9131”“? ., ,.. .zn . . 1i 7... x If... .m L. 3.1.. 52.. r: 1 $21. aha; 4.. guvpihguufi “a“ 4.251.... r, .Il..li.‘>l.. . rifllgt 1.5: khnrt .2..‘..sth,t) 5.1%»...11...‘ .4 . . : t .1 119:1..211111E1111 f. flhfi~s . $.23} .17: 51.11.1911... iv 1? Viv: 1...: Av HVP..1~..‘.§WF.N.1§TWH¥ ¢b~§-kflu.’ in 9.111.. 3.1mr 90.!!‘1 . o A 0.1191! . ‘Jfllfl: “4.7 4a.}77ffitritv‘mwflfl‘ 1.. kn. . ...1tt..ll,m.rtzsno ”£119,111,... (1.. if: 1 1.; 1 1 , £Ax L .. 13..."... 1:2,. . Pu. tubs”: v . ,h 1 . . .v 1.». $11.7. Iii . .1: I . 6... .05! “1 a .. V n . X .. 51 a in) J .l . .‘rttE-ugggk} rill-l Quilg . 1 . i . ‘ its. 1 s, ‘11)»)... \4‘33 4 3) I 1 1 a T u. . , . V 60.2“?!" 510.151.; .EE.‘ .1 s... A fit: LI...“ . l ‘1!.l£ln{lnh.-i..tlll 1 l1. 2.5.114... v a . - . 1.. 1.11.1.1..-) {111.1 {1311113331113 .1 -- 1.}... 11.11. I [ti-7.5;; . 501111.]! 1.11:}. ‘ x I .. .. \ I . I. . .\. \\ . . z x . l1 . 1 A. . n a. . K. . 3. A . f . . . . . 4 .V ‘l A \ 1 . V I\ . 1 . 1 . . i 1.. hf. 1.11 .o. .11; . . 7. . . _ .1 (I. - 131.13.11.11 111.111} LR... .7 .. a -11. .. a. . . .. )i’!‘ 1 . . . . (xi).- .vbll) .(r 1.. .1. in... . I stylirtlvvl EL‘IEIIDA. . . . I‘lllIIuluviifllllllu .31)! . . . §L:1.¥. Jill {(11,11).}, .31 I 1.1. rlll..r...ll.\l1‘..ttlllib\{lu inflict it’ll 1.1"! if, .3! . £11.15 £115.31. 11.1.2.1. 1... [$111.5 111171.111":vt. fi‘ifixiibfs \1. . . v . 1711145311.? (13.1.1111... . II . itigtllLlultx 1.51311th "n I.» .Ct.lra‘.{vnr;-\llloi¢1n! 1:41.. . it; . loll-1.91: . . 5.1 1.111115611111111. 31.1.1 . . n )1 n(.tO.l>1|.) 51115;; 130...! V {shill-r; (.5231: . 1"; , 1111.53.11.3111111... 1111\‘1‘11 “11:11! 11%wi‘igfllv. . . . ll. hug...” Willa . Jfibfuyi¢1111£1111 1...; r a." 11.... Lil.llll\lf'fvl"ro\v\ .IP» I; . "Ifi}iii§ . . . . :1 l .11 L11. 177.11» ,11.r1.}‘ 'III'GLIV .- v 1.111‘lni'iis15’ . I‘lllillll‘; g . it", i [Effilt 11": t. 11‘! E. 5:11 y 5111.... Lnixohlrli.1.-t.}s.1x $1. 111. 1.11.2.7 311111.111... x . as . . . 1..th ,- 1" I. .|\IVI|1.L.|’,I\1.I'I.IIII'1"1.§I I. , . l 1 1.- .. .. .....I u-. '11:... . V ' II! V Ill 1 ‘I . .' II . . \ . 1... , . ,. 1... l “ I \JA 1. 11 . \ 1 .3. l3? .laln . 1.x: I .l . ‘1 4.. x. 1‘ L .. .71 I . t I . 4 1 .l u 7 I'YMKI. \A. f . 1 vi A 1 s I 1 r. 1 £ (\AL. .11.... . 1 1A. . r r n .7 . \ \ .i’. l 1‘ ¢ . .7 II ll. 1‘ \L ‘1 L1! A. It. It!) 1. 1 II 11.1 4’?! it! i1££ltk§l§l$fllt slulApv\L..l§£tL£i.ILrIH4|i.L I... 11.1 I All. Ill-1181;} E‘E’(E’§ 1| 4|!" 16.111. 111.. tiyl‘lltfhn‘” . .. :11 p11 . 1.15;..1118.‘ , 4 )1! Ac . {1iiiiu}: .. £3.31|||l1£p|b||1lu§lb§lfi1t$tlxltrl§llléifiil I. .3. Kit Ulrtht‘i‘utn III.” '11 R1. Vittgi 1‘11. "1? “‘12... g‘1tttgi évl‘frgg if {)1.X|\.lfl1.1|£$sllrttt£fl E . {I‘ll-0.111) . ..)\11.D.r|.1l.fll E111 1.111.114.) frilgitiffl‘xl EL 31.!!! till. .[Lliftlrtgttbllvi‘ uiilbtlgl.‘ L- .i 1 ”1?..05111‘13 .1111“- ‘11.!)«171‘1‘E1gtcél.!l. 51$." .Illtru‘glfnll‘tiulaynll‘nttvf‘ All: Illllll. it‘ll EllazfiVHflrl .- «I \E1>llllln:\ 311.111....111! . .11... . . . . . 1| lfl‘kkl 1.1, ltl. . .91); .1.) . ’15-)?" ’31. 1§§ flit-111...?! . A )1. E1111. . 1.. . t 1... . I . 1 , : innit-rix‘f .11! to]. .. In .l. 1%” 00" (fir .n l. 1 (t: I...) nil. nzt. .1.1‘IJ1|.1; thri‘ .13., . . 11.11111111111 1151 .vbsanlI-Il. .199)... 11.11. 1 Ir? [0.111 r 11.1\ .\ Iv . , . . r . . . «.11..01MPUH.1..1. 1.11, 11.79: 1- AI .1 11.112.12.111? .. i? :1. .. . . .I 1 . 1. I») . THES‘S lllllllllllllll|||ll|IIINIHIHllllllllllllllllNllllllNlIHl 193 01797 5685 This is to certify that the thesis entitled Part I. An Investigation of the Mechanism of the Olefin Metathesis Reaction using a Heterogeneous Catalyst; Part II. Development of a Polymer—Bound Catalyst for the Olefin Metathesis Reaction. presented by Sandra Jean Swetnick has been accepted towards fulfillment of the requirements for Ph.D . degree in .Chemiatm Mafia/X Major professor Date jf/Z 3/7? 0-7639 OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to Book drop to remove this checkout from your record. PART I AN INVESTIGATION OF THE MECHANISM OF THE OLEFIN METATHESIS REACTION USING A HETEROGENEOUS CATALYST PART II DEVELOPMENT OF A POLYMER-BOUND CATALYST FOR THE OLEFIN METATHESIS REACTION BY Sandra Jean Swetnick A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1979 ,I’ ABSTRACT PART I AN INVESTIGATION OF THE MECHANISM OF THE OLEFIN METATHESIS REACTION USING A HETEROGENEOUS CATALYST PART II DEVELOPMENT OF A POLYMER-BOUND CATALYST FOR THE OLEFIN METATHESIS REACTION BY Sandra Jean Swetnick The mechanism of the olefin metathesis reaction is examined using mixtures of cis,cis-Z,8—decadiene and cis, cis-2,8—decadiene-1,l,l,lO,lO,10-d6 over two heterogeneous catalysts. The metathesis of these diene mixtures produces a ratio of labelled and unlabelled butenes which can be compared to butene ratios predicted for a number of mechanisms. Previous studies using homogeneous catalysts have indi- cated that a non-pairwise exchange of alkylidene units is occurring. A metal carbene, initially formed by the inter- action of the catalyst and the co-catalyst has been Sandra Jean Swetnick proposed as the active site for this reaction. Most heterogeneous catalysts do not require such alkylating co-catalysts and therefore require an alterante source of the initial carbene for a metal-carbene chain mechanism to occur. Evidence to be presented here supports a non-pairwise exchange mechanism for olefin metathesis. The experi- mental ratio of labelled butenes formed is that previously predicted for a chain mechanism. In addition,the iden- tification of transient products has allowed the proposal of an initiation scheme which would produce the required metal-carbene active site during the "break-in" period observed with heterogeneous catalysts. The combination of these two pieces of experimental evidence strongly supports a metal-carbene chain mechanism for olefin metathesis with a heterogeneous catalyst. The second part of this investigation has studied the feasibility of using a polymeric co-catalyst to develop a polymer-bound metathesis catalyst. This is an unusual development in that the co-catalyst would serve both to activate the transition metal and to link it to the polymer support. A polymer-bound metathesis catalyst was successfully prepared using polystyryltrimethyltin with tungstenhexa- chloride. To The MSU Tolkien Fellowship without whom I never would have survived ii ACKNOWLEDGMENTS I would like to thank Dr. Robert H. Grubbs for his guidance, support and patience during the completion of this work. I also thank my father, my mother and my sister who offered unlimited love and support when I needed it most. I would like to mention Mrs. Bernice Wallace who simply believed I could do it. My fellow graduate students, in particular, Bob DeVries, Mir Mohammed Mohaddes, Cindy Huang Su, and Gwen Goretsas made graduate school more enjoyable. Finally, I cannot express the love, acceptance and deeply moving friendships I have experienced as a member of the MSU Tolkien Fellowship. iii TABLE OF CONTENTS Chapter LIST OF TABLES. LIST OF FIGURES PART I. AN INVESTIGATION OF THE MECHANISM OF THE OLEFIN METATHESIS REACTION USING A HETEROGENEOUS CATALYST. . . . . . . . . Introduction. Results and Discussion. Experimental. PART II. DEVELOPMENT OF A POLYMER BOUND CATALYST FOR THE OLEFIN METATHESIS REACTION . . Introduction. Results and Discussion. Experimental. REFERENCES. iv Page .Viii 42 84 .101 .101 .106 .111 .115 LIST OF TABLES Table Page 1 Typical Metathesis Catalysts. . . . . . . . 3 2 Predicted Values for A and 1/A Ratios of d :d 1,7-octadienes. . . . . . . 39 m0 m4 3 Product of the Ratios of Butenes Proposed for Pairwise and Non— pairwise Mechanisms . . . . . . . . . . . . 4O 4 Mass Spectral Analysis of Butenes Produced in the Metathesis of Labelled 2,8-decadienes, Over 13% M003 on Gamma Alumina, in the Absence of a Solvent. . . . . . . .~. . . . 45 5 Comparison of Product Ratios Using Katz's Method . . . . . . . . . . . . . . . 47 6 Ratios of Ethene, Propene and Butene Produced During the Metathesis of Trans, Trans—2,8— decadiene in the Absence of a Solvent.. . . . . . . . . . . . . . . . . . 51 7 Approximate Percentages of Labelled Propenes formed in the Metathesis of a 1:1 Mixture of 2,8-Decadiene-go and "d6 . o o . o o o o o o o o o o . o o o o o 60 ’b Table 10 ll 12 13 14 Page Approximate Percentages of Labelled Propenes. . . . . . . . . . . . . . . . . . 63 Mass Spectral Analysis of Propene Formed During the Metathesis of cis, cis-2,8-decadiene Over MoO3/CoO/A1203 Treated with a Tetramethyltin Co- catalyst. . . . . . . . . . . . . . . . . . 68 Analysis of cis-butene Produced in the Metathesis of a 1.05:1 Mixture of cis, cis-2,8-decadiene-d0 and -d6 . . . . . . . . . . . . . . . . . . 70 Analysis of cis-butene Produced in the Metathesis of a 2.23:1 Mixture of cis, cis-2,8-decadiene-d0 and ‘m6 . . . . . . . . . . . . . . . . . . . . 72 Analysis of cis-butene Produced in the Metathesis of a 1:3.3 Mixture of cis, cis-2,8-decadiene—d0 and 'Q6 . . . . . . . . . . . . . . . . . . . . 74 Analysis of cis-butenes Produced in the Metathesis of a 1:2.4 Mixture of cis, cis-2,8-decadiene-d0 and d6. . . . 75 Analysis of butene-g6 Obtained from the Metathesis of cis, cis-2,8—deca- diene-g6 Over MoO3/CoO/A1203. . . . . . . . 76 vi Table 15 16 17 18 19 Butenes Formed Using a Catalyst Treated with Sn(CD3)4 Predicted Butene Ratios for Carbene and Pairwise Mechanisms Calculated Values for DS/D0 x D3/D6 Conversion of 1,7-octadiene to Cyclohexene Over a Polymer-supported Catalyst. Calculation of the Turnover Number for the Polymeric Catalyst. vii Page 77 82 83 108 110 Figure 10 LIST OF FIGURES The carbene initiated chain mechanism for the metathesis of olefins. Carbene initiated, ring-opening polymerization . . . . . . . . . . . . . . Carbene initiation via an allyl intermediate . . . . . . . Initiation via alkylation and a-elimination. . . . . . . . . . . . Labelling studies of carbene initiation . Carbene initiation via a n-allyl intermediate Preparation of cis,cis-2,8- decadiene-d6 . . . . . . . . . . . . . Comparison of the rates of formation of butenes and propenes (Run 1) ................................... .. Comparison of the rates of for- mation of butenes (Run 2). . . . . . . . Production of propene and butenes relative to a propane standard . . . . . . . . . . . . . . . . . . viii Page 10 21 22 23 37 43 49 50 54 Figure Page 11 Production of cis-Z-butene and propene relative to trans-2- butene . . . . . . . . . . . . . . . . . . . 56 12 Initiation of the active catalyst for heterogeneously catalyzed olefin metathesis. . . . . . . . . . . . . . 57 13 Formation of propene-d0 and propene-d3 . . . . . . . . . . . . . . . . . 61 14 Formation of propene -d1, -g2, ’44 and -QS. . . o o o . o o . o o . o . o o 62 15 Conversion of 2,8-decadiene to cyclohexene. . . . . . . . . . . . . . . . . 65 16 Propagation steps in olefin metathesis . . . . . . . . . . . . . . . . . 79 17 Preparation of polymer supported tungsten carbonyl metathesis catalysts. . . . . . . . . . . . . . . . . . 103 18 Preparation of a polymer-supported tin-osmium complex . . . . . . . . . . . . 104 19 Preparation 0f the polymer-supported metathesis catalyst. . - . . . - - . - . . . 106 ix INTRODUCTION Since its discovery in 1964, olefin metathesis has received an ever increasing effort directed toward under- standing and controlling this reaction. A reasonable mechanism has been proposed for metathesis when conducted over non-supported catalysts. This study represents an effort to study the mechanism of metathesis when promoted by a heterogeneous catalyst. Although the reaction does not necessarily proceed by the same mechanism in both cases, similarities have already been observed. In order to demonstrate these similarities and to point out differ- ences,the current research using homogeneous catalysts will be reviewed. This will be followed by an outline of relevant studies of the metathesis reaction promoted by heterogeneous catalysts. The disproportionation of olefins was first recog- 1 They found that propene, ex- nized by Banks and Bailey. posed to a heterogeneous catalyst, exchanged olefin sub- stituents to produce ethylene and butene (Eq. 1). 