. .. . s ., , ...§..Vq‘.4...$§uh> it“. t.“ .. . ‘3‘ ».v 7.. . . .. , , A. A» wé..mriuw ,. s . .., .s e . r . In...“ ... .. . . ..v . s s V ..A.‘ .r _ .. A ,. ..,., .. .. s s . . ,. ,..s. .H. ... . . a s . , ,l . . . v2.. ‘ ms . .._ H . u . ..: , A. .H. ...s. .22.: , .1 . s. . A, . .. q _ . I 1 D _,.. .. ;. .. . . 5... ,.s ...s..s.....; .. u R“, s: v .. .. s . .0 .: . . 9 . . . .. ....f. . . ,e in, w . IL s I , . a , ... . .. m Efl {CH as 1. n 15 155 mi ties EM m #6:. . J .. . 1 v .. s. it . 1r 5 3.0.3.1“ emu. 3&5 "f! ss PH I v v A . ‘s D . . s s . .. .. .I 2}: ., , . f , . s . i . . . . .. 2 IV} $3, V .....1:1.s Au 21:5,. 5: , H a?“signrgsé €35. .A.Es.........x.: £59155 3:: . s. 3 S: : a... k This is to certify that the thesis entitled Photochemical Reactions of Bis(n5-cyclopontad1enyl)titanium Dichloride presented by Edward W. V1 tz has been accepted towards fulfillment of the requirements for Ph . D. degree in ._._ "—1.8310th ‘y Major professor Date—flQmDQLlZW 9 74 0-7639 3" " ma Iv ' HUAB 8- SUNS 800K BINDERY INC. LIBRARY Bmoans . summit]. mama; / ABSTRACT OF BIS(n5-CYCLOPENTADIENYL)TITANIUM DICHLORIDE GOV2 w PHOTOCHEMICAL REACT IONS 0\ By (9 Edward Walter Vitz The exchange of cyclopentadienyl ligands between molecules of bis(n5-cvc10pentadienyl)titanium dichloride (titanocene dichloride) is exclusively photolytic. The increase in the (D-5)titanocene dichloride concentration in benzene solutions which initially contained only titanocene dichloride and (D-lO)titanocene dichloride was detected by mass spectrometry. A quantum yield of 0.23 moles/einstein was calcul- ated, and the equilibrium constant for the reaction was 4.0. Cyclo- pentadienyl ligand exchange between titanocene dichloride and titano- cene monochloride dimer also proved to be photolytic, and had a higher quantum yield than the above photoexchange. The alcoholysis of titanocene dichloride in benzene-methanol solutions was also shown to be exclusively photolytic at room temper- ature. The quantum yield was 0.44 moles/einstein for solutions 8.1 x 10"3 fl.in titanocene dichloride and 1.0 g’in methanol, and the product was n5«cyc10pentadienyl(methoxo)titanium dichloride by mass spectrometry. Studies of several nonphotolytic reactions are appended. These reactions included the halide exchange between titanocene dichloride and titanocene dibromide (which was complete, with an equilibrium constant of 0.5, in the time of mixing and quenching in the dark) and two electron exchange reactions. First, the electron exchange between titanocene dichloride and titanocene monochloride in tetrahydrofuran was complete in the time of mixing and separating (ca, 30 seconds) in the dark. Second, a minimum rate constant of 4 x 103 was established for the electron exchange between bis(benzene)chromium and its cation. PHOTOCHEMICAL REACTIONS OF BIS(nS—CYCLOPENTADIENYL)TITANIUM DICHLORIDE By Edward Walter Vitz A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1974 To Mimi, Yossarian and The Greenville Granfalloon ii ACKNOWLEDGEMENT I would like to acknowledge the contributions of Professor Carl H. Brubaker, whose professional assistance was most helpful and whose personal inspiration was indispensable. The financial support of the National Science Foundation and the Dow Chemical Company has been greatly appreciated. iii TABLE OF CONTENTS I NTRO DUCT ION O O O O O O O O O O O O 0 O O O O O O O O O O O 0 EH ERII'ENTAL O O O O O O O O O O O O O O O O O O O O O O O O 0 General . . . . . . . . . . . . . . . . . . . . . . Purification of Materials . . . . . . . . . . . . . Preparations . . . . . . . . . . . . . . . . . . . . Photolysis Apparatus . . . . . . . . . . . . . . . . Chemical Actinometer Solutions . . . . . . . . . . . Determination of Lamp Intensity . . . . . . . . . . Photolytic Cyc10pentadieny1 Ligand Exchange Between Titanocene Dichloride and Perdeuterotitanocene Di- chloride . . . . . . . . . . . . . . . . . . . . . . H. Photolytic Cyc10pentadienyl Ligand Exchange Between Perdeuterotitanocene Dichloride and Titanocene Monochloride Dimer . . . . . . . . . . . . . . . . . I. McKay Analysis of Mass Spectrographic Data for Ligand Exchange Processes . . . . . . . . . . . . . J. Photolytic Methanolysis of Titanocene Dichloride . . nmmuow> RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . . . . . A. Photolytic Cyclopentadienyl Ligand Exchange Between Titanocene Dichloride and Perdeuterotitanocene Di- chloride . . . . . . . . . . . . . . . . . . . . . . B. Photolytic Cyclopentadienyl Ligand Exchange Between Perdeuterotitanocene Dichloride and Titanocene Monochloride Dimer . . . . . . . . . . . . . . . . . C. Photolytic Methanolysis of Titanocene Dichloride . . D. Reaction Mechanisms . . . . . . . . . . . . . . . . APPENDIX A. The Halide Exchange Between Titanocene Dichloride and Titanocene Dibromide . . . . . . . . . . . . Experimental . . . . . . . . . . . . . . . . . . . . . . Dis CUSSion O O O O O O O O O O C C O O O O O O O O I O . APPENDIX B. The Electron Exchange Between Bis(arene)chromium Complexes and Their Cations . . . . . . . . . . . I. Introduction . . . . . . . . . . . . . . . . . . . . II. Quenched-Flow Study of the Electron Exchange Between Bis(benzene)chromium(0) and Bis(benzene)chromium(I). iv l3 l4 19 21 21 25 30 36 42 43 43 44 46 46 Experimental . . . . . . . . . . . .s. . . . . . A. Preparation of Materials . . . . . . . . B. Gamma Counting Procedures . . . . . . . C. Quantitative Analysis of Chromium . . . D. The Szilard-Chalmers Reaction of Bis- (benzene)chromium Iodide . . . . . . . . E. Quenched-Flow Experiments . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . III. Study of the Electron Exchange Between Bis(arene)- chromium Complexes and Their Cations by ESR Tech- niques . . . . . . . . . . . . . . . . . . . . . . Experimental . . . . . . . . . . . . . . . . . . A. Preparation and Purification of Materials . . . . . . . . . . . . . . . B. Exchange Experiments . . . . . . . . . . C. Photochemistry of Bis(bipheny1)chromium Iodide . . . . . . . . . . . . . . . . . Results and Discussion . . . . . . . . . . . . . IV. Study of the Electron Exchange Between Bis(arene)- chromium Complexes and Their Cations by NMR Tech- niques . . . . . . . . . . . . . . . . . . . . . . Experimental . . . . . . . . . . . . . . . . . . Results and Discussion . . . . . . . . . . . . . APPENDIX C. The Electron Exchange Between Titanocene Dichloride and Titanocene Monochloride . . . . . Experimental . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . BIBLIOGRAPHY O O O O O O O O O O O O O O O O O O O O O O O O O 49 49 50 51 51 52 54 56 58 58 59 59 6O 67 67 68 7O 70 72 76 LIST OF TABLES Table Page 1. Mass Spectrographic data for the perdeuterotitanocene dichloride - titanocene dichloride ligand exchange . . . 22 2. Mass Spectrographic data for the perdeuterotitanocene dichloride - titanocene monochloride exchange . . . . . 28 3. Mass Spectrographic data for the perdeuterotitanocene dichloride — titanocene monochloride equilibrium mixture 29 4. Photomethanolysis data for titanocene dichloride . . . . 33 5. Quenched-flow exchange experimental data . . . . . . . . 55 6. NMR linewidths for exchange solutions . . . . . . . . . 69 vi LIST OF FIGURES Figure Page 1. Photolysis apparatus . . . . . . . . . . . . . . . . . . l 2. Mass spectrum of titanocene dichloride . . . . . . . . . 15 3. Mass spectrum of (D-10)titanocene dichloride . . . . . . 16 4. Mass spectra of exchange solutions . . . . . . . . . . . 18 5. McKay plot for perdeuterotitanocene dichloride — titano- cene dichloride exchange . . . . . . . . . . . . . . . . 23 6. Direct mechanism for ligand exchange between titanocene dichloride and titanocene monochloride . . . . . . . . . 26 7. Oxidation state interchange mechanism for ligand exchange between titanocene dichloride and titanocene monochloride 27 8. The mass spectrum of the photomethanolysis product . . . 35 9. Coordinate axes chosen for titanocene dichloride . . . . 37 10. Titanocene dichloride molecular orbitals . . . . . . . . 37 ll. Ultraviolet-visible spectrum of titanocene dichloride . . 38 12. Proposed mechanism for photolytic reactions of titanocene diChloride. O O C O ‘0 O O . O I O O O O O O O O O O O O O I 41 13. Mass spectrum of the halide exchange products . . . . . . 45 14 O quenChed-f 10w apparatus 0 O O O O O C O O C O O O O O C O 53 15. Exchange broadened esr spectrum of bis(biphenyl)chromium iodide O C I O O O O O O O I O O O O O O O O I O C O O C 61 16. Weissman plot for electron exchange between bis(biphenyl)- chromium(0) and its univalent cation . . . . . . . . . . 62 17. Exchange broadened esr Spectra of bis(benzene)chromium(I) 63 18. Exchange broadened esr spectra of bis(benzene)chromium(l) 64 19. Apparatus for electron exchange experiment . . . . . . . 71 20. Mass Spectrum of titanocene dichloride from electron ex- change solution . . , , , O O O O O O O I O O O O O O O 73 vii INTRODUCTION This is the first report of exchange of n-bonded ligands between identical transition metal complexes, and the first report of photo- substitution reactions of the cyclopentadienyl ligands in titanocene derivatives. Studies of some electron exchange reactions in bis(arene)- chromium systems and between selected titanocene derivatives, and halide exchange reactions of titanocene dihalides, which are not photo- chemical and for which rate constants were not determined exactly, are appended. Ligand exchange processes involve isotopic equilibration between two chemically identical reactant complexes, only one of which init- ially contains the tracer ligand. The present study concerns the equilibration of deuterocyclopentadienyl ligands between (nS-C5D5)2TiCl2 5H5)2TiC12. The system is always at chemical equilibrium, and there is no free energy change except that which arises as the entrOpy and (nS-C of isotopic mixing. Studies of these processes and their kinetics yields fundamental information about molecular properties and reaction mechanisms which may in turn be useful in explaining more complex chemical behavior. Ligand substitution or transfer reactions, which involve chem- ically different reactants and generally have a large free energy of reaction, are important synthetic routes to many transition metal com— plexes, and consequently have been studied extensively. This report concerns the photosubstitution of methoxide for cyclopentadienyl l ligands in (ns—C5H5)2TiC12. Because of the inertness of the titanium to cyclOpentadienyl bond in titanocene dichloride few thermal substitu- tion reactions have been reported, and the only photosubstitution reaction of titanocene derivatives that has been reported involves the sigma bonded methyl groups in dimethyltitanocene, (nS-CSH5)2Ti(CH3)2.l The generalization that has been applied to ligand substitution pro- cesses is that thermal substitution in arene metal tricarbonyl com- plexes usually involves the arene ligand, while photosubstitution usually involves the carbonyl ligands, and in mixed cyclopentadienyl- (arene)-metal complexes, substitution involves the arene moiety.2 Substitution of n-bonded ligands in general has received much attention, and serves well as an historical introduction to the present photosubstitution and photoexchange studies. In 1956, Wilkinson, Cotton and Birmingham reported the first n-ligand transfer reaction, in which ferrocene was prepared from chromocene and ferrous chloride.3 Processes of this type were first termed n-ligand transfers by Maitlis.4 Two comprehensive reviews have recently been published by Avi Efraty5 and A. Z. Rubezhov and S. P. Gubin.6 The general field of ligand substitution has been reviewed by several writers.7’8 Following the paradigm of Wilkinson, g£_§1,, most n-ligand transfer reactions have been studied for their synthetic utility, not for mechanistic detail. However, kinetics of the ethyleneplatinum chloride exchange with deuteroethylene has been studied, and was found to be complete in 15 minutes.9 The rate of exchange of ethylene with Zeise's salt is greater than 70 sec—1,10 and ethylene exchange can proceed rapidly enough to be studied by nmr line—broadening methods.11 Also, the kinetics of n-ligand transfer between n-tetraphenylbutadienyl paladium dihalides and n-cyclOpentadienyliron dicarbonyl bromide has been studied as well as the transfer of the allyl ligand between (n-allyl)palladium chloride and iron carbonyl.12’13 A mechanism that is consistent with the kinetics data for these systems involves three steps that do not necessarily occur in the order given: 1. Coordination of the attacking reagent on the substrate. 