ABSTRACT POLYMER-SUPPORTED TRANSITION METAL CATALYSTS By Chandrasekaran. E.S. Homogeneous catalysts have so far found only limited use, chiefly because of the difficulty of their separation from the reaction products. Making homogeneous transition metal catalysts insoluble by attachment to various polymers is a significant step in improving their industrial applicability. While the polymer-attachment technique obviously prevents the loss of possibly expensive materials and contamination 6'13 the method also offers the opportunity to of reaction products, prepare a new class of catalyst systems with other desirable properties. For instance, the polymer-supported catalysts have been demonstrated to have selectivity towards molecules of different bulk and polarity.7’12 Also, the attachment of a saturated complex that is a potential catalyst to a rigid matrix, followed by reductive elimination of a ligand, should produce higher concentration of monomeric, coordinatively unsaturated species than is obtained in 6,10 solution. This would be reflected in the increase in catalytic activity of that complex when compared to a similar non-attached complex under the same conditions and has been found to be true in the recent publications!"5 Chandrasekaran. E.S. For the research discussed here, it was decided to use polystyrene-divinylbenzene copolymer beads as the supporting matrix. Beads ranging in size from 30 to 60 mesh and with a divinylbenzene content of two to twenty percent (600 A pore) were used. The reaction scheme used for the polymer-attachment was as follows: Cl CHz-O- Csz S" C'4 Polystyrene CH Cl 2 Na". THF Jr #CHa l.i THF ./CH, .l/CH’ (A) Chandrasekaran. E.S. Treatment of (A) with MCl4 (where M = Ti, Zr, and Hf) in benzene results in the formation of mono cyclopentadienyl metal trichlorides. The polymer-attached and non-attached species showed close resemblance in their colors. The polymer-attached species were identified by the metal and chloride analysis. The expected and the observed metal to chloride ratio agree reasonably well. Additional evidence to characterize the polymer-supported species was obtained from far infrared spectra. The spectra of the non-attached and polymer-attached species agreed closely. The knowledge of the physical nature of the metal catalyst dispersion is useful in understanding the chemical properties of the polymer-attached compound. An electron microprobe was used to study the physical nature of the metal catalyst dispersion on the polymer support. Electron microprobe X-ray fluorescence analysis of the sectioned beads suggests a uniform distribution of metal and chloride. The mono cyclopentadienyl titanium trichloride on reduction produces an active catalyst whose hydrogenation efficiency is about twenty times as great as the corresponding non-attached species. This was attributed to the ability of the rigid supporting matrix used (20% divinylbenzene) to keep the metal atoms apart, preventing dimerization and the accompanying loss in activity. The supported catalyst also has been shown to have selectivity towards molecules of different bulk. The general trend among the supported catalyst systems is toward slower rates for larger olefins as observed by T4 Kroll. This behavior is substantially different from ordinary heterogeneous catalysts. Chandrasekaran. E.S. Another area of interest with metallocenes (titanocene in particular) is the nitrogen fixation. Two titanocenes seem to be 3’15“]8 required for the formation of dinitrogen complex, followed by its reduction to ammonia as shown by the following reaction: @ N Q N Q _. 2: 13—2» 13:]:Ti\© %;.2NH3 But in the supported titanocene system, the titanium centers are too far apart to form the dinitrogen complex and so is not expected to be a nitrogen fixing catalyst. This was found to be 15-17 and Vol'pin-Shur18 true by Kroll,14 using both Van Tamelin nitrogen fixation schemes. With this problem in view, attempt was made in this work to see whether by using a polymer with a lower degree of crosslinking, one can bring the titanium centers close enough (as the polymer matrix would be less rigid or more flexible) to form the dinitrogen complex. The research work reported here consists of (A) the polymer- attachment of cyclopentadienyl metal chlorides of titanium, zirconium, and hafnium to the polystyrene-divinylbenzene copolymer beads; (B) their characterizations using low frequency infrared spectroscopy in the solid state; (C) study of the physical nature of the metal catalyst dispersion on the polymer support by use of electron microprobe X~ray fluorescence studies; (0) the catalytic activity of the polymer-attached species towards hydrogenation of unsaturated organic molecules and towards nitrogen fixation. POLYMER-SUPPORTED TRANSITION METAL CATALYSTS By Chandrasekaran. E.S. A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry l975 To my parents, for their understanding and moral support throughout the many years of my education; to my brothers and sister, whose encouragement has made these years good and worthwhile. ii ACKNOWLEDGEMENTS The author wishes to express his gratitude to Professor Carl H. Brubaker for his guidance, understanding and encouragement throughout the course of my studies and research. He would like to express his gratitude to Michigan State University for granting him teaching and research assistantships. Thanks to Dr. N. D. Bonds for his assistance during the early part of the work and many fruitful discussions. Thanks also go to the past and present members of the Brubaker research group, whose friendship made the life bearable while away from home. iii TABLE OF CONTENTS Page INTRODUCTION ......................... 1 EXPERIMENTAL ......................... 8 Preparation of Cyclopentadienyl-Substituted Copolymer . . 9 Preparation of Copolymer-Attached TiCpCl3 ........ 10 Preparation of Copolymer-Attached TiCpCl2 ........ 12 Preparation of Copolymer-Attached ZGCCl3 and HprCl3 . . 13 Preparation of Cyclopentadienyl Titanium Trichloride. . . l4 Preparation of 2% Crosslinked Copolymer Containint TiCDZClz ................... 15 Preparation of Copolymer-Attached TiCpZCl ........ 16 Preparation of Copolymer Containing TiCp2(CH3)2 ..... 17 Analytical Methods 18 ELECTRON MICROPROBE STUDIES ........... . . . . . . . . . . . l9 INFRARED STUDIES ....................... 21 HYDROGENATION STUDIES .................... 22 Homogeneous TiCpCl3 Hydrogenations ............ 22 Polymer-Supported TiCpCl3 Hydrogenations ......... 23 Polymer-Supported TiCpCl2 Hydrogenations ......... 24 Polymer-Supported ZGCCl3 and HprCl3 Hydrogenations , , 24 iv Table of Contents Continued RESULTS AND DISCUSSION Optical Properties ......... Analytical Results ................... Infrared Studies Electron Microprobe Studies ............... Chemical Reactions ................... Hydrogenation Studies . Nitrogen Fixation Studies REFERENCES 25 28 32 32 39 42 48 57 Table 10 LIST OF TABLES Colors of Non-Attached and Polymer-Attached Metal Cyclopentadienyl Compounds .......... Analytical Results ................. Analytical Results ................. Far Infrared Bands of Polymer-Supported TiCpZCl2 Far Infrared Bands of Polymer-Attached TiCpCl3 and Non-Attached TiCpCl Hydrogenation Rates with Polymer-Attached and Non-Attached Titanocene Species ......... Hydrogenation Rates with Polymer-Attached Titanocene Species ................. Hydrogenation Rates with Polymer-Attached TiCpCl3 and Non-Attached TiCpCl3 .......... Relative Hydrogenation Rates ............ Hydrogenation Rates with Polymer-Attached TiCpClz. . . . vi and Non-Attached TiCpZCl2 ......... 3 .......... Page 28 33 34 35 49 50 . . 