2 CH2=CHCH3 g:e CH3CH=CHCH3 + CH2=CH2 (1) Later, a variety of homogeneous systems were reported which catalyzed the reaction at significantly lower temperatures.2 Many of the homogeneous systems are very rapid, producing an equilibrium mixture of products in a matter of seconds. Prior to the discovery of this reaction, similar catalysts had been used to promote the polymerization of cyclic olefins. This ring-opening polymerization was later recognized as a special case of the metathesis reaction.3 The majority of metathesis-promoting catalysts contain molybdenum, tungsten or rhenium and a non-transition metal component. The homogeneous catalysts and the insoluble non- supported catalysts are generally prepared from soluble precursors. The heterogeneous catalysts are solid materials prepared by combining a soluble metal complex with a solid support. Homogeneous catalyst systems are generally prepared by reducing transition metal complexes with non-transition metal organometallic co-catalysts. A survey of typical homogeneous systems is given in Table I. The majority of these systems show a high activity and selectivity under very mild conditions. The variety of metal complexes and co-catalysts which can be combined to produce an active metathesis catalyst has lead to an intensive study of how these components interact. Some of the results of these investigations will be presented later. Heterogeneous catalysts were the first to be recognized and developed. These are generally prepared by impregnating high-surface area supports with solutions of transition metal oxides or carbonyls. The solvent is removed under vacuum and the solid catalyst activated by heating in air Table 1. Typical Metathesis Catalysts. Catalyst Temp. K Ref. WClé-(C2H5)3Al 243 6 WCl6-LiA1H4 Ambient 7 MoC12(NO)2[P(C6H5)3]2- ReClS—(NC4H9)4Sn Ambient 9 Supported Catalyst Temp. K Ref. Moos-CoO-AIZO3 436 l MoO3-SiOZ 811 10 Mo(CO)6-A1203 344 l WO3-SiOZ 700 ll ReZO7-A1203 298 12 R82(CO)10’A1203 373 13 at SOD-600°C. It is then cooled to the reaction tempera- ture in a stream of inert gas.4 As shown in Table 1, these catalysts are optimally active at higher tempera- tures than comparative homogeneous systems. Supported catalysts tend to show an induction period before reach— ing maximum activity. This induction period has been care- fully studied and will be discussed later. The earliest investigations determined that transalkyl- idenation, cleavage of the double bond itself, was occurring. Calderon 33 gl.26 demonstrated double bond cleavage in a study of the metathesis of a mixture of 2-butene and per- deutero—Z-butene over a homogeneous catalyst. The only new product was 2-butene-l,l,l,2-;\i4 (Eq. 2). CH3-CH = CH-CH3 + T CH3-CH = CD-CD3 (2) 4.. CD3-CD CD-CD3 14 Similar results were obtained by Mol t al. in the meta- thesis of propene-2—14C over a heterogeneous catalyst, R8207-A1203 (Eq. 3). + CHZ=*CH-CH3 + CH2=CH2=CH2 + CH3-*CH=*CH-CH3 (3) The ethene showed essentially no radioactivity whereas the butene showed a radioactivity twice as high as that of the starting material. These results were verified by Clark and Cook15 using Moo3-cOo-A1203. Experiments with 1—14C- propene and 3-14C-propene were conducted using a variety of heterogeneous catalysts in an effort to examine the pos- sibility of an allyl intermediate. Mol gt 31.16 used Re207 and observed that the two methyl groups retained their k15 catalyzed identity. A similar study by Clark and C00 by MoO3-CoO-A1203 was consistent with Mol's work at low tem- peratures. However, at higher temperatures the propene isomerized prior to disproportionation. This was confirmed 17 in a separate study by Woody 33 31. Woody noted however that isomerization of the Z-butene product was negligible (Eq. 4). low temp. CH: = CH-CH3 -—-—+> CH: = CH: + CH3-CH=CH-CH3 * 1 high temp. * * (4) CHS-CH=CH2 CH2 = CH2 + CHs-CH=CH-CH3 + cross products The authors concluded that an allylic intermediate was excluded at low temperatures but could not definitely be excluded at higher temperatures. The first mechanistic schemes involved a concerted 2 + 2 addition of 2 complexed olefins to form a "quasi- cyclobutane" metal complex which could then decompose to give two new olefins18 (Eq. 5) -.- R ~HCTCH-R1 Rl-CH----CH-R1 R -CH CH—Rl M ‘~-~ M : M (5) RZ-HC#CH-R2 RZ-CH----CH-R2 Rz-CH CH-R2 Experimentally it was found that cyclobutanes, were un- reactive under metathesis conditions and were not observed as metathesis side-products.19 Also recent work demon- strated that similar processes occurred by non—concerted pathways20 (Eq. 6). I I Rh _ -Rh @3 7 (6) The first non-concertednwmhanicsinvolved the coordination of two olefins to the metal center and a "pairwise" exchange of alkylidene units via a five—membered metallocycle.21 This proposal was based on the synthesis and decomposition of a tungsten metallocycle. Under metathesis conditions this metallocycle gave metathesis-like products (Eq. 7). D D D D D g g --9 2 ; -——A < S (7) . iv = w D L1 L1 C14 C14 . i l WCl6 2CH2=CDH CH2=CH2+CHD=CHD Further support for this proposal developed from the isola- tion and characterization of metallocycles of iridium,22 23 and titanium.24 platinum In such a pairwise exchange the fates of the two ends of an olefin are linked and should give predictable products. The observation of unexpected pro- ducts lead to the proposal of yet another mechanism. 25 proposed a chain re- In 1970, Herisson and Chauvin action mechanism which provides a pathway for the scrambling of the alkylidene units in a non-pairwise fashion (Figure 1). This metal-carbene, chain reaction mechanism was pro- posed to explain a variety of experimental observations which seemed inconsistent with a pairwise exchange. The ring—opening polymerization of cyclic olefins had been utilized even before the metathesis reaction was recog- nized by Banks and Bailey.3 A careful study of the products of these reactions revealed that high molecular weight linear polymers were formed early in the reaction. Cal- 26 gt al. reported a polymer of molecular weight deron 200,000-300,000 at 6% conversion of cyclooctene. Also formed in this reaction were low molecular weight macro- cyclic oligomers. A pairwise scheme could account for these products only if the growing polymer preferentially re- mained attached to the catalytic site. This would require that the exchange of olefins onto and off of the metal center be slower than the actual metathesis. However, recent studies have demonstrated that olefin exchange is much faster than [M] initiation [M]=CHR R'HC=CHR' R' R' R'HC CHR' I l —————9 <—-———- n1 R [M]=CHR ~R'HC=CHR HfiR' HR' HCR' <——-——- [M HR H 5 [M] R"HC=CHR" etc . Figure l. The carbene initiated chain mechanism for the metathesis of olefins. metathesis.27 k28 Dolgoplos suggested that a pairwise scheme ought to give low molecular weight products early in the reaction which slowly build to higher polymers. The observed products were more characteristic of a chain mechanism. Chauvin's carbene, chain mechanism could explain these unexpected results. High molecular weight products would 1M3 expected early in a chain mechanism and, as shown in Figure 2, these polymers would be expected to be linear. The lower molecular weight cyclic oligomers would be formed by a ”back-biting” process as also shown. 29 Scott verified that the high molecular weight pro— ducts are linear and more importantly that the macrocycles are formed by an intramolecular process. This is in agree- ment with the pr0posed "back-biting" scheme. A pairwise mechanism would be expected to form macro- cycles by the intermolecular reaction between two smaller rings. The only intramolecular pairwise reaction would be a "pinching-off" of a larger ring to form two small rings.30 Chauvin's carbene scheme was originally proposed to explain the products resulting from the metathesis of anfixr ture of a cyclic olefin and an unsymmetrical olefin (Eq. 8). RICH CHRI CHR1=CHR2 * ——’C C C R2CHfl CHR2 CHR1CHR2 + R1CH=CR1H + RZCH=CR2H (8) D E 10 1 /H [M "—C\R + G _.) [M]=C\H R R [M] = [M]=c: :C== —_9 H Figure 2. Carbene initiated, ring-opening polymerization. 11 In two independent studies, the symmetrical products B and 25’31 Pairwise C were observed early in the reaction. exchange of alkylidene units would yield only product A from the starting materials. B and C could only arise by a secondary metathesis between D and E and the cyclic olefin. This would again require olefin exchange onto and off the catalytic center to be slow compared to metathesis. On the other hand,Chauvin's carbene scheme would uncouple the alkylidene units of the starting olefin and allow the direct formation of B and C. If the products are allowed to equilibrate, either mechanism would be expected to give the observed products. These investigations prompted more exact studies into the mechanism of metathesis. Olefin reactants which would be less likely to equilibrate in secondary reactions were designed using derivatives of 1,7- octadiene. Metathesis of 1,7-octadiene produces an almost quantitative yield of cyclohexene and ethylene32 (Eq. 9). _ H2 ——) I + CH2=CH2 (9) ——H 2 Cyclohexene is one of the few simple olefins that does not undergo metathesis. The volatile ethene can be easily removed from the reaction vessel and analyzed. The inertness of cyclohexene and the rapid removal of ethylene allows equili- bration of these products to be minimized. The first of these systems was designed by Grubbs33 12 using mixtures of l,7-octadiene and l,7-octadiene-l,l,8,8- CD2 CH2 + + DZC CD2 34 A similar system was developed by Katz using mixtures of deuterated and non-deuterated 2,2'-divinylbiphenyl (Eq. 11). D C CD D C'-—CH 2 . 2 (11) Both of these diolefin mixtures produce a mixture of labelled and unlabelled ethenes. By a pairwise mechanism, formation of this ethene would be an intramolecular process and a majority of the fully deuterated and the non-deuterated products would be expected. Partially deuterated ethene would arise from the cross metathesis of products or from label scrambling in the reactants. Calculations based on a pairwise model indicate that the ratio of ethene-d0: ethene-g2:ethene-g4 products would be l:l.6:l from the meta- thesis of 1:1 mixture of labelled and unlabelled starting materials.”’34 This is the maximum amount of ethene-g2, the cross product, expected by any pairwise exchange of alkylidene units prior to equilibration. 13 A carbene chain mechanism would uncouple the alkylidene units and ethene formation would no longer be an intramo- lecularprocess. Such a process would randomly recombine the alkylidene units giving an equilibrium mixture of pro- ducts. Starting with a 1:1 mixture of labelled and unlabelled diolefins this random process would produce a 1:2:1 ratio of ethene-d0:ethene-dzzethene-d4. Experimental results from both Katz's and Grubbs' work indicated a random mixture of labelled and unlabelled products. The authors proposed that this supported a carbene mechanism. Criticism of both these experimental schemes centered on the fact that equilibration of product ethenes would produce the observed random labelling pattern regardless of which mechanism was operating. Although a variety of techniques were attempted, an irrefutable lack of scrambling was not demonstrated. A more elaborate experiment was designed to eliminate the possibility of product equilibration. Utilizing both the selectivity of a l,7-octadiene system and the stereo— selectivity exhibited by metathesis catalysts, Hoppin35 repeated these investigations with labelled and unlabelled cis,cis-2,8-decadienes. The products from this metathesis would be cyclohexene and a mixture of labelled and un- labelled cis- and trans-Z-butenes. The stereoselectivity of the metathesis reaction has been clearly demonstrated. In metathesis, cis olefins produce predominantly cis products and trans olefins form 14 predominantly trans products. Casey36 has proposed a puckered metallocyclobutane model to explain these obser- vations. Basset37 suggested that this stereoselectivity is actually controlled by the approach of the olefin to the metal carbene. As the reaction continues an equilibrium mixture of cis and trans products are formed. Hoppin and Grubbs proposed that this stereoselectivity could be used as an internal probe to check the extent of reaction and by analogy, the extent