2. Rearrangement of n-ligands on the substrate to a lower coor- dination number, with formation of 0 bonds as the extreme case. 3. Ligand removal, possibly concerted with rearrangement of the entering group bonding from o to n bonds. This sort of mechanism is suggested by Rubezhov and Gubin6 and is now supported by independent information. For example, the bonding in tris(cyc10pentadienyl)titanium, where one cyclopentadienyl ligand is considered to have only a 3 electron bond with titanium, might be sug- gested as evidence for step 214 and known 0 to n rearrangements15 in- crease the acceptability of steps 2 and 3. There are only two qualitative reports of n-ligand exchange which are related to the present work. First, Strohmeier16 reported that on untraviolet irradiation of arene tricarbonyls in the presence of the corresponding ll'C arenes, labelled arene metal carbonyls result. Second, King and Bisnette17 reported the potent thermal cyclopentadienylating agent, ((nS-C5H5)Fe(CO)2)2 yields [(CF3)2C282)2Fe on ultraviolet irrad- iation in the presence of the ligand. The kinetics of n-ligand substitution or exchange has presumably been ignored for the titanocene derivatives for several reasons. First, the cyclOpentadienyl ligand is unstable in the uncoordinated state, and complicated reactions might be expected. Mixtures of organic products are found, for instance, in the study of photosubstitution reactions of ferrocene. Second, the n-bond between the cyclOpentadienyl ligand and titanium is very stable, expecially in titanocene dichloride and its analogs. For instance, titanocene dichloride is not subject to thermal alcoholysis and other substitution reactions that occur readily in bis(alkoxo)titanocene derivatives.19 Titanocene dichloride will not yield ferrocene on reaction with ferrous chloride20 even after 8 hours in a sealed tube at 150°. Similarly, no nickelocene results from the reaction of titanocene dichloride and nickel carbonyl.21 Electrochemical reduction of titanocene dichloride destroys the titanium to chlorine bonds, leaving the titanium to cyclopentadienyl bonds intact.22—24 Chemical reduction of titanocene dichloride with aluminum in tetrahydrofuran yields titanocene monochloride dimer.25 Similar reactions occur in water with zinc, and in nonaqueous solvents with aluminum alkyls. The reactions of titanium n—complexes are systemmatically presented in reviews by R. S. P. Coutts and P. C. Wailes26 and by J. M. Birmingham.27 Reduction of titanocene dichloride to lower oxidation states can also be effected with metallic sodium28 or sodium naphthalene29 and a green compound results,which was originally thought to be titanocene, bis(nS-cyclopentadienyl)titanium(II) or its dimer.30 Recently, however, Brintzinger and his coworkers have shown that the compound must be a hydrogen bridged dimer,31 and Davison and Wreford have demonstrated that the structure contains a u—(n5,n5—fulvalene) ligand:32 The elusive "true" titanocene can be prepared only by stirring the polymer ((nS—C5H5)2TiH)x in ether33 or by irradiating dimethyl titanocene.1 This compound is important in the present work because the ring hydrogen atoms exchange with gas phase deuterium to give the necessary tracer compounds. There are two notable exceptions to the rule of the inertness of the bond between cyclopentadienyl ligands and titanium in titanocene dichloride. First, it is cleaved by ammonia and amines.34 Second, the important synthetic route to (nS-C5H5)T1C1 involves a ligand trans- 3 35 fer in boiling xylene: TiCl + (nS-C5H5)2TiCl 5- 4 + 2 (n C5H5)TiCl 2 3 The possibility of photochemical pathways for this reaction should of course be explored. The photochemistry of transition metal complexes is rapidly be- coming a well-developed field, whereas only five years ago authors were 36 lamenting its neglect. Balzani and Carassiti have authored a book devoted exclusively to that subject, which contains references to sources of background information as well as current research. Several review articles have appeared in recent years, including those by A. W. 37’38 A. W. Adamson £5 31.,39 P. D. Fleischauer g£_§1,,40’41 E. L. Wehry,42 J. F. Endicott,43 W. Strohmeier,15 and P. C. Ford.44 Adamson Photochemistry of the metallocenes in particular has been reviewed re- cently by R. E. Bozak.45 The only work on titanocene systems even remotely related to those in the present study was the (previously mentioned) recent discovery of the photodecomposition of dimethyltitanocene.1 While photosubstitution reactions are the main subject of transi- tion metal photochemistry, only a few photosubstitution reactionsl’l7 involve n-bonded ligands. Most attempted photochemistry of n-complexes leads to decomposition, sometimes with intractable products, as in the case of (C4H4)Fe(CO)346 and Zeise's salt.47 Many attempted photochemr ical reactions of ferrocene gave only cyclopentadiene (or organic mix- tures) and iron metal.48 Likewise, the photolysis of bis(benzene)- chromdum(I) yields Cr(III) and benzene.45 The present study therefore makes inroads into a little known type of reaction (n-ligand photoexchange and photosubstitution) and involves a system (titanocene dichloride and its analogs) which deserves further photochemical study. Complexes related to titanocene dichloride are, of course, extremely important in the areas of catalysis and nitrogen fixation. EXPERIMENTAL A. General All materials were manipulated under argon or vacuum in Schlenk type apparatus.49 Argon was purified by passage through, first, a 70 x 5 cm column of Aquasorb ,50 then a 70 x 6 cm column of BTS catalyst51 in its reduced form, heated to 180°, and finally, through a second 70 x 6 cm column of Aquasorb. When necessary, manipulations were executed in an argon filled glove box. The argon was continually recirculated through a column of Dow Q-l "oxygen grabber" and Linde 4A molecular sieves. The inert atmosphere was typically pure enough so that a 25 watt tungsten lamp with an exposed filament would burn for 300 hours. B. Purification of Materials Titanocene Dichloride: The impure titanocene dichloride obtained from Alfa Products was recrystallized in a Soxhlet apparatus from chloroform saturated with HCl, under HCl and argon. Finally, the titan- ocene dichloride was sublimed at 0.1 mm pressure and 100-1400, and stored in the dark. Benzene: Benzene was purified by shaking with concentrated sul— furic acid several times, washing with aqueous sodium hydroxide, dry- ing over calcium hydride, then distilling from sodium metal and benzo- phenone. Valerophenone: Valerophenone was obtained from the J. T. Baker Chemical Company. The impure liquid was passed through an alumina column, then vacuum distilled. Methanol: Methanol was stirred over calcium hydride overnight and distilled through a 30 cm Vigreaux column under argon. Residual air was removed by the vacuum freeze-thaw technique. Deuterium: C. P. deuterium gas was obtained from Matheson Gas Products and used without further purification. C. Preparations Perdeuterotitanocene dichloride: The exchange of the ring hydrogen atoms on titanocene with deuterium, discovered by Marvich,52 was utilized to prepare the perdeuterotitanocene dichloride tracer. Titanocene hydride polymer, ((nS-CSH TiH)x, (0.3 g) was added to 20 5)2 ml of distilled toluene in a 250 ml rould bottom flask containing a Teflon coated magnetic stirrer. The suspension was stirred for several minutes, which was usually adequate for the formation of (nS-C5H5)2Ti solutions. The flask was then partially evacuated and deuterium gas was added at one atmosphere pressure. The solution was then stirred overnight, and the flask was evacuated and refilled with deuterium several times to ensure complete deuteration. Finally, the solution was cooled in a dry ice/acetone bath, the flask was partially evacuated, and hydrogen chloride was added gradually while stirring was continued. Hydrogen chloride was added until the pressure in the flask reached one atmosphere and no gas evolution was evident. The reaction mixture was then allowed to warm to room temperature, and was stirred for ca, one hour. The toluene was then evaporated, and the product extracted with air saturated chloroform which was subsequently evapor- ated to leave reddish crystals. These were sublimed to give pure (nS-C5D5)2TiCl2 as indicated by the melting point and mass Spectrum. Titanocene hydride polymer: The gray polymer ((nS-CSH5)2TiH]x was prepared from (ns-C5H5)2Ti(CH3)2 which in turn was prepared accord- ing to the method of Clauss and Bestian.53 This method involves treat- ing titanocene dichloride with methyllithium (in this case a 2 M sol- ution in ether) at -70°. The dimethyltitanocene thus obtained as a £3. 0.1 M_solution in hexane is stirred under hydrogen gas at 00 to give the 54 polymer as reported by Brintzinger gt al. Titanocene monochloride (bis(us—cyclopentadienyl)titanium chlor- ide: A solution of about 5 g of titanocene dichloride in 50 ml of tetrahydrofuran was treated with aluminum foil (activated by treatment with mercurous nitrate solution) after the method of Coutts, Wailes, and Martin.55 The excess aluminum was removed by filtration, and the fil- trate was evaporated. The residue was extracted with ether to remove the aluminum halide, then vacuum sublimed to give the pure, greenish 5- brown ((n C5H5)2TiC1) 2. D. Photolysis Apparatus Photolyses were effected with a 450 watt medium pressure mercury lamp, Hanovia #679 A 0360, which was mounted in a quartz cooling jacket through which tap water circulated. This assemply was mounted inside a 70 mm quartz filter solution cell which contained 0.002 M potassium chromate in 1% aqueous potassium carbonate. The path length of the filter solution was 1.0 cm so that the 313 nm line of the lamp was ef- fectively isolated.56 The lamp assembly was mounted at the center of 10 a merry-go-round apparatus57 to assure the uniform illumination of all samples and actinometer solutions (see Figure l). The distance from the center of the lamp to the sample cell front surface was 8 cm, and the merry-go-round window size was 7 x 19 mm. The entire apparatus was mounted in an alcohol—water bath to moderate temperature fluctuations, and the temperature stayed in the range 25:20C. The bath was 20% by volume alcohol to retard the growth of microorganisms. Samples and actinometer solutions of a precisely known volume (typically 3.0 mJ) were added to 13 mm o.d. pyrex culture tubes by use of a syringe in the glove box. The necks of the tubes were constricted to facilitate sealing with an oxypropane torch. The tubes were stOpper- ed in the glove box, then removed so that they could be frozen with liquid nitrogen and sealed. E. Chemical Actinometer Solutions The valerophenone actinometer56 was chosen because it absorbs strongly (e = 50 l/mole cm) at 313 nm and the quantum yield is 0.33 mole/einstein, similar to that of the experimental samples. The con- version 0 0 II II C hv C \ \ ©/ C4“9 ——> CH3 was kept to less than 25% of the total valerOphenone to prevent product interference. Actinometer solutions were typically prepared by weighing valer- Ophenone (1.80 g) and tetradecane (0.1000 g) into a 50 ml volumetric ll H o H o 2 2 our y r)? IN F k- 3 -II rCOOLlNG JACKET F F FILTER SOLUTION CELL SAMPLE POSITION 2-: wmoow l I. .