PREPARATIEQN Q? SQME' NEW GERMANWE‘xé (N) B eBEKETGNATEfi ANQ A NEJCLEAE MAGNE‘FEC RESQNANCE 511be OF LEGAND EXCHANGE REACHGNS FQEE $Q¢M§ Rift-6 BeQE‘KETGNATG} TEYANIUMfiV) COMHEXES' ( r 'o Emmi {or fin anmn 02‘; M. 5. éiECBESAN STATE UH’WERSZTY Y‘ A. ?e€era 2969 mm 51m umvmsm EAST umsmc. Micrgncm 1ng!“ _.- ‘ LIBRARY I Michigan State ; University , I r_ Mia-mars 31m :: ‘1‘:€;:TY' :\ -".' ;‘ "."ukl ; a .' -- L!” [..:‘J!".\) i 11.; L.‘£E-'Z'r"\.'”\.r\lY Cl {Eix'xiS’E H ‘z’ [JUILDENG ABSTRACT PREPARATION OF SOME NEW GE RMANIUM(IV) B-DIKETONATES AND A NUCLEAR MAGNETIC RESONANCE STUDY OF LIGAND EXCHANGE REACTIONS FOR SOME TRIS(B-DIKETONATO)TITANIUM(IV) COMPLEXES BY Y. A. Peters Two new dihalobis(fi-diketonato)germanium complexes were prepared. They are: gig: and Ergnnge(dpm)2C12, where dpm is dipivaloylmethanate. This is the first trans isomer of the type M(dik)2X2, where M is tita- nium(IV), germanium(IV), or tin(IV) and X is a halogen, to be reported. The isomerization of Erggnge(dpm)2Clz in benzene is first order and reversible: k1 trans—Ge(dpm)2C12 :7——> Cis-Ge(dpm)2C12 2 at 440 k1 - 9.8 x 10‘3 min-1 and k2 = 4.2 x 10‘4 min-1. Dihalobis(B-diketonato)germanium complexes react with antimony(V) chloride in glacial acetic acid to give tris- (fi-diketonato)germanium(IV) hexachloroantimonate(v). Two new complexes Were prepared in this manner, [Ge(dpm)3][SbC16], and [Ge(bzbz)3][SbC16], where bzbz is dibenzoylmethanate. The reaction of ginge(dpm)2C12 with antimony(V) chloride in methylene chloride was found to be slow enough to isolate a reaction intermediate. The spectral Y. A. Peters observations and the analysis indicate that this inter— mediate is the adduct, [Ge(dpm)2C12,SbC15]. Attempts to observe ligand exchange reactions between tris(acetylacetonato)germanium(IV), [Ge(acac)3]+, and 1+ [Ge(dpm)3]+ or [Ge(bzbz)3 in dichloromethane solution were unsuccessful. However, the three titanium(IV) sys- tems [Ti(acac)3]+ - [Ti(dpm)3]+, [Ti(acac)3]+ - [Ti(bzbz)3]T and [Ti(dpm)3]+ - [Ti(bzbz)3]+ undergo facile ligand exchange reactions in methylene chloride. For the [Ti(acac)3]+ - [Ti(bzbz)3]+ system, which was studied in greatest detail, equilibrium constants for the formation of each mixed ligand complex from the parent complexes at 25° are equal to the values expected for a random statistical distribution of ligands. Enthalpy and entropy changes for the formation of [Ti(acac)(bzbz)2]+ and [Ti(acac)2(bzbz)]+ are, resPectively, -0.10 i 0.17 kcal/mole; 1.83 i 0.40 eu and -0.04 i 0.19 kcal/mole; 2.04 i 0.39 eu. A random scrambling of ligands predicts AH = 0.0 kcal/mole and AS = 2.18 eu. for the formation of each mixed ligand complex. PREPARATION OF SOME NEW GERMANIUM(Iv) B—DIKETONATES AND A NUCLEAR MAGNETIC RESONANCE STUDY OF LIGAND EXCHANGE REACTIONS FOR SOME TRIS(B-DIKETONATO)TITANIUM(IV) COMPLEXES BY . PM :. ( , .- \ ‘ r "‘ YT A; Peters A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1969 To Jim and Mary ii ACKNOWLEDGMENTS I wish to thank Dr. Thomas J. Pinnavaia, who sug- gested the research reported here, for his guidance and understanding during the course of this work. Thanks are also due to Dr. H. A. Eick, who was my academic advisor. iii TABLE OF CONTENTS Page I. INTRODUCTION . . . . . . . . . . . . . . . . . . 1 II. EXPERIMENTAL . . . . . . . . . . . . . . . . . . 20 A. Reagents and General Techniques . . . . . . . 20 B. Preparations . . . . . . . . . . . . . . . . 22 1. Cis-Dichlorobis(2,4-pentanedionato)- germanium(IV) . . . . . . . . . . . . . 23 2. cis-Dichlorobis(2,2,6,6-tetramethyl- heptanedionato)germanium(IV) . . . . . . 24 3. trans—Dichlorobis(2,2,6,6-tetramethyl- heptanedionato)germanium(IV) . . . . . . 25 4. cis—Dichlotobis(1,3-diphenyl-1,3- propanedionato)germanium(IV) . . . . . . 25 5. Attempted preparation of dichlorobis- (1,1,1,5,5,5-hexafluoro-2,4-pentane- dionato)germanium(IV) . . . . . . . . . 26 6. Tris(2,4-propanedionato)germanium(IV) hexachloroantimonate(v) . . . . . . . . 27 7. Mixed complex of dichlorobis(2,2,6,6- tetramethylheptanedionato)germanium(IV) with antimony(V) chloride . . . . . . . 28 8. Tris(2,2,6,6-tetramethylheptanedionato)- germanium(IV) hexachloroantimonate(v). . 29 9. Tris(l,3-diphenyl-1,3—propanedionato)- germanium(IV) hexachloroantimonate(v) . 29 10. Titanium Complexes . . . . . . . . . . . 30 C. Nuclear Magnetic Resonance Spectra . . . . . 30 III. RESULTS AND DISCUSSION . . . . . . . . . . . . . 32 A. Synthesis and Observations of Some New Germanium B-diketonates . . . . . . . . . . . 32 B. Isomerization of Egggnge(dpm)2C12 . . . . . 54 iv TABLE OF CONTENTS (Cont.) Page C. Exchange Reactions . . . . . . . . . . . . . 62 D. Equilibrium+Studies of [Ti(acac)3]+ - [Ti(bzbz)3] mixture . . . . . . . . . . . . 80 IV. BIBLIOGRAPHY . . . . . . . . . . . . . . . . . 86 TABLE II. III. IV. VI. VII. VIII. IX. LIST OF TABLES Nomenclature for fi-diketones used as ligands Reaction times for several M(dik)2C12 complexes Chemical shifts for germanium and titanium B-diketonates o o o o o o o o o o o o o o o 0 Kinetic data for isomerization of trans- Ge(dpm)2C12 in benzene at 43.80 . . . . . . . Equilibrium constant data for k trans—Ge(dpm)2Clz‘:L—> cis-Ge(dpm)3C12 . . . __——_. - k2 Dependence of equi ibrium constants for the system [Ti(acac)3] - [Ti(bzbz)3] on total solute molarity . . . . . . . . . . . . . . . Dependence of equilibrium constants on the ligand composition for the [Ti(acac)3] - [Ti(bzbz)3] system . . . . . . . . . . . . . Temperature depende ce of equilibr'um constants for the [Ti(acac)3] — [Ti(bzbz)3] system . . Thermodynamic data for formation of mixed ligand complexes in methylene chloride at 25°. vi Page 12 33 48 58 59 82 83 84 84 Figure 1. 8. 9. 10. 11. 12 0. LIST OF FIGURES Intermolecular ligand dissociation mechanism for the isomerization and racemization of [(+)-CiS-C0(en)2(H20)2] o o o o o o o o o o ~Intramolecular bond rupture mechanism for the linkage isomerization of tris(acetylacetyl- acetonato)cobalt(III) . . . . . . . . . . . Intramolecular trigonal twist mechanism for the racemization of [Co(EDTA)]' with the participa- tion of solvent water . . . . . . . . . . . Intramolecular rhombic twist mechanism for the racemization of [Co(bigH)3]+3 . . . . . . . Representation of the psuedo C3 axes of [Ge(acac)(dpm)(hfac)]+ . . . . . . . . . . . Page 14 Examples of rotation about various psuedo C3 axes of [Ge(acac)(dpm)(hfac)]+ . . . . . . . . cis-Ge(dpm)2C12 . . . . . . . . . . . . . . EfButyl proton resonance lines (60 MHz) of 0.4M_cis-Ge(dpm)2C12 in methylene chloride at 44.00 . . . . . . . . . . . . . . . . . . . tranS‘Ge(dpm)2Cla o o o . o . . o o o . o o tyButyl proton resonance line 60 MHz) of a saturated solution of trans-Ge dpm)3Cla in rela- 16 36 38 41 tion to TMS at 500 sweep width in benzene d:44.0° 43 The cation [Ge(dpm)3]+ . . . . . . . . . . . The Chan e in the Efbutyl proton resonance spectra 60 MHz) with time for 0.20M; [Ge(dpm)2C12,SbC15] in methylene chloride at 33.00 . . . . . . . . . . . . . . . . . . . vii 49 52 .3 hi LIST OF FIGURES (Cont.) Figure 13. 14. 15. 16. 17. 18. 19. Page Plot of log [ftrans] y§_time for the Isomeri- zation of trans-GeIdpm)2C12 in benzene at 43.80 56 Plot of log [ftrans] - (1—ftrans for the isomerization of trans-Ge(dpm)2C12 in benzene at 43.80 . . . . . . . . . . . . . . . 60 )/Keq] vg time Changes in relative signal intensities as a function of ligand composition for the acetyl- acetonate ring proton spectra (60 MHz) of 0.3M_ [Ti(acac)3] - [Ti(bzbz)3] at equilibrium in methylene chloride at 33.00 . . . . . . . . . 66 Proton resonance line spectrum (60 MHz) of an equilibrium mixture of 0.30M [Ti(acac)3] - [Ti(dpm)3] , f = 0.67 in methylene chloride acac at 33.00 . . . . . . . . . . . . . . . . . . . 69 Proton resonance lines (100 MHz) for an equi- librium mixture of 0.30M_[Ti(acac)]+ - [Ti(dpm)3]+, facac = 0.50, in methylene chloride at room temperature . . . . . . . . . . . . . 72 Ring proton and Efbutyl nmr lines (60 MHz) of an equilibrium mixture of 0.30M'[Ti(dpm)3]+ - [Ti(bzbz)3]+, f = 0.50, in methylene chloride d m at 33.00 . . . .p. . . . . . . . . . . . . . . . 76 Ring proton and methyl proton nmr lines (60 MHz for an e uilibrium mixture of 0.40M.[Ti(acac)3] - [Ti(bzbz?a]+, f = 0.50, in methylene chloride at 33.00 . . écéc. . . . . . . . . . . . . . . 