of scrambling (Eq. 12). /=\ Cit (it 0 __ + —) + + CH fl 3 CD3 (12) CH3 CD3 + CD3 CD3 A 1:1 mixture of labelled and unlabelled 2,8-decadienes was metathesized and the cis-Z-butenes analyzed for isotopic distribution. Using the transzcis 2-butene ratio as a measure of the extent of reaction, the isotopic labelling pattern was extrapolated to zero stereoisomeric scrambling. The authors suggest this extrapolated value is the isotopic ratio at zero label scrambling. As in the earlier investigations, the product ratio was that expected for a carbene mechanism. The experiments with l,7-octadiene derivatives were all conducted over homogeneous catalysts. In an effort to distinguish between a pairwise and non-pairwise mechanism 15 over a heterogeneous catalyst, the study using cis-, cis-2,8-decadienes has been repeated using a molybdenum on alumina catalyst. The results obtained allow insight to both the mechanism of such catalysts and the initiat- ing scheme necessary for such a mechanism. At this point, the characteristics of carbenes and carbene initiation schemes will be discussed. Following this will be an out- line of pertinent investigations over heterogeneous catalysts. As evidence built indicating a non-pairwise mechanism, ef- fort was directed at demonstrating a metal-carbene intermediate. In 1964, Fischer g£‘§;.38 reported a stable carbene complex prepared from tungsten hexacarbonyl and phenyllith- ium. In these early complexes, the carbenoid carbon was substituted with an electron-rich heteroatom (Eq. 13). R\ /OLioxEt20 R\ //ocn3 (13) co ¢_Li 9 l)H+/H20 9 w “"'9 if 2) > W ‘ EtO CHN co 2 (c0) 2 2 (c0) ( )5 s EtZO 5 Casey39 and Berkhardt prepared one of the first stable carbene complexes which was not heteroatom-stabilized (Eq. 14). ¢ . (C0)5W=C ‘ (C0)5W=C\\\ (14) \\ 2)HC1C7 OCHs -78°C ¢ a: {min 16 This was followed by Schrock's4O isolation and characteriza- tion of a simple carbene from the reaction of a metal alkyl with a transition metal halide. Analysis by X-ray and NMR indicated the presence of a true double bond (Eq. 15). C(CH3)3 (#TaCls Fm C Ta c=/\{ 3 4 The first demonstration of a carbene initiated meta- 159 (15) thesis was reported by Cardi using a mixture of substi- tuted unsaturated amines with a rhodium-phosphine catalyst. A heteroatom stabilized rhodium-carbene was isolated from the reaction (Eq. 16). R R 1— —. R R, N N R R N (16) + > N R Rf I I R R R J—N’ ’—' N R R + (L)3Rh R 41 Casey reported that the diphenylcarbene tungsten complex would undergo alkylidene exchange with simple olefins and carbene exchange with substituted olefins (Eq. 17). 17 RCH=CHR ¢ ‘; :rH=CHR (c0)5w=c/ \ ‘1’ s, ¢;CH=CHR R-O\\ ¢ CH=CHR (17) ¢// + //OR (c0)5w=c \¢ This exchange is an excellent demonstration of the capa- bility of such complexes to initiate metathesis like reac- tions. However, this exchange was not catalytic. In a related study, Katz4z has reported the use of the diphenyl- carbene tungsten complex at higher temperatures to meta- thesize simple olefins. However, an attempt to repeat this investigation has failed.43 Dolgoplosk44 decomposed phenyldiazonethane to nitro- gen and stilbene using tungsten hexachloride. The reaction mixture was shown to be an active ring-opening polymeriza- tion catalyst (Eq. 18) ¢\\ ¢\\ /,¢ active H/CN2 + wc12 -—————9 H/;:=c\ + (18) H catalyst Cyclopropanes are a common side product in many of these studies. Metallocyclobutanes are known to decom- pose to give mixtures of olefins and cyclopropanes. Puddphatt45 decomposed a platinocyclobutane and the 18 major products were the result of either reductive elimina- tion to produce cyclopropanes or hydrogen transfer to yield olefins (Eq. 19). Pyr\\\\ql A A: /P|t *1 CH2=CH'CH3 4' i: (19) Such studies suggest metallocyclobutanes as a secondary intermediate in the metathesis of olefins. 36,41,46 Casey prepared a series of substituted metal carbene complexes and studied their reactions with olefins. Sterically hindered carbenes produced mostly the expected olefins, but less hindered carbenes formed a high percen- tage of cyclopropanes. Casey suggested that the hindered carbenes might lose a carbonyl ligand and then add the olefin to form a tetracarbonyl metallocyclobutane. This would cleave to give a new olefin and a new metal carbene. On the other hand, less hindered complexes might add an olefin without the loss of a carbonyl ligand. This would produce a pentacarbonyl tungsten metallocyclobutane which could eliminate W(CO)5 producing the observed cyclo- propanes (Eq. 20). Gassman42 has demonstrated that metathesis catalysts can be used to convert cyclopropanes to olefins. However, as indicated by Casey's work, different active species may be involved in this process. 19 'CO (CO) 4W-Cl 12 > --—-9 H 0,0, (cm w + 2 2 4 HZC-CRZ (CO)SW'-C\R (20) wrcm 5 2 2 + i: Metallocyclobutanes have been isolated from the reac- H tion of cyclopropanes with platinum complexes. Crystal structure analysis of these compounds has indicated that the metallocycle is bent, an observation which has been used to explain the stereospecificity observed in the metathesis of olefins (structure l).36 M 1 "b As mentioned earlier, these platinocyclobutanes decompose upon heating to form olefins and cyclopropanes. A unique 45 may relate to the rearrangement observed by Puddephat actual process occurring during a metathesis alkylidene exchange (Eq. 21). Cl fl PYT Pyr \\ ‘ 21 ‘7’ Pt :;::::2 6;? Pt:::r‘¢ ( ) Pyr 1 PYT 1 C1 C1 20 Tungsten metallocyclobutanes have been prepared and de- composed to yield cyclopropanes and the expected metathesis products (Eq. 22).47 ¢ ¢ ¢ ¢ UV ¢ ¢ ¢ CH2 Recent work with metallocyclobutanes has implicated n-allyl complexes as possible precursors for some active 48 prepared tungsten and metathesis catalysts. Green molybdenum metallocyclobutanes by the reduction of h3-R- allyl complexes (Eq. 23). These metallocycles could undergo B-C-C bond cleavage to form metal-carbene complexes. The isomerization of a metal-olefin complex to an allyl-hydride complex is a well established reaction (Eq. 24). Green has proposed an initiation scheme based on these findings in which the initial carbene 2 could be formed from a metallo- cyclobutane as shown in Figure 3. + ” LiCH3 + w __~, w :>CH (23) ~ 3 2 2 H H R Igrg (24) 21 H / H {:12 c GIFT C\ —-r CH4 ‘\\CHz—’ HZC/ \CH 2 c 2 _H \ / M. ECH ‘/ 2 H 11 3 CH2 M 4 Figure 3. Carbene initiation via an allyl intermediate. The initiation of the carbene chain carrying species has been carefully investigated. For those catalyst systems which use an organometallic co-catalyst a reason— able pathway has been proposed. The co-catalyst,appears to alkylate the transition metal forming a o-bonded species which can transform to the active carbene. This pathway 49 who observed the production was suggested by Muetterties, of alkanes during the interaction of dimethyl zinc and tungsten hexachloride. The use of a deuterated solvent did not produce deuterated methane. In place of the simple complex shown here, Muetterties suggested that a bridged species involving both tungsten and zinc would be more likely as an intermediate for the alkylation and subse- quent a-H elimination shown in Figure 4. 22 H \ | /C”3 \ y -CH4 \ a) w\ ———> JW‘ ———9 / =cn2 H 1 CH3 |/HC1 H -- b) \W< --—-> \W (CSHS)ZW _CD3 <—' (CSHS)2W \D Hoppin35 recently published supporting evidence indi- cating that the initial carbene is formed from alkyl groups originating in the co-catalyst. Metathesis of 2,8- decadiene theoretically produces only cyclohexene and 2- butene. However the metathesis of 2,8-decadiene using WCl6/SnMe4 produced an initial burst of propene. Following 23 this initial propene production, only normal metathesis products were observed. Metathesis of 2,8 decadiene (l,l,l,lO,lO,lO-d6) with this same catalyst produced propene- 3,3,3-d6. Use of a deuterated co-catalyst produced pro- pene-1,l-d2, Figure 5. Similar results were obtained with H3 H wc16 / a) — ( : Hzc=C “' Sn CH ) ' 3 4 \ H3 CH3 D H __ 3 wc16 __ / SnCCHC)7 \ D3 3 H3 H WC16 / C) _ \ D2C=C — Sn(CD3?4 \ H3 CH3 Figure 5. Labelling studies of carbene initiation. These results are consistent with Meutterties' scheme in which the initial carbene formed from the co-catalyst interacts with the diolefin to produce propene in the first turnover. After this first turnover only the usual metathesis products were observed. 10 An alternate route was proposed by Farona to explain 24 small amounts of unusual products observed early in the metathesis of l,7-octadiene over Re(CO)5Cl/EtA1C12. Forona suggested alkylation of one of the carbonyl ligands as the source of the initial rhenium propylidene (Eq. 26). EtAlClz H C1(CO)4ReCO -—————$ C1(CO)4Re(CO-Et)-—9[Re]=C-Et (26) The majority of investigations presented thus far have used non—supported catalysts. The initiation schemes have involved organometallic alkylating agents mixed with soluble precursors. The present investigation was designed to ex- amine the mechanism of olefin metathesis promoted by sup- ported catalysts. The mechanisms outlined for homogen- eously catalyzed metathesis do not necessarily apply to supported catalysts. This study was undertaken to see if parallels exist between these two general classes of catalysts. Investigations into the nature of these heterogeneous catalysts will now be reviewed to set the proper framework for the new evidence to be presented in this Thesis. As each study is reviewed those results which were formally unexplained will be reexamined in the light of new findings to be presented later in this Thesis. As mentioned earlier, heterogeneous catalysts were 14,16 and the first to be developed and investigated. Mol Cook15 conducted experiments with 14C labelled propene over rhenium oxide-alumina and molybdenum oxide-alumina 25 catalysts. Their results demonstrated that metathesis involves a transalkylidenation and at low temperatures, does not involve an allylic intermediate. At higher tem- peratures Cook and Woody17 demonstrated that an allylic intermediate could not be definitely ruled out. 51 also used a rhenium oxide-alumina catalyst to Mol study the metathesis of Z-deuteropropene. He concluded that tranalkylidenation was the major reaction as supported by Olsthoom who observed CZHZDZ as the only new product in the metathesis of a mixture of C2H4 and C2D4 over Re207/A1203. A simple transalkylidenation would produce ethene that is totally free of deuterium and Z-butene with 2 deuterium atoms. Experimentally a small incorporation of deuterium into the ethene was observed. The authors suggested iso- merization as the source of this deuterium in the ethene. H D D D \ / \ / c=c -—> CH2 CH2 + c:c / \ expected / \ (27) H l CH3 CH3 CH3 HD EC C HDC= H at M .1. 2 1W M M=CH2 observed 26 However, results to be presented in this paper suggest the formation of an initial carbene via an initiation scheme similar to that proposed by Green would better explain the deuterium observed in the ethene, (Eq. 27). The majority of heterogeneous catalysts have been pre- pared by impregnating high surface area supports with solu- tions of the oxides or carbonyls of molybdenum, tungsten or rhenium. These catalysts generally require activation at high temperatures. A notable exception to this are the catalysts prepared from the allyls of dinuclear and mono- nuclear complexes. These catalysts do not require activa- tion at high temperatures but are immediately active for metathesis.52 A recent study90 has demonstrated that the interaction of Mow-C3145)4 with SiOz at 20° in pentane, followed by the removal of pentane at 20° produces a catalyst with a high initial rate of disproportionation. Either reduction in H2 or oxidation in 02 of this catalyst at high tempera- tures greatly decreased its initial activity. This investi- gation strongly supports metal-allyl complexes as possible precursors for the active site in some heterogeneous catalysts. The vast majority of heterogeneous catalysts require heating at high temperatures and are activated by controlled H2 treatment. A major barrier to understanding the role of the 27 catalyst in heterogeneous metathesis has been that these supported complexes are not well defined compounds. Many sophisticated analysis techniques have been applied in an effort to identify the active site on these catalysts. This investigation has been complicated by observations 62 such as those of Burwell and Brenner who observed that three different forms of active catalyst can be prepared by combining Mo(CO)6 and A1203 under different conditions. Each of the three forms evolve different gases during preparation and exhibit different catalytic activities. 