- . - -- . HANOVIA 679A 0360 \I,. LAMP fiN‘ Ib— ‘l/K2C704 FILTER SOLN. UI ! r— 5 -cl I 7 l Figure l. Photolysis apparatus. 12 flask and diluting with purified and deoxygenated benzene. The actin- ometer solution was sealed in 13 mm pyrex culture tubes so that all dimensions were identical to the experimental samples. F. Determination of Lamp Intensity After photolysis, the actinometer solution (23, 1 ul) was injec- ted onto the column of a Varian Aerograph Series 1200 gas chromatograph with a flame detector. The column material was 4% QFl, 1.2% carbowax 20 M on Chromasorb G 60/80, and it was maintained at 100°C. An Infotronics CRS 208 Automatic Digital Integrator was used to facilitate the determination of the peak area ratio of the acetophenone to the tetradecane peaks. The number of moles of photons, n, in einsteins can then be calculated from the following equation: I Iacet0phenone x 2.0 x [tetradecane] x V = 0.33 n tetradecane where I = the peak area integral for the acetophenone acetOphenone peak I = the peak area integral for the tetradecane tetradecane peak 2.0 = an empirical correction factor depending on the column material [tetradecane] = the concentration of tetradecane in the actinometer solution V = the actinometer sample volume and 0.33 = the quantum yield for the formation of acetophenone. 13 G. Photolytic CyclopentadienylgLigand Exchange Between Titanocene Di- chloride and Perdeuterotitanocene Dichloride Weighed amounts of titanocene dichloride and perdeuterotitanocene dichloride were stirred with benzene in 50 ml volumetric flasks in the glove box until the crystals were totally dissolved (this sometimes required several hours). The solutions were diluted to the mark with benzene to give the desired 10-.2 to 10-3 M solutions, which were then mixed in dim light and added in 3.0 ml aliquots to the 13 mm culture tubes to make about 30 samples. One sample was retained to determine the initial optical absorbance at 520 nm. The tubes were sealed with an oxypropane torch and mounted, along with 6 actinometer cells, in the merry-go-round apparatus. Generally, the lamp was turned on and allow- ed to warm up for at least an hour before the samples were introduced. After measured periods of illumination, 4 or 5 sample tubes and one actinometer tube were withdrawn from the merry-go-round. The actin- ometer was analyzed as described previously. The samples were pooled, evaporated to dryness, and sublimed (the sublimation step was used mainly as a collection device, as very little impurity remained unsub- limed. The mass spectrum was then obtained on a Hitachi Perkin Elmer EMU-6 mass spectrometer. The absence of side reactions was established by measuring the optical absorbance at 520 nm of a sample which had been photolyzed the maximum time. H. Photolytic Cyclopentadienyl Ligand Exchange Between Perdeutero- titanocene Dichloride and Titanocene Monochloride Dimer Weighed amounts of perdeuterotitanocene dichloride and titanocene monochloride were dissolved separately in benzene to give 50 m1 of each solution. The samples were then treated as above, except that after l4 photolysis, the titanocene monochloride was separated from the exchange mixture by precipitation with 2,2'-—bipyridine.58 The blue precipitate coagulated in about a minute to yield a mixture which was easily separ- ated by filtration through a fine frit. The filtrate was then evapor- ated and the residue sublimed as above. The bipyridine adduct was con- verted to titanocene dichloride by dissolving it in a minimum of oxygen free, deionized water to give a blue solution. The solution was treat- ed with excess hydrogen chloride at 0°, then stirred at room temperature for several hours under HCl until it turned orange. The water was then evaporated and the bipyridine sublimed from the product at 950 and 0.1 mm pressure. Finally the product was sublimed and the mass spectrum was obtained. As in the previous experiment, the absorbance of the exchange solution at 520 nm was measured before and after photolysis to demon- strate that no decomposition had occured. I. McKay Analysis of Mass Spectrographic Data for Ligand Exchange Processes Titanocene dichloride has a well characterized mass spectrumsg-61 with a parent ion peak at m/e = 248. The mass spectrum of the perdeu- tero compound is, of course, identical except that the peaks correspon- ding to fragments with one or two cyclopentadienyl moieties appear 5 or 10 units higher, respectively. Figures 2 and 3 show the mass spectra of the two derivatives. In the mass spectrum of an unphotolyzed mix- ture of titanocene dichloride and its perdeutero analog there are peaks at 248 and 258 with associated manifolds due to the isotopes 3701, 35Cl, 46Ti, 47Ti, 48Ti, 49Ti, 50Ti, 120 and 13C. The peak at m/e = 253 is nonzero but small (typically less than 5% of the 248 peak). Photolyzed 15 no a. 1. I 1. fl _ _: u du—d E _._ mowuoanowo ocmoocmufiu mo aauuumam mom: n1— n—N .N ouzmfim ova :1 16 ad— oofluoanoao ucoooaoufiuaoalnv mo aauuownm mam: n— o“— ! ___ n __— :1 .m muswwm 17 samples show increased intensity of the 253 peak (Figure 4), which 2+. Samples photolyzed for exten- corresponds to (us-C5H5)(nS-CSD5)T1C1 ded periods of time, typically ten times longer than the time required to reach equilibrium, show no change in the ratio of the m/e = 253, 258 and 248 peaks, and this ratio consequently can be used to indicate the relative concentrations of the corresponding species at t = w. Further- more, the ratio of the m/e = 253 and 248 peaks for various periods of illumination reflects the extent of the reaction which has been com- pleted. The fraction, F, of the exchange process which is complete after a particular time is the ratio of the m/e a 253 and 248 peaks for a sample withdrawn at that time divided by the same ratio for a sample which has reached equilibrium: (1253/248)t (1253/248Jw If 1n(1-F) is plotted ye, time (the McKay62 plot) a straight line will result. The most probable slope of the line was calculated by a standard least squares treatment of the data, and the rate was calcul- ated by using the equation:63 slope = _R [[reactant A] + [reactant B1] [reactant A] [reactant B] where the bracketed quantities represent the concentrations of the in- dicated species. The quantum yield was calculated by using the equation: lull- 50 .l. .1. “Wills all l l ‘L .II “II Ml. 4H1 I u H. a H -_ __ 222222222222 19 Rate (moles/liter hour) Intensity (einsteins/liter hour) It should be noted that this analysis requires the intensity of the lamp to be constant with time. The most probable intercept of the 1n(l—F) axis was calculated by standard linear least squares methods, and the standard deviation of residuals was calculated by using the equation, 8 d2 2 1n(l-F) N-l J. Photolytic Methanolysis of Titanocene Dichloride Weighed amounts of titanocene dichloride were added to about 60 ml of benzene in 100 m1 volumetric flasks in the glove box and stirred until dissolution was complete. Purified methanol was then added by use of a syringe and the solution was diluted to the mark with benzene. The solutions were exposed only to very dim light after the addition of the methanol. As before, the solutions were added in 3.0 ml aliquots to pyrex culture tubes and sealed with an oxypropane torch. One sample was retained for a determination of the Optical absorbance at 520 nm before photolysis, and constancy of this absorbance in the dark for up to one week. Other samples were mounted in the merry-go-round apparatus with actinometers, after the lamp had stabilized. The samples were withdrawn at about 10 minute intervals and the absorbance at 520 nm was measured to determine the extent of methanolysis. A Unicam SP 800 .4... 1A....Ll. l - 55. I_l*¢n4.~n. 20 recording spectrophotometer and matched quartz 1 cm cells were used for all measurements. The actinometer tubes were withdrawn after longer periods of illumination to allow more conversion and thus more accurate integration of the gas chromatogram. Some sample tubes were stored in the dark for varying periods of time and the optical absorbance at 520 nm was redetermined to check for secondary thermal reactions of the photolysis products. RESULTS AND DISCUSSION A. Photolytic CyclopentadienylgLigand Exchange Between Titanocene Dichloride and Perdeuterotitanocene Dichloride A new type of reaction, the exchange of fl-bonded cyclopentadienyl ligands between identical transition metal complexes, is evidenced by the increasing concentration of (D—5)titanocene dichloride in irradiated benzene solutions originally containing (D—lO)titanocene dichloride and titanocene dichloride. Table 1 shows the pertinent peaks in the mass spectrum Of the exchange mixture after selected periods of irradiation. The peaks be- tween m/e = 60 and 72 are included to demonstrate that ligand exchange, and not hydrogen exchange on the ligands, is the process which results in the increase of the peak at m/e = 253 relative to those at 248 and 258. The small fragment and large fragment peaks are normalized separ- ately in the table. The calculated values for F, the fraction of exchange after sel- ected periods of illumination, are also displayed in the table. They are calculated on the basis of the equilibrium value of 2.4 for the ratio of the m/e = 253 and 258 peaks as outlined in the experimental section. The ratio of the m/e = 253 and 248 peaks, of course, gives similar results with comparable scatter. The initial concentrations of titanocene dichloride and perdeuterotitanocene dichloride were 4.4 x 10-3 and 3.9 x 10—3 M respectively. The plot of 1n(l-F) gs, time shown in Figure 5 indicates that the half life for the exchange was 0.83 hour, 21 .u .J .I «I. «HOLIIILtliiflsile 1.41 . I I..|l. .rl 22 O.H OO.O O OH OH HH O O OO OOH OOH OOH HO O.OH OO.O OO.O O OO OO OH O O OO OOH OOH OOH OO O.O OO.O OO.O O OH OH HH O O HO OOH OOH HOH HO O.O OO.O HH.O O OH HH OH O O HO OOH OOH OOH HO O.O HO.O HO.H O HO OH HH O H OO OOH OOH OOH OO O.O OH.O H.H O HH OH HH O O OO OOH OOH OHH OO O.H OO.O O.H O OH OH HH O H OO OOH OOH OO HO O.H OO.O OO.O O OH OH HH O O OO OOH OOH OO HH O.O OOO.O OOO.O O HO HH HH O O HO OOH OOH O HH O.O O OOOHHOOOH OH HH OH OO OO HO OO OO OOO OOH OOH u «Ha OHOO oaHH mwfiuwmcmuaw oONHHmawoz .Owcmcoxu vcmwwa ou«uo~:o«p ocmoocmuOuIOOOuoanowo mcooocmufiuououampuoa man How dump vanamumouuouam mmmz .H canoe Figure 5. 23 (l-F) l l V V 1,0 2,0 3,0 TIME (HOURS) 4b McKay plot for perdeuterotitanocene dichloride- titanocene dichloride exchange. 77. 