52 55 55 LIST OF FIGURES Figure Page 1. Reactions of (n5-C H )ZTi(CH ) in which Titanocene Appears to Occur a? fin Interaeaiate .......... l 2. Reactions of [(ns-CSH ) TiH] in which Titanocene Appears to Occur as a3 Intermediate .......... 2 3. Reaction Scheme for Preparation of Cyclopentadienyl~Substituted Copolymer ,,,,,,,, ll 4. Reaction Scheme for Preparation of Polymer-Supported TiCpCl3 ............... 12 5. Reaction Scheme for Preparation of _ Polymer-Attached ZGCCl3 and HprCl3 ,,,,,,,,, l4 6. Reaction Scheme for Preparation of Polymer-Attached TiCpZCl2 ,,,,,,,,,,,,,,, 16 7. Reaction Scheme for Preparation of Polymer-Attached TiCpZCl ............... l7 8. Colors of Polymer-Attached and Non-Attached Titanium Species ................... 29 9. Color of Polymer-Attached TiCpZCl2 .......... 30 lo. Colors of Polymer-Attached Species .......... 3l ll. Far-Infrared Spectra of Non-Attached and Polymer-Attached TiCpZCl2 .............. 36 l2. Far-Infrared Spectra of Non-Attached and Polymer—Attached TiCpCl 38 3 ooooooooooooo 13. X-ray Fluorescence Distribution of Titanium and Chlorine in the Polymer-Supported TiCpCl3, 4O 14. X-ray Fluorescence Distribution of Titanium and Chlorine in the Polymer-Supported TiCpZCl2 , , , , 4T 15. Ka X-ray Micrographs of Polystyrene-Supported TiCpCl3 ....................... 43 vii List of Figures Continued Figure Page 16. Hydrogenation Curve for Polymer-Attached TiCpCl3 (ground beads) ................ 53 17. Relative Reduction Rates ............... 56 18. Reaction Scheme for Nitrogen Fixation ........ 57 viii INTRODUCTION The unusual reactivity of systems involving titanocenes and other metallocenes of early transition metals towards normally rather inactive molecules such as hydrogen and nitrogen has been of great interest. Bercaw's thesis.1 deactivation 1/2 H2. 0°C "d ryfl Q 1/2 H2. 0°C T1 hexane — 7 BO'C hexane This is illustrated by the figures 1 and 2 taken from (CSHSTIH)2C1038 1/2 ((hs-c5u5)zrm)2 llx [(hs-C585)2T1H]x N113 + '1'1(IV) 1/2 [C10H1011P(06H5)3]2 Figure 1. Reactions of (nu-C5H5)2Ti(CH3)2 in which Titanocene Appears to Occur as an Intermediate mm Lzooo op mgmmaa< A nxufiHvdk~Anxno-ncv ~.» 93:. N + «:2 N A o x ~_ogu~flnznu-ngv_uu «Aoavsa~1n=nu-mgv ~Huii, mzofiu~1=aam=nuv museumemecH cm mcmoocmpwh gown: cw NHITHN Amxmu- avg mo mcopuomwm .N mgzmwm anu ~ uco>aon Nz my away 00 ¢ newua>uuuuov These compounds have been shown to be particularly useful in the hydrogenation of unsaturated organic compounds and in the chemistry of nitrogen fixation. However, the complexes readily polymerize,2 according to the following reaction, to form catalytically inactive materials.3 .691. . (97 This dimerization could be prevented by attaching the metallocene precursor compounds to a rigid polymer support. Several ways to attach the transition metal compounds on to polystyrene-divinylbenzene 4’5 have been demonstrated. copolymer resin Homogeneous catalysts have so far found only limited use, chiefly because of the difficulty of their separation from the reaction products. Making homogeneous transition metal catalysts insoluble by attachment to various polymers is a significant step in improving their industrial applicability. While the polymer-attachment technique obviously prevents the loss of possibly expensive materials 6-13 the method also offers and contamination of reaction products, the opportunity to prepare a new class of catalyst systems with other desirable properties. For instance, the polymer-supported catalysts have been demonstrated to have selectivity towards molecules of 4 different bulk and polarity.7’12 Also, the attachment of a saturated complex that is a potential catalyst to a rigid support, followed by reductive elimination of a ligand, should produce higher concentration of monomeric, coordinatively unsaturated species than is obtained in solution.6’10 This would be reflected in the increase in catalytic activity of that complex when compared to a similar non-attached complex under the same conditions as has been reported in recent publications from this laboratory.4’5 For the research discussed here, it was decided to use polystyrene- divinylbenzene copolymer beads as the supporting matrix. Beads ranging in size from 30 to 60 mesh and with a divinylbenzene content of two to twenty percent (600 A pore) were used. The reaction scheme used for the polymer-attachment was as follows: CI CHz-O- C2H5 5n C'4 fir Polystyrene CH 20 Na" THF I, ASH, u THF CH2 CH, 5 Treatment of (A) with MCl4 (where M = Ti, Zr and Hf) in benzene results in the formation of mono cyclopentadienyl metal trichlorides. In general treatment of (A) with a metal chloride of the formula MCln proceeds as shown in the following reaction scheme: @ ___... g . CH, CH, MCI...1 (M= Ti,Zr, Hf, Nb, Mo,W) Initial studies with Nb, Mo, and W chlorides indicated the polymer-attachment of the cyclopentadienyl metal chloride. But the present work was directed mainly to Zr, Ti and Hf. The mono cyclopentadienyl titanium trichloride, on reduction, produces an active catalyst whose hydrogenation efficiency is about twenty times as great as the corresponding non-attached species. This was attributed to the ability of the rigid supporting matrix used (20% divinylbenzene) to keep metal atoms apart, preventing dimerization and the accompanying loss in activity. The supported catalyst also has been shown to have selectivity towards molecules of different bulk. The general trend among the supported catalyst systems is toward slower rates for larger substrates as observed by Kroll.14 This behavior is substantially different from ordinary heterogeneous catalysts. Study of the physical nature of the metal catalyst dispersion on the polymer support by using an electron microprobe X-ray fluorescence study indicates that the dispersion is uniform throughout the entire section of the bead. The diffusion of reagents into the interior of the beads was found to be slow and depended on the 12 The samples of attached catalyst polarity of the solvent. contain the catalytic sites surrounded by nonpolar-aromatic groups. Polar solvents tend to decrease the pore size and consequently decreases the rate of diffusion of the reagents. The reducing agent and the olefin used in the hydrogenations may not be able to diffuse into the interior of the polymer bead and reach all the catalytic sites. This could be avoided by grinding the beads prior to hydrogenation. The substrates would no longer have to diffuse through the pore structure of the polymer beads to reach the catalytic ‘4 The sites. This was observed to be very successful by Kroll. ground beads in this research were found to be about eight times more active than the whole beads. Another area of interest with metallocenes (titanocene in particular) is the nitrogen fixation. Two titanocenes seem to be 3915-18 required for the formation of dinitrogen complex, followed by its reduction to ammonia as shown by the following reaction. “@x 2@\ /N\ (Q l__.)2e" @f “@7‘ W © 2’” 2 NH3 + Ti (IV) But in the supported titanocene system, the titanium centers are too far apart to form the dinitrogen complex and so is not expected to be a nitrogen fixing catalyst. This was found to be true by Kroll,14 15-17 18 using both Van Tamelin and Vol'pin-Shur nitrogen fixation schemes. With this problem in view, an attempt was made in this work to see whether using a lesser crosslinked polymer matrix, one can bring the titanium centers close enough (as the polymer would be less rigid or more flexible) to form the dinitrogen complex. ExfERIMENTAL Manipulations involving air-sensitive materials were performed under argon in Schlenk-type (airless ware) vessels. Small amounts of polymer supported complexes were routinely treated in 50 ml septum stoppered erlenmeyer flasks with reagents drawn from needle-tipped burets. Where necessary, the transfers were made in an argon