63 Davie et al. reported the infrared spectra of a molybdenum hexacarbonyl alumina catalyst before and after activation at 1009 in vacuum. Their results indicated the loss of at least one carbonyl group. Before this acti- vation procedure the supported catalyst was not active for the metathesis of propene. The nature of the active site in a MoO3-A1203 catalyst 60 was investigated by Basset. His results indicate a loss of surface —OH groups leading to an active site consisting 6+ of "an M0 ion situated in an octahedral cavity of sym- metry C4VJ' An ESR study of the interaction of cis 2- pentene with this catalyst indicated either the formation of a charge transfer complex or in the author's opinion 6+ 2- O by the reactant with a more likely "reduction of M0 the formation of an allyl complex" (Eq. 28). 6+ 5+ - - + C5H10 + Mo C5H9 + OH (28) Mo 02- 28 Although his results did not allow a distinction to be made between these two possibilities, he did conclude that a "relatively stable complex” had formed which could then interact with a second molecule of olefin. The active sites on WOS/Alzo3 catalysts were investi- 55 gated by deVries. He suggested that dehydration of the catalyst during heat activation lead to structure 3 and that reduction with propene leads to structure 4 with the loss of water (Eq. 29). O + __ 0—. w6+/ \A13+_ o 0—— w4._ 0—— A1...— 0 ‘ \O/ | C; '0 i 0 Ci 5 \ l/ | (29) A1 —_.w._. \ Al/ -—W -— / |\ | / \ | w? 4, deVries was unable to detect such reduced sites using ESCA. He suggests that only0.l4% of the total number of tungsten atoms form active sites and that there may be no physical techniques capable of revealing the exact nature of such a low concentration of active species. An indication of tungsten carbonyl-alumina complexation has been reported by Basset, structure 5.82 29 [W]C=O-+ Al 4 The heat activation procedure just discussed does not produce a catalyst with maximum activity, rather the ac- tivity of the catalyst gradually increases after the intro- duction of the olefin reactants. This has been clearly 53 This catalyst significantly demonstrated for WO3-Si02. increased in activity during the first few hours of con- tact with the propylene. Concurrently, the catalyst changed in color from pale yellow to deep blue-violet. Such a color change indicates a reduction of W03 and sug- gests that the propene itself is reducing the catalyst. ESR studies of Moos/A1203 have also indicated a reduction of the M003 during the formation of the active species.54 An increase in the initial activity of several supported catalysts could be induced by pretreatment of the catalyst with a reducing gas such as H2 or CO. 55 From temperature-dependent studies, deVries concluded that active sites for metathesis are formed during this "break-in” period. It has been observed that the activity of a catalyst (one fully activated by propene) could be 56 greatly reduced by a 3 hr purge with helium. Retreat- ment with propylene induced another break-in period. 57 demonstrated that the activity of a tungsten Pennella silica catalyst could be greatly enhanced by adding small amounts of polyenes such as 1,5-cyclooctadiene to 30 58 suggested that the olefin starting materials. Pennella normal break-in may involve the formation of metal ligands from the reactant olefins. In a quantitative analysis of transient products formed during this break-in period, Luckner59 demonstrated a large initial production of ethene and some butadiene along with a small amount of 2-butene during propene metathe- sis. A fully activated catalyst gave approximately equal amounts of ethene and 2-butene and no butadiene. With a WOS-SiO2 catalyst he demonstrated that simple reduction with H2 did not eliminate the break-in period. Luckner suggested that some additional phenomenon is a part of this break—in process. This initial excess of ethene may be explained using the n-allyl-metallocyclobutane initiation scheme suggested by Green (Eq. 30) H H2 -H \\\14/’/ C C C I ’\ / \ [[12 CH2 (30) CH2/ \CH2 ‘9 CH2 H2 A H Unusual product ratios have been observed with other catalysts during the initial contact with olefins. 60 Basset observed an unusually high level of butene early in the metathesis of Z-pentene over MODS-A1203 and M003- 61 also noted a higher Co-AlZO3 catalysts. Davie 33 31. amount of ethene in comparison with butenes when propene was metathesized over Mo(CO)6-A1203. 31 The break-in period appears to involve some process which reduces the catalyst and generates active sites for metathesis. Reduction alone does not produce a fully active catalyst. These observations lend support to the n-allyl mechanism which will be presented in a later sec— tion. Kinetic studies on heterogeneous catalysts have given results with poor reproducibility. Several studies by Moffat gt al.64 have demonstrated the non-uniformity of the surface of these catalysts. As a result the number of active sites may vary with changes in temperature. This may invalidate calculated activation energies and heats of adsorption since these are usually derived from tempera- ture-dependent studies. This aspect of heterogeneous catalysts must be considered when evaluating such reported data. The activation energy for metathesis has been reported as 7.7 KCal/mole for Co/MoOS/A120365 and 21.6-18.6 KCal/ mole for a tungsten supported system.66 The activation energy during the break-in of a tungsten-silica catalyst has been reported to be 47 KCal/mole.59 Attempts to relate experimental kinetic data to model mechanisms has proved difficult. In 1967, Begley and 67 reported pseudo-first-order kinetics for the Wilson metathesis of propene over WOS-SiOZ. From this they pro- posed a Langmiur-Rideal model in which a single site 32 chemisorbed molecule interacts with a gas-phase molecule. This model was critized on the basis of mass transfer effects.53b The proposal of BradshawB quasicyclobutane mechanism lead to a number of studies which attempted to support a dual-site mechanism. Theoretical studies centered around a Langmiur-Hinshelwood modeL in which two propene mole- cules adsorbed on adjacent active sites,interact. 70 proposed a similar four-center mechanism based Giordano on the interaction of two adjacently adsorbed molecules. The experimental evidence supporting a carbene chain mechanism for homogeneously catalysed metathesis prompted a reevaluation of reported kinetic data for heterogeneously 169 established a correlation of catalysed metathesis. Mo rate data for propene metathesis with a model in which two propene molecules are successively adsorbed on the same active site. He then undertook the reevaluation of the published data which had been correlated to a dual site model. In several cases he demonstrated that a model which involves only one metal site could reconcile the data equally well. Such a single site mechanism might be ex- pected for a carbene chain reaction. More complex metal species have been postulated in which two transition metal atoms form an active site.52 Also proposed are bridged bis-metal species involving a transition metal and a non-transition metal from either 33 the co-catalyst or the heterogeneous support material. Muetterties has proposed a bridged bis-metal species for a homogeneous catalyst structure 6.49b H \WXM / / \c1/ x 9 Support for such a bridged complex is varied. Halo- genated olefins have been shown to increase the activity 71 The halogens may facili- of some supported catalysts. tate formation of such bridged species. Triethyl aluminum has been used to increase the activity of a Co-MoOS/ A1203 catalyst. Tributyl phosphine increased the activity 73 of a tungsten oxide/silica catalyst. Tetrabutyl tin has been used to promote the activity of a Re207/A1203 catalyst,74 Formation of the proposed bridged species would explain the activating effect of these additives. In addition, Giordano70 has provided evidence for dimeric Mo(V) species in MoOs/Alzo3 catalysts. He suggested that such configura- tions of bis-molybdenyl type are the active sites for this reaction. With the proposal of a carbene mechanism for the homo- geneous systems’interest began to turn toward investigat- ing a similar mechanism in heterogeneous catalysis. O'Neil and Rooney reported that a Moos-CoO-Alzos 7".“ “ |.‘l‘l “i.l“l 1|“‘1‘1 ‘ 34 catalyst decomposed diazomethane into nitrogen and ethene under the same conditions as for propene metathesis(Eq. 31).75 CHZNZ ; CH2=CHZ + N2 (31) Similar studies using phenyldiazomethane were used as sup- port for a carbene mechanism promoted by homogeneous cat- 28 alysts. This reaction has been questioned as support for such a mechanism because many catalysts will promote this conversion.76 In 196% Banks77 patented a process which uses a tung- sten oxide-silica catalyst to convert ethylene into propene. The assumed mechanism involved dimerization to yield butene followed by metathesis with another ethene to produce prOpene. More recently,O'Neil78 has published a direct conversion of ethene to propene over a Mo(CO)6/A1203 catalyst. Labelling studies have indicated that this is a direct conversion and does not proceed via prior di- merization to butene. O'Neil proposed several possible mechanisms which involve splitting an ethylene into methylenes and addition of a methylene to another ethylene on.as an alternative,the recombination of three methylenes. A noteworthy development from O'Neil's work is the observa— tion that ethene is slow to convert to propene over a fresh catalyst. However, if ethene is added to an identical system where propene has been reacting, the conversion of 35 ethene to propene is rapid. The authors suggested that a n-allyl intermediate might explain this apparent selectivity. As mentioned earlier, a variety of studies have indi— cated that the metathesis catalyst becomes active only after having been brought into contact with alkenes. From this it follows that some active species is formed from the interaction of the catalyst and the olefin. If this species is a metal-carbene then some pathway for the initial formation of this carbene is necessary. Several reasonable pathways have been suggested. One of the earliest schemes for carbene formation was proposed by Dolgoplosk28 based on a similar scheme developed by Clark79 for methoxycarbene platinum compounds. This process has not been observed directly in metathesis. In- deed, for heterogeneous catalysts activation by ethene is slow indicating that this process may be slow (Eq. 32). 1 KC) R\c :54] (.2) HM] __. I H/C\RZ I/C\H\R2 A more elaborate scheme begins with the B-H addition of a metal hydride to the olefin, followed by a-H elimina- tion as shown in Eq. 33.80 36 B-H [M 3—H + CH2=CH-R -2 [Ma—CHz-CHZ-R ~ -H l“ (33) H [11}:HCHZR *— [lit]=CH-CH2-R active catalyst H This method has been suggested for the initiation of sup- ported metathesis catalysts with the metal hydride origin- ating from a support surface hydroxyl(Eq. 34).78 (34) 3”“ * [1 -.