78 viii IN TRODUCT ION First order racemizations or configurational rearrange- ments of octahedral complexes may occur gig three possible routes, ligand dissociation (intermolecular), bond rupture (intramolecular, chelating ligand), and twists (intra— molecular). These three possibilities are discussed in detail by Basolo and Pearson (1). Intermolecular ligand dissociation requires the complete departure of a coordinated ligand and its random mixing with any free ligand in solu- tion. As a result the ligand exchange must be as fast or faster than the configurational changes. An example of an intermolecular ligand dissociation process is the isomeri- zation and subsequent racemization of the gig-diaquobiSP (ethylenediamine)cobalt(III) cation. This is illustrated in Figure 1. In general, metal compleXes containing easily replaceable monodentate ligands appear to change configura— tion by an intermolecular process, whereas, complexes con- taining only polydentate ligands usually change configura- tion by an intramolecular process; either bond rupture or twisting. Intramolecular mechansims require that there be no random mixing of coordinated ligand with free ligand. As a result the ligand exchange must be as slow or slower than the configurational changes. The bond rupture process 1 .m ANAOnmvnAGTVOUIMHUIA+VH mo coaumNHEmomH Cam COflumNHumEOmH may now amasmzowa COADmHUOmMHC osmmfla HMHDOUHOEHUDCH .H wndmfim To... 2:38....011m For: £55-22; 78.3 .23 8-20-: Cg. n+ . n... . n... m._ 3 - . 7 ~ m; E . I N If I mlmv n.~ 3 Nina, om: _ «.2: _ o :l..v1.l~ all _ ez _\ of... o I- :5 of: d 0...: im is. I N A} if C66 the a me involves an opening-Closing mechanism of a chelate ring. It has been shown to be the mechanism for the linkage isomerization which occurs in tris(acetylaceylacetonato)— cobalt(III), Figure 2. This type of linkage isomerization, if accomplished in an intramolecular fashion, can only pro- ceed through bond rupture. The twisting mechanisms require that no bonds be broken during the deformation. Two basic types of twist mechanisms have been proposed, the rhombic and the trigonal twist. The ethylenediaminetetraaceto- cobaltate(III) anion is believed to racemize gig a trigonal twist with the participation of solvent water, Figure 3. The rhombic twist has been proposed for the racemization of trichlorotris(biguanidinium)cobalt(III), Figure 4. The original objective of this research was to test the merits of the various twist mechanisms for the racemi- zation of a mixed fi—diketonate complex, M(dik)(dik')(dik"). Previous studies of octahedral metal B—diketonates have shown that an intramolecular path is favored in the isomeri- zation and racemization of metal B-diketonates (2-7). The metal moiety of the complexes studied was aluminum, gallium, indium, cobalt, rhodium, chromium, andotitangum. The B— diketonate ligands, represented as R-C-CHz-C-R', were both the symmetric variety, R - R', and the unsymmetric variety R #‘R'. The terminal groups R or R' represented methyl, trifluoromethyl or phenyl groups. All compounds were neutral tris(6-diketonato)metal complexes except for those with titanium as the metal moiety, in which case the complexes .AHHHVpamnooAoumcoumomamumom§fimomvmfiuu mo GOHDMNAHOEOMA wmmxsfla Tau How Emflsmnowfi mndumsu Uson HmasomaoamuusH .N musmflm .7 .. .HTDMB usm>aom mo coaummflowunmm may nufl3 HA Ge(dik)2C12 + 2HCl (g). b. 3Ge(dik)2c12 + 25bc15 > 2[Ge(dik)3][SbC16] + GeCl4. Most reactions were found to give more than one crop of crystals and the chemical analyses of products were per- formed by Galbraith Laboratories, Inc., Knoxville, Ten— nessee. All the bis B-diketonates were prepared gig reaction "a" and employed analogous techniques. A side arm flask equipped with a reflux condenser and a phosphorus pentoxide drying tube was placed on a hot plate-magnetic stirrer. 23 Nitrogen was passed through the side arm of the reaction flask containing the previously purified chloroform and the germanium(IV) chloride was added. The solution was stirred, and then the appropriate ligand was added. The resulting solution was refluxed and stirred until hydrion paper indi- cated that hydrogen chloride gas was no longer being evolved. The reaction was usually refluxed for an additional half hour. Crystallizations and filtrations were performed in an inert atmosphere and all products were dried in vacuo at room temperature. The tris complexes were made yia_reaction "b". Anti- mony pentachloride was added directly to a side arm flask which contained the bis complex and solvent. Mixing and refluxing when required were accomplished with a hot plate- magnetic stirrer. Glass systems were swept with nitrogen during the reaction. Crystallization and subsequent handling were done in a nitrogen atmosphere to prevent hydrolysis. All samples were dried in vacuo for two hours at room tem- perature. 1. cis-Dichlorobis(2,4-pentanedionato)germanium(IV). This compound was prepared in the manner described by Ong and Prince (20). Germanium tetrachloride (3.9 ml, 34.5 mmoles) was added to 25 ml of chloroform. Acetylacetone (18.0 ml, 18.0 mmoles) was then added to the germanium(IV) chloride solution at room temperature. The mixture was refluxed for three hours and was subsequently cooled in a 24 dry ice-acetone bath for 0.5 hour. The white crystalline product was filtered under nitrogen, washed with ten ml of cold hexane and dried. The yield was 74%, and the melting point was 228—232°, lit. 236—400 (20) and 238-244° (26). The compound was used in the preparation of the tris complex without further purification. 2. cis-Qichlorobis(2,2,6,6-tetramethylheptanedionato)- germanium(IV). Dipivaloylmethane (11.0 ml, 56.6 mmoles) was added to a solution of germanium tetrachloride (3.2 ml, 28.3 mmoles) in thirty ml of chloroform. The resulting solution was refluxed for 119 hours. The reaction vessel was period— ically swept with nitrogen gas. The compound was found to be very soluble in most solvents. The product was recovered when the volume of chloroform was reduced to half, then hexane was added until a slight turbidity in the warm solu- tion was noted. The flask was then placed in a dry ice- acetone bath for 1.5 hours, after which the fine white crystals were filtered and dried. The yield was 40.2% and the melting point was 135.8-137.8°. Anal. CEICd. for Ge(C11H1902)2C12: C, 51.79; H, 7.52; Cl, 13.90; Ge, 14.23. Found: C, 51.71, H, 7.73; Cl, 14.49; Ge, 14.65. 25 3. trans-Dichlorobis(2,2,6,6—tetramethylheptanedionato)- germanium(IV). The trans isomer of Ge(dpm)2C12 was recovered from the reaction mixture after slow crystallization of the gig isomer from methylene chloride—hexane or benzene—hexane solutions. This isomer was also obtained by digestion of the gig isomer in methylene chloride-hexane or benzene- hexane at room temperature. The melting point was 139—1400. 4. cis-Dichlorobis(1,3-diphenyl-1,3-propanedionato)- germanium(IV) Dibenzoylmethane (13.3 g, 59.2 mmoles) was dissolved in 100 ml of benzene, and then germanium(IV) chloride (3.4 m1, 29.6 mmoles) was added. The solution was refluxed for 110 hours and the system was periodically flushed with nitrogen. The volume was reduced to one-third by passage of nitrogen over the warm solution, at which point a pre— cipitate was formed. The crude product weighed 9.6 g (55% yield) and had a melting point of 259-269°. Recrystalliza- tion was attempted from methylene chloride, but after the product was heated in 220 ml of solvent much of the solid remained. The undissolved yellow powder was filtered and then dried. It weighed 5.78 g (33.3% yield) and had a melt- ing point of 274—278°. Osipov (21) has prepared this compound in a similar manner. Carbon tetrachloride was used as the reaction sol- vent and the reaction time was about 40 hours. The reported 26 melting point was 258° which corresponded to the lower limit of the melting point of the crude product described above. Anal. Calcd. for GE(C15H1102)2C123 C, 60.07; H, 3.30. Found: C, 59.68; H, 3.54. 