filled glove box. Far infrared spectra in the region lOO-GOO cm"1 were obtained on a Digilab model FTS-l6 Fourier Transform Spectrophotometer. Samples were prepared by crushing the polystyrene beads in a ball mill under anaerobic conditions and mulling the powder with dry nujol in a dry box. The spectra were recorded with the sample in a dry nitrogen atmosphere and mounted between polyethylene plates. An ARL-EMX/SM Electron microprobe19 was used to study the physical nature of the metal catalyst dispersion on the polymer support. The polystyrene beads embedded in wax block were sliced by using a microtome and mounted on a quartz plate (1" X 1"). MATERIALS The 20% crosslinked (600 A pore size) and 2% crosslinked macro- reticular polystyrene-divinylbenzene copolymer were gifts from the Dow Chemical Company and were washed before use to remove the impurities. Organolithium and organo aluminium reagents were obtained 8 from Alfa, while TiCpCl3 was prepared by previously published method21 with some modifications, as described in page 14. TiCl4 was obtained from J. T. Baker Chemical Co., ZrCl4 and HfCl4 were obtained from Alfa and TiCl3 was obtained from Research Organic/Inorganic Chemical Corp. Chloromethyl ethyl ether was obtained from Aldrich Chemical Co., and was distilled before use. Tetrahydrofuran (THF), hexane, benzene, and toluene were distilled over sodium-benzophenone complex under argon. Diethyl ether was refluxed over lithium aluminium hydride before distillation. Pyridine was stored over sodium hydroxide pellets before distillation over barium oxide under argon. Preparation of Cyclopentadienyl-Substituted Copolymer Two hundred grams of 20% crosslinked copolymer beads were washed with 10% HCl, 10% NaOH, H20. HZO-CH3OH, CH3OH, CH3OH-CH2C12 and CH2C12 as recommended by Pittman.22 They were then vacuum dried. Following 23 the chloromethylation method of Pepper, gt_al., l t of freshly distilled chloromethyl ethyl ether* was added to the beads in a 2 2 three-necked flask with a drying tube and an overhead stirrer. 'The flask was cooled in ice for 2 hours. A solution prepared by cautiously adding 35 m1 SnCl4 to 125 m1 of ice-chilled chloromethyl ethyl ether was then introduced slowly. After vigorous stirring of the reaction mixture at room temperature for 30 hours, the ether was removed by filtration. The beads were washed with four 1 2 portions *Caution is advised in chloromethylation and in the handling of chloromethyl ethyl ether because the related compound dichlorodimethyl ether is carcinogen as is the monochloro ether. 10 of 50% aqueous dioxane, aqueous dioxane containing 10% HCl (v/v), and, finally, with dry dioxane, until the washings were chloride free. Chloride analysis of the chloromethylated copolymer after it had been dried for two days ig_yggug_yielded 1.30 meq Cl'/g or a 14% chloromethylation of the styrene rings. The above obtained chloromethylated beads were treated with 300 m1 dry THF and 300 m1 of 1.6 M sodium cyclopentadienide in THF. After the mixture was stirred for 5 days at room temperature, excess sodium cyclopentadienide and THF were removed by filtration and the product was washed with 1:1 dry, airfree ethanol: THF until the washings were chloride free. The product was washed with three one liter portions of THF and then dried jn_yagug_for several days and yielded beads containing 1.1 meq Cp (C5H5)/g (by C1. difference). Preparation of the cyclopentadienyl substituted, 2% crosslinked copolymer beads, were done by following the above described method for 20% crosslinked beads. Figure 3 summarizes the reaction scheme used. Preparation of Copolymer-Attached TiCpCl3 Cyclopentadiene substituted copolymer was converted to the cyclopentadienide anion by treatment with a two-fold excess of butyl lithium in benzene. After being stirred overnight under argon, the solution was removed, washed with benzene five times. The beads were then treated with a two-fold excess of TiCl in benzene and were 4 stirred for 2 days. The beads were then separated by filtration and washed with benzene in a soxhlet extractor until excess chloride had been removed. Extraction with THF converted the dark colored beads to ll Ewexfioaou umpzuwumnam-Fxcmwnmpcmao_oxu we co_pmgmamga toe mamcum cowpummm fu t: 1 3:61 “.1... © +02 _U~IU 0:01:30; 1 .6 cm nzflo -025 _u .m mesmwm 12 yellowish beads. They were washed a few times with THF, and dried at room temperature in_gagug_overnight. The flask containing the beads was covered with aluminium foil and cooled in ice bath, chlorine gas was added and bright yellow beads were formed. Figure 4 shows the reaction scheme. Analysis: Ti 0.282 mmol/g of beads CI 0.811 mmol/g of beads Ti:Cl ratio 2.88 (found) 3.00 (calcd) Far Infrared Spectrum: (polymer-attached) 452, 422, 380, 333 cm-1 (nujol mull): (non-attached) 450, 418, 381, 331, 295 cm'] (nujol mull). C21: TI Cl4 +U©6L 6 6 j: @(‘i‘é' C” Figure 4. Reaction Scheme for Preparation of Polymer-Supported TiCpCl3 Preparation of Copolymer Containing TiCpCl2 Cyclopentadiene substituted copolymer was convered to the cyclo- pentadienide anion by treatment with a two-fold excess of methyl lithium in ether. After being stirred overnight under argon, the solution was removed and washed with THF five times. The beads were then treated 13 with a two-fold excess of TiCl3 dissolved in THF (was dissolved by using soxhlet extraction). The beads were separated by filtration and washed with THF in a soxhlet extractor until excess chloride had been removed. The violet blue beads were dried in_ygcug_and stored in a dry box. Analysis: Ti 0.110 mmol/g of beads Cl 0.209 mmol/g of beads Ti:C1 ratio 1.90 (found) 2.00 (calc) Preparation of Copolymer-Attached ZGCCl3 and HprCl3 ZrCl4 and HfCl4 are solids (unlike TiCl4 which is a liquid) and are sparingly soluble in most organic solvents. So following the procedure described for preparation of attached TiCpCl3 (Page 10), only a very small amount(m0.06 mmol M/gm) of the species MCpCl3 (where M = Zr, Hf) could be attached to the polymer support. The method was modified by using ZrCI4, 2Py and HfCl 2Py adducts. The 4. pyridine adducts were prepared, following the procedure described by Ray and Westland.24 ZrCl4 was suspended in dry benzene and slightly more than a two-fold excess of dry pyridine was added. The reaction mixture was cooled in ice and vigorous stirring for two hours produced a white suspension, but the stirring was continued for a further 24 hours to ensure complete reaction, at room temperature. The MC14. 2Py (M = Zr and Hf) adduct in benzene made above was then added to the polymer beads containing cyclopentadienide anion attached, in benzene and allowed to stir for 48 hours at room temperature. The cream colored beads were filtered, washed with dry l4 benzene, followed by dry THF, in a soxhlet extractor until excess chloride had been removed. The cream colored beads were dried jn_vacuo. Figure 5 illustrates the reaction scheme: Benzene _ met-2». .. a + ANddhct CH2 CH2) e NMZI (M = Zr and Hi) 3 Figure 5. Reaction Scheme for Preparation of Polymer-Attached ZGCCl3 and HprCl3 Analysis: Zr = 0.293 mmol/g Hf = 0.250 mmol/g C1 = 0.830 mmol/g C1 = 0.700 mmol/g Zr:Cl ratio = 2.83 (found) Hf:C1 ratio = 2.80 (found) = 3.00 (calcd) = 3.00 (calcd) Preparation of Cyclopentadienyl Titanium Trichloride A mixture of 36.0 g (0.096 moles) of bis(cyclopentadieny1) titanium dichloride, 50.2 g (0.272 moles) of titanium tetrachloride, and 150 ml of dry xylene was heated at the boiling point of xylene (m140°) for 2 1/2 hours. The reaction mixture was then cooled to room temperature, and yellow crystals of the product separated. These were filtered under dry nitrogen, washed with hexane, and dried briefly under nitrogen. 