-> >0-[MH1 -—’ This proposal is favored by the experimental evidence of increased activity in the presence of H2 or HC1. This dissertation will present evidence favoring an allylic initiation scheme such as that proposed by Green31 and later suggested by the work of O'Neil.78 This allylic intermediate is further supported by Olsthoorn's81 ob- servation-that "propene, butene and higher alkenes reduce a Re207/A1203 metathesis catalyst at ambient temperature 37 Figure 6. Carbene initiation via a n-allyl intermediate. and sub-atmospheric pressures, while ethene shows hardly any reduction capacity under these conditions. The metathesis activity is generated by this reduction". The early studies with 14C labelled propene excluded an allylic intermediate for the overall metathesis reaction at low temperature. However, these results do not exclude an allylic initiation scheme which would effect only the first catalytic turnover. This initiation scheme is strongly supported by the immediate activity reported for n-allyl metal complexes 52 It is not clear whether or not a supported on alumina. break-in effect is observed with these catalysts. The present investigation has repeated the work of 38 Hoppin with labelled 2,8-decadienes using supported catalysts. It was assumed that results similar to the homogeneous metathesis of 2,8-decadiene would be produced. That proved to be a false assumption but has resulted in an intriguing mechanistic problem. A notable difference between the two catalytic systems was the observation of an isotope effect when catalyzing the reaction with supported metal complexes. Katz34 and Grubbs considered the possible results of an isotope effect and judged such effects to be negligible with homogeneous catalysts. Grubbs noted that the catalyst formed from tungsten hexachloride and butyl lithium, which has been reported to be heterogeneous in nature, exhibited what might have been an isotope effect. However, he explained this apparent result by a mono- hydride scrambling scheme, supported by the observation of cyclohexene-l-dl and ethene-d1 and ~d3 (Eq. 35). D H D . CD2 H D ' D (35) -—-) ——> I“ Hi <- *- [H D CD2 [M D 2 In order to exclude the possibility of an isotope effect, Grubbs made use of the calculated ratio expected 39 2 where A is the ratio in a carbene mechanism namely 1:2A:A of starting l,7-octadiene-dozd4. His treatment is based on the observation that the values for the g4zdzzd0 ratio are exactly reversed when a dozd4 ratio of l/A is used in place of a Q0:84 ratio of A. An example is shown in Table 2. Table 2. Predicted Values for A and l/A Ratios of d 1% - m0 4 l,7-octad1enes. A=d0/d4 1 : 2A : A2 3 l : 6 : 9 1/3 1 : 2/3 : 1/9 (9 : 6 : l) An isotOpe effect would shift the product ratios away from this inverse symmetry. No such shift was observed experi- mentally from which Grubbs concluded the lack of such an isotope effect with homogeneous catalysts. Katz34 calculated the dependence of the ratio of the rates of formation of ethylene-d0, -d2 and '94 on an isotOpe effect R. He proposed that the product of the ratios d[C2D2Hz]/d[C2H4] and d[C2D2H2]/d[C2D4] should be inde- pendent of the actual isotope effect. As shown in Table 3 this product should be significantly different for pairwise and for carbene mechanisms. 40 Table 3. Product of the Ratios of Butenes Proposed for Pairwise and Non—pairwise Mechanisms. Pairwise 2.56 Carbene 4 Katz calculated this product for a pairwise exchange and found it to be ”very insensitive to R. It decreased slowly from 2.56 to 2.53 when R increased from 1 to 2.5." Assum- ing the validity of the mechanism used by Katz this should be a reasonable test to distinguish between the two mechan- isms even in the presence of an isotope effect. No other isotope effects have been reported for meta- thesis catalysts. This review will now cover isotope effects which have been reported in the catalysis of other reactions over heterogeneous systems. Cvetanovic and Duncan84 studied a silver ion-olefin com- plex for possible isotope effects. They found that an inverse isotope effect contributed to the stability of the various olefin complexes. Olefins which had deuterium atoms directly attached to the unsaturated carbons were stabilized the most and the effect of increased deutera- tion was additive. In this case deuteration lead to an 41 increase in complex stability. 85 proposed an allylic intermediate Adams and Jennings for the oxidation of propylene over Bi203/M003. Based on the observed isotope effect of KH/KD = 1.8, they postulated methyl hydrogen abstraction as the rate determining step. Cant and Hall86 proposed a similar mechanism for propylene oxidation over Au which demonstrated an isotope effect of 2.5. For metathesis’an overall mechanism which involves an allylic intermediate appears unlikely. However, as demon- strated throughout this introduction, and as will be shown in the next section, an allylic intermediate is a likely candidate for the initiation of metathesis over a hetero- geneous catalyst. RESULTS AND DISCUSSION The original goal of this investigation was to study the ratio of labelled butenes produced in the metathesis of a mixture of labelled and unlabelled dienes over a heterogeneous catalyst and from this ratio, to deduce the mechanism of olefin metathesis. In the course of this study a variety of unexpected observations were made which lead to an intriguing mechan- istic problem. These observations will be discussed one at a time and then placed into an overall scheme which gives insight into the chemical processes occurring in heterogeneously catalyzed metathesis. The preparation of pure cis,cis-Z,8-decadiene-l,l,l,- 10,10,10—d6 and cis,cis-2,8-decadiene was a prerequisite for conducting these experiments. Using the method devised 35 these compounds were prepared as outlined in by Hoppin, Figure 7. Commercially available l,7-octadiyne was treat- ed with n-butyl lithium yielding the dianion. Addition of methyl iodide-d3 led to the formation of 2,8-decadiyne- 1,1,1,10,10,10-,d6 which was isolated and then stereo- specifically hydrogenated. Cis,cis-2,8-decadiene-d0 was prepared in the same manner using unlabelled methyl iodide and also by hydrogenating commercially available 2,8-decadiyne. Trans,trans-2,8- decadiene-d0 was commercially available and trans,trans- 42 43 ZN-BuLi .+ iii—H ==='L1 8) —-> =~H 0° =Li+ \:)° 2CD31 3 ___ 5% Pd/BaSO4 b) t==='CD3 \ ==:-CD Pyridine Figure 7. Preparation of cis,cis—2,8-decadiene-d6. 2,8-decadiene-g6 was supplied by C. Hoppin. Preliminary investigations were conducted using a com- mercial Molybdena-alumina catalyst which contained approxi- 88 This was activated mately 13% M003 on gamma alumina. by heating in air at 560° for twelve hours followed by a 2 hour purge with argon at 560° ApproximatelleS g of catalyst was transferred under argon to a side-arm flask which was then sealed with a rubber septum. Without adding solvent,0.2 ml of a 1.1:1 mixture of 2,8-decadiene- SO and 2,8-decadiene-d6 was added. The white solid catalyst 44 turned black upon addition of the olefin. The reaction was followed by analyzing small ( 0.2 ml) gas samples by using gas chromatography. Larger gas samples ( 5 ml) were separated on a gas chromatograph and the butenes collected by passing the effluent gas through a gas col- lection tube cooled in liquid nitrogen. The collection tube was sealed off and the gas sample analyzed by mass spectroscopy. Some results of this initial study are presented in Table 4. In these preliminary studies, the reaction began slowly and the catalyst underwent a rapid deactivation. As a result, all of the gas samples were very small. In spite of this, the resultant product ratios are reasonably consistent. As has been noted in the Introduction, a specific ratio of 86:83:80 butenes should be produced depending on the mechanism. For a 1.1:1 $0186 mixture of 2,8-decadienes, a carbene chain mechanism was expected to produce a 1:2.2: 83 For a pairwise exchange 1.2 ratio of 86:83:80 butenes. of alkylidene units a ratio of l:l.73:l.18 was expected. As shown in Table 4, the observed product ratios did not correspond to either of these possibilities. An excess of both do and d3 butenes was produced in all four runs. This excess of products containing a (:CH-CHS) unit suggested that whatever mechanism was operating was influenced by an isotope effect. 45 Table 4. Mass Spectral Analysis of Butenes Produced in the Metathesis of Labelled 2,8-decadienes, Over 13% M003 on gamma alumina, in the Absence of a Solvent. a . b c trans A = 30/96 Temp. Time D6 : D3 D0 c15 trans-trans 1.1 85-13S° l/2 hr 1 2.46 : 1.67 1.9 1.1 70° 22 min 1 : 2.63 : 1.69 3.5 136 hr 1 : 2.66 : 1.74 1.4 cis cis-cis trans 1.1 57° <1/2 hr 1 2.5 1.58 3.5 1.1 68° 3 hr 20 min 1 : 2.47 : 1.58 1.5 D6 only 34-80° 50 min D6 only D0 only 48° 20 min D0 only 3) +0.05. b) Several samples collected and analyzed together. C) 10.05. d) +0.1. e)D3 and D0 values corrected for ionization peaks. 46 As mentioned earlier, Katz developed a simple method for calculating relationships between product ratios which would theoretically cancel the effect of an isotope effect. In his work, an experimental value of 3.8 was observed for the product (Dz/D0) x (DZ/D4). As shown in Table 5, the calculated products from these preliminary experiments were somewhat higher and well above the maximum value (2.56) expected for a pairwise alkylidene exchange. This would suggest that a non-pairwise exchange of alkylidene units may occur over this heterogeneous catalyst. The butene ratios were obtained by comparing the peak heights of butene-d6 (m/e 62)) butene-Q3 (m/e 59) and butene-d0 (m/e 56) in the mass spectra. Both the butene-d3 and butene-d0 peaks were corrected for overlapping ioniza- tion peaks. No such correction was necessary for the butene- 86 peak. The butene-d3 had only a small correction, there- fore the most accurate measurement of product ratios can be obtained by comparing the experimental butene-gs/ butene-d6 ratio to that expected theoretically. Each run in Table 4 indicated an excess of non-deuterated products. As mentioned earlier, a number of prOposed hetero- geneous initiation schemes involve an attack on the olefin by surface metal-hydrides. An attack of'thiskind might cause an excess of non-deuterated butenes. In order to rule out this possibility, 86' and d0-2,8 decadiene were treated with the metathesis catalyst in separate experiments. 47 Table 5. Comparison of Product Ratios Using Katz's Method. 12.3; D D6 D3 D0 D0 D4 1 2 44 1.57 3 79 l 2 46 1.67 3 62 1 2.63 1.69 4.09 1 2.66 1.74 4.06 1 2 5 1.58 3 95 l 2 47 1.58 3 86 Mass spectral analysis of the cis- and trans-butenes pro- duced by these separate runs indicated no contamination due to hydrogen sources of any kind. Cis, cis-2,8-decadiene-d6 produced only 2-butene-d6 (ml. wt. 62). Peaks at 59 and 56 corresponding to Z-butene-d3 and do were similar to litera- ture values for ionization peaks. Exact values are present- ed in the experimental section. Similarly cis, cis-2,8- decadiene-d0 produced no peaks at 62 or 59. The expected small ionization peaks were observed at m/e- 55 through m/e- 50. These results suggested that the excess Z-butene-d0 and -g3 might be due to an isotope effect. This was further supported by the fact that either the dO-diene or a mixture of the do- and d6-diene produced collectable amounts of 48 Z-butene faster than the dé-diene alone. As a check of this apparent rate difference, simultaneous experiments were conducted using equal amounts of catalyst, propane standard and either the 80' or d6-diene in separate flasks. The two flasks were stirred side by side in an oil bath at 30° and 0.5 m1 gas samples removed and analyzed by gas chromatography. Figure 8 shows the production of cis and trans butenes and propene relative to a propane standard at various time intervals. The butenes and propene were produced more rapidly from cis, cis-2,8-decadiene-d0. A repeat of this experiment is shown in Figure 9 and again the go—diene produced butenes more rapidly. Considering the sluggish reaction observed in the 81278 compared to the relative ease metathesis of ethene with prOpene or higher alkenes, the above results might suggest v-allyl formation as a rate-determining step (Eq. 36). (36) Table 4 also indicates the ratio of transzcis products 49 0 2,8-decadiene-g0 x 2,8-decadiene-d m6 5.0-- 4.0-- x propane standard 3.0“- ' 7 I f T 1 20 40 60 80 100 120 Minutes Figure 8. Comparison of the rates of formation of butenes and propenes (Run 1). 