5. Attempted Preparation of Dichlorobis(1,1,1,5,5,5—hexa- fluoro-2,4:pentanedionato)germanium(IV). All attempts to make Ge(hfac)2C12 were unsuccessful. Neither the method of preparation used for the other bis complexes nor the addition of germanium tetrachloride to 1,1,1,5,5,5fhexafluoro-2,4-pentanedione in the absence of solvent gave a discernible reaction after 120 hours. Fernelius and Bryant (27) have made several B-diketo— nate complexes by reaction of the metal chloride with the sodium salt of the desired ligand: 2Na + 2H(dik) > 2Na(dik) + H2(g). MCI4 + 2Na(dik) > M(dik)2cl2 + 2NaCl. Germanium tetrachloride (3.2 ml, 27.2 mmoles) was added to sodium hexafluoroacetylacetonate (12.4 g, 54.4 mmoles) in dioxane and the turbidity of the solution increased. After twelve hours the supernatant was decanted. Crystallization was attempted from dioxane without success. The solvent was then pumped off and a gritty brown and white residue remained. This residue was taken up in 40 ml of methylene chloride and gave a clear dark brown solution. Upon addi- tion of 95 ml of hexane the solution became cloudy. After 27 the flask cooled in the freezer overnight, the off-white precipitate was filtered and dried in vacuo. The light red- brown mother liquor was evaporated to dryness under vacuum to yield a very small amount of an off-white residue. The two residues were combined and placed in a vacuum sublimer. Sublimation at 80° gave a brown oily sublimate and a brown clay-like residue. Examination of the liquid nitrogen trap showed a light yellow oil had frozen out at the top of the trap and a colorless liquid was present in the bottom of the trap. Since none of the four components were sensitive to air, it was concluded that none of them were Ge(hfac)2C12. Of the three methods tried the most promising was the addi- tion of germanium tetrachloride to sodium hexafluoroacetyl- acetonate. 6. Tris(zyg-propanedionato)germanium(IV)hexachloro- antimonate(V). This compound was first prepared by Cox, Lewis, and Nyholm (19). The reaction solvent was Changed from chloro- form to methylene chloride. Ge(acac)2C12 (3.6 g, 10.5 mmoles) was added to 120 ml of methylene chloride, and then antimony pentachloride (1.9 ml, 10.5 mmoles) was added. The Ge(acac)2C12 disappeared immediately and the yellow solu- tion was reduced to twenty ml by passing nitrogen over the surface of the solution at room temperature. Crystalliza- tion was accomplished by addition of an equal volume of ether to the solution, which was subsequently cooled in a 28 dry ice-acetone bath. The product was washed with 10 ml of hexane and recrystallized once. A 50% yield with a melt— ing point of 165-167° was obtained, reported; 34%, 165-167° (23). 7. Mixed Complex of Dichlorobis(2,2,6,6-tetramethylheptane- dionato)germanium(IV)With Antimonngentachloride. Ge(dpm)2C12 (1.43 g, 2.82 mmoles) was dissolved in 100 ml of methylene chloride. Then antimony pentachloride (0.36 ml, 2.82 mmoles) was added to the solution at room temperature. The orange solution was evaporated to 15 ml under a stream of nitrogen then 25 ml of ether was added. After 0.5 hour the white needle crystals were collected and washed with 5 ml of hexane. The crystals were dried in the vacuum line for two hours, and then in an Abderhalden ap- paratus for four hours, at 80°. The yield was 80.5% and the melting point was 232-234° . Anal. Calcd. for Ge(C11H1902)2C12'SbC15: C, 32.65; H, 4.74; Cl, 30.67. Found: C, 32.85; H, 4.95; Cl, 28.40. The preparation was repeated and sent for analyses. Anal. Found: C, 32.36; H, 5.17; Cl, 27.53. Unfortunately, Galbraith was not notified that the com— pound was air sensitive. It is felt that this contributed to the poor correlation of calculated versus actual composi- tion. 29 8. Tris(2, 2, 6, 6— —tetramethylhep_anedionatolgermanium(IV) hexachloroantimonate(V). [Ge(dpm)3][SbC16] was obtained by reaction of Ge(dpm)2C12 and antimony(V) chloride in glacial acetic acid. Ge(dpm)2C12 (2.8 g, 5.5 mmoles) was dissolved in 40 ml of glacial acetic acid, then antimony pentachloride (0.71 ml, 5.50 mmoles) dissolved in ten ml of glacial acetic acid was added. A fine white precipitate was formed almost immediately. The {—U mixture was refluxed for two hours, then filtered and dried. Recrystallization was accomplished by dissolution of the product in the minimum amount of methylene chloride and addi- tion of hot hexane. The flask was then cooled in the freezer k overnight and the yellow, needlelike crystals were filtered and dried. The yield was 46% and the melting point was 275.5-277.5°(d). Anal. Calcd. for [Ge(C11H1902)3][SbC16]: C, 41.41; H, 6.02; Cl, 22.23. Found: C, 45.49; H, 6.66; Cl, 20.25. 9. Tris(l,3-diphenyl-1,3-propanedionato)germanium(IV) hexachloroantimonate(V). Ge(bzbz)2C12 (5.1 g, 8.7 mmoles) was dissolved in 200 ml of glacial acetic acid by gentle heating and stirring. The antimony(V) Chloride (1.4 ml, 8.7 mmoles) was added, at which point the solution became bright yellow, and a pre— cipitate formed. The mixture was allowed to reflux for two hours. The volume of the solution was reduced to one-third 30 and cooled, at which point a precipitate formed. The pre- cipitate was filtered, dried, and then dissolved in 100 ml of methylene chloride. The volume was reduced to half and ten ml of hexane was added. The mixture was cooled over— night in the freezer. The prismatic crystals were then filtered and dried. The yield was 79.8% and the melting point was 221—223°. Anal. Calcd. for [Ge(C15H1102)3][SbC16]: C, 50.19; H, 3.10; Cl, 19.75. Found: C, 49.22; H, 2.97; Cl, 20.25. 10. Titanium Complexes. The titanium analogs of the above germanium complexes were prepared in a similar manner by Luis Matienzo (24). C. Nuclear Magnetic Resonance Spectra Nuclear magnetic resonance spectra were obtained with a Varian A-60 analytical spectrometer (60.00 MHz) fitted with a Varian variable temperature controller, Model V-6040. The chemical shifts of methanol and ethylene glycol were used to determine the low and high temperatures respectively. A calibration standard containing seven compounds was used to correct the magnetic sweep width at 250 Hz and 500 Hz. This standard was also employed for chemical shift determin- ations of several new compounds. In this case calibrations of magnetic sweep widths were determined for sweep widths of 500, 250, and 50 Hz. 31 Signal areas used in the determination of equilibrium constants were measured by planimetry. In general, eight to ten spectra were measured for each equilibrium constant to minimize the error due to minor variations in the indi— vidual spectrum. Spectra were run at various radiofrequency fields and examined to avoid errors caused by saturation effects. The optimum radiofrequency at room temperature was found to have an amplitude of 0.10 mG. Signal areas used in the determination of the rate of isomerization of trans-Ge(dpm)§C12 were measured by elec- tronic integration. In general, three spectra were measured for the concentration of trans and gi§_isomer at half hour intervals for four hours. To insure the establishment of equilibrium, subsequent measurements were taken at six and fifteen hours. III. RESULTS AND DISCUSSION A. Synthesis and Observations of Some New Germanium B-diketonates The investigation of germanium complexes led to the prep— aration of four new compounds. They are: gig— and t_ra__n_s- Ge(dpm)2Clz, [Ge(dpm)3][SbC16], and [Ge(bzbz)3][SbC16]. Originally, the object of the preparative phase of this re- search was to prepare a mixed ligand complex from three different tris(B-diketonato)germanium(IV) complexes. These complexes are not formed directly from the metal halide and B—diketones. Their preparation requires the following sequence of reactions: a. GeCl4 + 2Hdik -—9 Ge(dik)2C12 + 2HCl(g). b. 3Ge(dik)2Clz + 2SbCl5 ——> 2[Ge(dik)3][SbC16] + GeCl4. As in the case of germanium, the reactions of titanium or tin tetrachloride with various B-diketones give the di- chlorobis(B-diketonato)metal complexes. However, tin and titanium B-diketonates usually take less than two hours to prepare. The reaction, in the first attempt to prepare §i§76e(dpm)2clg in methylene chloride, was allowed to reflux for eight hours. No product was recovered. The gig- Ge(dpm)2clz was successfully prepared in methylene chloride 32 33 when the reaction was allowed to reflux for approximately 120 hours. It was noted that the time required to prepare the dihalobis(B-diketonato)germanium(IV) complexes varied considerably as compared to their titanium and tin analogs. In Table II are tabulated reaction times and reaction sol- vents for several Metal(IVXdik)2C12 complexes. In all cases, the preparation involved the reaction of the metal chloride with the free ligand. The yields for most reactions were from 75% to 90%. Table II. Reaction times for several M(dik)2C12 complexes. M = Germanium M - Titanium Ligand ngz' Solvent . T532, Solvent Hacac 2 CHCl3 . 