15 The crude product was then dissolved in minimum amount of boiling benzene. Charcoal was then added to the solution and it was stirred for one hour. The solvent was removed by evaporation and the product was dried jg_yaggg, The crude product was transferred into a soxhlet extractor in a dry box under argon and extracted with 100 m1 dry benzene. After a couple of hours the solution was cooled to room temperature and the yellow crystals of cyclopentadienyl titanium trichloride were separated by filtration under argon and vacuum dried (m 75% yield). Preparation of 2% Crosslinked Copolymer Containing TiszCl2 100 g cyclopentadiene substituted 2% crosslinked beads (m 2 meq C5H5 per gm. of beads) was treated with a two-fold excess of methyl lithium in ether and allowed to stir for two days. The beads were filtered, washed with dry THF. The product was suspended in dry THF and a solution containing 70 g TiCpCl3 (0.32 moles) dissolved in dry benzene was introduced and the mixture stirred for three days. Excess TiCpCl3 was removed by extracting with benzene in a soxhlet extractor, followed by a similar extraction with THF. The product, red colored beads, was dried jn_yagu9_for two days. Figure 6 gives the reaction scheme: Analysis: Ti = 0.788 mmol/9 c1 = 1.499 mmol/9 Ti:C1 ratio = 1.90 (found) 2.00 (calcd) l6 Far_lnjrargg_Spectrum: (polymer-attached): 402, 362, 307, 280, 256 (broad) and 196 cm-1 (nujol mull) (non-attached): 400, 360, 303, 276, 247, 206 cm“ (nujol mull) —’ + U" CH2 Ti CH2 /1\ (:1 CH C] ’//(:| Ti or \.. Figure 6. Reaction Scheme for Preparation of Polymer-Attached TiCp2C12 Preparation of Copolymer Containing TiCpZCl To 2 g (1.72 mol Ti)of polymer-attached TiCpZCl2 in a 50 m1 septum stoppered vessel was added 10 ml tri-n-butyl aluminium.25 The reaction mixture was stirred at 110° for three hours to produce the dark red-brown aluminium-titanium adduct. After'the excess aluminium alkyl was removed, the beads were washed with diethyl ether to give a blue-grey species which became olive green when dried ig_vacuo for several hours. Analytical results are reported in Table 2, page 33. Analysis: Ti 0.274 mmol/g CI 0.285 mmol/g Ti:C1 ratio 1.04 (found) 1.00 (calcd) Figure 7 shows the reaction scheme for the preparation of copolymer containing TiCp2C1. I7 110°C, 4hr 2) 81,0 4 @K @7 \C‘ @7 Figure 7. Reaction Scheme for Preparation of Polymer—Attached TiszCl Preparation of Copolymer Containing TiCp2(CH3)2 Polymer-attached TiCp2(CH3)2 was prepared by the method described by Clauss and Bestian26 for monomeric TiCp2(CH3)2 with appropriate modifications. To 2 g (1.6nnml Ti/g) of polymer-attached (2% crosslinked) TiCpZCl2 maintained at dry ice-acetone temperatures was added 20 ml diethylether and 10 ml of 1.64 M methyl lithium in ether. The mixture was stirred and allowed to warm slowly to room temperature. After four hours, the excess methyl lithium solution was removed by syringe and the copolymer was extracted with THF until the washings were chloride free. The yellow product was stored under argon at -10°° 18 analytical results are reported in Table 2 and the far infrared spectrum discussed in page 35. Analytical Methods Halide from the chloromethyl groups was removed from the c0polymer with hot pyridine and determined by the Volhard technique.27 . . . . . . 2 T1tan1um, 21rc0n1um and hafnium were determ1ned 8 by ignition of the metal-containing polymer at 900° for eight hours and weighed as the oxide. Titanium was also analyzed by spectrophotometry of the titanium-peroxide complex. The two methods agree closely. Titanium complexed chloride was removed by digestion of the powdered polymer samples in 2N KOH solution at 80° for eight hours. Chloride was determined by the Volhard method following acidification of the aqueous supernant. ELECTRON MICROPROBE STUDIES Metal catalysts are often made in a high degree of dispersion on the surface of a high-area solid (polystyrene beads in this work). The knowledge of the metal catalyst dispersion on the polymer support surface will help understand the catalytic process taking place and the catalytic activity also.29 The investigation was carried out by using an electron microprobe X-ray fluorescence study of titanium, zirconium and chlorine, supported on the copolymer beads as the cyclopentadienyl metal chlorides. Experimental The polymer-supported TiCpC13, TiCp2C12, and ZGCCl3 were prepared according to the procedures described in pages 10-15 . The beads were sliced by using a microtome. Since the beads were fragile, they had to be coated with wax before slicing them. This was done by adding some beads to molten wax in a metal container and allowing it to cool, to solidify. The wax block containing the beads was mounted on a microtome and sliced. The sliced sections were about 10 microns in thickness. Selected thin sections of the beads were then placed on a l” X l“ quartz plate and most of the wax was removed by careful melting. The rest of the wax was removed by careful rinsing with xylene. The sections were then mounted on the quartz plate by using a contact cement, "Zipbond". A thin evaporated carbon coating was 19 20 applied to give good surface conductance. Analyses were done with three or four beads of each sample, selected at random. Radial profiles of Ti, Zr, and C1 in the sample were determined from the traverses of sectioned beads with the ARL-EMX/SM electron microprobe* at the Horticulture Department of Michigan State University.19 The microprobe conditions were the following: 25 KV accelerating voltage .01 pa sample current 1 um beam width KG X-ray lines were used for titanium and chlorine. La X-ray line was used for zirconium. The crystal used for titanium was LiF and (NH4)2HPO4 crystal for zirconium and chlorine. Results Electron microprobe X-ray fluorescence analysis of the sectioned beads gave radial distributions of the metal and chloride as exemplified by the results of figures 14 and 15 (pages 40 and 41). The metal and chloride distributions on the polymer bead support is uniform through the entire section of the bead. Similar results were obtained for the other polymer-supported cyclopentadienyl metal chlorides of zirconium and hafnium. Figure 16 shows the Ka X-ray micrographs of polystyrene-supported TiCpCl3 for titanium and chlorine. This also corresponds to a uniform distribution of the metal and chloride on the beads. *Technical assistance of Vivian E. Shull, Department of Horticulture is appreciated. INFRARED STUDIES The vibrational spectra of tetrahedral cyclopentadienyl halogeno complexes of Ti(IV) and Zr(IV), in the low frequency region of 100-600 30’3] The spectra cm"1 have been reported in two recent publications. of these complexes in the high frequency region (600-5000 cm']) are relatively simple to interpret, as they are not complicated and are similar to those of szRu and szFe, for which the theoretical bands 32 But, in the low frequency region their assignments are available. spectra are difficult to interpret because the metal-ring and the metal-halogen stretching frequencies lie in the same region. In this investigation, the low frequency infrared spectra in the 1, of the polymer-supported compounds in the solid region 50-600 cm- state was studied. The spectra obtained were substantially in agreement with those reported in the literature. The infrared spectra were measured on a Digilab model FTS-16 Fourier Transform Spectrophotometer. Samples prepared by crushing the polystyrene beads in a ball mill under argon in a dry box and their nujol mulls sandwiched between polyethylene plates. The spectra were recorded in a dry nitrogen atmosphere. The spectral resolution 1 1 was approximately 4 cm- , with an accuracy of :_1 cm' . 