50 .AN csmv monousn mo coflumEeom mo money may mo cemfipmeou mopscwz OVH ONH OOH ow oo ow ON _ .7 . n ._ b L ‘ i OWIH ) 1 ed - L a l op-& n ow-ocowpmoow-w.m x i ow-u 4 5852388-”; o .m opsmwm 1 m pumpcmum oaomopm x C 1 OH I ma 51 in the samples collected. This gives some indication of the extent of equilibration at the time of sampling. As noted earlier, one of the major criticisms of the early studies with labelled l,7-octadienes was that the product ethenes could equilibrate following productive metathesis. The relatively high stereospecificity observed in the early samples suggests that the scrambling of deuterium labels by secondary metathesis reactions should be minimal. Besides butenes, small amounts of both ethene and propene were produced in these experiments. Table 6 lists typical ratios of these products at various times. The Table 6. Ratios of Ethene, Propene and Butene Produced During the Metathesis of trans, trans-2,8-deca- diene in the Absence of a Solvent. Time Et Pr t-Butene c-Butene 2 min 45 sec 2 -- 61 -- 25 min 0.5 1 104 23 31 min ---- 1 110 39 62 min ---- 1 166 72 source of these unexpected products will be discussed pres- ently. These preliminary studies were hampered by the low 52 activity of the MoO3/A1203 catalyst. A far more satis- factory catalyst proved to be a cobalt-molybdenum-alumina 60 studied the behavior of both extruded catalyst. Basset MODS/A1203 and Moos/CoO/AlZOS and noted no significant dif- ference in the metathesis of pentene. However, the MO3/ CoO/A1203 was found to be more active for the metathesis of 2,8-decadienes. A chemical analysis of this commercial catalyst indicated 12.5% M003 and 3.5% CoO, supported on an alumina base. As in the initial tests, this catalyst was heated to 560° in air followed by an argon purge at 560° and then cooled to the reaction temperature under argon. Investigations into the source of the ethene and pro- pene have allowed some insight to the role of the catalyst and the formation of the active catalyst species. As with the preliminary studies, the Moos/CoO/AleS catalyst produced a tiny initial burst of ethene followed by a steady production of propene and cis- and trans-Z-butene. Figure 10 indicates the production of propene, trans- Z-butene and cis-Z-butene, relative to a 1/2 m1 propane standard at room temperature, early in the metathesis of cis, cis-2,8-decadiene-d0 over MoO3/CoO/A1203. Ethene was occasionally observed during the first few seconds of olefin contact with the catalyst. After this initial burst, ethene was only detected if the reaction was heated to above 90°. Propene was also formed seconds after the addition Figure 10. 53 Production of propene and butenes relative to a propane standard. 54 10 l l (\cis butene sta. trans butene 0.5.4 er‘propene 0001— 10 20 30 40 50 Minutes Figure 10 55 of the olefin. As shown in Figures 8 and 11 the level of propene decreased significantly after about 1 hour of olefin exposure relative to the amount of butene pro- duced. Neither of these products would be expected from the simple metathesis of 2,8-decadiene. Figure 11 shows the change in product ratios with time. The cisztrans butene ratio decreases as equilibration occurs and the relative level of propene, which is high during the first few minutes of olefin-catalyst contact, decreases steadily as the catalyst is ”broken—in". As mentioned previously, heterogeneous catalysts exhibit a "break-in" period during which the activity of the catalyst gradually increases. In a careful study of this break—in period over freshly 59 observed a large activated WO3 on silica gel, Luckner initial production of ethene in the metathesis of propene. 2-Butene production was negligible during this break-in period. The fully activated catalyst produced nearly equal amounts of ethene and butene. A similar break-in phenomena appears to be occurring in this study. Combining the observed isotope effect with the production of ethene and propene suggests the follow- ing initiation scheme (Figure 12). The isotope effect observed in the formation of the labelled butenes may be due to the C-H bond cleavage inherent in the formation of the n-allyl species. A detailed study of this isotope effect will be presented later. 56 .o:ou:n-~-m:weu ow o>wpmfioh ocoaoea paw o:ou:n-~-mflo mo :OMDUSpoem .HH oezwwm mousse: u: em ocH ova oNH CCH ow so ow ON ILJC _ P _ e _ _ _ H____rL _ _ .r x x x "Inn!" u o.H O o In]! D O 0 O . u.l o.~ X 0 O 0 III o.m . .ar. oceans -m-m:weu oIIK ocomopm x . fl o.e o:oa:p-~-mwo . .l .l.o.m 57 \l 3 CH3 MH l=HD D2 Eaj=wz Figure 12. Initiation of the active catalyst for hetero- geneously catalyzed olefin metathesis. The ethene observed early in the reaction could be pro- duced by the breakdown of the metallocyclobutane formed from this n-allyl, as shown in path a of Figure 12. The propene would then be formed by the cross-metathesis of ethene and 2,8-decadiene. Alternately, the metallocyclo- butane could rearrange as in path b of Figure 12 to produce a metal—methylene complex and l,7-nonadiene. This metal— methylene complex could undergo metathesis with 2,8-deca- diene to yield propene. The ethene could be formed by 58 the dimerization of two metal-methylene species. Experi- mental conditions more than likely dictate the preferred pathway for the decomposition of the proposed metallocyclo- butane. Either of these two pathways would yield nearly identical products and it would be difficult to distinguish between them. The formation of a D-allyl, rearrangement to a metallo- cyclobutane and decomposition to a metal-carbene-olefin complex could explain the break-in effect observed with a number of heterogeneous catalysts. This initial n-allyl formation could also explain why little or no "break-in" effect is observed with tetrakis-w~allylmolybdenum sup- ported on alumina. Moreover this scheme provides the initia- tion step necessary for a metal-carbene-catalyzed mechanism. The ethene and propene could also have been secondary products formed in a competing isomerization and meta- thesis sequence. Bradshaw93 demonstrated that a CoO/MoOS/ A1203 catalyst could selectively produce either ethene and 3-hexene or propene and n-pentenes from the metathesis and controlled isomerization of n-butenes. However, these investigations were conducted at temperatures of 120° and higher. Basset6O has noted that isomerization is significant only at temperatures above 120° with these catalysts. In addition, isomerization could not explain 59 the initial burst of ethene noted by Luckner in the metathesis of propene. 59 If such a competing isomerization were occurring, the 2,8-decadiene would first isomerize to 1,8-decadiene and then undergo metathesis to produce propene and cyclo- heptene (Eq. 37). (37) D3 D3 : ——9 —"> \+D2=\ D I 3 2 In order to look for these products, 3 ml of cis, cis- 2,8-decadiene-dO was treated with 3.4 g of MoO3/CoO/A1203 in the absence of a solvent. The liquid phase was analyzed by gas chromatography at 5 minutes and 30 minutes. Cyclo- hexene was clearly evident but no cycloheptene was observed. The absence of this product is most significant early in the reaction when the ethene and prOpene levels are highest relative to the butene products. Cycloheptene is known to undergo ring-opening polymerization, however some free monomer might be expected. The ethene occurred in such small amounts that re- peated efforts to trap it in a gas collection tube for mass spectral analysis failed. The propene was success- fully isolated and analyzed by mass spectroscopy. Table 7 lists the approximate percentages of propene-d0 60 Table 7. Approximate Percentages of Labelled Propenes Formed in the Metathesis of a 1:1 Mixture of 2,8-decadiene-d0 and 'é6’ M.W. M.W. Formula Percent 47 C3HD5 3% 46 C3H2D4 14% 45 C3H3D3 26% 44 C3H4D2 5% 43 C3H5D 19% 42 C3H6 33% 61 through propene-d5 obtained in the metathesis of a 1.05:1 mixture of do-zd6-diene at room temperature. From the range of propenes produced, it appears that the major source of these products is the cross metathesis of a methylene species with the 2,8-decadiene mixture. The methylene species might be either the ethene or the metal-methylene complex formed in the decomposition of the proposed metallo- cyclobutane. The largest fraction is that at molecular weight 42 which corresponds to propene-d0. The next largest fraction is that containing three deuterium atoms. These two pro- ducts comprise approximately half the total propene. As shown in Figure 13, the expected route for the formation of these propenes would be cross-metathesis between a non- deuterated ethene and either a -go or -d3 Z-alkene. H H ‘\ / R CH2 7 \CH3 M H 2 ~ / H H \. C:C R’/ ‘\CD3- Figure 13. Formation of propene-g0 and propene-d3. The preferential formation of products resulting from 62 ethene-dO cross-metathesis would be expected if the initia- tion scheme shown in Figure 12 were more rapid for non- deuterated 2,8-decadiene. The remainder of the sample was composed of propene with 1,2,4, or 5 deuterium atoms in- corporated. These would be expected to form as shown in Figure 14. RH: CH(CH ) . /H 3 D c\ M.W. 43 9 CH3 __ H RHC—CH(CD3) / M w 46 ) DHCZCE . . CHD D3 (LL RHC=CH(CH3) )4 2 > D c=c M.W. 44 2 \ CH3 RHC—— CH(CD3) )4 MW 47 ) ch::::C ° ° \CD 3 Figure 14. Formation of propene -d1, -d2, -d4 and —g5. Table 8 presents the percentages of labelled prOpenes collected in six separate experiments. The product per- centages were calculated by a stepwise procedure in which the peaks due to the heavier components were successively subtracted out.106 With six overlapping spectra, the 63 Hem --- Hem --- so HRH e.~ A we Hm com awe am sHH smfi am mm H mN.N we we saw Ame am smfi wmfi am RN H mN.N we om one man --- new we --- --- H mN.N see OH see smm --- we Hem we RAH H mo.H we N emu smm sea am Rem sea mm H mo.H we we beefiee< emmu mmzmo Nmezmu mmmzmu emszu mmzmu em on sage .HEee .mocomoem wofifioan wo mommucooeom opmeflxoemm< .w oHan 64 product ratios are only approximate, however experiments using the same DozD6 2,8-decadiene mixtures gave similar product ratios. Therefore, these product ratios should be valid for qualitative comparisons between runs. The most notable feature is the excess of products containing non-deuterated methylene units. As in earlier studies, the extent of reaction became a key factor in the analysis of other data. Basset60 had noted earlier that at the conditions used in this study, the catalyst was deactivated within a few hours. Since the goal of this study was to study the butenes formed early in the reaction, it was important to know the con- version level at short reaction times. Figure 15 graphs the production of cyclohexene from 2,8-decadiene as a function of time. An exponential increase in cyclohexene formation is clearly demonstrated for the first hour of reaction. This increase is charac- teristic of an induction period which in turn suggests a chain reaction. It is during this first hour of reac- tion that ethene is observed and that propene is formed in the highest levels. Following this induction period, the formation of cyclohexene is linear with time and levels off. This is presumably due to catalyst de- activation. The highest conversion observed in any of these experiments was approximately 20% and many runs were well below that point. In similar studies over a 65 20‘- with Sn(CH3)4 without Sn(CH3)4 l ‘r *1 l I 60 120 180 240 300 Figure 15. Conversion of 2,8-decadiene to cyclohexene. 66 molybdena-alumina catalyst, workers at the British Petro- leum Co., Ltd.96 noted conversions of less than 2% for propene metathesis at 25°C. This indicated that the turn- over number was probably low and that a large fraction of the butene formed would be influenced by the processes occurring during this induction period. In an effort to further identify the source of the induction period and the resulting unexpected products, another study which was originally used with a homogeneous catalyst was attempted. As mentioned in the Introduction, Muetterties49 has suggested alkylation followed by a- elimination as the source of the chain-carrying metal- 35 supported this theory with a carbene species. Hoppin series of metathesis experiments using deuterium labelled olefins and co-catalysts. Heterogeneous catalysts have similarly been activated by the use of various co-cata- lysts.72’74 An investigation of the effect of co-catalyst activation has given insight into the initiation process. Figure 15 also demonstrates the level of cyclohexene formation when the Moos-CoO-AlZOS catalyst was treated with a tetramethyltin co-catalyst in chlorobenzene followed by the addition of the diene. The catalyst was initially more active for the production of cyclohexene. A decreased induction period was observed during which additional active sites may have been formed by normal metal-olefin break-in processes. 