24 and the corresponding rate, calculated by the method outlined in the experimental section, is 1.8 x 10-3‘M_hr-1. The best intercept, cal- culated by standard linear least squares methods, is 1n(l-F) = 1.1, and the Standard deviation in 1n(l—F) is 0.08. The intensity of the incident radiation was 0.0078 einsteins/hr liter. The quantum yield, which is the ratio of the rate to the intensity, is 0.23 moles/einstein. In a typical experiment, the total concentration of titanocene dichloride was 10-2.M_(and about half of this was the deuterated species). The molar absorptivity for titanocene dichloride at 313 nm is 6 x 103‘M71 cmfl, so the sample absorbance at this wavelength, as- suming a 1 cm cell path length, was about 60, and all incident light was absorbed. The actinometers were 0.107 M in valerophenone, which has a molar absorptivity of SO‘Mfl cmfl. Again assuming a path length of 1 cm, the absorbance at 313 nm was 5, SO all incident light was absorbed within the limits of experimental error. Conversion of valerophenone to acetOphenone was never allowed to exceed 25%. The average intensity of the lamp, calculated from the degree of conversion of several actinometer samples, was 0.0078 einsteins/liter hr. The actinometers were exposed over different periods of time before, during, and after the sample photolysis, and established that the inten- sity was constant over the photolysis period. The equilibrium constant for the reaction, 5_ 5_ 5- 5- (n CSD5)2TiCl + (n C5H5)2TiCl + 2 (n C5H5)(n CSDS)TiC12 2 2 can be calculated from the intensities of the indicated peaks in the 25 mass spectrum of the equilibrated mixture by using the equation, 2 (1253) (1248)(1258) The empirical value of 4.0 agrees with that calculated65 by assigning symmetry numbers to the products and reactants and calculating the en- tropy of mixing. Since there is little or no enthalpy change for the exchange reaction, the free energy change is due to the entropy change, and the equilibrium constant can be calculated from it by simple argu- ments involving only classical thermodynamics.65 B. Photolytic Cyclopentadienyl Ligand Exchange Between Perdeutero- titanocene Dichloride and Titanocene Monochloride Dimer The ligand exchange between titanocene dichloride and titanocene monochloride may proceed by two mechanisms. The first is direct ligand exchange, as shown in Figure 6. The rapid oxidation state interchange illustrated in Figure 7 creates a second pathway, the ligand exchange between titanocene dichloride and itself. The mechanism illustrated in Figure 7 can be expected to contribute after a maximum induction time of less than a minute (see Appendix C). The kinetics of the overall ligand exchange reaction (via both mechanisms) was studied only briefly because it does not add significant novel information to that obtained from the titanocene dichloride - perdeuterotitanocene dichloride system, and it is much more complicated. 3 The data in Table 2 for the exchange in a solution initially 4.3 x 10- .M in perdeuterotitanocene dichloride and 2.6 x 10-3 M in titanocene 26 O O O O O O HO O O ./ \ / \ HO/ \HO/ OOOO\ HO OOOO O O O O O O O O /« Ha\ o/H H\ OOOO\H/H OH\ .H/OOOO Amuoaoma my ammo Ho memo / \ / \ \HO/ \HO/ OOOO HO OOOO O O O O O HO O / \ / \ HH/ \HO/ OOOO\ HO OOOO Opauoanoouoa mauooawuau vow OOHuoanOOo oomooamuwu cmmsuon owamnoxo oawwua wow smwamnooa uoouwa .o muowfim 27 ooauoanooaoa maoooaoOOu paw mouuoasowv wamooamuau ommzuon owcmsoxw vcmwfia you amwownoua mwamnoumuca Oumum doaumvfixo .m ouowam OO OO . OO HO O O HO O O HO O O / \ I / \ / \ \HO/ O .. HO/ + HO. HO OOOO HO\ OOOO HO\ /OOOO O O O O O O O O O O O O HO O O O O HO O O HO O O / \ / \ I / \ / \ / \ + HO HO .. HO. HO. + HO. I / OOOO\ /HO\ OOOO OO\ / \ /OO \ OO 28 Table 2. Mass Spectrographic data for the perdeuterotitanocene dichloride-titanocene monochloride ligand exchange Fri Irradiation Normalized peak intensities Lamp intensity* F j time (min) m/e - 248 253 258 (ei/l x1000) ; 10 1.1 0.31 1.0 1.2 0.17 I5 1.3 0.37 1.0 20 1.5 2.3 1.0 2.4 1.0 1.3 2.2 1.0 30 1.3 2.1 1.0 3.6 1.0 70 1.7 2.3 1.0 8.4 1.0 *Estimated from the calculated average intensity of 7.2 x 10-3 einstein/ liter hr over 100 minutes of actinometer photolysis time which spanned the sample irradiation period. 29 monochloride dimer clearly indicate that a simple second order treat- ment is not apprOpriate. The equilibrium constant was not calculated because most of the species present at equilibrium were indistinguishable by mass spectra- metry. The ligand exchange between titanocene dichloride and titano- cene monochloride has a larger entropy change than the ligand exchange be- tween titanocene dichloride and itself because a mixture of several 5—; isotOpes is produced. The equilibrium constant for the former reaction would be larger than that for the latter because it increases as the magnitude of the entropy change increases. The data in Table 3 were obtained for a solution which was ini— VI tially 4.0 x 10-3 M in perdeuterotitanocene dichloride and 2.3 x 10-3 _1V_l_ in titanocene monochloride dimer. The solution was photolyzed for 27 Table 3. Mass Spectrographic data for the perdeuterotitanocene dichlor- ide - titanocene monochloride equilibrium mixture. Sample Normalized peak intensities m/e = 248 253 258 Titanocene dichloride separated 1.7 2.5 1.0 from the exchange solution Titanocene monochloride separ— 2.0 2.5 1.0 ated from the exchange solution and converted to titanocene dichloride hours, but no variation in the peak ratios occured in samples taken at 0.75, 1.5, 2.3, 3.0, 4.0, and 27 hours. The incident intensity was ap- proximately 4 x 10-.3 einsteins/liter hr. No decrease in the absorbance at 520 nm.was observed for the photolyzed samples. 30 The air sensitivity of the titanocene monochloride bipyridyl ad- duct precluded a direct measurement of the extent of substitution by deuterated ligands by mass spectrometry. To confirm the exchange pro- cess, the titanocene monochloride bipyridyl adduct separated from the exchange mixtures was converted to titanocene dichloride, for which the mass spectrum could easily be obtained. Information was thus necessar- ily lost on the total number of deuterated ligands on the dimer. fr C. Photolytic Methanolysis of Titanocene Dichloride Extensive work has been done on the substitution reactions of titanocene dichloride by A. N. Nesmeyanov and his coworkers.66w7S It was found that the reactivity toward ferrous chloride (exchange of the n5—cyclopentadienyl ligand to give ferrocene), toward water (hydrolysis of the nS-cyclopentadienyl to titanium bond), and toward alcohols (alco- holysis of the n5-cyclopentadienyl to titanium bond) increases as the electron withdrawing tendencies of other substituents on the titanocene derivative decrease. Consequently, titanocene dichloride reacted with ethanol only to a small extent, even after eight hours at reflux tem- perature,68 while titanocene bis-acyl complexes react smoothly to pro- duce cyclopentadiene and (nS-cyclopentadienyl)(dichloro)(ethoxo)titan- ium(IV) .72 Because of the inertness of the cyclopentadienyl to titanium bond in titanocene dichloride, the synthetic route to compounds such as (nS-cyclopentadienyl)(dichloro)(alkoxo)titanium(IV) has been either the reaction of (nS-cyclopentadienyl)titanium trichloride with hot methanol,76 the reaction of (nS-cyclopentadienyl)tris(alkoxo)titanium complexes with acetyl chloride,67’69 31 S- 5 (n C5H5)Ti(OR)3 + 2 CH3COC1 + (n -C5H5)Ti(OR)Cl2 + 2 CH3COOR S 5- (n -C5H5)Ti(OEt)3 + CH3COC1 + (n C5H5)Ti(OEt)2Cl + CH3COOEt S_ 5- (n C5H5)Ti(OR)3 + 3 CH3COC1 + (n CSH5)TiCl3 + 3 CH3COOR or by the redistribution reaction between cyc10pentadienyltitanium tri— chloride and cyc10pentadienyltitanium triethoxide in a 2:1 ratio,73’74 5_ ' 5_ . S 2 (n C5H5)T1Cl + (n C5H5)T1(0Et)3 + 3 (n —C5H5)Ti(OEt)C12 3 Alcoholysis of the titanium to chlorine bond in titanocene di- chloride can be induced by the presence Of a tertiary amine,68 + ROH + (C H 5_ S . (n C5H5)2TiCl 2 5)3N + (n -C5H5)2Ti(OR)Cl + (C2H5)3N HCl 2 The bis(cyclOpentadienyl)(alkoxo)(chloro)titanium complexes are usually prepared by the action of sodium cyclopentadienide on alkoxo- titanium halides,67 5_ . Ti(OEt)C13 + Na(C5H5) + (n CSH5)2T1(OEt)Cl Ti(OEt)2Cl + Na(C5H5) + (nS-C5H5)Ti(OEt)3 + (nS-CSH Ti(OEt)C1 2 5)2 The reaction between (us—cyclopentadienyl)(dichloro)(methoxo)ti- tanium(IV) and sodium cyclopentadienide, however, does not yield the desired bis(nS—cyclOpentadienyl)(chloro)(methoxo)titanium(IV).76 The present work thus,provides a much-needed facile route to the (nS-cyclopentadienyl)(dihalo)(alkoxo)titanium(IV) complexes. The quan- tum yield is large, especially for solutions with high concentrations of both alcohol and titanocene dichloride, and gram quantities of pro- duct can be prepared in about an hour in a suitable preparative cell. This is also a convenient route to the (nS-cyclopentadienyl)titanium trihalides, which can be prepared from the (nS-cyclopentadienyl)(di- chloro)(alkoxo)titanium(IV) complexes: 32 S_ 5 (n C5H5)T1(0R)C12 4' HC]. -> (n -CSH5)T1C13 + HOR In the photolysis experiment, the lamp was allowed to stabilize for one day before the actinometer tubes were mounted in the merry-go- round (at time = 0). Methanolysis samples with concentrations of reac- tants as shown in Table 4 were mounted at 1800 seconds, then removed at 4200 and 6100 seconds. The actinometers were removed after 7500 and 10,400 seconds, so that they were exposed long enough to give an accurately measurable conversion. Table 4 shows the initial absorbance at 520 nm, and the absorbance of the solutions after each photolysis period. Control samples which were stored in the dark for 36 hours showed no change in absorbance at 520 nm, and photolyzed samples which were stored in the dark for 36 hours showed no change in absorbance from that measured immediately after photolysis. Thus no thermal reaction occurred at room temperature, and no secondary thermal reaction of the photolysis product occurred. Since the molar absorptivity of titanocene dichloride is 6 x 103 and the initial concentration of titanocene dichloride was always greater than 4 x 10-3 M, the initial absorbance was greater than 24, and all incident light was absorbed. Even after 4300 second photolysis per- iods, the titanocene dichloride concentration was at least 7.7 x 10-4 M5 so that the absorbance was 4.8 at 313 nm, and again all light was absor- bed, within experimental error. Neither benzene nor methanol absorb at 313 nm although they both absorb strongly below 280 nm. The product does absorb at 313 nm, and since it is produced in concentrations great- er than 2 x 10-3.!.1t might be expected to reduce quantum efficiency. The fact that the decrease in absorbance of the photolyzed samples was 33 OOOHOQ sowuomou ocooow coem onu so oommmxe sumo Ono SO muoon on woumIu0HOO>muuHD .HH owowwm E: .LH0:O_O>OZ 03 Dow onv oov onm . oom H H H H H H H. H 0n; p23:p 39 bond.45 The titanium to cyclopentadienyl bond may, of course, be broken in the present system, but this is unlikely for several reasons: First, this would probably lead to decomposition and intractible photo- lysis products. In ligand exchange experiments, however, decomposition was negligible compared to the primary process. Furthermore, a bound cyclopentadienyl radical would tend to abstract a proton from a methan- ol molecule attacking adjacent to it, while a dissociated cyclopenta- dienyl radical would tend to react with other dissociated radicals. Second, the titanium to cyclopentadienyl bond is a multiple bond and it is particularly inert to substitution in titanocene dichloride. Reduc- tion of bond order, rather than total bond scission, would be a more reasonable hypothesis with regard to reaction energetics. Third, ther- mal processes in which the nS-cyclOpentadienyl ligand probably undergoes a conversion to a o-bonded,nl-cyclopentadienyl intermediate are known for some titanocene species.80 Brintzinger and Bartell81 have shown that an electronic transition from ligand e n-orbitals to a titanium lg lg orbital leads to a distortion of elg symmetry.* A n5+n1 distortion has this symmetry. Thus the electronic transition which probably re- a sults from irradiation with 313 nm light is one which promotes a n5 to n1 bonding rearrangement. In light of the present experimental results, it is reasonable to propose reduction in bond order from n5 (to n} as a limiting case) as the first step in both the methanolysis and ligand exchange reactions of titanocene dichloride. The coordination number of titanium is thus reduced, allowing the attack of the entering group. The suggested *Notations appropriate to D symmetry were used because the titanocene 5d system was treated as a perturbed ferrocene system. 40 mechanism is illustrated in Figure 12. It must be noted that reduction of the coordination sphere may not be necessary for ligand substitution. Titanocene monochloride forms a 1:1.5 bipyridine adduct which must have a coordination number of at least 5, while Coutts and Wailes58 suggest a coordination number of 6. Attack of the entering group could thus occur before, or in concert with, cleavage of the leaving group bonds. These pathways are probably Fm“ experimentally indistinguishable from the suggested one. 41 reu + .ooHHoHnoHo ocmoocmuHu mo mcowuomou OOOOHoOoca OOO EmOcocoms ommoooum /; II” .fio/.