0.33 CHC135 Hetaca 8 CHC1317 .Hbzac 10 CHC1321 . 0.25 C6H66 Hpivacb 110 CHC1329 1 CHC1329 Hbzbz 110 CeHe 0.25 C6H66 Hdpm 119 CHCl3 2 CH3C02H29 o O aCH3CH25CH25CH3 0 Q b(CH3)3CCH2CCH3 Although Table II indicates that there is some variation in the polarity of the solvents used in the reactions of a given ligand, the longer reaction times for the germanium complexes 34 indicate that other factors may be involved. Table II lists no tin complexes, although Sn(acac)2C12 (17), Sn(bzac)2C12 (30), and Sn(bzbz)2Cl2 (30) have been prepared in methylene chloride. The preparation of Sn(acac)2Clz required less than an hour, but no times were reported for Sn(bzac)2C12 or Sn(bzbz)2C12. The ionic radii of germanium(IV), titanium(IV) and tin(IV) are 0.44 R, 0.64 R, and 0.74 8 respectively (31). Because germanium(IV) has the smallest radius it would be the most sensitive to steric factors. The ligands in Table II are listed in increasing order of size of the terminal R groups. The reaction times for the germanium compounds increase as the steric hindrance increases. The reluctance of germanium(IV) chloride to form etherates has been noted by Udovenko and Fialkov (18). The explanation given was that since germanium(IV) has the smallest radius of the three tetrachlorides, it is the most screened and therefore the slowest to react. The calcu— lated ratio of metal to chloride ionic radii are germanium, 0.24; titanium, 0.37; and tin, 0.41. Since both etherate and dihalobis formation require the formation of a metal— oxygen bond, the screening factor for germanium could also contribute to the difference in the time required for the preparation of complexes which differ only in their metal moiety. Both the screening factor and ionic radii of germanium(IV) chloride should affect the reaction rate if the mechanism 35 is SN2° This mechanism requires nucleophilic attack and subsequent formation of a five-coordinate intermediate. The 8N1 mechanism requires bond rupture and subsequent forma— tion of a three coordinate intermediate. This mechanism is influenced by the bond energy of the species which must dis— sociate, in this instance a metal-chlorine bond. Mean thermochemical bond energies in kcal/mole for the three metal chlorides in question are: germanium, 104 (32), tin, 76 (33), and titanium,o48 (34). An increase in reaction time with increase in bond energy is observed. These argu- ments predict that either an 5N1 or 3N2 mechanism would lead to lower reaction rates for germanium. In addition to the small ionic radius, large screening factor, and high bond energy of germanium(IV) chloride, all complexes are octahedral and the hybridization of germanium to 4sp3d2 requires more energy than the hybridization of titanium to 3d24Sp3. Since tin has the largest ionic radius, but also requires hybridization of outer d orbitals the reaction times for tin are expected to be somewhere between those of titanium and germanium. There are two possible isomers for the octahedral Ge(dpm)2C12, gig and trans. Figure 7 shows the structure of the gi§_isomer. Since this isomer has one two fold axis of rotation there are two nonequivalent sets of Efbutyl groups which give rise to a doublet in the tfbutyl region bf the nmr spectrum as shown in Figure 8. The peak heights of the cis isomer doublet in Figure 8 are not exactly equal 36 Figure 7. cis—Ge(dpm)2Clz. 38 Figure 8. 'E-Butyl proton resonance lines (60 MHz) of 0.4M_cis-Ge(dpm)2Clz in methylene chloride at 44.00. 78.85 18.74 40 due to the presence of some of the trans isomer. The Efbutyl resonance line for the trans isomer occurs at the same position as the downfield resonance line in the doublet for the gi§_isomer. The small peak between the two lines of the gi§_doublet is attributed to free ligand whose E- butyl resonance line occurs at approximately I 8.8. This small peak at T 8.8 appears after the solution has aged a few hours. The solvent for this spectrum was methylene chloride. Figure 9 shows the structure of the tgang isomer. In this isomer all Efbutyl groups are equivalent and the nmr spectrum gives rise to a singlet in the Efbutyl region. Figure 10 shows the singlet of the t£2n§.isomer in relation to tetramethylsilane for a saturated solution at 44°, taken in benzene where E£§n§12i§ isomerization is slow. The solid germanium complexes react slowly with atmo- spheric moisture. A few crystals of Ge(dpm)2C12 were ex- posed to the atmosphere and after twelve hours the melting point had dropped ten degrees, although no deterioration of the crystals was detected visually. After three days, the crystals were no longer hard and the compound only partially melted leaving a white residue. After eight days the crys— tals had changed to a white, wet powder. This powder parti— ally melted twenty degrees below the original melting point and left a white residue, probably germanium dioxide. The ginge(dpm)2Clz complex is very soluble in chlor- inated hydrocarbons but only slightly soluble in hexane. If a solution of cis-Ge(dpm)2Clz is allowed to stand for more 41 Figure 9. trans-Ge(dpm)2C12. 43 Figure 10. EfButyl proton resonance line (60 MHz) of a saturated solution of trans- Ge(dpm)2C12 in relation to TMS at 500 sweep width in benzene at 44.0°. TMS I! ‘. 1 1 ' T 1'88 1? I 0.0 45 than a few hours in a capped nmr tube, the solution develops a faint yellow color and the spectrum shows a weak peak between the two peaks of the doublet in the tfbutyl region. This peak corresponds to the chemical shift of the Efbutyl resonance line for the free ligand. Since freshly prepared samples of Ge(dpm)3clz did not show the spurious peak, the possibility that this weak peak was due to the E£§n§_complex and not decomposition was considered. Samples were dis- solved in gfdichlorobenzene and heated at 100°, 180°, and 210° for periods of fifteen minutes to twenty-four hours. The spectra showed that the area of the spurious peak re- mained constant within experimental error. The subsequent ‘4 -—m- wr- preparation of the gals isomer ruled out the assignment of this spurious peak as the resonance line for the trans isomer. Apparently, in pfdichlorobenzene an equilibrium is established between the complex and the dissociation products, one of which is the free ligand. The ££§n§_isomer of Ge(dpm)2Cla can be prepared by either slow crystallization of the Ge(dpm)2C12 reaction mixture or by digestion of the gig isomer. In either method the solvent pairs methylene chloride-hexane or benzene-hexane work well. In methylene chloride the Egggg isomer required about fifteen minutes to isomerize which was also the re- quired time for dissolution of the sample used, (10 mg in 0.4 ml). In benzene the isomerization was much slower and an equilibrium constant, K ' [gigj/[EEEQ§J, of 4.6 was eq found for this process. The rate of trans-cis isomerization 46 in benzene will be discussed in greater detail later. Steric and electronic factors predict that the trans isomer of M(dik)2X2, where }( is a halogen and M is metal, is favored. .Except for trans—Ge(dpm)2C12 reported here, no other trans isomers of the type M(dik)2C12 have been re- ported either in the solid state Or in solution. x-ray studies have been done on M(acac)2X2, where M is titanium or germanium and X is fluorine, chlorine, or bromine (35). IR studies of the solids have also been done for the above compounds and also the tin analogs (36). Nmr techniques have been used to study the geometry of these complexes in solution (6,26,37,38). In all cases, the M(dik)2X2 com- plexes have been found to possess gig symmetry. To explain this fact two opposing arguments have been put forward. Both invoke electronic effects as the cause of the stabili— zation of the ggg_configuration. In both arguments, the .g;g_stabilization is assumed to arise from the fact that in the gig isomer all three tzg orbitals can participate in p -> d w-bonding, whereas, in the trans isomer only two of the orbitals can participate. .In dispute, however, is whether the halogen or oxygen atoms are more important in p -> d w-bonding. From infrared spectral data, Nelson (39) has concluded that halogen-metal p -> d v-bonding is more important than the oxygen-metal p-—> d w-bonding in. Sn(acac)2C12 complexes. From nmr and IR data Bradley (39) concluded that electron donation of the oxygen atoms to the metal is the more important factor in Ti(dik)2X2 complexes. This latter View is also held by Fay (37). 