67 Some active sites for olefin metathesis appear to have been generated by the interaction of the solid catalyst with the organometallic agent. In analogy with homogeneous catalysts, this could suggest that the induction period was Specifically due to the formation of a metal-carbene species, and that prior formation via an alkylating agent shortened the induction period. Analysis of the gaseous products formed during meta- thesis in the presence of alkyl-tin compounds gave addi- tional support for this theory. As shown in Table 9, the propene formed in the presence of tin co-catalysts contains a higher proportion of deuterium. In each run, some of the propylene contained methylene units which could only originate in the co—catalyst (Eq. 37). ;[M]=CD2 % CD2=C\\ 9% CH3 [M] (37) D -diene H fi‘ [M]=CH2 5 % CH2=C< Sn(CH3)4 6% CD3 The majority of the propene appears to be due to normal break-in processes. The original goal of this investigation was to study the isotopic ratio of butenes produced prior to product equilibration. Typical product ratios will now be considered in light of the various 68 Table 9. Mass Spectral Analysis of Propene Formed During the Metathesis of cis,cis-2,8-decadiene Over MoO3/CoO/A1203 Treated with a Tetramethyltin Co-catalyst. Co-Catalyst: Sn(CD3)4 Sn(CD3)4 Sn(CD3)4 Sn(CH3)4 Olefin: D6 Only DO Only D6 : D0 D6 Only 2.4: 1 C3HD5 78% -——— 56% 71% C3H2D4 22% —— 21% 23% C3H3D3 -—— -—- 3% 6% C3H4D2 --—- 9% 19% -—-— CSHSD -———— ___. —*——- ————— C3H6 ———— 91% ———~ ———— Temp. 47°C 34°C 61° 90° Time 62 min 30 min 29 min 3.5 hr. 69 reactions which appear to precede the actual catalytic metathesis of olefins. Table 10 presents typical data obtained from an ap- proximate 1:1 mixture of -do and -d6-2,8-decadiene meta- thesized in chlorobenzene over a MoO3/CoO/A1203 catalyst. Chlorobenzene has been used frequently as a solvent for the homogeneous metathesis of olefins. Therefore, it was considered a reasonable choice for this investigation. The observed ratio of butene-d3 and butene-d0 is higher than that expected for either a carbene or a pairwise mechanism. However, even at short reaction times, the experimental butene ratio is consistent with the 1:2RA: RZAZ ratio of products expected for a carbene mechanism involving an isotope effect (R). This excess of non— deuterated species decreases at longer reaction times suggesting that the preference for deuterium-free olefins occurs early in the reaction. An initially high isotope selectivity which decreases as the reaction proceeds would be expected for the initiation processes previously pro- posed. The ratio of cis to trans butene is relatively high for these early samples suggesting that equilibration due to secondary metathesis processes should be minimal. With a minimum of label scrambling, these early product ratios support a non-pairwise alkylidene exchange mechanism such as Chauvin's chain-mechanism. 70 .zhuosopuoomm mmme kn wommecm H.oa mm NH.Hmv.H vo.owmm.m H m.o cHE mm H: N mH.Hwo.H eo.owoe.N H m.H cHE 5H H: H mm.Hun.NuH em.H NH.HNH.~ oo.ome.N H m.N :HE OH m mH.HHo.H no.OHNe.N H o.H 9: mm 0H.Hmm.m wosoamn.m H m.H H: n 0H.Hmn.N wo.owmm.m H w.H :HE OH H: H om.Hnoe.NuH NH.H NH.me.H wo.owme.N H m.H :HE om Doom < . . o m o mcmeu N~CH p(c H .) - CH -p -M (:0 z 6 5 U.v.-co7 2 <72 °( )5 Analysis indicated 0.5mequiv. Mo/g resin, and 0u4mequiv. phosphine/g resin. The polymeric molybdenum complex was then treated with C2H2A1C12/02 and used to metathesize cis-Z-pentene. The catalyst was weakly active giving a 3.4% conversion after 20 minutes. No mention of recycling was made. Subsequent to the present work, a similar catalyst was prepared using W(CO)6 and compared to a catalyst in which the tungsten was directly bound to the chloromethylated 104 polystyrene. The preparation of these two compounds is outlined in Figure 17. l) NaPO2 4. CH Pp W(CO) 2) W(CO)6 3 Q 2 Z 5 (A) >_©CH2C1 NaW(CO)3Cp H) H ifCH W(CO) (n-C H ) -NaCl 2 3 s s (B) Figure 17. Preparation of polymer supported tungsten carbonyl metathesis catalysts. v .11 104 BuAlClZ/Oz was used as a co-catalyst with each of the above polymer-supported species. Compound B formed an active catalyst which did not dissociate from the polymer support, and which demonstrated a constant activity after 8 cycles of washing, vacuum drying, and reuse. It appears that fresh co-catalyst was required during each reuse. The organo-aluminum co-catalyst cleaved the tungsten carbonyl complex from Compound A. Due to this cleavage, washing and reuse resulted in a dramatic loss of activity. The tungsten hexachloride-tetramethyltin system was chosen for study because this was known to be a very active homogeneous catalyst. Moreover several compounds contain- ing tin-transition metal bonds have been reported.105 In particular, a polymer-supported alkyltin compound106 has been used as a reagent in the preparation of bis(di- N-butylchlorotin)tetracarbonylosmium. The pertinent steps are outlined in Figure 18. Bu (nBu)ZSnC12 I >~©—Li _> >—©—Sn - c1 Bu Bu H205(C0)4 >—©—Sn-os (CO)4 EtZNI'I l Bu Figure 18. Preparation of a polymer supported tin-osmium complex. 105 The infrared spectrum of the supported complex was similar to that of cis-¢SSnOS(CO)4H. After further reactions, the polymer-Sn bond was cleaved by treatment with anhydrous HCl at ~15° for 30 min. The strong acid needed to cleave this polymer-Sn bond suggests that such polymer-bound tin- transition metal complexes would be stable to cleavage under normal metathesis conditions. A co-worker in this laboratory studied a polymer—sup- ported (Mo(03P)2(NO)2C12 complex linked through a polymeric phosphine.107 Activation with (CH3)3A1C13 gave an active metathesis catalyst with an activity similar to the homo- geneous system. However, after filtration and washing the catalyst proved inactive unless additional co-catalyst was added. These results are similar to those obtained using the polymeric-W(CO)3Cp complex. The necessity of adding fresh co-catalyst for each reuse would make these catalysts less convenient. The present study was undertaken in an effort to develop a polymer-bound catalyst which would not require the addi— tion of co-catalyst with each reuse. RESULTS AND DISCUSSION The polymeric Sn compound was prepared by treating lithiated polystyrene with trimethyltin chloride in THF. A large batch could be prepared and stored in a dry box until needed. The polymer supported metathesis catalyst was prepared by treating the polymeric tin compound with WCl6 in purified trichloroethylene. The reactions are outlined in Figure 19. nBuLi > Q TMEDA 7) Q ‘ THF 4; >0. Sn(CH3) 3 CH3 1 /wc16 Sn—{W]:CH2 1 CH3 Figure 19. Preparation of the polymer-supported catalyst. The active beads were washed with trichloroethylene and transferred to a glass column under argon. They were supported in the column on a pad of glass wool. A dilute 106 107 solution of 1,7 octadiene in trichloroethylene was slowly dripped through the beads in a continuous flow. The liquid was then collected in a side arm flask connected to the reaction column. A heating tape was wrapped around the reaction column to control the temperature. The entire system was sealed so gaseous products could be sampled. The reaction was followed by analyzing the conversion of 1,7 octadiene to cyclohexene in a gas chromatograph. Table 18 lists conversion data for a typical run. The activity of this catalyst was very low and much further development would be needed before it could be used for catalytic purposes. But the reasonably long lifetime (3.6 days) demonstrated by the run in Table 18 is compar- able to homogeneous systems. The catalyst did not appear to wash off the beads as the collected liquid was generally clear. The main difficulty appears to be the extremely clean conditions required for working with this catalyst. Once formed, the blue-black polymer supported catalyst was extremely air sensitive turning pale blue at the slightest exposure to air. This pale blue complex was inactive for metathesis. Highest catalytic activity was attained by purifying the argon with a BASF catalyst and double distilling both solvent and olefin reactants. The extreme air sensitivity of the active species was a major drawback to this catalyst and more than likely a direct cause for its low activity. 108 Table 18. Conversion of 1,7 Octadiene to Cyclohexene Over a Polymer Supported Catalyst. Temp. Time Total ml % 80° 1 min 1 30 79° 9 hr, 10 31 55 79° 9 hr, 30 38 30 80° 10 hr, 40 58 20 " 11 hr, 30 88 10 ” 11 hr, 40 95 7 " 12 hr, 40 105 15 89° 13 hr, 50 115 2 90° 15 hr, 5 117 2 100° 15 hr, 30 137 20 " 40 hr, 15 202 5 " 40 hr, 40 210 5 102° 41 hr, 15 211 55 104° 42 hr, 50 213 50 109° 44 hr, 40 225 20 101° 45 hr, 0 226 20 ” 45 hr, 15 227 25 100° 62 hr, 20 241 70 105° 64 hr, 30 243 65 80° 83 hr, 0 244 40 80° 88 hr, 50 250 5 109 Analysis of the polymeric tin reagent indicated .476 mmoles Sn/g polymer. An experiment was undertaken to see what percent of the available tin actually formed active metathesis sites. Meutterties has proposed an alkylation followed by an elimination as the major source of active carbene species. 'CH 4012 4 \ /' 2 CH3 H \| CH ' /w< 3——> >1 CH3 I By this scheme the number of active sites should equal the methane produced during the interaction of the poly- meric co-catalyst and tungsten catalyst. To examine the number of active sites 1.24 g (.59 mmoles) polymeric Sn was stirred in the presence of .07 g WC16 (.17 mmoles) in trichloroethylene at 81°C. A propane standard was in- jected and the methane produced measured relative to this standard. The mixture was allowed to stir 2 hours after which WCl6 was still in excess as noted by the dark solu- tion. Approximately .1 mmole of methane was produced. Assuming a ratio of 2 Sn:l W this would mean that .1/.59 = .169 i 2 = 8.4% percent of the tin available was actually activated. A major limitation may have been a limited number of sites where two polymer-attached tin atoms were in close enough proximity to achieve the alkylation and 110 subsequent elimination. Following this gas study the beads were carefully washed and then 2.0 ml 1,7 octadiene added. At 80°C a 33% conversion was attained after 1 hr, 44% after 2 hours and 53% after 3 hours. Table 19 presents the pertinent data. As indicated in Table 19 this catalyst was much more active than the earliest polymer-supported metathesis system developed by Basset. It is hoped that further work may develop systems which are highly active and less sensitive to impurities. Table 19. Calculation of the Turnover Number for the Polymeric Catalyst.