21 22 Use of the Reduced Copolymer-Attached Complexes as Catalyst (l) Olefin Reduction All solvents used were reagent grade, further purified and deoxygenated by refluxing and distilling under nitrogen with the sodium (or potassium)- benzophenone complex. n-Hexane was distilled over CaHz. All the liquid substrates were distilled from sodium under argon. The hydrogenations were carried out by using gas burets of 100 ml volume. The hydrogen uptake was measured at normal atmospheric pressure and at 20° : O.5°C. Alkenes were obtained from Aldrich Chemical Co., and from Chemical Samples Co. All reductions were carried out in a 100 ml round-bottomed flask, with a side arm. The catalyst was weighed into this flask, suspended in 10 ml of hexane, and treated with 1 m1 of 2.0 M BuLi in hexane for two hours. The excess BuLi was removed and the sample washed few times with hexane by using a syringe and a needle. The reduced catalyst was then taken in 9 ml hexane and the appropriate olefin was then added by means of a syringe. The rate of hydrogen uptake was measured by using the gas buret as mentioned earlier. Homogeneous CpTiCl3 Hydrogenations A 0.071 9 sample of TiCpCl (0.3238 mole) was weighed into a 3 100 ml sidearm round-bottomed flask under argon, attached to the hydrogenation unit. The line was flushed with hydrogen by alternate vacuum and hydrogen addition, cycled a few times to remove the oxygen from the system before solvent addition. BuLi (0.97 mmol) was then added and allowed to stir for two hours. The appropriate olefin was 23 then added and hydrogen uptake rate was measured, with the reaction vessel immersed in a 20° water bath. The results are presented in Table 8 on page 52. Polymer-Supported TiCpCl3 Hydrogenations A 0.0354 9 sample of polymer-attached TiCpCl3 (0.0085 mmol Ti) was weighed into a 100 ml sidearm round-bottomed flask under argon. After attachment to the hydrogenator and hydrogen flushing, the beads were treated with 10 ml of hexane and 1 m1 of BuLi in hexane. After two hours of stirring, the excess of BuLi was removed and the reduced beads were rinsed with two 10 m1 portions of hexane. Then the beads were placed in 9 m1 of hexane, the appropriate olefin substrate was added and the hydrogen uptake rate was measured. The substrates tested were cyclohexene, l-hexene and cyclooctene. The rate observed for cyclohexene was 30 ml/min- mmol Ti, for l-hexene it was 42 ml/min-mmol Ti and 7.6 ml/min-mmol Ti, for cyclooctene. One sample of the polymer-attached TiCpCl3 was tested with sodium naphthalene as the reducing agent. For this experiment, the same procedure as above was used, except that 1 m1 of 0.3 M NaNp in THF was added to the beads in 10 ml of THF. After being stirred for three hours, the excess NaNp was removed and the beads were rinsed with three 10 ml portions of hexane. Then the beads were placed in 9 m1 of hexane, the substrate was injected and the hydrogen uptake was measured. The rate observed for cyclohexene was 20 ml/min-mmo] Ti and 32 ml/min-mmol Ti, for l-hexene. In the above hydrogenations using BuLi and NaNp as the reducing agents, the sample was used for more than one substrate hydrogenation experiment. After completing the hydrogenation of one substrate, the 24 solvent was removed, the beads were rinsed with three 10 ml portions of hexane and the hydrogenation was repeated with the same substrate or another substrate. The results were reproducible. The same reduction procedure was repeated using the beads which were thoroughly ground in a ball mill under argon, before use. The ground beads were about eight times more active than the whole beads towards hydrogenation. The initial rate observed was 194 ml/min-mmol Ti, for cyclohexene, 390 ml/min-mmol Ti, for l-hexene and 60.5 ml/min-mmol Ti, for cyclooctene. Polymer-Supported TiCpClz Hydrogenations A 0.1 9 sample of polymer-attached TiCpCl2 (0.011 mmol Ti) was weighed into a 100 m1 sidearm round-bottomed flask under argon. After attachment to the hydrogenator and hydrogen flushing, the beads were treated with 10 ml hexane and 1 ml of BuLi in hexane. After two hours of stirring, the excess BuLi was removed and the beads were rinsed with two 10 m1 portions of hexane. Then the beads were taken in 9 m1 of hexane, the substrate was injected and the hydrogen uptake measured. The substrates used were l-hexene, cyclohexene and cyclooctene. The rate observed for l-hexene was 36.5 ml/min-mmol Ti, 22.7 ml/min-mmol Ti, f0r cyclohexene and 13.6 ml/min-mmol Ti, for cyclooctene. Polymer-Supported ZGCCl3 and HprCl3 The hydrogenations were carried out, following the above described procedure for polymer-attached TiCpCl3. The substrates used were l-hexene, and cyclohexene. The rates of hydrogenation was about 2 m1/min-mmol of metal in both cases. RESULTS AND DISCUSSION Homogeneous transition metal catalysts have so far found only limited industrial use, mainly because of their difficulty of separation from the reaction products. Heterogeneous catalysts have been widely used in industry for many years, and a great deal has been learned about them. A homogeneous catalyst can be heterogenized in a variety 8,33,34 The of ways, as discussed in recently published review articles. common method is to attach it to a solid support by adsorption or by an ionic or covalent chemical bond. Less common methods involve polymerizing the catalyst so that it becomes insoluble in the medium in which it is to be used, or by trapping it in a gel or other porous medium. The catalyst heterogenized by linking it to a polymer consists of the insoluble, polymeric portion, which is the catalyst support, and the catalytic portion, which projects into the solution and is solvated and, in a sense, dissolved in it. The polymer support must be inert to the reagents that are to be in contact with it and it must withstand the temperatures and pressures that are required reaction conditions. -Polystyrene-divinylbenzene seems to meet all these requirements and offers a convenient method to support the cyclopentadienyl metal chlorides. The main difficulty in the field of polymer-supported catalysts, is the determination of the structure and the exact nature of the active 25 26 catalyst. Elemental analysis of the complexes gives the amount of the elements that are present, but this information is of only limited use. Detailed information about the environment of the metal atom and how it changes when chemical reactions occur is generally not known. Attempts have been made in this work to get some information of the nature of the complex by using far infrared and electron microprobe X-ray fluorescence studies. A copolymer of 80 percent styrene and 20 percent divinylbenzene in the form of reticulated beads with a pore size of about 600 A was chosen. This highly crosslinked polymer is quite rigid, totally insoluble, and has a very high surface area to volume ratio, due to its porous nature. The first step in the polymer support scheme is the attachment of chloromethyl (-CH2C1) group to the aromatic rings in the polystyrene-divinylbenzene copolymer beads. The chloromethylation procedure was discussed in pages 9 and 10. The chloromethylated beads were found to have 1.30 mmol of chloride/g of beads by analysis. The analytical method used was to reflux the polymer in pyridine and then determine the chlorine content of the pyridine using standard 35 Volhard analysis. The following reaction illustrates the chloro- methylation step C1 H -O-C H S C1 1) dioxane H20 wash C 2 2 s/ n 4) a) 30 hours 2) anhy. dioxane wash Polystyrene/DVB 0120 1.30 meg/g C1; W14.0% phenyl rings (Pepper, 95.91:) chloromethylated 27 Substitution of the chloride of the chloromethyl groups attached to the polymer support, with cyclopentadienide ion was done by using sodium cyclopentadienide. The procedure is outlined in page 9. Analysis of the washed, dried beads