47 Another unusual property of germanium complexes was also noted when an attempt was made to find the coalescence temperature of Ge(dpm)2Clz. At room temperature this com- plex shows a doublet, which indicates there is Slow terminal group exchange. This observation is also true for Ge(acac)2C12 (26) and Sn(acac)2C12 (40), but not Ti(acac)2C12 (5). The coalescence temperature for Ti(acac)2C12 (5) and Sn(acac)2C12 (40) are -50° and 90°, respectively, but neither Ge(acac)2Cl2 (41) nor Ge(dpm)2Clz coalesce below 180°. The spectrum of 0.67M_Ge(dpm)zclz in gfdichloro— benzene indicated that the complex may be close to coales- cence at 180° since the lines of the doublet had begun to broaden and the peak separation had decreased from 7.9 Hz at room temperature to 5.6 Hz at 180°. The chemical shifts of some tris(5-diketonates) and gggfiGe(dpm)2C12 were determined and are reported in Table III. The solvent in all cases was methylene chloride and tetra— methylsilane was used as the internal standard. The concen— tration of the germanium complexes was 0.20M_and the concen- tration of the titanium complexes was 0.1QM. The proton between the two carbonyl groups in the B-diketonate anion, [RCOCHCOR]—, is referred to as the ring proton. The chemical shift of the ring proton of cis-Ge(dpm)2Cl2 is 0.1 ppm downfield from the shift for Al(dpm)3 in CDC13 (14). 'Ihis downfield shift for M(dik)2X2 as compared to the analo- gous M(dik)3 complex is typical and attributed to the fact that the cis-dihalobis(B-diketonates) have a dipole, whereas, the symmetric tris(B-diketonates) do not (26). - .qV': m v— 48 Table III. Chemical shifts for germanium and titanium fi-diketonatesa. Compound T, -CH T, -CH3 T, E7C4H9 ggnge(dpm)2C12 4.06 . 8.74, 8.85 [Ge(acac)3][SbC16] 4.03 7.76 [Ge(dpm)3][SbC16] 3.82 8.77 [Ti(acac)3][SbC16] 3.80 7.75 [Ti(dpm)3][SbC16] 3.61 8.78 aSee text for solvents and Concentrations used. The values for [Ge(acac)3][SbC16] and [Ti(acac)3][SbCl6] in Table III, are in good agreement with the values recently reported by Fay and Serpone (28). The ring proton resonances of the cationic complexes are shifted to lower fields than their analogous neutral complexes. Fay and Serpone attribute this fact to the positive charge and not to benzenoid ring currents. The trend of this downfield shift is germanium < tin < titanium. A comparison of the values of the chemical shifts reported in Table III illustrate this trend. The M(dik)2C12 compounds were used to prepare the [M(dik)3][SbC16] complexes. When Ge(dpm)2Cl2 was refluxed with antimony pentachloride in glacial acetic acid for two hours the analysis of the product corresponded to the ex- pected [Ge(dpm)3][SbC16] complex. Since all terminal groups of this complex are in an equivalent environment, see Figure 11, a singlet is expected in the gfbutyl region of the nmr 49 Figure 11. The cation [Ge(dpm)3]+. 51 spectrum. The expected singlet was observed at T 8.7. If however, antimony pentachloride is added to Ge(dpm)2C12 in methylene chloride in the absence of tetramethylsilane, and allowed to age for 0.5 hour at room temperature, the crystal- line product gives a set of two lines in the gfbutyl region. A very weak peak is seen at T 8.7, and corresponds to the expected resonance of [Ge(dpm%][SbC16]. 7A second, much larger peak is seen about seven cps downfield. As the solu- tion ages the downfield peak decreases and the upfield peak for [Ge(dpm)3][SbC16] increases. After thirty-two hours the peak heights are inverted. This change in spectra with time is illustrated in Figure 12. The spectral observations, the analysis, and the manner of preparation'indicate that the isolated compound may be a reaction intermediate in the preparation of [M(dik)3][M'Xn] from M(dik)2X2 and M'Xm, where n equals four to six and m equals two to five. This intermediate is postulated to be an adduct of the form M(dik)2X2,M'Xm if the octahedral symmetry is maintained in the adduct, then for [Ge(dpm)2C12,SbCl5] the ligands must be Egggg_since only one line is observed in the gfbutyl region which is attributed to the adduct. The reaction of dihalobis(B-diketonato)metal complexes with Lewis acids, in particular, ferric chloride, cupric chloride, auric chloride, and antimony pentachloride, to give the tris(B-diketonates) is well known. However, no 'work appears to have been done on the mechanism involved. During the preparation of [Ge(dpm)3][SbClG], it was observed 52 59mm pm 03.820 0:01:35 3 iaonméaofismswmg meme How .05... SDHB Aumz omv muuoomm wocmsommn cououm amusnlu 03p CH mmcmsu one .NH muowflm 2; mm... Camp 7{ .v I semi 2; N... seam... hm» Nmp . hm» 7r... 54 that the solvent plays an important role in this reaction. In glacial acetic acid [Ge(dpm)3][SbC16] precipitates im- mediately on addition of antimony(V) chloride to Ge(dpm)2C12. In methylene chloride the reaction is slow enough to detect a mixed compound, probably [Ge(dpm)2C12,SbC15], prior to formation of [Ge(dpm)3][SbC16]. Considerable effort and time was invested in an at- tempt to make Ge(hfac)2C12 (gf. experimental section). Neither the method of preparation used for the other bis- (B-diketonates), nor the addition of germanium tetrachloride to hexafluoroacetylacetone in the absence of solvent gave a discernible reaction. The addition of sodium hexafluoro- acetylacetoneate to germanium tetrachloride did give a reaction but no product was recovered. Since these complexes were to be used in the investigation of mixed ligand com— plexes, the preparation of the hexafluoro analog was especi- ally desirable. The reason for this is that the mixed ligand complexes are appreciably favored at the expense of the parent complexes when one complex contains very electro- negative terminal groups and the other does not. B. Isomerization of trans-Ge(dpm)2C12 The gfbutyl resonance line for trans-Ge(dpm)2C12 is located at the same position as the downfield peak of the doublet which corresponds to the gig isomer. When the trans isomer was dissolved in methylene chloride, it required less than fifteen minutes for the compound to isomerize. At this 55 point the spectrum was characteristic of an equilibrium mixture of the isomers. This is also about the length of time it took to dissolve the sample, (10 mg in 0.4 ml). In benzene the isomerization is much slower. Since trans-Ge(dpm)2C12 is only slightly soluble in benzene, a saturated solution was prepared and decanted after an hour. The change in concentration with time was followed by measuring the change in peak areas. Usually three measure- ments were taken at each time and the concentration of the lgig and trans isomers were determined by electronic inte- gration. Data was collected at half hour intervals for four hours and subsequent measurements were taken at six and fifteen hours to assure the establishment of equilibrium. The reaction was followed at 43.8°. If the isomerization is a simple first order reaction as written in reaction "c", c. trans > cis then a plot of the log (f ) versus time, where f trans -—————- mas is the fraction of trans isomer, should give a straight line. This plot is Figure 13 and shows that an equilibrium was established after approximately 260 minutes. Table IV lists the data used to calculate the rate of isomerization. Table V lists the data used to calculate the equilibrium constant, Keq = [figigl/[ftrans]° The equilibrium constant was found to be 4.64 i .31 calculated at the 95% confidence level. A better fit of the data is obtained if it is assumed that the reaction is reversible and first order as described in re- action "d". 56 .om.mv um mswncmn CA «HUNAEQCVTO udeuu mo coflumuflumEOmH may now mEAu m> a mCMHU LL moH to node .mH mAsmAm «E 0.9 02.. . con 08 oo. 0 To- O :3: , u u 050:? «no: Co .3202» of a. s 2...... C e 20 1|. ago: J. 3:2: .x a , o .06.... .u gaziom 232cm 5 . ~_o «1.335.. 26: 3 5:822:03 2: .2 as: 2.22.: co. 3 3:. .06... 