‘ .476 mmoles x .084 (% active) = .039 mmoles. .039 mmoles active catalyst:18.l mmoles olefin 1 hr 33% 5.9 mmoles = 153 turnover/hr 2 hr 44% 3 hr 53% 9.59 mmoles = 245 turnover/3 hr Sn : Mo : olefin (Assume) 7 : 1 : 464.1 EXPERIMENTAL The trichloroethene solvent was double distilled from P205 leaving the last 100 ml of solvent behind. Argon was then bubbled through the trichloroethene for 2 hours to remove trace oxygen. The 1,7 octadiene was double distilled from Na and degassed with argon. The argon was purified by passing through a BASF column heated to 100°C and then through molecular sieves. A significant increase in ac— tivity was observed for those experiments using these puri- fied reagents. All glassware was oven—dried and cooled in a stream of purified argon. A Varian series 1400 gas chromatograph equipped with a flame ionization detector was used to analyze the gas samples using a Durapak column at 60°C. Liquid samples were analyzed on a Varian Model 90-P gas chromatograph with a 25' 5% Carbowax 20M/Chromosorb W column at 135-150°C. Preparation ofgpolystyrene trimethyltin 20% divinylbenzene/styrene copolymer beads were cleaned by washing in cyclohexane under argon. In a 250 ml side arm round bottom flask the beads were stirred with 30 m1 cyclohexane (K/benzophenone), 16 ml TMEDA (Na) and 60 ml NBuLi. The mixture was stirred at 60°C for 12 hours followed by five washings with cyclohexane and three washings 111 112 in tetrahydrofuran (Na/benzophenone). Following this, 60 m1 of THF was added and trimethyltinchloride slowly added, under argon. The bright red lithiated beads slowly turned pink and then off white as the tin compound was added. A total of 3.27 g of trimethyltinchloride was used. The beads were allowed to stir 3 days to insure complete re- action. The beads were then washed with THF and vacuumed dry overnight. Analysis (Galbraith Laboratories) 6.65% Sn (.56 mmole/gram beads). The beads were stored in a dry box and placed in the reaction tube under argon for each separate catalyst preparation. Preparation of a Metathesis Catalyst From Polystyrene Tri- methyltin and Tungsten Hexachloride Inside the dry box, 1.5 g of the polystyrene trimethyl— tin beads prepared above were placed in a Schlenk tube with .04 g (.1 mmole) WC16. The tube was sealed under argon and removed from the dry box. 20 ml C12C=CC1 was added and the mixture stirred with warming. After two hours, the beads were carefully washed and transferred under argon to the reaction column. Set Up and Analysis of a Constant Flow System The reaction column consisted of a side arm round bottom flask topped by a 10 inch tube which was indented slightly at about the middle to support the glass wool 113 pad and catalyst beads. This was topped by an ”airless- ware" addition funnel which allowed the slow addition of the dilute olefin. This was topped by a flo-control stopper and an argon bubbler. The column was set up with argon flowing through while the beads were prepared. The beads were transferred to the column under argon and the reaction begun by dripping in the dilute olefin. The liquid phase was sampled from the round bottom flask through the side arm which was stoppered with a rubber septum. The results of a column using 250 m1 C12C=CC1/5 m1 1,7 octadiene are shown in Table 18. A heating table wrapped around the reaction column allowed the temperature to be controlled. Analysis of Gaseous Products in the Formation of the Polymer Supported Catalyst and Analysis of Conversion in the Metathesis of 1,7 Octadiene In the dry box 1.24 g polystyrene trimethyltin beads (.476 mmoles/g) was placed in a Schlenk tube with .07 g WC16' The tube was sealed and removed from the dry box. 10 m1 C12=C1 was added and the mixture stirred for 2 hours. A propane std was added (2 m1, .089 mmole) and the methane was compared to this standard on an FID gas chromatograph. Peak heights were corrected for the influence of molecular weight. The results are shown in Table 19. Following the above analysis, 2 m1 of 1,7 octadiene was added and the reaction of the liquid phase analyzed 114 on a gas chromatograph. The results are presented in Table 19. REFERENCES 10. 11. 12. 13. REFERENCES R. L. Banks and G. C. Bailey, Ind. Eng. Chem,, Prod. Res. Develpp., 3, 170 (1964). a) N. Calderon, H. Y. Chen, and K. W. Scott, Tetra- hedron Letters, 1967, 3327; b) N. Calderon, E. A. Ofstead, J. P. Ward, W. A. Judy, and K. W. Scott, J. Amer. Chem. Soc., 1968, 90, 4133. a) H. S. Eluterio, Ger. Offen., 1960, l 072 811 (Chem. Abs., 1961, 55, 16 005 h); b) G. Natta, G. Dall'Asta, G. Mazzanti, and G. Motroni, Makromol. Chem. 1963, 69, 163; c) G. Natta, G. DallTAsta, and_G} Mazzanti, Angew. Chem. Interat. Edn., 1964, 3, 723; G. Natta, G. Dall' Asta, I. W. Bassi, and G. Carella, Makromol. Chem., 1966, 91, 87. G. C. Bailey, Catal. Rev., 3, 37 (1969). J. C. Mol., and J. A. Mouljin, Adv. in Catal. 24, 131 (1975). Y. Uchida, M. Hidai, and T. Tatsumi, Bull. Chem. Soc. Jap. 45, 1158 (1972). J. Chatt, R. J. Haines, and G. J. Leigh, Chem. Commun. 1202 (1972). E. A. Zuech, W. B. Hughes, D. H. Kubicek and E. T. Kittleman, J. Amer. Chem. Soc., 92, 528 (1970). J. A. Moulijn and C. Boelhouwer, Chem. Commun. 1170 (1971). L. F. Heckelsberg, R. L. Banks, and G. C. Bailey, Ind. Eng. Chem., Prod. Res. Develop., 8, 259 (1969). L. F. Heckelsberg, R. L. Banks, and G. C. Bailey, Ind. Eng. Chem:, Prod. Res. Develop., Z, 29 (1968). E. J. Howman, L. Turner, and K. V. Williams, British Patent 1,106,015 (1968). K. V. Williams and L. Turner, British Patent 1,116,243 (1968). 115 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 116 J. C. Mol, J. A. Moulijn, and C. Boelhouwer, J. Chem. Soc. Chem. Comm., 663 (1968). A. Clark, and C. Cook, J. Catal. 33, 420 (1969). J. C. Mol, J. A. Moulijn, and C. Boelhouwer, J. Catal. 33, 87 (1968). - F. L. Woody, M. J. Lewis, and G. B. Wills, J. Catal. 32, 389 (1969). C. P. C. Bradshaw, E. J. Howman, and L. Turner, 3. Catal. Z, 269 (1967). G. S. Lewandos, and R. Pettit, J. Amer. Chem. Soc. 23, 7087 (1971). L. Cassar, and J. Halpern, J. Chem. Soc. Chem. Comm., 1082 (1970). R. H. Grubbs, and T. K. Brunck, J. Amer. Chem. Soc., 22, 2538 (1972). A. R. Fraser, P. H. Bird, S. A. Bezman, J. R. Shapley, R. White, and J. A. Osborn, J. Amer. Chem. Soc., ‘23, 597 (1973). C. G. Biefeld, H. A. Eick and R. H. Grubbs, Inorg. Chem., 32, 2166 (1973). J. X. McDermott and G. M. Whitesides, J. Amer. Chem. Soc., 23, 947 (1974). J. L. Herisson and Y. Chauvin, Makrmol. Chem., 141, 161 (1970). N. Calderon, E. A. Ofstead, and W. A. Judy, Paper presented at the Central Regional Meeting of the American Chemical Society, Akron, Ohio. May 1968. a) W. B. Hughes, J. Amer. Chem. Soc., 92 532 (1970); ) T. J. Katz andFJ. MCGinnis, J. AmerT-Chem. Soc., 22, 1903 (1977). B. A. Dolgolplosk, K. L. Makovetsky, T. G. Golenko, Y. U. Korshak, and E. I. Timyakova, Eur. Pol. J., 10, 901 (1974). a) K. W. Scott, N. Calderon, E. A. Ofstead, W. A. Judy and J. P. Ward. Rubber Chem. Technol. 44, 1341 (1971) '_— b) K. W. Scott. Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem. 33, 874 (1972). 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 117 N. Calderon, E. A. Ofstead, and W. A. Judy, Angew. Chem. Int. Ed. Engl., 33, 401 (1976). T. J. Katz and J. McGinnis, J. Amer. Chem. Soc., 97, 1592 (1975). ”— W. B. Hughes, J. Amer. Chem. Soc., 22, 532 (1970). R. H. Grubbs, P. L. Burk, and D. D. Carr, J. Amer. Chem. Soc., 21, 3265 (1975). T. J. Katz and R. Rothchild, J. Amer. Chem. Soc., 22, 2519 (1976). C. R. Hoppin, Ph. D. Thesis, Michigan State University, 1978. C. P. Casey and S. Polichnowski, J. Amer. Chem. Soc., .22, 6097 (1977). J. L. Bilhou, J. M. Basset, R. Mutin, W. F. Graydon, J. Amer. Chem. Soc., 22, 4083 (1977). E. O. Fischer, and A. Maasbol, Angew. Chem., 76, 645 (1964). —_ C. P. Casey and T. J. Burkhardt, J. Amer. Chem. Soc., 23, 5833 (1973). R. R. Schrock, J. Amer. Chem. Soc., 29, 6796 (1974). C. P. Casey and T. J. Burkhardt, J. Amer. Chem. Soc., 23, 7808 (1974). J. McGinnis, T. J. Katz, and S. Hurwitz, J. Amer. Chem. Soc., 22, 605 (1976). Private communication between C. P. Casey and R. Grubbs. B. A. Dolgoplosk, SE 23., Dokl, Chem., 216, 380 (1974). R. J. Puddphatt, M. A. Quyse, and C. F. H. Tipper, Chem. Commun., 626 (1976). C. P. Casey, L. D. Albin, T. J. Burkhardt, J. Amer. Chem. Soc., 22, 2533 (1977). M. Ephritikine, J. Levisalles, H. Rudler, D. V. Villemin, J. Organometal. Chem., 124, C1 (1977). 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 118 M. Ephritikhine, M. L. H. Green, and R. E. Mackenzie, Chem. Commun. 619 (1976). a) E. L. Muetterties and M. A. Busch, J. Chem. Soc. Chem. Commun., 754 (1974). E. L. Muetterties, Inorg. Chem., 33, 951 (1975). N. J. Cooper, M. L. H. Green, Chem., Commun., 76 (1974). J. C. M01, P. R. Visser, and C. Boelhouwer, J. Catal. 33, 114 (1970). a) A. Morris, H. Thomas, and C. J. Attridge, Ger. Offen., 1972, 2213948 (Chem. Abs., 1973, 32, 65, 76-h); b) J. P. Candlin, A. H. Mawby, and H. Thomas, Ger. Offen., 1972 2,213,947 (Chem. Abs., 1973, 32, 97, c) I. C. I. Ltd., Fr. P., 1972, 2,120,509 (Chem. Abs. 1973, 13, 115, 10411); —_ d) R. J. Haines, Chem. Soc. Rev. 3, 155 (1975). a) J. C. Mol, Thesis, University of Amsterdam (1971). b) J. C. M01, and J. A. Moulijn, Adv. Catal. 24 131 (1975). “"“““' ‘— G. Henrici-Olive', S. Olive', Angew. Chem. 85, 148 (1973). ‘—‘—“““‘— ‘— J. L. K. F. deVries, and G. T. Pott, Rec. Tr. Chim. 23, M115 (1977). G. B. Wills, J. Fathikalajahi, S. K. Gangwal, and S. Tang, Rec. Tr. Chim., 23, M110 (1977). F. Pennella, R. L. Banks, J. Catal., 33, 304 (1973). F. Pennella, R. B. Regier, R. L. Banks, J. Catal., 33, 52 (1974). R. C. Luckner, and G. B. Wills, J. Catal. 22, 83 (1973). A. Ismayel-Milanovic, J. M. Basset, H. Praliaud, M. Dufaux, and L. deMourges, J. Catal. 33, 408 (1973). E. S. Davie, D. A. Whan, and C. Kemball, J. Catal. 23, 272 (1972). ' R. L. Burwell, Jr. and A. Brenner, J. Mol. Catal., 3, 77 (1975). 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 119 E. S. Davie, D. A. Whan, C. Kemball, Chem. Commun. 1430 (1969). a) A. J. Moffat, and A. Clark, J. Catal. 17, 264 (1970); __ b) A. J. Moffat, M. M. Johnson, and A. Clark, J. Catal. 18, 345 (1970); — c! A. UT Moffat, A. Clark, and M. M. Johnson, J. Catal. 22, 379 (1971). _ A. Clark, paper presented at 155th Natl. Meeting Amer. Chem. Soc., April 1968 Abstr. 212. J. M. Begley and R. T. Wilson, J. Catal. 2, 375 (1967). J. W. Begley, and R. T. Wilson, J. Catal. 9, 375 (1967). U. R. Hattikudur, and G. Thodos, Advan. Chem. Ser. 133, 80 (1974). F. H. M. vanRijn, and J. C. Mol., Rec. Tr. Ch. 96, M96 (1977). __ N. Giordano, M. Padovan, A. Vaghi, J. C. J. Bar and A. Castellan, J. Catal. 33, 1 (1975). E. S. Davie, D. A. Whan, and C. Kemball, Chem. Comm., 1202 (1971). a) H. J. Alkema, R. VanHelden, British Patent 1,174,968; Belgian Patent 722,432; b) R. L. Banks, Forschr. Chem. Forsch., 23, 39 (1972). a) L. F. Heckelsberg, Belgian Patent 713,185; b) R. L. Banks, Forschr. Chem. Forsch., 23, 39 (1972). Yu. V. Maksimov, M. Ya, Kusnerev, J. A. Dumesic, A. E. Necitailo, R. A. Fridman, J. Catal., 33, 114 (1976). P. P. O'Neill, and J. J. Rooney, Chem. Commun. 104, (1972). G. Wittig, and K. Schwarzenbach, Justus Liebigs Ann. Chem., 650, 1 (1961). a) R. L. Banks, G. C. Bailey, J. Catal. 33, 276 (1969); b) U. 8. Patent 3,431,316. P. P. O'Neill and J. J. Rooney, J. Amer. Chem. Soc., 23, 4383 (1972). 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 120 M. H. Chisholm and H. C. Clark, Chem. Commun., 763 (1970). D. T. Laverty, J. J. Rooney and A. Stewart, J. Catal. 33, 110 (1976). A. A. Olsthoorn, and C. Boelhouwer, J. Catal., 33, 207 (1976). J. M. Basset, J. Molec. Catal., 3, 245 (1978). D. D. Carr, Ph.D. Thesis, Michigan State University, 1977. R. J. Cuetanovic, F. J. Duncan, W. E. Falconer, and R. S. Irwin, J. Amer. Chem. Soc., 32, 1827 (1965). C. R. Adams and T. J. Jennings, J. Catalysis, 33, 204 (1970). N. W. Cant and W. K. Hall, J. Phys. Chem., 75, 2914 (1971). __ C. G. Kreiter, An ew Chem., 80, 402 (1968). Angew. Chem., Int. Ed. Engl., 2, 390‘T1968). Supplied by Nalco Chemical Co. M. F. Lappert, D. J. Cardin, and M. J. Doyle, J. Chem. Soc. Chem. Comm., 927 (1972). Yu. I. Yermakov, B. N. Kuznetzov, Yu. P. Grabouski, A. N. Startzev, A. M. Lazutkin, V. A. Zakharov, A. I. Lazutkina, Catalysis, Heterogeneous and Homogeneous, Elsevier Sc1ent1f1c Pfibllshlng Co., New’York, 1975. T. T. Chang, Phys. Rev., 3 136, 1413 (1964). J. M. Basset, J. L. Bihou, R. Mutin, A. Theolier, 3. Amer. Chem. Soc., 22, 7376 (1975). C. P. C. Bradshaw, E. J. Howman, and L. Turner, J. Catal., 2, 269 (1967). _ M. Guisnet, G. Perot and R. Maurel, J. Catal., 33, 136 (1973). J. W. Hightower and W. K. Hall, J. Amer. Chem. Soc., 32, 778 (1967). a) British Petroleum Co Ltd., British Patent 1,106,016. b) C. C. Bailey, Catalysis Rev., 3, 37 (1969). 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 121 D. D. Carr, Ph.D. Dissertation, Michigan State University, 1977. a) F. Pennella, R. L. Banks, J. Catal., 33, 304 (1973); b F. Pennella, R. B. Regier, R. L. Banks, J. Catal., 33, 52 (1974). Supplied by C. Hoppin. R. B. Merrifield, J. Amer. Chem. Soc., 33, 2149 (1963). a) N. K. Mathur, R. E. Williams, J. Macromol. Sci. Rev. 333, 117 (1976). F. R. Hartley, P. N. Vezey, Adv. Opgmet. Chem., 15, 189 (1977); c J. C. Bailer, Jr., Cat. Rev.-Sci. Eng., 10, 17 (1974); —_ d) R. H. Grubbs, 333333333 512 (1977). a) R. H. Grubbs, S. Phisanbut, in press; b) W. Dumont, J. C. Poulin, T—P. Dang, and H. B. Kagan, J. Amer. Chem. Soc., 23, 8295 (1973). J. Basset, R. Mutin, G. Descotes, and D. Sinou, C. R. Hebd. Seances. Acad. Sci., Ser. C, 280, 1181 (1975). S. Warwel, P. Buschmeyer, Angew. Chem. Int. Ed., 33, 131 (1978). N. A. D. Carey and H. C. Clark, Inorg. Chem., 2, 94 (1968). J. M. Burlitch, R. C. Winterton, J. Amer. Chem. Soc., 22, 5606 (1975). R. H. Grubbs, S. Swetnick, S. C.—H. Su, J. Molecul. Catal. 3, 11 (1977/78). W. S. Greenlee and M. F. Farona, Inorg. Chem., 33, 2129 (1976). .11.