\@ _._\/_. ©\\/ .6 Hm? 6“— .OH OAOOHO .. Ox. @3. /_ 1. Q7. 1 @m... SUMMARY The present research has demonstrated the importance of photo- lytic processes in the chemistry of titanocene dichloride. Specific— ally, the ns-cyclOpentadienyl ligand exchange between molecules Of titanocene dichloride and between titanocene dichloride and monochlor- ide, and the methanolysis of titanocene dichloride are photolytic. A mechanism based on photolytic reduction of the bond order of the us- cyclopentadienyl to titanium bond has been proposed as the initial step of these processes. The discovery of the photolytic activity of titanocene deriva- tives indicates a new dimension which must be controlled in all re— lated work, and which, it is hoped, will be exploited to yield novel results in further work. 42 APPENDICES APPENDIX A THE HALIDE EXCHANGE BETWEEN TITANOCENE DICHLORIDE AND TITANOCENE DIBROMIDE A benzene solution initially 3.7 x 10-3‘M_in titanocene dichloride and 1.4 x 10-3‘M_in titanocene dibromide contained the mixed halide complex in the same molar ratio after 18 hours irradiation as it did when immediately frozen in the dark and evaporated to dryness. Thus the halogen exchange is rapid and thermal. An equilibrium constant of 0.5 was found for the reaction, 5_ 5_ + s_ (n C5H5)2TiCl2 + (n C5H5)2TiBr2 + 2 (n CSH5)2TiClBr Experimental Titanocene dibromide was prepared by a new method, which involved treating the titanocene hydride polymer54 in toluene with HBr at -70c. The sublimed product melted at 305°C. All other materials were prepared and/or purified in the usual manner. Separate solutions of titanocene dibromide and titanocene dichloride in about 75 ml of benzene were mixed in the dark. One aliquot (23, 30 ml) was frozen and the solvent was sublimed in the dark, and the mass spectrum of the titanocene derivative was obtained. The remainder of the exchange solution was sealed into pyrex culture tubes in the usual manner and mounted in the photolysis apparatus. The samples were photolyzed for up to 18 hours (over which time the intensity was about 0.010 einsteins/liter hr), evaporated to dryness, and the mass spectrum was obtained. 43 44 Discussion The mass spectra of the photolyzed and unphotolyzed samples were identical. The main peaks are shown in Figure 13. It appears that halide ligands on titanocene dihalides are thermally labile, in con- trast to the cyclopentadienyl ligands. The equilibrium constant was calculated from the intensities of the product and reactant peaks in the prOper ratio. The sum of the FF; intensities of all peaks in the peak manifold of the species of inter— est was taken to be proportional to the concentration of that species, and an equilibrium constant of 0.5 was calculated for the reaction: I t, 5_ 2 [(n C5H5)2TiC1Br] . S 5 [(n -C5H5)2T1C12] [(n -C5H5)2TiBr2] 45 .n'. _ .. Vi}... us I... I'Tr .uoooown ‘ oncogene oowau: onu mo anuuuomm now: .1008 19 an“ v00 .v—N kw #310216: a :mrt) .OH .OOOHH can APPENDIX B THE ELECTRON EXCHANGE BETWEEN BIS(ARENE)CHROMIUM COMPLEXES AND THEIR CATIONS I. Introduction Recent interest in measuring the rate of electron exchange between sandwich complexes and the corresponding cations (obtained by one elec- tron oxidation of the complex) has been intense. The electron exchange between ferrocene and the ferricenium ion, T+THT.T l as 1 OH is typical of this class of reaction, and has been studied extensively. The first study of electron exchange in a sandwich complex system was done by Voigt and Katz in 1958.82 The exchange between bis(cyclo- pentadienyl)cobalt(II) and bis(cyc10pentadienyl)cobalt(III) was found to be "very fast." In 1960, D. R. Stranks did a tracer study of the ferrocene-ferri- cenium exchange by using stopped flow techniques. By cooling methanol solutions of the reactants to -70°, he was able83 to estimate the ex- change rate constant at 1.7 x 106‘Mf1 sec-1. In 1963, Dietrich and Wahl studied the same system by nmr line- broadening techniques.84 They established that the rate constant for 1 - —l the reaction in acetone at room temperature was greater than 10 M; sec . 46 47 The acceleration of the diffusion of ferricenium ions through a solution of ferrocene in alcohol by the electron exchange process was used by I. Ruff, g£_§l, in 1971 to determine a rate constant of greater than 1.5 x lOg‘Mfl sec.1 at room temperature.85 Recently, Espenson and Pladziewicz used the Marcus relation for outer sphere electron transfer to calculate the rate of self-exchange between ferrocene and ferricenium from data on the rate of mixed exchange reaction (i. 5. between bis(methylcyclopentadienyl)iron(II) and ferricenium cation). These experiments were done in tetrahydrofuran- 6 l -1 water mdxtures and yielded a calculated value of 5.7 x 10 ML- sec for the rate constant for electron exchange between ferrocene and ferri- e cenium ion.86 The results on the ferrocene system are not consistent. The esr linebroadening technique cannot be applied to the ferrocene system be- cause of the short relaxation time characteristic of the ferricenium cation87 so further corroboration of the results above seems unlikely. The isoelectronic and isostructural bis(benzene)chromium(0) (ab- breviated bZZCr) - bis(benzene)chromium(1) [abbreviated szCr(I)] system escapes the difficulty of short relaxation time, and is a tantalizing system for electron exchange studies for this and a myriad of other reasons: Molecular orbital bonding schemes are highly refined.88-90 A polarographic wave is obtained at -0.97 V 223 SCE regardless of the direction of approach, which reflects the necessary reversibility.91-94 Methyl derivatives, like bis(toluene)chromium, have more negative re- duction potentials, while phenyl derivatives, like bis(l,3,5-triphenyl)- 95-97 chromdum, have more positive reduction potentials. Good theoreti— cal correlation is obtained between the reduction potential and the 48 localization of the highest occupied molecular orbital (HOMO) predicted by substituent effects,98 and correlations with exchange rates would be interesting. The core sandwich structure remains essentially the same through the series of derivatives,99 so free energy changes related to electronic reorganization and inner sphere solvent effects can be elim- inated for theoretical treatment. The fact that one of the reactants is neutral allows further simplification in theoretical calculations of the 3‘ exchange rate, since coulombic interaction is nonexistent. Theory has . not dealt extensively with ion-molecule reactions in solution. 5 Finally, the electron exchange would necessarily be simple, because associated chemical reaction of the ligands, ion transfer, or atom k transfer are unlikely. The chemistry of szCr(0) and szCr(I) is well 100, 101 102-104 known, and has been reviewed several times. The ultraviolet and infrared105 spectra have been analyzed. The structure has been 106-110 studied thoroughly and the symmetry established as D6 The esr h. spectrum is reminiscent of those of organic arene radicals, with 3.6 G hyperfine due to arene protons and similarly narrow hyperfine due to 53 111-113 Cr. For these reasons, the electron exchange between szCr(0) and szCr(I) was studied, first by tracer techniques which established a lower limit on the rate constant, second by ear linewidth measurements, and finally by the nmr technique. 49 II. Quenched-Flow Study of the Electron Exchange Between Bis(benzene)- chromium(0) and Bis(benzene)chromium(l). A quenched—flow experiment established the lower limit of 4 x 103 —l -1 M sec for the rate constant for the electron exchange between szCr(0) and szCr(I) in dimethyl sulfoxide. A 1.0 x 10.2 M solution of szCr(O) was rapidly mixed with a 9.8 x 10—3 M solution of his- (benzene)chromium iodide, bZZCrI, containing 51Cr at a specific activity of 5.3 x 105 Cpm/mmole. The molecule was separated from the cation g after a 0.07 second reaction time by spraying the exchange solution into water at 0°C. The bz2Cr(0) precipitated immediately and was removed by filtration. The specific activity of the initially inactive complex indicated that equilibrium with respect to exchange had been reached. It was found that szCrI undergoes Szilard—Chalmers reactions when it is bombarded with neutrons. The product of the Szilard-Chalmers reaction does not oxidize bZZCr(0), so electron transfer experiments run with tracer prepared by neutron bombardment of szCrI indicate that no exchange occurs. A true measurement of the exchange rate between szCr(0) and bZZCr(I) can be made only if the tracer is prepared from 51 CrC13. Experimental A. Preparation of Materials The szSICrI tracer was prepared from anhydrous chromium chloride which was bombarded £3, 1 h with a neutron flux of 2 x 1012 neutrons/cmz- sec in the MSU/TRIGA Reactor to give an activity of 2.8 x 105 cps/mg. The high chlorine and 55Cr activity which resulted was allowed to decay before the samples were handled. The szCrI was prepared by sealing 50 the active 51CrCl in a pyrex tube of about 100 ml capacity with sub- 3 limed AlCl3, aluminum powder, and benzene. The tube was rotated in an oven following the Friedel Crafts preparative method of Fischer}14 It was purified by recrystallization from an aqueous solution of KI. The zerovalent bz Cr was prepared by a similar method, except that a final 2 step involving reduction of the product with Na28204 was required,115 and the product was purified by vacuum sublimation. F1“ Dimethylsulfoxide was vacuum distilled from Ca0 at low temperature to avoid decomposition. Water was distilled from KMhOA and then vacuum distilled to remove air. B. Gamma Counting Procedures Solvent was evaporated from or added to samples obtained from the separated exchange mixture to yield solutions of about 1 ml volume, and they were sealed in 15 mm pyrex test tubes. The 0.32 MeV gamma radiation from the 51Cr (t1/2 - 27.8 d) isotopic tracer was counted by use of a sodium iodide scintillation detector (Packard Model 1212 WSP) and a Nuclear Data Series 130, 512 Channel Analyzer/Computer that automatically integrated the channels correspon- ding to the sample peak to give an output of total counts (sample plus background) for those channels. The background activity was calculated by adding the counts registered by the three channels before the first of the integrated channels, and the three after the last of the inte— grated channels and dividing by 6 to give an estimate of the average number of background counts per channel. This average was multiplied by the number of channels integrated to give the total background count and was subtracted from the total counts output. 51 About 104 counts were recorded for most samples, and the measured activity was adjusted for decay between collection time and counting time. C. Quantitative Analysis of Chromium Cr and bz CrI were washed The dimethylsulfoxide solutions of bz2 2 from the counting ampoules with 2 ml benzene, 2 ml methanol, 1 m1 con- Fe centrated H2804, and 4-5 ml concentrated HCl. The solution was heated I for several hours to digest the organometallics (and most of the DMSO). The samples turned green in this time. Sodium hydroxide (5 M) was then added to make the samples basic, and 10 ml of 30% H202 was added. The r samples turned yellow immediately as the chromium was oxidized to Cr(VI). ‘ They were boiled for 1-2 hours to destroy excess peroxide, diluted to known volumes, and added to quartz, 1 on cells so that the absorbance at 372 nm could be accurately measured on a Cary Mbdel l4 SpectrOpho- tometer. A calibration solution was prepared by dissolving 0.0303 g KZCrZO7 in 0.05 MLKOH to make 1.00 liter of solution with an absorbance of 1.0. D. The Szilard-Chalmers Reaction of Bis(benzene)chromium Iodide BzZCrI prepared from 51CrCl3 was split into two samples, one of which was subjected to further neutron bombardment. Both samples were dissolved in air free water, and added under nitrogen to 15 x 1 cm columns of Dowex 50WX4 cation exchange resin. Dilute, air free hydro- chloric acid was added to the columns as needed, and aliquots of about 1 ml volume were collected. The sample subjected to the second bombard- ment separated on the column to give a high specific activity fraction (aliquots 25 - 28) as well as a fraction common to both samples 52 (aliquots 36 - 40) which had a specific activity nearly identical to that of the bz CrI prepared from 51CrCl . 