0.0.... To... .m ' [won *1 «.o 0.0 58 Table IV. Kinetic data for isomerization of trans- Ge(dpm)2C12 in benzene at 43.8°. gifie ftrans fcis 1:9 ) log[ftrans" £33233 (1_ftrans);Keq] 0 0.826 0.174 -0.08328 -0.1033 0 0.857 0.143 -0.07308 -0.08291 27 0.591 0 .419 -0.1903 -0 .3004 0.620 0.380 -0.2076 -0.2691 0.610 0.390 -0.2145 -0.2791 33 0.596 0.404 -0.2246 -0.2925 0.592 0.408 -0.2274 -0.2975 37 0.598' 0.402 -0.2519 ~0.2998 0 .596 0 .404 -0 .2246 -0 .2925 60 0.443 0.557 -0.3536 -0.4908 0.461 0.539 -0.3364 -0.4622 77 0.404 0.596 -0.3940 -0.5607 0.417 0.583 -0.3796 -0.5361 105 0.322 0.678 -0.4917 -0.7545 0.316 0.654 -0.4604 -0.6883 0.323 0.677 -0.4914 -0.7520 133 0.269 0.731 -0.5711 -0.9547 0.277 0.723 -0.5576 -0.9172 0.277 0.723 -0.5571 -0.9172 165 0.229 0.771 -0.6395 -1.201 0.232 0.768 -0.6340 -1.179 0.229 0.771 -0.6395 -1.201 197 0.200 0.800 -0.6988 -1.656 0.222 0.778 -0.6527 -1.268 0.209 0.791 —0.6793 -1.420 59 Table V. Equilibrium constant data for trans-Ge(dpm)2Clz 2+§g> cis-Ge(dpm)2C12. Time f f . K min trans Cls eq 265 0.165 0.835 . 5.07a 0.214 0.786 3.67 0.172 0.828 4.81 411 0.178 0.822 4.62 0.192 0.808 4.21 0.183 0.817 4.47 15.5 ' 0.167 0.833 4.99a f— hrs 0.206 0.793 3.85 0.189 0.811 4.29 Average: 4.64 i .31b aDropped for 95% confidence level calculation. 1 b95% confidence level. ' a 1 . d. trans > CIS RE The integrated rate equation for this reaction is: log [ftrans - (l-ftranS)/Keq] -k1(1 + l/Keq)t, where Keq = k1/k2. k1 calculated from the slope was 1.21 x 10_2 min-1. 3 From k1 and Ke kg was calculated to be 2.61 x 10- q' _ . -1 - min . Although the plot in Figure 14 implies that there is good correlation between the rate equation for reaction "d" and the data, the Spectra showed that a small amount of free ligand (about 53% of the total concentration) had formed as the solution aged. It was not determined if the 60 .ow.mv um msmncmn CH «HONWEQCVOOI 0:» you msAu.mw .umx\ WCMH A en IHVI mcmuu msmuu mo COHumNHHmEOmA ma moa no uon .vH musmflm CON 5:; . 8. . ,8. 8 O... .x. m2 . use: 8: 8 .828 8.2: 755 no. x 5?... rs... N.o. . 5.... :. A22. . I . «x a. .m + ¢w .6 ad..." 3. T. a... + 1.5.5:.-. 3. £0 4.” mac: 3 0:0 «a Ancobafiv U , o omfit 3 «.0 ~28: no... 28.: 8 cozouzoeoa. of 8. as: .2, way: 15...... T. 22$ no. .3 8... (Do 8:23. 838m 5 . «.7... 0.0 62 free ligand resulted from isomerization or from hydrolysis and the formation of the free ligand was not taken into account when the rate equation was derived. In any case, the calculated rates are intended to be only a first approxi- mation. There are no examples of M(dik)3X2 isomerization to use as a comparison. Fay (3) has studied the cis-trans isomerization of gigftris(1,1,1-trifluoro-2,4-pentanedionato)- cobalt(III), Co(tfac)3. Fay found that the rate equation based on reaction “d" fit the data. The equilibrium con- stant was 0.25, k1 equaled 9.8 x 10.3 min-1, and k2 equaled 4.2 x 10.4 min.1 for the isomerization of gi§;Co(tfac)3 at 660 in methylene chloride. In both cases, the isomerization follows a reversible, first order rate law. Since, the isomerization of Eragnge(dpm)2Clz is re- versible and first order, three mechanisms are possible, (1) bond rupture of the ligand; (2) dissociation of chlorine; and (3) dissociation of ligand. Also the observations indicate that the rate of isomerization is influenced by the solvent. However, the data do not permit the evaluation of the three possible mechanisms. C. Exchange Reactions The relative inertness of germanium tris(fi—diketonate) complexes was illustrated when the compounds were mixed so that exchange might occur. These reactions, for neutral B—diketonates, are numerous and have been studied 63 extensively. The exchanges proceed smoothly and quickly to form a large variety of mixed ligand complexes. That the germanium systems examined did not undergo any perceptible ligand exchange reactions was, therefore, somewhat unusual. The systems studied were [Ge(acac)3]+ - [Ge(dpm)3]+ and [Ge(acac)3]+ - [Ge(bzbz)3]+. The expected spectra for the exchange reaction between [Ge(acac)3]+ and [Ge(dpm)3]+ may have a total of eight lines in the tfbutyl and methyl regions of the nmr spectrum. They are: two singlets which correspond to the parent com- plexes; a doublet in the tfbutyl region and a singlet in the methyl region which correspond to the mixed ligand com- ’3 plex [Ge(acac)(dpm)2]+ (the tfbutyl groups exist in two non— equivalent environments); and a doublet in the methyl region and a singlet in the Efbutyl region which correspond to the mixed ligand complex [Ge(acac)2(dpm)]+ (the methyl groups exist in two non-equivalent environments). An equimolar solution of [Ge(acac)3]+ and [Ge(dpm)3]+ in methylene chloride was allowed to age for ten hours at room temperature. The sample was periodically checked gig nmr but after ten hours the spectra indicated that no exchange had occurred. Only the two original singlets which corresponded to the parent compounds were present and unchanged. The system [Ge(acac)3]+ - [Ge(bzbz)3]+ was investigated next. »In this case, the electronegativity of the two ligands was much larger and, therefore, the possibility of exchange was enhanced. The spectrum may have a maximum of 64 four lines in the methyl region; two singlets which cor- respond to [Ge(acac)3]+ and [Ge(acac)(bzbz)2]+, and one doublet which corresponds to [Ge(acac)2(bzbz)]+. An equi- molar mixture of [Ge(acac)3]+ and [Ge(bzbz)3]+ in methylene chloride, was sealed in a heavy walled tube and heated for eight hours at 210°. .Since [Ge(bzbz)3]+ is only parti- ally soluble, the supernatant was decanted into a nmr tube in a nitrogen atmosphere. The spectra gave one line in the methyl region which corresponded to [Ge(acac)3]+. The presence of [Ge(bzbz)3]+ was indicated by a complicated peak in the phenyl region. Since the cationic germanium systems did not undergo exchange reactions a comparable set of cationic titanium complexes were investigated, they are: [Ti(acac)3][SbC16], [Ti(dpm)3][SbC16], and [Ti(bzbz)3][SbC16]. These complexes were studied in pairs and all three sets, [Ti(acac)3]+ — [Ti(dpm)3]+, [Ti(acac)3]+ - [Ti(bzbz)3]+, and [Ti(dpm)3]+ - [Ti(bzbz)3]+ were found to readily undergo exchange reac- tions. .An examination of the relative intensities of the nmr lines as a function of time showed that in all cases equilibrium was established in less than five minutes after dissolution of the complexes. The total molarity of titan- ium was varied from 0.1g;to 0.65 and the equilibrium was observed in a variety of solvent; methylene chloride, benzene, chlorobenzene, and Qfdichlorobenzene. ‘For each system there are four possible species in solution, the two parent com- plexes, and two mixed ligand complexes, [Ti(dik)2(dik')]+ 65 and [Ti(dik)(dik')2]+. The resonance lines of the complexes in both the ring proton and terminal group region were as- signed by examination of the changes in relative signal intensities as a function of ligand composition. Examples of spectra in the ring proton region of an equilibrium mix— ture are shown in Figure 15. The system illustrated is [Ti(acac)3]+ — [Ti(bzbz)3]+ in methylene chloride and at equilibrium. The term facac is defined as the fraction of total ligand present as acetylacetonate. The resonance lines at 13.67, 13.72, and 13.77 are assigned to the acetylacetonate ring protons of [Ti(acac)(bzbz)2]+, [Ti(acac)2(bzbz)]+, and [Ti(acac)3]+, respectively. The chemical shifts for both the parent and mixed com- plexes were determined in methylene chloride with tetra- methylsilane as an internal standard and facac = 0.50. They are reported in the appr0priate sections. Resonance lines for the ligand dibenzoylmethanate in either the parent complex or in the mixed ligand complexes were not examined. Both the terminal phenyl groups and the ring protons are much farther downfield than the resonances attributed to the other ligands. Some difficulty was encountered in the resolution of lines in the Efbutyl and methyl regions and the corresponding ring proton regions. This led to the spectral examination of these three systems in methylene chloride and chloro- benzene at 100 MHz. -The spectra were run by Dr. J..Heeschen of the Chemical Physics Lab, Dow Chemical Company, Midland. 66 Figure 15. Changes in relative signal intensities as a function of ligand composition for the acetylacetonate ring proton s ectra + (60 MHz) of 0.3Og_[Ti(acac)3] - [Ti(bzbz)3] at equilibrium in methylene chloride at 33.0 . fame = 0.67 all 1545 "‘35 Nant‘l!