syntheses, g, 28 (1967). K. Nakamoto, P. J. McCarthy, and A. E. Martell, Nature, 183, 459 (1959). S. I. Weissman, J. Chem. Phys., 18, 1258 (1950). H. S. Gutowsky and C. H. Holm, J. Chem. Phys., 25, 1228 (1956). E. C. Lingafelter and R. L. Brown, J. Am. Chem. Soc., 88, 2951 (1966) . ‘—