2 3 If DMSO solutions of neutron-bombarded szCrI were added to the exchange column without strict air exclusion, nearly all the activity was retained in the first few centimeters of resin even after 50 ml of elutate was collected. This top layer of the column turned green slow- ly during the elution. A yellow solution, identical in appearance to “7’1 bZZCrI solutions, but which contained no chromium, was obtained in aliquots 27 - 75 Of elutate. I E. Quenched-Flow Experiments r!“‘ A diagram of the quenched-flow apparatus is shown in Figure 14. A pneumatic plunger, driven by compressed air at 50 psi operated two sample syringes which could be filled with the reactant solutions under nitrogen. The mixing chamber was designed to give high turbulence. The time of the plunging stroke (0.4 seconds) was measured by using a Heath Universal Analog Computer with a Heath Timer MOdule which was activated by microswitches at the extremes of plunger travel. The reaction time was calculated as follows: Since the apparatus delivered 15.6 ml of solution in 0.4 seconds, it delivers 39 mllsec. The volume of the ap- paratus from the mixing chamber to the outlet in the quench solution was 2.8 m1. It follows that the reaction time was 2.8/39 - 0.07 seconds. Schlenk flasks containing the DMSO solutions of reactants under nitrogen, and the quench of air-free water, were attached to the appar- atus without admitting air. The plunger was activated to mix the reac- tants and spray them into the quench. The quench mixture was immediate- ly filtered, and the respective samples were prepared for counting as 53 N2 2%: N2 QUENCH V s I .. ‘ N N2 Y _ _ 2 ’0: W \IIJ REACTANT A REACTANT B PLUNGER Figure 14. Quenched—flow apparatus. 54 previously detailed. A tco sample was obtained by activating the plunger to deliver a sample which was not separated (quenched). Instead, this sample was subjected directly to counting and analysis to determine the total'amount of chromium and the total activity. From these totals, a specific activity can be calculated which should be equal to the specific activity of either species in the exchange solution at equil- ibrium. Discussion The rate constant for the electron exchange can be calculated from the standard equation, 0.693 (ti/2) ([szCr(O)] + [szCr(I)]) In the exchange experiment, the sum of the concentrations of the reac- tants, ([szCr(0)] + [szCr(I)]) was 9.9 x 10.3 M, If the reaction time (in which the reaction reached completion) is assumed to be at least four times the half life, t1,2, then the rate constant is greater 3‘Mfl sec-1. Pertinent data for the exchange experiment are than 4 x 10 displayed in Table 5. Some error arose from difficulties in decomposing the organometallic compounds and DMSO in the quantitative determination of chromium. The expected, well resolved esr Spectra were obtained for the szCrI and oxidized szCr(0) recovered from the quench mixture. The exchange was checked by preparing the bz251Cr(0) tracer (rather than szSICrI tracer) and measuring the activity of the (initially in- active) szCrI after mixing and separating the reactants. The exchange 55 ououxfia noooov moa x O.H m.~ ma Bonn mucuwowuoum CA N O.H m.H HH ououxwa nuoooo Bonn m Ououuaww mooooo< coauoaom z NIOH O O.H u HHOH OOH x 0.0 H.O OH HHOOOO Ha O.H soauoflom O N.OH A O.H u HHOO 0.0 O.O OH HOOHOOOO Ha O.H m mIcH on 0.0 I THE moa x H.~ OO.H 0.0 magma 8“. Ha O.H muooasoo oaoaa\amo OH x Emu OH x moaoaa coaumwuomon mamamm OOO>Ooo< OOwOooam mI.%uO>Huo< m .OOHEOHSU .oumo Hauooaauonxo owomsoxo aoavaonooooo .O OHOOO 56 was complete in less than 30 seconds, the shortest time checked in this "reverse direction" of exchange. 251Cr1 tracer prepared by neutron bombardment of bz2CrI (rather than from 51CrCl3) showed no exchange. Exchange experiments run with bz This proved to be a consequence of a Szilard—Chalmers reaction of szCrI that resulted in a product that did not oxidize szCr(0). The presence of the Szilard-Chalmers product was proved by preparing a sam- ple of szCrI from 51CrC13, dividing the szCrI into two samples, and subjecting one to further neutron bombardment. When aqueous solutions of both samples were added to Dowex 50WX4 ion exchange columns, the sample which was subjected to the second bombardment separated to give a high specific activity fraction as well as a fraction common to both samples with a specific activity nearly identical to that of the bZZCrI prepared from 51CrCl3. The high specific activity fraction (Szilard- Chalmers product) was unstable in air, and no characterization was at- tempted. III. Studygof the Electron Exchaggg Between Bis(arene)chromium Complexes and Their Cations by ESR Technique The width of esr signals is a measure of the lifetime of the elec- tron in a particular state. The rate of processes that change this lifetime may be studied by determining the effect of the process on the 116-120 esr linewidth. Electron exchange is, of course, such a process. The method developed by Weissman121 was applied to the study of the electron exchange in the systems szCr(0) - szCr(I) and bis(biphenyl)- chromium(0) - bis (bipheny1)chromium(I), biphZCr(0) -— biphZCrU). in dimethylsulfoxide and benzene. 57 While results were often consistent within a particular experi- ment, sets of experiments in which historical factors (methods of sam- ple preparation, of material purification, etc.) were varied were often inconsistent. In some experiments the concentration of the complex could be increased from zero to near the level of saturation without affecting the linewidth of the esr signal of the cation present; in others, broadening did occur in the expected way. If the Weissman method was applied, the rate constant for biphZCr(0) - biphZCr(I) electron 9 l 1 exchange was calculated to be 3.4 x 10 sec- in benzene and 2.7 x 109M.1 sec—1 in DMSO. In some experiments broadening occured to M- the extent that all hyperfine structure collapsed. Because of its occasional absence, however, it is likely that broadening was due to a state perturbing process (other than electron exchange) that occurs at a frequency near that corresponding to the sep- aration of the hyperfine components of the signal, which is due to (or promoted by) species other than the bis(arene)chromium complexes and their cations. If the broadening is Spurious, then the experiments showing no broadening limit the rate constant for the electron exchange to k < 3 x 108 M71 sec-1, and the experiments showing broadening are ir- relevant. In a search for possible complicating processes, it was found that biphZCrI undergoes photodecomposition yi§_a CTTM (charge transfer to metal) transition that resulted in prOduction of iodine and (probably) biph20r(0). To allow study of the electron exchange between bis(arene)chromium molecules and the corresponding cations in benzene, a new compound, bis(benzene)chromium tetrabutylaluminate, was prepared. This compound 58 had high solubility in benzene. The preparative.method appears to be one of general utility for creating benzene-soluble ionic species. Experimental A. Preparation and Purification of Materials Bis(benzene)chromium and its univalent cation were prepared as in the previous experiment. Bis(bipheny1)chromium(0) and bis(bipheny1)— : chromium iodide were prepared by standard techniques.122 Solvents were : purified as in the previous experiments. Bis(benzene)chromium tetrabutylaluminate was prepared by adding a solution of sodium tetrabutylaluminate (0.4 g in 10 m1 benzene) to a suspension of bz CrI (0.1 - 0.5 g) in benzene (10 m1). Additions must 2 be made drOpwise and with adequate stirring to prevent reduction of the complex. The resulting solution was filtered and evaporated. The pro- duct was triturated with ether, and further purified by subliming for- eign matter from it under high vacuum. The yellow product is soluble in benzene. Sodium tetrabutylaluminate123 was prepared by adding 50 m1 of tri— butylaluminum and 10 g. of sodium to 125 ml of n-heptane in a 250 ml Morton flask, then heating the mixture slowly to reflux temperature. A sodium dispersion was produced by high speed stirring of the reaction mixture. The mixture was allowed to reflux for two hours, then it was filtered and cooled to yield the product which was then recrystallized from.pentane. 59 B. Exchange Experiments The exchange solutions were prepared in the glove box by preparing a solution of the complex cation (typically 10-4 M), dividing it in half, and saturating one half with the zerovalent complex. Exchange solutions with varying concentrations of zerovalent complex, but constant in cation concentration, could then be prepared by mixing the portions. Samples were added, in the glove box, to 3 mm esr tubes with glass microstOpcocks. Tube size variation was checked by use of a standard ; paramagnetic solution and corrected for if necessary. Esr spectra were i obtained by use of a Varian E-4 esr spectrophotometer with a variable temperature controller. Stability was excellent over many hours, as the peak intensity for stable samples remained constant over the span of each experiment. Weissman's method121 was applied to the central peak of the his- (arene)chromium(I) spectrum centered at g = 1.99. The unbroadened spec- trum of bZZCr(I) had hyperfine Splitting of 3.6 G, while the spectrum of biphZCr(I) had hyperfine of 1.2 G. C. Photochemistry of Bis(biphenyl)chromium Iodide A solution of biphZCrI in benzene which gave the expected esr Sig- nal and a peak at 365 nm in the visible spectrum was divided in half. Half was exposed to intense but indirect sunlight for 20 minutes while the other half remained in the spectrophotometer cell compartment. The exposed sample had no esr signal and no peak at 365 nm. Instead, peaks had grown in at 500 and 300 nm, indicating the presence of I2 in benzene. No change occured in the sample that remained in the spectrOphotometer cell compartment, but when the samples were mixed the absorbance of the 60 365 nm peak rose from the initial value of 0.4 to 0.6. Addition of 12 to a solution of biphZCr(0) in benzene caused an intense esr signal, typical of biph CrI, to appear. 2 Results and Discussion Occasionally, experiments yielded data which indicated a monotonic decrease in the intensity (or increase in the linewidth) of the ear spectrum of the bis(arene)chromium cation with the addition of the res- pective zerovalent complex. Figure 15 shows the effects on the Spectrum 4 of biph CrI (2.8 x 10- 'M in DMSO) of the addition of biph2Cr(0) to give 2 concentrations of 0, 0.5, 1.0, 1.5, 2.0, and 3.0 x 10.3 M, respectively. The Weissman plot, shown in Figure 16, allows the calculation of a rate constant for the electron exchange of 2.7 x 109M-1 sec-l. Measuring the Spectrum at 230 and 500 revealed an inverse temperature dependence (higher temperatures gave less broadening) for the broadening. The hyperfine structure collapsed completely at high concentrations of added biphZCr(0) and high temperature (75°). Other experiments that spanned the concentration ranges of the experiments above showed no linewidth changes on addition of biphZCr(0). The same result was obtained for analogous experiments involving the szCr(0) - szCrI exchange, where the concentration of szCrI was 2.5 xlO-3 M and bZZCr(0) was varied up to 5 x 10.3 M, At the higher complex concentrations in one experiment carried out with bZZCt(0) and szCrAlBua [bis(benzene)chromium tetrabutylaluminate] in benzene, fur- ther splitting of the hyperfine structure into doublets was observed, as shown in Figure 17 and 18. It should be noted that what might be 61 0v .ooaoow aofiaowaufiahaonoanvowp mo anuuoomo woo oucmvooun swoonoxm .mH ouowum f O) H/ O H, OH /( 62 .cofiuoo unoam>woo OOO poo onasaaownoaazoonoflnvmfin awesome owowsoxm oouuooam wow uoan omammwo3 .oa ouowwm me: .14 :9 0.0 0.0 O.H u I q 63 = 15° r = 25° Figure 17. Exchange broadened esr spectra of bis(benzene)chromium(l). 