‘ a: ML!” “lg-1h tufirvrlx‘r . "V‘Ir‘lfijfilr . figs-f “Wig; . facac=o.5o as he 252 3: . [‘ Air ._' , ‘Q’ilflg. “k; *9 ,, ~ aqw‘ W fmfi 0.33 68 A variety of solvents were used in an attempt to resolve the expected peaks of the spectrum. The increase in resolu- tion with solvent was in the order: benzene < chlorobenzene < gfdichlorobenzene < methylene chloride. . The first system studied was [Ti(acac)3]+ - [Ti(dpm)3]+. A set of eight lines analogous to those described earlier for [Ge(acac)3]+ - [Ge(dpm)3]+ are possible in the methyl and tfbutyl regions. In the downfield region T3.0-4.0, two sets of three peaks should be present, with each set a representation of the ring protons which correspond to the mono, bis, and tris complexes of a given ligand.. This system was studied in benzene, chlorobenzene, Qfdichlorobenzene, and methylene chloride. The total titanium concentration was varied from 0.1§_to 0.4g, In benzene both ring proton regions and the methyl region showed a broad singlet with a shoulder. The Efbutyl region showed only a broad singlet. In chlorobenzene and gfdichloro- benzene the ring proton regions were obscured by the solvent or solvent impurities, and the methyl and Efbutyl regions showed broad peaks similar to those found when benzene was used as a solvent. A sample dissolved in chlorobenzene was examined in the Efbutyl and methyl regions at 100 MHz, but neither peak was further resolved. Figure 16 illustrates the results of the spectrum obtained in methylene chloride at 0.30§_titanium and facac = 0.67 at 60 MHz. The di- pivalolmethanate ring region was partially resolved and gave two greatly overlapped peaks. The half width of the 69 mamamnume CH bw.o H umom muouxflfi Esflunflaflsvm cm mo A m . _m Nmz+om m smovael . Eduuuwmm m .ao.mm um.mcHHoH;o .axumomvflel Som.o mo CHH QUGMGOmQH CODOHQ .®H wusmflm khmh skunk Sun... n5»... .. _ s .1 71 overlapped peaks is 0.8 cps. The two overlapped peaks are separated by 0.3 cps and the chemical shifts in T values for both is 13.61, in cps the chemical shifts are 383.4 and 383.1. The acetylacetonate ring proton region gave a set of three peaks, the two upfield peaks are slightly overlapped. The peak assignment and corresponding chemical shifts are: [Ti(acac)(dpm)2]+, 13.73; [Ti(acac)2(dpm)]+, 13.76; [Ti(acac)3]+, T3.77. The methyl region gave a broad singlet with a chemical shift of 17.74 and a half width of 0.77 cps. The Efbutyl region gave what appeared to be two peaks with a great deal of overlap. They are separated by only 0.3 cps and the half width is 0.69 cps. The chemical shifts for both peaks is 18.77, in cps the values are 73.7 and 73.4. The 100 MHz spectrum did not show a change in the tfbutyl region but the methyl peak was resolved into two greatly overlapped peaks, Figure 17. In both the Efbutyl and methyl regions, one of the two peaks which overlap should be a doublet which corresponds to the respective bis ligand complex. Other titanium com- plexes are known to undergo rapid terminal group exchange at room temperature and give singlets instead of the predicted doublets. Therefore, spectra of these systems were run at low temperature to see if any of the peaks could be resolved into doublets. Methylene chloride solutions which were 0.30g_in titanium and had a ligand composition of facac = 0.33 or 0.67 were examined at —62.8°. The results are: tfbutyl region, one broad line with a half width of 2.0 cps; 72 Figure 17. Proton resonance lines (100 MHz) for an equilibrium mixture of 0.30g.[Ti(acac)3] - [T1(dpm)3] , facac = 0.50 in methylene chloride at room temperature. 74 methyl region, one broad line with a slight shoulder and a half width of 2.0 cps; in both ring proton regions, broader peaks. Although the [Ti(acac)3]+ — [Ti(dpm)3]+ system did undergo ligand exchange, the system was found to be unsuit- able for quantitative work. The requirement that a set of peaks in at least one of the regions be complete, well de- fined, and separable was not met. Apparently, the chemical shifts in all regions for the system are so similar that the individual peaks are not separable. The system [Ti(dpm)3]+ - [Ti(bzbz)3]+, theoretically, should give rise to four lines in the Efbutyl region; two singlets which correspond to [Ti(dpm)3]+ and [Ti(dpm)(bzbz)2]T and a doublet which corresponds to [Ti(dpm)2(bzbz)]+. There should also be a set of three singlets in the dipivaloy1- methanate ring proton region which correspond to the mono, bis, and tris dipivaloylmethanate complexes. The system was insoluble'in benzene but fairly soluble in methylene chloride. The spectrum with methylene chloride as the solvent, fdpm = 0.50, and a titanium concentration of 0.30%, gave the fol— lowing results. The ring proton region gave a set of three lines with a small amount of overlap at the base line. The chemical shifts for the ring protons are: [Ti(dpm)(bzbz)2]+, + + 13-55. [Ti(dpm)3(bzbz)] . 13.57, and [Ti(dpm)3] . 73.59. The tfbutyl region gave two greatly overlapped peaks separ—. ated by 0.6 cps with chemical shifts of T8.75 and T8.76. The half width was 0.50 cps. The spectrum of an equilibrium 75 mixture of 0.309; [Ti(dpm)3]+ — [Ti(bzbz)3]+ in both the Erbutyl and dipivaloylmethanate ring proton regions is il- lustrated in Figure 18. The spectrum was also examined at 100 MHz. tAlthough the ring proton region was completely resolved, the Efbutyl region was unchanged. As in the case of the [Ti(acac)3]+ - [Ti(dpm)3]+ system, although ligand exchange did take place, the resulting spectra were not well enough resolved to permit quantitative study of the f— system. The system [Ti(acac)3]+ - [Ti(bzbz)3]+ theoretically should give rise to four lines in the methyl region; two singlets which correspond to [Ti(acac)3]+ and [Ti(acac)(bzbz)2]+, and a doublet which corresponds to [Ti(acac)2(bzbz)]+. There should also be a set of three singlets in the acetylacetonate ring proton region.which correspond to the mono, his and tris acetylacetonate com- plexes. The system was insoluble in benzene and only sparingly soluble in chlorobenzene and gfdichlorobenzene. The chemical shifts were done in methylene chloride with a total titanium concentration of 0.40§_and facac = 0.50. Figure 19 illustrates the spectrum obtained under these conditions. In the acetylacetonate ring proton region three well defined peaks were present: [Ti(acac)(bzbz)2]+, 13.67; [Ti(acac)2(bzbz)]+, 13.72; and [Ti(acac)3]+, 13.77. The methyl region gave three peaks which overlapped near the base line. The chemical shift assignments are: [Ti(acac)(bzbz)2]+, 17.66; [Ti(acac)2(bzbz)]+, 17.70; and [Ti(acac)3]+, 17.73. 76 Figure 18. Ring proton and t-butyl nmr lines (60 MHz) of an equilibrium mixture of 0.30M [Ti(dpm) 3] - [Ti(bzbz)3]+, fdpm— = 0. 50, in methylene chloride at 33.00. T3.55 T357 T359 T876 78 Figure 19. Ring proton and methyl proton nmr lines (60 MHz) for an equilibrium mixture of 0.40M_[Ti(acac)3] - [Ti(bzbz)3] , acac = 0.50, in methylene chloride at 33.00 b... - m“-u.-_~.xsm. ._,.__ _ 05.5.... 2......" 09...... . h . a _ _ 80 The "bis" line, methyl resonance line for [Ti(acac)2(bzbz)]+, should be a doublet. The sample was run at 100 MHz but there was no indication of a doublet for the "bis" line. The system was then studied at low temper- ature to see if the doublet could be resolved. Low temper- ature spectra were examined for samples with the following listed parameters, total concentration of titanium, acetyl- acetonate concentration, and the lowest temperature achieved. They are: 0-60l’l. Ti(IV), f = 0.33, T = —65.5°; 0-402’1. Ti(IV), acac f = 0.50, T = -86.9°; 0.3m Ti(IV), f = 0.67, T =- acac acac —89.0°. .All spectra gave the same results. Although there was a marked broadening of the peaks, the doublet which corresponds to [Ti(acac)2(bzbz)]+ was not observed. -This may be due to the chemical shifts of the two lines in the doublet. That is, the line separation may be very small, and not to the fact that at -89.00 the terminal groups are still exchanging. Although this system was not suitable for kinetic studies by nmr line broadening techniques, the sepctrum of the ring proton region was such that quantitative thermo- dynamic data could be obtained and this was done. ~D. Equilibrium Studies of [Ti(acac)3]+ - [Ti(bzbz)3]+ Mixture The equilibrium constants were determined for the sys- temei(acac)3]+ - (Ti(bzbz)3]+ in methylene chloride. The equilibria are described by two independent equilibrium 81 constants, K1 and K2, which were determined experimentally and are defined