indicated that the substitution was very good (about 80%). The following reaction illustrates this substitution. xss Nan/THF O N CI Sdays; R.T. 9l¢H2J”-1¢J+ 0 01,0 CH2C| O 0.20 meg/g 1.10 meg/g Cyclopentadiene substituted copolymer was then converted to the cyclopentadienide anion by treatment with alkyl lithium as described in page 10. This step eliminates the unreacted chloromethyl groups and completes the substitution reaction. Because the procedure completely eliminates - CHZCl, the halogen analyses reported in tables 2 and 3 do not reflect the presence of chloromethyl chloride after the conversion of cyclopentadiene to cyclopentadienide by treatment with methyl lithium. The treatment of cyclopentadienide anion attached to the polymer with MCln results in the formation of polymer-attached 28 MCpCln_1 (where M = Ti, Zr and Hf), as described in pages 10-14. The preparation of other metal cyclopentadienyl derivatives attached to the polymer were also described in pages 14-18. Optical Properties The optical properties of these cyclopentadienyl metal chlorides (especially Ti compounds) are striking. For example, titanocene dichloride is dark red in color, monocyclopentadienyl titanium trichloride is bright yellow, titanocene is dark grey, bis cyclopenta- dienyl titanium dimethyl is yellow, bis cyclopentadienyl titanium mono chloride is green, mono cyclopentadienyl titanium dichloride is purple, mono cyclopentadienyl zirconium and hafnium trichlorides are cream colored. So one would expect, to some extent the polymer- attached species involving the above described compounds to be similar in color, and they were found to be so. The following table summarizes the colors of the non-attached and polymer-attached species. Table 1 Colors of Non-Attached and Polymer-Attached Metal Cyclopentadienyl Compounds Species Non-Attached Polymer-Attached TiCpZCl2 Dark red Salmon to red (depending on the Ti concentration) TiCp2(CH3)2 Yellow Yellow TiCpCl3 Bright yellow Bright yellow TiCpCl2 Purple Violet-blue TiCp2C1 Green Olive Green ZGCCl3 Cream Cream HprCl Cream Cream 3 29 33QO 5323:. $532.82 En 353215238 mo 238 .m 953.... 95.3w»; :0 6.8; «iiiumxzaa :0 :3 3.3; 3:2.Aufx_:n¢ _: 30 N 595.: 8632-558 .6 .528 .m 28.; 31 3.83m 853218.58 «8 228 .op ae=m_a 32 The colors of the polymer—attached species are illustrated in Figures 8, 9 and 10. A close resemblance in color of the non-attached and polymer-attached species is observed. Also the intensity of the color of the polymer-attached species was found to be approximately proportional to the concentration of the attached species, as expected. Analytical Results The next step in the identification of the polymer-attached species was the metal and chloride analysis. Metals titanium, zirconium and hafnium were determined28 by ignition of the metal containing polymer at 900° for eight hours and weighed as the oxide. Titanium was also analyzed as peroxide complex spectrophotometrically, after the polymer matrix was decomposed and the sample taken in acid solution. The two methods agree closely. Metal complexed chloride was removed by digestion of the powdered polymer samples in 2N KOH solution at 80° for eight hours. Chloride was determined by the Volhard method following acidification of the aqueous supernant. The results of the analysis are summarized in Tables 2 and 3. The expected and the observed metal to chloride ratio agree reasonably well. Infrared Studies Additional evidence to characterize the polymer-supported species was obtained using a Digilab model FTS-l6 Fourier Transform Spectro- photometer (far infrared spectra). The experimental portion of this work was discussed in page 2]. Comparison was made between the spectra of non-attached and polymer-attached species. Figure 12 shows the far infrared spectra of non-attached TiCp2C12 and polymer-attached Substituent -CH C1 2 -CH CpTiCpCl 2 2 ~CH2CpTiCp -CH CH 2CpTiCpCl 2CpTiCl ~CpTiCpC1 3 2 33 Table 2 Analytical Results Chloride (mmol/g) Titanium 1.20 0.17 0.543 0.00 0.285 0.811 0.115 (mmol/q) 0.281 0.233 0.274 0.282 0.058 ClzTi Ratio Found 34 Table 3 Analytical Results Substituent Chloride (mmol/g) Metal M C1:M (mmol/g) Ratio Calcd Found -CH2C1 1.30 --- --- --- -CH2CpH 0.20 --- --- --- -CH2CpTiC13 0.811 0.282 3 2.88 -CHZCpT1012 0.209 0.110 2 1.90 -CH2CerCl3 0.830 0.293 3 2.83 -CHZCprCl3 0.700 0.250 3 2.80 -CH2CpTiCpCl2 1.499 0.788 2 1.90 (2% cross linked) 35 TiCp2C12. The bands agree closely. These bands are also in substantial agreement with the reported data in the literature.30’3] The band assignments were deduced from these reported data. The spectral data are summarized in Table 4. Table 4 Far Infrared Bands of Polymer-Supported TiCpZCl2 and Non-Attached TiCp2012 (Nujol Mull) in cm“ Non-Attached Polymer-Attached Assignments TiCp2C12 TiCpZCl2 400 402 o(M-Cl) 360 362 v(M-Cp) 303 307 ring tilt 276 280 na 247 broad ring tilt 206 196 6(Cl-M-C1) bending na - not assigned The assignments were made by comparison with those reported by 31 and E. Samuel gt_al.3o Edward Maslowsky, Jr., gt_gl,, Figure 11 illustrates the far-infrared spectra of non-attached TiCpZCl2 and polymer-supported TiCpZClz. The infrared spectrum of polymer-supported TiCp2(CH3)2 showed the absence of the intense (Ti-C1) stretching frequency and the presence of a strong 0(Ti-me) band at 475 cm'] as expected. The observed bands are at 265, 368, 410, and 475 cm“. 36 N Fumauwh omzoauumm-xo¥ .mp oezmvm AcmNXFme< xmm1x wnoenocowz :oeuome zm1x2m quoz 4m< Mxomwev mcweo_;u Eavempwe ' \jo , I'I. "...""“‘_.- I Ilzkr'l’gifi"; '! .. T '1 . q' I In T 5 .. - . .‘D .- n '7 4' . \ I l . .' .. . r .. 1' u ‘I . .. J l .-. y ‘ I ~« I 1.. . n... 1 - .'_ "1 E . . . -" I ". 1. '. . . 44 Polymér-attached TiCpZCl2 on 2% crosslinked copolymer, when treated with CH Li in ether, underwent reaction similar to non-attached 3 TiCpZClz, forming TiCp2(CH3)2 on the polymer. The experimental procedure was discussed in page 17. The reaction scheme is as follows: CH2 1) xss CHsLi; 0°C C 4 l"2 + 2 [id 2) THF wash (Egg: T1 T1 \ J55 HCI; R.T. \ Cl -2CH4 CH3 (Stable at R.T.) A similar reaction has been observed earlier by Bonds5 and by Gibbons,36 by using 20% crosslinked copolymer-supported TiszClz. Bercaw gt al,,3 have studied the reaction of homogeneous TiCp2C12 with CH3Li. Polymer- supported dimethyl titanocene, TiCp2(CH3)2, is significantly less reactive than free TiCp2(CH3)2 in solution. Attempts to decompose the yellow colored beads in a hydrogen atmosphere thermally were unsuccessful. But treatment of a mixture of polymer-supported TiCp2(CH3)2 and free TiCp2(CH3)2 by using the slurry technique described by Bercaw and Marrich,3 resulted in the formation of CH4. Treatment of polymer-attached TiCp2(CH3)2 with anhydrous HCl produces TiCp2C12 on the polymer, as shown in the above reaction scheme. Similar results were observed5 45 earlier in this laboratory, when polymer-attached (20% crosslinked copolymer) TiCp2(CH3)2, (TiCpZH)x and butyllithium reduced TiCpZCl2 were treated with anhydrous HCl results in the formation of TiCp2C12. This is particularly encouraging, because this observation is different from its homogeneous analogue. Homogeneous titanocene and its hydrides readily undergo a thermally induced rearrangement to form a dimer containing both 0- and n-bonded cyclopentadienyl residues.2’3 This complex is not converted to TiCpZClZ, when exposed to HCl, but rather forms a green chloride containing dimer. Therefore, it is concluded that these polymer-attached species retain their bis-n-cyclopentadienyl integrity throughout the reduction procedure. The formation of polymer-supported TiCp2(CH3)2 by treatment of polymer-attached TiCpZCl2 and CH3Li has been supported by the chemical analysis and by the far infrared spectrum. The chemical analysis