64 = 15° //\ (bz2Cr)= 0.002 T = 30° (bz2Cr) = 0.002 Figure 18. Exchange broadened spectra of bis(benzene)chromium(1). .mfi 65 interpreted as common anisotropy in Figure 15 could also be interpreted as the onset of the additional splitting of the hyperfine which is ap- parent in Figures 17 and 18. The Splitting pattern shown in Figures 17 and 18 has never been reported (even though the esr spectra of bis- (arene)chromium cations have been thoroughly investigated), and is dif- ficult to reproduce. It may be evidence for a Species which broadens the esr spectrum in the absence of electron exchange. The following observations were made: Increasing the concentration of the zerovalent complex increased the intensity of the set of peaks with the higher g value (see, for example, Figure 17A and Figure 18A), as did increasing the temperature (see Figures 17A and 17B). The bis(benzene)chromium tetrabutylaluminate which was used to allow the study of the szCr(0) - szCr(I) exchange in benzene is a novel Species that is paramagnetic (the usual bZZCr(I) esr signal is ob- tained) yet soluble in benzene. The tetrabutylaluminate ion will come bine with ferricenium cation to give a benzene soluble species which has an ear signal with hyperfine splitting. Because of the characteristic- ally short relaxation time for ferricenium, no hyperfine splitting was Observed in previous work in which solvents other than benzene were used. Several facts suggest that the broadening observed in the esr sig— nal for bis(arene)chromium(l) when bis(arene)chromium(0) is added in solution is due to processes other than intermolecular electron exchange. First, some broadening occurs even for frozen solutions, suggesting that the mechanism is intramolecular. Second, broadening was not reprodu- cible for solutions which had the same concentrations of reactants, so it probably was a result of a species whose concentration was not inten— tionally controlled. Finally, the solutions were of a very low 66 concentration compared to those required for the esr study of other rapid electron exchange systems. The concentrations used, however, were near the maximum possible because of the inherent solubilities of the species, and because the cation concentration should not exceed 10.3 M to prevent spin-spin interactions. The result of no broadening can be used to eliminate certain ran- ges of possible values for the rate constant. If the broadening of the FF‘ nominally 3.6 G szCr(I) hyperfine were considered "unobservable" if less than 0.1 G, and the concentration of szCr(0) was varied from zero to 5 x 10.3 M without causing observable broadening, then the rate con- 8‘Mfl sec.1 according to the equation,118 {H 1 . . stant would be less than 3 x 10 (AH) k = 5 "(2.83 x 106) I: ”Ch] [szCr(0)] where (AH)exch is the increase in linewidth due to exchange.121 67 IV. Study of the Electron Exchange Between Bis(arene)chromium Complexes and Their Cations by NMR Technique If a proton is located on a molecule that is undergoing electron exchange with others, the proton will experience a "paramagnetic pulse" each time the (usually diamagnetic) molecule accepts an electron from another and so becomes momentarily paramagnetic.124 The line Shape of the nmr signal is affected by the frequency of the pulses (and the in- fifih tensity of the hyperfine interaction). The nmr technique of measuring . the rate of electron exchange stems from the broadening of the nmr A spectrum of a diamagnetic compound when minute amounts of the exchange- jj ing paramagnetic compound are added.125-127 This technique has been i3 applied successfully to many systems, including inorganic],'28“131 organo- metallic,132 and organic133’ 134 ones. The nmr technique was applied to the study of electron exchange between bZZCr(0) and b22CrAlBu4 in benzene. As in the esr experiments, lineshapes appeared to be sensitive to competing processes to the ex- tent that electron exchange broadening was usually masked by more pro- nounced spurious broadening effects. Experimental B22Cr(0) was dissolved in benzene to give an approximately 2 x 10—2 M solution. This solution was divided into halves, and bz CrAlBu4 was 2 dissolved in one half to give approximately the concentration desired (the exact concentration was determined later by the usual spectrophoto- metric method, or by the intensity of the esr signal). Exchange solu- tions with constant concentration of bz Cr(O) but varying amounts of the 2 cation could then be prepared by mixing the two solutions. The esr 68 Spectrum of each exchange solution was checked to determine whether the cation signal was resolved into the expected hyperfine pattern, a pre- requisite for the application of the nmr method.130 The nmr spectrum of the exchange solution was obtained by use of a Varian A56-60 nmr Spectrometer with a temperature controller. The linewidth of the bZZCr(0) peak at 6= 4.3 was measured for solutions con- taining various concentrations of bz Cr(I). The natural linewidth of ffli 2 the peak was about 1.0 cps. Results and Discussion I The broadening observed for the exchange systems studied was in- t explicable in terms of only the szCr(0) and szCr(I) concentrations. For instance, extreme broadening was observed for initially narrow peaks as a result of slight air exposure of the szCr(0) solution. The amount of complex oxidized by this treatment was insufficient to give an esr Signal. If, however, bZZCr(I) was added to solutions of szCr(0) in amounts large enough to give a clear esr signal, there was often no broadening whatsoever. It is difficult to determine the cause of the spurious broadening, because the amount of material necessary to change the linewidth drastically is minute. The extreme sensitivity of these systems to oxygen and light are apparently only a small part of the dif- ficulty. Broadening was an artifact of many experimental variables. For instance, if a previously cleaned and dried nmr tube was filled with b22Cr(0) solution and capped with a teflon pressure cap, mere inversion of the tube broadened the spectrum. Flame sealing of tubes was not at- tempted because it broadened the esr Spectra of bz Cr(I) solutions 2 .s. LE Full-bill 69 contained in the tubes. The data for three samples are displayed in Table 6. Regular trends in the amount of broadening with added cation were always ob- literated by competing effects as noted above, so although the exchange probably belongs in the "weak pulse" category, no definite classifi- cation can be made, and thus no information can be extracted with cer- tainty from the data. Since it is very probable that the frequency of exchange is in the nmr range for some of the concentrations studied, it is perhaps most reasonable to propose that results of "no broadening" were misinterpretations of experiments where some degree of exchange narrowing actually occured. The natural linewidth, and the progressive broadening and then narrowing of the signal due to exchange, may very well have escaped observation because of the acute experimental diffi- culties. Table 6. NMR linewidths for exchange solutions. [szCr(0)] [bZZCr(I)] Linewidth M x 102 M x 103 at half max. (cps) 1.9 0.0 1.0 2.2 0.48 1.0 1.9 1.6 10.0 APPENDIX C THE ELECTRON EXCHANGE BETWEEN TITANOCENE DICHLORIDE AND TITANOCENE MONOCHLORIDE Tetrahydrofuran solutions initially 2.4 x 10'-3 M in (D-10)titan- ocene dichloride and 1.0 x 10-2 M_in titanocene monochloride show com- plete mixing of the (D-lO) tracer between the (III) and (IV) oxidation states within the time of mixing and separation, typically less than one minute. This indicates that the electron exchange, 5_ 5_ 5 5_ (n C5H5)2TiCl + (n C5D5)2TiCl2 I (n -C5D5)2TiCl + (n C5H5)2T1Cl2 is immeasurably fast when studied by this technique. Experimental Titanocene dichloride and titanocene monochloride were prepared and purified as described previously. Tetrahydrofuran (THF) was dis- tilled from CaH2 and then from sodium benzophenone. 2,2'-bipyridine was obtained from Eastman Organic Chemicals and used without further purifi- cation. The apparatus that was used for the exchange experiments is shown in Figure 19. Flask A contained a solution of (D-10)titanocene dichlor- ide (50 mg) in THF (40 m1), flask B contained a solution of titanocene monochloride (0.20 g) in THF (40 ml), and flask C contained excess 2,2'- bipyridine in THF (20 ml). To effect the mixing, argon was supplied through stopcock "a" to flask A at about 10 torr pressure, and a vacuum 70 71 O ‘I I 3,? b I lid 2 3: c Figure 19. Apparatus for electron exchange experiment. 72 was applied to flasks B and C. Stopcock l was opened to cause turbulent mixing of the reactant solutions. StOpcock 2 was opened after £3, 5 sec— onds so that all but 10 m1 of the reaction mixture was added to the bi- pyridine solution. A flocculent blue precipitate formed and could be vacuum filtered quickly by Opening StOpcock 3 which led to a filter tube. The filtrate was evaporated to dryness and sublimed before the mass Spectrum was obtained. The precipitate was dissolved in a minimum FE: (1—2 ml) of air free water and treated with HCl gas. The titanocene dichloride solution which resulted after an hour or so was evaporated . to dryness and sublimed before the mass Spectrum was obtained. A tO 5 sample was obtained by treating 10 m1 of the reaction mixture with HCl gas, evaporating it to dryness, and subliming the resulting titanocene dichloride. A mass Spectrum of this pure material was obtained. Discussion The mass spectrum of the titanocene dichloride species present in the exchange solution at the time of quenching is shown in Figure 20. A mass spectrum identical to that shown in Figure 20 was obtained for the t” sample, indicating that the electron exchange was complete in less than 1 minute. If this is assumed to be at least 4 t1,z, the rate constant is greater than 4 x 10 Z‘Mfl sec“1 as calculated by using the equation, ln 2 c1,2([(n5—csn5)2r101] + [(n5-0505)2r1c12]) The reaction mixture had a uv-visible absorption spectrum identi- cal to a combination of the spectra of the individual reactants, and the .ooOusaom owawnoxo couuooao aoum oowuoasouo oooooomuwu mo aouuuoom one: .ou muowam O. O. E O: OO. O.O 8H OOH . .I _ _ . . _ _ 73 H. __H :_ . :—w .. _ . . 74 est spectrum of the reaction mixture.was identical to that of titanocene monochloride. The clean electrongexchange between titanocene dichloride and titanocene monochloride demonstrates the reversibility of the reduc- tion of titanocene dichloride, which had come into doubt as a result of polarographic studies.22-24 There is little hope of improving on the experimental procedure used to study the rate of exchange by tracer techniques. The limiting factor is the slow precipitation of the titanocene monochloride-bipyri- dine adduct, which takes at least 30 seconds. Furthermore, Species that precipitate the (nS-CSH5)2Ti(III) Species from aqueous solutions usually precipitate the (nS—C5H5)2Ti(IV) species as well. These species include sulfate, phosphate, carbonate, phthalate, oxalate, picrate, and other ions. Sodium tetraphenylborate forms a blue precipitate with (n5-C5H5)2- Ti(III), but reacts with titanocene dichloride to give biphenyl and triphenylboron (identified by mass spectra) according to the equation,135 5. (n C5H5)2TiCl2 + NaB¢4 + NaCl + ¢3B + O2 + reduced titanium species There is one procedure that might allow the measurement of the rate of exchange by tracer techniques. Titanocene monochloride probably exists as a solvated monomer in THF,136 even though it is known to be dimeric in the solid form79 and in benzene solution.55 In aqueous sol- ution, the complex does show the conductance of a 1:1 electrolyte137 and it exists as a solvated monomeric cation. If titanocene monochloride is a monomer in THF, slower electron exchange might be encountered if ben- zene exchange solutions were studied. It would be interesting to see if the additional structural rearrangement that would be required for elec- tron exchange between the titanocene monochloride dimer and titanocene 75 dichloride would decrease the rate of that process enough to make it measurable by tracer techniques. BIBLIOGRAPHY 10. ll. 12. 