by the following reactions: K (1) [Ti(bzbz)3]+ + [Ti(acac)2(bzbz)]+‘—l> 2[Ti(acac)(bzbz)2]+ K (2) [Ti(acac)3]+ + [Ti(acac)(bzbz)2]+'—£> 2[Ti(acac)2(bzbz)]+. For the discussion of the equilibrium in terms of the forma— tion of one mole of the mixed ligand complexes from the parent complexes the equilibrium constants, Kf1 and Kfz’ are used and defined by the following reactions: + + Kfl + (3) 2/3[Ti(bzbz)3] + 1/3[Ti(acac)3] -¢ [Ti(acac)(bzbz)2] K (4) 1/3[Ti(bzbz)3]+ + 2/3[Ti(acac)3]+'-£§ [Ti(acac)2(bzbz)]+. K and Kf2 may be calculated from K1 and K2 from the £1 2 1 1 relationships; K = K1 /3K2 /3 and K 2 = K1 /3K22/3' f1 f For the statistical distribution_of ligands the depend- ence of the equilibrium molar fraction of each compound, fTi(acac)n(bzbz)3_ ’ ls given by: n Ti(acac)n(bzbz)3_n _ acac bzbz n1(3-n)£ ' where n = 0, l, 2, or 3; and f and f are the acac bzbz molar fractions of total ligand present as acetylacetonate and dibenzoylmethanate, respectively. The statistical value for both .K1 and K2 is 3.00. The statistical value for K and 'K is also 3.00. f1 £2 The concentration of the complexes which were used in the calculations of the equilibrium constants were determined 82 by integration of their acetylacetonate ring proton lines. The integration was done by planimetry. The concentration of [Ti(bzbz)3]+ was obtained by difference. The equilibrium constants, K1 and K2 were studied as a function of total solute molarity and ligand composi- tion at 33.30. The results are tabulated in Tables VI and VII . Table VI. Dependence of the equilibrium constant for the system [Ti(acac)3]+ - [Ti(bzbz)3]+ on total . a solute molarity. Total-Molarity Average Valuesb K1 K2 0.137 2.92 i .26c 3.11 i .13 0.215 2.69 i .24 2.86 i .18 0.304 2.65 i .12 -2.81 i .10 0.429 2.97 i .22 2.80 i .08 0.601 3.03 i .18 2.84 i .13 aIn methylene chloride; temperature is 33.30, facac = 0.50. bAverage of at least five spectral measurements. cAll errors estimated at the 95% confidence level. The equilibrium constants are independent of both total solute molarity and ligand composition within experimental error. This implies that the solutions are either close to ideality or that the activity coefficient quotient is constant over the concentration ranges investigated. Tables VI and VII also show that all equilibrium constants are close to the statistical value of 3.00. 83 Table VII. Dependence of equilibrium constants on ligand composition for the [Ti(acac)3] -[Ti(beZ)3]+ system.a b Average Values acac K1 K2 0.339 2.89 i .14C 0.507 2.65 i .12 2.81 i .10 0.665 3.12 i .20 aIn methylene chloride; temperature = 33.30; total solute molarity = 0.301. bAverage of at least eight spectral measurements. 9All errors estimated at 95% confidence level. The temperature dependence of K1 and K2 was studied so that the enthalpy and entropy of the exchange reactions might be determined. The temperature range was -12.0° to 52.00, facac = 0.33 and 0.67 for K1 and K2 respectively. Table VIII shows the temperature dependence of K1 and K2. Enthalpy and entropy changes for reactions 1 and 2 were determined from the slope and intercept of log K versus 1/T plots. The data were treated by least squares analysis, 29 points were used for K1 and 31 points were used for K2. The thermodynamic data, along with the extrapolated values of K1 and K2 at 250 are given in Table IX. Also included in Table IX are the enthalpies, entropies, and equilibrium constants for formation of the mixed ligand complexes from the parent complexes, as defined by reactions 3 and 4. These values were calculated from the experimentally 84 Table VIII. Temperature dependence of equilibrium constants for the [Ti(acac)3] - [Ti(bzbz)3] system.a Average Valuesb Tem erature c d poc K1 K2 —12.0 3.14 i .13e 2.92 i .14 9.3 2.94 i .16 3.08 i .12 33.3 2.92 i .14 3.12 i .20 52.0 2.98 i .19 2.91 i .14 aIn methylene chloride. bAverage of at least eight spectral measurements. Cfacac = 0.34; total solute molarity = 0.30. dfacac = 0.66; total solute molarity = 0.30. e All errors estimated at the 95% confidence level. Table IX. Thermodynamic data for formation of mixed ligand complexes in methylene chloride at 25°. Eggiiigitum AH kcal/mole AS eu K1 3.0 i .10 -0.16 i .16a 1.63 i .52 K2 2.97 i .10 0.02 i .19 2.24 i .33 Kf1 2.99 i .10 —0.10 i .17 1.83 i .46 Kf2 2.98 i .10 —0.04 i .19 2.04 i .39: stat. value 3.00 0.0 2.18 aAll errors estimated at the 95% confidence level. 85 observed parameters for reactions 1 and 2 with the use of f1 = K12/3K21/3 and Kfz = K11/3K22/3. .The results tabulated in Table IX indicate that the ligand the relationships: K exchange process for this system occurs in a statistical manner. That is, the cationic mixed ligand complexes are formed to the extent predicted by the statistical scrambling of ligands without regard for the chemical properties of the system. -These results are consistent with those found for the ligand exchange of neutral species where both of the a-diketonates are non-fluorinated. IV . BIBLIOGRAPHY 1. m 10. 11. 12. 13. 14. 15. IV. BIBLIOGRAPHY F. Basolo and R. G. Pearson,"Mechanisms of Inorganic Reactions", 2nd ed, John Wiley and Sons, Inc., New York, N.Y., London, and Sydney, 1967, chapter 4. R. C. Fay and T. S. Piper, J. Am. Chem. Soc., 88, 500 (1963). F— R. C. Fay and T. S. Piper, Inorg. Chem., 8, 348 (1964). A. Palmer. R. C. Fay. and T. S. Piper, Inorg. Chem., 8, . 875 (1964). R. C. Fay and R. N. Lowry, Inorg. Chem., 8, 1512 (1967). N. Serpone and R. C. Fay, Inorg. Chem., 8, 1835 (1967). T. J. Pinnavaia, J. M. Sebeson, and D. A. Case, Inorg. Chem., 8, 644 (1969). A. Werner, Ber., 28, 3061 (1912). D. Ra and N. K. Dutt, J. Indian Chem. Soc., 88, 81 (1943). J. C. Bailar, J. Inorg; Nucl. Chem., 8, 165 (1958). C. S. Springer and R. E. Sievers, Inorg. Chem., 8, 852 (1967). H. S. Gutowsky and C. H. Holm, J. Chem. Phys., 88, 1228 (1956). T. J. Pinnavaia and R. C. Fay, Inorg, Chem., 8, 233 (1966). J. J. Fortman and R. E. Sievers, Inorg. Chem., 8, 2022 (1967). T. J. Pinnavaia and S. O. NWeke, Inorg. Chem., 8, 639 (1969). 86 16. 17. 18. 19, 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 87 J. J. Fortman and R. E. Sievers,Abstracts, 155th National Meeting of the American Chemical Society, San Francisco, Calif., Apr. 1968, No. M57. G. T. Morgan and H. D. K. Drew, J. Chem. Soc., 125, 1261 (1924). V. V. Udovenko and Yu. Ya. Fialdov, Zh. Neor . Khim., 8, 434 (1957); J. Inorg. Chem. USSR, 8, 317 (1957). M. Cox, T. Lewis, and R. S. Nyholm. J. Chem. Soc., 6113 (1964). W. K. Ong and R. J. Prince, J. Inorg. Nucl. Chem., 21, 1037 (1965). O. A. Osipov, V. L. Shelepina, and O. E. Shelepin, Zh. 0bshch. Khim., 89‘, 264 (1966); J. Gen. Chem. USSR, §§, 274 (1966). L. S. Forster, J. W. Drenan, and A. F. Williston, Inorg. Syntheses, 8, 109 (1946). G. Brauer, "Handbook of Preparative Inorganic Chemistry", Vol. 1, 2nd ed, Academic Press, New York, N.Y. and London, 1963, pp. 715-16. R. C. Fay and T. J. Pinnavaia, Inorg. Syntheses, to be published. A. Weissberger, "Technique of Organic Chemistry”, Vol. 7, 2nd ed, Organic Solvents, Interscience,-New York, N.Y., and London, 1955. J. A. S. Smith and E. J. Wilkins, J. Chem. Soc., Sect. Q, 1749 (1966). W. C. Fernelius and B. E. Bryant, Inorgy Syntheses, 8, 105 (1957). R. C. Fay and N. Serpone, J. Am. Chem. Soc., 88, 5701 (1968). T. J. Pinnavaia and L. Matienzo, to be published. W. Dilthey, J. Prakt. Chem., 111, 147 (1925). V. M. Goldshmidt, Trans. Faraday Soc., 88, 253 (1929). L. Pauling, "Nature of the Chemical Bond", 2nd ed, Cornell Univ., Ithaca, N.Y., 1948, p. 53. 88 33. M. Klotz, "Chemical Thermodynamics", lst ed, Prentice— ‘Hall, Inc., New York, N.Y., 1948, p. 56. 34. "JANAF Thermochemical Tables", The Advanced Research Projects Agency, Air Force Contract AF 04(611)-7554, U.S. Government Printing Office, Washington, D. C., 1965. 35. M. Cox, R. J. H. Clark, and J. J. Milledye, Nature, 212, 1357 (1966). 36. y. Kawasaki, T. Tanaka, and R. Okawara, Spectrachim. Acta, 22, 1571 (1966). 37. R. c. Fay and R. N. Lowry, Inorg. Nucl. Chem. Letters, 8, 117 (1967). 38. W. H. Nelson, Inorg. Chem., 8, 1509 (1967). 39. D. c. Bradley and C. E. Hollowa , Official Digest Fed. Soc._gaint Tech., 81, 483 (1965K. 40. J. W. Faller and A. Davison, Inorg. Chem., 8, 182 (1967). 41. R. C. Fay, Cornell University, private communication, 1969. ' uuuuunmm nml uu11111111111111)!!!” 3 1293 0310