showed the absence of chloride as reported in Table 2, page 33. The far infrared spectrum showed the disappearance of Ti-Cl stretching frequency and the appearance of a strong Ti-CH3 stretching frequency around 475 cm-1. Treatment of polymer-attached TiCpZCl2 with tri-n-butyl aluminum proceeded according to the reaction reported25 for homogeneous titanocene dichloride. The experimental procedure is outlined in page 16. The reaction scheme is as follows: 46 1)A| (buty113 110° C, 4hr 2) 61,0 (IL @Ti/ Tim @T \CL @ \CI The olive green polymer-attached complex was identified by its color and chemical analysis. The product is very air-sensitive and it was not possible to obtain the far infrared spectrum. The analytical results are reported in Table 2, page 33. Treatment of the above product with anhydrous HCl at room temperature produced polymer-attached TiCpZClz, as shown in the following reaction: 9 9 C312 CI + HCI . / @g' \.. @ \.. + 1/2 H2 47 Reduction of polymer—supported TiszCl2 with BuLi in hexane has 36 been observed by Gibbons. The BuLi reduced species forms an active hydrogenation catalyst under hydrogen at room temperature. The above observations may be due to the following reaction scheme: Polymer-attached TiCpCl3 also was observed to be similar in its chemical behavior to non-attached TiCpCl Treatment of the polymer- 3. 3 with CH3Li in ether, produced the methyl derivative, as shown by the Ti-CH3 stretching frequency around 480 cm-1. attached TiCpCl Both TiCpCl3 and TiCp(CH3)3 are yellow and so the reactions could not be followed by color changes. One would expect the polymer-attached TiCp(CH to form TiCpC13, on treatment with anhydrous HCl, as 3)3 observed earlier in the case of TiCp2(CH3)2. Treatment of polymer-attached TiCpCl3 with BuLi produces a dark green product, similar to that observed for homogeneous TiCpClB. Treatment of this polymer bound complex with anhydrous HCl results 48 in the formation of bright yellow colored TiCpCl3. The BuLi reduced polymer-bound complex under hydrogen at room temperature produces an active hydrogenation catalyst. The polymer-attached complex on treatment with anhydrous HCl results in the formation of TiCpCl3. Hydrogenation Studies The polymer—attached cyclopentadienyl metal catalysts were then tested for their catalytic activity towards hydrogenation of unsaturated organic compounds and in the nitrogen fixation process. The reduction of olefins catalyzed by transition metal complexes generally requires 38 The experimental the presence of an open coordination site on the metal. evidence to date suggests that this is also true in the case of alkyl lithium reduced titanocene derivatives. Only monomeric titanocene species contains such a site. If this is the case, any increase in monomeric species concentration associated with attaching these catalysts to a polymer support should be reflected in an increase in catalytic activity of that complex when compared to a similar non-attached complex under the same conditions. Gibbons36 has studied the hydrogenation ability of polymer-attached titanocene dichloride and found it to be more active than the non-attached titanocene dichloride. The final results of the hydrogenation studies using polymer-supported TiCp2C12 reduced beads are reported recently.5 The results are summarized in Tables 6 and 7. From the electron microprobe study, it was seen that the catalyst is distributed uniformly within the bead. During the reduction process it was observed that some of the catalyst sites, which are within the bead, were unreduced. It was suggested by Kroll to be due 9 4 ego“ om“ u mezmmmeq Fepohu .cmmm m? newcmq cowpoaucw so we cmwm o: .mammu owes me» Como .cowpe>wuoe weommn Cmozoa mew; e on ucaocm mew: momma ago one a .pommmm gozm 0: 305m msmumxm omen mucmmmea we noweoa cowuoaucw cm was» xwwem> men; umpemqomo mEom o_ 111 mcmxmcopoao wcmxmzopoxu a-o_ cage mmo_ 1-- -erooe-_ mEom m.m 111 ocoxmnopoxo omEmm VFN op.F mcmxmgo_o»o meow m.mo~ mm.o mcwxmcopoxo owsem m.mm op._ mcmxmsopoxo mm mo.m 0F.F mcmxocopu»u Ezewxoz mepwcH N we Po§E-:wE\ I FE .mmpmm .cmocoo cwwmpo may way may one .o moo .0 So d moAU Phi—0&5 eoeoooo<1eoz muwcoFsowv mcooocmuwh nogomup<1coz moweopguPu memoocouprx~com umgoeuu<1coz wuwCoF50wo memooceuwppx~cmm cowuosuoe «Comma ocaoem empuwaoee16656eoo< m eoeooz eeaoeu-eoz N a cacao: - pumauee-eoeoeeo< Npumaoee eoeooo881eoz Lomeaooea pmapmumu ommwuoam memoocmuw» vogueup<1coz new umeompu<1emszpoa sow: 6mm um mmomm :owumcomoeuxz m mpnmh 50 Table 7 Hydrogenation Rates at 25° with Polymer-Attached Titanocene Species Olefin Hydrogenation Rate (0.5 M in hexane) (ml HZ/min-mmol Ti) l-hexene 213.0 styrene 243.0 cyclohexene 90.3 l-methylcyclohexene 1.0 1,2-dimethylcyclohexene 0.0 1,3-cyclooctadiene 216.0 1,5-cyclooctadiene 183.0 l-hexyne (polymer) 3-hexyne 149.0 diphenylacetylene 40.6 cholestenone 0.0 vinylacetate 0.0 51 14 If the bead structure is destroyed (by to the diffusion problem. grinding), however, a great increase in rate should be observed, because diffusion through the pores to get to the catalytic sites is then unnecessary. It was observed by Kroll that the rate of hydrogenation measured by using reduced polymer-supported titanocene dichloride is increased by a factor 8 to 10 by grinding the beads prior to hydrogenation. The results are seen in Table 6. The hydrogenation studies using polymer-attached TiCpCl3 are summarized in Table 8. The experimental part is discussed in page 22. The hydrogenation rate with the use of homogeneous TiCpCl3 reduced with BuLi was only 2.0 ml per minute per millimole of titanium for cyclohexene and 2.9 ml/min-mmol Ti for l-hexene. The hydrogenation rate with the polymer-attached TiCpCl3 reduced with BuLi was 30 ml/min-mmo] Ti for cyclohexene and 42 ml/min-mmol Ti for 1-hexene. There is a rate increase of about 15 times with polymer-supported TiCpCl3 compared to homogeneous TiCpCl3. The ground beads containing TiCpCl3 gave a hydrogenation rate of 194 ml/min-mmol Ti for cyclohexene and 390 ml/min-mmol Ti for 1-hexene. This is an increase in rate of about 7 to 10 times compared to the whole beads. This observation is similar to the earlier results found for polymer-attached TiCpZCl2 and attributed to the ease of diffusion of the substrates. So the increase in activation factor for polymer attachment is about 100 times compared to non-attached species. Also, on an operational basis, the non-attached catalyst required more than 40 hours to complete the reduction of cyclohexene, where as the attached catalyst completed the reduction in about 100 minutes. The attached catalyst shows good pseudo-first-order kinetics under a variety of conditions. Figure 16 shows the hydrogenation 52 m.oo oo.m m.o 00.0mm oo.~¢ om.m oo.¢mp oo.om oo.~ Faeoeee .fi See 155:5 933. m mp._ mF.F mP.P mp.