13. 14. 15. 16. 17. 18. 19. BIBLIOGRAPHY Alt, H., and M. D. Rausch, J. Amer. Chem. Soc., 24, 5936 (1974). Rubezhov, A. 2., and S. P. Gubin, "Advances in Organometallic Chemistry," F. G. A. Stone and R. West, Eds., Academic Press, New York, 1972, Vol. 9, p. 347. Wilkinson, G., F. A. Cotton, and J. M. Birmingham, J. Inorg. Nucl. Chem., 2, 95 (1956). Maitlis, P. M., J. Amer. Chem. Soc., 85, 1887 (1963). Efraty, A., J. Organometal. Chem., 51, l (1973). Rubezhov, A. 2., and S. P. Gubin, "Advances in Organometallic Chemistry," F. G. A. Stone and R. West, Eds., Academic Press, New York, 1972, Vol. 9, p. 347. Basolo, F., and R. G. Pearson, "Mechanisms of Inorganic Reactions," John Wiley and Sons, Inc., New York, 1967. Langford, C. H., and H. B. Gray, "Ligand Substitution Processes," The Benjamin CO., New York, 1965. Gow, A., and H. Heinemann, J. Phys. Chem., 44, 1574 (1960). Cramer, R., Inorg. Chem., 4, 445 (1965). Pearson, R. G., and M. M. Anderson, Angew. Chem., 4, 281 (1965). Nesmeyanov, A. N., A. Z. Rubezhov, and S. P. Gubin, J. Organometal. Chem., 19, 163 (1969). Cossee, P., Rec. Trav. Chim Pays-Baa, §2, 1151 (1966). Green, M., 22 gl,, Chem. Commun., 1973, 97. ‘Strohmeier, W., Angew. Chem., 3, 730 (1964). Strohmeier, W., and D. v.Hobe, Z. Naturforsch., B, 18, 981 (1963). King, R. B., and M. B. Bisnette, Inorg. Chem., 4, 469 (1967). Efraty, A., J. Organometal. Chem., 21, 19 (1973). Nesmeyanov, A. N., et a1., Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1962, 1481. 76 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 77 Nesmeyanov, A. N., g£_§l,, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 13, 2146 (1961). Maitlis, P. M., £8 El}: J. Amer. Chem. Soc., 81, 719 (1965). Hsiung, B., and G. H. Brown, J. Electrochem. Soc., 110, 1035 (1963). Dessey, R. E., and R. B. King, J. Amer. Chem. Soc., 88, 5112 (1966). Gubin, S. P., and S. A. Smirnova, J. Organometal. Chem., 88, 229 (1969). Coutts, R. S. P., P. C. Wailes, and R. L. Martin, J. Organometal. Chem., 81, 375 (1973). Coutts, R. S. P., and P. C. Wailes, "Advances in Organometallic Chemistry," F. G. A. Stone and R. West, Eds., Academic Press, New York, 1970, Vol. 9, p. 135. Birmingham, J. M.,"Advances in Organometallic Chemistry," F. G. A. Stone and R. West, Eds., Academic Press, New York, 1964, Vol. 2, p. 365. Salzmann, J. J., and P. Mosimann, Helv. Chim. Acta, 88, 1831 (1967). watt, G. W., SE 81,, J. Amer. Chem. Soc., 88, 1138 (1966). Fischer, A. K., and G. Wilkinson, J. Inorg. Nucl. Chem., 8, 149 (1956). Brintzinger, H. H., 25.28,, J. Amer. Chem. Soc., 88, 6182 (1970). Davison, A., and S. S. Wreford, J. Amer. Chem. Soc., 88, 3017 (1974). Bercaw, S. E., R. N. Marvich, L. C. Bell, and H. H. Brintzinger, J. Amer. Chem. Soc., 84, 1219 (1972). Baye, L. J., Syn. Inorg. Metal-Org. Chem., 8, 47 (1972). Gorsich, R. D., J. Amer. Chem. Soc., 88, S744 (1958). Balzani, V., and V. Carassiti, "Photochemistry of Coordination Compounds," Academic Press, New York, 1970. Adamson, A. W., Coord. Chem. Rev., 8, 169 (1968). Adamson, A. W., Rec. Chem. Progress, 88, 191 (1968). Adamson, A. W., 38 21,, Chem. Rev., 88, 541 (1968). Fleischauer, P. D., A. W. Adamson, and G. Sartori, "Progress in Inorganic Chemistry," 8. J. Lippard, Ed., Interscience Publishers, New York, 1972, Vol. 17, part 2, p. 1. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. S3. 54. 55. 56. 57. 58. 59. 60. 61. 62. 78 Fleischauer, P. D., and P. Fleischauer, Chem. Rev., 18, 199 (1970). Wehry, E. E., Quart. Rev. (London), 88, 213 (1967). Endicott, J. F., Israel J. Chem., 8, 209 (1970). Ford, P. C., J. D. Petersen, and R. E. Hintze, Coord. Chem. Rev., .84, 67 (1974). Bozak, R. E., "Advances in Photochemistry," J. N. Pitts, G. S. Hammond, and W. A. Noyes, Eds., John Wiley and Sons, New York, 1971, Vol. 8, p. 227. OOH Tyerman, W. J. R., 22.21:: Chem. Commun., 497 (1967). ; Natarajan, P., and A. W. Adamson, J. Amer. Chem. Soc., 88, 5599 5 (1971). Nesmeyanov, A. N., 25 21°: Dokl. Akad. Nauk SSSR, 188, 1130 (1964). E Herzog, S., and J. Dehnert, 2. Chem., 4, 1 (1964). 8, Aquasorb is P205 supported on an inert matrix, marketed by Mallin— ckrodt Chemical Works, St. Louis, Mo. Broadbent, A. D., J. Chem. Educ., 44, 145 (1967). Brintzinger, H. H., and R. H. Marvich, J. Amer. Chem. Soc. 88, 2048 (1971). Clause, K., and H. Bestian, Justus Liebigs Ann. Chem., 654, 8 (1962). Brintzinger, H. H., g£_§l,, J. Amer. Chem. Soc., 84, 1219 (1972). Coutts, R. S. P., P. C. Wailes, and R. L. Martin, J. Organometal. Chem., 41, 375 (1973). Wagner, P. J., J. Amer. Chem. Soc., 88, 5898 (1967). Moses, F. G., R. S. H. Liu, and B. M. Monroe, Mp1. Photochem.,.l, 245 (1969). Coutts, R. S. P., and P. C. Wailes, Aust. J. Chem., 88, 2199 (1968). Dillard, J. G., and R. W. Kiser, J. Organometal. Chem., 88, 265 (1969). Nesmeyanov, A. N., E£.El:: J. Organometal. Chem., 88, 225 (1973). Takegami, Y., £538... Bull. Chem. Soc. Jap., 4_1_, 2637 (1968). McKay, H. A. G., Nature, 148, 997 (1938). 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 79 Friedlander, G., J. W. Kennedy, and J. M. Miller, "Nuclear and Radiochemistry," 2nd ed., Wiley, New York, 1955, p. 198. Hoeschele, J. D., Ph.D. Thesis, Michigan State University, 1969, p. 78. Friedlander, G., J. W. Kennedy, and J. M. Miller, "Nuclear and Radiochemistry," 2nd ed., Wiley, New York, 1955, p. 202. Nesmeyanov, A. N., O. V. Nogina, and A- M. Berlin, Dokl. Akad. Nauk SSSR, 134, 607 (1960). Nesmeyanov, A. N., O. V. Nogina, and A. M. Berlin, Izv. Sib. Otd. Ffii Akad. Nauk SSSR, Ser. Khim. Nauk, 1961, 804. Nesmeyanov, A. N., O. V. Nogina, A. M. Berlin, A. S. Girshovich, and G. V. Shatalov, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1961, 2146. Nesmeyanov, A. N., O. V. Nogina, and V. A. Bubovitskii, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1962, 1481. fits-#1:... Nesmeyanov, A. N., G. G. Bvoryantseva, Y. N. Sheinker, O. V. Nogina, N. A. Lasareva, and V. A. Bubovitskii, Dokl. Akad. Nauk SSSR, 161, 603 (1965). Nesmeyanov, A. N., E. I. Fedin, P. V. Petrovskii, V. A. Bubovitskii, O. V. Nogina, and N. A. Lazareva, Dokl. Akad. Nauk, SSSR, 163, 659 (1965). Nesmeyanov, A. N., V. A. Dubovitskii, O. V. Nogina, and V. N. Bochkarev, Dokl. Akad. Nauk SSSR, 165, 125 (1965). Nesmeyanov, A. N., 0. V. Nogina, N. A. Lazareva, and V. A. Dubovitskii, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1967, 808 O Nesmeyanov, A. N., O. V. Nogina, and V. A. Bubovitskii, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1968, 527. Nesmeyanov, A. N., O. V. Nogina, B. V. Lokshin, and V. A. Bubovitskii, Dokl. Akad. Nauk SSSR, 182, 844 (1968). Gorsich, R. D., J. Amer. Chem. Soc., 88, 4211 (1960). Beynon, J. H., "Mass Spectrometry," Elsevier Publishing Co., New York, 1960, p. 298. Green, J. C.,‘M. L. H. Green, and C. K. Prout, Chem. Commun., 1972, 421. Martin, R. L., and G. Winter, J. Chem. Soc. A, 1965, 4709. Brintzinger, H. H., and L. S. Bartell, J. Amer. Chem. Soc., 88, 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94.. 95. 96. 97. 98. 99. 100. 101. 102. 80 6182 (1970). Brintzinger, H. H., and L. S. Bartell, J. Amer. Chem. Soc., 88, 1105 (1970). Voigt, A. F., and S. Katz, Dissertation Abstracts, 88, 2242 (1959). Stranks, D. R., Discuss. Faraday Soc., 88, 73 (1960). Wahl, A. C., and M. W. Dietrich, J. Chem. Phys., 88, 1591 (1963). Ruff, 1., V. J. Friedrich, K. Demeter, and K. Csillag, J. Phys. Chem., 88, 3303 (1971). Espenson, J. H., and J. R. Pladziewicz, J. Amer. Chem. Soc., 88, 56 (1973). Prins, R., and F. J. Reinders, J. Amer. Chem. Soc., 88, 4929 (1969). Cotton, F. A., "Chemical Applications of Group Theory," John Wiley and Sons, New York, 1967, p. 174. Bhustorovich, E. M., and M. E. Dyatkina, Zh. Neorg. Khim., 8, 493 (1961). Hillier, I. H., and R. M. Canadine, Discuss. Faraday Soc., 1969, 27. Dessy, R. B., R. B. King, and M. Waldrop, J. Amer. Chem. Soc., 88, 5112 (1966). Dessy, R. E., and R. L. Pohl, J. Amer. Chem. Soc., 88, 2005 (1968). Furlani, G., and E. O. Fischer, Z. Electrochem., 88, 481 (1957). Nesmeyanov, g£_§8,, J. Organometal. Chem., 88, 169 (1969). Korshunov, A., 58 88,, Dokl. Akad. Nauk SSSR, 888, 769 (1958). Furlani, C., and G. Sartori, Ric. Sci., 88, 973 (1958). Vertyulina, L. N., g ”Elf: Zh. Obschch. Khim., 88, 285 (1963). Vlcek, A. A., Z. Anorg. Allg. Chem., 884, 205 (1964). Vlcek, A. A., Z. Anorg. Allg. Chem., 884, 109 (1960). Cotton, F. A., Chem. Rev., 88, 551 (1955). Zeiss, H., P. J. Wheatley, and H. J. S. Winkler, "Benzenoid Metal Complexes," Ronald Press, New York, 1966. Sohn, Y. 8., D. N. Hendricksen, and H. B. Gray, J. Amer. Chem. Soc., .88, 3603 (1971). 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. Organometal. Chem._24, C13 (1970). 81 Berry, R. 8., J. Chem. Phys., 55, 2025 (1961). Levy, D. A., and L. E. Orgel, M01. Phys.,p4, 93 (1961). Fritz, H. P., and J. Manchot, J. Organometal. Chem., 2, 8 (1964). Jellinek, F., J. Organometal. Chem., 5, 490 (1966). Cotton, F. A., J. Amer. Chem. Soc., 55, 1543 (1963). Haaland, A., Acta Chem. Scand., 52, 41 (1965). Schafer, L., J. F. Southern, 8. J. Cyvin, and J. Brunvoll, J. fin‘ Cyvin, S. J., J. Brunvoll, and L. Schafer, J. Chem. Phys., 54, 1517 (1971). Olive, G. H., and S. Olive, Z. Phys. Chem., 55, 223 (1967). Hausser, K. H., Naturwissenschaften, fig, 666 (1964). Vetchinkin, S. 1., gg'gl,, Opt. Spektrosk., 5, 71 (1960). Fischer, E. 0., and W. Hafner, Z. Anorg. Allg. Chem., 286, 9 (1956). Fischer, E. 0., in "Inorganic Syntheses," E. G. Rochow, Ed., McGraw Hill Book Co., Inc., New York, 1960, Vol. 6, p. 132. Hudson, A., and G. R. Luckhurst, Chem. Rev., 1969, 191. Weissman, S. I., Z. Electrochem., 54, 47 (1960). Chang, R., J. Chem. Educ., 47, 563 (1970). Currin, J. D., Phys. Rev., 26, 1995 (1962). Weissman, S. 1., J. Amer. Chem. Soc., 84, 4269 (1962). Weissman, S. I., and R. L. Ward, J. Amer. Chem. Soc., 22, 2086 (1957). Brauer, G., Ed., "Handbook of Preparative Inorganic Chemistry," Vol. 2, Academic Press, New York, p. 1396. Hohn, E. G., J. A. Olander, and M. C. Day, J. Phys. Chem., 15, 3880 (1969). McConnell, H. M., and S. B. Berger, J. Chem. Phys., 51, 230 (1957). Johnson, C. 8., "Advances in Magnetic Resonance," J. S. Waugh, Ed., Academic Press, New York, 1965, Vol 1, p. 33. 82 126. Sack, R. A., Mol. Phys., 1, 163 (1958). 127. Bloembergen, N., E. M. Purcell, and R. V. Pound, Phys. Rev., 15, 679 (1948). 128. McConnell, H. M., and H. E. weaver, J. Chem. Phys., 25, 307 (1956). 129. Myers, 0. E., and Sheppard, J. C., J. Amer. Chem. Soc., 85, 4739 (1961). 130. McConnell, H. M., and C. R. Giuliano, J. Inorg. Nucl. Chem., 5, 171 (1959). 131. Britt, A. D., and W. M. Yen, J. Amer. Chem. Soc._§5, 4516 (1961). r: 132. Bietrich, M. W., and A. C. Wahl, J. Chem. Phys., 58, 1591 (1963). 133. DeBoer, E., and C. MacLean, J. Chem. Phys., 44, 1334 (1966). 134. Bruce, C. R., R. E. Norberg, and S. I. weissman, J. Chem. Phys., ; 24, 473 (1956). 3 135. Gerrard, W., "The Organic Chemistry of Boron," Academic Press, New York, 1961, p. 110. 136. Raynor, J. B., and A. W. L. Ball, Inorg. Chem. Acta, 1, 315 (1973). 137. Wilkinson, G., and J. M. Birmingham, J. Amer. Chem. Soc., 15, 4281 (1954).