~ mF.P _\mpoe .cmocoo Puauwh umzuwuu<1coz ucm m mcmpooopoxo mcmuuoopuxo mcmpooopoxo mcmxwg1F mcmxmz1p mcmxmc1p wcmxmzopoxu mcmxmeofiuxu mcmxmgo_uxo eweo_o mFo.o upo.o mm.o mmoo.o wppo.o nmmm.o mmoo.o mmpo.o om.o we Page m—uauwp ucaoem omcomuue1cmsxpoa mpuauee eaeoeooe Coea_oa m_uau_e uoeoaoea-eoz mpuauvh uczoem umzumupo1emsxpoa m_uauee eoeoeeoo-eoe»_oa mpuauee eoeooooe-eoz MFUQUPF uczoem omzompum1emsxpoa m_uau_e ooeoeooo-eoes_oa m_uauae uoeoeeoe-eoz comezumea uszmpmu _Uauee uoeoeop<-eoexpoa eo_z com om among coeoaeoGOCUAI m mpnmh 53 J—CH2CpTIC'3 1-HEXENE L L L V V : ""1— 20 40 60 801001201401601 1 1 ‘ Y '— 7 8‘1- Time in mins Figure 16. Hydrogenation Curve for Pol)’me'f‘AttaChed TiCpCl3 (ground beads) 54 curve for polymer-attached TiCpCl3. The major determinant in the rates of reduction with the attached catalysts was their size of the copolymer beads. Grinding the beads to a fine powder increased their activity. Another observation that can be made from the Table 8, is that the hydrogenation rate depends on the size of the substrate. The rate decreases as the size of the substrate increases. The order of decreasing rates of hydrogenation through the use of polymer-attached TiCpCl3 is 1-hexene > cyclohexene > cyclooctene. Figure 17 shows the plot of relative reduction rates observed and the results are summarized in Table 9. This observation is in accordance with the results obtained by Kroll,14 using Wilkinson's catalyst. The hydrogenation rates for polymer-attached TiCpCl2 are summarized in Table 10. The results agree with the earlier observations. Attempts to use the polymer-bound ZGCCl3 and HprCl3 were not very successful. Only about 3 m1 of hydrogen uptake per minute per milliequivalent of the metal was observed. At present, the reason for this observation is not clear. In the past the hydrogenations reported with cyclopentadienyl zirconium catalysts were carried out at high temperature and high hydrogen pressures. The absence of reported rates may be attributed to synthetic problems and probably also to very low hydrogenation rates. Recently some work has been reported 39’40 The reactivity of (n5-05H5)22r01(R) with on hydrozirconation. olefins have been studied. The alkyl zirconium(IV) complexes formed were treated with electrophiles like H+, Brz, I2 etc. The yields of the products formed were determined by vapor phase chromatography and were based on (nS-CSHS)ZZr(R)Cl species, and were between 90 and 100 percent. A similar experiment should be attempted with polymer-bound metal cyclopentadienyl species of zirconium and hafnium. 55 Table 9 Relative Hydrogenation Rates at 20° Substrate Used Hydrogenation Ratea Relative Rate 1—Hexene 42.0 1.00 Cyclohexene 30.0 0.71 Cyclooctene 7.6 0.18 aIn ml of hydrogen per minute :_0.05 ml of hydrogen per min. Table 10 Hydrogenation Rates at 20° with Polymer-Attached TiCpCl2 Catalyst Precursor mmol Ti Olefin Concen. Rates 11 ml/min-nnmfl 1i Polymer-attached TiCpCl .011 l-hexene 1.15 36.5 2 Polymer-attached TiCpCl .011 cyclohexene 1.15 22.7 2 Polymer-attached TiCpCl2 .014 cyclooctene 1.15 13.6 Relative Rates 1.0 0.8 0.6 0.4 0.2 l L L ' I fi l-Hexene Cyclohexene Cyclooctene Figure 17. Relative Reduction Rates 57 Nitrogen Fixation Studies The study of the catalytic activity of the polymer-attached species towards nitrogen fixation was attempted. Homogeneous titanocene has been used for nitrogen fixation by several workers. Van Tamelin's method”.17 uses sodium-napthalene as the reducing agent and nitrogen is continuously bubbled through the solution. The effluent ammonia may be trapped in an acid trap and quantitatively estimated. In the method of Vol'pin and snur,18 the catalyst is treated with organolithium or Grignard reagent in diethyl ether under 100 to 200 atmospheres of nitrogen. Ammonia was quantitatively estimated after the reaction mixture was hydrolyzed. Attempts were made by Kroll14 using 20% crosslinked copolymer- supported TiCp2C12 following Van Tamelin's and Vol'pin-Shur's methods. No significant amounts of ammonia were produced by the polymer-supported titanocene beads. This was not very surprising, since researchers in titanocene nitrogen fixation believe that a dinuclear titanium complex is involved in the process, as shown in Figure 13. N<§:;;Z\ 2VE:E;Z\1‘ zrqi~t,r @f @7‘“ Figure 18. Reaction Scheme for Nitrogen Fixation This leads to the expectation, that if the supporting matrix in the polymer-supported titanocene system is rigid enough to prevent the 58 association of the supported titanocene (titanium centers being far apart), then no ammonia formation should be observed in the nitrogen fixation attempts with the beads. The polymer-attached ligands should be mobile enough to have two titanocene centers for the formation of dinitrogen complex. Lower crosslinked (2% crosslinked) copolymer beads may have more mobile polymer matrices compared to the rigid 20% crosslinked copolymer support. Attaching the titanocene species to this 2% crosslinked beads, one would expect the titanium centers to be somewhat closer in distance, to aid the formation of dinitrogen complex. 2% crosslinked polymer-supported titanocenedichloride was prepared as described in page 15. The dark red colored beads looked very much like homogeneous titanocene dichloride (as seen in Figure 11) and the titanium content was 0.78 nmnfl per gram of the beads. An initial attempt by Van Tamelin's method, by using about 3.9 mmol of titanium gave about 10% amnonia and was very encouraging. REFERENCES 10. 11. 12. 13. 14. 15. 16. REFERENCES . E. Bercaw, Thesis, University of Michigan, 1971. . Davison and S. S. Wreford, J. Amer. Chem. Soc., 99, 3017 (1974). J A J. E. Bercaw, R. H. Marvich, L. C. Bell, and H. H. Brintzinger, J. Amer. Chem. Soc., 99, 1219 (1972), references cited therein. R C W . H. Grubbs, C. Gibbons, L. C. Kroll, W. D. Bonds, Jr., and . H. Brubaker, Jr., J. Amer. Chem. Soc., 99, 2373 (1973). . D. Bonds, Jr., C. H. Brubaker, Jr., E. S. Chandrasekaran. - Gibbons, R. H. Grubbs and L. C. Kroll, J. Amer. Chem. Soc., 91, 2128 (1975). n W. Haag and D. D. Whitehurst, Belg. Pat., 721, 686 (1969) and German offer, 800, 379 (1969). R. Grubbs and L. Kroll, J. Amer. Chem. Soc., 99, 3062 (1971). A. J. Moffatt, J. Catal., 19, 193 (1970). C. M. R. Davidson, Netherlands Patents, 9, 914, 754 and 9, 914, 685 (1970). J. P. Collman, L. S. Hegedus, M. P. Cooke, J. R. Norton, G. Dolcetti, and D. N. Marquardt, J; Amer. Chem. Soc., 99, 1789 (1972), and R. H. Grubbs and L. C. Kroll, unpublished results. M. Capka, P. Suobda, M. Cerny, and J. Hetfleje, Tetrehedron Lett., 4787 (1971). R. H. Grubbs, L. C. Kroll, and E. M. Sweet, J. Macromol. Sci., AZ_(5), 1047 (1973). Z. M. Michalska and D. E. Webster, Platinum metal Reviews, 19, 65 (1974). ' L. C. Kroll, Thesis, Michigan State University, 1974. E. E. Van Tamelin, et 1., J. Amer. Chem. Soc., 99, 5707 (1968). 1., ibid., 91, 1551 (1969). E. E. Van Tamelin, g3_ 59 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 60 E. E. Van Tamelin, gt_a_1_., ibid., 9_2, 5251 (1970). M. E. Vol'pin and V. B. Shur in "Organometallic Reactions", Vol. I, E. 1. Becker, M. Tsutsui, Eds., Wiley-Interscience, New York, N.Y., 1970, p. 102. H. P. Rasmussen and G. R. Hooper, Hort. Science, 9, 425 (1974). Abbreviations used herein include: Cp'; cyclopentadienyl anion; CpH; cyclopentadiene. R. D. Gorsich, J. Amer. Chem. Soc., 99, 4744 (1958). C. U. Pittman, Jr., R. M. Hanes, Private Communication. K. W. Pepper, H. M. Paisley, and M. A. Young, J. Chem. Soc., 4097 (1953). T. C. Ray and A. D. Westland, Inorg, Chem., 9, 1501 (1965). R. S. P. Coutts, P. C. Wailes, and R. L. Martin, J. Organomet. Chem., 51, 375 (1973). K. Clauss and H. Bestian, Justus Liebigs Ann. Chem., 654, 8 (1962). A. I. Vogel, in "A Textbook of Quantitative Inorganic Analysis", Lowe and Brydone (Printers) Ltd., London, 1969. R. Pribil and M. Malat, in ”Gravimetric Analysis", part II, . Erdey, Ed., Pergamon Press, New York, N.Y., 1965, p. 540. . L. Hanson, g§_91,, J. Cataly§is, 99, 204 (1974). . Maslowsky, Jr., and K. Nakamoto, App]. Spectroscgpy, 99, 87 (1971). L D E. Samuel, R. Ferner, and M. Bigorgne, Inorg. Chem., 19, 881 (1973). E 1 E. R. Lippincott and R. 0. Nelson, Spectro. Chim. Acta., 19, 307 (1958). . Pittman, Jr., and G. 0. Evans, Chem. Tech., 560 (1973). . Bailar, Jr., Cat. Rev. - Sci. Eng., 19, 17 (1974). U C . W. Steward, J. D. Young, "Solid Phase Peptide Synthesis", H. Freeman, San Francisco, CA, 1969, p. 55. Gibbons, Thesis, Michigan State University, 1972. . E. Bercaw, Thesis, University of Michigan, 1971, p. 11. 61 38. J. P. Collman, Accounts Chem. Rev., 1, 137 (1968). 39. D. W. Hart and J. Schwartz, J. Amer. Chem. Soc., 99, 8115 (1974). 40. C. A. Bertele and J. Schwartz, J. Amer. Chem. Soc., 91, 228 (1975).